| blob | ece0da1eee6aea15886206de5b50ddf7b870cdb2 |
1 2014-06-03 Andrew Pinski <apinski@cavium.com>
3 * config/aarch64/aarch64.c (aarch64_if_then_else_costs): Allow non comparisons
4 for OP0.
6 2014-06-03 Andrew Pinski <apinski@cavium.com>
8 * config/aarch64/aarch64.c (aarch64_if_then_else_costs): New function.
9 (aarch64_rtx_costs): Use aarch64_if_then_else_costs.
11 2014-06-03 Kai Tietz <ktietz@redhat.com>
13 * config/i386/i386.c (ix86_function_value_regno_p): Disallow DX_REG
14 for 64-bit ms-abi.
16 2014-06-03 Dehao Chen <dehao@google.com>
18 * tree-cfg.c (gimple_merge_blocks): Only reset count when BBs are in
19 the same loop.
21 2014-06-03 Marek Polacek <polacek@redhat.com>
23 PR c/60439
24 * doc/invoke.texi: Document -Wswitch-bool.
25 * function.c (stack_protect_epilogue): Cast controlling expression of
26 the switch to int.
27 * gengtype.c (walk_type): Generate switch expression with its
28 controlling expression cast to int.
30 2014-06-03 Vishnu K S <Vishnu.k_s@atmel.com>
32 * config/avr/avr-mcus.def: Add new avr25 devices attiny441, attiny828
33 and attiny841.
34 * config/avr/avr-tables.opt: Regenerate.
35 * config/avr/t-multilib: Regenerate.
36 * doc/avr-mmcu.texi: Regenerate.
38 2014-06-03 Vishnu K S <vishnu.k_s@atmel.com>
39 Pitchumani Sivanupandi <pitchumani.s@atmel.com>
41 * config/avr/avr-mcus.def (ata6616c): Add new avr25 device.
42 (ata6617c, ata664251): Add new avr35 devices.
43 (ata6612c): Add new avr4 device.
44 (ata6613c, ata6614q): Add new avr5 devices.
45 * config/avr/avr-tables.opt: Regenerate.
46 * config/avr/t-multilib: Regenerate.
47 * doc/avr-mmcu.texi: Regenerate.
49 2014-06-03 Alan Lawrence <alan.lawrence@arm.com>
51 * gcc/config/aarch64/aarch64-builtins.c
52 (aarch64_types_binop_ssu_qualifiers): New static data.
53 (TYPES_BINOP_SSU): Define.
54 * gcc/config/aarch64/aarch64-simd-builtins.def (suqadd, ushl, urshl,
55 urshr_n, ushll_n): Use appropriate unsigned qualifiers.
56 * gcc/config/aarch64/arm_neon.h (vrshl_u8, vrshl_u16, vrshl_u32,
57 vrshl_u64, vrshlq_u8, vrshlq_u16, vrshlq_u32, vrshlq_u64, vrshld_u64,
58 vrshr_n_u8, vrshr_n_u16, vrshr_n_u32, vrshr_n_u64, vrshrq_n_u8,
59 vrshrq_n_u16, vrshrq_n_u32, vrshrq_n_u64, vrshrd_n_u64, vshll_n_u8,
60 vshll_n_u16, vshll_n_u32, vuqadd_s8, vuqadd_s16, vuqadd_s32,
61 vuqadd_s64, vuqaddq_s8, vuqaddq_s16, vuqaddq_s32, vuqaddq_s64,
62 vuqaddb_s8, vuqaddh_s16, vuqadds_s32, vuqaddd_s64): Add signedness
63 suffix to builtin function name, remove cast.
64 (vshl_s8, vshl_s16, vshl_s32, vshl_s64, vshl_u8, vshl_u16, vshl_u32,
65 vshl_u64, vshlq_s8, vshlq_s16, vshlq_s32, vshlq_s64, vshlq_u8,
66 vshlq_u16, vshlq_u32, vshlq_u64, vshld_s64, vshld_u64): Remove cast.
68 2014-06-03 Alan Lawrence <alan.lawrence@arm.com>
70 * gcc/config/aarch64/aarch64-builtins.c
71 (aarch64_types_binop_uus_qualifiers,
72 aarch64_types_shift_to_unsigned_qualifiers,
73 aarch64_types_unsigned_shiftacc_qualifiers): Define.
74 * gcc/config/aarch64/aarch64-simd-builtins.def (uqshl, uqrshl, uqadd,
75 uqsub, usqadd, usra_n, ursra_n, uqshrn_n, uqrshrn_n, usri_n, usli_n,
76 sqshlu_n, uqshl_n): Update qualifiers.
77 * gcc/config/aarch64/arm_neon.h (vqadd_u8, vqadd_u16, vqadd_u32,
78 vqadd_u64, vqaddq_u8, vqaddq_u16, vqaddq_u32, vqaddq_u64, vqsub_u8,
79 vqsub_u16, vqsub_u32, vqsub_u64, vqsubq_u8, vqsubq_u16, vqsubq_u32,
80 vqsubq_u64, vqaddb_u8, vqaddh_u16, vqadds_u32, vqaddd_u64, vqrshl_u8,
81 vqrshl_u16, vqrshl_u32, vqrshl_u64, vqrshlq_u8, vqrshlq_u16,
82 vqrshlq_u32, vqrshlq_u64, vqrshlb_u8, vqrshlh_u16, vqrshls_u32,
83 vqrshld_u64, vqrshrn_n_u16, vqrshrn_n_u32, vqrshrn_n_u64,
84 vqrshrnh_n_u16, vqrshrns_n_u32, vqrshrnd_n_u64, vqshl_u8, vqshl_u16,
85 vqshl_u32, vqshl_u64, vqshlq_u8, vqshlq_u16, vqshlq_u32, vqshlq_u64,
86 vqshlb_u8, vqshlh_u16, vqshls_u32, vqshld_u64, vqshl_n_u8, vqshl_n_u16,
87 vqshl_n_u32, vqshl_n_u64, vqshlq_n_u8, vqshlq_n_u16, vqshlq_n_u32,
88 vqshlq_n_u64, vqshlb_n_u8, vqshlh_n_u16, vqshls_n_u32, vqshld_n_u64,
89 vqshlu_n_s8, vqshlu_n_s16, vqshlu_n_s32, vqshlu_n_s64, vqshluq_n_s8,
90 vqshluq_n_s16, vqshluq_n_s32, vqshluq_n_s64, vqshlub_n_s8,
91 vqshluh_n_s16, vqshlus_n_s32, vqshlud_n_s64, vqshrn_n_u16,
92 vqshrn_n_u32, vqshrn_n_u64, vqshrnh_n_u16, vqshrns_n_u32,
93 vqshrnd_n_u64, vqsubb_u8, vqsubh_u16, vqsubs_u32, vqsubd_u64,
94 vrsra_n_u8, vrsra_n_u16, vrsra_n_u32, vrsra_n_u64, vrsraq_n_u8,
95 vrsraq_n_u16, vrsraq_n_u32, vrsraq_n_u64, vrsrad_n_u64, vsli_n_u8,
96 vsli_n_u16, vsli_n_u32,vsli_n_u64, vsliq_n_u8, vsliq_n_u16,
97 vsliq_n_u32, vsliq_n_u64, vslid_n_u64, vsqadd_u8, vsqadd_u16,
98 vsqadd_u32, vsqadd_u64, vsqaddq_u8, vsqaddq_u16, vsqaddq_u32,
99 vsqaddq_u64, vsqaddb_u8, vsqaddh_u16, vsqadds_u32, vsqaddd_u64,
100 vsra_n_u8, vsra_n_u16, vsra_n_u32, vsra_n_u64, vsraq_n_u8,
101 vsraq_n_u16, vsraq_n_u32, vsraq_n_u64, vsrad_n_u64, vsri_n_u8,
102 vsri_n_u16, vsri_n_u32, vsri_n_u64, vsriq_n_u8, vsriq_n_u16,
103 vsriq_n_u32, vsriq_n_u64, vsrid_n_u64): Remove casts.
105 2014-06-03 Teresa Johnson <tejohnson@google.com>
107 * tree-sra.c (modify_function): Record caller nodes after rebuild.
109 2014-06-02 Jason Merrill <jason@redhat.com>
111 PR c++/61020
112 * varpool.c (ctor_for_folding): Handle uninitialized vtables.
114 2014-06-03 Alan Lawrence <alan.lawrence@arm.com>
116 * config/aarch64/aarch64.c (aarch64_evpc_ext): allow and handle
117 location == 0.
119 2014-06-03 Alan Lawrence <alan.lawrence@arm.com>
121 * config/aarch64/aarch64-simd.md (aarch64_rev<REVERSE:rev-op><mode>):
122 New pattern.
123 * config/aarch64/aarch64.c (aarch64_evpc_rev): New function.
124 (aarch64_expand_vec_perm_const_1): Add call to aarch64_evpc_rev.
125 * config/aarch64/iterators.md (REVERSE): New iterator.
126 (UNSPEC_REV64, UNSPEC_REV32, UNSPEC_REV16): New enum elements.
127 (rev_op): New int_attribute.
128 * config/aarch64/arm_neon.h (vrev16_p8, vrev16_s8, vrev16_u8,
129 vrev16q_p8, vrev16q_s8, vrev16q_u8, vrev32_p8, vrev32_p16, vrev32_s8,
130 vrev32_s16, vrev32_u8, vrev32_u16, vrev32q_p8, vrev32q_p16, vrev32q_s8,
131 vrev32q_s16, vrev32q_u8, vrev32q_u16, vrev64_f32, vrev64_p8,
132 vrev64_p16, vrev64_s8, vrev64_s16, vrev64_s32, vrev64_u8, vrev64_u16,
133 vrev64_u32, vrev64q_f32, vrev64q_p8, vrev64q_p16, vrev64q_s8,
134 vrev64q_s16, vrev64q_s32, vrev64q_u8, vrev64q_u16, vrev64q_u32):
135 Replace temporary __asm__ with __builtin_shuffle.
137 2014-06-03 Andrew Bennett <andrew.bennett@imgtec.com>
139 * config/mips/mips-cpus.def: Add mips32r3, mips32r5, mips64r3 and
140 mips64r5.
141 * config/mips/mips-tables.opt: Regenerate.
142 * config/mips/mips.c (mips_compute_frame_info): Changed if statement
143 to use mips_isa_rev rather than ISA_MIPS32R2.
144 * config/mips/mips.h (ISA_MIPS32R3): New define.
145 (ISA_MIPS32R5): New define.
146 (ISA_MIPS64R3): New define.
147 (ISA_MIPS64R5): New define.
148 (TARGET_CPU_CPP_BUILTINS): Added support for ISA_MIPS32R3, ISA_MIPS32R5,
149 ISA_MIPS64R3 and ISA_MIPS64R5.
150 (MIPS_ISA_LEVEL_SPEC): Added support for mips32r3, mips32r5, mips64r3
151 and mips64r5.
152 (MIPS_ISA_SYNCI_SPEC): Likewise.
153 (ISA_HAS_64BIT_REGS): Added ISA_MIPS64R3 and ISA_MIPS64R5.
154 (LINK_SPEC): Added mips32r3 and mips32r5.
155 * config/mips/t-isa3264 (MULTILIB_MATCHES): Map mips32r3 and mips32r5
156 to mips32r2; and mips64r3 and mips64r5 to mips64r2.
157 * config/mips/t-mti-elf (MULTILIB_MATCHES): Likewise.
158 * config/mips/t-mti-linux (MULTILIB_MATCHES): Likewise.
159 * config/mips/t-sde (MULTILIB_MATCHES): Likewise.
160 * config/mips/t-sdemtk (MULTILIB_MATCHES): New define.
161 * doc/invoke.texi: Document mips32r3, mips32r5, mips64r3 and mips64r5.
163 2014-06-03 Andrew Bennett <andrew.bennett@imgtec.com>
165 * doc/invoke.texi: Document -mxpa and -mno-xpa MIPS command line
166 options.
167 * config/mips/mips.opt (mxpa): New option.
168 * config/mips/mips.h (ASM_SPEC): Pass mxpa and mno-xpa to the
169 assembler.
171 2014-06-03 Martin Jambor <mjambor@suse.cz>
173 PR ipa/61160
174 * ipa-cp.c (cgraph_edge_brings_value_p): Handle edges leading to
175 thunks.
177 2014-06-03 Thomas Preud'homme <thomas.preudhomme@arm.com>
179 PR tree-optimization/61328
180 * tree-ssa-math-opts.c (init_symbolic_number): Extract symbolic number
181 initialization from find_bswap_or_nop_1.
182 (find_bswap_or_nop_1): Test return value of find_bswap_or_nop_1 stored
183 in source_expr2 before using the size value the function sets. Also
184 make use of init_symbolic_number () in both the old place and
185 find_bswap_or_nop_load () to avoid reading uninitialized memory when
186 doing recursion in the GIMPLE_BINARY_RHS case.
188 2014-06-03 Richard Biener <rguenther@suse.de>
190 PR tree-optimization/61383
191 * tree-ssa-ifcombine.c (bb_no_side_effects_p): Make sure
192 stmts can't trap.
194 2014-06-03 Richard Sandiford <rdsandiford@googlemail.com>
196 * defaults.h (USE_MD_CONSTRAINTS, EXTRA_MEMORY_CONSTRAINT)
197 (EXTRA_ADDRESS_CONSTRAINT, DEFAULT_CONSTRAINT_LEN, CONSTRAINT_LEN)
198 (CONST_OK_FOR_CONSTRAINT_P, CONST_DOUBLE_OK_FOR_LETTER_P)
199 (REG_CLASS_FROM_CONSTRAINT, EXTRA_CONSTRAINT_STR): Delete definitions
200 in this file.
201 (REG_CLASS_FROM_LETTER, CONST_OK_FOR_LETTER_P)
202 (CONST_DOUBLE_OK_FOR_LETTER_P, EXTRA_CONSTRAINT): Move poising to...
203 * system.h: ...here and make it unconditional.
204 * target.def (conditional_register_usage): Mention
205 define_register_constraint instead of old-style constraint macros.
206 * doc/tm.texi.in: Remove documentation for old-style constraint macros.
207 * doc/tm.texi: Regenerate.
208 * genoutput.c: Remove USE_MD_CONSTRAINTS conditions and all code
209 protected by !USE_MD_CONSTRAINTS.
210 * config/frv/frv.md: Remove quote from old version of documentation.
211 * config/frv/frv.c (frv_conditional_register_usage): Likewise.
212 * config/m32r/m32r.c (easy_di_const, easy_df_const): Avoid mentioning
213 CONST_DOUBLE_OK_FOR_LETTER.
214 * config/sh/constraints.md: Likewise EXTRA_CONSTRAINT.
216 2014-06-02 Andrew Pinski <apinski@cavium.com>
218 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER):
219 /lib/ld-linux32-aarch64.so.1 is used for ILP32.
220 (LINUX_TARGET_LINK_SPEC): Update linker script for ILP32.
221 file whose name depends on -mabi= and -mbig-endian.
222 * config/aarch64/t-aarch64-linux (MULTILIB_OSDIRNAMES):
223 Handle LP64 better and handle ilp32 too.
224 (MULTILIB_OPTIONS): Delete.
225 (MULTILIB_DIRNAMES): Delete.
227 2014-06-02 Andrew MacLeod <amacleod@redhat.com>
229 * expr.h: Remove prototypes of functions defined in builtins.c.
230 * tree.h: (build_call_expr_*, build_string_literal): Add prototypes.
231 Remove prototypes of functions defined in builtins.c.
232 * builtins.h: Update prototype list to include all exported functions.
233 * builtins.c: (default_libc_has_function, gnu_libc_has_function,
234 no_c99_libc_has_function): Move to targhooks.c
235 (build_string_literal, build_call_expr_loc_array,
236 build_call_expr_loc_vec, build_call_expr_loc, build_call_expr): Move
237 to tree.c.
238 (expand_builtin_object_size, fold_builtin_object_size): Make static.
239 * targhooks.c (default_libc_has_function, gnu_libc_has_function,
240 no_c99_libc_has_function): Relocate from builtins.c.
241 * tree.c: Include builtins.h.
242 (build_call_expr_loc_array, build_call_expr_loc_vec,
243 build_call_expr_loc, build_call_expr, build_string_literal): Relocate
244 from builtins.c.
245 * fold-const.h (fold_fma): Move prototype to builtins.h.
246 * realmpfr.h (do_mpc_arg2): Move prototype to builtins.h.
247 * asan.c: Include builtins.h.
248 * cfgexpand.c: Likewise.
249 * convert.c: Likewise.
250 * emit-rtl.c: Likewise.
251 * except.c: Likewise.
252 * expr.c: Likewise.
253 * fold-const.c: Likewise.
254 * gimple-fold.c: Likewise.
255 * gimple-ssa-strength-reduction.c: Likewise.
256 * gimplify.c: Likewise.
257 * ipa-inline.c: Likewise.
258 * ipa-prop.c: Likewise.
259 * lto-streamer-out.c: Likewise.
260 * stmt.c: Likewise.
261 * tree-inline.c: Likewise.
262 * tree-object-size.c: Likewise.
263 * tree-sra.c: Likewise.
264 * tree-ssa-ccp.c: Likewise.
265 * tree-ssa-forwprop.c: Likewise.
266 * tree-ssa-loop-ivcanon.c: Likewise.
267 * tree-ssa-loop-ivopts.c: Likewise.
268 * tree-ssa-math-opts.c: Likewise.
269 * tree-ssa-reassoc.c: Likewise.
270 * tree-ssa-threadedge.c: Likewise.
271 * tree-streamer-in.c: Likewise.
272 * tree-vect-data-refs.c: Likewise.
273 * tree-vect-patterns.c: Likewise.
274 * tree-vect-stmts.c: Likewise.
275 * config/aarch64/aarch64.c: Likewise.
276 * config/alpha/alpha.c: Likewise.
277 * config/arc/arc.c: Likewise.
278 * config/arm/arm.c: Likewise.
279 * config/avr/avr.c: Likewise.
280 * config/bfin/bfin.c: Likewise.
281 * config/c6x/c6x.c: Likewise.
282 * config/cr16/cr16.c: Likewise.
283 * config/cris/cris.c: Likewise.
284 * config/epiphany/epiphany.c: Likewise.
285 * config/fr30/fr30.c: Likewise.
286 * config/frv/frv.c: Likewise.
287 * config/h8300/h8300.c: Likewise.
288 * config/i386/i386.c: Likewise.
289 * config/i386/winnt.c: Likewise.
290 * config/ia64/ia64.c: Likewise.
291 * config/iq2000/iq2000.c: Likewise.
292 * config/lm32/lm32.c: Likewise.
293 * config/m32c/m32c.c: Likewise.
294 * config/m32r/m32r.c: Likewise.
295 * config/m68k/m68k.c: Likewise.
296 * config/mcore/mcore.c: Likewise.
297 * config/mep/mep.c: Likewise.
298 * config/microblaze/microblaze.c: Likewise.
299 * config/mips/mips.c: Likewise.
300 * config/mmix/mmix.c: Likewise.
301 * config/mn10300/mn10300.c: Likewise.
302 * config/moxie/moxie.c: Likewise.
303 * config/msp430/msp430.c: Likewise.
304 * config/nds32/nds32.c: Likewise.
305 * config/pa/pa.c: Likewise.
306 * config/pdp11/pdp11.c: Likewise.
307 * config/picochip/picochip.c: Likewise.
308 * config/rl78/rl78.c: Likewise.
309 * config/rs6000/rs6000.c: Likewise.
310 * config/rx/rx.c: Likewise.
311 * config/s390/s390.c: Likewise.
312 * config/score/score.c: Likewise.
313 * config/sh/sh.c: Likewise.
314 * config/sparc/sparc.c: Likewise.
315 * config/spu/spu.c: Likewise.
316 * config/stormy16/stormy16.c: Likewise.
317 * config/tilegx/tilegx.c: Likewise.
318 * config/tilepro/tilepro.c: Likewise.
319 * config/v850/v850.c: Likewise.
320 * config/vax/vax.c: Likewise.
321 * config/xtensa/xtensa.c: Likewise.
323 2014-06-02 Jeff Law <law@redhat.com>
325 PR rtl-optimization/61094
326 * ree.c (combine_reaching_defs): Do not reextend an insn if it
327 was marked as do_no_reextend. If a copy is needed to eliminate
328 an extension, then mark it as do_not_reextend.
330 2014-06-02 Marcus Shawcroft <marcus.shawcroft@arm.com>
332 * config/aarch64/aarch64.md (set_fpcr): Drop ISB after FPCR write.
334 2014-06-02 Richard Henderson <rth@redhat.com>
336 PR target/61336
337 * config/alpha/alpha.c (print_operand_address): Allow symbolic
338 addresses inside asms. Use output_operand_lossage instead of
339 gcc_unreachable.
341 2014-06-02 Uros Bizjak <ubizjak@gmail.com>
343 PR target/61239
344 * config/i386/i386.c (ix86_expand_vec_perm) [case V32QImode]: Use
345 GEN_INT (-128) instead of GEN_INT (128) to set MSB of QImode constant.
347 2014-06-02 Tom de Vries <tom@codesourcery.com>
349 * config/aarch64/aarch64.c (aarch64_float_const_representable_p): Handle
350 case that x has VOIDmode.
352 2014-06-02 Bernd Schmidt <bernds@codesourcery.com>
354 * varasm.c (copy_constant): Delete function.
355 (build_constant_desc): Don't call it.
357 2014-06-02 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
359 PR target/61154
360 * config/arm/arm.h (TARGET_SUPPORTS_WIDE_INT): Define.
361 * config/arm/arm.md (mov64 splitter): Replace const_double_operand
362 with immediate_operand.
364 2014-06-02 Andreas Schwab <schwab@suse.de>
366 * config/ia64/ia64.c
367 (ia64_first_cycle_multipass_dfa_lookahead_guard): Check
368 pending_data_specs first.
370 2014-06-02 Richard Biener <rguenther@suse.de>
372 PR tree-optimization/61378
373 * tree-ssa-sccvn.c (vn_reference_lookup_3): Initialize
374 valueized_anything.
376 2014-06-01 Uros Bizjak <ubizjak@gmail.com>
378 * config/i386/constraints.md (Bw): Rename from 'w'.
379 (Bz): Rename from 'z'.
380 * config/i386/i386.md: Change 'w' to 'Bw' and 'z' to 'Bz' globally.
382 2014-06-01 Kai Tietz <ktietz@redhat.com>
384 PR target/61377
385 * config/i386/constrains.md (define_constrain): New 'Bs' constraint.
386 * config/i386/i386.md (sibcall_insn_operand): Use Bs
387 instead of m constraint.
389 2014-05-31 Andreas Schwab <schwab@linux-m68k.org>
391 * config/m68k/m68k.md (beq0_di, bne0_di): Make the "o" constraint
392 a separate alternative where the scratch operand 2 is marked as
393 early clobber.
395 2014-05-31 Kugan Vivekanandarajah <kuganv@linaro.org>
397 * config/arm/arm.c (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New define.
398 (arm_builtins) : Add ARM_BUILTIN_GET_FPSCR and ARM_BUILTIN_SET_FPSCR.
399 (bdesc_2arg) : Add description for builtins __builtins_arm_set_fpscr
400 and __builtins_arm_get_fpscr.
401 (arm_init_builtins) : Initialize builtins __builtins_arm_set_fpscr and
402 __builtins_arm_get_fpscr.
403 (arm_expand_builtin) : Expand builtins __builtins_arm_set_fpscr and
404 __builtins_arm_ldfpscr.
405 (arm_atomic_assign_expand_fenv): New function.
406 * config/arm/vfp.md (set_fpscr): New pattern.
407 (get_fpscr) : Likewise.
408 * config/arm/unspecs.md (unspecv): Add VUNSPEC_GET_FPSCR and
409 VUNSPEC_SET_FPSCR.
410 * doc/extend.texi (AARCH64 Built-in Functions) : Document
411 __builtins_arm_set_fpscr, __builtins_arm_get_fpscr.
413 2014-05-30 Jakub Jelinek <jakub@redhat.com>
415 * asan.c (report_error_func): Add SLOW_P argument, use
416 BUILT_IN_ASAN_*_N if set.
417 (build_check_stmt): Likewise.
418 (instrument_derefs): If T has insufficient alignment,
419 force same handling as for odd sizes.
421 * sanitizer.def (BUILT_IN_ASAN_REPORT_LOAD_N,
422 BUILT_IN_ASAN_REPORT_STORE_N): New.
423 * asan.c (struct asan_mem_ref): Change access_size type to
424 HOST_WIDE_INT.
425 (asan_mem_ref_init, asan_mem_ref_new, get_mem_refs_of_builtin_call,
426 update_mem_ref_hash_table): Likewise.
427 (asan_mem_ref_hasher::hash): Hash in a HWI.
428 (report_error_func): Change size_in_bytes argument to HWI.
429 Use *_N builtins if size_in_bytes is larger than 16 or not power of
430 two.
431 (build_shadow_mem_access): New function.
432 (build_check_stmt): Use it. Change size_in_bytes argument to HWI.
433 Handle size_in_bytes not power of two or larger than 16.
434 (instrument_derefs): Don't give up if size_in_bytes is not
435 power of two or is larger than 16.
437 2014-05-30 Kai Tietz <ktietz@redhat.com>
439 PR target/60104
440 * config/i386/i386.c (x86_output_mi_thunk): Add memory case
441 for sibling-tail-calls.
442 * config/i386/i386.md (sibcall_insn_operand): Add memory-constrain
443 to its use.
444 * config/i386/predicates.md (sibcall_memory_operand): New predicate.
445 (sibcall_insn_operand): Add check for sibcall_memory_operand.
447 2014-05-30 Pitchumani Sivanupandi <pitchumani.s@atmel.com>
449 * config/avr/avr-mcus.def: Change ATA6289 ISA to AVR4
450 * config/avr/avr-tables.opt: Regenerate.
451 * config/avr/t-multilib: Regenerate.
452 * doc/avr-mmcu.texi: Regenerate.
454 2014-05-30 Ian Lance Taylor <iant@google.com>
456 * config/i386/xmmintrin.h (_mm_pause): Move out of scope of pragma
457 target("sse").
459 2014-05-30 Tom de Vries <tom@codesourcery.com>
461 * config/i386/i386.c (TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS):
462 Redefine as true.
464 2014-05-30 Tom de Vries <tom@codesourcery.com>
466 * lra-int.h (struct lra_reg): Add field actual_call_used_reg_set.
467 * lra.c (initialize_lra_reg_info_element): Add init of
468 actual_call_used_reg_set field.
469 (lra): Call lra_create_live_ranges before lra_inheritance for
470 -fuse-caller-save.
471 * lra-assigns.c (lra_assign): Allow call_used_regs to cross calls for
472 -fuse-caller-save.
473 * lra-constraints.c (need_for_call_save_p): Use actual_call_used_reg_set
474 instead of call_used_reg_set for -fuse-caller-save.
475 * lra-lives.c (process_bb_lives): Calculate actual_call_used_reg_set.
477 2014-05-30 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
479 * config/arm/thumb2.md (*thumb2_movhi_insn): Set type of movw
480 to mov_imm.
481 * config/arm/vfp.md (*thumb2_movsi_vfp): Likewise.
483 2014-05-30 Richard Sandiford <rdsandiford@googlemail.com>
485 * ira.c (ira_get_dup_out_num): Check for output operands at
486 the start of the loop. Handle cases where an included alternative
487 follows an excluded one.
489 2014-05-29 Mike Stump <mikestump@comcast.net>
491 PR debug/61352
492 * collect2.c (maybe_run_lto_and_relink): Be sure to always run
493 post ld passes when lto is used.
495 2014-05-29 Vladimir Makarov <vmakarov@redhat.com>
497 PR rtl-optimization/61325
498 * lra-constraints.c (process_address): Rename to process_address_1.
499 (process_address): New function.
501 2014-05-29 Alan Lawrence <alan.lawrence@arm.com>
503 * config/aarch64/aarch64-builtins.c (aarch64_types_binopv_qualifiers,
504 TYPES_BINOPV): New static data.
505 * config/aarch64/aarch64-simd-builtins.def (im_lane_bound):
506 New builtin.
507 * config/aarch64/aarch64-simd.md (aarch64_ext,
508 aarch64_im_lane_boundsi): New patterns.
509 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_const_1): Match
510 patterns for EXT.
511 (aarch64_evpc_ext): New function.
513 * config/aarch64/iterators.md (UNSPEC_EXT): New enum element.
515 * config/aarch64/arm_neon.h (vext_f32, vext_f64, vext_p8, vext_p16,
516 vext_s8, vext_s16, vext_s32, vext_s64, vext_u8, vext_u16, vext_u32,
517 vext_u64, vextq_f32, vextq_f64, vextq_p8, vextq_p16, vextq_s8,
518 vextq_s16, vextq_s32, vextq_s64, vextq_u8, vextq_u16, vextq_u32,
519 vextq_u64): Replace __asm with __builtin_shuffle and im_lane_boundsi.
521 2014-05-29 Tom de Vries <tom@codesourcery.com>
523 * rtl.h (BLOCK_SYMBOL_CHECK): Use SYMBOL_REF_FLAGS.
525 2014-05-29 Richard Earnshaw <rearnsha@arm.com>
526 Richard Sandiford <rdsandiford@googlemail.com>
528 * arm/iterators.md (shiftable_ops): New code iterator.
529 (t2_binop0, arith_shift_insn): New code attributes.
530 * arm/predicates.md (shift_nomul_operator): New predicate.
531 * arm/arm.md (insn_enabled): Delete.
532 (enabled): Remove insn_enabled test.
533 (*arith_shiftsi): Delete. Replace with ...
534 (*<arith_shift_insn>_multsi): ... new pattern.
535 (*<arith_shift_insn>_shiftsi): ... new pattern.
536 * config/arm/arm.c (arm_print_operand): Handle operand format 'b'.
538 2014-05-29 Radovan Obradovic <robradovic@mips.com>
539 Tom de Vries <tom@codesourcery.com>
541 * config/mips/mips.h (POST_CALL_TMP_REG): Define.
542 * config/mips/mips.c (mips_emit_call_insn): Add POST_CALL_TMP_REG
543 clobber.
544 (mips_split_call): Use POST_CALL_TMP_REG.
545 (TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): Redefine to true.
547 2014-05-29 Tom de Vries <tom@codesourcery.com>
549 * final.c (collect_fn_hard_reg_usage): Guard variable declaration
550 with #ifdef STACK_REGS.
552 2014-05-28 Jan Hubicka <hubicka@ucw.cz>
554 * varasm.c (get_variable_section): Walk aliases.
555 (place_block_symbol): Walk aliases.
557 2014-05-28 Tom de Vries <tom@codesourcery.com>
559 Revert:
560 2014-05-28 Tom de Vries <tom@codesourcery.com>
562 * lra-int.h (struct lra_reg): Add field actual_call_used_reg_set.
563 * lra.c (initialize_lra_reg_info_element): Add init of
564 actual_call_used_reg_set field.
565 (lra): Call lra_create_live_ranges before lra_inheritance for
566 -fuse-caller-save.
567 * lra-assigns.c (lra_assign): Allow call_used_regs to cross calls for
568 -fuse-caller-save.
569 * lra-constraints.c (need_for_call_save_p): Use
570 actual_call_used_reg_set instead of call_used_reg_set for
571 -fuse-caller-save.
572 * lra-lives.c (process_bb_lives): Calculate actual_call_used_reg_set.
574 2014-05-28 Richard Sandiford <rdsandiford@googlemail.com>
576 * doc/md.texi: Document that the % constraint character must
577 be at the beginning of the string.
578 * genoutput.c (validate_insn_alternatives): Check that '=',
579 '+' and '%' only appear at the beginning of a constraint.
580 * ira.c (commutative_constraint_p): Delete.
581 (ira_get_dup_out_num): Expect the '%' commutativity marker to be
582 at the start of the string.
583 * config/alpha/alpha.md (*movmemdi_1, *clrmemdi_1): Remove
584 duplicate '='s.
585 * config/arm/neon.md (bicdi3_neon): Likewise.
586 * config/iq2000/iq2000.md (addsi3_internal, subsi3_internal, sgt_si)
587 (slt_si, sltu_si): Likewise.
588 * config/vax/vax.md (sbcdi3): Likewise.
589 * config/h8300/h8300.md (*cmpstz): Remove duplicate '+'.
590 * config/arc/arc.md (mulsi_600, mulsidi_600, umulsidi_600)
591 (mul64): Move '%' to beginning of constraint.
592 * config/arm/arm.md (*xordi3_insn): Likewise.
593 * config/nds32/nds32.md (add<mode>3, mulsi3, andsi3, iorsi3)
594 (xorsi3): Likewise.
596 2014-05-28 Richard Sandiford <rdsandiford@googlemail.com>
598 * doc/md.texi: Document the restrictions on the "enabled" attribute.
600 2014-05-28 Jason Merrill <jason@redhat.com>
602 PR c++/47202
603 * cgraph.h (symtab_node::get_comdat_group_id): New.
604 * cgraphunit.c (analyze_functions): Call it.
605 * symtab.c (dump_symtab_node): Likewise.
606 * tree.c (decl_comdat_group_id): New.
607 * tree.h: Declare it.
608 * lto-streamer-out.c (write_symbol): Use it.
609 * trans-mem.c (ipa_tm_create_version_alias): Likewise.
611 2014-05-28 Francois-Xavier Coudert <fxcoudert@gcc.gnu.org>
613 PR bootstrap/PR61146
614 * wide-int.cc: Do not include longlong.h when compiling with clang.
616 2014-05-28 Richard Biener <rguenther@suse.de>
618 * tree-ssa-propagate.c (add_control_edge): Print less vertical space.
619 * tree-vrp.c (extract_range_from_ssa_name): Also copy VR_UNDEFINED.
620 (vrp_visit_assignment_or_call): Print less vertical space.
621 (vrp_visit_stmt): Likewise.
622 (vrp_visit_phi_node): Likewise. For a PHI argument with
623 VR_VARYING range consider recording it as copy.
625 2014-05-28 Richard Biener <rguenther@suse.de>
627 Revert
628 2014-05-28 Richard Biener <rguenther@suse.de>
630 * hwint.h (HOST_WIDE_INT_PRINT_*): Define in terms of PRI*64.
632 2014-05-28 Bernd Edlinger <bernd.edlinger@hotmail.de>
634 * expr.c (expand_assignment): Fold the bitpos in the to_rtx if
635 sufficiently aligned and an offset is used at the same time.
636 (expand_expr_real_1): Likewise.
638 2014-05-28 Richard Biener <rguenther@suse.de>
640 PR middle-end/61045
641 * fold-const.c (fold_comparison): When folding
642 X +- C1 CMP Y +- C2 to X CMP Y +- C2 +- C1 also ensure
643 the sign of the remaining constant operand stays the same.
645 2014-05-28 Kaushik Phatak <kaushik.phatak@kpit.com>
647 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
648 __RL78_64BIT_DOUBLES__ or __RL78_32BIT_DOUBLES__.
649 (ASM_SPEC): Pass -m64bit-doubles or -m32bit-doubles on
650 to the assembler.
651 (DOUBLE_TYPE_SIZE): Use 64 bit if TARGET_64BIT_DOUBLES is true.
652 * gcc/config/rl78/rl78.opt (m64bit-doubles): New option.
653 (m32bit-doubles) Likewise.
654 * gcc/config/rl78/t-rl78: Add 64-bit-double multilib.
655 * doc/invoke.texi: Document -m32bit-doubles and -m64bit-doubles
656 option for RL78.
658 2014-05-28 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
660 * configure.ac ($gcc_cv_ld_clearcap): New test.
661 * configure: Regenerate.
662 * config.in: Regenerate.
663 * config/sol2.opt (mclear-hwcap): New option.
664 * config/sol2.h (LINK_CLEARCAP_SPEC): Define.
665 * config/sol2-clearcap.map: Moved here from
666 testsuite/gcc.target/i386/clearcap.map.
667 * config/sol2-clearcapv2.map: Move here from
668 gcc.target/i386/clearcapv2.map.
669 * config/t-sol2 (install): Depend on install-clearcap-map.
670 (install-clearcap-map): New target.
671 * doc/invoke.texi (Option Summary, Solaris 2 Options): Document
672 -mclear-hwcap.
674 2014-05-28 Richard Biener <rguenther@suse.de>
676 * hwint.h (*_HALF_WIDE_INT*): Move to ...
677 * wide-int.cc (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT):
678 ... here and remove the rest.
679 * hwint.h (HOST_WIDE_INT_PRINT_*): Define in terms of PRI*64.
681 2014-05-28 Richard Biener <rguenther@suse.de>
683 PR tree-optimization/61335
684 * tree-vrp.c (vrp_visit_phi_node): If the compare of old and
685 new range fails, drop to varying.
687 2014-05-28 Tom de Vries <tom@codesourcery.com>
689 * lra-int.h (struct lra_reg): Add field actual_call_used_reg_set.
690 * lra.c (initialize_lra_reg_info_element): Add init of
691 actual_call_used_reg_set field.
692 (lra): Call lra_create_live_ranges before lra_inheritance for
693 -fuse-caller-save.
694 * lra-assigns.c (lra_assign): Allow call_used_regs to cross calls for
695 -fuse-caller-save.
696 * lra-constraints.c (need_for_call_save_p): Use
697 actual_call_used_reg_set instead of call_used_reg_set for
698 -fuse-caller-save.
699 * lra-lives.c (process_bb_lives): Calculate actual_call_used_reg_set.
701 2014-05-28 Radovan Obradovic <robradovic@mips.com>
702 Tom de Vries <tom@codesourcery.com>
704 * doc/invoke.texi (@item Optimization Options): Add -fuse-caller-save
705 to gccoptlist.
706 (@item -fuse-caller-save): New item.
708 2014-05-28 Radovan Obradovic <robradovic@mips.com>
709 Tom de Vries <tom@codesourcery.com>
711 * opts.c (default_options_table): Add OPT_LEVELS_2_PLUS entry with
712 OPT_fuse_caller_save.
714 2014-05-28 Radovan Obradovic <robradovic@mips.com>
715 Tom de Vries <tom@codesourcery.com>
717 * df-scan.c (df_get_call_refs): Use get_call_reg_set_usage.
718 * caller-save.c (setup_save_areas, save_call_clobbered_regs): Use
719 get_call_reg_set_usage.
720 * resource.c (mark_set_resources, mark_target_live_regs): Use
721 get_call_reg_set_usage.
722 * ira-int.h (struct ira_allocno): Add crossed_calls_clobbered_regs
723 field.
724 (ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS): Define.
725 * ira-lives.c (process_bb_node_lives): Use get_call_reg_set_usage.
726 Calculate ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS.
727 * ira-build.c (ira_create_allocno): Init
728 ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS.
729 (create_cap_allocno, propagate_allocno_info)
730 (propagate_some_info_from_allocno)
731 (copy_info_to_removed_store_destinations): Handle
732 ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS.
733 * ira-costs.c (ira_tune_allocno_costs): Use
734 ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS to adjust costs.
736 2014-05-28 Radovan Obradovic <robradovic@mips.com>
737 Tom de Vries <tom@codesourcery.com>
739 * cgraph.h (struct cgraph_rtl_info): Add function_used_regs
740 and function_used_regs_valid fields.
741 * final.c: Move include of hard-reg-set.h to before rtl.h to declare
742 find_all_hard_reg_sets.
743 (collect_fn_hard_reg_usage, get_call_fndecl, get_call_cgraph_rtl_info)
744 (get_call_reg_set_usage): New function.
745 (rest_of_handle_final): Use collect_fn_hard_reg_usage.
746 * regs.h (get_call_reg_set_usage): Declare.
748 2014-05-28 Georg-Johann Lay <avr@gjlay.de>
750 PR libgcc/61152
751 * config/dbx.h (License): Add Runtime Library Exception.
752 * config/newlib-stdint.h (License): Same.
753 * config/rtems.h (License): Same
754 * config/initfini-array.h (License): Same
755 * config/v850/v850.h (License): Same.
756 * config/v850/v850-opts.h (License): Same
757 * config/v850/rtems.h (License): Same.
759 2014-05-28 Georg-Johann Lay <avr@gjlay.de>
761 PR target/61044
762 * doc/extend.texi (Local Labels): Note that label differences are
763 not supported for AVR.
765 2014-05-27 Andrew Pinski <apinski@cavium.com>
767 * config/aarch64/aarch64.md (stack_protect_set_<mode>):
768 Use <w> for the register in assembly template.
769 (stack_protect_test): Use the mode of operands[0] for the result.
770 (stack_protect_test_<mode>): Use <w> for the register
771 in assembly template.
773 2014-05-27 DJ Delorie <dj@redhat.com>
775 * config/rx/rx.c (add_vector_labels): New.
776 (rx_output_function_prologue): Call it.
777 (rx_handle_func_attribute): Don't require empty arguments.
778 (rx_handle_vector_attribute): New.
779 (rx_attribute_table): Add "vector" attribute.
780 * doc/extend.texi (interrupt, vector): Document new/changed
781 RX-specific attributes.
783 * config/rx/rx.c (rx_adjust_insn_length): Skip for non-insns.
785 2014-05-27 Eric Botcazou <ebotcazou@adacore.com>
787 * double-int.c (div_and_round_double) <ROUND_DIV_EXPR>: Use the proper
788 predicate to detect a negative quotient.
790 2014-05-27 Eric Botcazou <ebotcazou@adacore.com>
792 * fold-const.c (fold_comparison): Clean up and extend X +- C1 CMP C2
793 to X CMP C2 -+ C1 transformation to EQ_EXPR/NE_EXPR.
794 Add X - Y CMP 0 to X CMP Y transformation.
795 (fold_binary_loc) <EQ_EXPR/NE_EXPR>: Remove same transformations.
797 2014-05-27 Segher Boessenkool <segher@kernel.crashing.org>
799 * stmt.c (dump_case_nodes): Don't convert values to HOST_WIDE_INT
800 before printing.
802 2014-05-27 Steve Ellcey <sellcey@mips.com>
804 * config/mips/mips.c: Add include of cgraph.h.
806 2014-05-27 Richard Biener <rguenther@suse.de>
808 * system.h (__STDC_FORMAT_MACROS): Define as very first thing.
810 2014-05-27 Georg-Johann Lay <avr@gjlay.de>
812 PR libgcc/61152
813 * config/arm/arm.h (License): Add note to COPYING.RUNTIME.
814 * config/arm/arm-cores.def (License): Same.
815 * config/arm/arm-opts.h (License): Same.
816 * config/arm/aout.h (License): Same.
817 * config/arm/bpabi.h (License): Same.
818 * config/arm/elf.h (License): Same.
819 * config/arm/linux-elf.h (License): Same.
820 * config/arm/linux-gas.h (License): Same.
821 * config/arm/netbsd-elf.h (License): Same.
822 * config/arm/uclinux-eabi.h (License): Same.
823 * config/arm/uclinux-elf.h (License): Same.
824 * config/arm/vxworks.h (License): Same.
826 2014-05-27 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
828 * config/arm/neon.md (neon_bswap<mode>): New pattern.
829 * config/arm/arm.c (neon_itype): Add NEON_BSWAP.
830 (arm_init_neon_builtins): Handle NEON_BSWAP.
831 Define required type nodes.
832 (arm_expand_neon_builtin): Handle NEON_BSWAP.
833 (arm_builtin_vectorized_function): Handle BUILTIN_BSWAP builtins.
834 * config/arm/arm_neon_builtins.def (bswap): Define builtins.
835 * config/arm/iterators.md (VDQHSD): New mode iterator.
837 2014-05-27 Richard Biener <rguenther@suse.de>
839 * tree-vrp.c (vrp_evaluate_conditional_warnv_with_ops_using_ranges):
840 Try using literal operands when comparing value-ranges failed.
842 2014-05-27 Richard Sandiford <rdsandiford@googlemail.com>
844 * ira.c (commutative_operand): Adjust for change to recog_data.
845 [Missing from previous commit.]
847 2014-05-27 Richard Sandiford <rdsandiford@googlemail.com>
849 * system.h (TEST_BIT): New macro.
850 * recog.h (alternative_mask): New type.
851 (ALL_ALTERNATIVES, ALTERNATIVE_BIT): New macros.
852 (recog_data_d): Replace alternative_enabled_p array with
853 enabled_alternatives.
854 (target_recog): New structure.
855 (default_target_recog, this_target_recog): Declare.
856 (get_enabled_alternatives, recog_init): Likewise.
857 * recog.c (default_target_recog, this_target_recog): New variables.
858 (get_enabled_alternatives): New function.
859 (extract_insn): Use it.
860 (recog_init): New function.
861 (preprocess_constraints, constrain_operands): Adjust for change to
862 recog_data.
863 * postreload.c (reload_cse_simplify_operands): Likewise.
864 * reload.c (find_reloads): Likewise.
865 * ira-costs.c (record_reg_classes): Likewise.
866 * ira-lives.c (single_reg_class): Likewise. Fix bug in which
867 all alternatives after a disabled one would be skipped.
868 (ira_implicitly_set_insn_hard_regs): Likewise.
869 * ira.c (ira_setup_alts): Adjust for change to recog_data.
870 * lra-int.h (lra_insn_recog_data): Replace alternative_enabled_p
871 with enabled_alternatives.
872 * lra.c (free_insn_recog_data): Update accordingly.
873 (lra_update_insn_recog_data): Likewise.
874 (lra_set_insn_recog_data): Likewise. Use get_enabled_alternatives.
875 * lra-constraints.c (process_alt_operands): Likewise. Handle
876 only_alternative as part of the enabled mask.
877 * target-globals.h (this_target_recog): Declare.
878 (target_globals): Add a recog field.
879 (restore_target_globals): Restore this_target_recog.
880 * target-globals.c: Include recog.h.
881 (default_target_globals): Initialize recog field.
882 (save_target_globals): Likewise.
883 * reginfo.c (reinit_regs): Call recog_init.
884 * toplev.c (backend_init_target): Likewise.
886 2014-05-27 Richard Sandiford <rdsandiford@googlemail.com>
888 * gencodes.c (main): Make LAST_INSN_CODE higher than any insn code,
889 rather than any named insn's code.
891 2014-05-27 Georg-Johann Lay <avr@gjlay.de>
893 PR libgcc/61152
894 * config/arm/arm-opts.h (License): Add GCC Runtime Library Exception.
895 * config/arm/arm-cores.def (License): Same.
897 2014-05-26 Jan Hubicka <hubicka@ucw.cz>
899 * tree.h (decl_comdat_group): Declare.
900 * cgraph.h (symtab_in_same_comdat_p): Move offline to ...
901 * tree.c (decl_comdat_group): Here.
903 2014-05-26 Richard Sandiford <r.sandiford@uk.ibm.com>
905 PR rtl-optimization/61222
906 * combine.c (simplify_shift_const_1): When moving a PLUS outside
907 the shift, truncate the PLUS operand to the result mode.
909 2014-05-26 Uros Bizjak <ubizjak@gmail.com>
911 PR target/61271
912 * config/i386/i386.c (ix86_rtx_costs)
913 <case CONST_INT, case CONST, case LABEL_REF, case SYMBOL_REF>:
914 Fix condition.
916 2014-05-26 Martin Jambor <mjambor@suse.cz>
918 * ira.c (split_live_ranges_for_shrink_wrap): Remove bailout on
919 subreg uses.
921 2014-05-26 Richard Biener <rguenther@suse.de>
923 * wide-int.h (wi::int_traits <long>, wi::int_traits <unsigned long>,
924 wi::int_traits <long long>, wi::int_traits <unsigned long long>):
925 Provide specializations.
926 (wi::int_traits <HOST_WIDE_INT>,
927 wi::int_traits <unsigned HOST_WIDE_INT>): Remove specializations.
929 2014-05-26 Alan Modra <amodra@gmail.com>
931 PR target/61098
932 * config/rs6000/rs6000.c (rs6000_emit_set_const): Remove unneeded
933 params and return a bool. Remove dead code. Update comment.
934 Assert we have a const_int source. Remove bogus code from
935 32-bit HWI days. Move !TARGET_POWERPC64 handling, and correct
936 handling of constants > 2G and reg_equal note, from..
937 (rs6000_emit_set_long_const): ..here. Remove unneeded param and
938 return value. Update comment. If we can, use a new pseudo
939 for intermediate calculations.
940 * config/rs6000/rs6000-protos.h (rs6000_emit_set_const): Update
941 prototype.
942 * config/rs6000/rs6000.md (movsi_internal1_single+1): Update
943 call to rs6000_emit_set_const in splitter.
944 (movdi_internal64+2, +3): Likewise.
946 2014-05-26 Richard Biener <rguenther@suse.de>
948 * system.h: Define __STDC_FORMAT_MACROS before
949 including inttypes.h.
950 * hwint.h (HOST_WIDEST_INT, HOST_BITS_PER_WIDEST_INT,
951 HOST_WIDEST_INT_PRINT, HOST_WIDEST_INT_PRINT_DEC,
952 HOST_WIDEST_INT_PRINT_DEC_C, HOST_WIDEST_INT_PRINT_UNSIGNED,
953 HOST_WIDEST_INT_PRINT_HEX, HOST_WIDEST_INT_PRINT_DOUBLE_HEX,
954 HOST_WIDEST_INT_C): Remove.
955 (PRId64, PRIi64, PRIo64, PRIu64, PRIx64, PRIX64): Define
956 if C99 inttypes.h is not available.
957 * coretypes.h (gcov_type, gcov_type_unsigned): Use [u]int64_t.
958 * gcov-io.h (gcov_type, gcov_type_unsigned): Likewise.
959 * gcov-io.c (gcov_histo_index): Drop non-64bit hwi case.
960 * cfgloop.h (struct niter_desc): Use uint64_t for niter field.
961 * bitmap.c (struct bitmap_descriptor_d): Use uint64_t for counters.
962 (struct output_info): Likewise.
963 (print_statistics): Adjust.
964 (dump_bitmap_statistics): Likewise.
965 * bt-load.c (migrate_btr_defs): Print with PRId64.
966 * cfg.c (dump_edge_info, dump_bb_info): Likewise.
967 (MAX_SAFE_MULTIPLIER): Adjust.
968 * cfghooks.c (dump_bb_for_graph): Print with PRId64.
969 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee,
970 dump_cgraph_node): Likewise.
971 * final.c (dump_basic_block_info): Likewise.
972 * gcov-dump.c (tag_counters, tag_summary, dump_working_sets): Likewise.
973 * gcov.c (format_gcov): Likewise.
974 * ipa-cp.c (good_cloning_opportunity_p): Likewise. Use int64_t
975 for calculation.
976 (get_clone_agg_value): Use HOST_WIDE_INT for offset.
977 * ipa-inline.c (compute_max_insns): Use int64_t for calcuation.
978 (inline_small_functions, dump_overall_stats, dump_inline_stats):
979 Use PRId64 for dumping.
980 * ipa-profile.c (dump_histogram, ipa_profile): Likewise.
981 * ira-color.c (struct allocno_hard_regs): Use int64_t for cost.
982 (add_allocno_hard_regs): Adjust.
983 * loop-doloop.c (doloop_modify): Print using PRId64.
984 * loop-iv.c (inverse): Compute in uint64_t.
985 (determine_max_iter, iv_number_of_iterations): Likewise.
986 * loop-unroll.c (decide_peel_completely, decide_peel_simple):
987 Print using PRId64.
988 * lto-streamer-out.c (write_symbol): Use uint64_t.
989 * mcf.c (CAP_INFINITY): Use int64_t maximum.
990 (dump_fixup_edge, create_fixup_graph, cancel_negative_cycle,
991 find_max_flow, adjust_cfg_counts): Use int64_t and dump with PRId64.
992 * modulo-sched.c (const_iteration_count): Use int64_t.
993 (sms_schedule): Dump using PRId64.
994 * predict.c (dump_prediction): Likewise.
995 * pretty-print.h (pp_widest_integer): Remove.
996 * profile.c (get_working_sets, is_edge_inconsistent,
997 is_inconsistent, read_profile_edge_counts): Dump using PRId64.
998 * tree-pretty-print.c (pp_double_int): Remove case handling
999 HOST_BITS_PER_DOUBLE_INT == HOST_BITS_PER_WIDEST_INT.
1000 * tree-ssa-math-opts.c (struct symbolic_number): Use uint64_t
1001 and adjust users.
1002 (pass_optimize_bswap::execute): Remove restriction on hosts.
1003 * tree-streamer-in.c (streamer_alloc_tree): Use HOST_WIDE_INT.
1004 * tree-streamer-out.c (streamer_write_tree_header): Likewise.
1005 * tree.c (widest_int_cst_value): Remove.
1006 * tree.h (widest_int_cst_value): Likewise.
1007 * value-prof.c (dump_histogram_value): Print using PRId64.
1008 * gengtype.c (main): Also inject int64_t.
1009 * ggc-page.c (struct max_alignment): Use int64_t.
1010 * alloc-pool.c (struct allocation_object_def): Likewise.
1011 * ira-conflicts.c (build_conflict_bit_table): Use uint64_t
1012 for computation.
1013 * doc/tm.texi.in: Remove reference to HOST_WIDEST_INT.
1014 * doc/tm.texi: Regenerated.
1015 * gengtype-lex.l (IWORD): Handle [u]int64_t.
1016 * config/sh/sh.c (expand_cbranchdi4): Use gcov_type.
1017 * config/mmix/mmix-protos.h (mmix_intval, mmix_shiftable_wyde_value,
1018 mmix_output_register_setting): Use [u]int64_t in prototypes.
1019 * config/mmix/mmix.c (mmix_print_operand, mmix_output_register_setting,
1020 mmix_shiftable_wyde_value, mmix_output_shiftvalue_op_from_str,
1021 mmix_output_octa, mmix_output_shifted_value): Adjust.
1022 (mmix_intval): Adjust. Remove unreachable case.
1023 * config/mmix/mmix.md (*nonlocal_goto_receiver_expanded): Use int64_t.
1025 2014-05-26 Richard Biener <rguenther@suse.de>
1027 * configure.ac: Drop __int64 type check. Insist that we
1028 found uint64_t and int64_t.
1029 * hwint.h (HOST_BITS_PER___INT64): Remove.
1030 (HOST_BITS_PER_WIDE_INT): Define to 64 and remove __int64 case.
1031 (HOST_WIDE_INT_PRINT_*): Remove 32bit case.
1032 (HOST_WIDEST_INT*): Define to HOST_WIDE_INT*.
1033 (HOST_WIDEST_FAST_INT): Remove __int64 case.
1034 * vmsdbg.h (struct _DST_SRC_COMMAND): Use int64_t
1035 for dst_q_src_df_rms_cdt.
1036 * configure: Regenerate.
1037 * config.in: Likewise.
1039 2014-05-26 Michael Tautschnig <mt@debian.org>
1041 PR target/61249
1042 * doc/extend.texi (X86 Built-in Functions): Fix parameter lists of
1043 __builtin_ia32_vfrczs[sd] and __builtin_ia32_mpsadbw256.
1045 2014-05-26 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1047 PR rtl-optimization/61278
1048 * shrink-wrap.c (move_insn_for_shrink_wrap): Check df_live.
1050 2014-05-26 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1052 PR rtl-optimization/61220
1053 Part of PR rtl-optimization/61225
1054 * shrink-wrap.c (move_insn_for_shrink_wrap): Skip SP and FP adjustment
1055 insn; skip split_edge for a block with only one successor.
1057 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
1059 * symtab.c (symtab_nonoverwritable_alias): Copy READONLY flag
1060 for variables.
1062 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
1064 * ipa-visibility.c (can_replace_by_local_alias_in_vtable): New function.
1065 (update_vtable_references): New function.
1066 (function_and_variable_visibility): Rewrite also vtable initializers.
1067 * varpool.c (cgraph_variable_initializer_availability): Remove assert.
1069 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
1071 * ggc.h (ggc_grow): New function.
1072 * ggc-none.c (ggc_grow): New function.
1073 * ggc-page.c (ggc_grow): Likewise.
1075 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
1077 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
1078 address_taken_from_non_vtable_p, comdat_can_be_unshared_p_1,
1079 comdat_can_be_unshared_p, cgraph_externally_visible_p,
1080 varpool_externally_visible_p, can_replace_by_local_alias,
1081 update_visibility_by_resolution_info, function_and_variable_visibility,
1082 pass_data_ipa_function_and_variable_visibility,
1083 make_pass_ipa_function_and_variable_visibility,
1084 whole_program_function_and_variable_visibility,
1085 pass_data_ipa_whole_program_visibility,
1086 make_pass_ipa_whole_program_visibility): Move to ipa-visibility.c
1087 * cgraph.h (cgraph_local_node_p): Declare.
1088 * ipa-visibility.c: New file.
1089 * Makefile.in (OBJS): Add ipa-visiblity.o
1091 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
1093 * gimple-fold.c (can_refer_decl_in_current_unit_p): Be sure
1094 that var decl is available.
1096 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
1098 * tree-core.h (tree_decl_with_vis): Replace comdat_group by
1099 symtab_node pointer.
1100 * tree.c (copy_node_stat): Be sure to not copy symtab_node pointer.
1101 (find_decls_types_r): Do not walk COMDAT_GROUP.
1102 * tree.h (DECL_COMDAT_GROUP): Revamp to use decl_comdat_group.
1103 * varasm.c (make_decl_one_only): Use set_comdat_group;
1104 create node if needed.
1105 * ipa-inline-transform.c (save_inline_function_body): Update
1106 way we decl->symtab mapping.
1107 * symtab.c (symtab_hash, hash_node, eq_node
1108 symtab_insert_node_to_hashtable): Remove.
1109 (symtab_register_node): Update.
1110 (symtab_unregister_node): Update.
1111 (symtab_get_node): Reimplement as inline function.
1112 (symtab_add_to_same_comdat_group): Update.
1113 (symtab_dissolve_same_comdat_group_list): Update.
1114 (dump_symtab_base): Update.
1115 (verify_symtab_base): Update.
1116 (symtab_make_decl_local): Update.
1117 (fixup_same_cpp_alias_visibility): Update.
1118 (symtab_nonoverwritable_alias): Update.
1119 * cgraphclones.c (set_new_clone_decl_and_node_flags): Update.
1120 * ipa.c (update_visibility_by_resolution_info): UPdate.
1121 * bb-reorder.c: Include cgraph.h
1122 * lto-streamer-out.c (DFS_write_tree_body, hash_tree): Do not deal
1123 with comdat groups.
1124 * ipa-comdats.c (set_comdat_group, ipa_comdats): Update.
1125 * cgraph.c (cgraph_get_create_node): Update.
1126 * cgraph.h (struct symtab_node): Add get_comdat_group, set_comdat_group
1127 and comdat_group_.
1128 (symtab_get_node): Make inline.
1129 (symtab_insert_node_to_hashtable): Remove.
1130 (symtab_can_be_discarded): Update.
1131 (decl_comdat_group): New function.
1132 * tree-streamer-in.c (lto_input_ts_decl_with_vis_tree_pointers):
1133 Update.
1134 * lto-cgraph.c (lto_output_node, lto_output_varpool_node): Stream out
1135 comdat group name.
1136 (read_comdat_group): New function.
1137 (input_node, input_varpool_node): Use it.
1138 * trans-mem.c (ipa_tm_create_version_alias): Update code creating
1139 comdat groups.
1140 * mips.c (mips_start_unique_function): Likewise.
1141 (ix86_code_end): Likewise.
1142 (rs6000_code_end): Likweise.
1143 * tree-streamer-out.c (DECL_COMDAT_GROUP): Do not stream comdat group.
1145 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
1147 * gengtype-state.c (fatal_reading_state): Bring offline.
1148 * optabs.c (widening_optab_handler): Bring offline.
1149 * optabs.h (widening_optab_handler): Likewise.
1150 * final.c (get_attr_length_1): Likewise.
1152 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
1154 * sched-int.h (sd_iterator_cond): Manually tail recurse.
1156 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1158 * config/rs6000/440.md (ppc440-integer): Include shift without dot.
1159 (ppc440-compare): Include shift with dot.
1160 * config/rs6000/e300c2c3.md (ppce300c3_iu): Include shift without dot.
1161 * config/rs6000/e5500.md (e5500_sfx2): Include constant shift
1162 without dot.
1163 * config/rs6000/e6500.md (e6500_sfx): Exclude constant shift
1164 without dot.
1165 (e6500_sfx2): Include it.
1166 * config/rs6000/rs6000.md ( *zero_extend<mode>di2_internal1,
1167 *zero_extend<mode>di2_internal2, *zero_extend<mode>di2_internal3,
1168 *zero_extendsidi2_lfiwzx, andsi3_mc, andsi3_nomc,
1169 andsi3_internal0_nomc, extzvsi_internal, extzvdi_internal,
1170 *extzvdi_internal1, *extzvdi_internal2, rotlsi3, *rotlsi3_64,
1171 *rotlsi3_internal4, *rotlsi3_internal7le, *rotlsi3_internal7be,
1172 *rotlsi3_internal10le, *rotlsi3_internal10be, rlwinm,
1173 *lshiftrt_internal1le, *lshiftrt_internal1be,
1174 *lshiftrt_internal4le, *lshiftrt_internal4be, rotldi3,
1175 *rotldi3_internal4, *rotldi3_internal7le, *rotldi3_internal7be,
1176 *rotldi3_internal10le, *rotldi3_internal10be,
1177 *rotldi3_internal13le, *rotldi3_internal13be, *ashldi3_internal4,
1178 ashldi3_internal5, *ashldi3_internal6, *ashldi3_internal7,
1179 ashldi3_internal8, *ashldi3_internal9, anddi3_mc, anddi3_nomc,
1180 *anddi3_internal2_mc, *anddi3_internal3_mc, and 4 anonymous
1181 define_insns): Use type "shift" in the appropriate alternatives.
1183 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1185 * config/rs6000/rs6000.md (type): Add "logical". Delete
1186 "fast_compare".
1187 (dot): Adjust comment.
1188 (andsi3_mc, *andsi3_internal2_mc, *andsi3_internal3_mc,
1189 *andsi3_internal4, *andsi3_internal5_mc, *boolsi3_internal2,
1190 *boolsi3_internal3, *boolccsi3_internal2, *boolccsi3_internal3,
1191 anddi3_mc, *anddi3_internal2_mc, *anddi3_internal3_mc,
1192 *booldi3_internal2, *booldi3_internal3, *boolcdi3_internal2,
1193 *boolcdi3_internal3, *boolccdi3_internal2, *boolccdi3_internal3,
1194 *mov<mode>_internal2, and 10 anonymous define_insns): Use "logical".
1195 * config/rs6000/rs6000.c (rs6000_adjust_cost): Adjust.
1197 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
1198 * config/rs6000/440.md (ppc440-integer, ppc440-compare): Adjust.
1199 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare): Adjust.
1200 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
1201 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
1202 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare): Adjust.
1203 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
1204 * config/rs6000/8540.md (ppc8540_su): Adjust.
1205 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
1206 cell-cmp-microcoded): Adjust.
1207 * config/rs6000/e300c2c3.md (ppce300c3_cmp, ppce300c3_iu): Adjust.
1208 * config/rs6000/e500mc.md (e500mc_su): Adjust.
1209 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2): Adjust.
1210 * config/rs6000/e5500.md (e5500_sfx, e5500_sfx2): Adjust.
1211 * config/rs6000/e6500.md (e6500_sfx, e6500_sfx2): Adjust.
1212 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
1213 * config/rs6000/power4.md (power4-integer, power4-cmp): Adjust.
1214 * config/rs6000/power5.md (power5-integer, power5-cmp): Adjust.
1215 * config/rs6000/power6.md (power6-integer, power6-fast-compare):
1216 Adjust.
1217 * config/rs6000/power7.md (power7-integer, power7-cmp): Adjust.
1218 * config/rs6000/power8.md (power8-1cyc, power8-fast-compare):
1219 Adjust. Adjust comment.
1220 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
1221 * config/rs6000/titan.md (titan_fxu_adder, titan_fxu_alu): Adjust.
1223 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1225 * config/rs6000/rs6000.md (type): Add "add".
1226 (*add<mode>3_internal1, addsi3_high, *add<mode>3_internal2,
1227 *add<mode>3_internal3, *neg<mode>2_internal, and 5 anonymous
1228 define_insns): Use it.
1229 * config/rs6000/rs6000.c (rs6000_adjust_cost): Adjust.
1231 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
1232 * config/rs6000/440.md (ppc440-integer, ppc440-compare): Adjust.
1233 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare): Adjust.
1234 * config/rs6000/601.md (ppc601-integer): Adjust.
1235 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
1236 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
1237 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare): Adjust.
1238 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
1239 * config/rs6000/8540.md (ppc8540_su): Adjust.
1240 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
1241 cell-cmp-microcoded): Adjust.
1242 * config/rs6000/e300c2c3.md (ppce300c3_cmp, ppce300c3_iu): Adjust.
1243 * config/rs6000/e500mc.md (e500mc_su): Adjust.
1244 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2): Adjust.
1245 * config/rs6000/e5500.md (e5500_sfx, e5500_sfx2): Adjust.
1246 * config/rs6000/e6500.md (e6500_sfx, e6500_sfx2): Adjust.
1247 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
1248 * config/rs6000/power4.md (power4-integer, power4-cmp): Adjust.
1249 * config/rs6000/power5.md (power5-integer, power5-cmp): Adjust.
1250 * config/rs6000/power6.md (power6-integer, power6-fast-compare):
1251 Adjust.
1252 * config/rs6000/power7.md (power7-integer, power7-cmp): Adjust.
1253 * config/rs6000/power8.md (power8-1cyc, power8-fast-compare): Adjust.
1254 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
1255 * config/rs6000/titan.md (titan_fxu_adder, titan_fxu_alu): Adjust.
1257 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1259 * config/rs6000/rs6000.md (type): Delete "var_shift_rotate",
1260 "delayed_compare", "var_delayed_compare".
1261 (var_shift): New attribute.
1262 (cell_micro): Adjust.
1263 (*andsi3_internal2_mc, *andsi3_internal3_mc, *andsi3_internal4,
1264 *andsi3_internal5_mc, *extzvsi_internal1, *extzvsi_internal2,
1265 rotlsi3, *rotlsi3_64, *rotlsi3_internal2, *rotlsi3_internal3,
1266 *rotlsi3_internal4, *rotlsi3_internal5, *rotlsi3_internal6,
1267 *rotlsi3_internal8le, *rotlsi3_internal8be, *rotlsi3_internal9le,
1268 *rotlsi3_internal9be, *rotlsi3_internal10le, *rotlsi3_internal10be,
1269 *rotlsi3_internal11le, *rotlsi3_internal11be, *rotlsi3_internal12le,
1270 *rotlsi3_internal12be, ashlsi3, *ashlsi3_64, lshrsi3, *lshrsi3_64,
1271 *lshiftrt_internal2le, *lshiftrt_internal2be, *lshiftrt_internal3le,
1272 *lshiftrt_internal3be, *lshiftrt_internal5le, *lshiftrt_internal5be,
1273 *lshiftrt_internal5le, *lshiftrt_internal5be, ashrsi3, *ashrsi3_64,
1274 rotldi3, *rotldi3_internal2, *rotldi3_internal3, *rotldi3_internal4,
1275 *rotldi3_internal5, *rotldi3_internal6, *rotldi3_internal7le,
1276 *rotldi3_internal7be, *rotldi3_internal8le, *rotldi3_internal8be,
1277 *rotldi3_internal9le, *rotldi3_internal9be, *rotldi3_internal10le,
1278 *rotldi3_internal10be, *rotldi3_internal11le, *rotldi3_internal11be,
1279 *rotldi3_internal12le, *rotldi3_internal12be, *rotldi3_internal13le,
1280 *rotldi3_internal13be, *rotldi3_internal14le, *rotldi3_internal14be,
1281 *rotldi3_internal15le, *rotldi3_internal15be, *ashldi3_internal1,
1282 *ashldi3_internal2, *ashldi3_internal3, *lshrdi3_internal1,
1283 *lshrdi3_internal2, *lshrdi3_internal3, *ashrdi3_internal1,
1284 *ashrdi3_internal2, *ashrdi3_internal3, *anddi3_internal2_mc,
1285 *anddi3_internal3_mc, as well as 11 anonymous define_insns): Adjust.
1286 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
1287 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
1289 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
1290 * config/rs6000/440.md (ppc440-integer): Adjust.
1291 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare): Adjust.
1292 * config/rs6000/601.md (ppc601-integer, ppc601-compare): Adjust.
1293 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
1294 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
1295 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare): Adjust.
1296 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
1297 * config/rs6000/8540.md (ppc8540_su): Adjust.
1298 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
1299 cell-cmp-microcoded): Adjust.
1300 * config/rs6000/e300c2c3.md (ppce300c3_cmp): Adjust.
1301 * config/rs6000/e500mc.md (e500mc_su): Adjust.
1302 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2,
1303 e500mc64_delayed): Adjust.
1304 * config/rs6000/e5500.md (e5500_sfx, e5500_delayed): Adjust.
1305 * config/rs6000/e6500.md (e6500_sfx, e6500_delayed): Adjust.
1306 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
1307 * config/rs6000/power4.md (power4-integer, power4-compare): Adjust.
1308 * config/rs6000/power5.md (power5-integer, power5-compare): Adjust.
1309 * config/rs6000/power6.md (power6-shift, power6-var-rotate,
1310 power6-delayed-compare, power6-var-delayed-compare): Adjust.
1311 * config/rs6000/power7.md (power7-integer, power7-compare): Adjust.
1312 * config/rs6000/power8.md (power8-1cyc, power8-compare): Adjust.
1313 Adjust comment.
1314 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
1315 * config/rs6000/titan.md (titan_fxu_shift_and_rotate): Adjust.
1317 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1319 * config/rs6000/rs6000.md (type): Delete "idiv", "ldiv". Add "div".
1320 (bits): New mode_attr.
1321 (idiv_ldiv): Delete mode_attr.
1322 (udiv<mode>3, *div<mode>3, div<div_extend>_<mode>): Adjust.
1323 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
1324 rs6000_adjust_priority, is_nonpipeline_insn,
1325 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
1327 * config/rs6000/40x.md (ppc403-idiv): Adjust.
1328 * config/rs6000/440.md (ppc440-idiv): Adjust.
1329 * config/rs6000/476.md (ppc476-idiv): Adjust.
1330 * config/rs6000/601.md (ppc601-idiv): Adjust.
1331 * config/rs6000/603.md (ppc603-idiv): Adjust.
1332 * config/rs6000/6xx.md (ppc604-idiv, ppc620-idiv, ppc630-idiv,
1333 ppc620-ldiv): Adjust.
1334 * config/rs6000/7450.md (ppc7450-idiv): Adjust.
1335 * config/rs6000/7xx.md (ppc750-idiv): Adjust.
1336 * config/rs6000/8540.md (ppc8540_divide): Adjust.
1337 * config/rs6000/a2.md (ppca2-idiv, ppca2-ldiv): Adjust.
1338 * config/rs6000/cell.md (cell-idiv, cell-ldiv): Adjust.
1339 * config/rs6000/e300c2c3.md (ppce300c3_divide): Adjust.
1340 * config/rs6000/e500mc.md (e500mc_divide): Adjust.
1341 * config/rs6000/e500mc64.md (e500mc64_divide): Adjust.
1342 * config/rs6000/e5500.md (e5500_divide, e5500_divide_d): Adjust.
1343 * config/rs6000/e6500.md (e6500_divide, e6500_divide_d): Adjust.
1344 * config/rs6000/mpc.md (mpccore-idiv): Adjust.
1345 * config/rs6000/power4.md (power4-idiv, power4-ldiv): Adjust.
1346 * config/rs6000/power5.md (power5-idiv, power5-ldiv): Adjust.
1347 * config/rs6000/power6.md (power6-idiv, power6-ldiv): Adjust.
1348 * config/rs6000/power7.md (power7-idiv, power7-ldiv): Adjust.
1349 * config/rs6000/power8.md (power8-idiv, power8-ldiv): Adjust.
1350 * config/rs6000/rs64.md (rs64a-idiv, rs64a-ldiv): Adjust.
1351 * config/rs6000/titan.md (titan_fxu_div): Adjust.
1353 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1355 * config/rs6000/rs6000.md (type): Delete "insert_word",
1356 "insert_dword". Add "insert".
1357 (size): Update comment.
1358 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
1359 insn_must_be_first_in_group): Adjust.
1360 (insvsi_internal, *insvsi_internal1, *insvsi_internal2,
1361 *insvsi_internal3, *insvsi_internal4, *insvsi_internal5,
1362 *insvsi_internal6, insvdi_internal): Adjust.
1364 * config/rs6000/40x.md (ppc403-integer): Adjust.
1365 * config/rs6000/440.md (ppc440-integer): Adjust.
1366 * config/rs6000/476.md (ppc476-simple-integer): Adjust.
1367 * config/rs6000/601.md (ppc601-integer): Adjust.
1368 * config/rs6000/603.md (ppc603-integer): Adjust.
1369 * config/rs6000/6xx.md (ppc604-integer): Adjust.
1370 * config/rs6000/7450.md (ppc7450-integer): Adjust.
1371 * config/rs6000/7xx.md (ppc750-integer): Adjust.
1372 * config/rs6000/8540.md (ppc8540_su): Adjust.
1373 * config/rs6000/cell.md (cell-integer, cell-insert): Adjust.
1374 * config/rs6000/e300c2c3.md (ppce300c3_iu): Adjust.
1375 * config/rs6000/e500mc.md (e500mc_su): Adjust.
1376 * config/rs6000/e500mc64.md (e500mc64_su): Adjust.
1377 * config/rs6000/e5500.md (e5500_sfx): Adjust.
1378 * config/rs6000/e6500.md (e6500_sfx): Adjust.
1379 * config/rs6000/mpc.md (mpccore-integer): Adjust.
1380 * config/rs6000/power4.md (power4-integer, power4-insert): Adjust.
1381 * config/rs6000/power5.md (power5-integer, power5-insert): Adjust.
1382 * config/rs6000/power6.md (power6-insert, power6-insert-dword): Adjust.
1383 * config/rs6000/power7.md (power7-integer): Adjust.
1384 * config/rs6000/power8.md (power8-1cyc): Adjust.
1385 * config/rs6000/rs64.md (rs64a-integer): Adjust.
1386 * config/rs6000/titan.md (titan_fxu_shift_and_rotate): Adjust.
1388 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1390 * config/rs6000/rs6000.md (type): Add "mul". Delete "imul",
1391 "imul2", "imul3", "lmul", "imul_compare", "lmul_compare".
1392 (size): New attribute.
1393 (dot): New attribute.
1394 (cell_micro): Adjust.
1395 (mulsi3, *mulsi3_internal1, *mulsi3_internal2, mulsidi3,
1396 umulsidi3, smulsi3_highpart, umulsi3_highpart, muldi3,
1397 *muldi3_internal1, *muldi3_internal2, smuldi3_highpart,
1398 umuldi3_highpart): Adjust.
1399 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
1400 rs6000_adjust_priority, is_nonpipeline_insn,
1401 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
1403 * config/rs6000/40x.md (ppc403-imul, ppc405-imul, ppc405-imul2,
1404 ppc405-imul3): Adjust.
1405 * config/rs6000/440.md (ppc440-imul, ppc440-imul2): Adjust.
1406 * config/rs6000/476.md (ppc476-imul): Adjust.
1407 * config/rs6000/601.md (ppc601-imul): Adjust.
1408 * config/rs6000/603.md (ppc603-imul, ppc603-imul2): Adjust.
1409 * config/rs6000/6xx.md (ppc604-imul, ppc604e-imul, ppc620-imul,
1410 ppc620-imul2, ppc620-imul3, ppc620-lmul): Adjust.
1411 * config/rs6000/7450.md (ppc7450-imul, ppc7450-imul2): Adjust.
1412 * config/rs6000/7xx.md (ppc750-imul, ppc750-imul2, ppc750-imul3):
1413 Adjust.
1414 * config/rs6000/8540.md (ppc8540_multiply): Adjust.
1415 * config/rs6000/a2.md (ppca2-imul, ppca2-lmul): Adjust.
1416 * config/rs6000/cell.md (cell-lmul, cell-lmul-cmp, cell-imul23,
1417 cell-imul): Adjust.
1418 * config/rs6000/e300c2c3.md (ppce300c3_multiply): Adjust.
1419 * config/rs6000/e500mc.md (e500mc_multiply): Adjust.
1420 * config/rs6000/e500mc64.md (e500mc64_multiply): Adjust.
1421 * config/rs6000/e5500.md (e5500_multiply, e5500_multiply_i): Adjust.
1422 * config/rs6000/e6500.md (e6500_multiply, e6500_multiply_i): Adjust.
1423 * config/rs6000/mpc.md (mpccore-imul): Adjust.
1424 * config/rs6000/power4.md (power4-lmul-cmp, power4-imul-cmp,
1425 power4-lmul, power4-imul, power4-imul3): Adjust.
1426 * config/rs6000/power5.md (power5-lmul-cmp, power5-imul-cmp,
1427 power5-lmul, power5-imul, power5-imul3): Adjust.
1428 * config/rs6000/power6.md (power6-lmul-cmp, power6-imul-cmp,
1429 power6-lmul, power6-imul, power6-imul3): Adjust.
1430 * config/rs6000/power7.md (power7-mul, power7-mul-compare): Adjust.
1431 * config/rs6000/power8.md (power8-mul, power8-mul-compare): Adjust.
1433 * config/rs6000/rs64.md (rs64a-imul, rs64a-imul2, rs64a-imul3,
1434 rs64a-lmul): Adjust.
1435 * config/rs6000/titan.md (titan_imul): Adjust.
1437 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1439 * config/rs6000/rs6000.md (type): Add new value "halfmul".
1440 (*macchwc, *macchw, *macchwuc, *macchwu, *machhwc, *machhw,
1441 *machhwuc, *machhwu, *maclhwc, *maclhw, *maclhwuc, *maclhwu,
1442 *nmacchwc, *nmacchw, *nmachhwc, *nmachhw, *nmaclhwc, *nmaclhw,
1443 *mulchwc, *mulchw, *mulchwuc, *mulchwu, *mulhhwc, *mulhhw,
1444 *mulhhwuc, *mulhhwu, *mullhwc, *mullhw, *mullhwuc, *mullhwu): Use it.
1445 * config/rs6000/40x.md (ppc405-imul3): Add type halfmul.
1446 * config/rs6000/440.md (ppc440-imul2): Add type halfmul.
1447 * config/rs6000/476.md (ppc476-imul): Add type halfmul.
1448 * config/rs6000/titan.md: Delete nonsensical comment.
1449 (titan_imul): Add type imul3.
1450 (titan_mulhw): Remove type imul3; add type halfmul.
1452 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
1454 * config/rs6000/rs6000.md (type): Reorder, reformat.
1456 2014-05-23 Martin Jambor <mjambor@suse.cz>
1458 PR tree-optimization/53787
1459 * params.def (PARAM_IPA_MAX_AA_STEPS): New param.
1460 * ipa-prop.h (ipa_node_params): Rename uses_analysis_done to
1461 analysis_done, update all uses.
1462 * ipa-prop.c: Include domwalk.h
1463 (param_analysis_info): Removed.
1464 (param_aa_status): New type.
1465 (ipa_bb_info): Likewise.
1466 (func_body_info): Likewise.
1467 (ipa_get_bb_info): New function.
1468 (aa_overwalked): Likewise.
1469 (find_dominating_aa_status): Likewise.
1470 (parm_bb_aa_status_for_bb): Likewise.
1471 (parm_preserved_before_stmt_p): Changed to use new param AA info.
1472 (load_from_unmodified_param): Accept func_body_info as a parameter
1473 instead of parms_ainfo.
1474 (parm_ref_data_preserved_p): Changed to use new param AA info.
1475 (parm_ref_data_pass_through_p): Likewise.
1476 (ipa_load_from_parm_agg_1): Likewise. Update callers.
1477 (compute_complex_assign_jump_func): Changed to use new param AA info.
1478 (compute_complex_ancestor_jump_func): Likewise.
1479 (ipa_compute_jump_functions_for_edge): Likewise.
1480 (ipa_compute_jump_functions): Removed.
1481 (ipa_compute_jump_functions_for_bb): New function.
1482 (ipa_analyze_indirect_call_uses): Likewise, moved variable
1483 declarations down.
1484 (ipa_analyze_virtual_call_uses): Accept func_body_info instead of node
1485 and info, moved variable declarations down.
1486 (ipa_analyze_call_uses): Accept and pass on func_body_info instead of
1487 node and info.
1488 (ipa_analyze_stmt_uses): Likewise.
1489 (ipa_analyze_params_uses): Removed.
1490 (ipa_analyze_params_uses_in_bb): New function.
1491 (ipa_analyze_controlled_uses): Likewise.
1492 (free_ipa_bb_info): Likewise.
1493 (analysis_dom_walker): New class.
1494 (ipa_analyze_node): Handle node-specific forbidden analysis,
1495 initialize and free func_body_info, use dominator walker.
1496 (ipcp_modif_dom_walker): New class.
1497 (ipcp_transform_function): Create and free func_body_info, use
1498 ipcp_modif_dom_walker, moved a lot of functionality there.
1500 2014-05-23 Marek Polacek <polacek@redhat.com>
1501 Jakub Jelinek <jakub@redhat.com>
1503 * builtins.def: Change SANITIZE_FLOAT_DIVIDE to SANITIZE_NONDEFAULT.
1504 * gcc.c (sanitize_spec_function): Likewise.
1505 * convert.c (convert_to_integer): Include "ubsan.h". Add
1506 floating-point to integer instrumentation.
1507 * doc/invoke.texi: Document -fsanitize=float-cast-overflow.
1508 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_CAST and
1509 SANITIZE_NONDEFAULT.
1510 * opts.c (common_handle_option): Handle -fsanitize=float-cast-overflow.
1511 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_FLOAT_CAST_OVERFLOW,
1512 BUILT_IN_UBSAN_HANDLE_FLOAT_CAST_OVERFLOW_ABORT): Add.
1513 * ubsan.c: Include "realmpfr.h" and "dfp.h".
1514 (get_ubsan_type_info_for_type): Handle REAL_TYPEs.
1515 (ubsan_type_descriptor): Set tkind to 0xffff for types other than
1516 float/double/long double.
1517 (ubsan_instrument_float_cast): New function.
1518 * ubsan.h (ubsan_instrument_float_cast): Declare.
1520 2014-05-23 Jiong Wang <jiong.wang@arm.com>
1522 * config/aarch64/predicates.md (aarch64_call_insn_operand): New
1523 predicate.
1524 * config/aarch64/constraints.md ("Ucs", "Usf"): New constraints.
1525 * config/aarch64/aarch64.md (*sibcall_insn, *sibcall_value_insn):
1526 Adjust for tailcalling through registers.
1527 * config/aarch64/aarch64.h (enum reg_class): New caller save
1528 register class.
1529 (REG_CLASS_NAMES): Likewise.
1530 (REG_CLASS_CONTENTS): Likewise.
1531 * config/aarch64/aarch64.c (aarch64_function_ok_for_sibcall):
1532 Allow tailcalling without decls.
1534 2014-05-23 Thomas Schwinge <thomas@codesourcery.com>
1536 * gimplify.c (omp_notice_variable) <case OMP_CLAUSE_DEFAULT_NONE>:
1537 Rewrite check for ORT_PARALLEL and ORT_COMBINED_PARALLEL.
1539 * omp-low.c (expand_omp_for_static_chunk): Rename variable si to
1540 gsi, and variables v_* to v*.
1542 2014-05-23 Eric Botcazou <ebotcazou@adacore.com>
1544 * varasm.c (output_constructor_bitfield): Fix thinkos in latest change.
1546 2014-05-23 Thomas Schwinge <thomas@codesourcery.com>
1548 * gimple.h (enum gf_mask): Add and use GF_OMP_FOR_SIMD.
1549 * omp-low.c: Update accordingly.
1551 * gimple.h (enum gf_mask): Rewrite "<< 0" shift expressions used
1552 for GF_OMP_FOR_KIND_MASK, GF_OMP_FOR_KIND_FOR,
1553 GF_OMP_FOR_KIND_DISTRIBUTE, GF_OMP_FOR_KIND_SIMD,
1554 GF_OMP_FOR_KIND_CILKSIMD, GF_OMP_TARGET_KIND_MASK,
1555 GF_OMP_TARGET_KIND_REGION, GF_OMP_TARGET_KIND_DATA,
1556 GF_OMP_TARGET_KIND_UPDATE.
1558 * gimplify.c (omp_notice_variable) <case OMP_CLAUSE_DEFAULT_NONE>:
1559 Explicitly enumerate the expected region types.
1561 2014-05-23 Paul Eggert <eggert@cs.ucla.edu>
1563 PR other/56955
1564 * doc/extend.texi (Function Attributes): Fix __attribute__ ((malloc))
1565 documentation; the old documentation didn't clearly state the
1566 constraints on the contents of the pointed-to storage.
1568 2014-05-23 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
1570 Fix bootstrap error on ia64
1571 * config/ia64/ia64.c (ia64_first_cycle_multipass_dfa_lookahead_guard):
1572 Return default value.
1574 2014-05-23 Thomas Preud'homme <thomas.preudhomme@arm.com>
1576 PR tree-optimization/54733
1577 * tree-ssa-math-opts.c (nop_stats): New "bswap_stats" structure.
1578 (CMPNOP): Define.
1579 (find_bswap_or_nop_load): New.
1580 (find_bswap_1): Renamed to ...
1581 (find_bswap_or_nop_1): This. Also add support for memory source.
1582 (find_bswap): Renamed to ...
1583 (find_bswap_or_nop): This. Also add support for memory source and
1584 detection of bitwise operations equivalent to load in host endianness.
1585 (execute_optimize_bswap): Likewise. Also move its leading comment back
1586 in place and split statement transformation into ...
1587 (bswap_replace): This.
1589 2014-05-22 Vladimir Makarov <vmakarov@redhat.com>
1591 PR rtl-optimization/61215
1592 * lra-elelimination.c (lra_eliminate_regs_1): Don't use
1593 simplify_gen_subreg until final substitution.
1595 2014-05-23 Alan Modra <amodra@gmail.com>
1597 PR target/61231
1598 * config/rs6000/rs6000.c (mem_operand_gpr): Handle SImode.
1599 * config/rs6000/rs6000.md (extendsidi2_lfiwax, extendsidi2_nocell):
1600 Use "Y" constraint rather than "m".
1602 2014-05-23 Kugan Vivekanandarajah <kuganv@linaro.org>
1604 * config/aarch64/aarch64.c (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New
1605 define.
1606 * config/aarch64/aarch64-protos.h (aarch64_atomic_assign_expand_fenv):
1607 New function declaration.
1608 * config/aarch64/aarch64-builtins.c (aarch64_builtins) : Add
1609 AARCH64_BUILTIN_GET_FPCR, AARCH64_BUILTIN_SET_FPCR.
1610 AARCH64_BUILTIN_GET_FPSR and AARCH64_BUILTIN_SET_FPSR.
1611 (aarch64_init_builtins) : Initialize builtins
1612 __builtins_aarch64_set_fpcr, __builtins_aarch64_get_fpcr.
1613 __builtins_aarch64_set_fpsr and __builtins_aarch64_get_fpsr.
1614 (aarch64_expand_builtin) : Expand builtins __builtins_aarch64_set_fpcr
1615 __builtins_aarch64_get_fpcr, __builtins_aarch64_get_fpsr,
1616 and __builtins_aarch64_set_fpsr.
1617 (aarch64_atomic_assign_expand_fenv): New function.
1618 * config/aarch64/aarch64.md (set_fpcr): New pattern.
1619 (get_fpcr) : Likewise.
1620 (set_fpsr) : Likewise.
1621 (get_fpsr) : Likewise.
1622 (unspecv): Add UNSPECV_GET_FPCR and UNSPECV_SET_FPCR, UNSPECV_GET_FPSR
1623 and UNSPECV_SET_FPSR.
1624 * doc/extend.texi (AARCH64 Built-in Functions) : Document
1625 __builtins_aarch64_set_fpcr, __builtins_aarch64_get_fpcr.
1626 __builtins_aarch64_set_fpsr and __builtins_aarch64_get_fpsr.
1628 2014-05-22 Vladimir Makarov <vmakarov@redhat.com>
1630 PR rtl-optimization/60969
1631 * ira-costs.c (record_reg_classes): Process NO_REGS for matching
1632 constraints. Set up mem cost for NO_REGS case.
1634 2014-05-22 Thomas Schwinge <thomas@codesourcery.com>
1636 * builtin-types.def: Simplify examples for DEF_FUNCTION_TYPE_*.
1638 2012-05-22 Bernd Schmidt <bernds@codesourcery.com>
1640 * config/darwin.c: Include "lto-section-names.h".
1641 (LTO_SEGMENT_NAME): Don't define.
1642 * config/i386/winnt.c: Include "lto-section-names.h".
1643 * lto-streamer.c: Include "lto-section-names.h".
1644 * lto-streamer.h (LTO_SECTION_NAME_PREFIX): Don't define.
1645 * lto-wrapper.c: Include "lto-section-names.h".
1646 (LTO_SECTION_NAME_PREFIX): Don't define.
1647 * lto-section-names.h: New file.
1648 * cgraphunit.c: Include "lto-section-names.h".
1650 2014-05-22 Peter Bergner <bergner@vnet.ibm.com>
1652 * config/rs6000/htm.md (ttest): Use correct shift value to get CR0.
1654 2014-05-22 Richard Earnshaw <rearnsha@arm.com>
1656 PR target/61208
1657 * arm.md (arm_cmpdi_unsigned): Fix length calculation for Thumb2.
1659 2014-05-22 Nick Clifton <nickc@redhat.com>
1661 * config/msp430/msp430.h (ASM_SPEC): Add spaces after inserted options.
1663 2014-05-22 Eric Botcazou <ebotcazou@adacore.com>
1665 * tree-ssa-forwprop.c (associate_plusminus): Extend (T)(P + A) - (T)P
1666 -> (T)A transformation to integer types.
1668 2014-05-22 Teresa Johnson <tejohnson@google.com>
1670 * gcov-io.c (gcov_position): Use gcov_nonruntime_assert.
1671 (gcov_is_error): Remove gcc_assert from IN_LIBGCOV code.
1672 (gcov_rewrite): Use gcov_nonruntime_assert.
1673 (gcov_open): Ditto.
1674 (gcov_write_words): Ditto.
1675 (gcov_write_length): Ditto.
1676 (gcov_read_words): Use gcov_nonruntime_assert, and remove
1677 gcc_assert from IN_LIBGCOV code.
1678 (gcov_read_summary): Use gcov_error to flag profile corruption.
1679 (gcov_sync): Use gcov_nonruntime_assert.
1680 (gcov_seek): Remove gcc_assert from IN_LIBGCOV code.
1681 (gcov_histo_index): Use gcov_nonruntime_assert.
1682 (static void gcov_histogram_merge): Ditto.
1683 (compute_working_sets): Ditto.
1684 * gcov-io.h (gcov_nonruntime_assert): Define.
1685 (gcov_error): Define for !IN_LIBGCOV
1687 2014-05-22 Richard Biener <rguenther@suse.de>
1689 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
1690 BUILT_IN_REALLOC like BUILT_IN_STRDUP.
1691 (call_may_clobber_ref_p_1): Handle BUILT_IN_REALLOC as allocation
1692 and deallocation site.
1693 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
1694 Handle BUILT_IN_REALLOC similar to BUILT_IN_STRDUP with also
1695 passing through the incoming points-to set.
1696 (handle_lhs_call): Use flags argument instead of recomputing it.
1697 (find_func_aliases_for_call): Call handle_lhs_call with proper
1698 call return flags.
1700 2014-05-22 Jakub Jelinek <jakub@redhat.com>
1702 * tree-streamer-in.c (unpack_ts_real_cst_value_fields): Make sure
1703 all padding bits in REAL_VALUE_TYPE are cleared.
1705 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
1707 Cleanup and improve multipass_dfa_lookahead_guard
1708 * config/i386/i386.c (core2i7_first_cycle_multipass_filter_ready_try,)
1709 (core2i7_first_cycle_multipass_begin,)
1710 (core2i7_first_cycle_multipass_issue,)
1711 (core2i7_first_cycle_multipass_backtrack): Update signature.
1712 * config/ia64/ia64.c
1713 (ia64_first_cycle_multipass_dfa_lookahead_guard_spec): Remove.
1714 (ia64_first_cycle_multipass_dfa_lookahead_guard): Update signature.
1715 (TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC): Remove
1716 hook definition.
1717 (ia64_first_cycle_multipass_dfa_lookahead_guard): Merge logic from
1718 ia64_first_cycle_multipass_dfa_lookahead_guard_spec. Update return
1719 values.
1720 * config/rs6000/rs6000.c (rs6000_use_sched_lookahead_guard): Update
1721 return values.
1722 * doc/tm.texi: Regenerate.
1723 * doc/tm.texi.in
1724 (TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC): Remove.
1725 * haifa-sched.c (ready_try): Make signed to allow negative values.
1726 (rebug_ready_list_1): Update.
1727 (choose_ready): Simplify.
1728 (sched_extend_ready_list): Update.
1730 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
1732 Remove IA64 speculation tweaking flags
1733 * config/ia64/ia64.c (ia64_set_sched_flags): Delete handling of
1734 speculation tuning flags.
1735 (msched-prefer-non-data-spec-insns,)
1736 (msched-prefer-non-control-spec-insns): Obsolete options.
1737 * haifa-sched.c (choose_ready): Remove handling of
1738 PREFER_NON_CONTROL_SPEC and PREFER_NON_DATA_SPEC.
1739 * sched-int.h (enum SPEC_SCHED_FLAGS): Remove PREFER_NON_CONTROL_SPEC
1740 and PREFER_NON_DATA_SPEC.
1741 * sel-sched.c (process_spec_exprs): Remove handling of
1742 PREFER_NON_CONTROL_SPEC and PREFER_NON_DATA_SPEC.
1744 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
1746 Improve scheduling debug output
1747 * haifa-sched.c (debug_ready_list): Remove unnecessary prototype.
1748 (advance_one_cycle): Update.
1749 (schedule_insn, queue_to_ready): Add debug printouts.
1750 (debug_ready_list_1): New static function.
1751 (debug_ready_list): Update.
1752 (max_issue): Add debug printouts.
1753 (dump_insn_stream): New static function.
1754 (schedule_block): Use it. Also better indent printouts.
1756 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
1758 Fix sched_insn debug counter
1759 * haifa-sched.c (schedule_insn): Update.
1760 (struct haifa_saved_data): Add nonscheduled_insns_begin.
1761 (save_backtrack_point, restore_backtrack_point): Update.
1762 (first_nonscheduled_insn): New static function.
1763 (queue_to_ready, choose_ready): Use it.
1764 (schedule_block): Init nonscheduled_insns_begin.
1765 (sched_emit_insn): Update.
1768 2014-05-22 Kugan Vivekanandarajah <kuganv@linaro.org>
1770 * config/aarch64/aarch64.c (aarch64_regno_regclass) : Change CORE_REGS
1771 to GENERAL_REGS.
1772 (aarch64_secondary_reload) : LikeWise.
1773 (aarch64_class_max_nregs) : Remove CORE_REGS.
1774 * config/aarch64/aarch64.h (enum reg_class) : Remove CORE_REGS.
1775 (REG_CLASS_NAMES) : Likewise.
1776 (REG_CLASS_CONTENTS) : LikeWise.
1777 (INDEX_REG_CLASS) : Change CORE_REGS to GENERAL_REGS.
1779 2014-05-21 Guozhi Wei <carrot@google.com>
1781 PR target/61202
1782 * config/aarch64/arm_neon.h (vqdmulh_n_s16): Change the last operand's
1783 constraint.
1784 (vqdmulhq_n_s16): Likewise.
1786 2014-05-21 Segher Boessenkool <segher@kernel.crashing.org>
1788 * config/rs6000/predicates.md (update_indexed_address_mem): Delete.
1790 2014-05-21 Marek Polacek <polacek@redhat.com>
1792 PR sanitizer/61272
1793 * ubsan.c (is_ubsan_builtin_p): Turn assert into a condition.
1795 2014-05-21 Martin Jambor <mjambor@suse.cz>
1797 * doc/invoke.texi (Optimize Options): Document parameters
1798 ipa-cp-eval-threshold, ipa-max-agg-items, ipa-cp-loop-hint-bonus and
1799 ipa-cp-array-index-hint-bonus.
1801 2014-05-21 Mark Wielaard <mjw@redhat.com>
1803 PR debug/16063
1804 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_type if DWARF
1805 version >= 3 or not strict DWARF.
1806 * langhooks.h (struct lang_hooks_for_types): Add
1807 enum_underlying_base_type.
1808 * langhooks.c (lhd_enum_underlying_base_type): New function.
1809 * gcc/langhooks.h (struct lang_hooks_for_types): Add
1810 enum_underlying_base_type.
1811 * langhooks-def.h (lhd_enum_underlying_base_type): New declaration.
1812 (LANG_HOOKS_ENUM_UNDERLYING_BASE_TYPE): New define.
1813 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add new lang hook.
1815 2014-05-21 Richard Biener <rguenther@suse.de>
1817 * doc/invoke.texi (-flto-partition=): Document one and none algorithms.
1819 2014-05-21 John Marino <gnugcc@marino.st>
1821 * config.gcc (*-*-dragonfly*): New target.
1822 * configure.ac: Detect dl_iterate_phdr (*freebsd*, *dragonfly*).
1823 * configure: Regenerate.
1824 * config/dragonfly-stdint.h: New.
1825 * config/dragonfly.h: New.
1826 * config/dragonfly.opt: New.
1827 * config/i386/dragonfly.h: New.
1828 * ginclude/stddef.h: Detect _PTRDIFF_T_DECLARED for DragonFly.
1830 2014-05-21 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1832 * tree.def (VOID_CST): New.
1833 * tree-core.h (TI_VOID): New.
1834 * tree.h (void_node): New.
1835 * tree.c (tree_node_structure_for_code, tree_code_size)
1836 (iterative_hash_expr): Handle VOID_CST.
1837 (build_common_tree_nodes): Initialize void_node.
1839 2014-05-21 Bernd Schmidt <bernds@codesourcery.com>
1841 * reload1.c (remove_init_insns, will_delete_init_insn_p): New static
1842 functions.
1843 (reload, calculate_needs_all_insns, reload_as_needed): Use them.
1845 * config/bfin/bfin.c (split_load_immediate): Use gen_int_mode in a few
1846 more places.
1848 * cfgrtl.c (cfg_layout_initialize): Weaken assert to only trigger if
1849 flag_reorder_blocks_and_partition.
1850 * hw-doloop.c (reorg_loops): Avoid reordering if that flag is set.
1852 2014-05-21 Oleg Endo <olegendo@gcc.gnu.org>
1854 PR target/54236
1855 * config/sh/sh.md (*addc_r_1): Rename to addc_t_r. Remove empty
1856 constraints.
1857 (*addc_r_t): Add new insn_and_split.
1859 2014-05-21 Jakub Jelinek <jakub@redhat.com>
1861 PR middle-end/61252
1862 * omp-low.c (handle_simd_reference): New function.
1863 (lower_rec_input_clauses): Use it. Defer adding reference
1864 initialization even for reduction without placeholder if in simd,
1865 handle it properly later on.
1867 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
1869 PR tree-optimization/60899
1870 * gimple-fold.c (can_refer_decl_in_current_unit_p): Cleanup;
1871 assume all static symbols will have definition wile parsing and
1872 check the do have definition later in compilation; check that
1873 variable referring symbol will be output before concluding that
1874 reference is safe; be conservative for referring local statics;
1875 be more precise about when comdat is output in other partition.
1877 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
1879 PR bootstrap/60984
1880 * ipa-inline-transform.c (inline_call): Use add CALLEE_REMOVED
1881 parameter.
1882 * ipa-inline.c (inline_to_all_callers): If callee was removed; return.
1883 (ipa_inline): Loop inline_to_all_callers until no more aliases
1884 are removed.
1886 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
1888 * ipa.c (ipa_discover_readonly_nonaddressable_var): Fix dumping;
1889 set writeonly flag only for vars actually written to.
1891 2014-05-20 Dehao Chen <dehao@google.com>
1893 * ipa-inline-transform.c (clone_inlined_nodes): Use min of edge count
1894 and callee count to get clone count.
1895 * tree-inline.c (expand_call_inline): Use callee count instead of bb
1896 count in copy_body.
1898 2014-05-20 Richard Sandiford <rdsandiford@googlemail.com>
1900 PR rtl-optimization/61243
1901 * emit-rtl.c (emit_copy_of_insn_after): Copy CROSSING_JUMP_P.
1903 2014-05-20 Xinliang David Li <davidxl@google.com>
1905 * cgraphunit.c (walk_polymorphic_call_targets): Add
1906 dbgcnt and fopt-info support.
1907 * ipa-prop.c (ipa_make_edge_direct_to_target): Ditto.
1908 * ipa-devirt.c (ipa_devirt): Ditto.
1909 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children): Ditto.
1910 * ipa.c (walk_polymorphic_call_targets): Ditto.
1911 * gimple-fold.c (fold_gimple_assign): Ditto.
1912 (gimple_fold_call): Ditto.
1913 * dbgcnt.def: New counter.
1915 2014-05-20 DJ Delorie <dj@redhat.com>
1917 * config/msp430/msp430.md (split): Don't allow subregs when
1918 splitting SImode adds.
1919 (andneghi): Fix subtraction logic.
1920 * config/msp430/predicates.md (msp430_nonsubreg_or_imm_operand): New.
1922 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
1924 * tree.h (DECL_ONE_ONLY): Return true only for externally visible
1925 symbols.
1926 * except.c (switch_to_exception_section, resolve_unique_section,
1927 get_named_text_section, default_function_rodata_section,
1928 align_variable, get_block_for_decl, default_section_type_flags):
1929 Use DECL_COMDAT_GROUP instead of DECL_ONE_ONLY.
1930 * symtab.c (symtab_add_to_same_comdat_group,
1931 symtab_make_decl_local, fixup_same_cpp_alias_visibility,
1932 symtab_nonoverwritable_alias, symtab_get_symbol_partitioning_class):
1933 Likewise.
1934 * cgraphclones.c (cgraph_create_virtual_clone): Likewise.
1935 * bb-reorder.c (pass_partition_blocks::gate): Likewise.
1936 * config/c6x/c6x.c (c6x_elf_unique_section): Likewise.
1937 (c6x_function_in_section_p): Likewise.
1938 * config/darwin.c (machopic_select_section): Likewise.
1939 * config/arm/arm.c (arm_function_in_section_p): Likewise.
1940 * config/mips/mips.c (mips_function_rodata_section): Likewise.
1941 * config/mep/mep.c (mep_select_section): LIkewise.
1942 * config/i386/i386.c (x86_64_elf_unique_section): Likewise.
1944 2014-05-20 Eric Botcazou <ebotcazou@adacore.com>
1946 * tree-ssa-dom.c (hashable_expr_equal_p) <EXPR_CALL>: Also compare the
1947 EH region of calls to pure functions that can throw an exception.
1948 * tree-ssa-sccvn.c (vn_reference_eq): Remove duplicated test.
1949 (copy_reference_ops_from_call): Also copy the EH region of the call if
1950 it can throw an exception.
1952 2014-05-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1954 * simplify-rtx.c (simplify_binary_operation_1): Optimize case of
1955 nested VEC_SELECTs that are inverses of each other.
1957 2014-05-20 Richard Biener <rguenther@suse.de>
1959 * tree-ssa-sccvn.c (process_scc): Dump SCC here, when iterating,
1960 (extract_and_process_scc_for_name): not here.
1961 (cond_dom_walker::before_dom_children): Only process
1962 stmts that end the BB in interesting ways.
1963 (run_scc_vn): Mark param uses as visited.
1965 2014-05-20 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1967 * config/arm/arm.md (arith_shiftsi): Do not predicate for
1968 arm_restrict_it.
1970 2014-05-20 Nick Clifton <nickc@redhat.com>
1972 * config/msp430/msp430.c (TARGET_GIMPLIFY_VA_ARG_EXPR): Define.
1973 (msp430_gimplify_va_arg_expr): New function.
1974 (msp430_print_operand): Handle (CONST (ZERO_EXTRACT)).
1976 * config/msp430/msp430.md (zero_extendpsisi2): Use + constraint on
1977 operand 0 in order to prevent confusion about the number of
1978 registers involved.
1980 2014-05-20 Richard Biener <rguenther@suse.de>
1982 PR tree-optimization/61221
1983 * tree-ssa-pre.c (el_to_update): Remove.
1984 (eliminate_dom_walker::before_dom_children): Handle released
1985 VDEFs by value-numbering them to the associated VUSE. Update
1986 stmt immediately for substituted call address.
1987 (eliminate): Remove delayed stmt updating code.
1988 * tree-ssa-sccvn.c (vuse_ssa_val): New function valueizing
1989 possibly late re-numbered vuses.
1990 (vn_reference_lookup_2): Adjust.
1991 (vn_reference_lookup_pieces): Likewise.
1992 (vn_reference_lookup): Likewise.
1994 2014-05-20 Richard Biener <rguenther@suse.de>
1996 * config.gcc: Remove need_64bit_hwint.
1997 * configure.ac: Do not define NEED_64BIT_HOST_WIDE_INT.
1998 * hwint.h: Do not check NEED_64BIT_HOST_WIDE_INT but assume
1999 it to be true.
2000 * config.in: Regenerate.
2001 * configure: Likewise.
2003 2014-05-19 David Wohlferd <dw@LimeGreenSocks.com>
2005 * doc/extend.texi: Create Label Attributes section,
2006 move all label attributes into it and reference it.
2008 2014-05-19 Richard Earnshaw <rearnsha@arm.com>
2010 * arm.c (thumb1_reorg): When scanning backwards skip anything
2011 that's not a proper insn.
2013 2014-05-19 Richard Biener <rguenther@suse.de>
2015 PR tree-optimization/61221
2016 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children):
2017 Do nothing for unreachable blocks.
2018 * tree-ssa-sccvn.c (cond_dom_walker::before_dom_children):
2019 Improve unreachability detection.
2021 2014-05-19 Richard Biener <rguenther@suse.de>
2023 PR tree-optimization/61209
2024 * tree-ssa-sccvn.c (visit_phi): Avoid setting expr to VN_TOP.
2026 2014-05-19 Nick Clifton <nickc@redhat.com>
2028 * except.c (init_eh): Fix computation of builtin setjmp buffer
2029 size to allow for targets where POINTER_SIZE > BITS_PER_WORD.
2031 2014-05-19 Richard Biener <rguenther@suse.de>
2033 PR tree-optimization/61184
2034 * tree-vrp.c (is_negative_overflow_infinity): Use
2035 TREE_OVERFLOW_P and do that check first.
2036 (is_positive_overflow_infinity): Likewise.
2037 (is_overflow_infinity): Likewise.
2038 (vrp_operand_equal_p): Properly treat operands with
2039 differing overflow as not equal.
2041 2014-05-19 Bernd Schmidt <bernds@codesourcery.com>
2043 * simplify-rtx.c (simplify_unary_operation_1): Use CONST_INT_P in
2044 shift simplification where it was intended.
2046 2014-05-19 Christian Bruel <christian.bruel@st.com>
2048 PR target/61195
2049 * config/sh/sh.md (movsf_ie): Unset fp_mode for fmov.
2051 2014-05-19 Richard Sandiford <r.sandiford@uk.ibm.com>
2053 PR target/61084
2054 * config/sparc/sparc.c (sparc_fold_builtin): Use widest_int rather
2055 than wide_int.
2057 2014-05-19 Richard Sandiford <rdsandiford@googlemail.com>
2059 * reg-notes.def (CROSSING_JUMP): Likewise.
2060 * rtl.h (rtx_def): Update comment for jump flag.
2061 (CROSSING_JUMP_P): Define.
2062 * cfgcleanup.c (try_forward_edges, try_optimize_cfg): Use it instead
2063 of a REG_CROSSING_JUMP note.
2064 * cfghooks.c (tidy_fallthru_edges): Likewise.
2065 * cfgrtl.c (fixup_partition_crossing, rtl_verify_edges): Likewise.
2066 * emit-rtl.c (try_split): Likewise.
2067 * haifa-sched.c (sched_create_recovery_edges): Likewise.
2068 * ifcvt.c (find_if_case_1, find_if_case_2): Likewise.
2069 * jump.c (redirect_jump_2): Likewise.
2070 * reorg.c (follow_jumps, fill_slots_from_thread): Likewise.
2071 (relax_delay_slots): Likewise.
2072 * config/arc/arc.md (jump_i, cbranchsi4_scratch, *bbit): Likewise.
2073 (bbit_di): Likewise.
2074 * config/arc/arc.c (arc_reorg, arc_can_follow_jump): Likewise.
2075 * config/sh/sh.md (jump_compact): Likewise.
2076 * bb-reorder.c (rotate_loop): Likewise.
2077 (pass_duplicate_computed_gotos::execute): Likewise.
2078 (add_reg_crossing_jump_notes): Rename to...
2079 (update_crossing_jump_flags): ...this.
2080 (pass_partition_blocks::execute): Update accordingly.
2082 2014-05-19 Richard Sandiford <rdsandiford@googlemail.com>
2084 * tree.h: Remove extraneous template <>.
2086 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
2088 * ipa.c (symtab_remove_unreachable_nodes): Remove
2089 symbol from comdat group if its body was eliminated.
2090 (comdat_can_be_unshared_p_1): Static symbols can always be privatized.
2091 * symtab.c (symtab_remove_from_same_comdat_group): Break out from ...
2092 (symtab_unregister_node): ... this one.
2093 (verify_symtab_base): More strict checking of comdats.
2094 * cgraph.h (symtab_remove_from_same_comdat_group): Declare.
2096 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
2098 * tree-pass.h (make_pass_ipa_comdats): New pass.
2099 * timevar.def (TV_IPA_COMDATS): New timevar.
2100 * passes.def (pass_ipa_comdats): Add.
2101 * Makefile.in (OBJS): Add ipa-comdats.o
2102 * ipa-comdats.c: New file.
2104 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
2106 * ipa.c (update_visibility_by_resolution_info): New function.
2107 (function_and_variable_visibility): Use it.
2109 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
2111 * cgraph.h (symtab_first_defined_symbol, symtab_next_defined_symbol):
2112 New functions.
2113 (FOR_EACH_DEFINED_SYMBOL): New macro.
2114 (varpool_first_static_initializer, varpool_next_static_initializer,
2115 varpool_first_defined_variable, varpool_next_defined_variable):
2116 Fix comments.
2117 (symtab_in_same_comdat_p): Correctly deal with inline functions.
2119 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
2121 * ggc-page.c (ggc_handle_finalizers): Add comment.
2123 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
2125 * ggc-common.c (ggc_internal_cleared_alloc): Adjust.
2126 * ggc-none.c (ggc_internal_alloc): Assert if a finalizer is passed.
2127 (ggc_internal_cleared_alloc): Likewise.
2128 * ggc-page.c (finalizer): New class.
2129 (vec_finalizer): Likewise.
2130 (globals::finalizers): New member.
2131 (globals::vec_finalizers): Likewise.
2132 (ggc_internal_alloc): Record the finalizer if any for the block being
2133 allocated.
2134 (ggc_handle_finalizers): New function.
2135 (ggc_collect): Call ggc_handle_finalizers.
2136 * ggc.h (ggc_internal_alloc): Add arguments to allow installing a
2137 finalizer.
2138 (ggc_internal_cleared_alloc): Likewise.
2139 (finalize): New function.
2140 (need_finalization_p): Likewise.
2141 (ggc_alloc): Install the type's destructor as the finalizer if it
2142 might do something.
2143 (ggc_cleared_alloc): Likewise.
2144 (ggc_vec_alloc): Likewise.
2145 (ggc_cleared_vec_alloc): Likewise.
2147 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
2149 * ggc.h (ggc_alloc_cleared_simd_clone_stat): Remove function.
2151 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
2153 * alias.c (record_alias_subset): Adjust.
2154 * bitmap.c (bitmap_element_allocate): Likewise.
2155 (bitmap_gc_alloc_stat): Likewise.
2156 * cfg.c (init_flow): Likewise.
2157 (alloc_block): Likewise.
2158 (unchecked_make_edge): Likewise.
2159 * cfgloop.c (alloc_loop): Likewise.
2160 (flow_loops_find): Likewise.
2161 (rescan_loop_exit): Likewise.
2162 * cfgrtl.c (init_rtl_bb_info): Likewise.
2163 * cgraph.c (insert_new_cgraph_node_version): Likewise.
2164 (cgraph_allocate_node): Likewise.
2165 (cgraph_create_edge_1): Likewise.
2166 (cgraph_allocate_init_indirect_info): Likewise.
2167 * cgraphclones.c (cgraph_clone_edge): Likewise.
2168 * cgraphunit.c (add_asm_node): Likewise.
2169 (init_lowered_empty_function): Likewise.
2170 * config/aarch64/aarch64.c (aarch64_init_machine_status): Likewise.
2171 * config/alpha/alpha.c (alpha_init_machine_status): Likewise.
2172 (alpha_use_linkage): Likewise.
2173 * config/arc/arc.c (arc_init_machine_status): Likewise.
2174 * config/arm/arm.c (arm_init_machine_status): Likewise.
2175 * config/avr/avr.c (avr_init_machine_status): Likewise.
2176 * config/bfin/bfin.c (bfin_init_machine_status): Likewise.
2177 * config/c6x/c6x.c (c6x_init_machine_status): Likewise.
2178 * config/cris/cris.c (cris_init_machine_status): Likewise.
2179 * config/darwin.c (machopic_indirection_name): Likewise.
2180 (darwin_build_constant_cfstring): Likewise.
2181 (darwin_enter_string_into_cfstring_table): Likewise.
2182 * config/epiphany/epiphany.c (epiphany_init_machine_status): Likewise.
2183 * config/frv/frv.c (frv_init_machine_status): Likewise.
2184 * config/i386/i386.c (get_dllimport_decl): Likewise.
2185 (ix86_init_machine_status): Likewise.
2186 (assign_386_stack_local): Likewise.
2187 * config/i386/winnt.c (i386_pe_record_external_function): Likewise.
2188 (i386_pe_maybe_record_exported_symbol): Likewise.
2189 (i386_pe_record_stub): Likewise.
2190 * config/ia64/ia64.c (ia64_init_machine_status): Likewise.
2191 * config/iq2000/iq2000.c (iq2000_init_machine_status): Likewise.
2192 * config/m32c/m32c.c (m32c_init_machine_status): Likewise.
2193 (m32c_note_pragma_address): Likewise.
2194 * config/mep/mep.c (mep_init_machine_status): Likewise.
2195 (mep_note_pragma_flag): Likewise.
2196 * config/mips/mips.c (mflip_mips16_use_mips16_p): Likewise.
2197 (mips16_local_alias): Likewise.
2198 (mips_init_machine_status): Likewise.
2199 * config/mmix/mmix.c (mmix_init_machine_status): Likewise.
2200 * config/moxie/moxie.c (moxie_init_machine_status): Likewise.
2201 * config/msp430/msp430.c (msp430_init_machine_status): Likewise.
2202 * config/nds32/nds32.c (nds32_init_machine_status): Likewise.
2203 * config/nios2/nios2.c (nios2_init_machine_status): Likewise.
2204 * config/pa/pa.c (pa_init_machine_status): Likewise.
2205 (pa_get_deferred_plabel): Likewise.
2206 * config/rl78/rl78.c (rl78_init_machine_status): Likewise.
2207 * config/rs6000/rs6000.c (builtin_function_type): Likewise.
2208 (rs6000_init_machine_status): Likewise.
2209 (output_toc): Likewise.
2210 * config/s390/s390.c (s390_init_machine_status): Likewise.
2211 * config/score/score.c (score_output_external): Likewise.
2212 * config/sparc/sparc.c (sparc_init_machine_status): Likewise.
2213 * config/spu/spu.c (spu_init_machine_status): Likewise.
2214 * config/tilegx/tilegx.c (tilegx_init_machine_status): Likewise.
2215 * config/tilepro/tilepro.c (tilepro_init_machine_status): Likewise.
2216 * config/xtensa/xtensa.c (xtensa_init_machine_status): Likewise.
2217 * coverage.c (coverage_end_function): Likewise.
2218 * dbxout.c (dbxout_init): Likewise.
2219 * doc/gty.texi: Don't mention variable_size attribute.
2220 * dwarf2cfi.c (new_cfi): Adjust.
2221 (new_cfi_row): Likewise.
2222 (copy_cfi_row): Likewise.
2223 (create_cie_data): Likewise.
2224 * dwarf2out.c (dwarf2out_alloc_current_fde): Likewise.
2225 (new_loc_descr): Likewise.
2226 (find_AT_string_in_table): Likewise.
2227 (add_addr_table_entry): Likewise.
2228 (new_die): Likewise.
2229 (add_var_loc_to_decl): Likewise.
2230 (clone_die): Likewise.
2231 (clone_as_declaration): Likewise.
2232 (break_out_comdat_types): Likewise.
2233 (new_loc_list): Likewise.
2234 (add_loc_descr_to_each): Likewise.
2235 (add_location_or_const_value_attribute): Likewise.
2236 (add_linkage_name): Likewise.
2237 (lookup_filename): Likewise.
2238 (dwarf2out_var_location): Likewise.
2239 (new_line_info_table): Likewise.
2240 (dwarf2out_init): Likewise.
2241 (mem_loc_descriptor): Likewise.
2242 (loc_descriptor): Likewise.
2243 (add_const_value_attribute): Likewise.
2244 (tree_add_const_value_attribute): Likewise.
2245 (comp_dir_string): Likewise.
2246 (dwarf2out_vms_debug_main_pointer): Likewise.
2247 (string_cst_pool_decl): Likewise.
2248 * emit-rtl.c (set_mem_attrs): Likewise.
2249 (get_reg_attrs): Likewise.
2250 (start_sequence): Likewise.
2251 (init_emit): Likewise.
2252 (init_emit_regs): Likewise.
2253 * except.c (init_eh_for_function): Likewise.
2254 (gen_eh_region): Likewise.
2255 (gen_eh_region_catch): Likewise.
2256 (gen_eh_landing_pad): Likewise.
2257 (add_call_site): Likewise.
2258 * function.c (add_frame_space): Likewise.
2259 (insert_temp_slot_address): Likewise.
2260 (assign_stack_temp_for_type): Likewise.
2261 (get_hard_reg_initial_val): Likewise.
2262 (allocate_struct_function): Likewise.
2263 (prepare_function_start): Likewise.
2264 (types_used_by_var_decl_insert): Likewise.
2265 * gengtype.c (variable_size_p): Remove function.
2266 (enum alloc_quantity): Remove enum.
2267 (write_typed_alloc_def): Remove function.
2268 (write_typed_struct_alloc_def): Likewise.
2269 (write_typed_typedef_alloc_def): Likewise.
2270 (write_typed_alloc_defns): Likewise.
2271 (main): Adjust.
2272 * ggc-common.c (ggc_cleared_alloc_htab_ignore_args): Adjust.
2273 (ggc_cleared_alloc_ptr_array_two_args): Likewise.
2274 * ggc.h (ggc_alloc): new function.
2275 (ggc_cleared_alloc): Likewise.
2276 (ggc_vec_alloc): Template on type of vector element, and remove
2277 element size argument.
2278 (ggc_cleared_vec_alloc): Likewise.
2279 * gimple.c (gimple_build_omp_for): Adjust.
2280 (gimple_copy): Likewise.
2281 * ipa-cp.c (get_replacement_map): Likewise.
2282 (find_aggregate_values_for_callers_subset): Likewise.
2283 (known_aggs_to_agg_replacement_list): Likewise.
2284 * ipa-devirt.c (get_odr_type): Likewise.
2285 * ipa-prop.c (ipa_node_duplication_hook): Likewise.
2286 (read_agg_replacement_chain): Likewise.
2287 * loop-iv.c (get_simple_loop_desc): Likewise.
2288 * lto-cgraph.c (input_node_opt_summary): Likewise.
2289 * lto-section-in.c (lto_new_in_decl_state): Likewise.
2290 * lto-streamer-in.c (lto_input_eh_catch_list): Likewise.
2291 (input_eh_region): Likewise.
2292 (input_eh_lp): Likewise.
2293 (input_cfg): Likewise.
2294 * optabs.c (set_optab_libfunc): Likewise.
2295 (init_tree_optimization_optabs): Likewise.
2296 (set_conv_libfunc): Likewise.
2297 * passes.c (do_per_function_toporder): Likewise.
2298 * rtl.h: Don't use variable_size gty attribute.
2299 * sese.c (if_region_set_false_region): Adjust.
2300 * stringpool.c (gt_pch_save_stringpool): Likewise.
2301 * target-globals.c (save_target_globals): Likewise.
2302 * toplev.c (general_init): Likewise.
2303 * trans-mem.c (record_tm_replacement): Likewise.
2304 (split_bb_make_tm_edge): Likewise.
2305 * tree-cfg.c (move_sese_region_to_fn): Likewise.
2306 * tree-data-ref.h (lambda_vector_new): Likewise.
2307 * tree-eh.c (add_stmt_to_eh_lp_fn): Likewise.
2308 * tree-iterator.c (tsi_link_before): Likewise.
2309 (tsi_link_after): Likewise.
2310 * tree-scalar-evolution.c (new_scev_info_str): Likewise.
2311 * tree-ssa-loop-niter.c (record_estimate): Likewise.
2312 * tree-ssa-operands.c (ssa_operand_alloc): Likewise.
2313 * tree-ssa-operands.h: Don't use variable_size gty attribute.
2314 * tree-ssa.c (init_tree_ssa): Adjust.
2315 * tree-ssanames.c (set_range_info): Likewise.
2316 (get_ptr_info): Likewise.
2317 (duplicate_ssa_name_ptr_info): Likewise.
2318 (duplicate_ssa_name_range_info): Likewise.
2319 * tree-streamer-in.c (unpack_ts_real_cst_value_fields): Likewise.
2320 (unpack_ts_fixed_cst_value_fields): Likewise.
2321 * tree.c (build_fixed): Likewise.
2322 (build_real): Likewise.
2323 (build_string): Likewise.
2324 (decl_priority_info): Likewise.
2325 (decl_debug_expr_insert): Likewise.
2326 (decl_value_expr_insert): Likewise.
2327 (decl_debug_args_insert): Likewise.
2328 (type_hash_add): Likewise.
2329 (build_omp_clause): Likewise.
2330 * ubsan.c (decl_for_type_insert): Likewise.
2331 * varasm.c (get_unnamed_section): Likewise.
2332 (get_noswitch_section): Likewise.
2333 (get_section): Likewise.
2334 (get_block_for_section): Likewise.
2335 (create_block_symbol): Likewise.
2336 (build_constant_desc): Likewise.
2337 (create_constant_pool): Likewise.
2338 (force_const_mem): Likewise.
2339 (record_tm_clone_pair): Likewise.
2340 * varpool.c (varpool_create_empty_node): Likewise.
2342 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
2344 * dwarf2out.c (tree_add_const_value_attribute): Call
2345 ggc_internal_cleared_alloc instead of ggc_alloc_cleared_atomic.
2346 * gengtype.c (write_typed_alloc_def): Call ggc_internal_<x>alloc
2347 instead of ggc_internal_<x>alloc_stat.
2348 * ggc-common.c (ggc_internal_cleared_alloc): Drop _stat suffix.
2349 (ggc_realloc): Likewise.
2350 * ggc-none.c (ggc_internal_alloc): Likewise.
2351 (ggc_internal_cleared_alloc): Likewise.
2352 * ggc-page.c: Likewise.
2353 * ggc.h (ggc_internal_alloc_stat): Likewise.
2354 (ggc_internal_alloc): Remove macro.
2355 (ggc_internal_cleared_alloc_stat): Drop _stat suffix.
2356 (ggc_internal_cleared_alloc): Remove macro.
2357 (GGC_RESIZEVEC): Adjust.
2358 (ggc_resizevar): Remove macro.
2359 (ggc_internal_vec_alloc_stat): Drop _stat suffix.
2360 (ggc_internal_cleared_vec_alloc_stat): Likewise.
2361 (ggc_internal_vec_cleared_alloc): Remove macro.
2362 (ggc_alloc_atomic_stat): Drop _stat suffix.
2363 (ggc_alloc_atomic): Remove macro.
2364 (ggc_alloc_cleared_atomic): Remove macro.
2365 (ggc_alloc_string_stat): Drop _stat suffix.
2366 (ggc_alloc_string): Remove macro.
2367 (ggc_alloc_rtx_def_stat): Adjust.
2368 (ggc_alloc_tree_node_stat): Likewise.
2369 (ggc_alloc_cleared_tree_node_stat): Likewise.
2370 (ggc_alloc_cleared_gimple_statement_stat): Likewise.
2371 (ggc_alloc_cleared_simd_clone_stat): Likewise.
2372 * gimple.c (gimple_build_omp_for): Likewise.
2373 (gimple_copy): Likewise.
2374 * stringpool.c (ggc_alloc_string_stat): Drop _stat suffix.
2375 * toplev.c (realloc_for_line_map): Adjust.
2376 * tree-data-ref.h (lambda_vector_new): Likewise.
2377 * tree-phinodes.c (allocate_phi_node): Likewise.
2378 * tree.c (grow_tree_vec_stat): Likewise.
2379 * vec.h (va_gc::reserve): Adjust.
2381 2014-05-17 Ajit Agarwal <ajitkum@xilinx.com>
2383 * config/microblaze/microblaze.c (break_handler): New Declaration.
2384 (microblaze_break_function_p,microblaze_is_break_handler): New.
2385 (compute_frame_size): Use microblaze_break_function_p.
2386 Add the test of break_handler.
2387 (microblaze_function_prologue) : Add the test of variable
2388 break_handler. Check the fnname by BREAK_HANDLER_NAME.
2389 (microblaze_function_epilogue) : Add the test of break_handler.
2390 (microblaze_globalize_label) : Add the test of break_handler.
2391 Check the name by BREAK_HANDLER_NAME.
2393 * config/microblaze/microblaze.h (BREAK_HANDLER_NAME): New macro
2395 * config/microblaze/microblaze.md (*<optab>,<optab>_internal): Add
2396 microblaze_is_break_handler test.
2397 (call_internal1,call_value_intern): Use microblaze_break_function_p.
2398 Use SYMBOL_REF_DECL.
2400 * config/microblaze/microblaze-protos.h
2401 (microblaze_break_function_p,microblaze_is_break_handler):
2402 New Declaration.
2404 * doc/extend.texi (MicroBlaze break_handler Functions): Document
2405 new MicroBlaze break_handler functions.
2407 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
2409 * doc/extend.texi (Size of an asm): Move node text according
2410 to its @menu entry position.
2412 2014-05-17 Marc Glisse <marc.glisse@inria.fr>
2414 PR tree-optimization/61140
2415 PR tree-optimization/61150
2416 PR tree-optimization/61197
2417 * tree-ssa-phiopt.c (value_replacement): Punt on multiple phis.
2419 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
2421 * doc/invoke.texi (free): Mention Alpha. Also enabled at -Os.
2423 2014-05-17 Richard Sandiford <r.sandiford@uk.ibm.com>
2425 * wide-int.cc: Only include longlong.h if W_TYPE_SIZE==32 or
2426 __SIZEOF_INT128__ is defined.
2428 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
2430 * config/rs6000/rs6000.c (rs6000_real_tls_symbol_ref_p): New function.
2431 (rs6000_delegitimize_address): Use it.
2433 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
2435 * emit-rtl.h (replace_equiv_address, replace_equiv_address_nv): Add an
2436 inplace argument. Store the new address in the original MEM when true.
2437 * emit-rtl.c (change_address_1): Likewise.
2438 (adjust_address_1, adjust_automodify_address_1, offset_address):
2439 Update accordingly.
2440 * rtl.h (plus_constant): Add an inplace argument.
2441 * explow.c (plus_constant): Likewise. Try to reuse the original PLUS
2442 when true. Avoid generating (plus X (const_int 0)).
2443 * function.c (instantiate_virtual_regs_in_rtx): Adjust the PLUS
2444 in-place. Pass true to plus_constant.
2445 (instantiate_virtual_regs_in_insn): Pass true to replace_equiv_address.
2447 2014-05-16 Dehao Chen <dehao@google.com>
2449 * tree-cfg.c (gimple_merge_blocks): Updates bb count with max count.
2451 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
2453 PR target/54089
2454 * config/sh/predicates.md (negt_reg_shl31_operand): Match additional
2455 patterns.
2456 * config/sh/sh.md (*negt_msb): Merge SH2A and non-SH2A variants.
2458 2014-05-16 Dehao Chen <dehao@google.com>
2460 * ira-int.h (REG_FREQ_FROM_EDGE_FREQ): Use
2461 optimize_function_for_size_p.
2462 * regs.h (REG_FREQ_FROM_BB): Likewise.
2464 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
2466 PR target/51244
2467 * config/sh/sh.c (sh_eval_treg_value): Handle t_reg_operand and
2468 negt_reg_operand cases.
2469 * config/sh/sh.md (*cset_zero): Likewise by using cbranch_treg_value
2470 predicate.
2471 * config/sh/predicates.md (cbranch_treg_value): Simplify.
2473 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
2475 * config/sh/sh.c (sh_option_override): Set branch cost to 2 for all
2476 target variants.
2478 2014-05-16 David Malcolm <dmalcolm@redhat.com>
2480 Revert:
2481 2014-04-29 David Malcolm <dmalcolm@redhat.com>
2483 * tree-cfg.c (dump_function_to_file): Dump the return type of
2484 functions, in a line to itself before the function body, mimicking
2485 the layout of a C function.
2487 2014-05-16 Dehao Chen <dehao@google.com>
2489 * cfghooks.c (make_forwarder_block): Use direct computation to
2490 get fall-through edge's count and frequency.
2492 2014-05-16 Benno Schulenberg <bensberg@justemail.net>
2494 * config/arc/arc.c (arc_init): Fix typo in error message.
2495 * config/i386/i386.c (ix86_expand_builtin): Likewise.
2496 (split_stack_prologue_scratch_regno): Likewise.
2497 * fortran/check.c (gfc_check_fn_rc2008): Remove duplicate
2498 word from error message.
2500 2014-05-16 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
2502 * ira-costs.c: Fix typo in comment.
2504 2014-05-16 David Wohlferd <dw@LimeGreenSocks.com>
2506 * doc/extend.texi: (Visibility Pragmas) Fix misplaced @xref
2508 2014-05-16 Jan Hubicka <hubicka@ucw.cz>
2510 * varpool.c (dump_varpool_node): Dump write-only flag.
2511 * lto-cgraph.c (lto_output_varpool_node, input_varpool_node): Stream
2512 write-only flag.
2513 * tree-cfg.c (execute_fixup_cfg): Remove statements setting
2514 write-only variables.
2515 * ipa.c (process_references): New function.
2516 (set_readonly_bit): New function.
2517 (set_writeonly_bit): New function.
2518 (clear_addressable_bit): New function.
2519 (ipa_discover_readonly_nonaddressable_var): Mark write only variables;
2520 fix handling of aliases.
2521 * cgraph.h (struct varpool_node): Add writeonly flag.
2523 2014-05-16 Vladimir Makarov <vmakarov@redhat.com>
2525 PR rtl-optimization/60969
2526 * ira-costs.c (record_reg_classes): Allow only memory for pseudo.
2527 Calculate costs for this case.
2529 2014-05-16 Eric Botcazou <ebotcazou@adacore.com>
2531 * fold-const (fold_unary_loc) <NON_LVALUE_EXPR>: New case.
2532 <CASE_CONVERT>: Pass arg0 instead of op0 to fold_convert_const.
2534 2014-05-16 Richard Biener <rguenther@suse.de>
2536 PR tree-optimization/61194
2537 * tree-vect-patterns.c (adjust_bool_pattern): Also handle
2538 bool patterns ending in a COND_EXPR.
2540 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2542 * config/aarch64/aarch64.c (aarch64_rtx_mult_cost): Fix FNMUL case.
2544 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2546 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle the case
2547 where we were unable to cost an RTX.
2549 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2551 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost SYMBOL_REF,
2552 HIGH, LO_SUM.
2554 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2555 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2557 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost TRUNCATE.
2559 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2560 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2562 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost FMA,
2563 FLOAT_EXTEND, FLOAT_TRUNCATE, ABS, SMAX, and SMIN.
2565 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2566 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2568 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost comparison
2569 operators.
2571 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2572 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2574 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
2575 DIV/MOD.
2577 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2578 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2580 * config/aarch64/aarch64.c (aarch64_rtx_arith_op_extract_p): New.
2581 (aarch64_rtx_costs): Improve costs for SIGN/ZERO_EXTRACT.
2583 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2584 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2586 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
2587 rotates and shifts.
2589 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2590 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2592 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost
2593 ZERO_EXTEND and SIGN_EXTEND better.
2595 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2596 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2598 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve cost for
2599 logical operations.
2601 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2602 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2604 * config/aarch64/aarch64.c (aarch64_rtx_costs): Use address
2605 costs when costing loads and stores to memory.
2607 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2608 Philip Tomsich <philipp.tomsich@theobroma-systems.com>
2610 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costing
2611 for SET RTX.
2613 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2615 * config/aarch64/aarch64.c (aarch64_rtx_costs): Set default costs.
2617 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2618 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2620 * config/aarch64/aarch64.c (aarch64_strip_shift_or_extend): Rename
2621 to...
2622 (aarch64_strip_extend): ...this, don't strip shifts, check RTX is
2623 well formed.
2624 (aarch64_rtx_mult_cost): New.
2625 (aarch64_rtx_costs): Use it, refactor as appropriate.
2627 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2628 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2630 * config/aarch64/aarch64.c (aarch64_build_constant): Conditionally
2631 emit instructions, return number of instructions which would
2632 be emitted.
2633 (aarch64_add_constant): Update call to aarch64_build_constant.
2634 (aarch64_output_mi_thunk): Likewise.
2635 (aarch64_rtx_costs): Estimate cost of a CONST_INT, cost of
2636 a CONST_DOUBLE.
2638 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2640 * config/aarch64/aarch64.c (aarch64_rtx_costs_wrapper): New.
2641 (TARGET_RTX_COSTS): Call it.
2643 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2645 * config/aarch64/aarch64.c (cortexa57_addrcost_table): New.
2646 (cortexa57_vector_cost): Likewise.
2647 (cortexa57_tunings): Use them.
2649 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2651 * config/aarch64/aarch64-protos.h (scale_addr_mode_cost): New.
2652 (cpu_addrcost_table): Use it.
2653 * config/aarch64/aarch64.c (generic_addrcost_table): Initialize it.
2654 (aarch64_address_cost): Rewrite using aarch64_classify_address,
2655 move it.
2657 2014-05-16 Richard Biener <rguenther@suse.de>
2659 * tree-ssa-sccvn.c: Include tree-cfg.h and domwalk.h.
2660 (set_ssa_val_to): Handle unexpected sets to VN_TOP.
2661 (visit_phi): Ignore edges marked as not executable.
2662 (class cond_dom_walker): New.
2663 (cond_dom_walker::before_dom_children): Value-number
2664 control statements and mark successor edges as not
2665 executable if possible.
2666 (run_scc_vn): First walk all control statements in
2667 dominator order, marking edges as not executable.
2668 * tree-inline.c (copy_edges_for_bb): Be not confused
2669 about random edge flags.
2671 2014-05-16 Richard Biener <rguenther@suse.de>
2673 * tree-ssa-sccvn.c (visit_use): Also constant-fold calls.
2675 2014-05-15 Peter Bergner <bergner@vnet.ibm.com>
2677 PR target/61193
2678 * config/rs6000/htmxlintrin.h (_HTM_TBEGIN_STARTED): New define.
2679 (__TM_simple_begin): Use it.
2680 (__TM_begin): Likewise.
2682 2014-05-15 Martin Jambor <mjambor@suse.cz>
2684 PR ipa/61085
2685 * ipa-prop.c (update_indirect_edges_after_inlining): Check
2686 type_preserved flag when the indirect edge is polymorphic.
2688 2014-05-15 Martin Jambor <mjambor@suse.cz>
2690 PR tree-optimization/61090
2691 * tree-sra.c (sra_modify_expr): Pass the current gsi to
2692 build_ref_for_model.
2694 2014-05-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2696 * config/arm/arm.c (arm_option_override): Use the SCHED_PRESSURE_MODEL
2697 enum name for PARAM_SCHED_PRESSURE_ALGORITHM.
2699 2014-05-15 Jakub Jelinek <jakub@redhat.com>
2701 PR tree-optimization/61158
2702 * fold-const.c (fold_binary_loc): If X is zero-extended and
2703 shiftc >= prec, make sure zerobits is all ones instead of
2704 invoking undefined behavior.
2706 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2708 * regcprop.h: New file.
2709 * regcprop.c (skip_debug_insn_p): New decl.
2710 (replace_oldest_value_reg): Check skip_debug_insn_p.
2711 (copyprop_hardreg_forward_bb_without_debug_insn): New function.
2712 * shrink-wrap.c: Include regcprop.h.
2713 (prepare_shrink_wrap): Call
2714 copyprop_hardreg_forward_bb_without_debug_insn.
2716 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2718 * shrink-wrap.h: Update comment.
2719 * shrink-wrap.c: Update comment.
2720 (next_block_for_reg): Rename to live_edge_for_reg.
2721 (live_edge_for_reg): Allow live_edge->dest has two predecessors.
2722 (move_insn_for_shrink_wrap): Split live_edge.
2723 (prepre_shrink_wrap): One more parameter for move_insn_for_shrink_wrap.
2725 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
2727 * config/sparc/sparc-protos.h (sparc_absnegfloat_split_legitimate):
2728 Delete.
2729 * config/sparc/sparc.c (sparc_absnegfloat_split_legitimate): Likewise.
2730 * config/sparc/sparc.md (fptype_ut699): New attribute.
2731 (in_branch_delay): Return false if -mfix-ut699 is specified and
2732 fptype_ut699 is set to single.
2733 (truncdfsf2): Add fptype_ut699 attribute.
2734 (fix_truncdfsi2): Likewise.
2735 (floatsisf2): Change fptype attribute.
2736 (fix_truncsfsi2): Likewise.
2737 (negtf2_notv9): Delete.
2738 (negtf2_v9): Likewise.
2739 (negtf2_hq): New instruction.
2740 (negtf2): New instruction and splitter.
2741 (negdf2_notv9): Rewrite.
2742 (abstf2_notv9): Delete.
2743 (abstf2_hq_v9): Likewise.
2744 (abstf2_v9): Likewise.
2745 (abstf2_hq): New instruction.
2746 (abstf2): New instruction and splitter.
2747 (absdf2_notv9): Rewrite.
2749 2014-05-14 Cary Coutant <ccoutant@google.com>
2751 PR debug/61013
2752 * opts.c (common_handle_option): Don't special-case "-g".
2753 (set_debug_level): Default to at least level 2 with "-g".
2755 2014-05-14 DJ Delorie <dj@redhat.com>
2757 * config/msp430/msp430.c (msp430_builtin): Add
2758 MSP430_BUILTIN_DELAY_CYCLES.
2759 (msp430_init_builtins): Register void __delay_cycles(long long).
2760 (msp430_builtin_decl): Add it.
2761 (cg_magic_constant): New.
2762 (msp430_expand_delay_cycles): New.
2763 (msp430_expand_builtin): Call it.
2764 (msp430_print_operand_raw): Change integer printing from "int" to
2765 HOST_WIDE_INT.
2766 * config/msp430/msp430.md (define_constants): Add delay_cycles tags.
2767 (delay_cycles_start): New.
2768 (delay_cycles_end): New.
2769 (delay_cycles_32): New.
2770 (delay_cycles_32x): New.
2771 (delay_cycles_16): New.
2772 (delay_cycles_16x): New.
2773 (delay_cycles_2): New.
2774 (delay_cycles_1): New.
2775 * doc/extend.texi: Document __delay_cycles().
2777 2014-05-14 Sandra Loosemore <sandra@codesourcery.com>
2779 * config/nios2/nios2.md (nios2_cbranch): Fix paste-o in
2780 length attribute computation.
2782 2014-05-14 Richard Sandiford <rdsandiford@googlemail.com>
2784 PR debug/61188
2785 * print-rtl.c (print_rtx): Suppress uids if flag_dump_unnumbered.
2787 2014-05-14 Richard Sandiford <r.sandiford@uk.ibm.com>
2789 PR target/61084
2790 * config/sparc/sparc.md: Fix types of low and high in DI constant
2791 splitter. Use gen_int_mode in some other splitters.
2793 2014-05-14 Martin Jambor <mjambor@suse.cz>
2795 PR ipa/60897
2796 * ipa-prop.c (ipa_modify_formal_parameters): Reset DECL_LANG_SPECIFIC.
2798 2014-05-14 James Norris <jnorris@codesourcery.com>
2800 * omp-low.c (expand_parallel_call): Remove shadow variable.
2801 (expand_omp_taskreg): Likewise.
2803 2014-05-14 Ilya Tocar <ilya.tocar@intel.com>
2805 * common/config/i386/i386-common.c
2806 (OPTION_MASK_ISA_CLFLUSHOPT_SET): Define.
2807 (OPTION_MASK_ISA_XSAVES_SET): Ditto.
2808 (OPTION_MASK_ISA_XSAVEC_SET): Ditto.
2809 (OPTION_MASK_ISA_CLFLUSHOPT_UNSET): Ditto.
2810 (OPTION_MASK_ISA_XSAVES_UNSET): Ditto.
2811 (OPTION_MASK_ISA_XSAVEC_UNSET): Ditto.
2812 (ix86_handle_option): Handle OPT_mxsavec, OPT_mxsaves, OPT_mclflushopt.
2813 * config.gcc (i[34567]86-*-*): Add clflushoptintrin.h,
2814 xsavecintrin.h, xsavesintrin.h.
2815 (x86_64-*-*): Ditto.
2816 * config/i386/clflushoptintrin.h: New.
2817 * config/i386/xsavecintrin.h: Ditto.
2818 * config/i386/xsavesintrin.h: Ditto.
2819 * config/i386/cpuid.h (bit_CLFLUSHOPT): Define.
2820 (bit_XSAVES): Ditto.
2821 (bit_XSAVES): Ditto.
2822 * config/i386/driver-i386.c (host_detect_local_cpu): Handle
2823 -mclflushopt, -mxsavec, -mxsaves, -mno-xsaves, -mno-xsavec,
2824 -mno-clflushopt.
2825 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
2826 OPTION_MASK_ISA_CLFLUSHOPT, OPTION_MASK_ISA_XSAVEC,
2827 OPTION_MASK_ISA_XSAVES.
2828 * config/i386/i386.c (ix86_target_string): Handle -mclflushopt,
2829 -mxsavec, -mxsaves.
2830 (PTA_CLFLUSHOPT) Define.
2831 (PTA_XSAVEC): Ditto.
2832 (PTA_XSAVES): Ditto.
2833 (ix86_option_override_internal): Handle new options.
2834 (ix86_valid_target_attribute_inner_p): Ditto.
2835 (ix86_builtins): Add IX86_BUILTIN_XSAVEC, IX86_BUILTIN_XSAVEC64,
2836 IX86_BUILTIN_XSAVES, IX86_BUILTIN_XRSTORS, IX86_BUILTIN_XSAVES64,
2837 IX86_BUILTIN_XRSTORS64, IX86_BUILTIN_CLFLUSHOPT.
2838 (bdesc_special_args): Add __builtin_ia32_xsaves,
2839 __builtin_ia32_xrstors, __builtin_ia32_xsavec, __builtin_ia32_xsaves64,
2840 __builtin_ia32_xrstors64, __builtin_ia32_xsavec64.
2841 (ix86_init_mmx_sse_builtins): Add __builtin_ia32_clflushopt.
2842 (ix86_expand_builtin): Handle new builtins.
2843 * config/i386/i386.h (TARGET_CLFLUSHOPT) Define.
2844 (TARGET_CLFLUSHOPT_P): Ditto.
2845 (TARGET_XSAVEC): Ditto.
2846 (TARGET_XSAVEC_P): Ditto.
2847 (TARGET_XSAVES): Ditto.
2848 (TARGET_XSAVES_P): Ditto.
2849 * config/i386/i386.md (ANY_XSAVE): Add UNSPECV_XSAVEC, UNSPECV_XSAVES.
2850 (ANY_XSAVE64)" Add UNSPECV_XSAVEC64, UNSPECV_XSAVES64.
2851 (attr xsave): Add xsavec, xsavec64, xsaves, xsaves64.
2852 (ANY_XRSTOR): New.
2853 (ANY_XRSTOR64): Ditto.
2854 (xrstor): Ditto.
2855 (xrstor): Change into <xrstor>.
2856 (xrstor_rex64): Change into <xrstor>_rex64.
2857 (xrstor64): Change into <xrstor>64
2858 (clflushopt): New.
2859 * config/i386/i386.opt (mclflushopt): New.
2860 (mxsavec): Ditto.
2861 (mxsaves): Ditto.
2862 * config/i386/x86intrin.h: Add clflushoptintrin.h, xsavesintrin.h,
2863 xsavecintrin.h.
2864 * doc/invoke.texi: Document new options.
2866 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
2868 PR rtl-optimization/60866
2869 * sel-sched-ir (sel_init_new_insn): New parameter old_seqno.
2870 Default it to -1. Pass it down to init_simplejump_data.
2871 (init_simplejump_data): New parameter old_seqno. Pass it down
2872 to get_seqno_for_a_jump.
2873 (get_seqno_for_a_jump): New parameter old_seqno. Use it for
2874 initializing new jump seqno as a last resort. Add comment.
2875 (sel_redirect_edge_and_branch): Save old seqno of the conditional
2876 jump and pass it down to sel_init_new_insn.
2877 (sel_redirect_edge_and_branch_force): Likewise.
2879 2014-05-14 Georg-Johann Lay <avr@gjlay.de>
2881 * config/avr/avr.h (REG_CLASS_CONTENTS): Use unsigned suffix for
2882 shifted values to avoid build warning.
2884 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
2886 * cfgcleanup.c (try_forward_edges): Use location_t for locations.
2887 * cfgrtl.c (rtl_merge_blocks): Fix comment.
2888 (cfg_layout_merge_blocks): Likewise.
2889 * except.c (emit_to_new_bb_before): Remove prev_bb local variable.
2891 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
2893 PR rtl-optimization/60901
2894 * config/i386/i386.c (ix86_dependencies_evaluation_hook): Check that
2895 bb predecessor belongs to the same scheduling region. Adjust comment.
2897 2014-05-13 Peter Bergner <bergner@vnet.ibm.com>
2899 * doc/sourcebuild.texi: (dfp_hw): Document.
2900 (p8vector_hw): Likewise.
2901 (powerpc_eabi_ok): Likewise.
2902 (powerpc_elfv2): Likewise.
2903 (powerpc_htm_ok): Likewise.
2904 (ppc_recip_hw): Likewise.
2905 (vsx_hw): Likewise.
2907 2014-05-13 Cary Coutant <ccoutant@google.com>
2909 * opts.c (finish_options): Use -ggnu-pubnames with -gsplit-dwarf.
2911 2014-05-13 David Malcolm <dmalcolm@redhat.com>
2913 * gengtype-parse.c (require3): Eliminate in favor of...
2914 (require4): New.
2915 (require_template_declaration): Update to support optional single *
2916 on a type.
2918 * gengtype.c (get_ultimate_base_class): Add a non-const overload.
2919 (create_user_defined_type): Handle a single level of explicit
2920 pointerness within template arguments.
2921 (struct write_types_data): Add field "kind".
2922 (filter_type_name): Handle "*" character.
2923 (write_user_func_for_structure_ptr): Require a write_types_data
2924 rather than just a prefix string, so that we can look up the kind
2925 of the wtd and use it as an index into wrote_user_func_for_ptr,
2926 ensuring that such functions are written at most once. Support
2927 subclasses by invoking the marking function of the ultimate base class.
2928 (write_user_func_for_structure_body): Require a write_types_data
2929 rather than just a prefix string, so that we can pass this to
2930 write_user_func_for_structure_ptr.
2931 (write_func_for_structure): Likewise.
2932 (ggc_wtd): Add initializer of new "kind" field.
2933 (pch_wtd): Likewise.
2935 * gengtype.h (enum write_types_kinds): New.
2936 (struct type): Add field wrote_user_func_for_ptr to the "s"
2937 union member.
2939 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
2941 * fold-const.c (optimize_bit_field_compare): Use wi:: operations
2942 instead of const_binop.
2943 (fold_binary_loc): Likewise.
2945 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
2947 * tree-dfa.h (get_addr_base_and_unit_offset_1): Update array index
2948 calculation to match get_ref_base_and_extent.
2950 2014-05-13 Catherine Moore <clm@codesourcery.com>
2951 Sandra Loosemore <sandra@codesourcery.com>
2953 * configure.ac: Fix assembly for explicit JALR relocation check.
2954 * configure: Regenerate.
2956 2014-05-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2958 * config/arm/arm.c (neon_itype): Remove NEON_RESULTPAIR.
2959 (arm_init_neon_builtins): Remove handling of NEON_RESULTPAIR.
2960 Remove associated type declarations and initialisations.
2961 (arm_expand_neon_builtin): Likewise.
2962 (neon_emit_pair_result_insn): Delete.
2963 * config/arm/arm_neon_builtins (vtrn, vzip, vuzp): Delete.
2964 * config/arm/neon.md (neon_vtrn<mode>): Delete.
2965 (neon_vzip<mode>): Likewise.
2966 (neon_vuzp<mode>): Likewise.
2968 2014-05-13 Richard Biener <rguenther@suse.de>
2970 PR ipa/60973
2971 * tree-inline.c (remap_gimple_stmt): Clear tail call flag,
2972 it needs revisiting whether the call still may be tail-called.
2974 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2976 * rtl.def (SYMBOL_REF): Remove middle "0" field.
2977 * rtl.h (block_symbol): Reduce number of fields to 2.
2978 (rtx_def): Add u2.symbol_ref_flags.
2979 (SYMBOL_REF_FLAGS): Use it.
2980 (SYMBOL_REF_DATA, SET_SYMBOL_REF_DECL, SYMBOL_REF_DECL)
2981 (SET_SYMBOL_REF_CONSTANT, SYMBOL_REF_CONSTANT): Lower index.
2982 * gengtype.c (adjust_field_rtx_def): Remove SYMBOL_REF_FLAGS handling.
2983 Lower index of SYMBOL_REF_DATA.
2984 * print-rtl.c (print_rtx): Lower index for SYMBOL_REF_DATA.
2985 Print SYMBOL_REF_FLAGS at the same time.
2986 * genattrtab.c (attr_rtx_1): Only initialize 1 "0" SYMBOL_REF field.
2988 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2990 * rtl.def (VAR_LOCATION): Remove "i" field.
2991 * rtl.h (rtx_def): Add u2.var_location_status.
2992 (PAT_VAR_LOCATION_STATUS): Use it.
2993 (gen_rtx_VAR_LOCATION): Declare.
2994 * gengenrtl.c (excluded_rtx): Add VAR_LOCATION.
2995 * emit-rtl.c (gen_rtx_VAR_LOCATION): New function.
2996 * var-tracking.c (emit_note_insn_var_location): Remove casts.
2998 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
3000 * rtl.def (scratch): Fix outdated comment and remove "0" field.
3001 * gengtype.c (adjust_field_rtx_def): Update accordingly.
3003 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
3005 * rtl.def (DEBUG_INSN, INSN, JUMP_INSN, CALL_INSN, JUMP_TABLE_DATA)
3006 (BARRIER, CODE_LABEL, NOTE): Remove first "i" field.
3007 * rtl.h (rtx_def): Add insn_uid to u2 field.
3008 (RTX_FLAG_CHECK8): Delete in favor of...
3009 (RTL_INSN_CHAIN_FLAG_CHECK): ...this new macro.
3010 (INSN_DELETED_P): Update accordingly.
3011 (INSN_UID): Use u2.insn_uid.
3012 (INSN_CHAIN_CODE_P): Define.
3013 (PREV_INSN, NEXT_INSN, BLOCK_FOR_INSN, PATTERN, INSN_LOCATION)
3014 (INSN_CODE, REG_NOTES, CALL_INSN_FUNCTION_USAGE, CODE_LABEL_NUMBER)
3015 (NOTE_DATA, NOTE_DELETED_LABEL_NAME, NOTE_BLOCK, NOTE_EH_HANDLER)
3016 (NOTE_BASIC_BLOCK, NOTE_VAR_LOCATION, NOTE_CFI, NOTE_LABEL_NUMBER)
3017 (NOTE_KIND, LABEL_NAME, LABEL_NUSES, JUMP_LABEL, LABEL_REFS): Lower
3018 indices accordingly.
3019 * print-rtl.c (print_rtx): Print INSN_UIDs before the main loop.
3020 Update indices for insn-chain rtxes.
3021 * gengtype.c (gen_rtx_next): Adjust test for insn-chain rtxes.
3022 (adjust_field_rtx_def): Lower '0' indices for all insn-chain rtxes.
3023 * emit-rtl.c (gen_label_rtx): Update gen_rtx_LABEL call.
3024 * caller-save.c (init_caller_save): Update gen_rtx_INSN calls.
3025 * combine.c (try_combine): Likewise.
3026 * ira.c (setup_prohibited_mode_move_regs): Likewise.
3028 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
3030 * rtl.def (REG): Remove middle field.
3031 * rtl.h (rtx_def): Add orignal_regno to u2.
3032 (ORIGINAL_REGNO): Use it instead of field 1.
3033 (REG_ATTRS): Lower field index accordingly.
3034 * gengtype.c (adjust_field_rtx_def): Remove handling of
3035 ORIGINAL_REGNO. Move REG_ATTRS index down.
3036 * print-rtl.c (print_rtx): Move ORIGINAL_REGNO handling to the
3037 code that prints the REGNO.
3039 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
3041 * print-rtl.c (print_rtx): Guard whole '0' block with ifndef
3042 GENERATOR_FILE.
3044 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
3046 * rtl.h (rtx_def): Mark u2 as GTY ((skip)).
3048 2014-05-13 Bin Cheng <bin.cheng@arm.com>
3050 * tree-ssa-loop-ivopts.c (contain_complex_addr_expr): New.
3051 (alloc_iv): Lower base expressions containing ADDR_EXPR.
3053 2014-05-13 Ian Bolton <ian.bolton@arm.com>
3055 * config/aarch64/aarch64-protos.h
3056 (aarch64_hard_regno_caller_save_mode): New prototype.
3057 * config/aarch64/aarch64.c (aarch64_hard_regno_caller_save_mode):
3058 New function.
3059 * config/aarch64/aarch64.h (HARD_REGNO_CALLER_SAVE_MODE): New macro.
3061 2014-05-13 Christian Bruel <christian.bruel@st.com>
3063 * target.def (mode_switching): New hook vector.
3064 (mode_emit, mode_needed, mode_after, mode_entry): New hooks.
3065 (mode_exit, modepriority_to_mode): Likewise.
3066 * mode-switching.c (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Hookify.
3067 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
3068 * target.h: Include tm.h and hard-reg-set.h.
3069 * doc/tm.texi.in (EMIT_MODE_SET, MODE_NEEDED, MODE_AFTER, MODE_ENTRY)
3070 (MODE_EXIT, MODE_PRIORITY_TO_MODE): Delete and hookify.
3071 * doc/tm.texi Regenerate.
3072 * config/sh/sh.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
3073 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
3074 * config/sh/sh.c (sh_emit_mode_set, sh_mode_priority): Hookify.
3075 (sh_mode_needed, sh_mode_after, sh_mode_entry, sh_mode_exit): Likewise.
3076 * config/i386/i386.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
3077 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
3078 * config/i386/i386-protos.h (ix86_mode_needed, ix86_mode_after)
3079 (ix86_mode_entrym, ix86_emit_mode_set): Remove external declaration.
3080 * config/i386/i386.c (ix86_mode_needed, ix86_mode_after,
3081 (ix86_mode_exit, ix86_mode_entry, ix86_mode_priority)
3082 (ix86_emit_mode_set): Hookify.
3083 * config/epiphany/epiphany.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY):
3084 Delete.
3085 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
3086 * config/epiphany/epiphany-protos.h (epiphany_mode_needed)
3087 (emit_set_fp_mode, epiphany_mode_entry_exit, epiphany_mode_after)
3088 (epiphany_mode_priority_to_mode): Remove declaration.
3089 * config/epiphany/epiphany.c (emit_set_fp_mode): Hookify.
3090 (epiphany_mode_needed, epiphany_mode_priority_to_mode): Likewise.
3091 (epiphany_mode_entry, epiphany_mode_exit, epiphany_mode_after):
3092 Likewise.
3093 (epiphany_mode_priority_to_mode): Change priority type. Hookify.
3094 (epiphany_mode_needed, epiphany_mode_entry_exit): Hookify.
3095 (epiphany_mode_after, epiphany_mode_entry, emit_set_fp_mode): Hookify.
3097 2014-05-13 Jakub Jelinek <jakub@redhat.com>
3099 PR target/61060
3100 * config/i386/i386.c (ix86_expand_set_or_movmem): If count_exp
3101 is const0_rtx, return immediately. Don't test count == 0 when
3102 it is always true.
3104 2014-05-13 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3106 * Makefile.in: add shrink-wrap.o.
3107 * config/i386/i386.c: include "shrink-wrap.h"
3108 * function.c: Likewise.
3109 (requires_stack_frame_p, next_block_for_reg,
3110 move_insn_for_shrink_wrap, prepare_shrink_wrap,
3111 dup_block_and_redirect): Move to shrink-wrap.c
3112 (thread_prologue_and_epilogue_insns): Extract three code segments
3113 as functions in shrink-wrap.c
3114 * function.h: Move #ifdef HAVE_simple_return ... #endif block to
3115 shrink-wrap.h
3116 * shrink-wrap.c: New file.
3117 * shrink-wrap.h: New file.
3119 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
3121 * doc/extend.texi: Reflect current numbers of pragmas. Remove
3122 reference to Solaris.
3124 2014-05-12 Mike Stump <mikestump@comcast.net>
3126 PR other/31778
3127 * genattrtab.c (filename): Add.
3128 (convert_set_attr_alternative): Improve error message.
3129 (check_defs): Restore read_md_filename for error messages.
3130 (gen_insn): Save filename.
3132 2014-05-12 Dimitris Papavasiliou <dpapavas@gmail.com>
3134 * doc/invoke.texi: Document new switches -Wno-shadow-ivar,
3135 -fno-local-ivars and -fivar-visibility.
3136 * c-family/c.opt: Make -Wshadow also implicitly enable
3137 -Wshadow-ivar.
3139 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
3141 * doc/tm.texi: Remove reference to deleted macro.
3142 * doc/tm.texi.in: Likewise.
3144 2014-05-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
3146 PR target/60991
3147 * config/avr/avr.c (avr_out_store_psi): Use correct constant
3148 to restore Y.
3150 2014-05-12 Georg-Johann Lay <avr@gjlay.de>
3152 PR libgcc/61152
3153 * config/arm/arm.h (License): Add GCC Runtime Library Exception.
3154 * config/arm/aout.h (License): Same.
3155 * config/arm/bpabi.h (License): Same.
3156 * config/arm/elf.h (License): Same.
3157 * config/arm/linux-elf.h (License): Same.
3158 * config/arm/linux-gas.h (License): Same.
3159 * config/arm/netbsd-elf.h (License): Same.
3160 * config/arm/uclinux-eabi.h (License): Same.
3161 * config/arm/uclinux-elf.h (License): Same.
3162 * config/arm/vxworks.h (License): Same.
3164 2014-05-11 Jakub Jelinek <jakub@redhat.com>
3166 * tree.h (OMP_CLAUSE_LINEAR_STMT): Define.
3167 * tree.c (omp_clause_num_ops): Increase OMP_CLAUSE_LINEAR
3168 number of operands to 3.
3169 (walk_tree_1): Walk all operands of OMP_CLAUSE_LINEAR.
3170 * tree-nested.c (convert_nonlocal_omp_clauses,
3171 convert_local_omp_clauses): Handle OMP_CLAUSE_DEPEND.
3172 * gimplify.c (gimplify_scan_omp_clauses): Handle
3173 OMP_CLAUSE_LINEAR_STMT.
3174 * omp-low.c (lower_rec_input_clauses): Fix typo.
3175 (maybe_add_implicit_barrier_cancel, lower_omp_1): Add
3176 cast between Fortran boolean_type_node and C _Bool if
3177 needed.
3179 2014-05-11 Richard Sandiford <rdsandiford@googlemail.com>
3181 PR tree-optimization/61136
3182 * wide-int.h (multiple_of_p): Define a version that doesn't return
3183 the quotient.
3184 * fold-const.c (extract_muldiv_1): Use wi::multiple_of_p instead of an
3185 integer_zerop/const_binop pair.
3186 (multiple_of_p): Likewise, converting both operands to widest_int
3187 precision.
3189 2014-05-09 Teresa Johnson <tejohnson@google.com>
3191 * cgraphunit.c (analyze_functions): Use correct dump file.
3193 2014-05-09 Florian Weimer <fweimer@redhat.com>
3195 * cfgexpand.c (stack_protect_decl_p): New function, extracted from
3196 expand_used_vars.
3197 (stack_protect_return_slot_p): New function.
3198 (expand_used_vars): Call stack_protect_decl_p and
3199 stack_protect_return_slot_p for -fstack-protector-strong.
3201 2014-05-09 David Wohlferd <LimeGreenSocks@yahoo.com>
3202 Andrew Haley <aph@redhat.com>
3203 Richard Sandiford <rdsandiford@googlemail.com>
3205 * doc/extend.texi: Rewrite inline asm page / re-org asm-related
3206 pages.
3208 2014-05-09 Kenneth Zadeck <zadeck@naturalbridge.com>
3210 PR middle-end/61111
3211 * fold-const.c (fold_binary_loc): Changed width of mask.
3213 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
3215 * config/avr/avr-fixed.md (round<mode>3): Use -1U instead of -1 in
3216 unsigned int initializers for regno_in, regno_out.
3218 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
3220 PR target/61055
3221 * config/avr/avr.md (cc): Add new attribute set_vzn.
3222 (addqi3, addqq3, adduqq3, subqi3, subqq3, subuqq3, negqi2) [cc]:
3223 Set cc insn attribute to set_vzn instead of set_zn for alternatives
3224 with INC, DEC or NEG.
3225 * config/avr/avr.c (avr_notice_update_cc): Handle SET_VZN.
3226 (avr_out_plus_1): ADIW sets cc0 to CC_SET_CZN.
3227 INC, DEC and ADD+ADC set cc0 to CC_CLOBBER.
3229 2014-05-09 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3231 Revert:
3232 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3234 * wide-int.cc (UTItype): Define.
3235 (UDWtype): Define for appropriate W_TYPE_SIZE.
3237 2014-05-09 Richard Biener <rguenther@suse.de>
3239 * Makefile.in (GTFILES): Remove tree-ssa-propagate.c.
3240 * tree-ssa-propagate.c: Do not include gt-tree-ssa-propagate.h.
3241 (interesting_ssa_edges, varying_ssa_edges): Move out of GC space.
3242 (add_ssa_edge, process_ssa_edge_worklist, ssa_prop_init,
3243 ssa_propagate): Adjust.
3245 2014-05-08 Jeff Law <law@redhat.com>
3247 PR tree-optimization/61009
3248 * tree-ssa-threadedge.c (thread_through_normal_block): Return a
3249 tri-state rather than a boolean. When a block is too big to
3250 thread through, inform caller via negative return value.
3251 (thread_across_edge): If a block was too big for normal threading,
3252 then it's too big for a joiner too, so remove temporary equivalences
3253 and return immediately.
3255 2014-05-08 Manuel López-Ibáñez <manu@gcc.gnu.org>
3256 Matthias Klose <doko@ubuntu.com>
3258 PR driver/61106
3259 * optc-gen.awk: Fix option handling for -Wunused-parameter.
3261 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
3263 PR target/59952
3264 * config/i386/i386.c (PTA_HASWELL): Remove PTA_RTM.
3266 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
3268 PR target/61092
3269 * config/alpha/alpha.c: Include gimple-iterator.h.
3270 (alpha_gimple_fold_builtin): New function. Move
3271 ALPHA_BUILTIN_UMULH folding from ...
3272 (alpha_fold_builtin): ... here.
3273 (TARGET_GIMPLE_FOLD_BUILTIN): New define.
3275 2014-05-08 Wei Mi <wmi@google.com>
3277 PR target/58066
3278 * config/i386/i386.c (ix86_compute_frame_layout): Update
3279 preferred_stack_boundary for call, expanded from tls descriptor.
3280 * config/i386/i386.md (*tls_global_dynamic_32_gnu): Update RTX
3281 to depend on SP register.
3282 (*tls_local_dynamic_base_32_gnu): Ditto.
3283 (*tls_local_dynamic_32_once): Ditto.
3284 (tls_global_dynamic_64_<mode>): Set
3285 ix86_tls_descriptor_calls_expanded_in_cfun.
3286 (tls_local_dynamic_base_64_<mode>): Ditto.
3287 (tls_global_dynamic_32): Set
3288 ix86_tls_descriptor_calls_expanded_in_cfun. Update RTX
3289 to depend on SP register.
3290 (tls_local_dynamic_base_32): Ditto.
3292 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3294 * config/arm/arm_neon.h: Update comment.
3295 * config/arm/neon-docgen.ml: Delete.
3296 * config/arm/neon-gen.ml: Delete.
3297 * doc/arm-neon-intrinsics.texi: Update comment.
3299 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3301 * config/arm/arm_neon_builtins.def (vadd, vsub): Only define the v2sf
3302 and v4sf versions.
3303 (vand, vorr, veor, vorn, vbic): Remove.
3304 * config/arm/neon.md (neon_vadd, neon_vsub, neon_vadd_unspec): Adjust
3305 iterator.
3306 (neon_vsub_unspec): Likewise.
3307 (neon_vorr, neon_vand, neon_vbic, neon_veor, neon_vorn): Remove.
3309 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3311 * config/arm/arm_neon.h (vadd_s8): GNU C implementation
3312 (vadd_s16): Likewise.
3313 (vadd_s32): Likewise.
3314 (vadd_f32): Likewise.
3315 (vadd_u8): Likewise.
3316 (vadd_u16): Likewise.
3317 (vadd_u32): Likewise.
3318 (vadd_s64): Likewise.
3319 (vadd_u64): Likewise.
3320 (vaddq_s8): Likewise.
3321 (vaddq_s16): Likewise.
3322 (vaddq_s32): Likewise.
3323 (vaddq_s64): Likewise.
3324 (vaddq_f32): Likewise.
3325 (vaddq_u8): Likewise.
3326 (vaddq_u16): Likewise.
3327 (vaddq_u32): Likewise.
3328 (vaddq_u64): Likewise.
3329 (vmul_s8): Likewise.
3330 (vmul_s16): Likewise.
3331 (vmul_s32): Likewise.
3332 (vmul_f32): Likewise.
3333 (vmul_u8): Likewise.
3334 (vmul_u16): Likewise.
3335 (vmul_u32): Likewise.
3336 (vmul_p8): Likewise.
3337 (vmulq_s8): Likewise.
3338 (vmulq_s16): Likewise.
3339 (vmulq_s32): Likewise.
3340 (vmulq_f32): Likewise.
3341 (vmulq_u8): Likewise.
3342 (vmulq_u16): Likewise.
3343 (vmulq_u32): Likewise.
3344 (vsub_s8): Likewise.
3345 (vsub_s16): Likewise.
3346 (vsub_s32): Likewise.
3347 (vsub_f32): Likewise.
3348 (vsub_u8): Likewise.
3349 (vsub_u16): Likewise.
3350 (vsub_u32): Likewise.
3351 (vsub_s64): Likewise.
3352 (vsub_u64): Likewise.
3353 (vsubq_s8): Likewise.
3354 (vsubq_s16): Likewise.
3355 (vsubq_s32): Likewise.
3356 (vsubq_s64): Likewise.
3357 (vsubq_f32): Likewise.
3358 (vsubq_u8): Likewise.
3359 (vsubq_u16): Likewise.
3360 (vsubq_u32): Likewise.
3361 (vsubq_u64): Likewise.
3362 (vand_s8): Likewise.
3363 (vand_s16): Likewise.
3364 (vand_s32): Likewise.
3365 (vand_u8): Likewise.
3366 (vand_u16): Likewise.
3367 (vand_u32): Likewise.
3368 (vand_s64): Likewise.
3369 (vand_u64): Likewise.
3370 (vandq_s8): Likewise.
3371 (vandq_s16): Likewise.
3372 (vandq_s32): Likewise.
3373 (vandq_s64): Likewise.
3374 (vandq_u8): Likewise.
3375 (vandq_u16): Likewise.
3376 (vandq_u32): Likewise.
3377 (vandq_u64): Likewise.
3378 (vorr_s8): Likewise.
3379 (vorr_s16): Likewise.
3380 (vorr_s32): Likewise.
3381 (vorr_u8): Likewise.
3382 (vorr_u16): Likewise.
3383 (vorr_u32): Likewise.
3384 (vorr_s64): Likewise.
3385 (vorr_u64): Likewise.
3386 (vorrq_s8): Likewise.
3387 (vorrq_s16): Likewise.
3388 (vorrq_s32): Likewise.
3389 (vorrq_s64): Likewise.
3390 (vorrq_u8): Likewise.
3391 (vorrq_u16): Likewise.
3392 (vorrq_u32): Likewise.
3393 (vorrq_u64): Likewise.
3394 (veor_s8): Likewise.
3395 (veor_s16): Likewise.
3396 (veor_s32): Likewise.
3397 (veor_u8): Likewise.
3398 (veor_u16): Likewise.
3399 (veor_u32): Likewise.
3400 (veor_s64): Likewise.
3401 (veor_u64): Likewise.
3402 (veorq_s8): Likewise.
3403 (veorq_s16): Likewise.
3404 (veorq_s32): Likewise.
3405 (veorq_s64): Likewise.
3406 (veorq_u8): Likewise.
3407 (veorq_u16): Likewise.
3408 (veorq_u32): Likewise.
3409 (veorq_u64): Likewise.
3410 (vbic_s8): Likewise.
3411 (vbic_s16): Likewise.
3412 (vbic_s32): Likewise.
3413 (vbic_u8): Likewise.
3414 (vbic_u16): Likewise.
3415 (vbic_u32): Likewise.
3416 (vbic_s64): Likewise.
3417 (vbic_u64): Likewise.
3418 (vbicq_s8): Likewise.
3419 (vbicq_s16): Likewise.
3420 (vbicq_s32): Likewise.
3421 (vbicq_s64): Likewise.
3422 (vbicq_u8): Likewise.
3423 (vbicq_u16): Likewise.
3424 (vbicq_u32): Likewise.
3425 (vbicq_u64): Likewise.
3426 (vorn_s8): Likewise.
3427 (vorn_s16): Likewise.
3428 (vorn_s32): Likewise.
3429 (vorn_u8): Likewise.
3430 (vorn_u16): Likewise.
3431 (vorn_u32): Likewise.
3432 (vorn_s64): Likewise.
3433 (vorn_u64): Likewise.
3434 (vornq_s8): Likewise.
3435 (vornq_s16): Likewise.
3436 (vornq_s32): Likewise.
3437 (vornq_s64): Likewise.
3438 (vornq_u8): Likewise.
3439 (vornq_u16): Likewise.
3440 (vornq_u32): Likewise.
3441 (vornq_u64): Likewise.
3443 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3445 * wide-int.cc (UTItype): Define.
3446 (UDWtype): Define for appropriate W_TYPE_SIZE.
3448 2014-05-08 Marc Glisse <marc.glisse@inria.fr>
3450 PR tree-optimization/59100
3451 * tree-ssa-phiopt.c: Include tree-inline.h.
3452 (neutral_element_p, absorbing_element_p): New functions.
3453 (value_replacement): Handle conditional binary operations with a
3454 neutral or absorbing element.
3456 2014-05-08 Richard Biener <rguenther@suse.de>
3458 * tree-ssa-sccvn.c (vn_get_expr_for): Valueize operands before
3459 folding the expression.
3460 (valueize_expr): Remove.
3461 (visit_reference_op_load): Do not valueize the result of
3462 vn_get_expr_for.
3463 (simplify_binary_expression): Likewise.
3464 (simplify_unary_expression): Likewise.
3466 2014-05-08 Richard Biener <rguenther@suse.de>
3468 * gimplify.c (gimplify_call_expr): Use saved fnptrtype for
3469 looking at TYPE_ARG_TYPES.
3471 2014-05-08 Richard Biener <rguenther@suse.de>
3473 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove
3474 pointer propagation special-case.
3476 2014-05-08 Bin Cheng <bin.cheng@arm.com>
3478 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
3479 core part of address expressions.
3481 2014-05-08 Alan Modra <amodra@gmail.com>
3483 PR target/60737
3484 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
3485 loads and stores when -mno-strict-align at any alignment.
3486 (expand_block_clear): Similarly. Also correct calculation of
3487 instruction count.
3489 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
3491 PR middle-end/39246
3492 * tree-complex.c (expand_complex_move): Keep line info when expanding
3493 complex move.
3494 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
3495 of complex expression. Use new argument to display correct location
3496 for values coming from phi statement.
3497 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
3498 (warn_uninitialized_phi): Pass location of phi argument to
3499 warn_uninit.
3500 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
3501 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
3503 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
3505 * config/rs6000/predicates.md (indexed_address_mem): New.
3506 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
3507 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
3508 fpstore_ux, fpstore_u.
3509 (sign_extend, indexed, update): New.
3510 (cell_micro): Adjust.
3511 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
3512 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
3513 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
3514 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
3515 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
3516 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
3517 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
3518 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
3519 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
3520 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
3521 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
3522 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
3523 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
3524 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
3525 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
3527 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
3528 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
3529 *vsx_extract_<mode>_store): Adjust.
3530 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
3531 is_cracked_insn, insn_must_be_first_in_group,
3532 insn_must_be_last_in_group): Adjust.
3534 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
3535 Adjust.
3536 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
3537 ppc440-fpstore): Adjust.
3538 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
3539 ppc476-fpstore): Adjust.
3540 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
3541 ppc601-fpstore): Adjust.
3542 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
3543 Adjust.
3544 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
3545 Adjust.
3546 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
3547 ppc7450-fpstore): Adjust.
3548 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
3549 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
3550 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
3551 Adjust.
3552 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
3553 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
3554 cell-fpstore, cell-fpstore-update): Adjust.
3555 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
3556 ppce300c3_store, ppce300c3_fpstore): Adjust.
3557 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
3558 e500mc_fpstore): Adjust.
3559 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
3560 e500mc64_store, e500mc64_fpstore): Adjust.
3561 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
3562 e5500_fpstore): Adjust.
3563 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
3564 e6500_fpstore): Adjust.
3565 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
3566 Adjust.
3567 * config/rs6000/power4.md (power4-load, power4-load-ext,
3568 power4-load-ext-update, power4-load-ext-update-indexed,
3569 power4-load-update-indexed, power4-load-update, power4-fpload,
3570 power4-fpload-update, power4-store, power4-store-update,
3571 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
3572 Adjust.
3573 * config/rs6000/power5.md (power5-load, power5-load-ext,
3574 power5-load-ext-update, power5-load-ext-update-indexed,
3575 power5-load-update-indexed, power5-load-update, power5-fpload,
3576 power5-fpload-update, power5-store, power5-store-update,
3577 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
3578 Adjust.
3579 * config/rs6000/power6.md (power6-load, power6-load-ext,
3580 power6-load-update, power6-load-update-indexed,
3581 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
3582 power6-fpload-update, power6-store, power6-store-update,
3583 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
3584 Adjust.
3585 * config/rs6000/power7.md (power7-load, power7-load-ext,
3586 power7-load-update, power7-load-update-indexed,
3587 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
3588 power7-fpload-update, power7-store, power7-store-update,
3589 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
3590 Adjust.
3591 * config/rs6000/power8.md (power8-load, power8-load-update,
3592 power8-load-ext, power8-load-ext-update, power8-fpload,
3593 power8-fpload-update, power8-store, power8-store-update-indexed,
3594 power8-fpstore, power8-fpstore-update): Adjust.
3595 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
3596 Adjust.
3597 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
3598 titan_lsu_store, titan_lsu_fpstore): Adjust.
3599 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
3601 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
3603 PR target/60884
3604 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
3605 unrolled byte insns. Emit address increments after move insns.
3607 2014-05-07 David Malcolm <dmalcolm@redhat.com>
3609 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
3610 const_gimple, rather than a gimple.
3611 (gimple_call_builtin_p): Likewise, for the three variants.
3613 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
3614 (gimple_call_builtin_p): Likewise, for the three variants.
3616 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
3618 PR tree-optimization/61095
3619 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
3621 2014-05-07 Richard Biener <rguenther@suse.de>
3623 PR tree-optimization/61034
3624 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
3625 (maybe_skip_until): Use translate to take into account
3626 lattices when trying to do disambiguations.
3627 (get_continuation_for_phi_1): Likewise.
3628 (get_continuation_for_phi): Adjust for added translate arguments.
3629 (walk_non_aliased_vuses): Likewise.
3630 * tree-ssa-alias.h (get_continuation_for_phi): Adjust prototype.
3631 (walk_non_aliased_vuses): Likewise.
3632 (call_may_clobber_ref_p_1): Declare.
3633 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also disambiguate against
3634 calls. Stop early if we are only supposed to disambiguate.
3635 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
3637 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
3639 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
3640 Emit an error when the function has arguments.
3642 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
3644 * cfgloop.h (unswitch_loops): Remove.
3645 * doc/passes.texi: Remove references to loop-unswitch.c
3646 * timevar.def (TV_LOOP_UNSWITCH): Remove.
3648 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
3650 * tree-vect-data-refs.c (vect_grouped_load_supported): New
3651 check for loads group of length 3.
3652 (vect_permute_load_chain): New permutations for loads group of
3653 length 3.
3654 * tree-vect-stmts.c (vect_model_load_cost): Change cost
3655 of vec_perm_shuffle for the new permutations.
3657 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
3659 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
3660 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
3661 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
3662 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
3663 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
3664 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
3665 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
3666 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
3668 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
3670 * loop-unswitch.c: Delete.
3672 2014-05-07 Richard Biener <rguenther@suse.de>
3674 * config.gcc: Always set need_64bit_hwint to yes.
3676 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
3678 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
3679 of using optimize_size.
3681 2014-05-06 Mike Stump <mikestump@comcast.net>
3683 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
3685 2014-05-06 Joseph Myers <joseph@codesourcery.com>
3687 * config/i386/sse.md (*mov<mode>_internal)
3688 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
3689 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
3690 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
3691 (*<code><mode>3, *andnot<mode>3<mask_name>)
3692 (<mask_codefor><code><mode>3<mask_name>): Only consider
3693 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
3695 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
3697 Revert:
3698 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
3700 * lra-constraints.c (valid_address_p): Move earlier in file.
3701 Add a constraint argument to the address_info version.
3702 (satisfies_memory_constraint_p): New function.
3703 (satisfies_address_constraint_p): Likewise.
3704 (process_alt_operands, curr_insn_transform): Use them.
3705 (process_address): Pass the constraint to valid_address_p when
3706 checking address operands.
3708 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
3710 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
3711 to their respective blocks. Fix inadvertent use of "node".
3713 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
3715 * emit-rtl.c (init_derived_machine_modes): New functionm, split
3716 out from...
3717 (init_emit_once): ...here.
3718 * rtl.h (init_derived_machine_modes): Declare.
3719 * toplev.c (do_compile): Call it even if no_backend.
3721 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
3722 Mike Stump <mikestump@comcast.net>
3723 Richard Sandiford <rdsandiford@googlemail.com>
3724 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3726 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
3727 (rtx_equal_for_memref_p): Update comment.
3728 (adjust_offset_for_component_ref): Use wide-int interfaces.
3729 * builtins.c (get_object_alignment_2): Likewise.
3730 (c_readstr): Likewise.
3731 (target_char_cast): Add comment.
3732 (determine_block_size): Use wide-int interfaces.
3733 (expand_builtin_signbit): Likewise.
3734 (fold_builtin_int_roundingfn): Likewise.
3735 (fold_builtin_bitop): Likewise.
3736 (fold_builtin_bswap): Likewise.
3737 (fold_builtin_logarithm): Use signop.
3738 (fold_builtin_pow): Likewise.
3739 (fold_builtin_memory_op): Use wide-int interfaces.
3740 (fold_builtin_object_size): Likewise.
3741 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
3742 nb_iterations_estimate.
3743 (record_niter_bound): Use wide-int interfaces.
3744 (get_estimated_loop_iterations_int): Likewise.
3745 (get_estimated_loop_iterations): Likewise.
3746 (get_max_loop_iterations): Likewise.
3747 * cfgloop.h: Include wide-int.h.
3748 (struct nb_iter_bound): Change bound to widest_int.
3749 (struct loop): Change nb_iterations_upper_bound and
3750 nb_iterations_estimate to widest_int.
3751 (record_niter_bound): Switch to use widest_int.
3752 (get_estimated_loop_iterations): Likewise.
3753 (get_max_loop_iterations): Likewise.
3754 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
3755 update for wide-int.
3756 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
3757 * combine.c (try_combine): Likewise.
3758 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
3759 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
3760 interfaces.
3761 (aarch64_float_const_representable_p): Likewise.
3762 * config/arc/arc.c: Include wide-int.h.
3763 (arc_can_use_doloop_p): Use wide-int interfaces.
3764 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
3765 (vfp3_const_double_index): Likewise.
3766 * config/avr/avr.c (avr_out_round): Likewise.
3767 (avr_fold_builtin): Likewise.
3768 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
3769 (bfin_can_use_doloop_p): Likewise.
3770 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
3771 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
3772 * config/i386/i386.c: Include wide-int.h.
3773 (ix86_data_alignment): Use wide-int interfaces.
3774 (ix86_local_alignment): Likewise.
3775 (ix86_emit_swsqrtsf): Update real_from_integer.
3776 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
3777 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
3778 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
3779 (zero_constant): Likewise.
3780 (input_operand): Likewise.
3781 (splat_input_operand): Likewise.
3782 (non_logical_cint_operand): Change const_double to const_wide_int.
3783 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
3784 (easy_altivec_constant): Remove comment.
3785 (paired_expand_vector_init): Use CONSTANT_P.
3786 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
3787 (rs6000_emit_move): Update checks.
3788 (rs6000_aggregate_candidate): Use wide-int interfaces.
3789 (rs6000_expand_ternop_builtin): Likewise.
3790 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
3791 (rs6000_assemble_integer): Likewise.
3792 (rs6000_hash_constant): Likewise.
3793 (output_toc): Likewise.
3794 (rs6000_rtx_costs): Likewise.
3795 (rs6000_emit_swrsqrt); Update call to real_from_integer.
3796 * config/rs6000/rs6000-c.c: Include wide-int.h.
3797 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
3798 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
3799 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
3800 Handle CONST_WIDE_INT.
3801 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
3802 Use tree_fits_uhwi_p.
3803 * config/sparc/sparc.c: Include wide-int.h.
3804 (sparc_fold_builtin): Use wide-int interfaces.
3805 * config/vax/vax.c: Include wide-int.h.
3806 (vax_float_literal): Use real_from_integer.
3807 * coretypes.h (struct hwivec_def): New.
3808 (hwivec): New.
3809 (const_hwivec): New.
3810 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
3811 (equiv_constant): Handle CONST_WIDE_INT.
3812 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
3813 (cselib_hash_rtx): Handle CONST_WIDE_INT.
3814 * dbxout.c (stabstr_U): Use wide-int interfaces.
3815 (dbxout_type): Update to use cst_fits_shwi_p.
3816 * defaults.h (LOG2_BITS_PER_UNIT): Define.
3817 (TARGET_SUPPORTS_WIDE_INT): Add default.
3818 * dfp.c: Include wide-int.h.
3819 (decimal_real_to_integer2): Use wide-int interfaces and rename to
3820 decimal_real_to_integer.
3821 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
3822 decimal_real_to_integer.
3823 * doc/generic.texi (Constant expressions): Update for wide_int.
3824 * doc/rtl.texi (const_double): Likewise.
3825 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
3826 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
3827 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
3828 (REAL_VALUE_FROM_INT): Remove.
3829 (TARGET_SUPPORTS_WIDE_INT): New.
3830 * doc/tm.texi: Regenerate.
3831 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
3832 * double-int.h: Include wide-int.h.
3833 (struct wi::int_traits): New.
3834 * dwarf2out.c (get_full_len): New.
3835 (dw_val_equal_p): Add case dw_val_class_wide_int.
3836 (size_of_loc_descr): Likewise.
3837 (output_loc_operands): Likewise.
3838 (insert_double): Remove.
3839 (insert_wide_int): New.
3840 (add_AT_wide): New.
3841 (print_die): Add case dw_val_class_wide_int.
3842 (attr_checksum): Likewise.
3843 (attr_checksum_ordered): Likewise.
3844 (same_dw_val_p): Likewise.
3845 (size_of_die): Likewise.
3846 (value_format): Likewise.
3847 (output_die): Likewise.
3848 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
3849 Use wide-int.
3850 (clz_loc_descriptor): Use wide-int interfaces.
3851 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
3852 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
3853 (round_up_to_align): Use wide-int interfaces.
3854 (field_byte_offset): Likewise.
3855 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
3856 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
3857 CONST_DOUBLE handling. Use wide-int interfaces.
3858 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
3859 (gen_enumeration_type_die): Use add_AT_wide.
3860 (hash_loc_operands): Add case dw_val_class_wide_int.
3861 (compare_loc_operands): Likewise.
3862 * dwarf2out.h: Include wide-int.h.
3863 (wide_int_ptr): New.
3864 (enum dw_val_class): Add dw_val_class_wide_int.
3865 (struct dw_val_struct): Add val_wide.
3866 * emit-rtl.c (const_wide_int_htab): New.
3867 (const_wide_int_htab_hash): New.
3868 (const_wide_int_htab_eq): New.
3869 (lookup_const_wide_int): New.
3870 (const_double_htab_hash): Use wide-int interfaces.
3871 (const_double_htab_eq): Likewise.
3872 (rtx_to_double_int): Conditionally compile for wide-int.
3873 (immed_double_int_const): Rename to immed_wide_int_const and
3874 update for wide-int.
3875 (immed_double_const): Conditionally compile for wide-int.
3876 (init_emit_once): Use wide-int interfaces.
3877 * explow.c (plus_constant): Likewise.
3878 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
3879 (lshift_value): Use wide-int interfaces.
3880 (expand_mult): Likewise.
3881 (choose_multiplier): Likewise.
3882 (expand_smod_pow2): Likewise.
3883 (make_tree): Likewise.
3884 * expr.c (convert_modes): Consolidate handling of constants.
3885 Use wide-int interfaces.
3886 (emit_group_load_1): Add note.
3887 (store_expr): Update comment.
3888 (get_inner_reference): Use wide-int interfaces.
3889 (expand_constructor): Update comment.
3890 (expand_expr_real_2): Use wide-int interfaces.
3891 (expand_expr_real_1): Likewise.
3892 (reduce_to_bit_field_precision): Likewise.
3893 (const_vector_from_tree): Likewise.
3894 * final.c: Include wide-int-print.h.
3895 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
3896 * fixed-value.c: Include wide-int.h.
3897 (fixed_from_string): Use wide-int interfaces.
3898 (fixed_to_decimal): Likewise.
3899 (fixed_convert_from_real): Likewise.
3900 (real_convert_from_fixed): Likewise.
3901 * fold-const.h (mem_ref_offset): Return an offset_int.
3902 (div_if_zero_remainder): Remove code parameter.
3903 * fold-const.c (div_if_zero_remainder): Remove code parameter.
3904 Use wide-int interfaces.
3905 (may_negate_without_overflow_p): Use wide-int interfaces.
3906 (negate_expr_p): Likewise.
3907 (fold_negate_expr): Likewise.
3908 (int_const_binop_1): Likewise.
3909 (const_binop): Likewise.
3910 (fold_convert_const_int_from_int): Likewise.
3911 (fold_convert_const_int_from_real): Likewise.
3912 (fold_convert_const_int_from_fixed): Likewise.
3913 (fold_convert_const_fixed_from_int): Likewise.
3914 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
3915 (sign_bit_p): Use wide-int interfaces.
3916 (make_range_step): Likewise.
3917 (build_range_check): Likewise. Pass an integer of the correct type
3918 instead of using integer_one_node.
3919 (range_predecessor): Pass an integer of the correct type instead
3920 of using integer_one_node.
3921 (range_successor): Likewise.
3922 (merge_ranges): Likewise.
3923 (unextend): Use wide-int interfaces.
3924 (extract_muldiv_1): Likewise.
3925 (fold_div_compare): Likewise.
3926 (fold_single_bit_test): Likewise.
3927 (fold_sign_changed_comparison): Likewise.
3928 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
3929 (fold_plusminus_mult_expr): Use wide-int interfaces.
3930 (native_encode_int): Likewise.
3931 (native_interpret_int): Likewise.
3932 (fold_unary_loc): Likewise.
3933 (pointer_may_wrap_p): Likewise.
3934 (size_low_cst): Likewise.
3935 (mask_with_tz): Likewise.
3936 (fold_binary_loc): Likewise.
3937 (fold_ternary_loc): Likewise.
3938 (multiple_of_p): Likewise.
3939 (tree_call_nonnegative_warnv_p): Update calls to
3940 tree_int_cst_min_precision and real_from_integer.
3941 (fold_negate_const): Use wide-int interfaces.
3942 (fold_abs_const): Likewise.
3943 (fold_relational_const): Use tree_int_cst_lt.
3944 (round_up_loc): Use wide-int interfaces.
3945 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
3946 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
3947 * gengtype.c: Remove include of double-int.h.
3948 (do_typedef): Use wide-int interfaces.
3949 (open_base_files): Add wide-int.h.
3950 (main): Add offset_int and widest_int typedefs.
3951 * gengtype-lex.l: Handle "^".
3952 (CXX_KEYWORD): Add "static".
3953 * gengtype-parse.c (require3): New.
3954 (require_template_declaration): Handle constant template arguments
3955 and nested templates.
3956 * gengtype-state.c: Don't include "double-int.h".
3957 * genpreds.c (write_one_predicate_function): Update comment.
3958 (write_tm_constrs_h): Add check for hval and lval use in
3959 CONST_WIDE_INT.
3960 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
3961 (add_to_sequence): Likewise.
3962 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
3963 and const_double_operand.
3964 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
3965 interfaces.
3966 * gimple-fold.c (get_base_constructor): Likewise.
3967 (fold_array_ctor_reference): Likewise.
3968 (fold_nonarray_ctor_reference): Likewise.
3969 (fold_const_aggregate_ref_1): Likewise.
3970 (gimple_val_nonnegative_real_p): Likewise.
3971 (gimple_fold_indirect_ref): Likewise.
3972 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
3973 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
3974 (struct slsr_cand_d): Change index to be widest_int.
3975 (struct incr_info_d): Change incr to be widest_int.
3976 (alloc_cand_and_find_basis): Use wide-int interfaces.
3977 (slsr_process_phi): Likewise.
3978 (backtrace_base_for_ref): Likewise. Return a widest_int.
3979 (restructure_reference): Take a widest_int instead of a double_int.
3980 (slsr_process_ref): Use wide-int interfaces.
3981 (create_mul_ssa_cand): Likewise.
3982 (create_mul_imm_cand): Likewise.
3983 (create_add_ssa_cand): Likewise.
3984 (create_add_imm_cand): Take a widest_int instead of a double_int.
3985 (slsr_process_add): Use wide-int interfaces.
3986 (slsr_process_cast): Likewise.
3987 (slsr_process_copy): Likewise.
3988 (dump_candidate): Likewise.
3989 (dump_incr_vec): Likewise.
3990 (replace_ref): Likewise.
3991 (cand_increment): Likewise. Return a widest_int.
3992 (cand_abs_increment): Likewise.
3993 (replace_mult_candidate): Take a widest_int instead of a double_int.
3994 (replace_unconditional_candidate): Use wide-int interfaces.
3995 (incr_vec_index): Take a widest_int instead of a double_int.
3996 (create_add_on_incoming_edge): Likewise.
3997 (create_phi_basis): Use wide-int interfaces.
3998 (replace_conditional_candidate): Likewise.
3999 (record_increment): Take a widest_int instead of a double_int.
4000 (record_phi_increments): Use wide-int interfaces.
4001 (phi_incr_cost): Take a widest_int instead of a double_int.
4002 (lowest_cost_path): Likewise.
4003 (total_savings): Likewise.
4004 (analyze_increments): Use wide-int interfaces.
4005 (ncd_with_phi): Take a widest_int instead of a double_int.
4006 (ncd_of_cand_and_phis): Likewise.
4007 (nearest_common_dominator_for_cands): Likewise.
4008 (insert_initializers): Use wide-int interfaces.
4009 (all_phi_incrs_profitable): Likewise.
4010 (replace_one_candidate): Likewise.
4011 (replace_profitable_candidates): Likewise.
4012 * godump.c: Include wide-int-print.h.
4013 (go_output_typedef): Use wide-int interfaces.
4014 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
4015 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
4016 (build_loop_iteration_domains): Likewise.
4017 * hooks.h: Include wide-int.h rather than double-int.h.
4018 (hook_bool_dint_dint_uint_bool_true): Delete.
4019 (hook_bool_wint_wint_uint_bool_true): Declare.
4020 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
4021 (hook_bool_wint_wint_uint_bool_true): New.
4022 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
4023 interfaces.
4024 (ubsan_expand_si_overflow_mul_check): Likewise.
4025 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
4026 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
4027 (get_ancestor_addr_info): Likewise.
4028 (ipa_modify_call_arguments): Likewise.
4029 * loop-doloop.c (doloop_modify): Likewise.
4030 (doloop_optimize): Likewise.
4031 * loop-iv.c (iv_number_of_iterations): Likewise.
4032 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
4033 (unroll_loop_constant_iterations): Likewise.
4034 (decide_unroll_runtime_iterations): Likewise.
4035 (unroll_loop_runtime_iterations): Likewise.
4036 (decide_peel_simple): Likewise.
4037 (decide_unroll_stupid): Likewise.
4038 * lto-streamer-in.c (streamer_read_wi): Add.
4039 (input_cfg): Use wide-int interfaces.
4040 (lto_input_tree_1): Likewise.
4041 * lto-streamer-out.c (streamer_write_wi): Add.
4042 (hash_tree): Use wide-int interfaces.
4043 (output_cfg): Likewise.
4044 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
4045 (GTFILES): Add wide-int.h and signop.h.
4046 (TAGS): Look for .cc files too.
4047 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
4048 * optabs.c (expand_subword_shift): Likewise.
4049 (expand_doubleword_shift): Likewise.
4050 (expand_absneg_bit): Likewise.
4051 (expand_copysign_absneg): Likewise.
4052 (expand_copysign_bit): Likewise.
4053 * postreload.c (reload_cse_simplify_set): Likewise.
4054 * predict.c (predict_iv_comparison): Likewise.
4055 * pretty-print.h: Include wide-int-print.h.
4056 (pp_wide_int) New.
4057 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
4058 * print-tree.c: Include wide-int-print.h.
4059 (print_node_brief): Use wide-int interfaces.
4060 (print_node): Likewise.
4061 * read-rtl.c (validate_const_wide_int): New.
4062 (read_rtx_code): Add CONST_WIDE_INT case.
4063 * real.c: Include wide-int.h.
4064 (real_to_integer2): Delete.
4065 (real_to_integer): New function, returning a wide_int.
4066 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
4067 (ten_to_ptwo): Update call to real_from_integer.
4068 (real_digit): Likewise.
4069 * real.h: Include signop.h, wide-int.h and insn-modes.h.
4070 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
4071 (REAL_VALUE_TO_INT): Delete.
4072 (real_to_integer): Declare a wide-int form.
4073 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
4074 * recog.c (const_int_operand): Improve comment.
4075 (const_scalar_int_operand): New.
4076 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
4077 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
4078 (split_double): Likewise.
4079 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
4080 (rtx_size): Likewise.
4081 (rtx_alloc_stat_v): New.
4082 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
4083 (cwi_output_hex): New.
4084 (iterative_hash_rtx): Handle CONST_WIDE_INT.
4085 (cwi_check_failed_bounds): New.
4086 * rtl.def (CONST_WIDE_INT): New.
4087 * rtl.h: Include <utility> and wide-int.h.
4088 (struct hwivec_def): New.
4089 (CWI_GET_NUM_ELEM): New.
4090 (CWI_PUT_NUM_ELEM): New.
4091 (struct rtx_def): Add num_elem and hwiv.
4092 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
4093 (CASE_CONST_UNIQUE): Likewise.
4094 (CASE_CONST_ANY): Likewise.
4095 (CONST_SCALAR_INT_P): Likewise.
4096 (CONST_WIDE_INT_P): New.
4097 (CWI_ELT): New.
4098 (HWIVEC_CHECK): New.
4099 (cwi_check_failed_bounds): New.
4100 (CWI_ELT): New.
4101 (HWIVEC_CHECK): New.
4102 (CONST_WIDE_INT_VEC) New.
4103 (CONST_WIDE_INT_NUNITS) New.
4104 (CONST_WIDE_INT_ELT) New.
4105 (rtx_mode_t): New type.
4106 (wi::int_traits <rtx_mode_t>): New.
4107 (wi::shwi): New.
4108 (wi::min_value): New.
4109 (wi::max_value): New.
4110 (rtx_alloc_v) New.
4111 (const_wide_int_alloc): New.
4112 (immed_wide_int_const): New.
4113 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
4114 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
4115 * signop.h: New file.
4116 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
4117 (simplify_const_unary_operation): Use wide-int interfaces.
4118 (simplify_binary_operation_1): Likewise.
4119 (simplify_const_binary_operation): Likewise.
4120 (simplify_const_relational_operation): Likewise.
4121 (simplify_immed_subreg): Likewise.
4122 * stmt.c (expand_case): Likewise.
4123 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
4124 signop rather than a bool.
4125 * stor-layout.c (layout_type): Use wide-int interfaces.
4126 (initialize_sizetypes): Update calls to
4127 set_min_and_max_values_for_integral_type.
4128 (set_min_and_max_values_for_integral_type): Take a signop rather
4129 than a bool. Use wide-int interfaces.
4130 (fixup_signed_type): Update accordingly. Remove
4131 HOST_BITS_PER_DOUBLE_INT limit.
4132 (fixup_unsigned_type): Likewise.
4133 * system.h (STATIC_CONSTANT_P): New.
4134 (STATIC_ASSERT): New.
4135 * target.def (can_use_doloop_p): Take widest_ints rather than
4136 double_ints.
4137 * target.h: Include wide-int.h rather than double-int.h.
4138 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
4139 than double_ints.
4140 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
4141 rather than INT_CST_LT_UNSIGNED.
4142 (can_use_doloop_if_innermost): Take widest_ints rather than
4143 double_ints.
4144 * tree-affine.c: Include wide-int-print.h.
4145 (double_int_ext_for_comb): Delete.
4146 (wide_int_ext_for_comb): New.
4147 (aff_combination_zero): Use wide-int interfaces.
4148 (aff_combination_const): Take a widest_int instead of a double_int.
4149 (aff_combination_elt): Use wide-int interfaces.
4150 (aff_combination_scale): Take a widest_int instead of a double_int.
4151 (aff_combination_add_elt): Likewise.
4152 (aff_combination_add_cst): Likewise.
4153 (aff_combination_add): Use wide-int interfaces.
4154 (aff_combination_convert): Likewise.
4155 (tree_to_aff_combination): Likewise.
4156 (add_elt_to_tree): Take a widest_int instead of a double_int.
4157 (aff_combination_to_tree): Use wide-int interfaces.
4158 (aff_combination_remove_elt): Likewise.
4159 (aff_combination_add_product): Take a widest_int instead of
4160 a double_int.
4161 (aff_combination_mult): Use wide-int interfaces.
4162 (aff_combination_expand): Likewise.
4163 (double_int_constant_multiple_p): Delete.
4164 (wide_int_constant_multiple_p): New.
4165 (aff_combination_constant_multiple_p): Take a widest_int pointer
4166 instead of a double_int pointer.
4167 (print_aff): Use wide-int interfaces.
4168 (get_inner_reference_aff): Take a widest_int pointer
4169 instead of a double_int pointer.
4170 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
4171 * tree-affine.h: Include wide-int.h.
4172 (struct aff_comb_elt): Change type of coef to widest_int.
4173 (struct affine_tree_combination): Change type of offset to widest_int.
4174 (double_int_ext_for_comb): Delete.
4175 (wide_int_ext_for_comb): New.
4176 (aff_combination_const): Use widest_int instead of double_int.
4177 (aff_combination_scale): Likewise.
4178 (aff_combination_add_elt): Likewise.
4179 (aff_combination_constant_multiple_p): Likewise.
4180 (get_inner_reference_aff): Likewise.
4181 (aff_comb_cannot_overlap_p): Likewise.
4182 (aff_combination_zero_p): Use wide-int interfaces.
4183 * tree.c: Include tree.h.
4184 (init_ttree): Use make_int_cst.
4185 (tree_code_size): Removed code for INTEGER_CST case.
4186 (tree_size): Add INTEGER_CST case.
4187 (make_node_stat): Update comment.
4188 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
4189 (build_int_cst_type): Use wide-int interfaces.
4190 (double_int_to_tree): Likewise.
4191 (double_int_fits_to_tree_p): Delete.
4192 (force_fit_type_double): Delete.
4193 (force_fit_type): New.
4194 (int_cst_hash_hash): Use wide-int interfaces.
4195 (int_cst_hash_eq): Likewise.
4196 (build_int_cst_wide): Delete.
4197 (wide_int_to_tree): New.
4198 (cache_integer_cst): Use wide-int interfaces.
4199 (build_low_bits_mask): Likewise.
4200 (cst_and_fits_in_hwi): Likewise.
4201 (real_value_from_int_cst): Likewise.
4202 (make_int_cst_stat): New.
4203 (integer_zerop): Use wide_int interfaces.
4204 (integer_onep): Likewise.
4205 (integer_all_onesp): Likewise.
4206 (integer_pow2p): Likewise.
4207 (integer_nonzerop): Likewise.
4208 (tree_log2): Likewise.
4209 (tree_floor_log2): Likewise.
4210 (tree_ctz): Likewise.
4211 (int_size_in_bytes): Likewise.
4212 (mem_ref_offset): Return an offset_int rather than a double_int.
4213 (build_type_attribute_qual_variant): Use wide_int interfaces.
4214 (type_hash_eq): Likewise
4215 (tree_int_cst_equal): Likewise.
4216 (tree_int_cst_lt): Delete.
4217 (tree_int_cst_compare): Likewise.
4218 (tree_fits_shwi_p): Use wide_int interfaces.
4219 (tree_fits_uhwi_p): Likewise.
4220 (tree_int_cst_sign_bit): Likewise.
4221 (tree_int_cst_sgn): Likewise.
4222 (tree_int_cst_min_precision): Take a signop rather than a bool.
4223 (simple_cst_equal): Use wide_int interfaces.
4224 (compare_tree_int): Likewise.
4225 (iterative_hash_expr): Likewise.
4226 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
4227 INT_CST_LT.
4228 (get_type_static_bounds): Use wide_int interfaces.
4229 (tree_int_cst_elt_check_failed): New.
4230 (build_common_tree_nodes): Reordered to set prec before filling in
4231 value.
4232 (int_cst_value): Check cst_and_fits_in_hwi.
4233 (widest_int_cst_value): Use wide_int interfaces.
4234 (upper_bound_in_type): Likewise.
4235 (lower_bound_in_type): Likewise.
4236 (num_ending_zeros): Likewise.
4237 (drop_tree_overflow): Likewise.
4238 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
4239 (gen_conditions_for_pow_cst_base): Likewise.
4240 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
4241 (group_case_labels_stmt): Use wide-int interfaces.
4242 (verify_gimple_assign_binary): Likewise.
4243 (print_loop): Likewise.
4244 * tree-chrec.c (tree_fold_binomial): Likewise.
4245 * tree-core.h (struct tree_base): Add int_length.
4246 (struct tree_int_cst): Change rep of value.
4247 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
4248 (dr_may_alias_p): Likewise.
4249 (max_stmt_executions_tree): Likewise.
4250 * tree.def (INTEGER_CST): Update comment.
4251 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
4252 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
4253 * tree-dump.c: Include wide-int.h and wide-int-print.h.
4254 (dequeue_and_dump): Use wide-int interfaces.
4255 * tree.h: Include wide-int.h.
4256 (NULL_TREE): Moved to earlier loc in file.
4257 (TREE_INT_CST_ELT_CHECK): New.
4258 (tree_int_cst_elt_check_failed): New.
4259 (TYPE_SIGN): New.
4260 (TREE_INT_CST): Delete.
4261 (TREE_INT_CST_LOW): Use wide-int interfaces.
4262 (TREE_INT_CST_HIGH): Delete.
4263 (TREE_INT_CST_NUNITS): New.
4264 (TREE_INT_CST_EXT_NUNITS): Likewise.
4265 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
4266 (TREE_INT_CST_ELT): Likewise.
4267 (INT_CST_LT): Delete.
4268 (tree_int_cst_elt_check): New (two forms).
4269 (type_code_size): Update comment.
4270 (make_int_cst_stat, make_int_cst): New.
4271 (tree_to_double_int): Delete.
4272 (double_int_fits_to_tree_p): Delete.
4273 (force_fit_type_double): Delete.
4274 (build_int_cstu): Replace with out-of-line function.
4275 (build_int_cst_wide): Delete.
4276 (tree_int_cst_lt): Define inline.
4277 (tree_int_cst_le): New.
4278 (tree_int_cst_compare): Define inline.
4279 (tree_int_cst_min_precision): Take a signop rather than a bool.
4280 (wi::int_traits <const_tree>): New.
4281 (wi::int_traits <tree>): New.
4282 (wi::extended_tree): New.
4283 (wi::int_traits <wi::extended_tree>): New.
4284 (wi::to_widest): New.
4285 (wi::to_offset): New.
4286 (wi::fits_to_tree_p): New.
4287 (wi::min_value): New.
4288 (wi::max_value): New.
4289 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
4290 (copy_tree_body_r): Likewise.
4291 * tree-object-size.c (compute_object_offset): Likewise.
4292 (addr_object_size): Likewise.
4293 * tree-predcom.c: Include wide-int-print.h.
4294 (struct dref_d): Change type of offset to widest_int.
4295 (dump_dref): Call wide-int printer.
4296 (aff_combination_dr_offset): Use wide-int interfaces.
4297 (determine_offset): Take a widest_int pointer rather than a
4298 double_int pointer.
4299 (split_data_refs_to_components): Use wide-int interfaces.
4300 (suitable_component_p): Likewise.
4301 (order_drefs): Likewise.
4302 (add_ref_to_chain): Likewise.
4303 (valid_initializer_p): Likewise.
4304 (determine_roots_comp): Likewise.
4305 * tree-pretty-print.c: Include wide-int-print.h.
4306 (dump_generic_node): Use wide-int interfaces.
4307 * tree-sra.c (sra_ipa_modify_expr): Likewise.
4308 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
4309 (move_fixed_address_to_symbol): Likewise.
4310 (move_hint_to_base): Likewise.
4311 (move_pointer_to_base): Likewise.
4312 (move_variant_to_index): Likewise.
4313 (most_expensive_mult_to_index): Likewise.
4314 (addr_to_parts): Likewise.
4315 (copy_ref_info): Likewise.
4316 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
4317 (indirect_refs_may_alias_p): Likewise.
4318 (stmt_kills_ref_p_1): Likewise.
4319 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
4320 * tree-ssa-ccp.c: Update comment at top of file. Include
4321 wide-int-print.h.
4322 (struct prop_value_d): Change type of mask to widest_int.
4323 (extend_mask): New function.
4324 (dump_lattice_value): Use wide-int interfaces.
4325 (get_default_value): Likewise.
4326 (set_constant_value): Likewise.
4327 (set_value_varying): Likewise.
4328 (valid_lattice_transition): Likewise.
4329 (set_lattice_value): Likewise.
4330 (value_to_double_int): Delete.
4331 (value_to_wide_int): New.
4332 (get_value_from_alignment): Use wide-int interfaces.
4333 (get_value_for_expr): Likewise.
4334 (do_dbg_cnt): Likewise.
4335 (ccp_finalize): Likewise.
4336 (ccp_lattice_meet): Likewise.
4337 (bit_value_unop_1): Use widest_ints rather than double_ints.
4338 (bit_value_binop_1): Likewise.
4339 (bit_value_unop): Use wide-int interfaces.
4340 (bit_value_binop): Likewise.
4341 (bit_value_assume_aligned): Likewise.
4342 (evaluate_stmt): Likewise.
4343 (ccp_fold_stmt): Likewise.
4344 (visit_cond_stmt): Likewise.
4345 (ccp_visit_stmt): Likewise.
4346 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
4347 (constant_pointer_difference): Likewise.
4348 (associate_pointerplus): Likewise.
4349 (combine_conversions): Likewise.
4350 * tree-ssa-loop.h: Include wide-int.h.
4351 (struct tree_niter_desc): Change type of max to widest_int.
4352 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
4353 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
4354 (remove_redundant_iv_tests): Likewise.
4355 (canonicalize_loop_induction_variables): Likewise.
4356 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
4357 (constant_multiple_of): Take a widest_int pointer instead of
4358 a double_int pointer.
4359 (get_computation_aff): Use wide-int interfaces.
4360 (ptr_difference_cost): Likewise.
4361 (difference_cost): Likewise.
4362 (get_loop_invariant_expr_id): Likewise.
4363 (get_computation_cost_at): Likewise.
4364 (iv_elimination_compare_lt): Likewise.
4365 (may_eliminate_iv): Likewise.
4366 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
4367 instead of double_int.
4368 (max_loop_iterations): Likewise.
4369 (max_stmt_executions): Likewise.
4370 (estimated_stmt_executions): Likewise.
4371 * tree-ssa-loop-niter.c: Include wide-int-print.h.
4372 (split_to_var_and_offset): Use wide-int interfaces.
4373 (determine_value_range): Likewise.
4374 (bound_difference_of_offsetted_base): Likewise.
4375 (bounds_add): Take a widest_int instead of a double_int.
4376 (number_of_iterations_ne_max): Use wide-int interfaces.
4377 (number_of_iterations_ne): Likewise.
4378 (number_of_iterations_lt_to_ne): Likewise.
4379 (assert_loop_rolls_lt): Likewise.
4380 (number_of_iterations_lt): Likewise.
4381 (number_of_iterations_le): Likewise.
4382 (number_of_iterations_cond): Likewise.
4383 (number_of_iterations_exit): Likewise.
4384 (finite_loop_p): Likewise.
4385 (derive_constant_upper_bound_assign): Likewise.
4386 (derive_constant_upper_bound): Return a widest_int.
4387 (derive_constant_upper_bound_ops): Likewise.
4388 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
4389 (record_estimate): Take a widest_int rather than a double_int.
4390 (record_nonwrapping_iv): Use wide-int interfaces.
4391 (double_int_cmp): Delete.
4392 (wide_int_cmp): New.
4393 (bound_index): Take a widest_int rather than a double_int.
4394 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
4395 (maybe_lower_iteration_bound): Likewise.
4396 (estimate_numbers_of_iterations_loop): Likewise.
4397 (estimated_loop_iterations): Take a widest_int pointer than than
4398 a double_int pointer.
4399 (estimated_loop_iterations_int): Use wide-int interfaces.
4400 (max_loop_iterations): Take a widest_int pointer than than
4401 a double_int pointer.
4402 (max_loop_iterations_int): Use wide-int interfaces.
4403 (max_stmt_executions): Take a widest_int pointer than than
4404 a double_int pointer.
4405 (estimated_stmt_executions): Likewise.
4406 (n_of_executions_at_most): Use wide-int interfaces.
4407 (scev_probably_wraps_p): Likewise.
4408 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
4409 to real_to_integer.
4410 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
4411 interfaces.
4412 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
4413 double_ints. Adjust for trailing_wide_ints <3> representation.
4414 (set_nonzero_bits): Likewise.
4415 (get_range_info): Return wide_ints rather than double_ints.
4416 Adjust for trailing_wide_ints <3> representation.
4417 (get_nonzero_bits): Likewise.
4418 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
4419 representation.
4420 * tree-ssanames.h (struct range_info_def): Replace min, max and
4421 nonzero_bits with a trailing_wide_ints <3>.
4422 (set_range_info): Use wide_int_refs rather than double_ints.
4423 (set_nonzero_bits): Likewise.
4424 (get_range_info): Return wide_ints rather than double_ints.
4425 (get_nonzero_bits): Likewise.
4426 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
4427 * tree-ssa-pre.c (phi_translate_1): Likewise.
4428 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
4429 (acceptable_pow_call): Likewise.
4430 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
4431 interfaces.
4432 (vn_reference_fold_indirect): Likewise.
4433 (vn_reference_maybe_forwprop_address): Likewise.
4434 (valueize_refs_1): Likewise.
4435 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
4436 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
4437 tree_int_cst_lt and tree_int_cst_le.
4438 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
4439 interfaces.
4440 (streamer_alloc_tree): Likewise.
4441 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
4442 (streamer_write_tree_header): Likewise.
4443 (streamer_write_integer_cst): Likewise.
4444 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
4445 (build_constructors): Likewise.
4446 (array_value_type): Likewise.
4447 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
4448 (vect_check_gather): Likewise.
4449 * tree-vect-generic.c (build_replicated_const): Likewise.
4450 (expand_vector_divmod): Likewise.
4451 * tree-vect-loop.c (vect_transform_loop): Likewise.
4452 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
4453 (vect_do_peeling_for_alignment): Likewise.
4454 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
4455 * tree-vrp.c: Include wide-int.h.
4456 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
4457 (extract_range_from_assert): Use wide-int interfaces.
4458 (vrp_int_const_binop): Likewise.
4459 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
4460 double_int pointers.
4461 (ranges_from_anti_range): Use wide-int interfaces.
4462 (quad_int_cmp): Delete.
4463 (quad_int_pair_sort): Likewise.
4464 (extract_range_from_binary_expr_1): Use wide-int interfaces.
4465 (extract_range_from_unary_expr_1): Likewise.
4466 (adjust_range_with_scev): Likewise.
4467 (masked_increment): Take and return wide_ints rather than double_ints.
4468 (register_edge_assert_for_2): Use wide-int interfaces.
4469 (check_array_ref): Likewise.
4470 (search_for_addr_array): Likewise.
4471 (maybe_set_nonzero_bits): Likewise.
4472 (union_ranges): Pass an integer of the correct type instead of
4473 using integer_one_node.
4474 (intersect_ranges): Likewise.
4475 (simplify_truth_ops_using_ranges): Likewise.
4476 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
4477 (range_fits_type_p): Likewise.
4478 (simplify_cond_using_ranges): Likewise. Take a signop rather than
4479 a bool.
4480 (simplify_conversion_using_ranges): Use wide-int interfaces.
4481 (simplify_float_conversion_using_ranges): Likewise.
4482 (vrp_finalize): Likewise.
4483 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
4484 (gimple_stringops_transform): Likewise.
4485 * varasm.c (decode_addr_const): Likewise.
4486 (const_hash_1): Likewise.
4487 (const_rtx_hash_1): Likewise
4488 (output_constant): Likewise.
4489 (array_size_for_constructor): Likewise.
4490 (output_constructor_regular_field): Likewise.
4491 (output_constructor_bitfield): Likewise.
4492 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
4493 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
4494 GENERATOR_FILEs.
4495 * gencheck.c: Define BITS_PER_UNIT.
4496 * wide-int.cc: New.
4497 * wide-int.h: New.
4498 * wide-int-print.cc: New.
4499 * wide-int-print.h: New.
4501 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4503 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
4505 2014-05-06 Richard Biener <rguenther@suse.de>
4507 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
4508 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
4509 (TODO_verify_all): Adjust.
4510 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
4511 TODO_verify_stmts and TODO_verify_rtl_sharing.
4512 * bb-reorder.c: Likewise.
4513 * cfgexpand.c: Likewise.
4514 * cprop.c: Likewise.
4515 * cse.c: Likewise.
4516 * function.c: Likewise.
4517 * fwprop.c: Likewise.
4518 * gcse.c: Likewise.
4519 * gimple-ssa-isolate-paths.c: Likewise.
4520 * gimple-ssa-strength-reduction.c: Likewise.
4521 * ipa-split.c: Likewise.
4522 * loop-init.c: Likewise.
4523 * loop-unroll.c: Likewise.
4524 * lower-subreg.c: Likewise.
4525 * modulo-sched.c: Likewise.
4526 * postreload-gcse.c: Likewise.
4527 * predict.c: Likewise.
4528 * recog.c: Likewise.
4529 * sched-rgn.c: Likewise.
4530 * store-motion.c: Likewise.
4531 * tracer.c: Likewise.
4532 * trans-mem.c: Likewise.
4533 * tree-call-cdce.c: Likewise.
4534 * tree-cfg.c: Likewise.
4535 * tree-cfgcleanup.c: Likewise.
4536 * tree-complex.c: Likewise.
4537 * tree-eh.c: Likewise.
4538 * tree-emutls.c: Likewise.
4539 * tree-if-conv.c: Likewise.
4540 * tree-into-ssa.c: Likewise.
4541 * tree-loop-distribution.c: Likewise.
4542 * tree-object-size.c: Likewise.
4543 * tree-parloops.c: Likewise.
4544 * tree-pass.h: Likewise.
4545 * tree-sra.c: Likewise.
4546 * tree-ssa-ccp.c: Likewise.
4547 * tree-ssa-copy.c: Likewise.
4548 * tree-ssa-copyrename.c: Likewise.
4549 * tree-ssa-dce.c: Likewise.
4550 * tree-ssa-dom.c: Likewise.
4551 * tree-ssa-dse.c: Likewise.
4552 * tree-ssa-forwprop.c: Likewise.
4553 * tree-ssa-ifcombine.c: Likewise.
4554 * tree-ssa-loop-ch.c: Likewise.
4555 * tree-ssa-loop-ivcanon.c: Likewise.
4556 * tree-ssa-loop.c: Likewise.
4557 * tree-ssa-math-opts.c: Likewise.
4558 * tree-ssa-phiopt.c: Likewise.
4559 * tree-ssa-phiprop.c: Likewise.
4560 * tree-ssa-pre.c: Likewise.
4561 * tree-ssa-reassoc.c: Likewise.
4562 * tree-ssa-sink.c: Likewise.
4563 * tree-ssa-strlen.c: Likewise.
4564 * tree-ssa-tail-merge.c: Likewise.
4565 * tree-ssa-uncprop.c: Likewise.
4566 * tree-switch-conversion.c: Likewise.
4567 * tree-tailcall.c: Likewise.
4568 * tree-vect-generic.c: Likewise.
4569 * tree-vectorizer.c: Likewise.
4570 * tree-vrp.c: Likewise.
4571 * tsan.c: Likewise.
4572 * var-tracking.c: Likewise.
4573 * bt-load.c: Likewise.
4574 * cfgcleanup.c: Likewise.
4575 * combine-stack-adj.c: Likewise.
4576 * combine.c: Likewise.
4577 * compare-elim.c: Likewise.
4578 * config/epiphany/resolve-sw-modes.c: Likewise.
4579 * config/i386/i386.c: Likewise.
4580 * config/mips/mips.c: Likewise.
4581 * config/s390/s390.c: Likewise.
4582 * config/sh/sh_treg_combine.cc: Likewise.
4583 * config/sparc/sparc.c: Likewise.
4584 * dce.c: Likewise.
4585 * dse.c: Likewise.
4586 * final.c: Likewise.
4587 * ifcvt.c: Likewise.
4588 * mode-switching.c: Likewise.
4589 * passes.c: Likewise.
4590 * postreload.c: Likewise.
4591 * ree.c: Likewise.
4592 * reg-stack.c: Likewise.
4593 * regcprop.c: Likewise.
4594 * regrename.c: Likewise.
4595 * web.c: Likewise.
4597 2014-05-06 Richard Biener <rguenther@suse.de>
4599 PR middle-end/61070
4600 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
4601 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
4603 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
4605 PR ipa/60965
4606 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
4608 2014-05-05 Radovan Obradovic <robradovic@mips.com>
4609 Tom de Vries <tom@codesourcery.com>
4611 * target.def (call_fusage_contains_non_callee_clobbers): New
4612 DEFHOOKPOD.
4613 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
4614 Hooks to @menu.
4615 (@node Miscellaneous Register Hooks): New node.
4616 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
4617 * doc/tm.texi: Regenerate.
4619 2014-05-05 Marek Polacek <polacek@redhat.com>
4621 PR driver/61065
4622 * opts.c (common_handle_option): Call error_at instead of warning_at.
4624 2014-05-05 Richard Biener <rguenther@suse.de>
4626 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
4627 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
4628 under the TODO_verify_il umbrella.
4630 2014-05-05 Richard Biener <rguenther@suse.de>
4632 * passes.c (execute_function_todo): Move TODO_verify_flow under
4633 the TODO_verify_ul umbrella.
4635 2014-05-05 Richard Biener <rguenther@suse.de>
4637 PR middle-end/61010
4638 * fold-const.c (fold_binary_loc): Consistently avoid canonicalizing
4639 X & CST away from a CST that is the mask of a mode.
4641 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4643 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
4644 int argument to enum machine_mode.
4645 (picochip_class_max_nregs): Ditto.
4646 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
4647 (picochip_class_max_nregs): Ditto.
4649 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4651 * target.def: Add new target hook.
4652 * doc/tm.texi: Regenerate.
4653 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
4654 * targhooks.c (default_keep_leaf_when_profiled): New function.
4656 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
4657 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
4659 2014-05-05 Bin Cheng <bin.cheng@arm.com>
4661 PR tree-optimization/60363
4662 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
4663 (copy_phi_args): New parameters. Call get_value_locus_in_path.
4664 (update_destination_phis): New parameter.
4665 (create_edge_and_update_destination_phis): Ditto.
4666 (ssa_fix_duplicate_block_edges): Pass new arguments.
4667 (thread_single_edge): Ditto.
4669 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
4671 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
4672 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
4673 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
4674 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
4675 Use RS6000_BTM_HARD_FLOAT.
4676 (BU_MISC_2): Likewise.
4677 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
4678 RS6000_BTM_HARD_FLOAT.
4679 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
4680 is explicitly used.
4681 (rs6000_invalid_builtin): Add hard floating builtin support.
4682 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
4683 hard float builtins.
4684 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
4686 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4688 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
4689 Add missing function* argument.
4691 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
4693 * lra-constraints.c (valid_address_p): Move earlier in file.
4694 Add a constraint argument to the address_info version.
4695 (satisfies_memory_constraint_p): New function.
4696 (satisfies_address_constraint_p): Likewise.
4697 (process_alt_operands, curr_insn_transform): Use them.
4698 (process_address): Pass the constraint to valid_address_p when
4699 checking address operands.
4701 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
4703 * config/mips/mips.c (mips_isa_rev): New variable.
4704 (mips_set_architecture): Set it.
4705 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
4706 from mips_isa_rev.
4707 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
4708 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
4709 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
4710 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
4711 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
4712 conditions in terms of mips_isa_rev.
4713 (mips_isa_rev): Declare.
4715 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4717 * config/sh/sh-mem.cc: Use tabs instead of spaces.
4718 (prob_unlikely, prob_likely): Make variables const.
4720 2014-05-03 Denis Chertykov <chertykov@gmail.com>
4722 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
4724 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4726 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
4728 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4730 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
4731 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
4732 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
4733 functions.
4734 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
4735 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
4736 sh_pass_in_reg_p.
4737 Replace usage of ROUND_REG with sh_round_reg.
4738 Use CEIL instead of ROUND_ADVANCE.
4740 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4742 PR target/61026
4743 * config/sh/sh.c: Include stdlib headers before everything else.
4745 2014-05-02 Jakub Jelinek <jakub@redhat.com>
4747 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
4748 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
4749 (gimplify_adjust_omp_clauses): Simd region is never
4750 directly nested in combined parallel. Instead, for linear
4751 with copyin/copyout, if in combined for simd loop, make decl
4752 firstprivate/lastprivate on OMP_FOR.
4753 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
4754 expand_omp_for_static_chunk): When setting endvar, also set
4755 fd->loop.v to the same value.
4757 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
4759 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
4761 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
4763 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
4764 expression.
4766 2014-05-02 Marek Polacek <polacek@redhat.com>
4768 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
4770 2014-05-02 Kito Cheng <kito@0xlab.org>
4772 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
4773 to a C expression marco.
4774 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
4775 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
4776 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
4777 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
4778 HONOR_REG_ALLOC_ORDER.
4779 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
4781 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4783 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
4785 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4787 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
4789 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
4791 * tree-if-conv.c (is_cond_scalar_reduction): New function.
4792 (convert_scalar_cond_reduction): Likewise.
4793 (predicate_scalar_phi): Add recognition and transformation
4794 of simple conditioanl reduction to be vectorizable.
4796 2014-05-01 Marek Polacek <polacek@redhat.com>
4798 PR c/43245
4799 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
4801 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
4803 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
4804 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
4805 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
4806 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
4807 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
4808 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
4809 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
4810 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
4812 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
4814 * config/arc/arc.opt (mlra): Move comment above option name
4815 to avoid mis-parsing as language options.
4817 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4819 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
4820 * config/sol2.h: ... here.
4821 * config/sol2-10.h: Remove.
4823 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
4824 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
4825 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
4826 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
4827 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
4828 * config/sol2.h: ... here.
4829 (SECTION_NAME_FORMAT): Don't redefine.
4830 (STARTFILE_ARCH32_SPEC): Rename to ...
4831 (STARTFILE_ARCH_SPEC): ... this.
4832 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
4833 * config/sparc/sol2.h: ... here.
4834 (SECTION_NAME_FORMAT): Don't undef.
4835 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
4836 (SUBTARGET_EXTRA_SPECS): Remove.
4837 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
4839 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
4840 (MD_STARTFILE_PREFIX): Remove.
4841 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
4842 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
4843 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
4844 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
4845 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
4846 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
4847 * config/i386/sol2.h: ... here.
4848 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
4849 * config/i386/sol2-bi.h: Remove.
4850 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
4851 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
4853 * config/i386/t-sol2-64: Rename to ...
4854 * config/i386/t-sol2: ... this.
4855 * config/sparc/t-sol2-64: Rename to ...
4856 * config/sparc/t-sol2: ... this.
4858 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
4859 sol2_tm_file_head, sol2_tm_file_tail.
4860 Include ${cpu_type}/sol2.h before sol2.h.
4861 Remove sol2-10.h.
4862 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
4863 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
4864 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
4865 Reflect i386/t-sol2-64 renaming.
4866 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
4867 Reflect sparc/t-sol2-64 renaming.
4869 2014-04-30 Richard Biener <rguenther@suse.de>
4871 * passes.c (execute_function_todo): Move TODO_verify_stmts
4872 and TODO_verify_ssa under the TODO_verify_il umbrella.
4873 * tree-ssa.h (verify_ssa): Adjust prototype.
4874 * tree-ssa.c (verify_ssa): Add parameter to tell whether
4875 we should verify SSA operands.
4876 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
4877 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
4878 whether we should verify whether not throwing stmts have EH info.
4879 * graphite-scop-detection.c (create_sese_edges): Adjust.
4880 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
4881 * tree-eh.c (lower_try_finally_switch): Do not add the
4882 default case label twice.
4884 2014-04-30 Marek Polacek <polacek@redhat.com>
4886 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
4887 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
4888 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
4889 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
4891 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
4893 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
4894 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
4895 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
4896 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
4897 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
4898 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
4899 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
4900 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
4902 2014-04-29 David Malcolm <dmalcolm@redhat.com>
4904 * tree-cfg.c (dump_function_to_file): Dump the return type of
4905 functions, in a line to itself before the function body, mimicking
4906 the layout of a C function.
4908 2014-04-29 Jakub Jelinek <jakub@redhat.com>
4910 PR tree-optimization/60971
4911 * tree-tailcall.c (process_assignment): Reject conversions which
4912 reduce precision.
4914 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
4916 * calls.c (initialize_argument_information): Always treat
4917 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
4918 (expand_call): Likewise.
4919 (emit_library_call_calue_1): Likewise.
4920 * expr.c (PUSH_ARGS_REVERSED): Do not define.
4921 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
4922 code accordingly.
4924 2014-04-29 Nick Clifton <nickc@redhat.com>
4926 * config/msp430/msp430.md (umulsidi): Fix typo.
4927 (mulhisi3): Enable even inside interrupt handlers.
4928 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
4929 bigger return address pushed in large mode.
4931 2014-04-29 Nick Clifton <nickc@redhat.com>
4933 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
4934 (arc_init_reg_tables): Use a machine_mode enum to iterate over
4935 available modes.
4936 * config/m32r/m32r.c (init_reg_tables): Likewise.
4937 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
4938 enum to hold the modes.
4940 2014-04-29 Richard Biener <rguenther@suse.de>
4942 * dominance.c (free_dominance_info): Add overload with
4943 function parameter.
4944 (dom_info_state): Likewise.
4945 (dom_info_available_p): Likewise.
4946 * basic-block.h (free_dominance_info, dom_info_state,
4947 dom_info_available_p): Declare overloads.
4948 * passes.c (execute_function_todo): Verify that verifiers
4949 don't change dominator info state. Drop dominator info
4950 for IPA pass invocations.
4951 * cgraph.c (release_function_body): Restore asserts that
4952 dominator information is released.
4954 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
4956 * doc/invoke.texi: Fix typo.
4957 * tree-vrp.c: Fix typos.
4958 * gimple.c (infer_nonnull_range): Reorder operands of an && condition.
4960 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
4962 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
4964 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
4966 * config/aarch64/aarch64-builtins.c
4967 (aarch64_types_storestruct_lane_qualifiers): New.
4968 (TYPES_STORESTRUCT_LANE): Likewise.
4969 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
4970 (st3_lane): Likewise.
4971 (st4_lane): Likewise.
4972 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
4973 (vec_store_lanesci_lane<mode>): Likewise.
4974 (vec_store_lanesxi_lane<mode>): Likewise.
4975 (aarch64_st2_lane<VQ:mode>): Likewise.
4976 (aarch64_st3_lane<VQ:mode>): Likewise.
4977 (aarch64_st4_lane<VQ:mode>): Likewise.
4978 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
4979 * config/aarch64/arm_neon.h
4980 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
4981 use new macro arguments.
4982 (__ST3_LANE_FUNC): Likewise.
4983 (__ST4_LANE_FUNC): Likewise.
4984 * config/aarch64/iterators.md (V_TWO_ELEM): New.
4985 (V_THREE_ELEM): Likewise.
4986 (V_FOUR_ELEM): Likewise.
4988 2014-04-28 David Malcolm <dmalcolm@redhat.com>
4990 * doc/gimple.texi: Replace the description of the now-defunct
4991 union gimple_statement_d with a diagram showing the
4992 gimple_statement_base class hierarchy and its relationships to
4993 the GSS_ and GIMPLE_ enums.
4995 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
4997 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
4998 * config/aarch64/aarch64.c
4999 (aarch64_cannot_change_mode_class): Weaken conditions.
5000 (aarch64_modes_tieable_p): New.
5001 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
5003 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
5005 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
5006 (loadsync_<mode>): Change mode.
5007 (load_quadpti, store_quadpti): New.
5008 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
5009 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
5011 2014-04-28 Martin Jambor <mjambor@suse.cz>
5013 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
5014 same alias type as the original statement.
5015 (subreplacement_assignment_data): New type.
5016 (handle_unscalarized_data_in_subtree): New type of parameter,
5017 generate new memory accesses with same alias type as the original
5018 statement.
5019 (load_assign_lhs_subreplacements): Likewise.
5020 (sra_modify_constructor_assign): Generate new memory accesses with
5021 same alias type as the original statement.
5023 2014-04-28 Richard Biener <rguenther@suse.de>
5025 * tree-pass.h (TODO_verify_il): Define.
5026 (TODO_verify_all): Complete properly.
5027 * passes.c (execute_function_todo): Move existing loop-closed
5028 SSA verification under TODO_verify_il.
5029 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
5030 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
5031 Fix tree sharing issue.
5033 2014-04-28 Richard Biener <rguenther@suse.de>
5035 PR middle-end/60092
5036 * builtins.def (DEF_C11_BUILTIN): Add.
5037 (BUILT_IN_ALIGNED_ALLOC): Likewise.
5038 * coretypes.h (enum function_class): Add function_c11_misc.
5039 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
5040 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
5041 (call_may_clobber_ref_p_1): Likewise.
5042 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
5043 (mark_all_reaching_defs_necessary_1): Likewise.
5044 (propagate_necessity): Likewise.
5045 (eliminate_unnecessary_stmts): Likewise.
5046 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
5048 2014-04-28 Richard Biener <rguenther@suse.de>
5050 * tree-vrp.c (vrp_var_may_overflow): Remove.
5051 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
5052 with overflow immediately bump to one before that value and
5053 let iteration figure out overflow status.
5055 2014-04-28 Richard Biener <rguenther@suse.de>
5057 * configure.ac: Do valgrind header checks unconditionally.
5058 Add --enable-valgrind-annotations.
5059 * system.h: Guard valgrind header inclusion with
5060 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
5061 * alloc-pool.c (pool_alloc, pool_free): Use
5062 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
5063 to guard possibly dead code.
5064 * config.in: Regenerated.
5065 * configure: Likewise.
5067 2014-04-28 Jeff Law <law@redhat.com>
5069 PR tree-optimization/60902
5070 * tree-ssa-threadedge.c
5071 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
5072 over real defs when invalidating outputs from statements that do not
5073 produce useful outputs for threading.
5075 2014-04-28 Richard Biener <rguenther@suse.de>
5077 PR tree-optimization/60979
5078 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
5079 SCOPs that end in a block with a successor with abnormal
5080 predecessors.
5082 2014-04-28 Richard Biener <rguenther@suse.de>
5084 * tree-pass.h (execute_pass_list): Adjust prototype.
5085 * passes.c (pass_manager::execute_early_local_passes): Adjust.
5086 (do_per_function): Change callback signature, push all actual
5087 work to the callbals.
5088 (do_per_function_toporder): Likewise.
5089 (execute_function_dump): Adjust.
5090 (execute_function_todo): Likewise.
5091 (clear_last_verified): Likewise.
5092 (verify_curr_properties): Likewise.
5093 (update_properties_after_pass): Likewise.
5094 (execute_pass_list_1): Split out from ...
5095 (execute_pass_list): ... here. Adjust.
5096 (execute_ipa_pass_list): Likewise.
5097 * cgraphunit.c (cgraph_add_new_function): Adjust.
5098 (analyze_function): Likewise.
5099 (expand_function): Likewise.
5100 * cgraph.c (release_function_body): Free dominance info
5101 here instead of asserting it was magically freed elsewhere.
5103 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
5105 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
5106 * configure: Regenerate.
5107 * config/sparc/sparc.opt (muser-mode): New option.
5108 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
5109 for LEON3.
5110 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
5111 * doc/invoke.texi (SPARC options): Document -muser-mode.
5113 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
5115 * cselib.c (find_slot_memmode): Delete.
5116 (cselib_hasher): Change compare_type to a struct.
5117 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
5118 constants.
5119 (preserve_constants_and_equivs): Adjust for new compare_type.
5120 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
5121 (wrap_constant): Delete.
5122 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
5124 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
5126 * doc/install.texi (Building with profile feedback): Remove
5127 outdated sentence.
5129 2014-04-26 Tom de Vries <tom@codesourcery.com>
5131 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
5132 array accesses.
5134 2014-04-25 Cary Coutant <ccoutant@google.com>
5136 PR debug/60929
5137 * dwarf2out.c (should_move_die_to_comdat): A type definition
5138 can contain a subprogram definition, but don't move it to a
5139 comdat unit.
5140 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
5141 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
5142 from original DIE.
5143 (clone_tree_hash): Rename to...
5144 (clone_tree_partial): ...this; change callers. Copy
5145 DW_TAG_subprogram DIEs as declarations.
5146 (copy_decls_walk): Don't copy children of a declaration into a
5147 type unit.
5149 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
5151 PR target/60969
5152 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
5153 alternative 12.
5155 2014-04-25 Jiong Wang <jiong.wang@arm.com>
5157 * config/arm/predicates.md (call_insn_operand): Add long_call check.
5158 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
5159 reg for long_call.
5160 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
5161 restriction.
5163 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5165 * config/arm/arm.c (arm_cortex_a8_tune): Initialise T16-related fields.
5167 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5169 PR tree-optimization/60930
5170 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
5171 creating a multiply candidate by folding two constant
5172 multiplicands when the result overflows.
5174 2014-04-25 Jakub Jelinek <jakub@redhat.com>
5176 PR tree-optimization/60960
5177 * tree-vect-generic.c (expand_vector_operation): Only call
5178 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
5180 2014-04-25 Tom de Vries <tom@codesourcery.com>
5182 * expr.c (clobber_reg_mode): New function.
5183 * expr.h (clobber_reg): New function.
5185 2014-04-25 Tom de Vries <tom@codesourcery.com>
5187 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
5188 clobbers.
5190 2014-04-25 Radovan Obradovic <robradovic@mips.com>
5191 Tom de Vries <tom@codesourcery.com>
5193 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
5194 handle.
5195 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
5196 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
5197 new argument to find_all_hard_reg_sets call.
5199 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5201 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
5202 Use HOST_WIDE_INT_C for mask literal.
5203 (aarch_rev16_shleft_mask_imm_p): Likewise.
5205 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
5207 PR target/60941
5208 * config/sparc/sparc.md (ashlsi3_extend): Delete.
5210 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
5212 PR preprocessor/56540
5213 * config/i386/i386-c.c (ix86_target_macros): Define
5214 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
5216 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5218 * configure.ac (tga_func): Remove.
5219 (LIB_TLS_SPEC): Remove.
5220 * configure: Regenerate.
5221 * config.in: Regenerate.
5222 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
5224 2014-04-25 Richard Biener <rguenther@suse.de>
5226 PR ipa/60912
5227 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
5228 call stmt use/clobber sets during stmt walk instead of
5229 walking the possibly incomplete set of caller edges.
5231 2014-04-25 Richard Biener <rguenther@suse.de>
5233 PR ipa/60911
5234 * passes.c (apply_ipa_transforms): Inline into only caller ...
5235 (execute_one_pass): ... here. Properly bring in function
5236 bodies for nodes we want to apply IPA transforms to.
5238 2014-04-24 Cong Hou <congh@google.com>
5240 PR tree-optimization/60896
5241 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
5242 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
5243 (vect_mark_pattern_stmts): Set the def type of all statements in
5244 PATTERN_DEF_SEQ as vect_internal_def.
5246 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
5248 * doc/extend.texi (PowerPC Built-in Functions): Document new
5249 powerpc extended divide, bcd, pack/unpack 128-bit, builtin functions.
5250 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
5252 * config/rs6000/predicates.md (const_0_to_3_operand): New
5253 predicate to match 0..3 integer constants.
5255 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
5256 to support adding miscellaneous builtin functions.
5257 (BU_DFP_MISC_2): Likewise.
5258 (BU_P7_MISC_1): Likewise.
5259 (BU_P7_MISC_2): Likewise.
5260 (BU_P8V_MISC_3): Likewise.
5261 (BU_MISC_1): Likewise.
5262 (BU_MISC_2): Likewise.
5263 (DIVWE): Add extended divide builtin functions.
5264 (DIVWEO): Likewise.
5265 (DIVWEU): Likewise.
5266 (DIVWEUO): Likewise.
5267 (DIVDE): Likewise.
5268 (DIVDEO): Likewise.
5269 (DIVDEU): Likewise.
5270 (DIVDEUO): Likewise.
5271 (DXEX): Add decimal floating-point builtin functions.
5272 (DXEXQ): Likewise.
5273 (DDEDPD): Likewise.
5274 (DDEDPDQ): Likewise.
5275 (DENBCD): Likewise.
5276 (DENBCDQ): Likewise.
5277 (DIEX): Likewise.
5278 (DIEXQ): Likewise.
5279 (DSCLI): Likewise.
5280 (DSCLIQ): Likewise.
5281 (DSCRI): Likewise.
5282 (DSCRIQ): Likewise.
5283 (CDTBCD): Add new BCD builtin functions.
5284 (CBCDTD): Likewise.
5285 (ADDG6S): Likewise.
5286 (BCDADD): Likewise.
5287 (BCDADD_LT): Likewise.
5288 (BCDADD_EQ): Likewise.
5289 (BCDADD_GT): Likewise.
5290 (BCDADD_OV): Likewise.
5291 (BCDSUB): Likewise.
5292 (BCDSUB_LT): Likewise.
5293 (BCDSUB_EQ): Likewise.
5294 (BCDSUB_GT): Likewise.
5295 (BCDSUB_OV): Likewise.
5296 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
5297 (UNPACK_TD): Likewise.
5298 (PACK_TF): Likewise.
5299 (UNPACK_TF): Likewise.
5300 (UNPACK_TF_0): Likewise.
5301 (UNPACK_TF_1): Likewise.
5302 (PACK_V1TI): Likewise.
5303 (UNPACK_V1TI): Likewise.
5305 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
5306 support for decimal floating point builtin functions.
5307 (rs6000_expand_ternop_builtin): Add checks for the new builtin
5308 functions that take constant arguments.
5309 (rs6000_invalid_builtin): Add decimal floating point builtin support.
5310 (rs6000_init_builtins): Setup long double, _Decimal64, and
5311 _Decimal128 types for new builtin functions.
5312 (builtin_function_type): Set the unsigned flags appropriately for
5313 the new builtin functions.
5314 (rs6000_opt_masks): Add support for decimal floating point builtin
5315 functions.
5317 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
5318 floating point builtin functions.
5319 (RS6000_BTM_COMMON): Likewise.
5320 (RS6000_BTI_long_double): Likewise.
5321 (RS6000_BTI_dfloat64): Likewise.
5322 (RS6000_BTI_dfloat128): Likewise.
5323 (long_double_type_internal_node): Likewise.
5324 (dfloat64_type_internal_node): Likewise.
5325 (dfloat128_type_internal_node): Likewise.
5327 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
5328 2.07 bcd arithmetic instructions.
5329 (UNSPEC_BCDSUB): Likewise.
5330 (UNSPEC_BCD_OVERFLOW): Likewise.
5331 (UNSPEC_BCD_ADD_SUB): Likewise.
5332 (bcd_add_sub): Likewise.
5333 (BCD_TEST): Likewise.
5334 (bcd<bcd_add_sub>): Likewise.
5335 (bcd<bcd_add_sub>_test): Likewise.
5336 (bcd<bcd_add_sub>_test2): Likewise.
5337 (bcd<bcd_add_sub>_<code>): Likewise.
5338 (peephole2 for combined bcd ops): Likewise.
5340 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
5341 decimal floating point builtin functions.
5342 (UNSPEC_DENBCD): Likewise.
5343 (UNSPEC_DXEX): Likewise.
5344 (UNSPEC_DIEX): Likewise.
5345 (UNSPEC_DSCLI): Likewise.
5346 (UNSPEC_DSCRI): Likewise.
5347 (D64_D128): Likewise.
5348 (dfp_suffix): Likewise.
5349 (dfp_ddedpd_<mode>): Likewise.
5350 (dfp_denbcd_<mode>): Likewise.
5351 (dfp_dxex_<mode>): Likewise.
5352 (dfp_diex_<mode>): Likewise.
5353 (dfp_dscli_<mode>): Likewise.
5354 (dfp_dscri_<mode>): Likewise.
5356 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
5357 builtin functions.
5358 (UNSPEC_CDTBCD): Likewise.
5359 (UNSPEC_CBCDTD): Likewise.
5360 (UNSPEC_DIVE): Add support for new extended divide builtin functions.
5361 (UNSPEC_DIVEO): Likewise.
5362 (UNSPEC_DIVEU): Likewise.
5363 (UNSPEC_DIVEUO): Likewise.
5364 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
5365 pack/unpack 128-bit types.
5366 (UNSPEC_PACK_128BIT): Likewise.
5367 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
5368 (udiv<mode>3): Use idiv_ldiv mode attribute.
5369 (div<mode>3): Likewise.
5370 (addg6s): Add new BCD builtin functions.
5371 (cdtbcd): Likewise.
5372 (cbcdtd): Likewise.
5373 (UNSPEC_DIV_EXTEND): Add support for new extended divide instructions.
5374 (div_extend): Likewise.
5375 (div<div_extend>_<mode>"): Likewise.
5376 (FP128_64): Add support for new builtin functions to pack/unpack
5377 128-bit types.
5378 (unpack<mode>): Likewise.
5379 (unpacktf_0): Likewise.
5380 (unpacktf_1): Likewise.
5381 (unpack<mode>_dm): Likewise.
5382 (unpack<mode>_nodm): Likewise.
5383 (pack<mode>): Likewise.
5384 (unpackv1ti): Likewise.
5385 (packv1ti): Likewise.
5387 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
5389 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
5390 is disabled.
5392 2014-04-24 Jakub Jelinek <jakub@redhat.com>
5394 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
5395 * gimplify.c (omp_is_private): Change last argument's type to int.
5396 Only diagnose lastprivate if the simd argument is 1, only diagnose
5397 linear if the simd argument is 2.
5398 (gimplify_omp_for): Adjust omp_is_private callers. When adding
5399 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
5400 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
5401 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
5402 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
5403 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
5404 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
5405 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
5406 * tree-nested.c (convert_nonlocal_omp_clauses,
5407 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
5409 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
5411 PR target/60822
5412 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
5413 operand 1.
5415 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
5417 * flag-types.h (enum ivar_visibility): Add.
5419 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
5421 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
5422 function * argument.
5424 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
5426 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
5428 2014-04-24 Radovan Obradovic <robradovic@mips.com>
5429 Tom de Vries <tom@codesourcery.com>
5431 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
5432 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
5433 reg-note.
5434 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
5435 * emit-rtl.c (try_split): Same.
5437 2014-04-24 Radovan Obradovic <robradovic@mips.com>
5438 Tom de Vries <tom@codesourcery.com>
5440 * common.opt (fuse-caller-save): New option.
5442 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
5444 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of
5445 elements for big-endian.
5447 2014-04-24 Richard Biener <rguenther@suse.de>
5449 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
5450 during TER and instead use the sepops interface for expanding
5451 non-GIMPLE_SINGLE_RHS.
5453 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5455 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
5456 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
5458 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5460 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
5461 assembler 64-bit option.
5462 * configure: Regenerate.
5464 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5466 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
5467 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
5468 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
5469 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
5470 (TARGET_CRYPTO): Take TARGET_SIMD into account.
5472 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5474 * config/aarch64/aarch64-builtins.c
5475 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
5476 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
5477 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
5478 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
5479 builtins.
5480 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
5481 (Vrevsuff): New mode attribute.
5483 2014-04-24 Terry Guo <terry.guo@arm.com>
5485 * config/arm/arm.h (machine_function): Define variable
5486 after_arm_reorg here.
5487 * config/arm/arm.c (after_arm_reorg): Remove the definition.
5488 (arm_split_constant): Update the way to access variable
5489 after_arm_reorg.
5490 (arm_reorg): Ditto.
5491 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
5493 2014-04-23 Tom de Vries <tom@codesourcery.com>
5495 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
5497 2014-04-23 David Malcolm <dmalcolm@redhat.com>
5499 * is-a.h: Update comments to reflect the following changes to the
5500 "pointerness" of the API, making the template parameter match the
5501 return type, allowing use of is-a.h with typedefs of pointers.
5502 (is_a_helper::cast): Return a T rather then a pointer to a T, so
5503 that the return type matches the parameter to the is_a_helper.
5504 (as_a): Likewise.
5505 (dyn_cast): Likewise.
5507 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
5508 pointer from the is-a.h API.
5510 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
5511 (is_a_helper <cgraph_node *>::test): ...this, matching change to
5512 is-a.h API.
5513 (is_a_helper <varpool_node>::test): Likewise, convert to...
5514 (is_a_helper <varpool_node *>::test): ...this.
5516 (varpool_first_variable): Update for removal of implicit pointer
5517 from the is-a.h API.
5518 (varpool_next_variable): Likewise.
5519 (varpool_first_static_initializer): Likewise.
5520 (varpool_next_static_initializer): Likewise.
5521 (varpool_first_defined_variable): Likewise.
5522 (varpool_next_defined_variable): Likewise.
5523 (cgraph_first_defined_function): Likewise.
5524 (cgraph_next_defined_function): Likewise.
5525 (cgraph_first_function): Likewise.
5526 (cgraph_next_function): Likewise.
5527 (cgraph_first_function_with_gimple_body): Likewise.
5528 (cgraph_next_function_with_gimple_body): Likewise.
5529 (cgraph_alias_target): Likewise.
5530 (varpool_alias_target): Likewise.
5531 (cgraph_function_or_thunk_node): Likewise.
5532 (varpool_variable_node): Likewise.
5533 (symtab_real_symbol_p): Likewise.
5534 * cgraphunit.c (referred_to_p): Likewise.
5535 (analyze_functions): Likewise.
5536 (handle_alias_pairs): Likewise.
5537 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
5538 * gimple-ssa.h (gimple_vuse_op): Likewise.
5539 (gimple_vdef_op): Likewise.
5540 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
5541 * gimple.c (gimple_build_asm_1): Likewise.
5542 (gimple_build_try): Likewise.
5543 (gimple_build_resx): Likewise.
5544 (gimple_build_eh_dispatch): Likewise.
5545 (gimple_build_omp_for): Likewise.
5546 (gimple_omp_for_set_clauses): Likewise.
5548 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
5549 (is_a_helper <gimple_statement_asm *>::test): ...this.
5550 (is_a_helper <gimple_statement_bind>::test): Convert to...
5551 (is_a_helper <gimple_statement_bind *>::test): ...this.
5552 (is_a_helper <gimple_statement_call>::test): Convert to...
5553 (is_a_helper <gimple_statement_call *>::test): ...this.
5554 (is_a_helper <gimple_statement_catch>::test): Convert to...
5555 (is_a_helper <gimple_statement_catch *>::test): ...this.
5556 (is_a_helper <gimple_statement_resx>::test): Convert to...
5557 (is_a_helper <gimple_statement_resx *>::test): ...this.
5558 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
5559 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
5560 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
5561 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
5562 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
5563 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
5564 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
5565 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
5566 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
5567 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
5568 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
5569 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
5570 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
5571 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
5572 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
5573 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
5574 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
5575 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
5576 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
5577 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
5578 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
5579 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
5580 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
5581 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
5582 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
5583 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
5584 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
5585 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
5586 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
5587 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
5588 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
5589 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
5590 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
5591 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
5592 (is_a_helper <gimple_statement_phi>::test): Convert to...
5593 (is_a_helper <gimple_statement_phi *>::test): ...this.
5594 (is_a_helper <gimple_statement_transaction>::test): Convert to...
5595 (is_a_helper <gimple_statement_transaction *>::test): ...this.
5596 (is_a_helper <gimple_statement_try>::test): Convert to...
5597 (is_a_helper <gimple_statement_try *>::test): ...this.
5598 (is_a_helper <gimple_statement_wce>::test): Convert to...
5599 (is_a_helper <gimple_statement_wce *>::test): ...this.
5600 (is_a_helper <const gimple_statement_asm>::test): Convert to...
5601 (is_a_helper <const gimple_statement_asm *>::test): ...this.
5602 (is_a_helper <const gimple_statement_bind>::test): Convert to...
5603 (is_a_helper <const gimple_statement_bind *>::test): ...this.
5604 (is_a_helper <const gimple_statement_call>::test): Convert to...
5605 (is_a_helper <const gimple_statement_call *>::test): ...this.
5606 (is_a_helper <const gimple_statement_catch>::test): Convert to...
5607 (is_a_helper <const gimple_statement_catch *>::test): ...this.
5608 (is_a_helper <const gimple_statement_resx>::test): Convert to...
5609 (is_a_helper <const gimple_statement_resx *>::test): ...this.
5610 (is_a_helper <const gimple_statement_eh_dispatch>::test): Convert to...
5611 (is_a_helper <const gimple_statement_eh_dispatch *>::test): ...this.
5612 (is_a_helper <const gimple_statement_eh_filter>::test): Convert to...
5613 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
5614 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
5615 Convert to...
5616 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
5617 ...this.
5618 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
5619 Convert to...
5620 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
5621 ...this.
5622 (is_a_helper <const gimple_statement_omp_return>::test): Convert to...
5623 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
5624 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
5625 to...
5626 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
5627 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
5628 to...
5629 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
5630 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
5631 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
5632 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert to...
5633 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
5634 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
5635 to...
5636 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
5637 (is_a_helper <const gimple_statement_omp_target>::test): Convert to...
5638 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
5639 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
5640 to...
5641 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
5642 (is_a_helper <const gimple_statement_omp_single>::test): Convert to...
5643 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
5644 (is_a_helper <const gimple_statement_omp_teams>::test): Convert to...
5645 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
5646 (is_a_helper <const gimple_statement_omp_task>::test): Convert to...
5647 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
5648 (is_a_helper <const gimple_statement_phi>::test): Convert to...
5649 (is_a_helper <const gimple_statement_phi *>::test): ...this.
5650 (is_a_helper <const gimple_statement_transaction>::test): Convert to...
5651 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
5652 (is_a_helper <const gimple_statement_with_ops>::test): Convert to...
5653 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
5654 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
5655 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
5656 (is_a_helper <const gimple_statement_with_memory_ops>::test): Convert
5657 to...
5658 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
5659 ...this.
5660 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert to...
5661 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
5663 (gimple_use_ops): Update for removal of implicit pointer from the
5664 is-a.h API.
5665 (gimple_set_use_ops): Likewise.
5666 (gimple_vuse): Likewise.
5667 (gimple_vdef): Likewise.
5668 (gimple_vuse_ptr): Likewise.
5669 (gimple_vdef_ptr): Likewise.
5670 (gimple_set_vuse): Likewise.
5671 (gimple_set_vdef): Likewise.
5672 (gimple_omp_return_set_lhs): Likewise.
5673 (gimple_omp_return_lhs): Likewise.
5674 (gimple_omp_return_lhs_ptr): Likewise.
5675 (gimple_call_fntype): Likewise.
5676 (gimple_call_set_fntype): Likewise.
5677 (gimple_call_set_internal_fn): Likewise.
5678 (gimple_call_use_set): Likewise.
5679 (gimple_call_clobber_set): Likewise.
5680 (gimple_bind_vars): Likewise.
5681 (gimple_bind_set_vars): Likewise.
5682 (gimple_bind_body_ptr): Likewise.
5683 (gimple_bind_set_body): Likewise.
5684 (gimple_bind_add_stmt): Likewise.
5685 (gimple_bind_block): Likewise.
5686 (gimple_bind_set_block): Likewise.
5687 (gimple_asm_ninputs): Likewise.
5688 (gimple_asm_noutputs): Likewise.
5689 (gimple_asm_nclobbers): Likewise.
5690 (gimple_asm_nlabels): Likewise.
5691 (gimple_asm_input_op): Likewise.
5692 (gimple_asm_input_op_ptr): Likewise.
5693 (gimple_asm_output_op): Likewise.
5694 (gimple_asm_output_op_ptr): Likewise.
5695 (gimple_asm_set_output_op): Likewise.
5696 (gimple_asm_clobber_op): Likewise.
5697 (gimple_asm_set_clobber_op): Likewise.
5698 (gimple_asm_label_op): Likewise.
5699 (gimple_asm_set_label_op): Likewise.
5700 (gimple_asm_string): Likewise.
5701 (gimple_catch_types): Likewise.
5702 (gimple_catch_types_ptr): Likewise.
5703 (gimple_catch_handler_ptr): Likewise.
5704 (gimple_catch_set_types): Likewise.
5705 (gimple_catch_set_handler): Likewise.
5706 (gimple_eh_filter_types): Likewise.
5707 (gimple_eh_filter_types_ptr): Likewise.
5708 (gimple_eh_filter_failure_ptr): Likewise.
5709 (gimple_eh_filter_set_types): Likewise.
5710 (gimple_eh_filter_set_failure): Likewise.
5711 (gimple_eh_must_not_throw_fndecl): Likewise.
5712 (gimple_eh_must_not_throw_set_fndecl): Likewise.
5713 (gimple_eh_else_n_body_ptr): Likewise.
5714 (gimple_eh_else_e_body_ptr): Likewise.
5715 (gimple_eh_else_set_n_body): Likewise.
5716 (gimple_eh_else_set_e_body): Likewise.
5717 (gimple_try_eval_ptr): Likewise.
5718 (gimple_try_cleanup_ptr): Likewise.
5719 (gimple_try_set_eval): Likewise.
5720 (gimple_try_set_cleanup): Likewise.
5721 (gimple_wce_cleanup_ptr): Likewise.
5722 (gimple_wce_set_cleanup): Likewise.
5723 (gimple_phi_capacity): Likewise.
5724 (gimple_phi_num_args): Likewise.
5725 (gimple_phi_result): Likewise.
5726 (gimple_phi_result_ptr): Likewise.
5727 (gimple_phi_set_result): Likewise.
5728 (gimple_phi_arg): Likewise.
5729 (gimple_phi_set_arg): Likewise.
5730 (gimple_resx_region): Likewise.
5731 (gimple_resx_set_region): Likewise.
5732 (gimple_eh_dispatch_region): Likewise.
5733 (gimple_eh_dispatch_set_region): Likewise.
5734 (gimple_omp_critical_name): Likewise.
5735 (gimple_omp_critical_name_ptr): Likewise.
5736 (gimple_omp_critical_set_name): Likewise.
5737 (gimple_omp_for_clauses): Likewise.
5738 (gimple_omp_for_clauses_ptr): Likewise.
5739 (gimple_omp_for_set_clauses): Likewise.
5740 (gimple_omp_for_collapse): Likewise.
5741 (gimple_omp_for_index): Likewise.
5742 (gimple_omp_for_index_ptr): Likewise.
5743 (gimple_omp_for_set_index): Likewise.
5744 (gimple_omp_for_initial): Likewise.
5745 (gimple_omp_for_initial_ptr): Likewise.
5746 (gimple_omp_for_set_initial): Likewise.
5747 (gimple_omp_for_final): Likewise.
5748 (gimple_omp_for_final_ptr): Likewise.
5749 (gimple_omp_for_set_final): Likewise.
5750 (gimple_omp_for_incr): Likewise.
5751 (gimple_omp_for_incr_ptr): Likewise.
5752 (gimple_omp_for_set_incr): Likewise.
5753 (gimple_omp_for_pre_body_ptr): Likewise.
5754 (gimple_omp_for_set_pre_body): Likewise.
5755 (gimple_omp_parallel_clauses): Likewise.
5756 (gimple_omp_parallel_clauses_ptr): Likewise.
5757 (gimple_omp_parallel_set_clauses): Likewise.
5758 (gimple_omp_parallel_child_fn): Likewise.
5759 (gimple_omp_parallel_child_fn_ptr): Likewise.
5760 (gimple_omp_parallel_set_child_fn): Likewise.
5761 (gimple_omp_parallel_data_arg): Likewise.
5762 (gimple_omp_parallel_data_arg_ptr): Likewise.
5763 (gimple_omp_parallel_set_data_arg): Likewise.
5764 (gimple_omp_task_clauses): Likewise.
5765 (gimple_omp_task_clauses_ptr): Likewise.
5766 (gimple_omp_task_set_clauses): Likewise.
5767 (gimple_omp_task_child_fn): Likewise.
5768 (gimple_omp_task_child_fn_ptr): Likewise.
5769 (gimple_omp_task_set_child_fn): Likewise.
5770 (gimple_omp_task_data_arg): Likewise.
5771 (gimple_omp_task_data_arg_ptr): Likewise.
5772 (gimple_omp_task_set_data_arg): Likewise.
5773 (gimple_omp_taskreg_clauses): Likewise.
5774 (gimple_omp_taskreg_clauses_ptr): Likewise.
5775 (gimple_omp_taskreg_set_clauses): Likewise.
5776 (gimple_omp_taskreg_child_fn): Likewise.
5777 (gimple_omp_taskreg_child_fn_ptr): Likewise.
5778 (gimple_omp_taskreg_set_child_fn): Likewise.
5779 (gimple_omp_taskreg_data_arg): Likewise.
5780 (gimple_omp_taskreg_data_arg_ptr): Likewise.
5781 (gimple_omp_taskreg_set_data_arg): Likewise.
5782 (gimple_omp_task_copy_fn): Likewise.
5783 (gimple_omp_task_copy_fn_ptr): Likewise.
5784 (gimple_omp_task_set_copy_fn): Likewise.
5785 (gimple_omp_task_arg_size): Likewise.
5786 (gimple_omp_task_arg_size_ptr): Likewise.
5787 (gimple_omp_task_set_arg_size): Likewise.
5788 (gimple_omp_task_arg_align): Likewise.
5789 (gimple_omp_task_arg_align_ptr): Likewise.
5790 (gimple_omp_task_set_arg_align): Likewise.
5791 (gimple_omp_single_clauses): Likewise.
5792 (gimple_omp_single_clauses_ptr): Likewise.
5793 (gimple_omp_single_set_clauses): Likewise.
5794 (gimple_omp_target_clauses): Likewise.
5795 (gimple_omp_target_clauses_ptr): Likewise.
5796 (gimple_omp_target_set_clauses): Likewise.
5797 (gimple_omp_target_child_fn): Likewise.
5798 (gimple_omp_target_child_fn_ptr): Likewise.
5799 (gimple_omp_target_set_child_fn): Likewise.
5800 (gimple_omp_target_data_arg): Likewise.
5801 (gimple_omp_target_data_arg_ptr): Likewise.
5802 (gimple_omp_target_set_data_arg): Likewise.
5803 (gimple_omp_teams_clauses): Likewise.
5804 (gimple_omp_teams_clauses_ptr): Likewise.
5805 (gimple_omp_teams_set_clauses): Likewise.
5806 (gimple_omp_sections_clauses): Likewise.
5807 (gimple_omp_sections_clauses_ptr): Likewise.
5808 (gimple_omp_sections_set_clauses): Likewise.
5809 (gimple_omp_sections_control): Likewise.
5810 (gimple_omp_sections_control_ptr): Likewise.
5811 (gimple_omp_sections_set_control): Likewise.
5812 (gimple_omp_for_set_cond): Likewise.
5813 (gimple_omp_for_cond): Likewise.
5814 (gimple_omp_atomic_store_set_val): Likewise.
5815 (gimple_omp_atomic_store_val): Likewise.
5816 (gimple_omp_atomic_store_val_ptr): Likewise.
5817 (gimple_omp_atomic_load_set_lhs): Likewise.
5818 (gimple_omp_atomic_load_lhs): Likewise.
5819 (gimple_omp_atomic_load_lhs_ptr): Likewise.
5820 (gimple_omp_atomic_load_set_rhs): Likewise.
5821 (gimple_omp_atomic_load_rhs): Likewise.
5822 (gimple_omp_atomic_load_rhs_ptr): Likewise.
5823 (gimple_omp_continue_control_def): Likewise.
5824 (gimple_omp_continue_control_def_ptr): Likewise.
5825 (gimple_omp_continue_set_control_def): Likewise.
5826 (gimple_omp_continue_control_use): Likewise.
5827 (gimple_omp_continue_control_use_ptr): Likewise.
5828 (gimple_omp_continue_set_control_use): Likewise.
5829 (gimple_transaction_body_ptr): Likewise.
5830 (gimple_transaction_label): Likewise.
5831 (gimple_transaction_label_ptr): Likewise.
5832 (gimple_transaction_set_body): Likewise.
5833 (gimple_transaction_set_label): Likewise.
5835 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
5836 * ipa-inline-analysis.c (inline_write_summary): Likewise.
5837 * ipa-ref.c (ipa_record_reference): Likewise.
5838 * ipa-reference.c (analyze_function): Likewise.
5839 (ipa_reference_write_optimization_summary): Likewise.
5840 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
5841 (address_taken_from_non_vtable_p): Likewise.
5842 (comdat_can_be_unshared_p_1): Likewise.
5843 * lto-cgraph.c (lto_output_ref): Likewise.
5844 (add_references): Likewise.
5845 (compute_ltrans_boundary): Likewise.
5846 (output_symtab): Likewise.
5847 (input_ref): Likewise.
5848 (input_cgraph_1): Likewise.
5849 (output_cgraph_opt_summary): Likewise.
5850 * lto-streamer-out.c (lto_output): Likewise.
5851 (output_symbol_p): Likewise.
5852 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
5853 (lsei_start_function_in_partition): Likewise.
5854 (lsei_next_variable_in_partition): Likewise.
5855 (lsei_start_variable_in_partition): Likewise.
5856 * symtab.c (insert_to_assembler_name_hash): Likewise.
5857 (unlink_from_assembler_name_hash): Likewise.
5858 (symtab_unregister_node): Likewise.
5859 (symtab_remove_node): Likewise.
5860 (dump_symtab_node): Likewise.
5861 (verify_symtab_base): Likewise.
5862 (verify_symtab_node): Likewise.
5863 (symtab_make_decl_local): Likewise.
5864 (symtab_alias_ultimate_target): Likewise.
5865 (symtab_resolve_alias): Likewise.
5866 (symtab_get_symbol_partitioning_class): Likewise.
5867 * tree-phinodes.c (allocate_phi_node): Likewise.
5868 (reserve_phi_args_for_new_edge): Likewise.
5869 (remove_phi_args): Likewise.
5870 * varpool.c (varpool_node_for_asm): Likewise.
5871 (varpool_remove_unreferenced_decls): Likewise.
5873 2014-04-23 Jeff Law <law@redhat.com>
5875 PR tree-optimization/60902
5876 * tree-ssa-threadedge.c
5877 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
5878 invalidate outputs from statements that do not produce useful
5879 outputs for threading.
5881 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
5883 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
5884 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
5885 machine descriptions for Stack Smashing Protector.
5887 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
5889 * aarch64.md (<optab>_rol<mode>3): New pattern.
5890 (<optab>_rolsi3_uxtw): Likewise.
5891 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
5893 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
5895 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
5896 (arm_cortex_a12_tune): Likewise.
5898 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5900 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
5902 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5904 * config/arm/arm.md (arm_rev16si2): New pattern.
5905 (arm_rev16si2_alt): Likewise.
5906 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
5908 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5910 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
5911 (rev16<mode>2_alt): Likewise.
5912 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
5913 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
5914 (aarch_rev16_shleft_mask_imm_p): Likewise.
5915 (aarch_rev16_p_1): Likewise.
5916 (aarch_rev16_p): Likewise.
5917 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
5918 (aarch_rev16_shright_mask_imm_p): Likewise.
5919 (aarch_rev16_shleft_mask_imm_p): Likewise.
5921 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5923 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
5924 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
5925 rev cost.
5926 (cortex_a53_extra_costs): Likewise.
5927 (cortex_a57_extra_costs): Likewise.
5928 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
5929 (cortexa7_extra_costs): Likewise.
5930 (cortexa8_extra_costs): Likewise.
5931 (cortexa12_extra_costs): Likewise.
5932 (cortexa15_extra_costs): Likewise.
5933 (v7m_extra_costs): Likewise.
5934 (arm_new_rtx_costs): Handle BSWAP.
5936 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5938 * config/arm/arm.c (cortexa8_extra_costs): New table.
5939 (arm_cortex_a8_tune): New tuning struct.
5940 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
5942 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5944 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
5946 2014-04-23 Richard Biener <rguenther@suse.de>
5948 * Makefile.in (OBJS): Remove loop-unswitch.o.
5949 * tree-pass.h (make_pass_rtl_unswitch): Remove.
5950 * passes.def (pass_rtl_unswitch): Likewise.
5951 * loop-init.c (gate_rtl_unswitch): Likewise.
5952 (rtl_unswitch): Likewise.
5953 (pass_data_rtl_unswitch): Likewise.
5954 (pass_rtl_unswitch): Likewise.
5955 (make_pass_rtl_unswitch): Likewise.
5956 * rtl.h (reversed_condition): Likewise.
5957 (compare_and_jump_seq): Likewise.
5958 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
5959 and make static.
5960 * loop-unroll.c (compare_and_jump_seq): Likewise.
5962 2014-04-23 Richard Biener <rguenther@suse.de>
5964 PR tree-optimization/60903
5965 * tree-ssa-loop-im.c (analyze_memory_references): Remove
5966 commented code block.
5967 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
5968 loop flags to newly created BBs and edges.
5970 2014-04-23 Nick Clifton <nickc@redhat.com>
5972 * config/msp430/msp430.c (msp430_handle_option): Move function
5973 to msp430-common.c
5974 (msp430_option_override): Simplify mcu and mcpu option handling.
5975 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
5976 support for -mhwmult command line option.
5977 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
5978 -mhwmult command line option.
5979 (msp430_hwmult_enabled): Delete.
5980 (msp43o_output_labelref): Add support for -mhwmult command line option.
5981 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
5982 (umulsidi3): Likewise.
5983 * config/msp430/msp430.opt (mmcu): Add Report attribute.
5984 (mcpu, mlarge, msmall): Likewise.
5985 (mhwmult): New option.
5986 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
5987 prototype.
5988 (msp430_is_f5_mcu): Remove prototype.
5989 (msp430_use_f5_series_hwmult): Add prototype.
5990 * config/msp430/msp430-opts.h: New file.
5991 * common/config/msp430: New directory.
5992 * common/config/msp430/msp430-common.c: New file.
5993 * config.gcc (msp430): Remove target_has_targetm_common.
5994 * doc/invoke.texi: Document -mhwmult command line option.
5996 2014-04-23 Nick Clifton <nickc@redhat.com>
5998 * config/i386/cygwin.h (ENDFILE_SPEC): Include
5999 default-manifest.o if it can be found in the search path.
6000 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
6002 2014-04-23 Terry Guo <terry.guo@arm.com>
6004 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
6006 2014-04-23 Richard Biener <rguenther@suse.de>
6008 PR middle-end/60895
6009 * tree-inline.c (declare_return_variable): Use mark_addressable.
6011 2014-04-23 Richard Biener <rguenther@suse.de>
6013 PR middle-end/60891
6014 * loop-init.c (loop_optimizer_init): Make sure to apply
6015 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
6017 2014-04-22 Jakub Jelinek <jakub@redhat.com>
6019 PR sanitizer/60275
6020 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
6021 New options.
6022 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
6023 if flag_sanitize_undefined_trap_on_error.
6024 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
6025 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
6026 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
6027 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
6028 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
6029 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
6030 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
6031 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
6032 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
6033 * ubsan.c (ubsan_instrument_unreachable): Return
6034 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
6035 (ubsan_expand_null_ifn): Emit __builtin_trap ()
6036 if flag_sanitize_undefined_trap_on_error and
6037 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
6038 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
6039 instrument_bool_enum_load): Emit __builtin_trap () if
6040 flag_sanitize_undefined_trap_on_error and
6041 __builtin_handle_*_abort () if !flag_sanitize_recover.
6042 * doc/invoke.texi (-fsanitize-recover,
6043 -fsanitize-undefined-trap-on-error): Document.
6045 2014-04-22 Christian Bruel <christian.bruel@st.com>
6047 * config/sh/sh.md (mov<mode>): Replace movQIHI.
6048 Force immediates to SImode.
6050 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
6052 * config/nios2/nios2.md (UNSPEC_ROUND): New.
6053 (lroundsfsi2): New.
6054 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
6055 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
6056 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
6057 (nios2_fpu_insn): Add entry for round.
6058 (N2FPU_NO_ERRNO_P): Define.
6059 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
6060 flag_errno_math.
6061 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
6063 2014-04-22 Richard Henderson <rth@redhat.com>
6065 * config/aarch64/aarch64 (addti3, subti3): New expanders.
6066 (add<GPI>3_compare0): Remove leading * from name.
6067 (add<GPI>3_carryin): Likewise.
6068 (sub<GPI>3_compare0): Likewise.
6069 (sub<GPI>3_carryin): Likewise.
6070 (<su_optab>mulditi3): New expander.
6071 (multi3): New expander.
6072 (madd<GPI>): Remove leading * from name.
6074 2014-04-22 Martin Jambor <mjambor@suse.cz>
6076 * cgraphclones.c (cgraph_function_versioning): Copy
6077 ipa_transforms_to_apply instead of asserting it is empty.
6079 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
6081 PR target/60868
6082 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
6083 on count_exp to get mode.
6085 2014-04-22 Andrew Pinski <apinski@cavium.com>
6087 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
6088 Handle TLS for ILP32.
6089 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
6090 (tlsie_small_<mode>): this and handle PTR.
6091 (tlsie_small_sidi): New pattern.
6092 (tlsle_small): Change to an expand to handle ILP32.
6093 (tlsle_small_<mode>): New pattern.
6094 (tlsdesc_small): Rename to ...
6095 (tlsdesc_small_<mode>): this and handle PTR.
6097 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6099 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
6101 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
6103 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
6104 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
6105 (aarch64_types_signed_poly_qualifiers): Likewise.
6106 (aarch64_types_unsigned_signed_qualifiers): Likewise.
6107 (aarch64_types_poly_signed_qualifiers): Likewise.
6108 (TYPES_REINTERP_SS): Type macro added.
6109 (TYPES_REINTERP_SU): Likewise.
6110 (TYPES_REINTERP_SP): Likewise.
6111 (TYPES_REINTERP_US): Likewise.
6112 (TYPES_REINTERP_PS): Likewise.
6113 (aarch64_fold_builtin): New expression folding added.
6114 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
6115 Declarations removed.
6116 (REINTERP_SS): Declarations added.
6117 (REINTERP_US): Likewise.
6118 (REINTERP_PS): Likewise.
6119 (REINTERP_SU): Likewise.
6120 (REINTERP_SP): Likewise.
6121 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
6122 (vreinterpretq_p8_f64): Likewise.
6123 (vreinterpret_p16_f64): Likewise.
6124 (vreinterpretq_p16_f64): Likewise.
6125 (vreinterpret_f32_f64): Likewise.
6126 (vreinterpretq_f32_f64): Likewise.
6127 (vreinterpret_f64_f32): Likewise.
6128 (vreinterpret_f64_p8): Likewise.
6129 (vreinterpret_f64_p16): Likewise.
6130 (vreinterpret_f64_s8): Likewise.
6131 (vreinterpret_f64_s16): Likewise.
6132 (vreinterpret_f64_s32): Likewise.
6133 (vreinterpret_f64_s64): Likewise.
6134 (vreinterpret_f64_u8): Likewise.
6135 (vreinterpret_f64_u16): Likewise.
6136 (vreinterpret_f64_u32): Likewise.
6137 (vreinterpret_f64_u64): Likewise.
6138 (vreinterpretq_f64_f32): Likewise.
6139 (vreinterpretq_f64_p8): Likewise.
6140 (vreinterpretq_f64_p16): Likewise.
6141 (vreinterpretq_f64_s8): Likewise.
6142 (vreinterpretq_f64_s16): Likewise.
6143 (vreinterpretq_f64_s32): Likewise.
6144 (vreinterpretq_f64_s64): Likewise.
6145 (vreinterpretq_f64_u8): Likewise.
6146 (vreinterpretq_f64_u16): Likewise.
6147 (vreinterpretq_f64_u32): Likewise.
6148 (vreinterpretq_f64_u64): Likewise.
6149 (vreinterpret_s64_f64): Likewise.
6150 (vreinterpretq_s64_f64): Likewise.
6151 (vreinterpret_u64_f64): Likewise.
6152 (vreinterpretq_u64_f64): Likewise.
6153 (vreinterpret_s8_f64): Likewise.
6154 (vreinterpretq_s8_f64): Likewise.
6155 (vreinterpret_s16_f64): Likewise.
6156 (vreinterpretq_s16_f64): Likewise.
6157 (vreinterpret_s32_f64): Likewise.
6158 (vreinterpretq_s32_f64): Likewise.
6159 (vreinterpret_u8_f64): Likewise.
6160 (vreinterpretq_u8_f64): Likewise.
6161 (vreinterpret_u16_f64): Likewise.
6162 (vreinterpretq_u16_f64): Likewise.
6163 (vreinterpret_u32_f64): Likewise.
6164 (vreinterpretq_u32_f64): Likewise.
6166 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
6168 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
6169 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
6170 (vreinterpret_p8_s8): Likewise.
6171 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
6172 (vreinterpret_p8_s16): Likewise.
6173 (vreinterpret_p8_s32): Likewise.
6174 (vreinterpret_p8_s64): Likewise.
6175 (vreinterpret_p8_f32): Likewise.
6176 (vreinterpret_p8_u8): Likewise.
6177 (vreinterpret_p8_u16): Likewise.
6178 (vreinterpret_p8_u32): Likewise.
6179 (vreinterpret_p8_u64): Likewise.
6180 (vreinterpret_p8_p16): Likewise.
6181 (vreinterpretq_p8_s8): Likewise.
6182 (vreinterpretq_p8_s16): Likewise.
6183 (vreinterpretq_p8_s32): Likewise.
6184 (vreinterpretq_p8_s64): Likewise.
6185 (vreinterpretq_p8_f32): Likewise.
6186 (vreinterpretq_p8_u8): Likewise.
6187 (vreinterpretq_p8_u16): Likewise.
6188 (vreinterpretq_p8_u32): Likewise.
6189 (vreinterpretq_p8_u64): Likewise.
6190 (vreinterpretq_p8_p16): Likewise.
6191 (vreinterpret_p16_s8): Likewise.
6192 (vreinterpret_p16_s16): Likewise.
6193 (vreinterpret_p16_s32): Likewise.
6194 (vreinterpret_p16_s64): Likewise.
6195 (vreinterpret_p16_f32): Likewise.
6196 (vreinterpret_p16_u8): Likewise.
6197 (vreinterpret_p16_u16): Likewise.
6198 (vreinterpret_p16_u32): Likewise.
6199 (vreinterpret_p16_u64): Likewise.
6200 (vreinterpret_p16_p8): Likewise.
6201 (vreinterpretq_p16_s8): Likewise.
6202 (vreinterpretq_p16_s16): Likewise.
6203 (vreinterpretq_p16_s32): Likewise.
6204 (vreinterpretq_p16_s64): Likewise.
6205 (vreinterpretq_p16_f32): Likewise.
6206 (vreinterpretq_p16_u8): Likewise.
6207 (vreinterpretq_p16_u16): Likewise.
6208 (vreinterpretq_p16_u32): Likewise.
6209 (vreinterpretq_p16_u64): Likewise.
6210 (vreinterpretq_p16_p8): Likewise.
6211 (vreinterpret_f32_s8): Likewise.
6212 (vreinterpret_f32_s16): Likewise.
6213 (vreinterpret_f32_s32): Likewise.
6214 (vreinterpret_f32_s64): Likewise.
6215 (vreinterpret_f32_u8): Likewise.
6216 (vreinterpret_f32_u16): Likewise.
6217 (vreinterpret_f32_u32): Likewise.
6218 (vreinterpret_f32_u64): Likewise.
6219 (vreinterpret_f32_p8): Likewise.
6220 (vreinterpret_f32_p16): Likewise.
6221 (vreinterpretq_f32_s8): Likewise.
6222 (vreinterpretq_f32_s16): Likewise.
6223 (vreinterpretq_f32_s32): Likewise.
6224 (vreinterpretq_f32_s64): Likewise.
6225 (vreinterpretq_f32_u8): Likewise.
6226 (vreinterpretq_f32_u16): Likewise.
6227 (vreinterpretq_f32_u32): Likewise.
6228 (vreinterpretq_f32_u64): Likewise.
6229 (vreinterpretq_f32_p8): Likewise.
6230 (vreinterpretq_f32_p16): Likewise.
6231 (vreinterpret_s64_s8): Likewise.
6232 (vreinterpret_s64_s16): Likewise.
6233 (vreinterpret_s64_s32): Likewise.
6234 (vreinterpret_s64_f32): Likewise.
6235 (vreinterpret_s64_u8): Likewise.
6236 (vreinterpret_s64_u16): Likewise.
6237 (vreinterpret_s64_u32): Likewise.
6238 (vreinterpret_s64_u64): Likewise.
6239 (vreinterpret_s64_p8): Likewise.
6240 (vreinterpret_s64_p16): Likewise.
6241 (vreinterpretq_s64_s8): Likewise.
6242 (vreinterpretq_s64_s16): Likewise.
6243 (vreinterpretq_s64_s32): Likewise.
6244 (vreinterpretq_s64_f32): Likewise.
6245 (vreinterpretq_s64_u8): Likewise.
6246 (vreinterpretq_s64_u16): Likewise.
6247 (vreinterpretq_s64_u32): Likewise.
6248 (vreinterpretq_s64_u64): Likewise.
6249 (vreinterpretq_s64_p8): Likewise.
6250 (vreinterpretq_s64_p16): Likewise.
6251 (vreinterpret_u64_s8): Likewise.
6252 (vreinterpret_u64_s16): Likewise.
6253 (vreinterpret_u64_s32): Likewise.
6254 (vreinterpret_u64_s64): Likewise.
6255 (vreinterpret_u64_f32): Likewise.
6256 (vreinterpret_u64_u8): Likewise.
6257 (vreinterpret_u64_u16): Likewise.
6258 (vreinterpret_u64_u32): Likewise.
6259 (vreinterpret_u64_p8): Likewise.
6260 (vreinterpret_u64_p16): Likewise.
6261 (vreinterpretq_u64_s8): Likewise.
6262 (vreinterpretq_u64_s16): Likewise.
6263 (vreinterpretq_u64_s32): Likewise.
6264 (vreinterpretq_u64_s64): Likewise.
6265 (vreinterpretq_u64_f32): Likewise.
6266 (vreinterpretq_u64_u8): Likewise.
6267 (vreinterpretq_u64_u16): Likewise.
6268 (vreinterpretq_u64_u32): Likewise.
6269 (vreinterpretq_u64_p8): Likewise.
6270 (vreinterpretq_u64_p16): Likewise.
6271 (vreinterpret_s8_s16): Likewise.
6272 (vreinterpret_s8_s32): Likewise.
6273 (vreinterpret_s8_s64): Likewise.
6274 (vreinterpret_s8_f32): Likewise.
6275 (vreinterpret_s8_u8): Likewise.
6276 (vreinterpret_s8_u16): Likewise.
6277 (vreinterpret_s8_u32): Likewise.
6278 (vreinterpret_s8_u64): Likewise.
6279 (vreinterpret_s8_p8): Likewise.
6280 (vreinterpret_s8_p16): Likewise.
6281 (vreinterpretq_s8_s16): Likewise.
6282 (vreinterpretq_s8_s32): Likewise.
6283 (vreinterpretq_s8_s64): Likewise.
6284 (vreinterpretq_s8_f32): Likewise.
6285 (vreinterpretq_s8_u8): Likewise.
6286 (vreinterpretq_s8_u16): Likewise.
6287 (vreinterpretq_s8_u32): Likewise.
6288 (vreinterpretq_s8_u64): Likewise.
6289 (vreinterpretq_s8_p8): Likewise.
6290 (vreinterpretq_s8_p16): Likewise.
6291 (vreinterpret_s16_s8): Likewise.
6292 (vreinterpret_s16_s32): Likewise.
6293 (vreinterpret_s16_s64): Likewise.
6294 (vreinterpret_s16_f32): Likewise.
6295 (vreinterpret_s16_u8): Likewise.
6296 (vreinterpret_s16_u16): Likewise.
6297 (vreinterpret_s16_u32): Likewise.
6298 (vreinterpret_s16_u64): Likewise.
6299 (vreinterpret_s16_p8): Likewise.
6300 (vreinterpret_s16_p16): Likewise.
6301 (vreinterpretq_s16_s8): Likewise.
6302 (vreinterpretq_s16_s32): Likewise.
6303 (vreinterpretq_s16_s64): Likewise.
6304 (vreinterpretq_s16_f32): Likewise.
6305 (vreinterpretq_s16_u8): Likewise.
6306 (vreinterpretq_s16_u16): Likewise.
6307 (vreinterpretq_s16_u32): Likewise.
6308 (vreinterpretq_s16_u64): Likewise.
6309 (vreinterpretq_s16_p8): Likewise.
6310 (vreinterpretq_s16_p16): Likewise.
6311 (vreinterpret_s32_s8): Likewise.
6312 (vreinterpret_s32_s16): Likewise.
6313 (vreinterpret_s32_s64): Likewise.
6314 (vreinterpret_s32_f32): Likewise.
6315 (vreinterpret_s32_u8): Likewise.
6316 (vreinterpret_s32_u16): Likewise.
6317 (vreinterpret_s32_u32): Likewise.
6318 (vreinterpret_s32_u64): Likewise.
6319 (vreinterpret_s32_p8): Likewise.
6320 (vreinterpret_s32_p16): Likewise.
6321 (vreinterpretq_s32_s8): Likewise.
6322 (vreinterpretq_s32_s16): Likewise.
6323 (vreinterpretq_s32_s64): Likewise.
6324 (vreinterpretq_s32_f32): Likewise.
6325 (vreinterpretq_s32_u8): Likewise.
6326 (vreinterpretq_s32_u16): Likewise.
6327 (vreinterpretq_s32_u32): Likewise.
6328 (vreinterpretq_s32_u64): Likewise.
6329 (vreinterpretq_s32_p8): Likewise.
6330 (vreinterpretq_s32_p16): Likewise.
6331 (vreinterpret_u8_s8): Likewise.
6332 (vreinterpret_u8_s16): Likewise.
6333 (vreinterpret_u8_s32): Likewise.
6334 (vreinterpret_u8_s64): Likewise.
6335 (vreinterpret_u8_f32): Likewise.
6336 (vreinterpret_u8_u16): Likewise.
6337 (vreinterpret_u8_u32): Likewise.
6338 (vreinterpret_u8_u64): Likewise.
6339 (vreinterpret_u8_p8): Likewise.
6340 (vreinterpret_u8_p16): Likewise.
6341 (vreinterpretq_u8_s8): Likewise.
6342 (vreinterpretq_u8_s16): Likewise.
6343 (vreinterpretq_u8_s32): Likewise.
6344 (vreinterpretq_u8_s64): Likewise.
6345 (vreinterpretq_u8_f32): Likewise.
6346 (vreinterpretq_u8_u16): Likewise.
6347 (vreinterpretq_u8_u32): Likewise.
6348 (vreinterpretq_u8_u64): Likewise.
6349 (vreinterpretq_u8_p8): Likewise.
6350 (vreinterpretq_u8_p16): Likewise.
6351 (vreinterpret_u16_s8): Likewise.
6352 (vreinterpret_u16_s16): Likewise.
6353 (vreinterpret_u16_s32): Likewise.
6354 (vreinterpret_u16_s64): Likewise.
6355 (vreinterpret_u16_f32): Likewise.
6356 (vreinterpret_u16_u8): Likewise.
6357 (vreinterpret_u16_u32): Likewise.
6358 (vreinterpret_u16_u64): Likewise.
6359 (vreinterpret_u16_p8): Likewise.
6360 (vreinterpret_u16_p16): Likewise.
6361 (vreinterpretq_u16_s8): Likewise.
6362 (vreinterpretq_u16_s16): Likewise.
6363 (vreinterpretq_u16_s32): Likewise.
6364 (vreinterpretq_u16_s64): Likewise.
6365 (vreinterpretq_u16_f32): Likewise.
6366 (vreinterpretq_u16_u8): Likewise.
6367 (vreinterpretq_u16_u32): Likewise.
6368 (vreinterpretq_u16_u64): Likewise.
6369 (vreinterpretq_u16_p8): Likewise.
6370 (vreinterpretq_u16_p16): Likewise.
6371 (vreinterpret_u32_s8): Likewise.
6372 (vreinterpret_u32_s16): Likewise.
6373 (vreinterpret_u32_s32): Likewise.
6374 (vreinterpret_u32_s64): Likewise.
6375 (vreinterpret_u32_f32): Likewise.
6376 (vreinterpret_u32_u8): Likewise.
6377 (vreinterpret_u32_u16): Likewise.
6378 (vreinterpret_u32_u64): Likewise.
6379 (vreinterpret_u32_p8): Likewise.
6380 (vreinterpret_u32_p16): Likewise.
6381 (vreinterpretq_u32_s8): Likewise.
6382 (vreinterpretq_u32_s16): Likewise.
6383 (vreinterpretq_u32_s32): Likewise.
6384 (vreinterpretq_u32_s64): Likewise.
6385 (vreinterpretq_u32_f32): Likewise.
6386 (vreinterpretq_u32_u8): Likewise.
6387 (vreinterpretq_u32_u16): Likewise.
6388 (vreinterpretq_u32_u64): Likewise.
6389 (vreinterpretq_u32_p8): Likewise.
6390 (vreinterpretq_u32_p16): Likewise.
6392 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
6394 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
6395 Pattern extended.
6396 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator extended.
6397 (sqabs): Likewise.
6398 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
6399 (vqnegd_s64): Likewise.
6400 (vqabs_s64): Likewise.
6401 (vqabsd_s64): Likewise.
6403 2014-04-22 Richard Henderson <rth@redhat.com>
6405 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
6406 computation to the top of the loop.
6408 2014-04-22 Renlin <renlin.li@arm.com>
6409 Jiong Wang <jiong.wang@arm.com>
6411 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
6412 * config/aarch64/aarch64.c (aarch64_layout_frame)
6413 (aarch64_initial_elimination_offset): Likewise.
6415 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
6417 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
6418 Fix indentation.
6420 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
6422 * machmode.h (bitwise_mode_for_mode): Declare.
6423 * stor-layout.h (bitwise_type_for_mode): Likewise.
6424 * stor-layout.c (bitwise_mode_for_mode): New function.
6425 (bitwise_type_for_mode): Likewise.
6426 * builtins.c (fold_builtin_memory_op): Use it instead of
6427 int_mode_for_mode and build_nonstandard_integer_type.
6429 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6431 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
6432 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
6433 (*-*-solaris2*): Simplify.
6434 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
6435 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
6436 *-*-solaris2.9* handling.
6438 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
6439 as bug.
6440 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
6441 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
6442 handling, simplify.
6443 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
6444 * configure: Regenerate.
6446 * config/i386/sol2-9.h: Remove.
6448 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
6449 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
6450 Remove Solaris 9 references.
6452 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
6454 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
6455 (floatuns<GPI:mode><GPF:mode>2): Remove.
6456 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
6457 and floatuns conversions.
6458 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
6459 and floatuns conversions.
6460 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
6461 (w1,w2): New mode attributes for inequal width conversions.
6463 2014-04-22 Renlin Li <Renlin.Li@arm.com>
6465 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
6466 the output asm format.
6468 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
6470 * config/aarch64/aarch64-simd.md
6471 (aarch64_cm<optab>di): Always split.
6472 (*aarch64_cm<optab>di): New.
6473 (aarch64_cmtstdi): Always split.
6474 (*aarch64_cmtstdi): New.
6476 2014-04-22 Jakub Jelinek <jakub@redhat.com>
6478 PR tree-optimization/60823
6479 * omp-low.c (ipa_simd_modify_function_body): Go through
6480 all SSA_NAMEs and for those refering to vector arguments
6481 which are going to be replaced adjust SSA_NAME_VAR and,
6482 if it is a default definition, change it into a non-default
6483 definition assigned at the beginning of function from new_decl.
6484 (ipa_simd_modify_stmt_ops): Rewritten.
6485 * tree-dfa.c (set_ssa_default_def): When removing default def,
6486 check for NULL loc instead of NULL *loc.
6488 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6490 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
6491 restrictions on core registers for DImode values in Thumb2.
6493 2014-04-22 Ian Bolton <ian.bolton@arm.com>
6495 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
6496 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
6498 2014-04-22 Ian Bolton <ian.bolton@arm.com>
6500 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
6501 (*iordi_notzesidi_di): Likewise.
6502 (*iordi_notsesidi_di): Likewise.
6504 2014-04-22 Ian Bolton <ian.bolton@arm.com>
6506 * config/arm/arm-protos.h (tune_params): New struct members.
6507 * config/arm/arm.c: Initialise tune_params per processor.
6508 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
6509 for speed, based on new tune_params.
6511 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
6513 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro added.
6514 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added macro.
6515 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment corrected.
6516 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
6517 * config/aarch64/arm_neon.h (vrnd_f64): Added.
6518 (vrnda_f64): Likewise.
6519 (vrndi_f64): Likewise.
6520 (vrndm_f64): Likewise.
6521 (vrndn_f64): Likewise.
6522 (vrndp_f64): Likewise.
6523 (vrndx_f64): Likewise.
6525 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6527 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
6528 GET_MODE_SIZE argument is enum machine_mode.
6530 2014-04-22 Jakub Jelinek <jakub@redhat.com>
6532 PR target/60910
6533 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
6534 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
6536 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
6538 PR middle-end/60281
6539 * asan.c (asan_emit_stack_protection): Force the base to align to
6540 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
6541 appropriate bits if STRICT_ALIGNMENT.
6542 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
6543 when asan is on.
6544 (expand_used_vars): Leave a space in the stack frame for alignment
6545 if STRICT_ALIGNMENT.
6547 2014-04-21 David Malcolm <dmalcolm@redhat.com>
6549 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
6550 than a gimple.
6551 (gimple_store_p): Likewise.
6552 (gimple_assign_load_p): Likewise.
6553 (gimple_assign_cast_p): Likewise.
6554 (gimple_clobber_p): Likewise.
6556 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
6557 rather than a gimple.
6558 (gimple_assign_cast_p): Likewise.
6560 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
6562 PR target/60735
6563 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
6564 If mode is DDmode and TARGET_E500_DOUBLE allow move.
6566 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
6567 more debug information for E500 if -mdebug=reg.
6569 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
6571 PR target/60909
6572 * config/i386/i386.c (ix86_expand_builtin)
6573 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
6574 register for target RTX.
6575 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
6577 2014-04-18 Cong Hou <congh@google.com>
6579 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
6580 the widen-mult pattern by handling two operands with different sizes,
6581 and operands whose size is smaller than half of the result type.
6583 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
6585 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
6586 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
6587 (do_estimate_edge_time): Compute it.
6588 * ipa-inline.c (want_inline_small_function_p): Bypass
6589 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
6591 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
6593 * ipa-inline.c (spec_rem): New static variable.
6594 (dump_overall_stats): New function.
6595 (dump_inline_stats): New function.
6597 2014-04-18 Richard Henderson <rth@redhat.com>
6599 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
6600 to GET_MODE_SIZE, not a reg_class_t.
6602 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6604 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
6605 (vsx_xxmrglw_<mode>): Likewise.
6607 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
6609 PR target/60876
6610 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
6611 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
6612 (rs6000_init_hard_regno_mode_ok): Likewise.
6614 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
6616 * ipa-inline.c (inline_small_functions): Account only non-cold
6617 functions.
6618 * doc/invoke.texi (inline-unit-growth): Update documentation.
6620 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
6622 * config/rs6000/rs6000.md (addti3, subti3): New.
6624 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
6626 PR target/60863
6627 * config/i386/i386.c (ix86_expand_clear): Remove outdated
6628 comment. Check optimize_insn_for_size_p instead of
6629 optimize_insn_for_speed_p.
6631 2014-04-17 Martin Jambor <mjambor@suse.cz>
6633 * gimple-iterator.c (gsi_start_edge): New function.
6634 * gimple-iterator.h (gsi_start_edge): Declare.
6635 * tree-sra.c (single_non_eh_succ): New function.
6636 (disqualify_ops_if_throwing_stmt): Renamed to
6637 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
6638 having one non-EH successor BB.
6639 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
6640 generate loads into replacements.
6641 (sra_modify_assign): Likewise and and also use the simple path for
6642 such statements.
6643 (sra_modify_function_body): Commit statements on edges.
6645 2014-04-17 Richard Biener <rguenther@suse.de>
6647 PR middle-end/60849
6648 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
6649 comparison results and add clarifying comment.
6651 2014-04-17 Jakub Jelinek <jakub@redhat.com>
6653 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
6654 (blank_mode): Initialize it.
6655 (emit_mode_size_inline, emit_mode_nunits_inline,
6656 emit_mode_inner_inline): New functions.
6657 (emit_insn_modes_h): Call them and surround their output with
6658 #if GCC_VERSION >= 4001 ... #endif.
6659 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
6660 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
6661 mode_* arrays if the argument is __builtin_constant_p.
6662 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
6663 is enum machine_mode.
6665 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6667 * passes.c (opt_pass::execute): Adjust.
6668 (pass_manager::execute_pass_mode_switching): Likewise.
6669 (early_local_passes::execute): Likewise.
6670 (execute_one_pass): Pass cfun to the pass's execute method.
6671 * tree-pass.h (opt_pass::execute): Add function * argument.
6672 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
6673 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
6674 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
6675 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
6676 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
6677 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
6678 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
6679 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
6680 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
6681 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
6682 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
6683 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
6684 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
6685 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
6686 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
6687 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
6688 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
6689 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
6690 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
6691 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
6692 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
6693 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
6694 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
6695 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
6696 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
6697 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
6698 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
6699 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
6700 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
6701 Adjust.
6703 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6705 * passes.c (opt_pass::gate): Take function * argument.
6706 (gate_all_early_local_passes): Merge into
6707 (early_local_passes::gate): this.
6708 (gate_all_early_optimizations): Merge into
6709 (all_early_optimizations::gate): this.
6710 (gate_all_optimizations): Mege into
6711 (all_optimizations::gate): this.
6712 (gate_all_optimizations_g): Merge into
6713 (all_optimizations_g::gate): this.
6714 (gate_rest_of_compilation): Mege into
6715 (rest_of_compilation::gate): this.
6716 (gate_postreload): Merge into
6717 (postreload::gate): this.
6718 (dump_one_pass): Pass cfun to the pass's gate method.
6719 (execute_ipa_summary_passes): Likewise.
6720 (execute_one_pass): Likewise.
6721 (ipa_write_summaries_2): Likewise.
6722 (ipa_write_optimization_summaries_1): Likewise.
6723 (ipa_read_summaries_1): Likewise.
6724 (ipa_read_optimization_summaries_1): Likewise.
6725 (execute_ipa_stmt_fixups): Likewise.
6726 * tree-pass.h (opt_pass::gate): Add function * argument.
6727 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
6728 combine-stack-adj.c, combine.c, compare-elim.c,
6729 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
6730 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
6731 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
6732 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
6733 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
6734 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
6735 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
6736 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
6737 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
6738 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
6739 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
6740 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
6741 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
6742 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
6743 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
6744 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
6745 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
6746 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
6747 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
6748 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
6749 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
6750 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
6751 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
6752 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
6753 var-tracking.c, vtable-verify.c, web.c: Adjust.
6755 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6757 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
6758 * configure: Regenerate.
6760 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6762 * passes.c (dump_one_pass): don't check pass->has_gate.
6763 (execute_ipa_summary_passes): Likewise.
6764 (execute_one_pass): Likewise.
6765 (ipa_write_summaries_2): Likewise.
6766 (ipa_write_optimization_summaries_1): Likewise.
6767 (ipa_read_optimization_summaries_1): Likewise.
6768 (execute_ipa_stmt_fixups): Likewise.
6769 * tree-pass.h (pass_data::has_gate): Remove.
6770 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
6771 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
6772 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
6773 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
6774 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
6775 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
6776 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
6777 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
6778 gimple-low.c, gimple-ssa-isolate-paths.c,
6779 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
6780 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
6781 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
6782 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
6783 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
6784 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
6785 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
6786 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
6787 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
6788 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
6789 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
6790 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
6791 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
6792 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
6793 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
6794 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
6795 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
6796 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
6797 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
6798 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
6799 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
6800 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
6801 Adjust.
6803 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6805 * pass_manager.h (pass_manager::register_dump_files_1): Remove
6806 declaration.
6807 * passes.c (pass_manager::register_dump_files_1): Merge into
6808 (pass_manager::register_dump_files): this, and remove its handling of
6809 properties since the pass always has the properties anyway.
6810 (pass_manager::pass_manager): Adjust.
6812 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6814 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
6815 * passes.c (pass_manager::register_dump_files_1): Remove dead code
6816 dealing with properties.
6817 (pass_manager::register_dump_files): Adjust.
6819 2014-03-20 Mark Wielaard <mjw@redhat.com>
6821 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
6822 then represent the bound as normal constant value.
6824 2014-04-17 Jakub Jelinek <jakub@redhat.com>
6826 PR target/60847
6827 Forward port from 4.8 branch
6828 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
6830 * config/i386/bmiintrin.h (_blsi_u32): New.
6831 (_blsi_u64): Ditto.
6832 (_blsr_u32): Ditto.
6833 (_blsr_u64): Ditto.
6834 (_blsmsk_u32): Ditto.
6835 (_blsmsk_u64): Ditto.
6836 (_tzcnt_u32): Ditto.
6837 (_tzcnt_u64): Ditto.
6839 2014-04-17 Kito Cheng <kito@0xlab.org>
6841 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
6843 2014-04-17 Richard Biener <rguenther@suse.de>
6845 PR middle-end/60849
6846 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
6847 boolean results for comparisons.
6849 2014-04-17 Richard Biener <rguenther@suse.de>
6851 PR tree-optimization/60836
6852 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
6853 initial PHI args to be gimple values.
6855 2014-04-17 Richard Biener <rguenther@suse.de>
6857 PR tree-optimization/60841
6858 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
6859 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
6860 of stmts to SLP build.
6861 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
6862 (vect_analyze_slp): Likewise.
6863 (vect_analyze_slp_instance): Likewise.
6864 (vect_build_slp_tree): Limit overall SLP tree growth.
6865 * tree-vectorizer.h (vect_analyze_data_refs,
6866 vect_analyze_slp): Adjust prototypes.
6868 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
6870 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
6871 Silvermont.
6873 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
6875 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
6876 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
6877 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
6878 for TARGET_SLOW_PSHUFB
6880 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
6882 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
6883 * config/i386/i386.c (intel_cost): Ditto.
6885 2014-04-17 Joey Ye <joey.ye@arm.com>
6887 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
6889 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
6891 * opts.c (common_handle_option): Disable -fipa-reference coorectly
6892 with -fuse-profile.
6894 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
6896 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
6897 (type_all_derivations_known_p): New predicate.
6898 (type_all_ctors_visible_p): New predicate.
6899 (type_possibly_instantiated_p): New predicate.
6900 (get_odr_type): Compute all_derivations_known.
6901 (dump_odr_type): Dump the flag.
6902 (maybe_record_type): Cleanup.
6903 (record_target_from_binfo): Add bases_to_consider array;
6904 record bases for types w/o instances and skip CXX destructor.
6905 (possible_polymorphic_call_targets_1): Add bases_to_consider
6906 and consider_construction parameters; check if type may have instance.
6907 (get_polymorphic_call_info): Set maybe_in_construction to true
6908 when we know nothing.
6909 (record_targets_from_bases): Skip CXX destructors; they are
6910 never called for types in construction.
6911 (possible_polymorphic_call_targets): Do not record target when
6912 type may not have instance.
6914 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
6916 PR ipa/60854
6917 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
6918 external aliases alive, too.
6920 2014-04-16 Andrew Pinski <apinski@cavium.com>
6922 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
6923 definition.
6925 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
6927 * final.c (compute_alignments): Do not apply loop alignment to a block
6928 falling through to the exit.
6930 2014-04-16 Catherine Moore <clm@codesourcery.com>
6932 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
6933 Adjust constraints for microMIPS store patterns.
6935 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
6937 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
6939 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
6941 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
6942 (append_use): Run at -O0.
6943 (append_vdef): Likewise.
6944 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
6945 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
6947 2014-04-16 Jakub Jelinek <jakub@redhat.com>
6949 PR tree-optimization/60844
6950 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
6951 (propagate_op_to_single_use, remove_visited_stmt_chain,
6952 linearize_expr, repropagate_negates, reassociate_bb): Use it
6953 instead of gsi_remove.
6955 2014-04-16 Martin Jambor <mjambor@suse.cz>
6957 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
6958 ipa_transforms_to_apply.
6959 (cgraph_function_versioning): Assert that old_node has empty
6960 ipa_transforms_to_apply.
6961 * trans-mem.c (ipa_tm_create_version): Likewise.
6962 * tree-inline.c (tree_function_versioning): Do not duplicate
6963 ipa_transforms_to_apply.
6965 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6967 PR target/60817
6968 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
6969 x86_64-*-* cases.
6970 Pass necessary as flags on 64-bit Solaris/x86.
6971 Use lowercase relocs for x86_64-*-*.
6972 * configure: Regenerate.
6974 2014-04-15 Jan Hubicka <jh@suse.cz>
6976 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
6977 (maybe_record_node, likely_target_p): Use it.
6979 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6981 PR target/60839
6982 Revert following patch
6984 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
6986 PR target/60735
6987 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
6988 software floating point or no floating point registers, do not
6989 allow any type in the FPRs. Eliminate a test for SPE SIMD types
6990 in GPRs that occurs after we tested for GPRs that would never be
6991 true.
6993 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
6994 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
6995 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
6996 specifically allow DDmode, since that does not use the SPE SIMD
6997 instructions.
6999 2014-03-21 Mark Wielaard <mjw@redhat.com>
7001 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
7002 as unsigned or int depending on type and value used.
7004 2014-04-15 Richard Biener <rguenther@suse.de>
7006 PR rtl-optimization/56965
7007 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
7008 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
7009 ... here.
7010 * alias.c (true_dependence_1): Do not call
7011 nonoverlapping_component_refs_p.
7012 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
7013 nonoverlapping_component_refs_p.
7014 (indirect_refs_may_alias_p): Likewise.
7016 2014-04-15 Teresa Johnson <tejohnson@google.com>
7018 * cfg.c (dump_bb_info): Fix flags check.
7019 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
7021 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7023 PR rtl-optimization/60663
7024 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
7025 avoid 0 cost.
7027 2014-04-15 Richard Biener <rguenther@suse.de>
7029 * lto-streamer.h (LTO_major_version): Bump to 4.
7031 2014-04-15 Richard Biener <rguenther@suse.de>
7033 * common.opt (lto_partition_model): New enum.
7034 (flto-partition=): Merge separate options with a single with argument,
7035 add -flto-partition=one support.
7036 * flag-types.h (enum lto_partition_model): Declare.
7037 * opts.c (finish_options): Remove duplicate -flto-partition=
7038 option check.
7039 * lto-wrapper.c (run_gcc): Adjust.
7041 2014-04-15 Richard Biener <rguenther@suse.de>
7043 * alias.c (ncr_compar): New function.
7044 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
7046 2014-04-15 Richard Biener <rguenther@suse.de>
7048 * alias.c (record_component_aliases): Do not walk BINFOs.
7050 2014-04-15 Richard Biener <rguenther@suse.de>
7052 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
7053 Add struct function argument and adjust.
7054 (find_func_aliases_for_call): Likewise.
7055 (find_func_aliases): Likewise.
7056 (find_func_clobbers): Likewise.
7057 (intra_create_variable_infos): Likewise.
7058 (compute_points_to_sets): Likewise.
7059 (ipa_pta_execute): Adjust. Do not push/pop cfun.
7061 2014-04-15 Richard Biener <rguenther@suse.de>
7063 * tree.c (iterative_hash_expr): Use enum tree_code_class
7064 to store TREE_CODE_CLASS.
7065 (tree_block): Likewise.
7066 (tree_set_block): Likewise.
7067 * tree.h (fold_build_pointer_plus_loc): Use
7068 convert_to_ptrofftype_loc.
7070 2014-04-15 Jakub Jelinek <jakub@redhat.com>
7072 PR plugins/59335
7073 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
7074 added in 4.9.
7076 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
7078 * cfgloop.h (struct loop): Move force_vectorize down.
7079 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
7080 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
7081 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
7082 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
7083 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
7084 * tree-core.h (enum annot_expr_kind): Add new kind values.
7085 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
7086 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
7087 kinds.
7088 * tree.def (ANNOTATE_EXPR): Tweak comment.
7090 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
7092 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
7093 cxa_pure_virtual).
7095 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
7097 * tree.h (TYPE_IDENTIFIER): Declare.
7098 * tree.c (subrange_type_for_debug_p): Use it.
7099 * godump.c (go_format_type): Likewise.
7100 * dwarf2out.c (is_cxx_auto, modified_type_die,
7101 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
7102 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
7104 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
7106 PR lto/60820
7107 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
7109 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
7111 * config/i386/i386.c (examine_argument): Return bool. Return true if
7112 parameter should be passed in memory.
7113 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
7114 (construct_container): Update calls to examine_argument.
7115 (function_arg_advance_64): Ditto.
7116 (return_in_memory_32): Merge with ix86_return_in_memory.
7117 (return_in_memory_64): Ditto.
7118 (return_in_memory_ms_64): Ditto.
7120 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
7122 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
7123 * coverage.c (coverage_compute_profile_id): Handle externally visible
7124 symbols.
7126 2014-04-14 Martin Jambor <mjambor@suse.cz>
7128 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
7129 DECL_DISREGARD_INLINE_LIMITS functions.
7131 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
7133 PR target/60827
7134 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
7136 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
7138 PR target/60827
7139 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
7140 optimize_insn_for_speed_p instead of
7141 optimize_function_for_speed_p.
7143 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
7145 * doc/invoke.texi (free): Document AArch64.
7147 2014-04-14 Richard Biener <rguenther@suse.de>
7149 PR tree-optimization/60042
7150 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
7151 (insert_into_preds_of_block): Do not prevent PHI insertion
7152 for REFERENCE exprs here ...
7153 (eliminate_dom_walker::before_dom_children): ... but prevent
7154 their use here under similar conditions when applied to the
7155 IL after PRE optimizations.
7157 2014-04-14 Richard Biener <rguenther@suse.de>
7159 * passes.def: Move early points-to after early SRA.
7161 2014-04-14 Richard Biener <rguenther@suse.de>
7163 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
7164 check for which sign-changes we allow when forwarding
7165 a converted value into a switch.
7167 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
7169 * stor-layout.c (place_field): Finalize non-constant offset for the
7170 field, if any.
7172 2014-04-14 Richard Biener <rguenther@suse.de>
7174 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
7175 as argument.
7176 (expand_switch_using_bit_tests_p): Likewise.
7177 (process_switch): Compute and pass on speed_p based on the
7178 switch stmt.
7179 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
7180 optimize_bb_for_speed_p.
7182 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
7184 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
7185 * function.h (struct function): Rename has_force_vect_loops into
7186 has_force_vectorize_loops.
7187 * lto-streamer-in.c (input_cfg): Adjust for renaming.
7188 (input_struct_function_base): Likewise.
7189 * lto-streamer-out.c (output_cfg): Likewise.
7190 (output_struct_function_base): Likewise.
7191 * omp-low.c (expand_omp_simd): Likewise.
7192 * tree-cfg.c (move_sese_region_to_fn): Likewise.
7193 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
7194 (version_loop_for_if_conversion): Likewise.
7195 (tree_if_conversion): Likewise.
7196 (main_tree_if_conversion): Likewise.
7197 (gate_tree_if_conversion): Likewise.
7198 * tree-inline.c (copy_loops): Likewise.
7199 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
7200 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
7201 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
7202 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
7203 * tree-vectorizer.c (vectorize_loops): Likewise.
7204 * tree-vectorizer.h (unlimited_cost_model): Likewise.
7206 2014-04-14 Richard Biener <rguenther@suse.de>
7208 PR lto/60720
7209 * lto-streamer-out.c (wrap_refs): New function.
7210 (lto_output): Wrap symbol references in global initializes in
7211 type-preserving MEM_REFs.
7213 2014-04-14 Christian Bruel <christian.bruel@st.com>
7215 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
7217 2014-04-14 Christian Bruel <christian.bruel@st.com>
7219 * config/sh/sh.md (setmemqi): New expand pattern.
7220 * config/sh/sh.h (CLEAR_RATIO): Define.
7221 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
7222 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
7224 2014-04-14 Richard Biener <rguenther@suse.de>
7226 PR middle-end/55022
7227 * fold-const.c (negate_expr_p): Don't negate directional rounding
7228 division.
7229 (fold_negate_expr): Likewise.
7231 2014-04-14 Richard Biener <rguenther@suse.de>
7233 PR tree-optimization/59817
7234 PR tree-optimization/60453
7235 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
7236 recursion to catch all CHRECs in the scalar evolution and restrict
7237 the predicate for the remains appropriately.
7239 2014-04-12 Catherine Moore <clm@codesourcery.com>
7241 * config/mips/constraints.md: Add new register constraint "kb".
7242 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
7243 (*movhi_internal): Likewise.
7244 (*movqi_internal): Likewise.
7245 * config/mips/mips.h (M16_STORE_REGS): New register class.
7246 (REG_CLASS_NAMES): Add M16_STORE_REGS.
7247 (REG_CLASS_CONTENTS): Likewise.
7248 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
7250 2014-04-11 Tobias Burnus <burnus@net-b.de>
7252 PR c/60194
7253 * doc/invoke.texi (-Wformat-signedness): Document it.
7254 (Wformat=2): Mention that this enables -Wformat-signedness.
7256 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
7258 * common/config/epiphany/epiphany-common.c
7259 (epiphany_option_optimization_table): Enable section anchors by
7260 default at -O1 or higher.
7261 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
7262 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
7263 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
7264 carries no extra cost.
7265 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
7266 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
7267 * config/epiphany/predicates.md (memclob_operand): New predicate.
7268 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
7269 Use memclob_operand predicate and X constraint for operand 3.
7271 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
7273 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
7274 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
7275 its operands.
7277 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
7279 PR rtl-optimization/60651
7280 * mode-switching.c (optimize_mode_switching): Make sure to emit
7281 sets of a lower numbered entity before sets of a higher numbered
7282 entity to a mode of the same or lower priority.
7283 When creating a seginfo for a basic block that starts with a code
7284 label, move the insertion point past the code label.
7285 (new_seginfo): Document and enforce requirement that
7286 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
7287 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
7288 * doc/tm.texi: Regenerate.
7290 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
7292 PR target/60811
7293 * config/arc/arc.c (arc_save_restore): Fix assert typo.
7295 2013-04-11 Jakub Jelinek <jakub@redhat.com>
7297 * BASE-VER: Set to 4.10.0.
7299 2014-04-11 Tobias Burnus <burnus@net-b.de>
7301 PR other/59055
7302 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
7303 * doc/gcc.texi (Service): Update description in the @menu
7304 * doc/invoke.texi (Option Summary): Remove misplaced and
7305 duplicated @menu.
7307 2014-04-11 Steve Ellcey <sellcey@mips.com>
7308 Jakub Jelinek <jakub@redhat.com>
7310 PR middle-end/60556
7311 * expr.c (convert_move): Use emit_store_flag_force instead of
7312 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
7313 argument to it.
7315 2014-04-11 Richard Biener <rguenther@suse.de>
7317 PR middle-end/60797
7318 * varasm.c (assemble_alias): Avoid endless error reporting
7319 recursion by setting TREE_ASM_WRITTEN.
7321 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7323 * config/s390/s390.md: Add a splitter for NOT rtx.
7325 2014-04-11 Jakub Jelinek <jakub@redhat.com>
7327 PR rtl-optimization/60663
7328 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
7330 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
7331 Jakub Jelinek <jakub@redhat.com>
7333 PR lto/60567
7334 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
7335 flag from decl_node to node.
7337 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7339 PR debug/60655
7340 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
7341 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
7342 ameliorating the cases where it can be.
7344 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
7346 Revert
7347 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
7349 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
7350 (loadsync_<mode>): Change mode.
7351 (load_quadpti, store_quadpti): New.
7352 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
7353 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
7354 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
7356 2014-04-09 Cong Hou <congh@google.com>
7358 PR testsuite/60773
7359 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
7360 documentation.
7362 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7364 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
7365 instead of vnor to exploit possible fusion opportunity in the
7366 future.
7367 (altivec_expand_vec_perm_const_le): Likewise.
7369 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
7371 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
7372 (loadsync_<mode>): Change mode.
7373 (load_quadpti, store_quadpti): New.
7374 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
7375 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
7377 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
7379 PR target/60763
7380 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
7381 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
7382 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
7384 2014-04-08 Richard Biener <rguenther@suse.de>
7386 PR middle-end/60706
7387 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
7388 a 64bit widest int print double-int similar to on HWI64 hosts.
7390 2014-04-08 Richard Biener <rguenther@suse.de>
7392 PR tree-optimization/60785
7393 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
7394 default defs properly.
7396 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
7398 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
7399 (Weffc++): Likewise.
7401 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
7403 * ipa-devirt.c (maybe_record_node): When node is not recorded,
7404 set completep to false rather than true.
7406 2014-04-07 Douglas B Rupp <rupp@adacore.com>
7408 PR target/60504
7409 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
7410 ARM_TARGET2_DWARF_FORMAT.
7412 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
7414 PR target/60609
7415 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
7416 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
7417 ADDR_DIFF_VEC.
7419 2014-04-07 Richard Biener <rguenther@suse.de>
7421 PR tree-optimization/60766
7422 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
7423 (may_eliminate_iv): Convert cand_value_at result to desired type.
7425 2014-04-07 Jason Merrill <jason@redhat.com>
7427 PR c++/60731
7428 * common.opt (-fno-gnu-unique): Add.
7429 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
7431 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7433 * haifa-sched.c: Fix outdated function reference and minor
7434 grammar errors in introductory comment.
7436 2014-04-07 Richard Biener <rguenther@suse.de>
7438 PR middle-end/60750
7439 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
7440 for noreturn calls.
7441 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
7443 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
7445 PR debug/55794
7446 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
7447 size accounting for thunks.
7448 (pa_asm_output_mi_thunk): Use final_start_function() and
7449 final_end_function() to output function start and end directives.
7451 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
7453 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
7454 device specific ISA/ feature information. Remove short_sp and
7455 errata_skip ds. Add avr_device_specific_features enum to have device
7456 specific info.
7457 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
7458 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
7459 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
7460 updated device specific info.
7461 * config/avr/avr-mcus.def: Merge device specific details to
7462 dev_attribute field.
7463 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
7464 errata_skip.
7465 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
7466 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
7467 assembler if RMW isa supported by current device.
7468 * config/avr/genmultilib.awk: Update as device info structure changed.
7469 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
7471 2014-04-04 Cong Hou <congh@google.com>
7473 PR tree-optimization/60656
7474 * tree-vect-stmts.c (supportable_widening_operation):
7475 Fix a bug that elements in a vector with vect_used_by_reduction
7476 property are incorrectly reordered when the operation on it is not
7477 consistant with the one in reduction operation.
7479 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
7481 PR rtl-optimization/60155
7482 * gcse.c (record_set_data): New function.
7483 (single_set_gcse): New function.
7484 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
7485 (hoist_code): Likewise.
7486 (get_pressure_class_and_nregs): Likewise.
7488 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
7490 * explow.c (probe_stack_range): Emit a final optimization blockage.
7492 2014-04-04 Anthony Green <green@moxielogic.com>
7494 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
7495 typos.
7497 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
7499 PR ipa/59626
7500 * lto-cgraph.c (input_overwrite_node): Check that partitioning
7501 flags are set only during streaming.
7502 * ipa.c (process_references, walk_polymorphic_call_targets,
7503 symtab_remove_unreachable_nodes): Drop bodies of always inline
7504 after early inlining.
7505 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
7507 2014-04-04 Jakub Jelinek <jakub@redhat.com>
7508 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7510 PR debug/60655
7511 * dwarf2out.c (const_ok_for_output_1): Reject expressions
7512 containing a NOT.
7514 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7516 PR bootstrap/60743
7517 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
7518 duration.
7519 (cortex_a53_fdivd): Likewise.
7521 2014-04-04 Martin Jambor <mjambor@suse.cz>
7523 PR ipa/60640
7524 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
7525 Adjust all callers.
7526 * cgraph.c (clone_of_p): Also return true if thunks match.
7527 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
7528 cgraph_function_or_thunk_node and an obsolete comment.
7529 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
7530 file.
7531 (build_function_decl_skip_args): Likewise.
7532 (set_new_clone_decl_and_node_flags): New function.
7533 (duplicate_thunk_for_node): Likewise.
7534 (redirect_edge_duplicating_thunks): Likewise.
7535 (cgraph_clone_node): New parameter args_to_skip, pass it to
7536 redirect_edge_duplicating_thunks which is called instead of
7537 cgraph_redirect_edge_callee.
7538 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
7539 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
7541 2014-04-04 Jeff Law <law@redhat.com>
7543 PR target/60657
7544 * config/arm/predicates.md (const_int_I_operand): New predicate.
7545 (const_int_M_operand): Similarly.
7546 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
7547 const_int_operand.
7548 (insv_t2, extv_reg, extzv_t2): Likewise.
7549 (load_multiple_with_writeback): Similarly for const_int_I_operand.
7550 (pop_multiple_with_writeback_and_return): Likewise.
7551 (vfp_pop_multiple_with_writeback): Likewise
7553 2014-04-04 Richard Biener <rguenther@suse.de>
7555 PR ipa/60746
7556 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
7557 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
7558 non-GIMPLE_LABELs.
7559 * gimplify.h (gimple_add_tmp_var_fn): Declare.
7560 * gimplify.c (gimple_add_tmp_var_fn): New function.
7561 * gimple-expr.h (create_tmp_reg_fn): Declare.
7562 * gimple-expr.c (create_tmp_reg_fn): New function.
7563 * gimple-low.c (record_vars_into): Don't change cfun.
7564 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
7565 code generation without cfun.
7567 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
7569 PR bootstrap/60719
7570 * Makefile.in (install-driver): Fix shell scripting.
7572 2014-04-03 Cong Hou <congh@google.com>
7574 PR tree-optimization/60505
7575 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
7576 threshold of number of iterations below which no vectorization
7577 will be done.
7578 * tree-vect-loop.c (new_loop_vec_info):
7579 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
7580 * tree-vect-loop.c (vect_analyze_loop_operations):
7581 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
7582 * tree-vect-loop.c (vect_transform_loop):
7583 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
7584 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
7585 of iterations of the loop and see if we should build the epilogue.
7587 2014-04-03 Richard Biener <rguenther@suse.de>
7589 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
7590 (streamer_tree_cache_create): Adjust.
7591 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
7592 to allow optional nodes array.
7593 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
7594 (streamer_tree_cache_append): Likewise.
7595 (streamer_tree_cache_create): Create nodes array optionally
7596 as specified by parameter.
7597 * lto-streamer-out.c (create_output_block): Avoid maintaining
7598 the node array in the writer cache.
7599 (DFS_write_tree): Remove assertion.
7600 (produce_asm_for_decls): Free the out decl state hash table early.
7601 * lto-streamer-in.c (lto_data_in_create): Adjust for
7602 streamer_tree_cache_create prototype change.
7604 2014-04-03 Richard Biener <rguenther@suse.de>
7606 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
7607 set TREE_CHAIN to NULL_TREE.
7609 2014-04-03 Richard Biener <rguenther@suse.de>
7611 PR tree-optimization/60740
7612 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
7613 over all GIMPLE_COND operands.
7615 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
7617 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
7618 (Weffc++): Remove Scott's numbering, merge lists and reference
7619 Wnon-virtual-dtor.
7621 2014-04-03 Nick Clifton <nickc@redhat.com>
7623 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
7624 properly.
7626 2014-04-03 Martin Jambor <mjambor@suse.cz>
7628 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
7629 mention gcc_unreachable before failing.
7630 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
7631 removed symbols.
7633 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
7635 PR ipa/60659
7636 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
7637 inconsistent code and instead mark the context inconsistent.
7638 (possible_polymorphic_call_targets): For inconsistent contexts
7639 return empty complete list.
7641 2014-04-02 Anthony Green <green@moxielogic.com>
7643 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
7644 (extendqisi2, extendhisi2): Define.
7645 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
7646 (WCHAR_TYPE): Change to unsigned int.
7648 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7650 PR tree-optimization/60733
7651 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
7652 insertion point for PHI candidates to be the end of the feeding
7653 block for the PHI argument.
7655 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
7657 PR rtl-optimization/60650
7658 * lra-constraints.c (process_alt_operands): Decrease reject for
7659 earlyclobber matching.
7661 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7663 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
7665 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7667 * config/spu/spu.c (pad_bb): Do not crash when the last
7668 insn is CODE_FOR_blockage.
7670 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7672 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
7673 lies outside the target mode.
7675 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
7677 PR target/60735
7678 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
7679 software floating point or no floating point registers, do not
7680 allow any type in the FPRs. Eliminate a test for SPE SIMD types
7681 in GPRs that occurs after we tested for GPRs that would never be
7682 true.
7684 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
7685 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
7686 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
7687 specifically allow DDmode, since that does not use the SPE SIMD
7688 instructions.
7690 2014-04-02 Richard Biener <rguenther@suse.de>
7692 PR middle-end/60729
7693 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
7694 MODE_INTs. Properly use negv_optab.
7695 (expand_abs): Likewise.
7697 2014-04-02 Richard Biener <rguenther@suse.de>
7699 PR bootstrap/60719
7700 * Makefile.in (install-driver): Guard extra installs with special
7701 names properly.
7703 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
7705 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
7706 Document vec_vgbbd.
7708 2014-04-01 Richard Henderson <rth@redhat.com>
7710 PR target/60704
7711 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
7712 alternative enabled before register allocation.
7714 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
7716 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
7717 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
7718 typo.
7719 (nios2_large_got_address): Remove unneeded 'sym' parameter.
7720 (nios2_got_address): Update nios2_large_got_address call site.
7721 (nios2_delegitimize_address): New function.
7722 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
7723 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
7724 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
7726 2014-04-01 Martin Husemann <martin@duskware.de>
7728 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
7729 for -mabi=32.
7731 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
7733 PR rtl-optimization/60604
7734 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
7735 check from register_operand.
7736 (register_operand): Redefine in terms of general_operand.
7737 (nonmemory_operand): Use register_operand for the non-constant cases.
7739 2014-04-01 Richard Biener <rguenther@suse.de>
7741 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
7743 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
7745 * doc/invoke.texi (mapp-regs): Clarify.
7747 2014-03-31 Ulrich Drepper <drepper@gmail.com>
7749 * config/i386/avx512fintrin.h (__v32hi): Define type.
7750 (__v64qi): Likewise.
7751 (_mm512_set1_epi8): Define.
7752 (_mm512_set1_epi16): Define.
7753 (_mm512_set4_epi32): Define.
7754 (_mm512_set4_epi64): Define.
7755 (_mm512_set4_pd): Define.
7756 (_mm512_set4_ps): Define.
7757 (_mm512_setr4_epi64): Define.
7758 (_mm512_setr4_epi32): Define.
7759 (_mm512_setr4_pd): Define.
7760 (_mm512_setr4_ps): Define.
7761 (_mm512_setzero_epi32): Define.
7763 2014-03-31 Martin Jambor <mjambor@suse.cz>
7765 PR middle-end/60647
7766 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
7767 callsite_arguments_match_p. Updated all callers. Also check types of
7768 corresponding formal parameters and actual arguments.
7769 (not_all_callers_have_enough_arguments_p) Renamed to
7770 some_callers_have_mismatched_arguments_p.
7772 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
7774 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
7776 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
7778 PR target/60034
7779 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
7780 section anchor.
7782 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
7784 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
7785 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
7786 Split out
7787 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
7788 Use FMAMODE_NOVF512 mode iterator.
7789 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
7790 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
7791 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
7792 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
7793 Split out
7794 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
7795 Use VF_128_256 mode iterator.
7796 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
7797 Ditto.
7799 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
7801 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
7802 static chain if needed.
7804 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
7806 PR target/60697
7807 * lra-constraints.c (index_part_to_reg): New.
7808 (process_address): Use it.
7810 2014-03-27 Jeff Law <law@redhat.com>
7811 Jakub Jelinek <jakub@redhat.com>
7813 PR target/60648
7814 * expr.c (do_tablejump): Use simplify_gen_binary rather than
7815 gen_rtx_{PLUS,MULT} to build up the address expression.
7817 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
7818 creating non-canonical RTL.
7820 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
7822 PR ipa/60243
7823 * ipa-inline.c (want_inline_small_function_p): Short circuit large
7824 functions; reorganize to make cheap checks first.
7825 (inline_small_functions): Do not estimate growth when dumping;
7826 it is expensive.
7827 * ipa-inline.h (inline_summary): Add min_size.
7828 (growth_likely_positive): New function.
7829 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
7830 (set_cond_stmt_execution_predicate): Cleanup.
7831 (estimate_edge_size_and_time): Compute min_size.
7832 (estimate_calls_size_and_time): Likewise.
7833 (estimate_node_size_and_time): Likewise.
7834 (inline_update_overall_summary): Update min_size.
7835 (do_estimate_edge_time): Likewise.
7836 (do_estimate_edge_size): Update.
7837 (do_estimate_edge_hints): Update.
7838 (growth_likely_positive): New function.
7840 2014-03-28 Jakub Jelinek <jakub@redhat.com>
7842 PR target/60693
7843 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
7844 also if addr has VOIDmode.
7846 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7848 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
7849 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
7850 Declare extern.
7851 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
7852 instructions as well as AdvancedSIMD loads.
7854 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7856 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
7857 Use crypto_aese type.
7858 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
7859 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
7860 crypto_aese, crypto_aesmc. Move to types.md.
7861 * config/arm/types.md (crypto_aes): Split into crypto_aese,
7862 crypto_aesmc.
7863 * config/arm/iterators.md (crypto_type): Likewise.
7865 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
7867 * cgraph.c: Include expr.h and tree-dfa.h.
7868 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
7869 remove LHS.
7871 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
7873 PR target/60675
7874 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
7875 regs from checking multi-reg pseudos.
7877 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7879 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
7881 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7883 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
7884 if it would clobber the stack pointer, even temporarily.
7886 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
7888 * mode-switching.c: Make small adjustments to the top comment.
7890 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
7892 * config/rs6000/constraints.md (wD constraint): New constraint to
7893 match the constant integer to get the top DImode/DFmode out of a
7894 vector in a VSX register.
7896 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
7897 match the constant integer to get the top DImode/DFmode out of a
7898 vector in a VSX register.
7900 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
7901 for ISA 2.07.
7903 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
7904 vbpermq builtins.
7906 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
7907 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
7909 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
7910 Optimize vec_extract of 64-bit values, where the value being
7911 extracted is in the top word, where we can use scalar
7912 instructions. Add direct move and store support. Combine the big
7913 endian/little endian vector select load support into a single insn.
7914 (vsx_extract_<mode>_internal1): Likewise.
7915 (vsx_extract_<mode>_internal2): Likewise.
7916 (vsx_extract_<mode>_load): Likewise.
7917 (vsx_extract_<mode>_store): Likewise.
7918 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
7919 combined into vsx_extract_<mode>_load.
7920 (vsx_extract_<mode>_one_le): Likewise.
7922 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
7923 define the top 64-bit vector element.
7925 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
7926 constraint.
7928 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
7929 Document vec_vbpermq builtin.
7931 PR target/60672
7932 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
7933 enable use of xxsldwi and xxpermdi builtin functions.
7934 (vec_xxpermdi): Likewise.
7936 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
7937 Document use of vec_xxsldwi and vec_xxpermdi builtins.
7939 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
7941 PR rtl-optimization/60650
7942 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
7943 first_p. Use it.
7944 (find_spills_for): New.
7945 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
7946 Spill all pseudos on the second iteration.
7948 2014-03-27 Marek Polacek <polacek@redhat.com>
7950 PR c/50347
7951 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
7952 types.
7954 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7956 * config/s390/s390.c (s390_can_use_return_insn): Check for
7957 call-saved FPRs on 31 bit.
7959 2014-03-27 Jakub Jelinek <jakub@redhat.com>
7961 PR middle-end/60682
7962 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
7963 if they need regimplification, just drop them instead of
7964 calling gimple_regimplify_operands on them.
7966 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
7968 PR target/60580
7969 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
7970 (aarch64_frame_pointer_required): Adjust logic.
7971 (aarch64_can_eliminate): Adjust logic.
7972 (aarch64_override_options_after_change): Adjust logic.
7974 2014-03-27 Dehao Chen <dehao@google.com>
7976 * ipa-inline.c (early_inliner): Update node's inline info.
7978 2014-03-26 Dehao Chen <dehao@google.com>
7980 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
7981 compiler inserted conditional jumps for NAN float check.
7983 2014-03-26 Jakub Jelinek <jakub@redhat.com>
7985 * ubsan.h (ubsan_create_data): Change second argument's type
7986 to const location_t *.
7987 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
7988 _("<unknown>").
7989 (ubsan_create_data): Change second argument to const location_t *PLOC.
7990 Create Loc field whenever PLOC is non-NULL.
7991 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
7992 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
7993 callers.
7995 PR other/59545
7996 * real.c (real_to_integer2): Change type of low to UHWI.
7998 2014-03-26 Tobias Burnus <burnus@net-b.de>
8000 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
8001 (CILK_SELF_SPECS): New define.
8002 (driver_self_specs): Use it.
8004 2014-03-26 Richard Biener <rguenther@suse.de>
8006 * tree-pretty-print.c (percent_K_format): Implement special
8007 case for LTO and its stripped down BLOCK tree.
8009 2014-03-26 Jakub Jelinek <jakub@redhat.com>
8011 PR sanitizer/60636
8012 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
8014 * tree-vrp.c (simplify_internal_call_using_ranges): If only
8015 one range is range_int_cst_p, but not both, at least optimize
8016 addition/subtraction of 0 and multiplication by 0 or 1.
8017 * gimple-fold.c (gimple_fold_call): Fold
8018 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
8019 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
8020 INTEGER_CSTs, try to fold at least x * 0 and y - y.
8022 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
8024 PR rtl-optimization/60452
8025 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
8026 <case REG>: Return 1 for invalid offsets from the frame pointer.
8028 2014-03-26 Marek Polacek <polacek@redhat.com>
8030 PR c/37428
8031 * doc/extend.texi (C Extensions): Mention variable-length arrays in
8032 a structure/union.
8034 2014-03-26 Marek Polacek <polacek@redhat.com>
8036 PR c/39525
8037 * doc/extend.texi (Designated Inits): Describe what happens to omitted
8038 field members.
8040 2014-03-26 Marek Polacek <polacek@redhat.com>
8042 PR other/59545
8043 * ira-color.c (update_conflict_hard_regno_costs): Perform the
8044 multiplication in unsigned type.
8046 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
8048 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
8050 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
8052 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
8054 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
8056 PR ipa/60315
8057 * cif-code.def (UNREACHABLE) New code.
8058 * ipa-inline.c (inline_small_functions): Skip edges to
8059 __builtlin_unreachable.
8060 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
8061 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
8062 predicate to __bulitin_unreachable.
8063 (set_cond_stmt_execution_predicate): Fix issue when
8064 invert_tree_comparison returns ERROR_MARK.
8065 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
8066 propagate to inline clones.
8067 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
8068 to unreachable.
8069 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
8070 * cgraphclones.c (cgraph_clone_node): If call destination is already
8071 ureachable, do not redirect it back.
8072 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
8073 unreachable.
8075 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
8077 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
8078 Do not modify inline clones.
8080 2014-03-25 Jakub Jelinek <jakub@redhat.com>
8082 * config/i386/i386.md (general_sext_operand): New mode attr.
8083 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
8084 don't generate (sign_extend (const_int)).
8085 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
8086 operands[2]. Use We constraint instead of <i> and
8087 <general_sext_operand> predicate instead of <general_operand>.
8088 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
8089 * config/i386/constraints.md (We): New constraint.
8090 * config/i386/predicates.md (x86_64_sext_operand,
8091 sext_operand): New predicates.
8093 2014-03-25 Martin Jambor <mjambor@suse.cz>
8095 PR ipa/60600
8096 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
8097 inconsistent devirtualizations to __builtin_unreachable.
8099 2014-03-25 Marek Polacek <polacek@redhat.com>
8101 PR c/35449
8102 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
8104 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
8106 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
8107 order of elements for big-endian.
8109 2014-03-25 Richard Biener <rguenther@suse.de>
8111 PR middle-end/60635
8112 * gimplify-me.c (gimple_regimplify_operands): Update the
8113 re-gimplifed stmt.
8115 2014-03-25 Martin Jambor <mjambor@suse.cz>
8117 PR ipa/59176
8118 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
8119 (lto_output_varpool_node): Likewise.
8120 (input_overwrite_node): Likewise.
8121 (input_varpool_node): Likewise.
8123 2014-03-25 Richard Biener <rguenther@suse.de>
8125 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
8126 (run_gcc): Likewise.
8128 2014-03-25 Jakub Jelinek <jakub@redhat.com>
8130 * combine.c (simplify_compare_const): Add MODE argument.
8131 Handle mode_width 0 as very large mode_width.
8132 (try_combine, simplify_comparison): Adjust callers.
8134 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
8135 type to avoid signed integer overflow.
8136 * explow.c (plus_constant): Likewise.
8138 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
8140 * doc/generic.texi: Correct typos.
8142 2014-03-24 Tobias Burnus <burnus@net-b.de>
8144 * doc/invoke.texi (-flto): Expand section about
8145 using static libraries with LTO.
8147 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8149 PR rtl-optimization/60501
8150 * optabs.def (addptr3_optab): New optab.
8151 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
8152 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
8153 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
8155 * lra.c (emit_add3_insn): Use the addptr pattern if available.
8157 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
8159 2014-03-24 Ulrich Drepper <drepper@gmail.com>
8161 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
8162 _mm512_set1_pd.
8164 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
8165 (_mm256_undefined_ps): Define.
8166 (_mm256_undefined_pd): Define.
8167 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
8168 (_mm_undefined_pd): Define.
8169 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
8170 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
8171 (_mm512_undefined_ps): Define.
8172 (_mm512_undefined_pd): Define.
8173 Use _mm*_undefined_*.
8174 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
8176 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
8178 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
8179 (lshr_simd): DI mode added.
8180 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
8181 (aarch64_ushr_simddi): Likewise.
8182 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
8183 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
8184 (vshrd_n_u64): Likewise.
8186 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8188 * Makefile.in (s-macro_list): Depend on cc1.
8190 2014-03-23 Teresa Johnson <tejohnson@google.com>
8192 * ipa-utils.c (ipa_print_order): Use specified dump file.
8194 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
8196 PR rtl-optimization/60601
8197 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
8199 * gcc.c (eval_spec_function): Initialize save_growing_value.
8201 2014-03-22 Jakub Jelinek <jakub@redhat.com>
8203 PR sanitizer/60613
8204 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
8205 code == MINUS_EXPR, never swap op0 with op1.
8207 * toplev.c (init_local_tick): Avoid signed integer multiplication
8208 overflow.
8209 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
8210 shift by first operand's bitsize.
8212 2014-03-21 Jakub Jelinek <jakub@redhat.com>
8214 PR target/60610
8215 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
8216 redefine to 1 or 0.
8217 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
8218 TARGET_ISA_64BIT_P(x).
8220 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8222 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
8223 pattern for vector nor instead of subtract from splat(-1).
8224 (altivec_expand_vec_perm_const_le): Likewise.
8226 2014-03-21 Richard Henderson <rth@twiddle.net>
8228 PR target/60598
8229 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
8230 related insns after epilogue_completed.
8232 2014-03-21 Martin Jambor <mjambor@suse.cz>
8234 PR ipa/59176
8235 * cgraph.h (symtab_node): New flag body_removed.
8236 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
8237 when removing bodies.
8238 * symtab.c (dump_symtab_base): Dump body_removed flag.
8239 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
8240 had their bodies removed.
8242 2014-03-21 Martin Jambor <mjambor@suse.cz>
8244 PR ipa/60419
8245 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
8246 in the border.
8248 2014-03-21 Richard Biener <rguenther@suse.de>
8250 PR tree-optimization/60577
8251 * tree-core.h (struct tree_base): Document nothrow_flag use
8252 in VAR_DECL_NONALIASED.
8253 * tree.h (VAR_DECL_NONALIASED): New.
8254 (may_be_aliased): Adjust.
8255 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
8257 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
8259 * expr.c (expand_expr_real_1): Remove outdated comment.
8261 2014-03-20 Jakub Jelinek <jakub@redhat.com>
8263 PR middle-end/60597
8264 * ira.c (adjust_cleared_regs): Call copy_rtx on
8265 *reg_equiv[REGNO (loc)].src_p before passing it to
8266 simplify_replace_fn_rtx.
8268 PR target/60568
8269 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
8270 into CONST, put pic register as first operand of PLUS. Use
8271 gen_const_mem for both 32-bit and 64-bit PIC got loads.
8273 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8275 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
8277 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
8279 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
8280 around for store forwarding issue in the FPU on the UT699.
8281 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
8282 loads and operations if -mfix-ut699 is specified.
8283 (divtf3_hq): Tweak attribute.
8284 (sqrttf2_hq): Likewise.
8286 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
8288 * calls.c (store_one_arg): Remove incorrect const qualification on the
8289 type of the temporary.
8290 * cfgexpand.c (expand_return): Likewise.
8291 * expr.c (expand_constructor): Likewise.
8292 (expand_expr_real_1): Likewise.
8294 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
8296 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
8297 of parts.
8299 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
8301 PR target/60039
8302 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
8304 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
8306 * config/arm/aarch-common-protos.h
8307 (alu_cost_table): Fix spelling of "extend".
8308 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
8310 2014-03-19 Richard Biener <rguenther@suse.de>
8312 PR middle-end/60553
8313 * tree-core.h (tree_type_common): Re-order pointer members
8314 to reduce recursion depth during GC walks.
8316 2014-03-19 Marek Polacek <polacek@redhat.com>
8318 PR sanitizer/60569
8319 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
8320 before accessing it.
8322 2014-03-19 Richard Biener <rguenther@suse.de>
8324 PR lto/59543
8325 * lto-streamer-in.c (input_function): In WPA stage do not drop
8326 debug stmts.
8328 2014-03-19 Jakub Jelinek <jakub@redhat.com>
8330 PR tree-optimization/60559
8331 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
8332 with build_zero_cst assignment.
8334 2014-03-18 Kai Tietz <ktietz@redhat.com>
8336 PR rtl-optimization/56356
8337 * sdbout.c (sdbout_parms): Verify that parms'
8338 incoming argument is valid.
8339 (sdbout_reg_parms): Likewise.
8341 2014-03-18 Richard Henderson <rth@redhat.com>
8343 PR target/60562
8344 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
8345 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
8346 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
8348 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
8350 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
8351 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
8352 Italicize plugin event names in description. Explain that
8353 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
8354 Remind that no GCC functions should be called after PLUGIN_FINISH.
8355 Explain what pragmas with expansion are.
8357 2014-03-18 Martin Liska <mliska@suse.cz>
8359 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
8360 gimple call statement is update.
8361 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
8362 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
8364 2014-03-18 Jakub Jelinek <jakub@redhat.com>
8366 PR sanitizer/60557
8367 * ubsan.c (ubsan_instrument_unreachable): Call
8368 initialize_sanitizer_builtins.
8369 (ubsan_pass): Likewise.
8371 PR sanitizer/60535
8372 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
8373 varpool_finalize_decl instead of rest_of_decl_compilation.
8375 2014-03-18 Richard Biener <rguenther@suse.de>
8377 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
8378 by using bitmap_and_compl instead of bitmap_and_compl_into.
8379 (df_rd_transfer_function): Likewise.
8381 2014-03-18 Richard Biener <rguenther@suse.de>
8383 * doc/lto.texi (fresolution): Fix typo.
8385 2014-03-18 Richard Biener <rguenther@suse.de>
8387 * doc/invoke.texi (flto): Update for changes in 4.9.
8389 2014-03-18 Richard Biener <rguenther@suse.de>
8391 * doc/loop.texi: Remove section on the removed lambda framework.
8392 Update loop docs with recent changes in preserving loop structure.
8394 2014-03-18 Richard Biener <rguenther@suse.de>
8396 * doc/lto.texi (-fresolution): Document.
8398 2014-03-18 Richard Biener <rguenther@suse.de>
8400 * doc/contrib.texi: Adjust my name.
8402 2014-03-18 Jakub Jelinek <jakub@redhat.com>
8404 PR ipa/58721
8405 * internal-fn.c: Include diagnostic-core.h.
8406 (expand_BUILTIN_EXPECT): New function.
8407 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
8408 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
8409 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
8410 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
8411 IFN_BUILTIN_EXPECT.
8412 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
8413 Revert 3 argument __builtin_expect code.
8414 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
8415 * gimple-fold.c (gimple_fold_call): Likewise.
8416 * tree.h (fold_builtin_expect): New prototype.
8417 * builtins.c (build_builtin_expect_predicate): Add predictor
8418 argument, if non-NULL, create 3 argument __builtin_expect.
8419 (fold_builtin_expect): No longer static. Add ARG2 argument,
8420 pass it through to build_builtin_expect_predicate.
8421 (fold_builtin_2): Adjust caller.
8422 (fold_builtin_3): Handle BUILT_IN_EXPECT.
8423 * internal-fn.def (BUILTIN_EXPECT): New.
8425 2014-03-18 Tobias Burnus <burnus@net-b.de>
8427 PR ipa/58721
8428 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
8429 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
8430 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
8432 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
8434 PR ipa/58721
8435 * predict.c (combine_predictions_for_bb): Fix up formatting.
8436 (expr_expected_value_1, expr_expected_value): Add predictor argument,
8437 fill what it points to if non-NULL.
8438 (tree_predict_by_opcode): Adjust caller, use the predictor.
8439 * predict.def (PRED_COMPARE_AND_SWAP): Add.
8441 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
8443 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
8444 proper constant for the store mode.
8446 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
8448 * symtab.c (change_decl_assembler_name): Fix transparent alias
8449 chain construction.
8451 2014-03-16 Renlin Li <Renlin.Li@arm.com>
8453 * config/aarch64/aarch64.c: Correct the comments about the
8454 aarch64 stack layout.
8456 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
8458 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
8459 check for GF_OMP_FOR_KIND_FOR.
8461 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
8463 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
8464 ymm and zmm register names.
8466 2014-03-17 Jakub Jelinek <jakub@redhat.com>
8468 PR target/60516
8469 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
8470 note creation for the 2010-08-31 changes.
8472 2014-03-17 Marek Polacek <polacek@redhat.com>
8474 PR middle-end/60534
8475 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
8476 as -fno-tree-loop-vectorize.
8477 (expand_omp_simd): Likewise.
8479 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
8481 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
8482 (eligible_for_call_delay): New prototype.
8483 * config/sparc/sparc.c (tls_call_delay): Rename into...
8484 (eligible_for_call_delay): ...this. Return false if the instruction
8485 cannot be put in the delay slot of a branch.
8486 (eligible_for_restore_insn): Simplify.
8487 (eligible_for_return_delay): Return false if the instruction cannot be
8488 put in the delay slot of a branch and simplify.
8489 (eligible_for_sibcall_delay): Return false if the instruction cannot be
8490 put in the delay slot of a branch.
8491 * config/sparc/sparc.md (fix_ut699): New attribute.
8492 (tls_call_delay): Delete.
8493 (in_call_delay): Reimplement.
8494 (eligible_for_sibcall_delay): Rename into...
8495 (in_sibcall_delay): ...this.
8496 (eligible_for_return_delay): Rename into...
8497 (in_return_delay): ...this.
8498 (in_branch_delay): Reimplement.
8499 (in_uncond_branch_delay): Delete.
8500 (in_annul_branch_delay): Delete.
8502 2014-03-14 Richard Henderson <rth@redhat.com>
8504 PR target/60525
8505 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
8506 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
8507 (*floathi<X87MODEF>2_i387_with_temp): Remove.
8508 (floathi splitters): Remove.
8509 (float<SWI48x>xf2): New pattern.
8510 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
8511 code that tried to handle DImode for 32-bit, but which was excluded
8512 by the pattern's condition. Drop allocation of stack temporary.
8513 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
8514 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
8515 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
8516 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
8517 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
8518 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
8519 (*float<SWI48><MODEF>2_sse_interunit): Remove.
8520 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
8521 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
8522 (*float<SWI48x><X87MODEF>2_i387): Remove.
8523 (all float _with_temp splitters): Remove.
8524 (*float<SWI48x><MODEF>2_i387): New pattern.
8525 (*float<SWI48><MODEF>2_sse): New pattern.
8526 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
8527 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
8529 2014-03-14 Jakub Jelinek <jakub@redhat.com>
8530 Marek Polacek <polacek@redhat.com>
8532 PR middle-end/60484
8533 * common.opt (dump_base_name_prefixed): New Variable.
8534 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
8535 if x_dump_base_name_prefixed is already set, set it at the end.
8537 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
8539 PR rtl-optimization/60508
8540 * lra-constraints.c (get_reload_reg): Add new parameter
8541 in_subreg_p.
8542 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
8543 Pass the new parameter values.
8545 2014-03-14 Richard Biener <rguenther@suse.de>
8547 * common.opt: Revert unintented changes from r205065.
8548 * opts.c: Likewise.
8550 2014-03-14 Richard Biener <rguenther@suse.de>
8552 PR middle-end/60518
8553 * cfghooks.c (split_block): Properly adjust all loops the
8554 block was a latch of.
8556 2014-03-14 Martin Jambor <mjambor@suse.cz>
8558 PR lto/60461
8559 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
8560 and simplify it.
8562 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
8564 PR target/59396
8565 * config/avr/avr.c (avr_set_current_function): Pass function name
8566 through default_strip_name_encoding before sanity checking instead
8567 of skipping the first char of the assembler name.
8569 2014-03-13 Richard Henderson <rth@redhat.com>
8571 PR debug/60438
8572 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
8573 (ix86_force_to_memory, ix86_free_from_memory): Remove.
8574 * config/i386/i386-protos.h: Likewise.
8575 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
8576 in the expander instead of a splitter.
8577 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
8578 any possibility of requiring a memory.
8579 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
8580 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
8581 (fp branch splitters): Update for ix86_split_fp_branch.
8582 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
8583 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
8584 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
8585 (*fop_<MODEF>_2_i387): Remove f/r alternative.
8586 (*fop_<MODEF>_3_i387): Likewise.
8587 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
8588 (splitters for the fop_* register patterns): Remove.
8589 (fscalexf4_i387): Rename from *fscalexf4_i387.
8590 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
8592 2014-03-13 Jakub Jelinek <jakub@redhat.com>
8594 PR tree-optimization/59779
8595 * tree-dfa.c (get_ref_base_and_extent): Use double_int
8596 type for bitsize and maxsize instead of HOST_WIDE_INT.
8598 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
8600 PR rtl-optimization/57320
8601 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
8602 the CFG after thread_prologue_and_epilogue_insns.
8604 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
8606 PR rtl-optimization/57189
8607 * lra-constraints.c (process_alt_operands): Disfavor spilling
8608 vector pseudos.
8610 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
8612 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
8614 2014-03-13 Jakub Jelinek <jakub@redhat.com>
8616 PR tree-optimization/59025
8617 PR middle-end/60418
8618 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
8619 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
8621 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
8623 PR target/60486
8624 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
8625 calls of avr_out_plus_1.
8627 2014-03-13 Bin Cheng <bin.cheng@arm.com>
8629 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
8630 BB's single pred and update the father loop's latch info later.
8632 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
8634 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
8635 (VEC_M): Likewise.
8636 (VEC_N): Likewise.
8637 (VEC_R): Likewise.
8638 (VEC_base): Likewise.
8639 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
8640 registers, we need to swap double words in little endian mode.
8642 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
8643 to be a container mode for 128-bit integer operations added in ISA
8644 2.07. Unlike TImode and PTImode, the preferred register set is
8645 the Altivec/VMX registers for the 128-bit operations.
8647 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
8648 declarations.
8649 (rs6000_split_128bit_ok_p): Likewise.
8651 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
8652 macros for creating ISA 2.07 normal and overloaded builtin
8653 functions with 3 arguments.
8654 (BU_P8V_OVERLOAD_3): Likewise.
8655 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
8656 for use as overloaded functions.
8657 (VPERM_1TI_UNS): Likewise.
8658 (VSEL_1TI): Likewise.
8659 (VSEL_1TI_UNS): Likewise.
8660 (ST_INTERNAL_1ti): Likewise.
8661 (LD_INTERNAL_1ti): Likewise.
8662 (XXSEL_1TI): Likewise.
8663 (XXSEL_1TI_UNS): Likewise.
8664 (VPERM_1TI): Likewise.
8665 (VPERM_1TI_UNS): Likewise.
8666 (XXPERMDI_1TI): Likewise.
8667 (SET_1TI): Likewise.
8668 (LXVD2X_V1TI): Likewise.
8669 (STXVD2X_V1TI): Likewise.
8670 (VEC_INIT_V1TI): Likewise.
8671 (VEC_SET_V1TI): Likewise.
8672 (VEC_EXT_V1TI): Likewise.
8673 (EQV_V1TI): Likewise.
8674 (NAND_V1TI): Likewise.
8675 (ORC_V1TI): Likewise.
8676 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
8677 added in ISA 2.07. Add both normal 'altivec' builtins, and the
8678 overloaded builtin.
8679 (VADDUQM): Likewise.
8680 (VSUBCUQ): Likewise.
8681 (VADDEUQM): Likewise.
8682 (VADDECUQ): Likewise.
8683 (VSUBEUQM): Likewise.
8684 (VSUBECUQ): Likewise.
8686 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
8687 __int128_t and __uint128_t types.
8688 (__uint128_type): Likewise.
8689 (altivec_categorize_keyword): Add support for vector __int128_t,
8690 vector __uint128_t, vector __int128, and vector unsigned __int128
8691 as a container type for TImode operations that need to be done in
8692 VSX/Altivec registers.
8693 (rs6000_macro_to_expand): Likewise.
8694 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
8695 to support 128-bit integer instructions vaddcuq, vadduqm,
8696 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
8697 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
8699 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
8700 for V1TImode, and set up preferences to use VSX/Altivec registers.
8701 Setup VSX reload handlers.
8702 (rs6000_debug_reg_global): Likewise.
8703 (rs6000_init_hard_regno_mode_ok): Likewise.
8704 (rs6000_preferred_simd_mode): Likewise.
8705 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
8706 (easy_altivec_constant): Likewise.
8707 (output_vec_const_move): Likewise.
8708 (rs6000_expand_vector_set): Convert V1TImode set and extract to
8709 simple move.
8710 (rs6000_expand_vector_extract): Likewise.
8711 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
8712 addressing.
8713 (rs6000_const_vec): Add support for V1TImode.
8714 (rs6000_emit_le_vsx_load): Swap double words when loading or
8715 storing TImode/V1TImode.
8716 (rs6000_emit_le_vsx_store): Likewise.
8717 (rs6000_emit_le_vsx_move): Likewise.
8718 (rs6000_emit_move): Add support for V1TImode.
8719 (altivec_expand_ld_builtin): Likewise.
8720 (altivec_expand_st_builtin): Likewise.
8721 (altivec_expand_vec_init_builtin): Likewise.
8722 (altivec_expand_builtin): Likewise.
8723 (rs6000_init_builtins): Add support for V1TImode type. Add
8724 support for ISA 2.07 128-bit integer builtins. Define type names
8725 for the VSX/Altivec vector types.
8726 (altivec_init_builtins): Add support for overloaded vector
8727 functions with V1TImode type.
8728 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
8729 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
8730 external function.
8731 (rs6000_split_128bit_ok_p): Likewise.
8732 (rs6000_handle_altivec_attribute): Create V1TImode from vector
8733 __int128_t and vector __uint128_t.
8735 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
8736 and mode attributes.
8737 (VSX_M): Likewise.
8738 (VSX_M2): Likewise.
8739 (VSm): Likewise.
8740 (VSs): Likewise.
8741 (VSr): Likewise.
8742 (VSv): Likewise.
8743 (VS_scalar): Likewise.
8744 (VS_double): Likewise.
8745 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
8747 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
8748 we support the ISA 2.07 128-bit integer arithmetic instructions.
8749 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
8750 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
8751 and TImode types for use with the builtin functions.
8752 (V1TI_type_node): Likewise.
8753 (unsigned_V1TI_type_node): Likewise.
8754 (intTI_type_internal_node): Likewise.
8755 (uintTI_type_internal_node): Likewise.
8757 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
8758 128-bit builtin functions.
8759 (UNSPEC_VADDEUQM): Likewise.
8760 (UNSPEC_VADDECUQ): Likewise.
8761 (UNSPEC_VSUBCUQ): Likewise.
8762 (UNSPEC_VSUBEUQM): Likewise.
8763 (UNSPEC_VSUBECUQ): Likewise.
8764 (VM): Add V1TImode to vector mode iterators.
8765 (VM2): Likewise.
8766 (VI_unit): Likewise.
8767 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
8768 (altivec_vaddcuq): Likewise.
8769 (altivec_vsubuqm): Likewise.
8770 (altivec_vsubcuq): Likewise.
8771 (altivec_vaddeuqm): Likewise.
8772 (altivec_vaddecuq): Likewise.
8773 (altivec_vsubeuqm): Likewise.
8774 (altivec_vsubecuq): Likewise.
8776 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
8777 mode iterators.
8778 (BOOL_128): Likewise.
8779 (BOOL_REGS_OUTPUT): Likewise.
8780 (BOOL_REGS_OP1): Likewise.
8781 (BOOL_REGS_OP2): Likewise.
8782 (BOOL_REGS_UNARY): Likewise.
8783 (BOOL_REGS_AND_CR0): Likewise.
8785 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
8786 128-bit integer builtin support.
8787 (vec_vadduqm): Likewise.
8788 (vec_vaddecuq): Likewise.
8789 (vec_vaddeuqm): Likewise.
8790 (vec_vsubecuq): Likewise.
8791 (vec_vsubeuqm): Likewise.
8792 (vec_vsubcuq): Likewise.
8793 (vec_vsubuqm): Likewise.
8795 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
8796 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
8797 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
8798 128-bit integer add/subtract to ISA 2.07.
8800 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
8802 * config/arc/arc.c (arc_predicate_delay_insns):
8803 Fix third argument passed to conditionalize_nonjump.
8805 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
8807 * config/aarch64/aarch64-builtins.c
8808 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
8809 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
8810 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
8811 instead of __builtin_lfloor.
8812 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
8814 2014-03-12 Jakub Jelinek <jakub@redhat.com>
8816 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
8817 (tree_ssa_ifcombine_bb_1): New function.
8818 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
8819 is an empty forwarder block to then_bb or vice versa and then_bb
8820 and else_bb are effectively swapped.
8822 2014-03-12 Christian Bruel <christian.bruel@st.com>
8824 PR target/60264
8825 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
8826 REG_CFA_DEF_CFA note.
8827 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
8828 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
8830 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
8832 PR tree-optimization/60454
8833 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
8835 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8837 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
8838 Do not define target_cpu_default2 to generic.
8839 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
8840 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
8841 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
8843 2014-03-12 Jakub Jelinek <jakub@redhat.com>
8844 Marc Glisse <marc.glisse@inria.fr>
8846 PR tree-optimization/60502
8847 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
8848 instead of build_low_bits_mask.
8850 2014-03-12 Jakub Jelinek <jakub@redhat.com>
8852 PR middle-end/60482
8853 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
8854 if there are multiple uses, but op doesn't live on E edge.
8855 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
8856 clobber stmts before __builtin_unreachable.
8858 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
8860 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
8861 hard_frame_pointer_rtx.
8862 * cse.c (cse_insn): Remove volatile check.
8863 * cselib.c (cselib_process_insn): Likewise.
8864 * dse.c (scan_insn): Likewise.
8866 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
8868 * config/arc/arc.c (conditionalize_nonjump): New function,
8869 broken out of ...
8870 (arc_ifcvt): ... this.
8871 (arc_predicate_delay_insns): Use it.
8873 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
8875 * config/arc/predicates.md (extend_operand): During/after reload,
8876 allow const_int_operand.
8877 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
8878 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
8879 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
8880 to "i".
8881 (umulsi3_highpart_i): Likewise.
8883 2014-03-11 Richard Biener <rguenther@suse.de>
8885 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
8886 Add asserts to guard possible wrong-code bugs.
8888 2014-03-11 Richard Biener <rguenther@suse.de>
8890 PR tree-optimization/60429
8891 PR tree-optimization/60485
8892 * tree-ssa-structalias.c (set_union_with_increment): Properly
8893 take into account all fields that overlap the shifted vars.
8894 (do_sd_constraint): Likewise.
8895 (do_ds_constraint): Likewise.
8896 (get_constraint_for_ptr_offset): Likewise.
8898 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
8900 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
8901 (nios2_compute_frame_layout):
8902 Add calculation of cfun->machine->fp_save_offset.
8903 (nios2_expand_prologue): Correct setting of frame pointer register
8904 in prologue.
8905 (nios2_expand_epilogue): Update recovery of stack pointer from
8906 frame pointer accordingly.
8907 (nios2_initial_elimination_offset): Update calculation of offset
8908 for eliminating to HARD_FRAME_POINTER_REGNUM.
8910 2014-03-10 Jakub Jelinek <jakub@redhat.com>
8912 PR ipa/60457
8913 * ipa.c (symtab_remove_unreachable_nodes): Don't call
8914 cgraph_get_create_node on VAR_DECLs.
8916 2014-03-10 Richard Biener <rguenther@suse.de>
8918 PR middle-end/60474
8919 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
8921 2014-03-08 Douglas B Rupp <rupp@gnat.com>
8923 * config/vms/vms.opt (vms_float_format): New variable.
8925 2014-03-08 Tobias Burnus <burnus@net-b.de>
8927 * doc/invoke.texi (-fcilkplus): Update implementation status.
8929 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
8930 Richard Biener <rguenther@suse.de>
8932 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
8933 consistently accross all TUs.
8934 (run_gcc): Enable -fshort-double automatically at link at link-time
8935 and disallow override.
8937 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
8939 PR target/58271
8940 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
8941 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
8942 if they can't be used.
8944 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8946 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
8947 for Solaris 11/x86 ld.
8948 * configure: Regenerate.
8950 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8952 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
8953 (LIB_TLS_SPEC): Save as ld_tls_libs.
8954 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
8955 (HAVE_AS_IX86_TLSLDM): New test.
8956 * configure, config.in: Regenerate.
8957 * config/i386/i386.c (legitimize_tls_address): Fall back to
8958 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
8959 cannot support TLS_MODEL_LOCAL_DYNAMIC.
8960 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
8961 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
8963 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
8965 * common.opt (fira-loop-pressure): Mark as optimization.
8967 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
8969 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
8970 an OpenMP mappable type.
8972 2014-03-06 Matthias Klose <doko@ubuntu.com>
8974 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
8975 MULTILIB_OSDIRNAMES is not defined.
8977 2014-03-06 Jakub Jelinek <jakub@redhat.com>
8978 Meador Inge <meadori@codesourcery.com>
8980 PR target/58595
8981 * config/arm/arm.c (arm_tls_symbol_p): Remove.
8982 (arm_legitimize_address): Call legitimize_tls_address for any
8983 arm_tls_referenced_p expression, handle constant addend. Call it
8984 before testing for !TARGET_ARM.
8985 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
8987 2014-03-06 Richard Biener <rguenther@suse.de>
8989 PR middle-end/60445
8990 PR lto/60424
8991 PR lto/60427
8992 Revert
8993 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
8995 * tree-streamer.c (record_common_node): Assert we don't record
8996 nodes with type double.
8997 (preload_common_node): Skip type double, complex double and double
8998 pointer since it is now frontend dependent due to fshort-double option.
9000 2014-03-06 Richard Biener <rguenther@suse.de>
9002 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
9003 or -fno-lto is specified and the linker has full plugin support.
9004 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
9005 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
9006 * lto-wrapper.c (merge_and_complain): Merge compile-time
9007 optimization levels.
9008 (run_gcc): And pass it through to the link options.
9010 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
9012 PR debug/60381
9013 Revert:
9014 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
9015 PR debug/59992
9016 * cselib.c (remove_useless_values): Skip to avoid quadratic
9017 behavior if the condition moved from...
9018 (cselib_process_insn): ... here holds.
9020 2014-03-05 Jakub Jelinek <jakub@redhat.com>
9022 PR plugins/59335
9023 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
9024 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
9026 PR plugins/59335
9027 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
9028 (TM_H): Add x86-tune.def.
9030 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9032 * config/aarch64/aarch64.c (generic_tunings):
9033 Use cortexa57_extra_costs.
9035 2014-03-05 Jakub Jelinek <jakub@redhat.com>
9037 PR lto/60404
9038 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
9039 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
9040 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
9041 cost for in_lto_p.
9043 2014-03-04 Heiher <r@hev.cc>
9045 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
9046 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
9048 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
9050 * config/i386/predicates.md (const2356_operand): Change to ...
9051 (const2367_operand): ... this.
9052 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
9053 const2367_operand.
9054 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
9055 (*avx512pf_scatterpf<mode>sf): Ditto.
9056 (avx512pf_scatterpf<mode>df): Ditto.
9057 (*avx512pf_scatterpf<mode>df_mask): Ditto.
9058 (*avx512pf_scatterpf<mode>df): Ditto.
9059 * config/i386/i386.c (ix86_expand_builtin): Update
9060 incorrect hint operand error message.
9062 2014-03-04 Richard Biener <rguenther@suse.de>
9064 * lto-section-in.c (lto_get_section_data): Fix const cast.
9066 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
9068 * tree-streamer.c (record_common_node): Assert we don't record
9069 nodes with type double.
9070 (preload_common_node): Skip type double, complex double and double
9071 pointer since it is now frontend dependent due to fshort-double option.
9073 2014-03-04 Richard Biener <rguenther@suse.de>
9075 PR lto/60405
9076 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
9077 (lto_input_toplevel_asms): Likewise.
9078 * lto-section-in.c (lto_get_section_data): Instead do it here
9079 for every section.
9081 2014-03-04 Richard Biener <rguenther@suse.de>
9083 PR tree-optimization/60382
9084 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
9085 dead PHIs a reduction.
9087 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
9089 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
9090 hint value.
9091 (_mm_prefetch): Move out of GCC target("sse") pragma.
9092 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
9093 GCC target("prfchw") pragma.
9094 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
9095 for locality <= 2.
9096 * config/i386/i386.c (ix86_option_override_internal): Enable
9097 -mprfchw with -mprefetchwt1.
9099 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
9101 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
9102 Mark as varying.
9104 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
9106 * opts.h (CL_PCH_IGNORE): Define.
9107 * targhooks.c (option_affects_pch_p):
9108 Return false for options that have CL_PCH_IGNORE set.
9109 * opt-functions.awk: Process PchIgnore.
9110 * doc/options.texi: Document PchIgnore.
9112 * config/arc/arc.opt (misize): Add PchIgnore property.
9114 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9116 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
9117 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
9118 constraint on constants to permit them being loaded into
9119 GENERAL_REGS or BASE_REGS.
9121 2014-03-03 Nick Clifton <nickc@redhat.com>
9123 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
9124 anti-cacnonical alternatives.
9125 (negandhi3_real): New pattern.
9126 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
9128 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
9130 * config/avr/avr-mcus.def: Remove atxmega16x1.
9131 * config/avr/avr-tables.opt: Regenerate.
9132 * config/avr/t-multilib: Regenerate.
9133 * doc/avr-mmcu.texi: Regenerate.
9135 2014-03-03 Tobias Grosser <tobias@grosser.es>
9136 Mircea Namolaru <mircea.namolaru@inria.fr>
9138 PR tree-optimization/58028
9139 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
9140 scalar dimensions.
9142 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
9144 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
9145 not handled by recognizers.
9147 2014-03-03 Jakub Jelinek <jakub@redhat.com>
9149 PR middle-end/60175
9150 * function.c (expand_function_end): Don't emit
9151 clobber_return_register sequence if clobber_after is a BARRIER.
9152 * cfgexpand.c (construct_exit_block): Append instructions before
9153 return_label to prev_bb.
9155 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9157 * config/rs6000/constraints.md: Document reserved use of "wc".
9159 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
9161 PR ipa/60150
9162 * ipa.c (function_and_variable_visibility): When dissolving comdat
9163 group, also set all symbols to local.
9165 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
9167 PR ipa/60306
9169 Revert:
9170 2013-12-14 Jan Hubicka <jh@suse.cz>
9171 PR middle-end/58477
9172 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
9174 2014-03-02 Jon Beniston <jon@beniston.com>
9176 PR bootstrap/48230
9177 PR bootstrap/50927
9178 PR bootstrap/52466
9179 PR target/46898
9180 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
9181 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
9182 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
9183 (simple_return, *simple_return): New patterns
9184 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
9185 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
9187 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
9189 * dwarf2out.c (gen_subprogram_die): Tidy.
9191 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
9193 PR target/60071
9194 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
9195 (*mov_t_msb_neg_negc): ... this new insn.
9197 2014-02-28 Jason Merrill <jason@redhat.com>
9199 PR c++/58678
9200 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
9201 function.
9203 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
9205 PR c++/60314
9206 * dwarf2out.c (decltype_auto_die): New static.
9207 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
9208 (gen_type_die_with_usage): Handle 'decltype(auto)'.
9209 (is_cxx_auto): Likewise.
9211 2014-02-28 Ian Bolton <ian.bolton@arm.com>
9213 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
9214 we are not using general regs only.
9216 2014-02-28 Richard Biener <rguenther@suse.de>
9218 PR target/60280
9219 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
9220 previous fix and only allow to remove trivial pre-headers
9221 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
9222 (remove_forwarder_block): Properly update the latch of a loop.
9224 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
9226 PR debug/59992
9227 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
9228 (cselib_preserved_hash_table): New.
9229 (preserve_constants_and_equivs): Move preserved vals to it.
9230 (cselib_find_slot): Look it up first.
9231 (cselib_init): Initialize it.
9232 (cselib_finish): Release it.
9233 (dump_cselib_table): Dump it.
9235 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
9237 PR debug/59992
9238 * cselib.c (remove_useless_values): Skip to avoid quadratic
9239 behavior if the condition moved from...
9240 (cselib_process_insn): ... here holds.
9242 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
9244 PR debug/57232
9245 * var-tracking.c (vt_initialize): Apply the same condition to
9246 preserve the CFA base value.
9248 2014-02-28 Joey Ye <joey.ye@arm.com>
9250 PR target/PR60169
9251 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
9252 if reload in progress or completed.
9254 2014-02-28 Tobias Burnus <burnus@net-b.de>
9256 PR middle-end/60147
9257 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
9258 NAMELIST_DECL.
9260 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
9262 * doc/tm.texi.in (Condition Code Status): Update documention for
9263 relative locations of cc0-setter and cc0-user.
9265 2014-02-27 Jeff Law <law@redhat.com>
9267 PR rtl-optimization/52714
9268 * combine.c (try_combine): When splitting an unrecognized PARALLEL
9269 into two independent simple sets, if I3 is a jump, ensure the
9270 pattern we place into I3 is a (set (pc) ...).
9272 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
9273 Jeff Law <law@redhat.com>
9275 PR rtl-optimization/49847
9276 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
9277 are in different blocks.
9278 * doc/tm.texi (Condition Code Status): Update documention for
9279 relative locations of cc0-setter and cc0-user.
9281 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
9283 PR target/59222
9284 * lra.c (lra_emit_add): Check SUBREG too.
9286 2014-02-27 Andreas Schwab <schwab@suse.de>
9288 * config/m68k/m68k.c (m68k_option_override): Disable
9289 -flive-range-shrinkage for classic m68k.
9290 (m68k_override_options_after_change): Likewise.
9292 2014-02-27 Marek Polacek <polacek@redhat.com>
9294 PR middle-end/59223
9295 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
9296 -Wmaybe-uninitialized.
9298 2014-02-27 Alan Modra <amodra@gmail.com>
9300 PR target/57936
9301 * reload1.c (emit_input_reload_insns): When reload_override_in,
9302 set old to rl->in_reg when rl->in_reg is a subreg.
9304 2014-02-26 Richard Biener <rguenther@suse.de>
9306 PR bootstrap/60343
9307 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
9309 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
9311 * common/config/i386/predicates.md (const1256_operand): Remove.
9312 (const2356_operand): New.
9313 (const_1_to_2_operand): Remove.
9314 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
9315 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
9316 (*avx512pf_gatherpf<mode>sf): Ditto.
9317 (avx512pf_gatherpf<mode>df): Ditto.
9318 (*avx512pf_gatherpf<mode>df_mask): Ditto.
9319 (*avx512pf_gatherpf<mode>df): Ditto.
9320 (avx512pf_scatterpf<mode>sf): Ditto.
9321 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
9322 (*avx512pf_scatterpf<mode>sf): Ditto.
9323 (avx512pf_scatterpf<mode>df): Ditto.
9324 (*avx512pf_scatterpf<mode>df_mask): Ditto.
9325 (*avx512pf_scatterpf<mode>df): Ditto.
9326 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
9328 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
9330 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
9331 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
9332 (_mm512_mask_testn_epi64_mask): Move to ...
9333 * config/i386/avx512cdintrin.h: Here.
9334 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
9335 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
9336 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
9337 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
9338 TARGET_AVX512F from TARGET_AVX512CD.
9340 2014-02-26 Richard Biener <rguenther@suse.de>
9342 PR ipa/60327
9343 * ipa.c (walk_polymorphic_call_targets): Properly guard
9344 call to inline_update_overall_summary.
9346 2014-02-26 Bin Cheng <bin.cheng@arm.com>
9348 PR target/60280
9349 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
9350 and latches only if requested. Fix latch if it is removed.
9351 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
9352 LOOPS_HAVE_PREHEADERS.
9354 2014-02-25 Andrew Pinski <apinski@cavium.com>
9356 * builtins.c (expand_builtin_thread_pointer): Create a new target
9357 when the target is NULL.
9359 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
9361 PR rtl-optimization/60317
9362 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
9363 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
9364 * lra-assigns.c: Include params.h.
9365 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
9366 other reload pseudos considerations.
9368 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9370 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
9371 to use canonical form for nor<mode>3.
9373 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9375 PR target/55426
9376 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
9377 conversions.
9379 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
9381 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
9382 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
9383 (ix86_handle_option): Handle OPT_mprefetchwt1.
9384 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
9385 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
9386 PREFETCHWT1 CPUID.
9387 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
9388 OPTION_MASK_ISA_PREFETCHWT1.
9389 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
9390 (PTA_PREFETCHWT1): New.
9391 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
9392 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
9393 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
9394 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
9395 (*prefetch_avx512pf_<mode>_: Change into ...
9396 (*prefetch_prefetchwt1_<mode>: This.
9397 * config/i386/i386.opt (mprefetchwt1): New.
9398 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
9399 (_mm_prefetch): Handle intent to write.
9400 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
9402 2014-02-25 Richard Biener <rguenther@suse.de>
9404 PR middle-end/60291
9405 * emit-rtl.c (mem_attrs_htab): Remove.
9406 (mem_attrs_htab_hash): Likewise.
9407 (mem_attrs_htab_eq): Likewise.
9408 (set_mem_attrs): Always allocate new mem-attrs when something changed.
9409 (init_emit_once): Do not allocate mem_attrs_htab.
9411 2014-02-25 Richard Biener <rguenther@suse.de>
9413 PR lto/60319
9414 * lto-opts.c (lto_write_options): Output non-explicit conservative
9415 -fwrapv, -fno-trapv and -fno-strict-overflow.
9416 * lto-wrapper.c (merge_and_complain): Handle merging those options.
9417 (run_gcc): And pass them through.
9419 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
9421 * sel-sched.c (calculate_new_fences): New parameter ptime.
9422 Calculate it as a maximum over all fence cycles.
9423 (sel_sched_region_2): Adjust the call to calculate_new_fences.
9424 Print the final schedule timing when sched_verbose.
9426 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
9428 PR rtl-optimization/60292
9429 * sel-sched.c (fill_vec_av_set): Do not reset target availability
9430 bit fot the fence instruction.
9432 2014-02-24 Alangi Derick <alangiderick@gmail.com>
9434 * calls.h: Fix typo in comment.
9436 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
9438 * config/pa/pa.c (pa_output_move_double): Don't valididate when
9439 adjusting offsetable addresses.
9441 2014-02-24 Guozhi Wei <carrot@google.com>
9443 * sparseset.h (sparseset_pop): Fix the wrong index.
9445 2014-02-24 Walter Lee <walt@tilera.com>
9447 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
9448 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
9449 triplet.
9450 * common/config/tilegx/tilegx-common.c
9451 (TARGET_DEFAULT_TARGET_FLAGS): Define.
9452 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
9453 (LINK_SPEC): Ditto.
9454 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
9455 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
9456 (tilegx_gimplify_va_arg_expr): Handle big endian.
9457 (tilegx_expand_unaligned_load): Ditto.
9458 (tilegx_expand_unaligned_store): Ditto.
9459 (TARGET_RETURN_IN_MSB): New.
9460 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
9461 (TARGET_ENDIAN_DEFAULT): New.
9462 (TARGET_BIG_ENDIAN): Handle big endian.
9463 (BYTES_BIG_ENDIAN): Ditto.
9464 (WORDS_BIG_ENDIAN): Ditto.
9465 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
9466 (ENDIAN_SPEC): New.
9467 (EXTRA_SPECS): New.
9468 * config/tilegx/tilegx.md (extv): Handle big endian.
9469 (extzv): Ditto.
9470 (insn_st<n>): Ditto.
9471 (insn_st<n>_add<bitsuffix>): Ditto.
9472 (insn_stnt<n>): Ditto.
9473 (insn_stnt<n>_add<bitsuffix>):Ditto.
9474 (vec_interleave_highv8qi): Handle big endian.
9475 (vec_interleave_highv8qi_be): New.
9476 (vec_interleave_highv8qi_le): New.
9477 (insn_v1int_h): Handle big endian.
9478 (vec_interleave_lowv8qi): Handle big endian.
9479 (vec_interleave_lowv8qi_be): New.
9480 (vec_interleave_lowv8qi_le): New.
9481 (insn_v1int_l): Handle big endian.
9482 (vec_interleave_highv4hi): Handle big endian.
9483 (vec_interleave_highv4hi_be): New.
9484 (vec_interleave_highv4hi_le): New.
9485 (insn_v2int_h): Handle big endian.
9486 (vec_interleave_lowv4hi): Handle big endian.
9487 (vec_interleave_lowv4hi_be): New.
9488 (vec_interleave_lowv4hi_le): New.
9489 (insn_v2int_l): Handle big endian.
9490 (vec_interleave_highv2si): Handle big endian.
9491 (vec_interleave_highv2si_be): New.
9492 (vec_interleave_highv2si_le): New.
9493 (insn_v4int_h): Handle big endian.
9494 (vec_interleave_lowv2si): Handle big endian.
9495 (vec_interleave_lowv2si_be): New.
9496 (vec_interleave_lowv2si_le): New.
9497 (insn_v4int_l): Handle big endian.
9498 * config/tilegx/tilegx.opt (mbig-endian): New option.
9499 (mlittle-endian): New option.
9500 * doc/install.texi: Document tilegxbe-linux.
9501 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
9503 2014-02-24 Martin Jambor <mjambor@suse.cz>
9505 PR ipa/60266
9506 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
9507 there are no parameter descriptors.
9509 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
9511 PR rtl-optimization/60268
9512 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
9513 initialization to ...
9514 (sched_rgn_init): ... here.
9515 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
9517 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
9519 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
9520 names.
9522 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
9524 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
9525 definition.
9527 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
9529 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
9530 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
9532 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
9534 * config/microblaze/predicates.md: Add cmp_op predicate.
9535 * config/microblaze/microblaze.md: Add branch_compare instruction
9536 which uses cmp_op predicate and emits cmp insn before branch.
9537 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
9538 to microblaze_expand_conditional_branch and consolidate logic.
9539 (microblaze_expand_conditional_branch): emit branch_compare
9540 insn instead of handling cmp op separate from branch insn.
9542 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9544 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
9545 to permit subregs.
9547 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9549 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
9550 define_insn with define_expand and new define_insn
9551 *altivec_lve<VI_char>x_internal.
9552 (altivec_stve<VI_char>x): Replace define_insn with define_expand
9553 and new define_insn *altivec_stve<VI_char>x_internal.
9554 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
9555 prototype.
9556 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
9557 lve*x built-ins.
9558 (altivec_expand_stvex_be): New function.
9560 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
9562 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
9563 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
9564 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
9565 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
9567 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
9569 PR target/60298
9570 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
9571 instead of emit_move_insn.
9573 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9575 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
9576 vspltw with vsldoi.
9577 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
9578 gen_altivec_vsumsws.
9580 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9582 * config/rs6000/altivec.md (altivec_lvxl): Rename as
9583 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
9584 (altivec_lvxl_<mode>): New define_expand incorporating
9585 -maltivec=be semantics where needed.
9586 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
9587 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
9588 semantics where needed.
9589 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
9590 (altivec_stvx_<mode>): New define_expand incorporating
9591 -maltivec=be semantics where needed.
9592 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
9593 VM2 iterator instead of V4SI.
9594 (altivec_stvxl_<mode>): New define_expand incorporating
9595 -maltivec=be semantics where needed.
9596 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
9597 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
9598 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
9599 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
9600 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
9601 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
9602 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
9603 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
9604 ALTIVEC_BUILTIN_STVXL.
9605 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
9606 (altivec_expand_stvx_be): Likewise.
9607 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
9608 (altivec_expand_lvx_be): Likewise.
9609 (altivec_expand_stvx_be): Likewise.
9610 (altivec_expand_builtin): Add cases for
9611 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
9612 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
9613 (altivec_init_builtins): Add definitions for
9614 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
9615 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
9617 2014-02-21 Catherine Moore <clm@codesourcery.com>
9619 * doc/invoke.texi (mvirt, mno-virt): Document.
9620 * config/mips/mips.opt (mvirt): New option.
9621 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
9623 2014-02-21 Richard Biener <rguenther@suse.de>
9625 PR tree-optimization/60276
9626 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
9627 (STMT_VINFO_MIN_NEG_DIST): New macro.
9628 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
9629 STMT_VINFO_MIN_NEG_DIST.
9630 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
9631 made for negative dependence distances still hold.
9633 2014-02-21 Richard Biener <rguenther@suse.de>
9635 PR middle-end/60291
9636 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
9637 DECL_INITIAL for globals not in the current function context.
9639 2014-02-21 Jakub Jelinek <jakub@redhat.com>
9641 PR tree-optimization/56490
9642 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
9643 * tree-ssa-uninit.c: Include params.h.
9644 (compute_control_dep_chain): Add num_calls argument, return false
9645 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
9646 num_calls to recursive call.
9647 (find_predicates): Change dep_chain into normal array,
9648 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
9649 variable and adjust compute_control_dep_chain caller.
9650 (find_def_preds): Likewise.
9652 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
9654 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
9655 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
9657 2014-02-21 Nick Clifton <nickc@redhat.com>
9659 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
9660 (pushhi1): Likewise.
9661 (popqi1): Add mode to pre_dec.
9662 (pophi1): Likewise.
9664 2014-02-21 Jakub Jelinek <jakub@redhat.com>
9666 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
9667 mode for mask of V8SFmode permutation.
9669 2014-02-20 Richard Henderson <rth@redhat.com>
9671 PR c++/60272
9672 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
9673 a new pseudo for OLDVAL.
9675 2014-02-20 Jakub Jelinek <jakub@redhat.com>
9677 PR target/57896
9678 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
9679 gen_reg_rtx if d->testing_p.
9680 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
9681 if d->testing_p and we will certainly return true.
9682 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
9683 if d->testing_p.
9685 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
9687 * emit-rtl.c (gen_reg_rtx): Assert that
9688 crtl->emit.regno_pointer_align_length is non-zero.
9690 2014-02-20 Richard Henderson <rth@redhat.com>
9692 PR c++/60272
9693 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
9694 on failure the store back into EXPECT.
9696 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
9697 Sandra Loosemore <sandra@codesourcery.com>
9699 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
9700 * config/nios2/nios2.c (nios2_function_profiler): Add
9701 -fPIC (flag_pic == 2) support.
9702 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
9703 (nios2_large_offset_p): New function.
9704 (nios2_unspec_reloc_p): Move up position, update to use
9705 nios2_large_offset_p.
9706 (nios2_unspec_address): Remove function.
9707 (nios2_unspec_offset): New function.
9708 (nios2_large_got_address): New function.
9709 (nios2_got_address): Add large offset support.
9710 (nios2_legitimize_tls_address): Update usage of removed and new
9711 functions.
9712 (nios2_symbol_binds_local_p): New function.
9713 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
9714 (nios2_legitimize_address): Update to use nios2_large_offset_p.
9715 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
9716 (nios2_print_operand): Merge H/L processing, add hiadj/lo
9717 processing for (const (unspec ...)).
9718 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
9720 2014-02-20 Richard Biener <rguenther@suse.de>
9722 * tree-cfg.c (replace_uses_by): Mark altered BBs before
9723 doing the substitution.
9724 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
9726 2014-02-20 Martin Jambor <mjambor@suse.cz>
9728 PR ipa/55260
9729 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
9730 info when checking whether lattices are bottom.
9732 2014-02-20 Richard Biener <rguenther@suse.de>
9734 PR middle-end/60221
9735 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
9736 regions at -O0.
9738 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
9740 PR ipa/58555
9741 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
9742 parameter specifying the scaling.
9743 (inline_call): Update.
9744 (want_inline_recursively): Guard division by zero.
9745 (recursive_inlining): Update.
9746 * ipa-inline.h (clone_inlined_nodes): Update.
9748 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
9750 PR target/60204
9751 * config/i386/i386.c (classify_argument): Pass structures of size
9752 64 bytes or less in register.
9754 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
9755 Kirill Yukhin <kirill.yukhin@intel.com>
9757 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
9758 (_mm_rcp28_round_ss): Ditto.
9759 (_mm_rsqrt28_round_sd): Ditto.
9760 (_mm_rsqrt28_round_ss): Ditto.
9761 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
9762 (_mm_rcp14_round_ss): Ditto.
9763 (_mm_rsqrt14_round_sd): Ditto.
9764 (_mm_rsqrt14_round_ss): Ditto.
9765 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
9766 the first input operand, get rid of match_dup.
9767 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
9768 attribute to sse.
9769 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
9770 Ditto.
9771 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
9772 operand as the first input operand, set type attribute.
9773 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
9774 Set type attribute.
9775 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
9776 operand as the first input operand, set type attribute.
9778 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9780 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
9781 bit of zero.
9783 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
9785 PR target/60207
9786 * config/i386/i386.c (construct_container): Remove TFmode check
9787 for X86_64_INTEGER_CLASS.
9789 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
9791 PR target/59794
9792 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
9793 only when -Wpsabi is enabled.
9795 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
9797 PR target/59799
9798 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
9799 passing arrays in registers are the same as for structs, so remove the
9800 special case for them.
9802 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
9804 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
9805 destination type, extract only the valid bits if the source type is not
9806 integral and has a different mode.
9808 2014-02-19 Richard Biener <rguenther@suse.de>
9810 PR ipa/60243
9811 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
9812 for all calls.
9814 2014-02-19 Richard Biener <rguenther@suse.de>
9816 PR ipa/60243
9817 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
9818 (ipa_modify_call_arguments): Emit an argument load explicitely and
9819 preserve virtual SSA form there and for the replacement call.
9820 Do not update SSA form nor free dominance info.
9822 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
9824 * ipa.c (function_and_variable_visibility): Also clear WEAK
9825 flag when disolving COMDAT_GROUP.
9827 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
9829 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
9830 * ipa-prop.c (ipa_set_jf_known_type): Return early when
9831 not devirtualizing.
9832 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
9833 do more sanity checks.
9834 (detect_type_change): Return true when giving up early.
9835 (compute_complex_assign_jump_func): Fix type parameter of
9836 ipa_set_ancestor_jf.
9837 (compute_complex_ancestor_jump_func): Likewise.
9838 (update_jump_functions_after_inlining): Fix updating of
9839 ancestor function.
9840 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
9842 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
9844 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
9845 inline clones when edge disappears.
9847 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
9849 PR target/60203
9850 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
9851 Split 64-bit moves into 2 patterns. Do not allow the use of
9852 direct move for TDmode in little endian, since the decimal value
9853 has little endian bytes within a word, but the 64-bit pieces are
9854 ordered in a big endian fashion, and normal subreg's of TDmode are
9855 not allowed.
9856 (mov<mode>_64bit_dm): Likewise.
9857 (movtd_64bit_nodm): Likewise.
9859 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
9861 PR tree-optimization/60174
9862 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
9863 statement of an SSA_NAME that occurs in an abnormal PHI node.
9865 2014-02-18 Jakub Jelinek <jakub@redhat.com>
9867 PR sanitizer/60142
9868 * final.c (SEEN_BB): Remove.
9869 (SEEN_NOTE, SEEN_EMITTED): Renumber.
9870 (final_scan_insn): Don't force_source_line on second
9871 NOTE_INSN_BASIC_BLOCK.
9873 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
9875 PR target/60205
9876 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
9877 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
9878 (type_natural_mode): Warn ABI change when %zmm register is not
9879 available for AVX512F vector value passing.
9881 2014-02-18 Kai Tietz <ktietz@redhat.com>
9883 PR target/60193
9884 * config/i386/i386.c (ix86_expand_prologue): Use value in
9885 rax register as displacement when restoring %r10 or %rax.
9886 Fix wrong offset when restoring both registers.
9888 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
9890 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
9891 assertion with conditional return.
9893 2014-02-18 Jakub Jelinek <jakub@redhat.com>
9894 Uros Bizjak <ubizjak@gmail.com>
9896 PR driver/60233
9897 * config/i386/driver-i386.c (host_detect_local_cpu): If
9898 YMM state is not saved by the OS, also clear has_f16c. Move
9899 CPUID 0x80000001 handling before YMM state saving checking.
9901 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
9903 PR rtl-optimization/58960
9904 * haifa-sched.c (alloc_global_sched_pressure_data): New,
9905 factored out from ...
9906 (sched_init): ... here.
9907 (free_global_sched_pressure_data): New, factored out from ...
9908 (sched_finish): ... here.
9909 * sched-int.h (free_global_sched_pressure_data): Declare.
9910 * sched-rgn.c (nr_regions_initial): New static global.
9911 (haifa_find_rgns): Initialize it.
9912 (schedule_region): Disable sched-pressure for the newly
9913 generated regions.
9915 2014-02-17 Richard Biener <rguenther@suse.de>
9917 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
9918 release SSA defs of pattern stmts.
9920 2014-02-17 Richard Biener <rguenther@suse.de>
9922 * tree-inline.c (expand_call_inline): Release the virtual
9923 operand defined by the call we are about to inline.
9925 2014-02-17 Richard Biener <rguenther@suse.de>
9927 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
9929 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
9930 Ilya Tocar <ilya.tocar@intel.com>
9932 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
9933 arguments order in builtin.
9934 (_mm512_permutexvar_epi64): Ditto.
9935 (_mm512_mask_permutexvar_epi64): Ditto
9936 (_mm512_maskz_permutexvar_epi32): Ditto
9937 (_mm512_permutexvar_epi32): Ditto
9938 (_mm512_mask_permutexvar_epi32): Ditto
9940 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9942 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
9943 (p8_vmrgow): Likewise.
9945 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9947 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
9948 endian targets.
9950 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
9952 PR target/60203
9953 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
9954 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
9955 into 64-bit and 32-bit moves. On 64-bit moves, add support for
9956 using direct move instructions on ISA 2.07. Also adjust
9957 instruction length for 64-bit.
9958 (mov<mode>_64bit, TFmode/TDmode): Likewise.
9959 (mov<mode>_32bit, TFmode/TDmode): Likewise.
9961 2014-02-15 Alan Modra <amodra@gmail.com>
9963 PR target/58675
9964 PR target/57935
9965 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
9966 find_replacement on parts of insn rtl that might be reloaded.
9968 2014-02-15 Richard Biener <rguenther@suse.de>
9970 PR tree-optimization/60183
9971 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
9972 (tree_ssa_phiprop): Calculate and free post-dominators.
9974 2014-02-14 Jeff Law <law@redhat.com>
9976 PR rtl-optimization/60131
9977 * ree.c (get_extended_src_reg): New function.
9978 (combine_reaching_defs): Use it rather than assuming location of REG.
9979 (find_and_remove_re): Verify first operand of extension is
9980 a REG before adding the insns to the copy list.
9982 2014-02-14 Roland McGrath <mcgrathr@google.com>
9984 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
9985 * configure: Regenerated.
9986 * config.in: Regenerated.
9987 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
9988 instead of ASM_SHORT.
9990 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
9991 Richard Earnshaw <rearnsha@arm.com>
9993 PR rtl-optimization/59535
9994 * lra-constraints.c (process_alt_operands): Encourage alternative
9995 when unassigned pseudo class is superset of the alternative class.
9996 (inherit_reload_reg): Don't inherit when optimizing for code size.
9997 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
9998 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
9999 modes not less than 4 for Thumb1.
10001 2014-02-14 Kyle McMartin <kyle@redhat.com>
10003 PR pch/60010
10004 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
10006 2014-02-14 Richard Biener <rguenther@suse.de>
10008 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
10009 (get_frame_arg): Drop the assert with langhook types_compatible_p.
10010 Do not strip INDIRECT_REFs.
10012 2014-02-14 Richard Biener <rguenther@suse.de>
10014 PR lto/60179
10015 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
10016 DECL_FUNCTION_SPECIFIC_TARGET.
10017 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
10018 * tree-streamer-out.c (pack_ts_target_option): Remove.
10019 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
10020 (write_ts_function_decl_tree_pointers): Do not stream
10021 DECL_FUNCTION_SPECIFIC_TARGET.
10022 * tree-streamer-in.c (unpack_ts_target_option): Remove.
10023 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
10024 (lto_input_ts_function_decl_tree_pointers): Do not stream
10025 DECL_FUNCTION_SPECIFIC_TARGET.
10027 2014-02-14 Jakub Jelinek <jakub@redhat.com>
10029 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
10030 (get_initial_def_for_induction, vectorizable_induction): Ignore
10031 debug stmts when looking for exit_phi.
10032 (vectorizable_live_operation): Fix up condition.
10034 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
10036 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
10037 nreverse() because it changes the content of original tree list.
10039 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
10041 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
10042 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
10044 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
10046 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
10047 GNU coding standards.
10049 2014-02-13 Jakub Jelinek <jakub@redhat.com>
10051 PR debug/60152
10052 * dwarf2out.c (gen_subprogram_die): Don't call
10053 add_calling_convention_attribute if subr_die is old_die.
10055 2014-02-13 Sharad Singhai <singhai@google.com>
10057 * doc/optinfo.texi: Fix order of nodes.
10059 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
10061 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
10062 operands[2], not operands[3].
10064 2014-02-13 Richard Biener <rguenther@suse.de>
10066 PR bootstrap/59878
10067 * doc/install.texi (ISL): Update recommended version to 0.12.2,
10068 mention the possibility of an in-tree build.
10069 (CLooG): Update recommended version to 0.18.1, mention the
10070 possibility of an in-tree build and clarify that the ISL
10071 bundled with CLooG does not work.
10073 2014-02-13 Jakub Jelinek <jakub@redhat.com>
10075 PR target/43546
10076 * expr.c (compress_float_constant): If x is a hard register,
10077 extend into a pseudo and then move to x.
10079 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
10081 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
10082 caused by bad second argument to warning_at() with -mhotpatch and
10083 nested functions (e.g. with gfortran).
10085 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
10087 * opts.c (option_name): Remove "enabled by default" rider.
10089 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
10091 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
10093 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
10094 Uros Bizjak <ubizjak@gmail.com>
10096 PR target/60151
10097 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
10098 * configure: Regenerated.
10100 2014-02-12 Richard Biener <rguenther@suse.de>
10102 * vec.c (vec_prefix::calculate_allocation): Move as
10103 inline variant to vec.h.
10104 (vec_prefix::calculate_allocation_1): New out-of-line version.
10105 * vec.h (vec_prefix::calculate_allocation_1): Declare.
10106 (vec_prefix::m_has_auto_buf): Rename to ...
10107 (vec_prefix::m_using_auto_storage): ... this.
10108 (vec_prefix::calculate_allocation): Inline the easy cases
10109 and dispatch to calculate_allocation_1 which doesn't need the
10110 prefix address.
10111 (va_heap::reserve): Use gcc_checking_assert.
10112 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
10113 m_using_auto_storage.
10114 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
10115 member and adjust.
10116 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
10117 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
10118 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
10120 2014-02-12 Richard Biener <rguenther@suse.de>
10122 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
10123 when we found a dependence.
10125 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
10127 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
10128 common code...
10129 (maybe_fold_stmt): ... into this new function.
10130 * omp-low.c (lower_omp): Update comment.
10132 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
10133 last use.
10135 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
10136 dereference.
10138 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
10140 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
10141 identifiers in comments.
10142 (cortexa53_extra_costs): Likewise.
10143 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
10144 (cortexa7_extra_costs): Likewise.
10145 (cortexa12_extra_costs): Likewise.
10146 (cortexa15_extra_costs): Likewise.
10147 (v7m_extra_costs): Likewise.
10149 2014-02-12 Richard Biener <rguenther@suse.de>
10151 PR middle-end/60092
10152 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
10153 of posix_memalign being successful.
10154 (lower_stmt): Restrict lowering of posix_memalign to when
10155 -ftree-bit-ccp is enabled.
10157 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
10159 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
10160 arg_loc.
10161 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
10163 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
10165 PR rtl-optimization/60116
10166 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
10167 other_insn once the combination has been validated.
10169 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
10171 PR lto/59468
10172 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
10173 and wrapper.
10174 * ipa-devirt.c: Include demangle.h
10175 (odr_violation_reported): New static variable.
10176 (add_type_duplicate): Update odr_violations.
10177 (maybe_record_node): Add completep parameter; update it.
10178 (record_target_from_binfo): Add COMPLETEP parameter;
10179 update it as needed.
10180 (possible_polymorphic_call_targets_1): Likewise.
10181 (struct polymorphic_call_target_d): Add nonconstruction_targets;
10182 rename FINAL to COMPLETE.
10183 (record_targets_from_bases): Sanity check we found the binfo;
10184 fix COMPLETEP updating.
10185 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
10186 parameter, fix computing of COMPLETEP.
10187 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
10188 at LTO time do demangling.
10189 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
10190 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
10191 parameter.
10192 (gimple_get_virt_method_for_binfo): Likewise.
10193 * gimple-fold.h (gimple_get_virt_method_for_binfo,
10194 gimple_get_virt_method_for_vtable): Update prototypes.
10196 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
10198 PR target/49008
10199 * genautomata.c (add_presence_absence): Fix typo with
10200 {final_}presence_list.
10202 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
10204 PR target/60137
10205 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
10206 for VSX/Altivec vectors that land in GPR registers.
10208 2014-02-11 Richard Henderson <rth@redhat.com>
10209 Jakub Jelinek <jakub@redhat.com>
10211 PR debug/59776
10212 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
10213 around drhs if type conversion to lacc->type is not useless.
10215 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10217 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
10218 tuning struct.
10219 (cortex-a57.cortex-a53): Likewise.
10220 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
10222 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10224 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
10225 arm_restrict_it.
10227 2014-02-11 Renlin Li <Renlin.Li@arm.com>
10229 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
10230 add_options_for_arm_vfp3.
10232 2014-02-11 Jeff Law <law@redhat.com>
10234 PR middle-end/54041
10235 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
10236 object with an undesirable mode.
10238 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
10240 PR libgomp/60107
10241 * config/i386/sol2-9.h: New file.
10242 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
10243 *-*-solaris2.9*): Use it.
10245 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
10247 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
10248 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
10250 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
10252 * config/microblaze/microblaze.c: Extend mcpu version format
10254 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
10256 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
10258 2014-02-10 Richard Henderson <rth@redhat.com>
10260 PR target/59927
10261 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
10262 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
10263 ms-abi vs -mno-accumulate-outgoing-args.
10264 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
10265 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
10266 respect to ms-abi.
10268 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
10270 PR middle-end/60080
10271 * cfgexpand.c (expand_asm_operands): Attach source location to
10272 ASM_INPUT rtx objects.
10273 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
10275 2014-02-10 Nick Clifton <nickc@redhat.com>
10277 * config/mn10300/mn10300.c (popcount): New function.
10278 (mn10300_expand_prologue): Include saved registers in stack usage
10279 count.
10281 2014-02-10 Jeff Law <law@redhat.com>
10283 PR middle-end/52306
10284 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
10285 when changing the SET_DEST of a prior insn to avoid an input reload.
10287 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
10289 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
10290 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
10291 -mcall-openbsd, or -mcall-linux.
10292 (CC1_ENDIAN_BIG_SPEC): Remove.
10293 (CC1_ENDIAN_LITTLE_SPEC): Remove.
10294 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
10295 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
10296 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
10297 and %cc1_endian_default.
10298 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
10300 2014-02-10 Richard Biener <rguenther@suse.de>
10302 PR tree-optimization/60115
10303 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
10304 MEM_REF handling. Properly verify that the accesses are not
10305 out of the objects bound.
10307 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10309 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
10310 coretex to cortex.
10312 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
10314 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
10315 proper constants and fix formatting.
10316 (possible_polymorphic_call_targets): Fix formatting.
10318 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
10319 Ilya Tocar <ilya.tocar@intel.com>
10321 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
10322 (_mm512_loadu_epi32): Renamed into...
10323 (_mm512_loadu_si512): This.
10324 (_mm512_storeu_epi32): Renamed into...
10325 (_mm512_storeu_si512): This.
10326 (_mm512_maskz_ceil_ps): Removed.
10327 (_mm512_maskz_ceil_pd): Ditto.
10328 (_mm512_maskz_floor_ps): Ditto.
10329 (_mm512_maskz_floor_pd): Ditto.
10330 (_mm512_floor_round_ps): Ditto.
10331 (_mm512_floor_round_pd): Ditto.
10332 (_mm512_ceil_round_ps): Ditto.
10333 (_mm512_ceil_round_pd): Ditto.
10334 (_mm512_mask_floor_round_ps): Ditto.
10335 (_mm512_mask_floor_round_pd): Ditto.
10336 (_mm512_mask_ceil_round_ps): Ditto.
10337 (_mm512_mask_ceil_round_pd): Ditto.
10338 (_mm512_maskz_floor_round_ps): Ditto.
10339 (_mm512_maskz_floor_round_pd): Ditto.
10340 (_mm512_maskz_ceil_round_ps): Ditto.
10341 (_mm512_maskz_ceil_round_pd): Ditto.
10342 (_mm512_expand_pd): Ditto.
10343 (_mm512_expand_ps): Ditto.
10344 * config/i386/i386.c (ix86_builtins): Remove
10345 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
10346 (bdesc_args): Ditto.
10347 * config/i386/predicates.md (const1256_operand): New.
10348 (const_1_to_2_operand): Ditto.
10349 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
10350 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
10351 (*avx512pf_gatherpf<mode>sf): Ditto.
10352 (avx512pf_gatherpf<mode>df): Ditto.
10353 (*avx512pf_gatherpf<mode>df_mask): Ditto.
10354 (*avx512pf_gatherpf<mode>df): Ditto.
10355 (avx512pf_scatterpf<mode>sf): Ditto.
10356 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
10357 (*avx512pf_scatterpf<mode>sf): Ditto.
10358 (avx512pf_scatterpf<mode>df): Ditto.
10359 (*avx512pf_scatterpf<mode>df_mask): Ditto.
10360 (*avx512pf_scatterpf<mode>df): Ditto.
10361 (avx512f_expand<mode>): Removed.
10362 (<shift_insn><mode>3<mask_name>): Change predicate type.
10364 2014-02-08 Jakub Jelinek <jakub@redhat.com>
10366 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
10367 not at the end of datarefs vector use ordered_remove to avoid
10368 reordering datarefs vector.
10370 PR c/59984
10371 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
10372 mark local addressable non-static vars as GOVD_PRIVATE
10373 instead of GOVD_LOCAL.
10374 * omp-low.c (lower_omp_for): Move gimple_bind_vars
10375 and BLOCK_VARS of gimple_bind_block to new_stmt rather
10376 than copying them.
10378 PR middle-end/60092
10379 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
10380 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
10381 assume_aligned or alloc_align attributes.
10382 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
10383 arguments. Handle also assume_aligned and alloc_align attributes.
10384 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
10385 calls to functions with assume_aligned or alloc_align attributes.
10386 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
10388 2014-02-08 Terry Guo <terry.guo@arm.com>
10390 * doc/invoke.texi: Document ARM -march=armv7e-m.
10392 2014-02-08 Jakub Jelinek <jakub@redhat.com>
10394 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
10395 flag on __cilkrts_rethrow builtin.
10397 PR ipa/60026
10398 * ipa-cp.c (determine_versionability): Fail at -O0
10399 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
10400 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
10402 Revert:
10403 2014-02-04 Jakub Jelinek <jakub@redhat.com>
10405 PR ipa/60026
10406 * tree-inline.c (copy_forbidden): Fail for
10407 __attribute__((optimize (0))) functions.
10409 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
10411 * varpool.c: Include pointer-set.h.
10412 (varpool_remove_unreferenced_decls): Variables in other partitions
10413 will not be output; be however careful to not lose information
10414 about partitioning.
10416 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
10418 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
10419 lookup in the vtable constructor.
10421 2014-02-07 Jeff Law <law@redhat.com>
10423 PR target/40977
10424 * config/m68k/m68k.md (ashldi_extsi): Turn into a
10425 define_insn_and_split.
10427 * ipa-inline.c (inline_small_functions): Fix typos.
10429 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10431 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
10432 (s390_can_use_return_insn): Declare.
10433 * config/s390/s390.h (EPILOGUE_USES): Define.
10434 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
10435 instructions.
10436 (s390_chunkify_start): Handle return JUMP_LABELs.
10437 (s390_early_mach): Emit a main_pool instruction on the entry edge.
10438 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
10439 (s390_can_use_return_insn): New functions.
10440 (s390_fix_long_loop_prediction): Handle conditional returns.
10441 (TARGET_SET_UP_BY_PROLOGUE): Define.
10442 * config/s390/s390.md (ANY_RETURN): New code iterator.
10443 (*creturn, *csimple_return, return, simple_return): New patterns.
10445 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10447 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
10448 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
10449 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
10450 REG_CFA_RESTORE list when deciding not to restore a register.
10452 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10454 * config/s390/s390.c: Include tree-pass.h and context.h.
10455 (s390_early_mach): New function, split out from...
10456 (s390_emit_prologue): ...here.
10457 (pass_data_s390_early_mach): New pass structure.
10458 (pass_s390_early_mach): New class.
10459 (s390_option_override): Create and register early_mach pass.
10460 Move to end of file.
10462 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10464 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
10465 to match for the exit block.
10467 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10469 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
10470 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
10471 Reject misaligned operands.
10473 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10475 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
10477 2014-02-07 Richard Biener <rguenther@suse.de>
10479 PR middle-end/60092
10480 * gimple-low.c (lower_builtin_posix_memalign): New function.
10481 (lower_stmt): Call it to lower posix_memalign in a way
10482 to make alignment info accessible.
10484 2014-02-07 Jakub Jelinek <jakub@redhat.com>
10486 PR c++/60082
10487 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
10488 __builtin_setjmp_receiver.
10490 2014-02-07 Richard Biener <rguenther@suse.de>
10492 PR middle-end/60092
10493 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
10494 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
10495 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
10496 Handle BUILT_IN_POSIX_MEMALIGN.
10497 (find_func_clobbers): Likewise.
10498 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
10499 (call_may_clobber_ref_p_1): Likewise.
10501 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
10503 PR ipa/59918
10504 * ipa-devirt.c (record_target_from_binfo): Remove overactive
10505 sanity check.
10507 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
10509 PR ipa/59469
10510 * lto-cgraph.c (lto_output_node): Use
10511 symtab_get_symbol_partitioning_class.
10512 (lto_output_varpool_node): likewise.
10513 (symtab_get_symbol_partitioning_class): Move here from
10514 lto/lto-partition.c
10515 * cgraph.h (symbol_partitioning_class): Likewise.
10516 (symtab_get_symbol_partitioning_class): Declare.
10518 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
10520 * ggc.h (ggc_internal_cleared_alloc): New macro.
10521 * vec.h (vec_safe_copy): Handle memory stats.
10522 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
10523 * target-globals.c (save_target_globals): Likewise.
10525 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
10527 PR target/60077
10528 * expr.c (emit_move_resolve_push): Export; be bit more selective
10529 on when to clear alias set.
10530 * expr.h (emit_move_resolve_push): Declare.
10531 * function.h (struct function): Add tail_call_marked.
10532 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
10533 * config/i386/i386-protos.h (ix86_expand_push): Remove.
10534 * config/i386/i386.md (TImode move expander): De not call
10535 ix86_expand_push.
10536 (FP push expanders): Preserve memory attributes.
10537 * config/i386/sse.md (push<mode>1): Remove.
10538 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
10539 (ix86_expand_push): Remove.
10540 * config/i386/mmx.md (push<mode>1): Remove.
10542 2014-02-06 Jakub Jelinek <jakub@redhat.com>
10544 PR rtl-optimization/60030
10545 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
10546 lopart with paradoxical subreg before shifting it up by hprec.
10548 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10550 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
10551 Remove extra newline at end of file.
10552 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
10553 (arm_issue_rate): Handle cortexa57.
10554 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
10555 (cortex-a57.cortex-a53): Likewise.
10557 2014-02-06 Jakub Jelinek <jakub@redhat.com>
10559 PR target/59575
10560 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
10561 don't record in REG_FRAME_RELATED_EXPR registers not set in that
10562 bitmask.
10563 (arm_expand_prologue): Adjust all callers.
10564 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
10565 info, registers also at the lowest numbered registers side. Use
10566 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
10567 XEXP.
10569 PR debug/59992
10570 * var-tracking.c (adjust_mems): Before adding a SET to
10571 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
10573 2014-02-06 Alan Modra <amodra@gmail.com>
10575 PR target/60032
10576 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
10577 change SDmode to DDmode when lra_in_progress.
10579 2014-02-06 Jakub Jelinek <jakub@redhat.com>
10581 PR middle-end/59150
10582 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
10583 free_data_ref on the dr first, and before goto again also set dr
10584 to the next dr. For simd_lane_access, free old datarefs[i] before
10585 overwriting it. For get_vectype_for_scalar_type failure, don't
10586 free_data_ref if simd_lane_access.
10588 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
10590 PR target/60062
10591 * tree.h (opts_for_fn): New inline function.
10592 (opt_for_fn): Define.
10593 * config/i386/i386.c (ix86_function_regparm): Use
10594 opt_for_fn (decl, optimize) instead of optimize.
10596 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
10598 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
10599 for SYMBOL_REF in large memory model.
10601 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10603 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
10604 and crypto support.
10605 (cortex-a57): Likewise.
10606 (cortex-a57.cortex-a53): Likewise.
10608 2014-02-06 Yury Gribov <y.gribov@samsung.com>
10609 Kugan Vivekanandarajah <kuganv@linaro.org>
10611 * config/arm/arm.c (arm_vector_alignment_reachable): Check
10612 unaligned_access.
10613 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
10615 2014-02-06 Richard Biener <rguenther@suse.de>
10617 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
10618 set_loop_copy and initialize_original_copy_tables.
10620 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
10622 * config/aarch64/aarch64-simd.md
10623 (aarch64_ashr_simddi): Change QI to SI.
10625 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
10626 Jakub Jelinek <jakub@redhat.com>
10628 PR middle-end/60013
10629 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
10630 of the dataflow.
10632 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10634 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
10635 CODE_FOR_altivec_vpku[hw]um to
10636 CODE_FOR_altivec_vpku[hw]um_direct.
10637 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
10638 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
10639 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
10640 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
10642 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10644 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
10645 generation for -maltivec=be.
10646 (altivec_vsumsws): Simplify redundant test.
10648 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10650 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
10651 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
10652 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
10653 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
10654 gen_altivec_vpkuwum.
10655 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
10656 BYTES_BIG_ENDIAN.
10657 (altivec_vpks<VI_char>ss): Likewise.
10658 (altivec_vpks<VI_char>us): Likewise.
10659 (altivec_vpku<VI_char>us): Likewise.
10660 (altivec_vpku<VI_char>um): Likewise.
10661 (altivec_vpku<VI_char>um_direct): New (copy of
10662 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
10663 internal use).
10664 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
10665 target is little endian and -maltivec=be is not specified.
10666 (*altivec_vupkhs<VU_char>_direct): New (copy of
10667 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
10668 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
10669 target is little endian and -maltivec=be is not specified.
10670 (*altivec_vupkls<VU_char>_direct): New (copy of
10671 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
10672 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
10673 little endian and -maltivec=be is not specified.
10674 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
10675 little endian and -maltivec=be is not specified.
10677 2014-02-05 Richard Henderson <rth@redhat.com>
10679 PR debug/52727
10680 * combine-stack-adj.c: Revert r206943.
10681 * sched-int.h (struct deps_desc): Add last_args_size.
10682 * sched-deps.c (init_deps): Initialize it.
10683 (sched_analyze_insn): Add OUTPUT dependencies between insns that
10684 contain REG_ARGS_SIZE notes.
10686 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
10688 * lto-cgraph.c (asm_nodes_output): Make global.
10689 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
10690 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
10691 (driver_handle_option): Handle OPT_fwpa.
10693 2014-02-05 Jakub Jelinek <jakub@redhat.com>
10695 PR ipa/59947
10696 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
10697 a comment typo and formatting issue. If odr_hash hasn't been
10698 created, return vNULL and set *completep to false.
10700 PR middle-end/57499
10701 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
10702 bb with no successors.
10704 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
10706 PR target/59718
10707 * doc/invoke.texi (-march): Clarify documentation for ARM.
10708 (-mtune): Likewise.
10709 (-mcpu): Likewise.
10711 2014-02-05 Richard Biener <rguenther@suse.de>
10713 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
10714 when not vectorizing because of too many alias checks.
10715 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
10716 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
10718 2014-02-05 Nick Clifton <nickc@redhat.com>
10720 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
10721 accept extended registers in any mode when compiling for the MN10300.
10723 2014-02-05 Yury Gribov <y.gribov@samsung.com>
10725 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
10726 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
10727 sanitization attributes.
10728 (can_inline_edge_p): Likewise.
10729 (sanitize_attrs_match_for_inline_p): New function.
10731 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
10733 * ipa-prop.c (detect_type_change): Shor circuit testing of
10734 type changes on THIS pointer.
10736 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
10738 PR target/59777
10739 * config/pa/pa.c (legitimize_tls_address): Return original address
10740 if not passed a SYMBOL_REF rtx.
10741 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
10742 addresses.
10743 (pa_emit_move_sequence): Simplify TLS source operands.
10744 (pa_legitimate_constant_p): Reject all TLS constants.
10745 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
10746 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
10748 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
10750 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
10751 groups when we know they are controlled by LTO.
10752 * varasm.c (default_binds_local_p_1): If object is in other partition,
10753 it will be resolved locally.
10755 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
10757 * config/host-linux.c (linux_gt_pch_use_address): Don't
10758 use SSIZE_MAX because it is not always defined.
10760 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
10762 PR bootstrap/59913
10763 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
10764 threshold for pseudo splitting.
10765 (update_ebb_live_info): Process call argument hard registers and
10766 hard registers from insn definition too.
10767 (max_small_class_regs_num): New constant.
10768 (inherit_in_ebb): Update live hard regs through EBBs. Update
10769 reloads_num only for small register classes. Don't split for
10770 outputs of jumps.
10772 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
10774 PR ipa/60058
10775 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
10776 is non-null.
10778 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
10780 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
10781 visibility is safe.
10783 2014-02-04 Marek Polacek <polacek@redhat.com>
10785 * gdbinit.in (pel): Define.
10787 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
10789 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
10790 behavior.
10792 2014-02-04 Richard Biener <rguenther@suse.de>
10794 PR lto/59723
10795 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
10796 in function context local.
10797 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
10798 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
10799 similar to LTO_imported_decl_ref.
10801 2014-02-04 Jakub Jelinek <jakub@redhat.com>
10803 PR tree-optimization/60002
10804 * cgraphclones.c (build_function_decl_skip_args): Clear
10805 DECL_LANG_SPECIFIC.
10807 PR tree-optimization/60023
10808 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
10809 false to gsi_replace.
10810 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
10811 has been in some EH region and vec_stmt could throw, add
10812 vec_stmt into the same EH region.
10813 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
10814 has no lhs, ignore it.
10815 * internal-fn.c (expand_MASK_LOAD): Likewise.
10817 PR ipa/60026
10818 * tree-inline.c (copy_forbidden): Fail for
10819 __attribute__((optimize (0))) functions.
10821 PR other/58712
10822 * omp-low.c (simd_clone_struct_copy): If from->inbranch
10823 is set, copy one less argument.
10824 (expand_simd_clones): Don't subtract clone_info->inbranch
10825 from simd_clone_struct_alloc argument.
10827 PR rtl-optimization/57915
10828 * recog.c (simplify_while_replacing): If all unary/binary/relational
10829 operation arguments are constant, attempt to simplify those.
10831 PR middle-end/59261
10832 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
10833 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
10835 2014-02-04 Richard Biener <rguenther@suse.de>
10837 PR tree-optimization/60012
10838 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
10839 TBAA disambiguation to all DDRs.
10841 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
10843 PR target/59788
10844 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
10845 (LINK_SPEC): Use it for -shared, -shared-libgcc.
10847 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10849 PR ipa/59882
10850 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
10852 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10854 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
10855 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
10857 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10859 PR ipa/59831
10860 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
10861 to figure out targets of polymorphic calls with known decl.
10862 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
10863 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
10864 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
10865 (get_polymorphic_call_info): ... here.
10866 (get_polymorphic_call_info_from_invariant): New function.
10868 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10870 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
10871 lookup via vtable pointer; check for type consistency
10872 and turn inconsitent facts into UNREACHABLE.
10873 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
10874 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
10875 type inconsistent querries; return UNREACHABLE instead.
10877 2014-02-03 Richard Henderson <rth@twiddle.net>
10879 PR tree-opt/59924
10880 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
10881 already processed this node.
10882 (normalize_one_pred_1): Pass along mark_set.
10883 (normalize_one_pred): Create and destroy a pointer_set_t.
10884 (normalize_one_pred_chain): Likewise.
10886 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
10888 PR gcov-profile/58602
10889 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
10891 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10893 PR ipa/59831
10894 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
10895 -fno-devirtualize; try to devirtualize by the knowledge of
10896 virtual table pointer given by aggregate propagation.
10897 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
10898 (ipa_print_node_jump_functions): Dump also offset that
10899 is relevant for polymorphic calls.
10900 (determine_known_aggregate_parts): Add arg_type parameter; use it
10901 instead of determining the type from pointer type.
10902 (ipa_compute_jump_functions_for_edge): Update call of
10903 determine_known_aggregate_parts.
10904 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
10905 (gimple_get_virt_method_for_binfo): ... here; simplify using
10906 vtable_pointer_value_to_vtable.
10907 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
10908 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
10909 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
10910 (vtable_pointer_value_to_vtable): Break out from ...; handle also
10911 POINTER_PLUS_EXPR.
10912 (vtable_pointer_value_to_binfo): ... here.
10913 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
10915 2014-02-03 Teresa Johnson <tejohnson@google.com>
10917 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
10918 redef of outer loop index variable.
10920 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
10922 PR c++/53017
10923 PR c++/59211
10924 * doc/extend.texi (Function Attributes): Typo.
10926 2014-02-03 Cong Hou <congh@google.com>
10928 PR tree-optimization/60000
10929 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
10930 if the vectorized statement is a store. A store statement can only
10931 appear at the end of pattern statements.
10933 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
10935 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
10936 (ix86_option_override_internal): Default long double to 64-bit for
10937 32-bit Bionic and to 128-bit for 64-bit Bionic.
10939 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
10940 TARGET_LONG_DOUBLE_128 is true.
10941 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
10943 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
10944 (mlong-double-64): Negate -mlong-double-128.
10945 (mlong-double-128): New option.
10947 * config/i386/i386-c.c (ix86_target_macros): Define
10948 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
10950 * doc/invoke.texi: Document -mlong-double-128.
10952 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
10954 PR rtl-optimization/60024
10955 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
10957 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
10959 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
10961 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
10963 PR rtl-optimization/57662
10964 * sel-sched.c (code_motion_path_driver): Do not mark already not
10965 existing blocks in the visiting bitmap.
10967 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
10969 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
10970 on the insn being emitted.
10972 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
10973 Will Deacon <will.deacon@arm.com>
10975 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
10977 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10979 * config/arm/arm-tables.opt: Regenerate.
10981 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10983 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
10984 for vector types other than V16QImode.
10985 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
10986 define_expand, and call altivec_expand_vec_perm_le when producing
10987 code with little endian element order.
10988 (*altivec_vperm_<mode>_internal): New insn having previous
10989 behavior of altivec_vperm_<mode>.
10990 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
10991 altivec_expand_vec_perm_le when producing code with little endian
10992 element order.
10993 (*altivec_vperm_<mode>_uns_internal): New insn having previous
10994 behavior of altivec_vperm_<mode>_uns.
10996 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10998 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
10999 (altivec_vsumsws): Add handling for -maltivec=be with a little
11000 endian target.
11001 (altivec_vsumsws_direct): New.
11002 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
11003 gen_altivec_vsumsws.
11005 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
11007 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
11008 vtable_pointer_value_to_binfo): New functions.
11009 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
11010 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
11012 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
11014 * config/nios2/nios2.md (load_got_register): Initialize GOT
11015 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
11016 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
11018 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
11020 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
11021 preserverd by passthrough, do not propagate the type.
11023 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
11025 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
11026 (mips_atomic_assign_expand_fenv): New function.
11027 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
11029 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
11031 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
11032 (__builtin_mips_set_fcsr): Likewise.
11033 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
11034 MIPS_USI_FTYPE_VOID.
11035 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
11036 (mips16_expand_set_fcsr): Likewise.
11037 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
11038 (mips16_set_fcsr_stub): Likewise.
11039 (mips16_get_fcsr_one_only_stub): New class.
11040 (mips16_set_fcsr_one_only_stub): Likewise.
11041 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
11042 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
11043 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
11044 (hard_float): New availability predicate.
11045 (mips_builtins): Add get_fcsr and set_fcsr.
11046 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
11047 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
11048 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
11049 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
11050 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
11051 patterns.
11053 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
11055 * config/mips/mips.c (mips_one_only_stub): New class.
11056 (mips_need_mips16_rdhwr_p): Replace with...
11057 (mips16_rdhwr_stub): ...this new variable.
11058 (mips16_stub_call_address): New function.
11059 (mips16_rdhwr_one_only_stub): New class.
11060 (mips_expand_thread_pointer): Use mips16_stub_call_address.
11061 (mips_output_mips16_rdhwr): Delete.
11062 (mips_finish_stub): New function.
11063 (mips_code_end): Use it to handle rdhwr stubs.
11065 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
11067 PR target/60017
11068 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
11069 when calculating size of integer atomic types.
11071 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
11073 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
11075 2014-02-01 Jakub Jelinek <jakub@redhat.com>
11077 PR tree-optimization/60003
11078 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
11079 * profile.c (branch_prob): Use gimple_call_builtin_p
11080 to check for BUILT_IN_SETJMP_RECEIVER.
11081 * tree-inline.c (copy_bb): Call notice_special_calls.
11083 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
11085 PR bootstrap/59985
11086 * lra-constraints.c (process_alt_operands): Update reload_sum only
11087 on the first pass.
11089 2014-01-31 Richard Henderson <rth@redhat.com>
11091 PR middle-end/60004
11092 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
11093 until after else_eh is processed.
11095 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
11097 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
11098 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
11099 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
11100 in smmintrin.h, remove them.
11101 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
11102 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
11103 * config/i386/i386.md (ROUND_SAE): Fix value.
11104 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
11105 (const48_operand): New.
11106 * config/i386/subst.md (round), (round_expand): Use
11107 const_4_or_8_to_11_operand.
11108 (round_saeonly), (round_saeonly_expand): Use const48_operand.
11110 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
11112 * config/i386/constraints.md (Yk): Swap meaning with k.
11113 * config/i386/i386.md (movhi_internal): Change Yk to k.
11114 (movqi_internal): Ditto.
11115 (*k<logic><mode>): Ditto.
11116 (*andhi_1): Ditto.
11117 (*andqi_1): Ditto.
11118 (kandn<mode>): Ditto.
11119 (*<code>hi_1): Ditto.
11120 (*<code>qi_1): Ditto.
11121 (kxnor<mode>): Ditto.
11122 (kortestzhi): Ditto.
11123 (kortestchi): Ditto.
11124 (kunpckhi): Ditto.
11125 (*one_cmplhi2_1): Ditto.
11126 (*one_cmplqi2_1): Ditto.
11127 * config/i386/sse.md (): Change k to Yk.
11128 (avx512f_load<mode>_mask): Ditto.
11129 (avx512f_blendm<mode>): Ditto.
11130 (avx512f_store<mode>_mask): Ditto.
11131 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
11132 (avx512f_storedqu<mode>_mask): Ditto.
11133 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
11134 Ditto.
11135 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
11136 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
11137 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
11138 (avx512f_maskcmp<mode>3): Ditto.
11139 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
11140 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
11141 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
11142 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
11143 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
11144 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
11145 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
11146 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
11147 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
11148 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
11149 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
11150 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
11151 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
11152 (vec_extract_lo_<mode>_maskm): Ditto.
11153 (vec_extract_hi_<mode>_maskm): Ditto.
11154 (avx512f_vternlog<mode>_mask): Ditto.
11155 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
11156 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
11157 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
11158 (avx512f_<code>v8div16qi2_mask): Ditto.
11159 (avx512f_<code>v8div16qi2_mask_store): Ditto.
11160 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
11161 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
11162 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
11163 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
11164 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
11165 (*avx512pf_gatherpf<mode>df_mask): Ditto.
11166 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
11167 (*avx512pf_scatterpf<mode>df_mask): Ditto.
11168 (avx512cd_maskb_vec_dupv8di): Ditto.
11169 (avx512cd_maskw_vec_dupv16si): Ditto.
11170 (avx512f_vpermi2var<mode>3_maskz): Ditto.
11171 (avx512f_vpermi2var<mode>3_mask): Ditto.
11172 (avx512f_vpermi2var<mode>3_mask): Ditto.
11173 (avx512f_vpermt2var<mode>3_maskz): Ditto.
11174 (*avx512f_gathersi<mode>): Ditto.
11175 (*avx512f_gathersi<mode>_2): Ditto.
11176 (*avx512f_gatherdi<mode>): Ditto.
11177 (*avx512f_gatherdi<mode>_2): Ditto.
11178 (*avx512f_scattersi<mode>): Ditto.
11179 (*avx512f_scatterdi<mode>): Ditto.
11180 (avx512f_compress<mode>_mask): Ditto.
11181 (avx512f_compressstore<mode>_mask): Ditto.
11182 (avx512f_expand<mode>_mask): Ditto.
11183 * config/i386/subst.md (mask): Change k to Yk.
11184 (mask_scalar_merge): Ditto.
11185 (sd): Ditto.
11187 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
11189 * doc/extend.texi (Vector Extensions): Document ?: in C++.
11191 2014-01-31 Richard Biener <rguenther@suse.de>
11193 PR middle-end/59990
11194 * builtins.c (fold_builtin_memory_op): Make sure to not
11195 use a floating-point mode or a boolean or enumeral type for
11196 the copy operation.
11198 2014-01-30 DJ Delorie <dj@redhat.com>
11200 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
11201 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
11202 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
11203 whenever main() has an epilogue.
11205 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11207 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
11208 unused variable "field".
11209 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
11210 (vsx_mergeh_<mode>): Likewise.
11211 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
11212 (altivec_vmrghh): Likewise.
11213 (altivec_vmrghw): Likewise.
11214 (altivec_vmrglb): Likewise.
11215 (altivec_vmrglh): Likewise.
11216 (altivec_vmrglw): Likewise.
11217 (altivec_vspltb): Add missing uses.
11218 (altivec_vsplth): Likewise.
11219 (altivec_vspltw): Likewise.
11220 (altivec_vspltsf): Likewise.
11222 2014-01-30 Jakub Jelinek <jakub@redhat.com>
11224 PR target/59923
11225 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
11226 frame related instructions.
11228 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
11230 PR rtl-optimization/59959
11231 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
11232 any reload of register whose subreg is invalid.
11234 2014-01-30 Jakub Jelinek <jakub@redhat.com>
11236 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
11237 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
11238 Add missing return type - void.
11240 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11242 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
11243 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
11244 remove element index adjustment for endian (now handled in vsx.md
11245 and altivec.md).
11246 (altivec_expand_vec_perm_const): Use
11247 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
11248 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
11249 (vsx_xxspltw_<mode>): Adjust element index for little endian.
11250 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
11251 define_expand and a new define_insn *altivec_vspltb_internal;
11252 adjust for -maltivec=be on a little endian target.
11253 (altivec_vspltb_direct): New.
11254 (altivec_vsplth): Divide into a define_expand and a new
11255 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
11256 little endian target.
11257 (altivec_vsplth_direct): New.
11258 (altivec_vspltw): Divide into a define_expand and a new
11259 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
11260 little endian target.
11261 (altivec_vspltw_direct): New.
11262 (altivec_vspltsf): Divide into a define_expand and a new
11263 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
11264 a little endian target.
11266 2014-01-30 Richard Biener <rguenther@suse.de>
11268 PR tree-optimization/59993
11269 * tree-ssa-forwprop.c (associate_pointerplus): Check we
11270 can propagate form the earlier stmt and avoid the transform
11271 when the intermediate result is needed.
11273 2014-01-30 Alangi Derick <alangiderick@gmail.com>
11275 * README.Portability: Fix typo.
11277 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
11279 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
11280 comparison_operator with ordered_comparison_operator.
11282 2014-01-30 Nick Clifton <nickc@redhat.com>
11284 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
11285 Rename to mn10300_store_multiple_regs.
11286 * config/mn10300/mn10300.c: Likewise.
11287 * config/mn10300/mn10300.md (store_movm): Fix typo: call
11288 store_multiple_regs.
11289 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
11290 Call mn10300_store_multiple_regs.
11292 2014-01-30 Nick Clifton <nickc@redhat.com>
11293 DJ Delorie <dj@redhat.com>
11295 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
11296 %fp 2 to keep registers after it properly word-aligned.
11297 (rl78_alloc_physical_registers_umul): Handle the case where both
11298 input operands are the same.
11300 2014-01-30 Richard Biener <rguenther@suse.de>
11302 PR tree-optimization/59903
11303 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
11304 check properly.
11306 2014-01-30 Jason Merrill <jason@redhat.com>
11308 PR c++/59633
11309 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
11311 PR c++/59645
11312 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
11314 2014-01-30 Richard Biener <rguenther@suse.de>
11316 PR tree-optimization/59951
11317 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
11319 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
11321 PR target/59784
11322 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
11323 SFmode to DFmode case.
11325 2014-01-29 DJ Delorie <dj@redhat.com>
11327 * config/msp430/msp430.opt (-minrt): New.
11328 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
11329 if -minrt given.
11330 (ENDFILE_SPEC): Likewise.
11332 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
11334 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
11335 (estimate_function_body_sizes): Use it.
11337 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
11339 PR c++/58561
11340 * dwarf2out.c (is_cxx_auto): New.
11341 (is_base_type): Use it.
11342 (gen_type_die_with_usage): Likewise.
11344 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11346 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
11347 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
11348 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
11349 -maltivec=be with LE targets.
11350 (vsx_mergeh_<mode>): Likewise.
11351 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
11352 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
11353 (altivec_vmrghb): Replace with define_expand and new
11354 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
11355 (altivec_vmrghb_direct): New define_insn.
11356 (altivec_vmrghh): Replace with define_expand and new
11357 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
11358 (altivec_vmrghh_direct): New define_insn.
11359 (altivec_vmrghw): Replace with define_expand and new
11360 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
11361 (altivec_vmrghw_direct): New define_insn.
11362 (*altivec_vmrghsf): Adjust for endianness.
11363 (altivec_vmrglb): Replace with define_expand and new
11364 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
11365 (altivec_vmrglb_direct): New define_insn.
11366 (altivec_vmrglh): Replace with define_expand and new
11367 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
11368 (altivec_vmrglh_direct): New define_insn.
11369 (altivec_vmrglw): Replace with define_expand and new
11370 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
11371 (altivec_vmrglw_direct): New define_insn.
11372 (*altivec_vmrglsf): Adjust for endianness.
11373 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
11374 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
11375 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
11376 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
11377 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
11378 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
11379 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
11380 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
11382 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
11384 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
11385 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
11386 whitespace.
11388 2014-01-29 Richard Biener <rguenther@suse.de>
11390 PR tree-optimization/58742
11391 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
11392 associate_pointerplus_align.
11393 (associate_pointerplus_diff): New function.
11394 (associate_pointerplus): Likewise. Call associate_pointerplus_align
11395 and associate_pointerplus_diff.
11397 2014-01-29 Richard Biener <rguenther@suse.de>
11399 * lto-streamer.h (LTO_major_version): Bump to 3.
11400 (LTO_minor_version): Reset to 0.
11402 2014-01-29 Renlin Li <Renlin.Li@arm.com>
11404 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
11405 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
11406 (arm_file_start): Generate correct asm header for armv7ve.
11407 * config/arm/bpabi.h: Add multilib support for armv7ve.
11408 * config/arm/driver-arm.c: Change the architectures of cortex-a7
11409 and cortex-a15 to armv7ve.
11410 * config/arm/t-aprofile: Add multilib support for armv7ve.
11411 * doc/invoke.texi: Document -march=armv7ve.
11413 2014-01-29 Richard Biener <rguenther@suse.de>
11415 PR tree-optimization/58742
11416 * tree-ssa-forwprop.c (associate_plusminus): Return true
11417 if we changed sth, defer EH cleanup to ...
11418 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
11419 (simplify_mult): New function.
11421 2014-01-29 Jakub Jelinek <jakub@redhat.com>
11423 PR middle-end/59917
11424 PR tree-optimization/59920
11425 * tree.c (build_common_builtin_nodes): Remove
11426 __builtin_setjmp_dispatcher initialization.
11427 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
11428 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
11429 instead of gsi_after_labels + manually skipping debug stmts.
11430 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
11431 ignore bbs with IFN_ABNORMAL_DISPATCHER.
11432 * tree-inline.c (copy_edges_for_bb): Remove
11433 can_make_abnormal_goto argument, instead add abnormal_goto_dest
11434 argument. Ignore computed_goto_p stmts. Don't call
11435 make_abnormal_goto_edges. If a call might need abnormal edges
11436 for non-local gotos, see if it already has an edge to
11437 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
11438 with true argument, don't do anything then, otherwise add
11439 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
11440 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
11441 caller.
11442 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
11443 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
11444 (lower_stmt): Don't set data->calls_builtin_setjmp.
11445 (lower_builtin_setjmp): Adjust comment.
11446 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
11447 * tree-cfg.c (found_computed_goto): Remove.
11448 (factor_computed_gotos): Remove.
11449 (make_goto_expr_edges): Return bool, true for computed gotos.
11450 Don't call make_abnormal_goto_edges.
11451 (build_gimple_cfg): Don't set found_computed_goto, don't call
11452 factor_computed_gotos.
11453 (computed_goto_p): No longer static.
11454 (make_blocks): Don't set found_computed_goto.
11455 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
11456 (make_edges): If make_goto_expr_edges returns true, push bb
11457 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
11458 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
11459 vector. Record mapping between bbs and OpenMP regions if there
11460 are any, adjust make_gimple_omp_edges caller. Call
11461 handle_abnormal_edges.
11462 (make_abnormal_goto_edges): Remove.
11463 * tree-cfg.h (make_abnormal_goto_edges): Remove.
11464 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
11465 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
11466 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
11467 * internal-fn.def (ABNORMAL_DISPATCHER): New.
11468 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
11469 filling *region also set *region_idx to (*region)->entry->index.
11471 PR other/58712
11472 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
11473 For REGs set ORIGINAL_REGNO.
11475 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
11477 * doc/md.texi: Mention that a target shouldn't implement
11478 vec_widen_(s|u)mul_even/odd pair if it is less efficient
11479 than hi/lo pair.
11481 2014-01-29 Jakub Jelinek <jakub@redhat.com>
11483 PR tree-optimization/59594
11484 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
11485 a copy of the datarefs vector rather than the vector itself.
11487 2014-01-28 Jason Merrill <jason@redhat.com>
11489 PR c++/53756
11490 * dwarf2out.c (auto_die): New static.
11491 (gen_type_die_with_usage): Handle C++1y 'auto'.
11492 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
11493 on definition.
11495 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
11497 PR target/59672
11498 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
11499 (SPEC_X32): Likewise.
11500 (SPEC_64): Likewise.
11501 * config/i386/i386.c (ix86_option_override_internal): Turn off
11502 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
11503 for TARGET_16BIT.
11504 (x86_file_start): Output .code16gcc for TARGET_16BIT.
11505 * config/i386/i386.h (TARGET_16BIT): New macro.
11506 (TARGET_16BIT_P): Likewise.
11507 * config/i386/i386.opt: Add m16.
11508 * doc/invoke.texi: Document -m16.
11510 2014-01-28 Jakub Jelinek <jakub@redhat.com>
11512 PR preprocessor/59935
11513 * input.c (location_get_source_line): Bail out on when line number
11514 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
11516 2014-01-28 Richard Biener <rguenther@suse.de>
11518 PR tree-optimization/58742
11519 * tree-ssa-forwprop.c (associate_plusminus): Handle
11520 pointer subtraction of the form (T)(P + A) - (T)P.
11522 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11524 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
11525 at const_int_cost.
11527 2014-01-28 Richard Biener <rguenther@suse.de>
11529 Revert
11530 2014-01-28 Richard Biener <rguenther@suse.de>
11532 PR rtl-optimization/45364
11533 PR rtl-optimization/59890
11534 * var-tracking.c (local_get_addr_clear_given_value): Handle
11535 already cleared slot.
11536 (val_reset): Handle not allocated local_get_addr_cache.
11537 (vt_find_locations): Use post-order on the inverted CFG.
11539 2014-01-28 Richard Biener <rguenther@suse.de>
11541 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
11543 2014-01-28 Richard Biener <rguenther@suse.de>
11545 PR rtl-optimization/45364
11546 PR rtl-optimization/59890
11547 * var-tracking.c (local_get_addr_clear_given_value): Handle
11548 already cleared slot.
11549 (val_reset): Handle not allocated local_get_addr_cache.
11550 (vt_find_locations): Use post-order on the inverted CFG.
11552 2014-01-28 Alan Modra <amodra@gmail.com>
11554 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
11555 * configure.ac <recursive call for build != host>: Define
11556 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
11557 and LD_FOR_BUILD too.
11558 * configure: Regenerate.
11560 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
11562 * config/i386/i386.c (get_builtin_code_for_version): Separate
11563 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
11564 Broadwell from Haswell.
11566 2014-01-27 Steve Ellcey <sellcey@mips.com>
11568 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
11569 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
11570 * config/mips/mips.c (mips_option_override): Change setting
11571 of TARGET_DSP.
11572 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
11573 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
11574 Change from Mask to Var.
11576 2014-01-27 Jeff Law <law@redhat.com>
11578 * ipa-inline.c (inline_small_functions): Fix typo.
11580 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
11582 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
11583 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
11584 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
11585 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
11586 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
11587 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
11588 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
11589 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
11590 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
11591 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
11592 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
11593 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
11594 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
11595 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
11596 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
11597 (_mm512_storeu_epi64): Ditto.
11598 (_mm512_cmpge_epi32_mask): Ditto.
11599 (_mm512_cmpge_epu32_mask): Ditto.
11600 (_mm512_cmpge_epi64_mask): Ditto.
11601 (_mm512_cmpge_epu64_mask): Ditto.
11602 (_mm512_cmple_epi32_mask): Ditto.
11603 (_mm512_cmple_epu32_mask): Ditto.
11604 (_mm512_cmple_epi64_mask): Ditto.
11605 (_mm512_cmple_epu64_mask): Ditto.
11606 (_mm512_cmplt_epi32_mask): Ditto.
11607 (_mm512_cmplt_epu32_mask): Ditto.
11608 (_mm512_cmplt_epi64_mask): Ditto.
11609 (_mm512_cmplt_epu64_mask): Ditto.
11610 (_mm512_cmpneq_epi32_mask): Ditto.
11611 (_mm512_cmpneq_epu32_mask): Ditto.
11612 (_mm512_cmpneq_epi64_mask): Ditto.
11613 (_mm512_cmpneq_epu64_mask): Ditto.
11614 (_mm512_expand_pd): Ditto.
11615 (_mm512_expand_ps): Ditto.
11616 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
11617 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
11618 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
11619 * config/i386/i386.c (ix86_builtins): Add
11620 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
11621 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
11622 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
11623 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
11624 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
11625 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
11626 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
11627 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
11628 IX86_BUILTIN_PMOVUSQW512_MEM.
11629 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
11630 __builtin_ia32_pmovsqd512mem_mask,
11631 __builtin_ia32_pmovqd512mem_mask,
11632 __builtin_ia32_pmovusqw512mem_mask,
11633 __builtin_ia32_pmovsqw512mem_mask,
11634 __builtin_ia32_pmovqw512mem_mask,
11635 __builtin_ia32_pmovusdw512mem_mask,
11636 __builtin_ia32_pmovsdw512mem_mask,
11637 __builtin_ia32_pmovdw512mem_mask,
11638 __builtin_ia32_pmovqb512mem_mask,
11639 __builtin_ia32_pmovusqb512mem_mask,
11640 __builtin_ia32_pmovsqb512mem_mask,
11641 __builtin_ia32_pmovusdb512mem_mask,
11642 __builtin_ia32_pmovsdb512mem_mask,
11643 __builtin_ia32_pmovdb512mem_mask.
11644 (bdesc_args): Add __builtin_ia32_expanddf512,
11645 __builtin_ia32_expandsf512.
11646 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
11647 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
11648 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
11649 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
11650 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
11651 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
11652 (avx512f_<code>v8div16qi2_mask_store): This.
11653 (avx512f_expand<mode>): New.
11655 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
11657 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
11658 New.
11659 (_mm512_mask_prefetch_i64gather_pd): Ditto.
11660 (_mm512_prefetch_i32scatter_pd): Ditto.
11661 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
11662 (_mm512_prefetch_i64scatter_pd): Ditto.
11663 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
11664 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
11665 (_mm512_mask_prefetch_i64gather_ps): Ditto.
11666 (_mm512_prefetch_i32scatter_ps): Ditto.
11667 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
11668 (_mm512_prefetch_i64scatter_ps): Ditto.
11669 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
11670 * config/i386/i386-builtin-types.def: Define
11671 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
11672 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
11673 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
11674 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
11675 IX86_BUILTIN_SCATTERPFQPD.
11676 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
11677 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
11678 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
11679 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
11680 __builtin_ia32_scatterpfqps.
11681 (ix86_expand_builtin): Expand new built-ins.
11682 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
11683 fix memory access data type.
11684 (*avx512pf_gatherpf<mode>_mask): Ditto.
11685 (*avx512pf_gatherpf<mode>): Ditto.
11686 (avx512pf_scatterpf<mode>): Ditto.
11687 (*avx512pf_scatterpf<mode>_mask): Ditto.
11688 (*avx512pf_scatterpf<mode>): Ditto.
11689 (GATHER_SCATTER_SF_MEM_MODE): New.
11690 (avx512pf_gatherpf<mode>df): Ditto.
11691 (*avx512pf_gatherpf<mode>df_mask): Ditto.
11692 (*avx512pf_scatterpf<mode>df): Ditto.
11694 2014-01-27 Jakub Jelinek <jakub@redhat.com>
11696 PR bootstrap/59934
11697 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
11698 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
11699 reached.
11701 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
11703 * common/config/arm/arm-common.c
11704 (arm_rewrite_mcpu): Handle multiple names.
11705 * config/arm/arm.h
11706 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
11708 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
11710 * gimple-builder.h (create_gimple_tmp): Delete.
11712 2014-01-27 Christian Bruel <christian.bruel@st.com>
11714 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
11715 words comparisons.
11717 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
11719 * config/pa/pa.md (call): Generate indirect long calls to non-local
11720 functions when outputing 32-bit code.
11721 (call_value): Likewise except for special call to buggy powf function.
11723 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
11724 portable runtime and PIC indirect calls.
11725 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
11726 and PIC call sequences. Use ldo instead of blr to set return register
11727 in PIC call sequence.
11729 2014-01-25 Walter Lee <walt@tilera.com>
11731 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
11732 avoid clobbering a live register.
11734 2014-01-25 Walter Lee <walt@tilera.com>
11736 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
11737 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
11738 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
11739 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
11741 2014-01-25 Walter Lee <walt@tilera.com>
11743 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
11744 arguments on even registers.
11745 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
11746 STACK_BOUNDARY.
11747 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
11748 (BIGGEST_ALIGNMENT): Ditto.
11749 (BIGGEST_FIELD_ALIGNMENT): Ditto.
11751 2014-01-25 Walter Lee <walt@tilera.com>
11753 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
11754 insns before bundling.
11755 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
11757 2014-01-25 Walter Lee <walt@tilera.com>
11759 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
11760 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
11761 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
11763 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
11765 * config/mips/constraints.md (kl): Delete.
11766 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
11767 define expands, using...
11768 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
11769 instructions for MIPS16.
11770 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
11771 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
11773 2014-01-25 Walter Lee <walt@tilera.com>
11775 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
11776 (clzdi2): Ditto.
11777 (ffsdi2): Ditto.
11779 2014-01-25 Walter Lee <walt@tilera.com>
11781 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
11782 (TARGET_EXPAND_TO_RTL_HOOK): Define.
11784 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
11786 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
11787 Handle XOR.
11789 2014-01-25 Jakub Jelinek <jakub@redhat.com>
11791 * print-rtl.c (in_call_function_usage): New var.
11792 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
11793 EXPR_LIST mode as mode and not as reg note name.
11795 PR middle-end/59561
11796 * cfgloopmanip.c (copy_loop_info): If
11797 loop->warned_aggressive_loop_optimizations, make sure
11798 the flag is set in target loop too.
11800 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
11802 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
11803 flag_cilkplus.
11804 * builtins.def: Likewise.
11805 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
11806 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
11807 * ira.c (ira_setup_eliminable_regset): Likewise.
11808 * omp-low.c (gate_expand_omp): Likewise.
11809 (execute_lower_omp): Likewise.
11810 (diagnose_sb_0): Likewise.
11811 (gate_diagnose_omp_blocks): Likewise.
11812 (simd_clone_clauses_extract): Likewise.
11813 (gate): Likewise.
11815 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11817 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
11818 correction for little endian...
11819 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
11820 here.
11822 2014-01-24 Jeff Law <law@redhat.com>
11824 PR tree-optimization/59919
11825 * tree-vrp.c (find_assert_locations_1): Do not register asserts
11826 for non-returning calls.
11828 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
11830 * common/config/aarch64/aarch64-common.c
11831 (aarch64_rewrite_mcpu): Handle multiple names.
11832 * config/aarch64/aarch64.h
11833 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
11835 2014-01-24 Dodji Seketeli <dodji@redhat.com>
11837 * input.c (add_file_to_cache_tab): Handle the case where fopen
11838 returns NULL.
11840 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
11842 PR target/59929
11843 * config/i386/i386.md (pushsf splitter): Get stack adjustment
11844 from push operand if code of push isn't PRE_DEC.
11846 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
11848 PR target/59909
11849 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
11850 -mquad-memory-atomic. Update -mquad-memory documentation to say
11851 it is only used for non-atomic loads/stores.
11853 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
11854 -mquad-memory or -mquad-memory-atomic switches.
11856 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
11857 -mquad-memory-atomic to ISA 2.07 support.
11859 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
11860 to separate support of normal quad word memory operations (ldq, stq)
11861 from the atomic quad word memory operations.
11863 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
11864 support to separate non-atomic quad word operations from atomic
11865 quad word operations. Disable non-atomic quad word operations in
11866 little endian mode so that we don't have to swap words after the
11867 load and before the store.
11868 (quad_load_store_p): Add comment about atomic quad word support.
11869 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
11870 options printed with -mdebug=reg.
11872 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
11873 -mquad-memory-atomic as the test for whether we have quad word
11874 atomic instructions.
11875 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
11876 or -mp8-vector are used, allow byte/half-word atomic operations.
11878 * config/rs6000/sync.md (load_lockedti): Insure that the address
11879 is a proper indexed or indirect address for the lqarx instruction.
11880 On little endian systems, swap the hi/lo registers after the lqarx
11881 instruction.
11882 (load_lockedpti): Use indexed_or_indirect_operand predicate to
11883 insure the address is valid for the lqarx instruction.
11884 (store_conditionalti): Insure that the address is a proper indexed
11885 or indirect address for the stqcrx. instruction. On little endian
11886 systems, swap the hi/lo registers before doing the stqcrx.
11887 instruction.
11888 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
11889 insure the address is valid for the stqcrx. instruction.
11891 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
11892 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
11893 type of quad memory support is available.
11895 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
11897 PR regression/59915
11898 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
11899 there is a danger of looping.
11901 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
11903 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
11904 force flag_ira_loop_pressure if set via command line.
11906 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
11908 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
11909 (ashr_simd): New builtin handling DI mode.
11910 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
11911 (aarch64_sshr_simddi): New match pattern.
11912 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
11913 (vshrd_n_s64): Likewise.
11914 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
11916 2014-01-23 Nick Clifton <nickc@redhat.com>
11918 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
11919 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
11920 favour of mcu specific scripts.
11921 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
11922 430x multilibs.
11924 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
11925 Alex Velenko <Alex.Velenko@arm.com>
11927 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
11928 (vaddv_s16): Likewise.
11929 (vaddv_s32): Likewise.
11930 (vaddv_u8): Likewise.
11931 (vaddv_u16): Likewise.
11932 (vaddv_u32): Likewise.
11933 (vaddvq_s8): Likewise.
11934 (vaddvq_s16): Likewise.
11935 (vaddvq_s32): Likewise.
11936 (vaddvq_s64): Likewise.
11937 (vaddvq_u8): Likewise.
11938 (vaddvq_u16): Likewise.
11939 (vaddvq_u32): Likewise.
11940 (vaddvq_u64): Likewise.
11941 (vaddv_f32): Likewise.
11942 (vaddvq_f32): Likewise.
11943 (vaddvq_f64): Likewise.
11944 (vmaxv_f32): Likewise.
11945 (vmaxv_s8): Likewise.
11946 (vmaxv_s16): Likewise.
11947 (vmaxv_s32): Likewise.
11948 (vmaxv_u8): Likewise.
11949 (vmaxv_u16): Likewise.
11950 (vmaxv_u32): Likewise.
11951 (vmaxvq_f32): Likewise.
11952 (vmaxvq_f64): Likewise.
11953 (vmaxvq_s8): Likewise.
11954 (vmaxvq_s16): Likewise.
11955 (vmaxvq_s32): Likewise.
11956 (vmaxvq_u8): Likewise.
11957 (vmaxvq_u16): Likewise.
11958 (vmaxvq_u32): Likewise.
11959 (vmaxnmv_f32): Likewise.
11960 (vmaxnmvq_f32): Likewise.
11961 (vmaxnmvq_f64): Likewise.
11962 (vminv_f32): Likewise.
11963 (vminv_s8): Likewise.
11964 (vminv_s16): Likewise.
11965 (vminv_s32): Likewise.
11966 (vminv_u8): Likewise.
11967 (vminv_u16): Likewise.
11968 (vminv_u32): Likewise.
11969 (vminvq_f32): Likewise.
11970 (vminvq_f64): Likewise.
11971 (vminvq_s8): Likewise.
11972 (vminvq_s16): Likewise.
11973 (vminvq_s32): Likewise.
11974 (vminvq_u8): Likewise.
11975 (vminvq_u16): Likewise.
11976 (vminvq_u32): Likewise.
11977 (vminnmv_f32): Likewise.
11978 (vminnmvq_f32): Likewise.
11979 (vminnmvq_f64): Likewise.
11981 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
11983 * config/aarch64/aarch64-simd.md
11984 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
11985 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
11986 (*aarch64_mul3_elt<mode>): Likewise.
11987 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
11988 (*aarch64_mul3_elt_to_64v2df): Likewise.
11989 (*aarch64_mla_elt<mode>): Likewise.
11990 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
11991 (*aarch64_mls_elt<mode>): Likewise.
11992 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
11993 (*aarch64_fma4_elt<mode>): Likewise.
11994 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
11995 (*aarch64_fma4_elt_to_64v2df): Likewise.
11996 (*aarch64_fnma4_elt<mode>): Likewise.
11997 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
11998 (*aarch64_fnma4_elt_to_64v2df): Likewise.
11999 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
12000 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
12001 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
12002 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
12003 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
12004 (aarch64_sqdmull_lane<mode>_internal): Likewise.
12005 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
12007 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
12009 * config/aarch64/aarch64-simd.md
12010 (aarch64_be_checked_get_lane<mode>): New define_expand.
12011 * config/aarch64/aarch64-simd-builtins.def
12012 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
12013 New builtin definition.
12014 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
12015 Use new safe be builtin.
12017 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
12019 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
12020 New define_insn.
12021 (aarch64_be_st1<mode>): Likewise.
12022 (aarch_ld1<VALL:mode>): Define_expand modified.
12023 (aarch_st1<VALL:mode>): Likewise.
12024 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
12025 (UNSPEC_ST1): Likewise.
12027 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
12029 * config/microblaze/microblaze.md: Add trap insn and attribute
12031 2014-01-23 Dodji Seketeli <dodji@redhat.com>
12033 PR preprocessor/58580
12034 * input.h (location_get_source_line): Take an additional line_size
12035 parameter.
12036 (void diagnostics_file_cache_fini): Declare new function.
12037 * input.c (struct fcache): New type.
12038 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
12039 New static constants.
12040 (diagnostic_file_cache_init, total_lines_num)
12041 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
12042 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
12043 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
12044 (get_next_line, read_next_line, goto_next_line, read_line_num):
12045 New static function definitions.
12046 (diagnostic_file_cache_fini): New function.
12047 (location_get_source_line): Take an additional output line_len
12048 parameter. Re-write using lookup_or_add_file_to_cache_tab and
12049 read_line_num.
12050 * diagnostic.c (diagnostic_finish): Call
12051 diagnostic_file_cache_fini.
12052 (adjust_line): Take an additional input parameter for the length
12053 of the line, rather than calculating it with strlen.
12054 (diagnostic_show_locus): Adjust the use of
12055 location_get_source_line and adjust_line with respect to their new
12056 signature. While displaying a line now, do not stop at the first
12057 null byte. Rather, display the zero byte as a space and keep
12058 going until we reach the size of the line.
12059 * Makefile.in: Add vec.o to OBJS-libcommon
12061 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
12062 Ilya Tocar <ilya.tocar@intel.com>
12064 * config/i386/avx512fintrin.h (_mm512_kmov): New.
12065 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
12066 (__builtin_ia32_kmov16): Ditto.
12067 * config/i386/i386.md (UNSPEC_KMOV): New.
12068 (kmovw): Ditto.
12070 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
12072 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
12073 (_mm512_storeu_si512): Ditto.
12075 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
12077 PR target/52125
12078 * rtl.h (get_referenced_operands): Declare.
12079 * recog.c (get_referenced_operands): New function.
12080 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
12081 operands have been referenced when recording LO_SUM references.
12083 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
12085 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
12087 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
12089 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
12090 Enable for generic and recent AMD targets.
12092 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
12094 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
12095 ARG_SIZE note when adjustment was eliminated.
12097 2014-01-22 Jeff Law <law@redhat.com>
12099 PR tree-optimization/59597
12100 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
12101 in file. Accept new argument REGISTERING and use it to modify
12102 dump output appropriately.
12103 (register_jump_thread): Corresponding changes.
12104 (mark_threaded_blocks): Reinstate code to cancel unprofitable
12105 thread paths involving joiner blocks. Add code to dump cancelled
12106 jump threading paths.
12108 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
12110 PR rtl-optimization/59477
12111 * lra-constraints.c (inherit_in_ebb): Process call for living hard
12112 regs. Update reloads_num and potential_reload_hard_regs for all insns.
12114 2014-01-22 Tom Tromey <tromey@redhat.com>
12116 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
12117 PARAMS.
12118 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
12120 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
12122 PR rtl-optimization/59896
12123 * lra-constraints.c (process_alt_operands): Check unused note for
12124 matched operands of insn with no output reloads.
12126 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
12128 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
12129 (mips_move_from_gpr_cost): Likewise.
12131 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
12133 PR rtl-optimization/59858
12134 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
12135 ira_class_hard_regs_num.
12136 (process_alt_operands): Increase reject for dying matched operand.
12138 2014-01-21 Jakub Jelinek <jakub@redhat.com>
12140 PR target/59003
12141 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
12142 smaller than size, perform several stores or loads and stores
12143 at dst + count - size to store or copy all of size bytes, rather
12144 than just last modesize bytes.
12146 2014-01-20 DJ Delorie <dj@redhat.com>
12148 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
12149 that CLOBBERs are REGs before propogating their values.
12151 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
12153 PR middle-end/59789
12154 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
12155 (cgraph_inline_failed_type): New function.
12156 * cgraph.h (DEFCIFCODE): Add type.
12157 (cgraph_inline_failed_type_t): New enum.
12158 (cgraph_inline_failed_type): New prototype.
12159 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
12160 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
12161 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
12162 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
12163 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
12164 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
12165 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
12166 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
12167 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
12168 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
12169 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
12170 OPTIMIZATION_MISMATCH.
12171 * tree-inline.c (expand_call_inline): Emit errors during
12172 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
12174 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
12176 PR target/59685
12177 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
12178 mode attribute in insn output.
12180 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
12182 * output.h (output_constant): Delete.
12183 * varasm.c (output_constant): Make private.
12185 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
12187 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
12189 2014-01-20 Jakub Jelinek <jakub@redhat.com>
12191 PR middle-end/59860
12192 * tree.h (fold_builtin_strcat): New prototype.
12193 * builtins.c (fold_builtin_strcat): No longer static. Add len
12194 argument, if non-NULL, don't call c_strlen. Optimize
12195 directly into __builtin_memcpy instead of __builtin_strcpy.
12196 (fold_builtin_2): Adjust fold_builtin_strcat caller.
12197 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
12199 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
12201 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
12202 for SImode_address_operand operands, having only a REG argument.
12204 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
12206 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
12207 loader name using mbig-endian.
12208 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
12210 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
12212 * doc/invoke.texi (-march): Clarify documentation for AArch64.
12213 (-mtune): Likewise.
12214 (-mcpu): Likewise.
12216 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
12218 * config/aarch64/aarch64-protos.h
12219 (aarch64_cannot_change_mode_class_ptr): Declare.
12220 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
12221 aarch64_cannot_change_mode_class_ptr): New.
12222 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
12223 backend hook aarch64_cannot_change_mode_class.
12225 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
12227 * common/config/aarch64/aarch64-common.c
12228 (aarch64_handle_option): Don't handle any option order logic here.
12229 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
12230 selected_cpu, warn on architecture version mismatch.
12231 (aarch64_override_options): Fix parsing order for option strings.
12233 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12234 Iain Sandoe <iain@codesourcery.com>
12236 PR bootstrap/59496
12237 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
12238 warning. Amend comment to reflect current functionality.
12240 2014-01-20 Richard Biener <rguenther@suse.de>
12242 PR middle-end/59860
12243 * builtins.c (fold_builtin_strcat): Remove case better handled
12244 by tree-ssa-strlen.c.
12246 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
12248 * config/aarch64/aarch64.opt
12249 (mcpu, march, mtune): Make case-insensitive.
12251 2014-01-20 Jakub Jelinek <jakub@redhat.com>
12253 PR target/59880
12254 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
12255 if operands[1] is a REG or ZERO_EXTEND of a REG.
12257 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
12259 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
12261 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
12263 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
12264 long non-pic millicode calls.
12266 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12268 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
12270 2014-01-19 Kito Cheng <kito@0xlab.org>
12272 * builtins.c (expand_movstr): Check movstr expand done or fail.
12274 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
12275 H.J. Lu <hongjiu.lu@intel.com>
12277 PR target/59379
12278 * config/i386/i386.md (*lea<mode>): Zero-extend return register
12279 to DImode for zero-extended addresses.
12281 2014-01-19 Jakub Jelinek <jakub@redhat.com>
12283 PR rtl-optimization/57763
12284 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
12285 on the new indirect jump_insn and increment LABEL_NUSES (label).
12287 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
12289 PR bootstrap/59580
12290 PR bootstrap/59583
12291 * config.gcc (x86_archs): New variable.
12292 (x86_64_archs): Likewise.
12293 (x86_cpus): Likewise.
12294 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
12295 --with-arch/--with-cpu= options.
12296 Support --with-arch=/--with-cpu={nehalem,westmere,
12297 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
12299 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
12301 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
12302 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
12304 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
12306 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
12308 2014-01-18 Jakub Jelinek <jakub@redhat.com>
12310 PR target/58944
12311 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
12312 clear cpp_get_options (parse_in)->warn_unused_macros for
12313 ix86_target_macros_internal with cpp_define.
12315 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
12317 * jump.c (delete_related_insns): Keep (use (insn))s.
12318 * reorg.c (redundant_insn): Check for barriers too.
12320 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
12322 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
12324 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
12326 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
12327 call to $$dyncall when TARGET_LONG_CALLS is true.
12329 2014-01-17 Jeff Law <law@redhat.com>
12331 * ree.c (combine_set_extension): Temporarily disable test for
12332 changing number of hard registers.
12334 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
12336 PR middle-end/58125
12337 * ipa-inline-analysis.c (inline_free_summary):
12338 Do not free summary of aliases.
12340 2014-01-17 Jakub Jelinek <jakub@redhat.com>
12342 PR middle-end/59706
12343 * gimplify.c (gimplify_expr): Use create_tmp_var
12344 instead of create_tmp_var_raw. If cond doesn't have
12345 integral type, don't add the IFN_ANNOTATE builtin at all.
12347 2014-01-17 Martin Jambor <mjambor@suse.cz>
12349 PR ipa/59736
12350 * ipa-cp.c (prev_edge_clone): New variable.
12351 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
12352 Also resize prev_edge_clone vector.
12353 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
12354 (ipcp_edge_removal_hook): New function.
12355 (ipcp_driver): Register ipcp_edge_removal_hook.
12357 2014-01-17 Andrew Pinski <apinski@cavium.com>
12358 Steve Ellcey <sellcey@mips.com>
12360 PR target/59462
12361 * config/mips/mips.c (mips_print_operand): Check operand mode instead
12362 of operator mode.
12364 2014-01-17 Jeff Law <law@redhat.com>
12366 PR middle-end/57904
12367 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
12368 so that pass_ccp runs first.
12370 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
12372 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
12373 (ix86_adjust_cost): Use !TARGET_XXX.
12374 (do_reorder_for_imul): Likewise.
12375 (swap_top_of_ready_list): Likewise.
12376 (ix86_sched_reorder): Likewise.
12378 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
12380 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
12381 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
12382 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
12383 (intel_memset): New. Duplicate slm_memset.
12384 (intel_cost): New. Duplicate slm_cost.
12385 (m_INTEL): New macro.
12386 (processor_target_table): Add "intel".
12387 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
12388 with PROCESSOR_INTEL for "intel".
12389 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
12390 PROCESSOR_SILVERMONT.
12391 (ix86_issue_rate): Likewise.
12392 (ix86_adjust_cost): Likewise.
12393 (ia32_multipass_dfa_lookahead): Likewise.
12394 (swap_top_of_ready_list): Likewise.
12395 (ix86_sched_reorder): Likewise.
12396 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
12397 instead of TARGET_OPT_AGU.
12398 * config/i386/i386.h (TARGET_INTEL): New.
12399 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
12400 (processor_type): Add PROCESSOR_INTEL.
12401 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
12402 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
12404 2014-01-17 Marek Polacek <polacek@redhat.com>
12406 PR c/58346
12407 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
12408 size is zero.
12410 2014-01-17 Richard Biener <rguenther@suse.de>
12412 PR tree-optimization/46590
12413 * opts.c (default_options_table): Add entries for
12414 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
12415 all enabled at -O1 but not for -Og.
12416 * common.opt (fbranch-count-reg): Remove Init(1).
12417 (fmove-loop-invariants): Likewise.
12418 (ftree-pta): Likewise.
12420 2014-01-17 Jakub Jelinek <jakub@redhat.com>
12422 * config/i386/i386.c (ix86_data_alignment): For compatibility with
12423 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
12424 decls to at least the GCC 4.8 used alignments.
12426 PR fortran/59440
12427 * tree-nested.c (convert_nonlocal_reference_stmt,
12428 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
12429 of GIMPLE_BIND stmts, adjust associated decls.
12431 2014-01-17 Richard Biener <rguenther@suse.de>
12433 PR tree-optimization/46590
12434 * vec.h (vec<>::bseach): New member function implementing
12435 binary search according to C89 bsearch.
12436 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
12437 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
12438 bitmap pointer again. Make accesses_in_loop a flat array.
12439 (mem_ref_obstack): New global.
12440 (outermost_indep_loop): Adjust for mem_ref->stored changes.
12441 (mark_ref_stored): Likewise.
12442 (ref_indep_loop_p_2): Likewise.
12443 (set_ref_stored_in_loop): New helper function.
12444 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
12445 (memref_free): Adjust.
12446 (record_mem_ref_loc): Simplify.
12447 (gather_mem_refs_stmt): Adjust.
12448 (sort_locs_in_loop_postorder_cmp): New function.
12449 (analyze_memory_references): Sort accesses_in_loop after
12450 loop postorder number.
12451 (find_ref_loc_in_loop_cmp): New function.
12452 (for_all_locs_in_loop): Find relevant cluster of locs in
12453 accesses_in_loop and iterate without recursion.
12454 (execute_sm): Avoid uninit warning.
12455 (struct ref_always_accessed): Simplify.
12456 (ref_always_accessed::operator ()): Likewise.
12457 (ref_always_accessed_p): Likewise.
12458 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
12459 loop postorder numbers here.
12460 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
12461 numbers.
12463 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
12465 PR c++/57945
12466 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
12467 on decls for which assemble_alias has been called.
12469 2014-01-17 Nick Clifton <nickc@redhat.com>
12471 * config/msp430/msp430.opt: (mcpu): New option.
12472 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
12473 (msp430_option_override): Parse target_cpu. If the MCU name
12474 matches a generic string, clear target_mcu.
12475 (msp430_attr): Allow numeric interrupt values up to 63.
12476 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
12477 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
12478 option.
12479 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
12480 Add mcpu matches.
12481 * config/msp430/msp430.md (popm): Use %J rather than %I.
12482 (addsi3): Use msp430_nonimmediate_operand for operand 2.
12483 (addhi_cy_i): Use immediate_operand for operand 2.
12484 * doc/invoke.texi: Document -mcpu option.
12486 2014-01-17 Richard Biener <rguenther@suse.de>
12488 PR rtl-optimization/38518
12489 * df.h (df_analyze_loop): Declare.
12490 * df-core.c: Include cfgloop.h.
12491 (df_analyze_1): Split out main part of df_analyze.
12492 (df_analyze): Adjust.
12493 (loop_inverted_post_order_compute): New function.
12494 (loop_post_order_compute): Likewise.
12495 (df_analyze_loop): New function avoiding whole-function
12496 postorder computes.
12497 * loop-invariant.c (find_defs): Use df_analyze_loop.
12498 (find_invariants): Adjust.
12499 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
12501 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
12503 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
12504 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
12506 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
12508 * ipa-ref.c (ipa_remove_stmt_references): Fix references
12509 traversal when removing references.
12511 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
12513 PR ipa/59775
12514 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
12516 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
12518 PR middle-end/56791
12519 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
12520 pushing a reload for an autoinc when we had previously reloaded an
12521 inner part of the address.
12523 2014-01-16 Jakub Jelinek <jakub@redhat.com>
12525 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
12526 field.
12527 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
12528 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
12529 when not giving up or versioning for alias only because of
12530 loop->safelen.
12531 (vect_analyze_data_ref_dependences): Set to true.
12532 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
12533 is a GIMPLE_PHI.
12534 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
12535 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
12536 to the condition.
12538 PR middle-end/58344
12539 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
12541 PR target/59839
12542 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
12543 operand 0 predicate for gathers, use a new pseudo as subtarget.
12545 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
12547 PR middle-end/59609
12548 * lra-constraints.c (process_alt_operands): Add printing debug info.
12549 Check absence of input/output reloads for matched operands too.
12551 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
12553 PR rtl-optimization/59835
12554 * ira.c (ira_init_register_move_cost): Increase cost for
12555 impossible modes.
12557 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
12559 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
12561 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
12563 PR target/59780
12564 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
12565 non-register objects. Use gen_(high/low)part more consistently.
12566 Fix assertions.
12568 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
12570 PR target/59844
12571 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
12572 endian support, remove tests for WORDS_BIG_ENDIAN.
12573 (p8_mfvsrd_3_<mode>): Likewise.
12574 (reload_gpr_from_vsx<mode>): Likewise.
12575 (reload_gpr_from_vsxsf): Likewise.
12576 (p8_mfvsrd_4_disf): Likewise.
12578 2014-01-16 Richard Biener <rguenther@suse.de>
12580 PR rtl-optimization/46590
12581 * lcm.c (compute_antinout_edge): Use postorder iteration.
12582 (compute_laterin): Use inverted postorder iteration.
12584 2014-01-16 Nick Clifton <nickc@redhat.com>
12586 PR middle-end/28865
12587 * varasm.c (output_constant): Return the number of bytes actually
12588 emitted.
12589 (output_constructor_array_range): Update the field size with the
12590 number of bytes emitted by output_constant.
12591 (output_constructor_regular_field): Likewise. Also do not
12592 complain if the total number of bytes emitted is now greater
12593 than the expected fieldpos.
12594 * output.h (output_constant): Update prototype and descriptive comment.
12596 2014-01-16 Marek Polacek <polacek@redhat.com>
12598 PR middle-end/59827
12599 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
12600 it is error_mark_node.
12602 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
12604 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
12605 VALID_AVX256_REG_OR_OI_MODE.
12607 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
12609 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
12610 current procedure should be profiled.
12612 2014-01-15 Andrew Pinski <apinski@cavium.com>
12614 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
12615 of moving from/to the STACK_REG register class.
12617 2014-01-15 Richard Henderson <rth@redhat.com>
12619 PR debug/54694
12620 * reginfo.c (global_regs_decl): Globalize.
12621 * rtl.h (global_regs_decl): Declare.
12622 * ira.c (do_reload): Diagnose frame_pointer_needed and it
12623 reserved via global_regs.
12625 2014-01-15 Teresa Johnson <tejohnson@google.com>
12627 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
12629 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12631 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
12632 and vmulosh rather than call gen_vec_widen_smult_*.
12633 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
12634 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
12635 (vec_widen_smult_even_v16qi): Likewise.
12636 (vec_widen_umult_even_v8hi): Likewise.
12637 (vec_widen_smult_even_v8hi): Likewise.
12638 (vec_widen_umult_odd_v16qi): Likewise.
12639 (vec_widen_smult_odd_v16qi): Likewise.
12640 (vec_widen_umult_odd_v8hi): Likewise.
12641 (vec_widen_smult_odd_v8hi): Likewise.
12642 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
12643 vmuloub rather than call gen_vec_widen_umult_*.
12644 (vec_widen_umult_lo_v16qi): Likewise.
12645 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
12646 vmulosb rather than call gen_vec_widen_smult_*.
12647 (vec_widen_smult_lo_v16qi): Likewise.
12648 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
12649 rather than call gen_vec_widen_umult_*.
12650 (vec_widen_umult_lo_v8hi): Likewise.
12651 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
12652 rather than call gen_vec_widen_smult_*.
12653 (vec_widen_smult_lo_v8hi): Likewise.
12655 2014-01-15 Jeff Law <law@redhat.com>
12657 PR tree-optimization/59747
12658 * ree.c (find_and_remove_re): Properly handle case where a second
12659 eliminated extension requires widening a copy created for elimination
12660 of a prior extension.
12661 (combine_set_extension): Ensure that the number of hard regs needed
12662 for a destination register does not change when we widen it.
12664 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
12666 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
12667 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
12668 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
12669 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
12670 (avr-*-rtems*): Likewise.
12671 (bfin*-rtems*): Likewise.
12672 (moxie-*-rtems*): Likewise.
12673 (h8300-*-rtems*): Likewise.
12674 (i[34567]86-*-rtems*): Likewise.
12675 (lm32-*-rtems*): Likewise.
12676 (m32r-*-rtems*): Likewise.
12677 (m68k-*-rtems*): Likewise.
12678 (microblaze*-*-rtems*): Likewise.
12679 (mips*-*-rtems*): Likewise.
12680 (powerpc-*-rtems*): Likewise.
12681 (sh-*-rtems*): Likewise.
12682 (sparc-*-rtems*): Likewise.
12683 (sparc64-*-rtems*): Likewise.
12684 (v850-*-rtems*): Likewise.
12685 (m32c-*-rtems*): Likewise.
12687 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
12689 PR rtl-optimization/59511
12690 * ira.c (ira_init_register_move_cost): Use memory costs for some
12691 cases of register move cost calculations.
12692 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
12693 instead of BB frequency.
12694 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
12695 * lra-assigns.c (find_hard_regno_for): Ditto.
12697 2014-01-15 Richard Biener <rguenther@suse.de>
12699 PR tree-optimization/59822
12700 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
12701 (vectorizable_load): Use it to hoist defs of uses of invariant
12702 loads out of the loop.
12704 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
12705 Kugan Vivekanandarajah <kuganv@linaro.org>
12707 PR target/59695
12708 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
12709 truncation.
12711 2014-01-15 Richard Biener <rguenther@suse.de>
12713 PR rtl-optimization/59802
12714 * lcm.c (compute_available): Use inverted postorder to seed
12715 the initial worklist.
12717 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12719 PR target/59803
12720 * config/s390/s390.c (s390_preferred_reload_class): Don't return
12721 ADDR_REGS for invalid symrefs in non-PIC code.
12723 2014-01-15 Jakub Jelinek <jakub@redhat.com>
12725 PR other/58712
12726 * builtins.c (determine_block_size): Initialize *probable_max_size
12727 even if len_rtx is CONST_INT.
12729 2014-01-14 Andrew Pinski <apinski@cavium.com>
12731 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
12732 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
12733 (cortexa53_tunings): Likewise.
12734 (aarch64_sched_issue_rate): New function.
12735 (TARGET_SCHED_ISSUE_RATE): Define.
12737 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
12739 * ira-costs.c (find_costs_and_classes): Add missed
12740 ira_init_register_move_cost_if_necessary.
12742 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
12744 PR target/59787
12745 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
12747 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
12749 PR target/59794
12750 * config/i386/i386.c (type_natural_mode): Add a bool parameter
12751 to indicate if type is used for function return value. Warn ABI
12752 change if the vector mode isn't available for function return value.
12753 (ix86_function_arg_advance): Pass false to type_natural_mode.
12754 (ix86_function_arg): Likewise.
12755 (ix86_gimplify_va_arg): Likewise.
12756 (function_arg_32): Don't warn ABI change.
12757 (ix86_function_value): Pass true to type_natural_mode.
12758 (ix86_return_in_memory): Likewise.
12759 (ix86_struct_value_rtx): Removed.
12760 (TARGET_STRUCT_VALUE_RTX): Likewise.
12762 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
12764 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
12765 converting a conditional jump into a conditional return.
12767 2014-01-14 Richard Biener <rguenther@suse.de>
12769 PR tree-optimization/58921
12770 PR tree-optimization/59006
12771 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
12772 hoisting invariant stmts.
12773 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
12774 invariant loads on the preheader edge if possible.
12776 2014-01-14 Joey Ye <joey.ye@arm.com>
12778 * doc/plugin.texi (Building GCC plugins): Update to C++.
12780 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
12782 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
12783 (_mm_rcp28_round_ss): Ditto.
12784 (_mm_rsqrt28_round_sd): Ditto.
12785 (_mm_rsqrt28_round_ss): Ditto.
12786 (_mm_rcp28_sd): Ditto.
12787 (_mm_rcp28_ss): Ditto.
12788 (_mm_rsqrt28_sd): Ditto.
12789 (_mm_rsqrt28_ss): Ditto.
12790 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
12791 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
12792 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
12793 (IX86_BUILTIN_RCP28SD): Ditto.
12794 (IX86_BUILTIN_RCP28SS): Ditto.
12795 (IX86_BUILTIN_RSQRT28SD): Ditto.
12796 (IX86_BUILTIN_RSQRT28SS): Ditto.
12797 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
12798 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
12799 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
12800 (ix86_expand_special_args_builtin): Expand new FTYPE.
12801 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
12802 (srcp14<mode>): Make insn unary.
12803 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
12804 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
12805 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
12806 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
12807 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
12808 Fix rounding: make it SAE only.
12809 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
12810 Ditto.
12811 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
12812 Ditto.
12813 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
12814 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
12815 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
12816 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
12817 (round_saeonly_mask_scalar_operand4): Ditto.
12818 (round_saeonly_mask_scalar_op3): Ditto.
12819 (round_saeonly_mask_scalar_op4): Ditto.
12821 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12823 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
12824 Implement -maltivec=be for vec_insert and vec_extract.
12826 2014-01-10 DJ Delorie <dj@redhat.com>
12828 * config/msp430/msp430.md (call_internal): Don't allow memory
12829 references with SP as the base register.
12830 (call_value_internal): Likewise.
12831 * config/msp430/constraints.md (Yc): New. For memory references
12832 that don't use SP as a base register.
12834 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
12835 "an integer without a # prefix"
12836 * config/msp430/msp430.md (epilogue_helper): Use it.
12838 2014-01-13 Jakub Jelinek <jakub@redhat.com>
12840 PR target/59617
12841 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
12842 AVX512F gather builtins.
12843 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
12844 on gather decls with INTEGER_TYPE masktype.
12845 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
12846 directly into the builtin rather than hoisting it before loop.
12848 PR tree-optimization/59387
12849 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
12850 (scev_const_prop): If folded_casts and type has undefined overflow,
12851 use force_gimple_operand instead of force_gimple_operand_gsi and
12852 for each added stmt if it is assign with
12853 arith_code_with_undefined_signed_overflow, call
12854 rewrite_to_defined_overflow.
12855 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
12856 gimple-fold.h instead.
12857 (arith_code_with_undefined_signed_overflow,
12858 rewrite_to_defined_overflow): Moved to ...
12859 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
12860 rewrite_to_defined_overflow): ... here. No longer static.
12861 Include gimplify-me.h.
12862 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
12863 rewrite_to_defined_overflow): New prototypes.
12865 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12867 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
12869 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
12871 * builtins.c (get_object_alignment_2): Minor tweak.
12872 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
12874 2014-01-13 Christian Bruel <christian.bruel@st.com>
12876 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
12877 optimized non constant lengths.
12879 2014-01-13 Jakub Jelinek <jakub@redhat.com>
12881 PR libgomp/59194
12882 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
12883 load as __atomic_load_N if possible.
12885 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
12887 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
12888 target parameter.
12889 (rs6000_expand_builtin): Adjust call.
12891 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
12893 PR target/58115
12894 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
12895 * config/rs6000/rs6000.c: Include target-globals.h.
12896 (rs6000_set_current_function): Instead of doing target_reinit
12897 unconditionally, use save_target_globals_default_opts and
12898 restore_target_globals.
12900 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
12901 FPSCR.
12902 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
12903 (rs6000_expand_builtin): Handle mffs and mtfsf.
12904 (rs6000_init_builtins): Define mffs and mtfsf.
12905 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
12906 (rs6000_mffs): New pattern.
12907 (rs6000_mtfsf): New pattern.
12909 2014-01-11 Bin Cheng <bin.cheng@arm.com>
12911 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
12912 Start narrowing with START. Apply candidate-use pair
12913 and check overall cost in narrowing.
12914 (iv_ca_prune): Pass new argument.
12916 2014-01-10 Jeff Law <law@redhat.com>
12918 PR middle-end/59743
12919 * ree.c (combine_reaching_defs): Ensure the defining statement
12920 occurs before the extension when optimizing extensions with
12921 different source and destination hard registers.
12923 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
12925 PR ipa/58585
12926 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
12927 vtables into the type inheritance graph.
12929 2014-01-10 Jakub Jelinek <jakub@redhat.com>
12931 PR rtl-optimization/59754
12932 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
12933 modes in the REGNO != REGNO case.
12935 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12937 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
12939 2014-01-10 Jakub Jelinek <jakub@redhat.com>
12941 PR tree-optimization/59745
12942 * tree-predcom.c (tree_predictive_commoning_loop): Call
12943 free_affine_expand_cache if giving up because components is NULL.
12945 * target-globals.c (save_target_globals): Allocate < 4KB structs using
12946 GC in payload of target_globals struct instead of allocating them on
12947 the heap and the larger structs separately using GC.
12948 * target-globals.h (struct target_globals): Make regs, hard_regs,
12949 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
12950 of GTY((skip)) and change type to void *.
12951 (reset_target_globals): Cast loads from those fields to corresponding
12952 types.
12954 2014-01-10 Steve Ellcey <sellcey@mips.com>
12956 PR plugins/59335
12957 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
12958 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
12959 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
12961 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
12963 PR target/59744
12964 * aarch64-modes.def (CC_Zmode): New flags mode.
12965 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
12966 represents an equality.
12967 (aarch64_get_condition_code): Handle CC_Zmode.
12968 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
12970 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12972 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
12973 extraction in good case.
12975 2014-01-10 Richard Biener <rguenther@suse.de>
12977 PR tree-optimization/59374
12978 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
12979 checking after SLP discovery. Mark stmts not participating
12980 in any SLP instance properly.
12982 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12984 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
12985 when handling a SET rtx.
12987 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12989 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
12990 (cortex-a57): Likewise.
12991 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
12993 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12995 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
12996 non-iwmmxt builtins.
12998 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
13000 PR ipa/58252
13001 PR ipa/59226
13002 * ipa-devirt.c record_target_from_binfo): Take as argument
13003 stack of binfos and lookup matching one for virtual inheritance.
13004 (possible_polymorphic_call_targets_1): Update.
13006 2014-01-10 Huacai Chen <chenhc@lemote.com>
13008 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
13009 kernel strings for Loongson-2E/2F/3A.
13011 2014-01-10 Jakub Jelinek <jakub@redhat.com>
13013 PR middle-end/59670
13014 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
13015 is_gimple_call before calling gimple_call_internal_p.
13017 2014-01-09 Steve Ellcey <sellcey@mips.com>
13019 * Makefile.in (TREE_FLOW_H): Remove.
13020 (TREE_SSA_H): Add file names from tree-flow.h.
13021 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
13022 * tree.h: Remove tree-flow.h reference.
13023 * hash-table.h: Remove tree-flow.h reference.
13024 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
13025 reference with tree-ssa-loop.h.
13027 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13029 * doc/invoke.texi: Add -maltivec={be,le} options, and document
13030 default element-order behavior for -maltivec.
13031 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
13032 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
13033 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
13034 when targeting big endian, at least for now.
13035 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
13037 2014-01-09 Jakub Jelinek <jakub@redhat.com>
13039 PR middle-end/47735
13040 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
13041 var satisfies use_register_for_decl, just take into account type
13042 alignment, rather than decl alignment.
13044 PR tree-optimization/59622
13045 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
13046 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
13047 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
13048 Don't devirtualize for inplace at all. For targets.length () == 1,
13049 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
13051 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
13053 * config/i386/i386.md (cpu): Remove the unused btver1.
13055 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
13057 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
13059 2014-01-09 Jakub Jelinek <jakub@redhat.com>
13061 PR target/58115
13062 * tree-core.h (struct target_globals): New forward declaration.
13063 (struct tree_target_option): Add globals field.
13064 * tree.h (TREE_TARGET_GLOBALS): Define.
13065 (prepare_target_option_nodes_for_pch): New prototype.
13066 * target-globals.h (struct target_globals): Define even if
13067 !SWITCHABLE_TARGET.
13068 * tree.c (prepare_target_option_node_for_pch,
13069 prepare_target_option_nodes_for_pch): New functions.
13070 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
13071 * config/i386/i386.c: Include target-globals.h.
13072 (ix86_set_current_function): Instead of doing target_reinit
13073 unconditionally, use save_target_globals_default_opts and
13074 restore_target_globals.
13076 2014-01-09 Richard Biener <rguenther@suse.de>
13078 PR tree-optimization/59715
13079 * tree-cfg.h (split_critical_edges): Declare.
13080 * tree-cfg.c (split_critical_edges): Export.
13081 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
13083 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
13085 * cfgexpand.c (expand_stack_vars): Optionally disable
13086 asan stack protection.
13087 (expand_used_vars): Likewise.
13088 (partition_stack_vars): Likewise.
13089 * asan.c (asan_emit_stack_protection): Optionally disable
13090 after return stack usage.
13091 (instrument_derefs): Optionally disable memory access instrumentation.
13092 (instrument_builtin_call): Likewise.
13093 (instrument_strlen_call): Likewise.
13094 (asan_protect_global): Optionally disable global variables protection.
13095 * doc/invoke.texi: Added doc for new options.
13096 * params.def: Added new options.
13097 * params.h: Likewise.
13099 2014-01-09 Jakub Jelinek <jakub@redhat.com>
13101 PR rtl-optimization/59724
13102 * ifcvt.c (cond_exec_process_if_block): Don't call
13103 flow_find_head_matching_sequence with 0 longest_match.
13104 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
13105 non-active insns if !stop_after.
13106 (try_head_merge_bb): Revert 2014-01-07 changes.
13108 2014-01-08 Jeff Law <law@redhat.com>
13110 * ree.c (get_sub_rtx): New function, extracted from...
13111 (merge_def_and_ext): Here.
13112 (combine_reaching_defs): Use get_sub_rtx.
13114 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
13116 * cgraph.h (varpool_variable_node): Do not choke on null node.
13118 2014-01-08 Catherine Moore <clm@codesourcery.com>
13120 * config/mips/mips.md (simple_return): Attempt to use JRC
13121 for microMIPS.
13122 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
13124 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
13126 PR rtl-optimization/59137
13127 * reorg.c (steal_delay_list_from_target): Call update_block for
13128 elided insns.
13129 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
13131 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13133 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
13134 two duplicate entries.
13136 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
13138 Revert:
13139 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
13141 * config/mips/mips.c (mips_truncated_op_cost): New function.
13142 (mips_rtx_costs): Adjust test for BADDU.
13143 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
13145 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
13147 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
13148 (*baddu_si): ...this new pattern.
13150 2014-01-08 Jakub Jelinek <jakub@redhat.com>
13152 PR ipa/59722
13153 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
13155 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
13157 PR middle-end/57748
13158 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
13159 inner_reference_p.
13160 (expand_expr, expand_normal): Adjust.
13161 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
13162 inner_reference_p. Use inner_reference_p to expand inner references.
13163 (store_expr): Adjust.
13164 * cfgexpand.c (expand_call_stmt): Adjust.
13166 2014-01-08 Rong Xu <xur@google.com>
13168 * gcov-io.c (gcov_var): Move from gcov-io.h.
13169 (gcov_position): Ditto.
13170 (gcov_is_error): Ditto.
13171 (gcov_rewrite): Ditto.
13172 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
13173 only part to libgcc/libgcov.h.
13175 2014-01-08 Marek Polacek <polacek@redhat.com>
13177 PR middle-end/59669
13178 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
13180 2014-01-08 Marek Polacek <polacek@redhat.com>
13182 PR sanitizer/59667
13183 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
13185 2014-01-08 Jakub Jelinek <jakub@redhat.com>
13187 PR rtl-optimization/59649
13188 * stor-layout.c (get_mode_bounds): For BImode return
13189 0 and STORE_FLAG_VALUE.
13191 2014-01-08 Richard Biener <rguenther@suse.de>
13193 PR middle-end/59630
13194 * gimple.h (is_gimple_builtin_call): Remove.
13195 (gimple_builtin_call_types_compatible_p): New.
13196 (gimple_call_builtin_p): New overload.
13197 * gimple.c (is_gimple_builtin_call): Remove.
13198 (validate_call): Rename to ...
13199 (gimple_builtin_call_types_compatible_p): ... this and export. Also
13200 check return types.
13201 (validate_type): New static function.
13202 (gimple_call_builtin_p): New overload and adjust.
13203 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
13204 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
13205 (gimple_fold_stmt_to_constant_1): Likewise.
13206 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
13208 2014-01-08 Richard Biener <rguenther@suse.de>
13210 PR middle-end/59471
13211 * gimplify.c (gimplify_expr): Gimplify register-register type
13212 VIEW_CONVERT_EXPRs to separate stmts.
13214 2014-01-07 Jeff Law <law@redhat.com>
13216 PR middle-end/53623
13217 * ree.c (combine_set_extension): Handle case where source
13218 and destination registers in an extension insn are different.
13219 (combine_reaching_defs): Allow source and destination registers
13220 in extension to be different under limited circumstances.
13221 (add_removable_extension): Remove restriction that the
13222 source and destination registers in the extension are the same.
13223 (find_and_remove_re): Emit a copy from the extension's
13224 destination to its source after the defining insn if
13225 the source and destination registers are different.
13227 PR middle-end/59285
13228 * ifcvt.c (merge_if_block): If we are merging a block with more than
13229 one successor with a block with no successors, remove any BARRIER
13230 after the second block.
13232 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
13234 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
13236 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
13238 PR target/59652
13239 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
13240 for 14-bit register offsets when INT14_OK_STRICT is false.
13242 2014-01-07 Roland Stigge <stigge@antcom.de>
13243 Michael Meissner <meissner@linux.vnet.ibm.com>
13245 PR 57386/target
13246 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
13247 Only check TFmode for SPE constants. Don't check TImode or TDmode.
13249 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
13251 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
13252 -mcpu.
13254 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
13256 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
13257 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
13258 rtx is const0_rtx or not.
13260 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
13262 PR target/58115
13263 * target-globals.c (save_target_globals): Remove this_fn_optab
13264 handling.
13265 * toplev.c: Include optabs.h.
13266 (target_reinit): Temporarily restore the global options if another
13267 set of options are in force.
13269 2014-01-07 Jakub Jelinek <jakub@redhat.com>
13271 PR rtl-optimization/58668
13272 * cfgcleanup.c (flow_find_cross_jump): Don't count
13273 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
13274 to determine what is counted.
13275 (flow_find_head_matching_sequence): Use active_insn_p to determine
13276 what is counted.
13277 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
13278 counting change.
13279 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
13280 determine what is counted.
13282 PR tree-optimization/59643
13283 * tree-predcom.c (split_data_refs_to_components): If one dr is
13284 read and one write, determine_offset fails and the write isn't
13285 in the bad component, just put the read into the bad component.
13287 2014-01-07 Mike Stump <mikestump@comcast.net>
13288 Jakub Jelinek <jakub@redhat.com>
13290 PR pch/59436
13291 * tree-core.h (struct tree_optimization_option): Change optabs
13292 type from unsigned char * to void *.
13293 * optabs.c (init_tree_optimization_optabs): Adjust
13294 TREE_OPTIMIZATION_OPTABS initialization.
13296 2014-01-06 Jakub Jelinek <jakub@redhat.com>
13298 PR target/59644
13299 * config/i386/i386.h (struct machine_function): Add
13300 no_drap_save_restore field.
13301 * config/i386/i386.c (ix86_save_reg): Use
13302 !cfun->machine->no_drap_save_restore instead of
13303 crtl->stack_realign_needed.
13304 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
13305 this function clears frame_pointer_needed. Set
13306 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
13307 and DRAP reg is needed.
13309 2014-01-06 Marek Polacek <polacek@redhat.com>
13311 PR c/57773
13312 * doc/implement-c.texi: Mention that other integer types are
13313 permitted as bit-field types in strictly conforming mode.
13315 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
13317 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
13318 is newly allocated.
13320 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
13322 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
13324 2014-01-06 Martin Jambor <mjambor@suse.cz>
13326 PR ipa/59008
13327 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
13328 to int.
13329 * ipa-prop.c (ipa_print_node_params): Fix indentation.
13331 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
13333 PR debug/59350
13334 PR debug/59510
13335 * var-tracking.c (add_stores): Preserve the value of the source even if
13336 we don't record the store.
13338 2014-01-06 Terry Guo <terry.guo@arm.com>
13340 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
13342 2014-01-05 Iain Sandoe <iain@codesourcery.com>
13344 PR bootstrap/59541
13345 * config/darwin.c (darwin_function_section): Adjust return values to
13346 correspond to optimisation changes made in r206070.
13348 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
13350 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
13351 from prefetch_block tune setting.
13352 (nocona_cost): Correct size of prefetch block to 64.
13354 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
13356 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
13357 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
13358 used to save the static chain register in the computation of the offset
13359 from which the FP registers need to be restored.
13361 2014-01-04 Jakub Jelinek <jakub@redhat.com>
13363 PR tree-optimization/59519
13364 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
13365 ICE if get_current_def (current_new_name) is already non-NULL, as long
13366 as it is a phi result of some other phi in *new_exit_bb that has
13367 the same argument.
13369 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
13370 or vmovdqu* for misaligned_operand.
13371 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
13372 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
13373 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
13374 aligned_mem for AVX512F masked aligned load and store builtins and for
13375 non-temporal moves.
13377 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
13379 PR tree-optimization/59651
13380 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
13381 Address range for negative step should be added by TYPE_SIZE_UNIT.
13383 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
13385 * config/m68k/m68k.c (handle_move_double): Handle pushes with
13386 overlapping registers also for registers other than the stack pointer.
13388 2014-01-03 Marek Polacek <polacek@redhat.com>
13390 PR other/59661
13391 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
13392 __builtin_FILE.
13394 2014-01-03 Jakub Jelinek <jakub@redhat.com>
13396 PR target/59625
13397 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
13398 asm goto as jump.
13400 * config/i386/i386.md (MODE_SIZE): New mode attribute.
13401 (push splitter): Use <P:MODE_SIZE> instead of
13402 GET_MODE_SIZE (<P:MODE>mode).
13403 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
13404 (mov -1, reg peephole2): Likewise.
13405 * config/i386/sse.md (*mov<mode>_internal,
13406 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
13407 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
13408 *<code><mode>3, *andnot<mode>3<mask_name>,
13409 <mask_codefor><code><mode>3<mask_name>): Likewise.
13410 * config/i386/subst.md (mask_mode512bit_condition,
13411 sd_mask_mode512bit_condition): Likewise.
13413 2014-01-02 Xinliang David Li <davidxl@google.com>
13415 PR tree-optimization/59303
13416 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
13417 (dump_predicates): Better output format.
13418 (pred_equal_p): New function.
13419 (is_neq_relop_p): Ditto.
13420 (is_neq_zero_form_p): Ditto.
13421 (pred_expr_equal_p): Ditto.
13422 (pred_neg_p): Ditto.
13423 (simplify_pred): Ditto.
13424 (simplify_preds_2): Ditto.
13425 (simplify_preds_3): Ditto.
13426 (simplify_preds_4): Ditto.
13427 (simplify_preds): Ditto.
13428 (push_pred): Ditto.
13429 (push_to_worklist): Ditto.
13430 (get_pred_info_from_cmp): Ditto.
13431 (is_degenerated_phi): Ditto.
13432 (normalize_one_pred_1): Ditto.
13433 (normalize_one_pred): Ditto.
13434 (normalize_one_pred_chain): Ditto.
13435 (normalize_preds): Ditto.
13436 (normalize_cond_1): Remove function.
13437 (normalize_cond): Ditto.
13438 (is_gcond_subset_of): Ditto.
13439 (is_subset_of_any): Ditto.
13440 (is_or_set_subset_of): Ditto.
13441 (is_and_set_subset_of): Ditto.
13442 (is_norm_cond_subset_of): Ditto.
13443 (pred_chain_length_cmp): Ditto.
13444 (convert_control_dep_chain_into_preds): Type change.
13445 (find_predicates): Ditto.
13446 (find_def_preds): Ditto.
13447 (destroy_predicates_vecs): Ditto.
13448 (find_matching_predicates_in_rest_chains): Ditto.
13449 (use_pred_not_overlap_with_undef_path_pred): Ditto.
13450 (is_pred_expr_subset): Ditto.
13451 (is_pred_chain_subset_of): Ditto.
13452 (is_included_in): Ditto.
13453 (is_superset_of): Ditto.
13455 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
13457 Update copyright years.
13459 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
13461 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
13462 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
13463 config/arc/arc.md, config/arc/arc.opt,
13464 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
13465 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
13466 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
13467 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
13468 config/linux-protos.h, config/linux.c, config/winnt-c.c,
13469 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
13470 vtable-verify.c, vtable-verify.h: Use the standard form for the
13471 copyright notice.
13473 2014-01-02 Tobias Burnus <burnus@net-b.de>
13475 * gcc.c (process_command): Update copyright notice dates.
13476 * gcov-dump.c: Ditto.
13477 * gcov.c: Ditto.
13478 * doc/cpp.texi: Bump @copying's copyright year.
13479 * doc/cppinternals.texi: Ditto.
13480 * doc/gcc.texi: Ditto.
13481 * doc/gccint.texi: Ditto.
13482 * doc/gcov.texi: Ditto.
13483 * doc/install.texi: Ditto.
13484 * doc/invoke.texi: Ditto.
13486 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
13488 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
13490 2014-01-01 Jakub Jelinek <jakub@redhat.com>
13492 * config/i386/sse.md (*mov<mode>_internal): Guard
13493 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
13495 PR rtl-optimization/59647
13496 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
13497 new_rtx into UNSIGNED_FLOAT rtxes.
13498 \f
13499 Copyright (C) 2014 Free Software Foundation, Inc.
13501 Copying and distribution of this file, with or without modification,
13502 are permitted in any medium without royalty provided the copyright
13503 notice and this notice are preserved.