1 2014-06-04 Martin Liska <mliska@suse.cz>
3 * coverage.h (coverage_compute_cfg_checksum): Argument added.
4 * coverage.c (coverage_compute_cfg_checksum): Likewise.
5 * profile.c (branch_prob): Likewise.
7 2014-06-04 Martin Jambor <mjambor@suse.cz>
10 * ipa-pure-const.c (propagate_pure_const): Add unreachable default
11 handler for switch on an ipa_ref_use enum.
12 * ipa-reference.c (analyze_function): Likewise.
14 2014-06-04 Kai Tietz <ktietz@redhat.com>
16 * recog.c (peep2_attempt): Copy SIBLING_CALL_P flag
17 from old call-instruction.
19 2014-06-04 Bin Cheng <bin.cheng@arm.com>
21 * config/aarch64/aarch64.c (aarch64_classify_address)
22 (aarch64_legitimize_reload_address): Support full addressing modes
24 * config/aarch64/aarch64.md (mov<mode>, movmisalign<mode>)
25 (*aarch64_simd_mov<mode>, *aarch64_simd_mov<mode>): Relax predicates.
27 2014-06-03 Andrew Pinski <apinski@cavium.com>
29 * config/aarch64/aarch64.c (aarch64_if_then_else_costs): Allow non comparisons
32 2014-06-03 Andrew Pinski <apinski@cavium.com>
34 * config/aarch64/aarch64.c (aarch64_if_then_else_costs): New function.
35 (aarch64_rtx_costs): Use aarch64_if_then_else_costs.
37 2014-06-03 Kai Tietz <ktietz@redhat.com>
39 * config/i386/i386.c (ix86_function_value_regno_p): Disallow DX_REG
42 2014-06-03 Dehao Chen <dehao@google.com>
44 * tree-cfg.c (gimple_merge_blocks): Only reset count when BBs are in
47 2014-06-03 Marek Polacek <polacek@redhat.com>
50 * doc/invoke.texi: Document -Wswitch-bool.
51 * function.c (stack_protect_epilogue): Cast controlling expression of
53 * gengtype.c (walk_type): Generate switch expression with its
54 controlling expression cast to int.
56 2014-06-03 Vishnu K S <Vishnu.k_s@atmel.com>
58 * config/avr/avr-mcus.def: Add new avr25 devices attiny441, attiny828
60 * config/avr/avr-tables.opt: Regenerate.
61 * config/avr/t-multilib: Regenerate.
62 * doc/avr-mmcu.texi: Regenerate.
64 2014-06-03 Vishnu K S <vishnu.k_s@atmel.com>
65 Pitchumani Sivanupandi <pitchumani.s@atmel.com>
67 * config/avr/avr-mcus.def (ata6616c): Add new avr25 device.
68 (ata6617c, ata664251): Add new avr35 devices.
69 (ata6612c): Add new avr4 device.
70 (ata6613c, ata6614q): Add new avr5 devices.
71 * config/avr/avr-tables.opt: Regenerate.
72 * config/avr/t-multilib: Regenerate.
73 * doc/avr-mmcu.texi: Regenerate.
75 2014-06-03 Alan Lawrence <alan.lawrence@arm.com>
77 * gcc/config/aarch64/aarch64-builtins.c
78 (aarch64_types_binop_ssu_qualifiers): New static data.
79 (TYPES_BINOP_SSU): Define.
80 * gcc/config/aarch64/aarch64-simd-builtins.def (suqadd, ushl, urshl,
81 urshr_n, ushll_n): Use appropriate unsigned qualifiers.
82 * gcc/config/aarch64/arm_neon.h (vrshl_u8, vrshl_u16, vrshl_u32,
83 vrshl_u64, vrshlq_u8, vrshlq_u16, vrshlq_u32, vrshlq_u64, vrshld_u64,
84 vrshr_n_u8, vrshr_n_u16, vrshr_n_u32, vrshr_n_u64, vrshrq_n_u8,
85 vrshrq_n_u16, vrshrq_n_u32, vrshrq_n_u64, vrshrd_n_u64, vshll_n_u8,
86 vshll_n_u16, vshll_n_u32, vuqadd_s8, vuqadd_s16, vuqadd_s32,
87 vuqadd_s64, vuqaddq_s8, vuqaddq_s16, vuqaddq_s32, vuqaddq_s64,
88 vuqaddb_s8, vuqaddh_s16, vuqadds_s32, vuqaddd_s64): Add signedness
89 suffix to builtin function name, remove cast.
90 (vshl_s8, vshl_s16, vshl_s32, vshl_s64, vshl_u8, vshl_u16, vshl_u32,
91 vshl_u64, vshlq_s8, vshlq_s16, vshlq_s32, vshlq_s64, vshlq_u8,
92 vshlq_u16, vshlq_u32, vshlq_u64, vshld_s64, vshld_u64): Remove cast.
94 2014-06-03 Alan Lawrence <alan.lawrence@arm.com>
96 * gcc/config/aarch64/aarch64-builtins.c
97 (aarch64_types_binop_uus_qualifiers,
98 aarch64_types_shift_to_unsigned_qualifiers,
99 aarch64_types_unsigned_shiftacc_qualifiers): Define.
100 * gcc/config/aarch64/aarch64-simd-builtins.def (uqshl, uqrshl, uqadd,
101 uqsub, usqadd, usra_n, ursra_n, uqshrn_n, uqrshrn_n, usri_n, usli_n,
102 sqshlu_n, uqshl_n): Update qualifiers.
103 * gcc/config/aarch64/arm_neon.h (vqadd_u8, vqadd_u16, vqadd_u32,
104 vqadd_u64, vqaddq_u8, vqaddq_u16, vqaddq_u32, vqaddq_u64, vqsub_u8,
105 vqsub_u16, vqsub_u32, vqsub_u64, vqsubq_u8, vqsubq_u16, vqsubq_u32,
106 vqsubq_u64, vqaddb_u8, vqaddh_u16, vqadds_u32, vqaddd_u64, vqrshl_u8,
107 vqrshl_u16, vqrshl_u32, vqrshl_u64, vqrshlq_u8, vqrshlq_u16,
108 vqrshlq_u32, vqrshlq_u64, vqrshlb_u8, vqrshlh_u16, vqrshls_u32,
109 vqrshld_u64, vqrshrn_n_u16, vqrshrn_n_u32, vqrshrn_n_u64,
110 vqrshrnh_n_u16, vqrshrns_n_u32, vqrshrnd_n_u64, vqshl_u8, vqshl_u16,
111 vqshl_u32, vqshl_u64, vqshlq_u8, vqshlq_u16, vqshlq_u32, vqshlq_u64,
112 vqshlb_u8, vqshlh_u16, vqshls_u32, vqshld_u64, vqshl_n_u8, vqshl_n_u16,
113 vqshl_n_u32, vqshl_n_u64, vqshlq_n_u8, vqshlq_n_u16, vqshlq_n_u32,
114 vqshlq_n_u64, vqshlb_n_u8, vqshlh_n_u16, vqshls_n_u32, vqshld_n_u64,
115 vqshlu_n_s8, vqshlu_n_s16, vqshlu_n_s32, vqshlu_n_s64, vqshluq_n_s8,
116 vqshluq_n_s16, vqshluq_n_s32, vqshluq_n_s64, vqshlub_n_s8,
117 vqshluh_n_s16, vqshlus_n_s32, vqshlud_n_s64, vqshrn_n_u16,
118 vqshrn_n_u32, vqshrn_n_u64, vqshrnh_n_u16, vqshrns_n_u32,
119 vqshrnd_n_u64, vqsubb_u8, vqsubh_u16, vqsubs_u32, vqsubd_u64,
120 vrsra_n_u8, vrsra_n_u16, vrsra_n_u32, vrsra_n_u64, vrsraq_n_u8,
121 vrsraq_n_u16, vrsraq_n_u32, vrsraq_n_u64, vrsrad_n_u64, vsli_n_u8,
122 vsli_n_u16, vsli_n_u32,vsli_n_u64, vsliq_n_u8, vsliq_n_u16,
123 vsliq_n_u32, vsliq_n_u64, vslid_n_u64, vsqadd_u8, vsqadd_u16,
124 vsqadd_u32, vsqadd_u64, vsqaddq_u8, vsqaddq_u16, vsqaddq_u32,
125 vsqaddq_u64, vsqaddb_u8, vsqaddh_u16, vsqadds_u32, vsqaddd_u64,
126 vsra_n_u8, vsra_n_u16, vsra_n_u32, vsra_n_u64, vsraq_n_u8,
127 vsraq_n_u16, vsraq_n_u32, vsraq_n_u64, vsrad_n_u64, vsri_n_u8,
128 vsri_n_u16, vsri_n_u32, vsri_n_u64, vsriq_n_u8, vsriq_n_u16,
129 vsriq_n_u32, vsriq_n_u64, vsrid_n_u64): Remove casts.
131 2014-06-03 Teresa Johnson <tejohnson@google.com>
133 * tree-sra.c (modify_function): Record caller nodes after rebuild.
135 2014-06-02 Jason Merrill <jason@redhat.com>
138 * varpool.c (ctor_for_folding): Handle uninitialized vtables.
140 2014-06-03 Alan Lawrence <alan.lawrence@arm.com>
142 * config/aarch64/aarch64.c (aarch64_evpc_ext): allow and handle
145 2014-06-03 Alan Lawrence <alan.lawrence@arm.com>
147 * config/aarch64/aarch64-simd.md (aarch64_rev<REVERSE:rev-op><mode>):
149 * config/aarch64/aarch64.c (aarch64_evpc_rev): New function.
150 (aarch64_expand_vec_perm_const_1): Add call to aarch64_evpc_rev.
151 * config/aarch64/iterators.md (REVERSE): New iterator.
152 (UNSPEC_REV64, UNSPEC_REV32, UNSPEC_REV16): New enum elements.
153 (rev_op): New int_attribute.
154 * config/aarch64/arm_neon.h (vrev16_p8, vrev16_s8, vrev16_u8,
155 vrev16q_p8, vrev16q_s8, vrev16q_u8, vrev32_p8, vrev32_p16, vrev32_s8,
156 vrev32_s16, vrev32_u8, vrev32_u16, vrev32q_p8, vrev32q_p16, vrev32q_s8,
157 vrev32q_s16, vrev32q_u8, vrev32q_u16, vrev64_f32, vrev64_p8,
158 vrev64_p16, vrev64_s8, vrev64_s16, vrev64_s32, vrev64_u8, vrev64_u16,
159 vrev64_u32, vrev64q_f32, vrev64q_p8, vrev64q_p16, vrev64q_s8,
160 vrev64q_s16, vrev64q_s32, vrev64q_u8, vrev64q_u16, vrev64q_u32):
161 Replace temporary __asm__ with __builtin_shuffle.
163 2014-06-03 Andrew Bennett <andrew.bennett@imgtec.com>
165 * config/mips/mips-cpus.def: Add mips32r3, mips32r5, mips64r3 and
167 * config/mips/mips-tables.opt: Regenerate.
168 * config/mips/mips.c (mips_compute_frame_info): Changed if statement
169 to use mips_isa_rev rather than ISA_MIPS32R2.
170 * config/mips/mips.h (ISA_MIPS32R3): New define.
171 (ISA_MIPS32R5): New define.
172 (ISA_MIPS64R3): New define.
173 (ISA_MIPS64R5): New define.
174 (TARGET_CPU_CPP_BUILTINS): Added support for ISA_MIPS32R3, ISA_MIPS32R5,
175 ISA_MIPS64R3 and ISA_MIPS64R5.
176 (MIPS_ISA_LEVEL_SPEC): Added support for mips32r3, mips32r5, mips64r3
178 (MIPS_ISA_SYNCI_SPEC): Likewise.
179 (ISA_HAS_64BIT_REGS): Added ISA_MIPS64R3 and ISA_MIPS64R5.
180 (LINK_SPEC): Added mips32r3 and mips32r5.
181 * config/mips/t-isa3264 (MULTILIB_MATCHES): Map mips32r3 and mips32r5
182 to mips32r2; and mips64r3 and mips64r5 to mips64r2.
183 * config/mips/t-mti-elf (MULTILIB_MATCHES): Likewise.
184 * config/mips/t-mti-linux (MULTILIB_MATCHES): Likewise.
185 * config/mips/t-sde (MULTILIB_MATCHES): Likewise.
186 * config/mips/t-sdemtk (MULTILIB_MATCHES): New define.
187 * doc/invoke.texi: Document mips32r3, mips32r5, mips64r3 and mips64r5.
189 2014-06-03 Andrew Bennett <andrew.bennett@imgtec.com>
191 * doc/invoke.texi: Document -mxpa and -mno-xpa MIPS command line
193 * config/mips/mips.opt (mxpa): New option.
194 * config/mips/mips.h (ASM_SPEC): Pass mxpa and mno-xpa to the
197 2014-06-03 Martin Jambor <mjambor@suse.cz>
200 * ipa-cp.c (cgraph_edge_brings_value_p): Handle edges leading to
203 2014-06-03 Thomas Preud'homme <thomas.preudhomme@arm.com>
205 PR tree-optimization/61328
206 * tree-ssa-math-opts.c (init_symbolic_number): Extract symbolic number
207 initialization from find_bswap_or_nop_1.
208 (find_bswap_or_nop_1): Test return value of find_bswap_or_nop_1 stored
209 in source_expr2 before using the size value the function sets. Also
210 make use of init_symbolic_number () in both the old place and
211 find_bswap_or_nop_load () to avoid reading uninitialized memory when
212 doing recursion in the GIMPLE_BINARY_RHS case.
214 2014-06-03 Richard Biener <rguenther@suse.de>
216 PR tree-optimization/61383
217 * tree-ssa-ifcombine.c (bb_no_side_effects_p): Make sure
220 2014-06-03 Richard Sandiford <rdsandiford@googlemail.com>
222 * defaults.h (USE_MD_CONSTRAINTS, EXTRA_MEMORY_CONSTRAINT)
223 (EXTRA_ADDRESS_CONSTRAINT, DEFAULT_CONSTRAINT_LEN, CONSTRAINT_LEN)
224 (CONST_OK_FOR_CONSTRAINT_P, CONST_DOUBLE_OK_FOR_LETTER_P)
225 (REG_CLASS_FROM_CONSTRAINT, EXTRA_CONSTRAINT_STR): Delete definitions
227 (REG_CLASS_FROM_LETTER, CONST_OK_FOR_LETTER_P)
228 (CONST_DOUBLE_OK_FOR_LETTER_P, EXTRA_CONSTRAINT): Move poising to...
229 * system.h: ...here and make it unconditional.
230 * target.def (conditional_register_usage): Mention
231 define_register_constraint instead of old-style constraint macros.
232 * doc/tm.texi.in: Remove documentation for old-style constraint macros.
233 * doc/tm.texi: Regenerate.
234 * genoutput.c: Remove USE_MD_CONSTRAINTS conditions and all code
235 protected by !USE_MD_CONSTRAINTS.
236 * config/frv/frv.md: Remove quote from old version of documentation.
237 * config/frv/frv.c (frv_conditional_register_usage): Likewise.
238 * config/m32r/m32r.c (easy_di_const, easy_df_const): Avoid mentioning
239 CONST_DOUBLE_OK_FOR_LETTER.
240 * config/sh/constraints.md: Likewise EXTRA_CONSTRAINT.
242 2014-06-02 Andrew Pinski <apinski@cavium.com>
244 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER):
245 /lib/ld-linux32-aarch64.so.1 is used for ILP32.
246 (LINUX_TARGET_LINK_SPEC): Update linker script for ILP32.
247 file whose name depends on -mabi= and -mbig-endian.
248 * config/aarch64/t-aarch64-linux (MULTILIB_OSDIRNAMES):
249 Handle LP64 better and handle ilp32 too.
250 (MULTILIB_OPTIONS): Delete.
251 (MULTILIB_DIRNAMES): Delete.
253 2014-06-02 Andrew MacLeod <amacleod@redhat.com>
255 * expr.h: Remove prototypes of functions defined in builtins.c.
256 * tree.h: (build_call_expr_*, build_string_literal): Add prototypes.
257 Remove prototypes of functions defined in builtins.c.
258 * builtins.h: Update prototype list to include all exported functions.
259 * builtins.c: (default_libc_has_function, gnu_libc_has_function,
260 no_c99_libc_has_function): Move to targhooks.c
261 (build_string_literal, build_call_expr_loc_array,
262 build_call_expr_loc_vec, build_call_expr_loc, build_call_expr): Move
264 (expand_builtin_object_size, fold_builtin_object_size): Make static.
265 * targhooks.c (default_libc_has_function, gnu_libc_has_function,
266 no_c99_libc_has_function): Relocate from builtins.c.
267 * tree.c: Include builtins.h.
268 (build_call_expr_loc_array, build_call_expr_loc_vec,
269 build_call_expr_loc, build_call_expr, build_string_literal): Relocate
271 * fold-const.h (fold_fma): Move prototype to builtins.h.
272 * realmpfr.h (do_mpc_arg2): Move prototype to builtins.h.
273 * asan.c: Include builtins.h.
274 * cfgexpand.c: Likewise.
275 * convert.c: Likewise.
276 * emit-rtl.c: Likewise.
277 * except.c: Likewise.
279 * fold-const.c: Likewise.
280 * gimple-fold.c: Likewise.
281 * gimple-ssa-strength-reduction.c: Likewise.
282 * gimplify.c: Likewise.
283 * ipa-inline.c: Likewise.
284 * ipa-prop.c: Likewise.
285 * lto-streamer-out.c: Likewise.
287 * tree-inline.c: Likewise.
288 * tree-object-size.c: Likewise.
289 * tree-sra.c: Likewise.
290 * tree-ssa-ccp.c: Likewise.
291 * tree-ssa-forwprop.c: Likewise.
292 * tree-ssa-loop-ivcanon.c: Likewise.
293 * tree-ssa-loop-ivopts.c: Likewise.
294 * tree-ssa-math-opts.c: Likewise.
295 * tree-ssa-reassoc.c: Likewise.
296 * tree-ssa-threadedge.c: Likewise.
297 * tree-streamer-in.c: Likewise.
298 * tree-vect-data-refs.c: Likewise.
299 * tree-vect-patterns.c: Likewise.
300 * tree-vect-stmts.c: Likewise.
301 * config/aarch64/aarch64.c: Likewise.
302 * config/alpha/alpha.c: Likewise.
303 * config/arc/arc.c: Likewise.
304 * config/arm/arm.c: Likewise.
305 * config/avr/avr.c: Likewise.
306 * config/bfin/bfin.c: Likewise.
307 * config/c6x/c6x.c: Likewise.
308 * config/cr16/cr16.c: Likewise.
309 * config/cris/cris.c: Likewise.
310 * config/epiphany/epiphany.c: Likewise.
311 * config/fr30/fr30.c: Likewise.
312 * config/frv/frv.c: Likewise.
313 * config/h8300/h8300.c: Likewise.
314 * config/i386/i386.c: Likewise.
315 * config/i386/winnt.c: Likewise.
316 * config/ia64/ia64.c: Likewise.
317 * config/iq2000/iq2000.c: Likewise.
318 * config/lm32/lm32.c: Likewise.
319 * config/m32c/m32c.c: Likewise.
320 * config/m32r/m32r.c: Likewise.
321 * config/m68k/m68k.c: Likewise.
322 * config/mcore/mcore.c: Likewise.
323 * config/mep/mep.c: Likewise.
324 * config/microblaze/microblaze.c: Likewise.
325 * config/mips/mips.c: Likewise.
326 * config/mmix/mmix.c: Likewise.
327 * config/mn10300/mn10300.c: Likewise.
328 * config/moxie/moxie.c: Likewise.
329 * config/msp430/msp430.c: Likewise.
330 * config/nds32/nds32.c: Likewise.
331 * config/pa/pa.c: Likewise.
332 * config/pdp11/pdp11.c: Likewise.
333 * config/picochip/picochip.c: Likewise.
334 * config/rl78/rl78.c: Likewise.
335 * config/rs6000/rs6000.c: Likewise.
336 * config/rx/rx.c: Likewise.
337 * config/s390/s390.c: Likewise.
338 * config/score/score.c: Likewise.
339 * config/sh/sh.c: Likewise.
340 * config/sparc/sparc.c: Likewise.
341 * config/spu/spu.c: Likewise.
342 * config/stormy16/stormy16.c: Likewise.
343 * config/tilegx/tilegx.c: Likewise.
344 * config/tilepro/tilepro.c: Likewise.
345 * config/v850/v850.c: Likewise.
346 * config/vax/vax.c: Likewise.
347 * config/xtensa/xtensa.c: Likewise.
349 2014-06-02 Jeff Law <law@redhat.com>
351 PR rtl-optimization/61094
352 * ree.c (combine_reaching_defs): Do not reextend an insn if it
353 was marked as do_no_reextend. If a copy is needed to eliminate
354 an extension, then mark it as do_not_reextend.
356 2014-06-02 Marcus Shawcroft <marcus.shawcroft@arm.com>
358 * config/aarch64/aarch64.md (set_fpcr): Drop ISB after FPCR write.
360 2014-06-02 Richard Henderson <rth@redhat.com>
363 * config/alpha/alpha.c (print_operand_address): Allow symbolic
364 addresses inside asms. Use output_operand_lossage instead of
367 2014-06-02 Uros Bizjak <ubizjak@gmail.com>
370 * config/i386/i386.c (ix86_expand_vec_perm) [case V32QImode]: Use
371 GEN_INT (-128) instead of GEN_INT (128) to set MSB of QImode constant.
373 2014-06-02 Tom de Vries <tom@codesourcery.com>
375 * config/aarch64/aarch64.c (aarch64_float_const_representable_p): Handle
376 case that x has VOIDmode.
378 2014-06-02 Bernd Schmidt <bernds@codesourcery.com>
380 * varasm.c (copy_constant): Delete function.
381 (build_constant_desc): Don't call it.
383 2014-06-02 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
386 * config/arm/arm.h (TARGET_SUPPORTS_WIDE_INT): Define.
387 * config/arm/arm.md (mov64 splitter): Replace const_double_operand
388 with immediate_operand.
390 2014-06-02 Andreas Schwab <schwab@suse.de>
393 (ia64_first_cycle_multipass_dfa_lookahead_guard): Check
394 pending_data_specs first.
396 2014-06-02 Richard Biener <rguenther@suse.de>
398 PR tree-optimization/61378
399 * tree-ssa-sccvn.c (vn_reference_lookup_3): Initialize
402 2014-06-01 Uros Bizjak <ubizjak@gmail.com>
404 * config/i386/constraints.md (Bw): Rename from 'w'.
405 (Bz): Rename from 'z'.
406 * config/i386/i386.md: Change 'w' to 'Bw' and 'z' to 'Bz' globally.
408 2014-06-01 Kai Tietz <ktietz@redhat.com>
411 * config/i386/constrains.md (define_constrain): New 'Bs' constraint.
412 * config/i386/i386.md (sibcall_insn_operand): Use Bs
413 instead of m constraint.
415 2014-05-31 Andreas Schwab <schwab@linux-m68k.org>
417 * config/m68k/m68k.md (beq0_di, bne0_di): Make the "o" constraint
418 a separate alternative where the scratch operand 2 is marked as
421 2014-05-31 Kugan Vivekanandarajah <kuganv@linaro.org>
423 * config/arm/arm.c (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New define.
424 (arm_builtins) : Add ARM_BUILTIN_GET_FPSCR and ARM_BUILTIN_SET_FPSCR.
425 (bdesc_2arg) : Add description for builtins __builtins_arm_set_fpscr
426 and __builtins_arm_get_fpscr.
427 (arm_init_builtins) : Initialize builtins __builtins_arm_set_fpscr and
428 __builtins_arm_get_fpscr.
429 (arm_expand_builtin) : Expand builtins __builtins_arm_set_fpscr and
430 __builtins_arm_ldfpscr.
431 (arm_atomic_assign_expand_fenv): New function.
432 * config/arm/vfp.md (set_fpscr): New pattern.
433 (get_fpscr) : Likewise.
434 * config/arm/unspecs.md (unspecv): Add VUNSPEC_GET_FPSCR and
436 * doc/extend.texi (AARCH64 Built-in Functions) : Document
437 __builtins_arm_set_fpscr, __builtins_arm_get_fpscr.
439 2014-05-30 Jakub Jelinek <jakub@redhat.com>
441 * asan.c (report_error_func): Add SLOW_P argument, use
442 BUILT_IN_ASAN_*_N if set.
443 (build_check_stmt): Likewise.
444 (instrument_derefs): If T has insufficient alignment,
445 force same handling as for odd sizes.
447 * sanitizer.def (BUILT_IN_ASAN_REPORT_LOAD_N,
448 BUILT_IN_ASAN_REPORT_STORE_N): New.
449 * asan.c (struct asan_mem_ref): Change access_size type to
451 (asan_mem_ref_init, asan_mem_ref_new, get_mem_refs_of_builtin_call,
452 update_mem_ref_hash_table): Likewise.
453 (asan_mem_ref_hasher::hash): Hash in a HWI.
454 (report_error_func): Change size_in_bytes argument to HWI.
455 Use *_N builtins if size_in_bytes is larger than 16 or not power of
457 (build_shadow_mem_access): New function.
458 (build_check_stmt): Use it. Change size_in_bytes argument to HWI.
459 Handle size_in_bytes not power of two or larger than 16.
460 (instrument_derefs): Don't give up if size_in_bytes is not
461 power of two or is larger than 16.
463 2014-05-30 Kai Tietz <ktietz@redhat.com>
466 * config/i386/i386.c (x86_output_mi_thunk): Add memory case
467 for sibling-tail-calls.
468 * config/i386/i386.md (sibcall_insn_operand): Add memory-constrain
470 * config/i386/predicates.md (sibcall_memory_operand): New predicate.
471 (sibcall_insn_operand): Add check for sibcall_memory_operand.
473 2014-05-30 Pitchumani Sivanupandi <pitchumani.s@atmel.com>
475 * config/avr/avr-mcus.def: Change ATA6289 ISA to AVR4
476 * config/avr/avr-tables.opt: Regenerate.
477 * config/avr/t-multilib: Regenerate.
478 * doc/avr-mmcu.texi: Regenerate.
480 2014-05-30 Ian Lance Taylor <iant@google.com>
482 * config/i386/xmmintrin.h (_mm_pause): Move out of scope of pragma
485 2014-05-30 Tom de Vries <tom@codesourcery.com>
487 * config/i386/i386.c (TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS):
490 2014-05-30 Tom de Vries <tom@codesourcery.com>
492 * lra-int.h (struct lra_reg): Add field actual_call_used_reg_set.
493 * lra.c (initialize_lra_reg_info_element): Add init of
494 actual_call_used_reg_set field.
495 (lra): Call lra_create_live_ranges before lra_inheritance for
497 * lra-assigns.c (lra_assign): Allow call_used_regs to cross calls for
499 * lra-constraints.c (need_for_call_save_p): Use actual_call_used_reg_set
500 instead of call_used_reg_set for -fuse-caller-save.
501 * lra-lives.c (process_bb_lives): Calculate actual_call_used_reg_set.
503 2014-05-30 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
505 * config/arm/thumb2.md (*thumb2_movhi_insn): Set type of movw
507 * config/arm/vfp.md (*thumb2_movsi_vfp): Likewise.
509 2014-05-30 Richard Sandiford <rdsandiford@googlemail.com>
511 * ira.c (ira_get_dup_out_num): Check for output operands at
512 the start of the loop. Handle cases where an included alternative
513 follows an excluded one.
515 2014-05-29 Mike Stump <mikestump@comcast.net>
518 * collect2.c (maybe_run_lto_and_relink): Be sure to always run
519 post ld passes when lto is used.
521 2014-05-29 Vladimir Makarov <vmakarov@redhat.com>
523 PR rtl-optimization/61325
524 * lra-constraints.c (process_address): Rename to process_address_1.
525 (process_address): New function.
527 2014-05-29 Alan Lawrence <alan.lawrence@arm.com>
529 * config/aarch64/aarch64-builtins.c (aarch64_types_binopv_qualifiers,
530 TYPES_BINOPV): New static data.
531 * config/aarch64/aarch64-simd-builtins.def (im_lane_bound):
533 * config/aarch64/aarch64-simd.md (aarch64_ext,
534 aarch64_im_lane_boundsi): New patterns.
535 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_const_1): Match
537 (aarch64_evpc_ext): New function.
539 * config/aarch64/iterators.md (UNSPEC_EXT): New enum element.
541 * config/aarch64/arm_neon.h (vext_f32, vext_f64, vext_p8, vext_p16,
542 vext_s8, vext_s16, vext_s32, vext_s64, vext_u8, vext_u16, vext_u32,
543 vext_u64, vextq_f32, vextq_f64, vextq_p8, vextq_p16, vextq_s8,
544 vextq_s16, vextq_s32, vextq_s64, vextq_u8, vextq_u16, vextq_u32,
545 vextq_u64): Replace __asm with __builtin_shuffle and im_lane_boundsi.
547 2014-05-29 Tom de Vries <tom@codesourcery.com>
549 * rtl.h (BLOCK_SYMBOL_CHECK): Use SYMBOL_REF_FLAGS.
551 2014-05-29 Richard Earnshaw <rearnsha@arm.com>
552 Richard Sandiford <rdsandiford@googlemail.com>
554 * arm/iterators.md (shiftable_ops): New code iterator.
555 (t2_binop0, arith_shift_insn): New code attributes.
556 * arm/predicates.md (shift_nomul_operator): New predicate.
557 * arm/arm.md (insn_enabled): Delete.
558 (enabled): Remove insn_enabled test.
559 (*arith_shiftsi): Delete. Replace with ...
560 (*<arith_shift_insn>_multsi): ... new pattern.
561 (*<arith_shift_insn>_shiftsi): ... new pattern.
562 * config/arm/arm.c (arm_print_operand): Handle operand format 'b'.
564 2014-05-29 Radovan Obradovic <robradovic@mips.com>
565 Tom de Vries <tom@codesourcery.com>
567 * config/mips/mips.h (POST_CALL_TMP_REG): Define.
568 * config/mips/mips.c (mips_emit_call_insn): Add POST_CALL_TMP_REG
570 (mips_split_call): Use POST_CALL_TMP_REG.
571 (TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): Redefine to true.
573 2014-05-29 Tom de Vries <tom@codesourcery.com>
575 * final.c (collect_fn_hard_reg_usage): Guard variable declaration
576 with #ifdef STACK_REGS.
578 2014-05-28 Jan Hubicka <hubicka@ucw.cz>
580 * varasm.c (get_variable_section): Walk aliases.
581 (place_block_symbol): Walk aliases.
583 2014-05-28 Tom de Vries <tom@codesourcery.com>
586 2014-05-28 Tom de Vries <tom@codesourcery.com>
588 * lra-int.h (struct lra_reg): Add field actual_call_used_reg_set.
589 * lra.c (initialize_lra_reg_info_element): Add init of
590 actual_call_used_reg_set field.
591 (lra): Call lra_create_live_ranges before lra_inheritance for
593 * lra-assigns.c (lra_assign): Allow call_used_regs to cross calls for
595 * lra-constraints.c (need_for_call_save_p): Use
596 actual_call_used_reg_set instead of call_used_reg_set for
598 * lra-lives.c (process_bb_lives): Calculate actual_call_used_reg_set.
600 2014-05-28 Richard Sandiford <rdsandiford@googlemail.com>
602 * doc/md.texi: Document that the % constraint character must
603 be at the beginning of the string.
604 * genoutput.c (validate_insn_alternatives): Check that '=',
605 '+' and '%' only appear at the beginning of a constraint.
606 * ira.c (commutative_constraint_p): Delete.
607 (ira_get_dup_out_num): Expect the '%' commutativity marker to be
608 at the start of the string.
609 * config/alpha/alpha.md (*movmemdi_1, *clrmemdi_1): Remove
611 * config/arm/neon.md (bicdi3_neon): Likewise.
612 * config/iq2000/iq2000.md (addsi3_internal, subsi3_internal, sgt_si)
613 (slt_si, sltu_si): Likewise.
614 * config/vax/vax.md (sbcdi3): Likewise.
615 * config/h8300/h8300.md (*cmpstz): Remove duplicate '+'.
616 * config/arc/arc.md (mulsi_600, mulsidi_600, umulsidi_600)
617 (mul64): Move '%' to beginning of constraint.
618 * config/arm/arm.md (*xordi3_insn): Likewise.
619 * config/nds32/nds32.md (add<mode>3, mulsi3, andsi3, iorsi3)
622 2014-05-28 Richard Sandiford <rdsandiford@googlemail.com>
624 * doc/md.texi: Document the restrictions on the "enabled" attribute.
626 2014-05-28 Jason Merrill <jason@redhat.com>
629 * cgraph.h (symtab_node::get_comdat_group_id): New.
630 * cgraphunit.c (analyze_functions): Call it.
631 * symtab.c (dump_symtab_node): Likewise.
632 * tree.c (decl_comdat_group_id): New.
633 * tree.h: Declare it.
634 * lto-streamer-out.c (write_symbol): Use it.
635 * trans-mem.c (ipa_tm_create_version_alias): Likewise.
637 2014-05-28 Francois-Xavier Coudert <fxcoudert@gcc.gnu.org>
640 * wide-int.cc: Do not include longlong.h when compiling with clang.
642 2014-05-28 Richard Biener <rguenther@suse.de>
644 * tree-ssa-propagate.c (add_control_edge): Print less vertical space.
645 * tree-vrp.c (extract_range_from_ssa_name): Also copy VR_UNDEFINED.
646 (vrp_visit_assignment_or_call): Print less vertical space.
647 (vrp_visit_stmt): Likewise.
648 (vrp_visit_phi_node): Likewise. For a PHI argument with
649 VR_VARYING range consider recording it as copy.
651 2014-05-28 Richard Biener <rguenther@suse.de>
654 2014-05-28 Richard Biener <rguenther@suse.de>
656 * hwint.h (HOST_WIDE_INT_PRINT_*): Define in terms of PRI*64.
658 2014-05-28 Bernd Edlinger <bernd.edlinger@hotmail.de>
660 * expr.c (expand_assignment): Fold the bitpos in the to_rtx if
661 sufficiently aligned and an offset is used at the same time.
662 (expand_expr_real_1): Likewise.
664 2014-05-28 Richard Biener <rguenther@suse.de>
667 * fold-const.c (fold_comparison): When folding
668 X +- C1 CMP Y +- C2 to X CMP Y +- C2 +- C1 also ensure
669 the sign of the remaining constant operand stays the same.
671 2014-05-28 Kaushik Phatak <kaushik.phatak@kpit.com>
673 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
674 __RL78_64BIT_DOUBLES__ or __RL78_32BIT_DOUBLES__.
675 (ASM_SPEC): Pass -m64bit-doubles or -m32bit-doubles on
677 (DOUBLE_TYPE_SIZE): Use 64 bit if TARGET_64BIT_DOUBLES is true.
678 * gcc/config/rl78/rl78.opt (m64bit-doubles): New option.
679 (m32bit-doubles) Likewise.
680 * gcc/config/rl78/t-rl78: Add 64-bit-double multilib.
681 * doc/invoke.texi: Document -m32bit-doubles and -m64bit-doubles
684 2014-05-28 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
686 * configure.ac ($gcc_cv_ld_clearcap): New test.
687 * configure: Regenerate.
688 * config.in: Regenerate.
689 * config/sol2.opt (mclear-hwcap): New option.
690 * config/sol2.h (LINK_CLEARCAP_SPEC): Define.
691 * config/sol2-clearcap.map: Moved here from
692 testsuite/gcc.target/i386/clearcap.map.
693 * config/sol2-clearcapv2.map: Move here from
694 gcc.target/i386/clearcapv2.map.
695 * config/t-sol2 (install): Depend on install-clearcap-map.
696 (install-clearcap-map): New target.
697 * doc/invoke.texi (Option Summary, Solaris 2 Options): Document
700 2014-05-28 Richard Biener <rguenther@suse.de>
702 * hwint.h (*_HALF_WIDE_INT*): Move to ...
703 * wide-int.cc (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT):
704 ... here and remove the rest.
705 * hwint.h (HOST_WIDE_INT_PRINT_*): Define in terms of PRI*64.
707 2014-05-28 Richard Biener <rguenther@suse.de>
709 PR tree-optimization/61335
710 * tree-vrp.c (vrp_visit_phi_node): If the compare of old and
711 new range fails, drop to varying.
713 2014-05-28 Tom de Vries <tom@codesourcery.com>
715 * lra-int.h (struct lra_reg): Add field actual_call_used_reg_set.
716 * lra.c (initialize_lra_reg_info_element): Add init of
717 actual_call_used_reg_set field.
718 (lra): Call lra_create_live_ranges before lra_inheritance for
720 * lra-assigns.c (lra_assign): Allow call_used_regs to cross calls for
722 * lra-constraints.c (need_for_call_save_p): Use
723 actual_call_used_reg_set instead of call_used_reg_set for
725 * lra-lives.c (process_bb_lives): Calculate actual_call_used_reg_set.
727 2014-05-28 Radovan Obradovic <robradovic@mips.com>
728 Tom de Vries <tom@codesourcery.com>
730 * doc/invoke.texi (@item Optimization Options): Add -fuse-caller-save
732 (@item -fuse-caller-save): New item.
734 2014-05-28 Radovan Obradovic <robradovic@mips.com>
735 Tom de Vries <tom@codesourcery.com>
737 * opts.c (default_options_table): Add OPT_LEVELS_2_PLUS entry with
738 OPT_fuse_caller_save.
740 2014-05-28 Radovan Obradovic <robradovic@mips.com>
741 Tom de Vries <tom@codesourcery.com>
743 * df-scan.c (df_get_call_refs): Use get_call_reg_set_usage.
744 * caller-save.c (setup_save_areas, save_call_clobbered_regs): Use
745 get_call_reg_set_usage.
746 * resource.c (mark_set_resources, mark_target_live_regs): Use
747 get_call_reg_set_usage.
748 * ira-int.h (struct ira_allocno): Add crossed_calls_clobbered_regs
750 (ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS): Define.
751 * ira-lives.c (process_bb_node_lives): Use get_call_reg_set_usage.
752 Calculate ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS.
753 * ira-build.c (ira_create_allocno): Init
754 ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS.
755 (create_cap_allocno, propagate_allocno_info)
756 (propagate_some_info_from_allocno)
757 (copy_info_to_removed_store_destinations): Handle
758 ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS.
759 * ira-costs.c (ira_tune_allocno_costs): Use
760 ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS to adjust costs.
762 2014-05-28 Radovan Obradovic <robradovic@mips.com>
763 Tom de Vries <tom@codesourcery.com>
765 * cgraph.h (struct cgraph_rtl_info): Add function_used_regs
766 and function_used_regs_valid fields.
767 * final.c: Move include of hard-reg-set.h to before rtl.h to declare
768 find_all_hard_reg_sets.
769 (collect_fn_hard_reg_usage, get_call_fndecl, get_call_cgraph_rtl_info)
770 (get_call_reg_set_usage): New function.
771 (rest_of_handle_final): Use collect_fn_hard_reg_usage.
772 * regs.h (get_call_reg_set_usage): Declare.
774 2014-05-28 Georg-Johann Lay <avr@gjlay.de>
777 * config/dbx.h (License): Add Runtime Library Exception.
778 * config/newlib-stdint.h (License): Same.
779 * config/rtems.h (License): Same
780 * config/initfini-array.h (License): Same
781 * config/v850/v850.h (License): Same.
782 * config/v850/v850-opts.h (License): Same
783 * config/v850/rtems.h (License): Same.
785 2014-05-28 Georg-Johann Lay <avr@gjlay.de>
788 * doc/extend.texi (Local Labels): Note that label differences are
789 not supported for AVR.
791 2014-05-27 Andrew Pinski <apinski@cavium.com>
793 * config/aarch64/aarch64.md (stack_protect_set_<mode>):
794 Use <w> for the register in assembly template.
795 (stack_protect_test): Use the mode of operands[0] for the result.
796 (stack_protect_test_<mode>): Use <w> for the register
797 in assembly template.
799 2014-05-27 DJ Delorie <dj@redhat.com>
801 * config/rx/rx.c (add_vector_labels): New.
802 (rx_output_function_prologue): Call it.
803 (rx_handle_func_attribute): Don't require empty arguments.
804 (rx_handle_vector_attribute): New.
805 (rx_attribute_table): Add "vector" attribute.
806 * doc/extend.texi (interrupt, vector): Document new/changed
807 RX-specific attributes.
809 * config/rx/rx.c (rx_adjust_insn_length): Skip for non-insns.
811 2014-05-27 Eric Botcazou <ebotcazou@adacore.com>
813 * double-int.c (div_and_round_double) <ROUND_DIV_EXPR>: Use the proper
814 predicate to detect a negative quotient.
816 2014-05-27 Eric Botcazou <ebotcazou@adacore.com>
818 * fold-const.c (fold_comparison): Clean up and extend X +- C1 CMP C2
819 to X CMP C2 -+ C1 transformation to EQ_EXPR/NE_EXPR.
820 Add X - Y CMP 0 to X CMP Y transformation.
821 (fold_binary_loc) <EQ_EXPR/NE_EXPR>: Remove same transformations.
823 2014-05-27 Segher Boessenkool <segher@kernel.crashing.org>
825 * stmt.c (dump_case_nodes): Don't convert values to HOST_WIDE_INT
828 2014-05-27 Steve Ellcey <sellcey@mips.com>
830 * config/mips/mips.c: Add include of cgraph.h.
832 2014-05-27 Richard Biener <rguenther@suse.de>
834 * system.h (__STDC_FORMAT_MACROS): Define as very first thing.
836 2014-05-27 Georg-Johann Lay <avr@gjlay.de>
839 * config/arm/arm.h (License): Add note to COPYING.RUNTIME.
840 * config/arm/arm-cores.def (License): Same.
841 * config/arm/arm-opts.h (License): Same.
842 * config/arm/aout.h (License): Same.
843 * config/arm/bpabi.h (License): Same.
844 * config/arm/elf.h (License): Same.
845 * config/arm/linux-elf.h (License): Same.
846 * config/arm/linux-gas.h (License): Same.
847 * config/arm/netbsd-elf.h (License): Same.
848 * config/arm/uclinux-eabi.h (License): Same.
849 * config/arm/uclinux-elf.h (License): Same.
850 * config/arm/vxworks.h (License): Same.
852 2014-05-27 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
854 * config/arm/neon.md (neon_bswap<mode>): New pattern.
855 * config/arm/arm.c (neon_itype): Add NEON_BSWAP.
856 (arm_init_neon_builtins): Handle NEON_BSWAP.
857 Define required type nodes.
858 (arm_expand_neon_builtin): Handle NEON_BSWAP.
859 (arm_builtin_vectorized_function): Handle BUILTIN_BSWAP builtins.
860 * config/arm/arm_neon_builtins.def (bswap): Define builtins.
861 * config/arm/iterators.md (VDQHSD): New mode iterator.
863 2014-05-27 Richard Biener <rguenther@suse.de>
865 * tree-vrp.c (vrp_evaluate_conditional_warnv_with_ops_using_ranges):
866 Try using literal operands when comparing value-ranges failed.
868 2014-05-27 Richard Sandiford <rdsandiford@googlemail.com>
870 * ira.c (commutative_operand): Adjust for change to recog_data.
871 [Missing from previous commit.]
873 2014-05-27 Richard Sandiford <rdsandiford@googlemail.com>
875 * system.h (TEST_BIT): New macro.
876 * recog.h (alternative_mask): New type.
877 (ALL_ALTERNATIVES, ALTERNATIVE_BIT): New macros.
878 (recog_data_d): Replace alternative_enabled_p array with
879 enabled_alternatives.
880 (target_recog): New structure.
881 (default_target_recog, this_target_recog): Declare.
882 (get_enabled_alternatives, recog_init): Likewise.
883 * recog.c (default_target_recog, this_target_recog): New variables.
884 (get_enabled_alternatives): New function.
885 (extract_insn): Use it.
886 (recog_init): New function.
887 (preprocess_constraints, constrain_operands): Adjust for change to
889 * postreload.c (reload_cse_simplify_operands): Likewise.
890 * reload.c (find_reloads): Likewise.
891 * ira-costs.c (record_reg_classes): Likewise.
892 * ira-lives.c (single_reg_class): Likewise. Fix bug in which
893 all alternatives after a disabled one would be skipped.
894 (ira_implicitly_set_insn_hard_regs): Likewise.
895 * ira.c (ira_setup_alts): Adjust for change to recog_data.
896 * lra-int.h (lra_insn_recog_data): Replace alternative_enabled_p
897 with enabled_alternatives.
898 * lra.c (free_insn_recog_data): Update accordingly.
899 (lra_update_insn_recog_data): Likewise.
900 (lra_set_insn_recog_data): Likewise. Use get_enabled_alternatives.
901 * lra-constraints.c (process_alt_operands): Likewise. Handle
902 only_alternative as part of the enabled mask.
903 * target-globals.h (this_target_recog): Declare.
904 (target_globals): Add a recog field.
905 (restore_target_globals): Restore this_target_recog.
906 * target-globals.c: Include recog.h.
907 (default_target_globals): Initialize recog field.
908 (save_target_globals): Likewise.
909 * reginfo.c (reinit_regs): Call recog_init.
910 * toplev.c (backend_init_target): Likewise.
912 2014-05-27 Richard Sandiford <rdsandiford@googlemail.com>
914 * gencodes.c (main): Make LAST_INSN_CODE higher than any insn code,
915 rather than any named insn's code.
917 2014-05-27 Georg-Johann Lay <avr@gjlay.de>
920 * config/arm/arm-opts.h (License): Add GCC Runtime Library Exception.
921 * config/arm/arm-cores.def (License): Same.
923 2014-05-26 Jan Hubicka <hubicka@ucw.cz>
925 * tree.h (decl_comdat_group): Declare.
926 * cgraph.h (symtab_in_same_comdat_p): Move offline to ...
927 * tree.c (decl_comdat_group): Here.
929 2014-05-26 Richard Sandiford <r.sandiford@uk.ibm.com>
931 PR rtl-optimization/61222
932 * combine.c (simplify_shift_const_1): When moving a PLUS outside
933 the shift, truncate the PLUS operand to the result mode.
935 2014-05-26 Uros Bizjak <ubizjak@gmail.com>
938 * config/i386/i386.c (ix86_rtx_costs)
939 <case CONST_INT, case CONST, case LABEL_REF, case SYMBOL_REF>:
942 2014-05-26 Martin Jambor <mjambor@suse.cz>
944 * ira.c (split_live_ranges_for_shrink_wrap): Remove bailout on
947 2014-05-26 Richard Biener <rguenther@suse.de>
949 * wide-int.h (wi::int_traits <long>, wi::int_traits <unsigned long>,
950 wi::int_traits <long long>, wi::int_traits <unsigned long long>):
951 Provide specializations.
952 (wi::int_traits <HOST_WIDE_INT>,
953 wi::int_traits <unsigned HOST_WIDE_INT>): Remove specializations.
955 2014-05-26 Alan Modra <amodra@gmail.com>
958 * config/rs6000/rs6000.c (rs6000_emit_set_const): Remove unneeded
959 params and return a bool. Remove dead code. Update comment.
960 Assert we have a const_int source. Remove bogus code from
961 32-bit HWI days. Move !TARGET_POWERPC64 handling, and correct
962 handling of constants > 2G and reg_equal note, from..
963 (rs6000_emit_set_long_const): ..here. Remove unneeded param and
964 return value. Update comment. If we can, use a new pseudo
965 for intermediate calculations.
966 * config/rs6000/rs6000-protos.h (rs6000_emit_set_const): Update
968 * config/rs6000/rs6000.md (movsi_internal1_single+1): Update
969 call to rs6000_emit_set_const in splitter.
970 (movdi_internal64+2, +3): Likewise.
972 2014-05-26 Richard Biener <rguenther@suse.de>
974 * system.h: Define __STDC_FORMAT_MACROS before
975 including inttypes.h.
976 * hwint.h (HOST_WIDEST_INT, HOST_BITS_PER_WIDEST_INT,
977 HOST_WIDEST_INT_PRINT, HOST_WIDEST_INT_PRINT_DEC,
978 HOST_WIDEST_INT_PRINT_DEC_C, HOST_WIDEST_INT_PRINT_UNSIGNED,
979 HOST_WIDEST_INT_PRINT_HEX, HOST_WIDEST_INT_PRINT_DOUBLE_HEX,
980 HOST_WIDEST_INT_C): Remove.
981 (PRId64, PRIi64, PRIo64, PRIu64, PRIx64, PRIX64): Define
982 if C99 inttypes.h is not available.
983 * coretypes.h (gcov_type, gcov_type_unsigned): Use [u]int64_t.
984 * gcov-io.h (gcov_type, gcov_type_unsigned): Likewise.
985 * gcov-io.c (gcov_histo_index): Drop non-64bit hwi case.
986 * cfgloop.h (struct niter_desc): Use uint64_t for niter field.
987 * bitmap.c (struct bitmap_descriptor_d): Use uint64_t for counters.
988 (struct output_info): Likewise.
989 (print_statistics): Adjust.
990 (dump_bitmap_statistics): Likewise.
991 * bt-load.c (migrate_btr_defs): Print with PRId64.
992 * cfg.c (dump_edge_info, dump_bb_info): Likewise.
993 (MAX_SAFE_MULTIPLIER): Adjust.
994 * cfghooks.c (dump_bb_for_graph): Print with PRId64.
995 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee,
996 dump_cgraph_node): Likewise.
997 * final.c (dump_basic_block_info): Likewise.
998 * gcov-dump.c (tag_counters, tag_summary, dump_working_sets): Likewise.
999 * gcov.c (format_gcov): Likewise.
1000 * ipa-cp.c (good_cloning_opportunity_p): Likewise. Use int64_t
1002 (get_clone_agg_value): Use HOST_WIDE_INT for offset.
1003 * ipa-inline.c (compute_max_insns): Use int64_t for calcuation.
1004 (inline_small_functions, dump_overall_stats, dump_inline_stats):
1005 Use PRId64 for dumping.
1006 * ipa-profile.c (dump_histogram, ipa_profile): Likewise.
1007 * ira-color.c (struct allocno_hard_regs): Use int64_t for cost.
1008 (add_allocno_hard_regs): Adjust.
1009 * loop-doloop.c (doloop_modify): Print using PRId64.
1010 * loop-iv.c (inverse): Compute in uint64_t.
1011 (determine_max_iter, iv_number_of_iterations): Likewise.
1012 * loop-unroll.c (decide_peel_completely, decide_peel_simple):
1014 * lto-streamer-out.c (write_symbol): Use uint64_t.
1015 * mcf.c (CAP_INFINITY): Use int64_t maximum.
1016 (dump_fixup_edge, create_fixup_graph, cancel_negative_cycle,
1017 find_max_flow, adjust_cfg_counts): Use int64_t and dump with PRId64.
1018 * modulo-sched.c (const_iteration_count): Use int64_t.
1019 (sms_schedule): Dump using PRId64.
1020 * predict.c (dump_prediction): Likewise.
1021 * pretty-print.h (pp_widest_integer): Remove.
1022 * profile.c (get_working_sets, is_edge_inconsistent,
1023 is_inconsistent, read_profile_edge_counts): Dump using PRId64.
1024 * tree-pretty-print.c (pp_double_int): Remove case handling
1025 HOST_BITS_PER_DOUBLE_INT == HOST_BITS_PER_WIDEST_INT.
1026 * tree-ssa-math-opts.c (struct symbolic_number): Use uint64_t
1028 (pass_optimize_bswap::execute): Remove restriction on hosts.
1029 * tree-streamer-in.c (streamer_alloc_tree): Use HOST_WIDE_INT.
1030 * tree-streamer-out.c (streamer_write_tree_header): Likewise.
1031 * tree.c (widest_int_cst_value): Remove.
1032 * tree.h (widest_int_cst_value): Likewise.
1033 * value-prof.c (dump_histogram_value): Print using PRId64.
1034 * gengtype.c (main): Also inject int64_t.
1035 * ggc-page.c (struct max_alignment): Use int64_t.
1036 * alloc-pool.c (struct allocation_object_def): Likewise.
1037 * ira-conflicts.c (build_conflict_bit_table): Use uint64_t
1039 * doc/tm.texi.in: Remove reference to HOST_WIDEST_INT.
1040 * doc/tm.texi: Regenerated.
1041 * gengtype-lex.l (IWORD): Handle [u]int64_t.
1042 * config/sh/sh.c (expand_cbranchdi4): Use gcov_type.
1043 * config/mmix/mmix-protos.h (mmix_intval, mmix_shiftable_wyde_value,
1044 mmix_output_register_setting): Use [u]int64_t in prototypes.
1045 * config/mmix/mmix.c (mmix_print_operand, mmix_output_register_setting,
1046 mmix_shiftable_wyde_value, mmix_output_shiftvalue_op_from_str,
1047 mmix_output_octa, mmix_output_shifted_value): Adjust.
1048 (mmix_intval): Adjust. Remove unreachable case.
1049 * config/mmix/mmix.md (*nonlocal_goto_receiver_expanded): Use int64_t.
1051 2014-05-26 Richard Biener <rguenther@suse.de>
1053 * configure.ac: Drop __int64 type check. Insist that we
1054 found uint64_t and int64_t.
1055 * hwint.h (HOST_BITS_PER___INT64): Remove.
1056 (HOST_BITS_PER_WIDE_INT): Define to 64 and remove __int64 case.
1057 (HOST_WIDE_INT_PRINT_*): Remove 32bit case.
1058 (HOST_WIDEST_INT*): Define to HOST_WIDE_INT*.
1059 (HOST_WIDEST_FAST_INT): Remove __int64 case.
1060 * vmsdbg.h (struct _DST_SRC_COMMAND): Use int64_t
1061 for dst_q_src_df_rms_cdt.
1062 * configure: Regenerate.
1063 * config.in: Likewise.
1065 2014-05-26 Michael Tautschnig <mt@debian.org>
1068 * doc/extend.texi (X86 Built-in Functions): Fix parameter lists of
1069 __builtin_ia32_vfrczs[sd] and __builtin_ia32_mpsadbw256.
1071 2014-05-26 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1073 PR rtl-optimization/61278
1074 * shrink-wrap.c (move_insn_for_shrink_wrap): Check df_live.
1076 2014-05-26 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1078 PR rtl-optimization/61220
1079 Part of PR rtl-optimization/61225
1080 * shrink-wrap.c (move_insn_for_shrink_wrap): Skip SP and FP adjustment
1081 insn; skip split_edge for a block with only one successor.
1083 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
1085 * symtab.c (symtab_nonoverwritable_alias): Copy READONLY flag
1088 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
1090 * ipa-visibility.c (can_replace_by_local_alias_in_vtable): New function.
1091 (update_vtable_references): New function.
1092 (function_and_variable_visibility): Rewrite also vtable initializers.
1093 * varpool.c (cgraph_variable_initializer_availability): Remove assert.
1095 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
1097 * ggc.h (ggc_grow): New function.
1098 * ggc-none.c (ggc_grow): New function.
1099 * ggc-page.c (ggc_grow): Likewise.
1101 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
1103 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
1104 address_taken_from_non_vtable_p, comdat_can_be_unshared_p_1,
1105 comdat_can_be_unshared_p, cgraph_externally_visible_p,
1106 varpool_externally_visible_p, can_replace_by_local_alias,
1107 update_visibility_by_resolution_info, function_and_variable_visibility,
1108 pass_data_ipa_function_and_variable_visibility,
1109 make_pass_ipa_function_and_variable_visibility,
1110 whole_program_function_and_variable_visibility,
1111 pass_data_ipa_whole_program_visibility,
1112 make_pass_ipa_whole_program_visibility): Move to ipa-visibility.c
1113 * cgraph.h (cgraph_local_node_p): Declare.
1114 * ipa-visibility.c: New file.
1115 * Makefile.in (OBJS): Add ipa-visiblity.o
1117 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
1119 * gimple-fold.c (can_refer_decl_in_current_unit_p): Be sure
1120 that var decl is available.
1122 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
1124 * tree-core.h (tree_decl_with_vis): Replace comdat_group by
1125 symtab_node pointer.
1126 * tree.c (copy_node_stat): Be sure to not copy symtab_node pointer.
1127 (find_decls_types_r): Do not walk COMDAT_GROUP.
1128 * tree.h (DECL_COMDAT_GROUP): Revamp to use decl_comdat_group.
1129 * varasm.c (make_decl_one_only): Use set_comdat_group;
1130 create node if needed.
1131 * ipa-inline-transform.c (save_inline_function_body): Update
1132 way we decl->symtab mapping.
1133 * symtab.c (symtab_hash, hash_node, eq_node
1134 symtab_insert_node_to_hashtable): Remove.
1135 (symtab_register_node): Update.
1136 (symtab_unregister_node): Update.
1137 (symtab_get_node): Reimplement as inline function.
1138 (symtab_add_to_same_comdat_group): Update.
1139 (symtab_dissolve_same_comdat_group_list): Update.
1140 (dump_symtab_base): Update.
1141 (verify_symtab_base): Update.
1142 (symtab_make_decl_local): Update.
1143 (fixup_same_cpp_alias_visibility): Update.
1144 (symtab_nonoverwritable_alias): Update.
1145 * cgraphclones.c (set_new_clone_decl_and_node_flags): Update.
1146 * ipa.c (update_visibility_by_resolution_info): UPdate.
1147 * bb-reorder.c: Include cgraph.h
1148 * lto-streamer-out.c (DFS_write_tree_body, hash_tree): Do not deal
1150 * ipa-comdats.c (set_comdat_group, ipa_comdats): Update.
1151 * cgraph.c (cgraph_get_create_node): Update.
1152 * cgraph.h (struct symtab_node): Add get_comdat_group, set_comdat_group
1154 (symtab_get_node): Make inline.
1155 (symtab_insert_node_to_hashtable): Remove.
1156 (symtab_can_be_discarded): Update.
1157 (decl_comdat_group): New function.
1158 * tree-streamer-in.c (lto_input_ts_decl_with_vis_tree_pointers):
1160 * lto-cgraph.c (lto_output_node, lto_output_varpool_node): Stream out
1162 (read_comdat_group): New function.
1163 (input_node, input_varpool_node): Use it.
1164 * trans-mem.c (ipa_tm_create_version_alias): Update code creating
1166 * mips.c (mips_start_unique_function): Likewise.
1167 (ix86_code_end): Likewise.
1168 (rs6000_code_end): Likweise.
1169 * tree-streamer-out.c (DECL_COMDAT_GROUP): Do not stream comdat group.
1171 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
1173 * gengtype-state.c (fatal_reading_state): Bring offline.
1174 * optabs.c (widening_optab_handler): Bring offline.
1175 * optabs.h (widening_optab_handler): Likewise.
1176 * final.c (get_attr_length_1): Likewise.
1178 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
1180 * sched-int.h (sd_iterator_cond): Manually tail recurse.
1182 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1184 * config/rs6000/440.md (ppc440-integer): Include shift without dot.
1185 (ppc440-compare): Include shift with dot.
1186 * config/rs6000/e300c2c3.md (ppce300c3_iu): Include shift without dot.
1187 * config/rs6000/e5500.md (e5500_sfx2): Include constant shift
1189 * config/rs6000/e6500.md (e6500_sfx): Exclude constant shift
1191 (e6500_sfx2): Include it.
1192 * config/rs6000/rs6000.md ( *zero_extend<mode>di2_internal1,
1193 *zero_extend<mode>di2_internal2, *zero_extend<mode>di2_internal3,
1194 *zero_extendsidi2_lfiwzx, andsi3_mc, andsi3_nomc,
1195 andsi3_internal0_nomc, extzvsi_internal, extzvdi_internal,
1196 *extzvdi_internal1, *extzvdi_internal2, rotlsi3, *rotlsi3_64,
1197 *rotlsi3_internal4, *rotlsi3_internal7le, *rotlsi3_internal7be,
1198 *rotlsi3_internal10le, *rotlsi3_internal10be, rlwinm,
1199 *lshiftrt_internal1le, *lshiftrt_internal1be,
1200 *lshiftrt_internal4le, *lshiftrt_internal4be, rotldi3,
1201 *rotldi3_internal4, *rotldi3_internal7le, *rotldi3_internal7be,
1202 *rotldi3_internal10le, *rotldi3_internal10be,
1203 *rotldi3_internal13le, *rotldi3_internal13be, *ashldi3_internal4,
1204 ashldi3_internal5, *ashldi3_internal6, *ashldi3_internal7,
1205 ashldi3_internal8, *ashldi3_internal9, anddi3_mc, anddi3_nomc,
1206 *anddi3_internal2_mc, *anddi3_internal3_mc, and 4 anonymous
1207 define_insns): Use type "shift" in the appropriate alternatives.
1209 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1211 * config/rs6000/rs6000.md (type): Add "logical". Delete
1213 (dot): Adjust comment.
1214 (andsi3_mc, *andsi3_internal2_mc, *andsi3_internal3_mc,
1215 *andsi3_internal4, *andsi3_internal5_mc, *boolsi3_internal2,
1216 *boolsi3_internal3, *boolccsi3_internal2, *boolccsi3_internal3,
1217 anddi3_mc, *anddi3_internal2_mc, *anddi3_internal3_mc,
1218 *booldi3_internal2, *booldi3_internal3, *boolcdi3_internal2,
1219 *boolcdi3_internal3, *boolccdi3_internal2, *boolccdi3_internal3,
1220 *mov<mode>_internal2, and 10 anonymous define_insns): Use "logical".
1221 * config/rs6000/rs6000.c (rs6000_adjust_cost): Adjust.
1223 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
1224 * config/rs6000/440.md (ppc440-integer, ppc440-compare): Adjust.
1225 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare): Adjust.
1226 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
1227 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
1228 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare): Adjust.
1229 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
1230 * config/rs6000/8540.md (ppc8540_su): Adjust.
1231 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
1232 cell-cmp-microcoded): Adjust.
1233 * config/rs6000/e300c2c3.md (ppce300c3_cmp, ppce300c3_iu): Adjust.
1234 * config/rs6000/e500mc.md (e500mc_su): Adjust.
1235 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2): Adjust.
1236 * config/rs6000/e5500.md (e5500_sfx, e5500_sfx2): Adjust.
1237 * config/rs6000/e6500.md (e6500_sfx, e6500_sfx2): Adjust.
1238 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
1239 * config/rs6000/power4.md (power4-integer, power4-cmp): Adjust.
1240 * config/rs6000/power5.md (power5-integer, power5-cmp): Adjust.
1241 * config/rs6000/power6.md (power6-integer, power6-fast-compare):
1243 * config/rs6000/power7.md (power7-integer, power7-cmp): Adjust.
1244 * config/rs6000/power8.md (power8-1cyc, power8-fast-compare):
1245 Adjust. Adjust comment.
1246 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
1247 * config/rs6000/titan.md (titan_fxu_adder, titan_fxu_alu): Adjust.
1249 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1251 * config/rs6000/rs6000.md (type): Add "add".
1252 (*add<mode>3_internal1, addsi3_high, *add<mode>3_internal2,
1253 *add<mode>3_internal3, *neg<mode>2_internal, and 5 anonymous
1254 define_insns): Use it.
1255 * config/rs6000/rs6000.c (rs6000_adjust_cost): Adjust.
1257 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
1258 * config/rs6000/440.md (ppc440-integer, ppc440-compare): Adjust.
1259 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare): Adjust.
1260 * config/rs6000/601.md (ppc601-integer): Adjust.
1261 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
1262 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
1263 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare): Adjust.
1264 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
1265 * config/rs6000/8540.md (ppc8540_su): Adjust.
1266 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
1267 cell-cmp-microcoded): Adjust.
1268 * config/rs6000/e300c2c3.md (ppce300c3_cmp, ppce300c3_iu): Adjust.
1269 * config/rs6000/e500mc.md (e500mc_su): Adjust.
1270 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2): Adjust.
1271 * config/rs6000/e5500.md (e5500_sfx, e5500_sfx2): Adjust.
1272 * config/rs6000/e6500.md (e6500_sfx, e6500_sfx2): Adjust.
1273 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
1274 * config/rs6000/power4.md (power4-integer, power4-cmp): Adjust.
1275 * config/rs6000/power5.md (power5-integer, power5-cmp): Adjust.
1276 * config/rs6000/power6.md (power6-integer, power6-fast-compare):
1278 * config/rs6000/power7.md (power7-integer, power7-cmp): Adjust.
1279 * config/rs6000/power8.md (power8-1cyc, power8-fast-compare): Adjust.
1280 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
1281 * config/rs6000/titan.md (titan_fxu_adder, titan_fxu_alu): Adjust.
1283 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1285 * config/rs6000/rs6000.md (type): Delete "var_shift_rotate",
1286 "delayed_compare", "var_delayed_compare".
1287 (var_shift): New attribute.
1288 (cell_micro): Adjust.
1289 (*andsi3_internal2_mc, *andsi3_internal3_mc, *andsi3_internal4,
1290 *andsi3_internal5_mc, *extzvsi_internal1, *extzvsi_internal2,
1291 rotlsi3, *rotlsi3_64, *rotlsi3_internal2, *rotlsi3_internal3,
1292 *rotlsi3_internal4, *rotlsi3_internal5, *rotlsi3_internal6,
1293 *rotlsi3_internal8le, *rotlsi3_internal8be, *rotlsi3_internal9le,
1294 *rotlsi3_internal9be, *rotlsi3_internal10le, *rotlsi3_internal10be,
1295 *rotlsi3_internal11le, *rotlsi3_internal11be, *rotlsi3_internal12le,
1296 *rotlsi3_internal12be, ashlsi3, *ashlsi3_64, lshrsi3, *lshrsi3_64,
1297 *lshiftrt_internal2le, *lshiftrt_internal2be, *lshiftrt_internal3le,
1298 *lshiftrt_internal3be, *lshiftrt_internal5le, *lshiftrt_internal5be,
1299 *lshiftrt_internal5le, *lshiftrt_internal5be, ashrsi3, *ashrsi3_64,
1300 rotldi3, *rotldi3_internal2, *rotldi3_internal3, *rotldi3_internal4,
1301 *rotldi3_internal5, *rotldi3_internal6, *rotldi3_internal7le,
1302 *rotldi3_internal7be, *rotldi3_internal8le, *rotldi3_internal8be,
1303 *rotldi3_internal9le, *rotldi3_internal9be, *rotldi3_internal10le,
1304 *rotldi3_internal10be, *rotldi3_internal11le, *rotldi3_internal11be,
1305 *rotldi3_internal12le, *rotldi3_internal12be, *rotldi3_internal13le,
1306 *rotldi3_internal13be, *rotldi3_internal14le, *rotldi3_internal14be,
1307 *rotldi3_internal15le, *rotldi3_internal15be, *ashldi3_internal1,
1308 *ashldi3_internal2, *ashldi3_internal3, *lshrdi3_internal1,
1309 *lshrdi3_internal2, *lshrdi3_internal3, *ashrdi3_internal1,
1310 *ashrdi3_internal2, *ashrdi3_internal3, *anddi3_internal2_mc,
1311 *anddi3_internal3_mc, as well as 11 anonymous define_insns): Adjust.
1312 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
1313 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
1315 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
1316 * config/rs6000/440.md (ppc440-integer): Adjust.
1317 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare): Adjust.
1318 * config/rs6000/601.md (ppc601-integer, ppc601-compare): Adjust.
1319 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
1320 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
1321 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare): Adjust.
1322 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
1323 * config/rs6000/8540.md (ppc8540_su): Adjust.
1324 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
1325 cell-cmp-microcoded): Adjust.
1326 * config/rs6000/e300c2c3.md (ppce300c3_cmp): Adjust.
1327 * config/rs6000/e500mc.md (e500mc_su): Adjust.
1328 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2,
1329 e500mc64_delayed): Adjust.
1330 * config/rs6000/e5500.md (e5500_sfx, e5500_delayed): Adjust.
1331 * config/rs6000/e6500.md (e6500_sfx, e6500_delayed): Adjust.
1332 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
1333 * config/rs6000/power4.md (power4-integer, power4-compare): Adjust.
1334 * config/rs6000/power5.md (power5-integer, power5-compare): Adjust.
1335 * config/rs6000/power6.md (power6-shift, power6-var-rotate,
1336 power6-delayed-compare, power6-var-delayed-compare): Adjust.
1337 * config/rs6000/power7.md (power7-integer, power7-compare): Adjust.
1338 * config/rs6000/power8.md (power8-1cyc, power8-compare): Adjust.
1340 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
1341 * config/rs6000/titan.md (titan_fxu_shift_and_rotate): Adjust.
1343 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1345 * config/rs6000/rs6000.md (type): Delete "idiv", "ldiv". Add "div".
1346 (bits): New mode_attr.
1347 (idiv_ldiv): Delete mode_attr.
1348 (udiv<mode>3, *div<mode>3, div<div_extend>_<mode>): Adjust.
1349 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
1350 rs6000_adjust_priority, is_nonpipeline_insn,
1351 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
1353 * config/rs6000/40x.md (ppc403-idiv): Adjust.
1354 * config/rs6000/440.md (ppc440-idiv): Adjust.
1355 * config/rs6000/476.md (ppc476-idiv): Adjust.
1356 * config/rs6000/601.md (ppc601-idiv): Adjust.
1357 * config/rs6000/603.md (ppc603-idiv): Adjust.
1358 * config/rs6000/6xx.md (ppc604-idiv, ppc620-idiv, ppc630-idiv,
1359 ppc620-ldiv): Adjust.
1360 * config/rs6000/7450.md (ppc7450-idiv): Adjust.
1361 * config/rs6000/7xx.md (ppc750-idiv): Adjust.
1362 * config/rs6000/8540.md (ppc8540_divide): Adjust.
1363 * config/rs6000/a2.md (ppca2-idiv, ppca2-ldiv): Adjust.
1364 * config/rs6000/cell.md (cell-idiv, cell-ldiv): Adjust.
1365 * config/rs6000/e300c2c3.md (ppce300c3_divide): Adjust.
1366 * config/rs6000/e500mc.md (e500mc_divide): Adjust.
1367 * config/rs6000/e500mc64.md (e500mc64_divide): Adjust.
1368 * config/rs6000/e5500.md (e5500_divide, e5500_divide_d): Adjust.
1369 * config/rs6000/e6500.md (e6500_divide, e6500_divide_d): Adjust.
1370 * config/rs6000/mpc.md (mpccore-idiv): Adjust.
1371 * config/rs6000/power4.md (power4-idiv, power4-ldiv): Adjust.
1372 * config/rs6000/power5.md (power5-idiv, power5-ldiv): Adjust.
1373 * config/rs6000/power6.md (power6-idiv, power6-ldiv): Adjust.
1374 * config/rs6000/power7.md (power7-idiv, power7-ldiv): Adjust.
1375 * config/rs6000/power8.md (power8-idiv, power8-ldiv): Adjust.
1376 * config/rs6000/rs64.md (rs64a-idiv, rs64a-ldiv): Adjust.
1377 * config/rs6000/titan.md (titan_fxu_div): Adjust.
1379 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1381 * config/rs6000/rs6000.md (type): Delete "insert_word",
1382 "insert_dword". Add "insert".
1383 (size): Update comment.
1384 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
1385 insn_must_be_first_in_group): Adjust.
1386 (insvsi_internal, *insvsi_internal1, *insvsi_internal2,
1387 *insvsi_internal3, *insvsi_internal4, *insvsi_internal5,
1388 *insvsi_internal6, insvdi_internal): Adjust.
1390 * config/rs6000/40x.md (ppc403-integer): Adjust.
1391 * config/rs6000/440.md (ppc440-integer): Adjust.
1392 * config/rs6000/476.md (ppc476-simple-integer): Adjust.
1393 * config/rs6000/601.md (ppc601-integer): Adjust.
1394 * config/rs6000/603.md (ppc603-integer): Adjust.
1395 * config/rs6000/6xx.md (ppc604-integer): Adjust.
1396 * config/rs6000/7450.md (ppc7450-integer): Adjust.
1397 * config/rs6000/7xx.md (ppc750-integer): Adjust.
1398 * config/rs6000/8540.md (ppc8540_su): Adjust.
1399 * config/rs6000/cell.md (cell-integer, cell-insert): Adjust.
1400 * config/rs6000/e300c2c3.md (ppce300c3_iu): Adjust.
1401 * config/rs6000/e500mc.md (e500mc_su): Adjust.
1402 * config/rs6000/e500mc64.md (e500mc64_su): Adjust.
1403 * config/rs6000/e5500.md (e5500_sfx): Adjust.
1404 * config/rs6000/e6500.md (e6500_sfx): Adjust.
1405 * config/rs6000/mpc.md (mpccore-integer): Adjust.
1406 * config/rs6000/power4.md (power4-integer, power4-insert): Adjust.
1407 * config/rs6000/power5.md (power5-integer, power5-insert): Adjust.
1408 * config/rs6000/power6.md (power6-insert, power6-insert-dword): Adjust.
1409 * config/rs6000/power7.md (power7-integer): Adjust.
1410 * config/rs6000/power8.md (power8-1cyc): Adjust.
1411 * config/rs6000/rs64.md (rs64a-integer): Adjust.
1412 * config/rs6000/titan.md (titan_fxu_shift_and_rotate): Adjust.
1414 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1416 * config/rs6000/rs6000.md (type): Add "mul". Delete "imul",
1417 "imul2", "imul3", "lmul", "imul_compare", "lmul_compare".
1418 (size): New attribute.
1419 (dot): New attribute.
1420 (cell_micro): Adjust.
1421 (mulsi3, *mulsi3_internal1, *mulsi3_internal2, mulsidi3,
1422 umulsidi3, smulsi3_highpart, umulsi3_highpart, muldi3,
1423 *muldi3_internal1, *muldi3_internal2, smuldi3_highpart,
1424 umuldi3_highpart): Adjust.
1425 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
1426 rs6000_adjust_priority, is_nonpipeline_insn,
1427 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
1429 * config/rs6000/40x.md (ppc403-imul, ppc405-imul, ppc405-imul2,
1430 ppc405-imul3): Adjust.
1431 * config/rs6000/440.md (ppc440-imul, ppc440-imul2): Adjust.
1432 * config/rs6000/476.md (ppc476-imul): Adjust.
1433 * config/rs6000/601.md (ppc601-imul): Adjust.
1434 * config/rs6000/603.md (ppc603-imul, ppc603-imul2): Adjust.
1435 * config/rs6000/6xx.md (ppc604-imul, ppc604e-imul, ppc620-imul,
1436 ppc620-imul2, ppc620-imul3, ppc620-lmul): Adjust.
1437 * config/rs6000/7450.md (ppc7450-imul, ppc7450-imul2): Adjust.
1438 * config/rs6000/7xx.md (ppc750-imul, ppc750-imul2, ppc750-imul3):
1440 * config/rs6000/8540.md (ppc8540_multiply): Adjust.
1441 * config/rs6000/a2.md (ppca2-imul, ppca2-lmul): Adjust.
1442 * config/rs6000/cell.md (cell-lmul, cell-lmul-cmp, cell-imul23,
1444 * config/rs6000/e300c2c3.md (ppce300c3_multiply): Adjust.
1445 * config/rs6000/e500mc.md (e500mc_multiply): Adjust.
1446 * config/rs6000/e500mc64.md (e500mc64_multiply): Adjust.
1447 * config/rs6000/e5500.md (e5500_multiply, e5500_multiply_i): Adjust.
1448 * config/rs6000/e6500.md (e6500_multiply, e6500_multiply_i): Adjust.
1449 * config/rs6000/mpc.md (mpccore-imul): Adjust.
1450 * config/rs6000/power4.md (power4-lmul-cmp, power4-imul-cmp,
1451 power4-lmul, power4-imul, power4-imul3): Adjust.
1452 * config/rs6000/power5.md (power5-lmul-cmp, power5-imul-cmp,
1453 power5-lmul, power5-imul, power5-imul3): Adjust.
1454 * config/rs6000/power6.md (power6-lmul-cmp, power6-imul-cmp,
1455 power6-lmul, power6-imul, power6-imul3): Adjust.
1456 * config/rs6000/power7.md (power7-mul, power7-mul-compare): Adjust.
1457 * config/rs6000/power8.md (power8-mul, power8-mul-compare): Adjust.
1459 * config/rs6000/rs64.md (rs64a-imul, rs64a-imul2, rs64a-imul3,
1460 rs64a-lmul): Adjust.
1461 * config/rs6000/titan.md (titan_imul): Adjust.
1463 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1465 * config/rs6000/rs6000.md (type): Add new value "halfmul".
1466 (*macchwc, *macchw, *macchwuc, *macchwu, *machhwc, *machhw,
1467 *machhwuc, *machhwu, *maclhwc, *maclhw, *maclhwuc, *maclhwu,
1468 *nmacchwc, *nmacchw, *nmachhwc, *nmachhw, *nmaclhwc, *nmaclhw,
1469 *mulchwc, *mulchw, *mulchwuc, *mulchwu, *mulhhwc, *mulhhw,
1470 *mulhhwuc, *mulhhwu, *mullhwc, *mullhw, *mullhwuc, *mullhwu): Use it.
1471 * config/rs6000/40x.md (ppc405-imul3): Add type halfmul.
1472 * config/rs6000/440.md (ppc440-imul2): Add type halfmul.
1473 * config/rs6000/476.md (ppc476-imul): Add type halfmul.
1474 * config/rs6000/titan.md: Delete nonsensical comment.
1475 (titan_imul): Add type imul3.
1476 (titan_mulhw): Remove type imul3; add type halfmul.
1478 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1480 * config/rs6000/rs6000.md (type): Reorder, reformat.
1482 2014-05-23 Martin Jambor <mjambor@suse.cz>
1484 PR tree-optimization/53787
1485 * params.def (PARAM_IPA_MAX_AA_STEPS): New param.
1486 * ipa-prop.h (ipa_node_params): Rename uses_analysis_done to
1487 analysis_done, update all uses.
1488 * ipa-prop.c: Include domwalk.h
1489 (param_analysis_info): Removed.
1490 (param_aa_status): New type.
1491 (ipa_bb_info): Likewise.
1492 (func_body_info): Likewise.
1493 (ipa_get_bb_info): New function.
1494 (aa_overwalked): Likewise.
1495 (find_dominating_aa_status): Likewise.
1496 (parm_bb_aa_status_for_bb): Likewise.
1497 (parm_preserved_before_stmt_p): Changed to use new param AA info.
1498 (load_from_unmodified_param): Accept func_body_info as a parameter
1499 instead of parms_ainfo.
1500 (parm_ref_data_preserved_p): Changed to use new param AA info.
1501 (parm_ref_data_pass_through_p): Likewise.
1502 (ipa_load_from_parm_agg_1): Likewise. Update callers.
1503 (compute_complex_assign_jump_func): Changed to use new param AA info.
1504 (compute_complex_ancestor_jump_func): Likewise.
1505 (ipa_compute_jump_functions_for_edge): Likewise.
1506 (ipa_compute_jump_functions): Removed.
1507 (ipa_compute_jump_functions_for_bb): New function.
1508 (ipa_analyze_indirect_call_uses): Likewise, moved variable
1510 (ipa_analyze_virtual_call_uses): Accept func_body_info instead of node
1511 and info, moved variable declarations down.
1512 (ipa_analyze_call_uses): Accept and pass on func_body_info instead of
1514 (ipa_analyze_stmt_uses): Likewise.
1515 (ipa_analyze_params_uses): Removed.
1516 (ipa_analyze_params_uses_in_bb): New function.
1517 (ipa_analyze_controlled_uses): Likewise.
1518 (free_ipa_bb_info): Likewise.
1519 (analysis_dom_walker): New class.
1520 (ipa_analyze_node): Handle node-specific forbidden analysis,
1521 initialize and free func_body_info, use dominator walker.
1522 (ipcp_modif_dom_walker): New class.
1523 (ipcp_transform_function): Create and free func_body_info, use
1524 ipcp_modif_dom_walker, moved a lot of functionality there.
1526 2014-05-23 Marek Polacek <polacek@redhat.com>
1527 Jakub Jelinek <jakub@redhat.com>
1529 * builtins.def: Change SANITIZE_FLOAT_DIVIDE to SANITIZE_NONDEFAULT.
1530 * gcc.c (sanitize_spec_function): Likewise.
1531 * convert.c (convert_to_integer): Include "ubsan.h". Add
1532 floating-point to integer instrumentation.
1533 * doc/invoke.texi: Document -fsanitize=float-cast-overflow.
1534 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_CAST and
1535 SANITIZE_NONDEFAULT.
1536 * opts.c (common_handle_option): Handle -fsanitize=float-cast-overflow.
1537 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_FLOAT_CAST_OVERFLOW,
1538 BUILT_IN_UBSAN_HANDLE_FLOAT_CAST_OVERFLOW_ABORT): Add.
1539 * ubsan.c: Include "realmpfr.h" and "dfp.h".
1540 (get_ubsan_type_info_for_type): Handle REAL_TYPEs.
1541 (ubsan_type_descriptor): Set tkind to 0xffff for types other than
1542 float/double/long double.
1543 (ubsan_instrument_float_cast): New function.
1544 * ubsan.h (ubsan_instrument_float_cast): Declare.
1546 2014-05-23 Jiong Wang <jiong.wang@arm.com>
1548 * config/aarch64/predicates.md (aarch64_call_insn_operand): New
1550 * config/aarch64/constraints.md ("Ucs", "Usf"): New constraints.
1551 * config/aarch64/aarch64.md (*sibcall_insn, *sibcall_value_insn):
1552 Adjust for tailcalling through registers.
1553 * config/aarch64/aarch64.h (enum reg_class): New caller save
1555 (REG_CLASS_NAMES): Likewise.
1556 (REG_CLASS_CONTENTS): Likewise.
1557 * config/aarch64/aarch64.c (aarch64_function_ok_for_sibcall):
1558 Allow tailcalling without decls.
1560 2014-05-23 Thomas Schwinge <thomas@codesourcery.com>
1562 * gimplify.c (omp_notice_variable) <case OMP_CLAUSE_DEFAULT_NONE>:
1563 Rewrite check for ORT_PARALLEL and ORT_COMBINED_PARALLEL.
1565 * omp-low.c (expand_omp_for_static_chunk): Rename variable si to
1566 gsi, and variables v_* to v*.
1568 2014-05-23 Eric Botcazou <ebotcazou@adacore.com>
1570 * varasm.c (output_constructor_bitfield): Fix thinkos in latest change.
1572 2014-05-23 Thomas Schwinge <thomas@codesourcery.com>
1574 * gimple.h (enum gf_mask): Add and use GF_OMP_FOR_SIMD.
1575 * omp-low.c: Update accordingly.
1577 * gimple.h (enum gf_mask): Rewrite "<< 0" shift expressions used
1578 for GF_OMP_FOR_KIND_MASK, GF_OMP_FOR_KIND_FOR,
1579 GF_OMP_FOR_KIND_DISTRIBUTE, GF_OMP_FOR_KIND_SIMD,
1580 GF_OMP_FOR_KIND_CILKSIMD, GF_OMP_TARGET_KIND_MASK,
1581 GF_OMP_TARGET_KIND_REGION, GF_OMP_TARGET_KIND_DATA,
1582 GF_OMP_TARGET_KIND_UPDATE.
1584 * gimplify.c (omp_notice_variable) <case OMP_CLAUSE_DEFAULT_NONE>:
1585 Explicitly enumerate the expected region types.
1587 2014-05-23 Paul Eggert <eggert@cs.ucla.edu>
1590 * doc/extend.texi (Function Attributes): Fix __attribute__ ((malloc))
1591 documentation; the old documentation didn't clearly state the
1592 constraints on the contents of the pointed-to storage.
1594 2014-05-23 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
1596 Fix bootstrap error on ia64
1597 * config/ia64/ia64.c (ia64_first_cycle_multipass_dfa_lookahead_guard):
1598 Return default value.
1600 2014-05-23 Thomas Preud'homme <thomas.preudhomme@arm.com>
1602 PR tree-optimization/54733
1603 * tree-ssa-math-opts.c (nop_stats): New "bswap_stats" structure.
1605 (find_bswap_or_nop_load): New.
1606 (find_bswap_1): Renamed to ...
1607 (find_bswap_or_nop_1): This. Also add support for memory source.
1608 (find_bswap): Renamed to ...
1609 (find_bswap_or_nop): This. Also add support for memory source and
1610 detection of bitwise operations equivalent to load in host endianness.
1611 (execute_optimize_bswap): Likewise. Also move its leading comment back
1612 in place and split statement transformation into ...
1613 (bswap_replace): This.
1615 2014-05-22 Vladimir Makarov <vmakarov@redhat.com>
1617 PR rtl-optimization/61215
1618 * lra-elelimination.c (lra_eliminate_regs_1): Don't use
1619 simplify_gen_subreg until final substitution.
1621 2014-05-23 Alan Modra <amodra@gmail.com>
1624 * config/rs6000/rs6000.c (mem_operand_gpr): Handle SImode.
1625 * config/rs6000/rs6000.md (extendsidi2_lfiwax, extendsidi2_nocell):
1626 Use "Y" constraint rather than "m".
1628 2014-05-23 Kugan Vivekanandarajah <kuganv@linaro.org>
1630 * config/aarch64/aarch64.c (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New
1632 * config/aarch64/aarch64-protos.h (aarch64_atomic_assign_expand_fenv):
1633 New function declaration.
1634 * config/aarch64/aarch64-builtins.c (aarch64_builtins) : Add
1635 AARCH64_BUILTIN_GET_FPCR, AARCH64_BUILTIN_SET_FPCR.
1636 AARCH64_BUILTIN_GET_FPSR and AARCH64_BUILTIN_SET_FPSR.
1637 (aarch64_init_builtins) : Initialize builtins
1638 __builtins_aarch64_set_fpcr, __builtins_aarch64_get_fpcr.
1639 __builtins_aarch64_set_fpsr and __builtins_aarch64_get_fpsr.
1640 (aarch64_expand_builtin) : Expand builtins __builtins_aarch64_set_fpcr
1641 __builtins_aarch64_get_fpcr, __builtins_aarch64_get_fpsr,
1642 and __builtins_aarch64_set_fpsr.
1643 (aarch64_atomic_assign_expand_fenv): New function.
1644 * config/aarch64/aarch64.md (set_fpcr): New pattern.
1645 (get_fpcr) : Likewise.
1646 (set_fpsr) : Likewise.
1647 (get_fpsr) : Likewise.
1648 (unspecv): Add UNSPECV_GET_FPCR and UNSPECV_SET_FPCR, UNSPECV_GET_FPSR
1649 and UNSPECV_SET_FPSR.
1650 * doc/extend.texi (AARCH64 Built-in Functions) : Document
1651 __builtins_aarch64_set_fpcr, __builtins_aarch64_get_fpcr.
1652 __builtins_aarch64_set_fpsr and __builtins_aarch64_get_fpsr.
1654 2014-05-22 Vladimir Makarov <vmakarov@redhat.com>
1656 PR rtl-optimization/60969
1657 * ira-costs.c (record_reg_classes): Process NO_REGS for matching
1658 constraints. Set up mem cost for NO_REGS case.
1660 2014-05-22 Thomas Schwinge <thomas@codesourcery.com>
1662 * builtin-types.def: Simplify examples for DEF_FUNCTION_TYPE_*.
1664 2012-05-22 Bernd Schmidt <bernds@codesourcery.com>
1666 * config/darwin.c: Include "lto-section-names.h".
1667 (LTO_SEGMENT_NAME): Don't define.
1668 * config/i386/winnt.c: Include "lto-section-names.h".
1669 * lto-streamer.c: Include "lto-section-names.h".
1670 * lto-streamer.h (LTO_SECTION_NAME_PREFIX): Don't define.
1671 * lto-wrapper.c: Include "lto-section-names.h".
1672 (LTO_SECTION_NAME_PREFIX): Don't define.
1673 * lto-section-names.h: New file.
1674 * cgraphunit.c: Include "lto-section-names.h".
1676 2014-05-22 Peter Bergner <bergner@vnet.ibm.com>
1678 * config/rs6000/htm.md (ttest): Use correct shift value to get CR0.
1680 2014-05-22 Richard Earnshaw <rearnsha@arm.com>
1683 * arm.md (arm_cmpdi_unsigned): Fix length calculation for Thumb2.
1685 2014-05-22 Nick Clifton <nickc@redhat.com>
1687 * config/msp430/msp430.h (ASM_SPEC): Add spaces after inserted options.
1689 2014-05-22 Eric Botcazou <ebotcazou@adacore.com>
1691 * tree-ssa-forwprop.c (associate_plusminus): Extend (T)(P + A) - (T)P
1692 -> (T)A transformation to integer types.
1694 2014-05-22 Teresa Johnson <tejohnson@google.com>
1696 * gcov-io.c (gcov_position): Use gcov_nonruntime_assert.
1697 (gcov_is_error): Remove gcc_assert from IN_LIBGCOV code.
1698 (gcov_rewrite): Use gcov_nonruntime_assert.
1700 (gcov_write_words): Ditto.
1701 (gcov_write_length): Ditto.
1702 (gcov_read_words): Use gcov_nonruntime_assert, and remove
1703 gcc_assert from IN_LIBGCOV code.
1704 (gcov_read_summary): Use gcov_error to flag profile corruption.
1705 (gcov_sync): Use gcov_nonruntime_assert.
1706 (gcov_seek): Remove gcc_assert from IN_LIBGCOV code.
1707 (gcov_histo_index): Use gcov_nonruntime_assert.
1708 (static void gcov_histogram_merge): Ditto.
1709 (compute_working_sets): Ditto.
1710 * gcov-io.h (gcov_nonruntime_assert): Define.
1711 (gcov_error): Define for !IN_LIBGCOV
1713 2014-05-22 Richard Biener <rguenther@suse.de>
1715 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
1716 BUILT_IN_REALLOC like BUILT_IN_STRDUP.
1717 (call_may_clobber_ref_p_1): Handle BUILT_IN_REALLOC as allocation
1718 and deallocation site.
1719 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
1720 Handle BUILT_IN_REALLOC similar to BUILT_IN_STRDUP with also
1721 passing through the incoming points-to set.
1722 (handle_lhs_call): Use flags argument instead of recomputing it.
1723 (find_func_aliases_for_call): Call handle_lhs_call with proper
1726 2014-05-22 Jakub Jelinek <jakub@redhat.com>
1728 * tree-streamer-in.c (unpack_ts_real_cst_value_fields): Make sure
1729 all padding bits in REAL_VALUE_TYPE are cleared.
1731 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
1733 Cleanup and improve multipass_dfa_lookahead_guard
1734 * config/i386/i386.c (core2i7_first_cycle_multipass_filter_ready_try,)
1735 (core2i7_first_cycle_multipass_begin,)
1736 (core2i7_first_cycle_multipass_issue,)
1737 (core2i7_first_cycle_multipass_backtrack): Update signature.
1738 * config/ia64/ia64.c
1739 (ia64_first_cycle_multipass_dfa_lookahead_guard_spec): Remove.
1740 (ia64_first_cycle_multipass_dfa_lookahead_guard): Update signature.
1741 (TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC): Remove
1743 (ia64_first_cycle_multipass_dfa_lookahead_guard): Merge logic from
1744 ia64_first_cycle_multipass_dfa_lookahead_guard_spec. Update return
1746 * config/rs6000/rs6000.c (rs6000_use_sched_lookahead_guard): Update
1748 * doc/tm.texi: Regenerate.
1750 (TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC): Remove.
1751 * haifa-sched.c (ready_try): Make signed to allow negative values.
1752 (rebug_ready_list_1): Update.
1753 (choose_ready): Simplify.
1754 (sched_extend_ready_list): Update.
1756 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
1758 Remove IA64 speculation tweaking flags
1759 * config/ia64/ia64.c (ia64_set_sched_flags): Delete handling of
1760 speculation tuning flags.
1761 (msched-prefer-non-data-spec-insns,)
1762 (msched-prefer-non-control-spec-insns): Obsolete options.
1763 * haifa-sched.c (choose_ready): Remove handling of
1764 PREFER_NON_CONTROL_SPEC and PREFER_NON_DATA_SPEC.
1765 * sched-int.h (enum SPEC_SCHED_FLAGS): Remove PREFER_NON_CONTROL_SPEC
1766 and PREFER_NON_DATA_SPEC.
1767 * sel-sched.c (process_spec_exprs): Remove handling of
1768 PREFER_NON_CONTROL_SPEC and PREFER_NON_DATA_SPEC.
1770 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
1772 Improve scheduling debug output
1773 * haifa-sched.c (debug_ready_list): Remove unnecessary prototype.
1774 (advance_one_cycle): Update.
1775 (schedule_insn, queue_to_ready): Add debug printouts.
1776 (debug_ready_list_1): New static function.
1777 (debug_ready_list): Update.
1778 (max_issue): Add debug printouts.
1779 (dump_insn_stream): New static function.
1780 (schedule_block): Use it. Also better indent printouts.
1782 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
1784 Fix sched_insn debug counter
1785 * haifa-sched.c (schedule_insn): Update.
1786 (struct haifa_saved_data): Add nonscheduled_insns_begin.
1787 (save_backtrack_point, restore_backtrack_point): Update.
1788 (first_nonscheduled_insn): New static function.
1789 (queue_to_ready, choose_ready): Use it.
1790 (schedule_block): Init nonscheduled_insns_begin.
1791 (sched_emit_insn): Update.
1794 2014-05-22 Kugan Vivekanandarajah <kuganv@linaro.org>
1796 * config/aarch64/aarch64.c (aarch64_regno_regclass) : Change CORE_REGS
1798 (aarch64_secondary_reload) : LikeWise.
1799 (aarch64_class_max_nregs) : Remove CORE_REGS.
1800 * config/aarch64/aarch64.h (enum reg_class) : Remove CORE_REGS.
1801 (REG_CLASS_NAMES) : Likewise.
1802 (REG_CLASS_CONTENTS) : LikeWise.
1803 (INDEX_REG_CLASS) : Change CORE_REGS to GENERAL_REGS.
1805 2014-05-21 Guozhi Wei <carrot@google.com>
1808 * config/aarch64/arm_neon.h (vqdmulh_n_s16): Change the last operand's
1810 (vqdmulhq_n_s16): Likewise.
1812 2014-05-21 Segher Boessenkool <segher@kernel.crashing.org>
1814 * config/rs6000/predicates.md (update_indexed_address_mem): Delete.
1816 2014-05-21 Marek Polacek <polacek@redhat.com>
1819 * ubsan.c (is_ubsan_builtin_p): Turn assert into a condition.
1821 2014-05-21 Martin Jambor <mjambor@suse.cz>
1823 * doc/invoke.texi (Optimize Options): Document parameters
1824 ipa-cp-eval-threshold, ipa-max-agg-items, ipa-cp-loop-hint-bonus and
1825 ipa-cp-array-index-hint-bonus.
1827 2014-05-21 Mark Wielaard <mjw@redhat.com>
1830 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_type if DWARF
1831 version >= 3 or not strict DWARF.
1832 * langhooks.h (struct lang_hooks_for_types): Add
1833 enum_underlying_base_type.
1834 * langhooks.c (lhd_enum_underlying_base_type): New function.
1835 * gcc/langhooks.h (struct lang_hooks_for_types): Add
1836 enum_underlying_base_type.
1837 * langhooks-def.h (lhd_enum_underlying_base_type): New declaration.
1838 (LANG_HOOKS_ENUM_UNDERLYING_BASE_TYPE): New define.
1839 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add new lang hook.
1841 2014-05-21 Richard Biener <rguenther@suse.de>
1843 * doc/invoke.texi (-flto-partition=): Document one and none algorithms.
1845 2014-05-21 John Marino <gnugcc@marino.st>
1847 * config.gcc (*-*-dragonfly*): New target.
1848 * configure.ac: Detect dl_iterate_phdr (*freebsd*, *dragonfly*).
1849 * configure: Regenerate.
1850 * config/dragonfly-stdint.h: New.
1851 * config/dragonfly.h: New.
1852 * config/dragonfly.opt: New.
1853 * config/i386/dragonfly.h: New.
1854 * ginclude/stddef.h: Detect _PTRDIFF_T_DECLARED for DragonFly.
1856 2014-05-21 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1858 * tree.def (VOID_CST): New.
1859 * tree-core.h (TI_VOID): New.
1860 * tree.h (void_node): New.
1861 * tree.c (tree_node_structure_for_code, tree_code_size)
1862 (iterative_hash_expr): Handle VOID_CST.
1863 (build_common_tree_nodes): Initialize void_node.
1865 2014-05-21 Bernd Schmidt <bernds@codesourcery.com>
1867 * reload1.c (remove_init_insns, will_delete_init_insn_p): New static
1869 (reload, calculate_needs_all_insns, reload_as_needed): Use them.
1871 * config/bfin/bfin.c (split_load_immediate): Use gen_int_mode in a few
1874 * cfgrtl.c (cfg_layout_initialize): Weaken assert to only trigger if
1875 flag_reorder_blocks_and_partition.
1876 * hw-doloop.c (reorg_loops): Avoid reordering if that flag is set.
1878 2014-05-21 Oleg Endo <olegendo@gcc.gnu.org>
1881 * config/sh/sh.md (*addc_r_1): Rename to addc_t_r. Remove empty
1883 (*addc_r_t): Add new insn_and_split.
1885 2014-05-21 Jakub Jelinek <jakub@redhat.com>
1888 * omp-low.c (handle_simd_reference): New function.
1889 (lower_rec_input_clauses): Use it. Defer adding reference
1890 initialization even for reduction without placeholder if in simd,
1891 handle it properly later on.
1893 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
1895 PR tree-optimization/60899
1896 * gimple-fold.c (can_refer_decl_in_current_unit_p): Cleanup;
1897 assume all static symbols will have definition wile parsing and
1898 check the do have definition later in compilation; check that
1899 variable referring symbol will be output before concluding that
1900 reference is safe; be conservative for referring local statics;
1901 be more precise about when comdat is output in other partition.
1903 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
1906 * ipa-inline-transform.c (inline_call): Use add CALLEE_REMOVED
1908 * ipa-inline.c (inline_to_all_callers): If callee was removed; return.
1909 (ipa_inline): Loop inline_to_all_callers until no more aliases
1912 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
1914 * ipa.c (ipa_discover_readonly_nonaddressable_var): Fix dumping;
1915 set writeonly flag only for vars actually written to.
1917 2014-05-20 Dehao Chen <dehao@google.com>
1919 * ipa-inline-transform.c (clone_inlined_nodes): Use min of edge count
1920 and callee count to get clone count.
1921 * tree-inline.c (expand_call_inline): Use callee count instead of bb
1924 2014-05-20 Richard Sandiford <rdsandiford@googlemail.com>
1926 PR rtl-optimization/61243
1927 * emit-rtl.c (emit_copy_of_insn_after): Copy CROSSING_JUMP_P.
1929 2014-05-20 Xinliang David Li <davidxl@google.com>
1931 * cgraphunit.c (walk_polymorphic_call_targets): Add
1932 dbgcnt and fopt-info support.
1933 * ipa-prop.c (ipa_make_edge_direct_to_target): Ditto.
1934 * ipa-devirt.c (ipa_devirt): Ditto.
1935 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children): Ditto.
1936 * ipa.c (walk_polymorphic_call_targets): Ditto.
1937 * gimple-fold.c (fold_gimple_assign): Ditto.
1938 (gimple_fold_call): Ditto.
1939 * dbgcnt.def: New counter.
1941 2014-05-20 DJ Delorie <dj@redhat.com>
1943 * config/msp430/msp430.md (split): Don't allow subregs when
1944 splitting SImode adds.
1945 (andneghi): Fix subtraction logic.
1946 * config/msp430/predicates.md (msp430_nonsubreg_or_imm_operand): New.
1948 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
1950 * tree.h (DECL_ONE_ONLY): Return true only for externally visible
1952 * except.c (switch_to_exception_section, resolve_unique_section,
1953 get_named_text_section, default_function_rodata_section,
1954 align_variable, get_block_for_decl, default_section_type_flags):
1955 Use DECL_COMDAT_GROUP instead of DECL_ONE_ONLY.
1956 * symtab.c (symtab_add_to_same_comdat_group,
1957 symtab_make_decl_local, fixup_same_cpp_alias_visibility,
1958 symtab_nonoverwritable_alias, symtab_get_symbol_partitioning_class):
1960 * cgraphclones.c (cgraph_create_virtual_clone): Likewise.
1961 * bb-reorder.c (pass_partition_blocks::gate): Likewise.
1962 * config/c6x/c6x.c (c6x_elf_unique_section): Likewise.
1963 (c6x_function_in_section_p): Likewise.
1964 * config/darwin.c (machopic_select_section): Likewise.
1965 * config/arm/arm.c (arm_function_in_section_p): Likewise.
1966 * config/mips/mips.c (mips_function_rodata_section): Likewise.
1967 * config/mep/mep.c (mep_select_section): LIkewise.
1968 * config/i386/i386.c (x86_64_elf_unique_section): Likewise.
1970 2014-05-20 Eric Botcazou <ebotcazou@adacore.com>
1972 * tree-ssa-dom.c (hashable_expr_equal_p) <EXPR_CALL>: Also compare the
1973 EH region of calls to pure functions that can throw an exception.
1974 * tree-ssa-sccvn.c (vn_reference_eq): Remove duplicated test.
1975 (copy_reference_ops_from_call): Also copy the EH region of the call if
1976 it can throw an exception.
1978 2014-05-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1980 * simplify-rtx.c (simplify_binary_operation_1): Optimize case of
1981 nested VEC_SELECTs that are inverses of each other.
1983 2014-05-20 Richard Biener <rguenther@suse.de>
1985 * tree-ssa-sccvn.c (process_scc): Dump SCC here, when iterating,
1986 (extract_and_process_scc_for_name): not here.
1987 (cond_dom_walker::before_dom_children): Only process
1988 stmts that end the BB in interesting ways.
1989 (run_scc_vn): Mark param uses as visited.
1991 2014-05-20 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1993 * config/arm/arm.md (arith_shiftsi): Do not predicate for
1996 2014-05-20 Nick Clifton <nickc@redhat.com>
1998 * config/msp430/msp430.c (TARGET_GIMPLIFY_VA_ARG_EXPR): Define.
1999 (msp430_gimplify_va_arg_expr): New function.
2000 (msp430_print_operand): Handle (CONST (ZERO_EXTRACT)).
2002 * config/msp430/msp430.md (zero_extendpsisi2): Use + constraint on
2003 operand 0 in order to prevent confusion about the number of
2006 2014-05-20 Richard Biener <rguenther@suse.de>
2008 PR tree-optimization/61221
2009 * tree-ssa-pre.c (el_to_update): Remove.
2010 (eliminate_dom_walker::before_dom_children): Handle released
2011 VDEFs by value-numbering them to the associated VUSE. Update
2012 stmt immediately for substituted call address.
2013 (eliminate): Remove delayed stmt updating code.
2014 * tree-ssa-sccvn.c (vuse_ssa_val): New function valueizing
2015 possibly late re-numbered vuses.
2016 (vn_reference_lookup_2): Adjust.
2017 (vn_reference_lookup_pieces): Likewise.
2018 (vn_reference_lookup): Likewise.
2020 2014-05-20 Richard Biener <rguenther@suse.de>
2022 * config.gcc: Remove need_64bit_hwint.
2023 * configure.ac: Do not define NEED_64BIT_HOST_WIDE_INT.
2024 * hwint.h: Do not check NEED_64BIT_HOST_WIDE_INT but assume
2026 * config.in: Regenerate.
2027 * configure: Likewise.
2029 2014-05-19 David Wohlferd <dw@LimeGreenSocks.com>
2031 * doc/extend.texi: Create Label Attributes section,
2032 move all label attributes into it and reference it.
2034 2014-05-19 Richard Earnshaw <rearnsha@arm.com>
2036 * arm.c (thumb1_reorg): When scanning backwards skip anything
2037 that's not a proper insn.
2039 2014-05-19 Richard Biener <rguenther@suse.de>
2041 PR tree-optimization/61221
2042 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children):
2043 Do nothing for unreachable blocks.
2044 * tree-ssa-sccvn.c (cond_dom_walker::before_dom_children):
2045 Improve unreachability detection.
2047 2014-05-19 Richard Biener <rguenther@suse.de>
2049 PR tree-optimization/61209
2050 * tree-ssa-sccvn.c (visit_phi): Avoid setting expr to VN_TOP.
2052 2014-05-19 Nick Clifton <nickc@redhat.com>
2054 * except.c (init_eh): Fix computation of builtin setjmp buffer
2055 size to allow for targets where POINTER_SIZE > BITS_PER_WORD.
2057 2014-05-19 Richard Biener <rguenther@suse.de>
2059 PR tree-optimization/61184
2060 * tree-vrp.c (is_negative_overflow_infinity): Use
2061 TREE_OVERFLOW_P and do that check first.
2062 (is_positive_overflow_infinity): Likewise.
2063 (is_overflow_infinity): Likewise.
2064 (vrp_operand_equal_p): Properly treat operands with
2065 differing overflow as not equal.
2067 2014-05-19 Bernd Schmidt <bernds@codesourcery.com>
2069 * simplify-rtx.c (simplify_unary_operation_1): Use CONST_INT_P in
2070 shift simplification where it was intended.
2072 2014-05-19 Christian Bruel <christian.bruel@st.com>
2075 * config/sh/sh.md (movsf_ie): Unset fp_mode for fmov.
2077 2014-05-19 Richard Sandiford <r.sandiford@uk.ibm.com>
2080 * config/sparc/sparc.c (sparc_fold_builtin): Use widest_int rather
2083 2014-05-19 Richard Sandiford <rdsandiford@googlemail.com>
2085 * reg-notes.def (CROSSING_JUMP): Likewise.
2086 * rtl.h (rtx_def): Update comment for jump flag.
2087 (CROSSING_JUMP_P): Define.
2088 * cfgcleanup.c (try_forward_edges, try_optimize_cfg): Use it instead
2089 of a REG_CROSSING_JUMP note.
2090 * cfghooks.c (tidy_fallthru_edges): Likewise.
2091 * cfgrtl.c (fixup_partition_crossing, rtl_verify_edges): Likewise.
2092 * emit-rtl.c (try_split): Likewise.
2093 * haifa-sched.c (sched_create_recovery_edges): Likewise.
2094 * ifcvt.c (find_if_case_1, find_if_case_2): Likewise.
2095 * jump.c (redirect_jump_2): Likewise.
2096 * reorg.c (follow_jumps, fill_slots_from_thread): Likewise.
2097 (relax_delay_slots): Likewise.
2098 * config/arc/arc.md (jump_i, cbranchsi4_scratch, *bbit): Likewise.
2099 (bbit_di): Likewise.
2100 * config/arc/arc.c (arc_reorg, arc_can_follow_jump): Likewise.
2101 * config/sh/sh.md (jump_compact): Likewise.
2102 * bb-reorder.c (rotate_loop): Likewise.
2103 (pass_duplicate_computed_gotos::execute): Likewise.
2104 (add_reg_crossing_jump_notes): Rename to...
2105 (update_crossing_jump_flags): ...this.
2106 (pass_partition_blocks::execute): Update accordingly.
2108 2014-05-19 Richard Sandiford <rdsandiford@googlemail.com>
2110 * tree.h: Remove extraneous template <>.
2112 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
2114 * ipa.c (symtab_remove_unreachable_nodes): Remove
2115 symbol from comdat group if its body was eliminated.
2116 (comdat_can_be_unshared_p_1): Static symbols can always be privatized.
2117 * symtab.c (symtab_remove_from_same_comdat_group): Break out from ...
2118 (symtab_unregister_node): ... this one.
2119 (verify_symtab_base): More strict checking of comdats.
2120 * cgraph.h (symtab_remove_from_same_comdat_group): Declare.
2122 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
2124 * tree-pass.h (make_pass_ipa_comdats): New pass.
2125 * timevar.def (TV_IPA_COMDATS): New timevar.
2126 * passes.def (pass_ipa_comdats): Add.
2127 * Makefile.in (OBJS): Add ipa-comdats.o
2128 * ipa-comdats.c: New file.
2130 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
2132 * ipa.c (update_visibility_by_resolution_info): New function.
2133 (function_and_variable_visibility): Use it.
2135 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
2137 * cgraph.h (symtab_first_defined_symbol, symtab_next_defined_symbol):
2139 (FOR_EACH_DEFINED_SYMBOL): New macro.
2140 (varpool_first_static_initializer, varpool_next_static_initializer,
2141 varpool_first_defined_variable, varpool_next_defined_variable):
2143 (symtab_in_same_comdat_p): Correctly deal with inline functions.
2145 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
2147 * ggc-page.c (ggc_handle_finalizers): Add comment.
2149 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
2151 * ggc-common.c (ggc_internal_cleared_alloc): Adjust.
2152 * ggc-none.c (ggc_internal_alloc): Assert if a finalizer is passed.
2153 (ggc_internal_cleared_alloc): Likewise.
2154 * ggc-page.c (finalizer): New class.
2155 (vec_finalizer): Likewise.
2156 (globals::finalizers): New member.
2157 (globals::vec_finalizers): Likewise.
2158 (ggc_internal_alloc): Record the finalizer if any for the block being
2160 (ggc_handle_finalizers): New function.
2161 (ggc_collect): Call ggc_handle_finalizers.
2162 * ggc.h (ggc_internal_alloc): Add arguments to allow installing a
2164 (ggc_internal_cleared_alloc): Likewise.
2165 (finalize): New function.
2166 (need_finalization_p): Likewise.
2167 (ggc_alloc): Install the type's destructor as the finalizer if it
2169 (ggc_cleared_alloc): Likewise.
2170 (ggc_vec_alloc): Likewise.
2171 (ggc_cleared_vec_alloc): Likewise.
2173 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
2175 * ggc.h (ggc_alloc_cleared_simd_clone_stat): Remove function.
2177 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
2179 * alias.c (record_alias_subset): Adjust.
2180 * bitmap.c (bitmap_element_allocate): Likewise.
2181 (bitmap_gc_alloc_stat): Likewise.
2182 * cfg.c (init_flow): Likewise.
2183 (alloc_block): Likewise.
2184 (unchecked_make_edge): Likewise.
2185 * cfgloop.c (alloc_loop): Likewise.
2186 (flow_loops_find): Likewise.
2187 (rescan_loop_exit): Likewise.
2188 * cfgrtl.c (init_rtl_bb_info): Likewise.
2189 * cgraph.c (insert_new_cgraph_node_version): Likewise.
2190 (cgraph_allocate_node): Likewise.
2191 (cgraph_create_edge_1): Likewise.
2192 (cgraph_allocate_init_indirect_info): Likewise.
2193 * cgraphclones.c (cgraph_clone_edge): Likewise.
2194 * cgraphunit.c (add_asm_node): Likewise.
2195 (init_lowered_empty_function): Likewise.
2196 * config/aarch64/aarch64.c (aarch64_init_machine_status): Likewise.
2197 * config/alpha/alpha.c (alpha_init_machine_status): Likewise.
2198 (alpha_use_linkage): Likewise.
2199 * config/arc/arc.c (arc_init_machine_status): Likewise.
2200 * config/arm/arm.c (arm_init_machine_status): Likewise.
2201 * config/avr/avr.c (avr_init_machine_status): Likewise.
2202 * config/bfin/bfin.c (bfin_init_machine_status): Likewise.
2203 * config/c6x/c6x.c (c6x_init_machine_status): Likewise.
2204 * config/cris/cris.c (cris_init_machine_status): Likewise.
2205 * config/darwin.c (machopic_indirection_name): Likewise.
2206 (darwin_build_constant_cfstring): Likewise.
2207 (darwin_enter_string_into_cfstring_table): Likewise.
2208 * config/epiphany/epiphany.c (epiphany_init_machine_status): Likewise.
2209 * config/frv/frv.c (frv_init_machine_status): Likewise.
2210 * config/i386/i386.c (get_dllimport_decl): Likewise.
2211 (ix86_init_machine_status): Likewise.
2212 (assign_386_stack_local): Likewise.
2213 * config/i386/winnt.c (i386_pe_record_external_function): Likewise.
2214 (i386_pe_maybe_record_exported_symbol): Likewise.
2215 (i386_pe_record_stub): Likewise.
2216 * config/ia64/ia64.c (ia64_init_machine_status): Likewise.
2217 * config/iq2000/iq2000.c (iq2000_init_machine_status): Likewise.
2218 * config/m32c/m32c.c (m32c_init_machine_status): Likewise.
2219 (m32c_note_pragma_address): Likewise.
2220 * config/mep/mep.c (mep_init_machine_status): Likewise.
2221 (mep_note_pragma_flag): Likewise.
2222 * config/mips/mips.c (mflip_mips16_use_mips16_p): Likewise.
2223 (mips16_local_alias): Likewise.
2224 (mips_init_machine_status): Likewise.
2225 * config/mmix/mmix.c (mmix_init_machine_status): Likewise.
2226 * config/moxie/moxie.c (moxie_init_machine_status): Likewise.
2227 * config/msp430/msp430.c (msp430_init_machine_status): Likewise.
2228 * config/nds32/nds32.c (nds32_init_machine_status): Likewise.
2229 * config/nios2/nios2.c (nios2_init_machine_status): Likewise.
2230 * config/pa/pa.c (pa_init_machine_status): Likewise.
2231 (pa_get_deferred_plabel): Likewise.
2232 * config/rl78/rl78.c (rl78_init_machine_status): Likewise.
2233 * config/rs6000/rs6000.c (builtin_function_type): Likewise.
2234 (rs6000_init_machine_status): Likewise.
2235 (output_toc): Likewise.
2236 * config/s390/s390.c (s390_init_machine_status): Likewise.
2237 * config/score/score.c (score_output_external): Likewise.
2238 * config/sparc/sparc.c (sparc_init_machine_status): Likewise.
2239 * config/spu/spu.c (spu_init_machine_status): Likewise.
2240 * config/tilegx/tilegx.c (tilegx_init_machine_status): Likewise.
2241 * config/tilepro/tilepro.c (tilepro_init_machine_status): Likewise.
2242 * config/xtensa/xtensa.c (xtensa_init_machine_status): Likewise.
2243 * coverage.c (coverage_end_function): Likewise.
2244 * dbxout.c (dbxout_init): Likewise.
2245 * doc/gty.texi: Don't mention variable_size attribute.
2246 * dwarf2cfi.c (new_cfi): Adjust.
2247 (new_cfi_row): Likewise.
2248 (copy_cfi_row): Likewise.
2249 (create_cie_data): Likewise.
2250 * dwarf2out.c (dwarf2out_alloc_current_fde): Likewise.
2251 (new_loc_descr): Likewise.
2252 (find_AT_string_in_table): Likewise.
2253 (add_addr_table_entry): Likewise.
2254 (new_die): Likewise.
2255 (add_var_loc_to_decl): Likewise.
2256 (clone_die): Likewise.
2257 (clone_as_declaration): Likewise.
2258 (break_out_comdat_types): Likewise.
2259 (new_loc_list): Likewise.
2260 (add_loc_descr_to_each): Likewise.
2261 (add_location_or_const_value_attribute): Likewise.
2262 (add_linkage_name): Likewise.
2263 (lookup_filename): Likewise.
2264 (dwarf2out_var_location): Likewise.
2265 (new_line_info_table): Likewise.
2266 (dwarf2out_init): Likewise.
2267 (mem_loc_descriptor): Likewise.
2268 (loc_descriptor): Likewise.
2269 (add_const_value_attribute): Likewise.
2270 (tree_add_const_value_attribute): Likewise.
2271 (comp_dir_string): Likewise.
2272 (dwarf2out_vms_debug_main_pointer): Likewise.
2273 (string_cst_pool_decl): Likewise.
2274 * emit-rtl.c (set_mem_attrs): Likewise.
2275 (get_reg_attrs): Likewise.
2276 (start_sequence): Likewise.
2277 (init_emit): Likewise.
2278 (init_emit_regs): Likewise.
2279 * except.c (init_eh_for_function): Likewise.
2280 (gen_eh_region): Likewise.
2281 (gen_eh_region_catch): Likewise.
2282 (gen_eh_landing_pad): Likewise.
2283 (add_call_site): Likewise.
2284 * function.c (add_frame_space): Likewise.
2285 (insert_temp_slot_address): Likewise.
2286 (assign_stack_temp_for_type): Likewise.
2287 (get_hard_reg_initial_val): Likewise.
2288 (allocate_struct_function): Likewise.
2289 (prepare_function_start): Likewise.
2290 (types_used_by_var_decl_insert): Likewise.
2291 * gengtype.c (variable_size_p): Remove function.
2292 (enum alloc_quantity): Remove enum.
2293 (write_typed_alloc_def): Remove function.
2294 (write_typed_struct_alloc_def): Likewise.
2295 (write_typed_typedef_alloc_def): Likewise.
2296 (write_typed_alloc_defns): Likewise.
2298 * ggc-common.c (ggc_cleared_alloc_htab_ignore_args): Adjust.
2299 (ggc_cleared_alloc_ptr_array_two_args): Likewise.
2300 * ggc.h (ggc_alloc): new function.
2301 (ggc_cleared_alloc): Likewise.
2302 (ggc_vec_alloc): Template on type of vector element, and remove
2303 element size argument.
2304 (ggc_cleared_vec_alloc): Likewise.
2305 * gimple.c (gimple_build_omp_for): Adjust.
2306 (gimple_copy): Likewise.
2307 * ipa-cp.c (get_replacement_map): Likewise.
2308 (find_aggregate_values_for_callers_subset): Likewise.
2309 (known_aggs_to_agg_replacement_list): Likewise.
2310 * ipa-devirt.c (get_odr_type): Likewise.
2311 * ipa-prop.c (ipa_node_duplication_hook): Likewise.
2312 (read_agg_replacement_chain): Likewise.
2313 * loop-iv.c (get_simple_loop_desc): Likewise.
2314 * lto-cgraph.c (input_node_opt_summary): Likewise.
2315 * lto-section-in.c (lto_new_in_decl_state): Likewise.
2316 * lto-streamer-in.c (lto_input_eh_catch_list): Likewise.
2317 (input_eh_region): Likewise.
2318 (input_eh_lp): Likewise.
2319 (input_cfg): Likewise.
2320 * optabs.c (set_optab_libfunc): Likewise.
2321 (init_tree_optimization_optabs): Likewise.
2322 (set_conv_libfunc): Likewise.
2323 * passes.c (do_per_function_toporder): Likewise.
2324 * rtl.h: Don't use variable_size gty attribute.
2325 * sese.c (if_region_set_false_region): Adjust.
2326 * stringpool.c (gt_pch_save_stringpool): Likewise.
2327 * target-globals.c (save_target_globals): Likewise.
2328 * toplev.c (general_init): Likewise.
2329 * trans-mem.c (record_tm_replacement): Likewise.
2330 (split_bb_make_tm_edge): Likewise.
2331 * tree-cfg.c (move_sese_region_to_fn): Likewise.
2332 * tree-data-ref.h (lambda_vector_new): Likewise.
2333 * tree-eh.c (add_stmt_to_eh_lp_fn): Likewise.
2334 * tree-iterator.c (tsi_link_before): Likewise.
2335 (tsi_link_after): Likewise.
2336 * tree-scalar-evolution.c (new_scev_info_str): Likewise.
2337 * tree-ssa-loop-niter.c (record_estimate): Likewise.
2338 * tree-ssa-operands.c (ssa_operand_alloc): Likewise.
2339 * tree-ssa-operands.h: Don't use variable_size gty attribute.
2340 * tree-ssa.c (init_tree_ssa): Adjust.
2341 * tree-ssanames.c (set_range_info): Likewise.
2342 (get_ptr_info): Likewise.
2343 (duplicate_ssa_name_ptr_info): Likewise.
2344 (duplicate_ssa_name_range_info): Likewise.
2345 * tree-streamer-in.c (unpack_ts_real_cst_value_fields): Likewise.
2346 (unpack_ts_fixed_cst_value_fields): Likewise.
2347 * tree.c (build_fixed): Likewise.
2348 (build_real): Likewise.
2349 (build_string): Likewise.
2350 (decl_priority_info): Likewise.
2351 (decl_debug_expr_insert): Likewise.
2352 (decl_value_expr_insert): Likewise.
2353 (decl_debug_args_insert): Likewise.
2354 (type_hash_add): Likewise.
2355 (build_omp_clause): Likewise.
2356 * ubsan.c (decl_for_type_insert): Likewise.
2357 * varasm.c (get_unnamed_section): Likewise.
2358 (get_noswitch_section): Likewise.
2359 (get_section): Likewise.
2360 (get_block_for_section): Likewise.
2361 (create_block_symbol): Likewise.
2362 (build_constant_desc): Likewise.
2363 (create_constant_pool): Likewise.
2364 (force_const_mem): Likewise.
2365 (record_tm_clone_pair): Likewise.
2366 * varpool.c (varpool_create_empty_node): Likewise.
2368 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
2370 * dwarf2out.c (tree_add_const_value_attribute): Call
2371 ggc_internal_cleared_alloc instead of ggc_alloc_cleared_atomic.
2372 * gengtype.c (write_typed_alloc_def): Call ggc_internal_<x>alloc
2373 instead of ggc_internal_<x>alloc_stat.
2374 * ggc-common.c (ggc_internal_cleared_alloc): Drop _stat suffix.
2375 (ggc_realloc): Likewise.
2376 * ggc-none.c (ggc_internal_alloc): Likewise.
2377 (ggc_internal_cleared_alloc): Likewise.
2378 * ggc-page.c: Likewise.
2379 * ggc.h (ggc_internal_alloc_stat): Likewise.
2380 (ggc_internal_alloc): Remove macro.
2381 (ggc_internal_cleared_alloc_stat): Drop _stat suffix.
2382 (ggc_internal_cleared_alloc): Remove macro.
2383 (GGC_RESIZEVEC): Adjust.
2384 (ggc_resizevar): Remove macro.
2385 (ggc_internal_vec_alloc_stat): Drop _stat suffix.
2386 (ggc_internal_cleared_vec_alloc_stat): Likewise.
2387 (ggc_internal_vec_cleared_alloc): Remove macro.
2388 (ggc_alloc_atomic_stat): Drop _stat suffix.
2389 (ggc_alloc_atomic): Remove macro.
2390 (ggc_alloc_cleared_atomic): Remove macro.
2391 (ggc_alloc_string_stat): Drop _stat suffix.
2392 (ggc_alloc_string): Remove macro.
2393 (ggc_alloc_rtx_def_stat): Adjust.
2394 (ggc_alloc_tree_node_stat): Likewise.
2395 (ggc_alloc_cleared_tree_node_stat): Likewise.
2396 (ggc_alloc_cleared_gimple_statement_stat): Likewise.
2397 (ggc_alloc_cleared_simd_clone_stat): Likewise.
2398 * gimple.c (gimple_build_omp_for): Likewise.
2399 (gimple_copy): Likewise.
2400 * stringpool.c (ggc_alloc_string_stat): Drop _stat suffix.
2401 * toplev.c (realloc_for_line_map): Adjust.
2402 * tree-data-ref.h (lambda_vector_new): Likewise.
2403 * tree-phinodes.c (allocate_phi_node): Likewise.
2404 * tree.c (grow_tree_vec_stat): Likewise.
2405 * vec.h (va_gc::reserve): Adjust.
2407 2014-05-17 Ajit Agarwal <ajitkum@xilinx.com>
2409 * config/microblaze/microblaze.c (break_handler): New Declaration.
2410 (microblaze_break_function_p,microblaze_is_break_handler): New.
2411 (compute_frame_size): Use microblaze_break_function_p.
2412 Add the test of break_handler.
2413 (microblaze_function_prologue) : Add the test of variable
2414 break_handler. Check the fnname by BREAK_HANDLER_NAME.
2415 (microblaze_function_epilogue) : Add the test of break_handler.
2416 (microblaze_globalize_label) : Add the test of break_handler.
2417 Check the name by BREAK_HANDLER_NAME.
2419 * config/microblaze/microblaze.h (BREAK_HANDLER_NAME): New macro
2421 * config/microblaze/microblaze.md (*<optab>,<optab>_internal): Add
2422 microblaze_is_break_handler test.
2423 (call_internal1,call_value_intern): Use microblaze_break_function_p.
2424 Use SYMBOL_REF_DECL.
2426 * config/microblaze/microblaze-protos.h
2427 (microblaze_break_function_p,microblaze_is_break_handler):
2430 * doc/extend.texi (MicroBlaze break_handler Functions): Document
2431 new MicroBlaze break_handler functions.
2433 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
2435 * doc/extend.texi (Size of an asm): Move node text according
2436 to its @menu entry position.
2438 2014-05-17 Marc Glisse <marc.glisse@inria.fr>
2440 PR tree-optimization/61140
2441 PR tree-optimization/61150
2442 PR tree-optimization/61197
2443 * tree-ssa-phiopt.c (value_replacement): Punt on multiple phis.
2445 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
2447 * doc/invoke.texi (free): Mention Alpha. Also enabled at -Os.
2449 2014-05-17 Richard Sandiford <r.sandiford@uk.ibm.com>
2451 * wide-int.cc: Only include longlong.h if W_TYPE_SIZE==32 or
2452 __SIZEOF_INT128__ is defined.
2454 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
2456 * config/rs6000/rs6000.c (rs6000_real_tls_symbol_ref_p): New function.
2457 (rs6000_delegitimize_address): Use it.
2459 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
2461 * emit-rtl.h (replace_equiv_address, replace_equiv_address_nv): Add an
2462 inplace argument. Store the new address in the original MEM when true.
2463 * emit-rtl.c (change_address_1): Likewise.
2464 (adjust_address_1, adjust_automodify_address_1, offset_address):
2466 * rtl.h (plus_constant): Add an inplace argument.
2467 * explow.c (plus_constant): Likewise. Try to reuse the original PLUS
2468 when true. Avoid generating (plus X (const_int 0)).
2469 * function.c (instantiate_virtual_regs_in_rtx): Adjust the PLUS
2470 in-place. Pass true to plus_constant.
2471 (instantiate_virtual_regs_in_insn): Pass true to replace_equiv_address.
2473 2014-05-16 Dehao Chen <dehao@google.com>
2475 * tree-cfg.c (gimple_merge_blocks): Updates bb count with max count.
2477 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
2480 * config/sh/predicates.md (negt_reg_shl31_operand): Match additional
2482 * config/sh/sh.md (*negt_msb): Merge SH2A and non-SH2A variants.
2484 2014-05-16 Dehao Chen <dehao@google.com>
2486 * ira-int.h (REG_FREQ_FROM_EDGE_FREQ): Use
2487 optimize_function_for_size_p.
2488 * regs.h (REG_FREQ_FROM_BB): Likewise.
2490 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
2493 * config/sh/sh.c (sh_eval_treg_value): Handle t_reg_operand and
2494 negt_reg_operand cases.
2495 * config/sh/sh.md (*cset_zero): Likewise by using cbranch_treg_value
2497 * config/sh/predicates.md (cbranch_treg_value): Simplify.
2499 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
2501 * config/sh/sh.c (sh_option_override): Set branch cost to 2 for all
2504 2014-05-16 David Malcolm <dmalcolm@redhat.com>
2507 2014-04-29 David Malcolm <dmalcolm@redhat.com>
2509 * tree-cfg.c (dump_function_to_file): Dump the return type of
2510 functions, in a line to itself before the function body, mimicking
2511 the layout of a C function.
2513 2014-05-16 Dehao Chen <dehao@google.com>
2515 * cfghooks.c (make_forwarder_block): Use direct computation to
2516 get fall-through edge's count and frequency.
2518 2014-05-16 Benno Schulenberg <bensberg@justemail.net>
2520 * config/arc/arc.c (arc_init): Fix typo in error message.
2521 * config/i386/i386.c (ix86_expand_builtin): Likewise.
2522 (split_stack_prologue_scratch_regno): Likewise.
2523 * fortran/check.c (gfc_check_fn_rc2008): Remove duplicate
2524 word from error message.
2526 2014-05-16 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
2528 * ira-costs.c: Fix typo in comment.
2530 2014-05-16 David Wohlferd <dw@LimeGreenSocks.com>
2532 * doc/extend.texi: (Visibility Pragmas) Fix misplaced @xref
2534 2014-05-16 Jan Hubicka <hubicka@ucw.cz>
2536 * varpool.c (dump_varpool_node): Dump write-only flag.
2537 * lto-cgraph.c (lto_output_varpool_node, input_varpool_node): Stream
2539 * tree-cfg.c (execute_fixup_cfg): Remove statements setting
2540 write-only variables.
2541 * ipa.c (process_references): New function.
2542 (set_readonly_bit): New function.
2543 (set_writeonly_bit): New function.
2544 (clear_addressable_bit): New function.
2545 (ipa_discover_readonly_nonaddressable_var): Mark write only variables;
2546 fix handling of aliases.
2547 * cgraph.h (struct varpool_node): Add writeonly flag.
2549 2014-05-16 Vladimir Makarov <vmakarov@redhat.com>
2551 PR rtl-optimization/60969
2552 * ira-costs.c (record_reg_classes): Allow only memory for pseudo.
2553 Calculate costs for this case.
2555 2014-05-16 Eric Botcazou <ebotcazou@adacore.com>
2557 * fold-const (fold_unary_loc) <NON_LVALUE_EXPR>: New case.
2558 <CASE_CONVERT>: Pass arg0 instead of op0 to fold_convert_const.
2560 2014-05-16 Richard Biener <rguenther@suse.de>
2562 PR tree-optimization/61194
2563 * tree-vect-patterns.c (adjust_bool_pattern): Also handle
2564 bool patterns ending in a COND_EXPR.
2566 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2568 * config/aarch64/aarch64.c (aarch64_rtx_mult_cost): Fix FNMUL case.
2570 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2572 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle the case
2573 where we were unable to cost an RTX.
2575 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2577 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost SYMBOL_REF,
2580 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2581 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2583 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost TRUNCATE.
2585 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2586 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2588 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost FMA,
2589 FLOAT_EXTEND, FLOAT_TRUNCATE, ABS, SMAX, and SMIN.
2591 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2592 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2594 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost comparison
2597 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2598 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2600 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
2603 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2604 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2606 * config/aarch64/aarch64.c (aarch64_rtx_arith_op_extract_p): New.
2607 (aarch64_rtx_costs): Improve costs for SIGN/ZERO_EXTRACT.
2609 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2610 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2612 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
2615 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2616 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2618 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost
2619 ZERO_EXTEND and SIGN_EXTEND better.
2621 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2622 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2624 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve cost for
2627 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2628 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2630 * config/aarch64/aarch64.c (aarch64_rtx_costs): Use address
2631 costs when costing loads and stores to memory.
2633 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2634 Philip Tomsich <philipp.tomsich@theobroma-systems.com>
2636 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costing
2639 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2641 * config/aarch64/aarch64.c (aarch64_rtx_costs): Set default costs.
2643 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2644 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2646 * config/aarch64/aarch64.c (aarch64_strip_shift_or_extend): Rename
2648 (aarch64_strip_extend): ...this, don't strip shifts, check RTX is
2650 (aarch64_rtx_mult_cost): New.
2651 (aarch64_rtx_costs): Use it, refactor as appropriate.
2653 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2654 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2656 * config/aarch64/aarch64.c (aarch64_build_constant): Conditionally
2657 emit instructions, return number of instructions which would
2659 (aarch64_add_constant): Update call to aarch64_build_constant.
2660 (aarch64_output_mi_thunk): Likewise.
2661 (aarch64_rtx_costs): Estimate cost of a CONST_INT, cost of
2664 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2666 * config/aarch64/aarch64.c (aarch64_rtx_costs_wrapper): New.
2667 (TARGET_RTX_COSTS): Call it.
2669 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2671 * config/aarch64/aarch64.c (cortexa57_addrcost_table): New.
2672 (cortexa57_vector_cost): Likewise.
2673 (cortexa57_tunings): Use them.
2675 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2677 * config/aarch64/aarch64-protos.h (scale_addr_mode_cost): New.
2678 (cpu_addrcost_table): Use it.
2679 * config/aarch64/aarch64.c (generic_addrcost_table): Initialize it.
2680 (aarch64_address_cost): Rewrite using aarch64_classify_address,
2683 2014-05-16 Richard Biener <rguenther@suse.de>
2685 * tree-ssa-sccvn.c: Include tree-cfg.h and domwalk.h.
2686 (set_ssa_val_to): Handle unexpected sets to VN_TOP.
2687 (visit_phi): Ignore edges marked as not executable.
2688 (class cond_dom_walker): New.
2689 (cond_dom_walker::before_dom_children): Value-number
2690 control statements and mark successor edges as not
2691 executable if possible.
2692 (run_scc_vn): First walk all control statements in
2693 dominator order, marking edges as not executable.
2694 * tree-inline.c (copy_edges_for_bb): Be not confused
2695 about random edge flags.
2697 2014-05-16 Richard Biener <rguenther@suse.de>
2699 * tree-ssa-sccvn.c (visit_use): Also constant-fold calls.
2701 2014-05-15 Peter Bergner <bergner@vnet.ibm.com>
2704 * config/rs6000/htmxlintrin.h (_HTM_TBEGIN_STARTED): New define.
2705 (__TM_simple_begin): Use it.
2706 (__TM_begin): Likewise.
2708 2014-05-15 Martin Jambor <mjambor@suse.cz>
2711 * ipa-prop.c (update_indirect_edges_after_inlining): Check
2712 type_preserved flag when the indirect edge is polymorphic.
2714 2014-05-15 Martin Jambor <mjambor@suse.cz>
2716 PR tree-optimization/61090
2717 * tree-sra.c (sra_modify_expr): Pass the current gsi to
2718 build_ref_for_model.
2720 2014-05-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2722 * config/arm/arm.c (arm_option_override): Use the SCHED_PRESSURE_MODEL
2723 enum name for PARAM_SCHED_PRESSURE_ALGORITHM.
2725 2014-05-15 Jakub Jelinek <jakub@redhat.com>
2727 PR tree-optimization/61158
2728 * fold-const.c (fold_binary_loc): If X is zero-extended and
2729 shiftc >= prec, make sure zerobits is all ones instead of
2730 invoking undefined behavior.
2732 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2734 * regcprop.h: New file.
2735 * regcprop.c (skip_debug_insn_p): New decl.
2736 (replace_oldest_value_reg): Check skip_debug_insn_p.
2737 (copyprop_hardreg_forward_bb_without_debug_insn): New function.
2738 * shrink-wrap.c: Include regcprop.h.
2739 (prepare_shrink_wrap): Call
2740 copyprop_hardreg_forward_bb_without_debug_insn.
2742 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2744 * shrink-wrap.h: Update comment.
2745 * shrink-wrap.c: Update comment.
2746 (next_block_for_reg): Rename to live_edge_for_reg.
2747 (live_edge_for_reg): Allow live_edge->dest has two predecessors.
2748 (move_insn_for_shrink_wrap): Split live_edge.
2749 (prepre_shrink_wrap): One more parameter for move_insn_for_shrink_wrap.
2751 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
2753 * config/sparc/sparc-protos.h (sparc_absnegfloat_split_legitimate):
2755 * config/sparc/sparc.c (sparc_absnegfloat_split_legitimate): Likewise.
2756 * config/sparc/sparc.md (fptype_ut699): New attribute.
2757 (in_branch_delay): Return false if -mfix-ut699 is specified and
2758 fptype_ut699 is set to single.
2759 (truncdfsf2): Add fptype_ut699 attribute.
2760 (fix_truncdfsi2): Likewise.
2761 (floatsisf2): Change fptype attribute.
2762 (fix_truncsfsi2): Likewise.
2763 (negtf2_notv9): Delete.
2764 (negtf2_v9): Likewise.
2765 (negtf2_hq): New instruction.
2766 (negtf2): New instruction and splitter.
2767 (negdf2_notv9): Rewrite.
2768 (abstf2_notv9): Delete.
2769 (abstf2_hq_v9): Likewise.
2770 (abstf2_v9): Likewise.
2771 (abstf2_hq): New instruction.
2772 (abstf2): New instruction and splitter.
2773 (absdf2_notv9): Rewrite.
2775 2014-05-14 Cary Coutant <ccoutant@google.com>
2778 * opts.c (common_handle_option): Don't special-case "-g".
2779 (set_debug_level): Default to at least level 2 with "-g".
2781 2014-05-14 DJ Delorie <dj@redhat.com>
2783 * config/msp430/msp430.c (msp430_builtin): Add
2784 MSP430_BUILTIN_DELAY_CYCLES.
2785 (msp430_init_builtins): Register void __delay_cycles(long long).
2786 (msp430_builtin_decl): Add it.
2787 (cg_magic_constant): New.
2788 (msp430_expand_delay_cycles): New.
2789 (msp430_expand_builtin): Call it.
2790 (msp430_print_operand_raw): Change integer printing from "int" to
2792 * config/msp430/msp430.md (define_constants): Add delay_cycles tags.
2793 (delay_cycles_start): New.
2794 (delay_cycles_end): New.
2795 (delay_cycles_32): New.
2796 (delay_cycles_32x): New.
2797 (delay_cycles_16): New.
2798 (delay_cycles_16x): New.
2799 (delay_cycles_2): New.
2800 (delay_cycles_1): New.
2801 * doc/extend.texi: Document __delay_cycles().
2803 2014-05-14 Sandra Loosemore <sandra@codesourcery.com>
2805 * config/nios2/nios2.md (nios2_cbranch): Fix paste-o in
2806 length attribute computation.
2808 2014-05-14 Richard Sandiford <rdsandiford@googlemail.com>
2811 * print-rtl.c (print_rtx): Suppress uids if flag_dump_unnumbered.
2813 2014-05-14 Richard Sandiford <r.sandiford@uk.ibm.com>
2816 * config/sparc/sparc.md: Fix types of low and high in DI constant
2817 splitter. Use gen_int_mode in some other splitters.
2819 2014-05-14 Martin Jambor <mjambor@suse.cz>
2822 * ipa-prop.c (ipa_modify_formal_parameters): Reset DECL_LANG_SPECIFIC.
2824 2014-05-14 James Norris <jnorris@codesourcery.com>
2826 * omp-low.c (expand_parallel_call): Remove shadow variable.
2827 (expand_omp_taskreg): Likewise.
2829 2014-05-14 Ilya Tocar <ilya.tocar@intel.com>
2831 * common/config/i386/i386-common.c
2832 (OPTION_MASK_ISA_CLFLUSHOPT_SET): Define.
2833 (OPTION_MASK_ISA_XSAVES_SET): Ditto.
2834 (OPTION_MASK_ISA_XSAVEC_SET): Ditto.
2835 (OPTION_MASK_ISA_CLFLUSHOPT_UNSET): Ditto.
2836 (OPTION_MASK_ISA_XSAVES_UNSET): Ditto.
2837 (OPTION_MASK_ISA_XSAVEC_UNSET): Ditto.
2838 (ix86_handle_option): Handle OPT_mxsavec, OPT_mxsaves, OPT_mclflushopt.
2839 * config.gcc (i[34567]86-*-*): Add clflushoptintrin.h,
2840 xsavecintrin.h, xsavesintrin.h.
2841 (x86_64-*-*): Ditto.
2842 * config/i386/clflushoptintrin.h: New.
2843 * config/i386/xsavecintrin.h: Ditto.
2844 * config/i386/xsavesintrin.h: Ditto.
2845 * config/i386/cpuid.h (bit_CLFLUSHOPT): Define.
2846 (bit_XSAVES): Ditto.
2847 (bit_XSAVES): Ditto.
2848 * config/i386/driver-i386.c (host_detect_local_cpu): Handle
2849 -mclflushopt, -mxsavec, -mxsaves, -mno-xsaves, -mno-xsavec,
2851 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
2852 OPTION_MASK_ISA_CLFLUSHOPT, OPTION_MASK_ISA_XSAVEC,
2853 OPTION_MASK_ISA_XSAVES.
2854 * config/i386/i386.c (ix86_target_string): Handle -mclflushopt,
2856 (PTA_CLFLUSHOPT) Define.
2857 (PTA_XSAVEC): Ditto.
2858 (PTA_XSAVES): Ditto.
2859 (ix86_option_override_internal): Handle new options.
2860 (ix86_valid_target_attribute_inner_p): Ditto.
2861 (ix86_builtins): Add IX86_BUILTIN_XSAVEC, IX86_BUILTIN_XSAVEC64,
2862 IX86_BUILTIN_XSAVES, IX86_BUILTIN_XRSTORS, IX86_BUILTIN_XSAVES64,
2863 IX86_BUILTIN_XRSTORS64, IX86_BUILTIN_CLFLUSHOPT.
2864 (bdesc_special_args): Add __builtin_ia32_xsaves,
2865 __builtin_ia32_xrstors, __builtin_ia32_xsavec, __builtin_ia32_xsaves64,
2866 __builtin_ia32_xrstors64, __builtin_ia32_xsavec64.
2867 (ix86_init_mmx_sse_builtins): Add __builtin_ia32_clflushopt.
2868 (ix86_expand_builtin): Handle new builtins.
2869 * config/i386/i386.h (TARGET_CLFLUSHOPT) Define.
2870 (TARGET_CLFLUSHOPT_P): Ditto.
2871 (TARGET_XSAVEC): Ditto.
2872 (TARGET_XSAVEC_P): Ditto.
2873 (TARGET_XSAVES): Ditto.
2874 (TARGET_XSAVES_P): Ditto.
2875 * config/i386/i386.md (ANY_XSAVE): Add UNSPECV_XSAVEC, UNSPECV_XSAVES.
2876 (ANY_XSAVE64)" Add UNSPECV_XSAVEC64, UNSPECV_XSAVES64.
2877 (attr xsave): Add xsavec, xsavec64, xsaves, xsaves64.
2879 (ANY_XRSTOR64): Ditto.
2881 (xrstor): Change into <xrstor>.
2882 (xrstor_rex64): Change into <xrstor>_rex64.
2883 (xrstor64): Change into <xrstor>64
2885 * config/i386/i386.opt (mclflushopt): New.
2888 * config/i386/x86intrin.h: Add clflushoptintrin.h, xsavesintrin.h,
2890 * doc/invoke.texi: Document new options.
2892 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
2894 PR rtl-optimization/60866
2895 * sel-sched-ir (sel_init_new_insn): New parameter old_seqno.
2896 Default it to -1. Pass it down to init_simplejump_data.
2897 (init_simplejump_data): New parameter old_seqno. Pass it down
2898 to get_seqno_for_a_jump.
2899 (get_seqno_for_a_jump): New parameter old_seqno. Use it for
2900 initializing new jump seqno as a last resort. Add comment.
2901 (sel_redirect_edge_and_branch): Save old seqno of the conditional
2902 jump and pass it down to sel_init_new_insn.
2903 (sel_redirect_edge_and_branch_force): Likewise.
2905 2014-05-14 Georg-Johann Lay <avr@gjlay.de>
2907 * config/avr/avr.h (REG_CLASS_CONTENTS): Use unsigned suffix for
2908 shifted values to avoid build warning.
2910 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
2912 * cfgcleanup.c (try_forward_edges): Use location_t for locations.
2913 * cfgrtl.c (rtl_merge_blocks): Fix comment.
2914 (cfg_layout_merge_blocks): Likewise.
2915 * except.c (emit_to_new_bb_before): Remove prev_bb local variable.
2917 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
2919 PR rtl-optimization/60901
2920 * config/i386/i386.c (ix86_dependencies_evaluation_hook): Check that
2921 bb predecessor belongs to the same scheduling region. Adjust comment.
2923 2014-05-13 Peter Bergner <bergner@vnet.ibm.com>
2925 * doc/sourcebuild.texi: (dfp_hw): Document.
2926 (p8vector_hw): Likewise.
2927 (powerpc_eabi_ok): Likewise.
2928 (powerpc_elfv2): Likewise.
2929 (powerpc_htm_ok): Likewise.
2930 (ppc_recip_hw): Likewise.
2933 2014-05-13 Cary Coutant <ccoutant@google.com>
2935 * opts.c (finish_options): Use -ggnu-pubnames with -gsplit-dwarf.
2937 2014-05-13 David Malcolm <dmalcolm@redhat.com>
2939 * gengtype-parse.c (require3): Eliminate in favor of...
2941 (require_template_declaration): Update to support optional single *
2944 * gengtype.c (get_ultimate_base_class): Add a non-const overload.
2945 (create_user_defined_type): Handle a single level of explicit
2946 pointerness within template arguments.
2947 (struct write_types_data): Add field "kind".
2948 (filter_type_name): Handle "*" character.
2949 (write_user_func_for_structure_ptr): Require a write_types_data
2950 rather than just a prefix string, so that we can look up the kind
2951 of the wtd and use it as an index into wrote_user_func_for_ptr,
2952 ensuring that such functions are written at most once. Support
2953 subclasses by invoking the marking function of the ultimate base class.
2954 (write_user_func_for_structure_body): Require a write_types_data
2955 rather than just a prefix string, so that we can pass this to
2956 write_user_func_for_structure_ptr.
2957 (write_func_for_structure): Likewise.
2958 (ggc_wtd): Add initializer of new "kind" field.
2959 (pch_wtd): Likewise.
2961 * gengtype.h (enum write_types_kinds): New.
2962 (struct type): Add field wrote_user_func_for_ptr to the "s"
2965 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
2967 * fold-const.c (optimize_bit_field_compare): Use wi:: operations
2968 instead of const_binop.
2969 (fold_binary_loc): Likewise.
2971 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
2973 * tree-dfa.h (get_addr_base_and_unit_offset_1): Update array index
2974 calculation to match get_ref_base_and_extent.
2976 2014-05-13 Catherine Moore <clm@codesourcery.com>
2977 Sandra Loosemore <sandra@codesourcery.com>
2979 * configure.ac: Fix assembly for explicit JALR relocation check.
2980 * configure: Regenerate.
2982 2014-05-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2984 * config/arm/arm.c (neon_itype): Remove NEON_RESULTPAIR.
2985 (arm_init_neon_builtins): Remove handling of NEON_RESULTPAIR.
2986 Remove associated type declarations and initialisations.
2987 (arm_expand_neon_builtin): Likewise.
2988 (neon_emit_pair_result_insn): Delete.
2989 * config/arm/arm_neon_builtins (vtrn, vzip, vuzp): Delete.
2990 * config/arm/neon.md (neon_vtrn<mode>): Delete.
2991 (neon_vzip<mode>): Likewise.
2992 (neon_vuzp<mode>): Likewise.
2994 2014-05-13 Richard Biener <rguenther@suse.de>
2997 * tree-inline.c (remap_gimple_stmt): Clear tail call flag,
2998 it needs revisiting whether the call still may be tail-called.
3000 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
3002 * rtl.def (SYMBOL_REF): Remove middle "0" field.
3003 * rtl.h (block_symbol): Reduce number of fields to 2.
3004 (rtx_def): Add u2.symbol_ref_flags.
3005 (SYMBOL_REF_FLAGS): Use it.
3006 (SYMBOL_REF_DATA, SET_SYMBOL_REF_DECL, SYMBOL_REF_DECL)
3007 (SET_SYMBOL_REF_CONSTANT, SYMBOL_REF_CONSTANT): Lower index.
3008 * gengtype.c (adjust_field_rtx_def): Remove SYMBOL_REF_FLAGS handling.
3009 Lower index of SYMBOL_REF_DATA.
3010 * print-rtl.c (print_rtx): Lower index for SYMBOL_REF_DATA.
3011 Print SYMBOL_REF_FLAGS at the same time.
3012 * genattrtab.c (attr_rtx_1): Only initialize 1 "0" SYMBOL_REF field.
3014 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
3016 * rtl.def (VAR_LOCATION): Remove "i" field.
3017 * rtl.h (rtx_def): Add u2.var_location_status.
3018 (PAT_VAR_LOCATION_STATUS): Use it.
3019 (gen_rtx_VAR_LOCATION): Declare.
3020 * gengenrtl.c (excluded_rtx): Add VAR_LOCATION.
3021 * emit-rtl.c (gen_rtx_VAR_LOCATION): New function.
3022 * var-tracking.c (emit_note_insn_var_location): Remove casts.
3024 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
3026 * rtl.def (scratch): Fix outdated comment and remove "0" field.
3027 * gengtype.c (adjust_field_rtx_def): Update accordingly.
3029 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
3031 * rtl.def (DEBUG_INSN, INSN, JUMP_INSN, CALL_INSN, JUMP_TABLE_DATA)
3032 (BARRIER, CODE_LABEL, NOTE): Remove first "i" field.
3033 * rtl.h (rtx_def): Add insn_uid to u2 field.
3034 (RTX_FLAG_CHECK8): Delete in favor of...
3035 (RTL_INSN_CHAIN_FLAG_CHECK): ...this new macro.
3036 (INSN_DELETED_P): Update accordingly.
3037 (INSN_UID): Use u2.insn_uid.
3038 (INSN_CHAIN_CODE_P): Define.
3039 (PREV_INSN, NEXT_INSN, BLOCK_FOR_INSN, PATTERN, INSN_LOCATION)
3040 (INSN_CODE, REG_NOTES, CALL_INSN_FUNCTION_USAGE, CODE_LABEL_NUMBER)
3041 (NOTE_DATA, NOTE_DELETED_LABEL_NAME, NOTE_BLOCK, NOTE_EH_HANDLER)
3042 (NOTE_BASIC_BLOCK, NOTE_VAR_LOCATION, NOTE_CFI, NOTE_LABEL_NUMBER)
3043 (NOTE_KIND, LABEL_NAME, LABEL_NUSES, JUMP_LABEL, LABEL_REFS): Lower
3044 indices accordingly.
3045 * print-rtl.c (print_rtx): Print INSN_UIDs before the main loop.
3046 Update indices for insn-chain rtxes.
3047 * gengtype.c (gen_rtx_next): Adjust test for insn-chain rtxes.
3048 (adjust_field_rtx_def): Lower '0' indices for all insn-chain rtxes.
3049 * emit-rtl.c (gen_label_rtx): Update gen_rtx_LABEL call.
3050 * caller-save.c (init_caller_save): Update gen_rtx_INSN calls.
3051 * combine.c (try_combine): Likewise.
3052 * ira.c (setup_prohibited_mode_move_regs): Likewise.
3054 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
3056 * rtl.def (REG): Remove middle field.
3057 * rtl.h (rtx_def): Add orignal_regno to u2.
3058 (ORIGINAL_REGNO): Use it instead of field 1.
3059 (REG_ATTRS): Lower field index accordingly.
3060 * gengtype.c (adjust_field_rtx_def): Remove handling of
3061 ORIGINAL_REGNO. Move REG_ATTRS index down.
3062 * print-rtl.c (print_rtx): Move ORIGINAL_REGNO handling to the
3063 code that prints the REGNO.
3065 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
3067 * print-rtl.c (print_rtx): Guard whole '0' block with ifndef
3070 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
3072 * rtl.h (rtx_def): Mark u2 as GTY ((skip)).
3074 2014-05-13 Bin Cheng <bin.cheng@arm.com>
3076 * tree-ssa-loop-ivopts.c (contain_complex_addr_expr): New.
3077 (alloc_iv): Lower base expressions containing ADDR_EXPR.
3079 2014-05-13 Ian Bolton <ian.bolton@arm.com>
3081 * config/aarch64/aarch64-protos.h
3082 (aarch64_hard_regno_caller_save_mode): New prototype.
3083 * config/aarch64/aarch64.c (aarch64_hard_regno_caller_save_mode):
3085 * config/aarch64/aarch64.h (HARD_REGNO_CALLER_SAVE_MODE): New macro.
3087 2014-05-13 Christian Bruel <christian.bruel@st.com>
3089 * target.def (mode_switching): New hook vector.
3090 (mode_emit, mode_needed, mode_after, mode_entry): New hooks.
3091 (mode_exit, modepriority_to_mode): Likewise.
3092 * mode-switching.c (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Hookify.
3093 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
3094 * target.h: Include tm.h and hard-reg-set.h.
3095 * doc/tm.texi.in (EMIT_MODE_SET, MODE_NEEDED, MODE_AFTER, MODE_ENTRY)
3096 (MODE_EXIT, MODE_PRIORITY_TO_MODE): Delete and hookify.
3097 * doc/tm.texi Regenerate.
3098 * config/sh/sh.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
3099 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
3100 * config/sh/sh.c (sh_emit_mode_set, sh_mode_priority): Hookify.
3101 (sh_mode_needed, sh_mode_after, sh_mode_entry, sh_mode_exit): Likewise.
3102 * config/i386/i386.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
3103 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
3104 * config/i386/i386-protos.h (ix86_mode_needed, ix86_mode_after)
3105 (ix86_mode_entrym, ix86_emit_mode_set): Remove external declaration.
3106 * config/i386/i386.c (ix86_mode_needed, ix86_mode_after,
3107 (ix86_mode_exit, ix86_mode_entry, ix86_mode_priority)
3108 (ix86_emit_mode_set): Hookify.
3109 * config/epiphany/epiphany.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY):
3111 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
3112 * config/epiphany/epiphany-protos.h (epiphany_mode_needed)
3113 (emit_set_fp_mode, epiphany_mode_entry_exit, epiphany_mode_after)
3114 (epiphany_mode_priority_to_mode): Remove declaration.
3115 * config/epiphany/epiphany.c (emit_set_fp_mode): Hookify.
3116 (epiphany_mode_needed, epiphany_mode_priority_to_mode): Likewise.
3117 (epiphany_mode_entry, epiphany_mode_exit, epiphany_mode_after):
3119 (epiphany_mode_priority_to_mode): Change priority type. Hookify.
3120 (epiphany_mode_needed, epiphany_mode_entry_exit): Hookify.
3121 (epiphany_mode_after, epiphany_mode_entry, emit_set_fp_mode): Hookify.
3123 2014-05-13 Jakub Jelinek <jakub@redhat.com>
3126 * config/i386/i386.c (ix86_expand_set_or_movmem): If count_exp
3127 is const0_rtx, return immediately. Don't test count == 0 when
3130 2014-05-13 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3132 * Makefile.in: add shrink-wrap.o.
3133 * config/i386/i386.c: include "shrink-wrap.h"
3134 * function.c: Likewise.
3135 (requires_stack_frame_p, next_block_for_reg,
3136 move_insn_for_shrink_wrap, prepare_shrink_wrap,
3137 dup_block_and_redirect): Move to shrink-wrap.c
3138 (thread_prologue_and_epilogue_insns): Extract three code segments
3139 as functions in shrink-wrap.c
3140 * function.h: Move #ifdef HAVE_simple_return ... #endif block to
3142 * shrink-wrap.c: New file.
3143 * shrink-wrap.h: New file.
3145 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
3147 * doc/extend.texi: Reflect current numbers of pragmas. Remove
3148 reference to Solaris.
3150 2014-05-12 Mike Stump <mikestump@comcast.net>
3153 * genattrtab.c (filename): Add.
3154 (convert_set_attr_alternative): Improve error message.
3155 (check_defs): Restore read_md_filename for error messages.
3156 (gen_insn): Save filename.
3158 2014-05-12 Dimitris Papavasiliou <dpapavas@gmail.com>
3160 * doc/invoke.texi: Document new switches -Wno-shadow-ivar,
3161 -fno-local-ivars and -fivar-visibility.
3162 * c-family/c.opt: Make -Wshadow also implicitly enable
3165 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
3167 * doc/tm.texi: Remove reference to deleted macro.
3168 * doc/tm.texi.in: Likewise.
3170 2014-05-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
3173 * config/avr/avr.c (avr_out_store_psi): Use correct constant
3176 2014-05-12 Georg-Johann Lay <avr@gjlay.de>
3179 * config/arm/arm.h (License): Add GCC Runtime Library Exception.
3180 * config/arm/aout.h (License): Same.
3181 * config/arm/bpabi.h (License): Same.
3182 * config/arm/elf.h (License): Same.
3183 * config/arm/linux-elf.h (License): Same.
3184 * config/arm/linux-gas.h (License): Same.
3185 * config/arm/netbsd-elf.h (License): Same.
3186 * config/arm/uclinux-eabi.h (License): Same.
3187 * config/arm/uclinux-elf.h (License): Same.
3188 * config/arm/vxworks.h (License): Same.
3190 2014-05-11 Jakub Jelinek <jakub@redhat.com>
3192 * tree.h (OMP_CLAUSE_LINEAR_STMT): Define.
3193 * tree.c (omp_clause_num_ops): Increase OMP_CLAUSE_LINEAR
3194 number of operands to 3.
3195 (walk_tree_1): Walk all operands of OMP_CLAUSE_LINEAR.
3196 * tree-nested.c (convert_nonlocal_omp_clauses,
3197 convert_local_omp_clauses): Handle OMP_CLAUSE_DEPEND.
3198 * gimplify.c (gimplify_scan_omp_clauses): Handle
3199 OMP_CLAUSE_LINEAR_STMT.
3200 * omp-low.c (lower_rec_input_clauses): Fix typo.
3201 (maybe_add_implicit_barrier_cancel, lower_omp_1): Add
3202 cast between Fortran boolean_type_node and C _Bool if
3205 2014-05-11 Richard Sandiford <rdsandiford@googlemail.com>
3207 PR tree-optimization/61136
3208 * wide-int.h (multiple_of_p): Define a version that doesn't return
3210 * fold-const.c (extract_muldiv_1): Use wi::multiple_of_p instead of an
3211 integer_zerop/const_binop pair.
3212 (multiple_of_p): Likewise, converting both operands to widest_int
3215 2014-05-09 Teresa Johnson <tejohnson@google.com>
3217 * cgraphunit.c (analyze_functions): Use correct dump file.
3219 2014-05-09 Florian Weimer <fweimer@redhat.com>
3221 * cfgexpand.c (stack_protect_decl_p): New function, extracted from
3223 (stack_protect_return_slot_p): New function.
3224 (expand_used_vars): Call stack_protect_decl_p and
3225 stack_protect_return_slot_p for -fstack-protector-strong.
3227 2014-05-09 David Wohlferd <LimeGreenSocks@yahoo.com>
3228 Andrew Haley <aph@redhat.com>
3229 Richard Sandiford <rdsandiford@googlemail.com>
3231 * doc/extend.texi: Rewrite inline asm page / re-org asm-related
3234 2014-05-09 Kenneth Zadeck <zadeck@naturalbridge.com>
3237 * fold-const.c (fold_binary_loc): Changed width of mask.
3239 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
3241 * config/avr/avr-fixed.md (round<mode>3): Use -1U instead of -1 in
3242 unsigned int initializers for regno_in, regno_out.
3244 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
3247 * config/avr/avr.md (cc): Add new attribute set_vzn.
3248 (addqi3, addqq3, adduqq3, subqi3, subqq3, subuqq3, negqi2) [cc]:
3249 Set cc insn attribute to set_vzn instead of set_zn for alternatives
3250 with INC, DEC or NEG.
3251 * config/avr/avr.c (avr_notice_update_cc): Handle SET_VZN.
3252 (avr_out_plus_1): ADIW sets cc0 to CC_SET_CZN.
3253 INC, DEC and ADD+ADC set cc0 to CC_CLOBBER.
3255 2014-05-09 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3258 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3260 * wide-int.cc (UTItype): Define.
3261 (UDWtype): Define for appropriate W_TYPE_SIZE.
3263 2014-05-09 Richard Biener <rguenther@suse.de>
3265 * Makefile.in (GTFILES): Remove tree-ssa-propagate.c.
3266 * tree-ssa-propagate.c: Do not include gt-tree-ssa-propagate.h.
3267 (interesting_ssa_edges, varying_ssa_edges): Move out of GC space.
3268 (add_ssa_edge, process_ssa_edge_worklist, ssa_prop_init,
3269 ssa_propagate): Adjust.
3271 2014-05-08 Jeff Law <law@redhat.com>
3273 PR tree-optimization/61009
3274 * tree-ssa-threadedge.c (thread_through_normal_block): Return a
3275 tri-state rather than a boolean. When a block is too big to
3276 thread through, inform caller via negative return value.
3277 (thread_across_edge): If a block was too big for normal threading,
3278 then it's too big for a joiner too, so remove temporary equivalences
3279 and return immediately.
3281 2014-05-08 Manuel López-Ibáñez <manu@gcc.gnu.org>
3282 Matthias Klose <doko@ubuntu.com>
3285 * optc-gen.awk: Fix option handling for -Wunused-parameter.
3287 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
3290 * config/i386/i386.c (PTA_HASWELL): Remove PTA_RTM.
3292 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
3295 * config/alpha/alpha.c: Include gimple-iterator.h.
3296 (alpha_gimple_fold_builtin): New function. Move
3297 ALPHA_BUILTIN_UMULH folding from ...
3298 (alpha_fold_builtin): ... here.
3299 (TARGET_GIMPLE_FOLD_BUILTIN): New define.
3301 2014-05-08 Wei Mi <wmi@google.com>
3304 * config/i386/i386.c (ix86_compute_frame_layout): Update
3305 preferred_stack_boundary for call, expanded from tls descriptor.
3306 * config/i386/i386.md (*tls_global_dynamic_32_gnu): Update RTX
3307 to depend on SP register.
3308 (*tls_local_dynamic_base_32_gnu): Ditto.
3309 (*tls_local_dynamic_32_once): Ditto.
3310 (tls_global_dynamic_64_<mode>): Set
3311 ix86_tls_descriptor_calls_expanded_in_cfun.
3312 (tls_local_dynamic_base_64_<mode>): Ditto.
3313 (tls_global_dynamic_32): Set
3314 ix86_tls_descriptor_calls_expanded_in_cfun. Update RTX
3315 to depend on SP register.
3316 (tls_local_dynamic_base_32): Ditto.
3318 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3320 * config/arm/arm_neon.h: Update comment.
3321 * config/arm/neon-docgen.ml: Delete.
3322 * config/arm/neon-gen.ml: Delete.
3323 * doc/arm-neon-intrinsics.texi: Update comment.
3325 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3327 * config/arm/arm_neon_builtins.def (vadd, vsub): Only define the v2sf
3329 (vand, vorr, veor, vorn, vbic): Remove.
3330 * config/arm/neon.md (neon_vadd, neon_vsub, neon_vadd_unspec): Adjust
3332 (neon_vsub_unspec): Likewise.
3333 (neon_vorr, neon_vand, neon_vbic, neon_veor, neon_vorn): Remove.
3335 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3337 * config/arm/arm_neon.h (vadd_s8): GNU C implementation
3338 (vadd_s16): Likewise.
3339 (vadd_s32): Likewise.
3340 (vadd_f32): Likewise.
3341 (vadd_u8): Likewise.
3342 (vadd_u16): Likewise.
3343 (vadd_u32): Likewise.
3344 (vadd_s64): Likewise.
3345 (vadd_u64): Likewise.
3346 (vaddq_s8): Likewise.
3347 (vaddq_s16): Likewise.
3348 (vaddq_s32): Likewise.
3349 (vaddq_s64): Likewise.
3350 (vaddq_f32): Likewise.
3351 (vaddq_u8): Likewise.
3352 (vaddq_u16): Likewise.
3353 (vaddq_u32): Likewise.
3354 (vaddq_u64): Likewise.
3355 (vmul_s8): Likewise.
3356 (vmul_s16): Likewise.
3357 (vmul_s32): Likewise.
3358 (vmul_f32): Likewise.
3359 (vmul_u8): Likewise.
3360 (vmul_u16): Likewise.
3361 (vmul_u32): Likewise.
3362 (vmul_p8): Likewise.
3363 (vmulq_s8): Likewise.
3364 (vmulq_s16): Likewise.
3365 (vmulq_s32): Likewise.
3366 (vmulq_f32): Likewise.
3367 (vmulq_u8): Likewise.
3368 (vmulq_u16): Likewise.
3369 (vmulq_u32): Likewise.
3370 (vsub_s8): Likewise.
3371 (vsub_s16): Likewise.
3372 (vsub_s32): Likewise.
3373 (vsub_f32): Likewise.
3374 (vsub_u8): Likewise.
3375 (vsub_u16): Likewise.
3376 (vsub_u32): Likewise.
3377 (vsub_s64): Likewise.
3378 (vsub_u64): Likewise.
3379 (vsubq_s8): Likewise.
3380 (vsubq_s16): Likewise.
3381 (vsubq_s32): Likewise.
3382 (vsubq_s64): Likewise.
3383 (vsubq_f32): Likewise.
3384 (vsubq_u8): Likewise.
3385 (vsubq_u16): Likewise.
3386 (vsubq_u32): Likewise.
3387 (vsubq_u64): Likewise.
3388 (vand_s8): Likewise.
3389 (vand_s16): Likewise.
3390 (vand_s32): Likewise.
3391 (vand_u8): Likewise.
3392 (vand_u16): Likewise.
3393 (vand_u32): Likewise.
3394 (vand_s64): Likewise.
3395 (vand_u64): Likewise.
3396 (vandq_s8): Likewise.
3397 (vandq_s16): Likewise.
3398 (vandq_s32): Likewise.
3399 (vandq_s64): Likewise.
3400 (vandq_u8): Likewise.
3401 (vandq_u16): Likewise.
3402 (vandq_u32): Likewise.
3403 (vandq_u64): Likewise.
3404 (vorr_s8): Likewise.
3405 (vorr_s16): Likewise.
3406 (vorr_s32): Likewise.
3407 (vorr_u8): Likewise.
3408 (vorr_u16): Likewise.
3409 (vorr_u32): Likewise.
3410 (vorr_s64): Likewise.
3411 (vorr_u64): Likewise.
3412 (vorrq_s8): Likewise.
3413 (vorrq_s16): Likewise.
3414 (vorrq_s32): Likewise.
3415 (vorrq_s64): Likewise.
3416 (vorrq_u8): Likewise.
3417 (vorrq_u16): Likewise.
3418 (vorrq_u32): Likewise.
3419 (vorrq_u64): Likewise.
3420 (veor_s8): Likewise.
3421 (veor_s16): Likewise.
3422 (veor_s32): Likewise.
3423 (veor_u8): Likewise.
3424 (veor_u16): Likewise.
3425 (veor_u32): Likewise.
3426 (veor_s64): Likewise.
3427 (veor_u64): Likewise.
3428 (veorq_s8): Likewise.
3429 (veorq_s16): Likewise.
3430 (veorq_s32): Likewise.
3431 (veorq_s64): Likewise.
3432 (veorq_u8): Likewise.
3433 (veorq_u16): Likewise.
3434 (veorq_u32): Likewise.
3435 (veorq_u64): Likewise.
3436 (vbic_s8): Likewise.
3437 (vbic_s16): Likewise.
3438 (vbic_s32): Likewise.
3439 (vbic_u8): Likewise.
3440 (vbic_u16): Likewise.
3441 (vbic_u32): Likewise.
3442 (vbic_s64): Likewise.
3443 (vbic_u64): Likewise.
3444 (vbicq_s8): Likewise.
3445 (vbicq_s16): Likewise.
3446 (vbicq_s32): Likewise.
3447 (vbicq_s64): Likewise.
3448 (vbicq_u8): Likewise.
3449 (vbicq_u16): Likewise.
3450 (vbicq_u32): Likewise.
3451 (vbicq_u64): Likewise.
3452 (vorn_s8): Likewise.
3453 (vorn_s16): Likewise.
3454 (vorn_s32): Likewise.
3455 (vorn_u8): Likewise.
3456 (vorn_u16): Likewise.
3457 (vorn_u32): Likewise.
3458 (vorn_s64): Likewise.
3459 (vorn_u64): Likewise.
3460 (vornq_s8): Likewise.
3461 (vornq_s16): Likewise.
3462 (vornq_s32): Likewise.
3463 (vornq_s64): Likewise.
3464 (vornq_u8): Likewise.
3465 (vornq_u16): Likewise.
3466 (vornq_u32): Likewise.
3467 (vornq_u64): Likewise.
3469 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3471 * wide-int.cc (UTItype): Define.
3472 (UDWtype): Define for appropriate W_TYPE_SIZE.
3474 2014-05-08 Marc Glisse <marc.glisse@inria.fr>
3476 PR tree-optimization/59100
3477 * tree-ssa-phiopt.c: Include tree-inline.h.
3478 (neutral_element_p, absorbing_element_p): New functions.
3479 (value_replacement): Handle conditional binary operations with a
3480 neutral or absorbing element.
3482 2014-05-08 Richard Biener <rguenther@suse.de>
3484 * tree-ssa-sccvn.c (vn_get_expr_for): Valueize operands before
3485 folding the expression.
3486 (valueize_expr): Remove.
3487 (visit_reference_op_load): Do not valueize the result of
3489 (simplify_binary_expression): Likewise.
3490 (simplify_unary_expression): Likewise.
3492 2014-05-08 Richard Biener <rguenther@suse.de>
3494 * gimplify.c (gimplify_call_expr): Use saved fnptrtype for
3495 looking at TYPE_ARG_TYPES.
3497 2014-05-08 Richard Biener <rguenther@suse.de>
3499 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove
3500 pointer propagation special-case.
3502 2014-05-08 Bin Cheng <bin.cheng@arm.com>
3504 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
3505 core part of address expressions.
3507 2014-05-08 Alan Modra <amodra@gmail.com>
3510 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
3511 loads and stores when -mno-strict-align at any alignment.
3512 (expand_block_clear): Similarly. Also correct calculation of
3515 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
3518 * tree-complex.c (expand_complex_move): Keep line info when expanding
3520 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
3521 of complex expression. Use new argument to display correct location
3522 for values coming from phi statement.
3523 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
3524 (warn_uninitialized_phi): Pass location of phi argument to
3526 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
3527 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
3529 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
3531 * config/rs6000/predicates.md (indexed_address_mem): New.
3532 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
3533 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
3534 fpstore_ux, fpstore_u.
3535 (sign_extend, indexed, update): New.
3536 (cell_micro): Adjust.
3537 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
3538 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
3539 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
3540 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
3541 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
3542 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
3543 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
3544 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
3545 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
3546 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
3547 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
3548 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
3549 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
3550 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
3551 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
3553 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
3554 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
3555 *vsx_extract_<mode>_store): Adjust.
3556 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
3557 is_cracked_insn, insn_must_be_first_in_group,
3558 insn_must_be_last_in_group): Adjust.
3560 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
3562 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
3563 ppc440-fpstore): Adjust.
3564 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
3565 ppc476-fpstore): Adjust.
3566 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
3567 ppc601-fpstore): Adjust.
3568 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
3570 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
3572 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
3573 ppc7450-fpstore): Adjust.
3574 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
3575 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
3576 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
3578 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
3579 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
3580 cell-fpstore, cell-fpstore-update): Adjust.
3581 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
3582 ppce300c3_store, ppce300c3_fpstore): Adjust.
3583 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
3584 e500mc_fpstore): Adjust.
3585 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
3586 e500mc64_store, e500mc64_fpstore): Adjust.
3587 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
3588 e5500_fpstore): Adjust.
3589 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
3590 e6500_fpstore): Adjust.
3591 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
3593 * config/rs6000/power4.md (power4-load, power4-load-ext,
3594 power4-load-ext-update, power4-load-ext-update-indexed,
3595 power4-load-update-indexed, power4-load-update, power4-fpload,
3596 power4-fpload-update, power4-store, power4-store-update,
3597 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
3599 * config/rs6000/power5.md (power5-load, power5-load-ext,
3600 power5-load-ext-update, power5-load-ext-update-indexed,
3601 power5-load-update-indexed, power5-load-update, power5-fpload,
3602 power5-fpload-update, power5-store, power5-store-update,
3603 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
3605 * config/rs6000/power6.md (power6-load, power6-load-ext,
3606 power6-load-update, power6-load-update-indexed,
3607 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
3608 power6-fpload-update, power6-store, power6-store-update,
3609 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
3611 * config/rs6000/power7.md (power7-load, power7-load-ext,
3612 power7-load-update, power7-load-update-indexed,
3613 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
3614 power7-fpload-update, power7-store, power7-store-update,
3615 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
3617 * config/rs6000/power8.md (power8-load, power8-load-update,
3618 power8-load-ext, power8-load-ext-update, power8-fpload,
3619 power8-fpload-update, power8-store, power8-store-update-indexed,
3620 power8-fpstore, power8-fpstore-update): Adjust.
3621 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
3623 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
3624 titan_lsu_store, titan_lsu_fpstore): Adjust.
3625 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
3627 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
3630 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
3631 unrolled byte insns. Emit address increments after move insns.
3633 2014-05-07 David Malcolm <dmalcolm@redhat.com>
3635 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
3636 const_gimple, rather than a gimple.
3637 (gimple_call_builtin_p): Likewise, for the three variants.
3639 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
3640 (gimple_call_builtin_p): Likewise, for the three variants.
3642 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
3644 PR tree-optimization/61095
3645 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
3647 2014-05-07 Richard Biener <rguenther@suse.de>
3649 PR tree-optimization/61034
3650 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
3651 (maybe_skip_until): Use translate to take into account
3652 lattices when trying to do disambiguations.
3653 (get_continuation_for_phi_1): Likewise.
3654 (get_continuation_for_phi): Adjust for added translate arguments.
3655 (walk_non_aliased_vuses): Likewise.
3656 * tree-ssa-alias.h (get_continuation_for_phi): Adjust prototype.
3657 (walk_non_aliased_vuses): Likewise.
3658 (call_may_clobber_ref_p_1): Declare.
3659 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also disambiguate against
3660 calls. Stop early if we are only supposed to disambiguate.
3661 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
3663 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
3665 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
3666 Emit an error when the function has arguments.
3668 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
3670 * cfgloop.h (unswitch_loops): Remove.
3671 * doc/passes.texi: Remove references to loop-unswitch.c
3672 * timevar.def (TV_LOOP_UNSWITCH): Remove.
3674 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
3676 * tree-vect-data-refs.c (vect_grouped_load_supported): New
3677 check for loads group of length 3.
3678 (vect_permute_load_chain): New permutations for loads group of
3680 * tree-vect-stmts.c (vect_model_load_cost): Change cost
3681 of vec_perm_shuffle for the new permutations.
3683 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
3685 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
3686 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
3687 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
3688 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
3689 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
3690 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
3691 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
3692 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
3694 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
3696 * loop-unswitch.c: Delete.
3698 2014-05-07 Richard Biener <rguenther@suse.de>
3700 * config.gcc: Always set need_64bit_hwint to yes.
3702 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
3704 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
3705 of using optimize_size.
3707 2014-05-06 Mike Stump <mikestump@comcast.net>
3709 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
3711 2014-05-06 Joseph Myers <joseph@codesourcery.com>
3713 * config/i386/sse.md (*mov<mode>_internal)
3714 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
3715 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
3716 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
3717 (*<code><mode>3, *andnot<mode>3<mask_name>)
3718 (<mask_codefor><code><mode>3<mask_name>): Only consider
3719 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
3721 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
3724 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
3726 * lra-constraints.c (valid_address_p): Move earlier in file.
3727 Add a constraint argument to the address_info version.
3728 (satisfies_memory_constraint_p): New function.
3729 (satisfies_address_constraint_p): Likewise.
3730 (process_alt_operands, curr_insn_transform): Use them.
3731 (process_address): Pass the constraint to valid_address_p when
3732 checking address operands.
3734 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
3736 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
3737 to their respective blocks. Fix inadvertent use of "node".
3739 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
3741 * emit-rtl.c (init_derived_machine_modes): New functionm, split
3743 (init_emit_once): ...here.
3744 * rtl.h (init_derived_machine_modes): Declare.
3745 * toplev.c (do_compile): Call it even if no_backend.
3747 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
3748 Mike Stump <mikestump@comcast.net>
3749 Richard Sandiford <rdsandiford@googlemail.com>
3750 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3752 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
3753 (rtx_equal_for_memref_p): Update comment.
3754 (adjust_offset_for_component_ref): Use wide-int interfaces.
3755 * builtins.c (get_object_alignment_2): Likewise.
3756 (c_readstr): Likewise.
3757 (target_char_cast): Add comment.
3758 (determine_block_size): Use wide-int interfaces.
3759 (expand_builtin_signbit): Likewise.
3760 (fold_builtin_int_roundingfn): Likewise.
3761 (fold_builtin_bitop): Likewise.
3762 (fold_builtin_bswap): Likewise.
3763 (fold_builtin_logarithm): Use signop.
3764 (fold_builtin_pow): Likewise.
3765 (fold_builtin_memory_op): Use wide-int interfaces.
3766 (fold_builtin_object_size): Likewise.
3767 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
3768 nb_iterations_estimate.
3769 (record_niter_bound): Use wide-int interfaces.
3770 (get_estimated_loop_iterations_int): Likewise.
3771 (get_estimated_loop_iterations): Likewise.
3772 (get_max_loop_iterations): Likewise.
3773 * cfgloop.h: Include wide-int.h.
3774 (struct nb_iter_bound): Change bound to widest_int.
3775 (struct loop): Change nb_iterations_upper_bound and
3776 nb_iterations_estimate to widest_int.
3777 (record_niter_bound): Switch to use widest_int.
3778 (get_estimated_loop_iterations): Likewise.
3779 (get_max_loop_iterations): Likewise.
3780 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
3781 update for wide-int.
3782 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
3783 * combine.c (try_combine): Likewise.
3784 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
3785 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
3787 (aarch64_float_const_representable_p): Likewise.
3788 * config/arc/arc.c: Include wide-int.h.
3789 (arc_can_use_doloop_p): Use wide-int interfaces.
3790 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
3791 (vfp3_const_double_index): Likewise.
3792 * config/avr/avr.c (avr_out_round): Likewise.
3793 (avr_fold_builtin): Likewise.
3794 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
3795 (bfin_can_use_doloop_p): Likewise.
3796 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
3797 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
3798 * config/i386/i386.c: Include wide-int.h.
3799 (ix86_data_alignment): Use wide-int interfaces.
3800 (ix86_local_alignment): Likewise.
3801 (ix86_emit_swsqrtsf): Update real_from_integer.
3802 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
3803 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
3804 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
3805 (zero_constant): Likewise.
3806 (input_operand): Likewise.
3807 (splat_input_operand): Likewise.
3808 (non_logical_cint_operand): Change const_double to const_wide_int.
3809 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
3810 (easy_altivec_constant): Remove comment.
3811 (paired_expand_vector_init): Use CONSTANT_P.
3812 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
3813 (rs6000_emit_move): Update checks.
3814 (rs6000_aggregate_candidate): Use wide-int interfaces.
3815 (rs6000_expand_ternop_builtin): Likewise.
3816 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
3817 (rs6000_assemble_integer): Likewise.
3818 (rs6000_hash_constant): Likewise.
3819 (output_toc): Likewise.
3820 (rs6000_rtx_costs): Likewise.
3821 (rs6000_emit_swrsqrt); Update call to real_from_integer.
3822 * config/rs6000/rs6000-c.c: Include wide-int.h.
3823 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
3824 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
3825 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
3826 Handle CONST_WIDE_INT.
3827 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
3828 Use tree_fits_uhwi_p.
3829 * config/sparc/sparc.c: Include wide-int.h.
3830 (sparc_fold_builtin): Use wide-int interfaces.
3831 * config/vax/vax.c: Include wide-int.h.
3832 (vax_float_literal): Use real_from_integer.
3833 * coretypes.h (struct hwivec_def): New.
3835 (const_hwivec): New.
3836 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
3837 (equiv_constant): Handle CONST_WIDE_INT.
3838 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
3839 (cselib_hash_rtx): Handle CONST_WIDE_INT.
3840 * dbxout.c (stabstr_U): Use wide-int interfaces.
3841 (dbxout_type): Update to use cst_fits_shwi_p.
3842 * defaults.h (LOG2_BITS_PER_UNIT): Define.
3843 (TARGET_SUPPORTS_WIDE_INT): Add default.
3844 * dfp.c: Include wide-int.h.
3845 (decimal_real_to_integer2): Use wide-int interfaces and rename to
3846 decimal_real_to_integer.
3847 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
3848 decimal_real_to_integer.
3849 * doc/generic.texi (Constant expressions): Update for wide_int.
3850 * doc/rtl.texi (const_double): Likewise.
3851 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
3852 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
3853 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
3854 (REAL_VALUE_FROM_INT): Remove.
3855 (TARGET_SUPPORTS_WIDE_INT): New.
3856 * doc/tm.texi: Regenerate.
3857 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
3858 * double-int.h: Include wide-int.h.
3859 (struct wi::int_traits): New.
3860 * dwarf2out.c (get_full_len): New.
3861 (dw_val_equal_p): Add case dw_val_class_wide_int.
3862 (size_of_loc_descr): Likewise.
3863 (output_loc_operands): Likewise.
3864 (insert_double): Remove.
3865 (insert_wide_int): New.
3867 (print_die): Add case dw_val_class_wide_int.
3868 (attr_checksum): Likewise.
3869 (attr_checksum_ordered): Likewise.
3870 (same_dw_val_p): Likewise.
3871 (size_of_die): Likewise.
3872 (value_format): Likewise.
3873 (output_die): Likewise.
3874 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
3876 (clz_loc_descriptor): Use wide-int interfaces.
3877 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
3878 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
3879 (round_up_to_align): Use wide-int interfaces.
3880 (field_byte_offset): Likewise.
3881 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
3882 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
3883 CONST_DOUBLE handling. Use wide-int interfaces.
3884 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
3885 (gen_enumeration_type_die): Use add_AT_wide.
3886 (hash_loc_operands): Add case dw_val_class_wide_int.
3887 (compare_loc_operands): Likewise.
3888 * dwarf2out.h: Include wide-int.h.
3889 (wide_int_ptr): New.
3890 (enum dw_val_class): Add dw_val_class_wide_int.
3891 (struct dw_val_struct): Add val_wide.
3892 * emit-rtl.c (const_wide_int_htab): New.
3893 (const_wide_int_htab_hash): New.
3894 (const_wide_int_htab_eq): New.
3895 (lookup_const_wide_int): New.
3896 (const_double_htab_hash): Use wide-int interfaces.
3897 (const_double_htab_eq): Likewise.
3898 (rtx_to_double_int): Conditionally compile for wide-int.
3899 (immed_double_int_const): Rename to immed_wide_int_const and
3900 update for wide-int.
3901 (immed_double_const): Conditionally compile for wide-int.
3902 (init_emit_once): Use wide-int interfaces.
3903 * explow.c (plus_constant): Likewise.
3904 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
3905 (lshift_value): Use wide-int interfaces.
3906 (expand_mult): Likewise.
3907 (choose_multiplier): Likewise.
3908 (expand_smod_pow2): Likewise.
3909 (make_tree): Likewise.
3910 * expr.c (convert_modes): Consolidate handling of constants.
3911 Use wide-int interfaces.
3912 (emit_group_load_1): Add note.
3913 (store_expr): Update comment.
3914 (get_inner_reference): Use wide-int interfaces.
3915 (expand_constructor): Update comment.
3916 (expand_expr_real_2): Use wide-int interfaces.
3917 (expand_expr_real_1): Likewise.
3918 (reduce_to_bit_field_precision): Likewise.
3919 (const_vector_from_tree): Likewise.
3920 * final.c: Include wide-int-print.h.
3921 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
3922 * fixed-value.c: Include wide-int.h.
3923 (fixed_from_string): Use wide-int interfaces.
3924 (fixed_to_decimal): Likewise.
3925 (fixed_convert_from_real): Likewise.
3926 (real_convert_from_fixed): Likewise.
3927 * fold-const.h (mem_ref_offset): Return an offset_int.
3928 (div_if_zero_remainder): Remove code parameter.
3929 * fold-const.c (div_if_zero_remainder): Remove code parameter.
3930 Use wide-int interfaces.
3931 (may_negate_without_overflow_p): Use wide-int interfaces.
3932 (negate_expr_p): Likewise.
3933 (fold_negate_expr): Likewise.
3934 (int_const_binop_1): Likewise.
3935 (const_binop): Likewise.
3936 (fold_convert_const_int_from_int): Likewise.
3937 (fold_convert_const_int_from_real): Likewise.
3938 (fold_convert_const_int_from_fixed): Likewise.
3939 (fold_convert_const_fixed_from_int): Likewise.
3940 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
3941 (sign_bit_p): Use wide-int interfaces.
3942 (make_range_step): Likewise.
3943 (build_range_check): Likewise. Pass an integer of the correct type
3944 instead of using integer_one_node.
3945 (range_predecessor): Pass an integer of the correct type instead
3946 of using integer_one_node.
3947 (range_successor): Likewise.
3948 (merge_ranges): Likewise.
3949 (unextend): Use wide-int interfaces.
3950 (extract_muldiv_1): Likewise.
3951 (fold_div_compare): Likewise.
3952 (fold_single_bit_test): Likewise.
3953 (fold_sign_changed_comparison): Likewise.
3954 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
3955 (fold_plusminus_mult_expr): Use wide-int interfaces.
3956 (native_encode_int): Likewise.
3957 (native_interpret_int): Likewise.
3958 (fold_unary_loc): Likewise.
3959 (pointer_may_wrap_p): Likewise.
3960 (size_low_cst): Likewise.
3961 (mask_with_tz): Likewise.
3962 (fold_binary_loc): Likewise.
3963 (fold_ternary_loc): Likewise.
3964 (multiple_of_p): Likewise.
3965 (tree_call_nonnegative_warnv_p): Update calls to
3966 tree_int_cst_min_precision and real_from_integer.
3967 (fold_negate_const): Use wide-int interfaces.
3968 (fold_abs_const): Likewise.
3969 (fold_relational_const): Use tree_int_cst_lt.
3970 (round_up_loc): Use wide-int interfaces.
3971 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
3972 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
3973 * gengtype.c: Remove include of double-int.h.
3974 (do_typedef): Use wide-int interfaces.
3975 (open_base_files): Add wide-int.h.
3976 (main): Add offset_int and widest_int typedefs.
3977 * gengtype-lex.l: Handle "^".
3978 (CXX_KEYWORD): Add "static".
3979 * gengtype-parse.c (require3): New.
3980 (require_template_declaration): Handle constant template arguments
3981 and nested templates.
3982 * gengtype-state.c: Don't include "double-int.h".
3983 * genpreds.c (write_one_predicate_function): Update comment.
3984 (write_tm_constrs_h): Add check for hval and lval use in
3986 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
3987 (add_to_sequence): Likewise.
3988 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
3989 and const_double_operand.
3990 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
3992 * gimple-fold.c (get_base_constructor): Likewise.
3993 (fold_array_ctor_reference): Likewise.
3994 (fold_nonarray_ctor_reference): Likewise.
3995 (fold_const_aggregate_ref_1): Likewise.
3996 (gimple_val_nonnegative_real_p): Likewise.
3997 (gimple_fold_indirect_ref): Likewise.
3998 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
3999 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
4000 (struct slsr_cand_d): Change index to be widest_int.
4001 (struct incr_info_d): Change incr to be widest_int.
4002 (alloc_cand_and_find_basis): Use wide-int interfaces.
4003 (slsr_process_phi): Likewise.
4004 (backtrace_base_for_ref): Likewise. Return a widest_int.
4005 (restructure_reference): Take a widest_int instead of a double_int.
4006 (slsr_process_ref): Use wide-int interfaces.
4007 (create_mul_ssa_cand): Likewise.
4008 (create_mul_imm_cand): Likewise.
4009 (create_add_ssa_cand): Likewise.
4010 (create_add_imm_cand): Take a widest_int instead of a double_int.
4011 (slsr_process_add): Use wide-int interfaces.
4012 (slsr_process_cast): Likewise.
4013 (slsr_process_copy): Likewise.
4014 (dump_candidate): Likewise.
4015 (dump_incr_vec): Likewise.
4016 (replace_ref): Likewise.
4017 (cand_increment): Likewise. Return a widest_int.
4018 (cand_abs_increment): Likewise.
4019 (replace_mult_candidate): Take a widest_int instead of a double_int.
4020 (replace_unconditional_candidate): Use wide-int interfaces.
4021 (incr_vec_index): Take a widest_int instead of a double_int.
4022 (create_add_on_incoming_edge): Likewise.
4023 (create_phi_basis): Use wide-int interfaces.
4024 (replace_conditional_candidate): Likewise.
4025 (record_increment): Take a widest_int instead of a double_int.
4026 (record_phi_increments): Use wide-int interfaces.
4027 (phi_incr_cost): Take a widest_int instead of a double_int.
4028 (lowest_cost_path): Likewise.
4029 (total_savings): Likewise.
4030 (analyze_increments): Use wide-int interfaces.
4031 (ncd_with_phi): Take a widest_int instead of a double_int.
4032 (ncd_of_cand_and_phis): Likewise.
4033 (nearest_common_dominator_for_cands): Likewise.
4034 (insert_initializers): Use wide-int interfaces.
4035 (all_phi_incrs_profitable): Likewise.
4036 (replace_one_candidate): Likewise.
4037 (replace_profitable_candidates): Likewise.
4038 * godump.c: Include wide-int-print.h.
4039 (go_output_typedef): Use wide-int interfaces.
4040 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
4041 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
4042 (build_loop_iteration_domains): Likewise.
4043 * hooks.h: Include wide-int.h rather than double-int.h.
4044 (hook_bool_dint_dint_uint_bool_true): Delete.
4045 (hook_bool_wint_wint_uint_bool_true): Declare.
4046 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
4047 (hook_bool_wint_wint_uint_bool_true): New.
4048 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
4050 (ubsan_expand_si_overflow_mul_check): Likewise.
4051 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
4052 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
4053 (get_ancestor_addr_info): Likewise.
4054 (ipa_modify_call_arguments): Likewise.
4055 * loop-doloop.c (doloop_modify): Likewise.
4056 (doloop_optimize): Likewise.
4057 * loop-iv.c (iv_number_of_iterations): Likewise.
4058 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
4059 (unroll_loop_constant_iterations): Likewise.
4060 (decide_unroll_runtime_iterations): Likewise.
4061 (unroll_loop_runtime_iterations): Likewise.
4062 (decide_peel_simple): Likewise.
4063 (decide_unroll_stupid): Likewise.
4064 * lto-streamer-in.c (streamer_read_wi): Add.
4065 (input_cfg): Use wide-int interfaces.
4066 (lto_input_tree_1): Likewise.
4067 * lto-streamer-out.c (streamer_write_wi): Add.
4068 (hash_tree): Use wide-int interfaces.
4069 (output_cfg): Likewise.
4070 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
4071 (GTFILES): Add wide-int.h and signop.h.
4072 (TAGS): Look for .cc files too.
4073 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
4074 * optabs.c (expand_subword_shift): Likewise.
4075 (expand_doubleword_shift): Likewise.
4076 (expand_absneg_bit): Likewise.
4077 (expand_copysign_absneg): Likewise.
4078 (expand_copysign_bit): Likewise.
4079 * postreload.c (reload_cse_simplify_set): Likewise.
4080 * predict.c (predict_iv_comparison): Likewise.
4081 * pretty-print.h: Include wide-int-print.h.
4083 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
4084 * print-tree.c: Include wide-int-print.h.
4085 (print_node_brief): Use wide-int interfaces.
4086 (print_node): Likewise.
4087 * read-rtl.c (validate_const_wide_int): New.
4088 (read_rtx_code): Add CONST_WIDE_INT case.
4089 * real.c: Include wide-int.h.
4090 (real_to_integer2): Delete.
4091 (real_to_integer): New function, returning a wide_int.
4092 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
4093 (ten_to_ptwo): Update call to real_from_integer.
4094 (real_digit): Likewise.
4095 * real.h: Include signop.h, wide-int.h and insn-modes.h.
4096 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
4097 (REAL_VALUE_TO_INT): Delete.
4098 (real_to_integer): Declare a wide-int form.
4099 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
4100 * recog.c (const_int_operand): Improve comment.
4101 (const_scalar_int_operand): New.
4102 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
4103 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
4104 (split_double): Likewise.
4105 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
4106 (rtx_size): Likewise.
4107 (rtx_alloc_stat_v): New.
4108 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
4109 (cwi_output_hex): New.
4110 (iterative_hash_rtx): Handle CONST_WIDE_INT.
4111 (cwi_check_failed_bounds): New.
4112 * rtl.def (CONST_WIDE_INT): New.
4113 * rtl.h: Include <utility> and wide-int.h.
4114 (struct hwivec_def): New.
4115 (CWI_GET_NUM_ELEM): New.
4116 (CWI_PUT_NUM_ELEM): New.
4117 (struct rtx_def): Add num_elem and hwiv.
4118 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
4119 (CASE_CONST_UNIQUE): Likewise.
4120 (CASE_CONST_ANY): Likewise.
4121 (CONST_SCALAR_INT_P): Likewise.
4122 (CONST_WIDE_INT_P): New.
4124 (HWIVEC_CHECK): New.
4125 (cwi_check_failed_bounds): New.
4127 (HWIVEC_CHECK): New.
4128 (CONST_WIDE_INT_VEC) New.
4129 (CONST_WIDE_INT_NUNITS) New.
4130 (CONST_WIDE_INT_ELT) New.
4131 (rtx_mode_t): New type.
4132 (wi::int_traits <rtx_mode_t>): New.
4134 (wi::min_value): New.
4135 (wi::max_value): New.
4137 (const_wide_int_alloc): New.
4138 (immed_wide_int_const): New.
4139 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
4140 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
4141 * signop.h: New file.
4142 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
4143 (simplify_const_unary_operation): Use wide-int interfaces.
4144 (simplify_binary_operation_1): Likewise.
4145 (simplify_const_binary_operation): Likewise.
4146 (simplify_const_relational_operation): Likewise.
4147 (simplify_immed_subreg): Likewise.
4148 * stmt.c (expand_case): Likewise.
4149 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
4150 signop rather than a bool.
4151 * stor-layout.c (layout_type): Use wide-int interfaces.
4152 (initialize_sizetypes): Update calls to
4153 set_min_and_max_values_for_integral_type.
4154 (set_min_and_max_values_for_integral_type): Take a signop rather
4155 than a bool. Use wide-int interfaces.
4156 (fixup_signed_type): Update accordingly. Remove
4157 HOST_BITS_PER_DOUBLE_INT limit.
4158 (fixup_unsigned_type): Likewise.
4159 * system.h (STATIC_CONSTANT_P): New.
4160 (STATIC_ASSERT): New.
4161 * target.def (can_use_doloop_p): Take widest_ints rather than
4163 * target.h: Include wide-int.h rather than double-int.h.
4164 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
4166 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
4167 rather than INT_CST_LT_UNSIGNED.
4168 (can_use_doloop_if_innermost): Take widest_ints rather than
4170 * tree-affine.c: Include wide-int-print.h.
4171 (double_int_ext_for_comb): Delete.
4172 (wide_int_ext_for_comb): New.
4173 (aff_combination_zero): Use wide-int interfaces.
4174 (aff_combination_const): Take a widest_int instead of a double_int.
4175 (aff_combination_elt): Use wide-int interfaces.
4176 (aff_combination_scale): Take a widest_int instead of a double_int.
4177 (aff_combination_add_elt): Likewise.
4178 (aff_combination_add_cst): Likewise.
4179 (aff_combination_add): Use wide-int interfaces.
4180 (aff_combination_convert): Likewise.
4181 (tree_to_aff_combination): Likewise.
4182 (add_elt_to_tree): Take a widest_int instead of a double_int.
4183 (aff_combination_to_tree): Use wide-int interfaces.
4184 (aff_combination_remove_elt): Likewise.
4185 (aff_combination_add_product): Take a widest_int instead of
4187 (aff_combination_mult): Use wide-int interfaces.
4188 (aff_combination_expand): Likewise.
4189 (double_int_constant_multiple_p): Delete.
4190 (wide_int_constant_multiple_p): New.
4191 (aff_combination_constant_multiple_p): Take a widest_int pointer
4192 instead of a double_int pointer.
4193 (print_aff): Use wide-int interfaces.
4194 (get_inner_reference_aff): Take a widest_int pointer
4195 instead of a double_int pointer.
4196 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
4197 * tree-affine.h: Include wide-int.h.
4198 (struct aff_comb_elt): Change type of coef to widest_int.
4199 (struct affine_tree_combination): Change type of offset to widest_int.
4200 (double_int_ext_for_comb): Delete.
4201 (wide_int_ext_for_comb): New.
4202 (aff_combination_const): Use widest_int instead of double_int.
4203 (aff_combination_scale): Likewise.
4204 (aff_combination_add_elt): Likewise.
4205 (aff_combination_constant_multiple_p): Likewise.
4206 (get_inner_reference_aff): Likewise.
4207 (aff_comb_cannot_overlap_p): Likewise.
4208 (aff_combination_zero_p): Use wide-int interfaces.
4209 * tree.c: Include tree.h.
4210 (init_ttree): Use make_int_cst.
4211 (tree_code_size): Removed code for INTEGER_CST case.
4212 (tree_size): Add INTEGER_CST case.
4213 (make_node_stat): Update comment.
4214 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
4215 (build_int_cst_type): Use wide-int interfaces.
4216 (double_int_to_tree): Likewise.
4217 (double_int_fits_to_tree_p): Delete.
4218 (force_fit_type_double): Delete.
4219 (force_fit_type): New.
4220 (int_cst_hash_hash): Use wide-int interfaces.
4221 (int_cst_hash_eq): Likewise.
4222 (build_int_cst_wide): Delete.
4223 (wide_int_to_tree): New.
4224 (cache_integer_cst): Use wide-int interfaces.
4225 (build_low_bits_mask): Likewise.
4226 (cst_and_fits_in_hwi): Likewise.
4227 (real_value_from_int_cst): Likewise.
4228 (make_int_cst_stat): New.
4229 (integer_zerop): Use wide_int interfaces.
4230 (integer_onep): Likewise.
4231 (integer_all_onesp): Likewise.
4232 (integer_pow2p): Likewise.
4233 (integer_nonzerop): Likewise.
4234 (tree_log2): Likewise.
4235 (tree_floor_log2): Likewise.
4236 (tree_ctz): Likewise.
4237 (int_size_in_bytes): Likewise.
4238 (mem_ref_offset): Return an offset_int rather than a double_int.
4239 (build_type_attribute_qual_variant): Use wide_int interfaces.
4240 (type_hash_eq): Likewise
4241 (tree_int_cst_equal): Likewise.
4242 (tree_int_cst_lt): Delete.
4243 (tree_int_cst_compare): Likewise.
4244 (tree_fits_shwi_p): Use wide_int interfaces.
4245 (tree_fits_uhwi_p): Likewise.
4246 (tree_int_cst_sign_bit): Likewise.
4247 (tree_int_cst_sgn): Likewise.
4248 (tree_int_cst_min_precision): Take a signop rather than a bool.
4249 (simple_cst_equal): Use wide_int interfaces.
4250 (compare_tree_int): Likewise.
4251 (iterative_hash_expr): Likewise.
4252 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
4254 (get_type_static_bounds): Use wide_int interfaces.
4255 (tree_int_cst_elt_check_failed): New.
4256 (build_common_tree_nodes): Reordered to set prec before filling in
4258 (int_cst_value): Check cst_and_fits_in_hwi.
4259 (widest_int_cst_value): Use wide_int interfaces.
4260 (upper_bound_in_type): Likewise.
4261 (lower_bound_in_type): Likewise.
4262 (num_ending_zeros): Likewise.
4263 (drop_tree_overflow): Likewise.
4264 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
4265 (gen_conditions_for_pow_cst_base): Likewise.
4266 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
4267 (group_case_labels_stmt): Use wide-int interfaces.
4268 (verify_gimple_assign_binary): Likewise.
4269 (print_loop): Likewise.
4270 * tree-chrec.c (tree_fold_binomial): Likewise.
4271 * tree-core.h (struct tree_base): Add int_length.
4272 (struct tree_int_cst): Change rep of value.
4273 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
4274 (dr_may_alias_p): Likewise.
4275 (max_stmt_executions_tree): Likewise.
4276 * tree.def (INTEGER_CST): Update comment.
4277 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
4278 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
4279 * tree-dump.c: Include wide-int.h and wide-int-print.h.
4280 (dequeue_and_dump): Use wide-int interfaces.
4281 * tree.h: Include wide-int.h.
4282 (NULL_TREE): Moved to earlier loc in file.
4283 (TREE_INT_CST_ELT_CHECK): New.
4284 (tree_int_cst_elt_check_failed): New.
4286 (TREE_INT_CST): Delete.
4287 (TREE_INT_CST_LOW): Use wide-int interfaces.
4288 (TREE_INT_CST_HIGH): Delete.
4289 (TREE_INT_CST_NUNITS): New.
4290 (TREE_INT_CST_EXT_NUNITS): Likewise.
4291 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
4292 (TREE_INT_CST_ELT): Likewise.
4293 (INT_CST_LT): Delete.
4294 (tree_int_cst_elt_check): New (two forms).
4295 (type_code_size): Update comment.
4296 (make_int_cst_stat, make_int_cst): New.
4297 (tree_to_double_int): Delete.
4298 (double_int_fits_to_tree_p): Delete.
4299 (force_fit_type_double): Delete.
4300 (build_int_cstu): Replace with out-of-line function.
4301 (build_int_cst_wide): Delete.
4302 (tree_int_cst_lt): Define inline.
4303 (tree_int_cst_le): New.
4304 (tree_int_cst_compare): Define inline.
4305 (tree_int_cst_min_precision): Take a signop rather than a bool.
4306 (wi::int_traits <const_tree>): New.
4307 (wi::int_traits <tree>): New.
4308 (wi::extended_tree): New.
4309 (wi::int_traits <wi::extended_tree>): New.
4310 (wi::to_widest): New.
4311 (wi::to_offset): New.
4312 (wi::fits_to_tree_p): New.
4313 (wi::min_value): New.
4314 (wi::max_value): New.
4315 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
4316 (copy_tree_body_r): Likewise.
4317 * tree-object-size.c (compute_object_offset): Likewise.
4318 (addr_object_size): Likewise.
4319 * tree-predcom.c: Include wide-int-print.h.
4320 (struct dref_d): Change type of offset to widest_int.
4321 (dump_dref): Call wide-int printer.
4322 (aff_combination_dr_offset): Use wide-int interfaces.
4323 (determine_offset): Take a widest_int pointer rather than a
4325 (split_data_refs_to_components): Use wide-int interfaces.
4326 (suitable_component_p): Likewise.
4327 (order_drefs): Likewise.
4328 (add_ref_to_chain): Likewise.
4329 (valid_initializer_p): Likewise.
4330 (determine_roots_comp): Likewise.
4331 * tree-pretty-print.c: Include wide-int-print.h.
4332 (dump_generic_node): Use wide-int interfaces.
4333 * tree-sra.c (sra_ipa_modify_expr): Likewise.
4334 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
4335 (move_fixed_address_to_symbol): Likewise.
4336 (move_hint_to_base): Likewise.
4337 (move_pointer_to_base): Likewise.
4338 (move_variant_to_index): Likewise.
4339 (most_expensive_mult_to_index): Likewise.
4340 (addr_to_parts): Likewise.
4341 (copy_ref_info): Likewise.
4342 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
4343 (indirect_refs_may_alias_p): Likewise.
4344 (stmt_kills_ref_p_1): Likewise.
4345 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
4346 * tree-ssa-ccp.c: Update comment at top of file. Include
4348 (struct prop_value_d): Change type of mask to widest_int.
4349 (extend_mask): New function.
4350 (dump_lattice_value): Use wide-int interfaces.
4351 (get_default_value): Likewise.
4352 (set_constant_value): Likewise.
4353 (set_value_varying): Likewise.
4354 (valid_lattice_transition): Likewise.
4355 (set_lattice_value): Likewise.
4356 (value_to_double_int): Delete.
4357 (value_to_wide_int): New.
4358 (get_value_from_alignment): Use wide-int interfaces.
4359 (get_value_for_expr): Likewise.
4360 (do_dbg_cnt): Likewise.
4361 (ccp_finalize): Likewise.
4362 (ccp_lattice_meet): Likewise.
4363 (bit_value_unop_1): Use widest_ints rather than double_ints.
4364 (bit_value_binop_1): Likewise.
4365 (bit_value_unop): Use wide-int interfaces.
4366 (bit_value_binop): Likewise.
4367 (bit_value_assume_aligned): Likewise.
4368 (evaluate_stmt): Likewise.
4369 (ccp_fold_stmt): Likewise.
4370 (visit_cond_stmt): Likewise.
4371 (ccp_visit_stmt): Likewise.
4372 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
4373 (constant_pointer_difference): Likewise.
4374 (associate_pointerplus): Likewise.
4375 (combine_conversions): Likewise.
4376 * tree-ssa-loop.h: Include wide-int.h.
4377 (struct tree_niter_desc): Change type of max to widest_int.
4378 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
4379 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
4380 (remove_redundant_iv_tests): Likewise.
4381 (canonicalize_loop_induction_variables): Likewise.
4382 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
4383 (constant_multiple_of): Take a widest_int pointer instead of
4384 a double_int pointer.
4385 (get_computation_aff): Use wide-int interfaces.
4386 (ptr_difference_cost): Likewise.
4387 (difference_cost): Likewise.
4388 (get_loop_invariant_expr_id): Likewise.
4389 (get_computation_cost_at): Likewise.
4390 (iv_elimination_compare_lt): Likewise.
4391 (may_eliminate_iv): Likewise.
4392 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
4393 instead of double_int.
4394 (max_loop_iterations): Likewise.
4395 (max_stmt_executions): Likewise.
4396 (estimated_stmt_executions): Likewise.
4397 * tree-ssa-loop-niter.c: Include wide-int-print.h.
4398 (split_to_var_and_offset): Use wide-int interfaces.
4399 (determine_value_range): Likewise.
4400 (bound_difference_of_offsetted_base): Likewise.
4401 (bounds_add): Take a widest_int instead of a double_int.
4402 (number_of_iterations_ne_max): Use wide-int interfaces.
4403 (number_of_iterations_ne): Likewise.
4404 (number_of_iterations_lt_to_ne): Likewise.
4405 (assert_loop_rolls_lt): Likewise.
4406 (number_of_iterations_lt): Likewise.
4407 (number_of_iterations_le): Likewise.
4408 (number_of_iterations_cond): Likewise.
4409 (number_of_iterations_exit): Likewise.
4410 (finite_loop_p): Likewise.
4411 (derive_constant_upper_bound_assign): Likewise.
4412 (derive_constant_upper_bound): Return a widest_int.
4413 (derive_constant_upper_bound_ops): Likewise.
4414 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
4415 (record_estimate): Take a widest_int rather than a double_int.
4416 (record_nonwrapping_iv): Use wide-int interfaces.
4417 (double_int_cmp): Delete.
4418 (wide_int_cmp): New.
4419 (bound_index): Take a widest_int rather than a double_int.
4420 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
4421 (maybe_lower_iteration_bound): Likewise.
4422 (estimate_numbers_of_iterations_loop): Likewise.
4423 (estimated_loop_iterations): Take a widest_int pointer than than
4424 a double_int pointer.
4425 (estimated_loop_iterations_int): Use wide-int interfaces.
4426 (max_loop_iterations): Take a widest_int pointer than than
4427 a double_int pointer.
4428 (max_loop_iterations_int): Use wide-int interfaces.
4429 (max_stmt_executions): Take a widest_int pointer than than
4430 a double_int pointer.
4431 (estimated_stmt_executions): Likewise.
4432 (n_of_executions_at_most): Use wide-int interfaces.
4433 (scev_probably_wraps_p): Likewise.
4434 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
4436 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
4438 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
4439 double_ints. Adjust for trailing_wide_ints <3> representation.
4440 (set_nonzero_bits): Likewise.
4441 (get_range_info): Return wide_ints rather than double_ints.
4442 Adjust for trailing_wide_ints <3> representation.
4443 (get_nonzero_bits): Likewise.
4444 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
4446 * tree-ssanames.h (struct range_info_def): Replace min, max and
4447 nonzero_bits with a trailing_wide_ints <3>.
4448 (set_range_info): Use wide_int_refs rather than double_ints.
4449 (set_nonzero_bits): Likewise.
4450 (get_range_info): Return wide_ints rather than double_ints.
4451 (get_nonzero_bits): Likewise.
4452 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
4453 * tree-ssa-pre.c (phi_translate_1): Likewise.
4454 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
4455 (acceptable_pow_call): Likewise.
4456 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
4458 (vn_reference_fold_indirect): Likewise.
4459 (vn_reference_maybe_forwprop_address): Likewise.
4460 (valueize_refs_1): Likewise.
4461 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
4462 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
4463 tree_int_cst_lt and tree_int_cst_le.
4464 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
4466 (streamer_alloc_tree): Likewise.
4467 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
4468 (streamer_write_tree_header): Likewise.
4469 (streamer_write_integer_cst): Likewise.
4470 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
4471 (build_constructors): Likewise.
4472 (array_value_type): Likewise.
4473 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
4474 (vect_check_gather): Likewise.
4475 * tree-vect-generic.c (build_replicated_const): Likewise.
4476 (expand_vector_divmod): Likewise.
4477 * tree-vect-loop.c (vect_transform_loop): Likewise.
4478 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
4479 (vect_do_peeling_for_alignment): Likewise.
4480 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
4481 * tree-vrp.c: Include wide-int.h.
4482 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
4483 (extract_range_from_assert): Use wide-int interfaces.
4484 (vrp_int_const_binop): Likewise.
4485 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
4486 double_int pointers.
4487 (ranges_from_anti_range): Use wide-int interfaces.
4488 (quad_int_cmp): Delete.
4489 (quad_int_pair_sort): Likewise.
4490 (extract_range_from_binary_expr_1): Use wide-int interfaces.
4491 (extract_range_from_unary_expr_1): Likewise.
4492 (adjust_range_with_scev): Likewise.
4493 (masked_increment): Take and return wide_ints rather than double_ints.
4494 (register_edge_assert_for_2): Use wide-int interfaces.
4495 (check_array_ref): Likewise.
4496 (search_for_addr_array): Likewise.
4497 (maybe_set_nonzero_bits): Likewise.
4498 (union_ranges): Pass an integer of the correct type instead of
4499 using integer_one_node.
4500 (intersect_ranges): Likewise.
4501 (simplify_truth_ops_using_ranges): Likewise.
4502 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
4503 (range_fits_type_p): Likewise.
4504 (simplify_cond_using_ranges): Likewise. Take a signop rather than
4506 (simplify_conversion_using_ranges): Use wide-int interfaces.
4507 (simplify_float_conversion_using_ranges): Likewise.
4508 (vrp_finalize): Likewise.
4509 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
4510 (gimple_stringops_transform): Likewise.
4511 * varasm.c (decode_addr_const): Likewise.
4512 (const_hash_1): Likewise.
4513 (const_rtx_hash_1): Likewise
4514 (output_constant): Likewise.
4515 (array_size_for_constructor): Likewise.
4516 (output_constructor_regular_field): Likewise.
4517 (output_constructor_bitfield): Likewise.
4518 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
4519 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
4521 * gencheck.c: Define BITS_PER_UNIT.
4524 * wide-int-print.cc: New.
4525 * wide-int-print.h: New.
4527 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4529 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
4531 2014-05-06 Richard Biener <rguenther@suse.de>
4533 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
4534 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
4535 (TODO_verify_all): Adjust.
4536 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
4537 TODO_verify_stmts and TODO_verify_rtl_sharing.
4538 * bb-reorder.c: Likewise.
4539 * cfgexpand.c: Likewise.
4540 * cprop.c: Likewise.
4542 * function.c: Likewise.
4543 * fwprop.c: Likewise.
4545 * gimple-ssa-isolate-paths.c: Likewise.
4546 * gimple-ssa-strength-reduction.c: Likewise.
4547 * ipa-split.c: Likewise.
4548 * loop-init.c: Likewise.
4549 * loop-unroll.c: Likewise.
4550 * lower-subreg.c: Likewise.
4551 * modulo-sched.c: Likewise.
4552 * postreload-gcse.c: Likewise.
4553 * predict.c: Likewise.
4554 * recog.c: Likewise.
4555 * sched-rgn.c: Likewise.
4556 * store-motion.c: Likewise.
4557 * tracer.c: Likewise.
4558 * trans-mem.c: Likewise.
4559 * tree-call-cdce.c: Likewise.
4560 * tree-cfg.c: Likewise.
4561 * tree-cfgcleanup.c: Likewise.
4562 * tree-complex.c: Likewise.
4563 * tree-eh.c: Likewise.
4564 * tree-emutls.c: Likewise.
4565 * tree-if-conv.c: Likewise.
4566 * tree-into-ssa.c: Likewise.
4567 * tree-loop-distribution.c: Likewise.
4568 * tree-object-size.c: Likewise.
4569 * tree-parloops.c: Likewise.
4570 * tree-pass.h: Likewise.
4571 * tree-sra.c: Likewise.
4572 * tree-ssa-ccp.c: Likewise.
4573 * tree-ssa-copy.c: Likewise.
4574 * tree-ssa-copyrename.c: Likewise.
4575 * tree-ssa-dce.c: Likewise.
4576 * tree-ssa-dom.c: Likewise.
4577 * tree-ssa-dse.c: Likewise.
4578 * tree-ssa-forwprop.c: Likewise.
4579 * tree-ssa-ifcombine.c: Likewise.
4580 * tree-ssa-loop-ch.c: Likewise.
4581 * tree-ssa-loop-ivcanon.c: Likewise.
4582 * tree-ssa-loop.c: Likewise.
4583 * tree-ssa-math-opts.c: Likewise.
4584 * tree-ssa-phiopt.c: Likewise.
4585 * tree-ssa-phiprop.c: Likewise.
4586 * tree-ssa-pre.c: Likewise.
4587 * tree-ssa-reassoc.c: Likewise.
4588 * tree-ssa-sink.c: Likewise.
4589 * tree-ssa-strlen.c: Likewise.
4590 * tree-ssa-tail-merge.c: Likewise.
4591 * tree-ssa-uncprop.c: Likewise.
4592 * tree-switch-conversion.c: Likewise.
4593 * tree-tailcall.c: Likewise.
4594 * tree-vect-generic.c: Likewise.
4595 * tree-vectorizer.c: Likewise.
4596 * tree-vrp.c: Likewise.
4598 * var-tracking.c: Likewise.
4599 * bt-load.c: Likewise.
4600 * cfgcleanup.c: Likewise.
4601 * combine-stack-adj.c: Likewise.
4602 * combine.c: Likewise.
4603 * compare-elim.c: Likewise.
4604 * config/epiphany/resolve-sw-modes.c: Likewise.
4605 * config/i386/i386.c: Likewise.
4606 * config/mips/mips.c: Likewise.
4607 * config/s390/s390.c: Likewise.
4608 * config/sh/sh_treg_combine.cc: Likewise.
4609 * config/sparc/sparc.c: Likewise.
4612 * final.c: Likewise.
4613 * ifcvt.c: Likewise.
4614 * mode-switching.c: Likewise.
4615 * passes.c: Likewise.
4616 * postreload.c: Likewise.
4618 * reg-stack.c: Likewise.
4619 * regcprop.c: Likewise.
4620 * regrename.c: Likewise.
4623 2014-05-06 Richard Biener <rguenther@suse.de>
4626 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
4627 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
4629 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
4632 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
4634 2014-05-05 Radovan Obradovic <robradovic@mips.com>
4635 Tom de Vries <tom@codesourcery.com>
4637 * target.def (call_fusage_contains_non_callee_clobbers): New
4639 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
4641 (@node Miscellaneous Register Hooks): New node.
4642 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
4643 * doc/tm.texi: Regenerate.
4645 2014-05-05 Marek Polacek <polacek@redhat.com>
4648 * opts.c (common_handle_option): Call error_at instead of warning_at.
4650 2014-05-05 Richard Biener <rguenther@suse.de>
4652 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
4653 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
4654 under the TODO_verify_il umbrella.
4656 2014-05-05 Richard Biener <rguenther@suse.de>
4658 * passes.c (execute_function_todo): Move TODO_verify_flow under
4659 the TODO_verify_ul umbrella.
4661 2014-05-05 Richard Biener <rguenther@suse.de>
4664 * fold-const.c (fold_binary_loc): Consistently avoid canonicalizing
4665 X & CST away from a CST that is the mask of a mode.
4667 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4669 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
4670 int argument to enum machine_mode.
4671 (picochip_class_max_nregs): Ditto.
4672 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
4673 (picochip_class_max_nregs): Ditto.
4675 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4677 * target.def: Add new target hook.
4678 * doc/tm.texi: Regenerate.
4679 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
4680 * targhooks.c (default_keep_leaf_when_profiled): New function.
4682 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
4683 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
4685 2014-05-05 Bin Cheng <bin.cheng@arm.com>
4687 PR tree-optimization/60363
4688 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
4689 (copy_phi_args): New parameters. Call get_value_locus_in_path.
4690 (update_destination_phis): New parameter.
4691 (create_edge_and_update_destination_phis): Ditto.
4692 (ssa_fix_duplicate_block_edges): Pass new arguments.
4693 (thread_single_edge): Ditto.
4695 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
4697 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
4698 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
4699 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
4700 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
4701 Use RS6000_BTM_HARD_FLOAT.
4702 (BU_MISC_2): Likewise.
4703 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
4704 RS6000_BTM_HARD_FLOAT.
4705 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
4707 (rs6000_invalid_builtin): Add hard floating builtin support.
4708 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
4709 hard float builtins.
4710 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
4712 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4714 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
4715 Add missing function* argument.
4717 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
4719 * lra-constraints.c (valid_address_p): Move earlier in file.
4720 Add a constraint argument to the address_info version.
4721 (satisfies_memory_constraint_p): New function.
4722 (satisfies_address_constraint_p): Likewise.
4723 (process_alt_operands, curr_insn_transform): Use them.
4724 (process_address): Pass the constraint to valid_address_p when
4725 checking address operands.
4727 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
4729 * config/mips/mips.c (mips_isa_rev): New variable.
4730 (mips_set_architecture): Set it.
4731 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
4733 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
4734 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
4735 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
4736 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
4737 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
4738 conditions in terms of mips_isa_rev.
4739 (mips_isa_rev): Declare.
4741 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4743 * config/sh/sh-mem.cc: Use tabs instead of spaces.
4744 (prob_unlikely, prob_likely): Make variables const.
4746 2014-05-03 Denis Chertykov <chertykov@gmail.com>
4748 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
4750 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4752 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
4754 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4756 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
4757 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
4758 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
4760 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
4761 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
4763 Replace usage of ROUND_REG with sh_round_reg.
4764 Use CEIL instead of ROUND_ADVANCE.
4766 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4769 * config/sh/sh.c: Include stdlib headers before everything else.
4771 2014-05-02 Jakub Jelinek <jakub@redhat.com>
4773 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
4774 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
4775 (gimplify_adjust_omp_clauses): Simd region is never
4776 directly nested in combined parallel. Instead, for linear
4777 with copyin/copyout, if in combined for simd loop, make decl
4778 firstprivate/lastprivate on OMP_FOR.
4779 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
4780 expand_omp_for_static_chunk): When setting endvar, also set
4781 fd->loop.v to the same value.
4783 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
4785 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
4787 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
4789 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
4792 2014-05-02 Marek Polacek <polacek@redhat.com>
4794 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
4796 2014-05-02 Kito Cheng <kito@0xlab.org>
4798 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
4799 to a C expression marco.
4800 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
4801 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
4802 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
4803 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
4804 HONOR_REG_ALLOC_ORDER.
4805 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
4807 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4809 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
4811 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4813 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
4815 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
4817 * tree-if-conv.c (is_cond_scalar_reduction): New function.
4818 (convert_scalar_cond_reduction): Likewise.
4819 (predicate_scalar_phi): Add recognition and transformation
4820 of simple conditioanl reduction to be vectorizable.
4822 2014-05-01 Marek Polacek <polacek@redhat.com>
4825 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
4827 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
4829 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
4830 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
4831 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
4832 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
4833 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
4834 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
4835 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
4836 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
4838 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
4840 * config/arc/arc.opt (mlra): Move comment above option name
4841 to avoid mis-parsing as language options.
4843 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4845 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
4846 * config/sol2.h: ... here.
4847 * config/sol2-10.h: Remove.
4849 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
4850 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
4851 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
4852 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
4853 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
4854 * config/sol2.h: ... here.
4855 (SECTION_NAME_FORMAT): Don't redefine.
4856 (STARTFILE_ARCH32_SPEC): Rename to ...
4857 (STARTFILE_ARCH_SPEC): ... this.
4858 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
4859 * config/sparc/sol2.h: ... here.
4860 (SECTION_NAME_FORMAT): Don't undef.
4861 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
4862 (SUBTARGET_EXTRA_SPECS): Remove.
4863 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
4865 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
4866 (MD_STARTFILE_PREFIX): Remove.
4867 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
4868 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
4869 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
4870 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
4871 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
4872 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
4873 * config/i386/sol2.h: ... here.
4874 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
4875 * config/i386/sol2-bi.h: Remove.
4876 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
4877 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
4879 * config/i386/t-sol2-64: Rename to ...
4880 * config/i386/t-sol2: ... this.
4881 * config/sparc/t-sol2-64: Rename to ...
4882 * config/sparc/t-sol2: ... this.
4884 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
4885 sol2_tm_file_head, sol2_tm_file_tail.
4886 Include ${cpu_type}/sol2.h before sol2.h.
4888 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
4889 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
4890 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
4891 Reflect i386/t-sol2-64 renaming.
4892 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
4893 Reflect sparc/t-sol2-64 renaming.
4895 2014-04-30 Richard Biener <rguenther@suse.de>
4897 * passes.c (execute_function_todo): Move TODO_verify_stmts
4898 and TODO_verify_ssa under the TODO_verify_il umbrella.
4899 * tree-ssa.h (verify_ssa): Adjust prototype.
4900 * tree-ssa.c (verify_ssa): Add parameter to tell whether
4901 we should verify SSA operands.
4902 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
4903 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
4904 whether we should verify whether not throwing stmts have EH info.
4905 * graphite-scop-detection.c (create_sese_edges): Adjust.
4906 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
4907 * tree-eh.c (lower_try_finally_switch): Do not add the
4908 default case label twice.
4910 2014-04-30 Marek Polacek <polacek@redhat.com>
4912 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
4913 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
4914 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
4915 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
4917 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
4919 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
4920 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
4921 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
4922 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
4923 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
4924 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
4925 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
4926 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
4928 2014-04-29 David Malcolm <dmalcolm@redhat.com>
4930 * tree-cfg.c (dump_function_to_file): Dump the return type of
4931 functions, in a line to itself before the function body, mimicking
4932 the layout of a C function.
4934 2014-04-29 Jakub Jelinek <jakub@redhat.com>
4936 PR tree-optimization/60971
4937 * tree-tailcall.c (process_assignment): Reject conversions which
4940 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
4942 * calls.c (initialize_argument_information): Always treat
4943 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
4944 (expand_call): Likewise.
4945 (emit_library_call_calue_1): Likewise.
4946 * expr.c (PUSH_ARGS_REVERSED): Do not define.
4947 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
4950 2014-04-29 Nick Clifton <nickc@redhat.com>
4952 * config/msp430/msp430.md (umulsidi): Fix typo.
4953 (mulhisi3): Enable even inside interrupt handlers.
4954 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
4955 bigger return address pushed in large mode.
4957 2014-04-29 Nick Clifton <nickc@redhat.com>
4959 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
4960 (arc_init_reg_tables): Use a machine_mode enum to iterate over
4962 * config/m32r/m32r.c (init_reg_tables): Likewise.
4963 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
4964 enum to hold the modes.
4966 2014-04-29 Richard Biener <rguenther@suse.de>
4968 * dominance.c (free_dominance_info): Add overload with
4970 (dom_info_state): Likewise.
4971 (dom_info_available_p): Likewise.
4972 * basic-block.h (free_dominance_info, dom_info_state,
4973 dom_info_available_p): Declare overloads.
4974 * passes.c (execute_function_todo): Verify that verifiers
4975 don't change dominator info state. Drop dominator info
4976 for IPA pass invocations.
4977 * cgraph.c (release_function_body): Restore asserts that
4978 dominator information is released.
4980 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
4982 * doc/invoke.texi: Fix typo.
4983 * tree-vrp.c: Fix typos.
4984 * gimple.c (infer_nonnull_range): Reorder operands of an && condition.
4986 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
4988 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
4990 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
4992 * config/aarch64/aarch64-builtins.c
4993 (aarch64_types_storestruct_lane_qualifiers): New.
4994 (TYPES_STORESTRUCT_LANE): Likewise.
4995 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
4996 (st3_lane): Likewise.
4997 (st4_lane): Likewise.
4998 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
4999 (vec_store_lanesci_lane<mode>): Likewise.
5000 (vec_store_lanesxi_lane<mode>): Likewise.
5001 (aarch64_st2_lane<VQ:mode>): Likewise.
5002 (aarch64_st3_lane<VQ:mode>): Likewise.
5003 (aarch64_st4_lane<VQ:mode>): Likewise.
5004 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
5005 * config/aarch64/arm_neon.h
5006 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
5007 use new macro arguments.
5008 (__ST3_LANE_FUNC): Likewise.
5009 (__ST4_LANE_FUNC): Likewise.
5010 * config/aarch64/iterators.md (V_TWO_ELEM): New.
5011 (V_THREE_ELEM): Likewise.
5012 (V_FOUR_ELEM): Likewise.
5014 2014-04-28 David Malcolm <dmalcolm@redhat.com>
5016 * doc/gimple.texi: Replace the description of the now-defunct
5017 union gimple_statement_d with a diagram showing the
5018 gimple_statement_base class hierarchy and its relationships to
5019 the GSS_ and GIMPLE_ enums.
5021 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
5023 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
5024 * config/aarch64/aarch64.c
5025 (aarch64_cannot_change_mode_class): Weaken conditions.
5026 (aarch64_modes_tieable_p): New.
5027 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
5029 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
5031 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
5032 (loadsync_<mode>): Change mode.
5033 (load_quadpti, store_quadpti): New.
5034 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
5035 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
5037 2014-04-28 Martin Jambor <mjambor@suse.cz>
5039 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
5040 same alias type as the original statement.
5041 (subreplacement_assignment_data): New type.
5042 (handle_unscalarized_data_in_subtree): New type of parameter,
5043 generate new memory accesses with same alias type as the original
5045 (load_assign_lhs_subreplacements): Likewise.
5046 (sra_modify_constructor_assign): Generate new memory accesses with
5047 same alias type as the original statement.
5049 2014-04-28 Richard Biener <rguenther@suse.de>
5051 * tree-pass.h (TODO_verify_il): Define.
5052 (TODO_verify_all): Complete properly.
5053 * passes.c (execute_function_todo): Move existing loop-closed
5054 SSA verification under TODO_verify_il.
5055 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
5056 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
5057 Fix tree sharing issue.
5059 2014-04-28 Richard Biener <rguenther@suse.de>
5062 * builtins.def (DEF_C11_BUILTIN): Add.
5063 (BUILT_IN_ALIGNED_ALLOC): Likewise.
5064 * coretypes.h (enum function_class): Add function_c11_misc.
5065 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
5066 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
5067 (call_may_clobber_ref_p_1): Likewise.
5068 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
5069 (mark_all_reaching_defs_necessary_1): Likewise.
5070 (propagate_necessity): Likewise.
5071 (eliminate_unnecessary_stmts): Likewise.
5072 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
5074 2014-04-28 Richard Biener <rguenther@suse.de>
5076 * tree-vrp.c (vrp_var_may_overflow): Remove.
5077 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
5078 with overflow immediately bump to one before that value and
5079 let iteration figure out overflow status.
5081 2014-04-28 Richard Biener <rguenther@suse.de>
5083 * configure.ac: Do valgrind header checks unconditionally.
5084 Add --enable-valgrind-annotations.
5085 * system.h: Guard valgrind header inclusion with
5086 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
5087 * alloc-pool.c (pool_alloc, pool_free): Use
5088 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
5089 to guard possibly dead code.
5090 * config.in: Regenerated.
5091 * configure: Likewise.
5093 2014-04-28 Jeff Law <law@redhat.com>
5095 PR tree-optimization/60902
5096 * tree-ssa-threadedge.c
5097 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
5098 over real defs when invalidating outputs from statements that do not
5099 produce useful outputs for threading.
5101 2014-04-28 Richard Biener <rguenther@suse.de>
5103 PR tree-optimization/60979
5104 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
5105 SCOPs that end in a block with a successor with abnormal
5108 2014-04-28 Richard Biener <rguenther@suse.de>
5110 * tree-pass.h (execute_pass_list): Adjust prototype.
5111 * passes.c (pass_manager::execute_early_local_passes): Adjust.
5112 (do_per_function): Change callback signature, push all actual
5113 work to the callbals.
5114 (do_per_function_toporder): Likewise.
5115 (execute_function_dump): Adjust.
5116 (execute_function_todo): Likewise.
5117 (clear_last_verified): Likewise.
5118 (verify_curr_properties): Likewise.
5119 (update_properties_after_pass): Likewise.
5120 (execute_pass_list_1): Split out from ...
5121 (execute_pass_list): ... here. Adjust.
5122 (execute_ipa_pass_list): Likewise.
5123 * cgraphunit.c (cgraph_add_new_function): Adjust.
5124 (analyze_function): Likewise.
5125 (expand_function): Likewise.
5126 * cgraph.c (release_function_body): Free dominance info
5127 here instead of asserting it was magically freed elsewhere.
5129 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
5131 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
5132 * configure: Regenerate.
5133 * config/sparc/sparc.opt (muser-mode): New option.
5134 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
5136 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
5137 * doc/invoke.texi (SPARC options): Document -muser-mode.
5139 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
5141 * cselib.c (find_slot_memmode): Delete.
5142 (cselib_hasher): Change compare_type to a struct.
5143 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
5145 (preserve_constants_and_equivs): Adjust for new compare_type.
5146 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
5147 (wrap_constant): Delete.
5148 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
5150 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
5152 * doc/install.texi (Building with profile feedback): Remove
5155 2014-04-26 Tom de Vries <tom@codesourcery.com>
5157 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
5160 2014-04-25 Cary Coutant <ccoutant@google.com>
5163 * dwarf2out.c (should_move_die_to_comdat): A type definition
5164 can contain a subprogram definition, but don't move it to a
5166 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
5167 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
5169 (clone_tree_hash): Rename to...
5170 (clone_tree_partial): ...this; change callers. Copy
5171 DW_TAG_subprogram DIEs as declarations.
5172 (copy_decls_walk): Don't copy children of a declaration into a
5175 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
5178 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
5181 2014-04-25 Jiong Wang <jiong.wang@arm.com>
5183 * config/arm/predicates.md (call_insn_operand): Add long_call check.
5184 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
5186 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
5189 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5191 * config/arm/arm.c (arm_cortex_a8_tune): Initialise T16-related fields.
5193 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5195 PR tree-optimization/60930
5196 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
5197 creating a multiply candidate by folding two constant
5198 multiplicands when the result overflows.
5200 2014-04-25 Jakub Jelinek <jakub@redhat.com>
5202 PR tree-optimization/60960
5203 * tree-vect-generic.c (expand_vector_operation): Only call
5204 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
5206 2014-04-25 Tom de Vries <tom@codesourcery.com>
5208 * expr.c (clobber_reg_mode): New function.
5209 * expr.h (clobber_reg): New function.
5211 2014-04-25 Tom de Vries <tom@codesourcery.com>
5213 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
5216 2014-04-25 Radovan Obradovic <robradovic@mips.com>
5217 Tom de Vries <tom@codesourcery.com>
5219 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
5221 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
5222 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
5223 new argument to find_all_hard_reg_sets call.
5225 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5227 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
5228 Use HOST_WIDE_INT_C for mask literal.
5229 (aarch_rev16_shleft_mask_imm_p): Likewise.
5231 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
5234 * config/sparc/sparc.md (ashlsi3_extend): Delete.
5236 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
5238 PR preprocessor/56540
5239 * config/i386/i386-c.c (ix86_target_macros): Define
5240 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
5242 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5244 * configure.ac (tga_func): Remove.
5245 (LIB_TLS_SPEC): Remove.
5246 * configure: Regenerate.
5247 * config.in: Regenerate.
5248 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
5250 2014-04-25 Richard Biener <rguenther@suse.de>
5253 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
5254 call stmt use/clobber sets during stmt walk instead of
5255 walking the possibly incomplete set of caller edges.
5257 2014-04-25 Richard Biener <rguenther@suse.de>
5260 * passes.c (apply_ipa_transforms): Inline into only caller ...
5261 (execute_one_pass): ... here. Properly bring in function
5262 bodies for nodes we want to apply IPA transforms to.
5264 2014-04-24 Cong Hou <congh@google.com>
5266 PR tree-optimization/60896
5267 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
5268 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
5269 (vect_mark_pattern_stmts): Set the def type of all statements in
5270 PATTERN_DEF_SEQ as vect_internal_def.
5272 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
5274 * doc/extend.texi (PowerPC Built-in Functions): Document new
5275 powerpc extended divide, bcd, pack/unpack 128-bit, builtin functions.
5276 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
5278 * config/rs6000/predicates.md (const_0_to_3_operand): New
5279 predicate to match 0..3 integer constants.
5281 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
5282 to support adding miscellaneous builtin functions.
5283 (BU_DFP_MISC_2): Likewise.
5284 (BU_P7_MISC_1): Likewise.
5285 (BU_P7_MISC_2): Likewise.
5286 (BU_P8V_MISC_3): Likewise.
5287 (BU_MISC_1): Likewise.
5288 (BU_MISC_2): Likewise.
5289 (DIVWE): Add extended divide builtin functions.
5292 (DIVWEUO): Likewise.
5296 (DIVDEUO): Likewise.
5297 (DXEX): Add decimal floating-point builtin functions.
5300 (DDEDPDQ): Likewise.
5302 (DENBCDQ): Likewise.
5309 (CDTBCD): Add new BCD builtin functions.
5313 (BCDADD_LT): Likewise.
5314 (BCDADD_EQ): Likewise.
5315 (BCDADD_GT): Likewise.
5316 (BCDADD_OV): Likewise.
5318 (BCDSUB_LT): Likewise.
5319 (BCDSUB_EQ): Likewise.
5320 (BCDSUB_GT): Likewise.
5321 (BCDSUB_OV): Likewise.
5322 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
5323 (UNPACK_TD): Likewise.
5324 (PACK_TF): Likewise.
5325 (UNPACK_TF): Likewise.
5326 (UNPACK_TF_0): Likewise.
5327 (UNPACK_TF_1): Likewise.
5328 (PACK_V1TI): Likewise.
5329 (UNPACK_V1TI): Likewise.
5331 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
5332 support for decimal floating point builtin functions.
5333 (rs6000_expand_ternop_builtin): Add checks for the new builtin
5334 functions that take constant arguments.
5335 (rs6000_invalid_builtin): Add decimal floating point builtin support.
5336 (rs6000_init_builtins): Setup long double, _Decimal64, and
5337 _Decimal128 types for new builtin functions.
5338 (builtin_function_type): Set the unsigned flags appropriately for
5339 the new builtin functions.
5340 (rs6000_opt_masks): Add support for decimal floating point builtin
5343 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
5344 floating point builtin functions.
5345 (RS6000_BTM_COMMON): Likewise.
5346 (RS6000_BTI_long_double): Likewise.
5347 (RS6000_BTI_dfloat64): Likewise.
5348 (RS6000_BTI_dfloat128): Likewise.
5349 (long_double_type_internal_node): Likewise.
5350 (dfloat64_type_internal_node): Likewise.
5351 (dfloat128_type_internal_node): Likewise.
5353 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
5354 2.07 bcd arithmetic instructions.
5355 (UNSPEC_BCDSUB): Likewise.
5356 (UNSPEC_BCD_OVERFLOW): Likewise.
5357 (UNSPEC_BCD_ADD_SUB): Likewise.
5358 (bcd_add_sub): Likewise.
5359 (BCD_TEST): Likewise.
5360 (bcd<bcd_add_sub>): Likewise.
5361 (bcd<bcd_add_sub>_test): Likewise.
5362 (bcd<bcd_add_sub>_test2): Likewise.
5363 (bcd<bcd_add_sub>_<code>): Likewise.
5364 (peephole2 for combined bcd ops): Likewise.
5366 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
5367 decimal floating point builtin functions.
5368 (UNSPEC_DENBCD): Likewise.
5369 (UNSPEC_DXEX): Likewise.
5370 (UNSPEC_DIEX): Likewise.
5371 (UNSPEC_DSCLI): Likewise.
5372 (UNSPEC_DSCRI): Likewise.
5373 (D64_D128): Likewise.
5374 (dfp_suffix): Likewise.
5375 (dfp_ddedpd_<mode>): Likewise.
5376 (dfp_denbcd_<mode>): Likewise.
5377 (dfp_dxex_<mode>): Likewise.
5378 (dfp_diex_<mode>): Likewise.
5379 (dfp_dscli_<mode>): Likewise.
5380 (dfp_dscri_<mode>): Likewise.
5382 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
5384 (UNSPEC_CDTBCD): Likewise.
5385 (UNSPEC_CBCDTD): Likewise.
5386 (UNSPEC_DIVE): Add support for new extended divide builtin functions.
5387 (UNSPEC_DIVEO): Likewise.
5388 (UNSPEC_DIVEU): Likewise.
5389 (UNSPEC_DIVEUO): Likewise.
5390 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
5391 pack/unpack 128-bit types.
5392 (UNSPEC_PACK_128BIT): Likewise.
5393 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
5394 (udiv<mode>3): Use idiv_ldiv mode attribute.
5395 (div<mode>3): Likewise.
5396 (addg6s): Add new BCD builtin functions.
5399 (UNSPEC_DIV_EXTEND): Add support for new extended divide instructions.
5400 (div_extend): Likewise.
5401 (div<div_extend>_<mode>"): Likewise.
5402 (FP128_64): Add support for new builtin functions to pack/unpack
5404 (unpack<mode>): Likewise.
5405 (unpacktf_0): Likewise.
5406 (unpacktf_1): Likewise.
5407 (unpack<mode>_dm): Likewise.
5408 (unpack<mode>_nodm): Likewise.
5409 (pack<mode>): Likewise.
5410 (unpackv1ti): Likewise.
5411 (packv1ti): Likewise.
5413 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
5415 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
5418 2014-04-24 Jakub Jelinek <jakub@redhat.com>
5420 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
5421 * gimplify.c (omp_is_private): Change last argument's type to int.
5422 Only diagnose lastprivate if the simd argument is 1, only diagnose
5423 linear if the simd argument is 2.
5424 (gimplify_omp_for): Adjust omp_is_private callers. When adding
5425 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
5426 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
5427 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
5428 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
5429 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
5430 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
5431 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
5432 * tree-nested.c (convert_nonlocal_omp_clauses,
5433 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
5435 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
5438 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
5441 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
5443 * flag-types.h (enum ivar_visibility): Add.
5445 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
5447 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
5448 function * argument.
5450 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
5452 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
5454 2014-04-24 Radovan Obradovic <robradovic@mips.com>
5455 Tom de Vries <tom@codesourcery.com>
5457 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
5458 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
5460 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
5461 * emit-rtl.c (try_split): Same.
5463 2014-04-24 Radovan Obradovic <robradovic@mips.com>
5464 Tom de Vries <tom@codesourcery.com>
5466 * common.opt (fuse-caller-save): New option.
5468 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
5470 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of
5471 elements for big-endian.
5473 2014-04-24 Richard Biener <rguenther@suse.de>
5475 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
5476 during TER and instead use the sepops interface for expanding
5477 non-GIMPLE_SINGLE_RHS.
5479 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5481 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
5482 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
5484 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5486 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
5487 assembler 64-bit option.
5488 * configure: Regenerate.
5490 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5492 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
5493 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
5494 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
5495 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
5496 (TARGET_CRYPTO): Take TARGET_SIMD into account.
5498 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5500 * config/aarch64/aarch64-builtins.c
5501 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
5502 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
5503 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
5504 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
5506 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
5507 (Vrevsuff): New mode attribute.
5509 2014-04-24 Terry Guo <terry.guo@arm.com>
5511 * config/arm/arm.h (machine_function): Define variable
5512 after_arm_reorg here.
5513 * config/arm/arm.c (after_arm_reorg): Remove the definition.
5514 (arm_split_constant): Update the way to access variable
5517 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
5519 2014-04-23 Tom de Vries <tom@codesourcery.com>
5521 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
5523 2014-04-23 David Malcolm <dmalcolm@redhat.com>
5525 * is-a.h: Update comments to reflect the following changes to the
5526 "pointerness" of the API, making the template parameter match the
5527 return type, allowing use of is-a.h with typedefs of pointers.
5528 (is_a_helper::cast): Return a T rather then a pointer to a T, so
5529 that the return type matches the parameter to the is_a_helper.
5531 (dyn_cast): Likewise.
5533 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
5534 pointer from the is-a.h API.
5536 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
5537 (is_a_helper <cgraph_node *>::test): ...this, matching change to
5539 (is_a_helper <varpool_node>::test): Likewise, convert to...
5540 (is_a_helper <varpool_node *>::test): ...this.
5542 (varpool_first_variable): Update for removal of implicit pointer
5543 from the is-a.h API.
5544 (varpool_next_variable): Likewise.
5545 (varpool_first_static_initializer): Likewise.
5546 (varpool_next_static_initializer): Likewise.
5547 (varpool_first_defined_variable): Likewise.
5548 (varpool_next_defined_variable): Likewise.
5549 (cgraph_first_defined_function): Likewise.
5550 (cgraph_next_defined_function): Likewise.
5551 (cgraph_first_function): Likewise.
5552 (cgraph_next_function): Likewise.
5553 (cgraph_first_function_with_gimple_body): Likewise.
5554 (cgraph_next_function_with_gimple_body): Likewise.
5555 (cgraph_alias_target): Likewise.
5556 (varpool_alias_target): Likewise.
5557 (cgraph_function_or_thunk_node): Likewise.
5558 (varpool_variable_node): Likewise.
5559 (symtab_real_symbol_p): Likewise.
5560 * cgraphunit.c (referred_to_p): Likewise.
5561 (analyze_functions): Likewise.
5562 (handle_alias_pairs): Likewise.
5563 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
5564 * gimple-ssa.h (gimple_vuse_op): Likewise.
5565 (gimple_vdef_op): Likewise.
5566 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
5567 * gimple.c (gimple_build_asm_1): Likewise.
5568 (gimple_build_try): Likewise.
5569 (gimple_build_resx): Likewise.
5570 (gimple_build_eh_dispatch): Likewise.
5571 (gimple_build_omp_for): Likewise.
5572 (gimple_omp_for_set_clauses): Likewise.
5574 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
5575 (is_a_helper <gimple_statement_asm *>::test): ...this.
5576 (is_a_helper <gimple_statement_bind>::test): Convert to...
5577 (is_a_helper <gimple_statement_bind *>::test): ...this.
5578 (is_a_helper <gimple_statement_call>::test): Convert to...
5579 (is_a_helper <gimple_statement_call *>::test): ...this.
5580 (is_a_helper <gimple_statement_catch>::test): Convert to...
5581 (is_a_helper <gimple_statement_catch *>::test): ...this.
5582 (is_a_helper <gimple_statement_resx>::test): Convert to...
5583 (is_a_helper <gimple_statement_resx *>::test): ...this.
5584 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
5585 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
5586 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
5587 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
5588 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
5589 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
5590 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
5591 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
5592 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
5593 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
5594 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
5595 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
5596 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
5597 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
5598 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
5599 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
5600 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
5601 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
5602 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
5603 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
5604 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
5605 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
5606 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
5607 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
5608 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
5609 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
5610 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
5611 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
5612 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
5613 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
5614 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
5615 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
5616 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
5617 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
5618 (is_a_helper <gimple_statement_phi>::test): Convert to...
5619 (is_a_helper <gimple_statement_phi *>::test): ...this.
5620 (is_a_helper <gimple_statement_transaction>::test): Convert to...
5621 (is_a_helper <gimple_statement_transaction *>::test): ...this.
5622 (is_a_helper <gimple_statement_try>::test): Convert to...
5623 (is_a_helper <gimple_statement_try *>::test): ...this.
5624 (is_a_helper <gimple_statement_wce>::test): Convert to...
5625 (is_a_helper <gimple_statement_wce *>::test): ...this.
5626 (is_a_helper <const gimple_statement_asm>::test): Convert to...
5627 (is_a_helper <const gimple_statement_asm *>::test): ...this.
5628 (is_a_helper <const gimple_statement_bind>::test): Convert to...
5629 (is_a_helper <const gimple_statement_bind *>::test): ...this.
5630 (is_a_helper <const gimple_statement_call>::test): Convert to...
5631 (is_a_helper <const gimple_statement_call *>::test): ...this.
5632 (is_a_helper <const gimple_statement_catch>::test): Convert to...
5633 (is_a_helper <const gimple_statement_catch *>::test): ...this.
5634 (is_a_helper <const gimple_statement_resx>::test): Convert to...
5635 (is_a_helper <const gimple_statement_resx *>::test): ...this.
5636 (is_a_helper <const gimple_statement_eh_dispatch>::test): Convert to...
5637 (is_a_helper <const gimple_statement_eh_dispatch *>::test): ...this.
5638 (is_a_helper <const gimple_statement_eh_filter>::test): Convert to...
5639 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
5640 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
5642 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
5644 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
5646 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
5648 (is_a_helper <const gimple_statement_omp_return>::test): Convert to...
5649 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
5650 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
5652 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
5653 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
5655 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
5656 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
5657 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
5658 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert to...
5659 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
5660 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
5662 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
5663 (is_a_helper <const gimple_statement_omp_target>::test): Convert to...
5664 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
5665 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
5667 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
5668 (is_a_helper <const gimple_statement_omp_single>::test): Convert to...
5669 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
5670 (is_a_helper <const gimple_statement_omp_teams>::test): Convert to...
5671 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
5672 (is_a_helper <const gimple_statement_omp_task>::test): Convert to...
5673 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
5674 (is_a_helper <const gimple_statement_phi>::test): Convert to...
5675 (is_a_helper <const gimple_statement_phi *>::test): ...this.
5676 (is_a_helper <const gimple_statement_transaction>::test): Convert to...
5677 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
5678 (is_a_helper <const gimple_statement_with_ops>::test): Convert to...
5679 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
5680 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
5681 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
5682 (is_a_helper <const gimple_statement_with_memory_ops>::test): Convert
5684 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
5686 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert to...
5687 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
5689 (gimple_use_ops): Update for removal of implicit pointer from the
5691 (gimple_set_use_ops): Likewise.
5692 (gimple_vuse): Likewise.
5693 (gimple_vdef): Likewise.
5694 (gimple_vuse_ptr): Likewise.
5695 (gimple_vdef_ptr): Likewise.
5696 (gimple_set_vuse): Likewise.
5697 (gimple_set_vdef): Likewise.
5698 (gimple_omp_return_set_lhs): Likewise.
5699 (gimple_omp_return_lhs): Likewise.
5700 (gimple_omp_return_lhs_ptr): Likewise.
5701 (gimple_call_fntype): Likewise.
5702 (gimple_call_set_fntype): Likewise.
5703 (gimple_call_set_internal_fn): Likewise.
5704 (gimple_call_use_set): Likewise.
5705 (gimple_call_clobber_set): Likewise.
5706 (gimple_bind_vars): Likewise.
5707 (gimple_bind_set_vars): Likewise.
5708 (gimple_bind_body_ptr): Likewise.
5709 (gimple_bind_set_body): Likewise.
5710 (gimple_bind_add_stmt): Likewise.
5711 (gimple_bind_block): Likewise.
5712 (gimple_bind_set_block): Likewise.
5713 (gimple_asm_ninputs): Likewise.
5714 (gimple_asm_noutputs): Likewise.
5715 (gimple_asm_nclobbers): Likewise.
5716 (gimple_asm_nlabels): Likewise.
5717 (gimple_asm_input_op): Likewise.
5718 (gimple_asm_input_op_ptr): Likewise.
5719 (gimple_asm_output_op): Likewise.
5720 (gimple_asm_output_op_ptr): Likewise.
5721 (gimple_asm_set_output_op): Likewise.
5722 (gimple_asm_clobber_op): Likewise.
5723 (gimple_asm_set_clobber_op): Likewise.
5724 (gimple_asm_label_op): Likewise.
5725 (gimple_asm_set_label_op): Likewise.
5726 (gimple_asm_string): Likewise.
5727 (gimple_catch_types): Likewise.
5728 (gimple_catch_types_ptr): Likewise.
5729 (gimple_catch_handler_ptr): Likewise.
5730 (gimple_catch_set_types): Likewise.
5731 (gimple_catch_set_handler): Likewise.
5732 (gimple_eh_filter_types): Likewise.
5733 (gimple_eh_filter_types_ptr): Likewise.
5734 (gimple_eh_filter_failure_ptr): Likewise.
5735 (gimple_eh_filter_set_types): Likewise.
5736 (gimple_eh_filter_set_failure): Likewise.
5737 (gimple_eh_must_not_throw_fndecl): Likewise.
5738 (gimple_eh_must_not_throw_set_fndecl): Likewise.
5739 (gimple_eh_else_n_body_ptr): Likewise.
5740 (gimple_eh_else_e_body_ptr): Likewise.
5741 (gimple_eh_else_set_n_body): Likewise.
5742 (gimple_eh_else_set_e_body): Likewise.
5743 (gimple_try_eval_ptr): Likewise.
5744 (gimple_try_cleanup_ptr): Likewise.
5745 (gimple_try_set_eval): Likewise.
5746 (gimple_try_set_cleanup): Likewise.
5747 (gimple_wce_cleanup_ptr): Likewise.
5748 (gimple_wce_set_cleanup): Likewise.
5749 (gimple_phi_capacity): Likewise.
5750 (gimple_phi_num_args): Likewise.
5751 (gimple_phi_result): Likewise.
5752 (gimple_phi_result_ptr): Likewise.
5753 (gimple_phi_set_result): Likewise.
5754 (gimple_phi_arg): Likewise.
5755 (gimple_phi_set_arg): Likewise.
5756 (gimple_resx_region): Likewise.
5757 (gimple_resx_set_region): Likewise.
5758 (gimple_eh_dispatch_region): Likewise.
5759 (gimple_eh_dispatch_set_region): Likewise.
5760 (gimple_omp_critical_name): Likewise.
5761 (gimple_omp_critical_name_ptr): Likewise.
5762 (gimple_omp_critical_set_name): Likewise.
5763 (gimple_omp_for_clauses): Likewise.
5764 (gimple_omp_for_clauses_ptr): Likewise.
5765 (gimple_omp_for_set_clauses): Likewise.
5766 (gimple_omp_for_collapse): Likewise.
5767 (gimple_omp_for_index): Likewise.
5768 (gimple_omp_for_index_ptr): Likewise.
5769 (gimple_omp_for_set_index): Likewise.
5770 (gimple_omp_for_initial): Likewise.
5771 (gimple_omp_for_initial_ptr): Likewise.
5772 (gimple_omp_for_set_initial): Likewise.
5773 (gimple_omp_for_final): Likewise.
5774 (gimple_omp_for_final_ptr): Likewise.
5775 (gimple_omp_for_set_final): Likewise.
5776 (gimple_omp_for_incr): Likewise.
5777 (gimple_omp_for_incr_ptr): Likewise.
5778 (gimple_omp_for_set_incr): Likewise.
5779 (gimple_omp_for_pre_body_ptr): Likewise.
5780 (gimple_omp_for_set_pre_body): Likewise.
5781 (gimple_omp_parallel_clauses): Likewise.
5782 (gimple_omp_parallel_clauses_ptr): Likewise.
5783 (gimple_omp_parallel_set_clauses): Likewise.
5784 (gimple_omp_parallel_child_fn): Likewise.
5785 (gimple_omp_parallel_child_fn_ptr): Likewise.
5786 (gimple_omp_parallel_set_child_fn): Likewise.
5787 (gimple_omp_parallel_data_arg): Likewise.
5788 (gimple_omp_parallel_data_arg_ptr): Likewise.
5789 (gimple_omp_parallel_set_data_arg): Likewise.
5790 (gimple_omp_task_clauses): Likewise.
5791 (gimple_omp_task_clauses_ptr): Likewise.
5792 (gimple_omp_task_set_clauses): Likewise.
5793 (gimple_omp_task_child_fn): Likewise.
5794 (gimple_omp_task_child_fn_ptr): Likewise.
5795 (gimple_omp_task_set_child_fn): Likewise.
5796 (gimple_omp_task_data_arg): Likewise.
5797 (gimple_omp_task_data_arg_ptr): Likewise.
5798 (gimple_omp_task_set_data_arg): Likewise.
5799 (gimple_omp_taskreg_clauses): Likewise.
5800 (gimple_omp_taskreg_clauses_ptr): Likewise.
5801 (gimple_omp_taskreg_set_clauses): Likewise.
5802 (gimple_omp_taskreg_child_fn): Likewise.
5803 (gimple_omp_taskreg_child_fn_ptr): Likewise.
5804 (gimple_omp_taskreg_set_child_fn): Likewise.
5805 (gimple_omp_taskreg_data_arg): Likewise.
5806 (gimple_omp_taskreg_data_arg_ptr): Likewise.
5807 (gimple_omp_taskreg_set_data_arg): Likewise.
5808 (gimple_omp_task_copy_fn): Likewise.
5809 (gimple_omp_task_copy_fn_ptr): Likewise.
5810 (gimple_omp_task_set_copy_fn): Likewise.
5811 (gimple_omp_task_arg_size): Likewise.
5812 (gimple_omp_task_arg_size_ptr): Likewise.
5813 (gimple_omp_task_set_arg_size): Likewise.
5814 (gimple_omp_task_arg_align): Likewise.
5815 (gimple_omp_task_arg_align_ptr): Likewise.
5816 (gimple_omp_task_set_arg_align): Likewise.
5817 (gimple_omp_single_clauses): Likewise.
5818 (gimple_omp_single_clauses_ptr): Likewise.
5819 (gimple_omp_single_set_clauses): Likewise.
5820 (gimple_omp_target_clauses): Likewise.
5821 (gimple_omp_target_clauses_ptr): Likewise.
5822 (gimple_omp_target_set_clauses): Likewise.
5823 (gimple_omp_target_child_fn): Likewise.
5824 (gimple_omp_target_child_fn_ptr): Likewise.
5825 (gimple_omp_target_set_child_fn): Likewise.
5826 (gimple_omp_target_data_arg): Likewise.
5827 (gimple_omp_target_data_arg_ptr): Likewise.
5828 (gimple_omp_target_set_data_arg): Likewise.
5829 (gimple_omp_teams_clauses): Likewise.
5830 (gimple_omp_teams_clauses_ptr): Likewise.
5831 (gimple_omp_teams_set_clauses): Likewise.
5832 (gimple_omp_sections_clauses): Likewise.
5833 (gimple_omp_sections_clauses_ptr): Likewise.
5834 (gimple_omp_sections_set_clauses): Likewise.
5835 (gimple_omp_sections_control): Likewise.
5836 (gimple_omp_sections_control_ptr): Likewise.
5837 (gimple_omp_sections_set_control): Likewise.
5838 (gimple_omp_for_set_cond): Likewise.
5839 (gimple_omp_for_cond): Likewise.
5840 (gimple_omp_atomic_store_set_val): Likewise.
5841 (gimple_omp_atomic_store_val): Likewise.
5842 (gimple_omp_atomic_store_val_ptr): Likewise.
5843 (gimple_omp_atomic_load_set_lhs): Likewise.
5844 (gimple_omp_atomic_load_lhs): Likewise.
5845 (gimple_omp_atomic_load_lhs_ptr): Likewise.
5846 (gimple_omp_atomic_load_set_rhs): Likewise.
5847 (gimple_omp_atomic_load_rhs): Likewise.
5848 (gimple_omp_atomic_load_rhs_ptr): Likewise.
5849 (gimple_omp_continue_control_def): Likewise.
5850 (gimple_omp_continue_control_def_ptr): Likewise.
5851 (gimple_omp_continue_set_control_def): Likewise.
5852 (gimple_omp_continue_control_use): Likewise.
5853 (gimple_omp_continue_control_use_ptr): Likewise.
5854 (gimple_omp_continue_set_control_use): Likewise.
5855 (gimple_transaction_body_ptr): Likewise.
5856 (gimple_transaction_label): Likewise.
5857 (gimple_transaction_label_ptr): Likewise.
5858 (gimple_transaction_set_body): Likewise.
5859 (gimple_transaction_set_label): Likewise.
5861 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
5862 * ipa-inline-analysis.c (inline_write_summary): Likewise.
5863 * ipa-ref.c (ipa_record_reference): Likewise.
5864 * ipa-reference.c (analyze_function): Likewise.
5865 (ipa_reference_write_optimization_summary): Likewise.
5866 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
5867 (address_taken_from_non_vtable_p): Likewise.
5868 (comdat_can_be_unshared_p_1): Likewise.
5869 * lto-cgraph.c (lto_output_ref): Likewise.
5870 (add_references): Likewise.
5871 (compute_ltrans_boundary): Likewise.
5872 (output_symtab): Likewise.
5873 (input_ref): Likewise.
5874 (input_cgraph_1): Likewise.
5875 (output_cgraph_opt_summary): Likewise.
5876 * lto-streamer-out.c (lto_output): Likewise.
5877 (output_symbol_p): Likewise.
5878 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
5879 (lsei_start_function_in_partition): Likewise.
5880 (lsei_next_variable_in_partition): Likewise.
5881 (lsei_start_variable_in_partition): Likewise.
5882 * symtab.c (insert_to_assembler_name_hash): Likewise.
5883 (unlink_from_assembler_name_hash): Likewise.
5884 (symtab_unregister_node): Likewise.
5885 (symtab_remove_node): Likewise.
5886 (dump_symtab_node): Likewise.
5887 (verify_symtab_base): Likewise.
5888 (verify_symtab_node): Likewise.
5889 (symtab_make_decl_local): Likewise.
5890 (symtab_alias_ultimate_target): Likewise.
5891 (symtab_resolve_alias): Likewise.
5892 (symtab_get_symbol_partitioning_class): Likewise.
5893 * tree-phinodes.c (allocate_phi_node): Likewise.
5894 (reserve_phi_args_for_new_edge): Likewise.
5895 (remove_phi_args): Likewise.
5896 * varpool.c (varpool_node_for_asm): Likewise.
5897 (varpool_remove_unreferenced_decls): Likewise.
5899 2014-04-23 Jeff Law <law@redhat.com>
5901 PR tree-optimization/60902
5902 * tree-ssa-threadedge.c
5903 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
5904 invalidate outputs from statements that do not produce useful
5905 outputs for threading.
5907 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
5909 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
5910 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
5911 machine descriptions for Stack Smashing Protector.
5913 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
5915 * aarch64.md (<optab>_rol<mode>3): New pattern.
5916 (<optab>_rolsi3_uxtw): Likewise.
5917 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
5919 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
5921 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
5922 (arm_cortex_a12_tune): Likewise.
5924 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5926 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
5928 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5930 * config/arm/arm.md (arm_rev16si2): New pattern.
5931 (arm_rev16si2_alt): Likewise.
5932 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
5934 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5936 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
5937 (rev16<mode>2_alt): Likewise.
5938 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
5939 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
5940 (aarch_rev16_shleft_mask_imm_p): Likewise.
5941 (aarch_rev16_p_1): Likewise.
5942 (aarch_rev16_p): Likewise.
5943 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
5944 (aarch_rev16_shright_mask_imm_p): Likewise.
5945 (aarch_rev16_shleft_mask_imm_p): Likewise.
5947 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5949 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
5950 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
5952 (cortex_a53_extra_costs): Likewise.
5953 (cortex_a57_extra_costs): Likewise.
5954 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
5955 (cortexa7_extra_costs): Likewise.
5956 (cortexa8_extra_costs): Likewise.
5957 (cortexa12_extra_costs): Likewise.
5958 (cortexa15_extra_costs): Likewise.
5959 (v7m_extra_costs): Likewise.
5960 (arm_new_rtx_costs): Handle BSWAP.
5962 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5964 * config/arm/arm.c (cortexa8_extra_costs): New table.
5965 (arm_cortex_a8_tune): New tuning struct.
5966 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
5968 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5970 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
5972 2014-04-23 Richard Biener <rguenther@suse.de>
5974 * Makefile.in (OBJS): Remove loop-unswitch.o.
5975 * tree-pass.h (make_pass_rtl_unswitch): Remove.
5976 * passes.def (pass_rtl_unswitch): Likewise.
5977 * loop-init.c (gate_rtl_unswitch): Likewise.
5978 (rtl_unswitch): Likewise.
5979 (pass_data_rtl_unswitch): Likewise.
5980 (pass_rtl_unswitch): Likewise.
5981 (make_pass_rtl_unswitch): Likewise.
5982 * rtl.h (reversed_condition): Likewise.
5983 (compare_and_jump_seq): Likewise.
5984 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
5986 * loop-unroll.c (compare_and_jump_seq): Likewise.
5988 2014-04-23 Richard Biener <rguenther@suse.de>
5990 PR tree-optimization/60903
5991 * tree-ssa-loop-im.c (analyze_memory_references): Remove
5992 commented code block.
5993 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
5994 loop flags to newly created BBs and edges.
5996 2014-04-23 Nick Clifton <nickc@redhat.com>
5998 * config/msp430/msp430.c (msp430_handle_option): Move function
6000 (msp430_option_override): Simplify mcu and mcpu option handling.
6001 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
6002 support for -mhwmult command line option.
6003 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
6004 -mhwmult command line option.
6005 (msp430_hwmult_enabled): Delete.
6006 (msp43o_output_labelref): Add support for -mhwmult command line option.
6007 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
6008 (umulsidi3): Likewise.
6009 * config/msp430/msp430.opt (mmcu): Add Report attribute.
6010 (mcpu, mlarge, msmall): Likewise.
6011 (mhwmult): New option.
6012 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
6014 (msp430_is_f5_mcu): Remove prototype.
6015 (msp430_use_f5_series_hwmult): Add prototype.
6016 * config/msp430/msp430-opts.h: New file.
6017 * common/config/msp430: New directory.
6018 * common/config/msp430/msp430-common.c: New file.
6019 * config.gcc (msp430): Remove target_has_targetm_common.
6020 * doc/invoke.texi: Document -mhwmult command line option.
6022 2014-04-23 Nick Clifton <nickc@redhat.com>
6024 * config/i386/cygwin.h (ENDFILE_SPEC): Include
6025 default-manifest.o if it can be found in the search path.
6026 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
6028 2014-04-23 Terry Guo <terry.guo@arm.com>
6030 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
6032 2014-04-23 Richard Biener <rguenther@suse.de>
6035 * tree-inline.c (declare_return_variable): Use mark_addressable.
6037 2014-04-23 Richard Biener <rguenther@suse.de>
6040 * loop-init.c (loop_optimizer_init): Make sure to apply
6041 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
6043 2014-04-22 Jakub Jelinek <jakub@redhat.com>
6046 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
6048 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
6049 if flag_sanitize_undefined_trap_on_error.
6050 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
6051 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
6052 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
6053 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
6054 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
6055 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
6056 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
6057 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
6058 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
6059 * ubsan.c (ubsan_instrument_unreachable): Return
6060 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
6061 (ubsan_expand_null_ifn): Emit __builtin_trap ()
6062 if flag_sanitize_undefined_trap_on_error and
6063 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
6064 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
6065 instrument_bool_enum_load): Emit __builtin_trap () if
6066 flag_sanitize_undefined_trap_on_error and
6067 __builtin_handle_*_abort () if !flag_sanitize_recover.
6068 * doc/invoke.texi (-fsanitize-recover,
6069 -fsanitize-undefined-trap-on-error): Document.
6071 2014-04-22 Christian Bruel <christian.bruel@st.com>
6073 * config/sh/sh.md (mov<mode>): Replace movQIHI.
6074 Force immediates to SImode.
6076 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
6078 * config/nios2/nios2.md (UNSPEC_ROUND): New.
6080 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
6081 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
6082 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
6083 (nios2_fpu_insn): Add entry for round.
6084 (N2FPU_NO_ERRNO_P): Define.
6085 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
6087 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
6089 2014-04-22 Richard Henderson <rth@redhat.com>
6091 * config/aarch64/aarch64 (addti3, subti3): New expanders.
6092 (add<GPI>3_compare0): Remove leading * from name.
6093 (add<GPI>3_carryin): Likewise.
6094 (sub<GPI>3_compare0): Likewise.
6095 (sub<GPI>3_carryin): Likewise.
6096 (<su_optab>mulditi3): New expander.
6097 (multi3): New expander.
6098 (madd<GPI>): Remove leading * from name.
6100 2014-04-22 Martin Jambor <mjambor@suse.cz>
6102 * cgraphclones.c (cgraph_function_versioning): Copy
6103 ipa_transforms_to_apply instead of asserting it is empty.
6105 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
6108 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
6109 on count_exp to get mode.
6111 2014-04-22 Andrew Pinski <apinski@cavium.com>
6113 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
6114 Handle TLS for ILP32.
6115 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
6116 (tlsie_small_<mode>): this and handle PTR.
6117 (tlsie_small_sidi): New pattern.
6118 (tlsle_small): Change to an expand to handle ILP32.
6119 (tlsle_small_<mode>): New pattern.
6120 (tlsdesc_small): Rename to ...
6121 (tlsdesc_small_<mode>): this and handle PTR.
6123 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6125 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
6127 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
6129 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
6130 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
6131 (aarch64_types_signed_poly_qualifiers): Likewise.
6132 (aarch64_types_unsigned_signed_qualifiers): Likewise.
6133 (aarch64_types_poly_signed_qualifiers): Likewise.
6134 (TYPES_REINTERP_SS): Type macro added.
6135 (TYPES_REINTERP_SU): Likewise.
6136 (TYPES_REINTERP_SP): Likewise.
6137 (TYPES_REINTERP_US): Likewise.
6138 (TYPES_REINTERP_PS): Likewise.
6139 (aarch64_fold_builtin): New expression folding added.
6140 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
6141 Declarations removed.
6142 (REINTERP_SS): Declarations added.
6143 (REINTERP_US): Likewise.
6144 (REINTERP_PS): Likewise.
6145 (REINTERP_SU): Likewise.
6146 (REINTERP_SP): Likewise.
6147 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
6148 (vreinterpretq_p8_f64): Likewise.
6149 (vreinterpret_p16_f64): Likewise.
6150 (vreinterpretq_p16_f64): Likewise.
6151 (vreinterpret_f32_f64): Likewise.
6152 (vreinterpretq_f32_f64): Likewise.
6153 (vreinterpret_f64_f32): Likewise.
6154 (vreinterpret_f64_p8): Likewise.
6155 (vreinterpret_f64_p16): Likewise.
6156 (vreinterpret_f64_s8): Likewise.
6157 (vreinterpret_f64_s16): Likewise.
6158 (vreinterpret_f64_s32): Likewise.
6159 (vreinterpret_f64_s64): Likewise.
6160 (vreinterpret_f64_u8): Likewise.
6161 (vreinterpret_f64_u16): Likewise.
6162 (vreinterpret_f64_u32): Likewise.
6163 (vreinterpret_f64_u64): Likewise.
6164 (vreinterpretq_f64_f32): Likewise.
6165 (vreinterpretq_f64_p8): Likewise.
6166 (vreinterpretq_f64_p16): Likewise.
6167 (vreinterpretq_f64_s8): Likewise.
6168 (vreinterpretq_f64_s16): Likewise.
6169 (vreinterpretq_f64_s32): Likewise.
6170 (vreinterpretq_f64_s64): Likewise.
6171 (vreinterpretq_f64_u8): Likewise.
6172 (vreinterpretq_f64_u16): Likewise.
6173 (vreinterpretq_f64_u32): Likewise.
6174 (vreinterpretq_f64_u64): Likewise.
6175 (vreinterpret_s64_f64): Likewise.
6176 (vreinterpretq_s64_f64): Likewise.
6177 (vreinterpret_u64_f64): Likewise.
6178 (vreinterpretq_u64_f64): Likewise.
6179 (vreinterpret_s8_f64): Likewise.
6180 (vreinterpretq_s8_f64): Likewise.
6181 (vreinterpret_s16_f64): Likewise.
6182 (vreinterpretq_s16_f64): Likewise.
6183 (vreinterpret_s32_f64): Likewise.
6184 (vreinterpretq_s32_f64): Likewise.
6185 (vreinterpret_u8_f64): Likewise.
6186 (vreinterpretq_u8_f64): Likewise.
6187 (vreinterpret_u16_f64): Likewise.
6188 (vreinterpretq_u16_f64): Likewise.
6189 (vreinterpret_u32_f64): Likewise.
6190 (vreinterpretq_u32_f64): Likewise.
6192 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
6194 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
6195 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
6196 (vreinterpret_p8_s8): Likewise.
6197 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
6198 (vreinterpret_p8_s16): Likewise.
6199 (vreinterpret_p8_s32): Likewise.
6200 (vreinterpret_p8_s64): Likewise.
6201 (vreinterpret_p8_f32): Likewise.
6202 (vreinterpret_p8_u8): Likewise.
6203 (vreinterpret_p8_u16): Likewise.
6204 (vreinterpret_p8_u32): Likewise.
6205 (vreinterpret_p8_u64): Likewise.
6206 (vreinterpret_p8_p16): Likewise.
6207 (vreinterpretq_p8_s8): Likewise.
6208 (vreinterpretq_p8_s16): Likewise.
6209 (vreinterpretq_p8_s32): Likewise.
6210 (vreinterpretq_p8_s64): Likewise.
6211 (vreinterpretq_p8_f32): Likewise.
6212 (vreinterpretq_p8_u8): Likewise.
6213 (vreinterpretq_p8_u16): Likewise.
6214 (vreinterpretq_p8_u32): Likewise.
6215 (vreinterpretq_p8_u64): Likewise.
6216 (vreinterpretq_p8_p16): Likewise.
6217 (vreinterpret_p16_s8): Likewise.
6218 (vreinterpret_p16_s16): Likewise.
6219 (vreinterpret_p16_s32): Likewise.
6220 (vreinterpret_p16_s64): Likewise.
6221 (vreinterpret_p16_f32): Likewise.
6222 (vreinterpret_p16_u8): Likewise.
6223 (vreinterpret_p16_u16): Likewise.
6224 (vreinterpret_p16_u32): Likewise.
6225 (vreinterpret_p16_u64): Likewise.
6226 (vreinterpret_p16_p8): Likewise.
6227 (vreinterpretq_p16_s8): Likewise.
6228 (vreinterpretq_p16_s16): Likewise.
6229 (vreinterpretq_p16_s32): Likewise.
6230 (vreinterpretq_p16_s64): Likewise.
6231 (vreinterpretq_p16_f32): Likewise.
6232 (vreinterpretq_p16_u8): Likewise.
6233 (vreinterpretq_p16_u16): Likewise.
6234 (vreinterpretq_p16_u32): Likewise.
6235 (vreinterpretq_p16_u64): Likewise.
6236 (vreinterpretq_p16_p8): Likewise.
6237 (vreinterpret_f32_s8): Likewise.
6238 (vreinterpret_f32_s16): Likewise.
6239 (vreinterpret_f32_s32): Likewise.
6240 (vreinterpret_f32_s64): Likewise.
6241 (vreinterpret_f32_u8): Likewise.
6242 (vreinterpret_f32_u16): Likewise.
6243 (vreinterpret_f32_u32): Likewise.
6244 (vreinterpret_f32_u64): Likewise.
6245 (vreinterpret_f32_p8): Likewise.
6246 (vreinterpret_f32_p16): Likewise.
6247 (vreinterpretq_f32_s8): Likewise.
6248 (vreinterpretq_f32_s16): Likewise.
6249 (vreinterpretq_f32_s32): Likewise.
6250 (vreinterpretq_f32_s64): Likewise.
6251 (vreinterpretq_f32_u8): Likewise.
6252 (vreinterpretq_f32_u16): Likewise.
6253 (vreinterpretq_f32_u32): Likewise.
6254 (vreinterpretq_f32_u64): Likewise.
6255 (vreinterpretq_f32_p8): Likewise.
6256 (vreinterpretq_f32_p16): Likewise.
6257 (vreinterpret_s64_s8): Likewise.
6258 (vreinterpret_s64_s16): Likewise.
6259 (vreinterpret_s64_s32): Likewise.
6260 (vreinterpret_s64_f32): Likewise.
6261 (vreinterpret_s64_u8): Likewise.
6262 (vreinterpret_s64_u16): Likewise.
6263 (vreinterpret_s64_u32): Likewise.
6264 (vreinterpret_s64_u64): Likewise.
6265 (vreinterpret_s64_p8): Likewise.
6266 (vreinterpret_s64_p16): Likewise.
6267 (vreinterpretq_s64_s8): Likewise.
6268 (vreinterpretq_s64_s16): Likewise.
6269 (vreinterpretq_s64_s32): Likewise.
6270 (vreinterpretq_s64_f32): Likewise.
6271 (vreinterpretq_s64_u8): Likewise.
6272 (vreinterpretq_s64_u16): Likewise.
6273 (vreinterpretq_s64_u32): Likewise.
6274 (vreinterpretq_s64_u64): Likewise.
6275 (vreinterpretq_s64_p8): Likewise.
6276 (vreinterpretq_s64_p16): Likewise.
6277 (vreinterpret_u64_s8): Likewise.
6278 (vreinterpret_u64_s16): Likewise.
6279 (vreinterpret_u64_s32): Likewise.
6280 (vreinterpret_u64_s64): Likewise.
6281 (vreinterpret_u64_f32): Likewise.
6282 (vreinterpret_u64_u8): Likewise.
6283 (vreinterpret_u64_u16): Likewise.
6284 (vreinterpret_u64_u32): Likewise.
6285 (vreinterpret_u64_p8): Likewise.
6286 (vreinterpret_u64_p16): Likewise.
6287 (vreinterpretq_u64_s8): Likewise.
6288 (vreinterpretq_u64_s16): Likewise.
6289 (vreinterpretq_u64_s32): Likewise.
6290 (vreinterpretq_u64_s64): Likewise.
6291 (vreinterpretq_u64_f32): Likewise.
6292 (vreinterpretq_u64_u8): Likewise.
6293 (vreinterpretq_u64_u16): Likewise.
6294 (vreinterpretq_u64_u32): Likewise.
6295 (vreinterpretq_u64_p8): Likewise.
6296 (vreinterpretq_u64_p16): Likewise.
6297 (vreinterpret_s8_s16): Likewise.
6298 (vreinterpret_s8_s32): Likewise.
6299 (vreinterpret_s8_s64): Likewise.
6300 (vreinterpret_s8_f32): Likewise.
6301 (vreinterpret_s8_u8): Likewise.
6302 (vreinterpret_s8_u16): Likewise.
6303 (vreinterpret_s8_u32): Likewise.
6304 (vreinterpret_s8_u64): Likewise.
6305 (vreinterpret_s8_p8): Likewise.
6306 (vreinterpret_s8_p16): Likewise.
6307 (vreinterpretq_s8_s16): Likewise.
6308 (vreinterpretq_s8_s32): Likewise.
6309 (vreinterpretq_s8_s64): Likewise.
6310 (vreinterpretq_s8_f32): Likewise.
6311 (vreinterpretq_s8_u8): Likewise.
6312 (vreinterpretq_s8_u16): Likewise.
6313 (vreinterpretq_s8_u32): Likewise.
6314 (vreinterpretq_s8_u64): Likewise.
6315 (vreinterpretq_s8_p8): Likewise.
6316 (vreinterpretq_s8_p16): Likewise.
6317 (vreinterpret_s16_s8): Likewise.
6318 (vreinterpret_s16_s32): Likewise.
6319 (vreinterpret_s16_s64): Likewise.
6320 (vreinterpret_s16_f32): Likewise.
6321 (vreinterpret_s16_u8): Likewise.
6322 (vreinterpret_s16_u16): Likewise.
6323 (vreinterpret_s16_u32): Likewise.
6324 (vreinterpret_s16_u64): Likewise.
6325 (vreinterpret_s16_p8): Likewise.
6326 (vreinterpret_s16_p16): Likewise.
6327 (vreinterpretq_s16_s8): Likewise.
6328 (vreinterpretq_s16_s32): Likewise.
6329 (vreinterpretq_s16_s64): Likewise.
6330 (vreinterpretq_s16_f32): Likewise.
6331 (vreinterpretq_s16_u8): Likewise.
6332 (vreinterpretq_s16_u16): Likewise.
6333 (vreinterpretq_s16_u32): Likewise.
6334 (vreinterpretq_s16_u64): Likewise.
6335 (vreinterpretq_s16_p8): Likewise.
6336 (vreinterpretq_s16_p16): Likewise.
6337 (vreinterpret_s32_s8): Likewise.
6338 (vreinterpret_s32_s16): Likewise.
6339 (vreinterpret_s32_s64): Likewise.
6340 (vreinterpret_s32_f32): Likewise.
6341 (vreinterpret_s32_u8): Likewise.
6342 (vreinterpret_s32_u16): Likewise.
6343 (vreinterpret_s32_u32): Likewise.
6344 (vreinterpret_s32_u64): Likewise.
6345 (vreinterpret_s32_p8): Likewise.
6346 (vreinterpret_s32_p16): Likewise.
6347 (vreinterpretq_s32_s8): Likewise.
6348 (vreinterpretq_s32_s16): Likewise.
6349 (vreinterpretq_s32_s64): Likewise.
6350 (vreinterpretq_s32_f32): Likewise.
6351 (vreinterpretq_s32_u8): Likewise.
6352 (vreinterpretq_s32_u16): Likewise.
6353 (vreinterpretq_s32_u32): Likewise.
6354 (vreinterpretq_s32_u64): Likewise.
6355 (vreinterpretq_s32_p8): Likewise.
6356 (vreinterpretq_s32_p16): Likewise.
6357 (vreinterpret_u8_s8): Likewise.
6358 (vreinterpret_u8_s16): Likewise.
6359 (vreinterpret_u8_s32): Likewise.
6360 (vreinterpret_u8_s64): Likewise.
6361 (vreinterpret_u8_f32): Likewise.
6362 (vreinterpret_u8_u16): Likewise.
6363 (vreinterpret_u8_u32): Likewise.
6364 (vreinterpret_u8_u64): Likewise.
6365 (vreinterpret_u8_p8): Likewise.
6366 (vreinterpret_u8_p16): Likewise.
6367 (vreinterpretq_u8_s8): Likewise.
6368 (vreinterpretq_u8_s16): Likewise.
6369 (vreinterpretq_u8_s32): Likewise.
6370 (vreinterpretq_u8_s64): Likewise.
6371 (vreinterpretq_u8_f32): Likewise.
6372 (vreinterpretq_u8_u16): Likewise.
6373 (vreinterpretq_u8_u32): Likewise.
6374 (vreinterpretq_u8_u64): Likewise.
6375 (vreinterpretq_u8_p8): Likewise.
6376 (vreinterpretq_u8_p16): Likewise.
6377 (vreinterpret_u16_s8): Likewise.
6378 (vreinterpret_u16_s16): Likewise.
6379 (vreinterpret_u16_s32): Likewise.
6380 (vreinterpret_u16_s64): Likewise.
6381 (vreinterpret_u16_f32): Likewise.
6382 (vreinterpret_u16_u8): Likewise.
6383 (vreinterpret_u16_u32): Likewise.
6384 (vreinterpret_u16_u64): Likewise.
6385 (vreinterpret_u16_p8): Likewise.
6386 (vreinterpret_u16_p16): Likewise.
6387 (vreinterpretq_u16_s8): Likewise.
6388 (vreinterpretq_u16_s16): Likewise.
6389 (vreinterpretq_u16_s32): Likewise.
6390 (vreinterpretq_u16_s64): Likewise.
6391 (vreinterpretq_u16_f32): Likewise.
6392 (vreinterpretq_u16_u8): Likewise.
6393 (vreinterpretq_u16_u32): Likewise.
6394 (vreinterpretq_u16_u64): Likewise.
6395 (vreinterpretq_u16_p8): Likewise.
6396 (vreinterpretq_u16_p16): Likewise.
6397 (vreinterpret_u32_s8): Likewise.
6398 (vreinterpret_u32_s16): Likewise.
6399 (vreinterpret_u32_s32): Likewise.
6400 (vreinterpret_u32_s64): Likewise.
6401 (vreinterpret_u32_f32): Likewise.
6402 (vreinterpret_u32_u8): Likewise.
6403 (vreinterpret_u32_u16): Likewise.
6404 (vreinterpret_u32_u64): Likewise.
6405 (vreinterpret_u32_p8): Likewise.
6406 (vreinterpret_u32_p16): Likewise.
6407 (vreinterpretq_u32_s8): Likewise.
6408 (vreinterpretq_u32_s16): Likewise.
6409 (vreinterpretq_u32_s32): Likewise.
6410 (vreinterpretq_u32_s64): Likewise.
6411 (vreinterpretq_u32_f32): Likewise.
6412 (vreinterpretq_u32_u8): Likewise.
6413 (vreinterpretq_u32_u16): Likewise.
6414 (vreinterpretq_u32_u64): Likewise.
6415 (vreinterpretq_u32_p8): Likewise.
6416 (vreinterpretq_u32_p16): Likewise.
6418 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
6420 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
6422 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator extended.
6424 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
6425 (vqnegd_s64): Likewise.
6426 (vqabs_s64): Likewise.
6427 (vqabsd_s64): Likewise.
6429 2014-04-22 Richard Henderson <rth@redhat.com>
6431 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
6432 computation to the top of the loop.
6434 2014-04-22 Renlin <renlin.li@arm.com>
6435 Jiong Wang <jiong.wang@arm.com>
6437 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
6438 * config/aarch64/aarch64.c (aarch64_layout_frame)
6439 (aarch64_initial_elimination_offset): Likewise.
6441 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
6443 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
6446 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
6448 * machmode.h (bitwise_mode_for_mode): Declare.
6449 * stor-layout.h (bitwise_type_for_mode): Likewise.
6450 * stor-layout.c (bitwise_mode_for_mode): New function.
6451 (bitwise_type_for_mode): Likewise.
6452 * builtins.c (fold_builtin_memory_op): Use it instead of
6453 int_mode_for_mode and build_nonstandard_integer_type.
6455 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6457 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
6458 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
6459 (*-*-solaris2*): Simplify.
6460 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
6461 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
6462 *-*-solaris2.9* handling.
6464 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
6466 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
6467 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
6469 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
6470 * configure: Regenerate.
6472 * config/i386/sol2-9.h: Remove.
6474 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
6475 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
6476 Remove Solaris 9 references.
6478 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
6480 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
6481 (floatuns<GPI:mode><GPF:mode>2): Remove.
6482 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
6483 and floatuns conversions.
6484 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
6485 and floatuns conversions.
6486 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
6487 (w1,w2): New mode attributes for inequal width conversions.
6489 2014-04-22 Renlin Li <Renlin.Li@arm.com>
6491 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
6492 the output asm format.
6494 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
6496 * config/aarch64/aarch64-simd.md
6497 (aarch64_cm<optab>di): Always split.
6498 (*aarch64_cm<optab>di): New.
6499 (aarch64_cmtstdi): Always split.
6500 (*aarch64_cmtstdi): New.
6502 2014-04-22 Jakub Jelinek <jakub@redhat.com>
6504 PR tree-optimization/60823
6505 * omp-low.c (ipa_simd_modify_function_body): Go through
6506 all SSA_NAMEs and for those refering to vector arguments
6507 which are going to be replaced adjust SSA_NAME_VAR and,
6508 if it is a default definition, change it into a non-default
6509 definition assigned at the beginning of function from new_decl.
6510 (ipa_simd_modify_stmt_ops): Rewritten.
6511 * tree-dfa.c (set_ssa_default_def): When removing default def,
6512 check for NULL loc instead of NULL *loc.
6514 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6516 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
6517 restrictions on core registers for DImode values in Thumb2.
6519 2014-04-22 Ian Bolton <ian.bolton@arm.com>
6521 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
6522 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
6524 2014-04-22 Ian Bolton <ian.bolton@arm.com>
6526 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
6527 (*iordi_notzesidi_di): Likewise.
6528 (*iordi_notsesidi_di): Likewise.
6530 2014-04-22 Ian Bolton <ian.bolton@arm.com>
6532 * config/arm/arm-protos.h (tune_params): New struct members.
6533 * config/arm/arm.c: Initialise tune_params per processor.
6534 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
6535 for speed, based on new tune_params.
6537 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
6539 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro added.
6540 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added macro.
6541 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment corrected.
6542 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
6543 * config/aarch64/arm_neon.h (vrnd_f64): Added.
6544 (vrnda_f64): Likewise.
6545 (vrndi_f64): Likewise.
6546 (vrndm_f64): Likewise.
6547 (vrndn_f64): Likewise.
6548 (vrndp_f64): Likewise.
6549 (vrndx_f64): Likewise.
6551 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6553 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
6554 GET_MODE_SIZE argument is enum machine_mode.
6556 2014-04-22 Jakub Jelinek <jakub@redhat.com>
6559 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
6560 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
6562 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
6565 * asan.c (asan_emit_stack_protection): Force the base to align to
6566 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
6567 appropriate bits if STRICT_ALIGNMENT.
6568 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
6570 (expand_used_vars): Leave a space in the stack frame for alignment
6571 if STRICT_ALIGNMENT.
6573 2014-04-21 David Malcolm <dmalcolm@redhat.com>
6575 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
6577 (gimple_store_p): Likewise.
6578 (gimple_assign_load_p): Likewise.
6579 (gimple_assign_cast_p): Likewise.
6580 (gimple_clobber_p): Likewise.
6582 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
6583 rather than a gimple.
6584 (gimple_assign_cast_p): Likewise.
6586 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
6589 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
6590 If mode is DDmode and TARGET_E500_DOUBLE allow move.
6592 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
6593 more debug information for E500 if -mdebug=reg.
6595 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
6598 * config/i386/i386.c (ix86_expand_builtin)
6599 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
6600 register for target RTX.
6601 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
6603 2014-04-18 Cong Hou <congh@google.com>
6605 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
6606 the widen-mult pattern by handling two operands with different sizes,
6607 and operands whose size is smaller than half of the result type.
6609 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
6611 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
6612 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
6613 (do_estimate_edge_time): Compute it.
6614 * ipa-inline.c (want_inline_small_function_p): Bypass
6615 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
6617 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
6619 * ipa-inline.c (spec_rem): New static variable.
6620 (dump_overall_stats): New function.
6621 (dump_inline_stats): New function.
6623 2014-04-18 Richard Henderson <rth@redhat.com>
6625 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
6626 to GET_MODE_SIZE, not a reg_class_t.
6628 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6630 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
6631 (vsx_xxmrglw_<mode>): Likewise.
6633 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
6636 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
6637 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
6638 (rs6000_init_hard_regno_mode_ok): Likewise.
6640 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
6642 * ipa-inline.c (inline_small_functions): Account only non-cold
6644 * doc/invoke.texi (inline-unit-growth): Update documentation.
6646 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
6648 * config/rs6000/rs6000.md (addti3, subti3): New.
6650 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
6653 * config/i386/i386.c (ix86_expand_clear): Remove outdated
6654 comment. Check optimize_insn_for_size_p instead of
6655 optimize_insn_for_speed_p.
6657 2014-04-17 Martin Jambor <mjambor@suse.cz>
6659 * gimple-iterator.c (gsi_start_edge): New function.
6660 * gimple-iterator.h (gsi_start_edge): Declare.
6661 * tree-sra.c (single_non_eh_succ): New function.
6662 (disqualify_ops_if_throwing_stmt): Renamed to
6663 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
6664 having one non-EH successor BB.
6665 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
6666 generate loads into replacements.
6667 (sra_modify_assign): Likewise and and also use the simple path for
6669 (sra_modify_function_body): Commit statements on edges.
6671 2014-04-17 Richard Biener <rguenther@suse.de>
6674 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
6675 comparison results and add clarifying comment.
6677 2014-04-17 Jakub Jelinek <jakub@redhat.com>
6679 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
6680 (blank_mode): Initialize it.
6681 (emit_mode_size_inline, emit_mode_nunits_inline,
6682 emit_mode_inner_inline): New functions.
6683 (emit_insn_modes_h): Call them and surround their output with
6684 #if GCC_VERSION >= 4001 ... #endif.
6685 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
6686 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
6687 mode_* arrays if the argument is __builtin_constant_p.
6688 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
6689 is enum machine_mode.
6691 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6693 * passes.c (opt_pass::execute): Adjust.
6694 (pass_manager::execute_pass_mode_switching): Likewise.
6695 (early_local_passes::execute): Likewise.
6696 (execute_one_pass): Pass cfun to the pass's execute method.
6697 * tree-pass.h (opt_pass::execute): Add function * argument.
6698 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
6699 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
6700 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
6701 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
6702 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
6703 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
6704 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
6705 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
6706 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
6707 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
6708 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
6709 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
6710 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
6711 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
6712 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
6713 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
6714 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
6715 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
6716 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
6717 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
6718 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
6719 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
6720 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
6721 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
6722 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
6723 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
6724 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
6725 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
6726 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
6729 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6731 * passes.c (opt_pass::gate): Take function * argument.
6732 (gate_all_early_local_passes): Merge into
6733 (early_local_passes::gate): this.
6734 (gate_all_early_optimizations): Merge into
6735 (all_early_optimizations::gate): this.
6736 (gate_all_optimizations): Mege into
6737 (all_optimizations::gate): this.
6738 (gate_all_optimizations_g): Merge into
6739 (all_optimizations_g::gate): this.
6740 (gate_rest_of_compilation): Mege into
6741 (rest_of_compilation::gate): this.
6742 (gate_postreload): Merge into
6743 (postreload::gate): this.
6744 (dump_one_pass): Pass cfun to the pass's gate method.
6745 (execute_ipa_summary_passes): Likewise.
6746 (execute_one_pass): Likewise.
6747 (ipa_write_summaries_2): Likewise.
6748 (ipa_write_optimization_summaries_1): Likewise.
6749 (ipa_read_summaries_1): Likewise.
6750 (ipa_read_optimization_summaries_1): Likewise.
6751 (execute_ipa_stmt_fixups): Likewise.
6752 * tree-pass.h (opt_pass::gate): Add function * argument.
6753 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
6754 combine-stack-adj.c, combine.c, compare-elim.c,
6755 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
6756 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
6757 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
6758 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
6759 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
6760 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
6761 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
6762 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
6763 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
6764 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
6765 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
6766 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
6767 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
6768 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
6769 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
6770 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
6771 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
6772 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
6773 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
6774 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
6775 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
6776 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
6777 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
6778 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
6779 var-tracking.c, vtable-verify.c, web.c: Adjust.
6781 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6783 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
6784 * configure: Regenerate.
6786 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6788 * passes.c (dump_one_pass): don't check pass->has_gate.
6789 (execute_ipa_summary_passes): Likewise.
6790 (execute_one_pass): Likewise.
6791 (ipa_write_summaries_2): Likewise.
6792 (ipa_write_optimization_summaries_1): Likewise.
6793 (ipa_read_optimization_summaries_1): Likewise.
6794 (execute_ipa_stmt_fixups): Likewise.
6795 * tree-pass.h (pass_data::has_gate): Remove.
6796 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
6797 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
6798 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
6799 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
6800 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
6801 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
6802 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
6803 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
6804 gimple-low.c, gimple-ssa-isolate-paths.c,
6805 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
6806 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
6807 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
6808 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
6809 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
6810 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
6811 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
6812 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
6813 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
6814 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
6815 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
6816 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
6817 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
6818 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
6819 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
6820 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
6821 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
6822 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
6823 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
6824 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
6825 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
6826 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
6829 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6831 * pass_manager.h (pass_manager::register_dump_files_1): Remove
6833 * passes.c (pass_manager::register_dump_files_1): Merge into
6834 (pass_manager::register_dump_files): this, and remove its handling of
6835 properties since the pass always has the properties anyway.
6836 (pass_manager::pass_manager): Adjust.
6838 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6840 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
6841 * passes.c (pass_manager::register_dump_files_1): Remove dead code
6842 dealing with properties.
6843 (pass_manager::register_dump_files): Adjust.
6845 2014-03-20 Mark Wielaard <mjw@redhat.com>
6847 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
6848 then represent the bound as normal constant value.
6850 2014-04-17 Jakub Jelinek <jakub@redhat.com>
6853 Forward port from 4.8 branch
6854 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
6856 * config/i386/bmiintrin.h (_blsi_u32): New.
6860 (_blsmsk_u32): Ditto.
6861 (_blsmsk_u64): Ditto.
6862 (_tzcnt_u32): Ditto.
6863 (_tzcnt_u64): Ditto.
6865 2014-04-17 Kito Cheng <kito@0xlab.org>
6867 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
6869 2014-04-17 Richard Biener <rguenther@suse.de>
6872 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
6873 boolean results for comparisons.
6875 2014-04-17 Richard Biener <rguenther@suse.de>
6877 PR tree-optimization/60836
6878 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
6879 initial PHI args to be gimple values.
6881 2014-04-17 Richard Biener <rguenther@suse.de>
6883 PR tree-optimization/60841
6884 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
6885 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
6886 of stmts to SLP build.
6887 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
6888 (vect_analyze_slp): Likewise.
6889 (vect_analyze_slp_instance): Likewise.
6890 (vect_build_slp_tree): Limit overall SLP tree growth.
6891 * tree-vectorizer.h (vect_analyze_data_refs,
6892 vect_analyze_slp): Adjust prototypes.
6894 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
6896 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
6899 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
6901 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
6902 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
6903 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
6904 for TARGET_SLOW_PSHUFB
6906 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
6908 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
6909 * config/i386/i386.c (intel_cost): Ditto.
6911 2014-04-17 Joey Ye <joey.ye@arm.com>
6913 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
6915 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
6917 * opts.c (common_handle_option): Disable -fipa-reference coorectly
6920 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
6922 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
6923 (type_all_derivations_known_p): New predicate.
6924 (type_all_ctors_visible_p): New predicate.
6925 (type_possibly_instantiated_p): New predicate.
6926 (get_odr_type): Compute all_derivations_known.
6927 (dump_odr_type): Dump the flag.
6928 (maybe_record_type): Cleanup.
6929 (record_target_from_binfo): Add bases_to_consider array;
6930 record bases for types w/o instances and skip CXX destructor.
6931 (possible_polymorphic_call_targets_1): Add bases_to_consider
6932 and consider_construction parameters; check if type may have instance.
6933 (get_polymorphic_call_info): Set maybe_in_construction to true
6934 when we know nothing.
6935 (record_targets_from_bases): Skip CXX destructors; they are
6936 never called for types in construction.
6937 (possible_polymorphic_call_targets): Do not record target when
6938 type may not have instance.
6940 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
6943 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
6944 external aliases alive, too.
6946 2014-04-16 Andrew Pinski <apinski@cavium.com>
6948 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
6951 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
6953 * final.c (compute_alignments): Do not apply loop alignment to a block
6954 falling through to the exit.
6956 2014-04-16 Catherine Moore <clm@codesourcery.com>
6958 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
6959 Adjust constraints for microMIPS store patterns.
6961 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
6963 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
6965 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
6967 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
6968 (append_use): Run at -O0.
6969 (append_vdef): Likewise.
6970 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
6971 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
6973 2014-04-16 Jakub Jelinek <jakub@redhat.com>
6975 PR tree-optimization/60844
6976 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
6977 (propagate_op_to_single_use, remove_visited_stmt_chain,
6978 linearize_expr, repropagate_negates, reassociate_bb): Use it
6979 instead of gsi_remove.
6981 2014-04-16 Martin Jambor <mjambor@suse.cz>
6983 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
6984 ipa_transforms_to_apply.
6985 (cgraph_function_versioning): Assert that old_node has empty
6986 ipa_transforms_to_apply.
6987 * trans-mem.c (ipa_tm_create_version): Likewise.
6988 * tree-inline.c (tree_function_versioning): Do not duplicate
6989 ipa_transforms_to_apply.
6991 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6994 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
6996 Pass necessary as flags on 64-bit Solaris/x86.
6997 Use lowercase relocs for x86_64-*-*.
6998 * configure: Regenerate.
7000 2014-04-15 Jan Hubicka <jh@suse.cz>
7002 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
7003 (maybe_record_node, likely_target_p): Use it.
7005 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7008 Revert following patch
7010 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
7013 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
7014 software floating point or no floating point registers, do not
7015 allow any type in the FPRs. Eliminate a test for SPE SIMD types
7016 in GPRs that occurs after we tested for GPRs that would never be
7019 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
7020 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
7021 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
7022 specifically allow DDmode, since that does not use the SPE SIMD
7025 2014-03-21 Mark Wielaard <mjw@redhat.com>
7027 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
7028 as unsigned or int depending on type and value used.
7030 2014-04-15 Richard Biener <rguenther@suse.de>
7032 PR rtl-optimization/56965
7033 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
7034 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
7036 * alias.c (true_dependence_1): Do not call
7037 nonoverlapping_component_refs_p.
7038 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
7039 nonoverlapping_component_refs_p.
7040 (indirect_refs_may_alias_p): Likewise.
7042 2014-04-15 Teresa Johnson <tejohnson@google.com>
7044 * cfg.c (dump_bb_info): Fix flags check.
7045 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
7047 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7049 PR rtl-optimization/60663
7050 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
7053 2014-04-15 Richard Biener <rguenther@suse.de>
7055 * lto-streamer.h (LTO_major_version): Bump to 4.
7057 2014-04-15 Richard Biener <rguenther@suse.de>
7059 * common.opt (lto_partition_model): New enum.
7060 (flto-partition=): Merge separate options with a single with argument,
7061 add -flto-partition=one support.
7062 * flag-types.h (enum lto_partition_model): Declare.
7063 * opts.c (finish_options): Remove duplicate -flto-partition=
7065 * lto-wrapper.c (run_gcc): Adjust.
7067 2014-04-15 Richard Biener <rguenther@suse.de>
7069 * alias.c (ncr_compar): New function.
7070 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
7072 2014-04-15 Richard Biener <rguenther@suse.de>
7074 * alias.c (record_component_aliases): Do not walk BINFOs.
7076 2014-04-15 Richard Biener <rguenther@suse.de>
7078 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
7079 Add struct function argument and adjust.
7080 (find_func_aliases_for_call): Likewise.
7081 (find_func_aliases): Likewise.
7082 (find_func_clobbers): Likewise.
7083 (intra_create_variable_infos): Likewise.
7084 (compute_points_to_sets): Likewise.
7085 (ipa_pta_execute): Adjust. Do not push/pop cfun.
7087 2014-04-15 Richard Biener <rguenther@suse.de>
7089 * tree.c (iterative_hash_expr): Use enum tree_code_class
7090 to store TREE_CODE_CLASS.
7091 (tree_block): Likewise.
7092 (tree_set_block): Likewise.
7093 * tree.h (fold_build_pointer_plus_loc): Use
7094 convert_to_ptrofftype_loc.
7096 2014-04-15 Jakub Jelinek <jakub@redhat.com>
7099 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
7102 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
7104 * cfgloop.h (struct loop): Move force_vectorize down.
7105 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
7106 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
7107 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
7108 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
7109 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
7110 * tree-core.h (enum annot_expr_kind): Add new kind values.
7111 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
7112 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
7114 * tree.def (ANNOTATE_EXPR): Tweak comment.
7116 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
7118 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
7121 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
7123 * tree.h (TYPE_IDENTIFIER): Declare.
7124 * tree.c (subrange_type_for_debug_p): Use it.
7125 * godump.c (go_format_type): Likewise.
7126 * dwarf2out.c (is_cxx_auto, modified_type_die,
7127 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
7128 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
7130 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
7133 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
7135 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
7137 * config/i386/i386.c (examine_argument): Return bool. Return true if
7138 parameter should be passed in memory.
7139 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
7140 (construct_container): Update calls to examine_argument.
7141 (function_arg_advance_64): Ditto.
7142 (return_in_memory_32): Merge with ix86_return_in_memory.
7143 (return_in_memory_64): Ditto.
7144 (return_in_memory_ms_64): Ditto.
7146 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
7148 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
7149 * coverage.c (coverage_compute_profile_id): Handle externally visible
7152 2014-04-14 Martin Jambor <mjambor@suse.cz>
7154 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
7155 DECL_DISREGARD_INLINE_LIMITS functions.
7157 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
7160 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
7162 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
7165 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
7166 optimize_insn_for_speed_p instead of
7167 optimize_function_for_speed_p.
7169 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
7171 * doc/invoke.texi (free): Document AArch64.
7173 2014-04-14 Richard Biener <rguenther@suse.de>
7175 PR tree-optimization/60042
7176 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
7177 (insert_into_preds_of_block): Do not prevent PHI insertion
7178 for REFERENCE exprs here ...
7179 (eliminate_dom_walker::before_dom_children): ... but prevent
7180 their use here under similar conditions when applied to the
7181 IL after PRE optimizations.
7183 2014-04-14 Richard Biener <rguenther@suse.de>
7185 * passes.def: Move early points-to after early SRA.
7187 2014-04-14 Richard Biener <rguenther@suse.de>
7189 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
7190 check for which sign-changes we allow when forwarding
7191 a converted value into a switch.
7193 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
7195 * stor-layout.c (place_field): Finalize non-constant offset for the
7198 2014-04-14 Richard Biener <rguenther@suse.de>
7200 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
7202 (expand_switch_using_bit_tests_p): Likewise.
7203 (process_switch): Compute and pass on speed_p based on the
7205 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
7206 optimize_bb_for_speed_p.
7208 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
7210 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
7211 * function.h (struct function): Rename has_force_vect_loops into
7212 has_force_vectorize_loops.
7213 * lto-streamer-in.c (input_cfg): Adjust for renaming.
7214 (input_struct_function_base): Likewise.
7215 * lto-streamer-out.c (output_cfg): Likewise.
7216 (output_struct_function_base): Likewise.
7217 * omp-low.c (expand_omp_simd): Likewise.
7218 * tree-cfg.c (move_sese_region_to_fn): Likewise.
7219 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
7220 (version_loop_for_if_conversion): Likewise.
7221 (tree_if_conversion): Likewise.
7222 (main_tree_if_conversion): Likewise.
7223 (gate_tree_if_conversion): Likewise.
7224 * tree-inline.c (copy_loops): Likewise.
7225 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
7226 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
7227 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
7228 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
7229 * tree-vectorizer.c (vectorize_loops): Likewise.
7230 * tree-vectorizer.h (unlimited_cost_model): Likewise.
7232 2014-04-14 Richard Biener <rguenther@suse.de>
7235 * lto-streamer-out.c (wrap_refs): New function.
7236 (lto_output): Wrap symbol references in global initializes in
7237 type-preserving MEM_REFs.
7239 2014-04-14 Christian Bruel <christian.bruel@st.com>
7241 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
7243 2014-04-14 Christian Bruel <christian.bruel@st.com>
7245 * config/sh/sh.md (setmemqi): New expand pattern.
7246 * config/sh/sh.h (CLEAR_RATIO): Define.
7247 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
7248 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
7250 2014-04-14 Richard Biener <rguenther@suse.de>
7253 * fold-const.c (negate_expr_p): Don't negate directional rounding
7255 (fold_negate_expr): Likewise.
7257 2014-04-14 Richard Biener <rguenther@suse.de>
7259 PR tree-optimization/59817
7260 PR tree-optimization/60453
7261 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
7262 recursion to catch all CHRECs in the scalar evolution and restrict
7263 the predicate for the remains appropriately.
7265 2014-04-12 Catherine Moore <clm@codesourcery.com>
7267 * config/mips/constraints.md: Add new register constraint "kb".
7268 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
7269 (*movhi_internal): Likewise.
7270 (*movqi_internal): Likewise.
7271 * config/mips/mips.h (M16_STORE_REGS): New register class.
7272 (REG_CLASS_NAMES): Add M16_STORE_REGS.
7273 (REG_CLASS_CONTENTS): Likewise.
7274 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
7276 2014-04-11 Tobias Burnus <burnus@net-b.de>
7279 * doc/invoke.texi (-Wformat-signedness): Document it.
7280 (Wformat=2): Mention that this enables -Wformat-signedness.
7282 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
7284 * common/config/epiphany/epiphany-common.c
7285 (epiphany_option_optimization_table): Enable section anchors by
7286 default at -O1 or higher.
7287 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
7288 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
7289 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
7290 carries no extra cost.
7291 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
7292 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
7293 * config/epiphany/predicates.md (memclob_operand): New predicate.
7294 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
7295 Use memclob_operand predicate and X constraint for operand 3.
7297 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
7299 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
7300 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
7303 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
7305 PR rtl-optimization/60651
7306 * mode-switching.c (optimize_mode_switching): Make sure to emit
7307 sets of a lower numbered entity before sets of a higher numbered
7308 entity to a mode of the same or lower priority.
7309 When creating a seginfo for a basic block that starts with a code
7310 label, move the insertion point past the code label.
7311 (new_seginfo): Document and enforce requirement that
7312 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
7313 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
7314 * doc/tm.texi: Regenerate.
7316 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
7319 * config/arc/arc.c (arc_save_restore): Fix assert typo.
7321 2013-04-11 Jakub Jelinek <jakub@redhat.com>
7323 * BASE-VER: Set to 4.10.0.
7325 2014-04-11 Tobias Burnus <burnus@net-b.de>
7328 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
7329 * doc/gcc.texi (Service): Update description in the @menu
7330 * doc/invoke.texi (Option Summary): Remove misplaced and
7333 2014-04-11 Steve Ellcey <sellcey@mips.com>
7334 Jakub Jelinek <jakub@redhat.com>
7337 * expr.c (convert_move): Use emit_store_flag_force instead of
7338 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
7341 2014-04-11 Richard Biener <rguenther@suse.de>
7344 * varasm.c (assemble_alias): Avoid endless error reporting
7345 recursion by setting TREE_ASM_WRITTEN.
7347 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7349 * config/s390/s390.md: Add a splitter for NOT rtx.
7351 2014-04-11 Jakub Jelinek <jakub@redhat.com>
7353 PR rtl-optimization/60663
7354 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
7356 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
7357 Jakub Jelinek <jakub@redhat.com>
7360 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
7361 flag from decl_node to node.
7363 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7366 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
7367 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
7368 ameliorating the cases where it can be.
7370 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
7373 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
7375 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
7376 (loadsync_<mode>): Change mode.
7377 (load_quadpti, store_quadpti): New.
7378 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
7379 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
7380 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
7382 2014-04-09 Cong Hou <congh@google.com>
7385 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
7388 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7390 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
7391 instead of vnor to exploit possible fusion opportunity in the
7393 (altivec_expand_vec_perm_const_le): Likewise.
7395 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
7397 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
7398 (loadsync_<mode>): Change mode.
7399 (load_quadpti, store_quadpti): New.
7400 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
7401 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
7403 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
7406 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
7407 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
7408 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
7410 2014-04-08 Richard Biener <rguenther@suse.de>
7413 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
7414 a 64bit widest int print double-int similar to on HWI64 hosts.
7416 2014-04-08 Richard Biener <rguenther@suse.de>
7418 PR tree-optimization/60785
7419 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
7420 default defs properly.
7422 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
7424 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
7425 (Weffc++): Likewise.
7427 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
7429 * ipa-devirt.c (maybe_record_node): When node is not recorded,
7430 set completep to false rather than true.
7432 2014-04-07 Douglas B Rupp <rupp@adacore.com>
7435 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
7436 ARM_TARGET2_DWARF_FORMAT.
7438 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
7441 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
7442 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
7445 2014-04-07 Richard Biener <rguenther@suse.de>
7447 PR tree-optimization/60766
7448 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
7449 (may_eliminate_iv): Convert cand_value_at result to desired type.
7451 2014-04-07 Jason Merrill <jason@redhat.com>
7454 * common.opt (-fno-gnu-unique): Add.
7455 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
7457 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7459 * haifa-sched.c: Fix outdated function reference and minor
7460 grammar errors in introductory comment.
7462 2014-04-07 Richard Biener <rguenther@suse.de>
7465 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
7467 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
7469 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
7472 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
7473 size accounting for thunks.
7474 (pa_asm_output_mi_thunk): Use final_start_function() and
7475 final_end_function() to output function start and end directives.
7477 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
7479 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
7480 device specific ISA/ feature information. Remove short_sp and
7481 errata_skip ds. Add avr_device_specific_features enum to have device
7483 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
7484 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
7485 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
7486 updated device specific info.
7487 * config/avr/avr-mcus.def: Merge device specific details to
7488 dev_attribute field.
7489 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
7491 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
7492 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
7493 assembler if RMW isa supported by current device.
7494 * config/avr/genmultilib.awk: Update as device info structure changed.
7495 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
7497 2014-04-04 Cong Hou <congh@google.com>
7499 PR tree-optimization/60656
7500 * tree-vect-stmts.c (supportable_widening_operation):
7501 Fix a bug that elements in a vector with vect_used_by_reduction
7502 property are incorrectly reordered when the operation on it is not
7503 consistant with the one in reduction operation.
7505 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
7507 PR rtl-optimization/60155
7508 * gcse.c (record_set_data): New function.
7509 (single_set_gcse): New function.
7510 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
7511 (hoist_code): Likewise.
7512 (get_pressure_class_and_nregs): Likewise.
7514 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
7516 * explow.c (probe_stack_range): Emit a final optimization blockage.
7518 2014-04-04 Anthony Green <green@moxielogic.com>
7520 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
7523 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
7526 * lto-cgraph.c (input_overwrite_node): Check that partitioning
7527 flags are set only during streaming.
7528 * ipa.c (process_references, walk_polymorphic_call_targets,
7529 symtab_remove_unreachable_nodes): Drop bodies of always inline
7530 after early inlining.
7531 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
7533 2014-04-04 Jakub Jelinek <jakub@redhat.com>
7534 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7537 * dwarf2out.c (const_ok_for_output_1): Reject expressions
7540 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7543 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
7545 (cortex_a53_fdivd): Likewise.
7547 2014-04-04 Martin Jambor <mjambor@suse.cz>
7550 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
7552 * cgraph.c (clone_of_p): Also return true if thunks match.
7553 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
7554 cgraph_function_or_thunk_node and an obsolete comment.
7555 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
7557 (build_function_decl_skip_args): Likewise.
7558 (set_new_clone_decl_and_node_flags): New function.
7559 (duplicate_thunk_for_node): Likewise.
7560 (redirect_edge_duplicating_thunks): Likewise.
7561 (cgraph_clone_node): New parameter args_to_skip, pass it to
7562 redirect_edge_duplicating_thunks which is called instead of
7563 cgraph_redirect_edge_callee.
7564 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
7565 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
7567 2014-04-04 Jeff Law <law@redhat.com>
7570 * config/arm/predicates.md (const_int_I_operand): New predicate.
7571 (const_int_M_operand): Similarly.
7572 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
7574 (insv_t2, extv_reg, extzv_t2): Likewise.
7575 (load_multiple_with_writeback): Similarly for const_int_I_operand.
7576 (pop_multiple_with_writeback_and_return): Likewise.
7577 (vfp_pop_multiple_with_writeback): Likewise
7579 2014-04-04 Richard Biener <rguenther@suse.de>
7582 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
7583 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
7585 * gimplify.h (gimple_add_tmp_var_fn): Declare.
7586 * gimplify.c (gimple_add_tmp_var_fn): New function.
7587 * gimple-expr.h (create_tmp_reg_fn): Declare.
7588 * gimple-expr.c (create_tmp_reg_fn): New function.
7589 * gimple-low.c (record_vars_into): Don't change cfun.
7590 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
7591 code generation without cfun.
7593 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
7596 * Makefile.in (install-driver): Fix shell scripting.
7598 2014-04-03 Cong Hou <congh@google.com>
7600 PR tree-optimization/60505
7601 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
7602 threshold of number of iterations below which no vectorization
7604 * tree-vect-loop.c (new_loop_vec_info):
7605 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
7606 * tree-vect-loop.c (vect_analyze_loop_operations):
7607 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
7608 * tree-vect-loop.c (vect_transform_loop):
7609 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
7610 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
7611 of iterations of the loop and see if we should build the epilogue.
7613 2014-04-03 Richard Biener <rguenther@suse.de>
7615 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
7616 (streamer_tree_cache_create): Adjust.
7617 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
7618 to allow optional nodes array.
7619 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
7620 (streamer_tree_cache_append): Likewise.
7621 (streamer_tree_cache_create): Create nodes array optionally
7622 as specified by parameter.
7623 * lto-streamer-out.c (create_output_block): Avoid maintaining
7624 the node array in the writer cache.
7625 (DFS_write_tree): Remove assertion.
7626 (produce_asm_for_decls): Free the out decl state hash table early.
7627 * lto-streamer-in.c (lto_data_in_create): Adjust for
7628 streamer_tree_cache_create prototype change.
7630 2014-04-03 Richard Biener <rguenther@suse.de>
7632 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
7633 set TREE_CHAIN to NULL_TREE.
7635 2014-04-03 Richard Biener <rguenther@suse.de>
7637 PR tree-optimization/60740
7638 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
7639 over all GIMPLE_COND operands.
7641 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
7643 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
7644 (Weffc++): Remove Scott's numbering, merge lists and reference
7647 2014-04-03 Nick Clifton <nickc@redhat.com>
7649 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
7652 2014-04-03 Martin Jambor <mjambor@suse.cz>
7654 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
7655 mention gcc_unreachable before failing.
7656 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
7659 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
7662 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
7663 inconsistent code and instead mark the context inconsistent.
7664 (possible_polymorphic_call_targets): For inconsistent contexts
7665 return empty complete list.
7667 2014-04-02 Anthony Green <green@moxielogic.com>
7669 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
7670 (extendqisi2, extendhisi2): Define.
7671 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
7672 (WCHAR_TYPE): Change to unsigned int.
7674 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7676 PR tree-optimization/60733
7677 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
7678 insertion point for PHI candidates to be the end of the feeding
7679 block for the PHI argument.
7681 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
7683 PR rtl-optimization/60650
7684 * lra-constraints.c (process_alt_operands): Decrease reject for
7685 earlyclobber matching.
7687 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7689 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
7691 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7693 * config/spu/spu.c (pad_bb): Do not crash when the last
7694 insn is CODE_FOR_blockage.
7696 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7698 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
7699 lies outside the target mode.
7701 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
7704 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
7705 software floating point or no floating point registers, do not
7706 allow any type in the FPRs. Eliminate a test for SPE SIMD types
7707 in GPRs that occurs after we tested for GPRs that would never be
7710 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
7711 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
7712 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
7713 specifically allow DDmode, since that does not use the SPE SIMD
7716 2014-04-02 Richard Biener <rguenther@suse.de>
7719 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
7720 MODE_INTs. Properly use negv_optab.
7721 (expand_abs): Likewise.
7723 2014-04-02 Richard Biener <rguenther@suse.de>
7726 * Makefile.in (install-driver): Guard extra installs with special
7729 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
7731 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
7734 2014-04-01 Richard Henderson <rth@redhat.com>
7737 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
7738 alternative enabled before register allocation.
7740 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
7742 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
7743 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
7745 (nios2_large_got_address): Remove unneeded 'sym' parameter.
7746 (nios2_got_address): Update nios2_large_got_address call site.
7747 (nios2_delegitimize_address): New function.
7748 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
7749 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
7750 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
7752 2014-04-01 Martin Husemann <martin@duskware.de>
7754 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
7757 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
7759 PR rtl-optimization/60604
7760 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
7761 check from register_operand.
7762 (register_operand): Redefine in terms of general_operand.
7763 (nonmemory_operand): Use register_operand for the non-constant cases.
7765 2014-04-01 Richard Biener <rguenther@suse.de>
7767 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
7769 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
7771 * doc/invoke.texi (mapp-regs): Clarify.
7773 2014-03-31 Ulrich Drepper <drepper@gmail.com>
7775 * config/i386/avx512fintrin.h (__v32hi): Define type.
7776 (__v64qi): Likewise.
7777 (_mm512_set1_epi8): Define.
7778 (_mm512_set1_epi16): Define.
7779 (_mm512_set4_epi32): Define.
7780 (_mm512_set4_epi64): Define.
7781 (_mm512_set4_pd): Define.
7782 (_mm512_set4_ps): Define.
7783 (_mm512_setr4_epi64): Define.
7784 (_mm512_setr4_epi32): Define.
7785 (_mm512_setr4_pd): Define.
7786 (_mm512_setr4_ps): Define.
7787 (_mm512_setzero_epi32): Define.
7789 2014-03-31 Martin Jambor <mjambor@suse.cz>
7792 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
7793 callsite_arguments_match_p. Updated all callers. Also check types of
7794 corresponding formal parameters and actual arguments.
7795 (not_all_callers_have_enough_arguments_p) Renamed to
7796 some_callers_have_mismatched_arguments_p.
7798 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
7800 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
7802 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
7805 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
7808 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
7810 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
7811 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
7813 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
7814 Use FMAMODE_NOVF512 mode iterator.
7815 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
7816 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
7817 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
7818 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
7820 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
7821 Use VF_128_256 mode iterator.
7822 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
7825 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
7827 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
7828 static chain if needed.
7830 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
7833 * lra-constraints.c (index_part_to_reg): New.
7834 (process_address): Use it.
7836 2014-03-27 Jeff Law <law@redhat.com>
7837 Jakub Jelinek <jakub@redhat.com>
7840 * expr.c (do_tablejump): Use simplify_gen_binary rather than
7841 gen_rtx_{PLUS,MULT} to build up the address expression.
7843 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
7844 creating non-canonical RTL.
7846 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
7849 * ipa-inline.c (want_inline_small_function_p): Short circuit large
7850 functions; reorganize to make cheap checks first.
7851 (inline_small_functions): Do not estimate growth when dumping;
7853 * ipa-inline.h (inline_summary): Add min_size.
7854 (growth_likely_positive): New function.
7855 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
7856 (set_cond_stmt_execution_predicate): Cleanup.
7857 (estimate_edge_size_and_time): Compute min_size.
7858 (estimate_calls_size_and_time): Likewise.
7859 (estimate_node_size_and_time): Likewise.
7860 (inline_update_overall_summary): Update min_size.
7861 (do_estimate_edge_time): Likewise.
7862 (do_estimate_edge_size): Update.
7863 (do_estimate_edge_hints): Update.
7864 (growth_likely_positive): New function.
7866 2014-03-28 Jakub Jelinek <jakub@redhat.com>
7869 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
7870 also if addr has VOIDmode.
7872 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7874 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
7875 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
7877 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
7878 instructions as well as AdvancedSIMD loads.
7880 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7882 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
7883 Use crypto_aese type.
7884 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
7885 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
7886 crypto_aese, crypto_aesmc. Move to types.md.
7887 * config/arm/types.md (crypto_aes): Split into crypto_aese,
7889 * config/arm/iterators.md (crypto_type): Likewise.
7891 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
7893 * cgraph.c: Include expr.h and tree-dfa.h.
7894 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
7897 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
7900 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
7901 regs from checking multi-reg pseudos.
7903 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7905 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
7907 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7909 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
7910 if it would clobber the stack pointer, even temporarily.
7912 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
7914 * mode-switching.c: Make small adjustments to the top comment.
7916 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
7918 * config/rs6000/constraints.md (wD constraint): New constraint to
7919 match the constant integer to get the top DImode/DFmode out of a
7920 vector in a VSX register.
7922 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
7923 match the constant integer to get the top DImode/DFmode out of a
7924 vector in a VSX register.
7926 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
7929 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
7932 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
7933 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
7935 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
7936 Optimize vec_extract of 64-bit values, where the value being
7937 extracted is in the top word, where we can use scalar
7938 instructions. Add direct move and store support. Combine the big
7939 endian/little endian vector select load support into a single insn.
7940 (vsx_extract_<mode>_internal1): Likewise.
7941 (vsx_extract_<mode>_internal2): Likewise.
7942 (vsx_extract_<mode>_load): Likewise.
7943 (vsx_extract_<mode>_store): Likewise.
7944 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
7945 combined into vsx_extract_<mode>_load.
7946 (vsx_extract_<mode>_one_le): Likewise.
7948 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
7949 define the top 64-bit vector element.
7951 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
7954 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
7955 Document vec_vbpermq builtin.
7958 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
7959 enable use of xxsldwi and xxpermdi builtin functions.
7960 (vec_xxpermdi): Likewise.
7962 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
7963 Document use of vec_xxsldwi and vec_xxpermdi builtins.
7965 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
7967 PR rtl-optimization/60650
7968 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
7970 (find_spills_for): New.
7971 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
7972 Spill all pseudos on the second iteration.
7974 2014-03-27 Marek Polacek <polacek@redhat.com>
7977 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
7980 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7982 * config/s390/s390.c (s390_can_use_return_insn): Check for
7983 call-saved FPRs on 31 bit.
7985 2014-03-27 Jakub Jelinek <jakub@redhat.com>
7988 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
7989 if they need regimplification, just drop them instead of
7990 calling gimple_regimplify_operands on them.
7992 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
7995 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
7996 (aarch64_frame_pointer_required): Adjust logic.
7997 (aarch64_can_eliminate): Adjust logic.
7998 (aarch64_override_options_after_change): Adjust logic.
8000 2014-03-27 Dehao Chen <dehao@google.com>
8002 * ipa-inline.c (early_inliner): Update node's inline info.
8004 2014-03-26 Dehao Chen <dehao@google.com>
8006 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
8007 compiler inserted conditional jumps for NAN float check.
8009 2014-03-26 Jakub Jelinek <jakub@redhat.com>
8011 * ubsan.h (ubsan_create_data): Change second argument's type
8012 to const location_t *.
8013 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
8015 (ubsan_create_data): Change second argument to const location_t *PLOC.
8016 Create Loc field whenever PLOC is non-NULL.
8017 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
8018 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
8022 * real.c (real_to_integer2): Change type of low to UHWI.
8024 2014-03-26 Tobias Burnus <burnus@net-b.de>
8026 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
8027 (CILK_SELF_SPECS): New define.
8028 (driver_self_specs): Use it.
8030 2014-03-26 Richard Biener <rguenther@suse.de>
8032 * tree-pretty-print.c (percent_K_format): Implement special
8033 case for LTO and its stripped down BLOCK tree.
8035 2014-03-26 Jakub Jelinek <jakub@redhat.com>
8038 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
8040 * tree-vrp.c (simplify_internal_call_using_ranges): If only
8041 one range is range_int_cst_p, but not both, at least optimize
8042 addition/subtraction of 0 and multiplication by 0 or 1.
8043 * gimple-fold.c (gimple_fold_call): Fold
8044 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
8045 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
8046 INTEGER_CSTs, try to fold at least x * 0 and y - y.
8048 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
8050 PR rtl-optimization/60452
8051 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
8052 <case REG>: Return 1 for invalid offsets from the frame pointer.
8054 2014-03-26 Marek Polacek <polacek@redhat.com>
8057 * doc/extend.texi (C Extensions): Mention variable-length arrays in
8060 2014-03-26 Marek Polacek <polacek@redhat.com>
8063 * doc/extend.texi (Designated Inits): Describe what happens to omitted
8066 2014-03-26 Marek Polacek <polacek@redhat.com>
8069 * ira-color.c (update_conflict_hard_regno_costs): Perform the
8070 multiplication in unsigned type.
8072 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
8074 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
8076 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
8078 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
8080 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
8083 * cif-code.def (UNREACHABLE) New code.
8084 * ipa-inline.c (inline_small_functions): Skip edges to
8085 __builtlin_unreachable.
8086 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
8087 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
8088 predicate to __bulitin_unreachable.
8089 (set_cond_stmt_execution_predicate): Fix issue when
8090 invert_tree_comparison returns ERROR_MARK.
8091 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
8092 propagate to inline clones.
8093 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
8095 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
8096 * cgraphclones.c (cgraph_clone_node): If call destination is already
8097 ureachable, do not redirect it back.
8098 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
8101 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
8103 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
8104 Do not modify inline clones.
8106 2014-03-25 Jakub Jelinek <jakub@redhat.com>
8108 * config/i386/i386.md (general_sext_operand): New mode attr.
8109 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
8110 don't generate (sign_extend (const_int)).
8111 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
8112 operands[2]. Use We constraint instead of <i> and
8113 <general_sext_operand> predicate instead of <general_operand>.
8114 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
8115 * config/i386/constraints.md (We): New constraint.
8116 * config/i386/predicates.md (x86_64_sext_operand,
8117 sext_operand): New predicates.
8119 2014-03-25 Martin Jambor <mjambor@suse.cz>
8122 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
8123 inconsistent devirtualizations to __builtin_unreachable.
8125 2014-03-25 Marek Polacek <polacek@redhat.com>
8128 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
8130 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
8132 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
8133 order of elements for big-endian.
8135 2014-03-25 Richard Biener <rguenther@suse.de>
8138 * gimplify-me.c (gimple_regimplify_operands): Update the
8141 2014-03-25 Martin Jambor <mjambor@suse.cz>
8144 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
8145 (lto_output_varpool_node): Likewise.
8146 (input_overwrite_node): Likewise.
8147 (input_varpool_node): Likewise.
8149 2014-03-25 Richard Biener <rguenther@suse.de>
8151 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
8152 (run_gcc): Likewise.
8154 2014-03-25 Jakub Jelinek <jakub@redhat.com>
8156 * combine.c (simplify_compare_const): Add MODE argument.
8157 Handle mode_width 0 as very large mode_width.
8158 (try_combine, simplify_comparison): Adjust callers.
8160 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
8161 type to avoid signed integer overflow.
8162 * explow.c (plus_constant): Likewise.
8164 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
8166 * doc/generic.texi: Correct typos.
8168 2014-03-24 Tobias Burnus <burnus@net-b.de>
8170 * doc/invoke.texi (-flto): Expand section about
8171 using static libraries with LTO.
8173 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8175 PR rtl-optimization/60501
8176 * optabs.def (addptr3_optab): New optab.
8177 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
8178 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
8179 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
8181 * lra.c (emit_add3_insn): Use the addptr pattern if available.
8183 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
8185 2014-03-24 Ulrich Drepper <drepper@gmail.com>
8187 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
8190 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
8191 (_mm256_undefined_ps): Define.
8192 (_mm256_undefined_pd): Define.
8193 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
8194 (_mm_undefined_pd): Define.
8195 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
8196 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
8197 (_mm512_undefined_ps): Define.
8198 (_mm512_undefined_pd): Define.
8199 Use _mm*_undefined_*.
8200 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
8202 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
8204 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
8205 (lshr_simd): DI mode added.
8206 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
8207 (aarch64_ushr_simddi): Likewise.
8208 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
8209 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
8210 (vshrd_n_u64): Likewise.
8212 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8214 * Makefile.in (s-macro_list): Depend on cc1.
8216 2014-03-23 Teresa Johnson <tejohnson@google.com>
8218 * ipa-utils.c (ipa_print_order): Use specified dump file.
8220 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
8222 PR rtl-optimization/60601
8223 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
8225 * gcc.c (eval_spec_function): Initialize save_growing_value.
8227 2014-03-22 Jakub Jelinek <jakub@redhat.com>
8230 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
8231 code == MINUS_EXPR, never swap op0 with op1.
8233 * toplev.c (init_local_tick): Avoid signed integer multiplication
8235 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
8236 shift by first operand's bitsize.
8238 2014-03-21 Jakub Jelinek <jakub@redhat.com>
8241 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
8243 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
8244 TARGET_ISA_64BIT_P(x).
8246 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8248 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
8249 pattern for vector nor instead of subtract from splat(-1).
8250 (altivec_expand_vec_perm_const_le): Likewise.
8252 2014-03-21 Richard Henderson <rth@twiddle.net>
8255 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
8256 related insns after epilogue_completed.
8258 2014-03-21 Martin Jambor <mjambor@suse.cz>
8261 * cgraph.h (symtab_node): New flag body_removed.
8262 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
8263 when removing bodies.
8264 * symtab.c (dump_symtab_base): Dump body_removed flag.
8265 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
8266 had their bodies removed.
8268 2014-03-21 Martin Jambor <mjambor@suse.cz>
8271 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
8274 2014-03-21 Richard Biener <rguenther@suse.de>
8276 PR tree-optimization/60577
8277 * tree-core.h (struct tree_base): Document nothrow_flag use
8278 in VAR_DECL_NONALIASED.
8279 * tree.h (VAR_DECL_NONALIASED): New.
8280 (may_be_aliased): Adjust.
8281 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
8283 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
8285 * expr.c (expand_expr_real_1): Remove outdated comment.
8287 2014-03-20 Jakub Jelinek <jakub@redhat.com>
8290 * ira.c (adjust_cleared_regs): Call copy_rtx on
8291 *reg_equiv[REGNO (loc)].src_p before passing it to
8292 simplify_replace_fn_rtx.
8295 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
8296 into CONST, put pic register as first operand of PLUS. Use
8297 gen_const_mem for both 32-bit and 64-bit PIC got loads.
8299 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8301 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
8303 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
8305 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
8306 around for store forwarding issue in the FPU on the UT699.
8307 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
8308 loads and operations if -mfix-ut699 is specified.
8309 (divtf3_hq): Tweak attribute.
8310 (sqrttf2_hq): Likewise.
8312 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
8314 * calls.c (store_one_arg): Remove incorrect const qualification on the
8315 type of the temporary.
8316 * cfgexpand.c (expand_return): Likewise.
8317 * expr.c (expand_constructor): Likewise.
8318 (expand_expr_real_1): Likewise.
8320 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
8322 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
8325 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
8328 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
8330 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
8332 * config/arm/aarch-common-protos.h
8333 (alu_cost_table): Fix spelling of "extend".
8334 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
8336 2014-03-19 Richard Biener <rguenther@suse.de>
8339 * tree-core.h (tree_type_common): Re-order pointer members
8340 to reduce recursion depth during GC walks.
8342 2014-03-19 Marek Polacek <polacek@redhat.com>
8345 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
8346 before accessing it.
8348 2014-03-19 Richard Biener <rguenther@suse.de>
8351 * lto-streamer-in.c (input_function): In WPA stage do not drop
8354 2014-03-19 Jakub Jelinek <jakub@redhat.com>
8356 PR tree-optimization/60559
8357 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
8358 with build_zero_cst assignment.
8360 2014-03-18 Kai Tietz <ktietz@redhat.com>
8362 PR rtl-optimization/56356
8363 * sdbout.c (sdbout_parms): Verify that parms'
8364 incoming argument is valid.
8365 (sdbout_reg_parms): Likewise.
8367 2014-03-18 Richard Henderson <rth@redhat.com>
8370 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
8371 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
8372 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
8374 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
8376 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
8377 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
8378 Italicize plugin event names in description. Explain that
8379 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
8380 Remind that no GCC functions should be called after PLUGIN_FINISH.
8381 Explain what pragmas with expansion are.
8383 2014-03-18 Martin Liska <mliska@suse.cz>
8385 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
8386 gimple call statement is update.
8387 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
8388 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
8390 2014-03-18 Jakub Jelinek <jakub@redhat.com>
8393 * ubsan.c (ubsan_instrument_unreachable): Call
8394 initialize_sanitizer_builtins.
8395 (ubsan_pass): Likewise.
8398 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
8399 varpool_finalize_decl instead of rest_of_decl_compilation.
8401 2014-03-18 Richard Biener <rguenther@suse.de>
8403 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
8404 by using bitmap_and_compl instead of bitmap_and_compl_into.
8405 (df_rd_transfer_function): Likewise.
8407 2014-03-18 Richard Biener <rguenther@suse.de>
8409 * doc/lto.texi (fresolution): Fix typo.
8411 2014-03-18 Richard Biener <rguenther@suse.de>
8413 * doc/invoke.texi (flto): Update for changes in 4.9.
8415 2014-03-18 Richard Biener <rguenther@suse.de>
8417 * doc/loop.texi: Remove section on the removed lambda framework.
8418 Update loop docs with recent changes in preserving loop structure.
8420 2014-03-18 Richard Biener <rguenther@suse.de>
8422 * doc/lto.texi (-fresolution): Document.
8424 2014-03-18 Richard Biener <rguenther@suse.de>
8426 * doc/contrib.texi: Adjust my name.
8428 2014-03-18 Jakub Jelinek <jakub@redhat.com>
8431 * internal-fn.c: Include diagnostic-core.h.
8432 (expand_BUILTIN_EXPECT): New function.
8433 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
8434 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
8435 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
8436 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
8438 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
8439 Revert 3 argument __builtin_expect code.
8440 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
8441 * gimple-fold.c (gimple_fold_call): Likewise.
8442 * tree.h (fold_builtin_expect): New prototype.
8443 * builtins.c (build_builtin_expect_predicate): Add predictor
8444 argument, if non-NULL, create 3 argument __builtin_expect.
8445 (fold_builtin_expect): No longer static. Add ARG2 argument,
8446 pass it through to build_builtin_expect_predicate.
8447 (fold_builtin_2): Adjust caller.
8448 (fold_builtin_3): Handle BUILT_IN_EXPECT.
8449 * internal-fn.def (BUILTIN_EXPECT): New.
8451 2014-03-18 Tobias Burnus <burnus@net-b.de>
8454 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
8455 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
8456 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
8458 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
8461 * predict.c (combine_predictions_for_bb): Fix up formatting.
8462 (expr_expected_value_1, expr_expected_value): Add predictor argument,
8463 fill what it points to if non-NULL.
8464 (tree_predict_by_opcode): Adjust caller, use the predictor.
8465 * predict.def (PRED_COMPARE_AND_SWAP): Add.
8467 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
8469 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
8470 proper constant for the store mode.
8472 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
8474 * symtab.c (change_decl_assembler_name): Fix transparent alias
8477 2014-03-16 Renlin Li <Renlin.Li@arm.com>
8479 * config/aarch64/aarch64.c: Correct the comments about the
8480 aarch64 stack layout.
8482 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
8484 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
8485 check for GF_OMP_FOR_KIND_FOR.
8487 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
8489 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
8490 ymm and zmm register names.
8492 2014-03-17 Jakub Jelinek <jakub@redhat.com>
8495 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
8496 note creation for the 2010-08-31 changes.
8498 2014-03-17 Marek Polacek <polacek@redhat.com>
8501 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
8502 as -fno-tree-loop-vectorize.
8503 (expand_omp_simd): Likewise.
8505 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
8507 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
8508 (eligible_for_call_delay): New prototype.
8509 * config/sparc/sparc.c (tls_call_delay): Rename into...
8510 (eligible_for_call_delay): ...this. Return false if the instruction
8511 cannot be put in the delay slot of a branch.
8512 (eligible_for_restore_insn): Simplify.
8513 (eligible_for_return_delay): Return false if the instruction cannot be
8514 put in the delay slot of a branch and simplify.
8515 (eligible_for_sibcall_delay): Return false if the instruction cannot be
8516 put in the delay slot of a branch.
8517 * config/sparc/sparc.md (fix_ut699): New attribute.
8518 (tls_call_delay): Delete.
8519 (in_call_delay): Reimplement.
8520 (eligible_for_sibcall_delay): Rename into...
8521 (in_sibcall_delay): ...this.
8522 (eligible_for_return_delay): Rename into...
8523 (in_return_delay): ...this.
8524 (in_branch_delay): Reimplement.
8525 (in_uncond_branch_delay): Delete.
8526 (in_annul_branch_delay): Delete.
8528 2014-03-14 Richard Henderson <rth@redhat.com>
8531 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
8532 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
8533 (*floathi<X87MODEF>2_i387_with_temp): Remove.
8534 (floathi splitters): Remove.
8535 (float<SWI48x>xf2): New pattern.
8536 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
8537 code that tried to handle DImode for 32-bit, but which was excluded
8538 by the pattern's condition. Drop allocation of stack temporary.
8539 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
8540 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
8541 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
8542 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
8543 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
8544 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
8545 (*float<SWI48><MODEF>2_sse_interunit): Remove.
8546 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
8547 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
8548 (*float<SWI48x><X87MODEF>2_i387): Remove.
8549 (all float _with_temp splitters): Remove.
8550 (*float<SWI48x><MODEF>2_i387): New pattern.
8551 (*float<SWI48><MODEF>2_sse): New pattern.
8552 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
8553 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
8555 2014-03-14 Jakub Jelinek <jakub@redhat.com>
8556 Marek Polacek <polacek@redhat.com>
8559 * common.opt (dump_base_name_prefixed): New Variable.
8560 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
8561 if x_dump_base_name_prefixed is already set, set it at the end.
8563 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
8565 PR rtl-optimization/60508
8566 * lra-constraints.c (get_reload_reg): Add new parameter
8568 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
8569 Pass the new parameter values.
8571 2014-03-14 Richard Biener <rguenther@suse.de>
8573 * common.opt: Revert unintented changes from r205065.
8576 2014-03-14 Richard Biener <rguenther@suse.de>
8579 * cfghooks.c (split_block): Properly adjust all loops the
8580 block was a latch of.
8582 2014-03-14 Martin Jambor <mjambor@suse.cz>
8585 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
8588 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
8591 * config/avr/avr.c (avr_set_current_function): Pass function name
8592 through default_strip_name_encoding before sanity checking instead
8593 of skipping the first char of the assembler name.
8595 2014-03-13 Richard Henderson <rth@redhat.com>
8598 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
8599 (ix86_force_to_memory, ix86_free_from_memory): Remove.
8600 * config/i386/i386-protos.h: Likewise.
8601 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
8602 in the expander instead of a splitter.
8603 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
8604 any possibility of requiring a memory.
8605 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
8606 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
8607 (fp branch splitters): Update for ix86_split_fp_branch.
8608 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
8609 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
8610 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
8611 (*fop_<MODEF>_2_i387): Remove f/r alternative.
8612 (*fop_<MODEF>_3_i387): Likewise.
8613 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
8614 (splitters for the fop_* register patterns): Remove.
8615 (fscalexf4_i387): Rename from *fscalexf4_i387.
8616 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
8618 2014-03-13 Jakub Jelinek <jakub@redhat.com>
8620 PR tree-optimization/59779
8621 * tree-dfa.c (get_ref_base_and_extent): Use double_int
8622 type for bitsize and maxsize instead of HOST_WIDE_INT.
8624 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
8626 PR rtl-optimization/57320
8627 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
8628 the CFG after thread_prologue_and_epilogue_insns.
8630 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
8632 PR rtl-optimization/57189
8633 * lra-constraints.c (process_alt_operands): Disfavor spilling
8636 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
8638 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
8640 2014-03-13 Jakub Jelinek <jakub@redhat.com>
8642 PR tree-optimization/59025
8644 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
8645 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
8647 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
8650 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
8651 calls of avr_out_plus_1.
8653 2014-03-13 Bin Cheng <bin.cheng@arm.com>
8655 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
8656 BB's single pred and update the father loop's latch info later.
8658 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
8660 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
8664 (VEC_base): Likewise.
8665 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
8666 registers, we need to swap double words in little endian mode.
8668 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
8669 to be a container mode for 128-bit integer operations added in ISA
8670 2.07. Unlike TImode and PTImode, the preferred register set is
8671 the Altivec/VMX registers for the 128-bit operations.
8673 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
8675 (rs6000_split_128bit_ok_p): Likewise.
8677 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
8678 macros for creating ISA 2.07 normal and overloaded builtin
8679 functions with 3 arguments.
8680 (BU_P8V_OVERLOAD_3): Likewise.
8681 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
8682 for use as overloaded functions.
8683 (VPERM_1TI_UNS): Likewise.
8684 (VSEL_1TI): Likewise.
8685 (VSEL_1TI_UNS): Likewise.
8686 (ST_INTERNAL_1ti): Likewise.
8687 (LD_INTERNAL_1ti): Likewise.
8688 (XXSEL_1TI): Likewise.
8689 (XXSEL_1TI_UNS): Likewise.
8690 (VPERM_1TI): Likewise.
8691 (VPERM_1TI_UNS): Likewise.
8692 (XXPERMDI_1TI): Likewise.
8693 (SET_1TI): Likewise.
8694 (LXVD2X_V1TI): Likewise.
8695 (STXVD2X_V1TI): Likewise.
8696 (VEC_INIT_V1TI): Likewise.
8697 (VEC_SET_V1TI): Likewise.
8698 (VEC_EXT_V1TI): Likewise.
8699 (EQV_V1TI): Likewise.
8700 (NAND_V1TI): Likewise.
8701 (ORC_V1TI): Likewise.
8702 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
8703 added in ISA 2.07. Add both normal 'altivec' builtins, and the
8705 (VADDUQM): Likewise.
8706 (VSUBCUQ): Likewise.
8707 (VADDEUQM): Likewise.
8708 (VADDECUQ): Likewise.
8709 (VSUBEUQM): Likewise.
8710 (VSUBECUQ): Likewise.
8712 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
8713 __int128_t and __uint128_t types.
8714 (__uint128_type): Likewise.
8715 (altivec_categorize_keyword): Add support for vector __int128_t,
8716 vector __uint128_t, vector __int128, and vector unsigned __int128
8717 as a container type for TImode operations that need to be done in
8718 VSX/Altivec registers.
8719 (rs6000_macro_to_expand): Likewise.
8720 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
8721 to support 128-bit integer instructions vaddcuq, vadduqm,
8722 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
8723 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
8725 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
8726 for V1TImode, and set up preferences to use VSX/Altivec registers.
8727 Setup VSX reload handlers.
8728 (rs6000_debug_reg_global): Likewise.
8729 (rs6000_init_hard_regno_mode_ok): Likewise.
8730 (rs6000_preferred_simd_mode): Likewise.
8731 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
8732 (easy_altivec_constant): Likewise.
8733 (output_vec_const_move): Likewise.
8734 (rs6000_expand_vector_set): Convert V1TImode set and extract to
8736 (rs6000_expand_vector_extract): Likewise.
8737 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
8739 (rs6000_const_vec): Add support for V1TImode.
8740 (rs6000_emit_le_vsx_load): Swap double words when loading or
8741 storing TImode/V1TImode.
8742 (rs6000_emit_le_vsx_store): Likewise.
8743 (rs6000_emit_le_vsx_move): Likewise.
8744 (rs6000_emit_move): Add support for V1TImode.
8745 (altivec_expand_ld_builtin): Likewise.
8746 (altivec_expand_st_builtin): Likewise.
8747 (altivec_expand_vec_init_builtin): Likewise.
8748 (altivec_expand_builtin): Likewise.
8749 (rs6000_init_builtins): Add support for V1TImode type. Add
8750 support for ISA 2.07 128-bit integer builtins. Define type names
8751 for the VSX/Altivec vector types.
8752 (altivec_init_builtins): Add support for overloaded vector
8753 functions with V1TImode type.
8754 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
8755 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
8757 (rs6000_split_128bit_ok_p): Likewise.
8758 (rs6000_handle_altivec_attribute): Create V1TImode from vector
8759 __int128_t and vector __uint128_t.
8761 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
8762 and mode attributes.
8769 (VS_scalar): Likewise.
8770 (VS_double): Likewise.
8771 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
8773 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
8774 we support the ISA 2.07 128-bit integer arithmetic instructions.
8775 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
8776 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
8777 and TImode types for use with the builtin functions.
8778 (V1TI_type_node): Likewise.
8779 (unsigned_V1TI_type_node): Likewise.
8780 (intTI_type_internal_node): Likewise.
8781 (uintTI_type_internal_node): Likewise.
8783 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
8784 128-bit builtin functions.
8785 (UNSPEC_VADDEUQM): Likewise.
8786 (UNSPEC_VADDECUQ): Likewise.
8787 (UNSPEC_VSUBCUQ): Likewise.
8788 (UNSPEC_VSUBEUQM): Likewise.
8789 (UNSPEC_VSUBECUQ): Likewise.
8790 (VM): Add V1TImode to vector mode iterators.
8792 (VI_unit): Likewise.
8793 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
8794 (altivec_vaddcuq): Likewise.
8795 (altivec_vsubuqm): Likewise.
8796 (altivec_vsubcuq): Likewise.
8797 (altivec_vaddeuqm): Likewise.
8798 (altivec_vaddecuq): Likewise.
8799 (altivec_vsubeuqm): Likewise.
8800 (altivec_vsubecuq): Likewise.
8802 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
8804 (BOOL_128): Likewise.
8805 (BOOL_REGS_OUTPUT): Likewise.
8806 (BOOL_REGS_OP1): Likewise.
8807 (BOOL_REGS_OP2): Likewise.
8808 (BOOL_REGS_UNARY): Likewise.
8809 (BOOL_REGS_AND_CR0): Likewise.
8811 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
8812 128-bit integer builtin support.
8813 (vec_vadduqm): Likewise.
8814 (vec_vaddecuq): Likewise.
8815 (vec_vaddeuqm): Likewise.
8816 (vec_vsubecuq): Likewise.
8817 (vec_vsubeuqm): Likewise.
8818 (vec_vsubcuq): Likewise.
8819 (vec_vsubuqm): Likewise.
8821 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
8822 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
8823 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
8824 128-bit integer add/subtract to ISA 2.07.
8826 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
8828 * config/arc/arc.c (arc_predicate_delay_insns):
8829 Fix third argument passed to conditionalize_nonjump.
8831 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
8833 * config/aarch64/aarch64-builtins.c
8834 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
8835 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
8836 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
8837 instead of __builtin_lfloor.
8838 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
8840 2014-03-12 Jakub Jelinek <jakub@redhat.com>
8842 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
8843 (tree_ssa_ifcombine_bb_1): New function.
8844 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
8845 is an empty forwarder block to then_bb or vice versa and then_bb
8846 and else_bb are effectively swapped.
8848 2014-03-12 Christian Bruel <christian.bruel@st.com>
8851 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
8852 REG_CFA_DEF_CFA note.
8853 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
8854 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
8856 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
8858 PR tree-optimization/60454
8859 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
8861 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8863 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
8864 Do not define target_cpu_default2 to generic.
8865 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
8866 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
8867 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
8869 2014-03-12 Jakub Jelinek <jakub@redhat.com>
8870 Marc Glisse <marc.glisse@inria.fr>
8872 PR tree-optimization/60502
8873 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
8874 instead of build_low_bits_mask.
8876 2014-03-12 Jakub Jelinek <jakub@redhat.com>
8879 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
8880 if there are multiple uses, but op doesn't live on E edge.
8881 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
8882 clobber stmts before __builtin_unreachable.
8884 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
8886 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
8887 hard_frame_pointer_rtx.
8888 * cse.c (cse_insn): Remove volatile check.
8889 * cselib.c (cselib_process_insn): Likewise.
8890 * dse.c (scan_insn): Likewise.
8892 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
8894 * config/arc/arc.c (conditionalize_nonjump): New function,
8896 (arc_ifcvt): ... this.
8897 (arc_predicate_delay_insns): Use it.
8899 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
8901 * config/arc/predicates.md (extend_operand): During/after reload,
8902 allow const_int_operand.
8903 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
8904 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
8905 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
8907 (umulsi3_highpart_i): Likewise.
8909 2014-03-11 Richard Biener <rguenther@suse.de>
8911 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
8912 Add asserts to guard possible wrong-code bugs.
8914 2014-03-11 Richard Biener <rguenther@suse.de>
8916 PR tree-optimization/60429
8917 PR tree-optimization/60485
8918 * tree-ssa-structalias.c (set_union_with_increment): Properly
8919 take into account all fields that overlap the shifted vars.
8920 (do_sd_constraint): Likewise.
8921 (do_ds_constraint): Likewise.
8922 (get_constraint_for_ptr_offset): Likewise.
8924 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
8926 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
8927 (nios2_compute_frame_layout):
8928 Add calculation of cfun->machine->fp_save_offset.
8929 (nios2_expand_prologue): Correct setting of frame pointer register
8931 (nios2_expand_epilogue): Update recovery of stack pointer from
8932 frame pointer accordingly.
8933 (nios2_initial_elimination_offset): Update calculation of offset
8934 for eliminating to HARD_FRAME_POINTER_REGNUM.
8936 2014-03-10 Jakub Jelinek <jakub@redhat.com>
8939 * ipa.c (symtab_remove_unreachable_nodes): Don't call
8940 cgraph_get_create_node on VAR_DECLs.
8942 2014-03-10 Richard Biener <rguenther@suse.de>
8945 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
8947 2014-03-08 Douglas B Rupp <rupp@gnat.com>
8949 * config/vms/vms.opt (vms_float_format): New variable.
8951 2014-03-08 Tobias Burnus <burnus@net-b.de>
8953 * doc/invoke.texi (-fcilkplus): Update implementation status.
8955 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
8956 Richard Biener <rguenther@suse.de>
8958 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
8959 consistently accross all TUs.
8960 (run_gcc): Enable -fshort-double automatically at link at link-time
8961 and disallow override.
8963 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
8966 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
8967 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
8968 if they can't be used.
8970 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8972 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
8973 for Solaris 11/x86 ld.
8974 * configure: Regenerate.
8976 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8978 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
8979 (LIB_TLS_SPEC): Save as ld_tls_libs.
8980 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
8981 (HAVE_AS_IX86_TLSLDM): New test.
8982 * configure, config.in: Regenerate.
8983 * config/i386/i386.c (legitimize_tls_address): Fall back to
8984 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
8985 cannot support TLS_MODEL_LOCAL_DYNAMIC.
8986 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
8987 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
8989 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
8991 * common.opt (fira-loop-pressure): Mark as optimization.
8993 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
8995 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
8996 an OpenMP mappable type.
8998 2014-03-06 Matthias Klose <doko@ubuntu.com>
9000 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
9001 MULTILIB_OSDIRNAMES is not defined.
9003 2014-03-06 Jakub Jelinek <jakub@redhat.com>
9004 Meador Inge <meadori@codesourcery.com>
9007 * config/arm/arm.c (arm_tls_symbol_p): Remove.
9008 (arm_legitimize_address): Call legitimize_tls_address for any
9009 arm_tls_referenced_p expression, handle constant addend. Call it
9010 before testing for !TARGET_ARM.
9011 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
9013 2014-03-06 Richard Biener <rguenther@suse.de>
9019 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
9021 * tree-streamer.c (record_common_node): Assert we don't record
9022 nodes with type double.
9023 (preload_common_node): Skip type double, complex double and double
9024 pointer since it is now frontend dependent due to fshort-double option.
9026 2014-03-06 Richard Biener <rguenther@suse.de>
9028 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
9029 or -fno-lto is specified and the linker has full plugin support.
9030 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
9031 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
9032 * lto-wrapper.c (merge_and_complain): Merge compile-time
9033 optimization levels.
9034 (run_gcc): And pass it through to the link options.
9036 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
9040 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
9042 * cselib.c (remove_useless_values): Skip to avoid quadratic
9043 behavior if the condition moved from...
9044 (cselib_process_insn): ... here holds.
9046 2014-03-05 Jakub Jelinek <jakub@redhat.com>
9049 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
9050 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
9053 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
9054 (TM_H): Add x86-tune.def.
9056 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9058 * config/aarch64/aarch64.c (generic_tunings):
9059 Use cortexa57_extra_costs.
9061 2014-03-05 Jakub Jelinek <jakub@redhat.com>
9064 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
9065 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
9066 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
9069 2014-03-04 Heiher <r@hev.cc>
9071 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
9072 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
9074 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
9076 * config/i386/predicates.md (const2356_operand): Change to ...
9077 (const2367_operand): ... this.
9078 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
9080 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
9081 (*avx512pf_scatterpf<mode>sf): Ditto.
9082 (avx512pf_scatterpf<mode>df): Ditto.
9083 (*avx512pf_scatterpf<mode>df_mask): Ditto.
9084 (*avx512pf_scatterpf<mode>df): Ditto.
9085 * config/i386/i386.c (ix86_expand_builtin): Update
9086 incorrect hint operand error message.
9088 2014-03-04 Richard Biener <rguenther@suse.de>
9090 * lto-section-in.c (lto_get_section_data): Fix const cast.
9092 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
9094 * tree-streamer.c (record_common_node): Assert we don't record
9095 nodes with type double.
9096 (preload_common_node): Skip type double, complex double and double
9097 pointer since it is now frontend dependent due to fshort-double option.
9099 2014-03-04 Richard Biener <rguenther@suse.de>
9102 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
9103 (lto_input_toplevel_asms): Likewise.
9104 * lto-section-in.c (lto_get_section_data): Instead do it here
9107 2014-03-04 Richard Biener <rguenther@suse.de>
9109 PR tree-optimization/60382
9110 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
9111 dead PHIs a reduction.
9113 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
9115 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
9117 (_mm_prefetch): Move out of GCC target("sse") pragma.
9118 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
9119 GCC target("prfchw") pragma.
9120 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
9122 * config/i386/i386.c (ix86_option_override_internal): Enable
9123 -mprfchw with -mprefetchwt1.
9125 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
9127 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
9130 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
9132 * opts.h (CL_PCH_IGNORE): Define.
9133 * targhooks.c (option_affects_pch_p):
9134 Return false for options that have CL_PCH_IGNORE set.
9135 * opt-functions.awk: Process PchIgnore.
9136 * doc/options.texi: Document PchIgnore.
9138 * config/arc/arc.opt (misize): Add PchIgnore property.
9140 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9142 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
9143 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
9144 constraint on constants to permit them being loaded into
9145 GENERAL_REGS or BASE_REGS.
9147 2014-03-03 Nick Clifton <nickc@redhat.com>
9149 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
9150 anti-cacnonical alternatives.
9151 (negandhi3_real): New pattern.
9152 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
9154 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
9156 * config/avr/avr-mcus.def: Remove atxmega16x1.
9157 * config/avr/avr-tables.opt: Regenerate.
9158 * config/avr/t-multilib: Regenerate.
9159 * doc/avr-mmcu.texi: Regenerate.
9161 2014-03-03 Tobias Grosser <tobias@grosser.es>
9162 Mircea Namolaru <mircea.namolaru@inria.fr>
9164 PR tree-optimization/58028
9165 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
9168 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
9170 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
9171 not handled by recognizers.
9173 2014-03-03 Jakub Jelinek <jakub@redhat.com>
9176 * function.c (expand_function_end): Don't emit
9177 clobber_return_register sequence if clobber_after is a BARRIER.
9178 * cfgexpand.c (construct_exit_block): Append instructions before
9179 return_label to prev_bb.
9181 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9183 * config/rs6000/constraints.md: Document reserved use of "wc".
9185 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
9188 * ipa.c (function_and_variable_visibility): When dissolving comdat
9189 group, also set all symbols to local.
9191 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
9196 2013-12-14 Jan Hubicka <jh@suse.cz>
9198 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
9200 2014-03-02 Jon Beniston <jon@beniston.com>
9206 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
9207 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
9208 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
9209 (simple_return, *simple_return): New patterns
9210 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
9211 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
9213 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
9215 * dwarf2out.c (gen_subprogram_die): Tidy.
9217 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
9220 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
9221 (*mov_t_msb_neg_negc): ... this new insn.
9223 2014-02-28 Jason Merrill <jason@redhat.com>
9226 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
9229 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
9232 * dwarf2out.c (decltype_auto_die): New static.
9233 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
9234 (gen_type_die_with_usage): Handle 'decltype(auto)'.
9235 (is_cxx_auto): Likewise.
9237 2014-02-28 Ian Bolton <ian.bolton@arm.com>
9239 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
9240 we are not using general regs only.
9242 2014-02-28 Richard Biener <rguenther@suse.de>
9245 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
9246 previous fix and only allow to remove trivial pre-headers
9247 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
9248 (remove_forwarder_block): Properly update the latch of a loop.
9250 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
9253 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
9254 (cselib_preserved_hash_table): New.
9255 (preserve_constants_and_equivs): Move preserved vals to it.
9256 (cselib_find_slot): Look it up first.
9257 (cselib_init): Initialize it.
9258 (cselib_finish): Release it.
9259 (dump_cselib_table): Dump it.
9261 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
9264 * cselib.c (remove_useless_values): Skip to avoid quadratic
9265 behavior if the condition moved from...
9266 (cselib_process_insn): ... here holds.
9268 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
9271 * var-tracking.c (vt_initialize): Apply the same condition to
9272 preserve the CFA base value.
9274 2014-02-28 Joey Ye <joey.ye@arm.com>
9277 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
9278 if reload in progress or completed.
9280 2014-02-28 Tobias Burnus <burnus@net-b.de>
9283 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
9286 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
9288 * doc/tm.texi.in (Condition Code Status): Update documention for
9289 relative locations of cc0-setter and cc0-user.
9291 2014-02-27 Jeff Law <law@redhat.com>
9293 PR rtl-optimization/52714
9294 * combine.c (try_combine): When splitting an unrecognized PARALLEL
9295 into two independent simple sets, if I3 is a jump, ensure the
9296 pattern we place into I3 is a (set (pc) ...).
9298 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
9299 Jeff Law <law@redhat.com>
9301 PR rtl-optimization/49847
9302 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
9303 are in different blocks.
9304 * doc/tm.texi (Condition Code Status): Update documention for
9305 relative locations of cc0-setter and cc0-user.
9307 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
9310 * lra.c (lra_emit_add): Check SUBREG too.
9312 2014-02-27 Andreas Schwab <schwab@suse.de>
9314 * config/m68k/m68k.c (m68k_option_override): Disable
9315 -flive-range-shrinkage for classic m68k.
9316 (m68k_override_options_after_change): Likewise.
9318 2014-02-27 Marek Polacek <polacek@redhat.com>
9321 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
9322 -Wmaybe-uninitialized.
9324 2014-02-27 Alan Modra <amodra@gmail.com>
9327 * reload1.c (emit_input_reload_insns): When reload_override_in,
9328 set old to rl->in_reg when rl->in_reg is a subreg.
9330 2014-02-26 Richard Biener <rguenther@suse.de>
9333 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
9335 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
9337 * common/config/i386/predicates.md (const1256_operand): Remove.
9338 (const2356_operand): New.
9339 (const_1_to_2_operand): Remove.
9340 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
9341 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
9342 (*avx512pf_gatherpf<mode>sf): Ditto.
9343 (avx512pf_gatherpf<mode>df): Ditto.
9344 (*avx512pf_gatherpf<mode>df_mask): Ditto.
9345 (*avx512pf_gatherpf<mode>df): Ditto.
9346 (avx512pf_scatterpf<mode>sf): Ditto.
9347 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
9348 (*avx512pf_scatterpf<mode>sf): Ditto.
9349 (avx512pf_scatterpf<mode>df): Ditto.
9350 (*avx512pf_scatterpf<mode>df_mask): Ditto.
9351 (*avx512pf_scatterpf<mode>df): Ditto.
9352 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
9354 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
9356 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
9357 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
9358 (_mm512_mask_testn_epi64_mask): Move to ...
9359 * config/i386/avx512cdintrin.h: Here.
9360 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
9361 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
9362 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
9363 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
9364 TARGET_AVX512F from TARGET_AVX512CD.
9366 2014-02-26 Richard Biener <rguenther@suse.de>
9369 * ipa.c (walk_polymorphic_call_targets): Properly guard
9370 call to inline_update_overall_summary.
9372 2014-02-26 Bin Cheng <bin.cheng@arm.com>
9375 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
9376 and latches only if requested. Fix latch if it is removed.
9377 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
9378 LOOPS_HAVE_PREHEADERS.
9380 2014-02-25 Andrew Pinski <apinski@cavium.com>
9382 * builtins.c (expand_builtin_thread_pointer): Create a new target
9383 when the target is NULL.
9385 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
9387 PR rtl-optimization/60317
9388 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
9389 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
9390 * lra-assigns.c: Include params.h.
9391 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
9392 other reload pseudos considerations.
9394 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9396 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
9397 to use canonical form for nor<mode>3.
9399 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9402 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
9405 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
9407 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
9408 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
9409 (ix86_handle_option): Handle OPT_mprefetchwt1.
9410 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
9411 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
9413 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
9414 OPTION_MASK_ISA_PREFETCHWT1.
9415 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
9416 (PTA_PREFETCHWT1): New.
9417 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
9418 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
9419 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
9420 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
9421 (*prefetch_avx512pf_<mode>_: Change into ...
9422 (*prefetch_prefetchwt1_<mode>: This.
9423 * config/i386/i386.opt (mprefetchwt1): New.
9424 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
9425 (_mm_prefetch): Handle intent to write.
9426 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
9428 2014-02-25 Richard Biener <rguenther@suse.de>
9431 * emit-rtl.c (mem_attrs_htab): Remove.
9432 (mem_attrs_htab_hash): Likewise.
9433 (mem_attrs_htab_eq): Likewise.
9434 (set_mem_attrs): Always allocate new mem-attrs when something changed.
9435 (init_emit_once): Do not allocate mem_attrs_htab.
9437 2014-02-25 Richard Biener <rguenther@suse.de>
9440 * lto-opts.c (lto_write_options): Output non-explicit conservative
9441 -fwrapv, -fno-trapv and -fno-strict-overflow.
9442 * lto-wrapper.c (merge_and_complain): Handle merging those options.
9443 (run_gcc): And pass them through.
9445 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
9447 * sel-sched.c (calculate_new_fences): New parameter ptime.
9448 Calculate it as a maximum over all fence cycles.
9449 (sel_sched_region_2): Adjust the call to calculate_new_fences.
9450 Print the final schedule timing when sched_verbose.
9452 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
9454 PR rtl-optimization/60292
9455 * sel-sched.c (fill_vec_av_set): Do not reset target availability
9456 bit fot the fence instruction.
9458 2014-02-24 Alangi Derick <alangiderick@gmail.com>
9460 * calls.h: Fix typo in comment.
9462 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
9464 * config/pa/pa.c (pa_output_move_double): Don't valididate when
9465 adjusting offsetable addresses.
9467 2014-02-24 Guozhi Wei <carrot@google.com>
9469 * sparseset.h (sparseset_pop): Fix the wrong index.
9471 2014-02-24 Walter Lee <walt@tilera.com>
9473 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
9474 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
9476 * common/config/tilegx/tilegx-common.c
9477 (TARGET_DEFAULT_TARGET_FLAGS): Define.
9478 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
9480 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
9481 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
9482 (tilegx_gimplify_va_arg_expr): Handle big endian.
9483 (tilegx_expand_unaligned_load): Ditto.
9484 (tilegx_expand_unaligned_store): Ditto.
9485 (TARGET_RETURN_IN_MSB): New.
9486 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
9487 (TARGET_ENDIAN_DEFAULT): New.
9488 (TARGET_BIG_ENDIAN): Handle big endian.
9489 (BYTES_BIG_ENDIAN): Ditto.
9490 (WORDS_BIG_ENDIAN): Ditto.
9491 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
9494 * config/tilegx/tilegx.md (extv): Handle big endian.
9496 (insn_st<n>): Ditto.
9497 (insn_st<n>_add<bitsuffix>): Ditto.
9498 (insn_stnt<n>): Ditto.
9499 (insn_stnt<n>_add<bitsuffix>):Ditto.
9500 (vec_interleave_highv8qi): Handle big endian.
9501 (vec_interleave_highv8qi_be): New.
9502 (vec_interleave_highv8qi_le): New.
9503 (insn_v1int_h): Handle big endian.
9504 (vec_interleave_lowv8qi): Handle big endian.
9505 (vec_interleave_lowv8qi_be): New.
9506 (vec_interleave_lowv8qi_le): New.
9507 (insn_v1int_l): Handle big endian.
9508 (vec_interleave_highv4hi): Handle big endian.
9509 (vec_interleave_highv4hi_be): New.
9510 (vec_interleave_highv4hi_le): New.
9511 (insn_v2int_h): Handle big endian.
9512 (vec_interleave_lowv4hi): Handle big endian.
9513 (vec_interleave_lowv4hi_be): New.
9514 (vec_interleave_lowv4hi_le): New.
9515 (insn_v2int_l): Handle big endian.
9516 (vec_interleave_highv2si): Handle big endian.
9517 (vec_interleave_highv2si_be): New.
9518 (vec_interleave_highv2si_le): New.
9519 (insn_v4int_h): Handle big endian.
9520 (vec_interleave_lowv2si): Handle big endian.
9521 (vec_interleave_lowv2si_be): New.
9522 (vec_interleave_lowv2si_le): New.
9523 (insn_v4int_l): Handle big endian.
9524 * config/tilegx/tilegx.opt (mbig-endian): New option.
9525 (mlittle-endian): New option.
9526 * doc/install.texi: Document tilegxbe-linux.
9527 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
9529 2014-02-24 Martin Jambor <mjambor@suse.cz>
9532 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
9533 there are no parameter descriptors.
9535 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
9537 PR rtl-optimization/60268
9538 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
9539 initialization to ...
9540 (sched_rgn_init): ... here.
9541 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
9543 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
9545 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
9548 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
9550 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
9553 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
9555 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
9556 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
9558 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
9560 * config/microblaze/predicates.md: Add cmp_op predicate.
9561 * config/microblaze/microblaze.md: Add branch_compare instruction
9562 which uses cmp_op predicate and emits cmp insn before branch.
9563 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
9564 to microblaze_expand_conditional_branch and consolidate logic.
9565 (microblaze_expand_conditional_branch): emit branch_compare
9566 insn instead of handling cmp op separate from branch insn.
9568 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9570 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
9573 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9575 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
9576 define_insn with define_expand and new define_insn
9577 *altivec_lve<VI_char>x_internal.
9578 (altivec_stve<VI_char>x): Replace define_insn with define_expand
9579 and new define_insn *altivec_stve<VI_char>x_internal.
9580 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
9582 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
9584 (altivec_expand_stvex_be): New function.
9586 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
9588 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
9589 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
9590 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
9591 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
9593 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
9596 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
9597 instead of emit_move_insn.
9599 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9601 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
9603 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
9604 gen_altivec_vsumsws.
9606 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9608 * config/rs6000/altivec.md (altivec_lvxl): Rename as
9609 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
9610 (altivec_lvxl_<mode>): New define_expand incorporating
9611 -maltivec=be semantics where needed.
9612 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
9613 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
9614 semantics where needed.
9615 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
9616 (altivec_stvx_<mode>): New define_expand incorporating
9617 -maltivec=be semantics where needed.
9618 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
9619 VM2 iterator instead of V4SI.
9620 (altivec_stvxl_<mode>): New define_expand incorporating
9621 -maltivec=be semantics where needed.
9622 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
9623 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
9624 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
9625 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
9626 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
9627 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
9628 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
9629 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
9630 ALTIVEC_BUILTIN_STVXL.
9631 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
9632 (altivec_expand_stvx_be): Likewise.
9633 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
9634 (altivec_expand_lvx_be): Likewise.
9635 (altivec_expand_stvx_be): Likewise.
9636 (altivec_expand_builtin): Add cases for
9637 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
9638 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
9639 (altivec_init_builtins): Add definitions for
9640 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
9641 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
9643 2014-02-21 Catherine Moore <clm@codesourcery.com>
9645 * doc/invoke.texi (mvirt, mno-virt): Document.
9646 * config/mips/mips.opt (mvirt): New option.
9647 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
9649 2014-02-21 Richard Biener <rguenther@suse.de>
9651 PR tree-optimization/60276
9652 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
9653 (STMT_VINFO_MIN_NEG_DIST): New macro.
9654 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
9655 STMT_VINFO_MIN_NEG_DIST.
9656 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
9657 made for negative dependence distances still hold.
9659 2014-02-21 Richard Biener <rguenther@suse.de>
9662 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
9663 DECL_INITIAL for globals not in the current function context.
9665 2014-02-21 Jakub Jelinek <jakub@redhat.com>
9667 PR tree-optimization/56490
9668 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
9669 * tree-ssa-uninit.c: Include params.h.
9670 (compute_control_dep_chain): Add num_calls argument, return false
9671 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
9672 num_calls to recursive call.
9673 (find_predicates): Change dep_chain into normal array,
9674 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
9675 variable and adjust compute_control_dep_chain caller.
9676 (find_def_preds): Likewise.
9678 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
9680 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
9681 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
9683 2014-02-21 Nick Clifton <nickc@redhat.com>
9685 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
9686 (pushhi1): Likewise.
9687 (popqi1): Add mode to pre_dec.
9690 2014-02-21 Jakub Jelinek <jakub@redhat.com>
9692 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
9693 mode for mask of V8SFmode permutation.
9695 2014-02-20 Richard Henderson <rth@redhat.com>
9698 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
9699 a new pseudo for OLDVAL.
9701 2014-02-20 Jakub Jelinek <jakub@redhat.com>
9704 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
9705 gen_reg_rtx if d->testing_p.
9706 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
9707 if d->testing_p and we will certainly return true.
9708 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
9711 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
9713 * emit-rtl.c (gen_reg_rtx): Assert that
9714 crtl->emit.regno_pointer_align_length is non-zero.
9716 2014-02-20 Richard Henderson <rth@redhat.com>
9719 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
9720 on failure the store back into EXPECT.
9722 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
9723 Sandra Loosemore <sandra@codesourcery.com>
9725 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
9726 * config/nios2/nios2.c (nios2_function_profiler): Add
9727 -fPIC (flag_pic == 2) support.
9728 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
9729 (nios2_large_offset_p): New function.
9730 (nios2_unspec_reloc_p): Move up position, update to use
9731 nios2_large_offset_p.
9732 (nios2_unspec_address): Remove function.
9733 (nios2_unspec_offset): New function.
9734 (nios2_large_got_address): New function.
9735 (nios2_got_address): Add large offset support.
9736 (nios2_legitimize_tls_address): Update usage of removed and new
9738 (nios2_symbol_binds_local_p): New function.
9739 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
9740 (nios2_legitimize_address): Update to use nios2_large_offset_p.
9741 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
9742 (nios2_print_operand): Merge H/L processing, add hiadj/lo
9743 processing for (const (unspec ...)).
9744 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
9746 2014-02-20 Richard Biener <rguenther@suse.de>
9748 * tree-cfg.c (replace_uses_by): Mark altered BBs before
9749 doing the substitution.
9750 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
9752 2014-02-20 Martin Jambor <mjambor@suse.cz>
9755 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
9756 info when checking whether lattices are bottom.
9758 2014-02-20 Richard Biener <rguenther@suse.de>
9761 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
9764 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
9767 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
9768 parameter specifying the scaling.
9769 (inline_call): Update.
9770 (want_inline_recursively): Guard division by zero.
9771 (recursive_inlining): Update.
9772 * ipa-inline.h (clone_inlined_nodes): Update.
9774 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
9777 * config/i386/i386.c (classify_argument): Pass structures of size
9778 64 bytes or less in register.
9780 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
9781 Kirill Yukhin <kirill.yukhin@intel.com>
9783 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
9784 (_mm_rcp28_round_ss): Ditto.
9785 (_mm_rsqrt28_round_sd): Ditto.
9786 (_mm_rsqrt28_round_ss): Ditto.
9787 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
9788 (_mm_rcp14_round_ss): Ditto.
9789 (_mm_rsqrt14_round_sd): Ditto.
9790 (_mm_rsqrt14_round_ss): Ditto.
9791 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
9792 the first input operand, get rid of match_dup.
9793 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
9795 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
9797 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
9798 operand as the first input operand, set type attribute.
9799 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
9801 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
9802 operand as the first input operand, set type attribute.
9804 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9806 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
9809 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
9812 * config/i386/i386.c (construct_container): Remove TFmode check
9813 for X86_64_INTEGER_CLASS.
9815 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
9818 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
9819 only when -Wpsabi is enabled.
9821 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
9824 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
9825 passing arrays in registers are the same as for structs, so remove the
9826 special case for them.
9828 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
9830 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
9831 destination type, extract only the valid bits if the source type is not
9832 integral and has a different mode.
9834 2014-02-19 Richard Biener <rguenther@suse.de>
9837 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
9840 2014-02-19 Richard Biener <rguenther@suse.de>
9843 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
9844 (ipa_modify_call_arguments): Emit an argument load explicitely and
9845 preserve virtual SSA form there and for the replacement call.
9846 Do not update SSA form nor free dominance info.
9848 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
9850 * ipa.c (function_and_variable_visibility): Also clear WEAK
9851 flag when disolving COMDAT_GROUP.
9853 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
9855 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
9856 * ipa-prop.c (ipa_set_jf_known_type): Return early when
9858 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
9859 do more sanity checks.
9860 (detect_type_change): Return true when giving up early.
9861 (compute_complex_assign_jump_func): Fix type parameter of
9862 ipa_set_ancestor_jf.
9863 (compute_complex_ancestor_jump_func): Likewise.
9864 (update_jump_functions_after_inlining): Fix updating of
9866 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
9868 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
9870 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
9871 inline clones when edge disappears.
9873 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
9876 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
9877 Split 64-bit moves into 2 patterns. Do not allow the use of
9878 direct move for TDmode in little endian, since the decimal value
9879 has little endian bytes within a word, but the 64-bit pieces are
9880 ordered in a big endian fashion, and normal subreg's of TDmode are
9882 (mov<mode>_64bit_dm): Likewise.
9883 (movtd_64bit_nodm): Likewise.
9885 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
9887 PR tree-optimization/60174
9888 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
9889 statement of an SSA_NAME that occurs in an abnormal PHI node.
9891 2014-02-18 Jakub Jelinek <jakub@redhat.com>
9894 * final.c (SEEN_BB): Remove.
9895 (SEEN_NOTE, SEEN_EMITTED): Renumber.
9896 (final_scan_insn): Don't force_source_line on second
9897 NOTE_INSN_BASIC_BLOCK.
9899 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
9902 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
9903 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
9904 (type_natural_mode): Warn ABI change when %zmm register is not
9905 available for AVX512F vector value passing.
9907 2014-02-18 Kai Tietz <ktietz@redhat.com>
9910 * config/i386/i386.c (ix86_expand_prologue): Use value in
9911 rax register as displacement when restoring %r10 or %rax.
9912 Fix wrong offset when restoring both registers.
9914 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
9916 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
9917 assertion with conditional return.
9919 2014-02-18 Jakub Jelinek <jakub@redhat.com>
9920 Uros Bizjak <ubizjak@gmail.com>
9923 * config/i386/driver-i386.c (host_detect_local_cpu): If
9924 YMM state is not saved by the OS, also clear has_f16c. Move
9925 CPUID 0x80000001 handling before YMM state saving checking.
9927 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
9929 PR rtl-optimization/58960
9930 * haifa-sched.c (alloc_global_sched_pressure_data): New,
9931 factored out from ...
9932 (sched_init): ... here.
9933 (free_global_sched_pressure_data): New, factored out from ...
9934 (sched_finish): ... here.
9935 * sched-int.h (free_global_sched_pressure_data): Declare.
9936 * sched-rgn.c (nr_regions_initial): New static global.
9937 (haifa_find_rgns): Initialize it.
9938 (schedule_region): Disable sched-pressure for the newly
9941 2014-02-17 Richard Biener <rguenther@suse.de>
9943 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
9944 release SSA defs of pattern stmts.
9946 2014-02-17 Richard Biener <rguenther@suse.de>
9948 * tree-inline.c (expand_call_inline): Release the virtual
9949 operand defined by the call we are about to inline.
9951 2014-02-17 Richard Biener <rguenther@suse.de>
9953 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
9955 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
9956 Ilya Tocar <ilya.tocar@intel.com>
9958 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
9959 arguments order in builtin.
9960 (_mm512_permutexvar_epi64): Ditto.
9961 (_mm512_mask_permutexvar_epi64): Ditto
9962 (_mm512_maskz_permutexvar_epi32): Ditto
9963 (_mm512_permutexvar_epi32): Ditto
9964 (_mm512_mask_permutexvar_epi32): Ditto
9966 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9968 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
9969 (p8_vmrgow): Likewise.
9971 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9973 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
9976 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
9979 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
9980 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
9981 into 64-bit and 32-bit moves. On 64-bit moves, add support for
9982 using direct move instructions on ISA 2.07. Also adjust
9983 instruction length for 64-bit.
9984 (mov<mode>_64bit, TFmode/TDmode): Likewise.
9985 (mov<mode>_32bit, TFmode/TDmode): Likewise.
9987 2014-02-15 Alan Modra <amodra@gmail.com>
9991 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
9992 find_replacement on parts of insn rtl that might be reloaded.
9994 2014-02-15 Richard Biener <rguenther@suse.de>
9996 PR tree-optimization/60183
9997 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
9998 (tree_ssa_phiprop): Calculate and free post-dominators.
10000 2014-02-14 Jeff Law <law@redhat.com>
10002 PR rtl-optimization/60131
10003 * ree.c (get_extended_src_reg): New function.
10004 (combine_reaching_defs): Use it rather than assuming location of REG.
10005 (find_and_remove_re): Verify first operand of extension is
10006 a REG before adding the insns to the copy list.
10008 2014-02-14 Roland McGrath <mcgrathr@google.com>
10010 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
10011 * configure: Regenerated.
10012 * config.in: Regenerated.
10013 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
10014 instead of ASM_SHORT.
10016 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
10017 Richard Earnshaw <rearnsha@arm.com>
10019 PR rtl-optimization/59535
10020 * lra-constraints.c (process_alt_operands): Encourage alternative
10021 when unassigned pseudo class is superset of the alternative class.
10022 (inherit_reload_reg): Don't inherit when optimizing for code size.
10023 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
10024 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
10025 modes not less than 4 for Thumb1.
10027 2014-02-14 Kyle McMartin <kyle@redhat.com>
10030 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
10032 2014-02-14 Richard Biener <rguenther@suse.de>
10034 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
10035 (get_frame_arg): Drop the assert with langhook types_compatible_p.
10036 Do not strip INDIRECT_REFs.
10038 2014-02-14 Richard Biener <rguenther@suse.de>
10041 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
10042 DECL_FUNCTION_SPECIFIC_TARGET.
10043 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
10044 * tree-streamer-out.c (pack_ts_target_option): Remove.
10045 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
10046 (write_ts_function_decl_tree_pointers): Do not stream
10047 DECL_FUNCTION_SPECIFIC_TARGET.
10048 * tree-streamer-in.c (unpack_ts_target_option): Remove.
10049 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
10050 (lto_input_ts_function_decl_tree_pointers): Do not stream
10051 DECL_FUNCTION_SPECIFIC_TARGET.
10053 2014-02-14 Jakub Jelinek <jakub@redhat.com>
10055 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
10056 (get_initial_def_for_induction, vectorizable_induction): Ignore
10057 debug stmts when looking for exit_phi.
10058 (vectorizable_live_operation): Fix up condition.
10060 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
10062 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
10063 nreverse() because it changes the content of original tree list.
10065 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
10067 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
10068 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
10070 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
10072 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
10073 GNU coding standards.
10075 2014-02-13 Jakub Jelinek <jakub@redhat.com>
10078 * dwarf2out.c (gen_subprogram_die): Don't call
10079 add_calling_convention_attribute if subr_die is old_die.
10081 2014-02-13 Sharad Singhai <singhai@google.com>
10083 * doc/optinfo.texi: Fix order of nodes.
10085 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
10087 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
10088 operands[2], not operands[3].
10090 2014-02-13 Richard Biener <rguenther@suse.de>
10093 * doc/install.texi (ISL): Update recommended version to 0.12.2,
10094 mention the possibility of an in-tree build.
10095 (CLooG): Update recommended version to 0.18.1, mention the
10096 possibility of an in-tree build and clarify that the ISL
10097 bundled with CLooG does not work.
10099 2014-02-13 Jakub Jelinek <jakub@redhat.com>
10102 * expr.c (compress_float_constant): If x is a hard register,
10103 extend into a pseudo and then move to x.
10105 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
10107 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
10108 caused by bad second argument to warning_at() with -mhotpatch and
10109 nested functions (e.g. with gfortran).
10111 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
10113 * opts.c (option_name): Remove "enabled by default" rider.
10115 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
10117 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
10119 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
10120 Uros Bizjak <ubizjak@gmail.com>
10123 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
10124 * configure: Regenerated.
10126 2014-02-12 Richard Biener <rguenther@suse.de>
10128 * vec.c (vec_prefix::calculate_allocation): Move as
10129 inline variant to vec.h.
10130 (vec_prefix::calculate_allocation_1): New out-of-line version.
10131 * vec.h (vec_prefix::calculate_allocation_1): Declare.
10132 (vec_prefix::m_has_auto_buf): Rename to ...
10133 (vec_prefix::m_using_auto_storage): ... this.
10134 (vec_prefix::calculate_allocation): Inline the easy cases
10135 and dispatch to calculate_allocation_1 which doesn't need the
10137 (va_heap::reserve): Use gcc_checking_assert.
10138 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
10139 m_using_auto_storage.
10140 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
10142 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
10143 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
10144 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
10146 2014-02-12 Richard Biener <rguenther@suse.de>
10148 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
10149 when we found a dependence.
10151 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
10153 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
10155 (maybe_fold_stmt): ... into this new function.
10156 * omp-low.c (lower_omp): Update comment.
10158 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
10161 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
10164 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
10166 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
10167 identifiers in comments.
10168 (cortexa53_extra_costs): Likewise.
10169 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
10170 (cortexa7_extra_costs): Likewise.
10171 (cortexa12_extra_costs): Likewise.
10172 (cortexa15_extra_costs): Likewise.
10173 (v7m_extra_costs): Likewise.
10175 2014-02-12 Richard Biener <rguenther@suse.de>
10177 PR middle-end/60092
10178 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
10179 of posix_memalign being successful.
10180 (lower_stmt): Restrict lowering of posix_memalign to when
10181 -ftree-bit-ccp is enabled.
10183 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
10185 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
10187 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
10189 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
10191 PR rtl-optimization/60116
10192 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
10193 other_insn once the combination has been validated.
10195 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
10198 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
10200 * ipa-devirt.c: Include demangle.h
10201 (odr_violation_reported): New static variable.
10202 (add_type_duplicate): Update odr_violations.
10203 (maybe_record_node): Add completep parameter; update it.
10204 (record_target_from_binfo): Add COMPLETEP parameter;
10205 update it as needed.
10206 (possible_polymorphic_call_targets_1): Likewise.
10207 (struct polymorphic_call_target_d): Add nonconstruction_targets;
10208 rename FINAL to COMPLETE.
10209 (record_targets_from_bases): Sanity check we found the binfo;
10210 fix COMPLETEP updating.
10211 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
10212 parameter, fix computing of COMPLETEP.
10213 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
10214 at LTO time do demangling.
10215 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
10216 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
10218 (gimple_get_virt_method_for_binfo): Likewise.
10219 * gimple-fold.h (gimple_get_virt_method_for_binfo,
10220 gimple_get_virt_method_for_vtable): Update prototypes.
10222 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
10225 * genautomata.c (add_presence_absence): Fix typo with
10226 {final_}presence_list.
10228 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
10231 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
10232 for VSX/Altivec vectors that land in GPR registers.
10234 2014-02-11 Richard Henderson <rth@redhat.com>
10235 Jakub Jelinek <jakub@redhat.com>
10238 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
10239 around drhs if type conversion to lacc->type is not useless.
10241 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10243 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
10245 (cortex-a57.cortex-a53): Likewise.
10246 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
10248 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10250 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
10253 2014-02-11 Renlin Li <Renlin.Li@arm.com>
10255 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
10256 add_options_for_arm_vfp3.
10258 2014-02-11 Jeff Law <law@redhat.com>
10260 PR middle-end/54041
10261 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
10262 object with an undesirable mode.
10264 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
10267 * config/i386/sol2-9.h: New file.
10268 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
10269 *-*-solaris2.9*): Use it.
10271 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
10273 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
10274 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
10276 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
10278 * config/microblaze/microblaze.c: Extend mcpu version format
10280 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
10282 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
10284 2014-02-10 Richard Henderson <rth@redhat.com>
10287 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
10288 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
10289 ms-abi vs -mno-accumulate-outgoing-args.
10290 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
10291 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
10294 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
10296 PR middle-end/60080
10297 * cfgexpand.c (expand_asm_operands): Attach source location to
10298 ASM_INPUT rtx objects.
10299 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
10301 2014-02-10 Nick Clifton <nickc@redhat.com>
10303 * config/mn10300/mn10300.c (popcount): New function.
10304 (mn10300_expand_prologue): Include saved registers in stack usage
10307 2014-02-10 Jeff Law <law@redhat.com>
10309 PR middle-end/52306
10310 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
10311 when changing the SET_DEST of a prior insn to avoid an input reload.
10313 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
10315 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
10316 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
10317 -mcall-openbsd, or -mcall-linux.
10318 (CC1_ENDIAN_BIG_SPEC): Remove.
10319 (CC1_ENDIAN_LITTLE_SPEC): Remove.
10320 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
10321 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
10322 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
10323 and %cc1_endian_default.
10324 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
10326 2014-02-10 Richard Biener <rguenther@suse.de>
10328 PR tree-optimization/60115
10329 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
10330 MEM_REF handling. Properly verify that the accesses are not
10331 out of the objects bound.
10333 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10335 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
10338 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
10340 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
10341 proper constants and fix formatting.
10342 (possible_polymorphic_call_targets): Fix formatting.
10344 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
10345 Ilya Tocar <ilya.tocar@intel.com>
10347 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
10348 (_mm512_loadu_epi32): Renamed into...
10349 (_mm512_loadu_si512): This.
10350 (_mm512_storeu_epi32): Renamed into...
10351 (_mm512_storeu_si512): This.
10352 (_mm512_maskz_ceil_ps): Removed.
10353 (_mm512_maskz_ceil_pd): Ditto.
10354 (_mm512_maskz_floor_ps): Ditto.
10355 (_mm512_maskz_floor_pd): Ditto.
10356 (_mm512_floor_round_ps): Ditto.
10357 (_mm512_floor_round_pd): Ditto.
10358 (_mm512_ceil_round_ps): Ditto.
10359 (_mm512_ceil_round_pd): Ditto.
10360 (_mm512_mask_floor_round_ps): Ditto.
10361 (_mm512_mask_floor_round_pd): Ditto.
10362 (_mm512_mask_ceil_round_ps): Ditto.
10363 (_mm512_mask_ceil_round_pd): Ditto.
10364 (_mm512_maskz_floor_round_ps): Ditto.
10365 (_mm512_maskz_floor_round_pd): Ditto.
10366 (_mm512_maskz_ceil_round_ps): Ditto.
10367 (_mm512_maskz_ceil_round_pd): Ditto.
10368 (_mm512_expand_pd): Ditto.
10369 (_mm512_expand_ps): Ditto.
10370 * config/i386/i386.c (ix86_builtins): Remove
10371 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
10372 (bdesc_args): Ditto.
10373 * config/i386/predicates.md (const1256_operand): New.
10374 (const_1_to_2_operand): Ditto.
10375 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
10376 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
10377 (*avx512pf_gatherpf<mode>sf): Ditto.
10378 (avx512pf_gatherpf<mode>df): Ditto.
10379 (*avx512pf_gatherpf<mode>df_mask): Ditto.
10380 (*avx512pf_gatherpf<mode>df): Ditto.
10381 (avx512pf_scatterpf<mode>sf): Ditto.
10382 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
10383 (*avx512pf_scatterpf<mode>sf): Ditto.
10384 (avx512pf_scatterpf<mode>df): Ditto.
10385 (*avx512pf_scatterpf<mode>df_mask): Ditto.
10386 (*avx512pf_scatterpf<mode>df): Ditto.
10387 (avx512f_expand<mode>): Removed.
10388 (<shift_insn><mode>3<mask_name>): Change predicate type.
10390 2014-02-08 Jakub Jelinek <jakub@redhat.com>
10392 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
10393 not at the end of datarefs vector use ordered_remove to avoid
10394 reordering datarefs vector.
10397 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
10398 mark local addressable non-static vars as GOVD_PRIVATE
10399 instead of GOVD_LOCAL.
10400 * omp-low.c (lower_omp_for): Move gimple_bind_vars
10401 and BLOCK_VARS of gimple_bind_block to new_stmt rather
10404 PR middle-end/60092
10405 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
10406 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
10407 assume_aligned or alloc_align attributes.
10408 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
10409 arguments. Handle also assume_aligned and alloc_align attributes.
10410 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
10411 calls to functions with assume_aligned or alloc_align attributes.
10412 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
10414 2014-02-08 Terry Guo <terry.guo@arm.com>
10416 * doc/invoke.texi: Document ARM -march=armv7e-m.
10418 2014-02-08 Jakub Jelinek <jakub@redhat.com>
10420 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
10421 flag on __cilkrts_rethrow builtin.
10424 * ipa-cp.c (determine_versionability): Fail at -O0
10425 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
10426 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
10429 2014-02-04 Jakub Jelinek <jakub@redhat.com>
10432 * tree-inline.c (copy_forbidden): Fail for
10433 __attribute__((optimize (0))) functions.
10435 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
10437 * varpool.c: Include pointer-set.h.
10438 (varpool_remove_unreferenced_decls): Variables in other partitions
10439 will not be output; be however careful to not lose information
10440 about partitioning.
10442 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
10444 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
10445 lookup in the vtable constructor.
10447 2014-02-07 Jeff Law <law@redhat.com>
10450 * config/m68k/m68k.md (ashldi_extsi): Turn into a
10451 define_insn_and_split.
10453 * ipa-inline.c (inline_small_functions): Fix typos.
10455 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10457 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
10458 (s390_can_use_return_insn): Declare.
10459 * config/s390/s390.h (EPILOGUE_USES): Define.
10460 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
10462 (s390_chunkify_start): Handle return JUMP_LABELs.
10463 (s390_early_mach): Emit a main_pool instruction on the entry edge.
10464 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
10465 (s390_can_use_return_insn): New functions.
10466 (s390_fix_long_loop_prediction): Handle conditional returns.
10467 (TARGET_SET_UP_BY_PROLOGUE): Define.
10468 * config/s390/s390.md (ANY_RETURN): New code iterator.
10469 (*creturn, *csimple_return, return, simple_return): New patterns.
10471 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10473 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
10474 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
10475 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
10476 REG_CFA_RESTORE list when deciding not to restore a register.
10478 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10480 * config/s390/s390.c: Include tree-pass.h and context.h.
10481 (s390_early_mach): New function, split out from...
10482 (s390_emit_prologue): ...here.
10483 (pass_data_s390_early_mach): New pass structure.
10484 (pass_s390_early_mach): New class.
10485 (s390_option_override): Create and register early_mach pass.
10486 Move to end of file.
10488 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10490 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
10491 to match for the exit block.
10493 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10495 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
10496 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
10497 Reject misaligned operands.
10499 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10501 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
10503 2014-02-07 Richard Biener <rguenther@suse.de>
10505 PR middle-end/60092
10506 * gimple-low.c (lower_builtin_posix_memalign): New function.
10507 (lower_stmt): Call it to lower posix_memalign in a way
10508 to make alignment info accessible.
10510 2014-02-07 Jakub Jelinek <jakub@redhat.com>
10513 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
10514 __builtin_setjmp_receiver.
10516 2014-02-07 Richard Biener <rguenther@suse.de>
10518 PR middle-end/60092
10519 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
10520 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
10521 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
10522 Handle BUILT_IN_POSIX_MEMALIGN.
10523 (find_func_clobbers): Likewise.
10524 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
10525 (call_may_clobber_ref_p_1): Likewise.
10527 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
10530 * ipa-devirt.c (record_target_from_binfo): Remove overactive
10533 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
10536 * lto-cgraph.c (lto_output_node): Use
10537 symtab_get_symbol_partitioning_class.
10538 (lto_output_varpool_node): likewise.
10539 (symtab_get_symbol_partitioning_class): Move here from
10540 lto/lto-partition.c
10541 * cgraph.h (symbol_partitioning_class): Likewise.
10542 (symtab_get_symbol_partitioning_class): Declare.
10544 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
10546 * ggc.h (ggc_internal_cleared_alloc): New macro.
10547 * vec.h (vec_safe_copy): Handle memory stats.
10548 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
10549 * target-globals.c (save_target_globals): Likewise.
10551 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
10554 * expr.c (emit_move_resolve_push): Export; be bit more selective
10555 on when to clear alias set.
10556 * expr.h (emit_move_resolve_push): Declare.
10557 * function.h (struct function): Add tail_call_marked.
10558 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
10559 * config/i386/i386-protos.h (ix86_expand_push): Remove.
10560 * config/i386/i386.md (TImode move expander): De not call
10562 (FP push expanders): Preserve memory attributes.
10563 * config/i386/sse.md (push<mode>1): Remove.
10564 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
10565 (ix86_expand_push): Remove.
10566 * config/i386/mmx.md (push<mode>1): Remove.
10568 2014-02-06 Jakub Jelinek <jakub@redhat.com>
10570 PR rtl-optimization/60030
10571 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
10572 lopart with paradoxical subreg before shifting it up by hprec.
10574 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10576 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
10577 Remove extra newline at end of file.
10578 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
10579 (arm_issue_rate): Handle cortexa57.
10580 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
10581 (cortex-a57.cortex-a53): Likewise.
10583 2014-02-06 Jakub Jelinek <jakub@redhat.com>
10586 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
10587 don't record in REG_FRAME_RELATED_EXPR registers not set in that
10589 (arm_expand_prologue): Adjust all callers.
10590 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
10591 info, registers also at the lowest numbered registers side. Use
10592 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
10596 * var-tracking.c (adjust_mems): Before adding a SET to
10597 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
10599 2014-02-06 Alan Modra <amodra@gmail.com>
10602 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
10603 change SDmode to DDmode when lra_in_progress.
10605 2014-02-06 Jakub Jelinek <jakub@redhat.com>
10607 PR middle-end/59150
10608 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
10609 free_data_ref on the dr first, and before goto again also set dr
10610 to the next dr. For simd_lane_access, free old datarefs[i] before
10611 overwriting it. For get_vectype_for_scalar_type failure, don't
10612 free_data_ref if simd_lane_access.
10614 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
10617 * tree.h (opts_for_fn): New inline function.
10618 (opt_for_fn): Define.
10619 * config/i386/i386.c (ix86_function_regparm): Use
10620 opt_for_fn (decl, optimize) instead of optimize.
10622 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
10624 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
10625 for SYMBOL_REF in large memory model.
10627 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10629 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
10630 and crypto support.
10631 (cortex-a57): Likewise.
10632 (cortex-a57.cortex-a53): Likewise.
10634 2014-02-06 Yury Gribov <y.gribov@samsung.com>
10635 Kugan Vivekanandarajah <kuganv@linaro.org>
10637 * config/arm/arm.c (arm_vector_alignment_reachable): Check
10639 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
10641 2014-02-06 Richard Biener <rguenther@suse.de>
10643 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
10644 set_loop_copy and initialize_original_copy_tables.
10646 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
10648 * config/aarch64/aarch64-simd.md
10649 (aarch64_ashr_simddi): Change QI to SI.
10651 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
10652 Jakub Jelinek <jakub@redhat.com>
10654 PR middle-end/60013
10655 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
10658 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10660 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
10661 CODE_FOR_altivec_vpku[hw]um to
10662 CODE_FOR_altivec_vpku[hw]um_direct.
10663 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
10664 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
10665 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
10666 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
10668 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10670 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
10671 generation for -maltivec=be.
10672 (altivec_vsumsws): Simplify redundant test.
10674 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10676 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
10677 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
10678 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
10679 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
10680 gen_altivec_vpkuwum.
10681 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
10683 (altivec_vpks<VI_char>ss): Likewise.
10684 (altivec_vpks<VI_char>us): Likewise.
10685 (altivec_vpku<VI_char>us): Likewise.
10686 (altivec_vpku<VI_char>um): Likewise.
10687 (altivec_vpku<VI_char>um_direct): New (copy of
10688 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
10690 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
10691 target is little endian and -maltivec=be is not specified.
10692 (*altivec_vupkhs<VU_char>_direct): New (copy of
10693 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
10694 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
10695 target is little endian and -maltivec=be is not specified.
10696 (*altivec_vupkls<VU_char>_direct): New (copy of
10697 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
10698 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
10699 little endian and -maltivec=be is not specified.
10700 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
10701 little endian and -maltivec=be is not specified.
10703 2014-02-05 Richard Henderson <rth@redhat.com>
10706 * combine-stack-adj.c: Revert r206943.
10707 * sched-int.h (struct deps_desc): Add last_args_size.
10708 * sched-deps.c (init_deps): Initialize it.
10709 (sched_analyze_insn): Add OUTPUT dependencies between insns that
10710 contain REG_ARGS_SIZE notes.
10712 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
10714 * lto-cgraph.c (asm_nodes_output): Make global.
10715 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
10716 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
10717 (driver_handle_option): Handle OPT_fwpa.
10719 2014-02-05 Jakub Jelinek <jakub@redhat.com>
10722 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
10723 a comment typo and formatting issue. If odr_hash hasn't been
10724 created, return vNULL and set *completep to false.
10726 PR middle-end/57499
10727 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
10728 bb with no successors.
10730 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
10733 * doc/invoke.texi (-march): Clarify documentation for ARM.
10734 (-mtune): Likewise.
10737 2014-02-05 Richard Biener <rguenther@suse.de>
10739 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
10740 when not vectorizing because of too many alias checks.
10741 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
10742 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
10744 2014-02-05 Nick Clifton <nickc@redhat.com>
10746 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
10747 accept extended registers in any mode when compiling for the MN10300.
10749 2014-02-05 Yury Gribov <y.gribov@samsung.com>
10751 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
10752 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
10753 sanitization attributes.
10754 (can_inline_edge_p): Likewise.
10755 (sanitize_attrs_match_for_inline_p): New function.
10757 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
10759 * ipa-prop.c (detect_type_change): Shor circuit testing of
10760 type changes on THIS pointer.
10762 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
10765 * config/pa/pa.c (legitimize_tls_address): Return original address
10766 if not passed a SYMBOL_REF rtx.
10767 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
10769 (pa_emit_move_sequence): Simplify TLS source operands.
10770 (pa_legitimate_constant_p): Reject all TLS constants.
10771 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
10772 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
10774 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
10776 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
10777 groups when we know they are controlled by LTO.
10778 * varasm.c (default_binds_local_p_1): If object is in other partition,
10779 it will be resolved locally.
10781 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
10783 * config/host-linux.c (linux_gt_pch_use_address): Don't
10784 use SSIZE_MAX because it is not always defined.
10786 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
10789 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
10790 threshold for pseudo splitting.
10791 (update_ebb_live_info): Process call argument hard registers and
10792 hard registers from insn definition too.
10793 (max_small_class_regs_num): New constant.
10794 (inherit_in_ebb): Update live hard regs through EBBs. Update
10795 reloads_num only for small register classes. Don't split for
10798 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
10801 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
10804 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
10806 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
10807 visibility is safe.
10809 2014-02-04 Marek Polacek <polacek@redhat.com>
10811 * gdbinit.in (pel): Define.
10813 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
10815 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
10818 2014-02-04 Richard Biener <rguenther@suse.de>
10821 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
10822 in function context local.
10823 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
10824 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
10825 similar to LTO_imported_decl_ref.
10827 2014-02-04 Jakub Jelinek <jakub@redhat.com>
10829 PR tree-optimization/60002
10830 * cgraphclones.c (build_function_decl_skip_args): Clear
10831 DECL_LANG_SPECIFIC.
10833 PR tree-optimization/60023
10834 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
10835 false to gsi_replace.
10836 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
10837 has been in some EH region and vec_stmt could throw, add
10838 vec_stmt into the same EH region.
10839 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
10840 has no lhs, ignore it.
10841 * internal-fn.c (expand_MASK_LOAD): Likewise.
10844 * tree-inline.c (copy_forbidden): Fail for
10845 __attribute__((optimize (0))) functions.
10848 * omp-low.c (simd_clone_struct_copy): If from->inbranch
10849 is set, copy one less argument.
10850 (expand_simd_clones): Don't subtract clone_info->inbranch
10851 from simd_clone_struct_alloc argument.
10853 PR rtl-optimization/57915
10854 * recog.c (simplify_while_replacing): If all unary/binary/relational
10855 operation arguments are constant, attempt to simplify those.
10857 PR middle-end/59261
10858 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
10859 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
10861 2014-02-04 Richard Biener <rguenther@suse.de>
10863 PR tree-optimization/60012
10864 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
10865 TBAA disambiguation to all DDRs.
10867 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
10870 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
10871 (LINK_SPEC): Use it for -shared, -shared-libgcc.
10873 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10876 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
10878 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10880 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
10881 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
10883 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10886 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
10887 to figure out targets of polymorphic calls with known decl.
10888 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
10889 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
10890 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
10891 (get_polymorphic_call_info): ... here.
10892 (get_polymorphic_call_info_from_invariant): New function.
10894 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10896 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
10897 lookup via vtable pointer; check for type consistency
10898 and turn inconsitent facts into UNREACHABLE.
10899 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
10900 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
10901 type inconsistent querries; return UNREACHABLE instead.
10903 2014-02-03 Richard Henderson <rth@twiddle.net>
10906 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
10907 already processed this node.
10908 (normalize_one_pred_1): Pass along mark_set.
10909 (normalize_one_pred): Create and destroy a pointer_set_t.
10910 (normalize_one_pred_chain): Likewise.
10912 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
10914 PR gcov-profile/58602
10915 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
10917 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10920 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
10921 -fno-devirtualize; try to devirtualize by the knowledge of
10922 virtual table pointer given by aggregate propagation.
10923 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
10924 (ipa_print_node_jump_functions): Dump also offset that
10925 is relevant for polymorphic calls.
10926 (determine_known_aggregate_parts): Add arg_type parameter; use it
10927 instead of determining the type from pointer type.
10928 (ipa_compute_jump_functions_for_edge): Update call of
10929 determine_known_aggregate_parts.
10930 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
10931 (gimple_get_virt_method_for_binfo): ... here; simplify using
10932 vtable_pointer_value_to_vtable.
10933 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
10934 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
10935 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
10936 (vtable_pointer_value_to_vtable): Break out from ...; handle also
10938 (vtable_pointer_value_to_binfo): ... here.
10939 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
10941 2014-02-03 Teresa Johnson <tejohnson@google.com>
10943 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
10944 redef of outer loop index variable.
10946 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
10950 * doc/extend.texi (Function Attributes): Typo.
10952 2014-02-03 Cong Hou <congh@google.com>
10954 PR tree-optimization/60000
10955 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
10956 if the vectorized statement is a store. A store statement can only
10957 appear at the end of pattern statements.
10959 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
10961 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
10962 (ix86_option_override_internal): Default long double to 64-bit for
10963 32-bit Bionic and to 128-bit for 64-bit Bionic.
10965 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
10966 TARGET_LONG_DOUBLE_128 is true.
10967 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
10969 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
10970 (mlong-double-64): Negate -mlong-double-128.
10971 (mlong-double-128): New option.
10973 * config/i386/i386-c.c (ix86_target_macros): Define
10974 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
10976 * doc/invoke.texi: Document -mlong-double-128.
10978 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
10980 PR rtl-optimization/60024
10981 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
10983 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
10985 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
10987 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
10989 PR rtl-optimization/57662
10990 * sel-sched.c (code_motion_path_driver): Do not mark already not
10991 existing blocks in the visiting bitmap.
10993 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
10995 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
10996 on the insn being emitted.
10998 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
10999 Will Deacon <will.deacon@arm.com>
11001 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
11003 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11005 * config/arm/arm-tables.opt: Regenerate.
11007 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11009 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
11010 for vector types other than V16QImode.
11011 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
11012 define_expand, and call altivec_expand_vec_perm_le when producing
11013 code with little endian element order.
11014 (*altivec_vperm_<mode>_internal): New insn having previous
11015 behavior of altivec_vperm_<mode>.
11016 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
11017 altivec_expand_vec_perm_le when producing code with little endian
11019 (*altivec_vperm_<mode>_uns_internal): New insn having previous
11020 behavior of altivec_vperm_<mode>_uns.
11022 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11024 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
11025 (altivec_vsumsws): Add handling for -maltivec=be with a little
11027 (altivec_vsumsws_direct): New.
11028 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
11029 gen_altivec_vsumsws.
11031 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
11033 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
11034 vtable_pointer_value_to_binfo): New functions.
11035 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
11036 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
11038 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
11040 * config/nios2/nios2.md (load_got_register): Initialize GOT
11041 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
11042 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
11044 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
11046 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
11047 preserverd by passthrough, do not propagate the type.
11049 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
11051 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
11052 (mips_atomic_assign_expand_fenv): New function.
11053 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
11055 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
11057 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
11058 (__builtin_mips_set_fcsr): Likewise.
11059 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
11060 MIPS_USI_FTYPE_VOID.
11061 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
11062 (mips16_expand_set_fcsr): Likewise.
11063 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
11064 (mips16_set_fcsr_stub): Likewise.
11065 (mips16_get_fcsr_one_only_stub): New class.
11066 (mips16_set_fcsr_one_only_stub): Likewise.
11067 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
11068 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
11069 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
11070 (hard_float): New availability predicate.
11071 (mips_builtins): Add get_fcsr and set_fcsr.
11072 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
11073 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
11074 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
11075 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
11076 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
11079 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
11081 * config/mips/mips.c (mips_one_only_stub): New class.
11082 (mips_need_mips16_rdhwr_p): Replace with...
11083 (mips16_rdhwr_stub): ...this new variable.
11084 (mips16_stub_call_address): New function.
11085 (mips16_rdhwr_one_only_stub): New class.
11086 (mips_expand_thread_pointer): Use mips16_stub_call_address.
11087 (mips_output_mips16_rdhwr): Delete.
11088 (mips_finish_stub): New function.
11089 (mips_code_end): Use it to handle rdhwr stubs.
11091 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
11094 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
11095 when calculating size of integer atomic types.
11097 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
11099 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
11101 2014-02-01 Jakub Jelinek <jakub@redhat.com>
11103 PR tree-optimization/60003
11104 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
11105 * profile.c (branch_prob): Use gimple_call_builtin_p
11106 to check for BUILT_IN_SETJMP_RECEIVER.
11107 * tree-inline.c (copy_bb): Call notice_special_calls.
11109 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
11112 * lra-constraints.c (process_alt_operands): Update reload_sum only
11115 2014-01-31 Richard Henderson <rth@redhat.com>
11117 PR middle-end/60004
11118 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
11119 until after else_eh is processed.
11121 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
11123 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
11124 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
11125 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
11126 in smmintrin.h, remove them.
11127 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
11128 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
11129 * config/i386/i386.md (ROUND_SAE): Fix value.
11130 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
11131 (const48_operand): New.
11132 * config/i386/subst.md (round), (round_expand): Use
11133 const_4_or_8_to_11_operand.
11134 (round_saeonly), (round_saeonly_expand): Use const48_operand.
11136 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
11138 * config/i386/constraints.md (Yk): Swap meaning with k.
11139 * config/i386/i386.md (movhi_internal): Change Yk to k.
11140 (movqi_internal): Ditto.
11141 (*k<logic><mode>): Ditto.
11144 (kandn<mode>): Ditto.
11145 (*<code>hi_1): Ditto.
11146 (*<code>qi_1): Ditto.
11147 (kxnor<mode>): Ditto.
11148 (kortestzhi): Ditto.
11149 (kortestchi): Ditto.
11151 (*one_cmplhi2_1): Ditto.
11152 (*one_cmplqi2_1): Ditto.
11153 * config/i386/sse.md (): Change k to Yk.
11154 (avx512f_load<mode>_mask): Ditto.
11155 (avx512f_blendm<mode>): Ditto.
11156 (avx512f_store<mode>_mask): Ditto.
11157 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
11158 (avx512f_storedqu<mode>_mask): Ditto.
11159 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
11161 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
11162 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
11163 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
11164 (avx512f_maskcmp<mode>3): Ditto.
11165 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
11166 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
11167 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
11168 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
11169 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
11170 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
11171 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
11172 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
11173 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
11174 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
11175 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
11176 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
11177 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
11178 (vec_extract_lo_<mode>_maskm): Ditto.
11179 (vec_extract_hi_<mode>_maskm): Ditto.
11180 (avx512f_vternlog<mode>_mask): Ditto.
11181 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
11182 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
11183 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
11184 (avx512f_<code>v8div16qi2_mask): Ditto.
11185 (avx512f_<code>v8div16qi2_mask_store): Ditto.
11186 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
11187 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
11188 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
11189 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
11190 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
11191 (*avx512pf_gatherpf<mode>df_mask): Ditto.
11192 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
11193 (*avx512pf_scatterpf<mode>df_mask): Ditto.
11194 (avx512cd_maskb_vec_dupv8di): Ditto.
11195 (avx512cd_maskw_vec_dupv16si): Ditto.
11196 (avx512f_vpermi2var<mode>3_maskz): Ditto.
11197 (avx512f_vpermi2var<mode>3_mask): Ditto.
11198 (avx512f_vpermi2var<mode>3_mask): Ditto.
11199 (avx512f_vpermt2var<mode>3_maskz): Ditto.
11200 (*avx512f_gathersi<mode>): Ditto.
11201 (*avx512f_gathersi<mode>_2): Ditto.
11202 (*avx512f_gatherdi<mode>): Ditto.
11203 (*avx512f_gatherdi<mode>_2): Ditto.
11204 (*avx512f_scattersi<mode>): Ditto.
11205 (*avx512f_scatterdi<mode>): Ditto.
11206 (avx512f_compress<mode>_mask): Ditto.
11207 (avx512f_compressstore<mode>_mask): Ditto.
11208 (avx512f_expand<mode>_mask): Ditto.
11209 * config/i386/subst.md (mask): Change k to Yk.
11210 (mask_scalar_merge): Ditto.
11213 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
11215 * doc/extend.texi (Vector Extensions): Document ?: in C++.
11217 2014-01-31 Richard Biener <rguenther@suse.de>
11219 PR middle-end/59990
11220 * builtins.c (fold_builtin_memory_op): Make sure to not
11221 use a floating-point mode or a boolean or enumeral type for
11222 the copy operation.
11224 2014-01-30 DJ Delorie <dj@redhat.com>
11226 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
11227 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
11228 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
11229 whenever main() has an epilogue.
11231 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11233 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
11234 unused variable "field".
11235 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
11236 (vsx_mergeh_<mode>): Likewise.
11237 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
11238 (altivec_vmrghh): Likewise.
11239 (altivec_vmrghw): Likewise.
11240 (altivec_vmrglb): Likewise.
11241 (altivec_vmrglh): Likewise.
11242 (altivec_vmrglw): Likewise.
11243 (altivec_vspltb): Add missing uses.
11244 (altivec_vsplth): Likewise.
11245 (altivec_vspltw): Likewise.
11246 (altivec_vspltsf): Likewise.
11248 2014-01-30 Jakub Jelinek <jakub@redhat.com>
11251 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
11252 frame related instructions.
11254 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
11256 PR rtl-optimization/59959
11257 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
11258 any reload of register whose subreg is invalid.
11260 2014-01-30 Jakub Jelinek <jakub@redhat.com>
11262 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
11263 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
11264 Add missing return type - void.
11266 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11268 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
11269 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
11270 remove element index adjustment for endian (now handled in vsx.md
11272 (altivec_expand_vec_perm_const): Use
11273 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
11274 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
11275 (vsx_xxspltw_<mode>): Adjust element index for little endian.
11276 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
11277 define_expand and a new define_insn *altivec_vspltb_internal;
11278 adjust for -maltivec=be on a little endian target.
11279 (altivec_vspltb_direct): New.
11280 (altivec_vsplth): Divide into a define_expand and a new
11281 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
11282 little endian target.
11283 (altivec_vsplth_direct): New.
11284 (altivec_vspltw): Divide into a define_expand and a new
11285 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
11286 little endian target.
11287 (altivec_vspltw_direct): New.
11288 (altivec_vspltsf): Divide into a define_expand and a new
11289 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
11290 a little endian target.
11292 2014-01-30 Richard Biener <rguenther@suse.de>
11294 PR tree-optimization/59993
11295 * tree-ssa-forwprop.c (associate_pointerplus): Check we
11296 can propagate form the earlier stmt and avoid the transform
11297 when the intermediate result is needed.
11299 2014-01-30 Alangi Derick <alangiderick@gmail.com>
11301 * README.Portability: Fix typo.
11303 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
11305 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
11306 comparison_operator with ordered_comparison_operator.
11308 2014-01-30 Nick Clifton <nickc@redhat.com>
11310 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
11311 Rename to mn10300_store_multiple_regs.
11312 * config/mn10300/mn10300.c: Likewise.
11313 * config/mn10300/mn10300.md (store_movm): Fix typo: call
11314 store_multiple_regs.
11315 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
11316 Call mn10300_store_multiple_regs.
11318 2014-01-30 Nick Clifton <nickc@redhat.com>
11319 DJ Delorie <dj@redhat.com>
11321 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
11322 %fp 2 to keep registers after it properly word-aligned.
11323 (rl78_alloc_physical_registers_umul): Handle the case where both
11324 input operands are the same.
11326 2014-01-30 Richard Biener <rguenther@suse.de>
11328 PR tree-optimization/59903
11329 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
11332 2014-01-30 Jason Merrill <jason@redhat.com>
11335 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
11338 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
11340 2014-01-30 Richard Biener <rguenther@suse.de>
11342 PR tree-optimization/59951
11343 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
11345 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
11348 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
11349 SFmode to DFmode case.
11351 2014-01-29 DJ Delorie <dj@redhat.com>
11353 * config/msp430/msp430.opt (-minrt): New.
11354 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
11356 (ENDFILE_SPEC): Likewise.
11358 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
11360 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
11361 (estimate_function_body_sizes): Use it.
11363 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
11366 * dwarf2out.c (is_cxx_auto): New.
11367 (is_base_type): Use it.
11368 (gen_type_die_with_usage): Likewise.
11370 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11372 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
11373 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
11374 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
11375 -maltivec=be with LE targets.
11376 (vsx_mergeh_<mode>): Likewise.
11377 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
11378 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
11379 (altivec_vmrghb): Replace with define_expand and new
11380 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
11381 (altivec_vmrghb_direct): New define_insn.
11382 (altivec_vmrghh): Replace with define_expand and new
11383 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
11384 (altivec_vmrghh_direct): New define_insn.
11385 (altivec_vmrghw): Replace with define_expand and new
11386 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
11387 (altivec_vmrghw_direct): New define_insn.
11388 (*altivec_vmrghsf): Adjust for endianness.
11389 (altivec_vmrglb): Replace with define_expand and new
11390 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
11391 (altivec_vmrglb_direct): New define_insn.
11392 (altivec_vmrglh): Replace with define_expand and new
11393 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
11394 (altivec_vmrglh_direct): New define_insn.
11395 (altivec_vmrglw): Replace with define_expand and new
11396 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
11397 (altivec_vmrglw_direct): New define_insn.
11398 (*altivec_vmrglsf): Adjust for endianness.
11399 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
11400 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
11401 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
11402 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
11403 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
11404 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
11405 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
11406 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
11408 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
11410 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
11411 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
11414 2014-01-29 Richard Biener <rguenther@suse.de>
11416 PR tree-optimization/58742
11417 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
11418 associate_pointerplus_align.
11419 (associate_pointerplus_diff): New function.
11420 (associate_pointerplus): Likewise. Call associate_pointerplus_align
11421 and associate_pointerplus_diff.
11423 2014-01-29 Richard Biener <rguenther@suse.de>
11425 * lto-streamer.h (LTO_major_version): Bump to 3.
11426 (LTO_minor_version): Reset to 0.
11428 2014-01-29 Renlin Li <Renlin.Li@arm.com>
11430 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
11431 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
11432 (arm_file_start): Generate correct asm header for armv7ve.
11433 * config/arm/bpabi.h: Add multilib support for armv7ve.
11434 * config/arm/driver-arm.c: Change the architectures of cortex-a7
11435 and cortex-a15 to armv7ve.
11436 * config/arm/t-aprofile: Add multilib support for armv7ve.
11437 * doc/invoke.texi: Document -march=armv7ve.
11439 2014-01-29 Richard Biener <rguenther@suse.de>
11441 PR tree-optimization/58742
11442 * tree-ssa-forwprop.c (associate_plusminus): Return true
11443 if we changed sth, defer EH cleanup to ...
11444 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
11445 (simplify_mult): New function.
11447 2014-01-29 Jakub Jelinek <jakub@redhat.com>
11449 PR middle-end/59917
11450 PR tree-optimization/59920
11451 * tree.c (build_common_builtin_nodes): Remove
11452 __builtin_setjmp_dispatcher initialization.
11453 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
11454 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
11455 instead of gsi_after_labels + manually skipping debug stmts.
11456 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
11457 ignore bbs with IFN_ABNORMAL_DISPATCHER.
11458 * tree-inline.c (copy_edges_for_bb): Remove
11459 can_make_abnormal_goto argument, instead add abnormal_goto_dest
11460 argument. Ignore computed_goto_p stmts. Don't call
11461 make_abnormal_goto_edges. If a call might need abnormal edges
11462 for non-local gotos, see if it already has an edge to
11463 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
11464 with true argument, don't do anything then, otherwise add
11465 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
11466 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
11468 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
11469 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
11470 (lower_stmt): Don't set data->calls_builtin_setjmp.
11471 (lower_builtin_setjmp): Adjust comment.
11472 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
11473 * tree-cfg.c (found_computed_goto): Remove.
11474 (factor_computed_gotos): Remove.
11475 (make_goto_expr_edges): Return bool, true for computed gotos.
11476 Don't call make_abnormal_goto_edges.
11477 (build_gimple_cfg): Don't set found_computed_goto, don't call
11478 factor_computed_gotos.
11479 (computed_goto_p): No longer static.
11480 (make_blocks): Don't set found_computed_goto.
11481 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
11482 (make_edges): If make_goto_expr_edges returns true, push bb
11483 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
11484 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
11485 vector. Record mapping between bbs and OpenMP regions if there
11486 are any, adjust make_gimple_omp_edges caller. Call
11487 handle_abnormal_edges.
11488 (make_abnormal_goto_edges): Remove.
11489 * tree-cfg.h (make_abnormal_goto_edges): Remove.
11490 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
11491 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
11492 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
11493 * internal-fn.def (ABNORMAL_DISPATCHER): New.
11494 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
11495 filling *region also set *region_idx to (*region)->entry->index.
11498 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
11499 For REGs set ORIGINAL_REGNO.
11501 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
11503 * doc/md.texi: Mention that a target shouldn't implement
11504 vec_widen_(s|u)mul_even/odd pair if it is less efficient
11507 2014-01-29 Jakub Jelinek <jakub@redhat.com>
11509 PR tree-optimization/59594
11510 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
11511 a copy of the datarefs vector rather than the vector itself.
11513 2014-01-28 Jason Merrill <jason@redhat.com>
11516 * dwarf2out.c (auto_die): New static.
11517 (gen_type_die_with_usage): Handle C++1y 'auto'.
11518 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
11521 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
11524 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
11525 (SPEC_X32): Likewise.
11526 (SPEC_64): Likewise.
11527 * config/i386/i386.c (ix86_option_override_internal): Turn off
11528 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
11530 (x86_file_start): Output .code16gcc for TARGET_16BIT.
11531 * config/i386/i386.h (TARGET_16BIT): New macro.
11532 (TARGET_16BIT_P): Likewise.
11533 * config/i386/i386.opt: Add m16.
11534 * doc/invoke.texi: Document -m16.
11536 2014-01-28 Jakub Jelinek <jakub@redhat.com>
11538 PR preprocessor/59935
11539 * input.c (location_get_source_line): Bail out on when line number
11540 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
11542 2014-01-28 Richard Biener <rguenther@suse.de>
11544 PR tree-optimization/58742
11545 * tree-ssa-forwprop.c (associate_plusminus): Handle
11546 pointer subtraction of the form (T)(P + A) - (T)P.
11548 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11550 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
11553 2014-01-28 Richard Biener <rguenther@suse.de>
11556 2014-01-28 Richard Biener <rguenther@suse.de>
11558 PR rtl-optimization/45364
11559 PR rtl-optimization/59890
11560 * var-tracking.c (local_get_addr_clear_given_value): Handle
11561 already cleared slot.
11562 (val_reset): Handle not allocated local_get_addr_cache.
11563 (vt_find_locations): Use post-order on the inverted CFG.
11565 2014-01-28 Richard Biener <rguenther@suse.de>
11567 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
11569 2014-01-28 Richard Biener <rguenther@suse.de>
11571 PR rtl-optimization/45364
11572 PR rtl-optimization/59890
11573 * var-tracking.c (local_get_addr_clear_given_value): Handle
11574 already cleared slot.
11575 (val_reset): Handle not allocated local_get_addr_cache.
11576 (vt_find_locations): Use post-order on the inverted CFG.
11578 2014-01-28 Alan Modra <amodra@gmail.com>
11580 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
11581 * configure.ac <recursive call for build != host>: Define
11582 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
11583 and LD_FOR_BUILD too.
11584 * configure: Regenerate.
11586 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
11588 * config/i386/i386.c (get_builtin_code_for_version): Separate
11589 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
11590 Broadwell from Haswell.
11592 2014-01-27 Steve Ellcey <sellcey@mips.com>
11594 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
11595 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
11596 * config/mips/mips.c (mips_option_override): Change setting
11598 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
11599 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
11600 Change from Mask to Var.
11602 2014-01-27 Jeff Law <law@redhat.com>
11604 * ipa-inline.c (inline_small_functions): Fix typo.
11606 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
11608 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
11609 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
11610 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
11611 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
11612 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
11613 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
11614 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
11615 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
11616 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
11617 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
11618 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
11619 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
11620 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
11621 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
11622 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
11623 (_mm512_storeu_epi64): Ditto.
11624 (_mm512_cmpge_epi32_mask): Ditto.
11625 (_mm512_cmpge_epu32_mask): Ditto.
11626 (_mm512_cmpge_epi64_mask): Ditto.
11627 (_mm512_cmpge_epu64_mask): Ditto.
11628 (_mm512_cmple_epi32_mask): Ditto.
11629 (_mm512_cmple_epu32_mask): Ditto.
11630 (_mm512_cmple_epi64_mask): Ditto.
11631 (_mm512_cmple_epu64_mask): Ditto.
11632 (_mm512_cmplt_epi32_mask): Ditto.
11633 (_mm512_cmplt_epu32_mask): Ditto.
11634 (_mm512_cmplt_epi64_mask): Ditto.
11635 (_mm512_cmplt_epu64_mask): Ditto.
11636 (_mm512_cmpneq_epi32_mask): Ditto.
11637 (_mm512_cmpneq_epu32_mask): Ditto.
11638 (_mm512_cmpneq_epi64_mask): Ditto.
11639 (_mm512_cmpneq_epu64_mask): Ditto.
11640 (_mm512_expand_pd): Ditto.
11641 (_mm512_expand_ps): Ditto.
11642 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
11643 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
11644 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
11645 * config/i386/i386.c (ix86_builtins): Add
11646 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
11647 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
11648 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
11649 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
11650 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
11651 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
11652 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
11653 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
11654 IX86_BUILTIN_PMOVUSQW512_MEM.
11655 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
11656 __builtin_ia32_pmovsqd512mem_mask,
11657 __builtin_ia32_pmovqd512mem_mask,
11658 __builtin_ia32_pmovusqw512mem_mask,
11659 __builtin_ia32_pmovsqw512mem_mask,
11660 __builtin_ia32_pmovqw512mem_mask,
11661 __builtin_ia32_pmovusdw512mem_mask,
11662 __builtin_ia32_pmovsdw512mem_mask,
11663 __builtin_ia32_pmovdw512mem_mask,
11664 __builtin_ia32_pmovqb512mem_mask,
11665 __builtin_ia32_pmovusqb512mem_mask,
11666 __builtin_ia32_pmovsqb512mem_mask,
11667 __builtin_ia32_pmovusdb512mem_mask,
11668 __builtin_ia32_pmovsdb512mem_mask,
11669 __builtin_ia32_pmovdb512mem_mask.
11670 (bdesc_args): Add __builtin_ia32_expanddf512,
11671 __builtin_ia32_expandsf512.
11672 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
11673 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
11674 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
11675 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
11676 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
11677 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
11678 (avx512f_<code>v8div16qi2_mask_store): This.
11679 (avx512f_expand<mode>): New.
11681 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
11683 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
11685 (_mm512_mask_prefetch_i64gather_pd): Ditto.
11686 (_mm512_prefetch_i32scatter_pd): Ditto.
11687 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
11688 (_mm512_prefetch_i64scatter_pd): Ditto.
11689 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
11690 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
11691 (_mm512_mask_prefetch_i64gather_ps): Ditto.
11692 (_mm512_prefetch_i32scatter_ps): Ditto.
11693 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
11694 (_mm512_prefetch_i64scatter_ps): Ditto.
11695 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
11696 * config/i386/i386-builtin-types.def: Define
11697 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
11698 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
11699 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
11700 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
11701 IX86_BUILTIN_SCATTERPFQPD.
11702 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
11703 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
11704 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
11705 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
11706 __builtin_ia32_scatterpfqps.
11707 (ix86_expand_builtin): Expand new built-ins.
11708 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
11709 fix memory access data type.
11710 (*avx512pf_gatherpf<mode>_mask): Ditto.
11711 (*avx512pf_gatherpf<mode>): Ditto.
11712 (avx512pf_scatterpf<mode>): Ditto.
11713 (*avx512pf_scatterpf<mode>_mask): Ditto.
11714 (*avx512pf_scatterpf<mode>): Ditto.
11715 (GATHER_SCATTER_SF_MEM_MODE): New.
11716 (avx512pf_gatherpf<mode>df): Ditto.
11717 (*avx512pf_gatherpf<mode>df_mask): Ditto.
11718 (*avx512pf_scatterpf<mode>df): Ditto.
11720 2014-01-27 Jakub Jelinek <jakub@redhat.com>
11723 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
11724 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
11727 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
11729 * common/config/arm/arm-common.c
11730 (arm_rewrite_mcpu): Handle multiple names.
11732 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
11734 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
11736 * gimple-builder.h (create_gimple_tmp): Delete.
11738 2014-01-27 Christian Bruel <christian.bruel@st.com>
11740 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
11743 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
11745 * config/pa/pa.md (call): Generate indirect long calls to non-local
11746 functions when outputing 32-bit code.
11747 (call_value): Likewise except for special call to buggy powf function.
11749 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
11750 portable runtime and PIC indirect calls.
11751 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
11752 and PIC call sequences. Use ldo instead of blr to set return register
11753 in PIC call sequence.
11755 2014-01-25 Walter Lee <walt@tilera.com>
11757 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
11758 avoid clobbering a live register.
11760 2014-01-25 Walter Lee <walt@tilera.com>
11762 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
11763 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
11764 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
11765 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
11767 2014-01-25 Walter Lee <walt@tilera.com>
11769 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
11770 arguments on even registers.
11771 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
11773 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
11774 (BIGGEST_ALIGNMENT): Ditto.
11775 (BIGGEST_FIELD_ALIGNMENT): Ditto.
11777 2014-01-25 Walter Lee <walt@tilera.com>
11779 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
11780 insns before bundling.
11781 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
11783 2014-01-25 Walter Lee <walt@tilera.com>
11785 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
11786 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
11787 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
11789 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
11791 * config/mips/constraints.md (kl): Delete.
11792 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
11793 define expands, using...
11794 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
11795 instructions for MIPS16.
11796 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
11797 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
11799 2014-01-25 Walter Lee <walt@tilera.com>
11801 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
11805 2014-01-25 Walter Lee <walt@tilera.com>
11807 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
11808 (TARGET_EXPAND_TO_RTL_HOOK): Define.
11810 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
11812 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
11815 2014-01-25 Jakub Jelinek <jakub@redhat.com>
11817 * print-rtl.c (in_call_function_usage): New var.
11818 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
11819 EXPR_LIST mode as mode and not as reg note name.
11821 PR middle-end/59561
11822 * cfgloopmanip.c (copy_loop_info): If
11823 loop->warned_aggressive_loop_optimizations, make sure
11824 the flag is set in target loop too.
11826 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
11828 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
11830 * builtins.def: Likewise.
11831 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
11832 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
11833 * ira.c (ira_setup_eliminable_regset): Likewise.
11834 * omp-low.c (gate_expand_omp): Likewise.
11835 (execute_lower_omp): Likewise.
11836 (diagnose_sb_0): Likewise.
11837 (gate_diagnose_omp_blocks): Likewise.
11838 (simd_clone_clauses_extract): Likewise.
11841 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11843 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
11844 correction for little endian...
11845 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
11848 2014-01-24 Jeff Law <law@redhat.com>
11850 PR tree-optimization/59919
11851 * tree-vrp.c (find_assert_locations_1): Do not register asserts
11852 for non-returning calls.
11854 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
11856 * common/config/aarch64/aarch64-common.c
11857 (aarch64_rewrite_mcpu): Handle multiple names.
11858 * config/aarch64/aarch64.h
11859 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
11861 2014-01-24 Dodji Seketeli <dodji@redhat.com>
11863 * input.c (add_file_to_cache_tab): Handle the case where fopen
11866 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
11869 * config/i386/i386.md (pushsf splitter): Get stack adjustment
11870 from push operand if code of push isn't PRE_DEC.
11872 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
11875 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
11876 -mquad-memory-atomic. Update -mquad-memory documentation to say
11877 it is only used for non-atomic loads/stores.
11879 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
11880 -mquad-memory or -mquad-memory-atomic switches.
11882 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
11883 -mquad-memory-atomic to ISA 2.07 support.
11885 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
11886 to separate support of normal quad word memory operations (ldq, stq)
11887 from the atomic quad word memory operations.
11889 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
11890 support to separate non-atomic quad word operations from atomic
11891 quad word operations. Disable non-atomic quad word operations in
11892 little endian mode so that we don't have to swap words after the
11893 load and before the store.
11894 (quad_load_store_p): Add comment about atomic quad word support.
11895 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
11896 options printed with -mdebug=reg.
11898 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
11899 -mquad-memory-atomic as the test for whether we have quad word
11900 atomic instructions.
11901 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
11902 or -mp8-vector are used, allow byte/half-word atomic operations.
11904 * config/rs6000/sync.md (load_lockedti): Insure that the address
11905 is a proper indexed or indirect address for the lqarx instruction.
11906 On little endian systems, swap the hi/lo registers after the lqarx
11908 (load_lockedpti): Use indexed_or_indirect_operand predicate to
11909 insure the address is valid for the lqarx instruction.
11910 (store_conditionalti): Insure that the address is a proper indexed
11911 or indirect address for the stqcrx. instruction. On little endian
11912 systems, swap the hi/lo registers before doing the stqcrx.
11914 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
11915 insure the address is valid for the stqcrx. instruction.
11917 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
11918 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
11919 type of quad memory support is available.
11921 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
11923 PR regression/59915
11924 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
11925 there is a danger of looping.
11927 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
11929 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
11930 force flag_ira_loop_pressure if set via command line.
11932 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
11934 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
11935 (ashr_simd): New builtin handling DI mode.
11936 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
11937 (aarch64_sshr_simddi): New match pattern.
11938 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
11939 (vshrd_n_s64): Likewise.
11940 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
11942 2014-01-23 Nick Clifton <nickc@redhat.com>
11944 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
11945 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
11946 favour of mcu specific scripts.
11947 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
11950 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
11951 Alex Velenko <Alex.Velenko@arm.com>
11953 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
11954 (vaddv_s16): Likewise.
11955 (vaddv_s32): Likewise.
11956 (vaddv_u8): Likewise.
11957 (vaddv_u16): Likewise.
11958 (vaddv_u32): Likewise.
11959 (vaddvq_s8): Likewise.
11960 (vaddvq_s16): Likewise.
11961 (vaddvq_s32): Likewise.
11962 (vaddvq_s64): Likewise.
11963 (vaddvq_u8): Likewise.
11964 (vaddvq_u16): Likewise.
11965 (vaddvq_u32): Likewise.
11966 (vaddvq_u64): Likewise.
11967 (vaddv_f32): Likewise.
11968 (vaddvq_f32): Likewise.
11969 (vaddvq_f64): Likewise.
11970 (vmaxv_f32): Likewise.
11971 (vmaxv_s8): Likewise.
11972 (vmaxv_s16): Likewise.
11973 (vmaxv_s32): Likewise.
11974 (vmaxv_u8): Likewise.
11975 (vmaxv_u16): Likewise.
11976 (vmaxv_u32): Likewise.
11977 (vmaxvq_f32): Likewise.
11978 (vmaxvq_f64): Likewise.
11979 (vmaxvq_s8): Likewise.
11980 (vmaxvq_s16): Likewise.
11981 (vmaxvq_s32): Likewise.
11982 (vmaxvq_u8): Likewise.
11983 (vmaxvq_u16): Likewise.
11984 (vmaxvq_u32): Likewise.
11985 (vmaxnmv_f32): Likewise.
11986 (vmaxnmvq_f32): Likewise.
11987 (vmaxnmvq_f64): Likewise.
11988 (vminv_f32): Likewise.
11989 (vminv_s8): Likewise.
11990 (vminv_s16): Likewise.
11991 (vminv_s32): Likewise.
11992 (vminv_u8): Likewise.
11993 (vminv_u16): Likewise.
11994 (vminv_u32): Likewise.
11995 (vminvq_f32): Likewise.
11996 (vminvq_f64): Likewise.
11997 (vminvq_s8): Likewise.
11998 (vminvq_s16): Likewise.
11999 (vminvq_s32): Likewise.
12000 (vminvq_u8): Likewise.
12001 (vminvq_u16): Likewise.
12002 (vminvq_u32): Likewise.
12003 (vminnmv_f32): Likewise.
12004 (vminnmvq_f32): Likewise.
12005 (vminnmvq_f64): Likewise.
12007 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
12009 * config/aarch64/aarch64-simd.md
12010 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
12011 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
12012 (*aarch64_mul3_elt<mode>): Likewise.
12013 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
12014 (*aarch64_mul3_elt_to_64v2df): Likewise.
12015 (*aarch64_mla_elt<mode>): Likewise.
12016 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
12017 (*aarch64_mls_elt<mode>): Likewise.
12018 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
12019 (*aarch64_fma4_elt<mode>): Likewise.
12020 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
12021 (*aarch64_fma4_elt_to_64v2df): Likewise.
12022 (*aarch64_fnma4_elt<mode>): Likewise.
12023 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
12024 (*aarch64_fnma4_elt_to_64v2df): Likewise.
12025 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
12026 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
12027 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
12028 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
12029 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
12030 (aarch64_sqdmull_lane<mode>_internal): Likewise.
12031 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
12033 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
12035 * config/aarch64/aarch64-simd.md
12036 (aarch64_be_checked_get_lane<mode>): New define_expand.
12037 * config/aarch64/aarch64-simd-builtins.def
12038 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
12039 New builtin definition.
12040 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
12041 Use new safe be builtin.
12043 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
12045 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
12047 (aarch64_be_st1<mode>): Likewise.
12048 (aarch_ld1<VALL:mode>): Define_expand modified.
12049 (aarch_st1<VALL:mode>): Likewise.
12050 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
12051 (UNSPEC_ST1): Likewise.
12053 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
12055 * config/microblaze/microblaze.md: Add trap insn and attribute
12057 2014-01-23 Dodji Seketeli <dodji@redhat.com>
12059 PR preprocessor/58580
12060 * input.h (location_get_source_line): Take an additional line_size
12062 (void diagnostics_file_cache_fini): Declare new function.
12063 * input.c (struct fcache): New type.
12064 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
12065 New static constants.
12066 (diagnostic_file_cache_init, total_lines_num)
12067 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
12068 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
12069 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
12070 (get_next_line, read_next_line, goto_next_line, read_line_num):
12071 New static function definitions.
12072 (diagnostic_file_cache_fini): New function.
12073 (location_get_source_line): Take an additional output line_len
12074 parameter. Re-write using lookup_or_add_file_to_cache_tab and
12076 * diagnostic.c (diagnostic_finish): Call
12077 diagnostic_file_cache_fini.
12078 (adjust_line): Take an additional input parameter for the length
12079 of the line, rather than calculating it with strlen.
12080 (diagnostic_show_locus): Adjust the use of
12081 location_get_source_line and adjust_line with respect to their new
12082 signature. While displaying a line now, do not stop at the first
12083 null byte. Rather, display the zero byte as a space and keep
12084 going until we reach the size of the line.
12085 * Makefile.in: Add vec.o to OBJS-libcommon
12087 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
12088 Ilya Tocar <ilya.tocar@intel.com>
12090 * config/i386/avx512fintrin.h (_mm512_kmov): New.
12091 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
12092 (__builtin_ia32_kmov16): Ditto.
12093 * config/i386/i386.md (UNSPEC_KMOV): New.
12096 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
12098 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
12099 (_mm512_storeu_si512): Ditto.
12101 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
12104 * rtl.h (get_referenced_operands): Declare.
12105 * recog.c (get_referenced_operands): New function.
12106 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
12107 operands have been referenced when recording LO_SUM references.
12109 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
12111 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
12113 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
12115 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
12116 Enable for generic and recent AMD targets.
12118 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
12120 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
12121 ARG_SIZE note when adjustment was eliminated.
12123 2014-01-22 Jeff Law <law@redhat.com>
12125 PR tree-optimization/59597
12126 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
12127 in file. Accept new argument REGISTERING and use it to modify
12128 dump output appropriately.
12129 (register_jump_thread): Corresponding changes.
12130 (mark_threaded_blocks): Reinstate code to cancel unprofitable
12131 thread paths involving joiner blocks. Add code to dump cancelled
12132 jump threading paths.
12134 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
12136 PR rtl-optimization/59477
12137 * lra-constraints.c (inherit_in_ebb): Process call for living hard
12138 regs. Update reloads_num and potential_reload_hard_regs for all insns.
12140 2014-01-22 Tom Tromey <tromey@redhat.com>
12142 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
12144 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
12146 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
12148 PR rtl-optimization/59896
12149 * lra-constraints.c (process_alt_operands): Check unused note for
12150 matched operands of insn with no output reloads.
12152 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
12154 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
12155 (mips_move_from_gpr_cost): Likewise.
12157 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
12159 PR rtl-optimization/59858
12160 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
12161 ira_class_hard_regs_num.
12162 (process_alt_operands): Increase reject for dying matched operand.
12164 2014-01-21 Jakub Jelinek <jakub@redhat.com>
12167 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
12168 smaller than size, perform several stores or loads and stores
12169 at dst + count - size to store or copy all of size bytes, rather
12170 than just last modesize bytes.
12172 2014-01-20 DJ Delorie <dj@redhat.com>
12174 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
12175 that CLOBBERs are REGs before propogating their values.
12177 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
12179 PR middle-end/59789
12180 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
12181 (cgraph_inline_failed_type): New function.
12182 * cgraph.h (DEFCIFCODE): Add type.
12183 (cgraph_inline_failed_type_t): New enum.
12184 (cgraph_inline_failed_type): New prototype.
12185 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
12186 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
12187 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
12188 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
12189 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
12190 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
12191 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
12192 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
12193 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
12194 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
12195 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
12196 OPTIMIZATION_MISMATCH.
12197 * tree-inline.c (expand_call_inline): Emit errors during
12198 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
12200 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
12203 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
12204 mode attribute in insn output.
12206 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
12208 * output.h (output_constant): Delete.
12209 * varasm.c (output_constant): Make private.
12211 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
12213 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
12215 2014-01-20 Jakub Jelinek <jakub@redhat.com>
12217 PR middle-end/59860
12218 * tree.h (fold_builtin_strcat): New prototype.
12219 * builtins.c (fold_builtin_strcat): No longer static. Add len
12220 argument, if non-NULL, don't call c_strlen. Optimize
12221 directly into __builtin_memcpy instead of __builtin_strcpy.
12222 (fold_builtin_2): Adjust fold_builtin_strcat caller.
12223 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
12225 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
12227 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
12228 for SImode_address_operand operands, having only a REG argument.
12230 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
12232 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
12233 loader name using mbig-endian.
12234 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
12236 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
12238 * doc/invoke.texi (-march): Clarify documentation for AArch64.
12239 (-mtune): Likewise.
12242 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
12244 * config/aarch64/aarch64-protos.h
12245 (aarch64_cannot_change_mode_class_ptr): Declare.
12246 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
12247 aarch64_cannot_change_mode_class_ptr): New.
12248 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
12249 backend hook aarch64_cannot_change_mode_class.
12251 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
12253 * common/config/aarch64/aarch64-common.c
12254 (aarch64_handle_option): Don't handle any option order logic here.
12255 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
12256 selected_cpu, warn on architecture version mismatch.
12257 (aarch64_override_options): Fix parsing order for option strings.
12259 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12260 Iain Sandoe <iain@codesourcery.com>
12263 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
12264 warning. Amend comment to reflect current functionality.
12266 2014-01-20 Richard Biener <rguenther@suse.de>
12268 PR middle-end/59860
12269 * builtins.c (fold_builtin_strcat): Remove case better handled
12270 by tree-ssa-strlen.c.
12272 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
12274 * config/aarch64/aarch64.opt
12275 (mcpu, march, mtune): Make case-insensitive.
12277 2014-01-20 Jakub Jelinek <jakub@redhat.com>
12280 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
12281 if operands[1] is a REG or ZERO_EXTEND of a REG.
12283 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
12285 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
12287 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
12289 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
12290 long non-pic millicode calls.
12292 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12294 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
12296 2014-01-19 Kito Cheng <kito@0xlab.org>
12298 * builtins.c (expand_movstr): Check movstr expand done or fail.
12300 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
12301 H.J. Lu <hongjiu.lu@intel.com>
12304 * config/i386/i386.md (*lea<mode>): Zero-extend return register
12305 to DImode for zero-extended addresses.
12307 2014-01-19 Jakub Jelinek <jakub@redhat.com>
12309 PR rtl-optimization/57763
12310 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
12311 on the new indirect jump_insn and increment LABEL_NUSES (label).
12313 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
12317 * config.gcc (x86_archs): New variable.
12318 (x86_64_archs): Likewise.
12319 (x86_cpus): Likewise.
12320 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
12321 --with-arch/--with-cpu= options.
12322 Support --with-arch=/--with-cpu={nehalem,westmere,
12323 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
12325 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
12327 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
12328 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
12330 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
12332 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
12334 2014-01-18 Jakub Jelinek <jakub@redhat.com>
12337 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
12338 clear cpp_get_options (parse_in)->warn_unused_macros for
12339 ix86_target_macros_internal with cpp_define.
12341 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
12343 * jump.c (delete_related_insns): Keep (use (insn))s.
12344 * reorg.c (redundant_insn): Check for barriers too.
12346 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
12348 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
12350 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
12352 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
12353 call to $$dyncall when TARGET_LONG_CALLS is true.
12355 2014-01-17 Jeff Law <law@redhat.com>
12357 * ree.c (combine_set_extension): Temporarily disable test for
12358 changing number of hard registers.
12360 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
12362 PR middle-end/58125
12363 * ipa-inline-analysis.c (inline_free_summary):
12364 Do not free summary of aliases.
12366 2014-01-17 Jakub Jelinek <jakub@redhat.com>
12368 PR middle-end/59706
12369 * gimplify.c (gimplify_expr): Use create_tmp_var
12370 instead of create_tmp_var_raw. If cond doesn't have
12371 integral type, don't add the IFN_ANNOTATE builtin at all.
12373 2014-01-17 Martin Jambor <mjambor@suse.cz>
12376 * ipa-cp.c (prev_edge_clone): New variable.
12377 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
12378 Also resize prev_edge_clone vector.
12379 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
12380 (ipcp_edge_removal_hook): New function.
12381 (ipcp_driver): Register ipcp_edge_removal_hook.
12383 2014-01-17 Andrew Pinski <apinski@cavium.com>
12384 Steve Ellcey <sellcey@mips.com>
12387 * config/mips/mips.c (mips_print_operand): Check operand mode instead
12390 2014-01-17 Jeff Law <law@redhat.com>
12392 PR middle-end/57904
12393 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
12394 so that pass_ccp runs first.
12396 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
12398 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
12399 (ix86_adjust_cost): Use !TARGET_XXX.
12400 (do_reorder_for_imul): Likewise.
12401 (swap_top_of_ready_list): Likewise.
12402 (ix86_sched_reorder): Likewise.
12404 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
12406 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
12407 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
12408 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
12409 (intel_memset): New. Duplicate slm_memset.
12410 (intel_cost): New. Duplicate slm_cost.
12411 (m_INTEL): New macro.
12412 (processor_target_table): Add "intel".
12413 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
12414 with PROCESSOR_INTEL for "intel".
12415 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
12416 PROCESSOR_SILVERMONT.
12417 (ix86_issue_rate): Likewise.
12418 (ix86_adjust_cost): Likewise.
12419 (ia32_multipass_dfa_lookahead): Likewise.
12420 (swap_top_of_ready_list): Likewise.
12421 (ix86_sched_reorder): Likewise.
12422 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
12423 instead of TARGET_OPT_AGU.
12424 * config/i386/i386.h (TARGET_INTEL): New.
12425 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
12426 (processor_type): Add PROCESSOR_INTEL.
12427 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
12428 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
12430 2014-01-17 Marek Polacek <polacek@redhat.com>
12433 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
12436 2014-01-17 Richard Biener <rguenther@suse.de>
12438 PR tree-optimization/46590
12439 * opts.c (default_options_table): Add entries for
12440 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
12441 all enabled at -O1 but not for -Og.
12442 * common.opt (fbranch-count-reg): Remove Init(1).
12443 (fmove-loop-invariants): Likewise.
12444 (ftree-pta): Likewise.
12446 2014-01-17 Jakub Jelinek <jakub@redhat.com>
12448 * config/i386/i386.c (ix86_data_alignment): For compatibility with
12449 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
12450 decls to at least the GCC 4.8 used alignments.
12453 * tree-nested.c (convert_nonlocal_reference_stmt,
12454 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
12455 of GIMPLE_BIND stmts, adjust associated decls.
12457 2014-01-17 Richard Biener <rguenther@suse.de>
12459 PR tree-optimization/46590
12460 * vec.h (vec<>::bseach): New member function implementing
12461 binary search according to C89 bsearch.
12462 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
12463 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
12464 bitmap pointer again. Make accesses_in_loop a flat array.
12465 (mem_ref_obstack): New global.
12466 (outermost_indep_loop): Adjust for mem_ref->stored changes.
12467 (mark_ref_stored): Likewise.
12468 (ref_indep_loop_p_2): Likewise.
12469 (set_ref_stored_in_loop): New helper function.
12470 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
12471 (memref_free): Adjust.
12472 (record_mem_ref_loc): Simplify.
12473 (gather_mem_refs_stmt): Adjust.
12474 (sort_locs_in_loop_postorder_cmp): New function.
12475 (analyze_memory_references): Sort accesses_in_loop after
12476 loop postorder number.
12477 (find_ref_loc_in_loop_cmp): New function.
12478 (for_all_locs_in_loop): Find relevant cluster of locs in
12479 accesses_in_loop and iterate without recursion.
12480 (execute_sm): Avoid uninit warning.
12481 (struct ref_always_accessed): Simplify.
12482 (ref_always_accessed::operator ()): Likewise.
12483 (ref_always_accessed_p): Likewise.
12484 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
12485 loop postorder numbers here.
12486 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
12489 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
12492 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
12493 on decls for which assemble_alias has been called.
12495 2014-01-17 Nick Clifton <nickc@redhat.com>
12497 * config/msp430/msp430.opt: (mcpu): New option.
12498 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
12499 (msp430_option_override): Parse target_cpu. If the MCU name
12500 matches a generic string, clear target_mcu.
12501 (msp430_attr): Allow numeric interrupt values up to 63.
12502 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
12503 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
12505 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
12507 * config/msp430/msp430.md (popm): Use %J rather than %I.
12508 (addsi3): Use msp430_nonimmediate_operand for operand 2.
12509 (addhi_cy_i): Use immediate_operand for operand 2.
12510 * doc/invoke.texi: Document -mcpu option.
12512 2014-01-17 Richard Biener <rguenther@suse.de>
12514 PR rtl-optimization/38518
12515 * df.h (df_analyze_loop): Declare.
12516 * df-core.c: Include cfgloop.h.
12517 (df_analyze_1): Split out main part of df_analyze.
12518 (df_analyze): Adjust.
12519 (loop_inverted_post_order_compute): New function.
12520 (loop_post_order_compute): Likewise.
12521 (df_analyze_loop): New function avoiding whole-function
12522 postorder computes.
12523 * loop-invariant.c (find_defs): Use df_analyze_loop.
12524 (find_invariants): Adjust.
12525 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
12527 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
12529 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
12530 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
12532 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
12534 * ipa-ref.c (ipa_remove_stmt_references): Fix references
12535 traversal when removing references.
12537 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
12540 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
12542 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
12544 PR middle-end/56791
12545 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
12546 pushing a reload for an autoinc when we had previously reloaded an
12547 inner part of the address.
12549 2014-01-16 Jakub Jelinek <jakub@redhat.com>
12551 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
12553 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
12554 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
12555 when not giving up or versioning for alias only because of
12557 (vect_analyze_data_ref_dependences): Set to true.
12558 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
12560 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
12561 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
12564 PR middle-end/58344
12565 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
12568 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
12569 operand 0 predicate for gathers, use a new pseudo as subtarget.
12571 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
12573 PR middle-end/59609
12574 * lra-constraints.c (process_alt_operands): Add printing debug info.
12575 Check absence of input/output reloads for matched operands too.
12577 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
12579 PR rtl-optimization/59835
12580 * ira.c (ira_init_register_move_cost): Increase cost for
12583 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
12585 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
12587 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
12590 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
12591 non-register objects. Use gen_(high/low)part more consistently.
12594 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
12597 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
12598 endian support, remove tests for WORDS_BIG_ENDIAN.
12599 (p8_mfvsrd_3_<mode>): Likewise.
12600 (reload_gpr_from_vsx<mode>): Likewise.
12601 (reload_gpr_from_vsxsf): Likewise.
12602 (p8_mfvsrd_4_disf): Likewise.
12604 2014-01-16 Richard Biener <rguenther@suse.de>
12606 PR rtl-optimization/46590
12607 * lcm.c (compute_antinout_edge): Use postorder iteration.
12608 (compute_laterin): Use inverted postorder iteration.
12610 2014-01-16 Nick Clifton <nickc@redhat.com>
12612 PR middle-end/28865
12613 * varasm.c (output_constant): Return the number of bytes actually
12615 (output_constructor_array_range): Update the field size with the
12616 number of bytes emitted by output_constant.
12617 (output_constructor_regular_field): Likewise. Also do not
12618 complain if the total number of bytes emitted is now greater
12619 than the expected fieldpos.
12620 * output.h (output_constant): Update prototype and descriptive comment.
12622 2014-01-16 Marek Polacek <polacek@redhat.com>
12624 PR middle-end/59827
12625 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
12626 it is error_mark_node.
12628 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
12630 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
12631 VALID_AVX256_REG_OR_OI_MODE.
12633 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
12635 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
12636 current procedure should be profiled.
12638 2014-01-15 Andrew Pinski <apinski@cavium.com>
12640 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
12641 of moving from/to the STACK_REG register class.
12643 2014-01-15 Richard Henderson <rth@redhat.com>
12646 * reginfo.c (global_regs_decl): Globalize.
12647 * rtl.h (global_regs_decl): Declare.
12648 * ira.c (do_reload): Diagnose frame_pointer_needed and it
12649 reserved via global_regs.
12651 2014-01-15 Teresa Johnson <tejohnson@google.com>
12653 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
12655 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12657 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
12658 and vmulosh rather than call gen_vec_widen_smult_*.
12659 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
12660 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
12661 (vec_widen_smult_even_v16qi): Likewise.
12662 (vec_widen_umult_even_v8hi): Likewise.
12663 (vec_widen_smult_even_v8hi): Likewise.
12664 (vec_widen_umult_odd_v16qi): Likewise.
12665 (vec_widen_smult_odd_v16qi): Likewise.
12666 (vec_widen_umult_odd_v8hi): Likewise.
12667 (vec_widen_smult_odd_v8hi): Likewise.
12668 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
12669 vmuloub rather than call gen_vec_widen_umult_*.
12670 (vec_widen_umult_lo_v16qi): Likewise.
12671 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
12672 vmulosb rather than call gen_vec_widen_smult_*.
12673 (vec_widen_smult_lo_v16qi): Likewise.
12674 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
12675 rather than call gen_vec_widen_umult_*.
12676 (vec_widen_umult_lo_v8hi): Likewise.
12677 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
12678 rather than call gen_vec_widen_smult_*.
12679 (vec_widen_smult_lo_v8hi): Likewise.
12681 2014-01-15 Jeff Law <law@redhat.com>
12683 PR tree-optimization/59747
12684 * ree.c (find_and_remove_re): Properly handle case where a second
12685 eliminated extension requires widening a copy created for elimination
12686 of a prior extension.
12687 (combine_set_extension): Ensure that the number of hard regs needed
12688 for a destination register does not change when we widen it.
12690 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
12692 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
12693 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
12694 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
12695 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
12696 (avr-*-rtems*): Likewise.
12697 (bfin*-rtems*): Likewise.
12698 (moxie-*-rtems*): Likewise.
12699 (h8300-*-rtems*): Likewise.
12700 (i[34567]86-*-rtems*): Likewise.
12701 (lm32-*-rtems*): Likewise.
12702 (m32r-*-rtems*): Likewise.
12703 (m68k-*-rtems*): Likewise.
12704 (microblaze*-*-rtems*): Likewise.
12705 (mips*-*-rtems*): Likewise.
12706 (powerpc-*-rtems*): Likewise.
12707 (sh-*-rtems*): Likewise.
12708 (sparc-*-rtems*): Likewise.
12709 (sparc64-*-rtems*): Likewise.
12710 (v850-*-rtems*): Likewise.
12711 (m32c-*-rtems*): Likewise.
12713 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
12715 PR rtl-optimization/59511
12716 * ira.c (ira_init_register_move_cost): Use memory costs for some
12717 cases of register move cost calculations.
12718 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
12719 instead of BB frequency.
12720 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
12721 * lra-assigns.c (find_hard_regno_for): Ditto.
12723 2014-01-15 Richard Biener <rguenther@suse.de>
12725 PR tree-optimization/59822
12726 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
12727 (vectorizable_load): Use it to hoist defs of uses of invariant
12728 loads out of the loop.
12730 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
12731 Kugan Vivekanandarajah <kuganv@linaro.org>
12734 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
12737 2014-01-15 Richard Biener <rguenther@suse.de>
12739 PR rtl-optimization/59802
12740 * lcm.c (compute_available): Use inverted postorder to seed
12741 the initial worklist.
12743 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12746 * config/s390/s390.c (s390_preferred_reload_class): Don't return
12747 ADDR_REGS for invalid symrefs in non-PIC code.
12749 2014-01-15 Jakub Jelinek <jakub@redhat.com>
12752 * builtins.c (determine_block_size): Initialize *probable_max_size
12753 even if len_rtx is CONST_INT.
12755 2014-01-14 Andrew Pinski <apinski@cavium.com>
12757 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
12758 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
12759 (cortexa53_tunings): Likewise.
12760 (aarch64_sched_issue_rate): New function.
12761 (TARGET_SCHED_ISSUE_RATE): Define.
12763 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
12765 * ira-costs.c (find_costs_and_classes): Add missed
12766 ira_init_register_move_cost_if_necessary.
12768 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
12771 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
12773 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
12776 * config/i386/i386.c (type_natural_mode): Add a bool parameter
12777 to indicate if type is used for function return value. Warn ABI
12778 change if the vector mode isn't available for function return value.
12779 (ix86_function_arg_advance): Pass false to type_natural_mode.
12780 (ix86_function_arg): Likewise.
12781 (ix86_gimplify_va_arg): Likewise.
12782 (function_arg_32): Don't warn ABI change.
12783 (ix86_function_value): Pass true to type_natural_mode.
12784 (ix86_return_in_memory): Likewise.
12785 (ix86_struct_value_rtx): Removed.
12786 (TARGET_STRUCT_VALUE_RTX): Likewise.
12788 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
12790 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
12791 converting a conditional jump into a conditional return.
12793 2014-01-14 Richard Biener <rguenther@suse.de>
12795 PR tree-optimization/58921
12796 PR tree-optimization/59006
12797 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
12798 hoisting invariant stmts.
12799 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
12800 invariant loads on the preheader edge if possible.
12802 2014-01-14 Joey Ye <joey.ye@arm.com>
12804 * doc/plugin.texi (Building GCC plugins): Update to C++.
12806 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
12808 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
12809 (_mm_rcp28_round_ss): Ditto.
12810 (_mm_rsqrt28_round_sd): Ditto.
12811 (_mm_rsqrt28_round_ss): Ditto.
12812 (_mm_rcp28_sd): Ditto.
12813 (_mm_rcp28_ss): Ditto.
12814 (_mm_rsqrt28_sd): Ditto.
12815 (_mm_rsqrt28_ss): Ditto.
12816 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
12817 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
12818 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
12819 (IX86_BUILTIN_RCP28SD): Ditto.
12820 (IX86_BUILTIN_RCP28SS): Ditto.
12821 (IX86_BUILTIN_RSQRT28SD): Ditto.
12822 (IX86_BUILTIN_RSQRT28SS): Ditto.
12823 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
12824 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
12825 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
12826 (ix86_expand_special_args_builtin): Expand new FTYPE.
12827 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
12828 (srcp14<mode>): Make insn unary.
12829 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
12830 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
12831 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
12832 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
12833 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
12834 Fix rounding: make it SAE only.
12835 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
12837 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
12839 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
12840 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
12841 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
12842 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
12843 (round_saeonly_mask_scalar_operand4): Ditto.
12844 (round_saeonly_mask_scalar_op3): Ditto.
12845 (round_saeonly_mask_scalar_op4): Ditto.
12847 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12849 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
12850 Implement -maltivec=be for vec_insert and vec_extract.
12852 2014-01-10 DJ Delorie <dj@redhat.com>
12854 * config/msp430/msp430.md (call_internal): Don't allow memory
12855 references with SP as the base register.
12856 (call_value_internal): Likewise.
12857 * config/msp430/constraints.md (Yc): New. For memory references
12858 that don't use SP as a base register.
12860 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
12861 "an integer without a # prefix"
12862 * config/msp430/msp430.md (epilogue_helper): Use it.
12864 2014-01-13 Jakub Jelinek <jakub@redhat.com>
12867 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
12868 AVX512F gather builtins.
12869 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
12870 on gather decls with INTEGER_TYPE masktype.
12871 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
12872 directly into the builtin rather than hoisting it before loop.
12874 PR tree-optimization/59387
12875 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
12876 (scev_const_prop): If folded_casts and type has undefined overflow,
12877 use force_gimple_operand instead of force_gimple_operand_gsi and
12878 for each added stmt if it is assign with
12879 arith_code_with_undefined_signed_overflow, call
12880 rewrite_to_defined_overflow.
12881 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
12882 gimple-fold.h instead.
12883 (arith_code_with_undefined_signed_overflow,
12884 rewrite_to_defined_overflow): Moved to ...
12885 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
12886 rewrite_to_defined_overflow): ... here. No longer static.
12887 Include gimplify-me.h.
12888 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
12889 rewrite_to_defined_overflow): New prototypes.
12891 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12893 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
12895 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
12897 * builtins.c (get_object_alignment_2): Minor tweak.
12898 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
12900 2014-01-13 Christian Bruel <christian.bruel@st.com>
12902 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
12903 optimized non constant lengths.
12905 2014-01-13 Jakub Jelinek <jakub@redhat.com>
12908 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
12909 load as __atomic_load_N if possible.
12911 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
12913 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
12915 (rs6000_expand_builtin): Adjust call.
12917 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
12920 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
12921 * config/rs6000/rs6000.c: Include target-globals.h.
12922 (rs6000_set_current_function): Instead of doing target_reinit
12923 unconditionally, use save_target_globals_default_opts and
12924 restore_target_globals.
12926 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
12928 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
12929 (rs6000_expand_builtin): Handle mffs and mtfsf.
12930 (rs6000_init_builtins): Define mffs and mtfsf.
12931 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
12932 (rs6000_mffs): New pattern.
12933 (rs6000_mtfsf): New pattern.
12935 2014-01-11 Bin Cheng <bin.cheng@arm.com>
12937 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
12938 Start narrowing with START. Apply candidate-use pair
12939 and check overall cost in narrowing.
12940 (iv_ca_prune): Pass new argument.
12942 2014-01-10 Jeff Law <law@redhat.com>
12944 PR middle-end/59743
12945 * ree.c (combine_reaching_defs): Ensure the defining statement
12946 occurs before the extension when optimizing extensions with
12947 different source and destination hard registers.
12949 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
12952 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
12953 vtables into the type inheritance graph.
12955 2014-01-10 Jakub Jelinek <jakub@redhat.com>
12957 PR rtl-optimization/59754
12958 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
12959 modes in the REGNO != REGNO case.
12961 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12963 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
12965 2014-01-10 Jakub Jelinek <jakub@redhat.com>
12967 PR tree-optimization/59745
12968 * tree-predcom.c (tree_predictive_commoning_loop): Call
12969 free_affine_expand_cache if giving up because components is NULL.
12971 * target-globals.c (save_target_globals): Allocate < 4KB structs using
12972 GC in payload of target_globals struct instead of allocating them on
12973 the heap and the larger structs separately using GC.
12974 * target-globals.h (struct target_globals): Make regs, hard_regs,
12975 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
12976 of GTY((skip)) and change type to void *.
12977 (reset_target_globals): Cast loads from those fields to corresponding
12980 2014-01-10 Steve Ellcey <sellcey@mips.com>
12983 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
12984 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
12985 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
12987 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
12990 * aarch64-modes.def (CC_Zmode): New flags mode.
12991 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
12992 represents an equality.
12993 (aarch64_get_condition_code): Handle CC_Zmode.
12994 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
12996 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12998 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
12999 extraction in good case.
13001 2014-01-10 Richard Biener <rguenther@suse.de>
13003 PR tree-optimization/59374
13004 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
13005 checking after SLP discovery. Mark stmts not participating
13006 in any SLP instance properly.
13008 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13010 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
13011 when handling a SET rtx.
13013 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13015 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
13016 (cortex-a57): Likewise.
13017 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
13019 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13021 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
13022 non-iwmmxt builtins.
13024 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
13028 * ipa-devirt.c record_target_from_binfo): Take as argument
13029 stack of binfos and lookup matching one for virtual inheritance.
13030 (possible_polymorphic_call_targets_1): Update.
13032 2014-01-10 Huacai Chen <chenhc@lemote.com>
13034 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
13035 kernel strings for Loongson-2E/2F/3A.
13037 2014-01-10 Jakub Jelinek <jakub@redhat.com>
13039 PR middle-end/59670
13040 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
13041 is_gimple_call before calling gimple_call_internal_p.
13043 2014-01-09 Steve Ellcey <sellcey@mips.com>
13045 * Makefile.in (TREE_FLOW_H): Remove.
13046 (TREE_SSA_H): Add file names from tree-flow.h.
13047 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
13048 * tree.h: Remove tree-flow.h reference.
13049 * hash-table.h: Remove tree-flow.h reference.
13050 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
13051 reference with tree-ssa-loop.h.
13053 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13055 * doc/invoke.texi: Add -maltivec={be,le} options, and document
13056 default element-order behavior for -maltivec.
13057 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
13058 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
13059 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
13060 when targeting big endian, at least for now.
13061 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
13063 2014-01-09 Jakub Jelinek <jakub@redhat.com>
13065 PR middle-end/47735
13066 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
13067 var satisfies use_register_for_decl, just take into account type
13068 alignment, rather than decl alignment.
13070 PR tree-optimization/59622
13071 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
13072 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
13073 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
13074 Don't devirtualize for inplace at all. For targets.length () == 1,
13075 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
13077 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
13079 * config/i386/i386.md (cpu): Remove the unused btver1.
13081 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
13083 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
13085 2014-01-09 Jakub Jelinek <jakub@redhat.com>
13088 * tree-core.h (struct target_globals): New forward declaration.
13089 (struct tree_target_option): Add globals field.
13090 * tree.h (TREE_TARGET_GLOBALS): Define.
13091 (prepare_target_option_nodes_for_pch): New prototype.
13092 * target-globals.h (struct target_globals): Define even if
13093 !SWITCHABLE_TARGET.
13094 * tree.c (prepare_target_option_node_for_pch,
13095 prepare_target_option_nodes_for_pch): New functions.
13096 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
13097 * config/i386/i386.c: Include target-globals.h.
13098 (ix86_set_current_function): Instead of doing target_reinit
13099 unconditionally, use save_target_globals_default_opts and
13100 restore_target_globals.
13102 2014-01-09 Richard Biener <rguenther@suse.de>
13104 PR tree-optimization/59715
13105 * tree-cfg.h (split_critical_edges): Declare.
13106 * tree-cfg.c (split_critical_edges): Export.
13107 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
13109 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
13111 * cfgexpand.c (expand_stack_vars): Optionally disable
13112 asan stack protection.
13113 (expand_used_vars): Likewise.
13114 (partition_stack_vars): Likewise.
13115 * asan.c (asan_emit_stack_protection): Optionally disable
13116 after return stack usage.
13117 (instrument_derefs): Optionally disable memory access instrumentation.
13118 (instrument_builtin_call): Likewise.
13119 (instrument_strlen_call): Likewise.
13120 (asan_protect_global): Optionally disable global variables protection.
13121 * doc/invoke.texi: Added doc for new options.
13122 * params.def: Added new options.
13123 * params.h: Likewise.
13125 2014-01-09 Jakub Jelinek <jakub@redhat.com>
13127 PR rtl-optimization/59724
13128 * ifcvt.c (cond_exec_process_if_block): Don't call
13129 flow_find_head_matching_sequence with 0 longest_match.
13130 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
13131 non-active insns if !stop_after.
13132 (try_head_merge_bb): Revert 2014-01-07 changes.
13134 2014-01-08 Jeff Law <law@redhat.com>
13136 * ree.c (get_sub_rtx): New function, extracted from...
13137 (merge_def_and_ext): Here.
13138 (combine_reaching_defs): Use get_sub_rtx.
13140 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
13142 * cgraph.h (varpool_variable_node): Do not choke on null node.
13144 2014-01-08 Catherine Moore <clm@codesourcery.com>
13146 * config/mips/mips.md (simple_return): Attempt to use JRC
13148 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
13150 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
13152 PR rtl-optimization/59137
13153 * reorg.c (steal_delay_list_from_target): Call update_block for
13155 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
13157 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13159 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
13160 two duplicate entries.
13162 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
13165 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
13167 * config/mips/mips.c (mips_truncated_op_cost): New function.
13168 (mips_rtx_costs): Adjust test for BADDU.
13169 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
13171 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
13173 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
13174 (*baddu_si): ...this new pattern.
13176 2014-01-08 Jakub Jelinek <jakub@redhat.com>
13179 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
13181 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
13183 PR middle-end/57748
13184 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
13186 (expand_expr, expand_normal): Adjust.
13187 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
13188 inner_reference_p. Use inner_reference_p to expand inner references.
13189 (store_expr): Adjust.
13190 * cfgexpand.c (expand_call_stmt): Adjust.
13192 2014-01-08 Rong Xu <xur@google.com>
13194 * gcov-io.c (gcov_var): Move from gcov-io.h.
13195 (gcov_position): Ditto.
13196 (gcov_is_error): Ditto.
13197 (gcov_rewrite): Ditto.
13198 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
13199 only part to libgcc/libgcov.h.
13201 2014-01-08 Marek Polacek <polacek@redhat.com>
13203 PR middle-end/59669
13204 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
13206 2014-01-08 Marek Polacek <polacek@redhat.com>
13209 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
13211 2014-01-08 Jakub Jelinek <jakub@redhat.com>
13213 PR rtl-optimization/59649
13214 * stor-layout.c (get_mode_bounds): For BImode return
13215 0 and STORE_FLAG_VALUE.
13217 2014-01-08 Richard Biener <rguenther@suse.de>
13219 PR middle-end/59630
13220 * gimple.h (is_gimple_builtin_call): Remove.
13221 (gimple_builtin_call_types_compatible_p): New.
13222 (gimple_call_builtin_p): New overload.
13223 * gimple.c (is_gimple_builtin_call): Remove.
13224 (validate_call): Rename to ...
13225 (gimple_builtin_call_types_compatible_p): ... this and export. Also
13226 check return types.
13227 (validate_type): New static function.
13228 (gimple_call_builtin_p): New overload and adjust.
13229 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
13230 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
13231 (gimple_fold_stmt_to_constant_1): Likewise.
13232 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
13234 2014-01-08 Richard Biener <rguenther@suse.de>
13236 PR middle-end/59471
13237 * gimplify.c (gimplify_expr): Gimplify register-register type
13238 VIEW_CONVERT_EXPRs to separate stmts.
13240 2014-01-07 Jeff Law <law@redhat.com>
13242 PR middle-end/53623
13243 * ree.c (combine_set_extension): Handle case where source
13244 and destination registers in an extension insn are different.
13245 (combine_reaching_defs): Allow source and destination registers
13246 in extension to be different under limited circumstances.
13247 (add_removable_extension): Remove restriction that the
13248 source and destination registers in the extension are the same.
13249 (find_and_remove_re): Emit a copy from the extension's
13250 destination to its source after the defining insn if
13251 the source and destination registers are different.
13253 PR middle-end/59285
13254 * ifcvt.c (merge_if_block): If we are merging a block with more than
13255 one successor with a block with no successors, remove any BARRIER
13256 after the second block.
13258 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
13260 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
13262 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
13265 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
13266 for 14-bit register offsets when INT14_OK_STRICT is false.
13268 2014-01-07 Roland Stigge <stigge@antcom.de>
13269 Michael Meissner <meissner@linux.vnet.ibm.com>
13272 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
13273 Only check TFmode for SPE constants. Don't check TImode or TDmode.
13275 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
13277 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
13280 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
13282 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
13283 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
13284 rtx is const0_rtx or not.
13286 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
13289 * target-globals.c (save_target_globals): Remove this_fn_optab
13291 * toplev.c: Include optabs.h.
13292 (target_reinit): Temporarily restore the global options if another
13293 set of options are in force.
13295 2014-01-07 Jakub Jelinek <jakub@redhat.com>
13297 PR rtl-optimization/58668
13298 * cfgcleanup.c (flow_find_cross_jump): Don't count
13299 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
13300 to determine what is counted.
13301 (flow_find_head_matching_sequence): Use active_insn_p to determine
13303 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
13305 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
13306 determine what is counted.
13308 PR tree-optimization/59643
13309 * tree-predcom.c (split_data_refs_to_components): If one dr is
13310 read and one write, determine_offset fails and the write isn't
13311 in the bad component, just put the read into the bad component.
13313 2014-01-07 Mike Stump <mikestump@comcast.net>
13314 Jakub Jelinek <jakub@redhat.com>
13317 * tree-core.h (struct tree_optimization_option): Change optabs
13318 type from unsigned char * to void *.
13319 * optabs.c (init_tree_optimization_optabs): Adjust
13320 TREE_OPTIMIZATION_OPTABS initialization.
13322 2014-01-06 Jakub Jelinek <jakub@redhat.com>
13325 * config/i386/i386.h (struct machine_function): Add
13326 no_drap_save_restore field.
13327 * config/i386/i386.c (ix86_save_reg): Use
13328 !cfun->machine->no_drap_save_restore instead of
13329 crtl->stack_realign_needed.
13330 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
13331 this function clears frame_pointer_needed. Set
13332 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
13333 and DRAP reg is needed.
13335 2014-01-06 Marek Polacek <polacek@redhat.com>
13338 * doc/implement-c.texi: Mention that other integer types are
13339 permitted as bit-field types in strictly conforming mode.
13341 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
13343 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
13344 is newly allocated.
13346 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
13348 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
13350 2014-01-06 Martin Jambor <mjambor@suse.cz>
13353 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
13355 * ipa-prop.c (ipa_print_node_params): Fix indentation.
13357 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
13361 * var-tracking.c (add_stores): Preserve the value of the source even if
13362 we don't record the store.
13364 2014-01-06 Terry Guo <terry.guo@arm.com>
13366 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
13368 2014-01-05 Iain Sandoe <iain@codesourcery.com>
13371 * config/darwin.c (darwin_function_section): Adjust return values to
13372 correspond to optimisation changes made in r206070.
13374 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
13376 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
13377 from prefetch_block tune setting.
13378 (nocona_cost): Correct size of prefetch block to 64.
13380 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
13382 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
13383 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
13384 used to save the static chain register in the computation of the offset
13385 from which the FP registers need to be restored.
13387 2014-01-04 Jakub Jelinek <jakub@redhat.com>
13389 PR tree-optimization/59519
13390 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
13391 ICE if get_current_def (current_new_name) is already non-NULL, as long
13392 as it is a phi result of some other phi in *new_exit_bb that has
13395 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
13396 or vmovdqu* for misaligned_operand.
13397 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
13398 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
13399 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
13400 aligned_mem for AVX512F masked aligned load and store builtins and for
13401 non-temporal moves.
13403 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
13405 PR tree-optimization/59651
13406 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
13407 Address range for negative step should be added by TYPE_SIZE_UNIT.
13409 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
13411 * config/m68k/m68k.c (handle_move_double): Handle pushes with
13412 overlapping registers also for registers other than the stack pointer.
13414 2014-01-03 Marek Polacek <polacek@redhat.com>
13417 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
13420 2014-01-03 Jakub Jelinek <jakub@redhat.com>
13423 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
13426 * config/i386/i386.md (MODE_SIZE): New mode attribute.
13427 (push splitter): Use <P:MODE_SIZE> instead of
13428 GET_MODE_SIZE (<P:MODE>mode).
13429 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
13430 (mov -1, reg peephole2): Likewise.
13431 * config/i386/sse.md (*mov<mode>_internal,
13432 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
13433 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
13434 *<code><mode>3, *andnot<mode>3<mask_name>,
13435 <mask_codefor><code><mode>3<mask_name>): Likewise.
13436 * config/i386/subst.md (mask_mode512bit_condition,
13437 sd_mask_mode512bit_condition): Likewise.
13439 2014-01-02 Xinliang David Li <davidxl@google.com>
13441 PR tree-optimization/59303
13442 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
13443 (dump_predicates): Better output format.
13444 (pred_equal_p): New function.
13445 (is_neq_relop_p): Ditto.
13446 (is_neq_zero_form_p): Ditto.
13447 (pred_expr_equal_p): Ditto.
13448 (pred_neg_p): Ditto.
13449 (simplify_pred): Ditto.
13450 (simplify_preds_2): Ditto.
13451 (simplify_preds_3): Ditto.
13452 (simplify_preds_4): Ditto.
13453 (simplify_preds): Ditto.
13454 (push_pred): Ditto.
13455 (push_to_worklist): Ditto.
13456 (get_pred_info_from_cmp): Ditto.
13457 (is_degenerated_phi): Ditto.
13458 (normalize_one_pred_1): Ditto.
13459 (normalize_one_pred): Ditto.
13460 (normalize_one_pred_chain): Ditto.
13461 (normalize_preds): Ditto.
13462 (normalize_cond_1): Remove function.
13463 (normalize_cond): Ditto.
13464 (is_gcond_subset_of): Ditto.
13465 (is_subset_of_any): Ditto.
13466 (is_or_set_subset_of): Ditto.
13467 (is_and_set_subset_of): Ditto.
13468 (is_norm_cond_subset_of): Ditto.
13469 (pred_chain_length_cmp): Ditto.
13470 (convert_control_dep_chain_into_preds): Type change.
13471 (find_predicates): Ditto.
13472 (find_def_preds): Ditto.
13473 (destroy_predicates_vecs): Ditto.
13474 (find_matching_predicates_in_rest_chains): Ditto.
13475 (use_pred_not_overlap_with_undef_path_pred): Ditto.
13476 (is_pred_expr_subset): Ditto.
13477 (is_pred_chain_subset_of): Ditto.
13478 (is_included_in): Ditto.
13479 (is_superset_of): Ditto.
13481 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
13483 Update copyright years.
13485 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
13487 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
13488 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
13489 config/arc/arc.md, config/arc/arc.opt,
13490 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
13491 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
13492 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
13493 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
13494 config/linux-protos.h, config/linux.c, config/winnt-c.c,
13495 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
13496 vtable-verify.c, vtable-verify.h: Use the standard form for the
13499 2014-01-02 Tobias Burnus <burnus@net-b.de>
13501 * gcc.c (process_command): Update copyright notice dates.
13502 * gcov-dump.c: Ditto.
13504 * doc/cpp.texi: Bump @copying's copyright year.
13505 * doc/cppinternals.texi: Ditto.
13506 * doc/gcc.texi: Ditto.
13507 * doc/gccint.texi: Ditto.
13508 * doc/gcov.texi: Ditto.
13509 * doc/install.texi: Ditto.
13510 * doc/invoke.texi: Ditto.
13512 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
13514 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
13516 2014-01-01 Jakub Jelinek <jakub@redhat.com>
13518 * config/i386/sse.md (*mov<mode>_internal): Guard
13519 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
13521 PR rtl-optimization/59647
13522 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
13523 new_rtx into UNSIGNED_FLOAT rtxes.
13525 Copyright (C) 2014 Free Software Foundation, Inc.
13527 Copying and distribution of this file, with or without modification,
13528 are permitted in any medium without royalty provided the copyright
13529 notice and this notice are preserved.