1 2014-06-03 Andrew Bennett <andrew.bennett@imgtec.com>
3 * config/mips/mips-cpus.def: Add mips32r3, mips32r5, mips64r3 and
5 * config/mips/mips-tables.opt: Regenerate.
6 * config/mips/mips.c (mips_compute_frame_info): Changed if statement
7 to use mips_isa_rev rather than ISA_MIPS32R2.
8 * config/mips/mips.h (ISA_MIPS32R3): New define.
9 (ISA_MIPS32R5): New define.
10 (ISA_MIPS64R3): New define.
11 (ISA_MIPS64R5): New define.
12 (TARGET_CPU_CPP_BUILTINS): Added support for ISA_MIPS32R3, ISA_MIPS32R5,
13 ISA_MIPS64R3 and ISA_MIPS64R5.
14 (MIPS_ISA_LEVEL_SPEC): Added support for mips32r3, mips32r5, mips64r3
16 (MIPS_ISA_SYNCI_SPEC): Likewise.
17 (ISA_HAS_64BIT_REGS): Added ISA_MIPS64R3 and ISA_MIPS64R5.
18 (LINK_SPEC): Added mips32r3 and mips32r5.
19 * config/mips/t-isa3264 (MULTILIB_MATCHES): Map mips32r3 and mips32r5
20 to mips32r2; and mips64r3 and mips64r5 to mips64r2.
21 * config/mips/t-mti-elf (MULTILIB_MATCHES): Likewise.
22 * config/mips/t-mti-linux (MULTILIB_MATCHES): Likewise.
23 * config/mips/t-sde (MULTILIB_MATCHES): Likewise.
24 * config/mips/t-sdemtk (MULTILIB_MATCHES): New define.
25 * doc/invoke.texi: Document mips32r3, mips32r5, mips64r3 and mips64r5.
27 2014-06-03 Andrew Bennett <andrew.bennett@imgtec.com>
29 * doc/invoke.texi: Document -mxpa and -mno-xpa MIPS command line
31 * config/mips/mips.opt (mxpa): New option.
32 * config/mips/mips.h (ASM_SPEC): Pass mxpa and mno-xpa to the
35 2014-06-03 Martin Jambor <mjambor@suse.cz>
38 * ipa-cp.c (cgraph_edge_brings_value_p): Handle edges leading to
41 2014-06-03 Thomas Preud'homme <thomas.preudhomme@arm.com>
43 PR tree-optimization/61328
44 * tree-ssa-math-opts.c (init_symbolic_number): Extract symbolic number
45 initialization from find_bswap_or_nop_1.
46 (find_bswap_or_nop_1): Test return value of find_bswap_or_nop_1 stored
47 in source_expr2 before using the size value the function sets. Also
48 make use of init_symbolic_number () in both the old place and
49 find_bswap_or_nop_load () to avoid reading uninitialized memory when
50 doing recursion in the GIMPLE_BINARY_RHS case.
52 2014-06-03 Richard Biener <rguenther@suse.de>
54 PR tree-optimization/61383
55 * tree-ssa-ifcombine.c (bb_no_side_effects_p): Make sure
58 2014-06-03 Richard Sandiford <rdsandiford@googlemail.com>
60 * defaults.h (USE_MD_CONSTRAINTS, EXTRA_MEMORY_CONSTRAINT)
61 (EXTRA_ADDRESS_CONSTRAINT, DEFAULT_CONSTRAINT_LEN, CONSTRAINT_LEN)
62 (CONST_OK_FOR_CONSTRAINT_P, CONST_DOUBLE_OK_FOR_LETTER_P)
63 (REG_CLASS_FROM_CONSTRAINT, EXTRA_CONSTRAINT_STR): Delete definitions
65 (REG_CLASS_FROM_LETTER, CONST_OK_FOR_LETTER_P)
66 (CONST_DOUBLE_OK_FOR_LETTER_P, EXTRA_CONSTRAINT): Move poising to...
67 * system.h: ...here and make it unconditional.
68 * target.def (conditional_register_usage): Mention
69 define_register_constraint instead of old-style constraint macros.
70 * doc/tm.texi.in: Remove documentation for old-style constraint macros.
71 * doc/tm.texi: Regenerate.
72 * genoutput.c: Remove USE_MD_CONSTRAINTS conditions and all code
73 protected by !USE_MD_CONSTRAINTS.
74 * config/frv/frv.md: Remove quote from old version of documentation.
75 * config/frv/frv.c (frv_conditional_register_usage): Likewise.
76 * config/m32r/m32r.c (easy_di_const, easy_df_const): Avoid mentioning
77 CONST_DOUBLE_OK_FOR_LETTER.
78 * config/sh/constraints.md: Likewise EXTRA_CONSTRAINT.
80 2014-06-02 Andrew Pinski <apinski@cavium.com>
82 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER):
83 /lib/ld-linux32-aarch64.so.1 is used for ILP32.
84 (LINUX_TARGET_LINK_SPEC): Update linker script for ILP32.
85 file whose name depends on -mabi= and -mbig-endian.
86 * config/aarch64/t-aarch64-linux (MULTILIB_OSDIRNAMES):
87 Handle LP64 better and handle ilp32 too.
88 (MULTILIB_OPTIONS): Delete.
89 (MULTILIB_DIRNAMES): Delete.
91 2014-06-02 Andrew MacLeod <amacleod@redhat.com>
93 * expr.h: Remove prototypes of functions defined in builtins.c.
94 * tree.h: (build_call_expr_*, build_string_literal): Add prototypes.
95 Remove prototypes of functions defined in builtins.c.
96 * builtins.h: Update prototype list to include all exported functions.
97 * builtins.c: (default_libc_has_function, gnu_libc_has_function,
98 no_c99_libc_has_function): Move to targhooks.c
99 (build_string_literal, build_call_expr_loc_array,
100 build_call_expr_loc_vec, build_call_expr_loc, build_call_expr): Move
102 (expand_builtin_object_size, fold_builtin_object_size): Make static.
103 * targhooks.c (default_libc_has_function, gnu_libc_has_function,
104 no_c99_libc_has_function): Relocate from builtins.c.
105 * tree.c: Include builtins.h.
106 (build_call_expr_loc_array, build_call_expr_loc_vec,
107 build_call_expr_loc, build_call_expr, build_string_literal): Relocate
109 * fold-const.h (fold_fma): Move prototype to builtins.h.
110 * realmpfr.h (do_mpc_arg2): Move prototype to builtins.h.
111 * asan.c: Include builtins.h.
112 * cfgexpand.c: Likewise.
113 * convert.c: Likewise.
114 * emit-rtl.c: Likewise.
115 * except.c: Likewise.
117 * fold-const.c: Likewise.
118 * gimple-fold.c: Likewise.
119 * gimple-ssa-strength-reduction.c: Likewise.
120 * gimplify.c: Likewise.
121 * ipa-inline.c: Likewise.
122 * ipa-prop.c: Likewise.
123 * lto-streamer-out.c: Likewise.
125 * tree-inline.c: Likewise.
126 * tree-object-size.c: Likewise.
127 * tree-sra.c: Likewise.
128 * tree-ssa-ccp.c: Likewise.
129 * tree-ssa-forwprop.c: Likewise.
130 * tree-ssa-loop-ivcanon.c: Likewise.
131 * tree-ssa-loop-ivopts.c: Likewise.
132 * tree-ssa-math-opts.c: Likewise.
133 * tree-ssa-reassoc.c: Likewise.
134 * tree-ssa-threadedge.c: Likewise.
135 * tree-streamer-in.c: Likewise.
136 * tree-vect-data-refs.c: Likewise.
137 * tree-vect-patterns.c: Likewise.
138 * tree-vect-stmts.c: Likewise.
139 * config/aarch64/aarch64.c: Likewise.
140 * config/alpha/alpha.c: Likewise.
141 * config/arc/arc.c: Likewise.
142 * config/arm/arm.c: Likewise.
143 * config/avr/avr.c: Likewise.
144 * config/bfin/bfin.c: Likewise.
145 * config/c6x/c6x.c: Likewise.
146 * config/cr16/cr16.c: Likewise.
147 * config/cris/cris.c: Likewise.
148 * config/epiphany/epiphany.c: Likewise.
149 * config/fr30/fr30.c: Likewise.
150 * config/frv/frv.c: Likewise.
151 * config/h8300/h8300.c: Likewise.
152 * config/i386/i386.c: Likewise.
153 * config/i386/winnt.c: Likewise.
154 * config/ia64/ia64.c: Likewise.
155 * config/iq2000/iq2000.c: Likewise.
156 * config/lm32/lm32.c: Likewise.
157 * config/m32c/m32c.c: Likewise.
158 * config/m32r/m32r.c: Likewise.
159 * config/m68k/m68k.c: Likewise.
160 * config/mcore/mcore.c: Likewise.
161 * config/mep/mep.c: Likewise.
162 * config/microblaze/microblaze.c: Likewise.
163 * config/mips/mips.c: Likewise.
164 * config/mmix/mmix.c: Likewise.
165 * config/mn10300/mn10300.c: Likewise.
166 * config/moxie/moxie.c: Likewise.
167 * config/msp430/msp430.c: Likewise.
168 * config/nds32/nds32.c: Likewise.
169 * config/pa/pa.c: Likewise.
170 * config/pdp11/pdp11.c: Likewise.
171 * config/picochip/picochip.c: Likewise.
172 * config/rl78/rl78.c: Likewise.
173 * config/rs6000/rs6000.c: Likewise.
174 * config/rx/rx.c: Likewise.
175 * config/s390/s390.c: Likewise.
176 * config/score/score.c: Likewise.
177 * config/sh/sh.c: Likewise.
178 * config/sparc/sparc.c: Likewise.
179 * config/spu/spu.c: Likewise.
180 * config/stormy16/stormy16.c: Likewise.
181 * config/tilegx/tilegx.c: Likewise.
182 * config/tilepro/tilepro.c: Likewise.
183 * config/v850/v850.c: Likewise.
184 * config/vax/vax.c: Likewise.
185 * config/xtensa/xtensa.c: Likewise.
187 2014-06-02 Jeff Law <law@redhat.com>
189 PR rtl-optimization/61094
190 * ree.c (combine_reaching_defs): Do not reextend an insn if it
191 was marked as do_no_reextend. If a copy is needed to eliminate
192 an extension, then mark it as do_not_reextend.
194 2014-06-02 Marcus Shawcroft <marcus.shawcroft@arm.com>
196 * config/aarch64/aarch64.md (set_fpcr): Drop ISB after FPCR write.
198 2014-06-02 Richard Henderson <rth@redhat.com>
201 * config/alpha/alpha.c (print_operand_address): Allow symbolic
202 addresses inside asms. Use output_operand_lossage instead of
205 2014-06-02 Uros Bizjak <ubizjak@gmail.com>
208 * config/i386/i386.c (ix86_expand_vec_perm) [case V32QImode]: Use
209 GEN_INT (-128) instead of GEN_INT (128) to set MSB of QImode constant.
211 2014-06-02 Tom de Vries <tom@codesourcery.com>
213 * config/aarch64/aarch64.c (aarch64_float_const_representable_p): Handle
214 case that x has VOIDmode.
216 2014-06-02 Bernd Schmidt <bernds@codesourcery.com>
218 * varasm.c (copy_constant): Delete function.
219 (build_constant_desc): Don't call it.
221 2014-06-02 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
224 * config/arm/arm.h (TARGET_SUPPORTS_WIDE_INT): Define.
225 * config/arm/arm.md (mov64 splitter): Replace const_double_operand
226 with immediate_operand.
228 2014-06-02 Andreas Schwab <schwab@suse.de>
231 (ia64_first_cycle_multipass_dfa_lookahead_guard): Check
232 pending_data_specs first.
234 2014-06-02 Richard Biener <rguenther@suse.de>
236 PR tree-optimization/61378
237 * tree-ssa-sccvn.c (vn_reference_lookup_3): Initialize
240 2014-06-01 Uros Bizjak <ubizjak@gmail.com>
242 * config/i386/constraints.md (Bw): Rename from 'w'.
243 (Bz): Rename from 'z'.
244 * config/i386/i386.md: Change 'w' to 'Bw' and 'z' to 'Bz' globally.
246 2014-06-01 Kai Tietz <ktietz@redhat.com>
249 * config/i386/constrains.md (define_constrain): New 'Bs' constraint.
250 * config/i386/i386.md (sibcall_insn_operand): Use Bs
251 instead of m constraint.
253 2014-05-31 Andreas Schwab <schwab@linux-m68k.org>
255 * config/m68k/m68k.md (beq0_di, bne0_di): Make the "o" constraint
256 a separate alternative where the scratch operand 2 is marked as
259 2014-05-31 Kugan Vivekanandarajah <kuganv@linaro.org>
261 * config/arm/arm.c (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New define.
262 (arm_builtins) : Add ARM_BUILTIN_GET_FPSCR and ARM_BUILTIN_SET_FPSCR.
263 (bdesc_2arg) : Add description for builtins __builtins_arm_set_fpscr
264 and __builtins_arm_get_fpscr.
265 (arm_init_builtins) : Initialize builtins __builtins_arm_set_fpscr and
266 __builtins_arm_get_fpscr.
267 (arm_expand_builtin) : Expand builtins __builtins_arm_set_fpscr and
268 __builtins_arm_ldfpscr.
269 (arm_atomic_assign_expand_fenv): New function.
270 * config/arm/vfp.md (set_fpscr): New pattern.
271 (get_fpscr) : Likewise.
272 * config/arm/unspecs.md (unspecv): Add VUNSPEC_GET_FPSCR and
274 * doc/extend.texi (AARCH64 Built-in Functions) : Document
275 __builtins_arm_set_fpscr, __builtins_arm_get_fpscr.
277 2014-05-30 Jakub Jelinek <jakub@redhat.com>
279 * asan.c (report_error_func): Add SLOW_P argument, use
280 BUILT_IN_ASAN_*_N if set.
281 (build_check_stmt): Likewise.
282 (instrument_derefs): If T has insufficient alignment,
283 force same handling as for odd sizes.
285 * sanitizer.def (BUILT_IN_ASAN_REPORT_LOAD_N,
286 BUILT_IN_ASAN_REPORT_STORE_N): New.
287 * asan.c (struct asan_mem_ref): Change access_size type to
289 (asan_mem_ref_init, asan_mem_ref_new, get_mem_refs_of_builtin_call,
290 update_mem_ref_hash_table): Likewise.
291 (asan_mem_ref_hasher::hash): Hash in a HWI.
292 (report_error_func): Change size_in_bytes argument to HWI.
293 Use *_N builtins if size_in_bytes is larger than 16 or not power of
295 (build_shadow_mem_access): New function.
296 (build_check_stmt): Use it. Change size_in_bytes argument to HWI.
297 Handle size_in_bytes not power of two or larger than 16.
298 (instrument_derefs): Don't give up if size_in_bytes is not
299 power of two or is larger than 16.
301 2014-05-30 Kai Tietz <ktietz@redhat.com>
304 * config/i386/i386.c (x86_output_mi_thunk): Add memory case
305 for sibling-tail-calls.
306 * config/i386/i386.md (sibcall_insn_operand): Add memory-constrain
308 * config/i386/predicates.md (sibcall_memory_operand): New predicate.
309 (sibcall_insn_operand): Add check for sibcall_memory_operand.
311 2014-05-30 Pitchumani Sivanupandi <pitchumani.s@atmel.com>
313 * config/avr/avr-mcus.def: Change ATA6289 ISA to AVR4
314 * config/avr/avr-tables.opt: Regenerate.
315 * config/avr/t-multilib: Regenerate.
316 * doc/avr-mmcu.texi: Regenerate.
318 2014-05-30 Ian Lance Taylor <iant@google.com>
320 * config/i386/xmmintrin.h (_mm_pause): Move out of scope of pragma
323 2014-05-30 Tom de Vries <tom@codesourcery.com>
325 * config/i386/i386.c (TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS):
328 2014-05-30 Tom de Vries <tom@codesourcery.com>
330 * lra-int.h (struct lra_reg): Add field actual_call_used_reg_set.
331 * lra.c (initialize_lra_reg_info_element): Add init of
332 actual_call_used_reg_set field.
333 (lra): Call lra_create_live_ranges before lra_inheritance for
335 * lra-assigns.c (lra_assign): Allow call_used_regs to cross calls for
337 * lra-constraints.c (need_for_call_save_p): Use actual_call_used_reg_set
338 instead of call_used_reg_set for -fuse-caller-save.
339 * lra-lives.c (process_bb_lives): Calculate actual_call_used_reg_set.
341 2014-05-30 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
343 * config/arm/thumb2.md (*thumb2_movhi_insn): Set type of movw
345 * config/arm/vfp.md (*thumb2_movsi_vfp): Likewise.
347 2014-05-30 Richard Sandiford <rdsandiford@googlemail.com>
349 * ira.c (ira_get_dup_out_num): Check for output operands at
350 the start of the loop. Handle cases where an included alternative
351 follows an excluded one.
353 2014-05-29 Mike Stump <mikestump@comcast.net>
356 * collect2.c (maybe_run_lto_and_relink): Be sure to always run
357 post ld passes when lto is used.
359 2014-05-29 Vladimir Makarov <vmakarov@redhat.com>
361 PR rtl-optimization/61325
362 * lra-constraints.c (process_address): Rename to process_address_1.
363 (process_address): New function.
365 2014-05-29 Alan Lawrence <alan.lawrence@arm.com>
367 * config/aarch64/aarch64-builtins.c (aarch64_types_binopv_qualifiers,
368 TYPES_BINOPV): New static data.
369 * config/aarch64/aarch64-simd-builtins.def (im_lane_bound):
371 * config/aarch64/aarch64-simd.md (aarch64_ext,
372 aarch64_im_lane_boundsi): New patterns.
373 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_const_1): Match
375 (aarch64_evpc_ext): New function.
377 * config/aarch64/iterators.md (UNSPEC_EXT): New enum element.
379 * config/aarch64/arm_neon.h (vext_f32, vext_f64, vext_p8, vext_p16,
380 vext_s8, vext_s16, vext_s32, vext_s64, vext_u8, vext_u16, vext_u32,
381 vext_u64, vextq_f32, vextq_f64, vextq_p8, vextq_p16, vextq_s8,
382 vextq_s16, vextq_s32, vextq_s64, vextq_u8, vextq_u16, vextq_u32,
383 vextq_u64): Replace __asm with __builtin_shuffle and im_lane_boundsi.
385 2014-05-29 Tom de Vries <tom@codesourcery.com>
387 * rtl.h (BLOCK_SYMBOL_CHECK): Use SYMBOL_REF_FLAGS.
389 2014-05-29 Richard Earnshaw <rearnsha@arm.com>
390 Richard Sandiford <rdsandiford@googlemail.com>
392 * arm/iterators.md (shiftable_ops): New code iterator.
393 (t2_binop0, arith_shift_insn): New code attributes.
394 * arm/predicates.md (shift_nomul_operator): New predicate.
395 * arm/arm.md (insn_enabled): Delete.
396 (enabled): Remove insn_enabled test.
397 (*arith_shiftsi): Delete. Replace with ...
398 (*<arith_shift_insn>_multsi): ... new pattern.
399 (*<arith_shift_insn>_shiftsi): ... new pattern.
400 * config/arm/arm.c (arm_print_operand): Handle operand format 'b'.
402 2014-05-29 Radovan Obradovic <robradovic@mips.com>
403 Tom de Vries <tom@codesourcery.com>
405 * config/mips/mips.h (POST_CALL_TMP_REG): Define.
406 * config/mips/mips.c (mips_emit_call_insn): Add POST_CALL_TMP_REG
408 (mips_split_call): Use POST_CALL_TMP_REG.
409 (TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): Redefine to true.
411 2014-05-29 Tom de Vries <tom@codesourcery.com>
413 * final.c (collect_fn_hard_reg_usage): Guard variable declaration
414 with #ifdef STACK_REGS.
416 2014-05-28 Jan Hubicka <hubicka@ucw.cz>
418 * varasm.c (get_variable_section): Walk aliases.
419 (place_block_symbol): Walk aliases.
421 2014-05-28 Tom de Vries <tom@codesourcery.com>
424 2014-05-28 Tom de Vries <tom@codesourcery.com>
426 * lra-int.h (struct lra_reg): Add field actual_call_used_reg_set.
427 * lra.c (initialize_lra_reg_info_element): Add init of
428 actual_call_used_reg_set field.
429 (lra): Call lra_create_live_ranges before lra_inheritance for
431 * lra-assigns.c (lra_assign): Allow call_used_regs to cross calls for
433 * lra-constraints.c (need_for_call_save_p): Use
434 actual_call_used_reg_set instead of call_used_reg_set for
436 * lra-lives.c (process_bb_lives): Calculate actual_call_used_reg_set.
438 2014-05-28 Richard Sandiford <rdsandiford@googlemail.com>
440 * doc/md.texi: Document that the % constraint character must
441 be at the beginning of the string.
442 * genoutput.c (validate_insn_alternatives): Check that '=',
443 '+' and '%' only appear at the beginning of a constraint.
444 * ira.c (commutative_constraint_p): Delete.
445 (ira_get_dup_out_num): Expect the '%' commutativity marker to be
446 at the start of the string.
447 * config/alpha/alpha.md (*movmemdi_1, *clrmemdi_1): Remove
449 * config/arm/neon.md (bicdi3_neon): Likewise.
450 * config/iq2000/iq2000.md (addsi3_internal, subsi3_internal, sgt_si)
451 (slt_si, sltu_si): Likewise.
452 * config/vax/vax.md (sbcdi3): Likewise.
453 * config/h8300/h8300.md (*cmpstz): Remove duplicate '+'.
454 * config/arc/arc.md (mulsi_600, mulsidi_600, umulsidi_600)
455 (mul64): Move '%' to beginning of constraint.
456 * config/arm/arm.md (*xordi3_insn): Likewise.
457 * config/nds32/nds32.md (add<mode>3, mulsi3, andsi3, iorsi3)
460 2014-05-28 Richard Sandiford <rdsandiford@googlemail.com>
462 * doc/md.texi: Document the restrictions on the "enabled" attribute.
464 2014-05-28 Jason Merrill <jason@redhat.com>
467 * cgraph.h (symtab_node::get_comdat_group_id): New.
468 * cgraphunit.c (analyze_functions): Call it.
469 * symtab.c (dump_symtab_node): Likewise.
470 * tree.c (decl_comdat_group_id): New.
471 * tree.h: Declare it.
472 * lto-streamer-out.c (write_symbol): Use it.
473 * trans-mem.c (ipa_tm_create_version_alias): Likewise.
475 2014-05-28 Francois-Xavier Coudert <fxcoudert@gcc.gnu.org>
478 * wide-int.cc: Do not include longlong.h when compiling with clang.
480 2014-05-28 Richard Biener <rguenther@suse.de>
482 * tree-ssa-propagate.c (add_control_edge): Print less vertical space.
483 * tree-vrp.c (extract_range_from_ssa_name): Also copy VR_UNDEFINED.
484 (vrp_visit_assignment_or_call): Print less vertical space.
485 (vrp_visit_stmt): Likewise.
486 (vrp_visit_phi_node): Likewise. For a PHI argument with
487 VR_VARYING range consider recording it as copy.
489 2014-05-28 Richard Biener <rguenther@suse.de>
492 2014-05-28 Richard Biener <rguenther@suse.de>
494 * hwint.h (HOST_WIDE_INT_PRINT_*): Define in terms of PRI*64.
496 2014-05-28 Bernd Edlinger <bernd.edlinger@hotmail.de>
498 * expr.c (expand_assignment): Fold the bitpos in the to_rtx if
499 sufficiently aligned and an offset is used at the same time.
500 (expand_expr_real_1): Likewise.
502 2014-05-28 Richard Biener <rguenther@suse.de>
505 * fold-const.c (fold_comparison): When folding
506 X +- C1 CMP Y +- C2 to X CMP Y +- C2 +- C1 also ensure
507 the sign of the remaining constant operand stays the same.
509 2014-05-28 Kaushik Phatak <kaushik.phatak@kpit.com>
511 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
512 __RL78_64BIT_DOUBLES__ or __RL78_32BIT_DOUBLES__.
513 (ASM_SPEC): Pass -m64bit-doubles or -m32bit-doubles on
515 (DOUBLE_TYPE_SIZE): Use 64 bit if TARGET_64BIT_DOUBLES is true.
516 * gcc/config/rl78/rl78.opt (m64bit-doubles): New option.
517 (m32bit-doubles) Likewise.
518 * gcc/config/rl78/t-rl78: Add 64-bit-double multilib.
519 * doc/invoke.texi: Document -m32bit-doubles and -m64bit-doubles
522 2014-05-28 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
524 * configure.ac ($gcc_cv_ld_clearcap): New test.
525 * configure: Regenerate.
526 * config.in: Regenerate.
527 * config/sol2.opt (mclear-hwcap): New option.
528 * config/sol2.h (LINK_CLEARCAP_SPEC): Define.
529 * config/sol2-clearcap.map: Moved here from
530 testsuite/gcc.target/i386/clearcap.map.
531 * config/sol2-clearcapv2.map: Move here from
532 gcc.target/i386/clearcapv2.map.
533 * config/t-sol2 (install): Depend on install-clearcap-map.
534 (install-clearcap-map): New target.
535 * doc/invoke.texi (Option Summary, Solaris 2 Options): Document
538 2014-05-28 Richard Biener <rguenther@suse.de>
540 * hwint.h (*_HALF_WIDE_INT*): Move to ...
541 * wide-int.cc (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT):
542 ... here and remove the rest.
543 * hwint.h (HOST_WIDE_INT_PRINT_*): Define in terms of PRI*64.
545 2014-05-28 Richard Biener <rguenther@suse.de>
547 PR tree-optimization/61335
548 * tree-vrp.c (vrp_visit_phi_node): If the compare of old and
549 new range fails, drop to varying.
551 2014-05-28 Tom de Vries <tom@codesourcery.com>
553 * lra-int.h (struct lra_reg): Add field actual_call_used_reg_set.
554 * lra.c (initialize_lra_reg_info_element): Add init of
555 actual_call_used_reg_set field.
556 (lra): Call lra_create_live_ranges before lra_inheritance for
558 * lra-assigns.c (lra_assign): Allow call_used_regs to cross calls for
560 * lra-constraints.c (need_for_call_save_p): Use
561 actual_call_used_reg_set instead of call_used_reg_set for
563 * lra-lives.c (process_bb_lives): Calculate actual_call_used_reg_set.
565 2014-05-28 Radovan Obradovic <robradovic@mips.com>
566 Tom de Vries <tom@codesourcery.com>
568 * doc/invoke.texi (@item Optimization Options): Add -fuse-caller-save
570 (@item -fuse-caller-save): New item.
572 2014-05-28 Radovan Obradovic <robradovic@mips.com>
573 Tom de Vries <tom@codesourcery.com>
575 * opts.c (default_options_table): Add OPT_LEVELS_2_PLUS entry with
576 OPT_fuse_caller_save.
578 2014-05-28 Radovan Obradovic <robradovic@mips.com>
579 Tom de Vries <tom@codesourcery.com>
581 * df-scan.c (df_get_call_refs): Use get_call_reg_set_usage.
582 * caller-save.c (setup_save_areas, save_call_clobbered_regs): Use
583 get_call_reg_set_usage.
584 * resource.c (mark_set_resources, mark_target_live_regs): Use
585 get_call_reg_set_usage.
586 * ira-int.h (struct ira_allocno): Add crossed_calls_clobbered_regs
588 (ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS): Define.
589 * ira-lives.c (process_bb_node_lives): Use get_call_reg_set_usage.
590 Calculate ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS.
591 * ira-build.c (ira_create_allocno): Init
592 ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS.
593 (create_cap_allocno, propagate_allocno_info)
594 (propagate_some_info_from_allocno)
595 (copy_info_to_removed_store_destinations): Handle
596 ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS.
597 * ira-costs.c (ira_tune_allocno_costs): Use
598 ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS to adjust costs.
600 2014-05-28 Radovan Obradovic <robradovic@mips.com>
601 Tom de Vries <tom@codesourcery.com>
603 * cgraph.h (struct cgraph_rtl_info): Add function_used_regs
604 and function_used_regs_valid fields.
605 * final.c: Move include of hard-reg-set.h to before rtl.h to declare
606 find_all_hard_reg_sets.
607 (collect_fn_hard_reg_usage, get_call_fndecl, get_call_cgraph_rtl_info)
608 (get_call_reg_set_usage): New function.
609 (rest_of_handle_final): Use collect_fn_hard_reg_usage.
610 * regs.h (get_call_reg_set_usage): Declare.
612 2014-05-28 Georg-Johann Lay <avr@gjlay.de>
615 * config/dbx.h (License): Add Runtime Library Exception.
616 * config/newlib-stdint.h (License): Same.
617 * config/rtems.h (License): Same
618 * config/initfini-array.h (License): Same
619 * config/v850/v850.h (License): Same.
620 * config/v850/v850-opts.h (License): Same
621 * config/v850/rtems.h (License): Same.
623 2014-05-28 Georg-Johann Lay <avr@gjlay.de>
626 * doc/extend.texi (Local Labels): Note that label differences are
627 not supported for AVR.
629 2014-05-27 Andrew Pinski <apinski@cavium.com>
631 * config/aarch64/aarch64.md (stack_protect_set_<mode>):
632 Use <w> for the register in assembly template.
633 (stack_protect_test): Use the mode of operands[0] for the result.
634 (stack_protect_test_<mode>): Use <w> for the register
635 in assembly template.
637 2014-05-27 DJ Delorie <dj@redhat.com>
639 * config/rx/rx.c (add_vector_labels): New.
640 (rx_output_function_prologue): Call it.
641 (rx_handle_func_attribute): Don't require empty arguments.
642 (rx_handle_vector_attribute): New.
643 (rx_attribute_table): Add "vector" attribute.
644 * doc/extend.texi (interrupt, vector): Document new/changed
645 RX-specific attributes.
647 * config/rx/rx.c (rx_adjust_insn_length): Skip for non-insns.
649 2014-05-27 Eric Botcazou <ebotcazou@adacore.com>
651 * double-int.c (div_and_round_double) <ROUND_DIV_EXPR>: Use the proper
652 predicate to detect a negative quotient.
654 2014-05-27 Eric Botcazou <ebotcazou@adacore.com>
656 * fold-const.c (fold_comparison): Clean up and extend X +- C1 CMP C2
657 to X CMP C2 -+ C1 transformation to EQ_EXPR/NE_EXPR.
658 Add X - Y CMP 0 to X CMP Y transformation.
659 (fold_binary_loc) <EQ_EXPR/NE_EXPR>: Remove same transformations.
661 2014-05-27 Segher Boessenkool <segher@kernel.crashing.org>
663 * stmt.c (dump_case_nodes): Don't convert values to HOST_WIDE_INT
666 2014-05-27 Steve Ellcey <sellcey@mips.com>
668 * config/mips/mips.c: Add include of cgraph.h.
670 2014-05-27 Richard Biener <rguenther@suse.de>
672 * system.h (__STDC_FORMAT_MACROS): Define as very first thing.
674 2014-05-27 Georg-Johann Lay <avr@gjlay.de>
677 * config/arm/arm.h (License): Add note to COPYING.RUNTIME.
678 * config/arm/arm-cores.def (License): Same.
679 * config/arm/arm-opts.h (License): Same.
680 * config/arm/aout.h (License): Same.
681 * config/arm/bpabi.h (License): Same.
682 * config/arm/elf.h (License): Same.
683 * config/arm/linux-elf.h (License): Same.
684 * config/arm/linux-gas.h (License): Same.
685 * config/arm/netbsd-elf.h (License): Same.
686 * config/arm/uclinux-eabi.h (License): Same.
687 * config/arm/uclinux-elf.h (License): Same.
688 * config/arm/vxworks.h (License): Same.
690 2014-05-27 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
692 * config/arm/neon.md (neon_bswap<mode>): New pattern.
693 * config/arm/arm.c (neon_itype): Add NEON_BSWAP.
694 (arm_init_neon_builtins): Handle NEON_BSWAP.
695 Define required type nodes.
696 (arm_expand_neon_builtin): Handle NEON_BSWAP.
697 (arm_builtin_vectorized_function): Handle BUILTIN_BSWAP builtins.
698 * config/arm/arm_neon_builtins.def (bswap): Define builtins.
699 * config/arm/iterators.md (VDQHSD): New mode iterator.
701 2014-05-27 Richard Biener <rguenther@suse.de>
703 * tree-vrp.c (vrp_evaluate_conditional_warnv_with_ops_using_ranges):
704 Try using literal operands when comparing value-ranges failed.
706 2014-05-27 Richard Sandiford <rdsandiford@googlemail.com>
708 * ira.c (commutative_operand): Adjust for change to recog_data.
709 [Missing from previous commit.]
711 2014-05-27 Richard Sandiford <rdsandiford@googlemail.com>
713 * system.h (TEST_BIT): New macro.
714 * recog.h (alternative_mask): New type.
715 (ALL_ALTERNATIVES, ALTERNATIVE_BIT): New macros.
716 (recog_data_d): Replace alternative_enabled_p array with
717 enabled_alternatives.
718 (target_recog): New structure.
719 (default_target_recog, this_target_recog): Declare.
720 (get_enabled_alternatives, recog_init): Likewise.
721 * recog.c (default_target_recog, this_target_recog): New variables.
722 (get_enabled_alternatives): New function.
723 (extract_insn): Use it.
724 (recog_init): New function.
725 (preprocess_constraints, constrain_operands): Adjust for change to
727 * postreload.c (reload_cse_simplify_operands): Likewise.
728 * reload.c (find_reloads): Likewise.
729 * ira-costs.c (record_reg_classes): Likewise.
730 * ira-lives.c (single_reg_class): Likewise. Fix bug in which
731 all alternatives after a disabled one would be skipped.
732 (ira_implicitly_set_insn_hard_regs): Likewise.
733 * ira.c (ira_setup_alts): Adjust for change to recog_data.
734 * lra-int.h (lra_insn_recog_data): Replace alternative_enabled_p
735 with enabled_alternatives.
736 * lra.c (free_insn_recog_data): Update accordingly.
737 (lra_update_insn_recog_data): Likewise.
738 (lra_set_insn_recog_data): Likewise. Use get_enabled_alternatives.
739 * lra-constraints.c (process_alt_operands): Likewise. Handle
740 only_alternative as part of the enabled mask.
741 * target-globals.h (this_target_recog): Declare.
742 (target_globals): Add a recog field.
743 (restore_target_globals): Restore this_target_recog.
744 * target-globals.c: Include recog.h.
745 (default_target_globals): Initialize recog field.
746 (save_target_globals): Likewise.
747 * reginfo.c (reinit_regs): Call recog_init.
748 * toplev.c (backend_init_target): Likewise.
750 2014-05-27 Richard Sandiford <rdsandiford@googlemail.com>
752 * gencodes.c (main): Make LAST_INSN_CODE higher than any insn code,
753 rather than any named insn's code.
755 2014-05-27 Georg-Johann Lay <avr@gjlay.de>
758 * config/arm/arm-opts.h (License): Add GCC Runtime Library Exception.
759 * config/arm/arm-cores.def (License): Same.
761 2014-05-26 Jan Hubicka <hubicka@ucw.cz>
763 * tree.h (decl_comdat_group): Declare.
764 * cgraph.h (symtab_in_same_comdat_p): Move offline to ...
765 * tree.c (decl_comdat_group): Here.
767 2014-05-26 Richard Sandiford <r.sandiford@uk.ibm.com>
769 PR rtl-optimization/61222
770 * combine.c (simplify_shift_const_1): When moving a PLUS outside
771 the shift, truncate the PLUS operand to the result mode.
773 2014-05-26 Uros Bizjak <ubizjak@gmail.com>
776 * config/i386/i386.c (ix86_rtx_costs)
777 <case CONST_INT, case CONST, case LABEL_REF, case SYMBOL_REF>:
780 2014-05-26 Martin Jambor <mjambor@suse.cz>
782 * ira.c (split_live_ranges_for_shrink_wrap): Remove bailout on
785 2014-05-26 Richard Biener <rguenther@suse.de>
787 * wide-int.h (wi::int_traits <long>, wi::int_traits <unsigned long>,
788 wi::int_traits <long long>, wi::int_traits <unsigned long long>):
789 Provide specializations.
790 (wi::int_traits <HOST_WIDE_INT>,
791 wi::int_traits <unsigned HOST_WIDE_INT>): Remove specializations.
793 2014-05-26 Alan Modra <amodra@gmail.com>
796 * config/rs6000/rs6000.c (rs6000_emit_set_const): Remove unneeded
797 params and return a bool. Remove dead code. Update comment.
798 Assert we have a const_int source. Remove bogus code from
799 32-bit HWI days. Move !TARGET_POWERPC64 handling, and correct
800 handling of constants > 2G and reg_equal note, from..
801 (rs6000_emit_set_long_const): ..here. Remove unneeded param and
802 return value. Update comment. If we can, use a new pseudo
803 for intermediate calculations.
804 * config/rs6000/rs6000-protos.h (rs6000_emit_set_const): Update
806 * config/rs6000/rs6000.md (movsi_internal1_single+1): Update
807 call to rs6000_emit_set_const in splitter.
808 (movdi_internal64+2, +3): Likewise.
810 2014-05-26 Richard Biener <rguenther@suse.de>
812 * system.h: Define __STDC_FORMAT_MACROS before
813 including inttypes.h.
814 * hwint.h (HOST_WIDEST_INT, HOST_BITS_PER_WIDEST_INT,
815 HOST_WIDEST_INT_PRINT, HOST_WIDEST_INT_PRINT_DEC,
816 HOST_WIDEST_INT_PRINT_DEC_C, HOST_WIDEST_INT_PRINT_UNSIGNED,
817 HOST_WIDEST_INT_PRINT_HEX, HOST_WIDEST_INT_PRINT_DOUBLE_HEX,
818 HOST_WIDEST_INT_C): Remove.
819 (PRId64, PRIi64, PRIo64, PRIu64, PRIx64, PRIX64): Define
820 if C99 inttypes.h is not available.
821 * coretypes.h (gcov_type, gcov_type_unsigned): Use [u]int64_t.
822 * gcov-io.h (gcov_type, gcov_type_unsigned): Likewise.
823 * gcov-io.c (gcov_histo_index): Drop non-64bit hwi case.
824 * cfgloop.h (struct niter_desc): Use uint64_t for niter field.
825 * bitmap.c (struct bitmap_descriptor_d): Use uint64_t for counters.
826 (struct output_info): Likewise.
827 (print_statistics): Adjust.
828 (dump_bitmap_statistics): Likewise.
829 * bt-load.c (migrate_btr_defs): Print with PRId64.
830 * cfg.c (dump_edge_info, dump_bb_info): Likewise.
831 (MAX_SAFE_MULTIPLIER): Adjust.
832 * cfghooks.c (dump_bb_for_graph): Print with PRId64.
833 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee,
834 dump_cgraph_node): Likewise.
835 * final.c (dump_basic_block_info): Likewise.
836 * gcov-dump.c (tag_counters, tag_summary, dump_working_sets): Likewise.
837 * gcov.c (format_gcov): Likewise.
838 * ipa-cp.c (good_cloning_opportunity_p): Likewise. Use int64_t
840 (get_clone_agg_value): Use HOST_WIDE_INT for offset.
841 * ipa-inline.c (compute_max_insns): Use int64_t for calcuation.
842 (inline_small_functions, dump_overall_stats, dump_inline_stats):
843 Use PRId64 for dumping.
844 * ipa-profile.c (dump_histogram, ipa_profile): Likewise.
845 * ira-color.c (struct allocno_hard_regs): Use int64_t for cost.
846 (add_allocno_hard_regs): Adjust.
847 * loop-doloop.c (doloop_modify): Print using PRId64.
848 * loop-iv.c (inverse): Compute in uint64_t.
849 (determine_max_iter, iv_number_of_iterations): Likewise.
850 * loop-unroll.c (decide_peel_completely, decide_peel_simple):
852 * lto-streamer-out.c (write_symbol): Use uint64_t.
853 * mcf.c (CAP_INFINITY): Use int64_t maximum.
854 (dump_fixup_edge, create_fixup_graph, cancel_negative_cycle,
855 find_max_flow, adjust_cfg_counts): Use int64_t and dump with PRId64.
856 * modulo-sched.c (const_iteration_count): Use int64_t.
857 (sms_schedule): Dump using PRId64.
858 * predict.c (dump_prediction): Likewise.
859 * pretty-print.h (pp_widest_integer): Remove.
860 * profile.c (get_working_sets, is_edge_inconsistent,
861 is_inconsistent, read_profile_edge_counts): Dump using PRId64.
862 * tree-pretty-print.c (pp_double_int): Remove case handling
863 HOST_BITS_PER_DOUBLE_INT == HOST_BITS_PER_WIDEST_INT.
864 * tree-ssa-math-opts.c (struct symbolic_number): Use uint64_t
866 (pass_optimize_bswap::execute): Remove restriction on hosts.
867 * tree-streamer-in.c (streamer_alloc_tree): Use HOST_WIDE_INT.
868 * tree-streamer-out.c (streamer_write_tree_header): Likewise.
869 * tree.c (widest_int_cst_value): Remove.
870 * tree.h (widest_int_cst_value): Likewise.
871 * value-prof.c (dump_histogram_value): Print using PRId64.
872 * gengtype.c (main): Also inject int64_t.
873 * ggc-page.c (struct max_alignment): Use int64_t.
874 * alloc-pool.c (struct allocation_object_def): Likewise.
875 * ira-conflicts.c (build_conflict_bit_table): Use uint64_t
877 * doc/tm.texi.in: Remove reference to HOST_WIDEST_INT.
878 * doc/tm.texi: Regenerated.
879 * gengtype-lex.l (IWORD): Handle [u]int64_t.
880 * config/sh/sh.c (expand_cbranchdi4): Use gcov_type.
881 * config/mmix/mmix-protos.h (mmix_intval, mmix_shiftable_wyde_value,
882 mmix_output_register_setting): Use [u]int64_t in prototypes.
883 * config/mmix/mmix.c (mmix_print_operand, mmix_output_register_setting,
884 mmix_shiftable_wyde_value, mmix_output_shiftvalue_op_from_str,
885 mmix_output_octa, mmix_output_shifted_value): Adjust.
886 (mmix_intval): Adjust. Remove unreachable case.
887 * config/mmix/mmix.md (*nonlocal_goto_receiver_expanded): Use int64_t.
889 2014-05-26 Richard Biener <rguenther@suse.de>
891 * configure.ac: Drop __int64 type check. Insist that we
892 found uint64_t and int64_t.
893 * hwint.h (HOST_BITS_PER___INT64): Remove.
894 (HOST_BITS_PER_WIDE_INT): Define to 64 and remove __int64 case.
895 (HOST_WIDE_INT_PRINT_*): Remove 32bit case.
896 (HOST_WIDEST_INT*): Define to HOST_WIDE_INT*.
897 (HOST_WIDEST_FAST_INT): Remove __int64 case.
898 * vmsdbg.h (struct _DST_SRC_COMMAND): Use int64_t
899 for dst_q_src_df_rms_cdt.
900 * configure: Regenerate.
901 * config.in: Likewise.
903 2014-05-26 Michael Tautschnig <mt@debian.org>
906 * doc/extend.texi (X86 Built-in Functions): Fix parameter lists of
907 __builtin_ia32_vfrczs[sd] and __builtin_ia32_mpsadbw256.
909 2014-05-26 Zhenqiang Chen <zhenqiang.chen@linaro.org>
911 PR rtl-optimization/61278
912 * shrink-wrap.c (move_insn_for_shrink_wrap): Check df_live.
914 2014-05-26 Zhenqiang Chen <zhenqiang.chen@linaro.org>
916 PR rtl-optimization/61220
917 Part of PR rtl-optimization/61225
918 * shrink-wrap.c (move_insn_for_shrink_wrap): Skip SP and FP adjustment
919 insn; skip split_edge for a block with only one successor.
921 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
923 * symtab.c (symtab_nonoverwritable_alias): Copy READONLY flag
926 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
928 * ipa-visibility.c (can_replace_by_local_alias_in_vtable): New function.
929 (update_vtable_references): New function.
930 (function_and_variable_visibility): Rewrite also vtable initializers.
931 * varpool.c (cgraph_variable_initializer_availability): Remove assert.
933 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
935 * ggc.h (ggc_grow): New function.
936 * ggc-none.c (ggc_grow): New function.
937 * ggc-page.c (ggc_grow): Likewise.
939 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
941 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
942 address_taken_from_non_vtable_p, comdat_can_be_unshared_p_1,
943 comdat_can_be_unshared_p, cgraph_externally_visible_p,
944 varpool_externally_visible_p, can_replace_by_local_alias,
945 update_visibility_by_resolution_info, function_and_variable_visibility,
946 pass_data_ipa_function_and_variable_visibility,
947 make_pass_ipa_function_and_variable_visibility,
948 whole_program_function_and_variable_visibility,
949 pass_data_ipa_whole_program_visibility,
950 make_pass_ipa_whole_program_visibility): Move to ipa-visibility.c
951 * cgraph.h (cgraph_local_node_p): Declare.
952 * ipa-visibility.c: New file.
953 * Makefile.in (OBJS): Add ipa-visiblity.o
955 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
957 * gimple-fold.c (can_refer_decl_in_current_unit_p): Be sure
958 that var decl is available.
960 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
962 * tree-core.h (tree_decl_with_vis): Replace comdat_group by
964 * tree.c (copy_node_stat): Be sure to not copy symtab_node pointer.
965 (find_decls_types_r): Do not walk COMDAT_GROUP.
966 * tree.h (DECL_COMDAT_GROUP): Revamp to use decl_comdat_group.
967 * varasm.c (make_decl_one_only): Use set_comdat_group;
968 create node if needed.
969 * ipa-inline-transform.c (save_inline_function_body): Update
970 way we decl->symtab mapping.
971 * symtab.c (symtab_hash, hash_node, eq_node
972 symtab_insert_node_to_hashtable): Remove.
973 (symtab_register_node): Update.
974 (symtab_unregister_node): Update.
975 (symtab_get_node): Reimplement as inline function.
976 (symtab_add_to_same_comdat_group): Update.
977 (symtab_dissolve_same_comdat_group_list): Update.
978 (dump_symtab_base): Update.
979 (verify_symtab_base): Update.
980 (symtab_make_decl_local): Update.
981 (fixup_same_cpp_alias_visibility): Update.
982 (symtab_nonoverwritable_alias): Update.
983 * cgraphclones.c (set_new_clone_decl_and_node_flags): Update.
984 * ipa.c (update_visibility_by_resolution_info): UPdate.
985 * bb-reorder.c: Include cgraph.h
986 * lto-streamer-out.c (DFS_write_tree_body, hash_tree): Do not deal
988 * ipa-comdats.c (set_comdat_group, ipa_comdats): Update.
989 * cgraph.c (cgraph_get_create_node): Update.
990 * cgraph.h (struct symtab_node): Add get_comdat_group, set_comdat_group
992 (symtab_get_node): Make inline.
993 (symtab_insert_node_to_hashtable): Remove.
994 (symtab_can_be_discarded): Update.
995 (decl_comdat_group): New function.
996 * tree-streamer-in.c (lto_input_ts_decl_with_vis_tree_pointers):
998 * lto-cgraph.c (lto_output_node, lto_output_varpool_node): Stream out
1000 (read_comdat_group): New function.
1001 (input_node, input_varpool_node): Use it.
1002 * trans-mem.c (ipa_tm_create_version_alias): Update code creating
1004 * mips.c (mips_start_unique_function): Likewise.
1005 (ix86_code_end): Likewise.
1006 (rs6000_code_end): Likweise.
1007 * tree-streamer-out.c (DECL_COMDAT_GROUP): Do not stream comdat group.
1009 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
1011 * gengtype-state.c (fatal_reading_state): Bring offline.
1012 * optabs.c (widening_optab_handler): Bring offline.
1013 * optabs.h (widening_optab_handler): Likewise.
1014 * final.c (get_attr_length_1): Likewise.
1016 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
1018 * sched-int.h (sd_iterator_cond): Manually tail recurse.
1020 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1022 * config/rs6000/440.md (ppc440-integer): Include shift without dot.
1023 (ppc440-compare): Include shift with dot.
1024 * config/rs6000/e300c2c3.md (ppce300c3_iu): Include shift without dot.
1025 * config/rs6000/e5500.md (e5500_sfx2): Include constant shift
1027 * config/rs6000/e6500.md (e6500_sfx): Exclude constant shift
1029 (e6500_sfx2): Include it.
1030 * config/rs6000/rs6000.md ( *zero_extend<mode>di2_internal1,
1031 *zero_extend<mode>di2_internal2, *zero_extend<mode>di2_internal3,
1032 *zero_extendsidi2_lfiwzx, andsi3_mc, andsi3_nomc,
1033 andsi3_internal0_nomc, extzvsi_internal, extzvdi_internal,
1034 *extzvdi_internal1, *extzvdi_internal2, rotlsi3, *rotlsi3_64,
1035 *rotlsi3_internal4, *rotlsi3_internal7le, *rotlsi3_internal7be,
1036 *rotlsi3_internal10le, *rotlsi3_internal10be, rlwinm,
1037 *lshiftrt_internal1le, *lshiftrt_internal1be,
1038 *lshiftrt_internal4le, *lshiftrt_internal4be, rotldi3,
1039 *rotldi3_internal4, *rotldi3_internal7le, *rotldi3_internal7be,
1040 *rotldi3_internal10le, *rotldi3_internal10be,
1041 *rotldi3_internal13le, *rotldi3_internal13be, *ashldi3_internal4,
1042 ashldi3_internal5, *ashldi3_internal6, *ashldi3_internal7,
1043 ashldi3_internal8, *ashldi3_internal9, anddi3_mc, anddi3_nomc,
1044 *anddi3_internal2_mc, *anddi3_internal3_mc, and 4 anonymous
1045 define_insns): Use type "shift" in the appropriate alternatives.
1047 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1049 * config/rs6000/rs6000.md (type): Add "logical". Delete
1051 (dot): Adjust comment.
1052 (andsi3_mc, *andsi3_internal2_mc, *andsi3_internal3_mc,
1053 *andsi3_internal4, *andsi3_internal5_mc, *boolsi3_internal2,
1054 *boolsi3_internal3, *boolccsi3_internal2, *boolccsi3_internal3,
1055 anddi3_mc, *anddi3_internal2_mc, *anddi3_internal3_mc,
1056 *booldi3_internal2, *booldi3_internal3, *boolcdi3_internal2,
1057 *boolcdi3_internal3, *boolccdi3_internal2, *boolccdi3_internal3,
1058 *mov<mode>_internal2, and 10 anonymous define_insns): Use "logical".
1059 * config/rs6000/rs6000.c (rs6000_adjust_cost): Adjust.
1061 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
1062 * config/rs6000/440.md (ppc440-integer, ppc440-compare): Adjust.
1063 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare): Adjust.
1064 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
1065 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
1066 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare): Adjust.
1067 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
1068 * config/rs6000/8540.md (ppc8540_su): Adjust.
1069 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
1070 cell-cmp-microcoded): Adjust.
1071 * config/rs6000/e300c2c3.md (ppce300c3_cmp, ppce300c3_iu): Adjust.
1072 * config/rs6000/e500mc.md (e500mc_su): Adjust.
1073 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2): Adjust.
1074 * config/rs6000/e5500.md (e5500_sfx, e5500_sfx2): Adjust.
1075 * config/rs6000/e6500.md (e6500_sfx, e6500_sfx2): Adjust.
1076 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
1077 * config/rs6000/power4.md (power4-integer, power4-cmp): Adjust.
1078 * config/rs6000/power5.md (power5-integer, power5-cmp): Adjust.
1079 * config/rs6000/power6.md (power6-integer, power6-fast-compare):
1081 * config/rs6000/power7.md (power7-integer, power7-cmp): Adjust.
1082 * config/rs6000/power8.md (power8-1cyc, power8-fast-compare):
1083 Adjust. Adjust comment.
1084 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
1085 * config/rs6000/titan.md (titan_fxu_adder, titan_fxu_alu): Adjust.
1087 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1089 * config/rs6000/rs6000.md (type): Add "add".
1090 (*add<mode>3_internal1, addsi3_high, *add<mode>3_internal2,
1091 *add<mode>3_internal3, *neg<mode>2_internal, and 5 anonymous
1092 define_insns): Use it.
1093 * config/rs6000/rs6000.c (rs6000_adjust_cost): Adjust.
1095 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
1096 * config/rs6000/440.md (ppc440-integer, ppc440-compare): Adjust.
1097 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare): Adjust.
1098 * config/rs6000/601.md (ppc601-integer): Adjust.
1099 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
1100 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
1101 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare): Adjust.
1102 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
1103 * config/rs6000/8540.md (ppc8540_su): Adjust.
1104 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
1105 cell-cmp-microcoded): Adjust.
1106 * config/rs6000/e300c2c3.md (ppce300c3_cmp, ppce300c3_iu): Adjust.
1107 * config/rs6000/e500mc.md (e500mc_su): Adjust.
1108 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2): Adjust.
1109 * config/rs6000/e5500.md (e5500_sfx, e5500_sfx2): Adjust.
1110 * config/rs6000/e6500.md (e6500_sfx, e6500_sfx2): Adjust.
1111 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
1112 * config/rs6000/power4.md (power4-integer, power4-cmp): Adjust.
1113 * config/rs6000/power5.md (power5-integer, power5-cmp): Adjust.
1114 * config/rs6000/power6.md (power6-integer, power6-fast-compare):
1116 * config/rs6000/power7.md (power7-integer, power7-cmp): Adjust.
1117 * config/rs6000/power8.md (power8-1cyc, power8-fast-compare): Adjust.
1118 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
1119 * config/rs6000/titan.md (titan_fxu_adder, titan_fxu_alu): Adjust.
1121 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1123 * config/rs6000/rs6000.md (type): Delete "var_shift_rotate",
1124 "delayed_compare", "var_delayed_compare".
1125 (var_shift): New attribute.
1126 (cell_micro): Adjust.
1127 (*andsi3_internal2_mc, *andsi3_internal3_mc, *andsi3_internal4,
1128 *andsi3_internal5_mc, *extzvsi_internal1, *extzvsi_internal2,
1129 rotlsi3, *rotlsi3_64, *rotlsi3_internal2, *rotlsi3_internal3,
1130 *rotlsi3_internal4, *rotlsi3_internal5, *rotlsi3_internal6,
1131 *rotlsi3_internal8le, *rotlsi3_internal8be, *rotlsi3_internal9le,
1132 *rotlsi3_internal9be, *rotlsi3_internal10le, *rotlsi3_internal10be,
1133 *rotlsi3_internal11le, *rotlsi3_internal11be, *rotlsi3_internal12le,
1134 *rotlsi3_internal12be, ashlsi3, *ashlsi3_64, lshrsi3, *lshrsi3_64,
1135 *lshiftrt_internal2le, *lshiftrt_internal2be, *lshiftrt_internal3le,
1136 *lshiftrt_internal3be, *lshiftrt_internal5le, *lshiftrt_internal5be,
1137 *lshiftrt_internal5le, *lshiftrt_internal5be, ashrsi3, *ashrsi3_64,
1138 rotldi3, *rotldi3_internal2, *rotldi3_internal3, *rotldi3_internal4,
1139 *rotldi3_internal5, *rotldi3_internal6, *rotldi3_internal7le,
1140 *rotldi3_internal7be, *rotldi3_internal8le, *rotldi3_internal8be,
1141 *rotldi3_internal9le, *rotldi3_internal9be, *rotldi3_internal10le,
1142 *rotldi3_internal10be, *rotldi3_internal11le, *rotldi3_internal11be,
1143 *rotldi3_internal12le, *rotldi3_internal12be, *rotldi3_internal13le,
1144 *rotldi3_internal13be, *rotldi3_internal14le, *rotldi3_internal14be,
1145 *rotldi3_internal15le, *rotldi3_internal15be, *ashldi3_internal1,
1146 *ashldi3_internal2, *ashldi3_internal3, *lshrdi3_internal1,
1147 *lshrdi3_internal2, *lshrdi3_internal3, *ashrdi3_internal1,
1148 *ashrdi3_internal2, *ashrdi3_internal3, *anddi3_internal2_mc,
1149 *anddi3_internal3_mc, as well as 11 anonymous define_insns): Adjust.
1150 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
1151 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
1153 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
1154 * config/rs6000/440.md (ppc440-integer): Adjust.
1155 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare): Adjust.
1156 * config/rs6000/601.md (ppc601-integer, ppc601-compare): Adjust.
1157 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
1158 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
1159 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare): Adjust.
1160 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
1161 * config/rs6000/8540.md (ppc8540_su): Adjust.
1162 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
1163 cell-cmp-microcoded): Adjust.
1164 * config/rs6000/e300c2c3.md (ppce300c3_cmp): Adjust.
1165 * config/rs6000/e500mc.md (e500mc_su): Adjust.
1166 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2,
1167 e500mc64_delayed): Adjust.
1168 * config/rs6000/e5500.md (e5500_sfx, e5500_delayed): Adjust.
1169 * config/rs6000/e6500.md (e6500_sfx, e6500_delayed): Adjust.
1170 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
1171 * config/rs6000/power4.md (power4-integer, power4-compare): Adjust.
1172 * config/rs6000/power5.md (power5-integer, power5-compare): Adjust.
1173 * config/rs6000/power6.md (power6-shift, power6-var-rotate,
1174 power6-delayed-compare, power6-var-delayed-compare): Adjust.
1175 * config/rs6000/power7.md (power7-integer, power7-compare): Adjust.
1176 * config/rs6000/power8.md (power8-1cyc, power8-compare): Adjust.
1178 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
1179 * config/rs6000/titan.md (titan_fxu_shift_and_rotate): Adjust.
1181 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1183 * config/rs6000/rs6000.md (type): Delete "idiv", "ldiv". Add "div".
1184 (bits): New mode_attr.
1185 (idiv_ldiv): Delete mode_attr.
1186 (udiv<mode>3, *div<mode>3, div<div_extend>_<mode>): Adjust.
1187 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
1188 rs6000_adjust_priority, is_nonpipeline_insn,
1189 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
1191 * config/rs6000/40x.md (ppc403-idiv): Adjust.
1192 * config/rs6000/440.md (ppc440-idiv): Adjust.
1193 * config/rs6000/476.md (ppc476-idiv): Adjust.
1194 * config/rs6000/601.md (ppc601-idiv): Adjust.
1195 * config/rs6000/603.md (ppc603-idiv): Adjust.
1196 * config/rs6000/6xx.md (ppc604-idiv, ppc620-idiv, ppc630-idiv,
1197 ppc620-ldiv): Adjust.
1198 * config/rs6000/7450.md (ppc7450-idiv): Adjust.
1199 * config/rs6000/7xx.md (ppc750-idiv): Adjust.
1200 * config/rs6000/8540.md (ppc8540_divide): Adjust.
1201 * config/rs6000/a2.md (ppca2-idiv, ppca2-ldiv): Adjust.
1202 * config/rs6000/cell.md (cell-idiv, cell-ldiv): Adjust.
1203 * config/rs6000/e300c2c3.md (ppce300c3_divide): Adjust.
1204 * config/rs6000/e500mc.md (e500mc_divide): Adjust.
1205 * config/rs6000/e500mc64.md (e500mc64_divide): Adjust.
1206 * config/rs6000/e5500.md (e5500_divide, e5500_divide_d): Adjust.
1207 * config/rs6000/e6500.md (e6500_divide, e6500_divide_d): Adjust.
1208 * config/rs6000/mpc.md (mpccore-idiv): Adjust.
1209 * config/rs6000/power4.md (power4-idiv, power4-ldiv): Adjust.
1210 * config/rs6000/power5.md (power5-idiv, power5-ldiv): Adjust.
1211 * config/rs6000/power6.md (power6-idiv, power6-ldiv): Adjust.
1212 * config/rs6000/power7.md (power7-idiv, power7-ldiv): Adjust.
1213 * config/rs6000/power8.md (power8-idiv, power8-ldiv): Adjust.
1214 * config/rs6000/rs64.md (rs64a-idiv, rs64a-ldiv): Adjust.
1215 * config/rs6000/titan.md (titan_fxu_div): Adjust.
1217 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1219 * config/rs6000/rs6000.md (type): Delete "insert_word",
1220 "insert_dword". Add "insert".
1221 (size): Update comment.
1222 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
1223 insn_must_be_first_in_group): Adjust.
1224 (insvsi_internal, *insvsi_internal1, *insvsi_internal2,
1225 *insvsi_internal3, *insvsi_internal4, *insvsi_internal5,
1226 *insvsi_internal6, insvdi_internal): Adjust.
1228 * config/rs6000/40x.md (ppc403-integer): Adjust.
1229 * config/rs6000/440.md (ppc440-integer): Adjust.
1230 * config/rs6000/476.md (ppc476-simple-integer): Adjust.
1231 * config/rs6000/601.md (ppc601-integer): Adjust.
1232 * config/rs6000/603.md (ppc603-integer): Adjust.
1233 * config/rs6000/6xx.md (ppc604-integer): Adjust.
1234 * config/rs6000/7450.md (ppc7450-integer): Adjust.
1235 * config/rs6000/7xx.md (ppc750-integer): Adjust.
1236 * config/rs6000/8540.md (ppc8540_su): Adjust.
1237 * config/rs6000/cell.md (cell-integer, cell-insert): Adjust.
1238 * config/rs6000/e300c2c3.md (ppce300c3_iu): Adjust.
1239 * config/rs6000/e500mc.md (e500mc_su): Adjust.
1240 * config/rs6000/e500mc64.md (e500mc64_su): Adjust.
1241 * config/rs6000/e5500.md (e5500_sfx): Adjust.
1242 * config/rs6000/e6500.md (e6500_sfx): Adjust.
1243 * config/rs6000/mpc.md (mpccore-integer): Adjust.
1244 * config/rs6000/power4.md (power4-integer, power4-insert): Adjust.
1245 * config/rs6000/power5.md (power5-integer, power5-insert): Adjust.
1246 * config/rs6000/power6.md (power6-insert, power6-insert-dword): Adjust.
1247 * config/rs6000/power7.md (power7-integer): Adjust.
1248 * config/rs6000/power8.md (power8-1cyc): Adjust.
1249 * config/rs6000/rs64.md (rs64a-integer): Adjust.
1250 * config/rs6000/titan.md (titan_fxu_shift_and_rotate): Adjust.
1252 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1254 * config/rs6000/rs6000.md (type): Add "mul". Delete "imul",
1255 "imul2", "imul3", "lmul", "imul_compare", "lmul_compare".
1256 (size): New attribute.
1257 (dot): New attribute.
1258 (cell_micro): Adjust.
1259 (mulsi3, *mulsi3_internal1, *mulsi3_internal2, mulsidi3,
1260 umulsidi3, smulsi3_highpart, umulsi3_highpart, muldi3,
1261 *muldi3_internal1, *muldi3_internal2, smuldi3_highpart,
1262 umuldi3_highpart): Adjust.
1263 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
1264 rs6000_adjust_priority, is_nonpipeline_insn,
1265 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
1267 * config/rs6000/40x.md (ppc403-imul, ppc405-imul, ppc405-imul2,
1268 ppc405-imul3): Adjust.
1269 * config/rs6000/440.md (ppc440-imul, ppc440-imul2): Adjust.
1270 * config/rs6000/476.md (ppc476-imul): Adjust.
1271 * config/rs6000/601.md (ppc601-imul): Adjust.
1272 * config/rs6000/603.md (ppc603-imul, ppc603-imul2): Adjust.
1273 * config/rs6000/6xx.md (ppc604-imul, ppc604e-imul, ppc620-imul,
1274 ppc620-imul2, ppc620-imul3, ppc620-lmul): Adjust.
1275 * config/rs6000/7450.md (ppc7450-imul, ppc7450-imul2): Adjust.
1276 * config/rs6000/7xx.md (ppc750-imul, ppc750-imul2, ppc750-imul3):
1278 * config/rs6000/8540.md (ppc8540_multiply): Adjust.
1279 * config/rs6000/a2.md (ppca2-imul, ppca2-lmul): Adjust.
1280 * config/rs6000/cell.md (cell-lmul, cell-lmul-cmp, cell-imul23,
1282 * config/rs6000/e300c2c3.md (ppce300c3_multiply): Adjust.
1283 * config/rs6000/e500mc.md (e500mc_multiply): Adjust.
1284 * config/rs6000/e500mc64.md (e500mc64_multiply): Adjust.
1285 * config/rs6000/e5500.md (e5500_multiply, e5500_multiply_i): Adjust.
1286 * config/rs6000/e6500.md (e6500_multiply, e6500_multiply_i): Adjust.
1287 * config/rs6000/mpc.md (mpccore-imul): Adjust.
1288 * config/rs6000/power4.md (power4-lmul-cmp, power4-imul-cmp,
1289 power4-lmul, power4-imul, power4-imul3): Adjust.
1290 * config/rs6000/power5.md (power5-lmul-cmp, power5-imul-cmp,
1291 power5-lmul, power5-imul, power5-imul3): Adjust.
1292 * config/rs6000/power6.md (power6-lmul-cmp, power6-imul-cmp,
1293 power6-lmul, power6-imul, power6-imul3): Adjust.
1294 * config/rs6000/power7.md (power7-mul, power7-mul-compare): Adjust.
1295 * config/rs6000/power8.md (power8-mul, power8-mul-compare): Adjust.
1297 * config/rs6000/rs64.md (rs64a-imul, rs64a-imul2, rs64a-imul3,
1298 rs64a-lmul): Adjust.
1299 * config/rs6000/titan.md (titan_imul): Adjust.
1301 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1303 * config/rs6000/rs6000.md (type): Add new value "halfmul".
1304 (*macchwc, *macchw, *macchwuc, *macchwu, *machhwc, *machhw,
1305 *machhwuc, *machhwu, *maclhwc, *maclhw, *maclhwuc, *maclhwu,
1306 *nmacchwc, *nmacchw, *nmachhwc, *nmachhw, *nmaclhwc, *nmaclhw,
1307 *mulchwc, *mulchw, *mulchwuc, *mulchwu, *mulhhwc, *mulhhw,
1308 *mulhhwuc, *mulhhwu, *mullhwc, *mullhw, *mullhwuc, *mullhwu): Use it.
1309 * config/rs6000/40x.md (ppc405-imul3): Add type halfmul.
1310 * config/rs6000/440.md (ppc440-imul2): Add type halfmul.
1311 * config/rs6000/476.md (ppc476-imul): Add type halfmul.
1312 * config/rs6000/titan.md: Delete nonsensical comment.
1313 (titan_imul): Add type imul3.
1314 (titan_mulhw): Remove type imul3; add type halfmul.
1316 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1318 * config/rs6000/rs6000.md (type): Reorder, reformat.
1320 2014-05-23 Martin Jambor <mjambor@suse.cz>
1322 PR tree-optimization/53787
1323 * params.def (PARAM_IPA_MAX_AA_STEPS): New param.
1324 * ipa-prop.h (ipa_node_params): Rename uses_analysis_done to
1325 analysis_done, update all uses.
1326 * ipa-prop.c: Include domwalk.h
1327 (param_analysis_info): Removed.
1328 (param_aa_status): New type.
1329 (ipa_bb_info): Likewise.
1330 (func_body_info): Likewise.
1331 (ipa_get_bb_info): New function.
1332 (aa_overwalked): Likewise.
1333 (find_dominating_aa_status): Likewise.
1334 (parm_bb_aa_status_for_bb): Likewise.
1335 (parm_preserved_before_stmt_p): Changed to use new param AA info.
1336 (load_from_unmodified_param): Accept func_body_info as a parameter
1337 instead of parms_ainfo.
1338 (parm_ref_data_preserved_p): Changed to use new param AA info.
1339 (parm_ref_data_pass_through_p): Likewise.
1340 (ipa_load_from_parm_agg_1): Likewise. Update callers.
1341 (compute_complex_assign_jump_func): Changed to use new param AA info.
1342 (compute_complex_ancestor_jump_func): Likewise.
1343 (ipa_compute_jump_functions_for_edge): Likewise.
1344 (ipa_compute_jump_functions): Removed.
1345 (ipa_compute_jump_functions_for_bb): New function.
1346 (ipa_analyze_indirect_call_uses): Likewise, moved variable
1348 (ipa_analyze_virtual_call_uses): Accept func_body_info instead of node
1349 and info, moved variable declarations down.
1350 (ipa_analyze_call_uses): Accept and pass on func_body_info instead of
1352 (ipa_analyze_stmt_uses): Likewise.
1353 (ipa_analyze_params_uses): Removed.
1354 (ipa_analyze_params_uses_in_bb): New function.
1355 (ipa_analyze_controlled_uses): Likewise.
1356 (free_ipa_bb_info): Likewise.
1357 (analysis_dom_walker): New class.
1358 (ipa_analyze_node): Handle node-specific forbidden analysis,
1359 initialize and free func_body_info, use dominator walker.
1360 (ipcp_modif_dom_walker): New class.
1361 (ipcp_transform_function): Create and free func_body_info, use
1362 ipcp_modif_dom_walker, moved a lot of functionality there.
1364 2014-05-23 Marek Polacek <polacek@redhat.com>
1365 Jakub Jelinek <jakub@redhat.com>
1367 * builtins.def: Change SANITIZE_FLOAT_DIVIDE to SANITIZE_NONDEFAULT.
1368 * gcc.c (sanitize_spec_function): Likewise.
1369 * convert.c (convert_to_integer): Include "ubsan.h". Add
1370 floating-point to integer instrumentation.
1371 * doc/invoke.texi: Document -fsanitize=float-cast-overflow.
1372 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_CAST and
1373 SANITIZE_NONDEFAULT.
1374 * opts.c (common_handle_option): Handle -fsanitize=float-cast-overflow.
1375 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_FLOAT_CAST_OVERFLOW,
1376 BUILT_IN_UBSAN_HANDLE_FLOAT_CAST_OVERFLOW_ABORT): Add.
1377 * ubsan.c: Include "realmpfr.h" and "dfp.h".
1378 (get_ubsan_type_info_for_type): Handle REAL_TYPEs.
1379 (ubsan_type_descriptor): Set tkind to 0xffff for types other than
1380 float/double/long double.
1381 (ubsan_instrument_float_cast): New function.
1382 * ubsan.h (ubsan_instrument_float_cast): Declare.
1384 2014-05-23 Jiong Wang <jiong.wang@arm.com>
1386 * config/aarch64/predicates.md (aarch64_call_insn_operand): New
1388 * config/aarch64/constraints.md ("Ucs", "Usf"): New constraints.
1389 * config/aarch64/aarch64.md (*sibcall_insn, *sibcall_value_insn):
1390 Adjust for tailcalling through registers.
1391 * config/aarch64/aarch64.h (enum reg_class): New caller save
1393 (REG_CLASS_NAMES): Likewise.
1394 (REG_CLASS_CONTENTS): Likewise.
1395 * config/aarch64/aarch64.c (aarch64_function_ok_for_sibcall):
1396 Allow tailcalling without decls.
1398 2014-05-23 Thomas Schwinge <thomas@codesourcery.com>
1400 * gimplify.c (omp_notice_variable) <case OMP_CLAUSE_DEFAULT_NONE>:
1401 Rewrite check for ORT_PARALLEL and ORT_COMBINED_PARALLEL.
1403 * omp-low.c (expand_omp_for_static_chunk): Rename variable si to
1404 gsi, and variables v_* to v*.
1406 2014-05-23 Eric Botcazou <ebotcazou@adacore.com>
1408 * varasm.c (output_constructor_bitfield): Fix thinkos in latest change.
1410 2014-05-23 Thomas Schwinge <thomas@codesourcery.com>
1412 * gimple.h (enum gf_mask): Add and use GF_OMP_FOR_SIMD.
1413 * omp-low.c: Update accordingly.
1415 * gimple.h (enum gf_mask): Rewrite "<< 0" shift expressions used
1416 for GF_OMP_FOR_KIND_MASK, GF_OMP_FOR_KIND_FOR,
1417 GF_OMP_FOR_KIND_DISTRIBUTE, GF_OMP_FOR_KIND_SIMD,
1418 GF_OMP_FOR_KIND_CILKSIMD, GF_OMP_TARGET_KIND_MASK,
1419 GF_OMP_TARGET_KIND_REGION, GF_OMP_TARGET_KIND_DATA,
1420 GF_OMP_TARGET_KIND_UPDATE.
1422 * gimplify.c (omp_notice_variable) <case OMP_CLAUSE_DEFAULT_NONE>:
1423 Explicitly enumerate the expected region types.
1425 2014-05-23 Paul Eggert <eggert@cs.ucla.edu>
1428 * doc/extend.texi (Function Attributes): Fix __attribute__ ((malloc))
1429 documentation; the old documentation didn't clearly state the
1430 constraints on the contents of the pointed-to storage.
1432 2014-05-23 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
1434 Fix bootstrap error on ia64
1435 * config/ia64/ia64.c (ia64_first_cycle_multipass_dfa_lookahead_guard):
1436 Return default value.
1438 2014-05-23 Thomas Preud'homme <thomas.preudhomme@arm.com>
1440 PR tree-optimization/54733
1441 * tree-ssa-math-opts.c (nop_stats): New "bswap_stats" structure.
1443 (find_bswap_or_nop_load): New.
1444 (find_bswap_1): Renamed to ...
1445 (find_bswap_or_nop_1): This. Also add support for memory source.
1446 (find_bswap): Renamed to ...
1447 (find_bswap_or_nop): This. Also add support for memory source and
1448 detection of bitwise operations equivalent to load in host endianness.
1449 (execute_optimize_bswap): Likewise. Also move its leading comment back
1450 in place and split statement transformation into ...
1451 (bswap_replace): This.
1453 2014-05-22 Vladimir Makarov <vmakarov@redhat.com>
1455 PR rtl-optimization/61215
1456 * lra-elelimination.c (lra_eliminate_regs_1): Don't use
1457 simplify_gen_subreg until final substitution.
1459 2014-05-23 Alan Modra <amodra@gmail.com>
1462 * config/rs6000/rs6000.c (mem_operand_gpr): Handle SImode.
1463 * config/rs6000/rs6000.md (extendsidi2_lfiwax, extendsidi2_nocell):
1464 Use "Y" constraint rather than "m".
1466 2014-05-23 Kugan Vivekanandarajah <kuganv@linaro.org>
1468 * config/aarch64/aarch64.c (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New
1470 * config/aarch64/aarch64-protos.h (aarch64_atomic_assign_expand_fenv):
1471 New function declaration.
1472 * config/aarch64/aarch64-builtins.c (aarch64_builtins) : Add
1473 AARCH64_BUILTIN_GET_FPCR, AARCH64_BUILTIN_SET_FPCR.
1474 AARCH64_BUILTIN_GET_FPSR and AARCH64_BUILTIN_SET_FPSR.
1475 (aarch64_init_builtins) : Initialize builtins
1476 __builtins_aarch64_set_fpcr, __builtins_aarch64_get_fpcr.
1477 __builtins_aarch64_set_fpsr and __builtins_aarch64_get_fpsr.
1478 (aarch64_expand_builtin) : Expand builtins __builtins_aarch64_set_fpcr
1479 __builtins_aarch64_get_fpcr, __builtins_aarch64_get_fpsr,
1480 and __builtins_aarch64_set_fpsr.
1481 (aarch64_atomic_assign_expand_fenv): New function.
1482 * config/aarch64/aarch64.md (set_fpcr): New pattern.
1483 (get_fpcr) : Likewise.
1484 (set_fpsr) : Likewise.
1485 (get_fpsr) : Likewise.
1486 (unspecv): Add UNSPECV_GET_FPCR and UNSPECV_SET_FPCR, UNSPECV_GET_FPSR
1487 and UNSPECV_SET_FPSR.
1488 * doc/extend.texi (AARCH64 Built-in Functions) : Document
1489 __builtins_aarch64_set_fpcr, __builtins_aarch64_get_fpcr.
1490 __builtins_aarch64_set_fpsr and __builtins_aarch64_get_fpsr.
1492 2014-05-22 Vladimir Makarov <vmakarov@redhat.com>
1494 PR rtl-optimization/60969
1495 * ira-costs.c (record_reg_classes): Process NO_REGS for matching
1496 constraints. Set up mem cost for NO_REGS case.
1498 2014-05-22 Thomas Schwinge <thomas@codesourcery.com>
1500 * builtin-types.def: Simplify examples for DEF_FUNCTION_TYPE_*.
1502 2012-05-22 Bernd Schmidt <bernds@codesourcery.com>
1504 * config/darwin.c: Include "lto-section-names.h".
1505 (LTO_SEGMENT_NAME): Don't define.
1506 * config/i386/winnt.c: Include "lto-section-names.h".
1507 * lto-streamer.c: Include "lto-section-names.h".
1508 * lto-streamer.h (LTO_SECTION_NAME_PREFIX): Don't define.
1509 * lto-wrapper.c: Include "lto-section-names.h".
1510 (LTO_SECTION_NAME_PREFIX): Don't define.
1511 * lto-section-names.h: New file.
1512 * cgraphunit.c: Include "lto-section-names.h".
1514 2014-05-22 Peter Bergner <bergner@vnet.ibm.com>
1516 * config/rs6000/htm.md (ttest): Use correct shift value to get CR0.
1518 2014-05-22 Richard Earnshaw <rearnsha@arm.com>
1521 * arm.md (arm_cmpdi_unsigned): Fix length calculation for Thumb2.
1523 2014-05-22 Nick Clifton <nickc@redhat.com>
1525 * config/msp430/msp430.h (ASM_SPEC): Add spaces after inserted options.
1527 2014-05-22 Eric Botcazou <ebotcazou@adacore.com>
1529 * tree-ssa-forwprop.c (associate_plusminus): Extend (T)(P + A) - (T)P
1530 -> (T)A transformation to integer types.
1532 2014-05-22 Teresa Johnson <tejohnson@google.com>
1534 * gcov-io.c (gcov_position): Use gcov_nonruntime_assert.
1535 (gcov_is_error): Remove gcc_assert from IN_LIBGCOV code.
1536 (gcov_rewrite): Use gcov_nonruntime_assert.
1538 (gcov_write_words): Ditto.
1539 (gcov_write_length): Ditto.
1540 (gcov_read_words): Use gcov_nonruntime_assert, and remove
1541 gcc_assert from IN_LIBGCOV code.
1542 (gcov_read_summary): Use gcov_error to flag profile corruption.
1543 (gcov_sync): Use gcov_nonruntime_assert.
1544 (gcov_seek): Remove gcc_assert from IN_LIBGCOV code.
1545 (gcov_histo_index): Use gcov_nonruntime_assert.
1546 (static void gcov_histogram_merge): Ditto.
1547 (compute_working_sets): Ditto.
1548 * gcov-io.h (gcov_nonruntime_assert): Define.
1549 (gcov_error): Define for !IN_LIBGCOV
1551 2014-05-22 Richard Biener <rguenther@suse.de>
1553 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
1554 BUILT_IN_REALLOC like BUILT_IN_STRDUP.
1555 (call_may_clobber_ref_p_1): Handle BUILT_IN_REALLOC as allocation
1556 and deallocation site.
1557 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
1558 Handle BUILT_IN_REALLOC similar to BUILT_IN_STRDUP with also
1559 passing through the incoming points-to set.
1560 (handle_lhs_call): Use flags argument instead of recomputing it.
1561 (find_func_aliases_for_call): Call handle_lhs_call with proper
1564 2014-05-22 Jakub Jelinek <jakub@redhat.com>
1566 * tree-streamer-in.c (unpack_ts_real_cst_value_fields): Make sure
1567 all padding bits in REAL_VALUE_TYPE are cleared.
1569 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
1571 Cleanup and improve multipass_dfa_lookahead_guard
1572 * config/i386/i386.c (core2i7_first_cycle_multipass_filter_ready_try,)
1573 (core2i7_first_cycle_multipass_begin,)
1574 (core2i7_first_cycle_multipass_issue,)
1575 (core2i7_first_cycle_multipass_backtrack): Update signature.
1576 * config/ia64/ia64.c
1577 (ia64_first_cycle_multipass_dfa_lookahead_guard_spec): Remove.
1578 (ia64_first_cycle_multipass_dfa_lookahead_guard): Update signature.
1579 (TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC): Remove
1581 (ia64_first_cycle_multipass_dfa_lookahead_guard): Merge logic from
1582 ia64_first_cycle_multipass_dfa_lookahead_guard_spec. Update return
1584 * config/rs6000/rs6000.c (rs6000_use_sched_lookahead_guard): Update
1586 * doc/tm.texi: Regenerate.
1588 (TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC): Remove.
1589 * haifa-sched.c (ready_try): Make signed to allow negative values.
1590 (rebug_ready_list_1): Update.
1591 (choose_ready): Simplify.
1592 (sched_extend_ready_list): Update.
1594 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
1596 Remove IA64 speculation tweaking flags
1597 * config/ia64/ia64.c (ia64_set_sched_flags): Delete handling of
1598 speculation tuning flags.
1599 (msched-prefer-non-data-spec-insns,)
1600 (msched-prefer-non-control-spec-insns): Obsolete options.
1601 * haifa-sched.c (choose_ready): Remove handling of
1602 PREFER_NON_CONTROL_SPEC and PREFER_NON_DATA_SPEC.
1603 * sched-int.h (enum SPEC_SCHED_FLAGS): Remove PREFER_NON_CONTROL_SPEC
1604 and PREFER_NON_DATA_SPEC.
1605 * sel-sched.c (process_spec_exprs): Remove handling of
1606 PREFER_NON_CONTROL_SPEC and PREFER_NON_DATA_SPEC.
1608 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
1610 Improve scheduling debug output
1611 * haifa-sched.c (debug_ready_list): Remove unnecessary prototype.
1612 (advance_one_cycle): Update.
1613 (schedule_insn, queue_to_ready): Add debug printouts.
1614 (debug_ready_list_1): New static function.
1615 (debug_ready_list): Update.
1616 (max_issue): Add debug printouts.
1617 (dump_insn_stream): New static function.
1618 (schedule_block): Use it. Also better indent printouts.
1620 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
1622 Fix sched_insn debug counter
1623 * haifa-sched.c (schedule_insn): Update.
1624 (struct haifa_saved_data): Add nonscheduled_insns_begin.
1625 (save_backtrack_point, restore_backtrack_point): Update.
1626 (first_nonscheduled_insn): New static function.
1627 (queue_to_ready, choose_ready): Use it.
1628 (schedule_block): Init nonscheduled_insns_begin.
1629 (sched_emit_insn): Update.
1632 2014-05-22 Kugan Vivekanandarajah <kuganv@linaro.org>
1634 * config/aarch64/aarch64.c (aarch64_regno_regclass) : Change CORE_REGS
1636 (aarch64_secondary_reload) : LikeWise.
1637 (aarch64_class_max_nregs) : Remove CORE_REGS.
1638 * config/aarch64/aarch64.h (enum reg_class) : Remove CORE_REGS.
1639 (REG_CLASS_NAMES) : Likewise.
1640 (REG_CLASS_CONTENTS) : LikeWise.
1641 (INDEX_REG_CLASS) : Change CORE_REGS to GENERAL_REGS.
1643 2014-05-21 Guozhi Wei <carrot@google.com>
1646 * config/aarch64/arm_neon.h (vqdmulh_n_s16): Change the last operand's
1648 (vqdmulhq_n_s16): Likewise.
1650 2014-05-21 Segher Boessenkool <segher@kernel.crashing.org>
1652 * config/rs6000/predicates.md (update_indexed_address_mem): Delete.
1654 2014-05-21 Marek Polacek <polacek@redhat.com>
1657 * ubsan.c (is_ubsan_builtin_p): Turn assert into a condition.
1659 2014-05-21 Martin Jambor <mjambor@suse.cz>
1661 * doc/invoke.texi (Optimize Options): Document parameters
1662 ipa-cp-eval-threshold, ipa-max-agg-items, ipa-cp-loop-hint-bonus and
1663 ipa-cp-array-index-hint-bonus.
1665 2014-05-21 Mark Wielaard <mjw@redhat.com>
1668 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_type if DWARF
1669 version >= 3 or not strict DWARF.
1670 * langhooks.h (struct lang_hooks_for_types): Add
1671 enum_underlying_base_type.
1672 * langhooks.c (lhd_enum_underlying_base_type): New function.
1673 * gcc/langhooks.h (struct lang_hooks_for_types): Add
1674 enum_underlying_base_type.
1675 * langhooks-def.h (lhd_enum_underlying_base_type): New declaration.
1676 (LANG_HOOKS_ENUM_UNDERLYING_BASE_TYPE): New define.
1677 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add new lang hook.
1679 2014-05-21 Richard Biener <rguenther@suse.de>
1681 * doc/invoke.texi (-flto-partition=): Document one and none algorithms.
1683 2014-05-21 John Marino <gnugcc@marino.st>
1685 * config.gcc (*-*-dragonfly*): New target.
1686 * configure.ac: Detect dl_iterate_phdr (*freebsd*, *dragonfly*).
1687 * configure: Regenerate.
1688 * config/dragonfly-stdint.h: New.
1689 * config/dragonfly.h: New.
1690 * config/dragonfly.opt: New.
1691 * config/i386/dragonfly.h: New.
1692 * ginclude/stddef.h: Detect _PTRDIFF_T_DECLARED for DragonFly.
1694 2014-05-21 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1696 * tree.def (VOID_CST): New.
1697 * tree-core.h (TI_VOID): New.
1698 * tree.h (void_node): New.
1699 * tree.c (tree_node_structure_for_code, tree_code_size)
1700 (iterative_hash_expr): Handle VOID_CST.
1701 (build_common_tree_nodes): Initialize void_node.
1703 2014-05-21 Bernd Schmidt <bernds@codesourcery.com>
1705 * reload1.c (remove_init_insns, will_delete_init_insn_p): New static
1707 (reload, calculate_needs_all_insns, reload_as_needed): Use them.
1709 * config/bfin/bfin.c (split_load_immediate): Use gen_int_mode in a few
1712 * cfgrtl.c (cfg_layout_initialize): Weaken assert to only trigger if
1713 flag_reorder_blocks_and_partition.
1714 * hw-doloop.c (reorg_loops): Avoid reordering if that flag is set.
1716 2014-05-21 Oleg Endo <olegendo@gcc.gnu.org>
1719 * config/sh/sh.md (*addc_r_1): Rename to addc_t_r. Remove empty
1721 (*addc_r_t): Add new insn_and_split.
1723 2014-05-21 Jakub Jelinek <jakub@redhat.com>
1726 * omp-low.c (handle_simd_reference): New function.
1727 (lower_rec_input_clauses): Use it. Defer adding reference
1728 initialization even for reduction without placeholder if in simd,
1729 handle it properly later on.
1731 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
1733 PR tree-optimization/60899
1734 * gimple-fold.c (can_refer_decl_in_current_unit_p): Cleanup;
1735 assume all static symbols will have definition wile parsing and
1736 check the do have definition later in compilation; check that
1737 variable referring symbol will be output before concluding that
1738 reference is safe; be conservative for referring local statics;
1739 be more precise about when comdat is output in other partition.
1741 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
1744 * ipa-inline-transform.c (inline_call): Use add CALLEE_REMOVED
1746 * ipa-inline.c (inline_to_all_callers): If callee was removed; return.
1747 (ipa_inline): Loop inline_to_all_callers until no more aliases
1750 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
1752 * ipa.c (ipa_discover_readonly_nonaddressable_var): Fix dumping;
1753 set writeonly flag only for vars actually written to.
1755 2014-05-20 Dehao Chen <dehao@google.com>
1757 * ipa-inline-transform.c (clone_inlined_nodes): Use min of edge count
1758 and callee count to get clone count.
1759 * tree-inline.c (expand_call_inline): Use callee count instead of bb
1762 2014-05-20 Richard Sandiford <rdsandiford@googlemail.com>
1764 PR rtl-optimization/61243
1765 * emit-rtl.c (emit_copy_of_insn_after): Copy CROSSING_JUMP_P.
1767 2014-05-20 Xinliang David Li <davidxl@google.com>
1769 * cgraphunit.c (walk_polymorphic_call_targets): Add
1770 dbgcnt and fopt-info support.
1771 * ipa-prop.c (ipa_make_edge_direct_to_target): Ditto.
1772 * ipa-devirt.c (ipa_devirt): Ditto.
1773 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children): Ditto.
1774 * ipa.c (walk_polymorphic_call_targets): Ditto.
1775 * gimple-fold.c (fold_gimple_assign): Ditto.
1776 (gimple_fold_call): Ditto.
1777 * dbgcnt.def: New counter.
1779 2014-05-20 DJ Delorie <dj@redhat.com>
1781 * config/msp430/msp430.md (split): Don't allow subregs when
1782 splitting SImode adds.
1783 (andneghi): Fix subtraction logic.
1784 * config/msp430/predicates.md (msp430_nonsubreg_or_imm_operand): New.
1786 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
1788 * tree.h (DECL_ONE_ONLY): Return true only for externally visible
1790 * except.c (switch_to_exception_section, resolve_unique_section,
1791 get_named_text_section, default_function_rodata_section,
1792 align_variable, get_block_for_decl, default_section_type_flags):
1793 Use DECL_COMDAT_GROUP instead of DECL_ONE_ONLY.
1794 * symtab.c (symtab_add_to_same_comdat_group,
1795 symtab_make_decl_local, fixup_same_cpp_alias_visibility,
1796 symtab_nonoverwritable_alias, symtab_get_symbol_partitioning_class):
1798 * cgraphclones.c (cgraph_create_virtual_clone): Likewise.
1799 * bb-reorder.c (pass_partition_blocks::gate): Likewise.
1800 * config/c6x/c6x.c (c6x_elf_unique_section): Likewise.
1801 (c6x_function_in_section_p): Likewise.
1802 * config/darwin.c (machopic_select_section): Likewise.
1803 * config/arm/arm.c (arm_function_in_section_p): Likewise.
1804 * config/mips/mips.c (mips_function_rodata_section): Likewise.
1805 * config/mep/mep.c (mep_select_section): LIkewise.
1806 * config/i386/i386.c (x86_64_elf_unique_section): Likewise.
1808 2014-05-20 Eric Botcazou <ebotcazou@adacore.com>
1810 * tree-ssa-dom.c (hashable_expr_equal_p) <EXPR_CALL>: Also compare the
1811 EH region of calls to pure functions that can throw an exception.
1812 * tree-ssa-sccvn.c (vn_reference_eq): Remove duplicated test.
1813 (copy_reference_ops_from_call): Also copy the EH region of the call if
1814 it can throw an exception.
1816 2014-05-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1818 * simplify-rtx.c (simplify_binary_operation_1): Optimize case of
1819 nested VEC_SELECTs that are inverses of each other.
1821 2014-05-20 Richard Biener <rguenther@suse.de>
1823 * tree-ssa-sccvn.c (process_scc): Dump SCC here, when iterating,
1824 (extract_and_process_scc_for_name): not here.
1825 (cond_dom_walker::before_dom_children): Only process
1826 stmts that end the BB in interesting ways.
1827 (run_scc_vn): Mark param uses as visited.
1829 2014-05-20 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1831 * config/arm/arm.md (arith_shiftsi): Do not predicate for
1834 2014-05-20 Nick Clifton <nickc@redhat.com>
1836 * config/msp430/msp430.c (TARGET_GIMPLIFY_VA_ARG_EXPR): Define.
1837 (msp430_gimplify_va_arg_expr): New function.
1838 (msp430_print_operand): Handle (CONST (ZERO_EXTRACT)).
1840 * config/msp430/msp430.md (zero_extendpsisi2): Use + constraint on
1841 operand 0 in order to prevent confusion about the number of
1844 2014-05-20 Richard Biener <rguenther@suse.de>
1846 PR tree-optimization/61221
1847 * tree-ssa-pre.c (el_to_update): Remove.
1848 (eliminate_dom_walker::before_dom_children): Handle released
1849 VDEFs by value-numbering them to the associated VUSE. Update
1850 stmt immediately for substituted call address.
1851 (eliminate): Remove delayed stmt updating code.
1852 * tree-ssa-sccvn.c (vuse_ssa_val): New function valueizing
1853 possibly late re-numbered vuses.
1854 (vn_reference_lookup_2): Adjust.
1855 (vn_reference_lookup_pieces): Likewise.
1856 (vn_reference_lookup): Likewise.
1858 2014-05-20 Richard Biener <rguenther@suse.de>
1860 * config.gcc: Remove need_64bit_hwint.
1861 * configure.ac: Do not define NEED_64BIT_HOST_WIDE_INT.
1862 * hwint.h: Do not check NEED_64BIT_HOST_WIDE_INT but assume
1864 * config.in: Regenerate.
1865 * configure: Likewise.
1867 2014-05-19 David Wohlferd <dw@LimeGreenSocks.com>
1869 * doc/extend.texi: Create Label Attributes section,
1870 move all label attributes into it and reference it.
1872 2014-05-19 Richard Earnshaw <rearnsha@arm.com>
1874 * arm.c (thumb1_reorg): When scanning backwards skip anything
1875 that's not a proper insn.
1877 2014-05-19 Richard Biener <rguenther@suse.de>
1879 PR tree-optimization/61221
1880 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children):
1881 Do nothing for unreachable blocks.
1882 * tree-ssa-sccvn.c (cond_dom_walker::before_dom_children):
1883 Improve unreachability detection.
1885 2014-05-19 Richard Biener <rguenther@suse.de>
1887 PR tree-optimization/61209
1888 * tree-ssa-sccvn.c (visit_phi): Avoid setting expr to VN_TOP.
1890 2014-05-19 Nick Clifton <nickc@redhat.com>
1892 * except.c (init_eh): Fix computation of builtin setjmp buffer
1893 size to allow for targets where POINTER_SIZE > BITS_PER_WORD.
1895 2014-05-19 Richard Biener <rguenther@suse.de>
1897 PR tree-optimization/61184
1898 * tree-vrp.c (is_negative_overflow_infinity): Use
1899 TREE_OVERFLOW_P and do that check first.
1900 (is_positive_overflow_infinity): Likewise.
1901 (is_overflow_infinity): Likewise.
1902 (vrp_operand_equal_p): Properly treat operands with
1903 differing overflow as not equal.
1905 2014-05-19 Bernd Schmidt <bernds@codesourcery.com>
1907 * simplify-rtx.c (simplify_unary_operation_1): Use CONST_INT_P in
1908 shift simplification where it was intended.
1910 2014-05-19 Christian Bruel <christian.bruel@st.com>
1913 * config/sh/sh.md (movsf_ie): Unset fp_mode for fmov.
1915 2014-05-19 Richard Sandiford <r.sandiford@uk.ibm.com>
1918 * config/sparc/sparc.c (sparc_fold_builtin): Use widest_int rather
1921 2014-05-19 Richard Sandiford <rdsandiford@googlemail.com>
1923 * reg-notes.def (CROSSING_JUMP): Likewise.
1924 * rtl.h (rtx_def): Update comment for jump flag.
1925 (CROSSING_JUMP_P): Define.
1926 * cfgcleanup.c (try_forward_edges, try_optimize_cfg): Use it instead
1927 of a REG_CROSSING_JUMP note.
1928 * cfghooks.c (tidy_fallthru_edges): Likewise.
1929 * cfgrtl.c (fixup_partition_crossing, rtl_verify_edges): Likewise.
1930 * emit-rtl.c (try_split): Likewise.
1931 * haifa-sched.c (sched_create_recovery_edges): Likewise.
1932 * ifcvt.c (find_if_case_1, find_if_case_2): Likewise.
1933 * jump.c (redirect_jump_2): Likewise.
1934 * reorg.c (follow_jumps, fill_slots_from_thread): Likewise.
1935 (relax_delay_slots): Likewise.
1936 * config/arc/arc.md (jump_i, cbranchsi4_scratch, *bbit): Likewise.
1937 (bbit_di): Likewise.
1938 * config/arc/arc.c (arc_reorg, arc_can_follow_jump): Likewise.
1939 * config/sh/sh.md (jump_compact): Likewise.
1940 * bb-reorder.c (rotate_loop): Likewise.
1941 (pass_duplicate_computed_gotos::execute): Likewise.
1942 (add_reg_crossing_jump_notes): Rename to...
1943 (update_crossing_jump_flags): ...this.
1944 (pass_partition_blocks::execute): Update accordingly.
1946 2014-05-19 Richard Sandiford <rdsandiford@googlemail.com>
1948 * tree.h: Remove extraneous template <>.
1950 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
1952 * ipa.c (symtab_remove_unreachable_nodes): Remove
1953 symbol from comdat group if its body was eliminated.
1954 (comdat_can_be_unshared_p_1): Static symbols can always be privatized.
1955 * symtab.c (symtab_remove_from_same_comdat_group): Break out from ...
1956 (symtab_unregister_node): ... this one.
1957 (verify_symtab_base): More strict checking of comdats.
1958 * cgraph.h (symtab_remove_from_same_comdat_group): Declare.
1960 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
1962 * tree-pass.h (make_pass_ipa_comdats): New pass.
1963 * timevar.def (TV_IPA_COMDATS): New timevar.
1964 * passes.def (pass_ipa_comdats): Add.
1965 * Makefile.in (OBJS): Add ipa-comdats.o
1966 * ipa-comdats.c: New file.
1968 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
1970 * ipa.c (update_visibility_by_resolution_info): New function.
1971 (function_and_variable_visibility): Use it.
1973 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
1975 * cgraph.h (symtab_first_defined_symbol, symtab_next_defined_symbol):
1977 (FOR_EACH_DEFINED_SYMBOL): New macro.
1978 (varpool_first_static_initializer, varpool_next_static_initializer,
1979 varpool_first_defined_variable, varpool_next_defined_variable):
1981 (symtab_in_same_comdat_p): Correctly deal with inline functions.
1983 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1985 * ggc-page.c (ggc_handle_finalizers): Add comment.
1987 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1989 * ggc-common.c (ggc_internal_cleared_alloc): Adjust.
1990 * ggc-none.c (ggc_internal_alloc): Assert if a finalizer is passed.
1991 (ggc_internal_cleared_alloc): Likewise.
1992 * ggc-page.c (finalizer): New class.
1993 (vec_finalizer): Likewise.
1994 (globals::finalizers): New member.
1995 (globals::vec_finalizers): Likewise.
1996 (ggc_internal_alloc): Record the finalizer if any for the block being
1998 (ggc_handle_finalizers): New function.
1999 (ggc_collect): Call ggc_handle_finalizers.
2000 * ggc.h (ggc_internal_alloc): Add arguments to allow installing a
2002 (ggc_internal_cleared_alloc): Likewise.
2003 (finalize): New function.
2004 (need_finalization_p): Likewise.
2005 (ggc_alloc): Install the type's destructor as the finalizer if it
2007 (ggc_cleared_alloc): Likewise.
2008 (ggc_vec_alloc): Likewise.
2009 (ggc_cleared_vec_alloc): Likewise.
2011 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
2013 * ggc.h (ggc_alloc_cleared_simd_clone_stat): Remove function.
2015 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
2017 * alias.c (record_alias_subset): Adjust.
2018 * bitmap.c (bitmap_element_allocate): Likewise.
2019 (bitmap_gc_alloc_stat): Likewise.
2020 * cfg.c (init_flow): Likewise.
2021 (alloc_block): Likewise.
2022 (unchecked_make_edge): Likewise.
2023 * cfgloop.c (alloc_loop): Likewise.
2024 (flow_loops_find): Likewise.
2025 (rescan_loop_exit): Likewise.
2026 * cfgrtl.c (init_rtl_bb_info): Likewise.
2027 * cgraph.c (insert_new_cgraph_node_version): Likewise.
2028 (cgraph_allocate_node): Likewise.
2029 (cgraph_create_edge_1): Likewise.
2030 (cgraph_allocate_init_indirect_info): Likewise.
2031 * cgraphclones.c (cgraph_clone_edge): Likewise.
2032 * cgraphunit.c (add_asm_node): Likewise.
2033 (init_lowered_empty_function): Likewise.
2034 * config/aarch64/aarch64.c (aarch64_init_machine_status): Likewise.
2035 * config/alpha/alpha.c (alpha_init_machine_status): Likewise.
2036 (alpha_use_linkage): Likewise.
2037 * config/arc/arc.c (arc_init_machine_status): Likewise.
2038 * config/arm/arm.c (arm_init_machine_status): Likewise.
2039 * config/avr/avr.c (avr_init_machine_status): Likewise.
2040 * config/bfin/bfin.c (bfin_init_machine_status): Likewise.
2041 * config/c6x/c6x.c (c6x_init_machine_status): Likewise.
2042 * config/cris/cris.c (cris_init_machine_status): Likewise.
2043 * config/darwin.c (machopic_indirection_name): Likewise.
2044 (darwin_build_constant_cfstring): Likewise.
2045 (darwin_enter_string_into_cfstring_table): Likewise.
2046 * config/epiphany/epiphany.c (epiphany_init_machine_status): Likewise.
2047 * config/frv/frv.c (frv_init_machine_status): Likewise.
2048 * config/i386/i386.c (get_dllimport_decl): Likewise.
2049 (ix86_init_machine_status): Likewise.
2050 (assign_386_stack_local): Likewise.
2051 * config/i386/winnt.c (i386_pe_record_external_function): Likewise.
2052 (i386_pe_maybe_record_exported_symbol): Likewise.
2053 (i386_pe_record_stub): Likewise.
2054 * config/ia64/ia64.c (ia64_init_machine_status): Likewise.
2055 * config/iq2000/iq2000.c (iq2000_init_machine_status): Likewise.
2056 * config/m32c/m32c.c (m32c_init_machine_status): Likewise.
2057 (m32c_note_pragma_address): Likewise.
2058 * config/mep/mep.c (mep_init_machine_status): Likewise.
2059 (mep_note_pragma_flag): Likewise.
2060 * config/mips/mips.c (mflip_mips16_use_mips16_p): Likewise.
2061 (mips16_local_alias): Likewise.
2062 (mips_init_machine_status): Likewise.
2063 * config/mmix/mmix.c (mmix_init_machine_status): Likewise.
2064 * config/moxie/moxie.c (moxie_init_machine_status): Likewise.
2065 * config/msp430/msp430.c (msp430_init_machine_status): Likewise.
2066 * config/nds32/nds32.c (nds32_init_machine_status): Likewise.
2067 * config/nios2/nios2.c (nios2_init_machine_status): Likewise.
2068 * config/pa/pa.c (pa_init_machine_status): Likewise.
2069 (pa_get_deferred_plabel): Likewise.
2070 * config/rl78/rl78.c (rl78_init_machine_status): Likewise.
2071 * config/rs6000/rs6000.c (builtin_function_type): Likewise.
2072 (rs6000_init_machine_status): Likewise.
2073 (output_toc): Likewise.
2074 * config/s390/s390.c (s390_init_machine_status): Likewise.
2075 * config/score/score.c (score_output_external): Likewise.
2076 * config/sparc/sparc.c (sparc_init_machine_status): Likewise.
2077 * config/spu/spu.c (spu_init_machine_status): Likewise.
2078 * config/tilegx/tilegx.c (tilegx_init_machine_status): Likewise.
2079 * config/tilepro/tilepro.c (tilepro_init_machine_status): Likewise.
2080 * config/xtensa/xtensa.c (xtensa_init_machine_status): Likewise.
2081 * coverage.c (coverage_end_function): Likewise.
2082 * dbxout.c (dbxout_init): Likewise.
2083 * doc/gty.texi: Don't mention variable_size attribute.
2084 * dwarf2cfi.c (new_cfi): Adjust.
2085 (new_cfi_row): Likewise.
2086 (copy_cfi_row): Likewise.
2087 (create_cie_data): Likewise.
2088 * dwarf2out.c (dwarf2out_alloc_current_fde): Likewise.
2089 (new_loc_descr): Likewise.
2090 (find_AT_string_in_table): Likewise.
2091 (add_addr_table_entry): Likewise.
2092 (new_die): Likewise.
2093 (add_var_loc_to_decl): Likewise.
2094 (clone_die): Likewise.
2095 (clone_as_declaration): Likewise.
2096 (break_out_comdat_types): Likewise.
2097 (new_loc_list): Likewise.
2098 (add_loc_descr_to_each): Likewise.
2099 (add_location_or_const_value_attribute): Likewise.
2100 (add_linkage_name): Likewise.
2101 (lookup_filename): Likewise.
2102 (dwarf2out_var_location): Likewise.
2103 (new_line_info_table): Likewise.
2104 (dwarf2out_init): Likewise.
2105 (mem_loc_descriptor): Likewise.
2106 (loc_descriptor): Likewise.
2107 (add_const_value_attribute): Likewise.
2108 (tree_add_const_value_attribute): Likewise.
2109 (comp_dir_string): Likewise.
2110 (dwarf2out_vms_debug_main_pointer): Likewise.
2111 (string_cst_pool_decl): Likewise.
2112 * emit-rtl.c (set_mem_attrs): Likewise.
2113 (get_reg_attrs): Likewise.
2114 (start_sequence): Likewise.
2115 (init_emit): Likewise.
2116 (init_emit_regs): Likewise.
2117 * except.c (init_eh_for_function): Likewise.
2118 (gen_eh_region): Likewise.
2119 (gen_eh_region_catch): Likewise.
2120 (gen_eh_landing_pad): Likewise.
2121 (add_call_site): Likewise.
2122 * function.c (add_frame_space): Likewise.
2123 (insert_temp_slot_address): Likewise.
2124 (assign_stack_temp_for_type): Likewise.
2125 (get_hard_reg_initial_val): Likewise.
2126 (allocate_struct_function): Likewise.
2127 (prepare_function_start): Likewise.
2128 (types_used_by_var_decl_insert): Likewise.
2129 * gengtype.c (variable_size_p): Remove function.
2130 (enum alloc_quantity): Remove enum.
2131 (write_typed_alloc_def): Remove function.
2132 (write_typed_struct_alloc_def): Likewise.
2133 (write_typed_typedef_alloc_def): Likewise.
2134 (write_typed_alloc_defns): Likewise.
2136 * ggc-common.c (ggc_cleared_alloc_htab_ignore_args): Adjust.
2137 (ggc_cleared_alloc_ptr_array_two_args): Likewise.
2138 * ggc.h (ggc_alloc): new function.
2139 (ggc_cleared_alloc): Likewise.
2140 (ggc_vec_alloc): Template on type of vector element, and remove
2141 element size argument.
2142 (ggc_cleared_vec_alloc): Likewise.
2143 * gimple.c (gimple_build_omp_for): Adjust.
2144 (gimple_copy): Likewise.
2145 * ipa-cp.c (get_replacement_map): Likewise.
2146 (find_aggregate_values_for_callers_subset): Likewise.
2147 (known_aggs_to_agg_replacement_list): Likewise.
2148 * ipa-devirt.c (get_odr_type): Likewise.
2149 * ipa-prop.c (ipa_node_duplication_hook): Likewise.
2150 (read_agg_replacement_chain): Likewise.
2151 * loop-iv.c (get_simple_loop_desc): Likewise.
2152 * lto-cgraph.c (input_node_opt_summary): Likewise.
2153 * lto-section-in.c (lto_new_in_decl_state): Likewise.
2154 * lto-streamer-in.c (lto_input_eh_catch_list): Likewise.
2155 (input_eh_region): Likewise.
2156 (input_eh_lp): Likewise.
2157 (input_cfg): Likewise.
2158 * optabs.c (set_optab_libfunc): Likewise.
2159 (init_tree_optimization_optabs): Likewise.
2160 (set_conv_libfunc): Likewise.
2161 * passes.c (do_per_function_toporder): Likewise.
2162 * rtl.h: Don't use variable_size gty attribute.
2163 * sese.c (if_region_set_false_region): Adjust.
2164 * stringpool.c (gt_pch_save_stringpool): Likewise.
2165 * target-globals.c (save_target_globals): Likewise.
2166 * toplev.c (general_init): Likewise.
2167 * trans-mem.c (record_tm_replacement): Likewise.
2168 (split_bb_make_tm_edge): Likewise.
2169 * tree-cfg.c (move_sese_region_to_fn): Likewise.
2170 * tree-data-ref.h (lambda_vector_new): Likewise.
2171 * tree-eh.c (add_stmt_to_eh_lp_fn): Likewise.
2172 * tree-iterator.c (tsi_link_before): Likewise.
2173 (tsi_link_after): Likewise.
2174 * tree-scalar-evolution.c (new_scev_info_str): Likewise.
2175 * tree-ssa-loop-niter.c (record_estimate): Likewise.
2176 * tree-ssa-operands.c (ssa_operand_alloc): Likewise.
2177 * tree-ssa-operands.h: Don't use variable_size gty attribute.
2178 * tree-ssa.c (init_tree_ssa): Adjust.
2179 * tree-ssanames.c (set_range_info): Likewise.
2180 (get_ptr_info): Likewise.
2181 (duplicate_ssa_name_ptr_info): Likewise.
2182 (duplicate_ssa_name_range_info): Likewise.
2183 * tree-streamer-in.c (unpack_ts_real_cst_value_fields): Likewise.
2184 (unpack_ts_fixed_cst_value_fields): Likewise.
2185 * tree.c (build_fixed): Likewise.
2186 (build_real): Likewise.
2187 (build_string): Likewise.
2188 (decl_priority_info): Likewise.
2189 (decl_debug_expr_insert): Likewise.
2190 (decl_value_expr_insert): Likewise.
2191 (decl_debug_args_insert): Likewise.
2192 (type_hash_add): Likewise.
2193 (build_omp_clause): Likewise.
2194 * ubsan.c (decl_for_type_insert): Likewise.
2195 * varasm.c (get_unnamed_section): Likewise.
2196 (get_noswitch_section): Likewise.
2197 (get_section): Likewise.
2198 (get_block_for_section): Likewise.
2199 (create_block_symbol): Likewise.
2200 (build_constant_desc): Likewise.
2201 (create_constant_pool): Likewise.
2202 (force_const_mem): Likewise.
2203 (record_tm_clone_pair): Likewise.
2204 * varpool.c (varpool_create_empty_node): Likewise.
2206 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
2208 * dwarf2out.c (tree_add_const_value_attribute): Call
2209 ggc_internal_cleared_alloc instead of ggc_alloc_cleared_atomic.
2210 * gengtype.c (write_typed_alloc_def): Call ggc_internal_<x>alloc
2211 instead of ggc_internal_<x>alloc_stat.
2212 * ggc-common.c (ggc_internal_cleared_alloc): Drop _stat suffix.
2213 (ggc_realloc): Likewise.
2214 * ggc-none.c (ggc_internal_alloc): Likewise.
2215 (ggc_internal_cleared_alloc): Likewise.
2216 * ggc-page.c: Likewise.
2217 * ggc.h (ggc_internal_alloc_stat): Likewise.
2218 (ggc_internal_alloc): Remove macro.
2219 (ggc_internal_cleared_alloc_stat): Drop _stat suffix.
2220 (ggc_internal_cleared_alloc): Remove macro.
2221 (GGC_RESIZEVEC): Adjust.
2222 (ggc_resizevar): Remove macro.
2223 (ggc_internal_vec_alloc_stat): Drop _stat suffix.
2224 (ggc_internal_cleared_vec_alloc_stat): Likewise.
2225 (ggc_internal_vec_cleared_alloc): Remove macro.
2226 (ggc_alloc_atomic_stat): Drop _stat suffix.
2227 (ggc_alloc_atomic): Remove macro.
2228 (ggc_alloc_cleared_atomic): Remove macro.
2229 (ggc_alloc_string_stat): Drop _stat suffix.
2230 (ggc_alloc_string): Remove macro.
2231 (ggc_alloc_rtx_def_stat): Adjust.
2232 (ggc_alloc_tree_node_stat): Likewise.
2233 (ggc_alloc_cleared_tree_node_stat): Likewise.
2234 (ggc_alloc_cleared_gimple_statement_stat): Likewise.
2235 (ggc_alloc_cleared_simd_clone_stat): Likewise.
2236 * gimple.c (gimple_build_omp_for): Likewise.
2237 (gimple_copy): Likewise.
2238 * stringpool.c (ggc_alloc_string_stat): Drop _stat suffix.
2239 * toplev.c (realloc_for_line_map): Adjust.
2240 * tree-data-ref.h (lambda_vector_new): Likewise.
2241 * tree-phinodes.c (allocate_phi_node): Likewise.
2242 * tree.c (grow_tree_vec_stat): Likewise.
2243 * vec.h (va_gc::reserve): Adjust.
2245 2014-05-17 Ajit Agarwal <ajitkum@xilinx.com>
2247 * config/microblaze/microblaze.c (break_handler): New Declaration.
2248 (microblaze_break_function_p,microblaze_is_break_handler): New.
2249 (compute_frame_size): Use microblaze_break_function_p.
2250 Add the test of break_handler.
2251 (microblaze_function_prologue) : Add the test of variable
2252 break_handler. Check the fnname by BREAK_HANDLER_NAME.
2253 (microblaze_function_epilogue) : Add the test of break_handler.
2254 (microblaze_globalize_label) : Add the test of break_handler.
2255 Check the name by BREAK_HANDLER_NAME.
2257 * config/microblaze/microblaze.h (BREAK_HANDLER_NAME): New macro
2259 * config/microblaze/microblaze.md (*<optab>,<optab>_internal): Add
2260 microblaze_is_break_handler test.
2261 (call_internal1,call_value_intern): Use microblaze_break_function_p.
2262 Use SYMBOL_REF_DECL.
2264 * config/microblaze/microblaze-protos.h
2265 (microblaze_break_function_p,microblaze_is_break_handler):
2268 * doc/extend.texi (MicroBlaze break_handler Functions): Document
2269 new MicroBlaze break_handler functions.
2271 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
2273 * doc/extend.texi (Size of an asm): Move node text according
2274 to its @menu entry position.
2276 2014-05-17 Marc Glisse <marc.glisse@inria.fr>
2278 PR tree-optimization/61140
2279 PR tree-optimization/61150
2280 PR tree-optimization/61197
2281 * tree-ssa-phiopt.c (value_replacement): Punt on multiple phis.
2283 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
2285 * doc/invoke.texi (free): Mention Alpha. Also enabled at -Os.
2287 2014-05-17 Richard Sandiford <r.sandiford@uk.ibm.com>
2289 * wide-int.cc: Only include longlong.h if W_TYPE_SIZE==32 or
2290 __SIZEOF_INT128__ is defined.
2292 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
2294 * config/rs6000/rs6000.c (rs6000_real_tls_symbol_ref_p): New function.
2295 (rs6000_delegitimize_address): Use it.
2297 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
2299 * emit-rtl.h (replace_equiv_address, replace_equiv_address_nv): Add an
2300 inplace argument. Store the new address in the original MEM when true.
2301 * emit-rtl.c (change_address_1): Likewise.
2302 (adjust_address_1, adjust_automodify_address_1, offset_address):
2304 * rtl.h (plus_constant): Add an inplace argument.
2305 * explow.c (plus_constant): Likewise. Try to reuse the original PLUS
2306 when true. Avoid generating (plus X (const_int 0)).
2307 * function.c (instantiate_virtual_regs_in_rtx): Adjust the PLUS
2308 in-place. Pass true to plus_constant.
2309 (instantiate_virtual_regs_in_insn): Pass true to replace_equiv_address.
2311 2014-05-16 Dehao Chen <dehao@google.com>
2313 * tree-cfg.c (gimple_merge_blocks): Updates bb count with max count.
2315 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
2318 * config/sh/predicates.md (negt_reg_shl31_operand): Match additional
2320 * config/sh/sh.md (*negt_msb): Merge SH2A and non-SH2A variants.
2322 2014-05-16 Dehao Chen <dehao@google.com>
2324 * ira-int.h (REG_FREQ_FROM_EDGE_FREQ): Use
2325 optimize_function_for_size_p.
2326 * regs.h (REG_FREQ_FROM_BB): Likewise.
2328 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
2331 * config/sh/sh.c (sh_eval_treg_value): Handle t_reg_operand and
2332 negt_reg_operand cases.
2333 * config/sh/sh.md (*cset_zero): Likewise by using cbranch_treg_value
2335 * config/sh/predicates.md (cbranch_treg_value): Simplify.
2337 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
2339 * config/sh/sh.c (sh_option_override): Set branch cost to 2 for all
2342 2014-05-16 David Malcolm <dmalcolm@redhat.com>
2345 2014-04-29 David Malcolm <dmalcolm@redhat.com>
2347 * tree-cfg.c (dump_function_to_file): Dump the return type of
2348 functions, in a line to itself before the function body, mimicking
2349 the layout of a C function.
2351 2014-05-16 Dehao Chen <dehao@google.com>
2353 * cfghooks.c (make_forwarder_block): Use direct computation to
2354 get fall-through edge's count and frequency.
2356 2014-05-16 Benno Schulenberg <bensberg@justemail.net>
2358 * config/arc/arc.c (arc_init): Fix typo in error message.
2359 * config/i386/i386.c (ix86_expand_builtin): Likewise.
2360 (split_stack_prologue_scratch_regno): Likewise.
2361 * fortran/check.c (gfc_check_fn_rc2008): Remove duplicate
2362 word from error message.
2364 2014-05-16 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
2366 * ira-costs.c: Fix typo in comment.
2368 2014-05-16 David Wohlferd <dw@LimeGreenSocks.com>
2370 * doc/extend.texi: (Visibility Pragmas) Fix misplaced @xref
2372 2014-05-16 Jan Hubicka <hubicka@ucw.cz>
2374 * varpool.c (dump_varpool_node): Dump write-only flag.
2375 * lto-cgraph.c (lto_output_varpool_node, input_varpool_node): Stream
2377 * tree-cfg.c (execute_fixup_cfg): Remove statements setting
2378 write-only variables.
2379 * ipa.c (process_references): New function.
2380 (set_readonly_bit): New function.
2381 (set_writeonly_bit): New function.
2382 (clear_addressable_bit): New function.
2383 (ipa_discover_readonly_nonaddressable_var): Mark write only variables;
2384 fix handling of aliases.
2385 * cgraph.h (struct varpool_node): Add writeonly flag.
2387 2014-05-16 Vladimir Makarov <vmakarov@redhat.com>
2389 PR rtl-optimization/60969
2390 * ira-costs.c (record_reg_classes): Allow only memory for pseudo.
2391 Calculate costs for this case.
2393 2014-05-16 Eric Botcazou <ebotcazou@adacore.com>
2395 * fold-const (fold_unary_loc) <NON_LVALUE_EXPR>: New case.
2396 <CASE_CONVERT>: Pass arg0 instead of op0 to fold_convert_const.
2398 2014-05-16 Richard Biener <rguenther@suse.de>
2400 PR tree-optimization/61194
2401 * tree-vect-patterns.c (adjust_bool_pattern): Also handle
2402 bool patterns ending in a COND_EXPR.
2404 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2406 * config/aarch64/aarch64.c (aarch64_rtx_mult_cost): Fix FNMUL case.
2408 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2410 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle the case
2411 where we were unable to cost an RTX.
2413 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2415 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost SYMBOL_REF,
2418 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2419 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2421 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost TRUNCATE.
2423 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2424 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2426 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost FMA,
2427 FLOAT_EXTEND, FLOAT_TRUNCATE, ABS, SMAX, and SMIN.
2429 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2430 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2432 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost comparison
2435 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2436 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2438 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
2441 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2442 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2444 * config/aarch64/aarch64.c (aarch64_rtx_arith_op_extract_p): New.
2445 (aarch64_rtx_costs): Improve costs for SIGN/ZERO_EXTRACT.
2447 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2448 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2450 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
2453 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2454 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2456 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost
2457 ZERO_EXTEND and SIGN_EXTEND better.
2459 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2460 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2462 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve cost for
2465 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2466 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2468 * config/aarch64/aarch64.c (aarch64_rtx_costs): Use address
2469 costs when costing loads and stores to memory.
2471 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2472 Philip Tomsich <philipp.tomsich@theobroma-systems.com>
2474 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costing
2477 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2479 * config/aarch64/aarch64.c (aarch64_rtx_costs): Set default costs.
2481 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2482 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2484 * config/aarch64/aarch64.c (aarch64_strip_shift_or_extend): Rename
2486 (aarch64_strip_extend): ...this, don't strip shifts, check RTX is
2488 (aarch64_rtx_mult_cost): New.
2489 (aarch64_rtx_costs): Use it, refactor as appropriate.
2491 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2492 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2494 * config/aarch64/aarch64.c (aarch64_build_constant): Conditionally
2495 emit instructions, return number of instructions which would
2497 (aarch64_add_constant): Update call to aarch64_build_constant.
2498 (aarch64_output_mi_thunk): Likewise.
2499 (aarch64_rtx_costs): Estimate cost of a CONST_INT, cost of
2502 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2504 * config/aarch64/aarch64.c (aarch64_rtx_costs_wrapper): New.
2505 (TARGET_RTX_COSTS): Call it.
2507 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2509 * config/aarch64/aarch64.c (cortexa57_addrcost_table): New.
2510 (cortexa57_vector_cost): Likewise.
2511 (cortexa57_tunings): Use them.
2513 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2515 * config/aarch64/aarch64-protos.h (scale_addr_mode_cost): New.
2516 (cpu_addrcost_table): Use it.
2517 * config/aarch64/aarch64.c (generic_addrcost_table): Initialize it.
2518 (aarch64_address_cost): Rewrite using aarch64_classify_address,
2521 2014-05-16 Richard Biener <rguenther@suse.de>
2523 * tree-ssa-sccvn.c: Include tree-cfg.h and domwalk.h.
2524 (set_ssa_val_to): Handle unexpected sets to VN_TOP.
2525 (visit_phi): Ignore edges marked as not executable.
2526 (class cond_dom_walker): New.
2527 (cond_dom_walker::before_dom_children): Value-number
2528 control statements and mark successor edges as not
2529 executable if possible.
2530 (run_scc_vn): First walk all control statements in
2531 dominator order, marking edges as not executable.
2532 * tree-inline.c (copy_edges_for_bb): Be not confused
2533 about random edge flags.
2535 2014-05-16 Richard Biener <rguenther@suse.de>
2537 * tree-ssa-sccvn.c (visit_use): Also constant-fold calls.
2539 2014-05-15 Peter Bergner <bergner@vnet.ibm.com>
2542 * config/rs6000/htmxlintrin.h (_HTM_TBEGIN_STARTED): New define.
2543 (__TM_simple_begin): Use it.
2544 (__TM_begin): Likewise.
2546 2014-05-15 Martin Jambor <mjambor@suse.cz>
2549 * ipa-prop.c (update_indirect_edges_after_inlining): Check
2550 type_preserved flag when the indirect edge is polymorphic.
2552 2014-05-15 Martin Jambor <mjambor@suse.cz>
2554 PR tree-optimization/61090
2555 * tree-sra.c (sra_modify_expr): Pass the current gsi to
2556 build_ref_for_model.
2558 2014-05-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2560 * config/arm/arm.c (arm_option_override): Use the SCHED_PRESSURE_MODEL
2561 enum name for PARAM_SCHED_PRESSURE_ALGORITHM.
2563 2014-05-15 Jakub Jelinek <jakub@redhat.com>
2565 PR tree-optimization/61158
2566 * fold-const.c (fold_binary_loc): If X is zero-extended and
2567 shiftc >= prec, make sure zerobits is all ones instead of
2568 invoking undefined behavior.
2570 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2572 * regcprop.h: New file.
2573 * regcprop.c (skip_debug_insn_p): New decl.
2574 (replace_oldest_value_reg): Check skip_debug_insn_p.
2575 (copyprop_hardreg_forward_bb_without_debug_insn): New function.
2576 * shrink-wrap.c: Include regcprop.h.
2577 (prepare_shrink_wrap): Call
2578 copyprop_hardreg_forward_bb_without_debug_insn.
2580 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2582 * shrink-wrap.h: Update comment.
2583 * shrink-wrap.c: Update comment.
2584 (next_block_for_reg): Rename to live_edge_for_reg.
2585 (live_edge_for_reg): Allow live_edge->dest has two predecessors.
2586 (move_insn_for_shrink_wrap): Split live_edge.
2587 (prepre_shrink_wrap): One more parameter for move_insn_for_shrink_wrap.
2589 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
2591 * config/sparc/sparc-protos.h (sparc_absnegfloat_split_legitimate):
2593 * config/sparc/sparc.c (sparc_absnegfloat_split_legitimate): Likewise.
2594 * config/sparc/sparc.md (fptype_ut699): New attribute.
2595 (in_branch_delay): Return false if -mfix-ut699 is specified and
2596 fptype_ut699 is set to single.
2597 (truncdfsf2): Add fptype_ut699 attribute.
2598 (fix_truncdfsi2): Likewise.
2599 (floatsisf2): Change fptype attribute.
2600 (fix_truncsfsi2): Likewise.
2601 (negtf2_notv9): Delete.
2602 (negtf2_v9): Likewise.
2603 (negtf2_hq): New instruction.
2604 (negtf2): New instruction and splitter.
2605 (negdf2_notv9): Rewrite.
2606 (abstf2_notv9): Delete.
2607 (abstf2_hq_v9): Likewise.
2608 (abstf2_v9): Likewise.
2609 (abstf2_hq): New instruction.
2610 (abstf2): New instruction and splitter.
2611 (absdf2_notv9): Rewrite.
2613 2014-05-14 Cary Coutant <ccoutant@google.com>
2616 * opts.c (common_handle_option): Don't special-case "-g".
2617 (set_debug_level): Default to at least level 2 with "-g".
2619 2014-05-14 DJ Delorie <dj@redhat.com>
2621 * config/msp430/msp430.c (msp430_builtin): Add
2622 MSP430_BUILTIN_DELAY_CYCLES.
2623 (msp430_init_builtins): Register void __delay_cycles(long long).
2624 (msp430_builtin_decl): Add it.
2625 (cg_magic_constant): New.
2626 (msp430_expand_delay_cycles): New.
2627 (msp430_expand_builtin): Call it.
2628 (msp430_print_operand_raw): Change integer printing from "int" to
2630 * config/msp430/msp430.md (define_constants): Add delay_cycles tags.
2631 (delay_cycles_start): New.
2632 (delay_cycles_end): New.
2633 (delay_cycles_32): New.
2634 (delay_cycles_32x): New.
2635 (delay_cycles_16): New.
2636 (delay_cycles_16x): New.
2637 (delay_cycles_2): New.
2638 (delay_cycles_1): New.
2639 * doc/extend.texi: Document __delay_cycles().
2641 2014-05-14 Sandra Loosemore <sandra@codesourcery.com>
2643 * config/nios2/nios2.md (nios2_cbranch): Fix paste-o in
2644 length attribute computation.
2646 2014-05-14 Richard Sandiford <rdsandiford@googlemail.com>
2649 * print-rtl.c (print_rtx): Suppress uids if flag_dump_unnumbered.
2651 2014-05-14 Richard Sandiford <r.sandiford@uk.ibm.com>
2654 * config/sparc/sparc.md: Fix types of low and high in DI constant
2655 splitter. Use gen_int_mode in some other splitters.
2657 2014-05-14 Martin Jambor <mjambor@suse.cz>
2660 * ipa-prop.c (ipa_modify_formal_parameters): Reset DECL_LANG_SPECIFIC.
2662 2014-05-14 James Norris <jnorris@codesourcery.com>
2664 * omp-low.c (expand_parallel_call): Remove shadow variable.
2665 (expand_omp_taskreg): Likewise.
2667 2014-05-14 Ilya Tocar <ilya.tocar@intel.com>
2669 * common/config/i386/i386-common.c
2670 (OPTION_MASK_ISA_CLFLUSHOPT_SET): Define.
2671 (OPTION_MASK_ISA_XSAVES_SET): Ditto.
2672 (OPTION_MASK_ISA_XSAVEC_SET): Ditto.
2673 (OPTION_MASK_ISA_CLFLUSHOPT_UNSET): Ditto.
2674 (OPTION_MASK_ISA_XSAVES_UNSET): Ditto.
2675 (OPTION_MASK_ISA_XSAVEC_UNSET): Ditto.
2676 (ix86_handle_option): Handle OPT_mxsavec, OPT_mxsaves, OPT_mclflushopt.
2677 * config.gcc (i[34567]86-*-*): Add clflushoptintrin.h,
2678 xsavecintrin.h, xsavesintrin.h.
2679 (x86_64-*-*): Ditto.
2680 * config/i386/clflushoptintrin.h: New.
2681 * config/i386/xsavecintrin.h: Ditto.
2682 * config/i386/xsavesintrin.h: Ditto.
2683 * config/i386/cpuid.h (bit_CLFLUSHOPT): Define.
2684 (bit_XSAVES): Ditto.
2685 (bit_XSAVES): Ditto.
2686 * config/i386/driver-i386.c (host_detect_local_cpu): Handle
2687 -mclflushopt, -mxsavec, -mxsaves, -mno-xsaves, -mno-xsavec,
2689 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
2690 OPTION_MASK_ISA_CLFLUSHOPT, OPTION_MASK_ISA_XSAVEC,
2691 OPTION_MASK_ISA_XSAVES.
2692 * config/i386/i386.c (ix86_target_string): Handle -mclflushopt,
2694 (PTA_CLFLUSHOPT) Define.
2695 (PTA_XSAVEC): Ditto.
2696 (PTA_XSAVES): Ditto.
2697 (ix86_option_override_internal): Handle new options.
2698 (ix86_valid_target_attribute_inner_p): Ditto.
2699 (ix86_builtins): Add IX86_BUILTIN_XSAVEC, IX86_BUILTIN_XSAVEC64,
2700 IX86_BUILTIN_XSAVES, IX86_BUILTIN_XRSTORS, IX86_BUILTIN_XSAVES64,
2701 IX86_BUILTIN_XRSTORS64, IX86_BUILTIN_CLFLUSHOPT.
2702 (bdesc_special_args): Add __builtin_ia32_xsaves,
2703 __builtin_ia32_xrstors, __builtin_ia32_xsavec, __builtin_ia32_xsaves64,
2704 __builtin_ia32_xrstors64, __builtin_ia32_xsavec64.
2705 (ix86_init_mmx_sse_builtins): Add __builtin_ia32_clflushopt.
2706 (ix86_expand_builtin): Handle new builtins.
2707 * config/i386/i386.h (TARGET_CLFLUSHOPT) Define.
2708 (TARGET_CLFLUSHOPT_P): Ditto.
2709 (TARGET_XSAVEC): Ditto.
2710 (TARGET_XSAVEC_P): Ditto.
2711 (TARGET_XSAVES): Ditto.
2712 (TARGET_XSAVES_P): Ditto.
2713 * config/i386/i386.md (ANY_XSAVE): Add UNSPECV_XSAVEC, UNSPECV_XSAVES.
2714 (ANY_XSAVE64)" Add UNSPECV_XSAVEC64, UNSPECV_XSAVES64.
2715 (attr xsave): Add xsavec, xsavec64, xsaves, xsaves64.
2717 (ANY_XRSTOR64): Ditto.
2719 (xrstor): Change into <xrstor>.
2720 (xrstor_rex64): Change into <xrstor>_rex64.
2721 (xrstor64): Change into <xrstor>64
2723 * config/i386/i386.opt (mclflushopt): New.
2726 * config/i386/x86intrin.h: Add clflushoptintrin.h, xsavesintrin.h,
2728 * doc/invoke.texi: Document new options.
2730 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
2732 PR rtl-optimization/60866
2733 * sel-sched-ir (sel_init_new_insn): New parameter old_seqno.
2734 Default it to -1. Pass it down to init_simplejump_data.
2735 (init_simplejump_data): New parameter old_seqno. Pass it down
2736 to get_seqno_for_a_jump.
2737 (get_seqno_for_a_jump): New parameter old_seqno. Use it for
2738 initializing new jump seqno as a last resort. Add comment.
2739 (sel_redirect_edge_and_branch): Save old seqno of the conditional
2740 jump and pass it down to sel_init_new_insn.
2741 (sel_redirect_edge_and_branch_force): Likewise.
2743 2014-05-14 Georg-Johann Lay <avr@gjlay.de>
2745 * config/avr/avr.h (REG_CLASS_CONTENTS): Use unsigned suffix for
2746 shifted values to avoid build warning.
2748 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
2750 * cfgcleanup.c (try_forward_edges): Use location_t for locations.
2751 * cfgrtl.c (rtl_merge_blocks): Fix comment.
2752 (cfg_layout_merge_blocks): Likewise.
2753 * except.c (emit_to_new_bb_before): Remove prev_bb local variable.
2755 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
2757 PR rtl-optimization/60901
2758 * config/i386/i386.c (ix86_dependencies_evaluation_hook): Check that
2759 bb predecessor belongs to the same scheduling region. Adjust comment.
2761 2014-05-13 Peter Bergner <bergner@vnet.ibm.com>
2763 * doc/sourcebuild.texi: (dfp_hw): Document.
2764 (p8vector_hw): Likewise.
2765 (powerpc_eabi_ok): Likewise.
2766 (powerpc_elfv2): Likewise.
2767 (powerpc_htm_ok): Likewise.
2768 (ppc_recip_hw): Likewise.
2771 2014-05-13 Cary Coutant <ccoutant@google.com>
2773 * opts.c (finish_options): Use -ggnu-pubnames with -gsplit-dwarf.
2775 2014-05-13 David Malcolm <dmalcolm@redhat.com>
2777 * gengtype-parse.c (require3): Eliminate in favor of...
2779 (require_template_declaration): Update to support optional single *
2782 * gengtype.c (get_ultimate_base_class): Add a non-const overload.
2783 (create_user_defined_type): Handle a single level of explicit
2784 pointerness within template arguments.
2785 (struct write_types_data): Add field "kind".
2786 (filter_type_name): Handle "*" character.
2787 (write_user_func_for_structure_ptr): Require a write_types_data
2788 rather than just a prefix string, so that we can look up the kind
2789 of the wtd and use it as an index into wrote_user_func_for_ptr,
2790 ensuring that such functions are written at most once. Support
2791 subclasses by invoking the marking function of the ultimate base class.
2792 (write_user_func_for_structure_body): Require a write_types_data
2793 rather than just a prefix string, so that we can pass this to
2794 write_user_func_for_structure_ptr.
2795 (write_func_for_structure): Likewise.
2796 (ggc_wtd): Add initializer of new "kind" field.
2797 (pch_wtd): Likewise.
2799 * gengtype.h (enum write_types_kinds): New.
2800 (struct type): Add field wrote_user_func_for_ptr to the "s"
2803 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
2805 * fold-const.c (optimize_bit_field_compare): Use wi:: operations
2806 instead of const_binop.
2807 (fold_binary_loc): Likewise.
2809 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
2811 * tree-dfa.h (get_addr_base_and_unit_offset_1): Update array index
2812 calculation to match get_ref_base_and_extent.
2814 2014-05-13 Catherine Moore <clm@codesourcery.com>
2815 Sandra Loosemore <sandra@codesourcery.com>
2817 * configure.ac: Fix assembly for explicit JALR relocation check.
2818 * configure: Regenerate.
2820 2014-05-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2822 * config/arm/arm.c (neon_itype): Remove NEON_RESULTPAIR.
2823 (arm_init_neon_builtins): Remove handling of NEON_RESULTPAIR.
2824 Remove associated type declarations and initialisations.
2825 (arm_expand_neon_builtin): Likewise.
2826 (neon_emit_pair_result_insn): Delete.
2827 * config/arm/arm_neon_builtins (vtrn, vzip, vuzp): Delete.
2828 * config/arm/neon.md (neon_vtrn<mode>): Delete.
2829 (neon_vzip<mode>): Likewise.
2830 (neon_vuzp<mode>): Likewise.
2832 2014-05-13 Richard Biener <rguenther@suse.de>
2835 * tree-inline.c (remap_gimple_stmt): Clear tail call flag,
2836 it needs revisiting whether the call still may be tail-called.
2838 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2840 * rtl.def (SYMBOL_REF): Remove middle "0" field.
2841 * rtl.h (block_symbol): Reduce number of fields to 2.
2842 (rtx_def): Add u2.symbol_ref_flags.
2843 (SYMBOL_REF_FLAGS): Use it.
2844 (SYMBOL_REF_DATA, SET_SYMBOL_REF_DECL, SYMBOL_REF_DECL)
2845 (SET_SYMBOL_REF_CONSTANT, SYMBOL_REF_CONSTANT): Lower index.
2846 * gengtype.c (adjust_field_rtx_def): Remove SYMBOL_REF_FLAGS handling.
2847 Lower index of SYMBOL_REF_DATA.
2848 * print-rtl.c (print_rtx): Lower index for SYMBOL_REF_DATA.
2849 Print SYMBOL_REF_FLAGS at the same time.
2850 * genattrtab.c (attr_rtx_1): Only initialize 1 "0" SYMBOL_REF field.
2852 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2854 * rtl.def (VAR_LOCATION): Remove "i" field.
2855 * rtl.h (rtx_def): Add u2.var_location_status.
2856 (PAT_VAR_LOCATION_STATUS): Use it.
2857 (gen_rtx_VAR_LOCATION): Declare.
2858 * gengenrtl.c (excluded_rtx): Add VAR_LOCATION.
2859 * emit-rtl.c (gen_rtx_VAR_LOCATION): New function.
2860 * var-tracking.c (emit_note_insn_var_location): Remove casts.
2862 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2864 * rtl.def (scratch): Fix outdated comment and remove "0" field.
2865 * gengtype.c (adjust_field_rtx_def): Update accordingly.
2867 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2869 * rtl.def (DEBUG_INSN, INSN, JUMP_INSN, CALL_INSN, JUMP_TABLE_DATA)
2870 (BARRIER, CODE_LABEL, NOTE): Remove first "i" field.
2871 * rtl.h (rtx_def): Add insn_uid to u2 field.
2872 (RTX_FLAG_CHECK8): Delete in favor of...
2873 (RTL_INSN_CHAIN_FLAG_CHECK): ...this new macro.
2874 (INSN_DELETED_P): Update accordingly.
2875 (INSN_UID): Use u2.insn_uid.
2876 (INSN_CHAIN_CODE_P): Define.
2877 (PREV_INSN, NEXT_INSN, BLOCK_FOR_INSN, PATTERN, INSN_LOCATION)
2878 (INSN_CODE, REG_NOTES, CALL_INSN_FUNCTION_USAGE, CODE_LABEL_NUMBER)
2879 (NOTE_DATA, NOTE_DELETED_LABEL_NAME, NOTE_BLOCK, NOTE_EH_HANDLER)
2880 (NOTE_BASIC_BLOCK, NOTE_VAR_LOCATION, NOTE_CFI, NOTE_LABEL_NUMBER)
2881 (NOTE_KIND, LABEL_NAME, LABEL_NUSES, JUMP_LABEL, LABEL_REFS): Lower
2882 indices accordingly.
2883 * print-rtl.c (print_rtx): Print INSN_UIDs before the main loop.
2884 Update indices for insn-chain rtxes.
2885 * gengtype.c (gen_rtx_next): Adjust test for insn-chain rtxes.
2886 (adjust_field_rtx_def): Lower '0' indices for all insn-chain rtxes.
2887 * emit-rtl.c (gen_label_rtx): Update gen_rtx_LABEL call.
2888 * caller-save.c (init_caller_save): Update gen_rtx_INSN calls.
2889 * combine.c (try_combine): Likewise.
2890 * ira.c (setup_prohibited_mode_move_regs): Likewise.
2892 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2894 * rtl.def (REG): Remove middle field.
2895 * rtl.h (rtx_def): Add orignal_regno to u2.
2896 (ORIGINAL_REGNO): Use it instead of field 1.
2897 (REG_ATTRS): Lower field index accordingly.
2898 * gengtype.c (adjust_field_rtx_def): Remove handling of
2899 ORIGINAL_REGNO. Move REG_ATTRS index down.
2900 * print-rtl.c (print_rtx): Move ORIGINAL_REGNO handling to the
2901 code that prints the REGNO.
2903 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2905 * print-rtl.c (print_rtx): Guard whole '0' block with ifndef
2908 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2910 * rtl.h (rtx_def): Mark u2 as GTY ((skip)).
2912 2014-05-13 Bin Cheng <bin.cheng@arm.com>
2914 * tree-ssa-loop-ivopts.c (contain_complex_addr_expr): New.
2915 (alloc_iv): Lower base expressions containing ADDR_EXPR.
2917 2014-05-13 Ian Bolton <ian.bolton@arm.com>
2919 * config/aarch64/aarch64-protos.h
2920 (aarch64_hard_regno_caller_save_mode): New prototype.
2921 * config/aarch64/aarch64.c (aarch64_hard_regno_caller_save_mode):
2923 * config/aarch64/aarch64.h (HARD_REGNO_CALLER_SAVE_MODE): New macro.
2925 2014-05-13 Christian Bruel <christian.bruel@st.com>
2927 * target.def (mode_switching): New hook vector.
2928 (mode_emit, mode_needed, mode_after, mode_entry): New hooks.
2929 (mode_exit, modepriority_to_mode): Likewise.
2930 * mode-switching.c (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Hookify.
2931 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
2932 * target.h: Include tm.h and hard-reg-set.h.
2933 * doc/tm.texi.in (EMIT_MODE_SET, MODE_NEEDED, MODE_AFTER, MODE_ENTRY)
2934 (MODE_EXIT, MODE_PRIORITY_TO_MODE): Delete and hookify.
2935 * doc/tm.texi Regenerate.
2936 * config/sh/sh.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
2937 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
2938 * config/sh/sh.c (sh_emit_mode_set, sh_mode_priority): Hookify.
2939 (sh_mode_needed, sh_mode_after, sh_mode_entry, sh_mode_exit): Likewise.
2940 * config/i386/i386.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
2941 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
2942 * config/i386/i386-protos.h (ix86_mode_needed, ix86_mode_after)
2943 (ix86_mode_entrym, ix86_emit_mode_set): Remove external declaration.
2944 * config/i386/i386.c (ix86_mode_needed, ix86_mode_after,
2945 (ix86_mode_exit, ix86_mode_entry, ix86_mode_priority)
2946 (ix86_emit_mode_set): Hookify.
2947 * config/epiphany/epiphany.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY):
2949 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
2950 * config/epiphany/epiphany-protos.h (epiphany_mode_needed)
2951 (emit_set_fp_mode, epiphany_mode_entry_exit, epiphany_mode_after)
2952 (epiphany_mode_priority_to_mode): Remove declaration.
2953 * config/epiphany/epiphany.c (emit_set_fp_mode): Hookify.
2954 (epiphany_mode_needed, epiphany_mode_priority_to_mode): Likewise.
2955 (epiphany_mode_entry, epiphany_mode_exit, epiphany_mode_after):
2957 (epiphany_mode_priority_to_mode): Change priority type. Hookify.
2958 (epiphany_mode_needed, epiphany_mode_entry_exit): Hookify.
2959 (epiphany_mode_after, epiphany_mode_entry, emit_set_fp_mode): Hookify.
2961 2014-05-13 Jakub Jelinek <jakub@redhat.com>
2964 * config/i386/i386.c (ix86_expand_set_or_movmem): If count_exp
2965 is const0_rtx, return immediately. Don't test count == 0 when
2968 2014-05-13 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2970 * Makefile.in: add shrink-wrap.o.
2971 * config/i386/i386.c: include "shrink-wrap.h"
2972 * function.c: Likewise.
2973 (requires_stack_frame_p, next_block_for_reg,
2974 move_insn_for_shrink_wrap, prepare_shrink_wrap,
2975 dup_block_and_redirect): Move to shrink-wrap.c
2976 (thread_prologue_and_epilogue_insns): Extract three code segments
2977 as functions in shrink-wrap.c
2978 * function.h: Move #ifdef HAVE_simple_return ... #endif block to
2980 * shrink-wrap.c: New file.
2981 * shrink-wrap.h: New file.
2983 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
2985 * doc/extend.texi: Reflect current numbers of pragmas. Remove
2986 reference to Solaris.
2988 2014-05-12 Mike Stump <mikestump@comcast.net>
2991 * genattrtab.c (filename): Add.
2992 (convert_set_attr_alternative): Improve error message.
2993 (check_defs): Restore read_md_filename for error messages.
2994 (gen_insn): Save filename.
2996 2014-05-12 Dimitris Papavasiliou <dpapavas@gmail.com>
2998 * doc/invoke.texi: Document new switches -Wno-shadow-ivar,
2999 -fno-local-ivars and -fivar-visibility.
3000 * c-family/c.opt: Make -Wshadow also implicitly enable
3003 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
3005 * doc/tm.texi: Remove reference to deleted macro.
3006 * doc/tm.texi.in: Likewise.
3008 2014-05-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
3011 * config/avr/avr.c (avr_out_store_psi): Use correct constant
3014 2014-05-12 Georg-Johann Lay <avr@gjlay.de>
3017 * config/arm/arm.h (License): Add GCC Runtime Library Exception.
3018 * config/arm/aout.h (License): Same.
3019 * config/arm/bpabi.h (License): Same.
3020 * config/arm/elf.h (License): Same.
3021 * config/arm/linux-elf.h (License): Same.
3022 * config/arm/linux-gas.h (License): Same.
3023 * config/arm/netbsd-elf.h (License): Same.
3024 * config/arm/uclinux-eabi.h (License): Same.
3025 * config/arm/uclinux-elf.h (License): Same.
3026 * config/arm/vxworks.h (License): Same.
3028 2014-05-11 Jakub Jelinek <jakub@redhat.com>
3030 * tree.h (OMP_CLAUSE_LINEAR_STMT): Define.
3031 * tree.c (omp_clause_num_ops): Increase OMP_CLAUSE_LINEAR
3032 number of operands to 3.
3033 (walk_tree_1): Walk all operands of OMP_CLAUSE_LINEAR.
3034 * tree-nested.c (convert_nonlocal_omp_clauses,
3035 convert_local_omp_clauses): Handle OMP_CLAUSE_DEPEND.
3036 * gimplify.c (gimplify_scan_omp_clauses): Handle
3037 OMP_CLAUSE_LINEAR_STMT.
3038 * omp-low.c (lower_rec_input_clauses): Fix typo.
3039 (maybe_add_implicit_barrier_cancel, lower_omp_1): Add
3040 cast between Fortran boolean_type_node and C _Bool if
3043 2014-05-11 Richard Sandiford <rdsandiford@googlemail.com>
3045 PR tree-optimization/61136
3046 * wide-int.h (multiple_of_p): Define a version that doesn't return
3048 * fold-const.c (extract_muldiv_1): Use wi::multiple_of_p instead of an
3049 integer_zerop/const_binop pair.
3050 (multiple_of_p): Likewise, converting both operands to widest_int
3053 2014-05-09 Teresa Johnson <tejohnson@google.com>
3055 * cgraphunit.c (analyze_functions): Use correct dump file.
3057 2014-05-09 Florian Weimer <fweimer@redhat.com>
3059 * cfgexpand.c (stack_protect_decl_p): New function, extracted from
3061 (stack_protect_return_slot_p): New function.
3062 (expand_used_vars): Call stack_protect_decl_p and
3063 stack_protect_return_slot_p for -fstack-protector-strong.
3065 2014-05-09 David Wohlferd <LimeGreenSocks@yahoo.com>
3066 Andrew Haley <aph@redhat.com>
3067 Richard Sandiford <rdsandiford@googlemail.com>
3069 * doc/extend.texi: Rewrite inline asm page / re-org asm-related
3072 2014-05-09 Kenneth Zadeck <zadeck@naturalbridge.com>
3075 * fold-const.c (fold_binary_loc): Changed width of mask.
3077 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
3079 * config/avr/avr-fixed.md (round<mode>3): Use -1U instead of -1 in
3080 unsigned int initializers for regno_in, regno_out.
3082 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
3085 * config/avr/avr.md (cc): Add new attribute set_vzn.
3086 (addqi3, addqq3, adduqq3, subqi3, subqq3, subuqq3, negqi2) [cc]:
3087 Set cc insn attribute to set_vzn instead of set_zn for alternatives
3088 with INC, DEC or NEG.
3089 * config/avr/avr.c (avr_notice_update_cc): Handle SET_VZN.
3090 (avr_out_plus_1): ADIW sets cc0 to CC_SET_CZN.
3091 INC, DEC and ADD+ADC set cc0 to CC_CLOBBER.
3093 2014-05-09 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3096 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3098 * wide-int.cc (UTItype): Define.
3099 (UDWtype): Define for appropriate W_TYPE_SIZE.
3101 2014-05-09 Richard Biener <rguenther@suse.de>
3103 * Makefile.in (GTFILES): Remove tree-ssa-propagate.c.
3104 * tree-ssa-propagate.c: Do not include gt-tree-ssa-propagate.h.
3105 (interesting_ssa_edges, varying_ssa_edges): Move out of GC space.
3106 (add_ssa_edge, process_ssa_edge_worklist, ssa_prop_init,
3107 ssa_propagate): Adjust.
3109 2014-05-08 Jeff Law <law@redhat.com>
3111 PR tree-optimization/61009
3112 * tree-ssa-threadedge.c (thread_through_normal_block): Return a
3113 tri-state rather than a boolean. When a block is too big to
3114 thread through, inform caller via negative return value.
3115 (thread_across_edge): If a block was too big for normal threading,
3116 then it's too big for a joiner too, so remove temporary equivalences
3117 and return immediately.
3119 2014-05-08 Manuel López-Ibáñez <manu@gcc.gnu.org>
3120 Matthias Klose <doko@ubuntu.com>
3123 * optc-gen.awk: Fix option handling for -Wunused-parameter.
3125 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
3128 * config/i386/i386.c (PTA_HASWELL): Remove PTA_RTM.
3130 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
3133 * config/alpha/alpha.c: Include gimple-iterator.h.
3134 (alpha_gimple_fold_builtin): New function. Move
3135 ALPHA_BUILTIN_UMULH folding from ...
3136 (alpha_fold_builtin): ... here.
3137 (TARGET_GIMPLE_FOLD_BUILTIN): New define.
3139 2014-05-08 Wei Mi <wmi@google.com>
3142 * config/i386/i386.c (ix86_compute_frame_layout): Update
3143 preferred_stack_boundary for call, expanded from tls descriptor.
3144 * config/i386/i386.md (*tls_global_dynamic_32_gnu): Update RTX
3145 to depend on SP register.
3146 (*tls_local_dynamic_base_32_gnu): Ditto.
3147 (*tls_local_dynamic_32_once): Ditto.
3148 (tls_global_dynamic_64_<mode>): Set
3149 ix86_tls_descriptor_calls_expanded_in_cfun.
3150 (tls_local_dynamic_base_64_<mode>): Ditto.
3151 (tls_global_dynamic_32): Set
3152 ix86_tls_descriptor_calls_expanded_in_cfun. Update RTX
3153 to depend on SP register.
3154 (tls_local_dynamic_base_32): Ditto.
3156 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3158 * config/arm/arm_neon.h: Update comment.
3159 * config/arm/neon-docgen.ml: Delete.
3160 * config/arm/neon-gen.ml: Delete.
3161 * doc/arm-neon-intrinsics.texi: Update comment.
3163 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3165 * config/arm/arm_neon_builtins.def (vadd, vsub): Only define the v2sf
3167 (vand, vorr, veor, vorn, vbic): Remove.
3168 * config/arm/neon.md (neon_vadd, neon_vsub, neon_vadd_unspec): Adjust
3170 (neon_vsub_unspec): Likewise.
3171 (neon_vorr, neon_vand, neon_vbic, neon_veor, neon_vorn): Remove.
3173 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3175 * config/arm/arm_neon.h (vadd_s8): GNU C implementation
3176 (vadd_s16): Likewise.
3177 (vadd_s32): Likewise.
3178 (vadd_f32): Likewise.
3179 (vadd_u8): Likewise.
3180 (vadd_u16): Likewise.
3181 (vadd_u32): Likewise.
3182 (vadd_s64): Likewise.
3183 (vadd_u64): Likewise.
3184 (vaddq_s8): Likewise.
3185 (vaddq_s16): Likewise.
3186 (vaddq_s32): Likewise.
3187 (vaddq_s64): Likewise.
3188 (vaddq_f32): Likewise.
3189 (vaddq_u8): Likewise.
3190 (vaddq_u16): Likewise.
3191 (vaddq_u32): Likewise.
3192 (vaddq_u64): Likewise.
3193 (vmul_s8): Likewise.
3194 (vmul_s16): Likewise.
3195 (vmul_s32): Likewise.
3196 (vmul_f32): Likewise.
3197 (vmul_u8): Likewise.
3198 (vmul_u16): Likewise.
3199 (vmul_u32): Likewise.
3200 (vmul_p8): Likewise.
3201 (vmulq_s8): Likewise.
3202 (vmulq_s16): Likewise.
3203 (vmulq_s32): Likewise.
3204 (vmulq_f32): Likewise.
3205 (vmulq_u8): Likewise.
3206 (vmulq_u16): Likewise.
3207 (vmulq_u32): Likewise.
3208 (vsub_s8): Likewise.
3209 (vsub_s16): Likewise.
3210 (vsub_s32): Likewise.
3211 (vsub_f32): Likewise.
3212 (vsub_u8): Likewise.
3213 (vsub_u16): Likewise.
3214 (vsub_u32): Likewise.
3215 (vsub_s64): Likewise.
3216 (vsub_u64): Likewise.
3217 (vsubq_s8): Likewise.
3218 (vsubq_s16): Likewise.
3219 (vsubq_s32): Likewise.
3220 (vsubq_s64): Likewise.
3221 (vsubq_f32): Likewise.
3222 (vsubq_u8): Likewise.
3223 (vsubq_u16): Likewise.
3224 (vsubq_u32): Likewise.
3225 (vsubq_u64): Likewise.
3226 (vand_s8): Likewise.
3227 (vand_s16): Likewise.
3228 (vand_s32): Likewise.
3229 (vand_u8): Likewise.
3230 (vand_u16): Likewise.
3231 (vand_u32): Likewise.
3232 (vand_s64): Likewise.
3233 (vand_u64): Likewise.
3234 (vandq_s8): Likewise.
3235 (vandq_s16): Likewise.
3236 (vandq_s32): Likewise.
3237 (vandq_s64): Likewise.
3238 (vandq_u8): Likewise.
3239 (vandq_u16): Likewise.
3240 (vandq_u32): Likewise.
3241 (vandq_u64): Likewise.
3242 (vorr_s8): Likewise.
3243 (vorr_s16): Likewise.
3244 (vorr_s32): Likewise.
3245 (vorr_u8): Likewise.
3246 (vorr_u16): Likewise.
3247 (vorr_u32): Likewise.
3248 (vorr_s64): Likewise.
3249 (vorr_u64): Likewise.
3250 (vorrq_s8): Likewise.
3251 (vorrq_s16): Likewise.
3252 (vorrq_s32): Likewise.
3253 (vorrq_s64): Likewise.
3254 (vorrq_u8): Likewise.
3255 (vorrq_u16): Likewise.
3256 (vorrq_u32): Likewise.
3257 (vorrq_u64): Likewise.
3258 (veor_s8): Likewise.
3259 (veor_s16): Likewise.
3260 (veor_s32): Likewise.
3261 (veor_u8): Likewise.
3262 (veor_u16): Likewise.
3263 (veor_u32): Likewise.
3264 (veor_s64): Likewise.
3265 (veor_u64): Likewise.
3266 (veorq_s8): Likewise.
3267 (veorq_s16): Likewise.
3268 (veorq_s32): Likewise.
3269 (veorq_s64): Likewise.
3270 (veorq_u8): Likewise.
3271 (veorq_u16): Likewise.
3272 (veorq_u32): Likewise.
3273 (veorq_u64): Likewise.
3274 (vbic_s8): Likewise.
3275 (vbic_s16): Likewise.
3276 (vbic_s32): Likewise.
3277 (vbic_u8): Likewise.
3278 (vbic_u16): Likewise.
3279 (vbic_u32): Likewise.
3280 (vbic_s64): Likewise.
3281 (vbic_u64): Likewise.
3282 (vbicq_s8): Likewise.
3283 (vbicq_s16): Likewise.
3284 (vbicq_s32): Likewise.
3285 (vbicq_s64): Likewise.
3286 (vbicq_u8): Likewise.
3287 (vbicq_u16): Likewise.
3288 (vbicq_u32): Likewise.
3289 (vbicq_u64): Likewise.
3290 (vorn_s8): Likewise.
3291 (vorn_s16): Likewise.
3292 (vorn_s32): Likewise.
3293 (vorn_u8): Likewise.
3294 (vorn_u16): Likewise.
3295 (vorn_u32): Likewise.
3296 (vorn_s64): Likewise.
3297 (vorn_u64): Likewise.
3298 (vornq_s8): Likewise.
3299 (vornq_s16): Likewise.
3300 (vornq_s32): Likewise.
3301 (vornq_s64): Likewise.
3302 (vornq_u8): Likewise.
3303 (vornq_u16): Likewise.
3304 (vornq_u32): Likewise.
3305 (vornq_u64): Likewise.
3307 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3309 * wide-int.cc (UTItype): Define.
3310 (UDWtype): Define for appropriate W_TYPE_SIZE.
3312 2014-05-08 Marc Glisse <marc.glisse@inria.fr>
3314 PR tree-optimization/59100
3315 * tree-ssa-phiopt.c: Include tree-inline.h.
3316 (neutral_element_p, absorbing_element_p): New functions.
3317 (value_replacement): Handle conditional binary operations with a
3318 neutral or absorbing element.
3320 2014-05-08 Richard Biener <rguenther@suse.de>
3322 * tree-ssa-sccvn.c (vn_get_expr_for): Valueize operands before
3323 folding the expression.
3324 (valueize_expr): Remove.
3325 (visit_reference_op_load): Do not valueize the result of
3327 (simplify_binary_expression): Likewise.
3328 (simplify_unary_expression): Likewise.
3330 2014-05-08 Richard Biener <rguenther@suse.de>
3332 * gimplify.c (gimplify_call_expr): Use saved fnptrtype for
3333 looking at TYPE_ARG_TYPES.
3335 2014-05-08 Richard Biener <rguenther@suse.de>
3337 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove
3338 pointer propagation special-case.
3340 2014-05-08 Bin Cheng <bin.cheng@arm.com>
3342 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
3343 core part of address expressions.
3345 2014-05-08 Alan Modra <amodra@gmail.com>
3348 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
3349 loads and stores when -mno-strict-align at any alignment.
3350 (expand_block_clear): Similarly. Also correct calculation of
3353 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
3356 * tree-complex.c (expand_complex_move): Keep line info when expanding
3358 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
3359 of complex expression. Use new argument to display correct location
3360 for values coming from phi statement.
3361 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
3362 (warn_uninitialized_phi): Pass location of phi argument to
3364 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
3365 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
3367 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
3369 * config/rs6000/predicates.md (indexed_address_mem): New.
3370 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
3371 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
3372 fpstore_ux, fpstore_u.
3373 (sign_extend, indexed, update): New.
3374 (cell_micro): Adjust.
3375 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
3376 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
3377 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
3378 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
3379 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
3380 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
3381 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
3382 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
3383 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
3384 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
3385 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
3386 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
3387 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
3388 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
3389 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
3391 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
3392 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
3393 *vsx_extract_<mode>_store): Adjust.
3394 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
3395 is_cracked_insn, insn_must_be_first_in_group,
3396 insn_must_be_last_in_group): Adjust.
3398 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
3400 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
3401 ppc440-fpstore): Adjust.
3402 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
3403 ppc476-fpstore): Adjust.
3404 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
3405 ppc601-fpstore): Adjust.
3406 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
3408 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
3410 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
3411 ppc7450-fpstore): Adjust.
3412 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
3413 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
3414 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
3416 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
3417 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
3418 cell-fpstore, cell-fpstore-update): Adjust.
3419 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
3420 ppce300c3_store, ppce300c3_fpstore): Adjust.
3421 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
3422 e500mc_fpstore): Adjust.
3423 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
3424 e500mc64_store, e500mc64_fpstore): Adjust.
3425 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
3426 e5500_fpstore): Adjust.
3427 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
3428 e6500_fpstore): Adjust.
3429 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
3431 * config/rs6000/power4.md (power4-load, power4-load-ext,
3432 power4-load-ext-update, power4-load-ext-update-indexed,
3433 power4-load-update-indexed, power4-load-update, power4-fpload,
3434 power4-fpload-update, power4-store, power4-store-update,
3435 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
3437 * config/rs6000/power5.md (power5-load, power5-load-ext,
3438 power5-load-ext-update, power5-load-ext-update-indexed,
3439 power5-load-update-indexed, power5-load-update, power5-fpload,
3440 power5-fpload-update, power5-store, power5-store-update,
3441 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
3443 * config/rs6000/power6.md (power6-load, power6-load-ext,
3444 power6-load-update, power6-load-update-indexed,
3445 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
3446 power6-fpload-update, power6-store, power6-store-update,
3447 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
3449 * config/rs6000/power7.md (power7-load, power7-load-ext,
3450 power7-load-update, power7-load-update-indexed,
3451 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
3452 power7-fpload-update, power7-store, power7-store-update,
3453 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
3455 * config/rs6000/power8.md (power8-load, power8-load-update,
3456 power8-load-ext, power8-load-ext-update, power8-fpload,
3457 power8-fpload-update, power8-store, power8-store-update-indexed,
3458 power8-fpstore, power8-fpstore-update): Adjust.
3459 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
3461 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
3462 titan_lsu_store, titan_lsu_fpstore): Adjust.
3463 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
3465 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
3468 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
3469 unrolled byte insns. Emit address increments after move insns.
3471 2014-05-07 David Malcolm <dmalcolm@redhat.com>
3473 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
3474 const_gimple, rather than a gimple.
3475 (gimple_call_builtin_p): Likewise, for the three variants.
3477 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
3478 (gimple_call_builtin_p): Likewise, for the three variants.
3480 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
3482 PR tree-optimization/61095
3483 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
3485 2014-05-07 Richard Biener <rguenther@suse.de>
3487 PR tree-optimization/61034
3488 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
3489 (maybe_skip_until): Use translate to take into account
3490 lattices when trying to do disambiguations.
3491 (get_continuation_for_phi_1): Likewise.
3492 (get_continuation_for_phi): Adjust for added translate arguments.
3493 (walk_non_aliased_vuses): Likewise.
3494 * tree-ssa-alias.h (get_continuation_for_phi): Adjust prototype.
3495 (walk_non_aliased_vuses): Likewise.
3496 (call_may_clobber_ref_p_1): Declare.
3497 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also disambiguate against
3498 calls. Stop early if we are only supposed to disambiguate.
3499 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
3501 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
3503 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
3504 Emit an error when the function has arguments.
3506 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
3508 * cfgloop.h (unswitch_loops): Remove.
3509 * doc/passes.texi: Remove references to loop-unswitch.c
3510 * timevar.def (TV_LOOP_UNSWITCH): Remove.
3512 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
3514 * tree-vect-data-refs.c (vect_grouped_load_supported): New
3515 check for loads group of length 3.
3516 (vect_permute_load_chain): New permutations for loads group of
3518 * tree-vect-stmts.c (vect_model_load_cost): Change cost
3519 of vec_perm_shuffle for the new permutations.
3521 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
3523 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
3524 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
3525 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
3526 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
3527 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
3528 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
3529 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
3530 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
3532 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
3534 * loop-unswitch.c: Delete.
3536 2014-05-07 Richard Biener <rguenther@suse.de>
3538 * config.gcc: Always set need_64bit_hwint to yes.
3540 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
3542 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
3543 of using optimize_size.
3545 2014-05-06 Mike Stump <mikestump@comcast.net>
3547 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
3549 2014-05-06 Joseph Myers <joseph@codesourcery.com>
3551 * config/i386/sse.md (*mov<mode>_internal)
3552 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
3553 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
3554 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
3555 (*<code><mode>3, *andnot<mode>3<mask_name>)
3556 (<mask_codefor><code><mode>3<mask_name>): Only consider
3557 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
3559 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
3562 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
3564 * lra-constraints.c (valid_address_p): Move earlier in file.
3565 Add a constraint argument to the address_info version.
3566 (satisfies_memory_constraint_p): New function.
3567 (satisfies_address_constraint_p): Likewise.
3568 (process_alt_operands, curr_insn_transform): Use them.
3569 (process_address): Pass the constraint to valid_address_p when
3570 checking address operands.
3572 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
3574 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
3575 to their respective blocks. Fix inadvertent use of "node".
3577 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
3579 * emit-rtl.c (init_derived_machine_modes): New functionm, split
3581 (init_emit_once): ...here.
3582 * rtl.h (init_derived_machine_modes): Declare.
3583 * toplev.c (do_compile): Call it even if no_backend.
3585 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
3586 Mike Stump <mikestump@comcast.net>
3587 Richard Sandiford <rdsandiford@googlemail.com>
3588 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3590 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
3591 (rtx_equal_for_memref_p): Update comment.
3592 (adjust_offset_for_component_ref): Use wide-int interfaces.
3593 * builtins.c (get_object_alignment_2): Likewise.
3594 (c_readstr): Likewise.
3595 (target_char_cast): Add comment.
3596 (determine_block_size): Use wide-int interfaces.
3597 (expand_builtin_signbit): Likewise.
3598 (fold_builtin_int_roundingfn): Likewise.
3599 (fold_builtin_bitop): Likewise.
3600 (fold_builtin_bswap): Likewise.
3601 (fold_builtin_logarithm): Use signop.
3602 (fold_builtin_pow): Likewise.
3603 (fold_builtin_memory_op): Use wide-int interfaces.
3604 (fold_builtin_object_size): Likewise.
3605 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
3606 nb_iterations_estimate.
3607 (record_niter_bound): Use wide-int interfaces.
3608 (get_estimated_loop_iterations_int): Likewise.
3609 (get_estimated_loop_iterations): Likewise.
3610 (get_max_loop_iterations): Likewise.
3611 * cfgloop.h: Include wide-int.h.
3612 (struct nb_iter_bound): Change bound to widest_int.
3613 (struct loop): Change nb_iterations_upper_bound and
3614 nb_iterations_estimate to widest_int.
3615 (record_niter_bound): Switch to use widest_int.
3616 (get_estimated_loop_iterations): Likewise.
3617 (get_max_loop_iterations): Likewise.
3618 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
3619 update for wide-int.
3620 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
3621 * combine.c (try_combine): Likewise.
3622 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
3623 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
3625 (aarch64_float_const_representable_p): Likewise.
3626 * config/arc/arc.c: Include wide-int.h.
3627 (arc_can_use_doloop_p): Use wide-int interfaces.
3628 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
3629 (vfp3_const_double_index): Likewise.
3630 * config/avr/avr.c (avr_out_round): Likewise.
3631 (avr_fold_builtin): Likewise.
3632 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
3633 (bfin_can_use_doloop_p): Likewise.
3634 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
3635 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
3636 * config/i386/i386.c: Include wide-int.h.
3637 (ix86_data_alignment): Use wide-int interfaces.
3638 (ix86_local_alignment): Likewise.
3639 (ix86_emit_swsqrtsf): Update real_from_integer.
3640 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
3641 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
3642 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
3643 (zero_constant): Likewise.
3644 (input_operand): Likewise.
3645 (splat_input_operand): Likewise.
3646 (non_logical_cint_operand): Change const_double to const_wide_int.
3647 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
3648 (easy_altivec_constant): Remove comment.
3649 (paired_expand_vector_init): Use CONSTANT_P.
3650 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
3651 (rs6000_emit_move): Update checks.
3652 (rs6000_aggregate_candidate): Use wide-int interfaces.
3653 (rs6000_expand_ternop_builtin): Likewise.
3654 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
3655 (rs6000_assemble_integer): Likewise.
3656 (rs6000_hash_constant): Likewise.
3657 (output_toc): Likewise.
3658 (rs6000_rtx_costs): Likewise.
3659 (rs6000_emit_swrsqrt); Update call to real_from_integer.
3660 * config/rs6000/rs6000-c.c: Include wide-int.h.
3661 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
3662 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
3663 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
3664 Handle CONST_WIDE_INT.
3665 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
3666 Use tree_fits_uhwi_p.
3667 * config/sparc/sparc.c: Include wide-int.h.
3668 (sparc_fold_builtin): Use wide-int interfaces.
3669 * config/vax/vax.c: Include wide-int.h.
3670 (vax_float_literal): Use real_from_integer.
3671 * coretypes.h (struct hwivec_def): New.
3673 (const_hwivec): New.
3674 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
3675 (equiv_constant): Handle CONST_WIDE_INT.
3676 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
3677 (cselib_hash_rtx): Handle CONST_WIDE_INT.
3678 * dbxout.c (stabstr_U): Use wide-int interfaces.
3679 (dbxout_type): Update to use cst_fits_shwi_p.
3680 * defaults.h (LOG2_BITS_PER_UNIT): Define.
3681 (TARGET_SUPPORTS_WIDE_INT): Add default.
3682 * dfp.c: Include wide-int.h.
3683 (decimal_real_to_integer2): Use wide-int interfaces and rename to
3684 decimal_real_to_integer.
3685 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
3686 decimal_real_to_integer.
3687 * doc/generic.texi (Constant expressions): Update for wide_int.
3688 * doc/rtl.texi (const_double): Likewise.
3689 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
3690 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
3691 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
3692 (REAL_VALUE_FROM_INT): Remove.
3693 (TARGET_SUPPORTS_WIDE_INT): New.
3694 * doc/tm.texi: Regenerate.
3695 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
3696 * double-int.h: Include wide-int.h.
3697 (struct wi::int_traits): New.
3698 * dwarf2out.c (get_full_len): New.
3699 (dw_val_equal_p): Add case dw_val_class_wide_int.
3700 (size_of_loc_descr): Likewise.
3701 (output_loc_operands): Likewise.
3702 (insert_double): Remove.
3703 (insert_wide_int): New.
3705 (print_die): Add case dw_val_class_wide_int.
3706 (attr_checksum): Likewise.
3707 (attr_checksum_ordered): Likewise.
3708 (same_dw_val_p): Likewise.
3709 (size_of_die): Likewise.
3710 (value_format): Likewise.
3711 (output_die): Likewise.
3712 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
3714 (clz_loc_descriptor): Use wide-int interfaces.
3715 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
3716 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
3717 (round_up_to_align): Use wide-int interfaces.
3718 (field_byte_offset): Likewise.
3719 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
3720 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
3721 CONST_DOUBLE handling. Use wide-int interfaces.
3722 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
3723 (gen_enumeration_type_die): Use add_AT_wide.
3724 (hash_loc_operands): Add case dw_val_class_wide_int.
3725 (compare_loc_operands): Likewise.
3726 * dwarf2out.h: Include wide-int.h.
3727 (wide_int_ptr): New.
3728 (enum dw_val_class): Add dw_val_class_wide_int.
3729 (struct dw_val_struct): Add val_wide.
3730 * emit-rtl.c (const_wide_int_htab): New.
3731 (const_wide_int_htab_hash): New.
3732 (const_wide_int_htab_eq): New.
3733 (lookup_const_wide_int): New.
3734 (const_double_htab_hash): Use wide-int interfaces.
3735 (const_double_htab_eq): Likewise.
3736 (rtx_to_double_int): Conditionally compile for wide-int.
3737 (immed_double_int_const): Rename to immed_wide_int_const and
3738 update for wide-int.
3739 (immed_double_const): Conditionally compile for wide-int.
3740 (init_emit_once): Use wide-int interfaces.
3741 * explow.c (plus_constant): Likewise.
3742 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
3743 (lshift_value): Use wide-int interfaces.
3744 (expand_mult): Likewise.
3745 (choose_multiplier): Likewise.
3746 (expand_smod_pow2): Likewise.
3747 (make_tree): Likewise.
3748 * expr.c (convert_modes): Consolidate handling of constants.
3749 Use wide-int interfaces.
3750 (emit_group_load_1): Add note.
3751 (store_expr): Update comment.
3752 (get_inner_reference): Use wide-int interfaces.
3753 (expand_constructor): Update comment.
3754 (expand_expr_real_2): Use wide-int interfaces.
3755 (expand_expr_real_1): Likewise.
3756 (reduce_to_bit_field_precision): Likewise.
3757 (const_vector_from_tree): Likewise.
3758 * final.c: Include wide-int-print.h.
3759 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
3760 * fixed-value.c: Include wide-int.h.
3761 (fixed_from_string): Use wide-int interfaces.
3762 (fixed_to_decimal): Likewise.
3763 (fixed_convert_from_real): Likewise.
3764 (real_convert_from_fixed): Likewise.
3765 * fold-const.h (mem_ref_offset): Return an offset_int.
3766 (div_if_zero_remainder): Remove code parameter.
3767 * fold-const.c (div_if_zero_remainder): Remove code parameter.
3768 Use wide-int interfaces.
3769 (may_negate_without_overflow_p): Use wide-int interfaces.
3770 (negate_expr_p): Likewise.
3771 (fold_negate_expr): Likewise.
3772 (int_const_binop_1): Likewise.
3773 (const_binop): Likewise.
3774 (fold_convert_const_int_from_int): Likewise.
3775 (fold_convert_const_int_from_real): Likewise.
3776 (fold_convert_const_int_from_fixed): Likewise.
3777 (fold_convert_const_fixed_from_int): Likewise.
3778 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
3779 (sign_bit_p): Use wide-int interfaces.
3780 (make_range_step): Likewise.
3781 (build_range_check): Likewise. Pass an integer of the correct type
3782 instead of using integer_one_node.
3783 (range_predecessor): Pass an integer of the correct type instead
3784 of using integer_one_node.
3785 (range_successor): Likewise.
3786 (merge_ranges): Likewise.
3787 (unextend): Use wide-int interfaces.
3788 (extract_muldiv_1): Likewise.
3789 (fold_div_compare): Likewise.
3790 (fold_single_bit_test): Likewise.
3791 (fold_sign_changed_comparison): Likewise.
3792 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
3793 (fold_plusminus_mult_expr): Use wide-int interfaces.
3794 (native_encode_int): Likewise.
3795 (native_interpret_int): Likewise.
3796 (fold_unary_loc): Likewise.
3797 (pointer_may_wrap_p): Likewise.
3798 (size_low_cst): Likewise.
3799 (mask_with_tz): Likewise.
3800 (fold_binary_loc): Likewise.
3801 (fold_ternary_loc): Likewise.
3802 (multiple_of_p): Likewise.
3803 (tree_call_nonnegative_warnv_p): Update calls to
3804 tree_int_cst_min_precision and real_from_integer.
3805 (fold_negate_const): Use wide-int interfaces.
3806 (fold_abs_const): Likewise.
3807 (fold_relational_const): Use tree_int_cst_lt.
3808 (round_up_loc): Use wide-int interfaces.
3809 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
3810 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
3811 * gengtype.c: Remove include of double-int.h.
3812 (do_typedef): Use wide-int interfaces.
3813 (open_base_files): Add wide-int.h.
3814 (main): Add offset_int and widest_int typedefs.
3815 * gengtype-lex.l: Handle "^".
3816 (CXX_KEYWORD): Add "static".
3817 * gengtype-parse.c (require3): New.
3818 (require_template_declaration): Handle constant template arguments
3819 and nested templates.
3820 * gengtype-state.c: Don't include "double-int.h".
3821 * genpreds.c (write_one_predicate_function): Update comment.
3822 (write_tm_constrs_h): Add check for hval and lval use in
3824 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
3825 (add_to_sequence): Likewise.
3826 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
3827 and const_double_operand.
3828 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
3830 * gimple-fold.c (get_base_constructor): Likewise.
3831 (fold_array_ctor_reference): Likewise.
3832 (fold_nonarray_ctor_reference): Likewise.
3833 (fold_const_aggregate_ref_1): Likewise.
3834 (gimple_val_nonnegative_real_p): Likewise.
3835 (gimple_fold_indirect_ref): Likewise.
3836 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
3837 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
3838 (struct slsr_cand_d): Change index to be widest_int.
3839 (struct incr_info_d): Change incr to be widest_int.
3840 (alloc_cand_and_find_basis): Use wide-int interfaces.
3841 (slsr_process_phi): Likewise.
3842 (backtrace_base_for_ref): Likewise. Return a widest_int.
3843 (restructure_reference): Take a widest_int instead of a double_int.
3844 (slsr_process_ref): Use wide-int interfaces.
3845 (create_mul_ssa_cand): Likewise.
3846 (create_mul_imm_cand): Likewise.
3847 (create_add_ssa_cand): Likewise.
3848 (create_add_imm_cand): Take a widest_int instead of a double_int.
3849 (slsr_process_add): Use wide-int interfaces.
3850 (slsr_process_cast): Likewise.
3851 (slsr_process_copy): Likewise.
3852 (dump_candidate): Likewise.
3853 (dump_incr_vec): Likewise.
3854 (replace_ref): Likewise.
3855 (cand_increment): Likewise. Return a widest_int.
3856 (cand_abs_increment): Likewise.
3857 (replace_mult_candidate): Take a widest_int instead of a double_int.
3858 (replace_unconditional_candidate): Use wide-int interfaces.
3859 (incr_vec_index): Take a widest_int instead of a double_int.
3860 (create_add_on_incoming_edge): Likewise.
3861 (create_phi_basis): Use wide-int interfaces.
3862 (replace_conditional_candidate): Likewise.
3863 (record_increment): Take a widest_int instead of a double_int.
3864 (record_phi_increments): Use wide-int interfaces.
3865 (phi_incr_cost): Take a widest_int instead of a double_int.
3866 (lowest_cost_path): Likewise.
3867 (total_savings): Likewise.
3868 (analyze_increments): Use wide-int interfaces.
3869 (ncd_with_phi): Take a widest_int instead of a double_int.
3870 (ncd_of_cand_and_phis): Likewise.
3871 (nearest_common_dominator_for_cands): Likewise.
3872 (insert_initializers): Use wide-int interfaces.
3873 (all_phi_incrs_profitable): Likewise.
3874 (replace_one_candidate): Likewise.
3875 (replace_profitable_candidates): Likewise.
3876 * godump.c: Include wide-int-print.h.
3877 (go_output_typedef): Use wide-int interfaces.
3878 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
3879 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
3880 (build_loop_iteration_domains): Likewise.
3881 * hooks.h: Include wide-int.h rather than double-int.h.
3882 (hook_bool_dint_dint_uint_bool_true): Delete.
3883 (hook_bool_wint_wint_uint_bool_true): Declare.
3884 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
3885 (hook_bool_wint_wint_uint_bool_true): New.
3886 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
3888 (ubsan_expand_si_overflow_mul_check): Likewise.
3889 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
3890 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
3891 (get_ancestor_addr_info): Likewise.
3892 (ipa_modify_call_arguments): Likewise.
3893 * loop-doloop.c (doloop_modify): Likewise.
3894 (doloop_optimize): Likewise.
3895 * loop-iv.c (iv_number_of_iterations): Likewise.
3896 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
3897 (unroll_loop_constant_iterations): Likewise.
3898 (decide_unroll_runtime_iterations): Likewise.
3899 (unroll_loop_runtime_iterations): Likewise.
3900 (decide_peel_simple): Likewise.
3901 (decide_unroll_stupid): Likewise.
3902 * lto-streamer-in.c (streamer_read_wi): Add.
3903 (input_cfg): Use wide-int interfaces.
3904 (lto_input_tree_1): Likewise.
3905 * lto-streamer-out.c (streamer_write_wi): Add.
3906 (hash_tree): Use wide-int interfaces.
3907 (output_cfg): Likewise.
3908 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
3909 (GTFILES): Add wide-int.h and signop.h.
3910 (TAGS): Look for .cc files too.
3911 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
3912 * optabs.c (expand_subword_shift): Likewise.
3913 (expand_doubleword_shift): Likewise.
3914 (expand_absneg_bit): Likewise.
3915 (expand_copysign_absneg): Likewise.
3916 (expand_copysign_bit): Likewise.
3917 * postreload.c (reload_cse_simplify_set): Likewise.
3918 * predict.c (predict_iv_comparison): Likewise.
3919 * pretty-print.h: Include wide-int-print.h.
3921 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
3922 * print-tree.c: Include wide-int-print.h.
3923 (print_node_brief): Use wide-int interfaces.
3924 (print_node): Likewise.
3925 * read-rtl.c (validate_const_wide_int): New.
3926 (read_rtx_code): Add CONST_WIDE_INT case.
3927 * real.c: Include wide-int.h.
3928 (real_to_integer2): Delete.
3929 (real_to_integer): New function, returning a wide_int.
3930 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
3931 (ten_to_ptwo): Update call to real_from_integer.
3932 (real_digit): Likewise.
3933 * real.h: Include signop.h, wide-int.h and insn-modes.h.
3934 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
3935 (REAL_VALUE_TO_INT): Delete.
3936 (real_to_integer): Declare a wide-int form.
3937 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
3938 * recog.c (const_int_operand): Improve comment.
3939 (const_scalar_int_operand): New.
3940 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
3941 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
3942 (split_double): Likewise.
3943 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
3944 (rtx_size): Likewise.
3945 (rtx_alloc_stat_v): New.
3946 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
3947 (cwi_output_hex): New.
3948 (iterative_hash_rtx): Handle CONST_WIDE_INT.
3949 (cwi_check_failed_bounds): New.
3950 * rtl.def (CONST_WIDE_INT): New.
3951 * rtl.h: Include <utility> and wide-int.h.
3952 (struct hwivec_def): New.
3953 (CWI_GET_NUM_ELEM): New.
3954 (CWI_PUT_NUM_ELEM): New.
3955 (struct rtx_def): Add num_elem and hwiv.
3956 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
3957 (CASE_CONST_UNIQUE): Likewise.
3958 (CASE_CONST_ANY): Likewise.
3959 (CONST_SCALAR_INT_P): Likewise.
3960 (CONST_WIDE_INT_P): New.
3962 (HWIVEC_CHECK): New.
3963 (cwi_check_failed_bounds): New.
3965 (HWIVEC_CHECK): New.
3966 (CONST_WIDE_INT_VEC) New.
3967 (CONST_WIDE_INT_NUNITS) New.
3968 (CONST_WIDE_INT_ELT) New.
3969 (rtx_mode_t): New type.
3970 (wi::int_traits <rtx_mode_t>): New.
3972 (wi::min_value): New.
3973 (wi::max_value): New.
3975 (const_wide_int_alloc): New.
3976 (immed_wide_int_const): New.
3977 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
3978 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
3979 * signop.h: New file.
3980 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
3981 (simplify_const_unary_operation): Use wide-int interfaces.
3982 (simplify_binary_operation_1): Likewise.
3983 (simplify_const_binary_operation): Likewise.
3984 (simplify_const_relational_operation): Likewise.
3985 (simplify_immed_subreg): Likewise.
3986 * stmt.c (expand_case): Likewise.
3987 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
3988 signop rather than a bool.
3989 * stor-layout.c (layout_type): Use wide-int interfaces.
3990 (initialize_sizetypes): Update calls to
3991 set_min_and_max_values_for_integral_type.
3992 (set_min_and_max_values_for_integral_type): Take a signop rather
3993 than a bool. Use wide-int interfaces.
3994 (fixup_signed_type): Update accordingly. Remove
3995 HOST_BITS_PER_DOUBLE_INT limit.
3996 (fixup_unsigned_type): Likewise.
3997 * system.h (STATIC_CONSTANT_P): New.
3998 (STATIC_ASSERT): New.
3999 * target.def (can_use_doloop_p): Take widest_ints rather than
4001 * target.h: Include wide-int.h rather than double-int.h.
4002 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
4004 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
4005 rather than INT_CST_LT_UNSIGNED.
4006 (can_use_doloop_if_innermost): Take widest_ints rather than
4008 * tree-affine.c: Include wide-int-print.h.
4009 (double_int_ext_for_comb): Delete.
4010 (wide_int_ext_for_comb): New.
4011 (aff_combination_zero): Use wide-int interfaces.
4012 (aff_combination_const): Take a widest_int instead of a double_int.
4013 (aff_combination_elt): Use wide-int interfaces.
4014 (aff_combination_scale): Take a widest_int instead of a double_int.
4015 (aff_combination_add_elt): Likewise.
4016 (aff_combination_add_cst): Likewise.
4017 (aff_combination_add): Use wide-int interfaces.
4018 (aff_combination_convert): Likewise.
4019 (tree_to_aff_combination): Likewise.
4020 (add_elt_to_tree): Take a widest_int instead of a double_int.
4021 (aff_combination_to_tree): Use wide-int interfaces.
4022 (aff_combination_remove_elt): Likewise.
4023 (aff_combination_add_product): Take a widest_int instead of
4025 (aff_combination_mult): Use wide-int interfaces.
4026 (aff_combination_expand): Likewise.
4027 (double_int_constant_multiple_p): Delete.
4028 (wide_int_constant_multiple_p): New.
4029 (aff_combination_constant_multiple_p): Take a widest_int pointer
4030 instead of a double_int pointer.
4031 (print_aff): Use wide-int interfaces.
4032 (get_inner_reference_aff): Take a widest_int pointer
4033 instead of a double_int pointer.
4034 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
4035 * tree-affine.h: Include wide-int.h.
4036 (struct aff_comb_elt): Change type of coef to widest_int.
4037 (struct affine_tree_combination): Change type of offset to widest_int.
4038 (double_int_ext_for_comb): Delete.
4039 (wide_int_ext_for_comb): New.
4040 (aff_combination_const): Use widest_int instead of double_int.
4041 (aff_combination_scale): Likewise.
4042 (aff_combination_add_elt): Likewise.
4043 (aff_combination_constant_multiple_p): Likewise.
4044 (get_inner_reference_aff): Likewise.
4045 (aff_comb_cannot_overlap_p): Likewise.
4046 (aff_combination_zero_p): Use wide-int interfaces.
4047 * tree.c: Include tree.h.
4048 (init_ttree): Use make_int_cst.
4049 (tree_code_size): Removed code for INTEGER_CST case.
4050 (tree_size): Add INTEGER_CST case.
4051 (make_node_stat): Update comment.
4052 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
4053 (build_int_cst_type): Use wide-int interfaces.
4054 (double_int_to_tree): Likewise.
4055 (double_int_fits_to_tree_p): Delete.
4056 (force_fit_type_double): Delete.
4057 (force_fit_type): New.
4058 (int_cst_hash_hash): Use wide-int interfaces.
4059 (int_cst_hash_eq): Likewise.
4060 (build_int_cst_wide): Delete.
4061 (wide_int_to_tree): New.
4062 (cache_integer_cst): Use wide-int interfaces.
4063 (build_low_bits_mask): Likewise.
4064 (cst_and_fits_in_hwi): Likewise.
4065 (real_value_from_int_cst): Likewise.
4066 (make_int_cst_stat): New.
4067 (integer_zerop): Use wide_int interfaces.
4068 (integer_onep): Likewise.
4069 (integer_all_onesp): Likewise.
4070 (integer_pow2p): Likewise.
4071 (integer_nonzerop): Likewise.
4072 (tree_log2): Likewise.
4073 (tree_floor_log2): Likewise.
4074 (tree_ctz): Likewise.
4075 (int_size_in_bytes): Likewise.
4076 (mem_ref_offset): Return an offset_int rather than a double_int.
4077 (build_type_attribute_qual_variant): Use wide_int interfaces.
4078 (type_hash_eq): Likewise
4079 (tree_int_cst_equal): Likewise.
4080 (tree_int_cst_lt): Delete.
4081 (tree_int_cst_compare): Likewise.
4082 (tree_fits_shwi_p): Use wide_int interfaces.
4083 (tree_fits_uhwi_p): Likewise.
4084 (tree_int_cst_sign_bit): Likewise.
4085 (tree_int_cst_sgn): Likewise.
4086 (tree_int_cst_min_precision): Take a signop rather than a bool.
4087 (simple_cst_equal): Use wide_int interfaces.
4088 (compare_tree_int): Likewise.
4089 (iterative_hash_expr): Likewise.
4090 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
4092 (get_type_static_bounds): Use wide_int interfaces.
4093 (tree_int_cst_elt_check_failed): New.
4094 (build_common_tree_nodes): Reordered to set prec before filling in
4096 (int_cst_value): Check cst_and_fits_in_hwi.
4097 (widest_int_cst_value): Use wide_int interfaces.
4098 (upper_bound_in_type): Likewise.
4099 (lower_bound_in_type): Likewise.
4100 (num_ending_zeros): Likewise.
4101 (drop_tree_overflow): Likewise.
4102 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
4103 (gen_conditions_for_pow_cst_base): Likewise.
4104 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
4105 (group_case_labels_stmt): Use wide-int interfaces.
4106 (verify_gimple_assign_binary): Likewise.
4107 (print_loop): Likewise.
4108 * tree-chrec.c (tree_fold_binomial): Likewise.
4109 * tree-core.h (struct tree_base): Add int_length.
4110 (struct tree_int_cst): Change rep of value.
4111 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
4112 (dr_may_alias_p): Likewise.
4113 (max_stmt_executions_tree): Likewise.
4114 * tree.def (INTEGER_CST): Update comment.
4115 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
4116 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
4117 * tree-dump.c: Include wide-int.h and wide-int-print.h.
4118 (dequeue_and_dump): Use wide-int interfaces.
4119 * tree.h: Include wide-int.h.
4120 (NULL_TREE): Moved to earlier loc in file.
4121 (TREE_INT_CST_ELT_CHECK): New.
4122 (tree_int_cst_elt_check_failed): New.
4124 (TREE_INT_CST): Delete.
4125 (TREE_INT_CST_LOW): Use wide-int interfaces.
4126 (TREE_INT_CST_HIGH): Delete.
4127 (TREE_INT_CST_NUNITS): New.
4128 (TREE_INT_CST_EXT_NUNITS): Likewise.
4129 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
4130 (TREE_INT_CST_ELT): Likewise.
4131 (INT_CST_LT): Delete.
4132 (tree_int_cst_elt_check): New (two forms).
4133 (type_code_size): Update comment.
4134 (make_int_cst_stat, make_int_cst): New.
4135 (tree_to_double_int): Delete.
4136 (double_int_fits_to_tree_p): Delete.
4137 (force_fit_type_double): Delete.
4138 (build_int_cstu): Replace with out-of-line function.
4139 (build_int_cst_wide): Delete.
4140 (tree_int_cst_lt): Define inline.
4141 (tree_int_cst_le): New.
4142 (tree_int_cst_compare): Define inline.
4143 (tree_int_cst_min_precision): Take a signop rather than a bool.
4144 (wi::int_traits <const_tree>): New.
4145 (wi::int_traits <tree>): New.
4146 (wi::extended_tree): New.
4147 (wi::int_traits <wi::extended_tree>): New.
4148 (wi::to_widest): New.
4149 (wi::to_offset): New.
4150 (wi::fits_to_tree_p): New.
4151 (wi::min_value): New.
4152 (wi::max_value): New.
4153 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
4154 (copy_tree_body_r): Likewise.
4155 * tree-object-size.c (compute_object_offset): Likewise.
4156 (addr_object_size): Likewise.
4157 * tree-predcom.c: Include wide-int-print.h.
4158 (struct dref_d): Change type of offset to widest_int.
4159 (dump_dref): Call wide-int printer.
4160 (aff_combination_dr_offset): Use wide-int interfaces.
4161 (determine_offset): Take a widest_int pointer rather than a
4163 (split_data_refs_to_components): Use wide-int interfaces.
4164 (suitable_component_p): Likewise.
4165 (order_drefs): Likewise.
4166 (add_ref_to_chain): Likewise.
4167 (valid_initializer_p): Likewise.
4168 (determine_roots_comp): Likewise.
4169 * tree-pretty-print.c: Include wide-int-print.h.
4170 (dump_generic_node): Use wide-int interfaces.
4171 * tree-sra.c (sra_ipa_modify_expr): Likewise.
4172 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
4173 (move_fixed_address_to_symbol): Likewise.
4174 (move_hint_to_base): Likewise.
4175 (move_pointer_to_base): Likewise.
4176 (move_variant_to_index): Likewise.
4177 (most_expensive_mult_to_index): Likewise.
4178 (addr_to_parts): Likewise.
4179 (copy_ref_info): Likewise.
4180 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
4181 (indirect_refs_may_alias_p): Likewise.
4182 (stmt_kills_ref_p_1): Likewise.
4183 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
4184 * tree-ssa-ccp.c: Update comment at top of file. Include
4186 (struct prop_value_d): Change type of mask to widest_int.
4187 (extend_mask): New function.
4188 (dump_lattice_value): Use wide-int interfaces.
4189 (get_default_value): Likewise.
4190 (set_constant_value): Likewise.
4191 (set_value_varying): Likewise.
4192 (valid_lattice_transition): Likewise.
4193 (set_lattice_value): Likewise.
4194 (value_to_double_int): Delete.
4195 (value_to_wide_int): New.
4196 (get_value_from_alignment): Use wide-int interfaces.
4197 (get_value_for_expr): Likewise.
4198 (do_dbg_cnt): Likewise.
4199 (ccp_finalize): Likewise.
4200 (ccp_lattice_meet): Likewise.
4201 (bit_value_unop_1): Use widest_ints rather than double_ints.
4202 (bit_value_binop_1): Likewise.
4203 (bit_value_unop): Use wide-int interfaces.
4204 (bit_value_binop): Likewise.
4205 (bit_value_assume_aligned): Likewise.
4206 (evaluate_stmt): Likewise.
4207 (ccp_fold_stmt): Likewise.
4208 (visit_cond_stmt): Likewise.
4209 (ccp_visit_stmt): Likewise.
4210 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
4211 (constant_pointer_difference): Likewise.
4212 (associate_pointerplus): Likewise.
4213 (combine_conversions): Likewise.
4214 * tree-ssa-loop.h: Include wide-int.h.
4215 (struct tree_niter_desc): Change type of max to widest_int.
4216 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
4217 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
4218 (remove_redundant_iv_tests): Likewise.
4219 (canonicalize_loop_induction_variables): Likewise.
4220 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
4221 (constant_multiple_of): Take a widest_int pointer instead of
4222 a double_int pointer.
4223 (get_computation_aff): Use wide-int interfaces.
4224 (ptr_difference_cost): Likewise.
4225 (difference_cost): Likewise.
4226 (get_loop_invariant_expr_id): Likewise.
4227 (get_computation_cost_at): Likewise.
4228 (iv_elimination_compare_lt): Likewise.
4229 (may_eliminate_iv): Likewise.
4230 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
4231 instead of double_int.
4232 (max_loop_iterations): Likewise.
4233 (max_stmt_executions): Likewise.
4234 (estimated_stmt_executions): Likewise.
4235 * tree-ssa-loop-niter.c: Include wide-int-print.h.
4236 (split_to_var_and_offset): Use wide-int interfaces.
4237 (determine_value_range): Likewise.
4238 (bound_difference_of_offsetted_base): Likewise.
4239 (bounds_add): Take a widest_int instead of a double_int.
4240 (number_of_iterations_ne_max): Use wide-int interfaces.
4241 (number_of_iterations_ne): Likewise.
4242 (number_of_iterations_lt_to_ne): Likewise.
4243 (assert_loop_rolls_lt): Likewise.
4244 (number_of_iterations_lt): Likewise.
4245 (number_of_iterations_le): Likewise.
4246 (number_of_iterations_cond): Likewise.
4247 (number_of_iterations_exit): Likewise.
4248 (finite_loop_p): Likewise.
4249 (derive_constant_upper_bound_assign): Likewise.
4250 (derive_constant_upper_bound): Return a widest_int.
4251 (derive_constant_upper_bound_ops): Likewise.
4252 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
4253 (record_estimate): Take a widest_int rather than a double_int.
4254 (record_nonwrapping_iv): Use wide-int interfaces.
4255 (double_int_cmp): Delete.
4256 (wide_int_cmp): New.
4257 (bound_index): Take a widest_int rather than a double_int.
4258 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
4259 (maybe_lower_iteration_bound): Likewise.
4260 (estimate_numbers_of_iterations_loop): Likewise.
4261 (estimated_loop_iterations): Take a widest_int pointer than than
4262 a double_int pointer.
4263 (estimated_loop_iterations_int): Use wide-int interfaces.
4264 (max_loop_iterations): Take a widest_int pointer than than
4265 a double_int pointer.
4266 (max_loop_iterations_int): Use wide-int interfaces.
4267 (max_stmt_executions): Take a widest_int pointer than than
4268 a double_int pointer.
4269 (estimated_stmt_executions): Likewise.
4270 (n_of_executions_at_most): Use wide-int interfaces.
4271 (scev_probably_wraps_p): Likewise.
4272 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
4274 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
4276 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
4277 double_ints. Adjust for trailing_wide_ints <3> representation.
4278 (set_nonzero_bits): Likewise.
4279 (get_range_info): Return wide_ints rather than double_ints.
4280 Adjust for trailing_wide_ints <3> representation.
4281 (get_nonzero_bits): Likewise.
4282 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
4284 * tree-ssanames.h (struct range_info_def): Replace min, max and
4285 nonzero_bits with a trailing_wide_ints <3>.
4286 (set_range_info): Use wide_int_refs rather than double_ints.
4287 (set_nonzero_bits): Likewise.
4288 (get_range_info): Return wide_ints rather than double_ints.
4289 (get_nonzero_bits): Likewise.
4290 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
4291 * tree-ssa-pre.c (phi_translate_1): Likewise.
4292 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
4293 (acceptable_pow_call): Likewise.
4294 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
4296 (vn_reference_fold_indirect): Likewise.
4297 (vn_reference_maybe_forwprop_address): Likewise.
4298 (valueize_refs_1): Likewise.
4299 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
4300 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
4301 tree_int_cst_lt and tree_int_cst_le.
4302 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
4304 (streamer_alloc_tree): Likewise.
4305 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
4306 (streamer_write_tree_header): Likewise.
4307 (streamer_write_integer_cst): Likewise.
4308 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
4309 (build_constructors): Likewise.
4310 (array_value_type): Likewise.
4311 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
4312 (vect_check_gather): Likewise.
4313 * tree-vect-generic.c (build_replicated_const): Likewise.
4314 (expand_vector_divmod): Likewise.
4315 * tree-vect-loop.c (vect_transform_loop): Likewise.
4316 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
4317 (vect_do_peeling_for_alignment): Likewise.
4318 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
4319 * tree-vrp.c: Include wide-int.h.
4320 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
4321 (extract_range_from_assert): Use wide-int interfaces.
4322 (vrp_int_const_binop): Likewise.
4323 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
4324 double_int pointers.
4325 (ranges_from_anti_range): Use wide-int interfaces.
4326 (quad_int_cmp): Delete.
4327 (quad_int_pair_sort): Likewise.
4328 (extract_range_from_binary_expr_1): Use wide-int interfaces.
4329 (extract_range_from_unary_expr_1): Likewise.
4330 (adjust_range_with_scev): Likewise.
4331 (masked_increment): Take and return wide_ints rather than double_ints.
4332 (register_edge_assert_for_2): Use wide-int interfaces.
4333 (check_array_ref): Likewise.
4334 (search_for_addr_array): Likewise.
4335 (maybe_set_nonzero_bits): Likewise.
4336 (union_ranges): Pass an integer of the correct type instead of
4337 using integer_one_node.
4338 (intersect_ranges): Likewise.
4339 (simplify_truth_ops_using_ranges): Likewise.
4340 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
4341 (range_fits_type_p): Likewise.
4342 (simplify_cond_using_ranges): Likewise. Take a signop rather than
4344 (simplify_conversion_using_ranges): Use wide-int interfaces.
4345 (simplify_float_conversion_using_ranges): Likewise.
4346 (vrp_finalize): Likewise.
4347 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
4348 (gimple_stringops_transform): Likewise.
4349 * varasm.c (decode_addr_const): Likewise.
4350 (const_hash_1): Likewise.
4351 (const_rtx_hash_1): Likewise
4352 (output_constant): Likewise.
4353 (array_size_for_constructor): Likewise.
4354 (output_constructor_regular_field): Likewise.
4355 (output_constructor_bitfield): Likewise.
4356 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
4357 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
4359 * gencheck.c: Define BITS_PER_UNIT.
4362 * wide-int-print.cc: New.
4363 * wide-int-print.h: New.
4365 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4367 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
4369 2014-05-06 Richard Biener <rguenther@suse.de>
4371 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
4372 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
4373 (TODO_verify_all): Adjust.
4374 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
4375 TODO_verify_stmts and TODO_verify_rtl_sharing.
4376 * bb-reorder.c: Likewise.
4377 * cfgexpand.c: Likewise.
4378 * cprop.c: Likewise.
4380 * function.c: Likewise.
4381 * fwprop.c: Likewise.
4383 * gimple-ssa-isolate-paths.c: Likewise.
4384 * gimple-ssa-strength-reduction.c: Likewise.
4385 * ipa-split.c: Likewise.
4386 * loop-init.c: Likewise.
4387 * loop-unroll.c: Likewise.
4388 * lower-subreg.c: Likewise.
4389 * modulo-sched.c: Likewise.
4390 * postreload-gcse.c: Likewise.
4391 * predict.c: Likewise.
4392 * recog.c: Likewise.
4393 * sched-rgn.c: Likewise.
4394 * store-motion.c: Likewise.
4395 * tracer.c: Likewise.
4396 * trans-mem.c: Likewise.
4397 * tree-call-cdce.c: Likewise.
4398 * tree-cfg.c: Likewise.
4399 * tree-cfgcleanup.c: Likewise.
4400 * tree-complex.c: Likewise.
4401 * tree-eh.c: Likewise.
4402 * tree-emutls.c: Likewise.
4403 * tree-if-conv.c: Likewise.
4404 * tree-into-ssa.c: Likewise.
4405 * tree-loop-distribution.c: Likewise.
4406 * tree-object-size.c: Likewise.
4407 * tree-parloops.c: Likewise.
4408 * tree-pass.h: Likewise.
4409 * tree-sra.c: Likewise.
4410 * tree-ssa-ccp.c: Likewise.
4411 * tree-ssa-copy.c: Likewise.
4412 * tree-ssa-copyrename.c: Likewise.
4413 * tree-ssa-dce.c: Likewise.
4414 * tree-ssa-dom.c: Likewise.
4415 * tree-ssa-dse.c: Likewise.
4416 * tree-ssa-forwprop.c: Likewise.
4417 * tree-ssa-ifcombine.c: Likewise.
4418 * tree-ssa-loop-ch.c: Likewise.
4419 * tree-ssa-loop-ivcanon.c: Likewise.
4420 * tree-ssa-loop.c: Likewise.
4421 * tree-ssa-math-opts.c: Likewise.
4422 * tree-ssa-phiopt.c: Likewise.
4423 * tree-ssa-phiprop.c: Likewise.
4424 * tree-ssa-pre.c: Likewise.
4425 * tree-ssa-reassoc.c: Likewise.
4426 * tree-ssa-sink.c: Likewise.
4427 * tree-ssa-strlen.c: Likewise.
4428 * tree-ssa-tail-merge.c: Likewise.
4429 * tree-ssa-uncprop.c: Likewise.
4430 * tree-switch-conversion.c: Likewise.
4431 * tree-tailcall.c: Likewise.
4432 * tree-vect-generic.c: Likewise.
4433 * tree-vectorizer.c: Likewise.
4434 * tree-vrp.c: Likewise.
4436 * var-tracking.c: Likewise.
4437 * bt-load.c: Likewise.
4438 * cfgcleanup.c: Likewise.
4439 * combine-stack-adj.c: Likewise.
4440 * combine.c: Likewise.
4441 * compare-elim.c: Likewise.
4442 * config/epiphany/resolve-sw-modes.c: Likewise.
4443 * config/i386/i386.c: Likewise.
4444 * config/mips/mips.c: Likewise.
4445 * config/s390/s390.c: Likewise.
4446 * config/sh/sh_treg_combine.cc: Likewise.
4447 * config/sparc/sparc.c: Likewise.
4450 * final.c: Likewise.
4451 * ifcvt.c: Likewise.
4452 * mode-switching.c: Likewise.
4453 * passes.c: Likewise.
4454 * postreload.c: Likewise.
4456 * reg-stack.c: Likewise.
4457 * regcprop.c: Likewise.
4458 * regrename.c: Likewise.
4461 2014-05-06 Richard Biener <rguenther@suse.de>
4464 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
4465 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
4467 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
4470 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
4472 2014-05-05 Radovan Obradovic <robradovic@mips.com>
4473 Tom de Vries <tom@codesourcery.com>
4475 * target.def (call_fusage_contains_non_callee_clobbers): New
4477 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
4479 (@node Miscellaneous Register Hooks): New node.
4480 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
4481 * doc/tm.texi: Regenerate.
4483 2014-05-05 Marek Polacek <polacek@redhat.com>
4486 * opts.c (common_handle_option): Call error_at instead of warning_at.
4488 2014-05-05 Richard Biener <rguenther@suse.de>
4490 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
4491 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
4492 under the TODO_verify_il umbrella.
4494 2014-05-05 Richard Biener <rguenther@suse.de>
4496 * passes.c (execute_function_todo): Move TODO_verify_flow under
4497 the TODO_verify_ul umbrella.
4499 2014-05-05 Richard Biener <rguenther@suse.de>
4502 * fold-const.c (fold_binary_loc): Consistently avoid canonicalizing
4503 X & CST away from a CST that is the mask of a mode.
4505 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4507 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
4508 int argument to enum machine_mode.
4509 (picochip_class_max_nregs): Ditto.
4510 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
4511 (picochip_class_max_nregs): Ditto.
4513 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4515 * target.def: Add new target hook.
4516 * doc/tm.texi: Regenerate.
4517 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
4518 * targhooks.c (default_keep_leaf_when_profiled): New function.
4520 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
4521 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
4523 2014-05-05 Bin Cheng <bin.cheng@arm.com>
4525 PR tree-optimization/60363
4526 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
4527 (copy_phi_args): New parameters. Call get_value_locus_in_path.
4528 (update_destination_phis): New parameter.
4529 (create_edge_and_update_destination_phis): Ditto.
4530 (ssa_fix_duplicate_block_edges): Pass new arguments.
4531 (thread_single_edge): Ditto.
4533 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
4535 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
4536 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
4537 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
4538 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
4539 Use RS6000_BTM_HARD_FLOAT.
4540 (BU_MISC_2): Likewise.
4541 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
4542 RS6000_BTM_HARD_FLOAT.
4543 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
4545 (rs6000_invalid_builtin): Add hard floating builtin support.
4546 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
4547 hard float builtins.
4548 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
4550 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4552 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
4553 Add missing function* argument.
4555 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
4557 * lra-constraints.c (valid_address_p): Move earlier in file.
4558 Add a constraint argument to the address_info version.
4559 (satisfies_memory_constraint_p): New function.
4560 (satisfies_address_constraint_p): Likewise.
4561 (process_alt_operands, curr_insn_transform): Use them.
4562 (process_address): Pass the constraint to valid_address_p when
4563 checking address operands.
4565 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
4567 * config/mips/mips.c (mips_isa_rev): New variable.
4568 (mips_set_architecture): Set it.
4569 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
4571 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
4572 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
4573 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
4574 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
4575 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
4576 conditions in terms of mips_isa_rev.
4577 (mips_isa_rev): Declare.
4579 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4581 * config/sh/sh-mem.cc: Use tabs instead of spaces.
4582 (prob_unlikely, prob_likely): Make variables const.
4584 2014-05-03 Denis Chertykov <chertykov@gmail.com>
4586 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
4588 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4590 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
4592 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4594 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
4595 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
4596 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
4598 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
4599 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
4601 Replace usage of ROUND_REG with sh_round_reg.
4602 Use CEIL instead of ROUND_ADVANCE.
4604 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4607 * config/sh/sh.c: Include stdlib headers before everything else.
4609 2014-05-02 Jakub Jelinek <jakub@redhat.com>
4611 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
4612 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
4613 (gimplify_adjust_omp_clauses): Simd region is never
4614 directly nested in combined parallel. Instead, for linear
4615 with copyin/copyout, if in combined for simd loop, make decl
4616 firstprivate/lastprivate on OMP_FOR.
4617 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
4618 expand_omp_for_static_chunk): When setting endvar, also set
4619 fd->loop.v to the same value.
4621 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
4623 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
4625 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
4627 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
4630 2014-05-02 Marek Polacek <polacek@redhat.com>
4632 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
4634 2014-05-02 Kito Cheng <kito@0xlab.org>
4636 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
4637 to a C expression marco.
4638 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
4639 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
4640 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
4641 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
4642 HONOR_REG_ALLOC_ORDER.
4643 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
4645 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4647 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
4649 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4651 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
4653 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
4655 * tree-if-conv.c (is_cond_scalar_reduction): New function.
4656 (convert_scalar_cond_reduction): Likewise.
4657 (predicate_scalar_phi): Add recognition and transformation
4658 of simple conditioanl reduction to be vectorizable.
4660 2014-05-01 Marek Polacek <polacek@redhat.com>
4663 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
4665 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
4667 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
4668 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
4669 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
4670 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
4671 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
4672 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
4673 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
4674 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
4676 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
4678 * config/arc/arc.opt (mlra): Move comment above option name
4679 to avoid mis-parsing as language options.
4681 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4683 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
4684 * config/sol2.h: ... here.
4685 * config/sol2-10.h: Remove.
4687 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
4688 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
4689 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
4690 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
4691 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
4692 * config/sol2.h: ... here.
4693 (SECTION_NAME_FORMAT): Don't redefine.
4694 (STARTFILE_ARCH32_SPEC): Rename to ...
4695 (STARTFILE_ARCH_SPEC): ... this.
4696 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
4697 * config/sparc/sol2.h: ... here.
4698 (SECTION_NAME_FORMAT): Don't undef.
4699 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
4700 (SUBTARGET_EXTRA_SPECS): Remove.
4701 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
4703 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
4704 (MD_STARTFILE_PREFIX): Remove.
4705 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
4706 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
4707 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
4708 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
4709 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
4710 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
4711 * config/i386/sol2.h: ... here.
4712 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
4713 * config/i386/sol2-bi.h: Remove.
4714 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
4715 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
4717 * config/i386/t-sol2-64: Rename to ...
4718 * config/i386/t-sol2: ... this.
4719 * config/sparc/t-sol2-64: Rename to ...
4720 * config/sparc/t-sol2: ... this.
4722 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
4723 sol2_tm_file_head, sol2_tm_file_tail.
4724 Include ${cpu_type}/sol2.h before sol2.h.
4726 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
4727 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
4728 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
4729 Reflect i386/t-sol2-64 renaming.
4730 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
4731 Reflect sparc/t-sol2-64 renaming.
4733 2014-04-30 Richard Biener <rguenther@suse.de>
4735 * passes.c (execute_function_todo): Move TODO_verify_stmts
4736 and TODO_verify_ssa under the TODO_verify_il umbrella.
4737 * tree-ssa.h (verify_ssa): Adjust prototype.
4738 * tree-ssa.c (verify_ssa): Add parameter to tell whether
4739 we should verify SSA operands.
4740 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
4741 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
4742 whether we should verify whether not throwing stmts have EH info.
4743 * graphite-scop-detection.c (create_sese_edges): Adjust.
4744 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
4745 * tree-eh.c (lower_try_finally_switch): Do not add the
4746 default case label twice.
4748 2014-04-30 Marek Polacek <polacek@redhat.com>
4750 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
4751 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
4752 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
4753 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
4755 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
4757 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
4758 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
4759 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
4760 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
4761 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
4762 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
4763 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
4764 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
4766 2014-04-29 David Malcolm <dmalcolm@redhat.com>
4768 * tree-cfg.c (dump_function_to_file): Dump the return type of
4769 functions, in a line to itself before the function body, mimicking
4770 the layout of a C function.
4772 2014-04-29 Jakub Jelinek <jakub@redhat.com>
4774 PR tree-optimization/60971
4775 * tree-tailcall.c (process_assignment): Reject conversions which
4778 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
4780 * calls.c (initialize_argument_information): Always treat
4781 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
4782 (expand_call): Likewise.
4783 (emit_library_call_calue_1): Likewise.
4784 * expr.c (PUSH_ARGS_REVERSED): Do not define.
4785 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
4788 2014-04-29 Nick Clifton <nickc@redhat.com>
4790 * config/msp430/msp430.md (umulsidi): Fix typo.
4791 (mulhisi3): Enable even inside interrupt handlers.
4792 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
4793 bigger return address pushed in large mode.
4795 2014-04-29 Nick Clifton <nickc@redhat.com>
4797 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
4798 (arc_init_reg_tables): Use a machine_mode enum to iterate over
4800 * config/m32r/m32r.c (init_reg_tables): Likewise.
4801 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
4802 enum to hold the modes.
4804 2014-04-29 Richard Biener <rguenther@suse.de>
4806 * dominance.c (free_dominance_info): Add overload with
4808 (dom_info_state): Likewise.
4809 (dom_info_available_p): Likewise.
4810 * basic-block.h (free_dominance_info, dom_info_state,
4811 dom_info_available_p): Declare overloads.
4812 * passes.c (execute_function_todo): Verify that verifiers
4813 don't change dominator info state. Drop dominator info
4814 for IPA pass invocations.
4815 * cgraph.c (release_function_body): Restore asserts that
4816 dominator information is released.
4818 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
4820 * doc/invoke.texi: Fix typo.
4821 * tree-vrp.c: Fix typos.
4822 * gimple.c (infer_nonnull_range): Reorder operands of an && condition.
4824 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
4826 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
4828 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
4830 * config/aarch64/aarch64-builtins.c
4831 (aarch64_types_storestruct_lane_qualifiers): New.
4832 (TYPES_STORESTRUCT_LANE): Likewise.
4833 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
4834 (st3_lane): Likewise.
4835 (st4_lane): Likewise.
4836 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
4837 (vec_store_lanesci_lane<mode>): Likewise.
4838 (vec_store_lanesxi_lane<mode>): Likewise.
4839 (aarch64_st2_lane<VQ:mode>): Likewise.
4840 (aarch64_st3_lane<VQ:mode>): Likewise.
4841 (aarch64_st4_lane<VQ:mode>): Likewise.
4842 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
4843 * config/aarch64/arm_neon.h
4844 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
4845 use new macro arguments.
4846 (__ST3_LANE_FUNC): Likewise.
4847 (__ST4_LANE_FUNC): Likewise.
4848 * config/aarch64/iterators.md (V_TWO_ELEM): New.
4849 (V_THREE_ELEM): Likewise.
4850 (V_FOUR_ELEM): Likewise.
4852 2014-04-28 David Malcolm <dmalcolm@redhat.com>
4854 * doc/gimple.texi: Replace the description of the now-defunct
4855 union gimple_statement_d with a diagram showing the
4856 gimple_statement_base class hierarchy and its relationships to
4857 the GSS_ and GIMPLE_ enums.
4859 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
4861 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
4862 * config/aarch64/aarch64.c
4863 (aarch64_cannot_change_mode_class): Weaken conditions.
4864 (aarch64_modes_tieable_p): New.
4865 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
4867 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
4869 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4870 (loadsync_<mode>): Change mode.
4871 (load_quadpti, store_quadpti): New.
4872 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4873 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4875 2014-04-28 Martin Jambor <mjambor@suse.cz>
4877 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
4878 same alias type as the original statement.
4879 (subreplacement_assignment_data): New type.
4880 (handle_unscalarized_data_in_subtree): New type of parameter,
4881 generate new memory accesses with same alias type as the original
4883 (load_assign_lhs_subreplacements): Likewise.
4884 (sra_modify_constructor_assign): Generate new memory accesses with
4885 same alias type as the original statement.
4887 2014-04-28 Richard Biener <rguenther@suse.de>
4889 * tree-pass.h (TODO_verify_il): Define.
4890 (TODO_verify_all): Complete properly.
4891 * passes.c (execute_function_todo): Move existing loop-closed
4892 SSA verification under TODO_verify_il.
4893 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
4894 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
4895 Fix tree sharing issue.
4897 2014-04-28 Richard Biener <rguenther@suse.de>
4900 * builtins.def (DEF_C11_BUILTIN): Add.
4901 (BUILT_IN_ALIGNED_ALLOC): Likewise.
4902 * coretypes.h (enum function_class): Add function_c11_misc.
4903 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
4904 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
4905 (call_may_clobber_ref_p_1): Likewise.
4906 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
4907 (mark_all_reaching_defs_necessary_1): Likewise.
4908 (propagate_necessity): Likewise.
4909 (eliminate_unnecessary_stmts): Likewise.
4910 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
4912 2014-04-28 Richard Biener <rguenther@suse.de>
4914 * tree-vrp.c (vrp_var_may_overflow): Remove.
4915 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
4916 with overflow immediately bump to one before that value and
4917 let iteration figure out overflow status.
4919 2014-04-28 Richard Biener <rguenther@suse.de>
4921 * configure.ac: Do valgrind header checks unconditionally.
4922 Add --enable-valgrind-annotations.
4923 * system.h: Guard valgrind header inclusion with
4924 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
4925 * alloc-pool.c (pool_alloc, pool_free): Use
4926 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
4927 to guard possibly dead code.
4928 * config.in: Regenerated.
4929 * configure: Likewise.
4931 2014-04-28 Jeff Law <law@redhat.com>
4933 PR tree-optimization/60902
4934 * tree-ssa-threadedge.c
4935 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
4936 over real defs when invalidating outputs from statements that do not
4937 produce useful outputs for threading.
4939 2014-04-28 Richard Biener <rguenther@suse.de>
4941 PR tree-optimization/60979
4942 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
4943 SCOPs that end in a block with a successor with abnormal
4946 2014-04-28 Richard Biener <rguenther@suse.de>
4948 * tree-pass.h (execute_pass_list): Adjust prototype.
4949 * passes.c (pass_manager::execute_early_local_passes): Adjust.
4950 (do_per_function): Change callback signature, push all actual
4951 work to the callbals.
4952 (do_per_function_toporder): Likewise.
4953 (execute_function_dump): Adjust.
4954 (execute_function_todo): Likewise.
4955 (clear_last_verified): Likewise.
4956 (verify_curr_properties): Likewise.
4957 (update_properties_after_pass): Likewise.
4958 (execute_pass_list_1): Split out from ...
4959 (execute_pass_list): ... here. Adjust.
4960 (execute_ipa_pass_list): Likewise.
4961 * cgraphunit.c (cgraph_add_new_function): Adjust.
4962 (analyze_function): Likewise.
4963 (expand_function): Likewise.
4964 * cgraph.c (release_function_body): Free dominance info
4965 here instead of asserting it was magically freed elsewhere.
4967 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
4969 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
4970 * configure: Regenerate.
4971 * config/sparc/sparc.opt (muser-mode): New option.
4972 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
4974 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
4975 * doc/invoke.texi (SPARC options): Document -muser-mode.
4977 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
4979 * cselib.c (find_slot_memmode): Delete.
4980 (cselib_hasher): Change compare_type to a struct.
4981 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
4983 (preserve_constants_and_equivs): Adjust for new compare_type.
4984 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
4985 (wrap_constant): Delete.
4986 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
4988 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
4990 * doc/install.texi (Building with profile feedback): Remove
4993 2014-04-26 Tom de Vries <tom@codesourcery.com>
4995 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
4998 2014-04-25 Cary Coutant <ccoutant@google.com>
5001 * dwarf2out.c (should_move_die_to_comdat): A type definition
5002 can contain a subprogram definition, but don't move it to a
5004 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
5005 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
5007 (clone_tree_hash): Rename to...
5008 (clone_tree_partial): ...this; change callers. Copy
5009 DW_TAG_subprogram DIEs as declarations.
5010 (copy_decls_walk): Don't copy children of a declaration into a
5013 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
5016 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
5019 2014-04-25 Jiong Wang <jiong.wang@arm.com>
5021 * config/arm/predicates.md (call_insn_operand): Add long_call check.
5022 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
5024 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
5027 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5029 * config/arm/arm.c (arm_cortex_a8_tune): Initialise T16-related fields.
5031 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5033 PR tree-optimization/60930
5034 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
5035 creating a multiply candidate by folding two constant
5036 multiplicands when the result overflows.
5038 2014-04-25 Jakub Jelinek <jakub@redhat.com>
5040 PR tree-optimization/60960
5041 * tree-vect-generic.c (expand_vector_operation): Only call
5042 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
5044 2014-04-25 Tom de Vries <tom@codesourcery.com>
5046 * expr.c (clobber_reg_mode): New function.
5047 * expr.h (clobber_reg): New function.
5049 2014-04-25 Tom de Vries <tom@codesourcery.com>
5051 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
5054 2014-04-25 Radovan Obradovic <robradovic@mips.com>
5055 Tom de Vries <tom@codesourcery.com>
5057 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
5059 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
5060 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
5061 new argument to find_all_hard_reg_sets call.
5063 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5065 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
5066 Use HOST_WIDE_INT_C for mask literal.
5067 (aarch_rev16_shleft_mask_imm_p): Likewise.
5069 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
5072 * config/sparc/sparc.md (ashlsi3_extend): Delete.
5074 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
5076 PR preprocessor/56540
5077 * config/i386/i386-c.c (ix86_target_macros): Define
5078 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
5080 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5082 * configure.ac (tga_func): Remove.
5083 (LIB_TLS_SPEC): Remove.
5084 * configure: Regenerate.
5085 * config.in: Regenerate.
5086 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
5088 2014-04-25 Richard Biener <rguenther@suse.de>
5091 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
5092 call stmt use/clobber sets during stmt walk instead of
5093 walking the possibly incomplete set of caller edges.
5095 2014-04-25 Richard Biener <rguenther@suse.de>
5098 * passes.c (apply_ipa_transforms): Inline into only caller ...
5099 (execute_one_pass): ... here. Properly bring in function
5100 bodies for nodes we want to apply IPA transforms to.
5102 2014-04-24 Cong Hou <congh@google.com>
5104 PR tree-optimization/60896
5105 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
5106 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
5107 (vect_mark_pattern_stmts): Set the def type of all statements in
5108 PATTERN_DEF_SEQ as vect_internal_def.
5110 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
5112 * doc/extend.texi (PowerPC Built-in Functions): Document new
5113 powerpc extended divide, bcd, pack/unpack 128-bit, builtin functions.
5114 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
5116 * config/rs6000/predicates.md (const_0_to_3_operand): New
5117 predicate to match 0..3 integer constants.
5119 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
5120 to support adding miscellaneous builtin functions.
5121 (BU_DFP_MISC_2): Likewise.
5122 (BU_P7_MISC_1): Likewise.
5123 (BU_P7_MISC_2): Likewise.
5124 (BU_P8V_MISC_3): Likewise.
5125 (BU_MISC_1): Likewise.
5126 (BU_MISC_2): Likewise.
5127 (DIVWE): Add extended divide builtin functions.
5130 (DIVWEUO): Likewise.
5134 (DIVDEUO): Likewise.
5135 (DXEX): Add decimal floating-point builtin functions.
5138 (DDEDPDQ): Likewise.
5140 (DENBCDQ): Likewise.
5147 (CDTBCD): Add new BCD builtin functions.
5151 (BCDADD_LT): Likewise.
5152 (BCDADD_EQ): Likewise.
5153 (BCDADD_GT): Likewise.
5154 (BCDADD_OV): Likewise.
5156 (BCDSUB_LT): Likewise.
5157 (BCDSUB_EQ): Likewise.
5158 (BCDSUB_GT): Likewise.
5159 (BCDSUB_OV): Likewise.
5160 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
5161 (UNPACK_TD): Likewise.
5162 (PACK_TF): Likewise.
5163 (UNPACK_TF): Likewise.
5164 (UNPACK_TF_0): Likewise.
5165 (UNPACK_TF_1): Likewise.
5166 (PACK_V1TI): Likewise.
5167 (UNPACK_V1TI): Likewise.
5169 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
5170 support for decimal floating point builtin functions.
5171 (rs6000_expand_ternop_builtin): Add checks for the new builtin
5172 functions that take constant arguments.
5173 (rs6000_invalid_builtin): Add decimal floating point builtin support.
5174 (rs6000_init_builtins): Setup long double, _Decimal64, and
5175 _Decimal128 types for new builtin functions.
5176 (builtin_function_type): Set the unsigned flags appropriately for
5177 the new builtin functions.
5178 (rs6000_opt_masks): Add support for decimal floating point builtin
5181 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
5182 floating point builtin functions.
5183 (RS6000_BTM_COMMON): Likewise.
5184 (RS6000_BTI_long_double): Likewise.
5185 (RS6000_BTI_dfloat64): Likewise.
5186 (RS6000_BTI_dfloat128): Likewise.
5187 (long_double_type_internal_node): Likewise.
5188 (dfloat64_type_internal_node): Likewise.
5189 (dfloat128_type_internal_node): Likewise.
5191 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
5192 2.07 bcd arithmetic instructions.
5193 (UNSPEC_BCDSUB): Likewise.
5194 (UNSPEC_BCD_OVERFLOW): Likewise.
5195 (UNSPEC_BCD_ADD_SUB): Likewise.
5196 (bcd_add_sub): Likewise.
5197 (BCD_TEST): Likewise.
5198 (bcd<bcd_add_sub>): Likewise.
5199 (bcd<bcd_add_sub>_test): Likewise.
5200 (bcd<bcd_add_sub>_test2): Likewise.
5201 (bcd<bcd_add_sub>_<code>): Likewise.
5202 (peephole2 for combined bcd ops): Likewise.
5204 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
5205 decimal floating point builtin functions.
5206 (UNSPEC_DENBCD): Likewise.
5207 (UNSPEC_DXEX): Likewise.
5208 (UNSPEC_DIEX): Likewise.
5209 (UNSPEC_DSCLI): Likewise.
5210 (UNSPEC_DSCRI): Likewise.
5211 (D64_D128): Likewise.
5212 (dfp_suffix): Likewise.
5213 (dfp_ddedpd_<mode>): Likewise.
5214 (dfp_denbcd_<mode>): Likewise.
5215 (dfp_dxex_<mode>): Likewise.
5216 (dfp_diex_<mode>): Likewise.
5217 (dfp_dscli_<mode>): Likewise.
5218 (dfp_dscri_<mode>): Likewise.
5220 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
5222 (UNSPEC_CDTBCD): Likewise.
5223 (UNSPEC_CBCDTD): Likewise.
5224 (UNSPEC_DIVE): Add support for new extended divide builtin functions.
5225 (UNSPEC_DIVEO): Likewise.
5226 (UNSPEC_DIVEU): Likewise.
5227 (UNSPEC_DIVEUO): Likewise.
5228 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
5229 pack/unpack 128-bit types.
5230 (UNSPEC_PACK_128BIT): Likewise.
5231 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
5232 (udiv<mode>3): Use idiv_ldiv mode attribute.
5233 (div<mode>3): Likewise.
5234 (addg6s): Add new BCD builtin functions.
5237 (UNSPEC_DIV_EXTEND): Add support for new extended divide instructions.
5238 (div_extend): Likewise.
5239 (div<div_extend>_<mode>"): Likewise.
5240 (FP128_64): Add support for new builtin functions to pack/unpack
5242 (unpack<mode>): Likewise.
5243 (unpacktf_0): Likewise.
5244 (unpacktf_1): Likewise.
5245 (unpack<mode>_dm): Likewise.
5246 (unpack<mode>_nodm): Likewise.
5247 (pack<mode>): Likewise.
5248 (unpackv1ti): Likewise.
5249 (packv1ti): Likewise.
5251 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
5253 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
5256 2014-04-24 Jakub Jelinek <jakub@redhat.com>
5258 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
5259 * gimplify.c (omp_is_private): Change last argument's type to int.
5260 Only diagnose lastprivate if the simd argument is 1, only diagnose
5261 linear if the simd argument is 2.
5262 (gimplify_omp_for): Adjust omp_is_private callers. When adding
5263 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
5264 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
5265 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
5266 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
5267 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
5268 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
5269 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
5270 * tree-nested.c (convert_nonlocal_omp_clauses,
5271 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
5273 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
5276 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
5279 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
5281 * flag-types.h (enum ivar_visibility): Add.
5283 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
5285 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
5286 function * argument.
5288 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
5290 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
5292 2014-04-24 Radovan Obradovic <robradovic@mips.com>
5293 Tom de Vries <tom@codesourcery.com>
5295 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
5296 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
5298 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
5299 * emit-rtl.c (try_split): Same.
5301 2014-04-24 Radovan Obradovic <robradovic@mips.com>
5302 Tom de Vries <tom@codesourcery.com>
5304 * common.opt (fuse-caller-save): New option.
5306 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
5308 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of
5309 elements for big-endian.
5311 2014-04-24 Richard Biener <rguenther@suse.de>
5313 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
5314 during TER and instead use the sepops interface for expanding
5315 non-GIMPLE_SINGLE_RHS.
5317 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5319 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
5320 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
5322 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5324 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
5325 assembler 64-bit option.
5326 * configure: Regenerate.
5328 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5330 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
5331 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
5332 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
5333 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
5334 (TARGET_CRYPTO): Take TARGET_SIMD into account.
5336 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5338 * config/aarch64/aarch64-builtins.c
5339 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
5340 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
5341 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
5342 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
5344 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
5345 (Vrevsuff): New mode attribute.
5347 2014-04-24 Terry Guo <terry.guo@arm.com>
5349 * config/arm/arm.h (machine_function): Define variable
5350 after_arm_reorg here.
5351 * config/arm/arm.c (after_arm_reorg): Remove the definition.
5352 (arm_split_constant): Update the way to access variable
5355 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
5357 2014-04-23 Tom de Vries <tom@codesourcery.com>
5359 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
5361 2014-04-23 David Malcolm <dmalcolm@redhat.com>
5363 * is-a.h: Update comments to reflect the following changes to the
5364 "pointerness" of the API, making the template parameter match the
5365 return type, allowing use of is-a.h with typedefs of pointers.
5366 (is_a_helper::cast): Return a T rather then a pointer to a T, so
5367 that the return type matches the parameter to the is_a_helper.
5369 (dyn_cast): Likewise.
5371 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
5372 pointer from the is-a.h API.
5374 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
5375 (is_a_helper <cgraph_node *>::test): ...this, matching change to
5377 (is_a_helper <varpool_node>::test): Likewise, convert to...
5378 (is_a_helper <varpool_node *>::test): ...this.
5380 (varpool_first_variable): Update for removal of implicit pointer
5381 from the is-a.h API.
5382 (varpool_next_variable): Likewise.
5383 (varpool_first_static_initializer): Likewise.
5384 (varpool_next_static_initializer): Likewise.
5385 (varpool_first_defined_variable): Likewise.
5386 (varpool_next_defined_variable): Likewise.
5387 (cgraph_first_defined_function): Likewise.
5388 (cgraph_next_defined_function): Likewise.
5389 (cgraph_first_function): Likewise.
5390 (cgraph_next_function): Likewise.
5391 (cgraph_first_function_with_gimple_body): Likewise.
5392 (cgraph_next_function_with_gimple_body): Likewise.
5393 (cgraph_alias_target): Likewise.
5394 (varpool_alias_target): Likewise.
5395 (cgraph_function_or_thunk_node): Likewise.
5396 (varpool_variable_node): Likewise.
5397 (symtab_real_symbol_p): Likewise.
5398 * cgraphunit.c (referred_to_p): Likewise.
5399 (analyze_functions): Likewise.
5400 (handle_alias_pairs): Likewise.
5401 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
5402 * gimple-ssa.h (gimple_vuse_op): Likewise.
5403 (gimple_vdef_op): Likewise.
5404 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
5405 * gimple.c (gimple_build_asm_1): Likewise.
5406 (gimple_build_try): Likewise.
5407 (gimple_build_resx): Likewise.
5408 (gimple_build_eh_dispatch): Likewise.
5409 (gimple_build_omp_for): Likewise.
5410 (gimple_omp_for_set_clauses): Likewise.
5412 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
5413 (is_a_helper <gimple_statement_asm *>::test): ...this.
5414 (is_a_helper <gimple_statement_bind>::test): Convert to...
5415 (is_a_helper <gimple_statement_bind *>::test): ...this.
5416 (is_a_helper <gimple_statement_call>::test): Convert to...
5417 (is_a_helper <gimple_statement_call *>::test): ...this.
5418 (is_a_helper <gimple_statement_catch>::test): Convert to...
5419 (is_a_helper <gimple_statement_catch *>::test): ...this.
5420 (is_a_helper <gimple_statement_resx>::test): Convert to...
5421 (is_a_helper <gimple_statement_resx *>::test): ...this.
5422 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
5423 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
5424 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
5425 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
5426 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
5427 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
5428 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
5429 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
5430 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
5431 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
5432 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
5433 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
5434 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
5435 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
5436 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
5437 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
5438 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
5439 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
5440 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
5441 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
5442 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
5443 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
5444 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
5445 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
5446 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
5447 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
5448 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
5449 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
5450 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
5451 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
5452 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
5453 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
5454 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
5455 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
5456 (is_a_helper <gimple_statement_phi>::test): Convert to...
5457 (is_a_helper <gimple_statement_phi *>::test): ...this.
5458 (is_a_helper <gimple_statement_transaction>::test): Convert to...
5459 (is_a_helper <gimple_statement_transaction *>::test): ...this.
5460 (is_a_helper <gimple_statement_try>::test): Convert to...
5461 (is_a_helper <gimple_statement_try *>::test): ...this.
5462 (is_a_helper <gimple_statement_wce>::test): Convert to...
5463 (is_a_helper <gimple_statement_wce *>::test): ...this.
5464 (is_a_helper <const gimple_statement_asm>::test): Convert to...
5465 (is_a_helper <const gimple_statement_asm *>::test): ...this.
5466 (is_a_helper <const gimple_statement_bind>::test): Convert to...
5467 (is_a_helper <const gimple_statement_bind *>::test): ...this.
5468 (is_a_helper <const gimple_statement_call>::test): Convert to...
5469 (is_a_helper <const gimple_statement_call *>::test): ...this.
5470 (is_a_helper <const gimple_statement_catch>::test): Convert to...
5471 (is_a_helper <const gimple_statement_catch *>::test): ...this.
5472 (is_a_helper <const gimple_statement_resx>::test): Convert to...
5473 (is_a_helper <const gimple_statement_resx *>::test): ...this.
5474 (is_a_helper <const gimple_statement_eh_dispatch>::test): Convert to...
5475 (is_a_helper <const gimple_statement_eh_dispatch *>::test): ...this.
5476 (is_a_helper <const gimple_statement_eh_filter>::test): Convert to...
5477 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
5478 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
5480 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
5482 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
5484 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
5486 (is_a_helper <const gimple_statement_omp_return>::test): Convert to...
5487 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
5488 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
5490 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
5491 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
5493 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
5494 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
5495 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
5496 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert to...
5497 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
5498 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
5500 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
5501 (is_a_helper <const gimple_statement_omp_target>::test): Convert to...
5502 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
5503 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
5505 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
5506 (is_a_helper <const gimple_statement_omp_single>::test): Convert to...
5507 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
5508 (is_a_helper <const gimple_statement_omp_teams>::test): Convert to...
5509 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
5510 (is_a_helper <const gimple_statement_omp_task>::test): Convert to...
5511 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
5512 (is_a_helper <const gimple_statement_phi>::test): Convert to...
5513 (is_a_helper <const gimple_statement_phi *>::test): ...this.
5514 (is_a_helper <const gimple_statement_transaction>::test): Convert to...
5515 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
5516 (is_a_helper <const gimple_statement_with_ops>::test): Convert to...
5517 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
5518 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
5519 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
5520 (is_a_helper <const gimple_statement_with_memory_ops>::test): Convert
5522 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
5524 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert to...
5525 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
5527 (gimple_use_ops): Update for removal of implicit pointer from the
5529 (gimple_set_use_ops): Likewise.
5530 (gimple_vuse): Likewise.
5531 (gimple_vdef): Likewise.
5532 (gimple_vuse_ptr): Likewise.
5533 (gimple_vdef_ptr): Likewise.
5534 (gimple_set_vuse): Likewise.
5535 (gimple_set_vdef): Likewise.
5536 (gimple_omp_return_set_lhs): Likewise.
5537 (gimple_omp_return_lhs): Likewise.
5538 (gimple_omp_return_lhs_ptr): Likewise.
5539 (gimple_call_fntype): Likewise.
5540 (gimple_call_set_fntype): Likewise.
5541 (gimple_call_set_internal_fn): Likewise.
5542 (gimple_call_use_set): Likewise.
5543 (gimple_call_clobber_set): Likewise.
5544 (gimple_bind_vars): Likewise.
5545 (gimple_bind_set_vars): Likewise.
5546 (gimple_bind_body_ptr): Likewise.
5547 (gimple_bind_set_body): Likewise.
5548 (gimple_bind_add_stmt): Likewise.
5549 (gimple_bind_block): Likewise.
5550 (gimple_bind_set_block): Likewise.
5551 (gimple_asm_ninputs): Likewise.
5552 (gimple_asm_noutputs): Likewise.
5553 (gimple_asm_nclobbers): Likewise.
5554 (gimple_asm_nlabels): Likewise.
5555 (gimple_asm_input_op): Likewise.
5556 (gimple_asm_input_op_ptr): Likewise.
5557 (gimple_asm_output_op): Likewise.
5558 (gimple_asm_output_op_ptr): Likewise.
5559 (gimple_asm_set_output_op): Likewise.
5560 (gimple_asm_clobber_op): Likewise.
5561 (gimple_asm_set_clobber_op): Likewise.
5562 (gimple_asm_label_op): Likewise.
5563 (gimple_asm_set_label_op): Likewise.
5564 (gimple_asm_string): Likewise.
5565 (gimple_catch_types): Likewise.
5566 (gimple_catch_types_ptr): Likewise.
5567 (gimple_catch_handler_ptr): Likewise.
5568 (gimple_catch_set_types): Likewise.
5569 (gimple_catch_set_handler): Likewise.
5570 (gimple_eh_filter_types): Likewise.
5571 (gimple_eh_filter_types_ptr): Likewise.
5572 (gimple_eh_filter_failure_ptr): Likewise.
5573 (gimple_eh_filter_set_types): Likewise.
5574 (gimple_eh_filter_set_failure): Likewise.
5575 (gimple_eh_must_not_throw_fndecl): Likewise.
5576 (gimple_eh_must_not_throw_set_fndecl): Likewise.
5577 (gimple_eh_else_n_body_ptr): Likewise.
5578 (gimple_eh_else_e_body_ptr): Likewise.
5579 (gimple_eh_else_set_n_body): Likewise.
5580 (gimple_eh_else_set_e_body): Likewise.
5581 (gimple_try_eval_ptr): Likewise.
5582 (gimple_try_cleanup_ptr): Likewise.
5583 (gimple_try_set_eval): Likewise.
5584 (gimple_try_set_cleanup): Likewise.
5585 (gimple_wce_cleanup_ptr): Likewise.
5586 (gimple_wce_set_cleanup): Likewise.
5587 (gimple_phi_capacity): Likewise.
5588 (gimple_phi_num_args): Likewise.
5589 (gimple_phi_result): Likewise.
5590 (gimple_phi_result_ptr): Likewise.
5591 (gimple_phi_set_result): Likewise.
5592 (gimple_phi_arg): Likewise.
5593 (gimple_phi_set_arg): Likewise.
5594 (gimple_resx_region): Likewise.
5595 (gimple_resx_set_region): Likewise.
5596 (gimple_eh_dispatch_region): Likewise.
5597 (gimple_eh_dispatch_set_region): Likewise.
5598 (gimple_omp_critical_name): Likewise.
5599 (gimple_omp_critical_name_ptr): Likewise.
5600 (gimple_omp_critical_set_name): Likewise.
5601 (gimple_omp_for_clauses): Likewise.
5602 (gimple_omp_for_clauses_ptr): Likewise.
5603 (gimple_omp_for_set_clauses): Likewise.
5604 (gimple_omp_for_collapse): Likewise.
5605 (gimple_omp_for_index): Likewise.
5606 (gimple_omp_for_index_ptr): Likewise.
5607 (gimple_omp_for_set_index): Likewise.
5608 (gimple_omp_for_initial): Likewise.
5609 (gimple_omp_for_initial_ptr): Likewise.
5610 (gimple_omp_for_set_initial): Likewise.
5611 (gimple_omp_for_final): Likewise.
5612 (gimple_omp_for_final_ptr): Likewise.
5613 (gimple_omp_for_set_final): Likewise.
5614 (gimple_omp_for_incr): Likewise.
5615 (gimple_omp_for_incr_ptr): Likewise.
5616 (gimple_omp_for_set_incr): Likewise.
5617 (gimple_omp_for_pre_body_ptr): Likewise.
5618 (gimple_omp_for_set_pre_body): Likewise.
5619 (gimple_omp_parallel_clauses): Likewise.
5620 (gimple_omp_parallel_clauses_ptr): Likewise.
5621 (gimple_omp_parallel_set_clauses): Likewise.
5622 (gimple_omp_parallel_child_fn): Likewise.
5623 (gimple_omp_parallel_child_fn_ptr): Likewise.
5624 (gimple_omp_parallel_set_child_fn): Likewise.
5625 (gimple_omp_parallel_data_arg): Likewise.
5626 (gimple_omp_parallel_data_arg_ptr): Likewise.
5627 (gimple_omp_parallel_set_data_arg): Likewise.
5628 (gimple_omp_task_clauses): Likewise.
5629 (gimple_omp_task_clauses_ptr): Likewise.
5630 (gimple_omp_task_set_clauses): Likewise.
5631 (gimple_omp_task_child_fn): Likewise.
5632 (gimple_omp_task_child_fn_ptr): Likewise.
5633 (gimple_omp_task_set_child_fn): Likewise.
5634 (gimple_omp_task_data_arg): Likewise.
5635 (gimple_omp_task_data_arg_ptr): Likewise.
5636 (gimple_omp_task_set_data_arg): Likewise.
5637 (gimple_omp_taskreg_clauses): Likewise.
5638 (gimple_omp_taskreg_clauses_ptr): Likewise.
5639 (gimple_omp_taskreg_set_clauses): Likewise.
5640 (gimple_omp_taskreg_child_fn): Likewise.
5641 (gimple_omp_taskreg_child_fn_ptr): Likewise.
5642 (gimple_omp_taskreg_set_child_fn): Likewise.
5643 (gimple_omp_taskreg_data_arg): Likewise.
5644 (gimple_omp_taskreg_data_arg_ptr): Likewise.
5645 (gimple_omp_taskreg_set_data_arg): Likewise.
5646 (gimple_omp_task_copy_fn): Likewise.
5647 (gimple_omp_task_copy_fn_ptr): Likewise.
5648 (gimple_omp_task_set_copy_fn): Likewise.
5649 (gimple_omp_task_arg_size): Likewise.
5650 (gimple_omp_task_arg_size_ptr): Likewise.
5651 (gimple_omp_task_set_arg_size): Likewise.
5652 (gimple_omp_task_arg_align): Likewise.
5653 (gimple_omp_task_arg_align_ptr): Likewise.
5654 (gimple_omp_task_set_arg_align): Likewise.
5655 (gimple_omp_single_clauses): Likewise.
5656 (gimple_omp_single_clauses_ptr): Likewise.
5657 (gimple_omp_single_set_clauses): Likewise.
5658 (gimple_omp_target_clauses): Likewise.
5659 (gimple_omp_target_clauses_ptr): Likewise.
5660 (gimple_omp_target_set_clauses): Likewise.
5661 (gimple_omp_target_child_fn): Likewise.
5662 (gimple_omp_target_child_fn_ptr): Likewise.
5663 (gimple_omp_target_set_child_fn): Likewise.
5664 (gimple_omp_target_data_arg): Likewise.
5665 (gimple_omp_target_data_arg_ptr): Likewise.
5666 (gimple_omp_target_set_data_arg): Likewise.
5667 (gimple_omp_teams_clauses): Likewise.
5668 (gimple_omp_teams_clauses_ptr): Likewise.
5669 (gimple_omp_teams_set_clauses): Likewise.
5670 (gimple_omp_sections_clauses): Likewise.
5671 (gimple_omp_sections_clauses_ptr): Likewise.
5672 (gimple_omp_sections_set_clauses): Likewise.
5673 (gimple_omp_sections_control): Likewise.
5674 (gimple_omp_sections_control_ptr): Likewise.
5675 (gimple_omp_sections_set_control): Likewise.
5676 (gimple_omp_for_set_cond): Likewise.
5677 (gimple_omp_for_cond): Likewise.
5678 (gimple_omp_atomic_store_set_val): Likewise.
5679 (gimple_omp_atomic_store_val): Likewise.
5680 (gimple_omp_atomic_store_val_ptr): Likewise.
5681 (gimple_omp_atomic_load_set_lhs): Likewise.
5682 (gimple_omp_atomic_load_lhs): Likewise.
5683 (gimple_omp_atomic_load_lhs_ptr): Likewise.
5684 (gimple_omp_atomic_load_set_rhs): Likewise.
5685 (gimple_omp_atomic_load_rhs): Likewise.
5686 (gimple_omp_atomic_load_rhs_ptr): Likewise.
5687 (gimple_omp_continue_control_def): Likewise.
5688 (gimple_omp_continue_control_def_ptr): Likewise.
5689 (gimple_omp_continue_set_control_def): Likewise.
5690 (gimple_omp_continue_control_use): Likewise.
5691 (gimple_omp_continue_control_use_ptr): Likewise.
5692 (gimple_omp_continue_set_control_use): Likewise.
5693 (gimple_transaction_body_ptr): Likewise.
5694 (gimple_transaction_label): Likewise.
5695 (gimple_transaction_label_ptr): Likewise.
5696 (gimple_transaction_set_body): Likewise.
5697 (gimple_transaction_set_label): Likewise.
5699 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
5700 * ipa-inline-analysis.c (inline_write_summary): Likewise.
5701 * ipa-ref.c (ipa_record_reference): Likewise.
5702 * ipa-reference.c (analyze_function): Likewise.
5703 (ipa_reference_write_optimization_summary): Likewise.
5704 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
5705 (address_taken_from_non_vtable_p): Likewise.
5706 (comdat_can_be_unshared_p_1): Likewise.
5707 * lto-cgraph.c (lto_output_ref): Likewise.
5708 (add_references): Likewise.
5709 (compute_ltrans_boundary): Likewise.
5710 (output_symtab): Likewise.
5711 (input_ref): Likewise.
5712 (input_cgraph_1): Likewise.
5713 (output_cgraph_opt_summary): Likewise.
5714 * lto-streamer-out.c (lto_output): Likewise.
5715 (output_symbol_p): Likewise.
5716 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
5717 (lsei_start_function_in_partition): Likewise.
5718 (lsei_next_variable_in_partition): Likewise.
5719 (lsei_start_variable_in_partition): Likewise.
5720 * symtab.c (insert_to_assembler_name_hash): Likewise.
5721 (unlink_from_assembler_name_hash): Likewise.
5722 (symtab_unregister_node): Likewise.
5723 (symtab_remove_node): Likewise.
5724 (dump_symtab_node): Likewise.
5725 (verify_symtab_base): Likewise.
5726 (verify_symtab_node): Likewise.
5727 (symtab_make_decl_local): Likewise.
5728 (symtab_alias_ultimate_target): Likewise.
5729 (symtab_resolve_alias): Likewise.
5730 (symtab_get_symbol_partitioning_class): Likewise.
5731 * tree-phinodes.c (allocate_phi_node): Likewise.
5732 (reserve_phi_args_for_new_edge): Likewise.
5733 (remove_phi_args): Likewise.
5734 * varpool.c (varpool_node_for_asm): Likewise.
5735 (varpool_remove_unreferenced_decls): Likewise.
5737 2014-04-23 Jeff Law <law@redhat.com>
5739 PR tree-optimization/60902
5740 * tree-ssa-threadedge.c
5741 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
5742 invalidate outputs from statements that do not produce useful
5743 outputs for threading.
5745 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
5747 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
5748 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
5749 machine descriptions for Stack Smashing Protector.
5751 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
5753 * aarch64.md (<optab>_rol<mode>3): New pattern.
5754 (<optab>_rolsi3_uxtw): Likewise.
5755 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
5757 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
5759 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
5760 (arm_cortex_a12_tune): Likewise.
5762 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5764 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
5766 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5768 * config/arm/arm.md (arm_rev16si2): New pattern.
5769 (arm_rev16si2_alt): Likewise.
5770 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
5772 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5774 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
5775 (rev16<mode>2_alt): Likewise.
5776 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
5777 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
5778 (aarch_rev16_shleft_mask_imm_p): Likewise.
5779 (aarch_rev16_p_1): Likewise.
5780 (aarch_rev16_p): Likewise.
5781 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
5782 (aarch_rev16_shright_mask_imm_p): Likewise.
5783 (aarch_rev16_shleft_mask_imm_p): Likewise.
5785 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5787 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
5788 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
5790 (cortex_a53_extra_costs): Likewise.
5791 (cortex_a57_extra_costs): Likewise.
5792 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
5793 (cortexa7_extra_costs): Likewise.
5794 (cortexa8_extra_costs): Likewise.
5795 (cortexa12_extra_costs): Likewise.
5796 (cortexa15_extra_costs): Likewise.
5797 (v7m_extra_costs): Likewise.
5798 (arm_new_rtx_costs): Handle BSWAP.
5800 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5802 * config/arm/arm.c (cortexa8_extra_costs): New table.
5803 (arm_cortex_a8_tune): New tuning struct.
5804 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
5806 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5808 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
5810 2014-04-23 Richard Biener <rguenther@suse.de>
5812 * Makefile.in (OBJS): Remove loop-unswitch.o.
5813 * tree-pass.h (make_pass_rtl_unswitch): Remove.
5814 * passes.def (pass_rtl_unswitch): Likewise.
5815 * loop-init.c (gate_rtl_unswitch): Likewise.
5816 (rtl_unswitch): Likewise.
5817 (pass_data_rtl_unswitch): Likewise.
5818 (pass_rtl_unswitch): Likewise.
5819 (make_pass_rtl_unswitch): Likewise.
5820 * rtl.h (reversed_condition): Likewise.
5821 (compare_and_jump_seq): Likewise.
5822 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
5824 * loop-unroll.c (compare_and_jump_seq): Likewise.
5826 2014-04-23 Richard Biener <rguenther@suse.de>
5828 PR tree-optimization/60903
5829 * tree-ssa-loop-im.c (analyze_memory_references): Remove
5830 commented code block.
5831 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
5832 loop flags to newly created BBs and edges.
5834 2014-04-23 Nick Clifton <nickc@redhat.com>
5836 * config/msp430/msp430.c (msp430_handle_option): Move function
5838 (msp430_option_override): Simplify mcu and mcpu option handling.
5839 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
5840 support for -mhwmult command line option.
5841 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
5842 -mhwmult command line option.
5843 (msp430_hwmult_enabled): Delete.
5844 (msp43o_output_labelref): Add support for -mhwmult command line option.
5845 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
5846 (umulsidi3): Likewise.
5847 * config/msp430/msp430.opt (mmcu): Add Report attribute.
5848 (mcpu, mlarge, msmall): Likewise.
5849 (mhwmult): New option.
5850 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
5852 (msp430_is_f5_mcu): Remove prototype.
5853 (msp430_use_f5_series_hwmult): Add prototype.
5854 * config/msp430/msp430-opts.h: New file.
5855 * common/config/msp430: New directory.
5856 * common/config/msp430/msp430-common.c: New file.
5857 * config.gcc (msp430): Remove target_has_targetm_common.
5858 * doc/invoke.texi: Document -mhwmult command line option.
5860 2014-04-23 Nick Clifton <nickc@redhat.com>
5862 * config/i386/cygwin.h (ENDFILE_SPEC): Include
5863 default-manifest.o if it can be found in the search path.
5864 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
5866 2014-04-23 Terry Guo <terry.guo@arm.com>
5868 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
5870 2014-04-23 Richard Biener <rguenther@suse.de>
5873 * tree-inline.c (declare_return_variable): Use mark_addressable.
5875 2014-04-23 Richard Biener <rguenther@suse.de>
5878 * loop-init.c (loop_optimizer_init): Make sure to apply
5879 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
5881 2014-04-22 Jakub Jelinek <jakub@redhat.com>
5884 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
5886 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
5887 if flag_sanitize_undefined_trap_on_error.
5888 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
5889 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
5890 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
5891 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
5892 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
5893 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
5894 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
5895 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
5896 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
5897 * ubsan.c (ubsan_instrument_unreachable): Return
5898 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
5899 (ubsan_expand_null_ifn): Emit __builtin_trap ()
5900 if flag_sanitize_undefined_trap_on_error and
5901 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
5902 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
5903 instrument_bool_enum_load): Emit __builtin_trap () if
5904 flag_sanitize_undefined_trap_on_error and
5905 __builtin_handle_*_abort () if !flag_sanitize_recover.
5906 * doc/invoke.texi (-fsanitize-recover,
5907 -fsanitize-undefined-trap-on-error): Document.
5909 2014-04-22 Christian Bruel <christian.bruel@st.com>
5911 * config/sh/sh.md (mov<mode>): Replace movQIHI.
5912 Force immediates to SImode.
5914 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
5916 * config/nios2/nios2.md (UNSPEC_ROUND): New.
5918 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
5919 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
5920 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
5921 (nios2_fpu_insn): Add entry for round.
5922 (N2FPU_NO_ERRNO_P): Define.
5923 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
5925 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
5927 2014-04-22 Richard Henderson <rth@redhat.com>
5929 * config/aarch64/aarch64 (addti3, subti3): New expanders.
5930 (add<GPI>3_compare0): Remove leading * from name.
5931 (add<GPI>3_carryin): Likewise.
5932 (sub<GPI>3_compare0): Likewise.
5933 (sub<GPI>3_carryin): Likewise.
5934 (<su_optab>mulditi3): New expander.
5935 (multi3): New expander.
5936 (madd<GPI>): Remove leading * from name.
5938 2014-04-22 Martin Jambor <mjambor@suse.cz>
5940 * cgraphclones.c (cgraph_function_versioning): Copy
5941 ipa_transforms_to_apply instead of asserting it is empty.
5943 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
5946 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
5947 on count_exp to get mode.
5949 2014-04-22 Andrew Pinski <apinski@cavium.com>
5951 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
5952 Handle TLS for ILP32.
5953 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
5954 (tlsie_small_<mode>): this and handle PTR.
5955 (tlsie_small_sidi): New pattern.
5956 (tlsle_small): Change to an expand to handle ILP32.
5957 (tlsle_small_<mode>): New pattern.
5958 (tlsdesc_small): Rename to ...
5959 (tlsdesc_small_<mode>): this and handle PTR.
5961 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5963 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
5965 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
5967 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
5968 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
5969 (aarch64_types_signed_poly_qualifiers): Likewise.
5970 (aarch64_types_unsigned_signed_qualifiers): Likewise.
5971 (aarch64_types_poly_signed_qualifiers): Likewise.
5972 (TYPES_REINTERP_SS): Type macro added.
5973 (TYPES_REINTERP_SU): Likewise.
5974 (TYPES_REINTERP_SP): Likewise.
5975 (TYPES_REINTERP_US): Likewise.
5976 (TYPES_REINTERP_PS): Likewise.
5977 (aarch64_fold_builtin): New expression folding added.
5978 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
5979 Declarations removed.
5980 (REINTERP_SS): Declarations added.
5981 (REINTERP_US): Likewise.
5982 (REINTERP_PS): Likewise.
5983 (REINTERP_SU): Likewise.
5984 (REINTERP_SP): Likewise.
5985 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
5986 (vreinterpretq_p8_f64): Likewise.
5987 (vreinterpret_p16_f64): Likewise.
5988 (vreinterpretq_p16_f64): Likewise.
5989 (vreinterpret_f32_f64): Likewise.
5990 (vreinterpretq_f32_f64): Likewise.
5991 (vreinterpret_f64_f32): Likewise.
5992 (vreinterpret_f64_p8): Likewise.
5993 (vreinterpret_f64_p16): Likewise.
5994 (vreinterpret_f64_s8): Likewise.
5995 (vreinterpret_f64_s16): Likewise.
5996 (vreinterpret_f64_s32): Likewise.
5997 (vreinterpret_f64_s64): Likewise.
5998 (vreinterpret_f64_u8): Likewise.
5999 (vreinterpret_f64_u16): Likewise.
6000 (vreinterpret_f64_u32): Likewise.
6001 (vreinterpret_f64_u64): Likewise.
6002 (vreinterpretq_f64_f32): Likewise.
6003 (vreinterpretq_f64_p8): Likewise.
6004 (vreinterpretq_f64_p16): Likewise.
6005 (vreinterpretq_f64_s8): Likewise.
6006 (vreinterpretq_f64_s16): Likewise.
6007 (vreinterpretq_f64_s32): Likewise.
6008 (vreinterpretq_f64_s64): Likewise.
6009 (vreinterpretq_f64_u8): Likewise.
6010 (vreinterpretq_f64_u16): Likewise.
6011 (vreinterpretq_f64_u32): Likewise.
6012 (vreinterpretq_f64_u64): Likewise.
6013 (vreinterpret_s64_f64): Likewise.
6014 (vreinterpretq_s64_f64): Likewise.
6015 (vreinterpret_u64_f64): Likewise.
6016 (vreinterpretq_u64_f64): Likewise.
6017 (vreinterpret_s8_f64): Likewise.
6018 (vreinterpretq_s8_f64): Likewise.
6019 (vreinterpret_s16_f64): Likewise.
6020 (vreinterpretq_s16_f64): Likewise.
6021 (vreinterpret_s32_f64): Likewise.
6022 (vreinterpretq_s32_f64): Likewise.
6023 (vreinterpret_u8_f64): Likewise.
6024 (vreinterpretq_u8_f64): Likewise.
6025 (vreinterpret_u16_f64): Likewise.
6026 (vreinterpretq_u16_f64): Likewise.
6027 (vreinterpret_u32_f64): Likewise.
6028 (vreinterpretq_u32_f64): Likewise.
6030 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
6032 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
6033 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
6034 (vreinterpret_p8_s8): Likewise.
6035 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
6036 (vreinterpret_p8_s16): Likewise.
6037 (vreinterpret_p8_s32): Likewise.
6038 (vreinterpret_p8_s64): Likewise.
6039 (vreinterpret_p8_f32): Likewise.
6040 (vreinterpret_p8_u8): Likewise.
6041 (vreinterpret_p8_u16): Likewise.
6042 (vreinterpret_p8_u32): Likewise.
6043 (vreinterpret_p8_u64): Likewise.
6044 (vreinterpret_p8_p16): Likewise.
6045 (vreinterpretq_p8_s8): Likewise.
6046 (vreinterpretq_p8_s16): Likewise.
6047 (vreinterpretq_p8_s32): Likewise.
6048 (vreinterpretq_p8_s64): Likewise.
6049 (vreinterpretq_p8_f32): Likewise.
6050 (vreinterpretq_p8_u8): Likewise.
6051 (vreinterpretq_p8_u16): Likewise.
6052 (vreinterpretq_p8_u32): Likewise.
6053 (vreinterpretq_p8_u64): Likewise.
6054 (vreinterpretq_p8_p16): Likewise.
6055 (vreinterpret_p16_s8): Likewise.
6056 (vreinterpret_p16_s16): Likewise.
6057 (vreinterpret_p16_s32): Likewise.
6058 (vreinterpret_p16_s64): Likewise.
6059 (vreinterpret_p16_f32): Likewise.
6060 (vreinterpret_p16_u8): Likewise.
6061 (vreinterpret_p16_u16): Likewise.
6062 (vreinterpret_p16_u32): Likewise.
6063 (vreinterpret_p16_u64): Likewise.
6064 (vreinterpret_p16_p8): Likewise.
6065 (vreinterpretq_p16_s8): Likewise.
6066 (vreinterpretq_p16_s16): Likewise.
6067 (vreinterpretq_p16_s32): Likewise.
6068 (vreinterpretq_p16_s64): Likewise.
6069 (vreinterpretq_p16_f32): Likewise.
6070 (vreinterpretq_p16_u8): Likewise.
6071 (vreinterpretq_p16_u16): Likewise.
6072 (vreinterpretq_p16_u32): Likewise.
6073 (vreinterpretq_p16_u64): Likewise.
6074 (vreinterpretq_p16_p8): Likewise.
6075 (vreinterpret_f32_s8): Likewise.
6076 (vreinterpret_f32_s16): Likewise.
6077 (vreinterpret_f32_s32): Likewise.
6078 (vreinterpret_f32_s64): Likewise.
6079 (vreinterpret_f32_u8): Likewise.
6080 (vreinterpret_f32_u16): Likewise.
6081 (vreinterpret_f32_u32): Likewise.
6082 (vreinterpret_f32_u64): Likewise.
6083 (vreinterpret_f32_p8): Likewise.
6084 (vreinterpret_f32_p16): Likewise.
6085 (vreinterpretq_f32_s8): Likewise.
6086 (vreinterpretq_f32_s16): Likewise.
6087 (vreinterpretq_f32_s32): Likewise.
6088 (vreinterpretq_f32_s64): Likewise.
6089 (vreinterpretq_f32_u8): Likewise.
6090 (vreinterpretq_f32_u16): Likewise.
6091 (vreinterpretq_f32_u32): Likewise.
6092 (vreinterpretq_f32_u64): Likewise.
6093 (vreinterpretq_f32_p8): Likewise.
6094 (vreinterpretq_f32_p16): Likewise.
6095 (vreinterpret_s64_s8): Likewise.
6096 (vreinterpret_s64_s16): Likewise.
6097 (vreinterpret_s64_s32): Likewise.
6098 (vreinterpret_s64_f32): Likewise.
6099 (vreinterpret_s64_u8): Likewise.
6100 (vreinterpret_s64_u16): Likewise.
6101 (vreinterpret_s64_u32): Likewise.
6102 (vreinterpret_s64_u64): Likewise.
6103 (vreinterpret_s64_p8): Likewise.
6104 (vreinterpret_s64_p16): Likewise.
6105 (vreinterpretq_s64_s8): Likewise.
6106 (vreinterpretq_s64_s16): Likewise.
6107 (vreinterpretq_s64_s32): Likewise.
6108 (vreinterpretq_s64_f32): Likewise.
6109 (vreinterpretq_s64_u8): Likewise.
6110 (vreinterpretq_s64_u16): Likewise.
6111 (vreinterpretq_s64_u32): Likewise.
6112 (vreinterpretq_s64_u64): Likewise.
6113 (vreinterpretq_s64_p8): Likewise.
6114 (vreinterpretq_s64_p16): Likewise.
6115 (vreinterpret_u64_s8): Likewise.
6116 (vreinterpret_u64_s16): Likewise.
6117 (vreinterpret_u64_s32): Likewise.
6118 (vreinterpret_u64_s64): Likewise.
6119 (vreinterpret_u64_f32): Likewise.
6120 (vreinterpret_u64_u8): Likewise.
6121 (vreinterpret_u64_u16): Likewise.
6122 (vreinterpret_u64_u32): Likewise.
6123 (vreinterpret_u64_p8): Likewise.
6124 (vreinterpret_u64_p16): Likewise.
6125 (vreinterpretq_u64_s8): Likewise.
6126 (vreinterpretq_u64_s16): Likewise.
6127 (vreinterpretq_u64_s32): Likewise.
6128 (vreinterpretq_u64_s64): Likewise.
6129 (vreinterpretq_u64_f32): Likewise.
6130 (vreinterpretq_u64_u8): Likewise.
6131 (vreinterpretq_u64_u16): Likewise.
6132 (vreinterpretq_u64_u32): Likewise.
6133 (vreinterpretq_u64_p8): Likewise.
6134 (vreinterpretq_u64_p16): Likewise.
6135 (vreinterpret_s8_s16): Likewise.
6136 (vreinterpret_s8_s32): Likewise.
6137 (vreinterpret_s8_s64): Likewise.
6138 (vreinterpret_s8_f32): Likewise.
6139 (vreinterpret_s8_u8): Likewise.
6140 (vreinterpret_s8_u16): Likewise.
6141 (vreinterpret_s8_u32): Likewise.
6142 (vreinterpret_s8_u64): Likewise.
6143 (vreinterpret_s8_p8): Likewise.
6144 (vreinterpret_s8_p16): Likewise.
6145 (vreinterpretq_s8_s16): Likewise.
6146 (vreinterpretq_s8_s32): Likewise.
6147 (vreinterpretq_s8_s64): Likewise.
6148 (vreinterpretq_s8_f32): Likewise.
6149 (vreinterpretq_s8_u8): Likewise.
6150 (vreinterpretq_s8_u16): Likewise.
6151 (vreinterpretq_s8_u32): Likewise.
6152 (vreinterpretq_s8_u64): Likewise.
6153 (vreinterpretq_s8_p8): Likewise.
6154 (vreinterpretq_s8_p16): Likewise.
6155 (vreinterpret_s16_s8): Likewise.
6156 (vreinterpret_s16_s32): Likewise.
6157 (vreinterpret_s16_s64): Likewise.
6158 (vreinterpret_s16_f32): Likewise.
6159 (vreinterpret_s16_u8): Likewise.
6160 (vreinterpret_s16_u16): Likewise.
6161 (vreinterpret_s16_u32): Likewise.
6162 (vreinterpret_s16_u64): Likewise.
6163 (vreinterpret_s16_p8): Likewise.
6164 (vreinterpret_s16_p16): Likewise.
6165 (vreinterpretq_s16_s8): Likewise.
6166 (vreinterpretq_s16_s32): Likewise.
6167 (vreinterpretq_s16_s64): Likewise.
6168 (vreinterpretq_s16_f32): Likewise.
6169 (vreinterpretq_s16_u8): Likewise.
6170 (vreinterpretq_s16_u16): Likewise.
6171 (vreinterpretq_s16_u32): Likewise.
6172 (vreinterpretq_s16_u64): Likewise.
6173 (vreinterpretq_s16_p8): Likewise.
6174 (vreinterpretq_s16_p16): Likewise.
6175 (vreinterpret_s32_s8): Likewise.
6176 (vreinterpret_s32_s16): Likewise.
6177 (vreinterpret_s32_s64): Likewise.
6178 (vreinterpret_s32_f32): Likewise.
6179 (vreinterpret_s32_u8): Likewise.
6180 (vreinterpret_s32_u16): Likewise.
6181 (vreinterpret_s32_u32): Likewise.
6182 (vreinterpret_s32_u64): Likewise.
6183 (vreinterpret_s32_p8): Likewise.
6184 (vreinterpret_s32_p16): Likewise.
6185 (vreinterpretq_s32_s8): Likewise.
6186 (vreinterpretq_s32_s16): Likewise.
6187 (vreinterpretq_s32_s64): Likewise.
6188 (vreinterpretq_s32_f32): Likewise.
6189 (vreinterpretq_s32_u8): Likewise.
6190 (vreinterpretq_s32_u16): Likewise.
6191 (vreinterpretq_s32_u32): Likewise.
6192 (vreinterpretq_s32_u64): Likewise.
6193 (vreinterpretq_s32_p8): Likewise.
6194 (vreinterpretq_s32_p16): Likewise.
6195 (vreinterpret_u8_s8): Likewise.
6196 (vreinterpret_u8_s16): Likewise.
6197 (vreinterpret_u8_s32): Likewise.
6198 (vreinterpret_u8_s64): Likewise.
6199 (vreinterpret_u8_f32): Likewise.
6200 (vreinterpret_u8_u16): Likewise.
6201 (vreinterpret_u8_u32): Likewise.
6202 (vreinterpret_u8_u64): Likewise.
6203 (vreinterpret_u8_p8): Likewise.
6204 (vreinterpret_u8_p16): Likewise.
6205 (vreinterpretq_u8_s8): Likewise.
6206 (vreinterpretq_u8_s16): Likewise.
6207 (vreinterpretq_u8_s32): Likewise.
6208 (vreinterpretq_u8_s64): Likewise.
6209 (vreinterpretq_u8_f32): Likewise.
6210 (vreinterpretq_u8_u16): Likewise.
6211 (vreinterpretq_u8_u32): Likewise.
6212 (vreinterpretq_u8_u64): Likewise.
6213 (vreinterpretq_u8_p8): Likewise.
6214 (vreinterpretq_u8_p16): Likewise.
6215 (vreinterpret_u16_s8): Likewise.
6216 (vreinterpret_u16_s16): Likewise.
6217 (vreinterpret_u16_s32): Likewise.
6218 (vreinterpret_u16_s64): Likewise.
6219 (vreinterpret_u16_f32): Likewise.
6220 (vreinterpret_u16_u8): Likewise.
6221 (vreinterpret_u16_u32): Likewise.
6222 (vreinterpret_u16_u64): Likewise.
6223 (vreinterpret_u16_p8): Likewise.
6224 (vreinterpret_u16_p16): Likewise.
6225 (vreinterpretq_u16_s8): Likewise.
6226 (vreinterpretq_u16_s16): Likewise.
6227 (vreinterpretq_u16_s32): Likewise.
6228 (vreinterpretq_u16_s64): Likewise.
6229 (vreinterpretq_u16_f32): Likewise.
6230 (vreinterpretq_u16_u8): Likewise.
6231 (vreinterpretq_u16_u32): Likewise.
6232 (vreinterpretq_u16_u64): Likewise.
6233 (vreinterpretq_u16_p8): Likewise.
6234 (vreinterpretq_u16_p16): Likewise.
6235 (vreinterpret_u32_s8): Likewise.
6236 (vreinterpret_u32_s16): Likewise.
6237 (vreinterpret_u32_s32): Likewise.
6238 (vreinterpret_u32_s64): Likewise.
6239 (vreinterpret_u32_f32): Likewise.
6240 (vreinterpret_u32_u8): Likewise.
6241 (vreinterpret_u32_u16): Likewise.
6242 (vreinterpret_u32_u64): Likewise.
6243 (vreinterpret_u32_p8): Likewise.
6244 (vreinterpret_u32_p16): Likewise.
6245 (vreinterpretq_u32_s8): Likewise.
6246 (vreinterpretq_u32_s16): Likewise.
6247 (vreinterpretq_u32_s32): Likewise.
6248 (vreinterpretq_u32_s64): Likewise.
6249 (vreinterpretq_u32_f32): Likewise.
6250 (vreinterpretq_u32_u8): Likewise.
6251 (vreinterpretq_u32_u16): Likewise.
6252 (vreinterpretq_u32_u64): Likewise.
6253 (vreinterpretq_u32_p8): Likewise.
6254 (vreinterpretq_u32_p16): Likewise.
6256 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
6258 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
6260 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator extended.
6262 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
6263 (vqnegd_s64): Likewise.
6264 (vqabs_s64): Likewise.
6265 (vqabsd_s64): Likewise.
6267 2014-04-22 Richard Henderson <rth@redhat.com>
6269 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
6270 computation to the top of the loop.
6272 2014-04-22 Renlin <renlin.li@arm.com>
6273 Jiong Wang <jiong.wang@arm.com>
6275 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
6276 * config/aarch64/aarch64.c (aarch64_layout_frame)
6277 (aarch64_initial_elimination_offset): Likewise.
6279 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
6281 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
6284 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
6286 * machmode.h (bitwise_mode_for_mode): Declare.
6287 * stor-layout.h (bitwise_type_for_mode): Likewise.
6288 * stor-layout.c (bitwise_mode_for_mode): New function.
6289 (bitwise_type_for_mode): Likewise.
6290 * builtins.c (fold_builtin_memory_op): Use it instead of
6291 int_mode_for_mode and build_nonstandard_integer_type.
6293 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6295 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
6296 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
6297 (*-*-solaris2*): Simplify.
6298 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
6299 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
6300 *-*-solaris2.9* handling.
6302 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
6304 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
6305 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
6307 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
6308 * configure: Regenerate.
6310 * config/i386/sol2-9.h: Remove.
6312 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
6313 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
6314 Remove Solaris 9 references.
6316 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
6318 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
6319 (floatuns<GPI:mode><GPF:mode>2): Remove.
6320 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
6321 and floatuns conversions.
6322 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
6323 and floatuns conversions.
6324 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
6325 (w1,w2): New mode attributes for inequal width conversions.
6327 2014-04-22 Renlin Li <Renlin.Li@arm.com>
6329 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
6330 the output asm format.
6332 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
6334 * config/aarch64/aarch64-simd.md
6335 (aarch64_cm<optab>di): Always split.
6336 (*aarch64_cm<optab>di): New.
6337 (aarch64_cmtstdi): Always split.
6338 (*aarch64_cmtstdi): New.
6340 2014-04-22 Jakub Jelinek <jakub@redhat.com>
6342 PR tree-optimization/60823
6343 * omp-low.c (ipa_simd_modify_function_body): Go through
6344 all SSA_NAMEs and for those refering to vector arguments
6345 which are going to be replaced adjust SSA_NAME_VAR and,
6346 if it is a default definition, change it into a non-default
6347 definition assigned at the beginning of function from new_decl.
6348 (ipa_simd_modify_stmt_ops): Rewritten.
6349 * tree-dfa.c (set_ssa_default_def): When removing default def,
6350 check for NULL loc instead of NULL *loc.
6352 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6354 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
6355 restrictions on core registers for DImode values in Thumb2.
6357 2014-04-22 Ian Bolton <ian.bolton@arm.com>
6359 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
6360 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
6362 2014-04-22 Ian Bolton <ian.bolton@arm.com>
6364 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
6365 (*iordi_notzesidi_di): Likewise.
6366 (*iordi_notsesidi_di): Likewise.
6368 2014-04-22 Ian Bolton <ian.bolton@arm.com>
6370 * config/arm/arm-protos.h (tune_params): New struct members.
6371 * config/arm/arm.c: Initialise tune_params per processor.
6372 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
6373 for speed, based on new tune_params.
6375 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
6377 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro added.
6378 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added macro.
6379 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment corrected.
6380 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
6381 * config/aarch64/arm_neon.h (vrnd_f64): Added.
6382 (vrnda_f64): Likewise.
6383 (vrndi_f64): Likewise.
6384 (vrndm_f64): Likewise.
6385 (vrndn_f64): Likewise.
6386 (vrndp_f64): Likewise.
6387 (vrndx_f64): Likewise.
6389 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6391 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
6392 GET_MODE_SIZE argument is enum machine_mode.
6394 2014-04-22 Jakub Jelinek <jakub@redhat.com>
6397 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
6398 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
6400 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
6403 * asan.c (asan_emit_stack_protection): Force the base to align to
6404 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
6405 appropriate bits if STRICT_ALIGNMENT.
6406 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
6408 (expand_used_vars): Leave a space in the stack frame for alignment
6409 if STRICT_ALIGNMENT.
6411 2014-04-21 David Malcolm <dmalcolm@redhat.com>
6413 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
6415 (gimple_store_p): Likewise.
6416 (gimple_assign_load_p): Likewise.
6417 (gimple_assign_cast_p): Likewise.
6418 (gimple_clobber_p): Likewise.
6420 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
6421 rather than a gimple.
6422 (gimple_assign_cast_p): Likewise.
6424 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
6427 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
6428 If mode is DDmode and TARGET_E500_DOUBLE allow move.
6430 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
6431 more debug information for E500 if -mdebug=reg.
6433 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
6436 * config/i386/i386.c (ix86_expand_builtin)
6437 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
6438 register for target RTX.
6439 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
6441 2014-04-18 Cong Hou <congh@google.com>
6443 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
6444 the widen-mult pattern by handling two operands with different sizes,
6445 and operands whose size is smaller than half of the result type.
6447 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
6449 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
6450 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
6451 (do_estimate_edge_time): Compute it.
6452 * ipa-inline.c (want_inline_small_function_p): Bypass
6453 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
6455 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
6457 * ipa-inline.c (spec_rem): New static variable.
6458 (dump_overall_stats): New function.
6459 (dump_inline_stats): New function.
6461 2014-04-18 Richard Henderson <rth@redhat.com>
6463 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
6464 to GET_MODE_SIZE, not a reg_class_t.
6466 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6468 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
6469 (vsx_xxmrglw_<mode>): Likewise.
6471 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
6474 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
6475 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
6476 (rs6000_init_hard_regno_mode_ok): Likewise.
6478 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
6480 * ipa-inline.c (inline_small_functions): Account only non-cold
6482 * doc/invoke.texi (inline-unit-growth): Update documentation.
6484 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
6486 * config/rs6000/rs6000.md (addti3, subti3): New.
6488 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
6491 * config/i386/i386.c (ix86_expand_clear): Remove outdated
6492 comment. Check optimize_insn_for_size_p instead of
6493 optimize_insn_for_speed_p.
6495 2014-04-17 Martin Jambor <mjambor@suse.cz>
6497 * gimple-iterator.c (gsi_start_edge): New function.
6498 * gimple-iterator.h (gsi_start_edge): Declare.
6499 * tree-sra.c (single_non_eh_succ): New function.
6500 (disqualify_ops_if_throwing_stmt): Renamed to
6501 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
6502 having one non-EH successor BB.
6503 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
6504 generate loads into replacements.
6505 (sra_modify_assign): Likewise and and also use the simple path for
6507 (sra_modify_function_body): Commit statements on edges.
6509 2014-04-17 Richard Biener <rguenther@suse.de>
6512 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
6513 comparison results and add clarifying comment.
6515 2014-04-17 Jakub Jelinek <jakub@redhat.com>
6517 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
6518 (blank_mode): Initialize it.
6519 (emit_mode_size_inline, emit_mode_nunits_inline,
6520 emit_mode_inner_inline): New functions.
6521 (emit_insn_modes_h): Call them and surround their output with
6522 #if GCC_VERSION >= 4001 ... #endif.
6523 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
6524 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
6525 mode_* arrays if the argument is __builtin_constant_p.
6526 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
6527 is enum machine_mode.
6529 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6531 * passes.c (opt_pass::execute): Adjust.
6532 (pass_manager::execute_pass_mode_switching): Likewise.
6533 (early_local_passes::execute): Likewise.
6534 (execute_one_pass): Pass cfun to the pass's execute method.
6535 * tree-pass.h (opt_pass::execute): Add function * argument.
6536 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
6537 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
6538 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
6539 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
6540 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
6541 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
6542 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
6543 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
6544 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
6545 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
6546 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
6547 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
6548 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
6549 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
6550 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
6551 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
6552 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
6553 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
6554 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
6555 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
6556 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
6557 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
6558 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
6559 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
6560 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
6561 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
6562 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
6563 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
6564 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
6567 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6569 * passes.c (opt_pass::gate): Take function * argument.
6570 (gate_all_early_local_passes): Merge into
6571 (early_local_passes::gate): this.
6572 (gate_all_early_optimizations): Merge into
6573 (all_early_optimizations::gate): this.
6574 (gate_all_optimizations): Mege into
6575 (all_optimizations::gate): this.
6576 (gate_all_optimizations_g): Merge into
6577 (all_optimizations_g::gate): this.
6578 (gate_rest_of_compilation): Mege into
6579 (rest_of_compilation::gate): this.
6580 (gate_postreload): Merge into
6581 (postreload::gate): this.
6582 (dump_one_pass): Pass cfun to the pass's gate method.
6583 (execute_ipa_summary_passes): Likewise.
6584 (execute_one_pass): Likewise.
6585 (ipa_write_summaries_2): Likewise.
6586 (ipa_write_optimization_summaries_1): Likewise.
6587 (ipa_read_summaries_1): Likewise.
6588 (ipa_read_optimization_summaries_1): Likewise.
6589 (execute_ipa_stmt_fixups): Likewise.
6590 * tree-pass.h (opt_pass::gate): Add function * argument.
6591 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
6592 combine-stack-adj.c, combine.c, compare-elim.c,
6593 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
6594 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
6595 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
6596 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
6597 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
6598 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
6599 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
6600 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
6601 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
6602 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
6603 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
6604 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
6605 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
6606 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
6607 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
6608 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
6609 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
6610 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
6611 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
6612 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
6613 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
6614 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
6615 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
6616 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
6617 var-tracking.c, vtable-verify.c, web.c: Adjust.
6619 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6621 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
6622 * configure: Regenerate.
6624 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6626 * passes.c (dump_one_pass): don't check pass->has_gate.
6627 (execute_ipa_summary_passes): Likewise.
6628 (execute_one_pass): Likewise.
6629 (ipa_write_summaries_2): Likewise.
6630 (ipa_write_optimization_summaries_1): Likewise.
6631 (ipa_read_optimization_summaries_1): Likewise.
6632 (execute_ipa_stmt_fixups): Likewise.
6633 * tree-pass.h (pass_data::has_gate): Remove.
6634 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
6635 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
6636 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
6637 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
6638 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
6639 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
6640 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
6641 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
6642 gimple-low.c, gimple-ssa-isolate-paths.c,
6643 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
6644 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
6645 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
6646 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
6647 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
6648 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
6649 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
6650 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
6651 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
6652 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
6653 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
6654 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
6655 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
6656 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
6657 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
6658 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
6659 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
6660 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
6661 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
6662 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
6663 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
6664 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
6667 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6669 * pass_manager.h (pass_manager::register_dump_files_1): Remove
6671 * passes.c (pass_manager::register_dump_files_1): Merge into
6672 (pass_manager::register_dump_files): this, and remove its handling of
6673 properties since the pass always has the properties anyway.
6674 (pass_manager::pass_manager): Adjust.
6676 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6678 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
6679 * passes.c (pass_manager::register_dump_files_1): Remove dead code
6680 dealing with properties.
6681 (pass_manager::register_dump_files): Adjust.
6683 2014-03-20 Mark Wielaard <mjw@redhat.com>
6685 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
6686 then represent the bound as normal constant value.
6688 2014-04-17 Jakub Jelinek <jakub@redhat.com>
6691 Forward port from 4.8 branch
6692 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
6694 * config/i386/bmiintrin.h (_blsi_u32): New.
6698 (_blsmsk_u32): Ditto.
6699 (_blsmsk_u64): Ditto.
6700 (_tzcnt_u32): Ditto.
6701 (_tzcnt_u64): Ditto.
6703 2014-04-17 Kito Cheng <kito@0xlab.org>
6705 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
6707 2014-04-17 Richard Biener <rguenther@suse.de>
6710 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
6711 boolean results for comparisons.
6713 2014-04-17 Richard Biener <rguenther@suse.de>
6715 PR tree-optimization/60836
6716 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
6717 initial PHI args to be gimple values.
6719 2014-04-17 Richard Biener <rguenther@suse.de>
6721 PR tree-optimization/60841
6722 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
6723 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
6724 of stmts to SLP build.
6725 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
6726 (vect_analyze_slp): Likewise.
6727 (vect_analyze_slp_instance): Likewise.
6728 (vect_build_slp_tree): Limit overall SLP tree growth.
6729 * tree-vectorizer.h (vect_analyze_data_refs,
6730 vect_analyze_slp): Adjust prototypes.
6732 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
6734 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
6737 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
6739 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
6740 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
6741 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
6742 for TARGET_SLOW_PSHUFB
6744 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
6746 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
6747 * config/i386/i386.c (intel_cost): Ditto.
6749 2014-04-17 Joey Ye <joey.ye@arm.com>
6751 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
6753 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
6755 * opts.c (common_handle_option): Disable -fipa-reference coorectly
6758 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
6760 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
6761 (type_all_derivations_known_p): New predicate.
6762 (type_all_ctors_visible_p): New predicate.
6763 (type_possibly_instantiated_p): New predicate.
6764 (get_odr_type): Compute all_derivations_known.
6765 (dump_odr_type): Dump the flag.
6766 (maybe_record_type): Cleanup.
6767 (record_target_from_binfo): Add bases_to_consider array;
6768 record bases for types w/o instances and skip CXX destructor.
6769 (possible_polymorphic_call_targets_1): Add bases_to_consider
6770 and consider_construction parameters; check if type may have instance.
6771 (get_polymorphic_call_info): Set maybe_in_construction to true
6772 when we know nothing.
6773 (record_targets_from_bases): Skip CXX destructors; they are
6774 never called for types in construction.
6775 (possible_polymorphic_call_targets): Do not record target when
6776 type may not have instance.
6778 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
6781 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
6782 external aliases alive, too.
6784 2014-04-16 Andrew Pinski <apinski@cavium.com>
6786 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
6789 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
6791 * final.c (compute_alignments): Do not apply loop alignment to a block
6792 falling through to the exit.
6794 2014-04-16 Catherine Moore <clm@codesourcery.com>
6796 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
6797 Adjust constraints for microMIPS store patterns.
6799 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
6801 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
6803 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
6805 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
6806 (append_use): Run at -O0.
6807 (append_vdef): Likewise.
6808 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
6809 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
6811 2014-04-16 Jakub Jelinek <jakub@redhat.com>
6813 PR tree-optimization/60844
6814 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
6815 (propagate_op_to_single_use, remove_visited_stmt_chain,
6816 linearize_expr, repropagate_negates, reassociate_bb): Use it
6817 instead of gsi_remove.
6819 2014-04-16 Martin Jambor <mjambor@suse.cz>
6821 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
6822 ipa_transforms_to_apply.
6823 (cgraph_function_versioning): Assert that old_node has empty
6824 ipa_transforms_to_apply.
6825 * trans-mem.c (ipa_tm_create_version): Likewise.
6826 * tree-inline.c (tree_function_versioning): Do not duplicate
6827 ipa_transforms_to_apply.
6829 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6832 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
6834 Pass necessary as flags on 64-bit Solaris/x86.
6835 Use lowercase relocs for x86_64-*-*.
6836 * configure: Regenerate.
6838 2014-04-15 Jan Hubicka <jh@suse.cz>
6840 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
6841 (maybe_record_node, likely_target_p): Use it.
6843 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6846 Revert following patch
6848 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
6851 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
6852 software floating point or no floating point registers, do not
6853 allow any type in the FPRs. Eliminate a test for SPE SIMD types
6854 in GPRs that occurs after we tested for GPRs that would never be
6857 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
6858 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
6859 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
6860 specifically allow DDmode, since that does not use the SPE SIMD
6863 2014-03-21 Mark Wielaard <mjw@redhat.com>
6865 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
6866 as unsigned or int depending on type and value used.
6868 2014-04-15 Richard Biener <rguenther@suse.de>
6870 PR rtl-optimization/56965
6871 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
6872 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
6874 * alias.c (true_dependence_1): Do not call
6875 nonoverlapping_component_refs_p.
6876 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
6877 nonoverlapping_component_refs_p.
6878 (indirect_refs_may_alias_p): Likewise.
6880 2014-04-15 Teresa Johnson <tejohnson@google.com>
6882 * cfg.c (dump_bb_info): Fix flags check.
6883 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
6885 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6887 PR rtl-optimization/60663
6888 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
6891 2014-04-15 Richard Biener <rguenther@suse.de>
6893 * lto-streamer.h (LTO_major_version): Bump to 4.
6895 2014-04-15 Richard Biener <rguenther@suse.de>
6897 * common.opt (lto_partition_model): New enum.
6898 (flto-partition=): Merge separate options with a single with argument,
6899 add -flto-partition=one support.
6900 * flag-types.h (enum lto_partition_model): Declare.
6901 * opts.c (finish_options): Remove duplicate -flto-partition=
6903 * lto-wrapper.c (run_gcc): Adjust.
6905 2014-04-15 Richard Biener <rguenther@suse.de>
6907 * alias.c (ncr_compar): New function.
6908 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
6910 2014-04-15 Richard Biener <rguenther@suse.de>
6912 * alias.c (record_component_aliases): Do not walk BINFOs.
6914 2014-04-15 Richard Biener <rguenther@suse.de>
6916 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
6917 Add struct function argument and adjust.
6918 (find_func_aliases_for_call): Likewise.
6919 (find_func_aliases): Likewise.
6920 (find_func_clobbers): Likewise.
6921 (intra_create_variable_infos): Likewise.
6922 (compute_points_to_sets): Likewise.
6923 (ipa_pta_execute): Adjust. Do not push/pop cfun.
6925 2014-04-15 Richard Biener <rguenther@suse.de>
6927 * tree.c (iterative_hash_expr): Use enum tree_code_class
6928 to store TREE_CODE_CLASS.
6929 (tree_block): Likewise.
6930 (tree_set_block): Likewise.
6931 * tree.h (fold_build_pointer_plus_loc): Use
6932 convert_to_ptrofftype_loc.
6934 2014-04-15 Jakub Jelinek <jakub@redhat.com>
6937 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
6940 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
6942 * cfgloop.h (struct loop): Move force_vectorize down.
6943 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
6944 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
6945 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
6946 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
6947 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
6948 * tree-core.h (enum annot_expr_kind): Add new kind values.
6949 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
6950 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
6952 * tree.def (ANNOTATE_EXPR): Tweak comment.
6954 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
6956 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
6959 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
6961 * tree.h (TYPE_IDENTIFIER): Declare.
6962 * tree.c (subrange_type_for_debug_p): Use it.
6963 * godump.c (go_format_type): Likewise.
6964 * dwarf2out.c (is_cxx_auto, modified_type_die,
6965 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
6966 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
6968 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
6971 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
6973 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
6975 * config/i386/i386.c (examine_argument): Return bool. Return true if
6976 parameter should be passed in memory.
6977 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
6978 (construct_container): Update calls to examine_argument.
6979 (function_arg_advance_64): Ditto.
6980 (return_in_memory_32): Merge with ix86_return_in_memory.
6981 (return_in_memory_64): Ditto.
6982 (return_in_memory_ms_64): Ditto.
6984 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
6986 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
6987 * coverage.c (coverage_compute_profile_id): Handle externally visible
6990 2014-04-14 Martin Jambor <mjambor@suse.cz>
6992 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
6993 DECL_DISREGARD_INLINE_LIMITS functions.
6995 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
6998 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
7000 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
7003 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
7004 optimize_insn_for_speed_p instead of
7005 optimize_function_for_speed_p.
7007 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
7009 * doc/invoke.texi (free): Document AArch64.
7011 2014-04-14 Richard Biener <rguenther@suse.de>
7013 PR tree-optimization/60042
7014 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
7015 (insert_into_preds_of_block): Do not prevent PHI insertion
7016 for REFERENCE exprs here ...
7017 (eliminate_dom_walker::before_dom_children): ... but prevent
7018 their use here under similar conditions when applied to the
7019 IL after PRE optimizations.
7021 2014-04-14 Richard Biener <rguenther@suse.de>
7023 * passes.def: Move early points-to after early SRA.
7025 2014-04-14 Richard Biener <rguenther@suse.de>
7027 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
7028 check for which sign-changes we allow when forwarding
7029 a converted value into a switch.
7031 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
7033 * stor-layout.c (place_field): Finalize non-constant offset for the
7036 2014-04-14 Richard Biener <rguenther@suse.de>
7038 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
7040 (expand_switch_using_bit_tests_p): Likewise.
7041 (process_switch): Compute and pass on speed_p based on the
7043 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
7044 optimize_bb_for_speed_p.
7046 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
7048 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
7049 * function.h (struct function): Rename has_force_vect_loops into
7050 has_force_vectorize_loops.
7051 * lto-streamer-in.c (input_cfg): Adjust for renaming.
7052 (input_struct_function_base): Likewise.
7053 * lto-streamer-out.c (output_cfg): Likewise.
7054 (output_struct_function_base): Likewise.
7055 * omp-low.c (expand_omp_simd): Likewise.
7056 * tree-cfg.c (move_sese_region_to_fn): Likewise.
7057 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
7058 (version_loop_for_if_conversion): Likewise.
7059 (tree_if_conversion): Likewise.
7060 (main_tree_if_conversion): Likewise.
7061 (gate_tree_if_conversion): Likewise.
7062 * tree-inline.c (copy_loops): Likewise.
7063 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
7064 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
7065 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
7066 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
7067 * tree-vectorizer.c (vectorize_loops): Likewise.
7068 * tree-vectorizer.h (unlimited_cost_model): Likewise.
7070 2014-04-14 Richard Biener <rguenther@suse.de>
7073 * lto-streamer-out.c (wrap_refs): New function.
7074 (lto_output): Wrap symbol references in global initializes in
7075 type-preserving MEM_REFs.
7077 2014-04-14 Christian Bruel <christian.bruel@st.com>
7079 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
7081 2014-04-14 Christian Bruel <christian.bruel@st.com>
7083 * config/sh/sh.md (setmemqi): New expand pattern.
7084 * config/sh/sh.h (CLEAR_RATIO): Define.
7085 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
7086 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
7088 2014-04-14 Richard Biener <rguenther@suse.de>
7091 * fold-const.c (negate_expr_p): Don't negate directional rounding
7093 (fold_negate_expr): Likewise.
7095 2014-04-14 Richard Biener <rguenther@suse.de>
7097 PR tree-optimization/59817
7098 PR tree-optimization/60453
7099 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
7100 recursion to catch all CHRECs in the scalar evolution and restrict
7101 the predicate for the remains appropriately.
7103 2014-04-12 Catherine Moore <clm@codesourcery.com>
7105 * config/mips/constraints.md: Add new register constraint "kb".
7106 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
7107 (*movhi_internal): Likewise.
7108 (*movqi_internal): Likewise.
7109 * config/mips/mips.h (M16_STORE_REGS): New register class.
7110 (REG_CLASS_NAMES): Add M16_STORE_REGS.
7111 (REG_CLASS_CONTENTS): Likewise.
7112 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
7114 2014-04-11 Tobias Burnus <burnus@net-b.de>
7117 * doc/invoke.texi (-Wformat-signedness): Document it.
7118 (Wformat=2): Mention that this enables -Wformat-signedness.
7120 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
7122 * common/config/epiphany/epiphany-common.c
7123 (epiphany_option_optimization_table): Enable section anchors by
7124 default at -O1 or higher.
7125 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
7126 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
7127 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
7128 carries no extra cost.
7129 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
7130 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
7131 * config/epiphany/predicates.md (memclob_operand): New predicate.
7132 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
7133 Use memclob_operand predicate and X constraint for operand 3.
7135 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
7137 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
7138 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
7141 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
7143 PR rtl-optimization/60651
7144 * mode-switching.c (optimize_mode_switching): Make sure to emit
7145 sets of a lower numbered entity before sets of a higher numbered
7146 entity to a mode of the same or lower priority.
7147 When creating a seginfo for a basic block that starts with a code
7148 label, move the insertion point past the code label.
7149 (new_seginfo): Document and enforce requirement that
7150 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
7151 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
7152 * doc/tm.texi: Regenerate.
7154 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
7157 * config/arc/arc.c (arc_save_restore): Fix assert typo.
7159 2013-04-11 Jakub Jelinek <jakub@redhat.com>
7161 * BASE-VER: Set to 4.10.0.
7163 2014-04-11 Tobias Burnus <burnus@net-b.de>
7166 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
7167 * doc/gcc.texi (Service): Update description in the @menu
7168 * doc/invoke.texi (Option Summary): Remove misplaced and
7171 2014-04-11 Steve Ellcey <sellcey@mips.com>
7172 Jakub Jelinek <jakub@redhat.com>
7175 * expr.c (convert_move): Use emit_store_flag_force instead of
7176 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
7179 2014-04-11 Richard Biener <rguenther@suse.de>
7182 * varasm.c (assemble_alias): Avoid endless error reporting
7183 recursion by setting TREE_ASM_WRITTEN.
7185 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7187 * config/s390/s390.md: Add a splitter for NOT rtx.
7189 2014-04-11 Jakub Jelinek <jakub@redhat.com>
7191 PR rtl-optimization/60663
7192 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
7194 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
7195 Jakub Jelinek <jakub@redhat.com>
7198 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
7199 flag from decl_node to node.
7201 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7204 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
7205 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
7206 ameliorating the cases where it can be.
7208 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
7211 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
7213 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
7214 (loadsync_<mode>): Change mode.
7215 (load_quadpti, store_quadpti): New.
7216 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
7217 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
7218 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
7220 2014-04-09 Cong Hou <congh@google.com>
7223 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
7226 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7228 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
7229 instead of vnor to exploit possible fusion opportunity in the
7231 (altivec_expand_vec_perm_const_le): Likewise.
7233 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
7235 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
7236 (loadsync_<mode>): Change mode.
7237 (load_quadpti, store_quadpti): New.
7238 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
7239 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
7241 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
7244 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
7245 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
7246 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
7248 2014-04-08 Richard Biener <rguenther@suse.de>
7251 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
7252 a 64bit widest int print double-int similar to on HWI64 hosts.
7254 2014-04-08 Richard Biener <rguenther@suse.de>
7256 PR tree-optimization/60785
7257 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
7258 default defs properly.
7260 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
7262 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
7263 (Weffc++): Likewise.
7265 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
7267 * ipa-devirt.c (maybe_record_node): When node is not recorded,
7268 set completep to false rather than true.
7270 2014-04-07 Douglas B Rupp <rupp@adacore.com>
7273 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
7274 ARM_TARGET2_DWARF_FORMAT.
7276 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
7279 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
7280 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
7283 2014-04-07 Richard Biener <rguenther@suse.de>
7285 PR tree-optimization/60766
7286 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
7287 (may_eliminate_iv): Convert cand_value_at result to desired type.
7289 2014-04-07 Jason Merrill <jason@redhat.com>
7292 * common.opt (-fno-gnu-unique): Add.
7293 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
7295 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7297 * haifa-sched.c: Fix outdated function reference and minor
7298 grammar errors in introductory comment.
7300 2014-04-07 Richard Biener <rguenther@suse.de>
7303 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
7305 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
7307 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
7310 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
7311 size accounting for thunks.
7312 (pa_asm_output_mi_thunk): Use final_start_function() and
7313 final_end_function() to output function start and end directives.
7315 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
7317 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
7318 device specific ISA/ feature information. Remove short_sp and
7319 errata_skip ds. Add avr_device_specific_features enum to have device
7321 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
7322 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
7323 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
7324 updated device specific info.
7325 * config/avr/avr-mcus.def: Merge device specific details to
7326 dev_attribute field.
7327 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
7329 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
7330 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
7331 assembler if RMW isa supported by current device.
7332 * config/avr/genmultilib.awk: Update as device info structure changed.
7333 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
7335 2014-04-04 Cong Hou <congh@google.com>
7337 PR tree-optimization/60656
7338 * tree-vect-stmts.c (supportable_widening_operation):
7339 Fix a bug that elements in a vector with vect_used_by_reduction
7340 property are incorrectly reordered when the operation on it is not
7341 consistant with the one in reduction operation.
7343 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
7345 PR rtl-optimization/60155
7346 * gcse.c (record_set_data): New function.
7347 (single_set_gcse): New function.
7348 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
7349 (hoist_code): Likewise.
7350 (get_pressure_class_and_nregs): Likewise.
7352 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
7354 * explow.c (probe_stack_range): Emit a final optimization blockage.
7356 2014-04-04 Anthony Green <green@moxielogic.com>
7358 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
7361 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
7364 * lto-cgraph.c (input_overwrite_node): Check that partitioning
7365 flags are set only during streaming.
7366 * ipa.c (process_references, walk_polymorphic_call_targets,
7367 symtab_remove_unreachable_nodes): Drop bodies of always inline
7368 after early inlining.
7369 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
7371 2014-04-04 Jakub Jelinek <jakub@redhat.com>
7372 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7375 * dwarf2out.c (const_ok_for_output_1): Reject expressions
7378 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7381 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
7383 (cortex_a53_fdivd): Likewise.
7385 2014-04-04 Martin Jambor <mjambor@suse.cz>
7388 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
7390 * cgraph.c (clone_of_p): Also return true if thunks match.
7391 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
7392 cgraph_function_or_thunk_node and an obsolete comment.
7393 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
7395 (build_function_decl_skip_args): Likewise.
7396 (set_new_clone_decl_and_node_flags): New function.
7397 (duplicate_thunk_for_node): Likewise.
7398 (redirect_edge_duplicating_thunks): Likewise.
7399 (cgraph_clone_node): New parameter args_to_skip, pass it to
7400 redirect_edge_duplicating_thunks which is called instead of
7401 cgraph_redirect_edge_callee.
7402 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
7403 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
7405 2014-04-04 Jeff Law <law@redhat.com>
7408 * config/arm/predicates.md (const_int_I_operand): New predicate.
7409 (const_int_M_operand): Similarly.
7410 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
7412 (insv_t2, extv_reg, extzv_t2): Likewise.
7413 (load_multiple_with_writeback): Similarly for const_int_I_operand.
7414 (pop_multiple_with_writeback_and_return): Likewise.
7415 (vfp_pop_multiple_with_writeback): Likewise
7417 2014-04-04 Richard Biener <rguenther@suse.de>
7420 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
7421 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
7423 * gimplify.h (gimple_add_tmp_var_fn): Declare.
7424 * gimplify.c (gimple_add_tmp_var_fn): New function.
7425 * gimple-expr.h (create_tmp_reg_fn): Declare.
7426 * gimple-expr.c (create_tmp_reg_fn): New function.
7427 * gimple-low.c (record_vars_into): Don't change cfun.
7428 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
7429 code generation without cfun.
7431 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
7434 * Makefile.in (install-driver): Fix shell scripting.
7436 2014-04-03 Cong Hou <congh@google.com>
7438 PR tree-optimization/60505
7439 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
7440 threshold of number of iterations below which no vectorization
7442 * tree-vect-loop.c (new_loop_vec_info):
7443 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
7444 * tree-vect-loop.c (vect_analyze_loop_operations):
7445 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
7446 * tree-vect-loop.c (vect_transform_loop):
7447 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
7448 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
7449 of iterations of the loop and see if we should build the epilogue.
7451 2014-04-03 Richard Biener <rguenther@suse.de>
7453 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
7454 (streamer_tree_cache_create): Adjust.
7455 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
7456 to allow optional nodes array.
7457 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
7458 (streamer_tree_cache_append): Likewise.
7459 (streamer_tree_cache_create): Create nodes array optionally
7460 as specified by parameter.
7461 * lto-streamer-out.c (create_output_block): Avoid maintaining
7462 the node array in the writer cache.
7463 (DFS_write_tree): Remove assertion.
7464 (produce_asm_for_decls): Free the out decl state hash table early.
7465 * lto-streamer-in.c (lto_data_in_create): Adjust for
7466 streamer_tree_cache_create prototype change.
7468 2014-04-03 Richard Biener <rguenther@suse.de>
7470 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
7471 set TREE_CHAIN to NULL_TREE.
7473 2014-04-03 Richard Biener <rguenther@suse.de>
7475 PR tree-optimization/60740
7476 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
7477 over all GIMPLE_COND operands.
7479 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
7481 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
7482 (Weffc++): Remove Scott's numbering, merge lists and reference
7485 2014-04-03 Nick Clifton <nickc@redhat.com>
7487 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
7490 2014-04-03 Martin Jambor <mjambor@suse.cz>
7492 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
7493 mention gcc_unreachable before failing.
7494 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
7497 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
7500 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
7501 inconsistent code and instead mark the context inconsistent.
7502 (possible_polymorphic_call_targets): For inconsistent contexts
7503 return empty complete list.
7505 2014-04-02 Anthony Green <green@moxielogic.com>
7507 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
7508 (extendqisi2, extendhisi2): Define.
7509 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
7510 (WCHAR_TYPE): Change to unsigned int.
7512 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7514 PR tree-optimization/60733
7515 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
7516 insertion point for PHI candidates to be the end of the feeding
7517 block for the PHI argument.
7519 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
7521 PR rtl-optimization/60650
7522 * lra-constraints.c (process_alt_operands): Decrease reject for
7523 earlyclobber matching.
7525 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7527 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
7529 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7531 * config/spu/spu.c (pad_bb): Do not crash when the last
7532 insn is CODE_FOR_blockage.
7534 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7536 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
7537 lies outside the target mode.
7539 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
7542 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
7543 software floating point or no floating point registers, do not
7544 allow any type in the FPRs. Eliminate a test for SPE SIMD types
7545 in GPRs that occurs after we tested for GPRs that would never be
7548 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
7549 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
7550 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
7551 specifically allow DDmode, since that does not use the SPE SIMD
7554 2014-04-02 Richard Biener <rguenther@suse.de>
7557 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
7558 MODE_INTs. Properly use negv_optab.
7559 (expand_abs): Likewise.
7561 2014-04-02 Richard Biener <rguenther@suse.de>
7564 * Makefile.in (install-driver): Guard extra installs with special
7567 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
7569 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
7572 2014-04-01 Richard Henderson <rth@redhat.com>
7575 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
7576 alternative enabled before register allocation.
7578 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
7580 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
7581 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
7583 (nios2_large_got_address): Remove unneeded 'sym' parameter.
7584 (nios2_got_address): Update nios2_large_got_address call site.
7585 (nios2_delegitimize_address): New function.
7586 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
7587 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
7588 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
7590 2014-04-01 Martin Husemann <martin@duskware.de>
7592 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
7595 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
7597 PR rtl-optimization/60604
7598 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
7599 check from register_operand.
7600 (register_operand): Redefine in terms of general_operand.
7601 (nonmemory_operand): Use register_operand for the non-constant cases.
7603 2014-04-01 Richard Biener <rguenther@suse.de>
7605 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
7607 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
7609 * doc/invoke.texi (mapp-regs): Clarify.
7611 2014-03-31 Ulrich Drepper <drepper@gmail.com>
7613 * config/i386/avx512fintrin.h (__v32hi): Define type.
7614 (__v64qi): Likewise.
7615 (_mm512_set1_epi8): Define.
7616 (_mm512_set1_epi16): Define.
7617 (_mm512_set4_epi32): Define.
7618 (_mm512_set4_epi64): Define.
7619 (_mm512_set4_pd): Define.
7620 (_mm512_set4_ps): Define.
7621 (_mm512_setr4_epi64): Define.
7622 (_mm512_setr4_epi32): Define.
7623 (_mm512_setr4_pd): Define.
7624 (_mm512_setr4_ps): Define.
7625 (_mm512_setzero_epi32): Define.
7627 2014-03-31 Martin Jambor <mjambor@suse.cz>
7630 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
7631 callsite_arguments_match_p. Updated all callers. Also check types of
7632 corresponding formal parameters and actual arguments.
7633 (not_all_callers_have_enough_arguments_p) Renamed to
7634 some_callers_have_mismatched_arguments_p.
7636 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
7638 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
7640 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
7643 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
7646 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
7648 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
7649 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
7651 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
7652 Use FMAMODE_NOVF512 mode iterator.
7653 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
7654 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
7655 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
7656 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
7658 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
7659 Use VF_128_256 mode iterator.
7660 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
7663 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
7665 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
7666 static chain if needed.
7668 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
7671 * lra-constraints.c (index_part_to_reg): New.
7672 (process_address): Use it.
7674 2014-03-27 Jeff Law <law@redhat.com>
7675 Jakub Jelinek <jakub@redhat.com>
7678 * expr.c (do_tablejump): Use simplify_gen_binary rather than
7679 gen_rtx_{PLUS,MULT} to build up the address expression.
7681 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
7682 creating non-canonical RTL.
7684 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
7687 * ipa-inline.c (want_inline_small_function_p): Short circuit large
7688 functions; reorganize to make cheap checks first.
7689 (inline_small_functions): Do not estimate growth when dumping;
7691 * ipa-inline.h (inline_summary): Add min_size.
7692 (growth_likely_positive): New function.
7693 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
7694 (set_cond_stmt_execution_predicate): Cleanup.
7695 (estimate_edge_size_and_time): Compute min_size.
7696 (estimate_calls_size_and_time): Likewise.
7697 (estimate_node_size_and_time): Likewise.
7698 (inline_update_overall_summary): Update min_size.
7699 (do_estimate_edge_time): Likewise.
7700 (do_estimate_edge_size): Update.
7701 (do_estimate_edge_hints): Update.
7702 (growth_likely_positive): New function.
7704 2014-03-28 Jakub Jelinek <jakub@redhat.com>
7707 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
7708 also if addr has VOIDmode.
7710 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7712 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
7713 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
7715 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
7716 instructions as well as AdvancedSIMD loads.
7718 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7720 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
7721 Use crypto_aese type.
7722 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
7723 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
7724 crypto_aese, crypto_aesmc. Move to types.md.
7725 * config/arm/types.md (crypto_aes): Split into crypto_aese,
7727 * config/arm/iterators.md (crypto_type): Likewise.
7729 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
7731 * cgraph.c: Include expr.h and tree-dfa.h.
7732 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
7735 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
7738 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
7739 regs from checking multi-reg pseudos.
7741 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7743 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
7745 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7747 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
7748 if it would clobber the stack pointer, even temporarily.
7750 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
7752 * mode-switching.c: Make small adjustments to the top comment.
7754 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
7756 * config/rs6000/constraints.md (wD constraint): New constraint to
7757 match the constant integer to get the top DImode/DFmode out of a
7758 vector in a VSX register.
7760 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
7761 match the constant integer to get the top DImode/DFmode out of a
7762 vector in a VSX register.
7764 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
7767 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
7770 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
7771 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
7773 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
7774 Optimize vec_extract of 64-bit values, where the value being
7775 extracted is in the top word, where we can use scalar
7776 instructions. Add direct move and store support. Combine the big
7777 endian/little endian vector select load support into a single insn.
7778 (vsx_extract_<mode>_internal1): Likewise.
7779 (vsx_extract_<mode>_internal2): Likewise.
7780 (vsx_extract_<mode>_load): Likewise.
7781 (vsx_extract_<mode>_store): Likewise.
7782 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
7783 combined into vsx_extract_<mode>_load.
7784 (vsx_extract_<mode>_one_le): Likewise.
7786 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
7787 define the top 64-bit vector element.
7789 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
7792 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
7793 Document vec_vbpermq builtin.
7796 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
7797 enable use of xxsldwi and xxpermdi builtin functions.
7798 (vec_xxpermdi): Likewise.
7800 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
7801 Document use of vec_xxsldwi and vec_xxpermdi builtins.
7803 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
7805 PR rtl-optimization/60650
7806 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
7808 (find_spills_for): New.
7809 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
7810 Spill all pseudos on the second iteration.
7812 2014-03-27 Marek Polacek <polacek@redhat.com>
7815 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
7818 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7820 * config/s390/s390.c (s390_can_use_return_insn): Check for
7821 call-saved FPRs on 31 bit.
7823 2014-03-27 Jakub Jelinek <jakub@redhat.com>
7826 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
7827 if they need regimplification, just drop them instead of
7828 calling gimple_regimplify_operands on them.
7830 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
7833 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
7834 (aarch64_frame_pointer_required): Adjust logic.
7835 (aarch64_can_eliminate): Adjust logic.
7836 (aarch64_override_options_after_change): Adjust logic.
7838 2014-03-27 Dehao Chen <dehao@google.com>
7840 * ipa-inline.c (early_inliner): Update node's inline info.
7842 2014-03-26 Dehao Chen <dehao@google.com>
7844 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
7845 compiler inserted conditional jumps for NAN float check.
7847 2014-03-26 Jakub Jelinek <jakub@redhat.com>
7849 * ubsan.h (ubsan_create_data): Change second argument's type
7850 to const location_t *.
7851 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
7853 (ubsan_create_data): Change second argument to const location_t *PLOC.
7854 Create Loc field whenever PLOC is non-NULL.
7855 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
7856 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
7860 * real.c (real_to_integer2): Change type of low to UHWI.
7862 2014-03-26 Tobias Burnus <burnus@net-b.de>
7864 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
7865 (CILK_SELF_SPECS): New define.
7866 (driver_self_specs): Use it.
7868 2014-03-26 Richard Biener <rguenther@suse.de>
7870 * tree-pretty-print.c (percent_K_format): Implement special
7871 case for LTO and its stripped down BLOCK tree.
7873 2014-03-26 Jakub Jelinek <jakub@redhat.com>
7876 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
7878 * tree-vrp.c (simplify_internal_call_using_ranges): If only
7879 one range is range_int_cst_p, but not both, at least optimize
7880 addition/subtraction of 0 and multiplication by 0 or 1.
7881 * gimple-fold.c (gimple_fold_call): Fold
7882 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
7883 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
7884 INTEGER_CSTs, try to fold at least x * 0 and y - y.
7886 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
7888 PR rtl-optimization/60452
7889 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
7890 <case REG>: Return 1 for invalid offsets from the frame pointer.
7892 2014-03-26 Marek Polacek <polacek@redhat.com>
7895 * doc/extend.texi (C Extensions): Mention variable-length arrays in
7898 2014-03-26 Marek Polacek <polacek@redhat.com>
7901 * doc/extend.texi (Designated Inits): Describe what happens to omitted
7904 2014-03-26 Marek Polacek <polacek@redhat.com>
7907 * ira-color.c (update_conflict_hard_regno_costs): Perform the
7908 multiplication in unsigned type.
7910 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
7912 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
7914 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
7916 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
7918 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
7921 * cif-code.def (UNREACHABLE) New code.
7922 * ipa-inline.c (inline_small_functions): Skip edges to
7923 __builtlin_unreachable.
7924 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
7925 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
7926 predicate to __bulitin_unreachable.
7927 (set_cond_stmt_execution_predicate): Fix issue when
7928 invert_tree_comparison returns ERROR_MARK.
7929 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
7930 propagate to inline clones.
7931 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
7933 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
7934 * cgraphclones.c (cgraph_clone_node): If call destination is already
7935 ureachable, do not redirect it back.
7936 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
7939 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
7941 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
7942 Do not modify inline clones.
7944 2014-03-25 Jakub Jelinek <jakub@redhat.com>
7946 * config/i386/i386.md (general_sext_operand): New mode attr.
7947 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
7948 don't generate (sign_extend (const_int)).
7949 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
7950 operands[2]. Use We constraint instead of <i> and
7951 <general_sext_operand> predicate instead of <general_operand>.
7952 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
7953 * config/i386/constraints.md (We): New constraint.
7954 * config/i386/predicates.md (x86_64_sext_operand,
7955 sext_operand): New predicates.
7957 2014-03-25 Martin Jambor <mjambor@suse.cz>
7960 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
7961 inconsistent devirtualizations to __builtin_unreachable.
7963 2014-03-25 Marek Polacek <polacek@redhat.com>
7966 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
7968 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
7970 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
7971 order of elements for big-endian.
7973 2014-03-25 Richard Biener <rguenther@suse.de>
7976 * gimplify-me.c (gimple_regimplify_operands): Update the
7979 2014-03-25 Martin Jambor <mjambor@suse.cz>
7982 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
7983 (lto_output_varpool_node): Likewise.
7984 (input_overwrite_node): Likewise.
7985 (input_varpool_node): Likewise.
7987 2014-03-25 Richard Biener <rguenther@suse.de>
7989 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
7990 (run_gcc): Likewise.
7992 2014-03-25 Jakub Jelinek <jakub@redhat.com>
7994 * combine.c (simplify_compare_const): Add MODE argument.
7995 Handle mode_width 0 as very large mode_width.
7996 (try_combine, simplify_comparison): Adjust callers.
7998 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
7999 type to avoid signed integer overflow.
8000 * explow.c (plus_constant): Likewise.
8002 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
8004 * doc/generic.texi: Correct typos.
8006 2014-03-24 Tobias Burnus <burnus@net-b.de>
8008 * doc/invoke.texi (-flto): Expand section about
8009 using static libraries with LTO.
8011 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8013 PR rtl-optimization/60501
8014 * optabs.def (addptr3_optab): New optab.
8015 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
8016 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
8017 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
8019 * lra.c (emit_add3_insn): Use the addptr pattern if available.
8021 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
8023 2014-03-24 Ulrich Drepper <drepper@gmail.com>
8025 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
8028 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
8029 (_mm256_undefined_ps): Define.
8030 (_mm256_undefined_pd): Define.
8031 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
8032 (_mm_undefined_pd): Define.
8033 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
8034 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
8035 (_mm512_undefined_ps): Define.
8036 (_mm512_undefined_pd): Define.
8037 Use _mm*_undefined_*.
8038 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
8040 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
8042 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
8043 (lshr_simd): DI mode added.
8044 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
8045 (aarch64_ushr_simddi): Likewise.
8046 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
8047 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
8048 (vshrd_n_u64): Likewise.
8050 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8052 * Makefile.in (s-macro_list): Depend on cc1.
8054 2014-03-23 Teresa Johnson <tejohnson@google.com>
8056 * ipa-utils.c (ipa_print_order): Use specified dump file.
8058 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
8060 PR rtl-optimization/60601
8061 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
8063 * gcc.c (eval_spec_function): Initialize save_growing_value.
8065 2014-03-22 Jakub Jelinek <jakub@redhat.com>
8068 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
8069 code == MINUS_EXPR, never swap op0 with op1.
8071 * toplev.c (init_local_tick): Avoid signed integer multiplication
8073 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
8074 shift by first operand's bitsize.
8076 2014-03-21 Jakub Jelinek <jakub@redhat.com>
8079 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
8081 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
8082 TARGET_ISA_64BIT_P(x).
8084 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8086 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
8087 pattern for vector nor instead of subtract from splat(-1).
8088 (altivec_expand_vec_perm_const_le): Likewise.
8090 2014-03-21 Richard Henderson <rth@twiddle.net>
8093 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
8094 related insns after epilogue_completed.
8096 2014-03-21 Martin Jambor <mjambor@suse.cz>
8099 * cgraph.h (symtab_node): New flag body_removed.
8100 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
8101 when removing bodies.
8102 * symtab.c (dump_symtab_base): Dump body_removed flag.
8103 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
8104 had their bodies removed.
8106 2014-03-21 Martin Jambor <mjambor@suse.cz>
8109 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
8112 2014-03-21 Richard Biener <rguenther@suse.de>
8114 PR tree-optimization/60577
8115 * tree-core.h (struct tree_base): Document nothrow_flag use
8116 in VAR_DECL_NONALIASED.
8117 * tree.h (VAR_DECL_NONALIASED): New.
8118 (may_be_aliased): Adjust.
8119 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
8121 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
8123 * expr.c (expand_expr_real_1): Remove outdated comment.
8125 2014-03-20 Jakub Jelinek <jakub@redhat.com>
8128 * ira.c (adjust_cleared_regs): Call copy_rtx on
8129 *reg_equiv[REGNO (loc)].src_p before passing it to
8130 simplify_replace_fn_rtx.
8133 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
8134 into CONST, put pic register as first operand of PLUS. Use
8135 gen_const_mem for both 32-bit and 64-bit PIC got loads.
8137 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8139 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
8141 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
8143 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
8144 around for store forwarding issue in the FPU on the UT699.
8145 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
8146 loads and operations if -mfix-ut699 is specified.
8147 (divtf3_hq): Tweak attribute.
8148 (sqrttf2_hq): Likewise.
8150 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
8152 * calls.c (store_one_arg): Remove incorrect const qualification on the
8153 type of the temporary.
8154 * cfgexpand.c (expand_return): Likewise.
8155 * expr.c (expand_constructor): Likewise.
8156 (expand_expr_real_1): Likewise.
8158 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
8160 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
8163 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
8166 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
8168 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
8170 * config/arm/aarch-common-protos.h
8171 (alu_cost_table): Fix spelling of "extend".
8172 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
8174 2014-03-19 Richard Biener <rguenther@suse.de>
8177 * tree-core.h (tree_type_common): Re-order pointer members
8178 to reduce recursion depth during GC walks.
8180 2014-03-19 Marek Polacek <polacek@redhat.com>
8183 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
8184 before accessing it.
8186 2014-03-19 Richard Biener <rguenther@suse.de>
8189 * lto-streamer-in.c (input_function): In WPA stage do not drop
8192 2014-03-19 Jakub Jelinek <jakub@redhat.com>
8194 PR tree-optimization/60559
8195 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
8196 with build_zero_cst assignment.
8198 2014-03-18 Kai Tietz <ktietz@redhat.com>
8200 PR rtl-optimization/56356
8201 * sdbout.c (sdbout_parms): Verify that parms'
8202 incoming argument is valid.
8203 (sdbout_reg_parms): Likewise.
8205 2014-03-18 Richard Henderson <rth@redhat.com>
8208 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
8209 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
8210 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
8212 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
8214 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
8215 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
8216 Italicize plugin event names in description. Explain that
8217 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
8218 Remind that no GCC functions should be called after PLUGIN_FINISH.
8219 Explain what pragmas with expansion are.
8221 2014-03-18 Martin Liska <mliska@suse.cz>
8223 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
8224 gimple call statement is update.
8225 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
8226 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
8228 2014-03-18 Jakub Jelinek <jakub@redhat.com>
8231 * ubsan.c (ubsan_instrument_unreachable): Call
8232 initialize_sanitizer_builtins.
8233 (ubsan_pass): Likewise.
8236 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
8237 varpool_finalize_decl instead of rest_of_decl_compilation.
8239 2014-03-18 Richard Biener <rguenther@suse.de>
8241 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
8242 by using bitmap_and_compl instead of bitmap_and_compl_into.
8243 (df_rd_transfer_function): Likewise.
8245 2014-03-18 Richard Biener <rguenther@suse.de>
8247 * doc/lto.texi (fresolution): Fix typo.
8249 2014-03-18 Richard Biener <rguenther@suse.de>
8251 * doc/invoke.texi (flto): Update for changes in 4.9.
8253 2014-03-18 Richard Biener <rguenther@suse.de>
8255 * doc/loop.texi: Remove section on the removed lambda framework.
8256 Update loop docs with recent changes in preserving loop structure.
8258 2014-03-18 Richard Biener <rguenther@suse.de>
8260 * doc/lto.texi (-fresolution): Document.
8262 2014-03-18 Richard Biener <rguenther@suse.de>
8264 * doc/contrib.texi: Adjust my name.
8266 2014-03-18 Jakub Jelinek <jakub@redhat.com>
8269 * internal-fn.c: Include diagnostic-core.h.
8270 (expand_BUILTIN_EXPECT): New function.
8271 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
8272 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
8273 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
8274 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
8276 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
8277 Revert 3 argument __builtin_expect code.
8278 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
8279 * gimple-fold.c (gimple_fold_call): Likewise.
8280 * tree.h (fold_builtin_expect): New prototype.
8281 * builtins.c (build_builtin_expect_predicate): Add predictor
8282 argument, if non-NULL, create 3 argument __builtin_expect.
8283 (fold_builtin_expect): No longer static. Add ARG2 argument,
8284 pass it through to build_builtin_expect_predicate.
8285 (fold_builtin_2): Adjust caller.
8286 (fold_builtin_3): Handle BUILT_IN_EXPECT.
8287 * internal-fn.def (BUILTIN_EXPECT): New.
8289 2014-03-18 Tobias Burnus <burnus@net-b.de>
8292 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
8293 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
8294 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
8296 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
8299 * predict.c (combine_predictions_for_bb): Fix up formatting.
8300 (expr_expected_value_1, expr_expected_value): Add predictor argument,
8301 fill what it points to if non-NULL.
8302 (tree_predict_by_opcode): Adjust caller, use the predictor.
8303 * predict.def (PRED_COMPARE_AND_SWAP): Add.
8305 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
8307 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
8308 proper constant for the store mode.
8310 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
8312 * symtab.c (change_decl_assembler_name): Fix transparent alias
8315 2014-03-16 Renlin Li <Renlin.Li@arm.com>
8317 * config/aarch64/aarch64.c: Correct the comments about the
8318 aarch64 stack layout.
8320 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
8322 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
8323 check for GF_OMP_FOR_KIND_FOR.
8325 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
8327 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
8328 ymm and zmm register names.
8330 2014-03-17 Jakub Jelinek <jakub@redhat.com>
8333 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
8334 note creation for the 2010-08-31 changes.
8336 2014-03-17 Marek Polacek <polacek@redhat.com>
8339 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
8340 as -fno-tree-loop-vectorize.
8341 (expand_omp_simd): Likewise.
8343 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
8345 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
8346 (eligible_for_call_delay): New prototype.
8347 * config/sparc/sparc.c (tls_call_delay): Rename into...
8348 (eligible_for_call_delay): ...this. Return false if the instruction
8349 cannot be put in the delay slot of a branch.
8350 (eligible_for_restore_insn): Simplify.
8351 (eligible_for_return_delay): Return false if the instruction cannot be
8352 put in the delay slot of a branch and simplify.
8353 (eligible_for_sibcall_delay): Return false if the instruction cannot be
8354 put in the delay slot of a branch.
8355 * config/sparc/sparc.md (fix_ut699): New attribute.
8356 (tls_call_delay): Delete.
8357 (in_call_delay): Reimplement.
8358 (eligible_for_sibcall_delay): Rename into...
8359 (in_sibcall_delay): ...this.
8360 (eligible_for_return_delay): Rename into...
8361 (in_return_delay): ...this.
8362 (in_branch_delay): Reimplement.
8363 (in_uncond_branch_delay): Delete.
8364 (in_annul_branch_delay): Delete.
8366 2014-03-14 Richard Henderson <rth@redhat.com>
8369 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
8370 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
8371 (*floathi<X87MODEF>2_i387_with_temp): Remove.
8372 (floathi splitters): Remove.
8373 (float<SWI48x>xf2): New pattern.
8374 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
8375 code that tried to handle DImode for 32-bit, but which was excluded
8376 by the pattern's condition. Drop allocation of stack temporary.
8377 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
8378 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
8379 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
8380 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
8381 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
8382 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
8383 (*float<SWI48><MODEF>2_sse_interunit): Remove.
8384 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
8385 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
8386 (*float<SWI48x><X87MODEF>2_i387): Remove.
8387 (all float _with_temp splitters): Remove.
8388 (*float<SWI48x><MODEF>2_i387): New pattern.
8389 (*float<SWI48><MODEF>2_sse): New pattern.
8390 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
8391 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
8393 2014-03-14 Jakub Jelinek <jakub@redhat.com>
8394 Marek Polacek <polacek@redhat.com>
8397 * common.opt (dump_base_name_prefixed): New Variable.
8398 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
8399 if x_dump_base_name_prefixed is already set, set it at the end.
8401 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
8403 PR rtl-optimization/60508
8404 * lra-constraints.c (get_reload_reg): Add new parameter
8406 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
8407 Pass the new parameter values.
8409 2014-03-14 Richard Biener <rguenther@suse.de>
8411 * common.opt: Revert unintented changes from r205065.
8414 2014-03-14 Richard Biener <rguenther@suse.de>
8417 * cfghooks.c (split_block): Properly adjust all loops the
8418 block was a latch of.
8420 2014-03-14 Martin Jambor <mjambor@suse.cz>
8423 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
8426 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
8429 * config/avr/avr.c (avr_set_current_function): Pass function name
8430 through default_strip_name_encoding before sanity checking instead
8431 of skipping the first char of the assembler name.
8433 2014-03-13 Richard Henderson <rth@redhat.com>
8436 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
8437 (ix86_force_to_memory, ix86_free_from_memory): Remove.
8438 * config/i386/i386-protos.h: Likewise.
8439 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
8440 in the expander instead of a splitter.
8441 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
8442 any possibility of requiring a memory.
8443 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
8444 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
8445 (fp branch splitters): Update for ix86_split_fp_branch.
8446 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
8447 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
8448 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
8449 (*fop_<MODEF>_2_i387): Remove f/r alternative.
8450 (*fop_<MODEF>_3_i387): Likewise.
8451 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
8452 (splitters for the fop_* register patterns): Remove.
8453 (fscalexf4_i387): Rename from *fscalexf4_i387.
8454 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
8456 2014-03-13 Jakub Jelinek <jakub@redhat.com>
8458 PR tree-optimization/59779
8459 * tree-dfa.c (get_ref_base_and_extent): Use double_int
8460 type for bitsize and maxsize instead of HOST_WIDE_INT.
8462 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
8464 PR rtl-optimization/57320
8465 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
8466 the CFG after thread_prologue_and_epilogue_insns.
8468 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
8470 PR rtl-optimization/57189
8471 * lra-constraints.c (process_alt_operands): Disfavor spilling
8474 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
8476 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
8478 2014-03-13 Jakub Jelinek <jakub@redhat.com>
8480 PR tree-optimization/59025
8482 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
8483 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
8485 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
8488 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
8489 calls of avr_out_plus_1.
8491 2014-03-13 Bin Cheng <bin.cheng@arm.com>
8493 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
8494 BB's single pred and update the father loop's latch info later.
8496 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
8498 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
8502 (VEC_base): Likewise.
8503 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
8504 registers, we need to swap double words in little endian mode.
8506 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
8507 to be a container mode for 128-bit integer operations added in ISA
8508 2.07. Unlike TImode and PTImode, the preferred register set is
8509 the Altivec/VMX registers for the 128-bit operations.
8511 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
8513 (rs6000_split_128bit_ok_p): Likewise.
8515 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
8516 macros for creating ISA 2.07 normal and overloaded builtin
8517 functions with 3 arguments.
8518 (BU_P8V_OVERLOAD_3): Likewise.
8519 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
8520 for use as overloaded functions.
8521 (VPERM_1TI_UNS): Likewise.
8522 (VSEL_1TI): Likewise.
8523 (VSEL_1TI_UNS): Likewise.
8524 (ST_INTERNAL_1ti): Likewise.
8525 (LD_INTERNAL_1ti): Likewise.
8526 (XXSEL_1TI): Likewise.
8527 (XXSEL_1TI_UNS): Likewise.
8528 (VPERM_1TI): Likewise.
8529 (VPERM_1TI_UNS): Likewise.
8530 (XXPERMDI_1TI): Likewise.
8531 (SET_1TI): Likewise.
8532 (LXVD2X_V1TI): Likewise.
8533 (STXVD2X_V1TI): Likewise.
8534 (VEC_INIT_V1TI): Likewise.
8535 (VEC_SET_V1TI): Likewise.
8536 (VEC_EXT_V1TI): Likewise.
8537 (EQV_V1TI): Likewise.
8538 (NAND_V1TI): Likewise.
8539 (ORC_V1TI): Likewise.
8540 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
8541 added in ISA 2.07. Add both normal 'altivec' builtins, and the
8543 (VADDUQM): Likewise.
8544 (VSUBCUQ): Likewise.
8545 (VADDEUQM): Likewise.
8546 (VADDECUQ): Likewise.
8547 (VSUBEUQM): Likewise.
8548 (VSUBECUQ): Likewise.
8550 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
8551 __int128_t and __uint128_t types.
8552 (__uint128_type): Likewise.
8553 (altivec_categorize_keyword): Add support for vector __int128_t,
8554 vector __uint128_t, vector __int128, and vector unsigned __int128
8555 as a container type for TImode operations that need to be done in
8556 VSX/Altivec registers.
8557 (rs6000_macro_to_expand): Likewise.
8558 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
8559 to support 128-bit integer instructions vaddcuq, vadduqm,
8560 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
8561 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
8563 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
8564 for V1TImode, and set up preferences to use VSX/Altivec registers.
8565 Setup VSX reload handlers.
8566 (rs6000_debug_reg_global): Likewise.
8567 (rs6000_init_hard_regno_mode_ok): Likewise.
8568 (rs6000_preferred_simd_mode): Likewise.
8569 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
8570 (easy_altivec_constant): Likewise.
8571 (output_vec_const_move): Likewise.
8572 (rs6000_expand_vector_set): Convert V1TImode set and extract to
8574 (rs6000_expand_vector_extract): Likewise.
8575 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
8577 (rs6000_const_vec): Add support for V1TImode.
8578 (rs6000_emit_le_vsx_load): Swap double words when loading or
8579 storing TImode/V1TImode.
8580 (rs6000_emit_le_vsx_store): Likewise.
8581 (rs6000_emit_le_vsx_move): Likewise.
8582 (rs6000_emit_move): Add support for V1TImode.
8583 (altivec_expand_ld_builtin): Likewise.
8584 (altivec_expand_st_builtin): Likewise.
8585 (altivec_expand_vec_init_builtin): Likewise.
8586 (altivec_expand_builtin): Likewise.
8587 (rs6000_init_builtins): Add support for V1TImode type. Add
8588 support for ISA 2.07 128-bit integer builtins. Define type names
8589 for the VSX/Altivec vector types.
8590 (altivec_init_builtins): Add support for overloaded vector
8591 functions with V1TImode type.
8592 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
8593 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
8595 (rs6000_split_128bit_ok_p): Likewise.
8596 (rs6000_handle_altivec_attribute): Create V1TImode from vector
8597 __int128_t and vector __uint128_t.
8599 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
8600 and mode attributes.
8607 (VS_scalar): Likewise.
8608 (VS_double): Likewise.
8609 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
8611 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
8612 we support the ISA 2.07 128-bit integer arithmetic instructions.
8613 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
8614 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
8615 and TImode types for use with the builtin functions.
8616 (V1TI_type_node): Likewise.
8617 (unsigned_V1TI_type_node): Likewise.
8618 (intTI_type_internal_node): Likewise.
8619 (uintTI_type_internal_node): Likewise.
8621 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
8622 128-bit builtin functions.
8623 (UNSPEC_VADDEUQM): Likewise.
8624 (UNSPEC_VADDECUQ): Likewise.
8625 (UNSPEC_VSUBCUQ): Likewise.
8626 (UNSPEC_VSUBEUQM): Likewise.
8627 (UNSPEC_VSUBECUQ): Likewise.
8628 (VM): Add V1TImode to vector mode iterators.
8630 (VI_unit): Likewise.
8631 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
8632 (altivec_vaddcuq): Likewise.
8633 (altivec_vsubuqm): Likewise.
8634 (altivec_vsubcuq): Likewise.
8635 (altivec_vaddeuqm): Likewise.
8636 (altivec_vaddecuq): Likewise.
8637 (altivec_vsubeuqm): Likewise.
8638 (altivec_vsubecuq): Likewise.
8640 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
8642 (BOOL_128): Likewise.
8643 (BOOL_REGS_OUTPUT): Likewise.
8644 (BOOL_REGS_OP1): Likewise.
8645 (BOOL_REGS_OP2): Likewise.
8646 (BOOL_REGS_UNARY): Likewise.
8647 (BOOL_REGS_AND_CR0): Likewise.
8649 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
8650 128-bit integer builtin support.
8651 (vec_vadduqm): Likewise.
8652 (vec_vaddecuq): Likewise.
8653 (vec_vaddeuqm): Likewise.
8654 (vec_vsubecuq): Likewise.
8655 (vec_vsubeuqm): Likewise.
8656 (vec_vsubcuq): Likewise.
8657 (vec_vsubuqm): Likewise.
8659 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
8660 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
8661 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
8662 128-bit integer add/subtract to ISA 2.07.
8664 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
8666 * config/arc/arc.c (arc_predicate_delay_insns):
8667 Fix third argument passed to conditionalize_nonjump.
8669 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
8671 * config/aarch64/aarch64-builtins.c
8672 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
8673 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
8674 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
8675 instead of __builtin_lfloor.
8676 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
8678 2014-03-12 Jakub Jelinek <jakub@redhat.com>
8680 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
8681 (tree_ssa_ifcombine_bb_1): New function.
8682 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
8683 is an empty forwarder block to then_bb or vice versa and then_bb
8684 and else_bb are effectively swapped.
8686 2014-03-12 Christian Bruel <christian.bruel@st.com>
8689 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
8690 REG_CFA_DEF_CFA note.
8691 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
8692 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
8694 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
8696 PR tree-optimization/60454
8697 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
8699 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8701 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
8702 Do not define target_cpu_default2 to generic.
8703 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
8704 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
8705 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
8707 2014-03-12 Jakub Jelinek <jakub@redhat.com>
8708 Marc Glisse <marc.glisse@inria.fr>
8710 PR tree-optimization/60502
8711 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
8712 instead of build_low_bits_mask.
8714 2014-03-12 Jakub Jelinek <jakub@redhat.com>
8717 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
8718 if there are multiple uses, but op doesn't live on E edge.
8719 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
8720 clobber stmts before __builtin_unreachable.
8722 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
8724 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
8725 hard_frame_pointer_rtx.
8726 * cse.c (cse_insn): Remove volatile check.
8727 * cselib.c (cselib_process_insn): Likewise.
8728 * dse.c (scan_insn): Likewise.
8730 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
8732 * config/arc/arc.c (conditionalize_nonjump): New function,
8734 (arc_ifcvt): ... this.
8735 (arc_predicate_delay_insns): Use it.
8737 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
8739 * config/arc/predicates.md (extend_operand): During/after reload,
8740 allow const_int_operand.
8741 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
8742 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
8743 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
8745 (umulsi3_highpart_i): Likewise.
8747 2014-03-11 Richard Biener <rguenther@suse.de>
8749 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
8750 Add asserts to guard possible wrong-code bugs.
8752 2014-03-11 Richard Biener <rguenther@suse.de>
8754 PR tree-optimization/60429
8755 PR tree-optimization/60485
8756 * tree-ssa-structalias.c (set_union_with_increment): Properly
8757 take into account all fields that overlap the shifted vars.
8758 (do_sd_constraint): Likewise.
8759 (do_ds_constraint): Likewise.
8760 (get_constraint_for_ptr_offset): Likewise.
8762 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
8764 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
8765 (nios2_compute_frame_layout):
8766 Add calculation of cfun->machine->fp_save_offset.
8767 (nios2_expand_prologue): Correct setting of frame pointer register
8769 (nios2_expand_epilogue): Update recovery of stack pointer from
8770 frame pointer accordingly.
8771 (nios2_initial_elimination_offset): Update calculation of offset
8772 for eliminating to HARD_FRAME_POINTER_REGNUM.
8774 2014-03-10 Jakub Jelinek <jakub@redhat.com>
8777 * ipa.c (symtab_remove_unreachable_nodes): Don't call
8778 cgraph_get_create_node on VAR_DECLs.
8780 2014-03-10 Richard Biener <rguenther@suse.de>
8783 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
8785 2014-03-08 Douglas B Rupp <rupp@gnat.com>
8787 * config/vms/vms.opt (vms_float_format): New variable.
8789 2014-03-08 Tobias Burnus <burnus@net-b.de>
8791 * doc/invoke.texi (-fcilkplus): Update implementation status.
8793 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
8794 Richard Biener <rguenther@suse.de>
8796 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
8797 consistently accross all TUs.
8798 (run_gcc): Enable -fshort-double automatically at link at link-time
8799 and disallow override.
8801 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
8804 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
8805 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
8806 if they can't be used.
8808 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8810 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
8811 for Solaris 11/x86 ld.
8812 * configure: Regenerate.
8814 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8816 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
8817 (LIB_TLS_SPEC): Save as ld_tls_libs.
8818 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
8819 (HAVE_AS_IX86_TLSLDM): New test.
8820 * configure, config.in: Regenerate.
8821 * config/i386/i386.c (legitimize_tls_address): Fall back to
8822 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
8823 cannot support TLS_MODEL_LOCAL_DYNAMIC.
8824 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
8825 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
8827 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
8829 * common.opt (fira-loop-pressure): Mark as optimization.
8831 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
8833 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
8834 an OpenMP mappable type.
8836 2014-03-06 Matthias Klose <doko@ubuntu.com>
8838 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
8839 MULTILIB_OSDIRNAMES is not defined.
8841 2014-03-06 Jakub Jelinek <jakub@redhat.com>
8842 Meador Inge <meadori@codesourcery.com>
8845 * config/arm/arm.c (arm_tls_symbol_p): Remove.
8846 (arm_legitimize_address): Call legitimize_tls_address for any
8847 arm_tls_referenced_p expression, handle constant addend. Call it
8848 before testing for !TARGET_ARM.
8849 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
8851 2014-03-06 Richard Biener <rguenther@suse.de>
8857 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
8859 * tree-streamer.c (record_common_node): Assert we don't record
8860 nodes with type double.
8861 (preload_common_node): Skip type double, complex double and double
8862 pointer since it is now frontend dependent due to fshort-double option.
8864 2014-03-06 Richard Biener <rguenther@suse.de>
8866 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
8867 or -fno-lto is specified and the linker has full plugin support.
8868 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
8869 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
8870 * lto-wrapper.c (merge_and_complain): Merge compile-time
8871 optimization levels.
8872 (run_gcc): And pass it through to the link options.
8874 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
8878 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
8880 * cselib.c (remove_useless_values): Skip to avoid quadratic
8881 behavior if the condition moved from...
8882 (cselib_process_insn): ... here holds.
8884 2014-03-05 Jakub Jelinek <jakub@redhat.com>
8887 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
8888 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
8891 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
8892 (TM_H): Add x86-tune.def.
8894 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8896 * config/aarch64/aarch64.c (generic_tunings):
8897 Use cortexa57_extra_costs.
8899 2014-03-05 Jakub Jelinek <jakub@redhat.com>
8902 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
8903 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
8904 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
8907 2014-03-04 Heiher <r@hev.cc>
8909 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
8910 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
8912 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
8914 * config/i386/predicates.md (const2356_operand): Change to ...
8915 (const2367_operand): ... this.
8916 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
8918 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
8919 (*avx512pf_scatterpf<mode>sf): Ditto.
8920 (avx512pf_scatterpf<mode>df): Ditto.
8921 (*avx512pf_scatterpf<mode>df_mask): Ditto.
8922 (*avx512pf_scatterpf<mode>df): Ditto.
8923 * config/i386/i386.c (ix86_expand_builtin): Update
8924 incorrect hint operand error message.
8926 2014-03-04 Richard Biener <rguenther@suse.de>
8928 * lto-section-in.c (lto_get_section_data): Fix const cast.
8930 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
8932 * tree-streamer.c (record_common_node): Assert we don't record
8933 nodes with type double.
8934 (preload_common_node): Skip type double, complex double and double
8935 pointer since it is now frontend dependent due to fshort-double option.
8937 2014-03-04 Richard Biener <rguenther@suse.de>
8940 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
8941 (lto_input_toplevel_asms): Likewise.
8942 * lto-section-in.c (lto_get_section_data): Instead do it here
8945 2014-03-04 Richard Biener <rguenther@suse.de>
8947 PR tree-optimization/60382
8948 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
8949 dead PHIs a reduction.
8951 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
8953 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
8955 (_mm_prefetch): Move out of GCC target("sse") pragma.
8956 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
8957 GCC target("prfchw") pragma.
8958 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
8960 * config/i386/i386.c (ix86_option_override_internal): Enable
8961 -mprfchw with -mprefetchwt1.
8963 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
8965 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
8968 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
8970 * opts.h (CL_PCH_IGNORE): Define.
8971 * targhooks.c (option_affects_pch_p):
8972 Return false for options that have CL_PCH_IGNORE set.
8973 * opt-functions.awk: Process PchIgnore.
8974 * doc/options.texi: Document PchIgnore.
8976 * config/arc/arc.opt (misize): Add PchIgnore property.
8978 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8980 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
8981 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
8982 constraint on constants to permit them being loaded into
8983 GENERAL_REGS or BASE_REGS.
8985 2014-03-03 Nick Clifton <nickc@redhat.com>
8987 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
8988 anti-cacnonical alternatives.
8989 (negandhi3_real): New pattern.
8990 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
8992 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
8994 * config/avr/avr-mcus.def: Remove atxmega16x1.
8995 * config/avr/avr-tables.opt: Regenerate.
8996 * config/avr/t-multilib: Regenerate.
8997 * doc/avr-mmcu.texi: Regenerate.
8999 2014-03-03 Tobias Grosser <tobias@grosser.es>
9000 Mircea Namolaru <mircea.namolaru@inria.fr>
9002 PR tree-optimization/58028
9003 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
9006 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
9008 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
9009 not handled by recognizers.
9011 2014-03-03 Jakub Jelinek <jakub@redhat.com>
9014 * function.c (expand_function_end): Don't emit
9015 clobber_return_register sequence if clobber_after is a BARRIER.
9016 * cfgexpand.c (construct_exit_block): Append instructions before
9017 return_label to prev_bb.
9019 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9021 * config/rs6000/constraints.md: Document reserved use of "wc".
9023 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
9026 * ipa.c (function_and_variable_visibility): When dissolving comdat
9027 group, also set all symbols to local.
9029 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
9034 2013-12-14 Jan Hubicka <jh@suse.cz>
9036 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
9038 2014-03-02 Jon Beniston <jon@beniston.com>
9044 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
9045 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
9046 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
9047 (simple_return, *simple_return): New patterns
9048 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
9049 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
9051 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
9053 * dwarf2out.c (gen_subprogram_die): Tidy.
9055 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
9058 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
9059 (*mov_t_msb_neg_negc): ... this new insn.
9061 2014-02-28 Jason Merrill <jason@redhat.com>
9064 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
9067 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
9070 * dwarf2out.c (decltype_auto_die): New static.
9071 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
9072 (gen_type_die_with_usage): Handle 'decltype(auto)'.
9073 (is_cxx_auto): Likewise.
9075 2014-02-28 Ian Bolton <ian.bolton@arm.com>
9077 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
9078 we are not using general regs only.
9080 2014-02-28 Richard Biener <rguenther@suse.de>
9083 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
9084 previous fix and only allow to remove trivial pre-headers
9085 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
9086 (remove_forwarder_block): Properly update the latch of a loop.
9088 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
9091 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
9092 (cselib_preserved_hash_table): New.
9093 (preserve_constants_and_equivs): Move preserved vals to it.
9094 (cselib_find_slot): Look it up first.
9095 (cselib_init): Initialize it.
9096 (cselib_finish): Release it.
9097 (dump_cselib_table): Dump it.
9099 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
9102 * cselib.c (remove_useless_values): Skip to avoid quadratic
9103 behavior if the condition moved from...
9104 (cselib_process_insn): ... here holds.
9106 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
9109 * var-tracking.c (vt_initialize): Apply the same condition to
9110 preserve the CFA base value.
9112 2014-02-28 Joey Ye <joey.ye@arm.com>
9115 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
9116 if reload in progress or completed.
9118 2014-02-28 Tobias Burnus <burnus@net-b.de>
9121 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
9124 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
9126 * doc/tm.texi.in (Condition Code Status): Update documention for
9127 relative locations of cc0-setter and cc0-user.
9129 2014-02-27 Jeff Law <law@redhat.com>
9131 PR rtl-optimization/52714
9132 * combine.c (try_combine): When splitting an unrecognized PARALLEL
9133 into two independent simple sets, if I3 is a jump, ensure the
9134 pattern we place into I3 is a (set (pc) ...).
9136 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
9137 Jeff Law <law@redhat.com>
9139 PR rtl-optimization/49847
9140 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
9141 are in different blocks.
9142 * doc/tm.texi (Condition Code Status): Update documention for
9143 relative locations of cc0-setter and cc0-user.
9145 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
9148 * lra.c (lra_emit_add): Check SUBREG too.
9150 2014-02-27 Andreas Schwab <schwab@suse.de>
9152 * config/m68k/m68k.c (m68k_option_override): Disable
9153 -flive-range-shrinkage for classic m68k.
9154 (m68k_override_options_after_change): Likewise.
9156 2014-02-27 Marek Polacek <polacek@redhat.com>
9159 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
9160 -Wmaybe-uninitialized.
9162 2014-02-27 Alan Modra <amodra@gmail.com>
9165 * reload1.c (emit_input_reload_insns): When reload_override_in,
9166 set old to rl->in_reg when rl->in_reg is a subreg.
9168 2014-02-26 Richard Biener <rguenther@suse.de>
9171 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
9173 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
9175 * common/config/i386/predicates.md (const1256_operand): Remove.
9176 (const2356_operand): New.
9177 (const_1_to_2_operand): Remove.
9178 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
9179 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
9180 (*avx512pf_gatherpf<mode>sf): Ditto.
9181 (avx512pf_gatherpf<mode>df): Ditto.
9182 (*avx512pf_gatherpf<mode>df_mask): Ditto.
9183 (*avx512pf_gatherpf<mode>df): Ditto.
9184 (avx512pf_scatterpf<mode>sf): Ditto.
9185 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
9186 (*avx512pf_scatterpf<mode>sf): Ditto.
9187 (avx512pf_scatterpf<mode>df): Ditto.
9188 (*avx512pf_scatterpf<mode>df_mask): Ditto.
9189 (*avx512pf_scatterpf<mode>df): Ditto.
9190 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
9192 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
9194 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
9195 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
9196 (_mm512_mask_testn_epi64_mask): Move to ...
9197 * config/i386/avx512cdintrin.h: Here.
9198 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
9199 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
9200 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
9201 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
9202 TARGET_AVX512F from TARGET_AVX512CD.
9204 2014-02-26 Richard Biener <rguenther@suse.de>
9207 * ipa.c (walk_polymorphic_call_targets): Properly guard
9208 call to inline_update_overall_summary.
9210 2014-02-26 Bin Cheng <bin.cheng@arm.com>
9213 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
9214 and latches only if requested. Fix latch if it is removed.
9215 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
9216 LOOPS_HAVE_PREHEADERS.
9218 2014-02-25 Andrew Pinski <apinski@cavium.com>
9220 * builtins.c (expand_builtin_thread_pointer): Create a new target
9221 when the target is NULL.
9223 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
9225 PR rtl-optimization/60317
9226 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
9227 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
9228 * lra-assigns.c: Include params.h.
9229 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
9230 other reload pseudos considerations.
9232 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9234 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
9235 to use canonical form for nor<mode>3.
9237 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9240 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
9243 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
9245 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
9246 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
9247 (ix86_handle_option): Handle OPT_mprefetchwt1.
9248 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
9249 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
9251 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
9252 OPTION_MASK_ISA_PREFETCHWT1.
9253 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
9254 (PTA_PREFETCHWT1): New.
9255 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
9256 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
9257 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
9258 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
9259 (*prefetch_avx512pf_<mode>_: Change into ...
9260 (*prefetch_prefetchwt1_<mode>: This.
9261 * config/i386/i386.opt (mprefetchwt1): New.
9262 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
9263 (_mm_prefetch): Handle intent to write.
9264 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
9266 2014-02-25 Richard Biener <rguenther@suse.de>
9269 * emit-rtl.c (mem_attrs_htab): Remove.
9270 (mem_attrs_htab_hash): Likewise.
9271 (mem_attrs_htab_eq): Likewise.
9272 (set_mem_attrs): Always allocate new mem-attrs when something changed.
9273 (init_emit_once): Do not allocate mem_attrs_htab.
9275 2014-02-25 Richard Biener <rguenther@suse.de>
9278 * lto-opts.c (lto_write_options): Output non-explicit conservative
9279 -fwrapv, -fno-trapv and -fno-strict-overflow.
9280 * lto-wrapper.c (merge_and_complain): Handle merging those options.
9281 (run_gcc): And pass them through.
9283 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
9285 * sel-sched.c (calculate_new_fences): New parameter ptime.
9286 Calculate it as a maximum over all fence cycles.
9287 (sel_sched_region_2): Adjust the call to calculate_new_fences.
9288 Print the final schedule timing when sched_verbose.
9290 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
9292 PR rtl-optimization/60292
9293 * sel-sched.c (fill_vec_av_set): Do not reset target availability
9294 bit fot the fence instruction.
9296 2014-02-24 Alangi Derick <alangiderick@gmail.com>
9298 * calls.h: Fix typo in comment.
9300 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
9302 * config/pa/pa.c (pa_output_move_double): Don't valididate when
9303 adjusting offsetable addresses.
9305 2014-02-24 Guozhi Wei <carrot@google.com>
9307 * sparseset.h (sparseset_pop): Fix the wrong index.
9309 2014-02-24 Walter Lee <walt@tilera.com>
9311 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
9312 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
9314 * common/config/tilegx/tilegx-common.c
9315 (TARGET_DEFAULT_TARGET_FLAGS): Define.
9316 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
9318 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
9319 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
9320 (tilegx_gimplify_va_arg_expr): Handle big endian.
9321 (tilegx_expand_unaligned_load): Ditto.
9322 (tilegx_expand_unaligned_store): Ditto.
9323 (TARGET_RETURN_IN_MSB): New.
9324 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
9325 (TARGET_ENDIAN_DEFAULT): New.
9326 (TARGET_BIG_ENDIAN): Handle big endian.
9327 (BYTES_BIG_ENDIAN): Ditto.
9328 (WORDS_BIG_ENDIAN): Ditto.
9329 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
9332 * config/tilegx/tilegx.md (extv): Handle big endian.
9334 (insn_st<n>): Ditto.
9335 (insn_st<n>_add<bitsuffix>): Ditto.
9336 (insn_stnt<n>): Ditto.
9337 (insn_stnt<n>_add<bitsuffix>):Ditto.
9338 (vec_interleave_highv8qi): Handle big endian.
9339 (vec_interleave_highv8qi_be): New.
9340 (vec_interleave_highv8qi_le): New.
9341 (insn_v1int_h): Handle big endian.
9342 (vec_interleave_lowv8qi): Handle big endian.
9343 (vec_interleave_lowv8qi_be): New.
9344 (vec_interleave_lowv8qi_le): New.
9345 (insn_v1int_l): Handle big endian.
9346 (vec_interleave_highv4hi): Handle big endian.
9347 (vec_interleave_highv4hi_be): New.
9348 (vec_interleave_highv4hi_le): New.
9349 (insn_v2int_h): Handle big endian.
9350 (vec_interleave_lowv4hi): Handle big endian.
9351 (vec_interleave_lowv4hi_be): New.
9352 (vec_interleave_lowv4hi_le): New.
9353 (insn_v2int_l): Handle big endian.
9354 (vec_interleave_highv2si): Handle big endian.
9355 (vec_interleave_highv2si_be): New.
9356 (vec_interleave_highv2si_le): New.
9357 (insn_v4int_h): Handle big endian.
9358 (vec_interleave_lowv2si): Handle big endian.
9359 (vec_interleave_lowv2si_be): New.
9360 (vec_interleave_lowv2si_le): New.
9361 (insn_v4int_l): Handle big endian.
9362 * config/tilegx/tilegx.opt (mbig-endian): New option.
9363 (mlittle-endian): New option.
9364 * doc/install.texi: Document tilegxbe-linux.
9365 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
9367 2014-02-24 Martin Jambor <mjambor@suse.cz>
9370 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
9371 there are no parameter descriptors.
9373 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
9375 PR rtl-optimization/60268
9376 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
9377 initialization to ...
9378 (sched_rgn_init): ... here.
9379 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
9381 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
9383 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
9386 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
9388 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
9391 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
9393 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
9394 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
9396 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
9398 * config/microblaze/predicates.md: Add cmp_op predicate.
9399 * config/microblaze/microblaze.md: Add branch_compare instruction
9400 which uses cmp_op predicate and emits cmp insn before branch.
9401 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
9402 to microblaze_expand_conditional_branch and consolidate logic.
9403 (microblaze_expand_conditional_branch): emit branch_compare
9404 insn instead of handling cmp op separate from branch insn.
9406 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9408 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
9411 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9413 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
9414 define_insn with define_expand and new define_insn
9415 *altivec_lve<VI_char>x_internal.
9416 (altivec_stve<VI_char>x): Replace define_insn with define_expand
9417 and new define_insn *altivec_stve<VI_char>x_internal.
9418 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
9420 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
9422 (altivec_expand_stvex_be): New function.
9424 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
9426 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
9427 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
9428 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
9429 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
9431 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
9434 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
9435 instead of emit_move_insn.
9437 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9439 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
9441 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
9442 gen_altivec_vsumsws.
9444 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9446 * config/rs6000/altivec.md (altivec_lvxl): Rename as
9447 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
9448 (altivec_lvxl_<mode>): New define_expand incorporating
9449 -maltivec=be semantics where needed.
9450 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
9451 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
9452 semantics where needed.
9453 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
9454 (altivec_stvx_<mode>): New define_expand incorporating
9455 -maltivec=be semantics where needed.
9456 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
9457 VM2 iterator instead of V4SI.
9458 (altivec_stvxl_<mode>): New define_expand incorporating
9459 -maltivec=be semantics where needed.
9460 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
9461 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
9462 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
9463 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
9464 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
9465 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
9466 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
9467 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
9468 ALTIVEC_BUILTIN_STVXL.
9469 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
9470 (altivec_expand_stvx_be): Likewise.
9471 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
9472 (altivec_expand_lvx_be): Likewise.
9473 (altivec_expand_stvx_be): Likewise.
9474 (altivec_expand_builtin): Add cases for
9475 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
9476 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
9477 (altivec_init_builtins): Add definitions for
9478 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
9479 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
9481 2014-02-21 Catherine Moore <clm@codesourcery.com>
9483 * doc/invoke.texi (mvirt, mno-virt): Document.
9484 * config/mips/mips.opt (mvirt): New option.
9485 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
9487 2014-02-21 Richard Biener <rguenther@suse.de>
9489 PR tree-optimization/60276
9490 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
9491 (STMT_VINFO_MIN_NEG_DIST): New macro.
9492 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
9493 STMT_VINFO_MIN_NEG_DIST.
9494 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
9495 made for negative dependence distances still hold.
9497 2014-02-21 Richard Biener <rguenther@suse.de>
9500 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
9501 DECL_INITIAL for globals not in the current function context.
9503 2014-02-21 Jakub Jelinek <jakub@redhat.com>
9505 PR tree-optimization/56490
9506 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
9507 * tree-ssa-uninit.c: Include params.h.
9508 (compute_control_dep_chain): Add num_calls argument, return false
9509 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
9510 num_calls to recursive call.
9511 (find_predicates): Change dep_chain into normal array,
9512 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
9513 variable and adjust compute_control_dep_chain caller.
9514 (find_def_preds): Likewise.
9516 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
9518 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
9519 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
9521 2014-02-21 Nick Clifton <nickc@redhat.com>
9523 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
9524 (pushhi1): Likewise.
9525 (popqi1): Add mode to pre_dec.
9528 2014-02-21 Jakub Jelinek <jakub@redhat.com>
9530 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
9531 mode for mask of V8SFmode permutation.
9533 2014-02-20 Richard Henderson <rth@redhat.com>
9536 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
9537 a new pseudo for OLDVAL.
9539 2014-02-20 Jakub Jelinek <jakub@redhat.com>
9542 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
9543 gen_reg_rtx if d->testing_p.
9544 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
9545 if d->testing_p and we will certainly return true.
9546 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
9549 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
9551 * emit-rtl.c (gen_reg_rtx): Assert that
9552 crtl->emit.regno_pointer_align_length is non-zero.
9554 2014-02-20 Richard Henderson <rth@redhat.com>
9557 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
9558 on failure the store back into EXPECT.
9560 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
9561 Sandra Loosemore <sandra@codesourcery.com>
9563 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
9564 * config/nios2/nios2.c (nios2_function_profiler): Add
9565 -fPIC (flag_pic == 2) support.
9566 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
9567 (nios2_large_offset_p): New function.
9568 (nios2_unspec_reloc_p): Move up position, update to use
9569 nios2_large_offset_p.
9570 (nios2_unspec_address): Remove function.
9571 (nios2_unspec_offset): New function.
9572 (nios2_large_got_address): New function.
9573 (nios2_got_address): Add large offset support.
9574 (nios2_legitimize_tls_address): Update usage of removed and new
9576 (nios2_symbol_binds_local_p): New function.
9577 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
9578 (nios2_legitimize_address): Update to use nios2_large_offset_p.
9579 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
9580 (nios2_print_operand): Merge H/L processing, add hiadj/lo
9581 processing for (const (unspec ...)).
9582 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
9584 2014-02-20 Richard Biener <rguenther@suse.de>
9586 * tree-cfg.c (replace_uses_by): Mark altered BBs before
9587 doing the substitution.
9588 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
9590 2014-02-20 Martin Jambor <mjambor@suse.cz>
9593 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
9594 info when checking whether lattices are bottom.
9596 2014-02-20 Richard Biener <rguenther@suse.de>
9599 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
9602 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
9605 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
9606 parameter specifying the scaling.
9607 (inline_call): Update.
9608 (want_inline_recursively): Guard division by zero.
9609 (recursive_inlining): Update.
9610 * ipa-inline.h (clone_inlined_nodes): Update.
9612 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
9615 * config/i386/i386.c (classify_argument): Pass structures of size
9616 64 bytes or less in register.
9618 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
9619 Kirill Yukhin <kirill.yukhin@intel.com>
9621 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
9622 (_mm_rcp28_round_ss): Ditto.
9623 (_mm_rsqrt28_round_sd): Ditto.
9624 (_mm_rsqrt28_round_ss): Ditto.
9625 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
9626 (_mm_rcp14_round_ss): Ditto.
9627 (_mm_rsqrt14_round_sd): Ditto.
9628 (_mm_rsqrt14_round_ss): Ditto.
9629 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
9630 the first input operand, get rid of match_dup.
9631 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
9633 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
9635 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
9636 operand as the first input operand, set type attribute.
9637 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
9639 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
9640 operand as the first input operand, set type attribute.
9642 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9644 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
9647 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
9650 * config/i386/i386.c (construct_container): Remove TFmode check
9651 for X86_64_INTEGER_CLASS.
9653 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
9656 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
9657 only when -Wpsabi is enabled.
9659 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
9662 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
9663 passing arrays in registers are the same as for structs, so remove the
9664 special case for them.
9666 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
9668 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
9669 destination type, extract only the valid bits if the source type is not
9670 integral and has a different mode.
9672 2014-02-19 Richard Biener <rguenther@suse.de>
9675 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
9678 2014-02-19 Richard Biener <rguenther@suse.de>
9681 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
9682 (ipa_modify_call_arguments): Emit an argument load explicitely and
9683 preserve virtual SSA form there and for the replacement call.
9684 Do not update SSA form nor free dominance info.
9686 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
9688 * ipa.c (function_and_variable_visibility): Also clear WEAK
9689 flag when disolving COMDAT_GROUP.
9691 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
9693 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
9694 * ipa-prop.c (ipa_set_jf_known_type): Return early when
9696 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
9697 do more sanity checks.
9698 (detect_type_change): Return true when giving up early.
9699 (compute_complex_assign_jump_func): Fix type parameter of
9700 ipa_set_ancestor_jf.
9701 (compute_complex_ancestor_jump_func): Likewise.
9702 (update_jump_functions_after_inlining): Fix updating of
9704 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
9706 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
9708 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
9709 inline clones when edge disappears.
9711 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
9714 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
9715 Split 64-bit moves into 2 patterns. Do not allow the use of
9716 direct move for TDmode in little endian, since the decimal value
9717 has little endian bytes within a word, but the 64-bit pieces are
9718 ordered in a big endian fashion, and normal subreg's of TDmode are
9720 (mov<mode>_64bit_dm): Likewise.
9721 (movtd_64bit_nodm): Likewise.
9723 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
9725 PR tree-optimization/60174
9726 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
9727 statement of an SSA_NAME that occurs in an abnormal PHI node.
9729 2014-02-18 Jakub Jelinek <jakub@redhat.com>
9732 * final.c (SEEN_BB): Remove.
9733 (SEEN_NOTE, SEEN_EMITTED): Renumber.
9734 (final_scan_insn): Don't force_source_line on second
9735 NOTE_INSN_BASIC_BLOCK.
9737 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
9740 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
9741 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
9742 (type_natural_mode): Warn ABI change when %zmm register is not
9743 available for AVX512F vector value passing.
9745 2014-02-18 Kai Tietz <ktietz@redhat.com>
9748 * config/i386/i386.c (ix86_expand_prologue): Use value in
9749 rax register as displacement when restoring %r10 or %rax.
9750 Fix wrong offset when restoring both registers.
9752 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
9754 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
9755 assertion with conditional return.
9757 2014-02-18 Jakub Jelinek <jakub@redhat.com>
9758 Uros Bizjak <ubizjak@gmail.com>
9761 * config/i386/driver-i386.c (host_detect_local_cpu): If
9762 YMM state is not saved by the OS, also clear has_f16c. Move
9763 CPUID 0x80000001 handling before YMM state saving checking.
9765 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
9767 PR rtl-optimization/58960
9768 * haifa-sched.c (alloc_global_sched_pressure_data): New,
9769 factored out from ...
9770 (sched_init): ... here.
9771 (free_global_sched_pressure_data): New, factored out from ...
9772 (sched_finish): ... here.
9773 * sched-int.h (free_global_sched_pressure_data): Declare.
9774 * sched-rgn.c (nr_regions_initial): New static global.
9775 (haifa_find_rgns): Initialize it.
9776 (schedule_region): Disable sched-pressure for the newly
9779 2014-02-17 Richard Biener <rguenther@suse.de>
9781 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
9782 release SSA defs of pattern stmts.
9784 2014-02-17 Richard Biener <rguenther@suse.de>
9786 * tree-inline.c (expand_call_inline): Release the virtual
9787 operand defined by the call we are about to inline.
9789 2014-02-17 Richard Biener <rguenther@suse.de>
9791 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
9793 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
9794 Ilya Tocar <ilya.tocar@intel.com>
9796 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
9797 arguments order in builtin.
9798 (_mm512_permutexvar_epi64): Ditto.
9799 (_mm512_mask_permutexvar_epi64): Ditto
9800 (_mm512_maskz_permutexvar_epi32): Ditto
9801 (_mm512_permutexvar_epi32): Ditto
9802 (_mm512_mask_permutexvar_epi32): Ditto
9804 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9806 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
9807 (p8_vmrgow): Likewise.
9809 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9811 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
9814 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
9817 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
9818 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
9819 into 64-bit and 32-bit moves. On 64-bit moves, add support for
9820 using direct move instructions on ISA 2.07. Also adjust
9821 instruction length for 64-bit.
9822 (mov<mode>_64bit, TFmode/TDmode): Likewise.
9823 (mov<mode>_32bit, TFmode/TDmode): Likewise.
9825 2014-02-15 Alan Modra <amodra@gmail.com>
9829 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
9830 find_replacement on parts of insn rtl that might be reloaded.
9832 2014-02-15 Richard Biener <rguenther@suse.de>
9834 PR tree-optimization/60183
9835 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
9836 (tree_ssa_phiprop): Calculate and free post-dominators.
9838 2014-02-14 Jeff Law <law@redhat.com>
9840 PR rtl-optimization/60131
9841 * ree.c (get_extended_src_reg): New function.
9842 (combine_reaching_defs): Use it rather than assuming location of REG.
9843 (find_and_remove_re): Verify first operand of extension is
9844 a REG before adding the insns to the copy list.
9846 2014-02-14 Roland McGrath <mcgrathr@google.com>
9848 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
9849 * configure: Regenerated.
9850 * config.in: Regenerated.
9851 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
9852 instead of ASM_SHORT.
9854 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
9855 Richard Earnshaw <rearnsha@arm.com>
9857 PR rtl-optimization/59535
9858 * lra-constraints.c (process_alt_operands): Encourage alternative
9859 when unassigned pseudo class is superset of the alternative class.
9860 (inherit_reload_reg): Don't inherit when optimizing for code size.
9861 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
9862 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
9863 modes not less than 4 for Thumb1.
9865 2014-02-14 Kyle McMartin <kyle@redhat.com>
9868 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
9870 2014-02-14 Richard Biener <rguenther@suse.de>
9872 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
9873 (get_frame_arg): Drop the assert with langhook types_compatible_p.
9874 Do not strip INDIRECT_REFs.
9876 2014-02-14 Richard Biener <rguenther@suse.de>
9879 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
9880 DECL_FUNCTION_SPECIFIC_TARGET.
9881 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
9882 * tree-streamer-out.c (pack_ts_target_option): Remove.
9883 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
9884 (write_ts_function_decl_tree_pointers): Do not stream
9885 DECL_FUNCTION_SPECIFIC_TARGET.
9886 * tree-streamer-in.c (unpack_ts_target_option): Remove.
9887 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
9888 (lto_input_ts_function_decl_tree_pointers): Do not stream
9889 DECL_FUNCTION_SPECIFIC_TARGET.
9891 2014-02-14 Jakub Jelinek <jakub@redhat.com>
9893 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
9894 (get_initial_def_for_induction, vectorizable_induction): Ignore
9895 debug stmts when looking for exit_phi.
9896 (vectorizable_live_operation): Fix up condition.
9898 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
9900 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
9901 nreverse() because it changes the content of original tree list.
9903 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
9905 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
9906 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
9908 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
9910 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
9911 GNU coding standards.
9913 2014-02-13 Jakub Jelinek <jakub@redhat.com>
9916 * dwarf2out.c (gen_subprogram_die): Don't call
9917 add_calling_convention_attribute if subr_die is old_die.
9919 2014-02-13 Sharad Singhai <singhai@google.com>
9921 * doc/optinfo.texi: Fix order of nodes.
9923 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
9925 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
9926 operands[2], not operands[3].
9928 2014-02-13 Richard Biener <rguenther@suse.de>
9931 * doc/install.texi (ISL): Update recommended version to 0.12.2,
9932 mention the possibility of an in-tree build.
9933 (CLooG): Update recommended version to 0.18.1, mention the
9934 possibility of an in-tree build and clarify that the ISL
9935 bundled with CLooG does not work.
9937 2014-02-13 Jakub Jelinek <jakub@redhat.com>
9940 * expr.c (compress_float_constant): If x is a hard register,
9941 extend into a pseudo and then move to x.
9943 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
9945 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
9946 caused by bad second argument to warning_at() with -mhotpatch and
9947 nested functions (e.g. with gfortran).
9949 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
9951 * opts.c (option_name): Remove "enabled by default" rider.
9953 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
9955 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
9957 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
9958 Uros Bizjak <ubizjak@gmail.com>
9961 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
9962 * configure: Regenerated.
9964 2014-02-12 Richard Biener <rguenther@suse.de>
9966 * vec.c (vec_prefix::calculate_allocation): Move as
9967 inline variant to vec.h.
9968 (vec_prefix::calculate_allocation_1): New out-of-line version.
9969 * vec.h (vec_prefix::calculate_allocation_1): Declare.
9970 (vec_prefix::m_has_auto_buf): Rename to ...
9971 (vec_prefix::m_using_auto_storage): ... this.
9972 (vec_prefix::calculate_allocation): Inline the easy cases
9973 and dispatch to calculate_allocation_1 which doesn't need the
9975 (va_heap::reserve): Use gcc_checking_assert.
9976 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
9977 m_using_auto_storage.
9978 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
9980 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
9981 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
9982 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
9984 2014-02-12 Richard Biener <rguenther@suse.de>
9986 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
9987 when we found a dependence.
9989 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
9991 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
9993 (maybe_fold_stmt): ... into this new function.
9994 * omp-low.c (lower_omp): Update comment.
9996 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
9999 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
10002 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
10004 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
10005 identifiers in comments.
10006 (cortexa53_extra_costs): Likewise.
10007 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
10008 (cortexa7_extra_costs): Likewise.
10009 (cortexa12_extra_costs): Likewise.
10010 (cortexa15_extra_costs): Likewise.
10011 (v7m_extra_costs): Likewise.
10013 2014-02-12 Richard Biener <rguenther@suse.de>
10015 PR middle-end/60092
10016 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
10017 of posix_memalign being successful.
10018 (lower_stmt): Restrict lowering of posix_memalign to when
10019 -ftree-bit-ccp is enabled.
10021 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
10023 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
10025 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
10027 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
10029 PR rtl-optimization/60116
10030 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
10031 other_insn once the combination has been validated.
10033 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
10036 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
10038 * ipa-devirt.c: Include demangle.h
10039 (odr_violation_reported): New static variable.
10040 (add_type_duplicate): Update odr_violations.
10041 (maybe_record_node): Add completep parameter; update it.
10042 (record_target_from_binfo): Add COMPLETEP parameter;
10043 update it as needed.
10044 (possible_polymorphic_call_targets_1): Likewise.
10045 (struct polymorphic_call_target_d): Add nonconstruction_targets;
10046 rename FINAL to COMPLETE.
10047 (record_targets_from_bases): Sanity check we found the binfo;
10048 fix COMPLETEP updating.
10049 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
10050 parameter, fix computing of COMPLETEP.
10051 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
10052 at LTO time do demangling.
10053 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
10054 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
10056 (gimple_get_virt_method_for_binfo): Likewise.
10057 * gimple-fold.h (gimple_get_virt_method_for_binfo,
10058 gimple_get_virt_method_for_vtable): Update prototypes.
10060 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
10063 * genautomata.c (add_presence_absence): Fix typo with
10064 {final_}presence_list.
10066 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
10069 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
10070 for VSX/Altivec vectors that land in GPR registers.
10072 2014-02-11 Richard Henderson <rth@redhat.com>
10073 Jakub Jelinek <jakub@redhat.com>
10076 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
10077 around drhs if type conversion to lacc->type is not useless.
10079 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10081 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
10083 (cortex-a57.cortex-a53): Likewise.
10084 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
10086 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10088 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
10091 2014-02-11 Renlin Li <Renlin.Li@arm.com>
10093 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
10094 add_options_for_arm_vfp3.
10096 2014-02-11 Jeff Law <law@redhat.com>
10098 PR middle-end/54041
10099 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
10100 object with an undesirable mode.
10102 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
10105 * config/i386/sol2-9.h: New file.
10106 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
10107 *-*-solaris2.9*): Use it.
10109 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
10111 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
10112 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
10114 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
10116 * config/microblaze/microblaze.c: Extend mcpu version format
10118 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
10120 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
10122 2014-02-10 Richard Henderson <rth@redhat.com>
10125 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
10126 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
10127 ms-abi vs -mno-accumulate-outgoing-args.
10128 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
10129 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
10132 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
10134 PR middle-end/60080
10135 * cfgexpand.c (expand_asm_operands): Attach source location to
10136 ASM_INPUT rtx objects.
10137 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
10139 2014-02-10 Nick Clifton <nickc@redhat.com>
10141 * config/mn10300/mn10300.c (popcount): New function.
10142 (mn10300_expand_prologue): Include saved registers in stack usage
10145 2014-02-10 Jeff Law <law@redhat.com>
10147 PR middle-end/52306
10148 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
10149 when changing the SET_DEST of a prior insn to avoid an input reload.
10151 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
10153 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
10154 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
10155 -mcall-openbsd, or -mcall-linux.
10156 (CC1_ENDIAN_BIG_SPEC): Remove.
10157 (CC1_ENDIAN_LITTLE_SPEC): Remove.
10158 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
10159 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
10160 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
10161 and %cc1_endian_default.
10162 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
10164 2014-02-10 Richard Biener <rguenther@suse.de>
10166 PR tree-optimization/60115
10167 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
10168 MEM_REF handling. Properly verify that the accesses are not
10169 out of the objects bound.
10171 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10173 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
10176 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
10178 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
10179 proper constants and fix formatting.
10180 (possible_polymorphic_call_targets): Fix formatting.
10182 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
10183 Ilya Tocar <ilya.tocar@intel.com>
10185 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
10186 (_mm512_loadu_epi32): Renamed into...
10187 (_mm512_loadu_si512): This.
10188 (_mm512_storeu_epi32): Renamed into...
10189 (_mm512_storeu_si512): This.
10190 (_mm512_maskz_ceil_ps): Removed.
10191 (_mm512_maskz_ceil_pd): Ditto.
10192 (_mm512_maskz_floor_ps): Ditto.
10193 (_mm512_maskz_floor_pd): Ditto.
10194 (_mm512_floor_round_ps): Ditto.
10195 (_mm512_floor_round_pd): Ditto.
10196 (_mm512_ceil_round_ps): Ditto.
10197 (_mm512_ceil_round_pd): Ditto.
10198 (_mm512_mask_floor_round_ps): Ditto.
10199 (_mm512_mask_floor_round_pd): Ditto.
10200 (_mm512_mask_ceil_round_ps): Ditto.
10201 (_mm512_mask_ceil_round_pd): Ditto.
10202 (_mm512_maskz_floor_round_ps): Ditto.
10203 (_mm512_maskz_floor_round_pd): Ditto.
10204 (_mm512_maskz_ceil_round_ps): Ditto.
10205 (_mm512_maskz_ceil_round_pd): Ditto.
10206 (_mm512_expand_pd): Ditto.
10207 (_mm512_expand_ps): Ditto.
10208 * config/i386/i386.c (ix86_builtins): Remove
10209 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
10210 (bdesc_args): Ditto.
10211 * config/i386/predicates.md (const1256_operand): New.
10212 (const_1_to_2_operand): Ditto.
10213 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
10214 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
10215 (*avx512pf_gatherpf<mode>sf): Ditto.
10216 (avx512pf_gatherpf<mode>df): Ditto.
10217 (*avx512pf_gatherpf<mode>df_mask): Ditto.
10218 (*avx512pf_gatherpf<mode>df): Ditto.
10219 (avx512pf_scatterpf<mode>sf): Ditto.
10220 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
10221 (*avx512pf_scatterpf<mode>sf): Ditto.
10222 (avx512pf_scatterpf<mode>df): Ditto.
10223 (*avx512pf_scatterpf<mode>df_mask): Ditto.
10224 (*avx512pf_scatterpf<mode>df): Ditto.
10225 (avx512f_expand<mode>): Removed.
10226 (<shift_insn><mode>3<mask_name>): Change predicate type.
10228 2014-02-08 Jakub Jelinek <jakub@redhat.com>
10230 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
10231 not at the end of datarefs vector use ordered_remove to avoid
10232 reordering datarefs vector.
10235 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
10236 mark local addressable non-static vars as GOVD_PRIVATE
10237 instead of GOVD_LOCAL.
10238 * omp-low.c (lower_omp_for): Move gimple_bind_vars
10239 and BLOCK_VARS of gimple_bind_block to new_stmt rather
10242 PR middle-end/60092
10243 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
10244 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
10245 assume_aligned or alloc_align attributes.
10246 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
10247 arguments. Handle also assume_aligned and alloc_align attributes.
10248 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
10249 calls to functions with assume_aligned or alloc_align attributes.
10250 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
10252 2014-02-08 Terry Guo <terry.guo@arm.com>
10254 * doc/invoke.texi: Document ARM -march=armv7e-m.
10256 2014-02-08 Jakub Jelinek <jakub@redhat.com>
10258 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
10259 flag on __cilkrts_rethrow builtin.
10262 * ipa-cp.c (determine_versionability): Fail at -O0
10263 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
10264 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
10267 2014-02-04 Jakub Jelinek <jakub@redhat.com>
10270 * tree-inline.c (copy_forbidden): Fail for
10271 __attribute__((optimize (0))) functions.
10273 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
10275 * varpool.c: Include pointer-set.h.
10276 (varpool_remove_unreferenced_decls): Variables in other partitions
10277 will not be output; be however careful to not lose information
10278 about partitioning.
10280 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
10282 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
10283 lookup in the vtable constructor.
10285 2014-02-07 Jeff Law <law@redhat.com>
10288 * config/m68k/m68k.md (ashldi_extsi): Turn into a
10289 define_insn_and_split.
10291 * ipa-inline.c (inline_small_functions): Fix typos.
10293 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10295 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
10296 (s390_can_use_return_insn): Declare.
10297 * config/s390/s390.h (EPILOGUE_USES): Define.
10298 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
10300 (s390_chunkify_start): Handle return JUMP_LABELs.
10301 (s390_early_mach): Emit a main_pool instruction on the entry edge.
10302 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
10303 (s390_can_use_return_insn): New functions.
10304 (s390_fix_long_loop_prediction): Handle conditional returns.
10305 (TARGET_SET_UP_BY_PROLOGUE): Define.
10306 * config/s390/s390.md (ANY_RETURN): New code iterator.
10307 (*creturn, *csimple_return, return, simple_return): New patterns.
10309 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10311 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
10312 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
10313 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
10314 REG_CFA_RESTORE list when deciding not to restore a register.
10316 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10318 * config/s390/s390.c: Include tree-pass.h and context.h.
10319 (s390_early_mach): New function, split out from...
10320 (s390_emit_prologue): ...here.
10321 (pass_data_s390_early_mach): New pass structure.
10322 (pass_s390_early_mach): New class.
10323 (s390_option_override): Create and register early_mach pass.
10324 Move to end of file.
10326 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10328 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
10329 to match for the exit block.
10331 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10333 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
10334 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
10335 Reject misaligned operands.
10337 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10339 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
10341 2014-02-07 Richard Biener <rguenther@suse.de>
10343 PR middle-end/60092
10344 * gimple-low.c (lower_builtin_posix_memalign): New function.
10345 (lower_stmt): Call it to lower posix_memalign in a way
10346 to make alignment info accessible.
10348 2014-02-07 Jakub Jelinek <jakub@redhat.com>
10351 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
10352 __builtin_setjmp_receiver.
10354 2014-02-07 Richard Biener <rguenther@suse.de>
10356 PR middle-end/60092
10357 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
10358 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
10359 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
10360 Handle BUILT_IN_POSIX_MEMALIGN.
10361 (find_func_clobbers): Likewise.
10362 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
10363 (call_may_clobber_ref_p_1): Likewise.
10365 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
10368 * ipa-devirt.c (record_target_from_binfo): Remove overactive
10371 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
10374 * lto-cgraph.c (lto_output_node): Use
10375 symtab_get_symbol_partitioning_class.
10376 (lto_output_varpool_node): likewise.
10377 (symtab_get_symbol_partitioning_class): Move here from
10378 lto/lto-partition.c
10379 * cgraph.h (symbol_partitioning_class): Likewise.
10380 (symtab_get_symbol_partitioning_class): Declare.
10382 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
10384 * ggc.h (ggc_internal_cleared_alloc): New macro.
10385 * vec.h (vec_safe_copy): Handle memory stats.
10386 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
10387 * target-globals.c (save_target_globals): Likewise.
10389 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
10392 * expr.c (emit_move_resolve_push): Export; be bit more selective
10393 on when to clear alias set.
10394 * expr.h (emit_move_resolve_push): Declare.
10395 * function.h (struct function): Add tail_call_marked.
10396 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
10397 * config/i386/i386-protos.h (ix86_expand_push): Remove.
10398 * config/i386/i386.md (TImode move expander): De not call
10400 (FP push expanders): Preserve memory attributes.
10401 * config/i386/sse.md (push<mode>1): Remove.
10402 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
10403 (ix86_expand_push): Remove.
10404 * config/i386/mmx.md (push<mode>1): Remove.
10406 2014-02-06 Jakub Jelinek <jakub@redhat.com>
10408 PR rtl-optimization/60030
10409 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
10410 lopart with paradoxical subreg before shifting it up by hprec.
10412 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10414 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
10415 Remove extra newline at end of file.
10416 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
10417 (arm_issue_rate): Handle cortexa57.
10418 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
10419 (cortex-a57.cortex-a53): Likewise.
10421 2014-02-06 Jakub Jelinek <jakub@redhat.com>
10424 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
10425 don't record in REG_FRAME_RELATED_EXPR registers not set in that
10427 (arm_expand_prologue): Adjust all callers.
10428 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
10429 info, registers also at the lowest numbered registers side. Use
10430 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
10434 * var-tracking.c (adjust_mems): Before adding a SET to
10435 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
10437 2014-02-06 Alan Modra <amodra@gmail.com>
10440 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
10441 change SDmode to DDmode when lra_in_progress.
10443 2014-02-06 Jakub Jelinek <jakub@redhat.com>
10445 PR middle-end/59150
10446 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
10447 free_data_ref on the dr first, and before goto again also set dr
10448 to the next dr. For simd_lane_access, free old datarefs[i] before
10449 overwriting it. For get_vectype_for_scalar_type failure, don't
10450 free_data_ref if simd_lane_access.
10452 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
10455 * tree.h (opts_for_fn): New inline function.
10456 (opt_for_fn): Define.
10457 * config/i386/i386.c (ix86_function_regparm): Use
10458 opt_for_fn (decl, optimize) instead of optimize.
10460 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
10462 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
10463 for SYMBOL_REF in large memory model.
10465 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10467 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
10468 and crypto support.
10469 (cortex-a57): Likewise.
10470 (cortex-a57.cortex-a53): Likewise.
10472 2014-02-06 Yury Gribov <y.gribov@samsung.com>
10473 Kugan Vivekanandarajah <kuganv@linaro.org>
10475 * config/arm/arm.c (arm_vector_alignment_reachable): Check
10477 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
10479 2014-02-06 Richard Biener <rguenther@suse.de>
10481 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
10482 set_loop_copy and initialize_original_copy_tables.
10484 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
10486 * config/aarch64/aarch64-simd.md
10487 (aarch64_ashr_simddi): Change QI to SI.
10489 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
10490 Jakub Jelinek <jakub@redhat.com>
10492 PR middle-end/60013
10493 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
10496 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10498 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
10499 CODE_FOR_altivec_vpku[hw]um to
10500 CODE_FOR_altivec_vpku[hw]um_direct.
10501 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
10502 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
10503 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
10504 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
10506 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10508 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
10509 generation for -maltivec=be.
10510 (altivec_vsumsws): Simplify redundant test.
10512 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10514 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
10515 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
10516 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
10517 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
10518 gen_altivec_vpkuwum.
10519 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
10521 (altivec_vpks<VI_char>ss): Likewise.
10522 (altivec_vpks<VI_char>us): Likewise.
10523 (altivec_vpku<VI_char>us): Likewise.
10524 (altivec_vpku<VI_char>um): Likewise.
10525 (altivec_vpku<VI_char>um_direct): New (copy of
10526 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
10528 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
10529 target is little endian and -maltivec=be is not specified.
10530 (*altivec_vupkhs<VU_char>_direct): New (copy of
10531 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
10532 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
10533 target is little endian and -maltivec=be is not specified.
10534 (*altivec_vupkls<VU_char>_direct): New (copy of
10535 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
10536 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
10537 little endian and -maltivec=be is not specified.
10538 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
10539 little endian and -maltivec=be is not specified.
10541 2014-02-05 Richard Henderson <rth@redhat.com>
10544 * combine-stack-adj.c: Revert r206943.
10545 * sched-int.h (struct deps_desc): Add last_args_size.
10546 * sched-deps.c (init_deps): Initialize it.
10547 (sched_analyze_insn): Add OUTPUT dependencies between insns that
10548 contain REG_ARGS_SIZE notes.
10550 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
10552 * lto-cgraph.c (asm_nodes_output): Make global.
10553 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
10554 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
10555 (driver_handle_option): Handle OPT_fwpa.
10557 2014-02-05 Jakub Jelinek <jakub@redhat.com>
10560 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
10561 a comment typo and formatting issue. If odr_hash hasn't been
10562 created, return vNULL and set *completep to false.
10564 PR middle-end/57499
10565 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
10566 bb with no successors.
10568 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
10571 * doc/invoke.texi (-march): Clarify documentation for ARM.
10572 (-mtune): Likewise.
10575 2014-02-05 Richard Biener <rguenther@suse.de>
10577 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
10578 when not vectorizing because of too many alias checks.
10579 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
10580 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
10582 2014-02-05 Nick Clifton <nickc@redhat.com>
10584 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
10585 accept extended registers in any mode when compiling for the MN10300.
10587 2014-02-05 Yury Gribov <y.gribov@samsung.com>
10589 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
10590 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
10591 sanitization attributes.
10592 (can_inline_edge_p): Likewise.
10593 (sanitize_attrs_match_for_inline_p): New function.
10595 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
10597 * ipa-prop.c (detect_type_change): Shor circuit testing of
10598 type changes on THIS pointer.
10600 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
10603 * config/pa/pa.c (legitimize_tls_address): Return original address
10604 if not passed a SYMBOL_REF rtx.
10605 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
10607 (pa_emit_move_sequence): Simplify TLS source operands.
10608 (pa_legitimate_constant_p): Reject all TLS constants.
10609 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
10610 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
10612 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
10614 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
10615 groups when we know they are controlled by LTO.
10616 * varasm.c (default_binds_local_p_1): If object is in other partition,
10617 it will be resolved locally.
10619 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
10621 * config/host-linux.c (linux_gt_pch_use_address): Don't
10622 use SSIZE_MAX because it is not always defined.
10624 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
10627 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
10628 threshold for pseudo splitting.
10629 (update_ebb_live_info): Process call argument hard registers and
10630 hard registers from insn definition too.
10631 (max_small_class_regs_num): New constant.
10632 (inherit_in_ebb): Update live hard regs through EBBs. Update
10633 reloads_num only for small register classes. Don't split for
10636 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
10639 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
10642 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
10644 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
10645 visibility is safe.
10647 2014-02-04 Marek Polacek <polacek@redhat.com>
10649 * gdbinit.in (pel): Define.
10651 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
10653 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
10656 2014-02-04 Richard Biener <rguenther@suse.de>
10659 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
10660 in function context local.
10661 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
10662 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
10663 similar to LTO_imported_decl_ref.
10665 2014-02-04 Jakub Jelinek <jakub@redhat.com>
10667 PR tree-optimization/60002
10668 * cgraphclones.c (build_function_decl_skip_args): Clear
10669 DECL_LANG_SPECIFIC.
10671 PR tree-optimization/60023
10672 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
10673 false to gsi_replace.
10674 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
10675 has been in some EH region and vec_stmt could throw, add
10676 vec_stmt into the same EH region.
10677 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
10678 has no lhs, ignore it.
10679 * internal-fn.c (expand_MASK_LOAD): Likewise.
10682 * tree-inline.c (copy_forbidden): Fail for
10683 __attribute__((optimize (0))) functions.
10686 * omp-low.c (simd_clone_struct_copy): If from->inbranch
10687 is set, copy one less argument.
10688 (expand_simd_clones): Don't subtract clone_info->inbranch
10689 from simd_clone_struct_alloc argument.
10691 PR rtl-optimization/57915
10692 * recog.c (simplify_while_replacing): If all unary/binary/relational
10693 operation arguments are constant, attempt to simplify those.
10695 PR middle-end/59261
10696 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
10697 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
10699 2014-02-04 Richard Biener <rguenther@suse.de>
10701 PR tree-optimization/60012
10702 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
10703 TBAA disambiguation to all DDRs.
10705 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
10708 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
10709 (LINK_SPEC): Use it for -shared, -shared-libgcc.
10711 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10714 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
10716 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10718 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
10719 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
10721 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10724 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
10725 to figure out targets of polymorphic calls with known decl.
10726 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
10727 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
10728 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
10729 (get_polymorphic_call_info): ... here.
10730 (get_polymorphic_call_info_from_invariant): New function.
10732 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10734 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
10735 lookup via vtable pointer; check for type consistency
10736 and turn inconsitent facts into UNREACHABLE.
10737 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
10738 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
10739 type inconsistent querries; return UNREACHABLE instead.
10741 2014-02-03 Richard Henderson <rth@twiddle.net>
10744 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
10745 already processed this node.
10746 (normalize_one_pred_1): Pass along mark_set.
10747 (normalize_one_pred): Create and destroy a pointer_set_t.
10748 (normalize_one_pred_chain): Likewise.
10750 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
10752 PR gcov-profile/58602
10753 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
10755 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10758 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
10759 -fno-devirtualize; try to devirtualize by the knowledge of
10760 virtual table pointer given by aggregate propagation.
10761 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
10762 (ipa_print_node_jump_functions): Dump also offset that
10763 is relevant for polymorphic calls.
10764 (determine_known_aggregate_parts): Add arg_type parameter; use it
10765 instead of determining the type from pointer type.
10766 (ipa_compute_jump_functions_for_edge): Update call of
10767 determine_known_aggregate_parts.
10768 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
10769 (gimple_get_virt_method_for_binfo): ... here; simplify using
10770 vtable_pointer_value_to_vtable.
10771 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
10772 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
10773 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
10774 (vtable_pointer_value_to_vtable): Break out from ...; handle also
10776 (vtable_pointer_value_to_binfo): ... here.
10777 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
10779 2014-02-03 Teresa Johnson <tejohnson@google.com>
10781 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
10782 redef of outer loop index variable.
10784 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
10788 * doc/extend.texi (Function Attributes): Typo.
10790 2014-02-03 Cong Hou <congh@google.com>
10792 PR tree-optimization/60000
10793 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
10794 if the vectorized statement is a store. A store statement can only
10795 appear at the end of pattern statements.
10797 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
10799 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
10800 (ix86_option_override_internal): Default long double to 64-bit for
10801 32-bit Bionic and to 128-bit for 64-bit Bionic.
10803 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
10804 TARGET_LONG_DOUBLE_128 is true.
10805 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
10807 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
10808 (mlong-double-64): Negate -mlong-double-128.
10809 (mlong-double-128): New option.
10811 * config/i386/i386-c.c (ix86_target_macros): Define
10812 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
10814 * doc/invoke.texi: Document -mlong-double-128.
10816 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
10818 PR rtl-optimization/60024
10819 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
10821 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
10823 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
10825 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
10827 PR rtl-optimization/57662
10828 * sel-sched.c (code_motion_path_driver): Do not mark already not
10829 existing blocks in the visiting bitmap.
10831 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
10833 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
10834 on the insn being emitted.
10836 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
10837 Will Deacon <will.deacon@arm.com>
10839 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
10841 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10843 * config/arm/arm-tables.opt: Regenerate.
10845 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10847 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
10848 for vector types other than V16QImode.
10849 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
10850 define_expand, and call altivec_expand_vec_perm_le when producing
10851 code with little endian element order.
10852 (*altivec_vperm_<mode>_internal): New insn having previous
10853 behavior of altivec_vperm_<mode>.
10854 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
10855 altivec_expand_vec_perm_le when producing code with little endian
10857 (*altivec_vperm_<mode>_uns_internal): New insn having previous
10858 behavior of altivec_vperm_<mode>_uns.
10860 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10862 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
10863 (altivec_vsumsws): Add handling for -maltivec=be with a little
10865 (altivec_vsumsws_direct): New.
10866 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
10867 gen_altivec_vsumsws.
10869 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
10871 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
10872 vtable_pointer_value_to_binfo): New functions.
10873 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
10874 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
10876 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
10878 * config/nios2/nios2.md (load_got_register): Initialize GOT
10879 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
10880 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
10882 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
10884 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
10885 preserverd by passthrough, do not propagate the type.
10887 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
10889 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
10890 (mips_atomic_assign_expand_fenv): New function.
10891 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
10893 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
10895 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
10896 (__builtin_mips_set_fcsr): Likewise.
10897 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
10898 MIPS_USI_FTYPE_VOID.
10899 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
10900 (mips16_expand_set_fcsr): Likewise.
10901 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
10902 (mips16_set_fcsr_stub): Likewise.
10903 (mips16_get_fcsr_one_only_stub): New class.
10904 (mips16_set_fcsr_one_only_stub): Likewise.
10905 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
10906 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
10907 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
10908 (hard_float): New availability predicate.
10909 (mips_builtins): Add get_fcsr and set_fcsr.
10910 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
10911 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
10912 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
10913 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
10914 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
10917 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
10919 * config/mips/mips.c (mips_one_only_stub): New class.
10920 (mips_need_mips16_rdhwr_p): Replace with...
10921 (mips16_rdhwr_stub): ...this new variable.
10922 (mips16_stub_call_address): New function.
10923 (mips16_rdhwr_one_only_stub): New class.
10924 (mips_expand_thread_pointer): Use mips16_stub_call_address.
10925 (mips_output_mips16_rdhwr): Delete.
10926 (mips_finish_stub): New function.
10927 (mips_code_end): Use it to handle rdhwr stubs.
10929 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
10932 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
10933 when calculating size of integer atomic types.
10935 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
10937 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
10939 2014-02-01 Jakub Jelinek <jakub@redhat.com>
10941 PR tree-optimization/60003
10942 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
10943 * profile.c (branch_prob): Use gimple_call_builtin_p
10944 to check for BUILT_IN_SETJMP_RECEIVER.
10945 * tree-inline.c (copy_bb): Call notice_special_calls.
10947 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
10950 * lra-constraints.c (process_alt_operands): Update reload_sum only
10953 2014-01-31 Richard Henderson <rth@redhat.com>
10955 PR middle-end/60004
10956 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
10957 until after else_eh is processed.
10959 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
10961 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
10962 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
10963 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
10964 in smmintrin.h, remove them.
10965 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
10966 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
10967 * config/i386/i386.md (ROUND_SAE): Fix value.
10968 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
10969 (const48_operand): New.
10970 * config/i386/subst.md (round), (round_expand): Use
10971 const_4_or_8_to_11_operand.
10972 (round_saeonly), (round_saeonly_expand): Use const48_operand.
10974 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
10976 * config/i386/constraints.md (Yk): Swap meaning with k.
10977 * config/i386/i386.md (movhi_internal): Change Yk to k.
10978 (movqi_internal): Ditto.
10979 (*k<logic><mode>): Ditto.
10982 (kandn<mode>): Ditto.
10983 (*<code>hi_1): Ditto.
10984 (*<code>qi_1): Ditto.
10985 (kxnor<mode>): Ditto.
10986 (kortestzhi): Ditto.
10987 (kortestchi): Ditto.
10989 (*one_cmplhi2_1): Ditto.
10990 (*one_cmplqi2_1): Ditto.
10991 * config/i386/sse.md (): Change k to Yk.
10992 (avx512f_load<mode>_mask): Ditto.
10993 (avx512f_blendm<mode>): Ditto.
10994 (avx512f_store<mode>_mask): Ditto.
10995 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
10996 (avx512f_storedqu<mode>_mask): Ditto.
10997 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
10999 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
11000 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
11001 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
11002 (avx512f_maskcmp<mode>3): Ditto.
11003 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
11004 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
11005 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
11006 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
11007 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
11008 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
11009 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
11010 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
11011 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
11012 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
11013 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
11014 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
11015 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
11016 (vec_extract_lo_<mode>_maskm): Ditto.
11017 (vec_extract_hi_<mode>_maskm): Ditto.
11018 (avx512f_vternlog<mode>_mask): Ditto.
11019 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
11020 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
11021 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
11022 (avx512f_<code>v8div16qi2_mask): Ditto.
11023 (avx512f_<code>v8div16qi2_mask_store): Ditto.
11024 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
11025 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
11026 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
11027 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
11028 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
11029 (*avx512pf_gatherpf<mode>df_mask): Ditto.
11030 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
11031 (*avx512pf_scatterpf<mode>df_mask): Ditto.
11032 (avx512cd_maskb_vec_dupv8di): Ditto.
11033 (avx512cd_maskw_vec_dupv16si): Ditto.
11034 (avx512f_vpermi2var<mode>3_maskz): Ditto.
11035 (avx512f_vpermi2var<mode>3_mask): Ditto.
11036 (avx512f_vpermi2var<mode>3_mask): Ditto.
11037 (avx512f_vpermt2var<mode>3_maskz): Ditto.
11038 (*avx512f_gathersi<mode>): Ditto.
11039 (*avx512f_gathersi<mode>_2): Ditto.
11040 (*avx512f_gatherdi<mode>): Ditto.
11041 (*avx512f_gatherdi<mode>_2): Ditto.
11042 (*avx512f_scattersi<mode>): Ditto.
11043 (*avx512f_scatterdi<mode>): Ditto.
11044 (avx512f_compress<mode>_mask): Ditto.
11045 (avx512f_compressstore<mode>_mask): Ditto.
11046 (avx512f_expand<mode>_mask): Ditto.
11047 * config/i386/subst.md (mask): Change k to Yk.
11048 (mask_scalar_merge): Ditto.
11051 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
11053 * doc/extend.texi (Vector Extensions): Document ?: in C++.
11055 2014-01-31 Richard Biener <rguenther@suse.de>
11057 PR middle-end/59990
11058 * builtins.c (fold_builtin_memory_op): Make sure to not
11059 use a floating-point mode or a boolean or enumeral type for
11060 the copy operation.
11062 2014-01-30 DJ Delorie <dj@redhat.com>
11064 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
11065 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
11066 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
11067 whenever main() has an epilogue.
11069 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11071 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
11072 unused variable "field".
11073 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
11074 (vsx_mergeh_<mode>): Likewise.
11075 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
11076 (altivec_vmrghh): Likewise.
11077 (altivec_vmrghw): Likewise.
11078 (altivec_vmrglb): Likewise.
11079 (altivec_vmrglh): Likewise.
11080 (altivec_vmrglw): Likewise.
11081 (altivec_vspltb): Add missing uses.
11082 (altivec_vsplth): Likewise.
11083 (altivec_vspltw): Likewise.
11084 (altivec_vspltsf): Likewise.
11086 2014-01-30 Jakub Jelinek <jakub@redhat.com>
11089 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
11090 frame related instructions.
11092 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
11094 PR rtl-optimization/59959
11095 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
11096 any reload of register whose subreg is invalid.
11098 2014-01-30 Jakub Jelinek <jakub@redhat.com>
11100 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
11101 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
11102 Add missing return type - void.
11104 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11106 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
11107 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
11108 remove element index adjustment for endian (now handled in vsx.md
11110 (altivec_expand_vec_perm_const): Use
11111 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
11112 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
11113 (vsx_xxspltw_<mode>): Adjust element index for little endian.
11114 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
11115 define_expand and a new define_insn *altivec_vspltb_internal;
11116 adjust for -maltivec=be on a little endian target.
11117 (altivec_vspltb_direct): New.
11118 (altivec_vsplth): Divide into a define_expand and a new
11119 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
11120 little endian target.
11121 (altivec_vsplth_direct): New.
11122 (altivec_vspltw): Divide into a define_expand and a new
11123 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
11124 little endian target.
11125 (altivec_vspltw_direct): New.
11126 (altivec_vspltsf): Divide into a define_expand and a new
11127 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
11128 a little endian target.
11130 2014-01-30 Richard Biener <rguenther@suse.de>
11132 PR tree-optimization/59993
11133 * tree-ssa-forwprop.c (associate_pointerplus): Check we
11134 can propagate form the earlier stmt and avoid the transform
11135 when the intermediate result is needed.
11137 2014-01-30 Alangi Derick <alangiderick@gmail.com>
11139 * README.Portability: Fix typo.
11141 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
11143 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
11144 comparison_operator with ordered_comparison_operator.
11146 2014-01-30 Nick Clifton <nickc@redhat.com>
11148 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
11149 Rename to mn10300_store_multiple_regs.
11150 * config/mn10300/mn10300.c: Likewise.
11151 * config/mn10300/mn10300.md (store_movm): Fix typo: call
11152 store_multiple_regs.
11153 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
11154 Call mn10300_store_multiple_regs.
11156 2014-01-30 Nick Clifton <nickc@redhat.com>
11157 DJ Delorie <dj@redhat.com>
11159 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
11160 %fp 2 to keep registers after it properly word-aligned.
11161 (rl78_alloc_physical_registers_umul): Handle the case where both
11162 input operands are the same.
11164 2014-01-30 Richard Biener <rguenther@suse.de>
11166 PR tree-optimization/59903
11167 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
11170 2014-01-30 Jason Merrill <jason@redhat.com>
11173 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
11176 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
11178 2014-01-30 Richard Biener <rguenther@suse.de>
11180 PR tree-optimization/59951
11181 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
11183 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
11186 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
11187 SFmode to DFmode case.
11189 2014-01-29 DJ Delorie <dj@redhat.com>
11191 * config/msp430/msp430.opt (-minrt): New.
11192 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
11194 (ENDFILE_SPEC): Likewise.
11196 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
11198 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
11199 (estimate_function_body_sizes): Use it.
11201 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
11204 * dwarf2out.c (is_cxx_auto): New.
11205 (is_base_type): Use it.
11206 (gen_type_die_with_usage): Likewise.
11208 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11210 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
11211 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
11212 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
11213 -maltivec=be with LE targets.
11214 (vsx_mergeh_<mode>): Likewise.
11215 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
11216 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
11217 (altivec_vmrghb): Replace with define_expand and new
11218 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
11219 (altivec_vmrghb_direct): New define_insn.
11220 (altivec_vmrghh): Replace with define_expand and new
11221 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
11222 (altivec_vmrghh_direct): New define_insn.
11223 (altivec_vmrghw): Replace with define_expand and new
11224 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
11225 (altivec_vmrghw_direct): New define_insn.
11226 (*altivec_vmrghsf): Adjust for endianness.
11227 (altivec_vmrglb): Replace with define_expand and new
11228 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
11229 (altivec_vmrglb_direct): New define_insn.
11230 (altivec_vmrglh): Replace with define_expand and new
11231 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
11232 (altivec_vmrglh_direct): New define_insn.
11233 (altivec_vmrglw): Replace with define_expand and new
11234 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
11235 (altivec_vmrglw_direct): New define_insn.
11236 (*altivec_vmrglsf): Adjust for endianness.
11237 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
11238 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
11239 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
11240 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
11241 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
11242 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
11243 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
11244 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
11246 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
11248 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
11249 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
11252 2014-01-29 Richard Biener <rguenther@suse.de>
11254 PR tree-optimization/58742
11255 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
11256 associate_pointerplus_align.
11257 (associate_pointerplus_diff): New function.
11258 (associate_pointerplus): Likewise. Call associate_pointerplus_align
11259 and associate_pointerplus_diff.
11261 2014-01-29 Richard Biener <rguenther@suse.de>
11263 * lto-streamer.h (LTO_major_version): Bump to 3.
11264 (LTO_minor_version): Reset to 0.
11266 2014-01-29 Renlin Li <Renlin.Li@arm.com>
11268 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
11269 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
11270 (arm_file_start): Generate correct asm header for armv7ve.
11271 * config/arm/bpabi.h: Add multilib support for armv7ve.
11272 * config/arm/driver-arm.c: Change the architectures of cortex-a7
11273 and cortex-a15 to armv7ve.
11274 * config/arm/t-aprofile: Add multilib support for armv7ve.
11275 * doc/invoke.texi: Document -march=armv7ve.
11277 2014-01-29 Richard Biener <rguenther@suse.de>
11279 PR tree-optimization/58742
11280 * tree-ssa-forwprop.c (associate_plusminus): Return true
11281 if we changed sth, defer EH cleanup to ...
11282 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
11283 (simplify_mult): New function.
11285 2014-01-29 Jakub Jelinek <jakub@redhat.com>
11287 PR middle-end/59917
11288 PR tree-optimization/59920
11289 * tree.c (build_common_builtin_nodes): Remove
11290 __builtin_setjmp_dispatcher initialization.
11291 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
11292 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
11293 instead of gsi_after_labels + manually skipping debug stmts.
11294 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
11295 ignore bbs with IFN_ABNORMAL_DISPATCHER.
11296 * tree-inline.c (copy_edges_for_bb): Remove
11297 can_make_abnormal_goto argument, instead add abnormal_goto_dest
11298 argument. Ignore computed_goto_p stmts. Don't call
11299 make_abnormal_goto_edges. If a call might need abnormal edges
11300 for non-local gotos, see if it already has an edge to
11301 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
11302 with true argument, don't do anything then, otherwise add
11303 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
11304 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
11306 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
11307 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
11308 (lower_stmt): Don't set data->calls_builtin_setjmp.
11309 (lower_builtin_setjmp): Adjust comment.
11310 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
11311 * tree-cfg.c (found_computed_goto): Remove.
11312 (factor_computed_gotos): Remove.
11313 (make_goto_expr_edges): Return bool, true for computed gotos.
11314 Don't call make_abnormal_goto_edges.
11315 (build_gimple_cfg): Don't set found_computed_goto, don't call
11316 factor_computed_gotos.
11317 (computed_goto_p): No longer static.
11318 (make_blocks): Don't set found_computed_goto.
11319 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
11320 (make_edges): If make_goto_expr_edges returns true, push bb
11321 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
11322 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
11323 vector. Record mapping between bbs and OpenMP regions if there
11324 are any, adjust make_gimple_omp_edges caller. Call
11325 handle_abnormal_edges.
11326 (make_abnormal_goto_edges): Remove.
11327 * tree-cfg.h (make_abnormal_goto_edges): Remove.
11328 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
11329 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
11330 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
11331 * internal-fn.def (ABNORMAL_DISPATCHER): New.
11332 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
11333 filling *region also set *region_idx to (*region)->entry->index.
11336 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
11337 For REGs set ORIGINAL_REGNO.
11339 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
11341 * doc/md.texi: Mention that a target shouldn't implement
11342 vec_widen_(s|u)mul_even/odd pair if it is less efficient
11345 2014-01-29 Jakub Jelinek <jakub@redhat.com>
11347 PR tree-optimization/59594
11348 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
11349 a copy of the datarefs vector rather than the vector itself.
11351 2014-01-28 Jason Merrill <jason@redhat.com>
11354 * dwarf2out.c (auto_die): New static.
11355 (gen_type_die_with_usage): Handle C++1y 'auto'.
11356 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
11359 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
11362 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
11363 (SPEC_X32): Likewise.
11364 (SPEC_64): Likewise.
11365 * config/i386/i386.c (ix86_option_override_internal): Turn off
11366 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
11368 (x86_file_start): Output .code16gcc for TARGET_16BIT.
11369 * config/i386/i386.h (TARGET_16BIT): New macro.
11370 (TARGET_16BIT_P): Likewise.
11371 * config/i386/i386.opt: Add m16.
11372 * doc/invoke.texi: Document -m16.
11374 2014-01-28 Jakub Jelinek <jakub@redhat.com>
11376 PR preprocessor/59935
11377 * input.c (location_get_source_line): Bail out on when line number
11378 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
11380 2014-01-28 Richard Biener <rguenther@suse.de>
11382 PR tree-optimization/58742
11383 * tree-ssa-forwprop.c (associate_plusminus): Handle
11384 pointer subtraction of the form (T)(P + A) - (T)P.
11386 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11388 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
11391 2014-01-28 Richard Biener <rguenther@suse.de>
11394 2014-01-28 Richard Biener <rguenther@suse.de>
11396 PR rtl-optimization/45364
11397 PR rtl-optimization/59890
11398 * var-tracking.c (local_get_addr_clear_given_value): Handle
11399 already cleared slot.
11400 (val_reset): Handle not allocated local_get_addr_cache.
11401 (vt_find_locations): Use post-order on the inverted CFG.
11403 2014-01-28 Richard Biener <rguenther@suse.de>
11405 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
11407 2014-01-28 Richard Biener <rguenther@suse.de>
11409 PR rtl-optimization/45364
11410 PR rtl-optimization/59890
11411 * var-tracking.c (local_get_addr_clear_given_value): Handle
11412 already cleared slot.
11413 (val_reset): Handle not allocated local_get_addr_cache.
11414 (vt_find_locations): Use post-order on the inverted CFG.
11416 2014-01-28 Alan Modra <amodra@gmail.com>
11418 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
11419 * configure.ac <recursive call for build != host>: Define
11420 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
11421 and LD_FOR_BUILD too.
11422 * configure: Regenerate.
11424 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
11426 * config/i386/i386.c (get_builtin_code_for_version): Separate
11427 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
11428 Broadwell from Haswell.
11430 2014-01-27 Steve Ellcey <sellcey@mips.com>
11432 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
11433 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
11434 * config/mips/mips.c (mips_option_override): Change setting
11436 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
11437 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
11438 Change from Mask to Var.
11440 2014-01-27 Jeff Law <law@redhat.com>
11442 * ipa-inline.c (inline_small_functions): Fix typo.
11444 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
11446 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
11447 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
11448 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
11449 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
11450 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
11451 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
11452 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
11453 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
11454 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
11455 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
11456 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
11457 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
11458 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
11459 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
11460 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
11461 (_mm512_storeu_epi64): Ditto.
11462 (_mm512_cmpge_epi32_mask): Ditto.
11463 (_mm512_cmpge_epu32_mask): Ditto.
11464 (_mm512_cmpge_epi64_mask): Ditto.
11465 (_mm512_cmpge_epu64_mask): Ditto.
11466 (_mm512_cmple_epi32_mask): Ditto.
11467 (_mm512_cmple_epu32_mask): Ditto.
11468 (_mm512_cmple_epi64_mask): Ditto.
11469 (_mm512_cmple_epu64_mask): Ditto.
11470 (_mm512_cmplt_epi32_mask): Ditto.
11471 (_mm512_cmplt_epu32_mask): Ditto.
11472 (_mm512_cmplt_epi64_mask): Ditto.
11473 (_mm512_cmplt_epu64_mask): Ditto.
11474 (_mm512_cmpneq_epi32_mask): Ditto.
11475 (_mm512_cmpneq_epu32_mask): Ditto.
11476 (_mm512_cmpneq_epi64_mask): Ditto.
11477 (_mm512_cmpneq_epu64_mask): Ditto.
11478 (_mm512_expand_pd): Ditto.
11479 (_mm512_expand_ps): Ditto.
11480 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
11481 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
11482 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
11483 * config/i386/i386.c (ix86_builtins): Add
11484 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
11485 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
11486 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
11487 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
11488 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
11489 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
11490 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
11491 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
11492 IX86_BUILTIN_PMOVUSQW512_MEM.
11493 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
11494 __builtin_ia32_pmovsqd512mem_mask,
11495 __builtin_ia32_pmovqd512mem_mask,
11496 __builtin_ia32_pmovusqw512mem_mask,
11497 __builtin_ia32_pmovsqw512mem_mask,
11498 __builtin_ia32_pmovqw512mem_mask,
11499 __builtin_ia32_pmovusdw512mem_mask,
11500 __builtin_ia32_pmovsdw512mem_mask,
11501 __builtin_ia32_pmovdw512mem_mask,
11502 __builtin_ia32_pmovqb512mem_mask,
11503 __builtin_ia32_pmovusqb512mem_mask,
11504 __builtin_ia32_pmovsqb512mem_mask,
11505 __builtin_ia32_pmovusdb512mem_mask,
11506 __builtin_ia32_pmovsdb512mem_mask,
11507 __builtin_ia32_pmovdb512mem_mask.
11508 (bdesc_args): Add __builtin_ia32_expanddf512,
11509 __builtin_ia32_expandsf512.
11510 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
11511 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
11512 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
11513 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
11514 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
11515 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
11516 (avx512f_<code>v8div16qi2_mask_store): This.
11517 (avx512f_expand<mode>): New.
11519 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
11521 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
11523 (_mm512_mask_prefetch_i64gather_pd): Ditto.
11524 (_mm512_prefetch_i32scatter_pd): Ditto.
11525 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
11526 (_mm512_prefetch_i64scatter_pd): Ditto.
11527 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
11528 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
11529 (_mm512_mask_prefetch_i64gather_ps): Ditto.
11530 (_mm512_prefetch_i32scatter_ps): Ditto.
11531 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
11532 (_mm512_prefetch_i64scatter_ps): Ditto.
11533 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
11534 * config/i386/i386-builtin-types.def: Define
11535 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
11536 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
11537 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
11538 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
11539 IX86_BUILTIN_SCATTERPFQPD.
11540 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
11541 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
11542 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
11543 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
11544 __builtin_ia32_scatterpfqps.
11545 (ix86_expand_builtin): Expand new built-ins.
11546 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
11547 fix memory access data type.
11548 (*avx512pf_gatherpf<mode>_mask): Ditto.
11549 (*avx512pf_gatherpf<mode>): Ditto.
11550 (avx512pf_scatterpf<mode>): Ditto.
11551 (*avx512pf_scatterpf<mode>_mask): Ditto.
11552 (*avx512pf_scatterpf<mode>): Ditto.
11553 (GATHER_SCATTER_SF_MEM_MODE): New.
11554 (avx512pf_gatherpf<mode>df): Ditto.
11555 (*avx512pf_gatherpf<mode>df_mask): Ditto.
11556 (*avx512pf_scatterpf<mode>df): Ditto.
11558 2014-01-27 Jakub Jelinek <jakub@redhat.com>
11561 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
11562 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
11565 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
11567 * common/config/arm/arm-common.c
11568 (arm_rewrite_mcpu): Handle multiple names.
11570 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
11572 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
11574 * gimple-builder.h (create_gimple_tmp): Delete.
11576 2014-01-27 Christian Bruel <christian.bruel@st.com>
11578 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
11581 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
11583 * config/pa/pa.md (call): Generate indirect long calls to non-local
11584 functions when outputing 32-bit code.
11585 (call_value): Likewise except for special call to buggy powf function.
11587 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
11588 portable runtime and PIC indirect calls.
11589 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
11590 and PIC call sequences. Use ldo instead of blr to set return register
11591 in PIC call sequence.
11593 2014-01-25 Walter Lee <walt@tilera.com>
11595 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
11596 avoid clobbering a live register.
11598 2014-01-25 Walter Lee <walt@tilera.com>
11600 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
11601 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
11602 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
11603 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
11605 2014-01-25 Walter Lee <walt@tilera.com>
11607 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
11608 arguments on even registers.
11609 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
11611 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
11612 (BIGGEST_ALIGNMENT): Ditto.
11613 (BIGGEST_FIELD_ALIGNMENT): Ditto.
11615 2014-01-25 Walter Lee <walt@tilera.com>
11617 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
11618 insns before bundling.
11619 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
11621 2014-01-25 Walter Lee <walt@tilera.com>
11623 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
11624 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
11625 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
11627 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
11629 * config/mips/constraints.md (kl): Delete.
11630 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
11631 define expands, using...
11632 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
11633 instructions for MIPS16.
11634 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
11635 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
11637 2014-01-25 Walter Lee <walt@tilera.com>
11639 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
11643 2014-01-25 Walter Lee <walt@tilera.com>
11645 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
11646 (TARGET_EXPAND_TO_RTL_HOOK): Define.
11648 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
11650 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
11653 2014-01-25 Jakub Jelinek <jakub@redhat.com>
11655 * print-rtl.c (in_call_function_usage): New var.
11656 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
11657 EXPR_LIST mode as mode and not as reg note name.
11659 PR middle-end/59561
11660 * cfgloopmanip.c (copy_loop_info): If
11661 loop->warned_aggressive_loop_optimizations, make sure
11662 the flag is set in target loop too.
11664 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
11666 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
11668 * builtins.def: Likewise.
11669 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
11670 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
11671 * ira.c (ira_setup_eliminable_regset): Likewise.
11672 * omp-low.c (gate_expand_omp): Likewise.
11673 (execute_lower_omp): Likewise.
11674 (diagnose_sb_0): Likewise.
11675 (gate_diagnose_omp_blocks): Likewise.
11676 (simd_clone_clauses_extract): Likewise.
11679 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11681 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
11682 correction for little endian...
11683 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
11686 2014-01-24 Jeff Law <law@redhat.com>
11688 PR tree-optimization/59919
11689 * tree-vrp.c (find_assert_locations_1): Do not register asserts
11690 for non-returning calls.
11692 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
11694 * common/config/aarch64/aarch64-common.c
11695 (aarch64_rewrite_mcpu): Handle multiple names.
11696 * config/aarch64/aarch64.h
11697 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
11699 2014-01-24 Dodji Seketeli <dodji@redhat.com>
11701 * input.c (add_file_to_cache_tab): Handle the case where fopen
11704 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
11707 * config/i386/i386.md (pushsf splitter): Get stack adjustment
11708 from push operand if code of push isn't PRE_DEC.
11710 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
11713 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
11714 -mquad-memory-atomic. Update -mquad-memory documentation to say
11715 it is only used for non-atomic loads/stores.
11717 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
11718 -mquad-memory or -mquad-memory-atomic switches.
11720 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
11721 -mquad-memory-atomic to ISA 2.07 support.
11723 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
11724 to separate support of normal quad word memory operations (ldq, stq)
11725 from the atomic quad word memory operations.
11727 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
11728 support to separate non-atomic quad word operations from atomic
11729 quad word operations. Disable non-atomic quad word operations in
11730 little endian mode so that we don't have to swap words after the
11731 load and before the store.
11732 (quad_load_store_p): Add comment about atomic quad word support.
11733 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
11734 options printed with -mdebug=reg.
11736 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
11737 -mquad-memory-atomic as the test for whether we have quad word
11738 atomic instructions.
11739 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
11740 or -mp8-vector are used, allow byte/half-word atomic operations.
11742 * config/rs6000/sync.md (load_lockedti): Insure that the address
11743 is a proper indexed or indirect address for the lqarx instruction.
11744 On little endian systems, swap the hi/lo registers after the lqarx
11746 (load_lockedpti): Use indexed_or_indirect_operand predicate to
11747 insure the address is valid for the lqarx instruction.
11748 (store_conditionalti): Insure that the address is a proper indexed
11749 or indirect address for the stqcrx. instruction. On little endian
11750 systems, swap the hi/lo registers before doing the stqcrx.
11752 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
11753 insure the address is valid for the stqcrx. instruction.
11755 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
11756 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
11757 type of quad memory support is available.
11759 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
11761 PR regression/59915
11762 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
11763 there is a danger of looping.
11765 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
11767 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
11768 force flag_ira_loop_pressure if set via command line.
11770 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
11772 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
11773 (ashr_simd): New builtin handling DI mode.
11774 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
11775 (aarch64_sshr_simddi): New match pattern.
11776 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
11777 (vshrd_n_s64): Likewise.
11778 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
11780 2014-01-23 Nick Clifton <nickc@redhat.com>
11782 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
11783 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
11784 favour of mcu specific scripts.
11785 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
11788 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
11789 Alex Velenko <Alex.Velenko@arm.com>
11791 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
11792 (vaddv_s16): Likewise.
11793 (vaddv_s32): Likewise.
11794 (vaddv_u8): Likewise.
11795 (vaddv_u16): Likewise.
11796 (vaddv_u32): Likewise.
11797 (vaddvq_s8): Likewise.
11798 (vaddvq_s16): Likewise.
11799 (vaddvq_s32): Likewise.
11800 (vaddvq_s64): Likewise.
11801 (vaddvq_u8): Likewise.
11802 (vaddvq_u16): Likewise.
11803 (vaddvq_u32): Likewise.
11804 (vaddvq_u64): Likewise.
11805 (vaddv_f32): Likewise.
11806 (vaddvq_f32): Likewise.
11807 (vaddvq_f64): Likewise.
11808 (vmaxv_f32): Likewise.
11809 (vmaxv_s8): Likewise.
11810 (vmaxv_s16): Likewise.
11811 (vmaxv_s32): Likewise.
11812 (vmaxv_u8): Likewise.
11813 (vmaxv_u16): Likewise.
11814 (vmaxv_u32): Likewise.
11815 (vmaxvq_f32): Likewise.
11816 (vmaxvq_f64): Likewise.
11817 (vmaxvq_s8): Likewise.
11818 (vmaxvq_s16): Likewise.
11819 (vmaxvq_s32): Likewise.
11820 (vmaxvq_u8): Likewise.
11821 (vmaxvq_u16): Likewise.
11822 (vmaxvq_u32): Likewise.
11823 (vmaxnmv_f32): Likewise.
11824 (vmaxnmvq_f32): Likewise.
11825 (vmaxnmvq_f64): Likewise.
11826 (vminv_f32): Likewise.
11827 (vminv_s8): Likewise.
11828 (vminv_s16): Likewise.
11829 (vminv_s32): Likewise.
11830 (vminv_u8): Likewise.
11831 (vminv_u16): Likewise.
11832 (vminv_u32): Likewise.
11833 (vminvq_f32): Likewise.
11834 (vminvq_f64): Likewise.
11835 (vminvq_s8): Likewise.
11836 (vminvq_s16): Likewise.
11837 (vminvq_s32): Likewise.
11838 (vminvq_u8): Likewise.
11839 (vminvq_u16): Likewise.
11840 (vminvq_u32): Likewise.
11841 (vminnmv_f32): Likewise.
11842 (vminnmvq_f32): Likewise.
11843 (vminnmvq_f64): Likewise.
11845 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
11847 * config/aarch64/aarch64-simd.md
11848 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
11849 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
11850 (*aarch64_mul3_elt<mode>): Likewise.
11851 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
11852 (*aarch64_mul3_elt_to_64v2df): Likewise.
11853 (*aarch64_mla_elt<mode>): Likewise.
11854 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
11855 (*aarch64_mls_elt<mode>): Likewise.
11856 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
11857 (*aarch64_fma4_elt<mode>): Likewise.
11858 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
11859 (*aarch64_fma4_elt_to_64v2df): Likewise.
11860 (*aarch64_fnma4_elt<mode>): Likewise.
11861 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
11862 (*aarch64_fnma4_elt_to_64v2df): Likewise.
11863 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
11864 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
11865 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
11866 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
11867 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
11868 (aarch64_sqdmull_lane<mode>_internal): Likewise.
11869 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
11871 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
11873 * config/aarch64/aarch64-simd.md
11874 (aarch64_be_checked_get_lane<mode>): New define_expand.
11875 * config/aarch64/aarch64-simd-builtins.def
11876 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
11877 New builtin definition.
11878 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
11879 Use new safe be builtin.
11881 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
11883 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
11885 (aarch64_be_st1<mode>): Likewise.
11886 (aarch_ld1<VALL:mode>): Define_expand modified.
11887 (aarch_st1<VALL:mode>): Likewise.
11888 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
11889 (UNSPEC_ST1): Likewise.
11891 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
11893 * config/microblaze/microblaze.md: Add trap insn and attribute
11895 2014-01-23 Dodji Seketeli <dodji@redhat.com>
11897 PR preprocessor/58580
11898 * input.h (location_get_source_line): Take an additional line_size
11900 (void diagnostics_file_cache_fini): Declare new function.
11901 * input.c (struct fcache): New type.
11902 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
11903 New static constants.
11904 (diagnostic_file_cache_init, total_lines_num)
11905 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
11906 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
11907 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
11908 (get_next_line, read_next_line, goto_next_line, read_line_num):
11909 New static function definitions.
11910 (diagnostic_file_cache_fini): New function.
11911 (location_get_source_line): Take an additional output line_len
11912 parameter. Re-write using lookup_or_add_file_to_cache_tab and
11914 * diagnostic.c (diagnostic_finish): Call
11915 diagnostic_file_cache_fini.
11916 (adjust_line): Take an additional input parameter for the length
11917 of the line, rather than calculating it with strlen.
11918 (diagnostic_show_locus): Adjust the use of
11919 location_get_source_line and adjust_line with respect to their new
11920 signature. While displaying a line now, do not stop at the first
11921 null byte. Rather, display the zero byte as a space and keep
11922 going until we reach the size of the line.
11923 * Makefile.in: Add vec.o to OBJS-libcommon
11925 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
11926 Ilya Tocar <ilya.tocar@intel.com>
11928 * config/i386/avx512fintrin.h (_mm512_kmov): New.
11929 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
11930 (__builtin_ia32_kmov16): Ditto.
11931 * config/i386/i386.md (UNSPEC_KMOV): New.
11934 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
11936 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
11937 (_mm512_storeu_si512): Ditto.
11939 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
11942 * rtl.h (get_referenced_operands): Declare.
11943 * recog.c (get_referenced_operands): New function.
11944 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
11945 operands have been referenced when recording LO_SUM references.
11947 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
11949 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
11951 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
11953 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
11954 Enable for generic and recent AMD targets.
11956 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
11958 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
11959 ARG_SIZE note when adjustment was eliminated.
11961 2014-01-22 Jeff Law <law@redhat.com>
11963 PR tree-optimization/59597
11964 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
11965 in file. Accept new argument REGISTERING and use it to modify
11966 dump output appropriately.
11967 (register_jump_thread): Corresponding changes.
11968 (mark_threaded_blocks): Reinstate code to cancel unprofitable
11969 thread paths involving joiner blocks. Add code to dump cancelled
11970 jump threading paths.
11972 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
11974 PR rtl-optimization/59477
11975 * lra-constraints.c (inherit_in_ebb): Process call for living hard
11976 regs. Update reloads_num and potential_reload_hard_regs for all insns.
11978 2014-01-22 Tom Tromey <tromey@redhat.com>
11980 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
11982 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
11984 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
11986 PR rtl-optimization/59896
11987 * lra-constraints.c (process_alt_operands): Check unused note for
11988 matched operands of insn with no output reloads.
11990 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
11992 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
11993 (mips_move_from_gpr_cost): Likewise.
11995 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
11997 PR rtl-optimization/59858
11998 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
11999 ira_class_hard_regs_num.
12000 (process_alt_operands): Increase reject for dying matched operand.
12002 2014-01-21 Jakub Jelinek <jakub@redhat.com>
12005 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
12006 smaller than size, perform several stores or loads and stores
12007 at dst + count - size to store or copy all of size bytes, rather
12008 than just last modesize bytes.
12010 2014-01-20 DJ Delorie <dj@redhat.com>
12012 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
12013 that CLOBBERs are REGs before propogating their values.
12015 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
12017 PR middle-end/59789
12018 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
12019 (cgraph_inline_failed_type): New function.
12020 * cgraph.h (DEFCIFCODE): Add type.
12021 (cgraph_inline_failed_type_t): New enum.
12022 (cgraph_inline_failed_type): New prototype.
12023 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
12024 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
12025 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
12026 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
12027 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
12028 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
12029 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
12030 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
12031 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
12032 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
12033 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
12034 OPTIMIZATION_MISMATCH.
12035 * tree-inline.c (expand_call_inline): Emit errors during
12036 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
12038 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
12041 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
12042 mode attribute in insn output.
12044 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
12046 * output.h (output_constant): Delete.
12047 * varasm.c (output_constant): Make private.
12049 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
12051 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
12053 2014-01-20 Jakub Jelinek <jakub@redhat.com>
12055 PR middle-end/59860
12056 * tree.h (fold_builtin_strcat): New prototype.
12057 * builtins.c (fold_builtin_strcat): No longer static. Add len
12058 argument, if non-NULL, don't call c_strlen. Optimize
12059 directly into __builtin_memcpy instead of __builtin_strcpy.
12060 (fold_builtin_2): Adjust fold_builtin_strcat caller.
12061 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
12063 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
12065 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
12066 for SImode_address_operand operands, having only a REG argument.
12068 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
12070 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
12071 loader name using mbig-endian.
12072 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
12074 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
12076 * doc/invoke.texi (-march): Clarify documentation for AArch64.
12077 (-mtune): Likewise.
12080 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
12082 * config/aarch64/aarch64-protos.h
12083 (aarch64_cannot_change_mode_class_ptr): Declare.
12084 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
12085 aarch64_cannot_change_mode_class_ptr): New.
12086 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
12087 backend hook aarch64_cannot_change_mode_class.
12089 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
12091 * common/config/aarch64/aarch64-common.c
12092 (aarch64_handle_option): Don't handle any option order logic here.
12093 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
12094 selected_cpu, warn on architecture version mismatch.
12095 (aarch64_override_options): Fix parsing order for option strings.
12097 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12098 Iain Sandoe <iain@codesourcery.com>
12101 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
12102 warning. Amend comment to reflect current functionality.
12104 2014-01-20 Richard Biener <rguenther@suse.de>
12106 PR middle-end/59860
12107 * builtins.c (fold_builtin_strcat): Remove case better handled
12108 by tree-ssa-strlen.c.
12110 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
12112 * config/aarch64/aarch64.opt
12113 (mcpu, march, mtune): Make case-insensitive.
12115 2014-01-20 Jakub Jelinek <jakub@redhat.com>
12118 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
12119 if operands[1] is a REG or ZERO_EXTEND of a REG.
12121 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
12123 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
12125 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
12127 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
12128 long non-pic millicode calls.
12130 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12132 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
12134 2014-01-19 Kito Cheng <kito@0xlab.org>
12136 * builtins.c (expand_movstr): Check movstr expand done or fail.
12138 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
12139 H.J. Lu <hongjiu.lu@intel.com>
12142 * config/i386/i386.md (*lea<mode>): Zero-extend return register
12143 to DImode for zero-extended addresses.
12145 2014-01-19 Jakub Jelinek <jakub@redhat.com>
12147 PR rtl-optimization/57763
12148 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
12149 on the new indirect jump_insn and increment LABEL_NUSES (label).
12151 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
12155 * config.gcc (x86_archs): New variable.
12156 (x86_64_archs): Likewise.
12157 (x86_cpus): Likewise.
12158 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
12159 --with-arch/--with-cpu= options.
12160 Support --with-arch=/--with-cpu={nehalem,westmere,
12161 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
12163 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
12165 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
12166 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
12168 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
12170 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
12172 2014-01-18 Jakub Jelinek <jakub@redhat.com>
12175 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
12176 clear cpp_get_options (parse_in)->warn_unused_macros for
12177 ix86_target_macros_internal with cpp_define.
12179 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
12181 * jump.c (delete_related_insns): Keep (use (insn))s.
12182 * reorg.c (redundant_insn): Check for barriers too.
12184 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
12186 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
12188 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
12190 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
12191 call to $$dyncall when TARGET_LONG_CALLS is true.
12193 2014-01-17 Jeff Law <law@redhat.com>
12195 * ree.c (combine_set_extension): Temporarily disable test for
12196 changing number of hard registers.
12198 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
12200 PR middle-end/58125
12201 * ipa-inline-analysis.c (inline_free_summary):
12202 Do not free summary of aliases.
12204 2014-01-17 Jakub Jelinek <jakub@redhat.com>
12206 PR middle-end/59706
12207 * gimplify.c (gimplify_expr): Use create_tmp_var
12208 instead of create_tmp_var_raw. If cond doesn't have
12209 integral type, don't add the IFN_ANNOTATE builtin at all.
12211 2014-01-17 Martin Jambor <mjambor@suse.cz>
12214 * ipa-cp.c (prev_edge_clone): New variable.
12215 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
12216 Also resize prev_edge_clone vector.
12217 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
12218 (ipcp_edge_removal_hook): New function.
12219 (ipcp_driver): Register ipcp_edge_removal_hook.
12221 2014-01-17 Andrew Pinski <apinski@cavium.com>
12222 Steve Ellcey <sellcey@mips.com>
12225 * config/mips/mips.c (mips_print_operand): Check operand mode instead
12228 2014-01-17 Jeff Law <law@redhat.com>
12230 PR middle-end/57904
12231 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
12232 so that pass_ccp runs first.
12234 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
12236 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
12237 (ix86_adjust_cost): Use !TARGET_XXX.
12238 (do_reorder_for_imul): Likewise.
12239 (swap_top_of_ready_list): Likewise.
12240 (ix86_sched_reorder): Likewise.
12242 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
12244 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
12245 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
12246 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
12247 (intel_memset): New. Duplicate slm_memset.
12248 (intel_cost): New. Duplicate slm_cost.
12249 (m_INTEL): New macro.
12250 (processor_target_table): Add "intel".
12251 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
12252 with PROCESSOR_INTEL for "intel".
12253 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
12254 PROCESSOR_SILVERMONT.
12255 (ix86_issue_rate): Likewise.
12256 (ix86_adjust_cost): Likewise.
12257 (ia32_multipass_dfa_lookahead): Likewise.
12258 (swap_top_of_ready_list): Likewise.
12259 (ix86_sched_reorder): Likewise.
12260 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
12261 instead of TARGET_OPT_AGU.
12262 * config/i386/i386.h (TARGET_INTEL): New.
12263 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
12264 (processor_type): Add PROCESSOR_INTEL.
12265 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
12266 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
12268 2014-01-17 Marek Polacek <polacek@redhat.com>
12271 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
12274 2014-01-17 Richard Biener <rguenther@suse.de>
12276 PR tree-optimization/46590
12277 * opts.c (default_options_table): Add entries for
12278 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
12279 all enabled at -O1 but not for -Og.
12280 * common.opt (fbranch-count-reg): Remove Init(1).
12281 (fmove-loop-invariants): Likewise.
12282 (ftree-pta): Likewise.
12284 2014-01-17 Jakub Jelinek <jakub@redhat.com>
12286 * config/i386/i386.c (ix86_data_alignment): For compatibility with
12287 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
12288 decls to at least the GCC 4.8 used alignments.
12291 * tree-nested.c (convert_nonlocal_reference_stmt,
12292 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
12293 of GIMPLE_BIND stmts, adjust associated decls.
12295 2014-01-17 Richard Biener <rguenther@suse.de>
12297 PR tree-optimization/46590
12298 * vec.h (vec<>::bseach): New member function implementing
12299 binary search according to C89 bsearch.
12300 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
12301 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
12302 bitmap pointer again. Make accesses_in_loop a flat array.
12303 (mem_ref_obstack): New global.
12304 (outermost_indep_loop): Adjust for mem_ref->stored changes.
12305 (mark_ref_stored): Likewise.
12306 (ref_indep_loop_p_2): Likewise.
12307 (set_ref_stored_in_loop): New helper function.
12308 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
12309 (memref_free): Adjust.
12310 (record_mem_ref_loc): Simplify.
12311 (gather_mem_refs_stmt): Adjust.
12312 (sort_locs_in_loop_postorder_cmp): New function.
12313 (analyze_memory_references): Sort accesses_in_loop after
12314 loop postorder number.
12315 (find_ref_loc_in_loop_cmp): New function.
12316 (for_all_locs_in_loop): Find relevant cluster of locs in
12317 accesses_in_loop and iterate without recursion.
12318 (execute_sm): Avoid uninit warning.
12319 (struct ref_always_accessed): Simplify.
12320 (ref_always_accessed::operator ()): Likewise.
12321 (ref_always_accessed_p): Likewise.
12322 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
12323 loop postorder numbers here.
12324 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
12327 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
12330 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
12331 on decls for which assemble_alias has been called.
12333 2014-01-17 Nick Clifton <nickc@redhat.com>
12335 * config/msp430/msp430.opt: (mcpu): New option.
12336 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
12337 (msp430_option_override): Parse target_cpu. If the MCU name
12338 matches a generic string, clear target_mcu.
12339 (msp430_attr): Allow numeric interrupt values up to 63.
12340 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
12341 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
12343 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
12345 * config/msp430/msp430.md (popm): Use %J rather than %I.
12346 (addsi3): Use msp430_nonimmediate_operand for operand 2.
12347 (addhi_cy_i): Use immediate_operand for operand 2.
12348 * doc/invoke.texi: Document -mcpu option.
12350 2014-01-17 Richard Biener <rguenther@suse.de>
12352 PR rtl-optimization/38518
12353 * df.h (df_analyze_loop): Declare.
12354 * df-core.c: Include cfgloop.h.
12355 (df_analyze_1): Split out main part of df_analyze.
12356 (df_analyze): Adjust.
12357 (loop_inverted_post_order_compute): New function.
12358 (loop_post_order_compute): Likewise.
12359 (df_analyze_loop): New function avoiding whole-function
12360 postorder computes.
12361 * loop-invariant.c (find_defs): Use df_analyze_loop.
12362 (find_invariants): Adjust.
12363 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
12365 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
12367 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
12368 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
12370 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
12372 * ipa-ref.c (ipa_remove_stmt_references): Fix references
12373 traversal when removing references.
12375 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
12378 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
12380 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
12382 PR middle-end/56791
12383 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
12384 pushing a reload for an autoinc when we had previously reloaded an
12385 inner part of the address.
12387 2014-01-16 Jakub Jelinek <jakub@redhat.com>
12389 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
12391 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
12392 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
12393 when not giving up or versioning for alias only because of
12395 (vect_analyze_data_ref_dependences): Set to true.
12396 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
12398 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
12399 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
12402 PR middle-end/58344
12403 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
12406 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
12407 operand 0 predicate for gathers, use a new pseudo as subtarget.
12409 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
12411 PR middle-end/59609
12412 * lra-constraints.c (process_alt_operands): Add printing debug info.
12413 Check absence of input/output reloads for matched operands too.
12415 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
12417 PR rtl-optimization/59835
12418 * ira.c (ira_init_register_move_cost): Increase cost for
12421 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
12423 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
12425 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
12428 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
12429 non-register objects. Use gen_(high/low)part more consistently.
12432 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
12435 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
12436 endian support, remove tests for WORDS_BIG_ENDIAN.
12437 (p8_mfvsrd_3_<mode>): Likewise.
12438 (reload_gpr_from_vsx<mode>): Likewise.
12439 (reload_gpr_from_vsxsf): Likewise.
12440 (p8_mfvsrd_4_disf): Likewise.
12442 2014-01-16 Richard Biener <rguenther@suse.de>
12444 PR rtl-optimization/46590
12445 * lcm.c (compute_antinout_edge): Use postorder iteration.
12446 (compute_laterin): Use inverted postorder iteration.
12448 2014-01-16 Nick Clifton <nickc@redhat.com>
12450 PR middle-end/28865
12451 * varasm.c (output_constant): Return the number of bytes actually
12453 (output_constructor_array_range): Update the field size with the
12454 number of bytes emitted by output_constant.
12455 (output_constructor_regular_field): Likewise. Also do not
12456 complain if the total number of bytes emitted is now greater
12457 than the expected fieldpos.
12458 * output.h (output_constant): Update prototype and descriptive comment.
12460 2014-01-16 Marek Polacek <polacek@redhat.com>
12462 PR middle-end/59827
12463 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
12464 it is error_mark_node.
12466 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
12468 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
12469 VALID_AVX256_REG_OR_OI_MODE.
12471 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
12473 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
12474 current procedure should be profiled.
12476 2014-01-15 Andrew Pinski <apinski@cavium.com>
12478 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
12479 of moving from/to the STACK_REG register class.
12481 2014-01-15 Richard Henderson <rth@redhat.com>
12484 * reginfo.c (global_regs_decl): Globalize.
12485 * rtl.h (global_regs_decl): Declare.
12486 * ira.c (do_reload): Diagnose frame_pointer_needed and it
12487 reserved via global_regs.
12489 2014-01-15 Teresa Johnson <tejohnson@google.com>
12491 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
12493 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12495 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
12496 and vmulosh rather than call gen_vec_widen_smult_*.
12497 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
12498 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
12499 (vec_widen_smult_even_v16qi): Likewise.
12500 (vec_widen_umult_even_v8hi): Likewise.
12501 (vec_widen_smult_even_v8hi): Likewise.
12502 (vec_widen_umult_odd_v16qi): Likewise.
12503 (vec_widen_smult_odd_v16qi): Likewise.
12504 (vec_widen_umult_odd_v8hi): Likewise.
12505 (vec_widen_smult_odd_v8hi): Likewise.
12506 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
12507 vmuloub rather than call gen_vec_widen_umult_*.
12508 (vec_widen_umult_lo_v16qi): Likewise.
12509 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
12510 vmulosb rather than call gen_vec_widen_smult_*.
12511 (vec_widen_smult_lo_v16qi): Likewise.
12512 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
12513 rather than call gen_vec_widen_umult_*.
12514 (vec_widen_umult_lo_v8hi): Likewise.
12515 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
12516 rather than call gen_vec_widen_smult_*.
12517 (vec_widen_smult_lo_v8hi): Likewise.
12519 2014-01-15 Jeff Law <law@redhat.com>
12521 PR tree-optimization/59747
12522 * ree.c (find_and_remove_re): Properly handle case where a second
12523 eliminated extension requires widening a copy created for elimination
12524 of a prior extension.
12525 (combine_set_extension): Ensure that the number of hard regs needed
12526 for a destination register does not change when we widen it.
12528 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
12530 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
12531 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
12532 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
12533 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
12534 (avr-*-rtems*): Likewise.
12535 (bfin*-rtems*): Likewise.
12536 (moxie-*-rtems*): Likewise.
12537 (h8300-*-rtems*): Likewise.
12538 (i[34567]86-*-rtems*): Likewise.
12539 (lm32-*-rtems*): Likewise.
12540 (m32r-*-rtems*): Likewise.
12541 (m68k-*-rtems*): Likewise.
12542 (microblaze*-*-rtems*): Likewise.
12543 (mips*-*-rtems*): Likewise.
12544 (powerpc-*-rtems*): Likewise.
12545 (sh-*-rtems*): Likewise.
12546 (sparc-*-rtems*): Likewise.
12547 (sparc64-*-rtems*): Likewise.
12548 (v850-*-rtems*): Likewise.
12549 (m32c-*-rtems*): Likewise.
12551 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
12553 PR rtl-optimization/59511
12554 * ira.c (ira_init_register_move_cost): Use memory costs for some
12555 cases of register move cost calculations.
12556 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
12557 instead of BB frequency.
12558 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
12559 * lra-assigns.c (find_hard_regno_for): Ditto.
12561 2014-01-15 Richard Biener <rguenther@suse.de>
12563 PR tree-optimization/59822
12564 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
12565 (vectorizable_load): Use it to hoist defs of uses of invariant
12566 loads out of the loop.
12568 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
12569 Kugan Vivekanandarajah <kuganv@linaro.org>
12572 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
12575 2014-01-15 Richard Biener <rguenther@suse.de>
12577 PR rtl-optimization/59802
12578 * lcm.c (compute_available): Use inverted postorder to seed
12579 the initial worklist.
12581 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12584 * config/s390/s390.c (s390_preferred_reload_class): Don't return
12585 ADDR_REGS for invalid symrefs in non-PIC code.
12587 2014-01-15 Jakub Jelinek <jakub@redhat.com>
12590 * builtins.c (determine_block_size): Initialize *probable_max_size
12591 even if len_rtx is CONST_INT.
12593 2014-01-14 Andrew Pinski <apinski@cavium.com>
12595 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
12596 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
12597 (cortexa53_tunings): Likewise.
12598 (aarch64_sched_issue_rate): New function.
12599 (TARGET_SCHED_ISSUE_RATE): Define.
12601 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
12603 * ira-costs.c (find_costs_and_classes): Add missed
12604 ira_init_register_move_cost_if_necessary.
12606 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
12609 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
12611 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
12614 * config/i386/i386.c (type_natural_mode): Add a bool parameter
12615 to indicate if type is used for function return value. Warn ABI
12616 change if the vector mode isn't available for function return value.
12617 (ix86_function_arg_advance): Pass false to type_natural_mode.
12618 (ix86_function_arg): Likewise.
12619 (ix86_gimplify_va_arg): Likewise.
12620 (function_arg_32): Don't warn ABI change.
12621 (ix86_function_value): Pass true to type_natural_mode.
12622 (ix86_return_in_memory): Likewise.
12623 (ix86_struct_value_rtx): Removed.
12624 (TARGET_STRUCT_VALUE_RTX): Likewise.
12626 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
12628 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
12629 converting a conditional jump into a conditional return.
12631 2014-01-14 Richard Biener <rguenther@suse.de>
12633 PR tree-optimization/58921
12634 PR tree-optimization/59006
12635 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
12636 hoisting invariant stmts.
12637 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
12638 invariant loads on the preheader edge if possible.
12640 2014-01-14 Joey Ye <joey.ye@arm.com>
12642 * doc/plugin.texi (Building GCC plugins): Update to C++.
12644 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
12646 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
12647 (_mm_rcp28_round_ss): Ditto.
12648 (_mm_rsqrt28_round_sd): Ditto.
12649 (_mm_rsqrt28_round_ss): Ditto.
12650 (_mm_rcp28_sd): Ditto.
12651 (_mm_rcp28_ss): Ditto.
12652 (_mm_rsqrt28_sd): Ditto.
12653 (_mm_rsqrt28_ss): Ditto.
12654 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
12655 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
12656 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
12657 (IX86_BUILTIN_RCP28SD): Ditto.
12658 (IX86_BUILTIN_RCP28SS): Ditto.
12659 (IX86_BUILTIN_RSQRT28SD): Ditto.
12660 (IX86_BUILTIN_RSQRT28SS): Ditto.
12661 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
12662 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
12663 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
12664 (ix86_expand_special_args_builtin): Expand new FTYPE.
12665 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
12666 (srcp14<mode>): Make insn unary.
12667 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
12668 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
12669 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
12670 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
12671 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
12672 Fix rounding: make it SAE only.
12673 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
12675 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
12677 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
12678 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
12679 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
12680 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
12681 (round_saeonly_mask_scalar_operand4): Ditto.
12682 (round_saeonly_mask_scalar_op3): Ditto.
12683 (round_saeonly_mask_scalar_op4): Ditto.
12685 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12687 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
12688 Implement -maltivec=be for vec_insert and vec_extract.
12690 2014-01-10 DJ Delorie <dj@redhat.com>
12692 * config/msp430/msp430.md (call_internal): Don't allow memory
12693 references with SP as the base register.
12694 (call_value_internal): Likewise.
12695 * config/msp430/constraints.md (Yc): New. For memory references
12696 that don't use SP as a base register.
12698 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
12699 "an integer without a # prefix"
12700 * config/msp430/msp430.md (epilogue_helper): Use it.
12702 2014-01-13 Jakub Jelinek <jakub@redhat.com>
12705 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
12706 AVX512F gather builtins.
12707 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
12708 on gather decls with INTEGER_TYPE masktype.
12709 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
12710 directly into the builtin rather than hoisting it before loop.
12712 PR tree-optimization/59387
12713 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
12714 (scev_const_prop): If folded_casts and type has undefined overflow,
12715 use force_gimple_operand instead of force_gimple_operand_gsi and
12716 for each added stmt if it is assign with
12717 arith_code_with_undefined_signed_overflow, call
12718 rewrite_to_defined_overflow.
12719 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
12720 gimple-fold.h instead.
12721 (arith_code_with_undefined_signed_overflow,
12722 rewrite_to_defined_overflow): Moved to ...
12723 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
12724 rewrite_to_defined_overflow): ... here. No longer static.
12725 Include gimplify-me.h.
12726 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
12727 rewrite_to_defined_overflow): New prototypes.
12729 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12731 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
12733 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
12735 * builtins.c (get_object_alignment_2): Minor tweak.
12736 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
12738 2014-01-13 Christian Bruel <christian.bruel@st.com>
12740 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
12741 optimized non constant lengths.
12743 2014-01-13 Jakub Jelinek <jakub@redhat.com>
12746 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
12747 load as __atomic_load_N if possible.
12749 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
12751 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
12753 (rs6000_expand_builtin): Adjust call.
12755 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
12758 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
12759 * config/rs6000/rs6000.c: Include target-globals.h.
12760 (rs6000_set_current_function): Instead of doing target_reinit
12761 unconditionally, use save_target_globals_default_opts and
12762 restore_target_globals.
12764 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
12766 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
12767 (rs6000_expand_builtin): Handle mffs and mtfsf.
12768 (rs6000_init_builtins): Define mffs and mtfsf.
12769 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
12770 (rs6000_mffs): New pattern.
12771 (rs6000_mtfsf): New pattern.
12773 2014-01-11 Bin Cheng <bin.cheng@arm.com>
12775 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
12776 Start narrowing with START. Apply candidate-use pair
12777 and check overall cost in narrowing.
12778 (iv_ca_prune): Pass new argument.
12780 2014-01-10 Jeff Law <law@redhat.com>
12782 PR middle-end/59743
12783 * ree.c (combine_reaching_defs): Ensure the defining statement
12784 occurs before the extension when optimizing extensions with
12785 different source and destination hard registers.
12787 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
12790 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
12791 vtables into the type inheritance graph.
12793 2014-01-10 Jakub Jelinek <jakub@redhat.com>
12795 PR rtl-optimization/59754
12796 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
12797 modes in the REGNO != REGNO case.
12799 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12801 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
12803 2014-01-10 Jakub Jelinek <jakub@redhat.com>
12805 PR tree-optimization/59745
12806 * tree-predcom.c (tree_predictive_commoning_loop): Call
12807 free_affine_expand_cache if giving up because components is NULL.
12809 * target-globals.c (save_target_globals): Allocate < 4KB structs using
12810 GC in payload of target_globals struct instead of allocating them on
12811 the heap and the larger structs separately using GC.
12812 * target-globals.h (struct target_globals): Make regs, hard_regs,
12813 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
12814 of GTY((skip)) and change type to void *.
12815 (reset_target_globals): Cast loads from those fields to corresponding
12818 2014-01-10 Steve Ellcey <sellcey@mips.com>
12821 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
12822 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
12823 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
12825 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
12828 * aarch64-modes.def (CC_Zmode): New flags mode.
12829 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
12830 represents an equality.
12831 (aarch64_get_condition_code): Handle CC_Zmode.
12832 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
12834 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12836 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
12837 extraction in good case.
12839 2014-01-10 Richard Biener <rguenther@suse.de>
12841 PR tree-optimization/59374
12842 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
12843 checking after SLP discovery. Mark stmts not participating
12844 in any SLP instance properly.
12846 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12848 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
12849 when handling a SET rtx.
12851 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12853 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
12854 (cortex-a57): Likewise.
12855 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
12857 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12859 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
12860 non-iwmmxt builtins.
12862 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
12866 * ipa-devirt.c record_target_from_binfo): Take as argument
12867 stack of binfos and lookup matching one for virtual inheritance.
12868 (possible_polymorphic_call_targets_1): Update.
12870 2014-01-10 Huacai Chen <chenhc@lemote.com>
12872 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
12873 kernel strings for Loongson-2E/2F/3A.
12875 2014-01-10 Jakub Jelinek <jakub@redhat.com>
12877 PR middle-end/59670
12878 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
12879 is_gimple_call before calling gimple_call_internal_p.
12881 2014-01-09 Steve Ellcey <sellcey@mips.com>
12883 * Makefile.in (TREE_FLOW_H): Remove.
12884 (TREE_SSA_H): Add file names from tree-flow.h.
12885 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
12886 * tree.h: Remove tree-flow.h reference.
12887 * hash-table.h: Remove tree-flow.h reference.
12888 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
12889 reference with tree-ssa-loop.h.
12891 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12893 * doc/invoke.texi: Add -maltivec={be,le} options, and document
12894 default element-order behavior for -maltivec.
12895 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
12896 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
12897 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
12898 when targeting big endian, at least for now.
12899 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
12901 2014-01-09 Jakub Jelinek <jakub@redhat.com>
12903 PR middle-end/47735
12904 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
12905 var satisfies use_register_for_decl, just take into account type
12906 alignment, rather than decl alignment.
12908 PR tree-optimization/59622
12909 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
12910 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
12911 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
12912 Don't devirtualize for inplace at all. For targets.length () == 1,
12913 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
12915 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
12917 * config/i386/i386.md (cpu): Remove the unused btver1.
12919 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
12921 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
12923 2014-01-09 Jakub Jelinek <jakub@redhat.com>
12926 * tree-core.h (struct target_globals): New forward declaration.
12927 (struct tree_target_option): Add globals field.
12928 * tree.h (TREE_TARGET_GLOBALS): Define.
12929 (prepare_target_option_nodes_for_pch): New prototype.
12930 * target-globals.h (struct target_globals): Define even if
12931 !SWITCHABLE_TARGET.
12932 * tree.c (prepare_target_option_node_for_pch,
12933 prepare_target_option_nodes_for_pch): New functions.
12934 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
12935 * config/i386/i386.c: Include target-globals.h.
12936 (ix86_set_current_function): Instead of doing target_reinit
12937 unconditionally, use save_target_globals_default_opts and
12938 restore_target_globals.
12940 2014-01-09 Richard Biener <rguenther@suse.de>
12942 PR tree-optimization/59715
12943 * tree-cfg.h (split_critical_edges): Declare.
12944 * tree-cfg.c (split_critical_edges): Export.
12945 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
12947 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
12949 * cfgexpand.c (expand_stack_vars): Optionally disable
12950 asan stack protection.
12951 (expand_used_vars): Likewise.
12952 (partition_stack_vars): Likewise.
12953 * asan.c (asan_emit_stack_protection): Optionally disable
12954 after return stack usage.
12955 (instrument_derefs): Optionally disable memory access instrumentation.
12956 (instrument_builtin_call): Likewise.
12957 (instrument_strlen_call): Likewise.
12958 (asan_protect_global): Optionally disable global variables protection.
12959 * doc/invoke.texi: Added doc for new options.
12960 * params.def: Added new options.
12961 * params.h: Likewise.
12963 2014-01-09 Jakub Jelinek <jakub@redhat.com>
12965 PR rtl-optimization/59724
12966 * ifcvt.c (cond_exec_process_if_block): Don't call
12967 flow_find_head_matching_sequence with 0 longest_match.
12968 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
12969 non-active insns if !stop_after.
12970 (try_head_merge_bb): Revert 2014-01-07 changes.
12972 2014-01-08 Jeff Law <law@redhat.com>
12974 * ree.c (get_sub_rtx): New function, extracted from...
12975 (merge_def_and_ext): Here.
12976 (combine_reaching_defs): Use get_sub_rtx.
12978 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
12980 * cgraph.h (varpool_variable_node): Do not choke on null node.
12982 2014-01-08 Catherine Moore <clm@codesourcery.com>
12984 * config/mips/mips.md (simple_return): Attempt to use JRC
12986 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
12988 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
12990 PR rtl-optimization/59137
12991 * reorg.c (steal_delay_list_from_target): Call update_block for
12993 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
12995 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12997 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
12998 two duplicate entries.
13000 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
13003 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
13005 * config/mips/mips.c (mips_truncated_op_cost): New function.
13006 (mips_rtx_costs): Adjust test for BADDU.
13007 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
13009 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
13011 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
13012 (*baddu_si): ...this new pattern.
13014 2014-01-08 Jakub Jelinek <jakub@redhat.com>
13017 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
13019 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
13021 PR middle-end/57748
13022 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
13024 (expand_expr, expand_normal): Adjust.
13025 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
13026 inner_reference_p. Use inner_reference_p to expand inner references.
13027 (store_expr): Adjust.
13028 * cfgexpand.c (expand_call_stmt): Adjust.
13030 2014-01-08 Rong Xu <xur@google.com>
13032 * gcov-io.c (gcov_var): Move from gcov-io.h.
13033 (gcov_position): Ditto.
13034 (gcov_is_error): Ditto.
13035 (gcov_rewrite): Ditto.
13036 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
13037 only part to libgcc/libgcov.h.
13039 2014-01-08 Marek Polacek <polacek@redhat.com>
13041 PR middle-end/59669
13042 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
13044 2014-01-08 Marek Polacek <polacek@redhat.com>
13047 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
13049 2014-01-08 Jakub Jelinek <jakub@redhat.com>
13051 PR rtl-optimization/59649
13052 * stor-layout.c (get_mode_bounds): For BImode return
13053 0 and STORE_FLAG_VALUE.
13055 2014-01-08 Richard Biener <rguenther@suse.de>
13057 PR middle-end/59630
13058 * gimple.h (is_gimple_builtin_call): Remove.
13059 (gimple_builtin_call_types_compatible_p): New.
13060 (gimple_call_builtin_p): New overload.
13061 * gimple.c (is_gimple_builtin_call): Remove.
13062 (validate_call): Rename to ...
13063 (gimple_builtin_call_types_compatible_p): ... this and export. Also
13064 check return types.
13065 (validate_type): New static function.
13066 (gimple_call_builtin_p): New overload and adjust.
13067 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
13068 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
13069 (gimple_fold_stmt_to_constant_1): Likewise.
13070 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
13072 2014-01-08 Richard Biener <rguenther@suse.de>
13074 PR middle-end/59471
13075 * gimplify.c (gimplify_expr): Gimplify register-register type
13076 VIEW_CONVERT_EXPRs to separate stmts.
13078 2014-01-07 Jeff Law <law@redhat.com>
13080 PR middle-end/53623
13081 * ree.c (combine_set_extension): Handle case where source
13082 and destination registers in an extension insn are different.
13083 (combine_reaching_defs): Allow source and destination registers
13084 in extension to be different under limited circumstances.
13085 (add_removable_extension): Remove restriction that the
13086 source and destination registers in the extension are the same.
13087 (find_and_remove_re): Emit a copy from the extension's
13088 destination to its source after the defining insn if
13089 the source and destination registers are different.
13091 PR middle-end/59285
13092 * ifcvt.c (merge_if_block): If we are merging a block with more than
13093 one successor with a block with no successors, remove any BARRIER
13094 after the second block.
13096 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
13098 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
13100 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
13103 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
13104 for 14-bit register offsets when INT14_OK_STRICT is false.
13106 2014-01-07 Roland Stigge <stigge@antcom.de>
13107 Michael Meissner <meissner@linux.vnet.ibm.com>
13110 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
13111 Only check TFmode for SPE constants. Don't check TImode or TDmode.
13113 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
13115 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
13118 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
13120 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
13121 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
13122 rtx is const0_rtx or not.
13124 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
13127 * target-globals.c (save_target_globals): Remove this_fn_optab
13129 * toplev.c: Include optabs.h.
13130 (target_reinit): Temporarily restore the global options if another
13131 set of options are in force.
13133 2014-01-07 Jakub Jelinek <jakub@redhat.com>
13135 PR rtl-optimization/58668
13136 * cfgcleanup.c (flow_find_cross_jump): Don't count
13137 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
13138 to determine what is counted.
13139 (flow_find_head_matching_sequence): Use active_insn_p to determine
13141 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
13143 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
13144 determine what is counted.
13146 PR tree-optimization/59643
13147 * tree-predcom.c (split_data_refs_to_components): If one dr is
13148 read and one write, determine_offset fails and the write isn't
13149 in the bad component, just put the read into the bad component.
13151 2014-01-07 Mike Stump <mikestump@comcast.net>
13152 Jakub Jelinek <jakub@redhat.com>
13155 * tree-core.h (struct tree_optimization_option): Change optabs
13156 type from unsigned char * to void *.
13157 * optabs.c (init_tree_optimization_optabs): Adjust
13158 TREE_OPTIMIZATION_OPTABS initialization.
13160 2014-01-06 Jakub Jelinek <jakub@redhat.com>
13163 * config/i386/i386.h (struct machine_function): Add
13164 no_drap_save_restore field.
13165 * config/i386/i386.c (ix86_save_reg): Use
13166 !cfun->machine->no_drap_save_restore instead of
13167 crtl->stack_realign_needed.
13168 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
13169 this function clears frame_pointer_needed. Set
13170 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
13171 and DRAP reg is needed.
13173 2014-01-06 Marek Polacek <polacek@redhat.com>
13176 * doc/implement-c.texi: Mention that other integer types are
13177 permitted as bit-field types in strictly conforming mode.
13179 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
13181 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
13182 is newly allocated.
13184 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
13186 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
13188 2014-01-06 Martin Jambor <mjambor@suse.cz>
13191 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
13193 * ipa-prop.c (ipa_print_node_params): Fix indentation.
13195 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
13199 * var-tracking.c (add_stores): Preserve the value of the source even if
13200 we don't record the store.
13202 2014-01-06 Terry Guo <terry.guo@arm.com>
13204 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
13206 2014-01-05 Iain Sandoe <iain@codesourcery.com>
13209 * config/darwin.c (darwin_function_section): Adjust return values to
13210 correspond to optimisation changes made in r206070.
13212 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
13214 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
13215 from prefetch_block tune setting.
13216 (nocona_cost): Correct size of prefetch block to 64.
13218 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
13220 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
13221 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
13222 used to save the static chain register in the computation of the offset
13223 from which the FP registers need to be restored.
13225 2014-01-04 Jakub Jelinek <jakub@redhat.com>
13227 PR tree-optimization/59519
13228 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
13229 ICE if get_current_def (current_new_name) is already non-NULL, as long
13230 as it is a phi result of some other phi in *new_exit_bb that has
13233 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
13234 or vmovdqu* for misaligned_operand.
13235 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
13236 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
13237 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
13238 aligned_mem for AVX512F masked aligned load and store builtins and for
13239 non-temporal moves.
13241 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
13243 PR tree-optimization/59651
13244 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
13245 Address range for negative step should be added by TYPE_SIZE_UNIT.
13247 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
13249 * config/m68k/m68k.c (handle_move_double): Handle pushes with
13250 overlapping registers also for registers other than the stack pointer.
13252 2014-01-03 Marek Polacek <polacek@redhat.com>
13255 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
13258 2014-01-03 Jakub Jelinek <jakub@redhat.com>
13261 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
13264 * config/i386/i386.md (MODE_SIZE): New mode attribute.
13265 (push splitter): Use <P:MODE_SIZE> instead of
13266 GET_MODE_SIZE (<P:MODE>mode).
13267 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
13268 (mov -1, reg peephole2): Likewise.
13269 * config/i386/sse.md (*mov<mode>_internal,
13270 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
13271 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
13272 *<code><mode>3, *andnot<mode>3<mask_name>,
13273 <mask_codefor><code><mode>3<mask_name>): Likewise.
13274 * config/i386/subst.md (mask_mode512bit_condition,
13275 sd_mask_mode512bit_condition): Likewise.
13277 2014-01-02 Xinliang David Li <davidxl@google.com>
13279 PR tree-optimization/59303
13280 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
13281 (dump_predicates): Better output format.
13282 (pred_equal_p): New function.
13283 (is_neq_relop_p): Ditto.
13284 (is_neq_zero_form_p): Ditto.
13285 (pred_expr_equal_p): Ditto.
13286 (pred_neg_p): Ditto.
13287 (simplify_pred): Ditto.
13288 (simplify_preds_2): Ditto.
13289 (simplify_preds_3): Ditto.
13290 (simplify_preds_4): Ditto.
13291 (simplify_preds): Ditto.
13292 (push_pred): Ditto.
13293 (push_to_worklist): Ditto.
13294 (get_pred_info_from_cmp): Ditto.
13295 (is_degenerated_phi): Ditto.
13296 (normalize_one_pred_1): Ditto.
13297 (normalize_one_pred): Ditto.
13298 (normalize_one_pred_chain): Ditto.
13299 (normalize_preds): Ditto.
13300 (normalize_cond_1): Remove function.
13301 (normalize_cond): Ditto.
13302 (is_gcond_subset_of): Ditto.
13303 (is_subset_of_any): Ditto.
13304 (is_or_set_subset_of): Ditto.
13305 (is_and_set_subset_of): Ditto.
13306 (is_norm_cond_subset_of): Ditto.
13307 (pred_chain_length_cmp): Ditto.
13308 (convert_control_dep_chain_into_preds): Type change.
13309 (find_predicates): Ditto.
13310 (find_def_preds): Ditto.
13311 (destroy_predicates_vecs): Ditto.
13312 (find_matching_predicates_in_rest_chains): Ditto.
13313 (use_pred_not_overlap_with_undef_path_pred): Ditto.
13314 (is_pred_expr_subset): Ditto.
13315 (is_pred_chain_subset_of): Ditto.
13316 (is_included_in): Ditto.
13317 (is_superset_of): Ditto.
13319 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
13321 Update copyright years.
13323 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
13325 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
13326 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
13327 config/arc/arc.md, config/arc/arc.opt,
13328 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
13329 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
13330 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
13331 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
13332 config/linux-protos.h, config/linux.c, config/winnt-c.c,
13333 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
13334 vtable-verify.c, vtable-verify.h: Use the standard form for the
13337 2014-01-02 Tobias Burnus <burnus@net-b.de>
13339 * gcc.c (process_command): Update copyright notice dates.
13340 * gcov-dump.c: Ditto.
13342 * doc/cpp.texi: Bump @copying's copyright year.
13343 * doc/cppinternals.texi: Ditto.
13344 * doc/gcc.texi: Ditto.
13345 * doc/gccint.texi: Ditto.
13346 * doc/gcov.texi: Ditto.
13347 * doc/install.texi: Ditto.
13348 * doc/invoke.texi: Ditto.
13350 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
13352 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
13354 2014-01-01 Jakub Jelinek <jakub@redhat.com>
13356 * config/i386/sse.md (*mov<mode>_internal): Guard
13357 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
13359 PR rtl-optimization/59647
13360 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
13361 new_rtx into UNSIGNED_FLOAT rtxes.
13363 Copyright (C) 2014 Free Software Foundation, Inc.
13365 Copying and distribution of this file, with or without modification,
13366 are permitted in any medium without royalty provided the copyright
13367 notice and this notice are preserved.