1 2013-10-10 Richard Biener <rguenther@suse.de>
3 PR tree-optimization/58656
4 * tree-ssa-pre.c (phi_translate): Do not cache failed
7 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
9 * gimplify.c: Include expr.h and tm_p.h for targets with special va-arg
12 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
14 * tree-flow.h: Move some prototypes to gimple.h.
15 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
16 * gimple.h: Relocate some prototypes from tree-flow.h
17 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
19 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
20 (build_va_arg_indirect_ref): Relocate and make static.
21 (std_gimplify_va_arg_expr): Relocate here.
22 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
23 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
25 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
27 * doc/md.texi: Document the mnemonic attribute.
29 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
32 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
33 (set_attr_alternative x ...) when searching for user defined
36 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
38 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
40 2013-10-09 Easwaran Raman <eraman@google.com>
42 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
43 * cfgexpand.c (defer_stack_allocation): ...use here
44 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
46 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
48 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
49 (operand_equal_for_value_replacement): New function, extracted from
50 value_replacement and enhanced to catch more cases.
51 (value_replacement): Use operand_equal_for_value_replacement.
53 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
55 * loop-doloop.c (doloop_modify, doloop_optimize): Use
56 get_max_loop_iterations.
58 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
60 * config/arm/aarch-common.c (arm_early_load_addr_dep):
61 Place comment above function.
63 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
65 * tree-flow.h: Remove all remaining prototypes, enums and structs that
66 are not related to tree-cfg.c.
67 * tree-ssa-address.h: New file. Relocate prototypes.
68 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
69 (addr_for_mem_ref): New. Combine call to get_address_description and
70 return addr_for_mem_ref.
71 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
72 * tree-ssa-live.h: Adjust prototypes.
73 * passes.c: Include tree-ssa-live.h.
74 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
75 * graphite.c (graphite_transform_loops): Make static.
76 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
77 make_pass_graphite, pass_data_graphite_transforms,
78 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
79 * ipa-pure-const.c (warn_function_noreturn): Make static.
80 (execute_warn_function_noreturn, gate_warn_function_noreturn,
81 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
82 Relocate from tree-cfg.c
83 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
85 (execute_warn_function_noreturn, gate_warn_function_noreturn,
86 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
87 Move to ipa-pure-const.c.
88 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg): Relocate
90 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
91 make_pass_fixup_cfg): Move to tree-cfg.c.
92 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
93 Relocate some prototypes.
94 * tree-data-ref.h (tree_check_data_deps) Add prototype.
95 * tree-dump.c (dump_function_to_file): Remove prototype.
96 Add tree-flow.h to the include file.
97 * tree-dump.h: Remove prototype.
98 * tree-parloops.h: New File. Add prototypes.
99 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
100 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
101 from tree-ssa-loop.c.
102 * tree-predcom.c (run_tree_predictive_commoning,
103 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
104 Relocate here from tree-ssa-loop.c.
105 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
106 ssa_name_values.release ().
107 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
108 (ssa_name_values): Relocate from tree-flow.h.
109 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
110 * tree-ssa-loop.c (run_tree_predictive_commoning,
111 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
112 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
113 make_pass_graphite, pass_data_graphite_transforms,
114 make_pass_graphite_transforms, gate_tree_parallelize_loops,
115 tree_parallelize_loops, pass_data_parallelize_loops,
116 make_pass_parallelize_loops): Move to other files.
117 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
119 * tree.h: Remove prototypes from tree-address.c.
121 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
123 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
124 (struct int_tree_map): Move to tree-hasher.h
125 (SCALE, LABEL, PERCENT): Move to gimple.h
126 * tree-flow-inline.h: Delete. Move functions to other files.
127 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
128 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
129 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
130 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
131 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
132 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
133 * tree-hasher.h: Don't include tree-flow.h.
134 (struct int_tree_map): Relocate from tree-flow.h.
135 * tree-sra.c (contains_view_convert_expr_p): Relocate from
136 tree-flow-inline.h and make static.
137 * tree-ssa-alias.h (ranges_overlap_p): Relocate from tree-flow-inline.h.
138 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
139 tree-flow-inline.h and make static.
140 * tree.h (is_global_var, may_be_aliased): Relocate from
142 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
143 * value-prof.c: No longer include tree-flow-inline.h.
144 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
146 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
148 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
149 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
150 (enum move_pos): Move to tree-ssa-loop-im.h
151 * cfgloop.h: Move some prototypes.
152 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
153 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
154 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
155 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
156 (loop_containing_stmt): Relocate from tree-flow-inline.h.
157 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
158 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
159 (enum move_pos): Relocate here.
160 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
162 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
163 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim, make_pass_lim):
164 Relocate here from tree-ssa-loop.c.
165 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
167 (loop_edge_to_cancel, unloop_loops): Make static.
168 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
169 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
170 (tree_complete_unroll, gate_tree_complete_unroll,
171 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
172 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
173 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
174 * tree-ssa-loop-ivopts.c: Remove local prototypes.
175 (stmt_invariant_in_loop_p): Remove unused function.
176 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
177 * tree-ssa-loop-manip.h: New file. Add prototypes.
178 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
179 (gcov_type_to_double_int): Move to cfgloop.h.
180 (double_int_cmp, bound_index,
181 estimate_numbers_of_iterations_loop): Make static.
182 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
183 (max_loop_iterations): Factor out get_max_loop_iterations.
184 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
186 * tree-ssa-loop-niter.h: New file. Add prototypes.
187 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
188 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
189 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
190 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
191 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
192 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
193 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
194 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
195 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
196 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
197 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
198 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
199 pass_data_complete_unroll, make_pass_complete_unroll,
200 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
201 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
202 tree-ssa-loop-ivcanon.c.
203 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
204 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
205 tree-ssa-loop-prefetch.c.
206 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
208 (get_lsm_tmp_name): Relocate and add suffix parameter.
209 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
210 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
211 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
212 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
213 (get_estimated_loop_iterations): Factor out accessor from
214 estimated_loop_iterations in tree-ssa-loop-niter.c.
215 (get_max_loop_iterations): Factor out accessor from _max_loop_iterations
217 * loop-unroll.c (decide_unroll_constant_iterations,
218 decide_unroll_runtime_iterations, decide_peel_simple,
219 decide_unroll_stupid): Use new get_* accessors.
221 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
223 PR tree-optimization/20318
224 * doc/extend.texi (returns_nonnull): New function attribute.
225 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
227 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
228 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
230 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
233 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
234 false if both components are bitfields.
236 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
238 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
239 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
240 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
241 * config/aarch64/aarch64.h
242 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
243 * config/aarch64/aarch64-simd-builtins.def
244 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
246 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
248 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
250 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
252 * config/aarch64/arm_neon.h (vdiv_f64): Added.
254 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
256 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
257 (vneg_f64): New intrinsic.
258 (vneg_s8): Asm replaced with C.
259 (vneg_s16): Likewise.
260 (vneg_s32): Likewise.
261 (vneg_s64): New intrinsic.
262 (vnegq_f32): Asm replaced with C.
263 (vnegq_f64): Likewise.
264 (vnegq_s8): Likewise.
265 (vnegq_s16): Likewise.
266 (vnegq_s32): Likewise.
267 (vnegq_s64): Likewise.
269 2013-10-09 Renlin Li <Renlin.Li@arm.com>
271 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
273 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
275 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
276 packed stack special handling.
277 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
278 back to fixed stack slots for FPRs saved due to stdarg.
280 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
282 * config/s390/s390.c (s390_frame_info): Restructure function.
284 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
286 * config/s390/s390.c (struct s390_frame_layout): New field
288 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
289 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
290 regs_ever_clobbered to char*.
291 (s390_regs_ever_clobbered): Check crtl->saves_all_registers
292 instead of cfun->has_nonlocal_label. Ignore frame related restore
294 (s390_register_info): Enable FPR save slots. Move/Copy some
295 functionality into ...
296 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
297 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
299 (s390_frame_info): Do gpr slot allocation here now. stdarg does
300 not imply a stack frame.
301 (s390_init_frame_layout): Remove variable clobbered_regs.
302 (s390_update_register_info): Remove function.
303 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
304 cannot be used for register renaming.
305 (s390_hard_regno_scratch_ok): New function.
306 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
307 (s390_initial_elimination_offset): Change offset calculation of
308 the return address pointer.
309 (save_gprs): Deal with only r6 being saved from the call-saved
311 (restore_gprs): Set frame related flag.
312 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New
314 (s390_emit_prologue): Call s390_register_info instead of
315 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
316 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
317 (s390_optimize_prologue): Call s390_optimize_register_info.
318 Try to remove also FPR slot save/restore INSNs. Remove frame
319 related flags from restore INSNs.
321 2013-10-08 DJ Delorie <dj@redhat.com>
323 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
326 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
327 avoid conflict with the MI use of %c.
328 * config/rl78/rl78-real.md: change %c to %C throughout.
329 * config/rl78/rl78-virt.md: Likewise.
331 2013-10-08 Jan Hubicka <jh@suse.cz>
333 * config/i386/i386.c (ix86_option_override_internal): Switch
334 to SSE math for -ffast-math when target ISA supports SSE2.
336 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
338 * tree-flow.h: Remove some prototypes.
339 * tree.h: Remove some protypes, add a couple.
340 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
341 using_eh_for_cleanups_p): Add interface routines for front ends.
342 * tree-eh.h: New file. Add protoptyes.
343 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
344 (add_stmt_to_eh_lp_fn): Make static.
345 (lower_try_finally): Use new using_eh_for_cleanups_p.
346 * emit-rtl.c: Include tree-eh.h.
347 * gimple.h: Include tree-eh.h.
349 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
351 PR tree-optimization/58480
352 * tree-vrp.c (infer_nonnull_range): New function.
353 (infer_value_range): Call infer_nonnull_range.
355 2013-10-08 Dehao Chen <dehao@google.com>
357 PR tree-optimization/58619
358 * tree-inline.c (copy_phis_for_bb): Combine location data
361 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
364 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
365 RTX_FRAME_RELATED_P on INSN.
367 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
369 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
370 (altivec_expand_vec_perm_const): Call it.
372 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
374 * config/rs6000/vector.md (mov<mode>): Emit permuted move
375 sequences for LE VSX loads and stores at expand time.
376 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
378 * config/rs6000/rs6000.c (rs6000_const_vec): New.
379 (rs6000_gen_le_vsx_permute): New.
380 (rs6000_gen_le_vsx_load): New.
381 (rs6000_gen_le_vsx_store): New.
382 (rs6000_gen_le_vsx_move): New.
383 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
384 (*vsx_le_perm_load_v4si): New.
385 (*vsx_le_perm_load_v8hi): New.
386 (*vsx_le_perm_load_v16qi): New.
387 (*vsx_le_perm_store_v2di): New.
388 (*vsx_le_perm_store_v4si): New.
389 (*vsx_le_perm_store_v8hi): New.
390 (*vsx_le_perm_store_v16qi): New.
391 (*vsx_xxpermdi2_le_<mode>): New.
392 (*vsx_xxpermdi4_le_<mode>): New.
393 (*vsx_xxpermdi8_le_V8HI): New.
394 (*vsx_xxpermdi16_le_V16QI): New.
395 (*vsx_lxvd2x2_le_<mode>): New.
396 (*vsx_lxvd2x4_le_<mode>): New.
397 (*vsx_lxvd2x8_le_V8HI): New.
398 (*vsx_lxvd2x16_le_V16QI): New.
399 (*vsx_stxvd2x2_le_<mode>): New.
400 (*vsx_stxvd2x4_le_<mode>): New.
401 (*vsx_stxvd2x8_le_V8HI): New.
402 (*vsx_stxvd2x16_le_V16QI): New.
404 2013-10-07 Renlin Li <Renlin.Li@arm.com>
406 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
408 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
410 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
411 loop to work also for 31bit ABI.
412 Save the stack pointer for frame_size > 0.
414 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
416 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
417 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
418 constraint letters from expanders.
419 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
420 retry count to general_operand.
421 ("tabort"): Give operand 0 a mode.
422 ("tabort_1"): Add mode and constraint letter for operand 0.
423 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
425 2013-10-04 Jeff Law <law@redhat.com>
427 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
429 * tree-ssa-threadedge.c (thread_through_normal_block): Broken out of ...
430 (thread_across_edge): Here. Call it.
432 2013-10-04 Cary Coutant <ccoutant@google.com>
434 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
435 discarding a location list expression (or a piece of one).
437 2013-10-03 Jan Hubicka <jh@suse.cz>
439 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
440 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
441 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
443 2013-10-03 Jan Hubicka <jh@suse.cz>
445 * config/i386/i386.c (ix86_option_override_internal): Do not enable
446 accumulate-outgoing-args when producing unwind info.
448 2013-10-03 Wei Mi <wmi@google.com>
450 * lra-constraints.c (insert_move_for_subreg): New function
451 extracted from simplify_operand_subreg.
452 (simplify_operand_subreg): Add reload for paradoxical subreg.
454 2013-10-03 Rong Xu <xur@google.com>
456 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
457 the candidate of builtin_expect such that we should fix the
458 size/time estimation.
459 (estimate_function_body_sizes): Do the acutally size/time fix-up
462 2013-10-03 Rong Xu <xur@google.com>
464 * predict.c (tree_predict_by_opcode): Get the probability
465 for builtin_expect from param builtin_expect_probability.
466 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
467 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
468 * doc/invoke.texi: Add documentation for builtin-expect-probability.
470 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
473 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
474 * calls.c (alloca_call_p): Use get_callee_fndecl.
475 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
476 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
478 (vrp_visit_stmt): Remove duplicated code.
480 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
482 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
483 ceildf2, btruncdf2, instead of vsx_* name.
485 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
486 iterators to only do V2DF and V4SF here. Move the DF code to
487 rs6000.md where it is combined with SF mode. Replace <VSv> with
488 just 'v' since only vector operations are handled with these insns
489 after moving the DF support to rs6000.md.
490 (vsx_sub<mode>3): Likewise.
491 (vsx_mul<mode>3): Likewise.
492 (vsx_div<mode>3): Likewise.
493 (vsx_fre<mode>2): Likewise.
494 (vsx_neg<mode>2): Likewise.
495 (vsx_abs<mode>2): Likewise.
496 (vsx_nabs<mode>2): Likewise.
497 (vsx_smax<mode>3): Likewise.
498 (vsx_smin<mode>3): Likewise.
499 (vsx_sqrt<mode>2): Likewise.
500 (vsx_rsqrte<mode>2): Likewise.
501 (vsx_fms<mode>4): Likewise.
502 (vsx_nfma<mode>4): Likewise.
503 (vsx_copysign<mode>3): Likewise.
504 (vsx_btrunc<mode>2): Likewise.
505 (vsx_floor<mode>2): Likewise.
506 (vsx_ceil<mode>2): Likewise.
507 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
508 (vsx_sminsf3): Likewise.
509 (vsx_fmadf4): Likewise.
510 (vsx_fmsdf4): Likewise.
511 (vsx_nfmadf4): Likewise.
512 (vsx_nfmsdf4): Likewise.
513 (vsx_cmpdf_internal1): Likewise.
515 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
516 simpler to select whether a target has SPE or traditional floating
517 point support in iterators.
518 (TARGET_DF_SPE): Likewise.
519 (TARGET_SF_FPR): Likewise.
520 (TARGET_DF_FPR): Likewise.
521 (TARGET_SF_INSN): Macros to say whether floating point support
522 exists for a given operation for expanders.
523 (TARGET_DF_INSN): Likewise.
525 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
526 combining of SF/DF mode operations, using both traditional and VSX
534 (abs<mode>2): Combine SF/DF modes using traditional floating point
535 instructions. Add support for using the upper DF registers with
536 VSX support, and SF registers with power8-vector support. Update
537 expanders for operations supported by both the SPE and traditional
538 floating point units.
539 (abs<mode>2_fpr): Likewise.
540 (nabs<mode>2): Likewise.
541 (nabs<mode>2_fpr): Likewise.
542 (neg<mode>2): Likewise.
543 (neg<mode>2_fpr): Likewise.
544 (add<mode>3): Likewise.
545 (add<mode>3_fpr): Likewise.
546 (sub<mode>3): Likewise.
547 (sub<mode>3_fpr): Likewise.
548 (mul<mode>3): Likewise.
549 (mul<mode>3_fpr): Likewise.
550 (div<mode>3): Likewise.
551 (div<mode>3_fpr): Likewise.
552 (sqrt<mode>3): Likewise.
553 (sqrt<mode>3_fpr): Likewise.
555 (rsqrt<mode>2): Likewise.
556 (cmp<mode>_fpr): Likewise.
557 (smax<mode>3): Likewise.
558 (smin<mode>3): Likewise.
559 (smax<mode>3_vsx): Likewise.
560 (smin<mode>3_vsx): Likewise.
561 (negsf2): Delete SF operations that are merged with DF.
568 (fmasf4_fpr): Likewise.
569 (fmssf4_fpr): Likewise.
570 (nfmasf4_fpr): Likewise.
571 (nfmssf4_fpr): Likewise.
573 (rsqrtsf_internal1): Likewise.
576 (cmpsf_internal1): Likewise.
577 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
578 (negdf2): Delete DF operations that are merged with SF.
586 (rsqrtdf_internal1): Likewise.
587 (fmadf4_fpr): Likewise.
588 (fmsdf4_fpr): Likewise.
589 (nfmadf4_fpr): Likewise.
590 (nfmsdf4_fpr): Likewise.
594 (cmpdf_internal1): Likewise.
595 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
596 (btrunc<mode>2): Delete separate expander, and combine with the
597 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
598 (btrunc<mode>2_fpr): Likewise.
599 (ceil<mode>2): Likewise.
600 (ceil<mode>2_fpr): Likewise.
601 (floor<mode>2): Likewise.
602 (floor<mode>2_fpr): Likewise.
603 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
604 Add support for using the upper registers with VSX and
605 power8-vector. Move insns to be closer to the define_expands. On
606 VSX systems, prefer the traditional form of FMA over the VSX
607 version, since the traditional form allows the target not to
608 overlap with the inputs.
609 (fms<mode>4_fpr): Likewise.
610 (nfma<mode>4_fpr): Likewise.
611 (nfms<mode>4_fpr): Likewise.
613 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
614 Richard Earnshaw <richard.earnshaw@arm.com>
616 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
617 (struct mult_cost_table): Likewise.
618 (struct mem_cost_table): Likewise.
619 (struct fp_cost_table): Likewise.
620 (struct vector_cost_table): Likewise.
621 (cpu_cost_table): Likewise.
622 * config/arm/arm.opt (mold-rts-costs): New option.
623 (mnew-generic-costs): Likewise.
624 * config/arm/arm.c (generic_extra_costs): New table.
625 (cortexa15_extra_costs): Likewise.
626 (arm_slowmul_tune): Use NULL as new costs.
627 (arm_fastmul_tune): Likewise.
628 (arm_strongarm_tune): Likewise.
629 (arm_xscale_tune): Likewise.
630 (arm_9e_tune): Likewise.
631 (arm_v6t2_tune): Likewise.
632 (arm_cortex_a5_tune): Likewise.
633 (arm_cortex_a9_tune): Likewise.
634 (arm_v6m_tune): Likewise.
635 (arm_fa726te_tune): Likewise.
636 (arm_cortex_a15_tune): Use cortex15_extra_costs.
637 (arm_cortex_tune): Use generict_extra_costs.
638 (shifter_op_p): New function.
639 (arm_unspec_cost): Likewise.
640 (LIBCALL_COST): Define.
641 (arm_new_rtx_costs): New function.
642 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
643 table is available. Use old costs otherwise unless mnew-generic-costs
645 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
646 (cpu_cost_table): Declare.
648 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
651 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
652 (*subs_mul_imm_<mode>)
653 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
654 (*sub_<shift>_<mode>)
655 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
658 2013-10-03 Ian Bolton <ian.bolton@arm.com>
660 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
662 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
664 2013-10-02 Teresa Johnson <tejohnson@google.com>
666 * predict.c (probably_never_executed): New function.
667 (probably_never_executed_bb_p): Invoke probably_never_executed.
668 (probably_never_executed_edge_p): Ditto.
669 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
670 Treat profile insanities conservatively.
672 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
674 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
676 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
678 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
679 use it. Use smaller increase for scratch. Don't increase reject
680 for early clobber scratch.
681 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
682 setting eliminated regs except setting fp from hfp.
683 (lra_eliminate): Check lra_insn_recog_data on NULL.
685 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
688 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
689 setting -mvsx-timode by default until the underlying problem is fixed.
690 (RS6000_CPU, power7 defaults): Likewise.
692 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
694 * config/x-linux (host-linux.o): Remove header dependencies.
695 Use $(COMPILE) and $(POSTCOMPILE).
696 * config/t-linux-android (linux-android.o): Ditto.
698 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
700 * Makefile.in (expmed.o-warn): Remove.
702 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
704 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
705 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
707 2013-10-02 Teresa Johnson <tejohnson@google.com>
709 * dojump.c (do_jump_1): Divide probability between
710 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
712 2013-10-02 Tom Tromey <tromey@redhat.com>
714 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
716 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
718 * tree-flow.h: Remove some prototypes.
719 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
720 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
721 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
722 mark_virtual_phi_result_for_renaming): Relocate here.
723 * tree-into-ssa.h: Add prototypes.
724 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
725 single_pred_before_succ_order.
726 (blocks_in_phiopt_order): Rename and move to cfganal.c.
727 (nonfreeing_call_p) Move to gimple.c.
728 * cfganal.c (single_pred_before_succ_order): Move and renamed from
730 * basic-block.h (single_pred_before_succ_order): Add prototype.
731 * gimple.c (nonfreeing_call_p): Relocate here.
732 * gimple.h: Add prototype.
733 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
734 * tree-ssa-dom.h: New file. Relocate prototypes here.
735 * tree-ssa.h: Include tree-ssa-dom.h.
737 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
739 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
740 Use $(COMPILE) and $(POSTCOMPILE).
742 * config/alpha/x-alpha (driver-alpha.o): Ditto.
744 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
746 * tree-flow.h: Remove some prototypes.
747 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
748 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
749 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
750 propagate_tree_value*): Move from here to...
751 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
752 propagate_tree_value*): Relocate here.
753 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
754 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
755 * gimple-fold.c: Remove gimple-fold.h from include list.
756 * tree-vrp.c: Remove gimple-fold.h from include list.
757 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
758 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
759 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
760 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
761 * sese.c: Add tree-ssa-propagate.h to include list.
763 2013-10-02 Richard Biener <rguenther@suse.de>
765 * tree-loop-distribution.c: Include tree-vectorizer.h for
767 (enum partition_kind): Remove PKIND_REDUCTION.
768 (struct partition_s): Remove has_writes member, add reduction_p member.
769 (partition_alloc): Adjust.
770 (partition_builtin_p): Likewise.
771 (partition_has_writes): Remove.
772 (partition_reduction_p): New function.
773 (partition_merge_into): Likewise.
774 (generate_code_for_partition): Commonize builtin partition
776 (rdg_cannot_recompute_vertex_p): Remove.
777 (already_processed_vertex_p): Likewise.
778 (rdg_flag_vertex): Do not set has_writes.
779 (classify_partition): Adjust.
780 (rdg_build_partitions): Do not set has_writes, treat all
781 partitions as useful.
782 (distribute_loop): Record number of library calls generated. Adjust.
783 (tree_loop_distribution): Report number of loops and library
784 calls generated as opt-info.
786 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
788 * tree-flow.h: Include new .h files. Move prototypes.
789 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
790 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
791 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
792 * tree-pretty-print.h: Add prototypes from tree-flow.h.
793 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
794 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
795 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
796 * tree.h (get_ref_base_and_extent): Move prototype out.
797 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
799 * gimple-low.h: New File. Add prototypes from tree-flow.h.
800 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
801 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
802 * tree-scalar-evolution.c: Include tree.h.
803 * sese.c: Include tree.h.
804 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
805 * dwarf2out.c: Include tree-dfa.h.
806 * tree-chrec.c: Include tree.h.
807 * tree-data-ref.c: Include tree.h.
809 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
811 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
814 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
816 * config/t-sol2 (sol2-c.o): Remove header dependencies.
817 Use $(COMPILE) and $(POSTCOMPILE).
818 (sol2-cxx.o): Likewise.
819 (sol2-stubs.o): Likewise.
821 * config/x-solaris (host-solaris.o): Likewise.
823 * config/sparc/t-sparc (sparc.o): Remove.
824 (sparc-c.o): Remove header dependencies.
825 Use $(COMPILE) and $(POSTCOMPILE).
826 * config/sparc/x-sparc: Likewise.
828 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
830 * config/arc/arc-opts.h: Add 2013 to Copyright years.
831 * config/arc/arc700.md: Likewise.
832 * config/arc/arc-modes.def: Likewise.
833 * config/arc/arc-simd.h: Likewise.
834 * config/arc/t-arc-uClibc: Likewise.
835 * config/arc/t-arc-newlib: Likewise.
837 2013-10-02 Renlin Li <renlin.li@arm.com>
839 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
841 (aarch64_expand_epilogue): Likewise.
843 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
844 Yufeng Zhang <yufeng.zhang@arm.com>
846 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
848 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
849 'base_in' represent a conversion and legal_cast_p_1 holds; set
850 'base_in' with the returned value from get_unwidened.
852 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
854 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
855 plus_constant is not used.
857 2013-10-01 Wei Mi <wmi@google.com>
859 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
860 * config/i386/i386.md: Add define_peephole2 to
861 break partial reg stall for cvtss2sd/cvtsd2ss.
863 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
865 * config/arc/arc.c (pass_arc_ifcvt::clone):
866 Update for ctxt_ -> m_ctxt change.
868 2013-10-01 Jeff Law <law@redhat.com>
870 * tree-ssa-threadupdate.c (struct redirection_data): Delete
871 outgoing_edge and intermediate_edge fields. Instead store the path.
872 (redirection_data::hash): Hash on the last edge's destination index.
873 (redirection_data::equal): Check the entire thread path.
874 (lookup_redirectio_data): Corresponding changes.
875 (create_edge_and_update_destination_phis): Likewise.
876 (thread_single_edge): Likewise.
878 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
879 Diego Novillo <dnovillo@google.com>
881 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
882 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
883 (vld32wh_insn, vld32wl_insn): Delete commented-out old
884 versions of these patterns.
886 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
887 (__builtin_arc_aligned): Likewise.
889 * config/arc/arc.md: Expand adc_0 comment stating the intended
890 purpose and why it isn't ready.
891 Replace commented out call_value_via_label_mixed with a
892 plain comment about bl_s.
894 * config/arc/arc.c (stdio.h): Don't include directly.
895 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
896 Remove if (1) condition.
897 (arc_encode_section_info): Fix comment.
899 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
901 * config/arc/arc.c (arc_conditional_register_usage):
902 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
903 Also set reg_alloc_order for DMA config regs.
905 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
906 Jeremy Bennett <jeremy.bennett@embecosm.com>
908 * doc/install.texi (--with-cpu): Mention ARC.
909 (arc-*-elf32): New paragraph.
910 (arc-linux-uclibc): Likewise.
911 * doc/md.texi (Machine Constraints): Add ARC part.
912 * doc/invoke.texi: (menu): Add ARC Options.
913 (Machine Dependent Options) <ARC Options>: Add synopsis.
914 (node ARC Options): Add.
915 * doc/extend.texi (long_call / short_call attribute): Add ARC.
916 (ARC Built-in Functions): New section defining
917 generic ARC built-in functions.
918 (ARC SIMD Built-in Functions): New section defining SIMD specific
920 (Declaring Attributes of Functions): Extended
921 description of short_call and long_call attributes for ARC and
924 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
925 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
926 Joern Rennecke <joern.rennecke@embecosm.com>
927 Muhammad Khurram Riaz <khurram.riaz@arc.com>
928 Brendan Kehoe <brendan@zen.org>
929 Michael Eager <eager@eagercon.com>
930 Simon Cook <simon.cook@embecosm.com>
931 Jeremy Bennett <jeremy.bennett@embecosm.com>
933 * config/arc, common/config/arc: New directories.
935 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
936 Brendan Kehoe <brendan@zen.org>
937 Simon Cook <simon.cook@embecosm.com>
939 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
941 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
943 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
944 * tree-ssa-coalesce.h: New. Move prototype to here.
945 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
946 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
947 (gimple_can_coalesce_p): Move to...
948 * gimple.c (gimple_can_coalesce_p): Here.
950 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
952 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
953 (dump_decl_set): Move to gimple.c.
954 * gimple.h: Don't include tree-ssa-operands.h.
955 (dump_decl_set): Add prototype.
956 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
957 Move to gimple-ssa.h.
958 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
959 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
960 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
961 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
962 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
964 (dump_decl_set): Relocate here.
965 * gimple-ssa.h: New file.
966 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
967 Relocate from gimple.h.
968 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
969 * tree-ssa-operands.c (swap_ssa_operands): Rename from
970 swap_tree_operands and remove non-ssa path.
971 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
972 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
974 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
975 vect_is_simple_reduction_1): Use swap_ssa_operands.
976 * tree-flow.h: Move various prototypes to tree-phinodes.h.
977 (enum need_phi_state): Move to tree-into-ssa.c.
978 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
979 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
980 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
981 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
982 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
983 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
984 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
985 num_imm_uses): Move to ssa-iterators.h.
986 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
987 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
989 (op_iter_done, op_iter_next_def, op_iter_next_tree,
990 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
991 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
992 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
993 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
994 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
995 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
996 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
997 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
998 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
999 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
1000 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
1001 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
1002 (set_phi_nodes): Move to tree-phinodes.h.
1003 * tree-ssa-operands.h (enum ssa_op_iter_type,
1004 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
1005 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
1006 (dump_decl_set): Remove prototype.
1007 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
1008 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
1009 (set_phi_nodes): Relocate from tree-flow-inline.h.
1010 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
1012 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
1013 include list. Temporarily add gimple.h to include list.
1014 * ssa-iterators.h: New file.
1015 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
1016 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
1017 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
1018 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
1019 Relocate from tree-ssa-operands.h.
1020 (delink_imm_use, link_imm_use_to_list, link_imm_use,
1021 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
1022 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
1023 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
1024 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
1025 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
1026 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
1027 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
1028 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
1029 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
1030 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
1031 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
1032 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
1033 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
1034 Relocate from tree-flow-inline.h.
1035 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
1037 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
1040 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
1041 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
1042 Insert '\t' to output template.
1043 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
1044 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
1045 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
1046 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
1048 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
1050 * doc/install.texi (Host/target specific installation notes for GCC):
1051 Put @anchor before @heading.
1052 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
1053 Use @email for email addresses.
1055 2013-10-01 Jeff Law <law@redhat.com>
1057 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
1058 a vec. Only delete the path if we create one without successfully
1059 registering a jump thread.
1060 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
1062 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
1063 (paths): New vector of jump threading paths.
1064 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
1065 (THREAD_PATH): New accessor macro for the entire thread path.
1066 (lookup_redirection_data): Get intermediate and final outgoing edge
1067 from the thread path.
1068 (create_edge_and_update_destination_phis): Copy the threading path.
1069 (ssa_fix_duplicate_block_edges): Get edges and block types from the
1070 jump threading path.
1071 (ssa_redirect_edges): Get edges and block types from the jump threading
1072 path. Free the path vector.
1073 (thread_block): Get edges from the jump threading path. Look at the
1074 entire path to see if we thread to a loop exit. If we cancel a jump
1075 thread request, then free the path vector.
1076 (thread_single_edge): Get edges and block types from the jump threading
1077 path. Free the path vector.
1078 (thread_through_loop_header): Get edges and block types from the jump
1079 threading path. Free the path vector.
1080 (mark_threaded_blocks): Iterate over the vector of paths and store
1081 the path on the appropriate edge. Get edges and block types from the
1082 jump threading path.
1083 (mark_threaded_blocks): Get edges and block types from the jump
1084 threading path. Free the path vector.
1085 (thread_through_all_blocks): Use the vector of paths rather than
1086 a vector of 3-edge sets.
1087 (register_jump_thread): Accept pointer to a path vector rather
1088 than the path vector itself. Store the path vector for later use.
1091 2013-10-01 Jakub Jelinek <jakub@redhat.com>
1092 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1095 * config/s390/s390.c (s390_split_branches): Modify check for table
1097 (s390_chunkify_start): Rearrange table jump insn check in order to
1098 deal with compare and branch insns correctly.
1100 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
1104 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
1105 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
1106 define_insn_and_split.
1107 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
1108 (shiftsi3_compare): New pattern.
1110 * config/arm/unspecs.md (UNSPEC_RRX): New.
1112 2013-10-01 Alan Modra <amodra@gmail.com>
1114 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
1115 casing inout operands.
1117 2013-10-01 Richard Biener <rguenther@suse.de>
1119 PR tree-optimization/58553
1120 * tree-loop-distribution.c (struct partition_s): Add niter member.
1121 (classify_partition): Populate niter member for the partition
1122 and properly identify whether the relevant store happens before
1123 or after the loop exit.
1124 (generate_memset_builtin): Use niter member from the partition.
1125 (generate_memcpy_builtin): Likewise.
1127 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
1129 * vec.h (vec_prefix, vec): Prefix member names with "m_".
1130 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
1132 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
1134 * basic-block.h (edge_list): Prefix member names with "m_".
1135 * context.h (context): Likewise.
1136 * domwalk.h (dom_walker): Likewise.
1137 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
1138 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
1139 * hash-table.h (hash_table): Likewise.
1140 * machmode.h (bit_field_mode_iterator): Likewise.
1141 * pass_manager.h (pass_list): Likewise.
1142 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
1143 * tree-pass.h (pass_data): Likewise.
1144 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
1145 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
1146 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
1147 * asan.c (pass_data_asan): Update accordingly.
1148 * cfganal.c (control_dependences::find_control_dependence): Likewise.
1149 (control_dependences::control_dependences): Likewise.
1150 (control_dependences::~control_dependences): Likewise.
1151 (control_dependences::~control_dependences): Likewise.
1152 (control_dependences::get_edges_dependent_on): Likewise.
1153 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
1154 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
1155 * context.c (gcc::context::context): Likewise.
1156 * cprop.c (pass_rtl_cprop::clone): Likewise.
1157 * domwalk.c (dom_walker::walk): Likewise.
1158 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
1159 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
1160 * mode-switching.c (pass_mode_switching::clone): Likewise.
1161 * passes.c (opt_pass::opt_pass): Likewise.
1162 (pass_manager::pass_manager): Likewise.
1163 * predict.c (pass_strip_predict_hints::clone): Likewise.
1164 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
1165 (pass_split_all_insns::clone): Likewise.
1166 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
1168 (bit_field_mode_iterator::next_mode): Likewise.
1169 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
1170 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
1171 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
1172 * tree-complex.c (pass_lower_complex::clone): Likewise.
1173 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
1174 * tree-object-size.c (pass_object_sizes::clone): Likewise.
1175 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
1176 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
1177 (pass_fold_builtins::clone): Likewise.
1178 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
1179 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
1180 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
1181 (pass_cd_dce::clone): Likewise.
1182 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
1183 (pass_phi_only_cprop::clone): Likewise.
1184 * tree-ssa-dse.c (pass_dse::clone): Likewise.
1185 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
1186 * tree-ssa-loop.c (pass_lim::clone): Likewise.
1187 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
1188 * tree-ssa-pre.c (pass_fre::clone): Likewise.
1189 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
1190 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
1191 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
1192 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
1193 * tree-vrp.c (pass_vrp::clone): Likewise.
1194 * tsan.c (pass_tsan::clone): Likewise.
1196 2013-09-30 Jakub Jelinek <jakub@redhat.com>
1199 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
1200 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
1203 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
1204 optimization, punt if sign_bit_p looked through any zero extension.
1206 2013-09-30 Teresa Johnson <tejohnson@google.com>
1208 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
1209 Update redirected out edge count in joiner case.
1210 (ssa_redirect_edges): Common the joiner and non-joiner cases
1211 so that joiner case gets profile updates.
1213 2013-09-30 Richard Biener <rguenther@suse.de>
1215 PR tree-optimization/58554
1216 * tree-loop-distribution.c (classify_partition): Require
1217 unconditionally executed stores for memcpy and memset recognition.
1218 (tree_loop_distribution): Calculate dominance info.
1220 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
1222 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
1223 (NO_PROFILE_COUNTERS): Likewise.
1224 (PROFILE_HOOK): Likewise.
1225 (FUNCTION_PROFILER): Likewise.
1226 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
1228 2013-09-30 Iain Sandoe <iain@codesourcery.com>
1230 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
1231 calls and defines in TARGET_MACHO conditional.
1232 (load_macho_picbase_di): Likewise.
1233 (reload_macho_picbase): Likewise.
1234 (reload_macho_picbase_si): Likewise.
1235 (reload_macho_picbase_di): Likewise.
1236 (nonlocal_goto_receiver): Likewise.
1238 2013-09-30 Nick Clifton <nickc@redhat.com>
1240 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
1241 that use the MSP430X ISA.
1242 (msp430_option_override): Scan -mmcu command line option for any
1243 MCU name that supports the MSP430X ISA.
1244 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
1245 -mmcu options which enable the MSP430X ISA.
1247 2013-09-30 Richard Biener <rguenther@suse.de>
1250 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
1251 before looking for setjmp-like calls.
1253 2013-09-29 Iain Sandoe <iain@codesourcery.com>
1256 * config/darwin-protos.h (machopic_get_function_picbase): New.
1257 * config/darwin.c (machopic_get_function_picbase): New.
1258 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
1259 label for a new func. (load_macho_picbase_di): Likewise.
1260 (reload_macho_picbase): New expand.
1261 (reload_macho_picbase_si): New insn.
1262 (reload_macho_picbase_di): New insn.
1263 (nonlocal_goto_receiver): New define and split.
1264 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
1265 (unspecv enum): Add UNSPECV_NLGR.
1267 2013-09-29 Iain Sandoe <iain@codesourcery.com>
1269 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
1270 that altivec registers are correctly sized on Darwin.
1272 2013-09-29 Iain Sandoe <iain@codesourcery.com>
1274 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
1275 darwin-driver.o): Use COMPILE and POSTCOMPILE.
1276 * config/x-darwin (host-darwin.o): Likewise.
1277 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
1278 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
1279 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
1281 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
1283 * doc/invoke.texi: Fix usage of @tie{} command.
1285 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
1287 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
1289 2013-09-28 Jan Hubicka <jh@suse.cz>
1291 * config/i386/x86-tune.def: Add documentation for each of the options;
1294 2013-09-28 Jan Hubicka <jh@suse.cz>
1296 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
1298 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
1299 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
1300 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
1301 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
1302 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
1304 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
1306 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
1307 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
1308 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
1309 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
1310 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
1311 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
1312 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
1313 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
1314 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
1315 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
1316 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
1317 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
1318 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
1319 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
1320 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
1321 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
1322 gimple.h, godump.c, graphite-clast-to-gimple.c,
1323 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
1324 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
1325 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
1326 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
1327 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
1328 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
1329 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
1330 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
1331 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
1332 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
1333 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
1334 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
1335 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
1336 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
1337 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
1338 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
1339 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
1340 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
1341 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
1342 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
1343 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
1344 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
1345 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
1346 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
1347 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
1348 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1349 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
1350 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
1351 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
1352 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1353 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1354 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
1355 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
1356 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
1357 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
1358 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
1359 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
1360 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
1361 whitespace before "(".
1363 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
1365 * expr.h (extract_bit_field): Remove packedp parameter.
1366 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
1367 from forward declaration.
1368 (store_split_bit_field): Remove packedp arg from calls to
1369 extract_fixed_bit_field.
1370 (extract_bit_field_1): Remove packedp parameter and packedp
1371 argument from recursive calls and calls to extract_fixed_bit_field.
1372 (extract_bit_field): Remove packedp parameter and corresponding
1373 arg to extract_bit_field_1.
1374 (extract_fixed_bit_field): Remove packedp parameter. Remove code
1376 (extract_split_bit_field): Remove packedp arg from call to
1377 extract_fixed_bit_field.
1378 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
1379 (copy_blkmode_from_reg): Likewise.
1380 (copy_blkmode_to_reg): Likewise.
1381 (read_complex_part): Likewise.
1382 (store_field): Likewise.
1383 (expand_expr_real_1): Likewise.
1384 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
1385 to extract_bit_field.
1386 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
1387 call to extract_bit_field.
1388 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
1389 call to extract_bit_field.
1390 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
1391 and special packedp behavior from -fstrict-volatile-bitfields
1394 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1396 * lra-eliminations.c (init_elim_table): Guard value_p.
1398 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
1400 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
1401 DFmode, DImode, and SFmode in the upper VSX registers based on the
1402 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
1403 if -mpower8-vector. Combine -mvsx-timode handling with the rest
1404 of the VSX register handling.
1406 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
1408 (zero_extendsidi2_lfiwzx): Add support for loading into the
1409 Altivec registers with -mpower8-vector. Use wu/wv constraints to
1410 only do VSX memory options on Altivec registers.
1411 (extendsidi2_lfiwax): Likewise.
1412 (extendsfdf2_fpr): Likewise.
1413 (mov<mode>_hardfloat, SF/SD modes): Likewise.
1414 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
1415 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
1416 (movdi_internal64): Likewise.
1418 2013-09-27 Xinliang David Li <davidxl@google.com>
1420 * opts.c (finish_options): Adjust parameters
1421 according to vect cost model.
1422 (common_handle_option): Set dynamic vect cost
1424 targhooks.c (default_add_stmt_cost): Compute stmt cost
1426 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
1427 Use helper function.
1428 * tree-vectorizer.h (unlimited_cost_model): New function.
1429 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
1430 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
1432 (vect_enhance_data_refs_alignment): Ditto.
1433 * flag-types.h: New enum.
1434 * common/config/i386/i386-common.c (ix86_option_init_struct):
1435 No need to initialize vect_cost_model flag.
1436 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
1439 2013-09-27 Diego Novillo <dnovillo@google.com>
1441 * gimple.h (enum ssa_mode): Remove.
1443 2013-09-27 Paulo Matos <pmatos@broadcom.com>
1446 * gcc.dg/pr58463.c: New test.
1448 2013-09-27 Paulo Matos <pmatos@broadcom.com>
1450 * cfgloop.h (number_of_loops): Fix typo in check for null.
1452 2013-09-27 Jakub Jelinek <jakub@redhat.com>
1455 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
1456 are children of outermost saved_cfun's loop, and set it up to
1457 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
1458 if loop != loop0 and SESE region contains bbs that belong to loop0.
1460 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
1462 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
1463 (binary_scale_code_p, get_base_term, get_index_term): New functions.
1464 (set_address_segment, set_address_base, set_address_index)
1465 (set_address_disp): Accept the argument unconditionally.
1466 (baseness): Remove must_be_base_p and must_be_index_p checks.
1467 (decompose_normal_address): Classify as much as possible in the
1470 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
1472 * cse.c (count_reg_usage): Handle INT_LIST.
1473 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
1474 * reginfo.c (reg_scan_mark_refs): Likewise.
1475 * reload1.c (eliminate_regs_1): Likewise.
1477 2013-09-27 Iain Sandoe <iain@codesourcery.com>
1480 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
1483 2013-09-27 Richard Biener <rguenther@suse.de>
1485 PR tree-optimization/58459
1486 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
1487 restriction not propagating into loops.
1489 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
1491 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
1492 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
1493 * doc/tree-ssa.texi (Walking use-def chains): Delete.
1495 2013-09-26 Richard Biener <rguenther@suse.de>
1497 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
1499 2013-09-26 Richard Biener <rguenther@suse.de>
1501 * alias.h (component_uses_parent_alias_set): Rename to ...
1502 (component_uses_parent_alias_set_from): ... this.
1503 * alias.c (component_uses_parent_alias_set): Rename to ...
1504 (component_uses_parent_alias_set_from): ... this and return
1506 (reference_alias_ptr_type_1): Use the result from
1507 component_uses_parent_alias_set_from instead of stripping
1508 components one at a time.
1509 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
1511 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
1513 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
1514 Move prototypes to...
1515 * tree-ssa-ter.h: New File. Move prototypes here.
1516 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
1517 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
1518 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
1519 from is_replaceable_p.
1520 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
1521 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
1522 (process_replaceable): Use ter_is_replaceable_p.
1523 (find_replaceable_in_bb): Use ter_is_replaceable_p.
1524 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
1525 newly refactored ssa_is_replaceable_p.
1526 * cfgexpand.c: Include tree-outof-ssa.h.
1527 * ssaexpand.h: Delete.
1529 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
1531 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
1532 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
1534 (create_gimple_tmp): Delete.
1535 (get_expr_type, build_assign, build_type_cast): Move to...
1536 * gimple-builder.c: New File.
1537 (get_expr_type): Relocate from gimple.c.
1538 (build_assign, build_type_cast): Change to only create ssanames.
1539 * gimple.h: Move prototypes to...
1540 * gimple-builder.h: New File. Here.
1541 * tree-ssa.h: And here.
1542 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
1543 count_uses_and_derefs): Relocate from gimple.c.
1544 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
1545 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
1546 * tree-ssa-math-opts (execute_cse_reciprocals): Use
1547 gimple_replace_ssa_lhs.
1548 * asan.c: Include gimple-builder.h.
1549 * Makefile.in: Add gimple-builder.o.
1551 2013-09-26 Richard Biener <rguenther@suse.de>
1553 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
1554 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
1555 (loe_visit_block): Use gcc_checking_assert.
1556 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
1557 gimple_assign_ssa_name_copy_p.
1558 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
1561 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
1563 * config/rs6000/t-rs6000 (rs6000.o): Remove.
1564 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
1566 2013-09-26 Richard Biener <rguenther@suse.de>
1568 PR tree-optimization/58539
1569 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
1570 the fact that debug statements are not taking part in loop-closed
1573 2013-09-26 Nick Clifton <nickc@redhat.com>
1575 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
1576 time warning message.
1577 (msp430_print_operand_raw): Delete unused letter parameter.
1578 (TARGET_PRINT_OPERAND_ADDRESS): Define.
1579 (msp430_print_operand_address): New function.
1580 (msp430_print_operand): Move address printing code from here to
1582 * config/msp430/msp430.md (movsipsi2): Add comment in generated
1584 (zero_extendpsisi2): Likewise.
1585 (extendpsisi2): New pattern.
1586 (andneghi3): New pattern.
1588 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
1590 * config/aarch64/aarch64.opt (mlra): New option.
1591 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
1592 (TARGET_LRA_P): Define.
1594 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
1596 * expr.c (expand_assignment): Remove obsolete comment.
1598 2013-09-25 Jeff Law <law@redhat.com>
1600 * tree-flow.h (thread_through_all_blocks): Prototype moved into
1601 tree-ssa-threadupdate.h.
1602 (register_jump_thread): Similarly.
1603 * tree-ssa-threadupdate.h: New header file.
1604 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
1605 * tree-vrp.c: Likewise.
1606 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
1607 (thread_around_empty_blocks): Change type of path vector argument to
1608 an edge,type pair from just an edge. Initialize both elements when
1609 appending to a jump threading path. Tweak references to elements
1611 (thread_across_edge): Similarly. Release memory for the elements
1613 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
1614 (dump_jump_thread_path): New function broken out from
1615 register_jump_thread.
1616 (register_jump_thread): Use dump_jump_thread_path. Change type of
1617 path vector entries. Search the path for NULL edges and dump
1618 the path if one is found. Tweak the conversion of path to 3-edge
1619 form to use the block copy type information embedded in the path.
1621 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
1623 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
1625 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
1626 Vladimir Makarov <vmakarov@redhat.com>
1628 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
1629 from the least significant bit.
1630 (strip_address_mutations): Add bitfield operations handling.
1631 (must_be_index_p): Add shifting and rotate operations handling.
1632 (set_address_base): Use must_be_base_p predicate.
1633 (set_address_index): Use must_be_index_p predicate.
1635 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
1636 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1637 Sergey Lega <sergey.s.lega@intel.com>
1638 Anna Tikhonova <anna.tikhonova@intel.com>
1639 Ilya Tocar <ilya.tocar@intel.com>
1640 Andrey Turetskiy <andrey.turetskiy@intel.com>
1641 Ilya Verbin <ilya.verbin@intel.com>
1642 Kirill Yukhin <kirill.yukhin@intel.com>
1643 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1645 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
1647 (ix86_expand_vector_move_misalign): Support new unaligned load and
1648 stores and use new names.
1649 (CODE_FOR_sse2_storedqu): Rename to ...
1650 (CODE_FOR_sse2_storedquv16qi): ... this.
1651 (CODE_FOR_sse2_loaddqu): Rename to ...
1652 (CODE_FOR_sse2_loaddquv16qi): ... this.
1653 (CODE_FOR_avx_loaddqu256): Rename to ...
1654 (CODE_FOR_avx_loaddquv32qi): ... this.
1655 (CODE_FOR_avx_storedqu256): Rename to ...
1656 (CODE_FOR_avx_storedquv32qi): ... this.
1657 * config/i386/i386.md (fpint_logic): New.
1658 * config/i386/sse.md (VMOVE): Extend for AVX512.
1662 (VI_UNALIGNED_LOADSTORE): Ditto.
1663 (sse2_avx_avx512f): Ditto.
1664 (sse2_avx2): Extend for AVX512.
1665 (sse4_1_avx2): Ditto.
1666 (avx2_avx512f): New.
1667 (sse): Extend for AVX512.
1670 (avxsizesuffix): Ditto.
1671 (sseintvecmode): Ditto.
1673 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
1674 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
1675 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
1676 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
1677 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
1678 (<sse2_avx_avx512f>_storedqu<mode): ... this.
1679 (<sse>_movnt<mode>): Replace constraint "x" with "v".
1680 (STORENT_MODE): Extend for AVX512.
1681 (*absneg<mode>2): Replace constraint "x" with "v".
1682 (*mul<mode>3): Ditto.
1683 (*ieee_smin<mode>3): Ditto.
1684 (*ieee_smax<mode>3): Ditto.
1685 (avx_cmp<mode>3): Replace VF with VF_128_256.
1686 (*<sse>_maskcmp<mode>3_comm): Ditto.
1687 (<sse>_maskcmp<mode>3): Ditto.
1688 (<sse>_andnot<mode>3): Extend for AVX512.
1689 (<code><mode>3, anylogic): Replace VF with VF_128_256.
1690 (<code><mode>3, fpint_logic): New.
1691 (*<code><mode>3): Extend for AVX512.
1692 (avx512flogicsuff): New.
1693 (avx512f_<logic><mode>): Ditto.
1694 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
1696 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
1697 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
1698 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
1699 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
1700 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
1701 (xop_vpermil2<mode>3): Ditto.
1702 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
1703 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
1704 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
1705 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
1707 2013-09-25 Tom Tromey <tromey@redhat.com>
1709 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
1710 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
1711 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
1712 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
1713 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
1714 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
1715 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
1716 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
1717 (GRAPHITE_HTAB_H): Remove.
1719 2013-09-25 Tom Tromey <tromey@redhat.com>
1721 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
1723 2013-09-25 Tom Tromey <tromey@redhat.com>
1725 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
1727 2013-09-25 Tom Tromey <tromey@redhat.com>
1729 * config/i386/t-i386 (i386.o): Remove.
1730 (i386-c.o): Use COMPILE and POSTCOMPILE.
1732 2013-09-25 Tom Tromey <tromey@redhat.com>
1734 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
1736 2013-09-25 Tom Tromey <tromey@redhat.com>
1738 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
1739 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
1740 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
1741 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
1742 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
1743 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
1744 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
1745 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
1746 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
1747 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
1748 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
1749 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
1750 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
1751 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
1752 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
1753 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
1754 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
1755 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
1756 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
1757 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
1758 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
1759 (tree-ssa-pre.o, tree-ssa-sccvn.o)
1760 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
1761 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
1762 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
1763 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
1764 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
1765 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
1766 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
1767 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
1768 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
1769 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
1770 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
1771 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
1772 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
1773 (graphite-clast-to-gimple.o, graphite-dependences.o)
1774 (graphite-interchange.o, graphite-poly.o)
1775 (graphite-scop-detection.o, graphite-sese-to-poly.o)
1776 (graphite-optimize-isl.o, tree-vect-loop.o)
1777 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
1778 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
1779 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
1780 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
1781 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
1782 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
1783 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
1784 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
1785 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
1786 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
1787 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
1788 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
1789 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
1790 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
1791 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
1792 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
1793 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
1794 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
1795 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
1796 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
1797 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
1798 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
1799 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
1800 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
1801 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
1802 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
1803 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
1804 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
1805 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
1806 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
1807 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
1808 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
1809 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
1810 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
1811 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
1812 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
1813 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
1814 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
1815 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
1816 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
1817 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
1818 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
1819 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
1820 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
1821 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
1822 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
1823 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
1824 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
1825 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
1826 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
1827 (gcov-dump.o): Remove.
1828 (default-c.o): Use COMPILE and POSTCOMPILE.
1829 (CFLAGS-gcc.o): New variable.
1830 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
1832 2013-09-25 Tom Tromey <tromey@redhat.com>
1834 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
1835 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
1836 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
1837 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
1838 (c-family/c-pragma.o, c-family/c-pretty-print.o)
1839 (c-family/c-semantics.o, c-family/c-ada-spec.o)
1840 (c-family/array-notation-common.o, c-family/stub-objc.o)
1841 (c-family/c-ubsan.o): Remove.
1843 2013-09-25 Tom Tromey <tromey@redhat.com>
1845 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
1847 2013-09-25 Tom Tromey <tromey@redhat.com>
1849 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
1850 to add -DENABLE_SHARED_LIBGCC.
1851 (gcc.o): Don't use subshell.
1853 2013-09-25 Tom Tromey <tromey@redhat.com>
1855 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
1856 * configure.ac: Don't invoke AM_PROG_CC_C_O.
1857 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
1858 * configure, config.in: Rebuild.
1860 2013-09-25 Tom Tromey <tromey@redhat.com>
1862 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
1863 (COMPILE, POSTCOMPILE): New variables.
1864 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
1865 (DEPFILES): New variable.
1866 Include ".Po" files.
1867 * configure.ac: Add checks for dependency checking.
1868 * configure, aclocal.m4: Regenerate.
1870 2013-09-25 Tom Tromey <tromey@redhat.com>
1872 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
1873 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
1875 2013-09-25 Tom Tromey <tromey@redhat.com>
1877 * Makefile.in (generated_files): Add options.h,
1878 target-hooks-def.h, insn-opinit.h,
1879 common/common-target-hooks-def.h, pass-instances.def,
1880 c-family/c-target-hooks-def.h.
1882 2013-09-25 Jeff Law <law@redhat.com>
1884 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
1885 than foo[foo.length () - 1] to access last member in a vec.
1886 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
1888 2013-09-25 Richard Biener <rguenther@suse.de>
1891 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
1893 2013-09-25 Jan Hubicka <jh@suse.cz>
1895 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
1898 2013-09-25 Marek Polacek <polacek@redhat.com>
1901 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
1902 when determining the type name.
1904 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
1906 * config/sh/sh.md: Fix formatting.
1908 2013-09-24 Xinliang David Li <davidxl@google.com>
1910 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
1911 max peel iterations parameter.
1912 * param.def: New parameter.
1913 * doc/invoke.texi: Document New parameter.
1915 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
1917 * gimple-pretty-print.c: Various whitespace tweaks.
1918 * tree-core.h: Likewise.
1919 * tree-pretty-print.c: Likewise.
1920 * tree-ssa-alias.c: Likewise.
1921 * tree-ssa-copy.c: Likewise.
1922 * tree-ssanames.c: Likewise.
1923 * tree-ssanames.h: Likewise.
1924 * tree-vrp.c: Likewise.
1926 2013-09-24 Alan Modra <amodra@gmail.com>
1930 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
1931 for output operands that disallow regs. Don't use EXPAND_WRITE on
1934 2013-09-24 Richard Biener <rguenther@suse.de>
1937 * tree.c (reference_alias_ptr_type): Move ...
1938 * alias.c (reference_alias_ptr_type): ... here and implement
1939 in terms of the new reference_alias_ptr_type_1.
1940 (ref_all_alias_ptr_type_p): New helper.
1941 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
1942 use ref_all_alias_ptr_type_p.
1943 (get_deref_alias_set): Add flag_strict_aliasing check here.
1944 (reference_alias_ptr_type_1): New function, split out from ...
1945 (get_alias_set): ... here.
1946 (alias_ptr_types_compatible_p): New function.
1947 * alias.h (reference_alias_ptr_type): Declare.
1948 (alias_ptr_types_compatible_p): Likewise.
1949 * tree.h (reference_alias_ptr_type): Remove.
1950 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
1951 to compare MEM_REF alias types.
1953 2013-09-24 Richard Biener <rguenther@suse.de>
1955 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
1957 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
1959 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
1960 reload helper function arrays into a single array reg_addr.
1961 (reload_fpr_gpr): Likewise.
1962 (reload_gpr_vsx): Likewise.
1963 (reload_vsx_gpr): Likewise.
1964 (struct rs6000_reg_addr): Likewise.
1965 (reg_addr): Likewise.
1966 (rs6000_debug_reg_global): Change rs6000_vector_reload,
1967 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
1968 (rs6000_init_hard_regno_mode_ok): Likewise.
1969 (rs6000_secondary_reload_direct_move): Likewise.
1970 (rs6000_secondary_reload): Likewise.
1972 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
1973 constraints: wu, ww, and wy. Repurpose wv constraint added during
1974 power8 changes. Put wg constraint in alphabetical order.
1976 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
1977 for future work to add ISA 2.07 VSX single precision support.
1978 (-mvsx-scalar-double): Change default from -1 to 1, update
1979 documentation comment.
1980 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
1981 (-mupper-regs-df): New debug switch to control whether DF values
1982 can go in the traditional Altivec registers.
1983 (-mupper-regs-sf): New debug switch to control whether SF values
1984 can go in the traditional Altivec registers.
1986 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
1988 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
1989 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
1990 Add new constraints, wu/ww/wy. Repurpose wv constraint.
1991 (rs6000_debug_legitimate_address_p): Print if we are running
1992 before, during, or after reload.
1993 (rs6000_secondary_reload): Add a comment.
1994 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
1996 * config/rs6000/constraints.md (wa constraint): Sort w<x>
1997 constraints. Update documentation string.
1998 (wd constraint): Likewise.
1999 (wf constraint): Likewise.
2000 (wg constraint): Likewise.
2001 (wn constraint): Likewise.
2002 (ws constraint): Likewise.
2003 (wt constraint): Likewise.
2004 (wx constraint): Likewise.
2005 (wz constraint): Likewise.
2006 (wu constraint): New constraint for ISA 2.07 SFmode scalar
2008 (ww constraint): Likewise.
2009 (wy constraint): Likewise.
2010 (wv constraint): Repurpose ISA 2.07 constraint that we not used in
2011 the previous submissions.
2012 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
2014 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
2016 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
2017 (REG_BR_PROB): Say that the probability is stored in an int_list.
2018 * reg-notes.def: Update commentary to mention INT_LIST.
2019 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
2020 (INT_LIST): New rtx.
2021 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
2022 * rtlanal.c (int_reg_note_p): New function.
2023 (alloc_reg_note): Assert that the note does not have an int argument.
2024 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
2025 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
2026 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
2027 rather than an INSN_LIST. Handle INT_LIST.
2028 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
2029 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
2030 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
2031 Manipulate them as ints rather than rtxes.
2032 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
2033 * regmove.c (copy_src_to_dest): Likewise.
2034 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
2036 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
2037 into the cases that need it.
2038 * config/arm/arm.c (arm_unwind_emit): Likewise.
2040 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
2041 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
2042 * loop-doloop.c (add_test, doloop_modify): Likewise.
2043 * loop-unswitch.c (compare_and_jump_seq): Likewise.
2044 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
2045 * predict.c (combine_predictions_for_insn): Likewise.
2046 * print-rtl.c (print_rtx): Handle INT_LIST.
2047 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
2048 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
2049 * config/arm/arm.c (emit_unlikely_jump): Likewise.
2050 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
2051 (ix86_split_fp_branch, predict_jump): Likewise.
2052 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
2053 * config/sh/sh.c (expand_cbranchsi4): Likewise.
2054 * config/spu/spu.c (ea_load_store_inline): Likewise.
2056 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
2057 value of a REG_BR_PROB note.
2058 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
2059 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
2060 * emit-rtl.c (try_split): Likewise.
2061 * predict.c (br_prob_note_reliable_p): Likewise.
2062 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
2063 * reorg.c (mostly_true_jump): Likewise.
2064 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
2065 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
2066 * config/i386/i386.c (ix86_print_operand): Likewise.
2067 * config/ia64/ia64.c (ia64_print_operand): Likewise.
2068 * config/mmix/mmix.c (mmix_print_operand): Likewise.
2069 * config/rs6000/rs6000.c (output_cbranch): Likewise.
2070 * config/s390/s390.c (s390_expand_tbegin): Likewise.
2071 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
2072 * config/sparc/sparc.c (output_cbranch): Likewise.
2073 * config/spu/spu.c (get_branch_target): Likewise.
2074 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
2075 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
2077 2013-09-23 Jan Hubicka <jh@suse.cz>
2079 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
2081 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
2082 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
2084 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
2085 (gimple_fold_call): Dump inconsistent devirtualizations; add
2086 sanity check for type based devirtualizations.
2087 * ipa-prop.c: Include ipa-utils.h
2088 (ipa_intraprocedural_devirtualization): Add sanity check.
2089 (try_make_edge_direct_virtual_call): Likewise.
2091 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
2093 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
2094 assignment statements.
2096 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
2098 * gimple-pretty-print.c (dump_ssaname_info): New function.
2099 (dump_gimple_phi): Call it.
2100 (pp_gimple_stmt_1): Likewise.
2101 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
2102 (range_info_def): Declare.
2103 * tree-pretty-print.c (pp_double_int): New function.
2104 (dump_generic_node): Call it.
2105 * tree-pretty-print.h (pp_double_int): Declare.
2106 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
2107 * tree-ssanames.h (range_info_def): New structure.
2108 (value_range_type): Move definition here.
2109 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
2111 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
2113 (set_range_info): New function.
2114 (get_range_info): Likewise.
2115 (duplicate_ssa_name_range_info): Likewise.
2116 (duplicate_ssa_name_fn): Check pointer type and call
2117 duplicate_ssa_name_range_info.
2118 * tree-ssa-copy.c (fini_copy_prop): Likewise.
2119 * tree-vrp.c (value_range_type): Remove definition, now in
2121 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
2122 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
2123 (SSA_NAME_RANGE_INFO): New macro.
2125 2013-09-23 Richard Biener <rguenther@suse.de>
2127 PR tree-optimization/58464
2128 * tree-ssa-pre.c (phi_trans_lookup): Remove.
2129 (phi_trans_add): Change to add conditionally on being not present.
2130 (phi_translate_1): Remove recursion detection here.
2131 (phi_translate): Pre-seed the cache with NULL to catch
2132 recursion here in a more generic way.
2133 (bitmap_find_leader): Adjust comment.
2134 (get_representative_for): Dump value-numbers.
2135 (create_expression_by_pieces): Likewise.
2136 (insert_into_preds_of_block): Likewise.
2138 2013-09-23 Christian Bruel <christian.bruel@st.com>
2141 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
2142 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
2144 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
2147 2013-09-20 Renlin Li <renlin.li@arm.com>
2149 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
2151 (aarch64_expand_epilogue): Likewise.
2152 (aarch64_legitimize_reload_address): Likewise.
2154 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
2156 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
2157 NULL_TREE before pushing them onto the vector. Likewise for labels.
2159 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
2161 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
2162 comparison operators when -fno-trapping-math is in effect.
2163 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
2164 comparison operators in TFmode and assert that unsupported operators
2166 (ia64_print_operand): Likewise.
2168 2013-09-21 Jan Hubicka <jh@suse.cz>
2170 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
2171 (sse_typeless_stores): Enable for core
2172 (sse_load0_by_pxor): Likewise.
2173 (four_jump_limit): Disable for core.
2174 (pad_returns): Likewise.
2175 (avoid_vector_decode): Likewise.
2176 (fuse_cmp_and_branch): Enable for cores.
2177 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
2179 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
2182 * config/pa/pa.c (pa_option_override): Disable auto increment and
2183 decrement instructions until reload is completed.
2185 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
2186 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
2187 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
2189 2013-09-20 DJ Delorie <dj@redhat.com>
2190 Nick Clifton <nickc@redhat.com>
2192 * config/rl78/rl78.c: Various whitespace and comment tweaks.
2193 (need_to_save): Save bank 0 on interrupts.
2194 (characterize_address): Strip far address wrappers.
2195 (rl78_as_legitimate_address): Likewise.
2196 (transcode_memory_rtx): Likewise.
2197 (rl78_peep_movhi_p): Disable this peephole after devirt.
2198 (rl78_propogate_register_origins): Forget all origins when a
2200 * config/rl78/rl78-virt.md: Various whitespace tweaks.
2201 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
2203 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
2204 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
2205 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
2207 * config/rl78/constraints.md: Various whitespace and paren tweaks.
2209 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
2211 * config/pa/pa.md: In "scc" insn patterns, change output template to
2212 handle const0_rtx in reg_or_0_operand operands.
2214 2013-09-20 Martin Husemann <martin@NetBSD.org>
2217 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
2218 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
2220 2013-09-20 Richard Biener <rguenther@suse.de>
2223 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
2224 remembering SSA name version and block index.
2225 (new_scev_info_str): Adjust.
2226 (hash_scev_info): Likewise. Also hash the block index.
2227 (eq_scev_info): Adjust.
2228 (find_var_scev_info): Likewise.
2229 (struct instantiate_cache_entry): Remove.
2230 (struct instantiate_cache_type): Use a htab to map name, block
2232 (instantiate_cache_type::~instantiate_cache_type): Adjust.
2233 (get_instantiated_value_entry): Likewise.
2234 (hash_idx_scev_info, eq_idx_scev_info): New functions.
2235 (instantiate_scev_name): Adjust.
2237 2013-09-20 Jeff Law <law@redhat.com>
2239 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
2241 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
2243 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
2244 Call aarch64_simd_expand_args to update op[argc].
2246 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
2248 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
2251 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
2253 * gengtype.c (file_rules): Added rule for *.cc files.
2254 (get_output_file_with_visibility): Give fatal message when no
2257 2013-09-20 Renlin Li <renlin.li@arm.com>
2259 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
2260 (aarch64_expand_epilogue): Likewise.
2261 (aarch64_legitimize_reload_address): Likewise.
2263 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
2266 * expr.c (expand_assignment): Remove misalignp code path.
2268 2013-09-20 Marek Polacek <polacek@redhat.com>
2271 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
2272 TYPE_PRECISION. Add asserts.
2274 2013-09-20 Richard Biener <rguenther@suse.de>
2276 PR tree-optimization/58453
2277 * tree-loop-distribution.c (distribute_loop): Apply the cost
2278 model for -ftree-loop-distribute-patterns, too.
2280 2013-09-20 Richard Biener <rguenther@suse.de>
2283 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
2284 make type comparison less strict.
2286 2013-09-20 Alan Modra <amodra@gmail.com>
2288 * configure: Regenerate.
2289 * aclocal.m4: Regenerate.
2291 2013-09-20 Marek Polacek <polacek@redhat.com>
2294 * doc/extend.texi: Document that attribute used is meant to be used
2295 on variables with static storage duration.
2297 2013-09-19 Jakub Jelinek <jakub@redhat.com>
2299 PR tree-optimization/58472
2300 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
2301 simd_lane_access set inv_p = false.
2302 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
2303 the simduid magic VAR_DECL.
2305 2013-09-19 Jan Hubicka <jh@suse.cz>
2307 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
2309 2013-09-17 Jeff Law <law@redhat.com>
2311 * tree-ssa-dom.c (record_temporary_equivalences): New function
2312 split out of dom_opt_dom_walker::after_dom_children.
2313 (dom_opt_dom_walker::thread_across_edge): Move common code
2314 in here from dom_opt_dom_walker::after_dom_children.
2315 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
2317 2013-09-19 Jan Hubicka <jh@suse.cz>
2319 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
2320 (TARGET_GENERIC): Use PROCESOR_GENERIC
2321 (enum processor_type): Unify generic32 and 64.
2322 * i386.md (cpu): Likewise.
2323 * x86-tune.def (use_leave): Enable for generic32.
2324 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
2325 * athlon.md: Change generic64 to generic in all occurences.
2326 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
2327 (ix86_target_macros_internal): Likewise.
2328 * driver-i386.c (host_detect_local_cpu): Likewise.
2329 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
2331 (generic_memcpy, generic_memset, generic_cost): This one.
2332 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
2333 (m_GENERIC32, m_GENERIC64): Remove.
2334 (m_GENERIC): Turn into one flag.
2335 (processor_target): Unify generic tunnings.
2336 (ix86_option_override_internal): Replace generic32/64 by generic.
2337 (ix86_issue_rate): Likewise.
2338 (ix86_adjust_cost): Likewise.
2340 2013-09-19 Jan Hubicka <jh@suse.cz>
2342 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
2343 of speculative flag.
2345 2013-09-19 Jakub Jelinek <jakub@redhat.com>
2347 * omp-low.c (expand_omp_sections): Always pass len - 1 to
2348 GOMP_sections_start, even if !exit_reachable.
2350 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
2352 * lra-constraints.c (need_for_all_save_p): Use macro
2353 HARD_REGNO_CALL_PART_CLOBBERED.
2354 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
2355 set up pseudo conflict hard regs.
2357 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
2360 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
2363 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
2365 PR rtl-optimization/58438
2366 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
2367 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
2368 from previous subpasses.
2370 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
2372 * arm.c (arm_get_frame_offsets): Validate architecture supports
2373 LDRD/STRD before accepting the tuning preference.
2374 (arm_expand_prologue): Likewise.
2375 (arm_expand_epilogue): Likewise.
2377 2013-09-18 Richard Biener <rguenther@suse.de>
2379 PR tree-optimization/58417
2380 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
2381 have chrecs with symbols defined in the loop as operands.
2382 (chrec_fold_multiply): Likewise.
2383 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
2384 parameters before folding binary operations.
2385 (struct instantiate_cache_entry_hasher): Remove.
2386 (struct instantiate_cache_type): Use a pointer-map.
2387 (instantiate_cache_type::instantiate_cache_type): New function.
2388 (instantiate_cache_type::get): Likewise.
2389 (instantiate_cache_type::set): Likewise.
2390 (instantiate_cache_type::~instantiate_cache_type): Adjust.
2391 (get_instantiated_value_entry): Likewise.
2392 (global_cache): New global.
2393 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
2394 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
2395 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
2396 (instantiate_scev_name): Adjust.
2397 (instantiate_scev): Construct global instead of local cache.
2398 (resolve_mixers): Likewise.
2400 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
2401 Paolo Carlini <paolo.carlini@oracle.com>
2404 * doc/implement-cxx.texi: Fix references to the C++ standards.
2406 2013-09-18 Jakub Jelinek <jakub@redhat.com>
2408 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
2409 * tree-vect-data-refs.c (vect_analyze_data_refs): For
2410 simd_lane_access drs, update also DR_ALIGNED_TO.
2412 2013-09-18 Marek Polacek <polacek@redhat.com>
2415 * doc/extend.texi: Document no_sanitize_undefined attribute.
2416 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
2417 no_sanitize_undefined attribute.
2419 2013-09-18 Nick Clifton <nickc@redhat.com>
2421 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
2422 (ASM_DECLARE_FUNCTION_NAME): Define.
2424 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
2426 * compare-elim.c (find_comparison_dom_walker): New class
2427 (find_comparisons_in_bb): Rename to
2428 find_comparison_dom_walker::before_dom_children
2429 (find_comparisons): Adjust
2430 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
2432 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
2433 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
2434 (init_walk_dominator_tree): Remove declaration.
2435 (fini_walk_dominator_tree): Remove declaration.
2436 * fwprop.c (single_def_use_dom_walker): New class
2437 (single_def_use_enter_block): Convert to
2438 single_def_use_dom_walker::before_dom_children.
2439 (single_def_use_leave_block): Convert to
2440 single_def_use_dom_walker::after_dom_children.
2441 (build_single_def_use_links): Adjust.
2442 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
2444 (find_candidates_in_block): Convert to
2445 find_candidates_dom_walker::before_dom_children.
2446 (execute_strength_reduction): Adjust.
2447 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
2448 (sese_dom_walker): New class.
2449 (sese_dom_walker::sese_dom_walker): New constructor.
2450 (sese_dom_walker::~sese_dom_walker): New destructor.
2451 (build_sese_conditions_before): Convert to
2452 sese_dom_walker::before_dom_children.
2453 (build_sese_conditions_after): Convert to
2454 sese_dom_walker::after_dom_children.
2455 (build_poly_scop): Adjust
2456 * tree-into-ssa.c (rewrite_dom_walker): New class
2457 (rewrite_enter_block): Convert to
2458 rewrite_dom_walker::before_dom_children.
2459 (rewrite_leave_block): Convert to
2460 rewrite_dom_walker::after_dom_children.
2461 (rewrite_update_dom_walker): New class.
2462 (rewrite_update_enter_block): Convert to
2463 rewrite_update_dom_walker::before_dom_children.
2464 (rewrite_update_leave_block): Convert to
2465 rewrite_update_dom_walker::after_dom_children.
2466 (rewrite_blocks, rewrite_into_ssa): Adjust.
2467 (mark_def_dom_walker): New class.
2468 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
2469 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
2470 (mark_def_sites_blocks): Convert to
2471 mark_def_dom_walker::before_dom_children.
2472 (mark_def_site_blocks): Remove.
2473 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
2474 (tree_ssa_dominator_optimize): Adjust.
2475 (dom_thread_across_edge): Convert to method
2476 dom_opt_dom_walker::thread_across_edge.
2477 (dom_opt_enter_block): Convert to member function
2478 dom_opt_dom_walker::before_dom_children.
2479 (dom_opt_leave_block): Convert to member function
2480 dom_opt_dom_walker::after_dom_children.
2481 * tree-ssa-dse.c (dse_dom_walker): New class.
2482 (dse_enter_block): Convert to member function
2483 dse_dom_walker::before_dom_children.
2484 (tree_ssa_dse): Adjust.
2485 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
2486 (determine_invariantness_stmt): Convert to method
2487 invariantness_dom_walker::before_dom_children.
2488 (determine_invariantness): Remove
2489 (move_computations_dom_walker): New class.
2490 (move_computations_stmt): Convert to method
2491 move_computations_dom_walker::before_dom_children.
2492 (move_computations, tree_ssa_lim): Adjust.
2493 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
2494 (nt_init_block): Make method
2495 notrappping_dom_walker::before_dom_children.
2496 (nt_fini_block): Make
2497 method nontrapping_dom_walker::after_dom_children.
2498 (get_non_trapping): Adjust.
2499 * tree-ssa-pre.c (eliminate_dom_walker): New class.
2500 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
2501 (eliminate_leave_block): Make method.
2502 eliminate_dom_walker::after_dom_children.
2504 * tree-ssa-strlen.c (strlen_dom_walker): New class.
2505 (strlen_enter_block): Make method
2506 strlen_dom_walker::before_dom_children.
2507 (strlen_leave_block): Make
2508 method strlen_dom_walker::after_dom_children.
2509 (tree_ssa_strlen): Adjust.
2510 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
2511 (tree_ssa_uncprop): Adjust.
2512 (uncprop_leave_block): Make method
2513 uncprop_dom_walker::after_dom_children.
2514 (uncprop_leave_block): Make method
2515 uncprop_dom_walker::before_dom_children.
2517 2013-09-18 Bin Cheng <bin.cheng@arm.com>
2519 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
2520 branch in same basic block. Check both src and dest of the move insn.
2522 2013-09-17 Nick Clifton <nickc@redhat.com>
2524 * config/rl78/rl78-real.md (bf): New pattern.
2526 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
2527 (rl78_print_operand): Do not put a # before a %B.
2528 * config/rl78/rl78.opt: Tweak doc strings.
2530 2013-09-17 DJ Delorie <dj@redhat.com>
2532 * config/rl78/constraints.md (Wcv): Allow up to $r31.
2533 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
2534 (rl78_option_override): Likewise, if -mallregs.
2535 (is_virtual_register): Likewise.
2536 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
2537 (REGNO_OK_FOR_BASE_P): Likewise.
2538 * config/rl78/rl78.opt (-mallregs): New.
2540 2013-09-17 Nick Clifton <nickc@redhat.com>
2542 * config/rl78/rl78.c (need_to_save): Change return type to bool.
2543 For interrupt functions: save all call clobbered registers if the
2544 interrupt handler is not a leaf function.
2545 (rl78_expand_prologue): Always recompute the frame information.
2546 For interrupt functions: only select bank 0 if one of the bank 0
2547 registers is going to be psuhed.
2549 2013-09-17 DJ Delorie <dj@redhat.com>
2551 * config/rl78/constraints.md: For each W* constraint, rename to C*
2552 and create a W* constraint that checks for an optional ES: prefix
2554 * config/rl78/rl78.md (UNS_ES_ADDR): New.
2555 (es_addr): New. Used to wrap far addresses.
2556 * config/rl78/rl78-protos.h (rl78_es_addr): New.
2557 (rl78_es_base): New.
2558 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
2559 wrapped far addresses.
2560 (rl78_print_operand_1): Unwrap far addresses before processing.
2561 (rl78_lo16): Wrap far addresses in unspecs.
2562 (rl78_es_addr): New.
2563 (rl78_es_base): New.
2564 (insn_ok_now): Check for not-yet-wrapped far addresses.
2565 (transcode_memory_rtx): Properly re-wrap far addresses.
2567 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
2569 * config/sparc/t-rtems: Add leon3 multilibs.
2571 2013-09-17 Cong Hou <congh@google.com>
2573 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
2574 when checking the dot production pattern. The type of rhs operand
2575 of multiply is now checked correctly.
2577 2013-09-17 Jeff Law <law@redhat.com>
2579 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
2580 edge implied equivalences into successor phis.
2581 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
2582 here from tree-ssa-threadedge.c.
2583 (mark_threaded_blocks): When threading through a joiner, if both
2584 successors of the joiner's clone reach the same block, verify the
2585 PHI arguments are equal. If not, cancel the jump threading request.
2586 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
2587 tree-ssa-threadupdate.c
2588 (thread_across_edge): Don't check PHI argument equality when
2589 threading through joiner block here.
2591 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
2593 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
2594 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
2595 (warn_uninit, warn_uninitialized_vars,
2596 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
2597 Move to tree-ssa-uninit.c.
2598 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
2599 (has_undefined_value_p): New. Pass dependant parts of
2600 ssa_undefined_value_p.
2601 (uninit_undefined_value_p): Use has_undefined_value_p.
2602 (warn_uninit, warn_uninitialized_vars,
2603 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
2604 Move from tree-ssa.c.
2605 * tree-ssa.h: Adjust prototypes.
2607 2013-09-17 Jan Hubicka <jh@suse.cz>
2610 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
2611 * ipa-inline.c (can_inline_edge_p): Do not downgrade
2612 FUNCTION_NOT_OPTIMIZED.
2613 * ipa-inline-analysis.c (compute_inline_parameters): Function
2614 not optimized is not inlinable unless it is alwaysinline.
2615 (inline_analyze_function): Force calls in not optimized
2616 function not inlinable.
2618 2013-09-17 Jan Hubicka <jh@suse.cz>
2621 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
2623 * ipa.c (function_and_variable_visibility): Likewise.
2624 * ipa-profile.c (ipa_profile): Likewise.
2626 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
2629 * cgraph.c (cgraph_function_body_availability): Check for ifunc
2630 attribute, and don't inline the resolver in this case.
2632 2013-09-17 Teresa Johnson <tejohnson@google.com>
2634 * coverage.c (get_coverage_counts): Add missing newline.
2636 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2638 PR tree-optimization/58088
2639 * fold-const.c (mask_with_trailing_zeros): New function.
2640 (fold_binary_loc): Make sure we don't recurse infinitely
2641 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
2642 Use mask_with_trailing_zeros where appropriate.
2644 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
2646 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
2647 of 'prev' var to get better distance estimation.
2649 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
2651 * tree-inline.h (struct copy_body_data): Add transform_parameter.
2652 * tree-inline.c (is_parameter_of): New predicate.
2653 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
2654 a parameter has been remapped.
2655 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
2656 (optimize_inline_calls): Initialize transform_parameter.
2657 (copy_gimple_seq_and_replace_locals): Likewise.
2658 (tree_function_versioning): Likewise.
2659 (maybe_inline_call_in_expr): Likewise.
2661 2013-09-17 Nick Clifton <nickc@redhat.com>
2663 * config/msp430/msp430-protos.h: Add prototypes for new functions.
2664 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
2666 (is_attr_func): New function.
2667 (msp430_is_interrupt_func): New function.
2668 (is_naked_func): New function.
2669 (is_reentrant_func): New function.
2670 (is_critical_func): New function.
2671 (msp430_start_function): Add annotations for function attributes.
2672 (msp430_attr): New function.
2673 (msp430_attribute_table): New.
2674 (msp430_function_section): New function.
2675 (TARGET_ASM_FUNCTION_SECTION): Define.
2676 (msp430_builtin): New enum.
2677 (msp430_init_builtins): New function.
2678 (msp430_builtin_devl): New function.
2679 (msp430_expand_builtin): New function.
2680 (TARGET_INIT_BUILTINS): Define.
2681 (TARGET_EXPAND_BUILTINS): Define.
2682 (TARGET_BUILTIN_DECL): Define.
2683 (msp430_expand_prologue): Add support for naked, interrupt,
2684 critical and reentrant functions.
2685 (msp430_expand_epilogue): Likewise.
2686 (msp430_print_operand): Handle 'O' character.
2687 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
2689 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
2690 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
2691 (pushm): Use a 'n' rather than an 'i' constraint.
2692 (msp_return): Add generation of the interrupt return instruction.
2693 (disable_interrupts): New pattern.
2694 (enable_interrupts): New pattern.
2695 (push_intr_state): New pattern.
2696 (pop_intr_state): New pattern.
2697 (bic_SR): New pattern.
2698 (bis_SR): New pattern.
2699 * doc/extend.texi: Document MSP430 function attributes and builtin
2702 2013-09-17 Richard Biener <rguenther@suse.de>
2704 PR tree-optimization/58432
2705 * tree-loop-distribution.c (tree_loop_distribution): Also
2706 scan PHIs for outside loop uses and seed a partition from them.
2708 2013-09-17 Bin Cheng <bin.cheng@arm.com>
2710 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
2711 (restructure_reference): Call backtrace_base_for_ref.
2713 2013-09-17 Alan Modra <amodra@gmail.com>
2716 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
2719 2013-09-16 DJ Delorie <dj@redhat.com>
2721 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
2722 vregs location for RL78/G10.
2723 (rl78_expand_prologue): Avoid SEL on G10.
2724 (rl78_expand_epilogue): Likewise.
2725 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
2726 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
2727 __RL78_G10__ when appropriate.
2728 (ASM_SPEC): Pass -mg10 along to the assembler.
2729 * config/rl78/rl78.md (sel_rb): Disable for G10.
2730 * config/rl78/rl78.opt: Add -mg10 option.
2731 * config/rl78/t-rl78: Add -mg10 multilib.
2733 2013-09-16 Xinliang David Li <davidxl@google.com>
2735 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
2736 * omp-low.c (omp_max_vf): Ditto.
2737 (expand_omp_simd): Ditto.
2738 * tree-vectorizer.c (vectorize_loops): Ditto.
2739 (gate_vect_slp): Ditto.
2740 (gate_increase_alignment): Ditto.
2741 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
2742 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
2743 (gate_tree_vectorize): Name change.
2744 (tree_vectorize): Ditto.
2745 (pass_vectorize::gate): Call new function.
2746 (pass_vectorize::execute): Ditto.
2747 * opts.c: O3 default setting change.
2748 (finish_options): Check new flag.
2749 * doc/invoke.texi: Document new flags.
2750 * common.opt: New flags.
2752 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
2754 * doc/tm.texi.in (Cond Exec Macros): Remove node.
2755 (Condition Code): Don't reference it.
2756 * doc/tm.texi: Regenerate.
2758 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
2761 * lra-constraints.c (undo_optional_reloads): Consider all optional
2762 reload even if it did not get a hard reg.
2764 2013-09-16 Teresa Johnson <tejohnson@google.com>
2766 * dumpfile.c (dump_loc): Remove newline emission.
2767 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
2768 emission to dump_printf_loc calls where missing.
2769 (vect_mark_for_runtime_alias_test): Ditto.
2770 (vect_analyze_data_ref_dependence): Ditto.
2771 (vect_analyze_data_ref_dependences): Ditto.
2772 (vect_slp_analyze_data_ref_dependence): Ditto.
2773 (vect_slp_analyze_data_ref_dependences): Ditto.
2774 (vect_compute_data_ref_alignment): Ditto.
2775 (vect_update_misalignment_for_peel): Ditto.
2776 (vect_verify_datarefs_alignment): Ditto.
2777 (vector_alignment_reachable_p): Ditto.
2778 (vect_get_data_access_cost): Ditto.
2779 (vect_enhance_data_refs_alignment): Ditto.
2780 (vect_find_same_alignment_drs): Ditto.
2781 (vect_analyze_data_refs_alignment): Ditto.
2782 (vect_analyze_group_access): Ditto.
2783 (vect_analyze_data_ref_access): Ditto.
2784 (vect_analyze_data_ref_accesses): Ditto.
2785 (vect_prune_runtime_alias_test_list): Ditto.
2786 (vect_analyze_data_refs): Ditto.
2787 (vect_create_addr_base_for_vector_ref): Ditto.
2788 (vect_create_data_ref_ptr): Ditto.
2789 (vect_grouped_store_supported): Ditto.
2790 (vect_grouped_load_supported): Ditto.
2791 * value-prof.c (check_counter): Ditto.
2792 (check_ic_target): Ditto.
2793 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
2794 (vect_recog_widen_mult_pattern): Ditto.
2795 (vect_recog_widen_sum_pattern): Ditto.
2796 (vect_recog_over_widening_pattern): Ditto.
2797 (vect_recog_widen_shift_pattern): Ditto.
2798 (vect_recog_rotate_pattern): Ditto.
2799 (vect_recog_vector_vector_shift_pattern): Ditto.
2800 (vect_recog_divmod_pattern): Ditto.
2801 (vect_recog_mixed_size_cond_pattern): Ditto.
2802 (vect_recog_bool_pattern): Ditto.
2803 (vect_pattern_recog_1): Ditto.
2804 (vect_pattern_recog): Ditto.
2805 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
2806 (vect_is_simple_iv_evolution): Ditto.
2807 (vect_analyze_scalar_cycles_1): Ditto.
2808 (vect_get_loop_niters): Ditto.
2809 (vect_analyze_loop_1): Ditto.
2810 (vect_analyze_loop_form): Ditto.
2811 (vect_analyze_loop_operations): Ditto.
2812 (vect_analyze_loop_2): Ditto.
2813 (vect_analyze_loop): Ditto.
2814 (report_vect_op): Ditto.
2815 (vect_is_slp_reduction): Ditto.
2816 (vect_is_simple_reduction_1): Ditto.
2817 (vect_get_known_peeling_cost): Ditto.
2818 (vect_estimate_min_profitable_iters): Ditto.
2819 (vect_model_reduction_cost): Ditto.
2820 (vect_model_induction_cost): Ditto.
2821 (get_initial_def_for_induction): Ditto.
2822 (vect_create_epilog_for_reduction): Ditto.
2823 (vectorizable_reduction): Ditto.
2824 (vectorizable_induction): Ditto.
2825 (vectorizable_live_operation): Ditto.
2826 (vect_loop_kill_debug_uses): Ditto.
2827 (vect_transform_loop): Ditto.
2828 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
2829 (vect_stmt_relevant_p): Ditto.
2830 (process_use): Ditto.
2831 (vect_mark_stmts_to_be_vectorized): Ditto.
2832 (vect_model_simple_cost): Ditto.
2833 (vect_model_promotion_demotion_cost): Ditto.
2834 (vect_model_store_cost): Ditto.
2835 (vect_get_store_cost): Ditto.
2836 (vect_model_load_cost): Ditto.
2837 (vect_get_load_cost): Ditto.
2838 (vect_init_vector_1): Ditto.
2839 (vect_get_vec_def_for_operand): Ditto.
2840 (vect_finish_stmt_generation): Ditto.
2841 (vectorizable_call): Ditto.
2842 (vectorizable_conversion): Ditto.
2843 (vectorizable_assignment): Ditto.
2844 (vectorizable_shift): Ditto.
2845 (vectorizable_operation): Ditto.
2846 (vectorizable_store): Ditto.
2847 (vectorizable_load): Ditto.
2848 (vectorizable_condition): Ditto.
2849 (vect_analyze_stmt): Ditto.
2850 (vect_transform_stmt): Ditto.
2851 (vect_is_simple_use): Ditto.
2852 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
2853 (vect_can_advance_ivs_p): Ditto.
2854 (vect_update_ivs_after_vectorizer): Ditto.
2855 (vect_do_peeling_for_loop_bound): Ditto.
2856 (vect_gen_niters_for_prolog_loop): Ditto.
2857 (vect_update_inits_of_drs): Ditto.
2858 (vect_create_cond_for_alias_checks): Ditto.
2859 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
2860 (vect_build_slp_tree_1): Ditto.
2861 (vect_supported_load_permutation_p): Ditto.
2862 (vect_analyze_slp_instance): Ditto.
2863 (vect_analyze_slp): Ditto.
2864 (vect_make_slp_decision): Ditto.
2865 (vect_detect_hybrid_slp): Ditto.
2866 (vect_bb_vectorization_profitable_p): Ditto.
2867 (vect_slp_analyze_bb_1): Ditto.
2868 (vect_update_slp_costs_according_to_vf): Ditto.
2869 (vect_get_mask_element): Ditto.
2870 (vect_transform_slp_perm_load): Ditto.
2871 (vect_schedule_slp_instance): Ditto.
2872 (vect_schedule_slp): Ditto.
2873 (vect_slp_transform_bb): Ditto.
2874 * profile.c (read_profile_edge_counts): Ditto.
2875 (compute_branch_probabilities): Ditto.
2876 * coverage.c (get_coverage_counts): Ditto.
2878 2013-09-16 Diego Novillo <dnovillo@google.com>
2880 * tree-core.h: Add missing comment lines from refactoring of tree.h.
2882 2013-09-16 Jan Hubicka <jh@suse.cz>
2884 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
2885 abstract functions; for static functions check the presence of body.
2887 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
2889 * config/aarch64/aarch64-simd-builtins.def (fma): New.
2890 * config/aarch64/aarch64-simd.md
2891 (aarch64_mla_elt<mode>): New.
2892 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
2893 (aarch64_mls_elt<mode>): Likewise.
2894 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
2895 (aarch64_fma4_elt<mode>): Likewise.
2896 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
2897 (aarch64_fma4_elt_to_128v2df): Likewise.
2898 (aarch64_fma4_elt_to_64df): Likewise.
2899 (fnma<mode>4): Likewise.
2900 (aarch64_fnma4_elt<mode>): Likewise.
2901 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
2902 (aarch64_fnma4_elt_to_128v2df): Likewise.
2903 (aarch64_fnma4_elt_to_64df): Likewise.
2904 * config/aarch64/iterators.md (VDQSF): New.
2905 * config/aarch64/arm_neon.h
2906 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
2907 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
2909 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
2911 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
2912 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
2913 (aarch64_mul3_elt_to_128df): Likewise.
2914 (aarch64_mul3_elt_to_64v2df): Likewise.
2915 * config/aarch64/iterators.md (VEL): Also handle DFmode.
2917 (VMUL_CHANGE_NLANES) Likewise.
2920 * config/aarch64/arm_neon.h
2921 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
2923 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
2925 * config/aarch64/arm_neon.h
2926 (vcvtx_high_f32_f64): Fix parameters.
2928 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2929 Uros Bizjak <ubizjak@gmail.com>
2931 * config/alpha.c: Include tree-ssanames.h.
2933 2013-09-16 Richard Biener <rguenther@suse.de>
2935 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
2936 (dot_rdg_1): Handle control_dd.
2937 (create_edge_for_control_dependence): New function.
2938 (create_rdg_edges): Add control dependences if asked for.
2939 (build_rdg): Likewise.
2940 (generate_loops_for_partition): If there are not necessary
2941 control stmts remove all their dependencies.
2942 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
2943 (distribute_loop): Pass on control dependences.
2944 (tree_loop_distribution): Compute control dependences and remove
2945 restriction on number of loop nodes.
2947 2013-09-16 Jakub Jelinek <jakub@redhat.com>
2949 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
2952 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
2954 * cse.c (try_const_anchors): Punt on CC modes.
2956 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2958 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
2960 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
2963 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
2966 2013-09-14 Iain Sandoe <iain@codesourcery.com>
2969 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
2971 (darwin_objc1_section): Likewise.
2972 (darwin_file_end): Emit Image Info section when required.
2974 2013-09-14 Jan Hubicka <jh@suse.cz>
2976 * tree-into-ssa.c (gate_into_ssa): New.
2977 (pass_data_build_ssa): Use it.
2978 * cgraph.h (expand_thunk): Update prototype.
2979 * cgraphunit.c (analyze_function): Expand thunks early.
2980 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
2981 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
2982 set lowered flag; do not add new function.
2983 (assemble_thunks_and_aliases): Update.
2984 * tree-ssa.c (gate_init_datastructures): New gate.
2985 (pass_data_init_datastructures): Use it.
2987 2013-09-14 Iain Sandoe <iain@codesourcery.com>
2990 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
2991 xmm register set described in the psABI.
2993 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
2995 * dwarf2out.c (should_emit_struct_debug): Add check
2996 for type_decl variable is not NULL.
2998 2013-09-13 Jacek Caban <jacek@codeweavers.com>
3000 * config.gcc: Use new winnt-c.c target hooks
3001 * config/t-winnt: New file
3002 * config/winnt-c.c: New file
3003 * doc/tm.texi.in: Document new hook
3004 * doc/tm.texi: Regenerated
3006 2013-09-13 Jan Hubicka <jh@suse.cz>
3009 * ipa-inline.c (check_callers): New function.
3010 (check_caller_edge): Remove.
3011 (want_inline_function_to_all_callers_p): Also permit alises that are
3013 (inline_to_all_callers): Terminate the walk when devirtualization
3014 introduce new calls.
3016 2013-09-13 Jan Hubicka <jh@suse.cz>
3018 * ipa-inline-analysis.c (struct growth_data): Add node.
3019 (do_estimate_growth_1): Fix detection of recursion.
3021 2013-09-13 Jakub Jelinek <jakub@redhat.com>
3023 PR tree-optimization/58392
3024 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
3025 to avoid shadowing of outer loop variable. If
3026 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
3027 replace_by_duplicate_decl simduid of loops that have it set and
3028 set dest_cfun->has_simduid_loops and/or
3029 dest_cfun->has_force_vect_loops.
3030 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
3031 instead of maybe_lookup_decl.
3032 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
3033 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
3034 call. Copy over force_vect and copy and remap simduid. Set
3035 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
3036 (copy_cfg_body): Remove blocks_to_copy argument. Use
3037 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
3038 caller. Don't set cfun->has_simduid_loops and/or
3039 cfun->has_force_vect_loops here.
3040 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
3042 (expand_call_inline, tree_function_versioning): Adjust copy_body
3045 2013-09-13 Martin Jambor <mjambor@suse.cz>
3048 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
3049 the assert if the edge was a speculative one.
3051 2013-09-13 Richard Biener <rguenther@suse.de>
3053 * tree-data-ref.h (known_dependences_p): Move here ...
3054 * tree-loop-distribution.c (known_dependences_p): ... from here.
3055 (dump_rdg_component, debug_rdg_component): Remove.
3057 (generate_loops_for_partition): Use gimple_uid instead of
3058 relying on matching stmt visit order.
3059 (rdg_build_partitions): Take starting stmt vector.
3060 (ldist_gen): Merge into ...
3061 (distribute_loop): ... this function. Do not compute starting
3063 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
3065 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3067 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
3068 Set type attribute correctly. Set predicable_short_it attribute.
3069 (cmpsi_shiftsi): Remove %? from output template.
3071 2013-09-13 Richard Biener <rguenther@suse.de>
3073 * tree-loop-distribution.c (struct rdg_component,
3074 rdg_defs_used_in_other_loops_p, free_rdg_components,
3075 rdg_build_components): Remove.
3076 (stmts_from_loop): Do not record virtual PHIs.
3077 (generate_loops_for_partition): Skip virtual PHIs.
3078 (build_rdg_partition_for_component): Rename to ...
3079 (build_rdg_partition_for_vertex): ... this and adjust.
3080 (rdg_build_partitions): Take a vector of starting vertices
3081 instead of components. Remove unnecessary leftover handling.
3082 (ldist_gen): Do not build components or record other stores.
3083 (distribute_loop): Do not distribute loops containing stmts
3086 2013-09-13 Christian Bruel <christian.bruel@st.com>
3089 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
3091 2013-09-13 Kai Tietz <ktietz@redhat.com>
3093 * config.gcc: Separate cases for mingw and cygwin targets,
3094 and add 64-bit cygwin target case.
3096 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
3097 dll-export inline-functions.
3098 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
3101 2013-09-13 Jeff Law <law@redhat.com>
3105 2013-09-06 Jeff Law <law@redhat.com>
3107 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
3108 edge implied equivalences into successor phis.
3110 2013-09-12 DJ Delorie <dj@redhat.com>
3112 * config/rl78/rl78-virt.md: Change from | to \; for asm line
3115 2013-09-12 Brooks Moses <bmoses@google.com>
3118 * Makefile.in: Do not install driver binaries in $(target)/bin.
3120 2013-09-12 DJ Delorie <dj@redhat.com>
3122 * config/rl78/rl78.opt (mrelax): New.
3123 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
3124 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
3126 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
3128 (rl78_expand_epilogue): Likewise.
3130 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
3133 * lra-eliminations.c (remove_reg_equal_offset_note): New.
3134 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
3135 pointer elimination with using remove_reg_equal_offset_note.
3137 2013-09-12 DJ Delorie <dj@redhat.com>
3139 * config/msp430/: New port.
3140 * config.gcc (msp430): Added.
3141 * doc/invoke.texi: Document MSP430 options.
3142 * doc/install.texi: Document msp430-elf
3143 * doc/md.texi: Document msp430-elf
3144 * doc/contrib.texi: Document msp430-elf
3146 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
3149 2013-09-12 Martin Jambor <mjambor@suse.cz>
3152 * ipa-prop.c (remove_described_reference): Give up if the edge in the
3153 reference descriptor is NULL.
3154 (ipa_edge_removal_hook): If owning a reference descriptor, set its
3157 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
3159 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
3160 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
3161 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
3162 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
3163 make_temp_ssa_name): move to tree-ssanames.h
3164 * tree-ssa-alias.h: Move prototype.
3165 * tree-ssa.h: Include tree-ssanames.h.
3166 * tree-ssanames.c (FREE_SSANAMES): Move to here.
3167 * tree-ssanames.h: New. Move items from tree-flow*.h
3168 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
3170 2013-09-12 Richard Biener <rguenther@suse.de>
3172 PR tree-optimization/58404
3173 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
3174 propagate non-invariant addresses into dereferences wrapped
3175 in component references.
3177 2013-09-12 Richard Biener <rguenther@suse.de>
3179 PR tree-optimization/58402
3180 * passes.def: Move pass_late_warn_uninitialized later.
3182 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
3184 * tree-ssa.h: New. Move content from tree-flow.h and
3186 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
3187 Move prototypes belonging to tree-ssa.c.
3188 * tree-flow-inline.h (redirect_edge_var_map_def,
3189 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
3191 * gimple.h: Adjust prototypes.
3192 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
3194 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
3195 * tree.h: Move prototype to tree-ssa.h.
3196 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
3197 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
3198 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
3199 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
3200 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
3201 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
3202 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
3203 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
3204 graphite-blocking.c, graphite-clast-to-gimple.c,
3205 graphite-dependences.c, graphite-interchange.c,
3206 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
3207 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
3208 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
3209 ipa-reference.c, ipa-split.c, ipa-utils.c,
3210 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
3211 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
3212 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
3213 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
3214 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
3215 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
3216 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
3217 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
3218 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
3219 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
3220 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
3221 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
3222 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
3223 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
3224 value-prof.c, var-tracking.c,
3225 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
3227 2013-09-12 Richard Biener <rguenther@suse.de>
3229 PR tree-optimization/58396
3230 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
3231 (build_rdg): Take a loop-nest parameter, fix memleaks.
3232 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
3234 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
3236 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
3239 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
3241 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
3242 The FMA instruction names should have a 'v' prefix.
3244 2013-09-12 Richard Biener <rguenther@suse.de>
3246 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
3247 (dot_rdg): Use popen instead of system in optional code.
3248 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
3249 (already_processed_vertex_p): Adjust.
3250 (has_anti_or_output_dependence, predecessor_has_mem_write,
3251 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
3252 rdg_flag_uses): Remove.
3253 (rdg_flag_vertex): Simplify.
3254 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
3256 (build_rdg_partition_for_component): Process the first vertex
3257 of a component only.
3258 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
3260 2013-09-12 Alan Modra <amodra@gmail.com>
3262 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
3264 2013-09-11 DJ Delorie <dj@redhat.com>
3265 Nick Clifton <nickc@redhat.com>
3267 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
3268 (rl78_stack_based_mem): New.
3269 * config/rl78/constraints.md (Iv08): New.
3278 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
3280 (movsi): Change from expand to insn-and-split.
3281 (ashrsi3): Clobber AX.
3285 * config/rl78/rl78.md (CC_REG): Fix.
3286 (addsi3): Allow memory and immediate operands.
3287 (addsi3_internal): Split into...
3288 (addsi3_internal_virt): ...new, and ...
3289 (addsi3_internal_real): ...new.
3291 (subsi3_internal_virt): New.
3292 (subsi3_internal_real): New.
3293 (mulsi3): Add memory operand.
3294 (mulsi3_rl78): Likewise.
3295 (mulsi3_g13): Likewise.
3296 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
3297 (cbranchqi4_real): Add more constraint options.
3298 (cbranchhi4_real): Expand pattern.
3299 (cbranchhi4_real_signed): New.
3300 (cbranchhi4_real_inverted): New.
3301 (cbranchsi4_real_lt): New.
3302 (cbranchsi4_real_ge): New.
3303 (cbranchsi4_real_signed): New.
3304 (cbranchsi4_real): New.
3306 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
3308 (lshrsi3_virt): Likewise.
3309 (ashlsi3_virt): Likewise.
3310 (cbranchqi4_virt_signed): New.
3311 (cbranchhi4_virt_signed): New.
3312 (cbranchsi4_virt): New.
3313 * config/rl78/rl78.c: Whitespace fixes throughout.
3314 (move_elim_pass): New.
3315 (pass_data_rl78_move_elim): New.
3316 (pass_rl78_move_elim): New.
3317 (make_pass_rl78_move_elim): New.
3318 (rl78_devirt_info): Run devirt earlier.
3319 (rl78_move_elim_info): New.
3320 (rl78_asm_file_start): Register it.
3321 (rl78_split_movsi): New.
3322 (rl78_as_legitimate_address): Allow virtual base registers when
3324 (rl78_addr_space_convert): Remove spurious debug stuff.
3325 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
3326 (rl78_print_operand): More cases for not printing '#'.
3327 (rl78_expand_compare): Remove most of the logic.
3328 (content_memory): New.
3329 (clear_content_memory): New.
3330 (get_content_index): New.
3331 (get_content_name): New.
3332 (display_content_memory): New.
3333 (update_content): New.
3334 (record_content): New.
3335 (already_contains): New.
3336 (insn_ok_now): Re-recog insns with virtual registers.
3337 (add_postponed_content_update): New.
3338 (process_postponed_content_update): New.
3339 (gen_and_emit_move): New.
3340 (transcode_memory_rtx): Record new location content.
3341 Use gen_and_emit_move.
3342 (force_into_acc): New.
3343 (move_to_acc): Use gen_and_emit_move.
3344 (move_from_acc): Likewise.
3345 (move_acc_to_reg): Likewise.
3346 (move_to_x): Likewise.
3347 (move_to_hl): Likewise.
3348 (move_to_de): Likewise.
3349 (rl78_alloc_physical_registers_op1): Record location content.
3350 (has_constraint): New.
3351 (rl78_alloc_physical_registers_op2): Record location content.
3353 (rl78_alloc_physical_registers_ro1): Likewise.
3354 (rl78_alloc_physical_registers_cmp): Likewise.
3355 (rl78_alloc_physical_registers_umul): Likewise.
3356 (rl78_alloc_address_registers_macax): New.
3357 (rl78_alloc_physical_registers): Initialize and set location
3358 content memory as needed.
3359 (rl78_reorg): Make sure split2 is called.
3360 (rl78_rtx_costs): New.
3362 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
3364 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
3365 for (not (neg ...)) and (neg (not ...)) cases.
3367 2013-09-11 Richard Biener <rguenther@suse.de>
3370 * passes.def: Split critical edges before late uninit warning passes.
3371 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
3373 2013-09-11 Jakub Jelinek <jakub@redhat.com>
3375 PR tree-optimization/58385
3376 * fold-const.c (build_range_check): If both low and high are NULL,
3377 use omit_one_operand_loc to preserve exp side-effects.
3379 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3381 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
3383 2013-09-11 Richard Biener <rguenther@suse.de>
3385 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
3386 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
3387 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
3388 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
3389 stmts_from_loop, known_dependences_p, build_empty_rdg,
3390 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
3391 * tree-loop-distribution.c: ... here.
3392 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
3393 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
3394 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
3395 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
3396 * tree-loop-distribution.c: ... here.
3397 * tree-loop-distribution.c: Include gimple-pretty-print.h.
3398 (struct partition_s): Add loops member.
3399 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
3400 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
3401 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
3403 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
3404 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3405 Sergey Lega <sergey.s.lega@intel.com>
3406 Anna Tikhonova <anna.tikhonova@intel.com>
3407 Ilya Tocar <ilya.tocar@intel.com>
3408 Andrey Turetskiy <andrey.turetskiy@intel.com>
3409 Ilya Verbin <ilya.verbin@intel.com>
3410 Kirill Yukhin <kirill.yukhin@intel.com>
3411 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3413 * config/i386/constraints.md (k): New.
3415 * config/i386/i386.c (const regclass_map): Add new mask registers.
3416 (dbx_register_map): Ditto.
3417 (dbx64_register_map): Ditto.
3418 (svr4_dbx_register_map): Ditto.
3419 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
3421 (ix86_preferred_reload_class): Disable constants for mask registers.
3422 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
3423 (ix86_hard_regno_mode_ok): Support new mask registers.
3424 (x86_order_regs_for_local_alloc): Ditto.
3425 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
3426 (FIXED_REGISTERS): Add new mask registers.
3427 (CALL_USED_REGISTERS): Ditto.
3428 (REG_ALLOC_ORDER): Ditto.
3429 (VALID_MASK_REG_MODE): New.
3430 (FIRST_MASK_REG): Ditto.
3431 (LAST_MASK_REG): Ditto.
3432 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
3433 (MAYBE_MASK_CLASS_P): New.
3434 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
3435 (REG_CLASS_CONTENTS): Ditto.
3436 (MASK_REGNO_P): New.
3437 (ANY_MASK_REG_P): Ditto.
3438 (HI_REGISTER_NAMES): Add new mask registers.
3439 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
3440 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
3442 (attribute "type"): Add mskmov, msklog.
3443 (attribute "length_immediate"): Support them.
3444 (attribute "memory"): Ditto.
3445 (attribute "prefix_0f"): Ditto.
3446 (*movhi_internal): Support new mask registers.
3447 (*movqi_internal): Ditto.
3448 (define_split): Split out clobber pattern is a logic
3449 insn on mask registers.
3450 (*k<logic><mode>): New.
3451 (*andhi_1): Extend to support mask regs.
3452 (*andqi_1): Extend to support mask regs.
3454 (define_split): Split and-not to and and not if operands
3456 (*<code><mode>_1): Separate HI mode to new pattern...
3457 (*<code>hi_1): This.
3458 (*<code>qi_1): Extend to support mask regs.
3460 (kortestzhi): Ditto.
3461 (kortestchi): Ditto.
3463 (*one_cmpl<mode>2_1): Remove HImode and handle it...
3464 (*one_cmplhi2_1): ...Here, now with mask registers support.
3465 (*one_cmplqi2_1): Support new mask registers.
3466 (HI/QImode arithmetics splitter): Don't split if mask registers
3468 (HI/QImode not splitter): Ditto.
3469 * config/i386/predicated.md (mask_reg_operand): New.
3470 (general_reg_operand): Ditto.
3472 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
3474 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
3475 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
3477 2013-09-10 Jeff Law <law@redhat.com>
3479 PR tree-optimization/58380
3480 * tree-ssa-threadupdate.c (thread_block): Recognize another case
3481 of threading through a buried loop header.
3483 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
3484 return value for single successor case.
3486 2013-09-10 Jan Hubicka <jh@suse.cz>
3488 * ipa-devirt.c (ipa_devirt): Enable with LTO.
3490 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
3493 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
3494 support conditional execution.
3495 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
3497 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
3499 * lra.c (lra): Clear lra_optional_reload_pseudos before every
3501 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
3502 Check destination too to check move insn.
3503 (undo_optional_reloads): Add check that the original peudo did not
3504 changed its allocation and the optional reload was inherited on last
3505 inheritance pass. Break loop after deciding to keep optional reload.
3506 (lra_undo_inheritance): Add check that inherited pseudo still in
3509 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
3511 * config/aarch64/aarch64.md (generic_sched): New.
3512 * config/aarch64/aarch64-generic.md (load): Make conditional
3513 on generic_sched attribute.
3514 (nonload): Likewise.
3516 2013-09-10 Jan Hubicka <jh@suse.cz>
3518 * lto-cgraph.c: Include ipa-utils.h.
3519 (compute_ltrans_boundary): Also add possible targets into the boundary.
3521 2013-09-10 Jan Hubicka <jh@suse.cz>
3523 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
3524 VAR_DECL of vtable rather than full expression.
3526 2013-09-10 Jan Hubicka <jh@suse.cz>
3527 Paolo Carlini <paolo.carlini@oracle.com>
3529 * cgraphunit.c (analyze_functions): Save input_location, set it
3530 to UNKNOWN_LOCATION and restore it at the end.
3532 2013-09-10 Martin Jambor <mjambor@suse.cz>
3534 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
3535 represented by a thunk.
3537 2013-09-10 Jeff Law <law@redhat.com>
3539 PR tree-optimization/58343
3540 * tree-ssa-threadupdate.c (thread_block): Identify and disable
3541 jump threading requests through loop headers buried in the middle
3542 of a jump threading path.
3544 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
3545 in return value/type.
3547 2013-09-10 Jakub Jelinek <jakub@redhat.com>
3549 PR rtl-optimization/58365
3550 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
3551 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
3554 2013-09-10 Richard Biener <rguenther@suse.de>
3556 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
3557 * tree-data-ref.c (create_rdg_vertices): Collect all data
3558 references, signal failure to the caller, use data-ref API.
3559 (build_rdg): Compute data references only once. Maintain lifetime
3560 of data references and data dependences from within RDG.
3561 (free_rdg): Free dependence relations.
3562 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
3563 inventing extra dependences.
3564 (distribute_loop): Update for RDG API changes.
3566 2013-09-10 Kai Tietz <ktietz@redhat.com>
3568 * doc/invoke.texi (fms-extensions): Document changed
3569 behavior for ms-abi targets.
3570 * config/i386/i386.c (ix86_option_override_internal):
3571 Set default value of option -fms-extension for ms-abi targets.
3573 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
3575 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
3577 2013-09-10 Alan Modra <amodra@gmail.com>
3580 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
3582 2013-09-10 Alan Modra <amodra@gmail.com>
3584 * config/rs6000/predicates.md (add_cint_operand): New.
3585 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
3586 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
3587 using add_cint_operand.
3588 (largetoc_high_plus_aix): Likewise.
3590 2013-09-09 Jakub Jelinek <jakub@redhat.com>
3592 PR tree-optimization/58364
3593 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
3594 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
3595 the current range can't be an unconditional true or false.
3597 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
3599 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
3601 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
3603 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
3605 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
3608 * doc/invoke.texi (-Wdelete-incomplete): Document it.
3610 2013-09-09 Ian Bolton <ian.bolton@arm.com>
3612 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
3613 NO_REGS for immediate that can't be moved directly into FP_REGS.
3615 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3617 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
3618 comparison with negated operand.
3619 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
3622 2013-09-09 Richard Biener <rguenther@suse.de>
3625 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
3626 of a subloop record all its block as affecting loop-closed SSA form.
3628 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
3630 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
3631 of an rtx/bitpos pair.
3632 (store_fixed_bit_field): Update accordingly.
3634 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
3636 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
3638 * builtins.c (expand_errno_check): Likewise.
3639 * dwarf2cfi.c (init_return_column_size): Likewise.
3640 * except.c (sjlj_mark_call_sites): Likewise.
3641 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
3642 * lra-constraints.c (emit_inc): Likewise.
3643 * ree.c (combine_set_extension): Likewise.
3644 * regmove.c (fixup_match_2): Likewise.
3645 * reload1.c (inc_for_reload): Likewise.
3647 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
3649 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
3650 (force_to_mode, simplify_shift_const_1, simplify_comparison):
3651 Use gen_int_mode with the mode of the associated simplify_* call.
3652 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
3653 * expmed.c (expand_shift_1): Likewise.
3654 * function.c (instantiate_virtual_regs_in_insn): Likewise.
3655 * loop-iv.c (iv_number_of_iterations): Likewise.
3656 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
3657 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
3659 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
3661 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
3662 of the associated expand_* call.
3663 (asan_emit_stack_protection): Likewise.
3664 * builtins.c (round_trampoline_addr): Likewise.
3665 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
3666 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
3667 (emit_store_flag): Likewise.
3668 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
3669 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
3671 * function.c (instantiate_virtual_regs_in_insn): Likewise.
3672 * ifcvt.c (noce_try_store_flag_constants): Likewise.
3673 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
3674 * modulo-sched.c (generate_prolog_epilog): Likewise.
3675 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
3676 (expand_ctz, expand_ffs, expand_unop): Likewise.
3678 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
3680 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
3681 of the associated gen_rtx_* call.
3682 * caller-save.c (init_caller_save): Likewise.
3683 * combine.c (find_split_point, make_extraction): Likewise.
3684 (make_compound_operation): Likewise.
3685 * dwarf2out.c (mem_loc_descriptor): Likewise.
3686 * explow.c (plus_constant, probe_stack_range): Likewise.
3687 * expmed.c (expand_mult_const): Likewise.
3688 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
3689 * reload1.c (init_reload): Likewise.
3690 * valtrack.c (cleanup_auto_inc_dec): Likewise.
3691 * var-tracking.c (adjust_mems): Likewise.
3692 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
3693 rather than gen_rtx_fmt_ee.
3695 2013-09-09 Jan Hubicka <jh@suse.cz>
3698 * value-prof.c (gimple_ic): Copy also abnormal edges.
3700 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
3702 * asan.c (asan_shadow_cst): Use gen_int_mode.
3704 2013-09-08 Jan Hubicka <jh@suse.cz>
3706 * ipa-profile.c: Add toplevel comment.
3707 (ipa_propagate_frequency_1): Be more conservative when profile is read.
3708 (contains_hot_call_p): New function.
3709 (ipa_propagate_frequency): Set frequencies based on counts when
3711 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
3712 profile; do not tamper with profile after inlining if it is read.
3714 2013-09-08 Jan Hubicka <jh@suse.cz>
3716 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
3719 2013-09-08 Jan Hubicka <jh@suse.cz>
3721 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
3724 2013-09-08 Jeff Law <law@redhat.com>
3727 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
3730 2013-09-08 Andi Kleen <ak@linux.intel.com>
3732 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
3734 2013-09-08 Jan Hubicka <jh@suse.cz>
3736 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
3738 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
3740 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
3741 for non-debug insns.
3742 * lra.c (new_insn_reg): Take the containing insn as a parameter.
3743 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
3744 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
3747 2013-09-08 Jan Hubicka <jh@suse.cz>
3749 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
3750 targets and devirtualize to BUILT_IN_UNREACHABLE.
3751 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
3752 * ipa.c (walk_polymorphic_call_targets): New function.
3753 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
3754 functions; use the new timevar.
3755 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
3756 was removed from the program.
3757 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
3758 not consider it in the walk when its vtable is dead.
3759 (possible_polymorphic_call_targets_1): Pass anonymous flag to
3761 (devirt_variable_node_removal_hook): New function.
3762 (possible_polymorphic_call_targets): Also register
3763 devirt_variable_node_removal_hook.
3764 (ipa_devirt): Do not do non-speculative devirtualization.
3765 (gate_ipa_devirt): One execute if devirtualizing speculatively.
3767 2013-09-08 Jan Hubicka <jh@suse.cz>
3769 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
3770 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
3771 varpool_remove_variable_insertion_hook): Declare.
3772 * varpool.c (varpool_node_hook_list): New structure.
3773 (first_varpool_node_removal_hook,
3774 first_varpool_variable_insertion_hook): New variables.
3775 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
3776 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
3777 varpool_remove_variable_insertion_hook,
3778 varpool_call_variable_insertion_hooks): New functions.
3779 (varpool_remove_node): Use it.
3781 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
3784 * diagnostic.c (diagnostic_build_prefix): When s.file is
3785 "<built-in>" don't output line and column numbers.
3787 2013-09-06 Jan Hubicka <jh@suse.cz>
3789 * cgraphunit.c (expand_thunk): Get body before touching arguments.
3790 * lto-streamer-out.c: Stream thunks, too.
3791 * lto-streamer-in.c (input_function): Pop cfun here
3792 (lto_read_body): Instead of here.
3794 2013-09-06 Caroline Tice <cmtice@google.com>
3796 * doc/install.texi: Add documentation for the --enable-vtable-verify
3797 and the --disable-libvtv configure options.
3799 2013-09-06 Jeff Law <law@redhat.com>
3801 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
3802 edge implied equivalences into successor phis.
3804 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
3806 * resource.c (mark_referenced_resources): Handle COND_EXEC.
3808 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
3810 * resource.c (mark_target_live_regs): Compute resources taking
3811 into account if a call is predicated or not.
3813 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
3815 * toplev.c (output_stack_usage): Be prepared for suffixes created by
3816 the compiler in the function names.
3818 2013-09-06 Jan Hubicka <jh@suse.cz>
3821 * ipa-inline.c (has_caller_p): New function.
3822 (want_inline_function_to_all_callers_p): Use it.
3823 (sum_callers, inline_to_all_callers): Break out from ...
3824 (ipa_inline): ... here.
3826 2013-09-06 Jan Hubicka <jh@suse.cz>
3828 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
3829 only when AVX is enabled.
3831 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3833 * config/aarch64/aarch64.md
3834 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
3835 is fpsimd_<load/store>2.
3836 (load_pair<mode>): Likewise.
3837 (store_pair<mode>): Likewise.
3839 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3841 * config/arm/types.md (type): Add "mrs" type.
3842 * config/aarch64/aarch64.md
3843 (aarch64_load_tp_hard): Make type "mrs".
3845 (load_tp_hard): Make type "mrs".
3846 * config/arm/cortex-a15.md: Update with new attributes.
3847 * config/arm/cortex-a5.md: Update with new attributes.
3848 * config/arm/cortex-a53.md: Update with new attributes.
3849 * config/arm/cortex-a7.md: Update with new attributes.
3850 * config/arm/cortex-a8.md: Update with new attributes.
3851 * config/arm/cortex-a9.md: Update with new attributes.
3852 * config/arm/cortex-m4.md: Update with new attributes.
3853 * config/arm/cortex-r4.md: Update with new attributes.
3854 * config/arm/fa526.md: Update with new attributes.
3855 * config/arm/fa606te.md: Update with new attributes.
3856 * config/arm/fa626te.md: Update with new attributes.
3857 * config/arm/fa726te.md: Update with new attributes.
3859 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3861 * config/aarch64/aarch64.md
3862 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
3863 (*movtf_aarch64): Likewise.
3865 (thumb1_movdi_insn): Use "multiple" for type where more than one
3866 instruction is used for a move.
3867 (*arm32_movhf): Likewise.
3868 (*thumb_movdf_insn): Likewise.
3870 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3872 * config/arm/types.md (type): Rename fcpys to fmov.
3874 (*arm_movsi_vfp): Rename type fcpys as fmov.
3875 (*thumb2_movsi_vfp): Likewise
3876 (*movhf_vfp_neon): Likewise
3877 (*movhf_vfp): Likewise
3878 (*movsf_vfp): Likewise
3879 (*thumb2_movsf_vfp): Likewise
3880 (*movsfcc_vfp): Likewise
3881 (*thumb2_movsfcc_vfp): Likewise
3882 * config/aarch64/aarch64-simd.md
3883 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
3884 * config/aarch64/aarch64.md
3885 (*movsi_aarch64): Replace type mov_reg with fmovs.
3886 (*movdi_aarch64): Likewise
3887 (*movsf_aarch64): Likewise
3888 (*movdf_aarch64): Likewise
3890 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
3891 * config/arm/iwmmxt.md
3892 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
3893 * config/arm/arm1020e.md: Update with new attributes.
3894 * config/arm/cortex-a15-neon.md: Update with new attributes.
3895 * config/arm/cortex-a5.md: Update with new attributes.
3896 * config/arm/cortex-a53.md: Update with new attributes.
3897 * config/arm/cortex-a7.md: Update with new attributes.
3898 * config/arm/cortex-a8-neon.md: Update with new attributes.
3899 * config/arm/cortex-a9.md: Update with new attributes.
3900 * config/arm/cortex-m4-fpu.md: Update with new attributes.
3901 * config/arm/cortex-r4f.md: Update with new attributes.
3902 * config/arm/marvell-pj4.md: Update with new attributes.
3903 * config/arm/vfp11.md: Update with new attributes.
3905 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3907 * config/aarch64/aarch64.md
3908 (*madd<mode>): Fix type attribute.
3909 (*maddsi_uxtw): Likewise.
3910 (*msub<mode>): Likewise.
3911 (*msubsi_uxtw): Likewise.
3912 (<su_optab>maddsidi4): Likewise.
3913 (<su_optab>msubsidi4): Likewise.
3915 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3917 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
3918 * config/arm/arm.md (core_cycles): Remove fdiv.
3919 * config/arm/vfp.md:
3920 (*sqrtsf2_vfp): Update for attribute changes.
3921 (*sqrtdf2_vfp): Likewise.
3922 * config/aarch64/aarch64.md:
3923 (sqrt<mode>2): Update for attribute changes.
3924 * config/arm/arm1020e.md: Update with new attributes.
3925 * config/arm/cortex-a15-neon.md: Update with new attributes.
3926 * config/arm/cortex-a5.md: Update with new attributes.
3927 * config/arm/cortex-a53.md: Update with new attributes.
3928 * config/arm/cortex-a7.md: Update with new attributes.
3929 * config/arm/cortex-a8-neon.md: Update with new attributes.
3930 * config/arm/cortex-a9.md: Update with new attributes.
3931 * config/arm/cortex-m4-fpu.md: Update with new attributes.
3932 * config/arm/cortex-r4f.md: Update with new attributes.
3933 * config/arm/marvell-pj4.md: Update with new attributes.
3934 * config/arm/vfp11.md: Update with new attributes.
3936 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3938 * config/arm/types.md
3939 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
3940 * config/aarch64/aarch64.md
3941 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
3943 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
3944 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
3945 (float<GPI:mode><GPF:mode>2): Likewise.
3947 (*truncsisf2_vfp): Update with new attributes.
3948 (*truncsidf2_vfp): Likewise.
3949 (fixuns_truncsfsi2): Likewise.
3950 (fixuns_truncdfsi2): Likewise.
3951 (*floatsisf2_vfp): Likewise.
3952 (*floatsidf2_vfp): Likewise.
3953 (floatunssisf2): Likewise.
3954 (floatunssidf2): Likewise.
3955 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
3956 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
3957 * config/arm/arm1020e.md: Update with new attributes.
3958 * config/arm/cortex-a15-neon.md: Update with new attributes.
3959 * config/arm/cortex-a5.md: Update with new attributes.
3960 * config/arm/cortex-a53.md: Update with new attributes.
3961 * config/arm/cortex-a7.md: Update with new attributes.
3962 * config/arm/cortex-a8-neon.md: Update with new attributes.
3963 * config/arm/cortex-a9.md: Update with new attributes.
3964 * config/arm/cortex-m4-fpu.md: Update with new attributes.
3965 * config/arm/cortex-r4f.md: Update with new attributes.
3966 * config/arm/marvell-pj4.md: Update with new attributes.
3967 * config/arm/vfp11.md: Update with new attributes.
3969 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3971 * config/aarch64/arm_neon.h
3972 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
3973 (vqtbx<1,2,3,4><q>_s8): Likewise.
3975 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
3977 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
3978 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
3979 (add<mode>3): Add type attribute.
3980 (add<mode>3): Likewise.
3981 (usadd<mode>3): Likewise.
3982 (ssadd<mode>3): Likewise.
3983 (sub<mode>3): Likewise.
3984 (sub<mode>3): Likewise.
3985 (ussub<mode>3): Likewise.
3986 (sssub<mode>3): Likewise.
3987 (ssmulsa3): Likewise.
3988 (usmulusa3): Likewise.
3989 (arm_usatsihi): Likewise.
3991 (*movdi_vfp): Add types for all instructions.
3992 (*movdi_vfp_cortexa8): Likewise.
3993 (*movhf_vfp_neon): Likewise.
3994 (*movhf_vfp): Likewise.
3995 (*movdf_vfp): Likewise.
3996 (*thumb2_movdf_vfp): Likewise.
3997 (*thumb2_movdfcc_vfp): Likewise.
3998 * config/arm/arm.md: Add type attribute to all insn patterns.
3999 (*thumb1_adddi3): Add type attribute.
4000 (*arm_adddi3): Likewise.
4001 (*adddi_sesidi_di): Likewise.
4002 (*adddi_zesidi_di): Likewise.
4003 (*thumb1_addsi3): Likewise.
4004 (addsi3_compare0): Likewise.
4005 (*addsi3_compare0_scratch): Likewise.
4006 (*compare_negsi_si): Likewise.
4007 (cmpsi2_addneg): Likewise.
4008 (*addsi3_carryin_<optab>): Likewise.
4009 (*addsi3_carryin_alt2_<optab>): Likewise.
4010 (*addsi3_carryin_clobercc_<optab>): Likewise.
4011 (*subsi3_carryin): Likewise.
4012 (*subsi3_carryin_const): Likewise.
4013 (*subsi3_carryin_compare): Likewise.
4014 (*subsi3_carryin_compare_const): Likewise.
4015 (*arm_subdi3): Likewise.
4016 (*thumb_subdi3): Likewise.
4017 (*subdi_di_zesidi): Likewise.
4018 (*subdi_di_sesidi): Likewise.
4019 (*subdi_zesidi_di): Likewise.
4020 (*subdi_sesidi_di): Likewise.
4021 (*subdi_zesidi_ze): Likewise.
4022 (thumb1_subsi3_insn): Likewise.
4023 (*arm_subsi3_insn): Likewise.
4024 (*anddi3_insn): Likewise.
4025 (*anddi_zesidi_di): Likewise.
4026 (*anddi_sesdi_di): Likewise.
4027 (*ne_zeroextracts): Likewise.
4028 (*ne_zeroextracts): Likewise.
4029 (*ite_ne_zeroextr): Likewise.
4030 (*ite_ne_zeroextr): Likewise.
4031 (*anddi_notdi_di): Likewise.
4032 (*anddi_notzesidi): Likewise.
4033 (*anddi_notsesidi): Likewise.
4034 (andsi_notsi_si): Likewise.
4035 (thumb1_bicsi3): Likewise.
4036 (*iordi3_insn): Likewise.
4037 (*iordi_zesidi_di): Likewise.
4038 (*iordi_sesidi_di): Likewise.
4039 (*thumb1_iorsi3_insn): Likewise.
4040 (*xordi3_insn): Likewise.
4041 (*xordi_zesidi_di): Likewise.
4042 (*xordi_sesidi_di): Likewise.
4043 (*arm_xorsi3): Likewise.
4044 (*andsi_iorsi3_no): Likewise.
4045 (*smax_0): Likewise.
4046 (*smax_m1): Likewise.
4047 (*arm_smax_insn): Likewise.
4048 (*smin_0): Likewise.
4049 (*arm_smin_insn): Likewise.
4050 (*arm_umaxsi3): Likewise.
4051 (*arm_uminsi3): Likewise.
4052 (*minmax_arithsi): Likewise.
4053 (*minmax_arithsi_): Likewise.
4054 (*satsi_<SAT:code>): Likewise.
4055 (arm_ashldi3_1bit): Likewise.
4056 (arm_ashrdi3_1bit): Likewise.
4057 (arm_lshrdi3_1bit): Likewise.
4058 (*arm_negdi2): Likewise.
4059 (*thumb1_negdi2): Likewise.
4060 (*arm_negsi2): Likewise.
4061 (*thumb1_negsi2): Likewise.
4062 (*negdi_extendsid): Likewise.
4063 (*negdi_zero_extend): Likewise.
4064 (*arm_abssi2): Likewise.
4065 (*thumb1_abssi2): Likewise.
4066 (*arm_neg_abssi2): Likewise.
4067 (*thumb1_neg_abss): Likewise.
4068 (one_cmpldi2): Likewise.
4069 (extend<mode>di2): Likewise.
4070 (*compareqi_eq0): Likewise.
4071 (*arm_extendhisi2addsi): Likewise.
4072 (*arm_movdi): Likewise.
4073 (*thumb1_movdi_insn): Likewise.
4074 (*arm_movt): Likewise.
4075 (*thumb1_movsi_insn): Likewise.
4076 (pic_add_dot_plus_four): Likewise.
4077 (pic_add_dot_plus_eight): Likewise.
4078 (tls_load_dot_plus_eight): Likewise.
4079 (*thumb1_movhi_insn): Likewise.
4080 (*thumb1_movsf_insn): Likewise.
4081 (*movdf_soft_insn): Likewise.
4082 (*thumb_movdf_insn): Likewise.
4083 (cbranchsi4_insn): Likewise.
4084 (cbranchsi4_scratch): Likewise.
4085 (*negated_cbranchsi4): Likewise.
4086 (*tbit_cbranch): Likewise.
4087 (*tlobits_cbranch): Likewise.
4088 (*tstsi3_cbranch): Likewise.
4089 (*cbranchne_decr1): Likewise.
4090 (*addsi3_cbranch): Likewise.
4091 (*addsi3_cbranch_scratch): Likewise.
4092 (*arm_cmpdi_insn): Likewise.
4093 (*arm_cmpdi_unsig): Likewise.
4094 (*arm_cmpdi_zero): Likewise.
4095 (*thumb_cmpdi_zero): Likewise.
4096 (*deleted_compare): Likewise.
4097 (*mov_scc): Likewise.
4098 (*mov_negscc): Likewise.
4099 (*mov_notscc): Likewise.
4100 (*cstoresi_eq0_thumb1_insn): Likewise.
4101 (cstoresi_nltu_thumb1): Likewise.
4102 (cstoresi_ltu_thu): Likewise.
4103 (thumb1_addsi3_addgeu): Likewise.
4104 (*arm_jump): Likewise.
4105 (*thumb_jump): Likewise.
4106 (*check_arch2): Likewise.
4107 (arm_casesi_internal): Likewise.
4108 (thumb1_casesi_dispatch): Likewise.
4109 (*arm_indirect_jump): Likewise.
4110 (*thumb1_indirect_jump): Likewise.
4112 (*and_scc): Likewise.
4113 (*ior_scc): Likewise.
4114 (*compare_scc): Likewise.
4115 (*cond_move): Likewise.
4116 (*cond_arith): Likewise.
4117 (*cond_sub): Likewise.
4118 (*cmp_ite0): Likewise.
4119 (*cmp_ite1): Likewise.
4120 (*cmp_and): Likewise.
4121 (*cmp_ior): Likewise.
4122 (*ior_scc_scc): Likewise.
4123 (*ior_scc_scc_cmp): Likewise.
4124 (*and_scc_scc): Likewise.
4125 (*and_scc_scc_cmp): Likewise.
4126 (*and_scc_scc_nod): Likewise.
4127 (*negscc): Likewise.
4128 (movcond_addsi): Likewise.
4129 (movcond): Likewise.
4130 (*ifcompare_plus_move): Likewise.
4131 (*if_plus_move): Likewise.
4132 (*ifcompare_move_plus): Likewise.
4133 (*if_move_plus): Likewise.
4134 (*ifcompare_arith_arith): Likewise.
4135 (*if_arith_arith): Likewise.
4136 (*ifcompare_arith_move): Likewise.
4137 (*if_arith_move): Likewise.
4138 (*ifcompare_move_arith): Likewise.
4139 (*if_move_arith): Likewise.
4140 (*ifcompare_move_not): Likewise.
4141 (*if_move_not): Likewise.
4142 (*ifcompare_not_move): Likewise.
4143 (*if_not_move): Likewise.
4144 (*ifcompare_shift_move): Likewise.
4145 (*if_shift_move): Likewise.
4146 (*ifcompare_move_shift): Likewise.
4147 (*if_move_shift): Likewise.
4148 (*ifcompare_shift_shift): Likewise.
4149 (*ifcompare_not_arith): Likewise.
4150 (*ifcompare_arith_not): Likewise.
4151 (*if_arith_not): Likewise.
4152 (*ifcompare_neg_move): Likewise.
4153 (*if_neg_move): Likewise.
4154 (*ifcompare_move_neg): Likewise.
4155 (*if_move_neg): Likewise.
4156 (prologue_thumb1_interwork): Likewise.
4157 (*cond_move_not): Likewise.
4158 (*sign_extract_onebit): Likewise.
4159 (*not_signextract_onebit): Likewise.
4160 (stack_tie): Likewise.
4161 (align_4): Likewise.
4162 (align_8): Likewise.
4163 (consttable_end): Likewise.
4164 (consttable_1): Likewise.
4165 (consttable_2): Likewise.
4166 (consttable_4): Likewise.
4167 (consttable_8): Likewise.
4168 (consttable_16): Likewise.
4169 (*thumb1_tablejump): Likewise.
4170 (prefetch): Likewise.
4171 (force_register_use): Likewise.
4172 (thumb_eh_return): Likewise.
4173 (load_tp_hard): Likewise.
4174 (load_tp_soft): Likewise.
4175 (tlscall): Likewise.
4176 (*arm_movtas_ze): Likewise.
4177 (*arm_rev): Likewise.
4178 (*arm_revsh): Likewise.
4179 (*arm_rev16): Likewise.
4180 * config/arm/thumb2.md
4181 (*thumb2_smaxsi3): Likewise.
4182 (*thumb2_sminsi3): Likewise.
4183 (*thumb32_umaxsi3): Likewise.
4184 (*thumb2_uminsi3): Likewise.
4185 (*thumb2_negdi2): Likewise.
4186 (*thumb2_abssi2): Likewise.
4187 (*thumb2_neg_abss): Likewise.
4188 (*thumb2_movsi_insn): Likewise.
4189 (tls_load_dot_plus_four): Likewise.
4190 (*thumb2_movhi_insn): Likewise.
4191 (*thumb2_mov_scc): Likewise.
4192 (*thumb2_mov_negs): Likewise.
4193 (*thumb2_mov_negs): Likewise.
4194 (*thumb2_mov_nots): Likewise.
4195 (*thumb2_mov_nots): Likewise.
4196 (*thumb2_movsicc_): Likewise.
4197 (*thumb2_movsfcc_soft_insn): Likewise.
4198 (*thumb2_indirect_jump): Likewise.
4199 (*thumb2_and_scc): Likewise.
4200 (*thumb2_ior_scc): Likewise.
4201 (*thumb2_ior_scc_strict_it): Likewise.
4202 (*thumb2_cond_move): Likewise.
4203 (*thumb2_cond_arith): Likewise.
4204 (*thumb2_cond_ari): Likewise.
4205 (*thumb2_cond_sub): Likewise.
4206 (*thumb2_negscc): Likewise.
4207 (*thumb2_movcond): Likewise.
4208 (thumb2_casesi_internal): Likewise.
4209 (thumb2_casesi_internal_pic): Likewise.
4210 (*thumb2_alusi3_short): Likewise.
4211 (*thumb2_mov<mode>_shortim): Likewise.
4212 (*thumb2_addsi_short): Likewise.
4213 (*thumb2_subsi_short): Likewise.
4214 (thumb2_addsi3_compare0): Likewise.
4215 (*thumb2_cbz): Likewise.
4216 (*thumb2_cbnz): Likewise.
4217 (*thumb2_one_cmplsi2_short): Likewise.
4218 (*thumb2_negsi2_short): Likewise.
4219 (*orsi_notsi_si): Likewise.
4220 * config/arm/arm1020e.md: Update with new attributes.
4221 * config/arm/arm1026ejs.md: Update with new attributes.
4222 * config/arm/arm1136jfs.md: Update with new attributes.
4223 * config/arm/arm926ejs.md: Update with new attributes.
4224 * config/arm/cortex-a15.md: Update with new attributes.
4225 * config/arm/cortex-a5.md: Update with new attributes.
4226 * config/arm/cortex-a53.md: Update with new attributes.
4227 * config/arm/cortex-a7.md: Update with new attributes.
4228 * config/arm/cortex-a8.md: Update with new attributes.
4229 * config/arm/cortex-a9.md: Update with new attributes.
4230 * config/arm/cortex-m4.md: Update with new attributes.
4231 * config/arm/cortex-r4.md: Update with new attributes.
4232 * config/arm/fa526.md: Update with new attributes.
4233 * config/arm/fa606te.md: Update with new attributes.
4234 * config/arm/fa626te.md: Update with new attributes.
4235 * config/arm/fa726te.md: Update with new attributes.
4237 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
4239 * config/aarch64/aarch64-simd.md
4240 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
4241 <vwx> iterator to ensure correct register choice.
4242 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
4243 (aarch64_sqdmull_n<mode>): Likewise.
4244 (aarch64_sqdmull2_n<mode>_internal): Likewise.
4245 * config/aarch64/arm_neon.h
4246 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
4247 (vml<as><q>_n_<su>16): Likewise.
4248 (vml<as>l_high_lane<q>_<su>16): Likewise.
4249 (vml<as>l_high_n_<su>16): Likewise.
4250 (vml<as>l_lane<q>_<su>16): Likewise.
4251 (vml<as>l_n_<su>16): Likewise.
4252 (vmul<q>_lane<q>_<su>16): Likewise.
4253 (vmul<q>_n_<su>16): Likewise.
4254 (vmull_lane<q>_<su>16): Likewise.
4255 (vmull_n_<su>16): Likewise.
4256 (vmull_high_lane<q>_<su>16): Likewise.
4257 (vmull_high_n_<su>16): Likewise.
4258 (vqrdmulh<q>_n_s16): Likewise.
4260 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
4262 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
4263 have the correct lane parameter.
4265 2013-09-06 Richard Biener <rguenther@suse.de>
4267 * cfganal.c (control_dependences::~control_dependences):
4268 Properly free all of the vector.
4270 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
4273 * config/i386/i386.c (ix86_conditional_register_usage):
4274 Proper initialize extended SSE registers.
4276 2013-09-06 Jan Hubicka <jh@suse.cz>
4278 PR tree-optimization/58311
4279 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
4281 2013-09-06 Jan Hubicka <jh@suse.cz>
4283 * Makefile.in (tree-sra.o): Update dependencies.
4284 * tree-sra.c: Include ipa-utils.h
4285 (scan_function): Use recursive_call_p.
4286 (has_caller_p): New function.
4287 (cgraph_for_node_and_aliases): Count also callers of aliases.
4289 2013-09-06 Jan Hubicka <jh@suse.cz>
4292 * cgraph.h (symtab_semantically_equivalent_p): Declare.
4293 * tree-tailcall.c: Include ipa-utils.h.
4294 (find_tail_calls): Use it.
4295 * ipa-pure-const.c (check_call): Likewise.
4296 * ipa-utils.c (recursive_call_p): New function.
4297 * ipa-utils.h (recursive_call_p): Dclare.
4298 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
4299 (symtab_semantically_equivalent_p): New function.
4300 * Makefile.in (tree-tailcall.o): Update dependencies.
4302 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
4304 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
4307 2013-09-06 Richard Biener <rguenther@suse.de>
4309 * lto-streamer.h (lto_global_var_decls): Remove.
4310 * Makefile.in (OBJS): Remove lto-symtab.o.
4311 (lto-symtab.o): Remove.
4312 (GTFILES): Remove lto-symtab.c
4313 * lto-symtab.c: Move to lto/
4315 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4317 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
4318 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
4319 (UNSPEC_FPINT_RINT): New constant definitions.
4320 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
4321 definition with 2 attributes.
4322 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
4323 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
4326 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4328 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
4329 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
4330 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
4331 attribute to "z196_alone".
4332 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
4335 2013-09-06 Richard Biener <rguenther@suse.de>
4337 * basic-block.h (class control_dependences): New.
4338 * tree-ssa-dce.c (control_dependence_map): Remove.
4340 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
4341 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
4342 find_pdom, find_control_dependence, find_all_control_dependences):
4344 (mark_control_dependent_edges_necessary,
4345 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
4346 tree_dce_done, perform_tree_ssa_dce): Adjust.
4347 * cfganal.c (set_control_dependence_map_bit,
4348 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
4349 find_all_control_dependences): Move from tree-ssa-dce.c and
4350 implement as methods of control_dependences class.
4351 (control_dependences::control_dependences): New.
4352 (control_dependences::~control_dependences): Likewise.
4353 (control_dependences::get_edges_dependent_on): Likewise.
4354 (control_dependences::get_edge): Likewise.
4356 2013-09-04 Jan Hubicka <jh@suse.cz>
4358 * tree.c (types_same_for_odr): Drop overactive check.
4359 * ipa-devirt.c (hash_type_name): Likewise.
4361 2013-09-04 Jan Hubicka <jh@suse.cz>
4363 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
4364 (analyze_functions): ... here.
4366 2013-09-04 Jan Hubicka <jh@suse.cz>
4369 * cgraphunit.c (analyze_functions): Clear AUX fields
4370 after processing; initialize assembler name has.
4372 2013-09-05 Jeff Law <law@redhat.com>
4374 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
4375 from thread_around_empty_block. Record threading path into PATH.
4376 Recurse if threading through the initial block is successful.
4377 (thread_across_edge): Corresponding changes to slightly simplify.
4379 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
4381 * config/aarch64/aarch64.md
4382 (type): Remove frecpe, frecps, frecpx.
4383 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
4384 fix to be a TARGET_SIMD instruction.
4385 (aarch64_frecps): Remove.
4386 * config/aarch64/aarch64-simd.md
4387 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
4388 (aarch64_frecps<mode>): Handle all float/vector of float modes.
4390 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
4391 Sofiane Naci <sofiane.naci@arm.com>
4393 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
4394 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
4395 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
4396 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
4397 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
4398 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
4399 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
4400 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
4401 into "clz, "rbit". Rename "shift" to "shift_imm".
4402 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
4403 changes. Update for attribute changes all occurrences of arlo_* and
4405 * config/arm/arm-fixed.md: Update for attribute changes
4406 all occurrences of arlo_* types.
4407 * config/arm/thumb2.md: Update for attribute changes all occurrences
4409 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
4410 (cortexa7_older_only): Likewise.
4411 (cortexa7_younger): Likewise.
4412 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
4413 (1020alu_shift_op): Likewise.
4414 (1020alu_shift_reg_op): Likewise.
4415 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
4416 (alu_shift_op): Likewise.
4417 (alu_shift_reg_op): Likewise.
4418 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
4419 (11_alu_shift_op): Likewise.
4420 (11_alu_shift_reg_op): Likewise.
4421 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
4422 (9_alu_shift_reg_op): Likewise.
4423 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
4425 (cortex_a15_alu_shift): Likewise.
4426 (cortex_a15_alu_shift_reg): Likewise.
4427 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
4429 (cortex_a5_alu_shift): Likewise.
4430 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
4432 (cortex_a53_alu_shift): Likewise.
4433 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
4435 (cortex_a7_alu_reg): Likewise.
4436 (cortex_a7_alu_shift): Likewise.
4437 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
4439 (cortex_a8_alu_shift): Likewise.
4440 (cortex_a8_alu_shift_reg): Likewise.
4441 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
4442 (cortex_a9_dp_shift): Likewise.
4443 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
4445 * config/arm/cortex-r4.md
4446 (cortex_r4_alu): Update for attribute changes.
4447 (cortex_r4_mov): Likewise.
4448 (cortex_r4_alu_shift_reg): Likewise.
4449 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
4450 (526_alu_shift_op): Likewise.
4451 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
4452 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
4453 (626te_alu_shift_op): Likewise.
4454 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
4455 (726te_alu_shift_op): Likewise.
4456 (726te_alu_shift_reg_op): Likewise.
4457 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
4458 (mp626_alu_shift_op): Likewise.
4459 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
4460 (pj4_alu_conds): Likewise.
4461 (pj4_shift): Likewise.
4462 (pj4_shift_conds): Likewise.
4463 (pj4_alu_shift): Likewise.
4464 (pj4_alu_shift_conds): Likewise.
4465 * config/aarch64/aarch64.md: Update for attribute change
4466 all occurrences of arlo_* and shift* types.
4468 2013-09-05 Mike Stump <mikestump@comcast.net>
4470 * tree.h: Move documentation for tree_function_decl to tree-core.h
4471 with the declaration.
4473 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
4476 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
4477 looking for widest mode.
4479 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
4481 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
4483 2013-09-05 Richard Biener <rguenther@suse.de>
4485 PR tree-optimization/58137
4486 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
4487 Do not create vectors of pointers.
4488 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
4489 types for the components of the vector initializer.
4490 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
4491 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
4493 2013-09-05 Martin Jambor <mjambor@suse.cz>
4495 * ipa-prop.c (remove_described_reference): Accept missing references,
4496 return false if that hppens, otherwise return true.
4497 (cgraph_node_for_jfunc): New function.
4498 (try_decrement_rdesc_refcount): Likewise.
4499 (try_make_edge_direct_simple_call): Use them.
4500 (ipa_edge_removal_hook): Remove references from rdescs.
4501 (ipa_edge_duplication_hook): Clone rdescs and their references
4502 when the new edge has the same caller as the old one.
4503 * cgraph.c (cgraph_resolve_speculation): Remove speculative
4504 reference before removing any edges.
4506 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
4508 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
4510 * thumb2.md (thumb2_storewb_parisi): New pattern.
4512 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
4514 * config/aarch64/aarch64-option-extensions.def: Add
4515 AARCH64_OPT_EXTENSION of 'crc'.
4516 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
4517 (AARCH64_ISA_CRC): Ditto.
4518 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
4519 description of the CRC extension.
4521 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
4523 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
4524 * config/rs6000/linux.h: Ditto.
4525 * alpha/linux.h: Ditto.
4526 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
4527 no_c99_libc_has_function.
4528 * config/c6x/uclinux-elf.h: Ditto.
4529 * config/lm32/uclinux-elf.h: Ditto.
4530 * config/m68k/uclinux.h: Ditto.
4531 * config/moxie/uclinux.h: Ditto.
4532 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
4533 (crisv32-*-linux*, cris-*-linux*): Ditto.
4534 * config/bfin/bfin.c: Include "tm_p.h".
4536 2013-09-05 Richard Biener <rguenther@suse.de>
4538 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
4539 check for a definition without a basic-block.
4541 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
4542 Sofiane Naci <sofiane.naci@arm.com>
4544 * config/aarch64/aarch64.md
4545 (*movti_aarch64): Rename r_2_f and f_2_r.
4546 (*movsf_aarch64): Likewise.
4547 (*movdf_aarch64): Likewise.
4548 (*movtf_aarch64): Likewise.
4549 (aarch64_movdi_<mode>low): Likewise.
4550 (aarch64_movdi_<mode>high): Likewise.
4551 (aarch64_mov<mode>high_di): Likewise.
4552 (aarch64_mov<mode>low_di): Likewise.
4553 (aarch64_movtilow_tilow): Likewise.
4554 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
4555 values to config/arm/types.md
4556 (attribute "conds"): Update for attribute change.
4557 (anddi3_insn): Likewise.
4558 (iordi3_insn): Likewise.
4559 (xordi3_insn): Likewise.
4560 (one_cmpldi2): Likewise.
4561 * config/arm/types.md (type): Add Neon types.
4562 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
4563 use "type" attribute.
4564 (movmisalign<mode>_neon_store): Likewise.
4565 (movmisalign<mode>_neon_load): Likewise.
4566 (vec_set<mode>_internal): Likewise.
4567 (vec_setv2di_internal): Likewise.
4568 (vec_extract<mode>): Likewise.
4569 (vec_extractv2di): Likewise.
4570 (add<mode>3_neon): Likewise.
4571 (adddi3_neon): Likewise.
4572 (sub<mode>3_neon): Likewise.
4573 (subdi3_neon): Likewise.
4574 (mul<mode>3_neon): Likewise.
4575 (mul<mode>3add<mode>_neon): Likewise.
4576 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
4577 (fma<VCVTF:mode>4)): Likewise.
4578 (fma<VCVTF:mode>4_intrinsic): Likewise.
4579 (fmsub<VCVTF:mode>4)): Likewise.
4580 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
4581 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
4582 (ior<mode>3): Likewise.
4583 (and<mode>3): Likewise.
4584 (anddi3_neon): Likewise.
4585 (orn<mode>3_neon): Likewise.
4586 (orndi3_neon): Likewise.
4587 (bic<mode>3_neon): Likewise.
4588 (bicdi3_neon): Likewise.
4589 (xor<mode>3): Likewise.
4590 (one_cmpl<mode>2): Likewise.
4591 (abs<mode>2): Likewise.
4592 (neg<mode>2): Likewise.
4593 (umin<mode>3_neon): Likewise.
4594 (umax<mode>3_neon): Likewise.
4595 (smin<mode>3_neon): Likewise.
4596 (smax<mode>3_neon): Likewise.
4597 (vashl<mode>3): Likewise.
4598 (vashr<mode>3_imm): Likewise.
4599 (vlshr<mode>3_imm): Likewise.
4600 (ashl<mode>3_signed): Likewise.
4601 (ashl<mode>3_unsigned): Likewise.
4602 (neon_load_count): Likewise.
4603 (ashldi3_neon_noclobber): Likewise.
4604 (signed_shift_di3_neon): Likewise.
4605 (unsigned_shift_di3_neon): Likewise.
4606 (ashrdi3_neon_imm_noclobber): Likewise.
4607 (lshrdi3_neon_imm_noclobber): Likewise.
4608 (widen_ssum<mode>3): Likewise.
4609 (widen_usum<mode>3): Likewise.
4610 (quad_halves_<code>v4si): Likewise.
4611 (quad_halves_<code>v4sf): Likewise.
4612 (quad_halves_<code>v8hi): Likewise.
4613 (quad_halves_<code>v16qi): Likewise.
4614 (reduc_splus_v2di): Likewise.
4615 (neon_vpadd_internal<mode>): Likewise.
4616 (neon_vpsmin<mode>): Likewise.
4617 (neon_vpsmax<mode>): Likewise.
4618 (neon_vpumin<mode>): Likewise.
4619 (neon_vpumax<mode>): Likewise.
4620 (ss_add<mode>_neon): Likewise.
4621 (us_add<mode>_neon): Likewise.
4622 (ss_sub<mode>_neon): Likewise.
4623 (us_sub<mode>_neon): Likewise.
4624 (neon_vadd<mode>_unspec): Likewise.
4625 (neon_vaddl<mode>): Likewise.
4626 (neon_vaddw<mode>): Likewise.
4627 (neon_vhadd<mode>): Likewise.
4628 (neon_vqadd<mode>): Likewise.
4629 (neon_vaddhn<mode>): Likewise.
4630 (neon_vmul<mode>): Likewise.
4631 (neon_vmla<mode>): Likewise.
4632 (neon_vmlal<mode>): Likewise.
4633 (neon_vmls<mode>): Likewise.
4634 (neon_vmlsl<mode>): Likewise.
4635 (neon_vqdmulh<mode>): Likewise.
4636 (neon_vqdmlal<mode>): Likewise.
4637 (neon_vqdmlsl<mode>): Likewise.
4638 (neon_vmull<mode>): Likewise.
4639 (neon_vqdmull<mode>): Likewise.
4640 (neon_vsub<mode>_unspec): Likewise.
4641 (neon_vsubl<mode>): Likewise.
4642 (neon_vsubw<mode>): Likewise.
4643 (neon_vqsub<mode>): Likewise.
4644 (neon_vhsub<mode>): Likewise.
4645 (neon_vsubhn<mode>): Likewise.
4646 (neon_vceq<mode>): Likewise.
4647 (neon_vcge<mode>): Likewise.
4648 (neon_vcgeu<mode>): Likewise.
4649 (neon_vcgt<mode>): Likewise.
4650 (neon_vcgtu<mode>): Likewise.
4651 (neon_vcle<mode>): Likewise.
4652 (neon_vclt<mode>): Likewise.
4653 (neon_vcage<mode>): Likewise.
4654 (neon_vcagt<mode>): Likewise.
4655 (neon_vtst<mode>): Likewise.
4656 (neon_vabd<mode>): Likewise.
4657 (neon_vabdl<mode>): Likewise.
4658 (neon_vaba<mode>): Likewise.
4659 (neon_vabal<mode>): Likewise.
4660 (neon_vmax<mode>): Likewise.
4661 (neon_vmin<mode>): Likewise.
4662 (neon_vpaddl<mode>): Likewise.
4663 (neon_vpadal<mode>): Likewise.
4664 (neon_vpmax<mode>): Likewise.
4665 (neon_vpmin<mode>): Likewise.
4666 (neon_vrecps<mode>): Likewise.
4667 (neon_vrsqrts<mode>): Likewise.
4668 (neon_vqabs<mode>): Likewise.
4669 (neon_vqneg<mode>): Likewise.
4670 (neon_vcls<mode>): Likewise.
4671 (clz<mode>2): Likewise.
4672 (popcount<mode>2): Likewise.
4673 (neon_vrecpe): Likewise.
4674 (neon_vrsqrte): Likewise.
4675 (neon_vget_lane<mode>_sext_internal): Likewise.
4676 (neon_vget_lane<mode>_zext_internal): Likewise.
4677 (neon_vdup_n<mode>): Likewise.
4678 (neon_vdup_nv2di): Likewise.
4679 (neon_vdpu_lane<mode>_internal): Likewise.
4680 (neon_vswp<mode>): Likewise.
4681 (float<mode><V_cvtto>2): Likewise.
4682 (floatuns<mode><V_cvtto>2): Likewise.
4683 (fix_trunc<mode><V_cvtto>)2): Likewise
4684 (fixuns_trunc<mode><V_cvtto)2): Likewise.
4685 (neon_vcvt<mode>): Likewise.
4686 (neon_vcvtv4sfv4hf): Likewise.
4687 (neon_vcvtv4hfv4sf): Likewise.
4688 (neon_vcvt_n<mode>): Likewise.
4689 (neon_vmovn<mode>): Likewise.
4690 (neon_vqmovn<mode>): Likewise.
4691 (neon_vqmovun<mode>): Likewise.
4692 (neon_vmovl<mode>): Likewise.
4693 (neon_vmul_lane<mode>): Likewise.
4694 (neon_vmull_lane<mode>): Likewise.
4695 (neon_vqdmull_lane<mode>): Likewise.
4696 (neon_vqdmulh_lane<mode>): Likewise.
4697 (neon_vmla_lane<mode>): Likewise.
4698 (neon_vmlal_lane<mode>): Likewise.
4699 (neon_vqdmlal_lane<mode>): Likewise.
4700 (neon_vmls_lane<mode>): Likewise.
4701 (neon_vmlsl_lane<mode>): Likewise.
4702 (neon_vqdmlsl_lane<mode>): Likewise.
4703 (neon_vext<mode>): Likewise.
4704 (neon_vrev64<mode>): Likewise.
4705 (neon_vrev32<mode>): Likewise.
4706 (neon_vrev16<mode>): Likewise.
4707 (neon_vbsl<mode>_internal): Likewise.
4708 (neon_vshl<mode>): Likewise.
4709 (neon_vqshl<mode>): Likewise.
4710 (neon_vshr_n<mode>): Likewise.
4711 (neon_vshrn_n<mode>): Likewise.
4712 (neon_vqshrn_n<mode>): Likewise.
4713 (neon_vqshrun_n<mode>): Likewise.
4714 (neon_vshl_n<mode>): Likewise.
4715 (neon_vqshl_n<mode>): Likewise.
4716 (neon_vqshlu_n<mode>): Likewise.
4717 (neon_vshll_n<mode>): Likewise.
4718 (neon_vsra_n<mode>): Likewise.
4719 (neon_vsri_n<mode>): Likewise.
4720 (neon_vsli_n<mode>): Likewise.
4721 (neon_vtbl1v8qi): Likewise.
4722 (neon_vtbl2v8qi): Likewise.
4723 (neon_vtbl3v8qi): Likewise.
4724 (neon_vtbl4v8qi): Likewise.
4725 (neon_vtbx1v8qi): Likewise.
4726 (neon_vtbx2v8qi): Likewise.
4727 (neon_vtbx3v8qi): Likewise.
4728 (neon_vtbx4v8qi): Likewise.
4729 (neon_vtrn<mode>_internal): Likewise.
4730 (neon_vzip<mode>_internal): Likewise.
4731 (neon_vuzp<mode>_internal): Likewise.
4732 (neon_vld1<mode>): Likewise.
4733 (neon_vld1_lane<mode>): Likewise.
4734 (neon_vld1_dup<mode>): Likewise.
4735 (neon_vld1_dupv2di): Likewise.
4736 (neon_vst1<mode>): Likewise.
4737 (neon_vst1_lane<mode>): Likewise.
4738 (neon_vld2<mode>): Likewise.
4739 (neon_vld2_lane<mode>): Likewise.
4740 (neon_vld2_dup<mode>): Likewise.
4741 (neon_vst2<mode>): Likewise.
4742 (neon_vst2_lane<mode>): Likewise.
4743 (neon_vld3<mode>): Likewise.
4744 (neon_vld3qa<mode>): Likewise.
4745 (neon_vld3qb<mode>): Likewise.
4746 (neon_vld3_lane<mode>): Likewise.
4747 (neon_vld3_dup<mode>): Likewise.
4748 (neon_vst3<mode>): Likewise.
4749 (neon_vst3qa<mode>): Likewise.
4750 (neon_vst3qb<mode>): Likewise.
4751 (neon_vst3_lane<mode>): Likewise.
4752 (neon_vld4<mode>): Likewise.
4753 (neon_vld4qa<mode>): Likewise.
4754 (neon_vld4qb<mode>): Likewise.
4755 (neon_vld4_lane<mode>): Likewise.
4756 (neon_vld4_dup<mode>): Likewise.
4757 (neon_vst4<mode>): Likewise.
4758 (neon_vst4qa<mode>): Likewise.
4759 (neon_vst4qb<mode>): Likewise.
4760 (neon_vst4_lane<mode>): Likewise.
4761 (neon_vec_unpack<US>_lo_<mode>): Likewise.
4762 (neon_vec_unpack<US>_hi_<mode>): Likewise.
4763 (neon_vec_<US>mult_lo_<mode>): Likewise.
4764 (neon_vec_<US>mult_hi_<mode>): Likewise.
4765 (neon_vec_<US>shiftl_<mode>): Likewise.
4766 (neon_unpack<US>_<mode>): Likewise.
4767 (neon_vec_<US>mult_<mode>): Likewise.
4768 (vec_pack_trunc_<mode>): Likewise.
4769 (neon_vec_pack_trunk_<mode>): Likewise.
4770 (neon_vabd<mode>_2): Likewise.
4771 (neon_vabd<mode>_3): Likewise.
4772 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
4773 (thumb2_movsi_vfp): Likewise.
4774 (movdi_vfp): Likewise.
4775 (movdi_vfp_cortexa8): Likewise.
4776 (movhf_vfp_neon): Likewise.
4777 (movhf_vfp): Likewiwse.
4778 (movsf_vfp): Likewiwse.
4779 (thumb2_movsf_vfp): Likewiwse.
4780 (movdf_vfp): Likewise.
4781 (thumb2_movdf_vfp): Likewise.
4782 (movsfcc_vfp): Likewise.
4783 (thumb2_movsfcc_vfp): Likewise.
4784 (movdfcc_vfp): Likewise.
4785 (thumb2_movdfcc_vfp): Likewise.
4786 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
4787 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
4788 (v10_v2c): Likewise.
4789 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
4791 (cortex_a15_neon_int_2): Likewise.
4792 (cortex_a15_neon_int_3): Likewise.
4793 (cortex_a15_neon_int_4): Likewise.
4794 (cortex_a15_neon_int_5): Likewise.
4795 (cortex_a15_neon_vqneg_vqabs): Likewise.
4796 (cortex_a15_neon_vmov): Likewise.
4797 (cortex_a15_neon_vaba): Likewise.
4798 (cortex_a15_neon_vaba_qqq): Likewise.
4799 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4800 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
4801 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
4802 scalar_64_32_long_scalar): Likewise.
4803 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4804 (cortex_a15_neon_mla_qqq_8_16): Likewise.
4805 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
4806 lotype_qdd_64_32_long): Likewise.
4807 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
4808 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
4809 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
4810 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
4811 (cortex_a15_neon_shift_1): Likewise.
4812 (cortex_a15_neon_shift_2): Likewise.
4813 (cortex_a15_neon_shift_3): Likewise.
4814 (cortex_a15_neon_vshl_ddd): Likewise.
4815 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
4816 (cortex_a15_neon_vsra_vrsra): Likewise.
4817 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
4818 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
4819 (cortex_a15_neon_fp_vmul_ddd): Likewise.
4820 (cortex_a15_neon_fp_vmul_qqd): Likewise.
4821 (cortex_a15_neon_fp_vmla_ddd): Likewise.
4822 (cortex_a15_neon_fp_vmla_qqq): Likewise.
4823 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
4824 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
4825 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
4826 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
4827 (cortex_a15_neon_bp_simple): Likewise.
4828 (cortex_a15_neon_bp_2cycle): Likewise.
4829 (cortex_a15_neon_bp_3cycle): Likewise.
4830 (cortex_a15_neon_vld1_1_2_regs): Likewise.
4831 (cortex_a15_neon_vld1_3_4_regs): Likewise.
4832 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
4833 (cortex_a15_neon_vld2_4_regs): Likewise.
4834 (cortex_a15_neon_vld3_vld4): Likewise.
4835 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
4836 (cortex_a15_neon_vst1_3_4_regs): Likewise.
4837 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
4838 (cortex_a15_neon_vst3_vst4): Likewise.
4839 (cortex_a15_neon_vld1_vld2_lane): Likewise.
4840 (cortex_a15_neon_vld3_vld4_lane" 10
4841 (cortex_a15_neon_vst1_vst2_lane): Likewise.
4842 (cortex_a15_neon_vst3_vst4_lane): Likewise.
4843 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
4844 (cortex_a15_neon_ldm_2): Likewise.0
4845 (cortex_a15_neon_stm_2): Likewise.
4846 (cortex_a15_neon_mcr): Likewise.
4847 (cortex_a15_neon_mcr_2_mcrr): Likewise.
4848 (cortex_a15_neon_mrc): Likewise.
4849 (cortex_a15_neon_mrrc): Likewise.
4850 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
4852 (cortex_a15_alu_shift): Likewise.
4853 (cortex_a15_alu_shift_reg): Likewise.
4854 (cortex_a15_mult32): Likewise.
4855 (cortex_a15_mult64): Likewise.
4856 (cortex_a15_block): Likewise.
4857 (cortex_a15_branch): Likewise.
4858 (cortex_a15_load1): Likewise.
4859 (cortex_a15_load3): Likewise.
4860 (cortex_a15_store1): Likewise.
4861 (cortex_a15_store3): Likewise.
4862 (cortex_a15_call): Likewise.
4863 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
4864 (cortex_a5_f2r): Likewise.
4865 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
4867 (cortex_a53_f2r): Likewise.
4868 * config/arm/cortex-a7.md
4869 (cortex_a7_branch): Update for attribute change.
4870 (cortex_a7_call): Likewise.
4871 (cortex_a7_alu_imm): Likewise.
4872 (cortex_a7_alu_reg): Likewise.
4873 (cortex_a7_alu_shift): Likewise.
4874 (cortex_a7_mul): Likewise.
4875 (cortex_a7_load1): Likewise.
4876 (cortex_a7_store1): Likewise.
4877 (cortex_a7_load2): Likewise.
4878 (cortex_a7_store2): Likewise.
4879 (cortex_a7_load3): Likewise.
4880 (cortex_a7_store3): Likewise.
4881 (cortex_a7_load4): Likewise.
4882 (cortex_a7_store4): Likewise.
4883 (cortex_a7_fpalu): Likewise.
4884 (cortex_a7_fconst): Likewise.
4885 (cortex_a7_fpmuls): Likewise.
4886 (cortex_a7_neon_mul): Likewise.
4887 (cortex_a7_fpmacs): Likewise.
4888 (cortex_a7_neon_mla: Likewise.
4889 (cortex_a7_fpmuld: Likewise.
4890 (cortex_a7_fpmacd: Likewise.
4891 (cortex_a7_fpfmad: Likewise.
4892 (cortex_a7_fdivs: Likewise.
4893 (cortex_a7_fdivd: Likewise.
4894 (cortex_a7_r2f: Likewise.
4895 (cortex_a7_f2r: Likewise.
4896 (cortex_a7_f_flags: Likewise.
4897 (cortex_a7_f_loads: Likewise.
4898 (cortex_a7_f_loadd: Likewise.
4899 (cortex_a7_f_stores: Likewise.
4900 (cortex_a7_f_stored: Likewise.
4901 (cortex_a7_neon): Likewise.
4902 * config/arm/cortex-a8-neon.md
4903 (cortex_a8_neon_mrc): Update for attribute change.
4904 (cortex_a8_neon_mrrc): Likewise.
4905 (cortex_a8_neon_int_1): Likewise.
4906 (cortex_a8_neon_int_2): Likewise.
4907 (cortex_a8_neon_int_3): Likewise.
4908 (cortex_a8_neon_int_4): Likewise.
4909 (cortex_a8_neon_int_5): Likewise.
4910 (cortex_a8_neon_vqneg_vqabs): Likewise.
4911 (cortex_a8_neon_vmov): Likewise.
4912 (cortex_a8_neon_vaba): Likewise.
4913 (cortex_a8_neon_vaba_qqq): Likewise.
4914 (cortex_a8_neon_vsma): Likewise.
4915 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4916 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
4917 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
4918 long_scalar): Likewise.
4919 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4920 (cortex_a8_neon_mla_qqq_8_16): Likewise.
4921 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
4922 long_scalar_qdd_64_32_long): Likewise.
4923 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
4924 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
4925 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
4926 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
4927 (cortex_a8_neon_shift_1): Likewise.
4928 (cortex_a8_neon_shift_2): Likewise.
4929 (cortex_a8_neon_shift_3): Likewise.
4930 (cortex_a8_neon_vshl_ddd): Likewise.
4931 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
4932 (cortex_a8_neon_vsra_vrsra): Likewise.
4933 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
4934 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
4935 (cortex_a8_neon_fp_vsum): Likewise.
4936 (cortex_a8_neon_fp_vmul_ddd): Likewise.
4937 (cortex_a8_neon_fp_vmul_qqd): Likewise.
4938 (cortex_a8_neon_fp_vmla_ddd): Likewise.
4939 (cortex_a8_neon_fp_vmla_qqq): Likewise.
4940 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
4941 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
4942 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
4943 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
4944 (cortex_a8_neon_bp_simple): Likewise.
4945 (cortex_a8_neon_bp_2cycle): Likewise.
4946 (cortex_a8_neon_bp_3cycle): Likewise.
4947 (cortex_a8_neon_ldr): Likewise.
4948 (cortex_a8_neon_str): Likewise.
4949 (cortex_a8_neon_vld1_1_2_regs): Likewise.
4950 (cortex_a8_neon_vld1_3_4_regs): Likewise.
4951 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
4952 (cortex_a8_neon_vld2_4_regs): Likewise.
4953 (cortex_a8_neon_vld3_vld4): Likewise.
4954 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
4955 (cortex_a8_neon_vst1_3_4_regs): Likewise.
4956 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
4957 (cortex_a8_neon_vst3_vst4): Likewise.
4958 (cortex_a8_neon_vld1_vld2_lane): Likewise.
4959 (cortex_a8_neon_vld3_vld4_lane): Likewise.
4960 (cortex_a8_neon_vst1_vst2_lane): Likewise.
4961 (cortex_a8_neon_vst3_vst4_lane): Likewise.
4962 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
4963 (cortex_a8_neon_mcr): Likewise.
4964 (cortex_a8_neon_mcr_2_mcrr): Likewise.
4965 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
4966 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
4968 (ca9_neon_mrrc): Likewise.
4969 (cortex_a9_neon_int_1): Likewise.
4970 (cortex_a9_neon_int_2): Likewise.
4971 (cortex_a9_neon_int_3): Likewise.
4972 (cortex_a9_neon_int_4): Likewise.
4973 (cortex_a9_neon_int_5): Likewise.
4974 (cortex_a9_neon_vqneg_vqabs): Likewise.
4975 (cortex_a9_neon_vmov): Likewise.
4976 (cortex_a9_neon_vaba): Likewise.
4977 (cortex_a9_neon_vaba_qqq): Likewise.
4978 (cortex_a9_neon_vsma): Likewise.
4979 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4980 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
4981 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
4982 long_scalar): Likewise.
4983 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4984 (cortex_a9_neon_mla_qqq_8_16): Likewise.
4985 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
4986 long_scalar_qdd_64_32_long): Likewise.
4987 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
4988 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
4989 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
4990 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
4991 (cortex_a9_neon_shift_1): Likewise.
4992 (cortex_a9_neon_shift_2): Likewise.
4993 (cortex_a9_neon_shift_3): Likewise.
4994 (cortex_a9_neon_vshl_ddd): Likewise.
4995 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
4996 (cortex_a9_neon_vsra_vrsra): Likewise.
4997 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
4998 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
4999 (cortex_a9_neon_fp_vsum): Likewise.
5000 (cortex_a9_neon_fp_vmul_ddd): Likewise.
5001 (cortex_a9_neon_fp_vmul_qqd): Likewise.
5002 (cortex_a9_neon_fp_vmla_ddd): Likewise.
5003 (cortex_a9_neon_fp_vmla_qqq): Likewise.
5004 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
5005 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
5006 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
5007 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
5008 (cortex_a9_neon_bp_simple): Likewise.
5009 (cortex_a9_neon_bp_2cycle): Likewise.
5010 (cortex_a9_neon_bp_3cycle): Likewise.
5011 (cortex_a9_neon_ldr): Likewise.
5012 (cortex_a9_neon_str): Likewise.
5013 (cortex_a9_neon_vld1_1_2_regs): Likewise.
5014 (cortex_a9_neon_vld1_3_4_regs): Likewise.
5015 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
5016 (cortex_a9_neon_vld2_4_regs): Likewise.
5017 (cortex_a9_neon_vld3_vld4): Likewise.
5018 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
5019 (cortex_a9_neon_vst1_3_4_regs): Likewise.
5020 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
5021 (cortex_a9_neon_vst3_vst4): Likewise.
5022 (cortex_a9_neon_vld1_vld2_lane): Likewise.
5023 (cortex_a9_neon_vld3_vld4_lane): Likewise.
5024 (cortex_a9_neon_vst1_vst2_lane): Likewise.
5025 (cortex_a9_neon_vst3_vst4_lane): Likewise.
5026 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
5027 (cortex_a9_neon_mcr): Likewise.
5028 (cortex_a9_neon_mcr_2_mcrr): Likewise.
5029 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
5030 (cortex_a9_fps): Likewise.
5031 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
5033 (cortex_m4_fmuls): Likewise.
5034 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
5036 (cortex_r4_mrc): Likewise.
5037 * config/arm/iterators.md: Update comment referring to neon_type.
5038 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
5039 (iwmmxt_movsi_insn): Likewise.
5040 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
5042 (pj4_core_to_vfp): Likewise.
5043 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
5045 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
5046 (vfp_fstore): Likewise.
5047 * doc/md.texi: Change references to neon_type to refer to type.
5049 2013-09-04 Dodji Seketeli <dodji@redhat.com>
5051 * tree.h (DECL_BUILT_IN): Fix typo in comment.
5053 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
5055 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
5058 2013-09-04 Easwaran Raman <eraman@google.com>
5061 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
5062 (build_and_add_sum): Use it.
5063 (appears_later_in_bb): Simplify code.
5065 2013-09-04 Teresa Johnson <tejohnson@google.com>
5067 * dumpfile.c (dump_finish): Don't close stderr/stdout.
5069 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
5071 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
5073 2013-09-04 Jan Hubicka <jh@suse.cz>
5075 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
5076 * ipa-devirt.c: Include diganostic.h
5077 (odr_type_d): Add types and types_set.
5078 (hash_type_name): Work for types with vtables during LTO.
5079 (odr_hasher::remove): Fix comment; destroy types_set.
5080 (add_type_duplicate): New function,
5081 (get_odr_type): Use it.
5082 (dump_type_inheritance_graph): Dump type duplicates.
5083 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
5085 * tree.c (types_same_for_odr): Give exact answers on types with
5088 2013-09-04 Dodji Seketeli <dodji@redhat.com>
5090 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
5091 explaining their differences.
5093 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
5095 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
5097 2013-09-03 Jeff Law <law@redhat.com>
5099 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
5100 when not threading through a joiner block. Pass joiner/no joiner
5101 state to register_jump_thread.
5102 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
5103 state from argument rather than implying on path length.
5104 Dump the entire jump thread path into debugging dump.
5105 * tree-flow.h (register_jump_thread): Update prototype.
5107 2013-08-29 Xinliang David Li <davidxl@google.com>
5109 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
5110 Remove a trivial gcc_assert.
5112 2013-08-29 Xinliang David Li <davidxl@google.com>
5114 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
5115 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
5116 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
5117 Delay base decl alignment adjustment.
5118 * tree-vectorizer.c (vect_destroy_datarefs): New function.
5119 * tree-vectorizer.h: New data structure.
5120 (set_dr_misalignment): New function.
5121 (dr_misalignment): Ditto.
5122 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
5123 (vectorizable_load): Ditto.
5124 (ensure_base_align): New function.
5125 (vectorize_loops): Add dbg_cnt support.
5126 (execute_vect_slp): Ditto.
5127 * dbgcnt.def: New debug counter.
5128 * Makefile: New dependency.
5130 2013-09-03 Meador Inge <meadori@codesourcery.com>
5134 2013-08-30 Meador Inge <meadori@codesourcery.com>
5136 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
5138 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
5140 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
5141 function descriptor.
5143 2013-09-03 Richard Biener <rguenther@suse.de>
5145 * tree-affine.c (add_elt_to_tree): Fix association issue,
5146 avoid useless converts and make sure to always return a
5147 properly typed result.
5149 2013-09-03 Richard Biener <rguenther@suse.de>
5152 * fold-const.c (negate_expr_p): Fix division case.
5153 (negate_expr): Likewise.
5155 2013-09-03 Richard Biener <rguenther@suse.de>
5158 * tree-streamer-out.c: Include tm.h.
5159 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
5161 2013-09-03 Jan Hubicka <jh@suse.cz>
5163 * tree-profile.c (tree_profiling): Cleanup CFG when done.
5165 2013-09-03 Alan Modra <amodra@gmail.com>
5167 * config.gcc (powerpc*-*-linux*): Add support for little-endian
5168 multilibs to big-endian target and vice versa.
5169 * config/rs6000/t-linux64: Use := assignment on all vars.
5170 (MULTILIB_EXTRA_OPTS): Remove fPIC.
5171 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
5172 * config/rs6000/t-linux64le: New file.
5173 * config/rs6000/t-linux64bele: New file.
5174 * config/rs6000/t-linux64lebe: New file.
5176 2013-09-02 Jan Hubicka <jh@suse.cz>
5178 * ipa-inline-transform.c (inline_transform): Do not
5179 optimize_inline_calls when not optimizing.
5181 2013-09-02 Jan Hubicka <jh@suse.cz>
5183 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
5184 duplicated nodes for assembler names.
5185 * symtab.c (symtab_unregister_node): Do not attempt to unlink
5186 hard registers from assembler name hash.
5188 2013-09-02 Jan Hubicka <jh@suse.cz>
5190 * ipa-split.c (execute_split_functions): Split externally visible
5191 functions called once.
5193 2013-09-02 Martin Jambor <mjambor@suse.cz>
5196 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
5197 linked list. When finding the correct duplicate, also consider also
5198 the caller in additon to its inlined_to node.
5200 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
5202 * config/aarch64/aarch64-simd-builtins.def
5203 (dup_lane_scalar): Remove.
5204 * config/aarch64/aarch64-simd.md
5205 (aarch64_simd_dup): Add 'w->w' alternative.
5206 (aarch64_dup_lane<mode>): Allow for VALL.
5207 (aarch64_dup_lane_scalar<mode>): Remove.
5208 (aarch64_dup_lane_<vswap_width_name><mode>): New.
5209 (aarch64_get_lane_signed<mode>): Add w->w altenative.
5210 (aarch64_get_lane_unsigned<mode>): Likewise.
5211 (aarch64_get_lane<mode>): Likewise.
5212 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
5213 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
5214 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
5215 (VCON): Change container of V2SF.
5216 (vswap_width_name): Likewise.
5217 * config/aarch64/arm_neon.h
5218 (__aarch64_vdup_lane_any): New.
5219 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
5220 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
5221 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
5223 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
5226 * expr.c (emit_move_complex): Do not move complex FP values as parts if
5227 the source or the destination is a single hard register.
5229 2013-09-02 Richard Biener <rguenther@suse.de>
5232 * tree-scalar-evolution.c (instantiate_scev_name): Allow
5235 2013-09-02 Richard Biener <rguenther@suse.de>
5237 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
5238 arithmetic to sizetype.
5240 2013-09-02 Bin Cheng <bin.cheng@arm.com>
5242 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
5243 Find auto-increment use both before and after candidate.
5245 2013-09-02 Marek Polacek <polacek@redhat.com>
5247 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
5249 2013-09-01 Jan Hubicka <jh@suse.cz>
5251 * Makefile.in: Add ipa-profile.o
5252 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
5253 * cgraph.c (struct cgraph_propagate_frequency_data,
5254 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
5255 ipa-profile.c; replace cgraph_ by ipa_ prefix.
5256 * cgraph.h (cgraph_propagate_frequency): Remove.
5257 * ipa-inline-analysis.c: Include ipa-utils.h;
5258 drop duplicated cfgloop.h.
5259 (inline_update_callee_summaries): Update.
5260 * ipa-profile.c: New file.
5261 * ipa-utils.h (ipa_propagate_frequency): Declare.
5262 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
5263 data-streamer.h, value-prof.h.
5264 (symtab_remove_unreachable_nodes): Update profile.
5265 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
5266 account_time_size, cmp_counts, dump_histogram,
5267 ipa_profile_generate_summary, ipa_profile_write_summary,
5268 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
5269 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
5270 Move to ipa-profile.c.
5272 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
5274 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
5276 2013-09-01 Jan Hubicka <jh@suse.cz>
5278 * common.opt (fdevirtualize-speculatively): New function.
5279 * invoke.texi (fdevirtualize-speculatively): Document.
5280 * ipa-devirt.c: Include ipa-inline.h
5281 (likely_target_p): New function.
5282 (ipa_devirt): New function.
5283 (gate_ipa_devirt): New function.
5284 (pass_data_ipa_devirt): New static var.
5285 (pass_ipa_devirt): Likewise.
5286 (make_pass_ipa_devirt): New function.
5287 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
5288 (common_handle_option): Disable devirtualization when
5289 value range profiling is available.
5290 * passes.def (pass_ipa_devirt): Add.
5291 * timever.def (TV_IPA_DEVIRT): New timevar.
5292 * tree-pass.h (make_pass_ipa_devirt):
5294 2013-09-01 Iain Sandoe <iain@codesourcery.com>
5296 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
5297 include sanitize(undefined).
5299 2013-08-31 Diego Novillo <dnovillo@google.com>
5301 * Makefile.in (TREE_CORE_H): Define.
5303 (GTFILES): Add tree-core.h.
5304 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
5306 * tree-core.h: New file.
5307 Move all data structures, enum, typedefs, global
5308 declarations and constants from ...
5311 2013-08-31 Jan Hubicka <jh@suse.cz>
5313 * bulitins.c (expand_builtin): Do not early exit for gcov
5314 instrumented functions.
5316 2013-08-31 Marek Polacek <polacek@redhat.com>
5318 * ubsan.c: Include tm_p.h.
5320 2013-08-31 Jan Hubicka <jh@suse.cz>
5322 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
5325 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
5326 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
5327 * tree-cfg.c (verify_gimple_label): ... here.
5328 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
5329 (ipa_merge_profiles): New function.
5330 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
5331 (lto_input_function_body): Likewise.
5332 * ipa-utils.h (ipa_merge_profiles): Declare.
5333 * lto-streamer.h (lto_input_function_body): Update prototype.
5334 (emit_label_in_global_context_p): Remove.
5335 * lto-symtab.c: Include ipa-utils.h
5336 (lto_cgraph_replace_node): Use ipa_merge_profiles.
5338 2013-08-31 Jan Hubicka <jh@suse.cz>
5340 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
5342 2013-08-31 Jan Hubicka <jh@suse.cz>
5344 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
5346 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
5348 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
5349 "cmp" RTX before signed_comparison_operator check to account
5352 2013-08-30 Jan Hubicka <jh@suse.cz>
5354 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
5355 (detect_type_change_1): Rename to ...
5356 (detect_type_change): ... this one; early return on non-polymorphic
5358 (detect_type_change_ssa): Add comp_type parameter; update
5359 use of detect_type_change.
5360 (compute_complex_assign_jump_func): Add param_type parameter;
5361 update use of detect_type_change_ssa.
5362 (compute_complex_ancestor_jump_func): Likewise.
5363 (ipa_get_callee_param_type): New function.
5364 (ipa_compute_jump_functions_for_edge): Compute parameter type;
5365 update calls to the jump function computation functions.
5367 2013-08-30 Teresa Johnson <tejohnson@google.com>
5368 Steven Bosscher <steven@gcc.gnu.org>
5370 * cfgrtl.c (fixup_new_cold_bb): New routine.
5371 (commit_edge_insertions): Invoke fixup_partitions.
5372 (find_partition_fixes): New routine.
5373 (fixup_partitions): Ditto.
5374 (verify_hot_cold_block_grouping): Update comments.
5375 (rtl_verify_edges): Invoke find_partition_fixes.
5376 (rtl_verify_bb_pointers): Update comments.
5377 (rtl_verify_bb_layout): Ditto.
5378 * basic-block.h (probably_never_executed_edge_p): Declare.
5379 (fixup_partitions): Ditto.
5380 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
5381 * bb-reorder.c (sanitize_hot_paths): New function.
5382 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
5384 * predict.c (probably_never_executed_edge_p): New routine.
5385 * cfg.c (check_bb_profile): Add partition insanity warnings.
5387 2013-08-30 Meador Inge <meadori@codesourcery.com>
5389 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
5391 2013-08-30 Marek Polacek <polacek@redhat.com>
5393 * Makefile.in (ubsan.o): Add.
5394 (c-family/c-ubsan.o): Add.
5395 (builtins.o): Add ubsan.h dependency.
5396 * ubsan.h: New file.
5397 * ubsan.c: New file.
5398 * common.opt: Add -fsanitize=undefined option.
5399 (flag_sanitize): Add variable.
5400 (fsanitize=): Add option. Add Driver.
5401 (fsanitize=thread): Remove option.
5402 (fsanitize=address): Likewise.
5403 (static-libubsan): New option.
5404 * doc/invoke.texi: Document the new flag and -static-libubsan.
5405 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
5406 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
5407 * builtin-attrs.def (ATTR_COLD): Define.
5408 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
5409 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
5410 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
5411 * flag-types.h (sanitize_code): New enum.
5412 * opts.c (common_handle_option): Parse command line arguments
5413 of -fsanitize=. Add -fsanitize=unreachable option.
5414 * varasm.c (get_variable_section): Adjust.
5415 (assemble_noswitch_variable): Likewise.
5416 (assemble_variable): Likewise.
5417 (output_constant_def_contents): Likewise.
5418 (categorize_decl_for_section): Likewise.
5419 (place_block_symbol): Likewise.
5420 (output_object_block): Likewise.
5421 * builtins.def: Likewise.
5422 * toplev.c (compile_file): Likewise.
5423 (process_options): Likewise.
5424 * cppbuiltin.c: Likewise.
5425 * tsan.c (tsan_pass): Likewise.
5426 (tsan_gate): Likewise.
5427 (tsan_gate_O0): Likewise.
5428 * cfgexpand.c (partition_stack_vars): Likewise.
5429 (expand_stack_vars): Likewise.
5430 (defer_stack_allocation): Likewise.
5431 (expand_used_vars): Likewise.
5432 * cfgcleanup.c (old_insns_match_p): Likewise.
5433 * asan.c (asan_finish_file): Likewise.
5434 (asan_instrument): Likewise.
5435 (gate_asan): Likewise.
5436 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
5437 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
5438 (asan_global_struct): Use pointer_sized_int_node instead
5439 calling build_nonstandard_integer_type.
5440 (initialize_sanitizer_builtins): Likewise.
5441 (asan_finish_file): Likewise.
5442 * gcc.c: Document %{%:function(args):X}.
5443 (static_spec_functions): Add sanitize.
5444 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
5445 store funcval != NULL there.
5446 (do_spec_1): Adjust handle_spec_function caller.
5447 (handle_braces): Allow %:function(args) as condition.
5448 (sanitize_spec_function): New function.
5449 (ADD_STATIC_LIBUBSAN_LIBS): Define.
5450 (LIBUBSAN_SPEC): Likewise.
5451 (LIBUBSAN_EARLY_SPEC): Likewise.
5452 (SANITIZER_SPEC): Handle libubsan.
5453 (SANITIZER_EARLY_SPEC): Likewise.
5454 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
5455 instead of fsanitize=address.
5456 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
5457 instead of fsanitize=address*.
5458 * builtins.c: Include ubsan.h.
5459 (fold_builtin_0): Instrument __builtin_unreachable.
5460 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
5461 instead of flag_asan.
5462 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
5463 (pointer_sized_int_node): Define.
5464 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
5466 2013-08-30 Mike Stump <mikestump@comcast.net>
5468 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
5471 2013-08-29 Jan Hubicka <jh@suse.cz>
5473 * cgraph.c (cgraph_function_body_availability): Handle weakref
5475 * passes.def: Remove pass_fixup_cfg.
5476 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
5477 track when we need to remove functions.
5478 (gate_ipa_inline): Execute inlining always; add comment why.
5479 (pass_data_ipa_inline): Remove TODO_remove_functions.
5480 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
5481 do not produce summaries.
5482 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
5483 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
5484 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
5485 constant pool and vtable.
5487 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
5489 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
5490 New arm_neon.h's internal macros to specify 64-bit constants.
5491 Avoid using stdint.h's macros.
5493 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
5495 * recog.c (verify_changes): Verify that changes[i].old is non-zero
5496 before applying REG_P.
5498 2013-08-30 Jakub Jelinek <jakub@redhat.com>
5500 PR tree-optimization/58277
5501 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
5502 after seeing too many stmts with vdef in between dombb and current
5503 bb, invalidate everything.
5505 2013-08-30 Richard Biener <rguenther@suse.de>
5507 * fold-const.c (fold_single_bit_test): Fix overflow test.
5509 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
5511 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
5512 and which can live in a register, always retrieve the value on entry.
5513 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
5514 passed by invisible reference specially.
5515 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
5516 (vt_add_function_parameter): Correctly deal with a parameter passed by
5517 invisible reference.
5519 2013-08-30 Jan Hubicka <jh@suse.cz>
5521 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
5523 2013-08-30 Richard Biener <rguenther@suse.de>
5525 PR tree-optimization/58228
5526 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
5527 allow invariant loads in nested loop vectorization.
5529 2013-08-30 Richard Biener <rguenther@suse.de>
5531 PR tree-optimization/58223
5532 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
5533 (has_anti_or_output_dependence): ... this and adjust to also
5534 look for output dependences.
5535 (mark_nodes_having_upstream_mem_writes): Adjust.
5536 (rdg_flag_uses): Likewise.
5538 2013-08-30 Richard Biener <rguenther@suse.de>
5540 PR tree-optimization/58010
5541 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
5542 assert that we have a loop-closed PHI.
5544 2013-08-29 Jan Hubicka <jh@suse.cz>
5546 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
5547 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
5548 * lto-section-in.c (lto_free_function_in_decl_state): New function.
5549 (lto_free_function_in_decl_state_for_node): New function.
5551 2013-08-29 Xinliang David Li <davidxl@google.com>
5553 * loop-unroll.c (report_unroll_peel): Minor message change.
5554 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
5555 Emit alignment peeling message with default -fopt-info.
5556 (vect_loop_versioning): Emit loop version info message.
5557 * tree-vectorizer.c (vectorize_loops): Minor message change.
5558 (execute_vect_slp): Ditto.
5560 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
5562 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
5563 of the clone from the DECL_NAME of the original function.
5565 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
5567 * passes.c (register_pass): Add overload.
5568 * tree-pass.h (register_pass): Forward declare it. Add comment.
5570 2013-08-29 Jan Hubicka <jh@suse.cz>
5572 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
5573 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
5574 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
5575 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
5576 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
5577 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
5578 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
5579 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
5581 2013-08-29 Teresa Johnson <tejohnson@google.com>
5583 * dumpfile.c (dump_loc): Output column number.
5584 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
5585 * doc/invoke.texi: Document optall -fopt-info flag.
5586 * profile.c (read_profile_edge_counts): Use new dump framework.
5587 (compute_branch_probabilities): Ditto.
5588 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
5589 when pass not in any opt group.
5590 * pass_manager.h (pass_manager::get_pass_profile): New method.
5591 * value-prof.c (check_counter): Use new dump framework.
5592 (check_ic_target): Ditto.
5593 * coverage.c (get_coverage_counts): Ditto.
5594 (coverage_init): Setup new dump framework.
5596 2013-08-29 Richard Biener <rguenther@suse.de>
5598 PR tree-optimization/58246
5599 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
5600 handle the dominance check inside a basic-block.
5602 2013-08-29 Richard Biener <rguenther@suse.de>
5605 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
5606 of default defs that appear in abnormal PHI nodes.
5608 2013-08-29 Richard Biener <rguenther@suse.de>
5610 PR tree-optimization/57685
5611 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
5612 single-use operands to avoid exponential complexity.
5614 2013-08-28 Dehao Chen <dehao@google.com>
5616 * ipa-inline.c (edge_badness): Fix integer underflow.
5618 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
5620 * gtm-builtins.def (_ITM_free): Declare leaf.
5622 2013-08-28 Jakub Jelinek <jakub@redhat.com>
5625 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
5626 (*tls_local_dynamic_base_64_largepic): Likewise.
5627 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
5628 Remove predicate from call operand.
5629 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
5630 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
5632 2013-08-28 Jeff Law <law@redhat.com>
5634 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
5635 checks for the number of predecessors and successors allowed.
5636 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
5637 which require copying a joiner block if there is a request which
5638 is a subpath that requires no joiner block copying.
5640 2013-08-28 Jan Hubicka <jh@suse.cz>
5642 * lto-streamer-out.c (DFS_write_tree_body): Drop
5643 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
5644 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
5645 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
5646 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
5647 Do not read DECL_ERROR_ISSUED.
5648 (unpack_ts_decl_with_vis_value_fields): Do not read
5650 (lto_input_ts_binfo_tree_pointers): Do not read
5651 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
5652 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
5653 write DECL_ERROR_ISSUED..
5654 (pack_ts_decl_with_vis_value_fields): Do not write
5656 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
5657 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
5658 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
5659 * tree.h (tree_decl_common): Update comment.
5660 (DECL_ERROR_ISSUED): Remove.
5662 2013-08-28 Jakub Jelinek <jakub@redhat.com>
5665 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
5667 2013-08-28 Jan Hubicka <jh@suse.cz>
5669 * builtins.def (free): Declare leaf.
5671 2013-08-27 David Malcolm <dmalcolm@redhat.com>
5674 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
5675 * configure: Regenerate.
5677 2013-08-27 Martin Jambor <mjambor@suse.cz>
5679 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
5680 (ipa_ancestor_jf_data): Likewise.
5681 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
5682 (ipa_get_jf_pass_through_type_preserved): New function.
5683 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
5684 (ipa_get_jf_ancestor_type_preserved): New function.
5685 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
5686 (ipa_get_jf_ancestor_result): Likewise.
5687 (propagate_vals_accross_pass_through): Use
5688 ipa_get_jf_pass_through_result to do all the value mappings.
5689 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
5690 type_preserved flag.
5691 (ipa_set_jf_cst_copy): New function.
5692 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
5693 (ipa_set_jf_arith_pass_through): Likewise.
5694 (ipa_set_ancestor_jf): Likewise.
5695 (compute_complex_assign_jump_func): Set type_preserved instead of
5697 (ipa_compute_jump_functions_for_edge): Likewise.
5698 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
5699 (update_jump_functions_after_inlining): Update type_preserved.
5700 Explicitely create jump functions when combining one with pass_through.
5701 (ipa_write_jump_function): Stream the type_preserved flags.
5702 (ipa_read_jump_function): Likewise.
5704 2013-08-27 Jakub Jelinek <jakub@redhat.com>
5705 Aldy Hernandez <aldyh@redhat.com>
5707 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
5708 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
5709 * function.h (struct function): Add has_force_vect_loops and
5711 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
5712 * gimple.c (gimple_build_omp_critical): Add KIND argument and
5714 * gimple.def: Update CLAUSES comments.
5715 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
5716 (gimple_build_omp_for): Add argument to prototype.
5717 (gimple_omp_for_kind): New.
5718 (gimple_omp_for_set_kind): New.
5719 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
5720 GOVD_DATA_SHARE_CLASS.
5721 (enum omp_region_type): Add ORT_SIMD.
5722 (gimple_add_tmp_var): Handle ORT_SIMD.
5723 (gimplify_var_or_parm_decl): Same.
5724 (is_gimple_stmt): Same.
5725 (omp_firstprivatize_variable): Same.
5726 (omp_add_variable): Only use splay_tree_insert if lookup failed.
5727 (omp_notice_variable): Handle ORT_SIMD.
5728 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
5729 (omp_check_private): Handle ORT_SIMD.
5730 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
5732 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
5733 Handle OMP_CLAUSE_LASTPRIVATE.
5734 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
5736 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
5737 (gimplify_expr): Handle OMP_SIMD.
5738 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
5739 (expand_GOMP_SIMD_VF): New.
5740 (expand_GOMP_SIMD_LAST_LANE): New.
5741 * internal-fn.def (GOMP_SIMD_LANE): New.
5742 (GOMP_SIMD_VF): New.
5743 (GOMP_SIMD_LAST_LANE): New.
5744 * omp-low.c: Include target.h.
5745 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
5747 (check_omp_nesting_restrictions): Same.
5749 (lower_rec_simd_input_clauses): New.
5750 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
5752 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
5753 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
5754 (expand_omp_build_assign): New.
5755 (expand_omp_for_init_counts): New.
5756 (expand_omp_for_init_vars): New.
5757 (extract_omp_for_update_vars): New.
5758 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
5759 and rewrite accordingly.
5760 (expand_omp_simd): New.
5761 (expand_omp_for): Use expand_omp_simd.
5762 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
5763 (lower_omp_for): Do not lower the body.
5764 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
5766 * tree-flow.h (find_omp_clause): Remove prototype.
5767 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
5768 forcing vectorization of the loop, or if flag_tree_vectorize.
5769 (gate_tree_if_conversion): Similarly.
5770 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
5771 gimple_build_omp_for.
5772 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
5773 * tree-parloops (create_parallel_loop): Pass kind argument to
5774 gimple_build_omp_for.
5775 * tree-pretty-print.c (dump_omp_clause): Add cases for
5776 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
5777 OMP_CLAUSE__SIMDUID_.
5778 (dump_generic_node): Handle OMP_SIMD.
5779 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
5780 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
5781 unroll OMP_SIMD loops here.
5782 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
5783 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
5785 (vect_analyze_data_refs): Handle simd loops.
5786 * tree-vect-loop.c (vectorizable_live_operation): Handle
5788 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
5789 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
5790 (vectorizable_load): Same.
5791 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
5792 (struct simduid_to_vf): New.
5793 (simduid_to_vf::hash): New.
5794 (simduid_to-vf::equal): New.
5795 (struct simd_array_to_simduid): New.
5796 (simd_array_to_simduid::hash): New.
5797 (simd_array_to_simduid::equal): New.
5798 (adjust_simduid_builtins): New.
5799 (struct note_simd_array_uses_struct): New.
5800 (note_simd_array_uses_cb): New.
5801 (note_simd_array_uses): New.
5802 (vectorize_loops): Handle simd hints and adjust simd builtins
5804 * tree-vectorizer.h (struct _stmt_vec_info): Add
5805 simd_lane_access_p field.
5806 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
5807 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
5808 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
5809 (omp_clause_code_name): Same.
5810 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
5811 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
5812 * tree.def (OMP_SIMD): New entry.
5813 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
5814 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
5815 (OMP_CLAUSE_DECL): Adjust range for new clauses.
5816 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
5817 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
5818 (OMP_CLAUSE_LINEAR_STEP): New.
5819 (OMP_CLAUSE_SAFELEN_EXPR): New.
5820 (OMP_CLAUSE__SIMDUID__DECL): New.
5821 (find_omp_clause): New prototype.
5823 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
5825 * config/i386/driver-i386.c (host_detect_local_cpu): Update
5826 Haswell processor detection.
5828 2013-08-27 Christian Widmer <shadow@umbrox.de>
5831 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
5832 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
5833 AVX2 capable processors.
5835 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
5837 * config/aarch64/arm_neon.h: Replace all inline asm implementations
5838 of vget_low_* with implementations in terms of other intrinsics.
5840 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
5843 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
5844 values to -1, 0 and 1.
5846 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
5848 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
5849 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
5850 (<optab><mode>3_insn): Remove.
5851 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
5852 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
5853 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
5854 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
5855 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
5856 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
5857 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
5858 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
5859 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
5860 (ror<mode>3_insn): Likewise.
5861 * config/aarch64/predicates.md (aarch64_simd_register): New.
5863 2013-08-27 Richard Biener <rguenther@suse.de>
5865 PR tree-optimization/57521
5866 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
5867 one edge is non-critical.
5868 (find_phi_replacement_condition): Make sure to use a non-critical
5869 edge. Cleanup and remove old bug workarounds.
5870 (bb_postdominates_preds): Remove.
5871 (if_convertible_loop_p_1): Do not compute post-dominators.
5872 (combine_blocks): Do not free post-dominators.
5873 (main_tree_if_conversion): Likewise.
5874 (pass_data_if_conversion): Add TODO_verify_ssa.
5876 2013-08-27 DJ Delorie <dj@redhat.com>
5878 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
5880 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
5882 * function.c (assign_parm_find_data_types): Set passed_mode and
5883 nominal_mode to the TYPE_MODE of nominal_type for the built
5884 pointer type in case of the struct-pass-by-reference.
5886 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
5888 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
5889 (UINT16_TYPE): Change default to "unsigned int".
5891 * config/avr/avr.opt (mfract-convert-truncate): New option.
5892 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
5893 is set, round negative fractional integers according to n1169
5894 when converting to integer types.
5896 2013-08-26 Jan Hubicka <jh@suse.cz>
5898 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
5899 methods can not be called indirectly when their address is not taken.
5901 2013-08-26 Jan Hubicka <jh@suse.cz>
5903 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
5906 2013-08-26 Jan Hubicka <jh@suse.cz>
5908 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
5911 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
5913 * reload.c (find_valid_class): Allow classes that do not include
5914 FIRST_PSEUDO_REGISTER - 1.
5916 2013-08-26 Jan Hubicka <jh@suse.cz>
5918 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
5919 fix edge count/frequency when speculation failed; fix type check
5920 for the direct call.
5922 2013-08-26 Jan Hubicka <jh@suse.cz>
5924 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
5926 2013-08-26 Jan Hubicka <jh@suse.cz>
5928 * ipa-inline-transform.c (inline_transform): Be ready for basic block
5929 to be changed by edge redirection.
5931 2013-08-26 Jan Hubicka <jh@suse.cz>
5933 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
5934 formating; add sanity check.
5935 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
5936 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
5937 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
5938 (resolve_noninline_speculation): Update callee keys, too.
5940 2013-08-26 Jan Hubicka <jh@suse.cz>
5942 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
5943 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
5945 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
5947 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
5950 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
5952 * config/i386/i386.c (ix86_debug_options): Remove prototype.
5953 (x86_64_elf_select_section): Ditto.
5954 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
5956 (ix86_pass_by_reference): Ditto.
5957 (output_set_got): Ditto.
5958 (ix86_unary_operator_ok): Ditto.
5959 (ix86_expand_builtin): Ditto.
5961 2013-08-23 Jan Hubicka <jh@suse.cz>
5963 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
5965 2013-08-23 Jan Hubicka <jh@suse.cz>
5967 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
5968 (tree_decl_with_vis): Add FINAL field.
5970 2013-08-23 Jeff Law <law@redhat.com>
5972 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
5973 the debugging dump when the expression is fully redundant.
5975 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
5977 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
5978 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
5979 * pretty-print.c (pp_formatted_text_data): Likewise.
5980 (pp_write_text_to_stream): Likewise.
5981 (pp_write_text_as_dot_label_to_stream): Likewise.
5982 (pp_append_r): Likewise.
5983 (pp_format): Likewise.
5984 (pp_flush): Likewise.
5985 (pp_clear_output_area): Likewise.
5986 (pp_append_text): Likewise.
5987 (pp_formatted_text): Likewise.
5988 (pp_remaining_character_count_for_line): Likewise.
5989 (pp_newline): Likewise.
5990 (pp_character): Likewise.
5991 (output_buffer::~output_buffer): Define.
5992 (pretty_printer::~pretty_printer): Destruct output buffer.
5993 * pretty-print.h (output_buffer::~output_buffer): Declare.
5994 (pretty_printer::~pretty_printer): Declare virtual.
5996 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
5999 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
6000 HOST_WIDE_INT_M1U): New macros.
6001 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
6002 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
6003 unsigned -1 for lshift.
6004 * cse.c (cse_insn): Likewise.
6005 * double-int.c (rshift_double, lshift_double): Likewise.
6006 * builtins.c (fold_builtin_bitop): Likewise.
6007 * combine.c (force_to_mode): Likewise.
6008 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
6009 * simplify-rtx.c (simplify_const_unary_operation,
6010 simplify_const_binary_operation): Likewise.
6011 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
6012 check_va_list_escapes): Likewise.
6013 * rtlanal.c (nonzero_bits1): Likewise.
6014 * expmed.c (expand_smod_pow2): Likewise.
6015 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
6017 2013-08-23 Jan Hubicka <jh@suse.cz>
6019 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
6020 as having address taken.
6022 2013-08-23 Jan Hubicka <jh@suse.cz>
6024 * ipa-utils.h (method_class_type): Declare.
6025 * ipa-devirt.c (method_class_type): Export.
6027 * cgraphunit.c (analyze_functions): Do basic devirtualization;
6028 do not walk base classes of anonymous types.
6030 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
6032 PR rtl-optimization/58220
6034 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
6035 handle SEQUENCE insns properly.
6037 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
6039 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
6041 (pp_newline_and_indent): Likewise.
6042 (pp_separate_with): Likewise.
6043 * pretty-print.c (pp_newline_and_flush): Define.
6044 (pp_newline_and_indent): Likewise.
6045 (pp_separate_with): Likewise.
6047 2013-08-23 Jakub Jelinek <jakub@redhat.com>
6050 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
6051 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
6053 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
6055 * config/i386/predicates.md (ext_sse_reg_operand): New.
6056 * config/i386/i386.md (*movti_internal): Use
6057 predicate to determine if EVEX is needed.
6058 (*movsi_internal): Ditto.
6059 (*movdf_internal): Ditto.
6060 (*movsf_internal): Ditto.
6061 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
6063 2013-08-23 Jakub Jelinek <jakub@redhat.com>
6065 PR tree-optimization/58209
6066 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
6067 (find_tail_calls): Give up for pointer result types if m is non-NULL.
6068 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
6069 emit POINTER_PLUS_EXPR.
6070 (create_tailcall_accumulator): For pointer result type accumulate in
6073 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
6075 * configure.ac: Add backslashes missing from the last change.
6076 * configure: Regenerate.
6078 2013-08-22 Jan Hubicka <jh@susue.cz>
6080 * ipa.c (function_and_variable_visibility): First remember function
6081 was global and then make it local.
6083 2013-08-22 Julian Brown <julian@codesourcery.com>
6085 * configure.ac: Add aarch64 to list of arches which use "nop" in
6087 * configure: Regenerate.
6089 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6091 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
6092 gnu_libc_has_function.
6093 * config/s390/tpf.h: Likewise.
6095 2013-08-22 Jan Hubicka <jh@susue.cz>
6097 * timevar.c (validate_phases): Add cast.
6099 2013-08-22 Jan Hubicka <jh@susue.cz>
6101 * timevar.c (validate_phases): Use size_t for memory.
6102 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
6104 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
6106 * pretty-print.h (output_buffer::output_buffer): Declare.
6107 (pretty_printer::pretty_printer): Likewise.
6108 (pp_construct): Remove.
6109 * pretty-print.c (output_buffer::output_buffer): Define.
6110 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
6111 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
6112 (print_gimple_expr): Likewise.
6113 (print_gimple_seq): Likewise.
6114 (gimple_dump_bb): Likewise.
6115 * sched-vis.c (dump_value_slim): Likewise.
6116 (dump_insn_slim): Likewise.
6117 (dump_rtl_slim): Likewise.
6118 (str_pattern_slim): Likewise.
6119 * tree-mudflap.c (mf_varname_tree): Likewise.
6120 * graph.c (print_graph_cfg): Likewise.
6121 (start_graph_dump): Likewise.
6122 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
6124 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
6125 pretty printer initialization.
6126 * coretypes.h (diagnostic_context): Remove superflous type alias
6128 (pretty_printer): Likewise. Declare directly as a class.
6129 (pretty_print_info): Remove declaration as class.
6130 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
6131 and pp_clear_output_area.
6132 (asan_add_global): Likewise.
6134 2013-08-22 Jan Hubicka <jh@suse.cz>
6136 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
6137 * ipa-utils.h (update_type_inheritance_graph): Declare.
6138 (possible_polymorphic_call_target_p): Declare.
6139 (possible_polymorphic_call_target_p): New.
6140 * ipa-devirt.c: Update toplevel comments.
6141 (cached_polymorphic_call_targets): Move up.
6142 (odr_type_d): Move ID down.
6143 (polymorphic_type_binfo_p): Update comment.
6144 (odr_hasher::remove): Likewise;
6145 (get_odr_type): Set anonymous_namespace.
6146 (dump_odr_type): Dump it.
6147 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
6148 (maybe_record_node): Record node in cached_polymorphic_call_targets.
6149 (record_binfo): Add comment.
6150 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
6152 (devirt_node_removal_hook): Do not iCE when cache is freed.
6153 (possible_polymorphic_call_target_p): New predicate.
6154 (update_type_inheritance_graph): New function.
6156 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
6157 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6158 Sergey Lega <sergey.s.lega@intel.com>
6159 Anna Tikhonova <anna.tikhonova@intel.com>
6160 Ilya Tocar <ilya.tocar@intel.com>
6161 Andrey Turetskiy <andrey.turetskiy@intel.com>
6162 Ilya Verbin <ilya.verbin@intel.com>
6163 Kirill Yukhin <kirill.yukhin@intel.com>
6164 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6166 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
6167 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
6168 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
6169 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
6170 (OPTION_MASK_ISA_AVX2_UNSET): Update.
6171 (OPTION_MASK_ISA_AVX512F_UNSET): New.
6172 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
6173 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
6174 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
6175 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
6176 OPT_mavx512pf, OPT_mavx512er cases.
6177 * config/i386/constraints.md (v): New constraint.
6178 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
6179 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
6180 (bit_AVX512CD): New.
6181 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
6182 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
6183 * config/i386/i386-c.c (ix86_target_macros_internal):
6184 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
6186 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
6187 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
6188 * config/i386/i386.c (regclass_map, dbx_register_map)
6189 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
6190 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
6191 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
6193 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
6194 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
6195 -mavx512pf options. Fix formatting.
6196 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
6197 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
6198 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
6199 -mavx512cd, -mavx512pf options.
6200 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
6201 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
6202 (ix86_preferred_output_reload_class): Replace SSE_REGS with
6204 (ix86_hard_regno_mode_ok): Support 512-bit registers.
6205 (ix86_set_reg_reg_cost): Ditto.
6206 (x86_order_regs_for_local_alloc): Ditto.
6207 (MAX_VECT_LEN): Extend to 64-byte.
6208 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
6209 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
6210 (TARGET_AVX512ER, TARGET_AVX512CD): New.
6211 (BIGGEST_ALIGNMENT): Extend to 512-bits.
6212 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
6213 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
6214 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
6215 (SSE_REG_MODE_P): Support new modes.
6216 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
6217 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
6218 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
6219 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
6220 (REG_CLASS_CONTENTS): Add new registers.
6221 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
6222 (EXT_REX_SSE_REGNO_P): New.
6223 (HI_REGISTER_NAMES): Add new registers.
6224 * config/i386/i386.md: Define constants for new registers.
6225 (mode): Add new 512-bit modes.
6226 (prefix): Support evex prefix.
6227 (isa): Support avx512f, noavx512f, fma_avx512f.
6228 (ssemodesuffix): Add new 512-bit modes.
6230 (*movxi_internal_avx512f): Ditto.
6231 (*movdi_internal): Replace constraint "x" with the new constraint "v".
6233 (*movsi_internal): Likewise.
6234 (*movdf_internal): Likewise.
6235 (*movsf_internal): Likewise.
6236 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
6237 (<code><mode>3): Likewise.
6238 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
6240 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
6241 with the new constraint "v".
6242 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
6244 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
6245 with the new constraint "v".
6246 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
6247 (<sse2>_storedqu<avxsizesuffix>): Likewise.
6248 (*<plusminus_insn><mode>3): Likewise.
6249 (<sse>_vm<plusminus_insn><mode>3): Likewise.
6250 (*mul<mode>3): Likewise.
6251 (<sse>_vmmul<mode>3): Likewise.
6252 (<sse>_div<mode>3): Likewise.
6253 (<sse>_vmdiv<mode>3): Likewise.
6254 (<sse>_sqrt<mode>2): Likewise.
6255 (<sse>_vmsqrt<mode>2): Likewise.
6256 (*<code><mode>3_finite): Likewise.
6257 (*<code><mode>3) <smaxmin>: Likewise.
6258 (<sse>_vm<code><mode>3): Likewise.
6259 (*<code><mode>3) <any_logic>: Likewise.
6260 (*fma_fmadd_<mode>): Likewise.
6261 (*fma_fmsub_<mode>): Likewise.
6262 (*fma_fnmadd_<mode>): Likewise.
6263 (*fma_fnmsub_<mode>): Likewise.
6264 (*fma_fmaddsub_<mode>): Likewise.
6265 (*fma_fmsubadd_<mode>): Likewise.
6266 (*fmai_fmadd_<mode>): Likewise.
6267 (*fmai_fmsub_<mode>): Likewise.
6268 (*fmai_fnmadd_<mode>): Likewise.
6269 (*fmai_fnmsub_<mode>): Likewise.
6270 (sse_cvtsi2ss): Likewise.
6271 (sse_cvtsi2ssq): Likewise.
6272 (sse_cvtss2si): Likewise.
6273 (sse_cvtss2si_2): Likewise.
6274 (sse_cvtss2siq): Likewise.
6275 (sse_cvtss2siq_2): Likewise.
6276 (sse_cvttss2si): Likewise.
6277 (sse_cvtss2siq_2): Likewise.
6278 (float<sseintvecmodelower><mode>2): Likewise.
6279 (sse2_cvtsd2si_2): Likewise.
6280 (sse2_cvtsd2siq_2): Likewise.
6281 (*<plusminus_insn><mode>3): Likewise.
6282 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
6283 (*<sse4_1_avx2>_mul<mode>3): Likewise.
6284 (ashr<mode>3): Likewise.
6285 (<shift_insn><mode>3): Likewise.
6286 (avx2_<code><mode>3): Likewise.
6287 (*avx2_<code><mode>3): Likewise.
6288 (*andnot<mode>3): Likewise.
6289 (*<code><mode>3) <any_logic>: Likewise.
6290 (abs<mode>2): Likewise.
6291 (avx2_permvar<mode>): Likewise.
6292 (avx2_perm<mode>_1): Likewise.
6293 (*avx_vpermilp<mode>): Likewise.
6294 (avx_vpermilvar<mode>3): Likewise.
6295 (avx2_ashrv<mode>): Likewise.
6296 (avx2_<shift_insn>v<mode>): Likewise.
6297 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
6299 * doc/rtl.texi: Document XImode.
6301 2013-08-21 Jeff Law <law@redhat.com>
6303 * tree-flow.h (register_jump_thread): Pass vector of edges
6304 instead of each important edge.
6305 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
6306 thread path into a vector and pass that to register_jump_thread.
6307 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
6308 passed in edge vector to the current 3-edge form.
6311 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
6313 * dce.c (fini_dce): Call df_analyze again just in case
6314 delete_unmarked_insns removed anything.
6316 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
6318 * reload.h (struct reg_equivs): Rename to ..
6319 (struct reg_equivs_s): .. this.
6321 2013-08-20 Martin Liska <marxin.liska@gmail.com>
6323 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
6325 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6327 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
6329 2013-08-21 Jeff Law <law@redhat.com>
6331 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
6332 simplify assignments too. If the RHS collapses to a singleton
6333 range, then return the value for the range.
6335 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
6337 * config/i386/sse.md (V16): Rename to...
6339 (mov<mode>): Update iterator name.
6340 (*mov<mode>_internal): Ditto.
6341 (push<mode>1): Ditto.
6342 (movmisalign<mode>): Ditto.
6344 2013-08-20 Jan Hubicka <jh@suse.cz>
6347 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
6348 entry for direct edges.
6349 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
6351 2013-08-20 David Malcolm <dmalcolm@redhat.com>
6353 Revert my last two changes, r201865 and r201864:
6356 2013-08-20 David Malcolm <dmalcolm@redhat.com>
6358 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
6359 instances can own GC refs.
6361 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
6362 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
6363 (gcc::context::gt_pch_nx): Likewise.
6364 (gcc::context::gt_pch_nx): Likewise.
6365 * ggc.h (gt_ggc_mx <T>): New.
6366 (gt_pch_nx_with_op <T>): New.
6367 (gt_pch_nx <T>): New.
6368 * passes.c (opt_pass::gt_ggc_mx): New.
6369 (opt_pass::gt_pch_nx): New.
6370 (opt_pass::gt_pch_nx_with_op): New.
6371 (pass_manager::gt_ggc_mx): New.
6372 (pass_manager::gt_pch_nx): New.
6373 (pass_manager::gt_pch_nx_with_op): New.
6374 (pass_manager::operator new): Use
6375 ggc_internal_cleared_alloc_stat rather than xcalloc.
6376 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
6377 (pass_manager::gt_ggc_mx): New.
6378 (pass_manager::gt_pch_nx): New.
6379 (pass_manager::gt_pch_nx_with_op): New.
6380 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
6381 (opt_pass::operator new): New.
6382 (opt_pass::gt_ggc_mx): New.
6383 (opt_pass::gt_pch_nx): New.
6384 (opt_pass::gt_pch_nx_with_op): New.
6387 2013-08-20 David Malcolm <dmalcolm@redhat.com>
6389 * Makefile.in (GTFILES): Add context.h.
6390 * context.c (gcc::context::operator new): New.
6391 (gcc::context::gt_ggc_mx): New.
6392 (gcc::context::gt_pch_nx): New.
6393 (gcc::context::gt_pch_nx): New.
6394 * context.h (gcc::context): Add GTY((user)) marking.
6395 (gcc::context::operator new): New.
6396 (gcc::context::gt_ggc_mx): New.
6397 (gcc::context::gt_pch_nx): New.
6398 (gcc::context::gt_pch_nx): New.
6399 (g): Add GTY marking.
6400 (gt_ggc_mx (gcc::context *)): New.
6401 (gt_pch_nx (gcc::context *)): New.
6402 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
6403 void *cookie)): New.
6404 * gengtype.c (open_base_files) <ifiles>: Add context.h.
6406 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
6408 * dce.c (fini_dce): Call df_analyze again just in case
6409 delete_unmarked_insns removed anything.
6411 2013-08-20 Teresa Johnson <tejohnson@google.com>
6413 PR rtl-optimizations/57451
6414 * final.c (reemit_insn_block_notes): Prevent lexical blocks
6415 from crossing split section boundaries.
6417 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
6419 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
6420 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
6421 with MULTLIB_DEFAULTS.
6423 2013-08-20 Nick Clifton <nickc@redhat.com>
6425 * target.def (narrow_volatile_bitfield): Note that the default
6426 value is false, not !TARGET_STRICT_ALIGN.
6427 * doc/tm.texi: Regenerate.
6429 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
6431 Fix LIB_SPEC for systems without libpthread.
6433 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
6434 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
6436 * config/i386/linux-common.h: Likewise.
6437 * config/mips/linux-common.h: Likewise.
6439 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
6441 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
6444 2013-08-20 David Malcolm <dmalcolm@redhat.com>
6446 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
6447 instances can own GC refs.
6449 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
6450 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
6451 (gcc::context::gt_pch_nx): Likewise.
6452 (gcc::context::gt_pch_nx): Likewise.
6453 * ggc.h (gt_ggc_mx <T>): New.
6454 (gt_pch_nx_with_op <T>): New.
6455 (gt_pch_nx <T>): New.
6456 * passes.c (opt_pass::gt_ggc_mx): New.
6457 (opt_pass::gt_pch_nx): New.
6458 (opt_pass::gt_pch_nx_with_op): New.
6459 (pass_manager::gt_ggc_mx): New.
6460 (pass_manager::gt_pch_nx): New.
6461 (pass_manager::gt_pch_nx_with_op): New.
6462 (pass_manager::operator new): Use
6463 ggc_internal_cleared_alloc_stat rather than xcalloc.
6464 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
6465 (pass_manager::gt_ggc_mx): New.
6466 (pass_manager::gt_pch_nx): New.
6467 (pass_manager::gt_pch_nx_with_op): New.
6468 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
6469 (opt_pass::operator new): New.
6470 (opt_pass::gt_ggc_mx): New.
6471 (opt_pass::gt_pch_nx): New.
6472 (opt_pass::gt_pch_nx_with_op): New.
6474 2013-08-20 David Malcolm <dmalcolm@redhat.com>
6476 * Makefile.in (GTFILES): Add context.h.
6477 * context.c (gcc::context::operator new): New.
6478 (gcc::context::gt_ggc_mx): New.
6479 (gcc::context::gt_pch_nx): New.
6480 (gcc::context::gt_pch_nx): New.
6481 * context.h (gcc::context): Add GTY((user)) marking.
6482 (gcc::context::operator new): New.
6483 (gcc::context::gt_ggc_mx): New.
6484 (gcc::context::gt_pch_nx): New.
6485 (gcc::context::gt_pch_nx): New.
6486 (g): Add GTY marking.
6487 (gt_ggc_mx (gcc::context *)): New.
6488 (gt_pch_nx (gcc::context *)): New.
6489 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
6490 void *cookie)): New.
6491 * gengtype.c (open_base_files) <ifiles>: Add context.h.
6493 2013-08-20 Alan Modra <amodra@gmail.com>
6496 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
6497 (rs6000_emit_epilogue): Likewise.
6499 2013-08-19 Dehao Chen <dehao@google.com>
6501 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
6503 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
6504 Jakub Jelinek <jakub@redhat.com>
6506 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
6507 (BUILT_IN_FABSD64): Likewise.
6508 (BUILT_IN_FABSD128): Likewise.
6509 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
6510 (fold_builtin_1): Likewise.
6511 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
6512 destination and source operands.
6513 (*abstd2_fpr): Likewise.
6514 (*nabstd2_fpr): Likewise.
6516 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
6518 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
6521 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
6523 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
6526 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
6528 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
6529 * builtins.c (default_libc_has_function): New.
6530 (gnu_libc_has_function): Ditto.
6531 (no_c99_libc_has_function): Ditto.
6532 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
6533 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
6534 (fold_builtin_sincos): Likewise.
6535 (fold_builtin_cexp): Likewise.
6536 * builtins.def (DEF_C94_BUILTIN): Likewise.
6537 (DEF_C99_BUILTIN): Likewise.
6538 (DEF_C99_C90RES_BUILTIN): Likewise.
6539 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
6540 definitions to using this define.
6541 * config/darwin-protos.h (darwin_libc_has_function): New.
6542 * config/darwin.c (darwin_libc_has_function): Ditto.
6543 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
6544 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
6545 * config/darwin.h: Ditto.
6546 * config/elfos.h: Ditto.
6547 * config/freebsd.h: Ditto.
6548 * config/i386/cygming.h: Ditto.
6549 * config/i386/djgpp.h: Ditto.
6550 * config/i386/i386-interix.h: Ditto.
6551 * config/microblaze/microblaze.h: Ditto.
6552 * config/mmix/mmix.h: Ditto.
6553 * config/gnu-user.h: Ditto.
6554 * config/ia64/hpux.h: Ditto.
6555 * config/pa/pa-hpux.h: Ditto.
6556 * config/pdp11/pdp11.h: Ditto.
6557 * config/picochip/picochip.h: Ditto.
6558 * config/linux.h: Ditto.
6559 * config/netbsd.h: Ditto.
6560 * config/openbsd.h: Ditto.
6561 * config/rs6000/aix43.h: Ditto.
6562 * config/rs6000/aix51.h: Ditto.
6563 * config/rs6000/aix52.h: Ditto.
6564 * config/rs6000/aix53.h: Ditto.
6565 * config/rs6000/aix61.h: Ditto.
6566 * config/rs6000/darwin.h: Ditto.
6567 * config/rs6000/linux.h: Ditto.
6568 * config/rs6000/linux64.h: Ditto.
6569 * config/s390/tpf.h: Ditto.
6570 * config/sol2-10.h: Ditto.
6571 * config/sol2.h: Ditto.
6572 * config/vms/vms.h: Ditto.
6573 * config/vxworks.h: Ditto.
6574 * config/linux-android.c (linux_android_libc_has_function):
6575 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
6576 * config/linux-protos.h (linux_android_libc_has_function):
6578 * config/i386/i386.c (ix86_libc_has_function): New.
6579 * config/i386/i386-protos.h
6580 (ix86_libc_has_function): New declaration.
6581 * config/i386/i386.md
6582 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
6583 ("isinf<mode>2): Likewise.
6584 * convert.c (convert_to_integer): Using new target hook
6585 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
6586 TARGET_C99_FUNCTIONS.
6587 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
6588 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
6589 * coretypes.h (function_class): New enum for different
6590 classes of functions.
6591 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
6592 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
6593 (TARGET_HAS_SINCOS): Likewise.
6594 (TARGET_LIBC_HAS_FUNCTION): New.
6595 * doc/tm.texi: Regenerated.
6596 * targhooks.h (default_libc_has_function): New declaration.
6597 (no_c99_libc_has_function): Ditto.
6598 (gnu_libc_has_function): Ditto.
6599 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
6600 and TARGET_HAS_SINCOS.
6602 2013-08-18 Jan Hubicka <jh@suse.cz>
6604 * Makeifle-in (ipa-devirt.o): New.
6605 (GTFILES): Add ipa-utils.h and ipa-devirt.c
6606 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
6607 (analyze_functions): Look into possible targets of polymorphic call.
6608 * dumpfile.c (dump_files): Add type-inheritance dump.
6609 * dumpfile.h (TDI_inheritance): New.
6610 * ipa-devirt.c: New file.
6611 * ipa-utils.h (odr_type_d): Forward declare.
6612 (odr_type): New type.
6613 (build_type_inheritance_graph): Declare.
6614 (possible_polymorphic_call_targets): Declare and introduce inline
6615 variant when only edge is pased.
6616 (dump_possible_polymorphic_call_targets): Likewise.
6617 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
6618 * tree.c (type_in_anonymous_namespace_p): Break out from ...
6619 (types_same_for_odr): ... here.
6620 * tree.h (type_in_anonymous_namespace_p): Declare.
6622 2013-08-18 Jakub Jelinek <jakub@redhat.com>
6624 PR tree-optimization/58006
6625 * tree-parloops.c (take_address_of): Don't ICE if get_name
6627 (eliminate_local_variables_stmt): Remove clobber stmts.
6629 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
6631 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
6632 error message is issued for an alias to undefined symbol.
6634 2013-08-18 Jan Hubicka <jh@suse.cz>
6636 * cgraph.c (cgraph_create_indirect_edge): Discover
6637 polymorphic calls and record basic info into indirect_info.
6638 * gimple-fold.c (gimple_fold_call): When doing BINFO based
6639 devirtualization, ignore objc function calls.
6640 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
6641 call with no parm index info.
6642 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
6643 * tree.c (virtual_method_call_p): New function.
6644 * tree.h (virtual_method_call_p): Declare.
6646 2013-08-16 Jan Hubicka <jh@suse.cz>
6649 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
6651 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
6653 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
6656 2013-08-16 David Malcolm <dmalcolm@redhat.com>
6658 * gengtype.c (type_for_name): Add special-case support for
6659 locating types within the "gcc::" namespace.
6660 (open_base_files): Emit a "using namespace gcc" directive.
6662 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
6665 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
6666 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
6668 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
6669 array instead of each individual operand as a separate argument.
6670 (emit_fusion_gpr_load): Likewise.
6671 (expand_fusion_gpr_load): Add new function declaration.
6673 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
6674 signature to have the operands passed as an array, instead of as
6675 separate arguments. Allow ZERO_EXTEND to be in the memory
6676 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
6677 depend on the register live/dead flags when peepholes are run.
6678 (expand_fusion_gpr_load): New function to be called from the
6679 peephole2 pass, to change the register that addis sets to be the
6681 (emit_fusion_gpr_load): Change the calling signature to have the
6682 operands passed as an array, instead of as separate arguments.
6683 Allow ZERO_EXTEND to be in the memory address, and also
6684 SIGN_EXTEND if -mpower8-fusion-sign.
6686 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
6688 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
6689 adjust the register addis loads up in the peephole2 pass. Do not
6690 depend on the register live/dead state when the peephole pass is done.
6692 2013-08-16 David Malcolm <dmalcolm@redhat.com>
6694 * gengtype.c (create_user_defined_type): Ensure that the kind
6695 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
6696 declaration is seen before the GTY((user)) marking.
6698 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
6701 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
6703 2013-08-16 Jan Hubicka <jh@suse.cz>
6705 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
6706 arugment expected_type.
6707 (gimple_fold_call): Use it.
6708 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
6709 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
6710 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
6711 (try_make_edge_direct_virtual_call): Likewise.
6712 * tree.c (obj_type_ref_class): New.
6713 * tree.h (obj_type_ref_class): Use it.
6715 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
6717 * sched-vis.c (rtl_slim_pp_initialized): Remove.
6718 (rtl_slim_pp): Likewise.
6719 (init_rtl_slim_pretty_print): Likewise.
6720 (dump_value_slim): Don't call it. Use local pretty printer.
6721 (dump_insn_slim): Likewise.
6722 (dump_rtl_slim): Likewise.
6723 (str_pattern_slim): Likewise.
6724 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
6727 2013-08-16 Jakub Jelinek <jakub@redhat.com>
6729 PR tree-optimization/58164
6730 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
6731 walk gimple_goto_dest of GIMPLE_GOTO.
6733 PR tree-optimization/58165
6734 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
6735 bi_call must be the last stmt in a bb, don't split_block, instead
6736 use fallthru edge from it and give up if there is none.
6737 Release conds vector when returning early.
6739 2013-08-14 Xinliang David Li <davidxl@google.com>
6741 * config/i386/i386.c (ix86_option_override_internal):
6742 Remove unused variable and field.
6744 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6747 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
6748 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
6749 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
6750 and Linux, correct BLKmode alignment when 128-bit alignment is
6751 required and compatibility flag is not set.
6752 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
6753 for zero-size arguments when compatibility flag is not set.
6755 2013-08-14 Jakub Jelinek <jakub@redhat.com>
6757 PR tree-optimization/58145
6758 * tree-sra.c (build_ref_for_offset): If prev_base has
6759 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
6761 2013-08-14 Xinliang David Li <davidxl@google.com>
6763 * config/i386/i386.c (ix86_option_override_internal):
6764 Fix uninitialized variable error.
6766 2013-08-14 Xinliang David Li <davidxl@google.com>
6768 * config/i386/i386.opt: Define two new options.
6769 * config/i386/x86-tune.def: Add arch selector field in macros.
6770 * config/i386/i386.h: Adjust macro definition.
6771 * config/i386/i386.c (ix86_option_override_internal):
6773 (parse_mtune_ctrl_str): New function.
6774 (set_ix86_tune_features): New function.
6775 (ix86_function_specific_restore): Call the new helper function.
6777 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
6779 PR rtl-optimization/57662
6780 * sel-sched.c (code_motion_process_successors): When the current insn
6781 is removed after the recursive traversal, break from the loop.
6782 Add comments and debug printouts.
6784 2013-08-14 Jakub Jelinek <jakub@redhat.com>
6785 Alexandre Oliva <aoliva@redhat.com>
6788 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
6789 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
6790 there also UNSPEC_PLTOFF.
6792 2013-08-14 Marek Polacek <polacek@redhat.com>
6794 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
6795 NUM_CONDITIONS bit positions.
6797 2013-08-13 Cary Coutant <ccoutant@google.com>
6799 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
6800 (attr_checksum): Hash vector contents instead of pointer.
6801 (attr_checksum_ordered): Likewise.
6803 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
6805 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
6806 when Pmode != word_mode. Add length_address attribute.
6807 (sse3_monitor_<mode>): Merge from sse3_monitor and
6808 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
6809 Pmode != word_mode. Update insn length attribute.
6810 * config/i386/i386.c (ix86_option_override_internal): Update
6811 ix86_gen_monitor selection for merged sse3_monitor insn.
6813 2013-08-13 Julian Brown <julian@codesourcery.com>
6815 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
6816 perform invalid legitimization on greater-than-word-size modes for
6819 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
6821 * ira.c (setup_class_translate_array): Use aclass instead of cl
6822 for classes not fully covered by allocno classes.
6824 2013-08-13 Jakub Jelinek <jakub@redhat.com>
6826 PR tree-optimization/57661
6827 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
6828 * tree-inline.c (tree_function_versioning): Initialize it.
6829 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
6830 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
6831 that is not being copied.
6834 * asan.c (instrument_strlen_call): Fix typo in comment.
6835 Use char * type even for the lhs of POINTER_PLUS_EXPR.
6837 2013-08-13 Steve Ellcey <sellcey@mips.com>
6839 * config/mips/mips.md (prefetch): Use lw instead of ld on
6840 loongson in 32bit mode.
6842 2013-08-13 Nick Clifton <nickc@redhat.com>
6844 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
6846 2013-08-13 Jan Hubicka <jh@suse.cz>
6848 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
6849 introduced edge; fix typo in sanity check.
6850 (cgraph_resolve_speculation): Export; improve diagnostic.
6851 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
6852 speculation at type mismatch.
6853 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
6854 (cgraph_resolve_speculation): Declare.
6855 (symtab_can_be_discarded): New function.
6856 * value-prof.c (gimple_ic_transform): Remove actual transform code.
6857 * ipa-inline-transform.c (speculation_removed): New global var.
6858 (clone_inlined_nodes): See if speculation can be removed.
6859 (inline_call): If speculations was removed, we growths may not match.
6860 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
6861 (speculation_useful_p): New function.
6862 (resolve_noninline_speculation): New function.
6863 (inline_small_functions): Resolve useless speculations.
6864 * ipa-inline.h (speculation_useful_p): Declare
6865 * ipa.c (can_replace_by_local_alias): Simplify.
6866 (ipa_profile): Produce speculative calls in non-lto, too;
6867 add simple cost model; produce local aliases.
6869 2013-08-13 David Malcolm <dmalcolm@redhat.com>
6871 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
6874 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
6876 * config/i386/i386.c (ix86_function_versions): Use error + inform.
6878 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
6880 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
6881 iterator instead of X87MODEF.
6883 2013-08-12 Perez Read <netfirewall@gmail.com>
6886 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
6887 operand 0 for intel asm alternative.
6888 (*movabs<mode>_2): Ditto for operand 1.
6890 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
6892 * config/aarch64/arm_none.h
6893 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
6895 2013-08-12 Nick Clifton <nickc@redhat.com>
6897 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
6899 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
6901 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
6902 expand for QI/HImode operand to produce more effictive code for
6903 unsigned char(short) --> float(double) conversion.
6905 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
6907 * doc/invoke.texi: Mention that -ftls-model does not force the final
6910 2013-08-12 Marek Polacek <polacek@redhat.com>
6911 Marc Glisse <marc.glisse@inria.fr>
6913 PR tree-optimization/57980
6914 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
6915 when creating -1 constant.
6917 2013-08-10 Jan Hubicka <jh@suse.cz>
6919 Workaround binutils PR14342.
6920 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
6921 (gimple_init_edge_profiler): Likewise.
6922 (gimple_gen_ic_func_profiler): Likewise.
6924 2013-08-09 Jan Hubicka <jh@suse.cz>
6926 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
6928 2013-08-09 Xinliang David Li <davidxl@google.com>
6930 * config/i386/stringop.def: New file.
6931 * config/i386/stringop.opt: New file.
6932 * config/i386/i386-opts.h: Include stringopt.def.
6933 * config/i386/i386.opt: Include stringopt.opt.
6934 * config/i386/i386.c (ix86_option_override_internal):
6935 Override default size based stringop inline strategies with options.
6936 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
6939 2013-08-09 Jan Hubicka <jh@suse.cz>
6941 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
6943 2013-08-09 Jan Hubicka <jh@suse.cz>
6945 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
6947 (dump_cgraph_node): Dump profile-id.
6948 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
6949 and common_target_probability.
6950 * lto-cgraph.c (lto_output_edge): Stream common targets.
6951 (lto_output_node): Stream profile ids.
6952 (input_node): Stream profile ids.
6953 (input_edge): Stream common targets.
6954 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
6955 * ipa.c: Include value-prof.h
6956 (ipa_profile_generate_summary): Turn indirect call statement histograms
6957 into common targets.
6958 (ipa_profile): Turn common targets into speculative edges.
6960 2013-08-09 Jan Hubicka <jh@suse.cz>
6962 * cgraph.h (cgraph_node): Add profile_id.
6963 * value-prof.c (cgraph_node_map): Turn into pointer_map.
6964 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
6965 (del_node_map): Update.
6966 (find_func_by_funcdef_no): Replace by ...
6967 (find_func_by_profile_id): ... this one.
6968 (gimple_ic_transform): Do not remove useful histograms when
6969 speculation is not done; dump info when indirect call removal
6971 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
6972 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
6973 (__gcov_indirect_call_profiler_v2): .. this one.
6974 * profile.h (init_node_map): Update.
6975 * coverage.c (coverage_compute_profile_id): New function.
6976 * coverage.h (coverage_compute_profile_id): Declare.
6977 * tree-profile.c (init_ic_make_global_vars): Make
6978 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
6979 (gimple_init_edge_profiler): Update prototype of
6980 __gcov_indirect_call_profiler.
6981 (gimple_gen_ic_func_profiler): Simplify.
6982 (tree_profiling): Use init_node_map
6984 2013-08-09 Jan Hubicka <jh@suse.cz>
6986 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
6987 non-speculative refs.
6988 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
6989 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
6990 (cgraph_set_call_stmt): Likewise.
6991 (cgraph_create_edge_1): Fix release checking compilatoin;
6993 (cgraph_free_edge): Free indirect info.
6994 (cgraph_turn_edge_to_speculative): New function.
6995 (cgraph_speculative_call_info): New function.
6996 (cgraph_make_edge_direct): Return direct edge; handle speculation.
6997 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
6998 (dump_cgraph_node): Dump speculation.
6999 (verify_edge_count_and_frequency): Accept speculative edges.
7000 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
7001 (verify_cgraph_node): Handle speculation.
7002 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
7003 (cgraph_set_call_stmt): Update prototype.
7004 (cgraph_make_edge_direct): Update prototype.
7005 (cgraph_speculative_call_info): Declare.
7006 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
7007 to change; update call of ipa_find_references.
7008 * ipa-ref.c (ipa_record_reference): Fix return value; clear
7009 lto_stmt_uid and speculative flags.
7010 (ipa_dump_references): Dump speculation.
7011 (ipa_clone_references): Clone speculative flag.
7012 (ipa_clone_referring): Likewise.
7013 (ipa_clone_ref): New function.
7014 (ipa_find_reference): Look into lto_stmt_uids
7015 (ipa_clear_stmts_in_references): Do not clear speculative calls.
7016 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
7017 (ipa_find_reference): Update declaration.
7018 (ipa_clone_ref): Declare.
7019 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
7020 stream speculative flag.
7021 (lto_output_ref): Stream statements uids and speculation.
7022 (input_ref): Likewise.
7023 (input_edge): Stream speuclation.
7024 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
7025 (cgraph_set_call_stmt_including_clones): Handle speculation.
7026 * ipa-inline.c (heap_edge_removal_hook): New function.
7027 (inline_small_functions): Register it.
7028 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
7029 also initialize refs.
7030 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
7032 (try_make_edge_direct_simple_call): Likewise.
7033 (try_make_edge_direct_simple_call): Likewise.
7034 (update_indirect_edges_after_inlining): Likewise.
7035 (remove_described_reference): Look proper lto_stmt_uid.
7036 (propagate_controlled_uses): Likewise.
7037 (propagate_controlled_uses): Liekwise.
7038 * tree-inline.c (copy_bb): Copy speculative edges.
7039 (redirect_all_calls): New function.
7040 (copy_cfg_body): Do redirection after loop info is updated.
7041 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
7043 2013-08-09 Jan Hubicka <jh@suse.cz>
7045 * lto-streamer-out.c (output_function): Renumber PHIs.
7046 * lto-streamer-in.c (input_function): Likewise.
7048 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
7050 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
7051 (get_lane_unsigned): Likewise.
7052 (dup_lane_scalar): Likewise.
7053 (get_lane): enable for VALL.
7054 * config/aarch64/aarch64-simd.md
7055 (aarch64_dup_lane_scalar<mode>): Remove.
7056 (aarch64_get_lane_signed<mode>): Likewise.
7057 (aarch64_get_lane_unsigned<mode>): Likewise.
7058 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
7059 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
7060 (aarch64_get_lane<mode>): Enable for all vector modes.
7061 (aarch64_get_lanedi): Remove misleading constraints.
7062 * config/aarch64/arm_neon.h
7063 (__aarch64_vget_lane_any): Define.
7064 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
7065 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
7066 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
7067 * config/aarch64/iterators.md (VDQQH): New.
7071 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
7073 * configure.ac: Add GAS check for LEON instructions on SPARC.
7074 * configure: Regenerate.
7075 * config.in: Likewise.
7076 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
7078 * config/sparc/sparc.opt (LEON, LEON3): New masks.
7079 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
7081 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
7082 (AS_LEON_FLAG): New macro.
7083 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
7084 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
7085 Deal with LEON and LEON3 for the memory model.
7086 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
7087 (atomic_compare_and_swap<mode>_1): Likewise.
7088 (*atomic_compare_and_swap<mode>_1): Likewise.
7090 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
7092 * config/arm/neon.md (vcond): Fix floating-point vector
7093 comparisons against 0.
7095 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
7097 * lra-constraints.c (emit_spill_move): Remove assert.
7098 (process_alt_operands): Add more debugging
7099 output. Increase reject for spilling into memory. Decrease
7100 reject for reloading scratch.
7101 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
7103 2013-08-08 Steve Ellcey <sellcey@mips.com>
7105 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
7106 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
7107 micromips incompatible. Add nan2008.
7108 (MULTILIB_DIRNAMES): Add nan2008.
7109 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
7110 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
7111 and micromips incompatible. Add nan2008.
7112 (MULTILIB_DIRNAMES): Add nan2008.
7113 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
7115 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
7117 PR rtl-optimization/58079
7118 * combine.c (combine_simplify_rtx): Avoid using SUBST if
7119 simplify_comparison has widened a comparison with an integer.
7121 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7123 * config/arm/neon.md (movmisalign<mode>): Disable when we
7124 don't allow unaligned accesses.
7125 (*movmisalign<mode>_neon_store): Likewise.
7126 (*movmisalign<mode>_neon_load): Likewise.
7127 (*movmisalign<mode>_neon_store): Likewise.
7128 (*movmisalign<mode>_neon_load): Likewise.
7130 2013-08-08 Jan Hubicka <jh@suse.cz>
7132 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
7133 (make_pass_rebuild_cgraph_edges): Also clear references.
7134 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
7135 * ipa-inline-transform.c (inline_transform): Remove all references
7137 * cgraphunit.c (expand_function): Remove all references after
7139 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
7140 (ipa_find_reference): Rewrite to iterator.
7141 (remove_stmt_references): Likewise.
7142 (ipa_clear_stmts_in_references): New function.
7143 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
7144 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
7146 * ipa-split.c (split_function): Remove references in split function.
7148 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
7151 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
7152 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
7154 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
7157 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
7158 suggested mode for the assignment isn't compatible with the
7161 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
7164 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
7166 2013-08-07 Xinliang David Li <davidxl@google.com>
7168 * config/i386/i386.opt: New option -mtune-ctrl=.
7169 * config/i386/x86-tune.def: New file.
7170 * config/i386/i386.h: include x86-tune.def.
7171 * config/i386/i386.c (ix86_option_override_internal):
7172 Parsing -mtune-ctrl= option and set tune features.
7174 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
7177 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
7178 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
7181 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
7183 * rtl.h (update_alignments): Declare.
7184 * final.c (grow_label_align): New function extracted from...
7185 (shorten_branches): ...here. Call it.
7186 (update_alignments): New function.
7187 * reorg.c (sibling_labels): New variable.
7188 (get_label_before): Add SIBLING parameter. If it is non-zero, push
7189 the new label along with it onto the sibling_labels vector.
7190 (fill_simple_delay_slots): Adjust call to get_label_before.
7191 (fill_slots_from_thread): Likewise.
7192 (relax_delay_slots): Likewise.
7193 (make_return_insns): Likewise.
7194 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
7196 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
7198 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
7199 document its semantics.
7200 (diagnostic_report_diagnostic): Adjust accordingly.
7202 2013-08-07 David Malcolm <dmalcolm@redhat.com>
7204 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
7205 (sparc_option_override): ...and port to new C++ pass API.
7206 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
7208 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
7210 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
7212 2013-08-06 Caroline Tice <cmtice@google.com>
7214 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
7215 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
7216 * tree-pass.h: Add pass_vtable_verify.
7217 * varasm.c (assemble_variable): Add code to properly set the comdat
7218 section and name for the .vtable_map_vars section.
7219 (assemble_vtyv_preinit_initializer): New function.
7220 (default_sectin_type_flags): Make sure .vtable_map_vars section has
7222 * output.h: Add function decl for assemble_vtv_preinit_initializer.
7223 * vtable-verify.c: New file.
7224 * vtable-verify.h: New file.
7225 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
7226 initialiation levels.
7227 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
7228 * passes.def: Insert pass_vtable_verify.
7229 * aclocal.m4: Reorder includes.
7230 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
7231 -fvtv-counts options.
7232 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
7233 as appropriate, if -fvtable-verify=... is used.
7234 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
7235 -fvtable-verify=... is used.
7236 * Makefile.in (OBJS): Add vtable-verify.o to list.
7237 (vtable-verify.o): Add new build rule.
7238 (GTFILES): Add vtable-verify.c to list.
7239 * common.opt (fvtable-verify=): New flag.
7240 (vtv_priority): Values for fvtable-verify= flag.
7241 (fvtv-counts): New flag.
7242 (fvtv-debug): New flag.
7243 * tree.h (save_vtable_map_decl): New extern function decl.
7245 2013-08-07 David Malcolm <dmalcolm@redhat.com>
7247 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
7248 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
7249 (pass_data_rl78_devirt): ...new pass_data instance and...
7250 (make_pass_rl78_devirt): ...new function.
7251 (rl78_asm_file_start): Port pass registration to new C++ API.
7253 2013-08-07 David Malcolm <dmalcolm@redhat.com>
7255 * coretypes.h (rtl_opt_pass): Add.
7256 (gcc::context): Add.
7257 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
7258 (epiphany_init): Port to new C++ pass API.
7259 (epiphany_optimize_mode_switching): Likewise.
7260 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
7261 (pass_manager::get_pass_mode_switching): New.
7262 (pass_manager::get_pass_peephole2): New.
7263 * mode-switching.c (pass_mode_switching): Add clone method.
7264 * recog.c (pass_peephole2): Add clone method.
7265 (pass_split_all_insns): Add clone method.
7267 2013-08-06 David Malcolm <dmalcolm@redhat.com>
7269 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
7270 (mips_option_override): ...here, porting to new C++ API for passes.
7272 2013-08-06 Jan Hubicka <jh@suse.cz>
7274 * cgraph.c (cgraph_get_body): New function based on lto.c
7276 * cgraph.h (cgraph_get_body): Declare.
7277 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
7279 * cgraphunit.c (expand_function): Get body prior expanding.
7280 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
7281 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
7283 * passes.c (do_per_function_toporder): Get body.
7284 * tree-inline.c (expand_call_inline): Get body prior inlining it.
7285 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
7287 2013-08-06 Martin Jambor <mjambor@suse.cz>
7290 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
7291 is not re-finalized. Rename second parameter to no_collect.
7293 2013-08-06 Martin Jambor <mjambor@suse.cz>
7296 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
7297 MEM_REF has proper alignment information.
7299 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
7302 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
7304 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
7305 rtx (*) (rtx, ...) with insn_gen_fn.
7306 * genoutput.c (output_insn_data): Cast gen_? function pointers to
7307 insn_gen_fn::stored_funcptr. Add initializer braces.
7309 2013-08-05 David Malcolm <dmalcolm@redhat.com>
7311 Rewrite how instances of passes are cloned to remove assumptions
7312 about their sizes (thus allowing pass subclasses to have
7313 additional data fields, albeit non-GC-managed ones at this point).
7315 * passes.c (make_pass_instance): Now that passes have clone
7316 methods, rewrite this function to eliminate XNEW and memcpy
7317 calls that used hardcoded sizes. Since this function no longer
7318 creates pass instances, rename it to...
7319 (add_pass_instance): ...this. Document the old way that passes were
7320 numbered and flagged, and rework this function to continue using it.
7321 (next_pass_1): Add an initial_pass argument for use by
7323 (position_pass): When adding multiple instances of a pass, use
7324 the pass's clone method, rather than relying on the XNEW/memcpy
7325 within the former make_pass_instance (now add_pass_instance).
7326 (pass_manager::pass_manager): When invoking next_pass_1, also supply
7327 the initial instance of the current pass within the pass manager.
7329 2013-08-05 David Malcolm <dmalcolm@redhat.com>
7331 This is the automated part of the conversion of passes from C
7332 structs to C++ classes.
7334 Patch autogenerated by refactor_passes.py from
7335 https://github.com/davidmalcolm/gcc-refactoring-scripts
7336 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
7338 * asan.c (pass_asan): Convert from a global struct to a subclass of
7339 gimple_opt_pass along with...
7340 (pass_data_asan): ...new pass_data instance and...
7341 (make_pass_asan): ...new function.
7342 (pass_asan_O0): Convert from a global struct to a subclass of
7343 gimple_opt_pass along with...
7344 (pass_data_asan_O0): ...new pass_data instance and...
7345 (make_pass_asan_O0): ...new function.
7346 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
7347 subclass of rtl_opt_pass along with...
7348 (pass_data_inc_dec): ...new pass_data instance and...
7349 (make_pass_inc_dec): ...new function.
7350 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
7351 a subclass of rtl_opt_pass along with...
7352 (pass_data_reorder_blocks): ...new pass_data instance and...
7353 (make_pass_reorder_blocks): ...new function.
7354 (pass_duplicate_computed_gotos): Convert from a global struct to a
7355 subclass of rtl_opt_pass along with...
7356 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
7357 (make_pass_duplicate_computed_gotos): ...new function.
7358 (pass_partition_blocks): Convert from a global struct to a subclass of
7359 rtl_opt_pass along with...
7360 (pass_data_partition_blocks): ...new pass_data instance and...
7361 (make_pass_partition_blocks): ...new function.
7362 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
7363 struct to a subclass of rtl_opt_pass along with...
7364 (pass_data_branch_target_load_optimize1): ...new pass_data instance
7366 (make_pass_branch_target_load_optimize1): ...new function.
7367 (pass_branch_target_load_optimize2): Convert from a global struct to a
7368 subclass of rtl_opt_pass along with...
7369 (pass_data_branch_target_load_optimize2): ...new pass_data instance
7371 (make_pass_branch_target_load_optimize2): ...new function.
7372 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
7373 of rtl_opt_pass along with...
7374 (pass_data_jump): ...new pass_data instance and...
7375 (make_pass_jump): ...new function.
7376 (pass_jump2): Convert from a global struct to a subclass of
7377 rtl_opt_pass along with...
7378 (pass_data_jump2): ...new pass_data instance and...
7379 (make_pass_jump2): ...new function.
7380 * cfgexpand.c (pass_expand): Convert from a global struct to a
7381 subclass of rtl_opt_pass along with...
7382 (pass_data_expand): ...new pass_data instance and...
7383 (make_pass_expand): ...new function.
7384 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
7385 of rtl_opt_pass along with...
7386 (pass_data_free_cfg): ...new pass_data instance and...
7387 (make_pass_free_cfg): ...new function.
7388 (pass_into_cfg_layout_mode): Convert from a global struct to a
7389 subclass of rtl_opt_pass along with...
7390 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
7391 (make_pass_into_cfg_layout_mode): ...new function.
7392 (pass_outof_cfg_layout_mode): Convert from a global struct to a
7393 subclass of rtl_opt_pass along with...
7394 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
7395 (make_pass_outof_cfg_layout_mode): ...new function.
7396 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
7397 struct to a subclass of gimple_opt_pass along with...
7398 (pass_data_build_cgraph_edges): ...new pass_data instance and...
7399 (make_pass_build_cgraph_edges): ...new function.
7400 (pass_rebuild_cgraph_edges): Convert from a global struct to a
7401 subclass of gimple_opt_pass along with...
7402 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
7403 (make_pass_rebuild_cgraph_edges): ...new function.
7404 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
7405 subclass of gimple_opt_pass along with...
7406 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
7408 (make_pass_remove_cgraph_callee_edges): ...new function.
7409 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
7410 struct to a subclass of rtl_opt_pass along with...
7411 (pass_data_stack_adjustments): ...new pass_data instance and...
7412 (make_pass_stack_adjustments): ...new function.
7413 * combine.c (pass_combine): Convert from a global struct to a subclass
7414 of rtl_opt_pass along with...
7415 (pass_data_combine): ...new pass_data instance and...
7416 (make_pass_combine): ...new function.
7417 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
7418 global struct to a subclass of rtl_opt_pass along with...
7419 (pass_data_compare_elim_after_reload): ...new pass_data instance
7421 (make_pass_compare_elim_after_reload): ...new function.
7422 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
7423 of rtl_opt_pass along with...
7424 (pass_data_rtl_cprop): ...new pass_data instance and...
7425 (make_pass_rtl_cprop): ...new function.
7426 * cse.c (pass_cse): Convert from a global struct to a subclass of
7427 rtl_opt_pass along with...
7428 (pass_data_cse): ...new pass_data instance and...
7429 (make_pass_cse): ...new function.
7430 (pass_cse2): Convert from a global struct to a subclass of
7431 rtl_opt_pass along with...
7432 (pass_data_cse2): ...new pass_data instance and...
7433 (make_pass_cse2): ...new function.
7434 (pass_cse_after_global_opts): Convert from a global struct to a
7435 subclass of rtl_opt_pass along with...
7436 (pass_data_cse_after_global_opts): ...new pass_data instance and...
7437 (make_pass_cse_after_global_opts): ...new function.
7438 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
7439 of rtl_opt_pass along with...
7440 (pass_data_ud_rtl_dce): ...new pass_data instance and...
7441 (make_pass_ud_rtl_dce): ...new function.
7442 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
7443 rtl_opt_pass along with...
7444 (pass_data_fast_rtl_dce): ...new pass_data instance and...
7445 (make_pass_fast_rtl_dce): ...new function.
7446 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
7447 a subclass of rtl_opt_pass along with...
7448 (pass_data_df_initialize_opt): ...new pass_data instance and...
7449 (make_pass_df_initialize_opt): ...new function.
7450 (pass_df_initialize_no_opt): Convert from a global struct to a
7451 subclass of rtl_opt_pass along with...
7452 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
7453 (make_pass_df_initialize_no_opt): ...new function.
7454 (pass_df_finish): Convert from a global struct to a subclass of
7455 rtl_opt_pass along with...
7456 (pass_data_df_finish): ...new pass_data instance and...
7457 (make_pass_df_finish): ...new function.
7458 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
7459 rtl_opt_pass along with...
7460 (pass_data_rtl_dse1): ...new pass_data instance and...
7461 (make_pass_rtl_dse1): ...new function.
7462 (pass_rtl_dse2): Convert from a global struct to a subclass of
7463 rtl_opt_pass along with...
7464 (pass_data_rtl_dse2): ...new pass_data instance and...
7465 (make_pass_rtl_dse2): ...new function.
7466 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
7467 subclass of rtl_opt_pass along with...
7468 (pass_data_dwarf2_frame): ...new pass_data instance and...
7469 (make_pass_dwarf2_frame): ...new function.
7470 * except.c (pass_set_nothrow_function_flags): Convert from a global
7471 struct to a subclass of rtl_opt_pass along with...
7472 (pass_data_set_nothrow_function_flags): ...new pass_data instance
7474 (make_pass_set_nothrow_function_flags): ...new function.
7475 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
7476 subclass of rtl_opt_pass along with...
7477 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
7479 (make_pass_convert_to_eh_region_ranges): ...new function.
7480 * final.c (pass_compute_alignments): Convert from a global struct to a
7481 subclass of rtl_opt_pass along with...
7482 (pass_data_compute_alignments): ...new pass_data instance and...
7483 (make_pass_compute_alignments): ...new function.
7484 (pass_final): Convert from a global struct to a subclass of
7485 rtl_opt_pass along with...
7486 (pass_data_final): ...new pass_data instance and...
7487 (make_pass_final): ...new function.
7488 (pass_shorten_branches): Convert from a global struct to a subclass of
7489 rtl_opt_pass along with...
7490 (pass_data_shorten_branches): ...new pass_data instance and...
7491 (make_pass_shorten_branches): ...new function.
7492 (pass_clean_state): Convert from a global struct to a subclass of
7493 rtl_opt_pass along with...
7494 (pass_data_clean_state): ...new pass_data instance and...
7495 (make_pass_clean_state): ...new function.
7496 * function.c (pass_instantiate_virtual_regs): Convert from a global
7497 struct to a subclass of rtl_opt_pass along with...
7498 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
7499 (make_pass_instantiate_virtual_regs): ...new function.
7500 (pass_leaf_regs): Convert from a global struct to a subclass of
7501 rtl_opt_pass along with...
7502 (pass_data_leaf_regs): ...new pass_data instance and...
7503 (make_pass_leaf_regs): ...new function.
7504 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
7505 subclass of rtl_opt_pass along with...
7506 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
7508 (make_pass_thread_prologue_and_epilogue): ...new function.
7509 (pass_match_asm_constraints): Convert from a global struct to a
7510 subclass of rtl_opt_pass along with...
7511 (pass_data_match_asm_constraints): ...new pass_data instance and...
7512 (make_pass_match_asm_constraints): ...new function.
7513 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
7514 subclass of rtl_opt_pass along with...
7515 (pass_data_rtl_fwprop): ...new pass_data instance and...
7516 (make_pass_rtl_fwprop): ...new function.
7517 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
7518 rtl_opt_pass along with...
7519 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
7520 (make_pass_rtl_fwprop_addr): ...new function.
7521 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
7522 rtl_opt_pass along with...
7523 (pass_data_rtl_pre): ...new pass_data instance and...
7524 (make_pass_rtl_pre): ...new function.
7525 (pass_rtl_hoist): Convert from a global struct to a subclass of
7526 rtl_opt_pass along with...
7527 (pass_data_rtl_hoist): ...new pass_data instance and...
7528 (make_pass_rtl_hoist): ...new function.
7529 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
7530 subclass of gimple_opt_pass along with...
7531 (pass_data_lower_cf): ...new pass_data instance and...
7532 (make_pass_lower_cf): ...new function.
7533 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
7534 from a global struct to a subclass of gimple_opt_pass along with...
7535 (pass_data_strength_reduction): ...new pass_data instance and...
7536 (make_pass_strength_reduction): ...new function.
7537 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
7538 of rtl_opt_pass along with...
7539 (pass_data_rtl_ifcvt): ...new pass_data instance and...
7540 (make_pass_rtl_ifcvt): ...new function.
7541 (pass_if_after_combine): Convert from a global struct to a subclass of
7542 rtl_opt_pass along with...
7543 (pass_data_if_after_combine): ...new pass_data instance and...
7544 (make_pass_if_after_combine): ...new function.
7545 (pass_if_after_reload): Convert from a global struct to a subclass of
7546 rtl_opt_pass along with...
7547 (pass_data_if_after_reload): ...new pass_data instance and...
7548 (make_pass_if_after_reload): ...new function.
7549 * init-regs.c (pass_initialize_regs): Convert from a global struct to
7550 a subclass of rtl_opt_pass along with...
7551 (pass_data_initialize_regs): ...new pass_data instance and...
7552 (make_pass_initialize_regs): ...new function.
7553 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
7554 of ipa_opt_pass_d along with...
7555 (pass_data_ipa_cp): ...new pass_data instance and...
7556 (make_pass_ipa_cp): ...new function.
7557 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
7558 global struct to a subclass of gimple_opt_pass along with...
7559 (pass_data_inline_parameters): ...new pass_data instance and...
7560 (make_pass_inline_parameters): ...new function.
7561 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
7562 subclass of gimple_opt_pass along with...
7563 (pass_data_early_inline): ...new pass_data instance and...
7564 (make_pass_early_inline): ...new function.
7565 (pass_ipa_inline): Convert from a global struct to a subclass of
7566 ipa_opt_pass_d along with...
7567 (pass_data_ipa_inline): ...new pass_data instance and...
7568 (make_pass_ipa_inline): ...new function.
7569 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
7570 struct to a subclass of gimple_opt_pass along with...
7571 (pass_data_local_pure_const): ...new pass_data instance and...
7572 (make_pass_local_pure_const): ...new function.
7573 (pass_ipa_pure_const): Convert from a global struct to a subclass of
7574 ipa_opt_pass_d along with...
7575 (pass_data_ipa_pure_const): ...new pass_data instance and...
7576 (make_pass_ipa_pure_const): ...new function.
7577 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
7578 to a subclass of ipa_opt_pass_d along with...
7579 (pass_data_ipa_reference): ...new pass_data instance and...
7580 (make_pass_ipa_reference): ...new function.
7581 * ipa-split.c (pass_split_functions): Convert from a global struct to
7582 a subclass of gimple_opt_pass along with...
7583 (pass_data_split_functions): ...new pass_data instance and...
7584 (make_pass_split_functions): ...new function.
7585 (pass_feedback_split_functions): Convert from a global struct to a
7586 subclass of gimple_opt_pass along with...
7587 (pass_data_feedback_split_functions): ...new pass_data instance and...
7588 (make_pass_feedback_split_functions): ...new function.
7589 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
7590 global struct to a subclass of simple_ipa_opt_pass along with...
7591 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
7593 (make_pass_ipa_function_and_variable_visibility): ...new function.
7594 (pass_ipa_free_inline_summary): Convert from a global struct to a
7595 subclass of simple_ipa_opt_pass along with...
7596 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
7597 (make_pass_ipa_free_inline_summary): ...new function.
7598 (pass_ipa_whole_program_visibility): Convert from a global struct to a
7599 subclass of ipa_opt_pass_d along with...
7600 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
7602 (make_pass_ipa_whole_program_visibility): ...new function.
7603 (pass_ipa_profile): Convert from a global struct to a subclass of
7604 ipa_opt_pass_d along with...
7605 (pass_data_ipa_profile): ...new pass_data instance and...
7606 (make_pass_ipa_profile): ...new function.
7607 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
7608 ipa_opt_pass_d along with...
7609 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
7610 (make_pass_ipa_cdtor_merge): ...new function.
7611 * ira.c (pass_ira): Convert from a global struct to a subclass of
7612 rtl_opt_pass along with...
7613 (pass_data_ira): ...new pass_data instance and...
7614 (make_pass_ira): ...new function.
7615 (pass_reload): Convert from a global struct to a subclass of
7616 rtl_opt_pass along with...
7617 (pass_data_reload): ...new pass_data instance and...
7618 (make_pass_reload): ...new function.
7619 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
7620 subclass of rtl_opt_pass along with...
7621 (pass_data_cleanup_barriers): ...new pass_data instance and...
7622 (make_pass_cleanup_barriers): ...new function.
7623 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
7624 of rtl_opt_pass along with...
7625 (pass_data_loop2): ...new pass_data instance and...
7626 (make_pass_loop2): ...new function.
7627 (pass_rtl_loop_init): Convert from a global struct to a subclass of
7628 rtl_opt_pass along with...
7629 (pass_data_rtl_loop_init): ...new pass_data instance and...
7630 (make_pass_rtl_loop_init): ...new function.
7631 (pass_rtl_loop_done): Convert from a global struct to a subclass of
7632 rtl_opt_pass along with...
7633 (pass_data_rtl_loop_done): ...new pass_data instance and...
7634 (make_pass_rtl_loop_done): ...new function.
7635 (pass_rtl_move_loop_invariants): Convert from a global struct to a
7636 subclass of rtl_opt_pass along with...
7637 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
7638 (make_pass_rtl_move_loop_invariants): ...new function.
7639 (pass_rtl_unswitch): Convert from a global struct to a subclass of
7640 rtl_opt_pass along with...
7641 (pass_data_rtl_unswitch): ...new pass_data instance and...
7642 (make_pass_rtl_unswitch): ...new function.
7643 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
7644 subclass of rtl_opt_pass along with...
7645 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
7647 (make_pass_rtl_unroll_and_peel_loops): ...new function.
7648 (pass_rtl_doloop): Convert from a global struct to a subclass of
7649 rtl_opt_pass along with...
7650 (pass_data_rtl_doloop): ...new pass_data instance and...
7651 (make_pass_rtl_doloop): ...new function.
7652 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
7653 a subclass of rtl_opt_pass along with...
7654 (pass_data_lower_subreg): ...new pass_data instance and...
7655 (make_pass_lower_subreg): ...new function.
7656 (pass_lower_subreg2): Convert from a global struct to a subclass of
7657 rtl_opt_pass along with...
7658 (pass_data_lower_subreg2): ...new pass_data instance and...
7659 (make_pass_lower_subreg2): ...new function.
7660 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
7661 struct to a subclass of ipa_opt_pass_d along with...
7662 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
7663 (make_pass_ipa_lto_gimple_out): ...new function.
7664 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
7665 of ipa_opt_pass_d along with...
7666 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
7667 (make_pass_ipa_lto_finish_out): ...new function.
7668 * mode-switching.c (pass_mode_switching): Convert from a global struct
7669 to a subclass of rtl_opt_pass along with...
7670 (pass_data_mode_switching): ...new pass_data instance and...
7671 (make_pass_mode_switching): ...new function.
7672 * modulo-sched.c (pass_sms): Convert from a global struct to a
7673 subclass of rtl_opt_pass along with...
7674 (pass_data_sms): ...new pass_data instance and...
7675 (make_pass_sms): ...new function.
7676 * omp-low.c (pass_expand_omp): Convert from a global struct to a
7677 subclass of gimple_opt_pass along with...
7678 (pass_data_expand_omp): ...new pass_data instance and...
7679 (make_pass_expand_omp): ...new function.
7680 (pass_lower_omp): Convert from a global struct to a subclass of
7681 gimple_opt_pass along with...
7682 (pass_data_lower_omp): ...new pass_data instance and...
7683 (make_pass_lower_omp): ...new function.
7684 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
7685 of gimple_opt_pass along with...
7686 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
7687 (make_pass_diagnose_omp_blocks): ...new function.
7688 * passes.c (pass_early_local_passes): Convert from a global struct to
7689 a subclass of simple_ipa_opt_pass along with...
7690 (pass_data_early_local_passes): ...new pass_data instance and...
7691 (make_pass_early_local_passes): ...new function.
7692 (pass_all_early_optimizations): Convert from a global struct to a
7693 subclass of gimple_opt_pass along with...
7694 (pass_data_all_early_optimizations): ...new pass_data instance and...
7695 (make_pass_all_early_optimizations): ...new function.
7696 (pass_all_optimizations): Convert from a global struct to a subclass
7697 of gimple_opt_pass along with...
7698 (pass_data_all_optimizations): ...new pass_data instance and...
7699 (make_pass_all_optimizations): ...new function.
7700 (pass_all_optimizations_g): Convert from a global struct to a subclass
7701 of gimple_opt_pass along with...
7702 (pass_data_all_optimizations_g): ...new pass_data instance and...
7703 (make_pass_all_optimizations_g): ...new function.
7704 (pass_rest_of_compilation): Convert from a global struct to a subclass
7705 of rtl_opt_pass along with...
7706 (pass_data_rest_of_compilation): ...new pass_data instance and...
7707 (make_pass_rest_of_compilation): ...new function.
7708 (pass_postreload): Convert from a global struct to a subclass of
7709 rtl_opt_pass along with...
7710 (pass_data_postreload): ...new pass_data instance and...
7711 (make_pass_postreload): ...new function.
7712 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
7713 subclass of rtl_opt_pass along with...
7714 (pass_data_gcse2): ...new pass_data instance and...
7715 (make_pass_gcse2): ...new function.
7716 * postreload.c (pass_postreload_cse): Convert from a global struct to
7717 a subclass of rtl_opt_pass along with...
7718 (pass_data_postreload_cse): ...new pass_data instance and...
7719 (make_pass_postreload_cse): ...new function.
7720 * predict.c (pass_profile): Convert from a global struct to a subclass
7721 of gimple_opt_pass along with...
7722 (pass_data_profile): ...new pass_data instance and...
7723 (make_pass_profile): ...new function.
7724 (pass_strip_predict_hints): Convert from a global struct to a subclass
7725 of gimple_opt_pass along with...
7726 (pass_data_strip_predict_hints): ...new pass_data instance and...
7727 (make_pass_strip_predict_hints): ...new function.
7728 * recog.c (pass_peephole2): Convert from a global struct to a subclass
7729 of rtl_opt_pass along with...
7730 (pass_data_peephole2): ...new pass_data instance and...
7731 (make_pass_peephole2): ...new function.
7732 (pass_split_all_insns): Convert from a global struct to a subclass of
7733 rtl_opt_pass along with...
7734 (pass_data_split_all_insns): ...new pass_data instance and...
7735 (make_pass_split_all_insns): ...new function.
7736 (pass_split_after_reload): Convert from a global struct to a subclass
7737 of rtl_opt_pass along with...
7738 (pass_data_split_after_reload): ...new pass_data instance and...
7739 (make_pass_split_after_reload): ...new function.
7740 (pass_split_before_regstack): Convert from a global struct to a
7741 subclass of rtl_opt_pass along with...
7742 (pass_data_split_before_regstack): ...new pass_data instance and...
7743 (make_pass_split_before_regstack): ...new function.
7744 (pass_split_before_sched2): Convert from a global struct to a subclass
7745 of rtl_opt_pass along with...
7746 (pass_data_split_before_sched2): ...new pass_data instance and...
7747 (make_pass_split_before_sched2): ...new function.
7748 (pass_split_for_shorten_branches): Convert from a global struct to a
7749 subclass of rtl_opt_pass along with...
7750 (pass_data_split_for_shorten_branches): ...new pass_data instance
7752 (make_pass_split_for_shorten_branches): ...new function.
7753 * ree.c (pass_ree): Convert from a global struct to a subclass of
7754 rtl_opt_pass along with...
7755 (pass_data_ree): ...new pass_data instance and...
7756 (make_pass_ree): ...new function.
7757 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
7758 subclass of rtl_opt_pass along with...
7759 (pass_data_stack_regs): ...new pass_data instance and...
7760 (make_pass_stack_regs): ...new function.
7761 (pass_stack_regs_run): Convert from a global struct to a subclass of
7762 rtl_opt_pass along with...
7763 (pass_data_stack_regs_run): ...new pass_data instance and...
7764 (make_pass_stack_regs_run): ...new function.
7765 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
7766 subclass of rtl_opt_pass along with...
7767 (pass_data_cprop_hardreg): ...new pass_data instance and...
7768 (make_pass_cprop_hardreg): ...new function.
7769 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
7770 subclass of rtl_opt_pass along with...
7771 (pass_data_reginfo_init): ...new pass_data instance and...
7772 (make_pass_reginfo_init): ...new function.
7773 * regmove.c (pass_regmove): Convert from a global struct to a subclass
7774 of rtl_opt_pass along with...
7775 (pass_data_regmove): ...new pass_data instance and...
7776 (make_pass_regmove): ...new function.
7777 * regrename.c (pass_regrename): Convert from a global struct to a
7778 subclass of rtl_opt_pass along with...
7779 (pass_data_regrename): ...new pass_data instance and...
7780 (make_pass_regrename): ...new function.
7781 * reorg.c (pass_delay_slots): Convert from a global struct to a
7782 subclass of rtl_opt_pass along with...
7783 (pass_data_delay_slots): ...new pass_data instance and...
7784 (make_pass_delay_slots): ...new function.
7785 (pass_machine_reorg): Convert from a global struct to a subclass of
7786 rtl_opt_pass along with...
7787 (pass_data_machine_reorg): ...new pass_data instance and...
7788 (make_pass_machine_reorg): ...new function.
7789 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
7790 of rtl_opt_pass along with...
7791 (pass_data_sched): ...new pass_data instance and...
7792 (make_pass_sched): ...new function.
7793 (pass_sched2): Convert from a global struct to a subclass of
7794 rtl_opt_pass along with...
7795 (pass_data_sched2): ...new pass_data instance and...
7796 (make_pass_sched2): ...new function.
7797 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
7798 to a subclass of rtl_opt_pass along with...
7799 (pass_data_stack_ptr_mod): ...new pass_data instance and...
7800 (make_pass_stack_ptr_mod): ...new function.
7801 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
7802 to a subclass of rtl_opt_pass along with...
7803 (pass_data_rtl_store_motion): ...new pass_data instance and...
7804 (make_pass_rtl_store_motion): ...new function.
7805 * tracer.c (pass_tracer): Convert from a global struct to a subclass
7806 of gimple_opt_pass along with...
7807 (pass_data_tracer): ...new pass_data instance and...
7808 (make_pass_tracer): ...new function.
7809 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
7810 to a subclass of gimple_opt_pass along with...
7811 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
7812 (make_pass_diagnose_tm_blocks): ...new function.
7813 (pass_lower_tm): Convert from a global struct to a subclass of
7814 gimple_opt_pass along with...
7815 (pass_data_lower_tm): ...new pass_data instance and...
7816 (make_pass_lower_tm): ...new function.
7817 (pass_tm_init): Convert from a global struct to a subclass of
7818 gimple_opt_pass along with...
7819 (pass_data_tm_init): ...new pass_data instance and...
7820 (make_pass_tm_init): ...new function.
7821 (pass_tm_mark): Convert from a global struct to a subclass of
7822 gimple_opt_pass along with...
7823 (pass_data_tm_mark): ...new pass_data instance and...
7824 (make_pass_tm_mark): ...new function.
7825 (pass_tm_edges): Convert from a global struct to a subclass of
7826 gimple_opt_pass along with...
7827 (pass_data_tm_edges): ...new pass_data instance and...
7828 (make_pass_tm_edges): ...new function.
7829 (pass_tm_memopt): Convert from a global struct to a subclass of
7830 gimple_opt_pass along with...
7831 (pass_data_tm_memopt): ...new pass_data instance and...
7832 (make_pass_tm_memopt): ...new function.
7833 (pass_ipa_tm): Convert from a global struct to a subclass of
7834 simple_ipa_opt_pass along with...
7835 (pass_data_ipa_tm): ...new pass_data instance and...
7836 (make_pass_ipa_tm): ...new function.
7837 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
7838 subclass of gimple_opt_pass along with...
7839 (pass_data_call_cdce): ...new pass_data instance and...
7840 (make_pass_call_cdce): ...new function.
7841 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
7842 subclass of gimple_opt_pass along with...
7843 (pass_data_build_cfg): ...new pass_data instance and...
7844 (make_pass_build_cfg): ...new function.
7845 (pass_split_crit_edges): Convert from a global struct to a subclass of
7846 gimple_opt_pass along with...
7847 (pass_data_split_crit_edges): ...new pass_data instance and...
7848 (make_pass_split_crit_edges): ...new function.
7849 (pass_warn_function_return): Convert from a global struct to a
7850 subclass of gimple_opt_pass along with...
7851 (pass_data_warn_function_return): ...new pass_data instance and...
7852 (make_pass_warn_function_return): ...new function.
7853 (pass_warn_function_noreturn): Convert from a global struct to a
7854 subclass of gimple_opt_pass along with...
7855 (pass_data_warn_function_noreturn): ...new pass_data instance and...
7856 (make_pass_warn_function_noreturn): ...new function.
7857 (pass_warn_unused_result): Convert from a global struct to a subclass
7858 of gimple_opt_pass along with...
7859 (pass_data_warn_unused_result): ...new pass_data instance and...
7860 (make_pass_warn_unused_result): ...new function.
7861 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
7862 a subclass of gimple_opt_pass along with...
7863 (pass_data_merge_phi): ...new pass_data instance and...
7864 (make_pass_merge_phi): ...new function.
7865 * tree-complex.c (pass_lower_complex): Convert from a global struct to
7866 a subclass of gimple_opt_pass along with...
7867 (pass_data_lower_complex): ...new pass_data instance and...
7868 (make_pass_lower_complex): ...new function.
7869 (pass_lower_complex_O0): Convert from a global struct to a subclass of
7870 gimple_opt_pass along with...
7871 (pass_data_lower_complex_O0): ...new pass_data instance and...
7872 (make_pass_lower_complex_O0): ...new function.
7873 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
7874 subclass of gimple_opt_pass along with...
7875 (pass_data_lower_eh): ...new pass_data instance and...
7876 (make_pass_lower_eh): ...new function.
7877 (pass_refactor_eh): Convert from a global struct to a subclass of
7878 gimple_opt_pass along with...
7879 (pass_data_refactor_eh): ...new pass_data instance and...
7880 (make_pass_refactor_eh): ...new function.
7881 (pass_lower_resx): Convert from a global struct to a subclass of
7882 gimple_opt_pass along with...
7883 (pass_data_lower_resx): ...new pass_data instance and...
7884 (make_pass_lower_resx): ...new function.
7885 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
7886 of gimple_opt_pass along with...
7887 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
7888 (make_pass_lower_eh_dispatch): ...new function.
7889 (pass_cleanup_eh): Convert from a global struct to a subclass of
7890 gimple_opt_pass along with...
7891 (pass_data_cleanup_eh): ...new pass_data instance and...
7892 (make_pass_cleanup_eh): ...new function.
7893 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
7894 to a subclass of simple_ipa_opt_pass along with...
7895 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
7896 (make_pass_ipa_lower_emutls): ...new function.
7897 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
7898 a subclass of gimple_opt_pass along with...
7899 (pass_data_if_conversion): ...new pass_data instance and...
7900 (make_pass_if_conversion): ...new function.
7901 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
7902 subclass of gimple_opt_pass along with...
7903 (pass_data_build_ssa): ...new pass_data instance and...
7904 (make_pass_build_ssa): ...new function.
7905 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
7906 global struct to a subclass of gimple_opt_pass along with...
7907 (pass_data_loop_distribution): ...new pass_data instance and...
7908 (make_pass_loop_distribution): ...new function.
7909 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
7910 subclass of gimple_opt_pass along with...
7911 (pass_data_mudflap_1): ...new pass_data instance and...
7912 (make_pass_mudflap_1): ...new function.
7913 (pass_mudflap_2): Convert from a global struct to a subclass of
7914 gimple_opt_pass along with...
7915 (pass_data_mudflap_2): ...new pass_data instance and...
7916 (make_pass_mudflap_2): ...new function.
7917 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
7918 subclass of gimple_opt_pass along with...
7919 (pass_data_mudflap_1): ...new pass_data instance and...
7920 (make_pass_mudflap_1): ...new function.
7921 (pass_mudflap_2): Convert from a global struct to a subclass of
7922 gimple_opt_pass along with...
7923 (pass_data_mudflap_2): ...new pass_data instance and...
7924 (make_pass_mudflap_2): ...new function.
7925 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
7926 gimple_opt_pass along with...
7927 (pass_data_nrv): ...new pass_data instance and...
7928 (make_pass_nrv): ...new function.
7929 (pass_return_slot): Convert from a global struct to a subclass of
7930 gimple_opt_pass along with...
7931 (pass_data_return_slot): ...new pass_data instance and...
7932 (make_pass_return_slot): ...new function.
7933 * tree-object-size.c (pass_object_sizes): Convert from a global struct
7934 to a subclass of gimple_opt_pass along with...
7935 (pass_data_object_sizes): ...new pass_data instance and...
7936 (make_pass_object_sizes): ...new function.
7937 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
7938 global struct to a subclass of gimple_opt_pass along with...
7939 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
7941 (make_pass_cleanup_cfg_post_optimizing): ...new function.
7942 (pass_fixup_cfg): Convert from a global struct to a subclass of
7943 gimple_opt_pass along with...
7944 (pass_data_fixup_cfg): ...new pass_data instance and...
7945 (make_pass_fixup_cfg): ...new function.
7946 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
7947 (make_pass_mudflap_1): ...new function.
7948 (pass_mudflap_2): Replace declaration with that of...
7949 (make_pass_mudflap_2): ...new function.
7950 (pass_asan): Replace declaration with that of...
7951 (make_pass_asan): ...new function.
7952 (pass_asan_O0): Replace declaration with that of...
7953 (make_pass_asan_O0): ...new function.
7954 (pass_tsan): Replace declaration with that of...
7955 (make_pass_tsan): ...new function.
7956 (pass_tsan_O0): Replace declaration with that of...
7957 (make_pass_tsan_O0): ...new function.
7958 (pass_lower_cf): Replace declaration with that of...
7959 (make_pass_lower_cf): ...new function.
7960 (pass_refactor_eh): Replace declaration with that of...
7961 (make_pass_refactor_eh): ...new function.
7962 (pass_lower_eh): Replace declaration with that of...
7963 (make_pass_lower_eh): ...new function.
7964 (pass_lower_eh_dispatch): Replace declaration with that of...
7965 (make_pass_lower_eh_dispatch): ...new function.
7966 (pass_lower_resx): Replace declaration with that of...
7967 (make_pass_lower_resx): ...new function.
7968 (pass_build_cfg): Replace declaration with that of...
7969 (make_pass_build_cfg): ...new function.
7970 (pass_early_tree_profile): Replace declaration with that of...
7971 (make_pass_early_tree_profile): ...new function.
7972 (pass_cleanup_eh): Replace declaration with that of...
7973 (make_pass_cleanup_eh): ...new function.
7974 (pass_sra): Replace declaration with that of...
7975 (make_pass_sra): ...new function.
7976 (pass_sra_early): Replace declaration with that of...
7977 (make_pass_sra_early): ...new function.
7978 (pass_early_ipa_sra): Replace declaration with that of...
7979 (make_pass_early_ipa_sra): ...new function.
7980 (pass_tail_recursion): Replace declaration with that of...
7981 (make_pass_tail_recursion): ...new function.
7982 (pass_tail_calls): Replace declaration with that of...
7983 (make_pass_tail_calls): ...new function.
7984 (pass_tree_loop): Replace declaration with that of...
7985 (make_pass_tree_loop): ...new function.
7986 (pass_tree_loop_init): Replace declaration with that of...
7987 (make_pass_tree_loop_init): ...new function.
7988 (pass_lim): Replace declaration with that of...
7989 (make_pass_lim): ...new function.
7990 (pass_tree_unswitch): Replace declaration with that of...
7991 (make_pass_tree_unswitch): ...new function.
7992 (pass_predcom): Replace declaration with that of...
7993 (make_pass_predcom): ...new function.
7994 (pass_iv_canon): Replace declaration with that of...
7995 (make_pass_iv_canon): ...new function.
7996 (pass_scev_cprop): Replace declaration with that of...
7997 (make_pass_scev_cprop): ...new function.
7998 (pass_empty_loop): Replace declaration with that of...
7999 (make_pass_empty_loop): ...new function.
8000 (pass_record_bounds): Replace declaration with that of...
8001 (make_pass_record_bounds): ...new function.
8002 (pass_graphite): Replace declaration with that of...
8003 (make_pass_graphite): ...new function.
8004 (pass_graphite_transforms): Replace declaration with that of...
8005 (make_pass_graphite_transforms): ...new function.
8006 (pass_if_conversion): Replace declaration with that of...
8007 (make_pass_if_conversion): ...new function.
8008 (pass_loop_distribution): Replace declaration with that of...
8009 (make_pass_loop_distribution): ...new function.
8010 (pass_vectorize): Replace declaration with that of...
8011 (make_pass_vectorize): ...new function.
8012 (pass_slp_vectorize): Replace declaration with that of...
8013 (make_pass_slp_vectorize): ...new function.
8014 (pass_complete_unroll): Replace declaration with that of...
8015 (make_pass_complete_unroll): ...new function.
8016 (pass_complete_unrolli): Replace declaration with that of...
8017 (make_pass_complete_unrolli): ...new function.
8018 (pass_parallelize_loops): Replace declaration with that of...
8019 (make_pass_parallelize_loops): ...new function.
8020 (pass_loop_prefetch): Replace declaration with that of...
8021 (make_pass_loop_prefetch): ...new function.
8022 (pass_iv_optimize): Replace declaration with that of...
8023 (make_pass_iv_optimize): ...new function.
8024 (pass_tree_loop_done): Replace declaration with that of...
8025 (make_pass_tree_loop_done): ...new function.
8026 (pass_ch): Replace declaration with that of...
8027 (make_pass_ch): ...new function.
8028 (pass_ccp): Replace declaration with that of...
8029 (make_pass_ccp): ...new function.
8030 (pass_phi_only_cprop): Replace declaration with that of...
8031 (make_pass_phi_only_cprop): ...new function.
8032 (pass_build_ssa): Replace declaration with that of...
8033 (make_pass_build_ssa): ...new function.
8034 (pass_build_alias): Replace declaration with that of...
8035 (make_pass_build_alias): ...new function.
8036 (pass_build_ealias): Replace declaration with that of...
8037 (make_pass_build_ealias): ...new function.
8038 (pass_dominator): Replace declaration with that of...
8039 (make_pass_dominator): ...new function.
8040 (pass_dce): Replace declaration with that of...
8041 (make_pass_dce): ...new function.
8042 (pass_dce_loop): Replace declaration with that of...
8043 (make_pass_dce_loop): ...new function.
8044 (pass_cd_dce): Replace declaration with that of...
8045 (make_pass_cd_dce): ...new function.
8046 (pass_call_cdce): Replace declaration with that of...
8047 (make_pass_call_cdce): ...new function.
8048 (pass_merge_phi): Replace declaration with that of...
8049 (make_pass_merge_phi): ...new function.
8050 (pass_split_crit_edges): Replace declaration with that of...
8051 (make_pass_split_crit_edges): ...new function.
8052 (pass_pre): Replace declaration with that of...
8053 (make_pass_pre): ...new function.
8054 (pass_profile): Replace declaration with that of...
8055 (make_pass_profile): ...new function.
8056 (pass_strip_predict_hints): Replace declaration with that of...
8057 (make_pass_strip_predict_hints): ...new function.
8058 (pass_lower_complex_O0): Replace declaration with that of...
8059 (make_pass_lower_complex_O0): ...new function.
8060 (pass_lower_complex): Replace declaration with that of...
8061 (make_pass_lower_complex): ...new function.
8062 (pass_lower_vector): Replace declaration with that of...
8063 (make_pass_lower_vector): ...new function.
8064 (pass_lower_vector_ssa): Replace declaration with that of...
8065 (make_pass_lower_vector_ssa): ...new function.
8066 (pass_lower_omp): Replace declaration with that of...
8067 (make_pass_lower_omp): ...new function.
8068 (pass_diagnose_omp_blocks): Replace declaration with that of...
8069 (make_pass_diagnose_omp_blocks): ...new function.
8070 (pass_expand_omp): Replace declaration with that of...
8071 (make_pass_expand_omp): ...new function.
8072 (pass_expand_omp_ssa): Replace declaration with that of...
8073 (make_pass_expand_omp_ssa): ...new function.
8074 (pass_object_sizes): Replace declaration with that of...
8075 (make_pass_object_sizes): ...new function.
8076 (pass_strlen): Replace declaration with that of...
8077 (make_pass_strlen): ...new function.
8078 (pass_fold_builtins): Replace declaration with that of...
8079 (make_pass_fold_builtins): ...new function.
8080 (pass_stdarg): Replace declaration with that of...
8081 (make_pass_stdarg): ...new function.
8082 (pass_early_warn_uninitialized): Replace declaration with that of...
8083 (make_pass_early_warn_uninitialized): ...new function.
8084 (pass_late_warn_uninitialized): Replace declaration with that of...
8085 (make_pass_late_warn_uninitialized): ...new function.
8086 (pass_cse_reciprocals): Replace declaration with that of...
8087 (make_pass_cse_reciprocals): ...new function.
8088 (pass_cse_sincos): Replace declaration with that of...
8089 (make_pass_cse_sincos): ...new function.
8090 (pass_optimize_bswap): Replace declaration with that of...
8091 (make_pass_optimize_bswap): ...new function.
8092 (pass_optimize_widening_mul): Replace declaration with that of...
8093 (make_pass_optimize_widening_mul): ...new function.
8094 (pass_warn_function_return): Replace declaration with that of...
8095 (make_pass_warn_function_return): ...new function.
8096 (pass_warn_function_noreturn): Replace declaration with that of...
8097 (make_pass_warn_function_noreturn): ...new function.
8098 (pass_cselim): Replace declaration with that of...
8099 (make_pass_cselim): ...new function.
8100 (pass_phiopt): Replace declaration with that of...
8101 (make_pass_phiopt): ...new function.
8102 (pass_forwprop): Replace declaration with that of...
8103 (make_pass_forwprop): ...new function.
8104 (pass_phiprop): Replace declaration with that of...
8105 (make_pass_phiprop): ...new function.
8106 (pass_tree_ifcombine): Replace declaration with that of...
8107 (make_pass_tree_ifcombine): ...new function.
8108 (pass_dse): Replace declaration with that of...
8109 (make_pass_dse): ...new function.
8110 (pass_nrv): Replace declaration with that of...
8111 (make_pass_nrv): ...new function.
8112 (pass_rename_ssa_copies): Replace declaration with that of...
8113 (make_pass_rename_ssa_copies): ...new function.
8114 (pass_sink_code): Replace declaration with that of...
8115 (make_pass_sink_code): ...new function.
8116 (pass_fre): Replace declaration with that of...
8117 (make_pass_fre): ...new function.
8118 (pass_check_data_deps): Replace declaration with that of...
8119 (make_pass_check_data_deps): ...new function.
8120 (pass_copy_prop): Replace declaration with that of...
8121 (make_pass_copy_prop): ...new function.
8122 (pass_vrp): Replace declaration with that of...
8123 (make_pass_vrp): ...new function.
8124 (pass_uncprop): Replace declaration with that of...
8125 (make_pass_uncprop): ...new function.
8126 (pass_return_slot): Replace declaration with that of...
8127 (make_pass_return_slot): ...new function.
8128 (pass_reassoc): Replace declaration with that of...
8129 (make_pass_reassoc): ...new function.
8130 (pass_rebuild_cgraph_edges): Replace declaration with that of...
8131 (make_pass_rebuild_cgraph_edges): ...new function.
8132 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
8133 (make_pass_remove_cgraph_callee_edges): ...new function.
8134 (pass_build_cgraph_edges): Replace declaration with that of...
8135 (make_pass_build_cgraph_edges): ...new function.
8136 (pass_local_pure_const): Replace declaration with that of...
8137 (make_pass_local_pure_const): ...new function.
8138 (pass_tracer): Replace declaration with that of...
8139 (make_pass_tracer): ...new function.
8140 (pass_warn_unused_result): Replace declaration with that of...
8141 (make_pass_warn_unused_result): ...new function.
8142 (pass_diagnose_tm_blocks): Replace declaration with that of...
8143 (make_pass_diagnose_tm_blocks): ...new function.
8144 (pass_lower_tm): Replace declaration with that of...
8145 (make_pass_lower_tm): ...new function.
8146 (pass_tm_init): Replace declaration with that of...
8147 (make_pass_tm_init): ...new function.
8148 (pass_tm_mark): Replace declaration with that of...
8149 (make_pass_tm_mark): ...new function.
8150 (pass_tm_memopt): Replace declaration with that of...
8151 (make_pass_tm_memopt): ...new function.
8152 (pass_tm_edges): Replace declaration with that of...
8153 (make_pass_tm_edges): ...new function.
8154 (pass_split_functions): Replace declaration with that of...
8155 (make_pass_split_functions): ...new function.
8156 (pass_feedback_split_functions): Replace declaration with that of...
8157 (make_pass_feedback_split_functions): ...new function.
8158 (pass_strength_reduction): Replace declaration with that of...
8159 (make_pass_strength_reduction): ...new function.
8160 (pass_ipa_lower_emutls): Replace declaration with that of...
8161 (make_pass_ipa_lower_emutls): ...new function.
8162 (pass_ipa_function_and_variable_visibility): Replace declaration with
8164 (make_pass_ipa_function_and_variable_visibility): ...new function.
8165 (pass_ipa_tree_profile): Replace declaration with that of...
8166 (make_pass_ipa_tree_profile): ...new function.
8167 (pass_early_local_passes): Replace declaration with that of...
8168 (make_pass_early_local_passes): ...new function.
8169 (pass_ipa_whole_program_visibility): Replace declaration with that
8171 (make_pass_ipa_whole_program_visibility): ...new function.
8172 (pass_ipa_lto_gimple_out): Replace declaration with that of...
8173 (make_pass_ipa_lto_gimple_out): ...new function.
8174 (pass_ipa_increase_alignment): Replace declaration with that of...
8175 (make_pass_ipa_increase_alignment): ...new function.
8176 (pass_ipa_inline): Replace declaration with that of...
8177 (make_pass_ipa_inline): ...new function.
8178 (pass_ipa_free_lang_data): Replace declaration with that of...
8179 (make_pass_ipa_free_lang_data): ...new function.
8180 (pass_ipa_free_inline_summary): Replace declaration with that of...
8181 (make_pass_ipa_free_inline_summary): ...new function.
8182 (pass_ipa_cp): Replace declaration with that of...
8183 (make_pass_ipa_cp): ...new function.
8184 (pass_ipa_reference): Replace declaration with that of...
8185 (make_pass_ipa_reference): ...new function.
8186 (pass_ipa_pure_const): Replace declaration with that of...
8187 (make_pass_ipa_pure_const): ...new function.
8188 (pass_ipa_pta): Replace declaration with that of...
8189 (make_pass_ipa_pta): ...new function.
8190 (pass_ipa_lto_finish_out): Replace declaration with that of...
8191 (make_pass_ipa_lto_finish_out): ...new function.
8192 (pass_ipa_tm): Replace declaration with that of...
8193 (make_pass_ipa_tm): ...new function.
8194 (pass_ipa_profile): Replace declaration with that of...
8195 (make_pass_ipa_profile): ...new function.
8196 (pass_ipa_cdtor_merge): Replace declaration with that of...
8197 (make_pass_ipa_cdtor_merge): ...new function.
8198 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
8200 (make_pass_cleanup_cfg_post_optimizing): ...new function.
8201 (pass_init_datastructures): Replace declaration with that of...
8202 (make_pass_init_datastructures): ...new function.
8203 (pass_fixup_cfg): Replace declaration with that of...
8204 (make_pass_fixup_cfg): ...new function.
8205 (pass_expand): Replace declaration with that of...
8206 (make_pass_expand): ...new function.
8207 (pass_instantiate_virtual_regs): Replace declaration with that of...
8208 (make_pass_instantiate_virtual_regs): ...new function.
8209 (pass_rtl_fwprop): Replace declaration with that of...
8210 (make_pass_rtl_fwprop): ...new function.
8211 (pass_rtl_fwprop_addr): Replace declaration with that of...
8212 (make_pass_rtl_fwprop_addr): ...new function.
8213 (pass_jump): Replace declaration with that of...
8214 (make_pass_jump): ...new function.
8215 (pass_jump2): Replace declaration with that of...
8216 (make_pass_jump2): ...new function.
8217 (pass_lower_subreg): Replace declaration with that of...
8218 (make_pass_lower_subreg): ...new function.
8219 (pass_cse): Replace declaration with that of...
8220 (make_pass_cse): ...new function.
8221 (pass_fast_rtl_dce): Replace declaration with that of...
8222 (make_pass_fast_rtl_dce): ...new function.
8223 (pass_ud_rtl_dce): Replace declaration with that of...
8224 (make_pass_ud_rtl_dce): ...new function.
8225 (pass_rtl_dce): Replace declaration with that of...
8226 (make_pass_rtl_dce): ...new function.
8227 (pass_rtl_dse1): Replace declaration with that of...
8228 (make_pass_rtl_dse1): ...new function.
8229 (pass_rtl_dse2): Replace declaration with that of...
8230 (make_pass_rtl_dse2): ...new function.
8231 (pass_rtl_dse3): Replace declaration with that of...
8232 (make_pass_rtl_dse3): ...new function.
8233 (pass_rtl_cprop): Replace declaration with that of...
8234 (make_pass_rtl_cprop): ...new function.
8235 (pass_rtl_pre): Replace declaration with that of...
8236 (make_pass_rtl_pre): ...new function.
8237 (pass_rtl_hoist): Replace declaration with that of...
8238 (make_pass_rtl_hoist): ...new function.
8239 (pass_rtl_store_motion): Replace declaration with that of...
8240 (make_pass_rtl_store_motion): ...new function.
8241 (pass_cse_after_global_opts): Replace declaration with that of...
8242 (make_pass_cse_after_global_opts): ...new function.
8243 (pass_rtl_ifcvt): Replace declaration with that of...
8244 (make_pass_rtl_ifcvt): ...new function.
8245 (pass_into_cfg_layout_mode): Replace declaration with that of...
8246 (make_pass_into_cfg_layout_mode): ...new function.
8247 (pass_outof_cfg_layout_mode): Replace declaration with that of...
8248 (make_pass_outof_cfg_layout_mode): ...new function.
8249 (pass_loop2): Replace declaration with that of...
8250 (make_pass_loop2): ...new function.
8251 (pass_rtl_loop_init): Replace declaration with that of...
8252 (make_pass_rtl_loop_init): ...new function.
8253 (pass_rtl_move_loop_invariants): Replace declaration with that of...
8254 (make_pass_rtl_move_loop_invariants): ...new function.
8255 (pass_rtl_unswitch): Replace declaration with that of...
8256 (make_pass_rtl_unswitch): ...new function.
8257 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
8258 (make_pass_rtl_unroll_and_peel_loops): ...new function.
8259 (pass_rtl_doloop): Replace declaration with that of...
8260 (make_pass_rtl_doloop): ...new function.
8261 (pass_rtl_loop_done): Replace declaration with that of...
8262 (make_pass_rtl_loop_done): ...new function.
8263 (pass_web): Replace declaration with that of...
8264 (make_pass_web): ...new function.
8265 (pass_cse2): Replace declaration with that of...
8266 (make_pass_cse2): ...new function.
8267 (pass_df_initialize_opt): Replace declaration with that of...
8268 (make_pass_df_initialize_opt): ...new function.
8269 (pass_df_initialize_no_opt): Replace declaration with that of...
8270 (make_pass_df_initialize_no_opt): ...new function.
8271 (pass_reginfo_init): Replace declaration with that of...
8272 (make_pass_reginfo_init): ...new function.
8273 (pass_inc_dec): Replace declaration with that of...
8274 (make_pass_inc_dec): ...new function.
8275 (pass_stack_ptr_mod): Replace declaration with that of...
8276 (make_pass_stack_ptr_mod): ...new function.
8277 (pass_initialize_regs): Replace declaration with that of...
8278 (make_pass_initialize_regs): ...new function.
8279 (pass_combine): Replace declaration with that of...
8280 (make_pass_combine): ...new function.
8281 (pass_if_after_combine): Replace declaration with that of...
8282 (make_pass_if_after_combine): ...new function.
8283 (pass_ree): Replace declaration with that of...
8284 (make_pass_ree): ...new function.
8285 (pass_partition_blocks): Replace declaration with that of...
8286 (make_pass_partition_blocks): ...new function.
8287 (pass_match_asm_constraints): Replace declaration with that of...
8288 (make_pass_match_asm_constraints): ...new function.
8289 (pass_regmove): Replace declaration with that of...
8290 (make_pass_regmove): ...new function.
8291 (pass_split_all_insns): Replace declaration with that of...
8292 (make_pass_split_all_insns): ...new function.
8293 (pass_fast_rtl_byte_dce): Replace declaration with that of...
8294 (make_pass_fast_rtl_byte_dce): ...new function.
8295 (pass_lower_subreg2): Replace declaration with that of...
8296 (make_pass_lower_subreg2): ...new function.
8297 (pass_mode_switching): Replace declaration with that of...
8298 (make_pass_mode_switching): ...new function.
8299 (pass_sms): Replace declaration with that of...
8300 (make_pass_sms): ...new function.
8301 (pass_sched): Replace declaration with that of...
8302 (make_pass_sched): ...new function.
8303 (pass_ira): Replace declaration with that of...
8304 (make_pass_ira): ...new function.
8305 (pass_reload): Replace declaration with that of...
8306 (make_pass_reload): ...new function.
8307 (pass_clean_state): Replace declaration with that of...
8308 (make_pass_clean_state): ...new function.
8309 (pass_branch_prob): Replace declaration with that of...
8310 (make_pass_branch_prob): ...new function.
8311 (pass_value_profile_transformations): Replace declaration with that
8313 (make_pass_value_profile_transformations): ...new function.
8314 (pass_postreload_cse): Replace declaration with that of...
8315 (make_pass_postreload_cse): ...new function.
8316 (pass_gcse2): Replace declaration with that of...
8317 (make_pass_gcse2): ...new function.
8318 (pass_split_after_reload): Replace declaration with that of...
8319 (make_pass_split_after_reload): ...new function.
8320 (pass_branch_target_load_optimize1): Replace declaration with that
8322 (make_pass_branch_target_load_optimize1): ...new function.
8323 (pass_thread_prologue_and_epilogue): Replace declaration with that
8325 (make_pass_thread_prologue_and_epilogue): ...new function.
8326 (pass_stack_adjustments): Replace declaration with that of...
8327 (make_pass_stack_adjustments): ...new function.
8328 (pass_peephole2): Replace declaration with that of...
8329 (make_pass_peephole2): ...new function.
8330 (pass_if_after_reload): Replace declaration with that of...
8331 (make_pass_if_after_reload): ...new function.
8332 (pass_regrename): Replace declaration with that of...
8333 (make_pass_regrename): ...new function.
8334 (pass_cprop_hardreg): Replace declaration with that of...
8335 (make_pass_cprop_hardreg): ...new function.
8336 (pass_reorder_blocks): Replace declaration with that of...
8337 (make_pass_reorder_blocks): ...new function.
8338 (pass_branch_target_load_optimize2): Replace declaration with that
8340 (make_pass_branch_target_load_optimize2): ...new function.
8341 (pass_leaf_regs): Replace declaration with that of...
8342 (make_pass_leaf_regs): ...new function.
8343 (pass_split_before_sched2): Replace declaration with that of...
8344 (make_pass_split_before_sched2): ...new function.
8345 (pass_compare_elim_after_reload): Replace declaration with that of...
8346 (make_pass_compare_elim_after_reload): ...new function.
8347 (pass_sched2): Replace declaration with that of...
8348 (make_pass_sched2): ...new function.
8349 (pass_stack_regs): Replace declaration with that of...
8350 (make_pass_stack_regs): ...new function.
8351 (pass_stack_regs_run): Replace declaration with that of...
8352 (make_pass_stack_regs_run): ...new function.
8353 (pass_df_finish): Replace declaration with that of...
8354 (make_pass_df_finish): ...new function.
8355 (pass_compute_alignments): Replace declaration with that of...
8356 (make_pass_compute_alignments): ...new function.
8357 (pass_duplicate_computed_gotos): Replace declaration with that of...
8358 (make_pass_duplicate_computed_gotos): ...new function.
8359 (pass_variable_tracking): Replace declaration with that of...
8360 (make_pass_variable_tracking): ...new function.
8361 (pass_free_cfg): Replace declaration with that of...
8362 (make_pass_free_cfg): ...new function.
8363 (pass_machine_reorg): Replace declaration with that of...
8364 (make_pass_machine_reorg): ...new function.
8365 (pass_cleanup_barriers): Replace declaration with that of...
8366 (make_pass_cleanup_barriers): ...new function.
8367 (pass_delay_slots): Replace declaration with that of...
8368 (make_pass_delay_slots): ...new function.
8369 (pass_split_for_shorten_branches): Replace declaration with that of...
8370 (make_pass_split_for_shorten_branches): ...new function.
8371 (pass_split_before_regstack): Replace declaration with that of...
8372 (make_pass_split_before_regstack): ...new function.
8373 (pass_convert_to_eh_region_ranges): Replace declaration with that
8375 (make_pass_convert_to_eh_region_ranges): ...new function.
8376 (pass_shorten_branches): Replace declaration with that of...
8377 (make_pass_shorten_branches): ...new function.
8378 (pass_set_nothrow_function_flags): Replace declaration with that of...
8379 (make_pass_set_nothrow_function_flags): ...new function.
8380 (pass_dwarf2_frame): Replace declaration with that of...
8381 (make_pass_dwarf2_frame): ...new function.
8382 (pass_final): Replace declaration with that of...
8383 (make_pass_final): ...new function.
8384 (pass_rtl_seqabstr): Replace declaration with that of...
8385 (make_pass_rtl_seqabstr): ...new function.
8386 (pass_release_ssa_names): Replace declaration with that of...
8387 (make_pass_release_ssa_names): ...new function.
8388 (pass_early_inline): Replace declaration with that of...
8389 (make_pass_early_inline): ...new function.
8390 (pass_inline_parameters): Replace declaration with that of...
8391 (make_pass_inline_parameters): ...new function.
8392 (pass_update_address_taken): Replace declaration with that of...
8393 (make_pass_update_address_taken): ...new function.
8394 (pass_convert_switch): Replace declaration with that of...
8395 (make_pass_convert_switch): ...new function.
8396 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
8397 to a subclass of simple_ipa_opt_pass along with...
8398 (pass_data_ipa_tree_profile): ...new pass_data instance and...
8399 (make_pass_ipa_tree_profile): ...new function.
8400 * tree-sra.c (pass_sra_early): Convert from a global struct to a
8401 subclass of gimple_opt_pass along with...
8402 (pass_data_sra_early): ...new pass_data instance and...
8403 (make_pass_sra_early): ...new function.
8404 (pass_sra): Convert from a global struct to a subclass of
8405 gimple_opt_pass along with...
8406 (pass_data_sra): ...new pass_data instance and...
8407 (make_pass_sra): ...new function.
8408 (pass_early_ipa_sra): Convert from a global struct to a subclass of
8409 gimple_opt_pass along with...
8410 (pass_data_early_ipa_sra): ...new pass_data instance and...
8411 (make_pass_early_ipa_sra): ...new function.
8412 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
8413 subclass of gimple_opt_pass along with...
8414 (pass_data_ccp): ...new pass_data instance and...
8415 (make_pass_ccp): ...new function.
8416 (pass_fold_builtins): Convert from a global struct to a subclass of
8417 gimple_opt_pass along with...
8418 (pass_data_fold_builtins): ...new pass_data instance and...
8419 (make_pass_fold_builtins): ...new function.
8420 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
8421 subclass of gimple_opt_pass along with...
8422 (pass_data_copy_prop): ...new pass_data instance and...
8423 (make_pass_copy_prop): ...new function.
8424 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
8425 global struct to a subclass of gimple_opt_pass along with...
8426 (pass_data_rename_ssa_copies): ...new pass_data instance and...
8427 (make_pass_rename_ssa_copies): ...new function.
8428 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
8429 subclass of gimple_opt_pass along with...
8430 (pass_data_dce): ...new pass_data instance and...
8431 (make_pass_dce): ...new function.
8432 (pass_dce_loop): Convert from a global struct to a subclass of
8433 gimple_opt_pass along with...
8434 (pass_data_dce_loop): ...new pass_data instance and...
8435 (make_pass_dce_loop): ...new function.
8436 (pass_cd_dce): Convert from a global struct to a subclass of
8437 gimple_opt_pass along with...
8438 (pass_data_cd_dce): ...new pass_data instance and...
8439 (make_pass_cd_dce): ...new function.
8440 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
8441 subclass of gimple_opt_pass along with...
8442 (pass_data_dominator): ...new pass_data instance and...
8443 (make_pass_dominator): ...new function.
8444 (pass_phi_only_cprop): Convert from a global struct to a subclass of
8445 gimple_opt_pass along with...
8446 (pass_data_phi_only_cprop): ...new pass_data instance and...
8447 (make_pass_phi_only_cprop): ...new function.
8448 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
8449 subclass of gimple_opt_pass along with...
8450 (pass_data_dse): ...new pass_data instance and...
8451 (make_pass_dse): ...new function.
8452 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
8453 a subclass of gimple_opt_pass along with...
8454 (pass_data_forwprop): ...new pass_data instance and...
8455 (make_pass_forwprop): ...new function.
8456 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
8457 struct to a subclass of gimple_opt_pass along with...
8458 (pass_data_tree_ifcombine): ...new pass_data instance and...
8459 (make_pass_tree_ifcombine): ...new function.
8460 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
8461 subclass of gimple_opt_pass along with...
8462 (pass_data_ch): ...new pass_data instance and...
8463 (make_pass_ch): ...new function.
8464 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
8465 subclass of gimple_opt_pass along with...
8466 (pass_data_tree_loop): ...new pass_data instance and...
8467 (make_pass_tree_loop): ...new function.
8468 (pass_tree_loop_init): Convert from a global struct to a subclass of
8469 gimple_opt_pass along with...
8470 (pass_data_tree_loop_init): ...new pass_data instance and...
8471 (make_pass_tree_loop_init): ...new function.
8472 (pass_lim): Convert from a global struct to a subclass of
8473 gimple_opt_pass along with...
8474 (pass_data_lim): ...new pass_data instance and...
8475 (make_pass_lim): ...new function.
8476 (pass_tree_unswitch): Convert from a global struct to a subclass of
8477 gimple_opt_pass along with...
8478 (pass_data_tree_unswitch): ...new pass_data instance and...
8479 (make_pass_tree_unswitch): ...new function.
8480 (pass_predcom): Convert from a global struct to a subclass of
8481 gimple_opt_pass along with...
8482 (pass_data_predcom): ...new pass_data instance and...
8483 (make_pass_predcom): ...new function.
8484 (pass_vectorize): Convert from a global struct to a subclass of
8485 gimple_opt_pass along with...
8486 (pass_data_vectorize): ...new pass_data instance and...
8487 (make_pass_vectorize): ...new function.
8488 (pass_graphite): Convert from a global struct to a subclass of
8489 gimple_opt_pass along with...
8490 (pass_data_graphite): ...new pass_data instance and...
8491 (make_pass_graphite): ...new function.
8492 (pass_graphite_transforms): Convert from a global struct to a subclass
8493 of gimple_opt_pass along with...
8494 (pass_data_graphite_transforms): ...new pass_data instance and...
8495 (make_pass_graphite_transforms): ...new function.
8496 (pass_check_data_deps): Convert from a global struct to a subclass of
8497 gimple_opt_pass along with...
8498 (pass_data_check_data_deps): ...new pass_data instance and...
8499 (make_pass_check_data_deps): ...new function.
8500 (pass_iv_canon): Convert from a global struct to a subclass of
8501 gimple_opt_pass along with...
8502 (pass_data_iv_canon): ...new pass_data instance and...
8503 (make_pass_iv_canon): ...new function.
8504 (pass_scev_cprop): Convert from a global struct to a subclass of
8505 gimple_opt_pass along with...
8506 (pass_data_scev_cprop): ...new pass_data instance and...
8507 (make_pass_scev_cprop): ...new function.
8508 (pass_record_bounds): Convert from a global struct to a subclass of
8509 gimple_opt_pass along with...
8510 (pass_data_record_bounds): ...new pass_data instance and...
8511 (make_pass_record_bounds): ...new function.
8512 (pass_complete_unroll): Convert from a global struct to a subclass of
8513 gimple_opt_pass along with...
8514 (pass_data_complete_unroll): ...new pass_data instance and...
8515 (make_pass_complete_unroll): ...new function.
8516 (pass_complete_unrolli): Convert from a global struct to a subclass of
8517 gimple_opt_pass along with...
8518 (pass_data_complete_unrolli): ...new pass_data instance and...
8519 (make_pass_complete_unrolli): ...new function.
8520 (pass_parallelize_loops): Convert from a global struct to a subclass
8521 of gimple_opt_pass along with...
8522 (pass_data_parallelize_loops): ...new pass_data instance and...
8523 (make_pass_parallelize_loops): ...new function.
8524 (pass_loop_prefetch): Convert from a global struct to a subclass of
8525 gimple_opt_pass along with...
8526 (pass_data_loop_prefetch): ...new pass_data instance and...
8527 (make_pass_loop_prefetch): ...new function.
8528 (pass_iv_optimize): Convert from a global struct to a subclass of
8529 gimple_opt_pass along with...
8530 (pass_data_iv_optimize): ...new pass_data instance and...
8531 (make_pass_iv_optimize): ...new function.
8532 (pass_tree_loop_done): Convert from a global struct to a subclass of
8533 gimple_opt_pass along with...
8534 (pass_data_tree_loop_done): ...new pass_data instance and...
8535 (make_pass_tree_loop_done): ...new function.
8536 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
8537 struct to a subclass of gimple_opt_pass along with...
8538 (pass_data_cse_reciprocals): ...new pass_data instance and...
8539 (make_pass_cse_reciprocals): ...new function.
8540 (pass_cse_sincos): Convert from a global struct to a subclass of
8541 gimple_opt_pass along with...
8542 (pass_data_cse_sincos): ...new pass_data instance and...
8543 (make_pass_cse_sincos): ...new function.
8544 (pass_optimize_bswap): Convert from a global struct to a subclass of
8545 gimple_opt_pass along with...
8546 (pass_data_optimize_bswap): ...new pass_data instance and...
8547 (make_pass_optimize_bswap): ...new function.
8548 (pass_optimize_widening_mul): Convert from a global struct to a
8549 subclass of gimple_opt_pass along with...
8550 (pass_data_optimize_widening_mul): ...new pass_data instance and...
8551 (make_pass_optimize_widening_mul): ...new function.
8552 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
8553 subclass of gimple_opt_pass along with...
8554 (pass_data_phiopt): ...new pass_data instance and...
8555 (make_pass_phiopt): ...new function.
8556 (pass_cselim): Convert from a global struct to a subclass of
8557 gimple_opt_pass along with...
8558 (pass_data_cselim): ...new pass_data instance and...
8559 (make_pass_cselim): ...new function.
8560 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
8561 subclass of gimple_opt_pass along with...
8562 (pass_data_phiprop): ...new pass_data instance and...
8563 (make_pass_phiprop): ...new function.
8564 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
8565 subclass of gimple_opt_pass along with...
8566 (pass_data_pre): ...new pass_data instance and...
8567 (make_pass_pre): ...new function.
8568 (pass_fre): Convert from a global struct to a subclass of
8569 gimple_opt_pass along with...
8570 (pass_data_fre): ...new pass_data instance and...
8571 (make_pass_fre): ...new function.
8572 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
8573 subclass of gimple_opt_pass along with...
8574 (pass_data_reassoc): ...new pass_data instance and...
8575 (make_pass_reassoc): ...new function.
8576 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
8577 subclass of gimple_opt_pass along with...
8578 (pass_data_sink_code): ...new pass_data instance and...
8579 (make_pass_sink_code): ...new function.
8580 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
8581 subclass of gimple_opt_pass along with...
8582 (pass_data_strlen): ...new pass_data instance and...
8583 (make_pass_strlen): ...new function.
8584 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
8585 struct to a subclass of gimple_opt_pass along with...
8586 (pass_data_build_alias): ...new pass_data instance and...
8587 (make_pass_build_alias): ...new function.
8588 (pass_build_ealias): Convert from a global struct to a subclass of
8589 gimple_opt_pass along with...
8590 (pass_data_build_ealias): ...new pass_data instance and...
8591 (make_pass_build_ealias): ...new function.
8592 (pass_ipa_pta): Convert from a global struct to a subclass of
8593 simple_ipa_opt_pass along with...
8594 (pass_data_ipa_pta): ...new pass_data instance and...
8595 (make_pass_ipa_pta): ...new function.
8596 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
8597 subclass of gimple_opt_pass along with...
8598 (pass_data_uncprop): ...new pass_data instance and...
8599 (make_pass_uncprop): ...new function.
8600 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
8601 global struct to a subclass of gimple_opt_pass along with...
8602 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
8603 (make_pass_late_warn_uninitialized): ...new function.
8604 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
8605 to a subclass of gimple_opt_pass along with...
8606 (pass_data_init_datastructures): ...new pass_data instance and...
8607 (make_pass_init_datastructures): ...new function.
8608 (pass_early_warn_uninitialized): Convert from a global struct to a
8609 subclass of gimple_opt_pass along with...
8610 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
8611 (make_pass_early_warn_uninitialized): ...new function.
8612 (pass_update_address_taken): Convert from a global struct to a
8613 subclass of gimple_opt_pass along with...
8614 (pass_data_update_address_taken): ...new pass_data instance and...
8615 (make_pass_update_address_taken): ...new function.
8616 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
8617 struct to a subclass of gimple_opt_pass along with...
8618 (pass_data_release_ssa_names): ...new pass_data instance and...
8619 (make_pass_release_ssa_names): ...new function.
8620 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
8621 subclass of gimple_opt_pass along with...
8622 (pass_data_stdarg): ...new pass_data instance and...
8623 (make_pass_stdarg): ...new function.
8624 * tree-switch-conversion.c (pass_convert_switch): Convert from a
8625 global struct to a subclass of gimple_opt_pass along with...
8626 (pass_data_convert_switch): ...new pass_data instance and...
8627 (make_pass_convert_switch): ...new function.
8628 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
8629 to a subclass of gimple_opt_pass along with...
8630 (pass_data_tail_recursion): ...new pass_data instance and...
8631 (make_pass_tail_recursion): ...new function.
8632 (pass_tail_calls): Convert from a global struct to a subclass of
8633 gimple_opt_pass along with...
8634 (pass_data_tail_calls): ...new pass_data instance and...
8635 (make_pass_tail_calls): ...new function.
8636 * tree-vect-generic.c (pass_lower_vector): Convert from a global
8637 struct to a subclass of gimple_opt_pass along with...
8638 (pass_data_lower_vector): ...new pass_data instance and...
8639 (make_pass_lower_vector): ...new function.
8640 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
8641 gimple_opt_pass along with...
8642 (pass_data_lower_vector_ssa): ...new pass_data instance and...
8643 (make_pass_lower_vector_ssa): ...new function.
8644 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
8645 to a subclass of gimple_opt_pass along with...
8646 (pass_data_slp_vectorize): ...new pass_data instance and...
8647 (make_pass_slp_vectorize): ...new function.
8648 (pass_ipa_increase_alignment): Convert from a global struct to a
8649 subclass of simple_ipa_opt_pass along with...
8650 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
8651 (make_pass_ipa_increase_alignment): ...new function.
8652 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
8653 gimple_opt_pass along with...
8654 (pass_data_vrp): ...new pass_data instance and...
8655 (make_pass_vrp): ...new function.
8656 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
8657 subclass of simple_ipa_opt_pass along with...
8658 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
8659 (make_pass_ipa_free_lang_data): ...new function.
8660 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
8661 gimple_opt_pass along with...
8662 (pass_data_tsan): ...new pass_data instance and...
8663 (make_pass_tsan): ...new function.
8664 (pass_tsan_O0): Convert from a global struct to a subclass of
8665 gimple_opt_pass along with...
8666 (pass_data_tsan_O0): ...new pass_data instance and...
8667 (make_pass_tsan_O0): ...new function.
8668 * var-tracking.c (pass_variable_tracking): Convert from a global
8669 struct to a subclass of rtl_opt_pass along with...
8670 (pass_data_variable_tracking): ...new pass_data instance and...
8671 (make_pass_variable_tracking): ...new function.
8672 * web.c (pass_web): Convert from a global struct to a subclass of
8673 rtl_opt_pass along with...
8674 (pass_data_web): ...new pass_data instance and...
8675 (make_pass_web): ...new function.
8676 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
8677 declaration with that of...
8678 (make_pass_mode_switch_use): ...new function.
8679 (pass_resolve_sw_modes): Replace declaration with that of...
8680 (make_pass_resolve_sw_modes): ...new function.
8681 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
8682 from a global struct to a subclass of rtl_opt_pass along with...
8683 (pass_data_mode_switch_use): ...new pass_data instance and...
8684 (make_pass_mode_switch_use): ...new function.
8685 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
8686 from a global struct to a subclass of rtl_opt_pass along with...
8687 (pass_data_resolve_sw_modes): ...new pass_data instance and...
8688 (make_pass_resolve_sw_modes): ...new function.
8689 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
8690 struct to a subclass of rtl_opt_pass along with...
8691 (pass_data_insert_vzeroupper): ...new pass_data instance and...
8692 (make_pass_insert_vzeroupper): ...new function.
8693 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
8694 global struct to a subclass of rtl_opt_pass along with...
8695 (pass_data_work_around_errata): ...new pass_data instance and...
8696 (make_pass_work_around_errata): ...new function.
8697 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
8698 struct to a subclass of rtl_opt_pass along with...
8699 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
8700 (make_pass_mips_machine_reorg2): ...new function.
8702 2013-08-05 David Malcolm <dmalcolm@redhat.com>
8704 * passes.c (pass_manager::operator new): New.
8706 2013-08-05 David Malcolm <dmalcolm@redhat.com>
8708 Handwritten part of conversion of passes to C++ classes.
8710 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
8711 (toplev.o): Add dep on PASS_MANAGER_H.
8712 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
8713 of early local pases to reflect this moving from a global to a
8714 member of gcc::pass_manager.
8715 (cgraph_add_new_function): Likewise.
8716 * lto-cgraph.c (lto_output_node): Update for conversion of
8717 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
8718 * passes.c (opt_pass::clone): New.
8719 (opt_pass::gate): New.
8720 (opt_pass::execute): New.
8721 (opt_pass::opt_pass): New.
8722 (pass_manager::execute_early_local_passes): New.
8723 (pass_manager::execute_pass_mode_switching): new.
8724 (finish_optimization_passes): Convert to...
8725 (pass_manager::finish_optimization_passes): ...this.
8726 (finish_optimization_passes): Update for conversion of passes to
8728 (register_dump_files_1): Use has_gate since we cannot portably
8729 check a vtable entry against NULL.
8730 (dump_one_pass): Likewise.
8731 (ipa_write_summaries_2): Likewise.
8732 (ipa_write_optimization_summaries_1): Likewise.
8733 (ipa_read_summaries_1): Likewise.
8734 (ipa_read_optimization_summaries_1): Likewise.
8735 (execute_ipa_stmt_fixups): Likewise.
8736 (pass_manager::pass_manager): Rewrite pass-creation, invoking
8737 pass-creation functions rather than wiring up globals, and
8738 storing the results in fields of pass_manager generated using
8740 (pass_manager::dump_profile_report): Update for conversion of
8741 passes to C++ classes.
8742 (pass_manager::execute_ipa_summary_passes): Likewise.
8743 (execute_one_ipa_transform_pass): Likewise.
8744 (execute_one_pass): Use has_gate and has_execute since we cannot
8745 portably check a vtable entry against NULL.
8746 * pass_manager.h (pass_manager::finish_optimization_passes): New.
8747 (pass_manager): Use pass-instances.def to add fields for the
8748 various pass instances.
8749 * toplev.c (finalize): Update for move of
8750 finish_optimization_passes to a method of gcc::pass_manager.
8751 * toplev.h (finish_optimization_passes): Move to method of class
8753 * tree-pass.h (struct pass_data): New.
8754 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
8755 (opt_pass::gate): Convert to virtual function.
8756 (opt_pass::~opt_pass): New.
8757 (opt_pass::clone): New.
8758 (opt_pass::execute): Convert to virtual function.
8759 (opt_pass::opt_pass): New.
8760 (opt_pass::ctxt_): new.
8761 (gimple_opt_pass): Convert to subclass of opt_pass.
8762 (gimple_opt_pass::gimple_opt_pass): New.
8763 (rtl_opt_pass): Convert to subclass of opt_pass.
8764 (rtl_opt_pass::rtl_opt_pass): New.
8765 (ipa_opt_pass_d): Convert to subclass of opt_pass.
8766 (ipa_opt_pass_d::ipa_opt_pass_d): New.
8767 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
8768 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
8769 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
8770 invocation of pass_mode_switching to reflect this moving from a
8771 global to a member of gcc::pass_manager.
8772 (ix86_option_override): Rework how pass_insert_vzeroupper is
8773 added to the pass_manager to reflect autogenerated changes.
8774 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
8776 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
8778 PR rtl-optimization/57708
8779 * recog.c (peep2_find_free_register): Validate all regs in a
8782 2013-08-05 Jan Hubicka <jh@suse.cz>
8785 * cgraph.c (verify_cgraph_node): Accept local flags from other
8787 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
8788 (function_and_variable_visibility): Likewise.
8789 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
8791 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
8793 * graph.c (init_graph_slim_pretty_print): Remove.
8794 (print_graph_cfg): Do not call it. Use local pretty printer.
8795 (start_graph_dump): Likewise.
8797 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
8799 * gimple-pretty-print.c (buffer): Remove.
8800 (initialized): Likewise.
8801 (maybe_init_pretty_print): Likewise.
8802 (print_gimple_stmt): Do not call it. Use non-static local
8803 pretty_printer variable.
8804 (print_gimple_expr): Likewise.
8805 (print_gimple_seq): Likewise.
8806 (gimple_dump_bb): Likewise.
8808 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
8810 * asan.c (asan_pp): Remove.
8811 (asan_pp_initialized): Likewise.
8812 (asan_pp_initialize): Likewise.
8813 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
8814 (asan_emit_stack_protection): Tidy. Use local pretty printer.
8815 (asan_add_global): Likewise.
8817 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
8819 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
8820 * diagnostic.h (diagnostic_flush_buffer): Adjust.
8821 * pretty-print.c (pp_formatted_text_data): Likewise.
8822 (pp_indent): Rename from pp_base_indent.
8823 (pp_format): Rename from pp_base_format.
8824 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
8825 (pp_format_verbatim): Rename from pp_base_format_verbatim.
8826 (pp_flush): Rename from pp_base_flush.
8827 (pp_set_line_maximum_length): Rename from
8828 pp_base_set_line_maximum_length.
8829 (pp_clear_output_area): Rename from pp_base_clear_output_area.
8830 (pp_set_prefix): Rename from pp_base_set_prefix.
8831 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
8832 (pp_emit_prefix): Rename from pp_base_emit_prefix.
8833 (pp_append_text): Rename from pp_base_append_text.
8834 (pp_formatted_text): Rename from pp_base_formatted_text.
8835 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
8836 (pp_remaining_character_count_for_line): Rename from
8837 pp_base_remaining_character_count_for_line.
8838 (pp_newline): Rename from pp_base_newline.
8839 (pp_character): Rename from pp_base_character.
8840 (pp_string): Rename from pp_base_string.
8841 (pp_maybe_space): Rename from pp_base_maybe_space.
8842 * asan.c (asan_pp_string): Adjust.
8843 (asan_emit_stack_protection): Likewise.
8844 (asan_add_global): Likewise.
8845 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
8846 * tree-mudflap.c (mf_varname_tree): Likewise.
8847 * tree-pretty-print.c (pp_tree_identifier): Rename from
8848 pp_base_tree_identifier.
8849 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
8850 Declare as function.
8852 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
8854 * pretty-print.h (pp_bar_bar): New.
8855 (pp_ampersand_ampersand): Likewise.
8856 (pp_less_equal): Likewise.
8857 (pp_greater_equal): Likewise.
8858 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
8859 printer functions instead of pp_string or operators and punctuators.
8860 (dump_gimple_call): Likewise.
8861 (dump_gimple_omp_for): Likewise.
8862 (dump_gimple_transaction): Likewise.
8863 (dump_gimple_phi): Likewise.
8864 (pp_gimple_stmt_1): Likewise.
8865 * sched-vis.c (print_insn): Likewise.
8866 * tree-mudflap.c (mf_varname_tree): Likewise.
8867 * tree-pretty-print.c (dump_block_node): Likewise.
8868 (dump_generic_node): Likewise.
8870 2013-08-02 Jan Hubicka <jh@suse.cz>
8872 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
8874 * lto-streamer-out.c (tree_is_indexable): Results decls and
8875 parm decls are not indexable.
8876 (DFS_write_tree_body): Do not follow args and results.
8877 (hash_tree): Likewise.
8878 (output_functions): Rearrange so struct function is needed
8879 only when real body is output; be able to also ouptut abstract
8880 functions; output DECL_ARGUMENTS and DECL_RESULT.
8881 (lto_output): When not in WPA, ale store abstract functions.
8882 (write_symbol): Do not care about RESULT_DECL.
8883 (output_symbol_p): Handle correctly sbtract decls.
8884 * lto-streamer-in.c (input_function): Rearrange so struct
8885 function can be NULL at entry; allow streaming of
8886 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
8887 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
8888 sanity check during LTO.
8889 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
8890 RESULT_DECl and DECL_ARGUMENTS.
8891 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
8894 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
8896 * pretty-print.h (pp_underscore): New.
8898 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
8899 printer functions instead of pp_character.
8900 (dump_binary_rhs): Likewise.
8901 (dump_ternary_rhs): Likewise.
8902 (dump_gimple_call_args): Likewise.
8903 (pp_points_to_solution): Likewise.
8904 (dump_gimple_call): Likewise.
8905 (dump_gimple_switch): Likewise.
8906 (dump_gimple_cond): Likewise.
8907 (dump_gimple_bind): Likewise.
8908 (dump_gimple_try): Likewise.
8909 (dump_gimple_omp_for): Likewise.
8910 (dump_gimple_omp_continue): Likewise.
8911 (dump_gimple_omp_single): Likewise.
8912 (dump_gimple_omp_sections): Likewise.
8913 (dump_gimple_omp_block): Likewise.
8914 (dump_gimple_omp_critical): Likewise.
8915 (dump_gimple_transaction): Likewise.
8916 (dump_gimple_asm): Likewise.
8917 (dump_gimple_phi): Likewise.
8918 (dump_gimple_omp_parallel): Likewise.
8919 (dump_gimple_omp_task): Likewise.
8920 (dump_gimple_omp_atomic_load): Likewise.
8921 (dump_gimple_omp_atomic_store): Likewise.
8922 (dump_gimple_mem_ops): Likewise.
8923 (pp_gimple_stmt_1): Likewise.
8924 (pp_cfg_jump): Likewise.
8925 (dump_implicit_edges): Likewise.
8926 (gimple_dump_bb_for_graph): Likewise.
8927 * graph.c (draw_cfg_node): Likewise.
8928 * langhooks.c (lhd_print_error_function): Likewise.
8929 * sched-vis.c (print_exp): Likewise.
8930 (print_value): Likewise.
8931 (print_pattern): Likewise.
8932 (print_insn): Likewise.
8933 (rtl_dump_bb_for_graph): Likewise.
8934 * tree-pretty-print.c (dump_function_declaration): Likewise.
8935 (dump_array_domain): Likewise.
8936 (dump_omp_clause): Likewise.
8937 (dump_location): Likewise.
8938 (dump_generic_node): Likewise.
8939 (print_struct_decl): Likewise.
8940 * diagnostic.c (diagnostic_show_locus): Use pp_space.
8942 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
8944 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
8945 candidate table when replacing a candidate statement.
8946 (replace_rhs_if_not_dup): Likewise.
8947 (replace_one_candidate): Likewise.
8949 2013-08-02 Jan Hubicka <jh@suse.cz>
8950 Martin Liska <marxin.liska@gmail.com>
8952 * cgraphunit.c (add_new_function): Fix logic when adding from
8954 (assemble_thunk): Rename to ...
8955 (expand_thunk); .. this one; export; get it working with
8956 general functions; make produced gimple valid.
8957 * cgraph.h (expand_thunk): Declare.
8959 2013-08-02 Jan Hubicka <jh@suse.cz>
8961 * ipa-cp.c (gather_context_independent_values): Use
8962 ipa_get_param_move_cost.
8963 (get_replacement_map): Remove PARAM; move parameter folding
8965 (create_specialized_node): Update.
8966 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
8967 assert that we have gimple body; update move_cost.
8968 (count_formal_params): Assert that we have gimple body.
8969 (ipa_dump_param): New function.
8970 (ipa_alloc_node_params): Break out from ...
8971 (ipa_initialize_node_params): ... here.
8972 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
8973 (ipa_write_node_info): Stream move costs.
8974 (ipa_read_node_info): Read move costs.
8975 (ipa_update_after_lto_read): Do not recompute node params.
8976 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
8977 (ipa_get_param): Check we are not in WPA.
8978 (ipa_get_param_move_cost): New.
8979 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
8980 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
8981 parm numbers to be present.
8983 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
8985 PR rtl-optimization/58048
8986 * lra-constraints.c (process_alt_operands): Don't check asm
8987 operand on register.
8989 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
8991 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
8992 the implied StoreLoad barrier for atomic operations if before.
8994 2013-08-02 Jan Hubicka <jh@suse.cz>
8995 Martin Liska <marxin.liska@gmail.com>
8997 * cgraph.c (cgraph_function_body_availability): Do not check
8999 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
9000 symtab_node_availability): Declare.
9001 * ipa.c (can_replace_by_local_alias): New.
9002 (function_and_variable_visibility): Use it.
9003 * symtab.c (symtab_for_node_and_aliases,
9004 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
9006 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
9008 PR rtl-optimization/57963
9009 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
9010 (lra_constraints): Use them.
9012 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
9014 * config/arm/types.md (define_attr "type"): Add "load_acq"
9016 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
9018 (cortex_a53_store1): Likewise.
9020 2013-08-01 Jan Hubicka <jh@suse.cz>
9022 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
9023 partitions are not needed.
9025 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
9027 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
9028 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
9029 MAYBE_NON_Q_CLASS_P where appropriate.
9031 2013-08-01 Jan Hubicka <jh@suse.cz>
9033 * cgraph.h (release_function_body): Declare.
9034 * tree.c (free_lang_data_in_decl): Free, parameters and return values
9035 of unused delcarations.
9037 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9039 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
9040 RTL form when subtracting a constant.
9042 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9044 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
9045 Generate canonical plus rtx with negated immediate instead of minus
9047 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
9049 2013-08-01 Jan Hubicka <jh@suse.cz>
9051 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
9052 (cgraph_release_function_body): Likewise.
9053 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
9054 * cgraph.h (cgrpah_node): Rename abstract_and_needed
9055 to used_as_abstract_origin.
9056 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
9057 symbols used as abstract origins.
9058 * cgraphunit.c (analyze_functions): Update.
9059 * ipa.c (symtab_remove_unreachable_nodes): Recompute
9060 used_as_abstract_origin.
9061 * tree-inline.c (tree_function_versioning): Update
9062 used_as_abstract_origin; be ready for DECL_RESULT and
9063 DECL_ARGUMENTS to be NULL.
9065 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
9066 for abstract functions.
9067 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
9070 2013-08-01 Jan Hubicka <jh@suse.cz>
9072 * profile.c (compute_value_histograms): Fix thinko.
9074 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
9076 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
9077 aarch-common-protos.h to extra_headers.
9078 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
9079 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
9080 * config/aarch64/t-aarch64 (aarch-common.o): Define.
9082 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
9084 * config/aarch64/aarch64.md (define_attr "type"): Delete.
9085 Include "../arm/types.md". Define "type" attribute for all patterns.
9086 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
9089 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
9091 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
9092 to support power8 load fusion.
9093 (fusion_gpr_mem_load): Likewise.
9095 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
9097 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
9098 declarations for power8 load fusion.
9099 (emit_fusion_gpr_load): Likewise.
9101 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
9102 tuning for power8, turn on fusion mode by default. Turn on sign
9103 extending fusion mode if normal fusion mode is on, and we are at
9105 (fusion_gpr_load_p): New function, return true if we can fuse an
9106 addis instruction with a dependent load to a GPR.
9107 (emit_fusion_gpr_load): Emit the instructions for power8 load
9110 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
9111 (VSX load fusion peepholes): New peepholes to fuse together an
9112 addi instruction with a VSX load instruction.
9114 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
9115 peepholes to fuse an addis instruction with a load to a GPR base
9116 register. If we are supporting sign extending fusions, convert
9117 sign extending loads to zero extending loads and add an explicit
9120 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
9122 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
9123 aarch-common-protos.h to extra_headers.
9124 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
9125 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
9126 (arm_early_store_addr_dep): Likewise.
9127 (arm_no_early_alu_shift_dep): Likewise.
9128 (arm_no_early_alu_shift_value_dep): Likewise.
9129 (arm_no_early_mul_dep): Likewise.
9130 (arm_no_early_store_addr_dep): Likewise.
9131 (arm_mac_accumulator_is_mul_result): Likewise.
9132 (arm_mac_accumulator_is_result): Likewise.
9133 * config/arm/aarch-common.c: ... here. New file.
9134 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
9136 (arm_early_store_addr_dep): Likewise.
9137 (arm_no_early_alu_shift_dep): Likewise.
9138 (arm_no_early_alu_shift_value_dep): Likewise.
9139 (arm_no_early_mul_dep): Likewise.
9140 (arm_no_early_store_addr_dep): Likewise.
9141 (arm_mac_accumulator_is_mul_result): Likewise.
9142 (arm_mac_accumulator_is_result): Likewise.
9143 * config/arm/aarch-common-protos.h: ... here. New file.
9144 * config/arm/t-arm (aarch-common.o): Define.
9146 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
9148 * config/arm/arm.md: Include new file "types.md".
9149 (define_attr "type"): Move from here to ...
9150 (define_attr "mul32"): Likewise.
9151 (define_attr "mul64"): Likewise.
9152 * config/arm/types.md: ... here. New file.
9154 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
9156 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
9157 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
9159 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9161 * gen-pass-instances.awk: Fix offset of substr().
9163 2013-07-31 David Malcolm <dmalcolm@redhat.com>
9165 * Makefile.in (pass-instances.def): New.
9166 (passes.o): Replace dependency on passes.def with one on
9169 * gen-pass-instances.awk: New.
9171 * passes.c (pass_manager::pass_manager): Use pass-instances.def
9172 rather than passes.def, updating local definition of NEXT_PASS
9173 macro to add an extra NUM parameter (currently unused).
9175 2013-07-30 David Malcolm <dmalcolm@redhat.com>
9177 * Makefile.in (PASS_MANAGER_H): New.
9178 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
9179 (passes.o): Likewise.
9180 (statistics.o): Likewise.
9181 (cgraphunit.o): Likewise.
9182 (context.o): Depend on PASS_MANAGER_H.
9184 * pass_manager.h: New.
9186 * cgraphunit.c (cgraph_add_new_function): Update for moves
9187 of globals to fields of pass_manager.
9188 (analyze_function): Likewise.
9189 (expand_function): Likewise.
9190 (ipa_passes): Likewise.
9191 (compile): Likewise.
9193 * context.c (context::context): New.
9194 * context.h (context::context): New.
9195 (context::get_passes): New.
9196 (context::passes_): New.
9198 * lto-cgraph.c (input_node): Update for moves of globals to
9199 fields of pass_manager.
9201 * passes.c (all_passes): Remove, in favor of a field of the
9202 same name within the new class pass_manager.
9203 (all_small_ipa_passes): Likewise.
9204 (all_lowering_passes): Likewise.
9205 (all_regular_ipa_passes): Likewise.
9206 (all_late_ipa_passes): Likewise.
9207 (all_lto_gen_passes): Likewise.
9208 (passes_by_id): Likewise.
9209 (passes_by_id_size): Likewise.
9210 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
9211 the new class pass_manager.
9212 (set_pass_for_id): Convert to...
9213 (pass_manager::set_pass_for_id): ...method.
9214 (get_pass_for_id): Convert to...
9215 (pass_manager::get_pass_for_id): ...method.
9216 (register_one_dump_file): Move body of implementation into...
9217 (pass_manager::register_one_dump_file): ...here.
9218 (register_dump_files_1): Convert to...
9219 (pass_manager::register_dump_files_1): ...method.
9220 (register_dump_files): Convert to...
9221 (pass_manager::register_dump_files): ...method.
9222 (create_pass_tab): Update for moves of globals to fields of
9224 (dump_passes): Move body of implementation into...
9225 (pass_manager::dump_passes): ...here.
9226 (register_pass): Move body of implementation into...
9227 (pass_manager::register_pass): ...here.
9228 (init_optimization_passes): Convert into...
9229 (pass_manager::pass_manager): ...constructor for new
9230 pass_manager class, and initialize the pass_lists array.
9231 (check_profile_consistency): Update for moves of globals to
9232 fields of pass_manager.
9233 (dump_profile_report): Move body of implementation into...
9234 (pass_manager::dump_profile_report): ...here.
9235 (ipa_write_summaries_1): Update for moves of pass lists from
9236 being globals to fields of pass_manager.
9237 (ipa_write_optimization_summaries): Likewise.
9238 (ipa_read_summaries): Likewise.
9239 (ipa_read_optimization_summaries): Likewise.
9240 (execute_all_ipa_stmt_fixups): Likewise.
9242 * statistics.c (statistics_fini): Update for moves of globals to
9243 fields of pass_manager.
9245 * toplev.c (general_init): Replace call to
9246 init_optimization_passes with construction of the pass_manager
9249 * tree-pass.h (all_passes): Remove, in favor of a field of the
9250 same name within the new class pass_manager.
9251 (all_small_ipa_passes): Likewise.
9252 (all_lowering_passes): Likewise.
9253 (all_regular_ipa_passes): Likewise.
9254 (all_lto_gen_passes): Likewise.
9255 (all_late_ipa_passes): Likewise.
9256 (passes_by_id): Likewise.
9257 (passes_by_id_size): Likewise.
9258 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
9259 the new class pass_manager.
9260 (get_pass_for_id): Remove.
9262 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
9264 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
9267 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
9269 * arm.md (mulhi3): New expand pattern.
9271 2013-07-30 Jan Hubicka <jh@suse.cz>
9272 Martin Liska <marxin.liska@gmail.com>
9274 * profile.c (compute_value_histograms): Do not ICE when
9275 there is mismatch only on some counters.
9277 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
9279 PR rtl-optimization/57637
9280 * function.c (move_insn_for_shrink_wrap): Also check the
9281 GEN set of the LIVE problem for the liveness analysis
9282 if it exists, otherwise give up.
9284 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
9286 PR tree-optimization/57993
9287 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
9288 replaced statement in the candidate table.
9289 (phi_add_costs): Return infinite cost when the hidden basis does
9290 not dominate all phis on which the candidate is dependent.
9291 (replace_one_candidate): Record replaced statement in the
9294 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
9296 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
9297 (ashlv2si3): New expander.
9298 (*ashlv2si3_i): New define_insn_and_split.
9299 * predicates.md (float_operation): Allow patterns with three
9302 PR rtl-optimization/58021
9303 * mode-switching.c (create_pre_exit): Always split off preceding
9304 insns if we are not at the basic block head.
9306 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
9308 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
9309 (UCLIBC_DYNAMIC_LINKER): New macro.
9310 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
9312 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
9313 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
9315 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
9316 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
9317 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
9318 for SF and DF modes. Use ieee_quad_format for TF mode.
9319 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
9320 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
9321 (mips_option_override): Handle `-mnan=legacy'.
9322 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
9323 `-mabs=2008' and `-mnan=2008'.
9324 (OPTION_DEFAULT_SPECS): Add "nan" default.
9325 (ASM_SPEC): Handle `-mnan='.
9326 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
9327 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
9328 comment accordingly.
9329 (neg<mode>2): Likewise.
9330 * config/mips/mips.opt (mabs, mnan): New options.
9331 * doc/install.texi (Configuration): Document `--with-nan=' option.
9332 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
9334 (MIPS Options): Document them.
9335 * config.gcc <mips*-*-*>: Handle `--with-nan='.
9336 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
9337 * configure: Regenerate.
9338 * config.in: Regenerate.
9340 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
9342 * config/i386/i386.md (float post-reload splitters): Do not check
9343 for subregs of SSE registers.
9345 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
9346 H.J. Lu <hongjiu.lu@intel.com>
9350 * config/i386/i386.md (post-reload splitter
9351 to avoid partial SSE reg dependency stalls): New pattern.
9353 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
9355 * config/s390/s390.md ("movcc"): Swap load and store instructions.
9357 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
9359 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
9360 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
9362 2013-07-26 Cary Coutant <ccoutant@google.com>
9364 * dwarf2out.c (die_checksum_ordered): Don't include template
9365 instantiations in signature.
9366 (is_template_parameter): New function.
9367 (is_template_instantiation): New function.
9368 (generate_skeleton_bottom_up): Don't include template instantiations
9370 (generate_skeleton): Likewise.
9371 (break_out_comdat_types): Move recursive call to break out nested
9373 (prune_unused_types_mark_generic_parms_dies): Call
9374 is_template_parameter.
9376 2013-07-26 Ian Bolton <ian.bolton@arm.com>
9378 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
9379 uses vector registers.
9380 * config/aarch64/iterators.md: Add attributes rtn and vas.
9382 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9383 Richard Earnshaw <richard.earnshaw@arm.com>
9385 * combine.c (simplify_comparison): Re-canonicalize operands
9387 * config/arm/arm.md (movcond_addsi): New splitter.
9389 2013-07-25 Sterling Augustine <saugustine@google.com>
9391 * dwarf2out.c (size_of_pubnames): Move code to...
9392 (include_pubname_in_output): ...here. New.
9393 (want_pubnames): Rearrange.
9394 (output_pubnames): Call include_pubname_in_output. Move assertion.
9396 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
9398 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
9400 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
9403 * doc/extend.texi: Remove obsolete builtins. Fix
9404 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
9406 2013-07-25 Jan Hubicka <jh@suse.cz>
9408 * cgraph.c (release_function_body): Break out from ...
9409 (cgraph_release_function_body): ... this one; also release DECL_RESULT
9411 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
9412 old_tree in the map.
9413 (create_specialized_node): Update.
9414 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
9416 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
9417 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
9418 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
9419 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
9422 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9424 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
9425 addsi3_carryin_alt2_<optab>): Correct output template.
9427 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9429 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
9430 Adjust for arm_restrict_it.
9431 Remove trailing whitespace.
9433 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
9435 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
9436 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
9438 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
9440 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
9442 PR rtl-optimization/57960
9443 * lra-constraints.c (process_alt_operands): Use the right mode
9444 when checking strict_low.
9446 2013-07-25 Jan Hubicka <jh@suse.cz>
9448 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
9449 * cgraph.c (cgraph_remove_node): Do not release function body
9450 when in cgraph streaming.
9451 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
9452 in other partitions are not considered reachable; fix handling of
9455 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
9457 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
9459 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
9464 * config/arm/arm.md ("*sibcall_insn): Replace use of
9465 Ss with US. Adjust output for v5 and v4t.
9466 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
9468 * config/arm/constraints.md ("Ss"): Rename to US.
9470 2013-07-25 Terry Guo <terry.guo@arm.com>
9472 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
9473 shift_add/shift_sub0/shift_sub1 RTXs.
9475 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
9476 Anton Blanchard <anton@au1.ibm.com>
9478 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
9479 (altivec_vpks<VI_char>ss): Likewise.
9480 (altivec_vpks<VI_char>us): Likewise.
9481 (altivec_vpku<VI_char>us): Likewise.
9482 (altivec_vpku<VI_char>um): Likewise.
9484 2013-07-24 David Malcolm <dmalcolm@redhat.com>
9486 Introduce context class.
9488 * Makefile.in (CONTEXT_H): New.
9489 (OBJS): Add context.o.
9490 (toplev.o): Add CONTEXT_H to dependencies.
9493 * toplev.c (general_init): Create the singleton gcc::context instance.
9499 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
9501 PR rtl-optimization/57968
9502 * mode-switching.c (create_pre_exit): Allow instructions that
9503 don't set a return register to need a non-exit mode.
9505 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
9506 Anton Blanchard <anton@au1.ibm.com>
9508 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
9509 operands to vperm for little endian.
9510 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
9511 of lvsl to create the control mask for a vperm for little endian.
9513 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9514 Anton Blanchard <anton@au1.ibm.com>
9516 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
9517 two operands for little-endian.
9519 2013-07-23 Steve Ellcey <sellcey@mips.com>
9521 * config/mips/mips.c (mips_case_values_threshold): New.
9522 (TARGET_CASE_VALUES_THRESHOLD): Define.
9524 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9525 Anton Blanchard <anton@au1.ibm.com>
9527 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
9528 selection of field for vector splat in little endian mode.
9530 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
9532 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
9533 expanders to rs6000.md.
9534 (ior<mode>3): Likewise.
9535 (and<mode>3): Likewise.
9536 (one_cmpl<mode>2): Likewise.
9537 (nor<mode>3): Likewise.
9538 (andc<mode>3): Likewise.
9539 (eqv<mode>3): Likewise.
9540 (nand<mode>3): Likewise.
9541 (orc<mode>3): Likewise.
9543 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
9546 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
9547 to split multi-word logical operations.
9548 (rs6000_split_logical_di): Likewise.
9549 (rs6000_split_logical): Likewise.
9551 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
9552 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
9553 and allow TImode operations in 32-bit.
9554 (vsx_and<mode>3_64bit): Likewise.
9555 (vsx_ior<mode>3_32bit): Likewise.
9556 (vsx_ior<mode>3_64bit): Likewise.
9557 (vsx_xor<mode>3_32bit): Likewise.
9558 (vsx_xor<mode>3_64bit): Likewise.
9559 (vsx_one_cmpl<mode>2_32bit): Likewise.
9560 (vsx_one_cmpl<mode>2_64bit): Likewise.
9561 (vsx_nor<mode>3_32bit): Likewise.
9562 (vsx_nor<mode>3_64bit): Likewise.
9563 (vsx_andc<mode>3_32bit): Likewise.
9564 (vsx_andc<mode>3_64bit): Likewise.
9565 (vsx_eqv<mode>3_32bit): Likewise.
9566 (vsx_eqv<mode>3_64bit): Likewise.
9567 (vsx_nand<mode>3_32bit): Likewise.
9568 (vsx_nand<mode>3_64bit): Likewise.
9569 (vsx_orc<mode>3_32bit): Likewise.
9570 (vsx_orc<mode>3_64bit): Likewise.
9572 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
9573 logical types in GPRs.
9575 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
9576 logical insns to rs6000.md, and allow TImode operations in
9578 (altivec_ior<mode>3): Likewise.
9579 (altivec_xor<mode>3): Likewise.
9580 (altivec_one_cmpl<mode>2): Likewise.
9581 (altivec_nor<mode>3): Likewise.
9582 (altivec_andc<mode>3): Likewise.
9584 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
9585 attributes for moving the 128-bit logical operations into
9587 (BOOL_REGS_OUTPUT): Likewise.
9588 (BOOL_REGS_OP1): Likewise.
9589 (BOOL_REGS_OP2): Likewise.
9590 (BOOL_REGS_UNARY): Likewise.
9591 (BOOL_REGS_AND_CR0): Likewise.
9592 (one_cmpl<mode>2): Add support for DI logical operations on
9593 32-bit, splitting the operations to 32-bit.
9597 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
9598 changes to combine the 32/64-bit code, allow logical operations on
9599 TI mode in 32-bit, and to use similar match_operator patterns like
9600 scalar mode uses. Combine the Altivec and VSX code for logical
9601 operations, and move it here.
9602 (ior<mode>3, 128-bit types): Likewise.
9603 (xor<mode>3, 128-bit types): Likewise.
9604 (one_cmpl<mode>3, 128-bit types): Likewise.
9605 (nor<mode>3, 128-bit types): Likewise.
9606 (andc<mode>3, 128-bit types): Likewise.
9607 (eqv<mode>3, 128-bit types): Likewise.
9608 (nand<mode>3, 128-bit types): Likewise.
9609 (orc<mode>3, 128-bit types): Likewise.
9610 (and<mode>3_internal): Likewise.
9611 (bool<mode>3_internal): Likewise.
9612 (boolc<mode>3_internal1): Likewise.
9613 (boolc<mode>3_internal2): Likewise.
9614 (boolcc<mode>3_internal1): Likewise.
9615 (boolcc<mode>3_internal2): Likewise.
9616 (eqv<mode>3_internal1): Likewise.
9617 (eqv<mode>3_internal2): Likewise.
9618 (one_cmpl1<mode>3_internal): Likewise.
9620 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
9622 * config/microblaze/microblaze.c (microblaze_expand_prologue):
9623 Rename flag_stack_usage to flag_stack_usage_info.
9625 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
9627 * config/microblaze/sync.md: New file.
9628 * config/microblaze/microblaze.md: Include sync.md
9629 * config/microblaze/microblaze.c: Add print_operand 'y'.
9630 * config/microblaze/constraints.md: Add memory_contraint
9631 'Q' which is a single register.
9633 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
9635 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
9637 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
9639 * reload.c (find_reloads): Exit loop once we find this operand
9640 cannot be reloaded somehow for this alternative.
9642 * reload.c (find_reloads): Exit loop once we find a hard register.
9644 * rtlanal.c (computed_jump_p): Exit loop once we find label
9647 * i386.c (ix86_pad_returns): Exit loop after setting replace.
9649 * cfgloopmanip.c (remove_path): Exit loop after setting
9652 * gensupport.c (subst_dup): Avoid loop if code is not
9653 MATCH_DUP nor MATCH_OP_DUP.
9655 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
9657 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
9659 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
9661 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
9662 true for SP_REGNUM if mode == ptr_mode.
9663 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
9664 with value R0_REGNUM + 31.
9666 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
9668 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
9669 pad pointer-typed argument downward.
9671 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
9673 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
9674 and __ILP32__ when the ILP32 model is in use.
9676 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
9678 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
9679 (aarch64_load_symref_appropriately): In the case of
9680 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
9681 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
9682 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
9683 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
9684 if 'mode' doesn't equal to 'ptr_mode'.
9685 (aarch64_output_mi_thunk): Add an assertion on the alignment of
9686 'vcall_offset'; change to call aarch64_emit_move differently depending
9687 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
9688 to calculate the upper bound of 'vcall_offset'.
9689 (aarch64_cannot_force_const_mem): Change to also return true if
9691 (aarch64_legitimize_reload_address): In the case of large
9692 displacements, add new local variable 'xmode' and an assertion
9693 based on it; change to use 'xmode' to generate the new rtx and
9695 (aarch64_asm_trampoline_template): Change to generate the template
9696 differently depending on TARGET_ILP32 or not; change to use
9697 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
9698 (aarch64_trampoline_size): Removed.
9699 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
9700 and replace immediate literals with it. Change to use 'ptr_mode'
9701 instead of 'DImode' and call convert_memory_address if the mode
9702 of 'fnaddr' doesn't equal to 'ptr_mode'.
9703 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
9705 (aarch64_elf_asm_destructor): Likewise.
9706 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
9707 on TARGET_ILP32 instead of aarch64_trampoline_size.
9708 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
9709 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
9710 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
9711 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
9712 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
9713 (storewb_pair<GPI:mode>_<P:mode>): ... this.
9714 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
9715 depending on the value of 'mode'.
9716 (add_losym_<mode>): New.
9717 (ldr_got_small_<mode>): New, based on ldr_got_small.
9718 (ldr_got_small): Remove.
9719 (ldr_got_small_sidi): New.
9720 * config/aarch64/iterators.md (P): New.
9721 (PTR): Change to 'ptr_mode' in the condition.
9723 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
9725 * config.gcc (aarch64*-*-*): Support --with-abi.
9726 (aarch64*-*-elf): Support --with-multilib-list.
9727 (aarch64*-*-linux*): Likewise.
9728 (supported_defaults): Add abi to aarch64*-*-*.
9729 * configure.ac: Mention AArch64 for --with-multilib-list.
9730 * configure: Re-generated.
9731 * config/aarch64/biarchilp32.h: New file.
9732 * config/aarch64/biarchlp64.h: New file.
9733 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
9735 (MULTILIB_DEFAULTS): Ditto.
9736 (DRIVER_SELF_SPECS): Ditto.
9737 (ASM_SPEC): Update to also substitute -mabi.
9738 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
9739 file whose name depends on -mabi= and -mbig-endian.
9740 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
9742 (POINTER_SIZE): New define.
9743 (POINTERS_EXTEND_UNSIGNED): Ditto.
9744 (enum aarch64_abi_type): New enumeration tag.
9745 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
9746 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
9747 (TARGET_ILP32): New define.
9748 * config/aarch64/aarch64.opt (mabi): New.
9750 (ilp32, lp64): New values for -mabi.
9751 * config/aarch64/t-aarch64 (comma): New define.
9752 (MULTILIB_OPTIONS): Ditto.
9753 (MULTILIB_DIRNAMES): Ditto.
9754 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
9755 * doc/invoke.texi: Document -mabi for AArch64.
9757 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
9759 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
9761 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
9762 Anton Blanchard <anton@au1.ibm.com>
9764 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
9765 endianness when selecting field to splat.
9767 2013-07-22 Eric Christopher <echristo@gmail.com>
9769 * dwarf2out.c (die_odr_checksum): New function to use
9770 CHECKSUM_ macros and ULEB128 for DIE tag.
9771 (generate_type_signature): Use.
9773 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
9775 * config.gcc (sparc*-*-*): Accept leon3 processor.
9776 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
9777 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
9778 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
9779 * config/sparc/sparc.opt (enum processor_type): Add leon3.
9780 (mfix-ut699): Adjust comment.
9781 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
9782 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
9783 (CPP_CPU_SPEC): Likewise.
9784 (ASM_CPU_SPEC): Likewise.
9785 * config/sparc/sparc.c (leon3_cost): New constant.
9786 (sparc_option_override): Add leon3 support.
9787 (mem_ref): New function.
9788 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
9789 (sparc_do_work_around_errata): Look into the instruction in the delay
9790 slot and adjust accordingly. Add fix for the data cache nullify issues
9791 of the UT699. Change insertion position for the NOP.
9792 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
9793 (leon3_load): New reservation.
9794 (leon_store): Bump latency to 2.
9795 (grfpu): New automaton.
9796 (grfpu_alu): New unit.
9797 (grfpu_ds): Likewise.
9798 (leon_fp_alu): Adjust.
9799 (leon_fp_mult): Delete.
9800 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
9801 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
9802 * config/sparc/sparc.md (cpu): Add leon3.
9803 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
9805 (atomic_test_and_set): Likewise.
9808 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
9810 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
9811 default for R5900 targets.
9812 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
9813 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
9814 * config/mips/mips.c (mips_option_override): Report an error for
9815 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
9816 for -march=r5900 -mhard-float.
9818 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
9820 * df-problems.c (can_move_insns_across): Exit loop once we
9821 find a non-fixed, non-global register.
9823 * ipa-pure-const.c (propagate_nothrow): Exit loop after
9826 * omega.c (omega_eliminate_red): Break after setting red_found.
9827 (omega_problem_has_red_equations): Similarly after setting found.
9828 (omega_query_variable): Similarly after setting coupled.
9830 2013-07-22 Marek Polacek <polacek@redhat.com>
9832 * gimplify.c: Don't include gimple.h twice.
9834 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9836 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
9837 instead of TARGET_THUMB1.
9838 (Pz): New constraint.
9839 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
9841 (compare_negsi_si): Likewise.
9842 (compare_addsi2_op0): Likewise.
9843 (compare_addsi2_op1): Likewise.
9844 (addsi3_carryin_<optab>): Likewise.
9845 (addsi3_carryin_alt2_<optab>): Likewise.
9846 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
9847 for arm_restrict_it.
9848 (subsi3_carryin): Likewise.
9849 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
9850 (minmax_arithsi): Disable for arm_restrict_it.
9851 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
9852 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
9853 (satsi_<SAT:code>_shift): Likewise.
9854 (arm_shiftsi3): Add alternative for 16-bit encoding.
9855 (arm32_movhf): Disable for arm_restrict_it.
9856 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
9857 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
9859 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
9861 * config/arm/arm.md (attribute "insn"): Delete.
9862 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
9863 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
9864 (not_shiftsi): Update for attribute change.
9865 (not_shiftsi_compare0): Likewise.
9866 (not_shiftsi_compare0_scratch): Likewise.
9867 (arm_one_cmplsi2): Likewise.
9868 (thumb1_one_cmplsi2): Likewise.
9869 (notsi_compare0): Likewise.
9870 (notsi_compare0_scratch): Likewise.
9871 (thumb1_movdi_insn): Likewise.
9872 (arm_movsi_insn): Likewise.
9873 (movhi_insn_arch4): Likewise.
9874 (movhi_bytes): Likewise.
9875 (arm_movqi_insn): Likewise.
9876 (thumb1_movqi_insn): Likewise.
9877 (arm32_movhf): Likewise.
9878 (thumb1_movhf): Likewise.
9879 (arm_movsf_soft_insn): Likewise.
9880 (thumb1_movsf_insn): Likewise.
9881 (thumb_movdf_insn): Likewise.
9882 (movsicc_insn): Likewise.
9883 (movsfcc_soft_insn): Likewise.
9884 (and_scc): Likewise.
9885 (cond_move): Likewise.
9886 (if_move_not): Likewise.
9887 (if_not_move): Likewise.
9888 (if_shift_move): Likewise.
9889 (if_move_shift): Likewise.
9890 (if_shift_shift): Likewise.
9891 (if_not_arith): Likewise.
9892 (if_arith_not): Likewise.
9893 (cond_move_not): Likewise.
9894 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
9895 (neon_mov<mode>): Likewise.
9896 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
9897 (thumb2_movsi_vfp): Likewise.
9898 (movsf_vfp): Likewise.
9899 (thumb2_movsf_vfp): Likewise.
9900 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
9902 (cortexa7_older_only): Likewise.
9903 (cortexa7_younger): Likewise.
9904 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
9905 (1020alu_shift_op): Likewise.
9906 (1020alu_shift_reg_op): Likewise.
9907 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
9908 (alu_shift_op): Likewise.
9909 (alu_shift_reg_op): Likewise.
9910 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
9911 (11_alu_shift_op): Likewise.
9912 (11_alu_shift_reg_op): Likewise.
9913 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
9914 (9_alu_shift_reg_op): Likewise.
9915 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
9917 (cortex_a15_alu_shift): Likewise.
9918 (cortex_a15_alu_shift_reg): Likewise.
9919 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
9920 (cortex_a5_alu_shift): Likewise.
9921 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
9923 (cortex_a53_alu_shift): Likewise.
9924 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
9926 (cortex_a7_alu_reg): Likewise.
9927 (cortex_a7_alu_shift): Likewise.
9928 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
9929 (cortex_a8_alu_shift): Likewise.
9930 (cortex_a8_alu_shift_reg): Likewise.
9931 (cortex_a8_mov): Likewise.
9932 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
9933 (cortex_a9_dp_shift): Likewise.
9934 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
9935 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
9936 (cortex_r4_mov): Likewise.
9937 (cortex_r4_alu_shift): Likewise.
9938 (cortex_r4_alu_shift_reg): Likewise.
9939 * config/arm/fa526.md (526_alu_op): Update for attribute change.
9940 (526_alu_shift_op): Likewise.
9941 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
9942 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
9943 (626te_alu_shift_op): Likewise.
9944 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
9945 (726te_alu_op): Likewise.
9946 (726te_alu_shift_op): Likewise.
9947 (726te_alu_shift_reg_op): Likewise.
9948 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
9949 (mp626_alu_shift_op): Likewise.
9950 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
9951 (pj4_alu_e1_conds): Likewise.
9952 (pj4_alu): Likewise.
9953 (pj4_alu_conds): Likewise.
9954 (pj4_shift): Likewise.
9955 (pj4_shift_conds): Likewise.
9956 (pj4_alu_shift): Likewise.
9957 (pj4_alu_shift_conds): Likewise.
9959 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9961 * config/arm/predicates.md (shiftable_operator_strict_it):
9963 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
9964 Disable cond_exec version for arm_restrict_it.
9965 (thumb2_smaxsi3): Convert to generate cond_exec.
9966 (thumb2_sminsi3): Likewise.
9967 (thumb32_umaxsi3): Likewise.
9968 (thumb2_uminsi3): Likewise.
9969 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
9970 (thumb2_neg_abssi2): Likewise.
9971 (thumb2_mov_scc): Add alternative for 16-bit encoding.
9972 (thumb2_movsicc_insn): Adjust alternatives.
9973 (thumb2_mov_negscc): Disable for arm_restrict_it.
9974 (thumb2_mov_negscc_strict_it): New pattern.
9975 (thumb2_mov_notscc_strict_it): New pattern.
9976 (thumb2_mov_notscc): Disable for arm_restrict_it.
9977 (thumb2_ior_scc): Likewise.
9978 (thumb2_ior_scc_strict_it): New pattern.
9979 (thumb2_cond_move): Adjust for arm_restrict_it.
9980 (thumb2_cond_arith): Disable for arm_restrict_it.
9981 (thumb2_cond_arith_strict_it): New pattern.
9982 (thumb2_cond_sub): Adjust for arm_restrict_it.
9983 (thumb2_movcond): Likewise.
9984 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
9985 (thumb2_zero_extendhisi2_v6): Likewise.
9986 (thumb2_zero_extendqisi2_v6): Likewise.
9987 (orsi_notsi_si): Likewise.
9988 (orsi_not_shiftsi_si): Likewise.
9990 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
9992 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
9993 instruction sequence is 1 byte shorter.
9995 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
9997 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
9998 it is not needed after split.
10000 2013-07-20 Iain Sandoe <iain@codesourcery.com>
10003 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
10004 second label for nonlocal goto receivers. Don't output pic base labels
10005 unless we're producing PIC; mark that action unreachable().
10006 (ix86_save_reg): If the function contains a nonlocal label, save the
10008 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
10009 * config/darwin.c (emitted_pic_label_num): New GTY.
10010 (update_pic_label_number_if_needed): New.
10011 (machopic_output_function_base_name): Adjust for nonlocal receiver
10013 (machopic_should_output_picbase_label): New.
10014 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
10015 (nonlocal_goto_receiver): New insn and split.
10017 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
10019 * config/aarch64/aarch64-builtins.c
10020 (aarch64_fold_builtin): Fold abs in all modes.
10021 * config/aarch64/aarch64-simd-builtins.def
10022 (abs): Enable for all modes.
10023 * config/aarch64/arm_neon.h
10024 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
10025 (vabs_f64): Add missing intrinsic.
10027 2013-07-19 Ian Bolton <ian.bolton@arm.com>
10029 * config/aarch64/arm_neon.h (vabs_s64): New function
10031 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
10034 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
10035 * config/avr/avr.md (adjust_len): Add `round'.
10036 * config/avr/avr-protos.h (avr_out_round): New prototype.
10037 (avr_out_plus): Add `out_label' argument.
10038 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
10039 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
10040 Handle the case where `insn' is just a pattern.
10041 (avr_out_bitop): Handle the case where `insn' is just a pattern.
10042 (avr_out_round): New function.
10043 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
10045 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
10047 * config/microblaze/microblaze.c (microblaze_expand_prologue):
10048 Add check for flag_stack_usage to handle -fstack-usage support
10050 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
10052 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
10053 interaction for new Power8 flags and VSX.
10055 2013-07-18 Sriraman Tallam <tmsriram@google.com>
10057 PR middle-end/57698
10058 * tree-inline.c (expand_call_inline): Emit errors during
10059 early_inlining only if optimization is not turned on.
10061 2013-07-18 David Malcolm <dmalcolm@redhat.com>
10065 * passes.c (init_optimization_passes): Move the construction of
10066 the pass hierarchy into a new passes.def file.
10068 * Makefile.in (passes.o): Add dependency on passes.def.
10070 2013-07-18 David Malcolm <dmalcolm@redhat.com>
10072 * passes.c (init_optimization_passes): Introduce macros for
10073 constructing the tree of passes (INSERT_PASSES_AFTER,
10074 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
10075 TERMINATE_PASS_LIST).
10077 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
10078 Wei Mi <wmi@google.com>
10080 PR rtl-optimization/57878
10081 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
10083 (reload_pseudo_compare_func): Check nregs first for reload
10086 2013-07-18 David Malcolm <dmalcolm@redhat.com>
10088 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
10090 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
10092 * read-rtl.c (validate_const_int): Once an invalid character is
10093 seen, quit the loop.
10095 * gengtype.c (write_roots): Similarly once we find the "deletable"
10096 or "if_marked" option.
10098 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
10100 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
10101 "xtab" and "sat". Move value "clz" from here to ...
10102 (attriubte "type"): ... here.
10103 (satsi_<SAT:code>): Delete "insn" attribute.
10104 (satsi_<SAT:code>_shift): Likewise.
10105 (arm_zero_extendqisi2addsi): Likewise.
10106 (arm_extendqisi2addsi): Likewise.
10107 (clzsi2): Update for attribute changes.
10108 (rbitsi2): Likewise.
10109 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
10111 (arm_usatsihi): Likewise.
10112 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
10114 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
10116 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
10117 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
10118 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
10119 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
10120 in alphabetical order.
10121 (attribute "core_cycles"): Update for attribute changes.
10122 (arm_addsi3): Likewise.
10123 (addsi3_compare0): Likewise.
10124 (addsi3_compare0_scratch): Likewise.
10125 (addsi3_compare_op1): Likewise.
10126 (addsi3_compare_op2): Likewise.
10127 (compare_addsi2_op0): Likewise.
10128 (compare_addsi2_op1): Likewise.
10129 (addsi3_carryin_shift_<optab>): Likewise.
10130 (subsi3_carryin_shift): Likewise.
10131 (rsbsi3_carryin_shift): Likewise.
10132 (arm_subsi3_insn): Likewise.
10133 (subsi3_compare0): Likewise.
10134 (subsi3_compare): Likewise.
10135 (arm_andsi3_insn): Likewise.
10136 (thumb1_andsi3_insn): Likewise.
10137 (andsi3_compare0): Likewise.
10138 (andsi3_compare0_scratch): Likewise.
10139 (zeroextractsi_compare0_scratch
10140 (andsi_not_shiftsi_si): Likewise.
10141 (iorsi3_insn): Likewise.
10142 (iorsi3_compare0): Likewise.
10143 (iorsi3_compare0_scratch): Likewise.
10144 (arm_xorsi3): Likewise.
10145 (thumb1_xorsi3_insn): Likewise.
10146 (xorsi3_compare0): Likewise.
10147 (xorsi3_compare0_scratch): Likewise.
10148 (satsi_<SAT:code>_shift): Likewise.
10150 (arm_shiftsi3): Likewise.
10151 (shiftsi3_compare0): Likewise.
10152 (not_shiftsi): Likewise.
10153 (not_shiftsi_compare0): Likewise.
10154 (not_shiftsi_compare0_scratch): Likewise.
10155 (arm_one_cmplsi2): Likewise.
10156 (thumb_one_complsi2): Likewise.
10157 (notsi_compare0): Likewise.
10158 (notsi_compare0_scratch): Likewise.
10159 (thumb1_zero_extendhisi2): Likewise.
10160 (arm_zero_extendhisi2): Likewise.
10161 (arm_zero_extendhisi2_v6): Likewise.
10162 (arm_zero_extendhisi2addsi): Likewise.
10163 (thumb1_zero_extendqisi2): Likewise.
10164 (thumb1_zero_extendqisi2_v6): Likewise.
10165 (arm_zero_extendqisi2): Likewise.
10166 (arm_zero_extendqisi2_v6): Likewise.
10167 (arm_zero_extendqisi2addsi): Likewise.
10168 (thumb1_extendhisi2): Likewise.
10169 (arm_extendhisi2): Likewise.
10170 (arm_extendhisi2_v6): Likewise.
10171 (arm_extendqisi): Likewise.
10172 (arm_extendqisi_v6): Likewise.
10173 (arm_extendqisi2addsi): Likewise.
10174 (thumb1_extendqisi2): Likewise.
10175 (thumb1_movdi_insn): Likewise.
10176 (arm_movsi_insn): Likewise.
10177 (movsi_compare0): Likewise.
10178 (movhi_insn_arch4): Likewise.
10179 (movhi_bytes): Likewise.
10180 (arm_movqi_insn): Likewise.
10181 (thumb1_movqi_insn): Likewise.
10182 (arm32_movhf): Likewise.
10183 (thumb1_movhf): Likewise.
10184 (arm_movsf_soft_insn): Likewise.
10185 (thumb1_movsf_insn): Likewise.
10186 (movdf_soft_insn): Likewise.
10187 (thumb_movdf_insn): Likewise.
10188 (arm_cmpsi_insn): Likewise.
10189 (cmpsi_shiftsi): Likewise.
10190 (cmpsi_shiftsi_swp): Likewise.
10191 (arm_cmpsi_negshiftsi_si): Likewise.
10192 (movsicc_insn): Likewise.
10193 (movsfcc_soft_insn): Likewise.
10194 (arith_shiftsi): Likewise.
10195 (arith_shiftsi_compare0
10196 (arith_shiftsi_compare0_scratch
10197 (sub_shiftsi): Likewise.
10198 (sub_shiftsi_compare0
10199 (sub_shiftsi_compare0_scratch
10200 (and_scc): Likewise.
10201 (cond_move): Likewise.
10202 (if_plus_move): Likewise.
10203 (if_move_plus): Likewise.
10204 (if_move_not): Likewise.
10205 (if_not_move): Likewise.
10206 (if_shift_move): Likewise.
10207 (if_move_shift): Likewise.
10208 (if_shift_shift): Likewise.
10209 (if_not_arith): Likewise.
10210 (if_arith_not): Likewise.
10211 (cond_move_not): Likewise.
10212 (thumb1_ashlsi3): Set type attribute.
10213 (thumb1_ashrsi3): Likewise.
10214 (thumb1_lshrsi3): Likewise.
10215 (thumb1_rotrsi3): Likewise.
10216 (shiftsi3_compare0_scratch): Likewise.
10217 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
10218 (neon_mov<mode>): Likewise.
10219 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
10221 (thumb2_movsi_insn): Likewise.
10222 (thumb2_cmpsi_neg_shiftsi): Likewise.
10223 (thumb2_extendqisi_v6): Likewise.
10224 (thumb2_zero_extendhisi2_v6): Likewise.
10225 (thumb2_zero_extendqisi2_v6): Likewise.
10226 (thumb2_shiftsi3_short): Likewise.
10227 (thumb2_addsi3_compare0_scratch): Likewise.
10228 (orsi_not_shiftsi_si): Likewise.
10229 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
10230 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
10232 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
10233 (1020alu_shift_op): Likewise.
10234 (1020alu_shift_reg_op): Likewise.
10235 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
10236 (alu_shift_op): Likewise.
10237 (alu_shift_reg_op): Likewise.
10238 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
10239 (11_alu_shift_op): Likewise.
10240 (11_alu_shift_reg_op): Likewise.
10241 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
10242 (9_alu_shift_reg_op): Likewise.
10243 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
10245 (cortex_a15_alu_shift): Likewise.
10246 (cortex_a15_alu_shift_reg): Likewise.
10247 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
10249 (cortex_a5_alu_shift): Likewise.
10250 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
10252 (cortex_a53_alu_shift): Likewise.
10253 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
10255 (cortex_a7_alu_reg): Likewise.
10256 (cortex_a7_alu_shift): Likewise.
10257 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
10259 (cortex_a8_alu_shift): Likewise.
10260 (cortex_a8_alu_shift_reg): Likewise.
10261 (cortex_a8_mov): Likewise.
10262 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
10263 (cortex_a9_dp_shift): Likewise.
10264 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
10266 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
10268 (cortex_r4_mov): Likewise.
10269 (cortex_r4_alu_shift): Likewise.
10270 (cortex_r4_alu_shift_reg): Likewise.
10271 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
10272 (526_alu_shift_op): Likewise.
10273 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
10274 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
10275 (626te_alu_shift_op): Likewise.
10276 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
10277 (726te_alu_op): Likewise.
10278 (726te_alu_shift_op): Likewise.
10279 (726te_alu_shift_reg_op): Likewise.
10280 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
10281 (mp626_alu_shift_op): Likewise.
10282 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
10283 (pj4_alu_e1_conds): Likewise.
10284 (pj4_alu): Likewise.
10285 (pj4_alu_conds): Likewise.
10286 (pj4_shift): Likewise.
10287 (pj4_shift_conds): Likewise.
10288 (pj4_alu_shift): Likewise.
10289 (pj4_alu_shift_conds): Likewise.
10290 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
10292 (cortexa7_older_only): Likewise.
10293 (cortexa7_younger): Likewise.
10295 2013-07-18 David Malcolm <dmalcolm@redhat.com>
10297 * ipa-pure-const.c (generate_summary): Rename to...
10298 (pure_const_generate_summary): ... this.
10300 2013-07-17 Iain Sandoe <iain@codesourcery.com>
10302 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
10304 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
10307 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
10310 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10312 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
10313 enabled without -march=zEC12.
10314 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
10317 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
10319 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
10320 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
10321 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
10322 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
10323 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
10324 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
10325 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
10327 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
10328 and ISA_HAS_NMADD3_NMSUB3.
10329 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
10330 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
10331 (nmsub4<mode>, nmsub3<mode>): Likewise.
10332 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
10334 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
10336 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
10337 TARGET_MIPS5400 checking.
10339 2013-07-16 Jakub Jelinek <jakub@redhat.com>
10340 Peter Bergner <bergner@vnet.ibm.com>
10342 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
10343 registers in the comment.
10344 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
10345 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
10346 rather than FIRST_PSEUDO_REGISTERS.
10348 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
10350 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
10351 enable extra ISA flags with TARGET_HTM.
10353 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
10355 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
10358 2013-07-15 Cong Hou <congh@google.com>
10360 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
10361 in compare function for sorting.
10363 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
10365 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
10366 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
10367 * config/rs6000/rs6000.opt: Add -mhtm option.
10368 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
10369 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
10370 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
10371 __HTM__ if the HTM instructions are available.
10372 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
10373 htm_spr_reg_operand): New define_predicates.
10374 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
10375 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
10377 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
10378 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
10379 HTM builtin functions.
10380 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
10381 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
10382 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
10383 (rs6000_builtin_mask_calculate): Likewise.
10384 (rs6000_option_override_internal): Likewise.
10385 (bdesc_htm): Add new HTM builtin support.
10386 (htm_spr_num): New function.
10387 (htm_spr_regno): Likewise.
10388 (rs6000_htm_spr_icode): Likewise.
10389 (htm_expand_builtin): Likewise.
10390 (htm_init_builtins): Likewise.
10391 (rs6000_expand_builtin): Add support for HTM builtin functions.
10392 (rs6000_init_builtins): Likewise.
10393 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
10395 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
10396 (TARGET_HTM, MASK_HTM): Define macros.
10397 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
10398 (FIXED_REGISTERS): Likewise.
10399 (CALL_USED_REGISTERS): Likewise.
10400 (CALL_REALLY_USED_REGISTERS): Likewise.
10401 (REG_ALLOC_ORDER): Likewise.
10402 (enum reg_class): Likewise.
10403 (REG_CLASS_NAMES): Likewise.
10404 (REG_CLASS_CONTENTS): Likewise.
10405 (REGISTER_NAMES): Likewise.
10406 (ADDITIONAL_REGISTER_NAMES): Likewise.
10407 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
10408 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
10409 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
10410 * config/rs6000/htm.md: New file.
10411 * config/rs6000/htmintrin.h: New file.
10412 * config/rs6000/htmxlintrin.h: New file.
10414 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
10416 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
10417 Define SYMBOL_TINY_GOT, update comment.
10418 * config/aarch64/aarch64.c
10419 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
10420 (aarch64_expand_mov_immediate): Likewise.
10421 (aarch64_print_operand): Likewise.
10422 (aarch64_classify_symbol): Likewise.
10423 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
10424 (ldr_got_tiny): Define.
10426 2013-07-13 Tobias Grosser <tobias@grosser.es>
10428 PR tree-optimization/54094
10429 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
10430 scheduling dimension for the parallelism check from the polyhedral
10431 information in the AST.
10432 * graphite-dependences.c (carries_deps): Do not assume the schedule is
10435 2013-07-13 Jason Merrill <jason@redhat.com>
10437 * print-tree.c (debug_vec_tree): Use debug_raw.
10438 (debug_raw (vec<tree, va_gc> &)): New.
10439 (debug_raw (vec<tree, va_gc> *)): New.
10440 * tree.h: Declare them.
10442 2013-07-13 Bin Cheng <bin.cheng@arm.com>
10444 * ifcvt.c (ifcvt_after_combine): New static variable.
10445 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
10447 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
10448 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
10449 rest_of_handle_if_after_reload): Pass new argument for if_convert.
10451 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
10453 * config/mips/mips.c (mips_expand_call): Remove empty statement.
10455 2013-07-12 Michael Matz <matz@suse.de>
10457 PR middle-end/55771
10458 * convert.c (convert_to_real): Reject non-float inner types.
10460 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
10462 * config/aarch64/aarch64-protos.h
10463 (aarch64_simd_immediate_valid_for_move): Remove.
10464 * config/aarch64/aarch64.c (simd_immediate_info): New member.
10465 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
10467 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
10469 2013-07-11 Steve Ellcey <sellcey@mips.com>
10471 * config/mips/mips.c (mips_conditional_register_usage): Do not
10472 use t[0-7] registers in MIPS16 mode when optimizing for size.
10474 2013-07-11 Sriraman Tallam <tmsriram@google.com>
10476 * config/i386/i386.c (dispatch_function_versions): Fix array
10477 indexing of function_version_info to match actual_versions.
10479 2013-07-11 Teresa Johnson <tejohnson@google.com>
10481 * vec.h (struct va_gc): Move release out-of-line.
10482 (va_gc::release): Call ggc_free on released vec.
10484 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
10486 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
10487 Require GOT register as additional operand in UNSPEC.
10488 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
10489 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
10490 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
10491 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
10492 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
10493 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
10494 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
10496 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
10499 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
10500 name seen by assembler/linker rather if available.
10502 2013-07-11 Andreas Schwab <schwab@suse.de>
10504 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
10506 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
10508 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
10510 2013-07-10 Joseph Myers <joseph@codesourcery.com>
10512 * doc/tm.texi.in: Move hook documentation to ....
10513 * target.def: ... here.
10515 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
10516 text on @hook line.
10517 * doc/tm.texi: Regenerate.
10519 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
10522 * doc/invoke.texi: Document -Wconditionally-supported.
10524 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
10527 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
10530 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
10533 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
10534 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
10535 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
10536 Remove duplicate devices.
10537 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
10538 * config/avr/t-multilib: Regenerate.
10539 * config/avr/avr-tables.opt: Regenerate.
10540 * doc/avr-mmcu.texi: Regenerate.
10542 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
10545 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
10547 2013-07-10 Graham Stott <graham.stott@btinternet.com>
10549 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
10550 the cost of MULT when optimizing for size.
10552 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10554 * config/cr16/cr16-protos.h: Don't include target.h.
10556 2013-07-09 Joseph Myers <joseph@codesourcery.com>
10558 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
10559 adjust register size for TDmode and TFmode for VSX registers.
10561 2013-07-08 Kai Tietz <ktietz@redhat.com>
10564 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
10565 hook_bool_const_tree_true.
10567 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10569 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
10570 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
10571 * config/s390/s390.md: Define FPR*_REGNUM constants.
10572 Fix FPR2_REGNUM constant (18 -> 17).
10573 ("*trunc<BFP:mode><DFP_ALL:mode>2")
10574 ("*trunc<DFP_ALL:mode><BFP:mode>2")
10575 ("trunc<BFP:mode><DFP_ALL:mode>2")
10576 ("trunc<DFP_ALL:mode><BFP:mode>2")
10577 ("*extend<BFP:mode><DFP_ALL:mode>2")
10578 ("*extend<DFP_ALL:mode><BFP:mode>2")
10579 ("extend<BFP:mode><DFP_ALL:mode>2")
10580 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
10583 2013-07-08 Graham Stott <graham.stott@btinternet.com>
10585 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
10587 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10589 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
10590 and cfun_fpr_bit_p to cfun_fpr_save_p.
10591 (s390_frame_area, s390_register_info, s390_frame_info)
10592 (s390_emit_prologue, s390_emit_epilogue)
10593 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
10595 * config/s390/s390.h: Define *_REGNUM macros for floating point
10598 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
10600 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
10602 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
10604 PR rtl-optimization/57786
10605 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
10606 and break out of the loop when it is set to false.
10608 2013-07-08 Jakub Jelinek <jakub@redhat.com>
10611 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
10612 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
10613 (const_int 63)) 0)).
10614 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
10615 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
10616 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
10618 PR rtl-optimization/57829
10619 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
10620 mask bits outside of mode are just sign-extension from mode to HWI.
10622 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
10624 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
10625 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
10626 adjust_address instead of change_address to keep info about alignment.
10627 (emit_strmov): Remove.
10628 (emit_memmov): New function.
10629 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
10630 (expand_movmem_epilogue): Likewise and return updated rtx for
10632 (expand_constant_movmem_prologue): Likewise and return updated rtx for
10633 destination and source.
10634 (decide_alignment): Refactor, handle vector_loop.
10635 (ix86_expand_movmem): Likewise.
10636 (ix86_expand_setmem): Likewise.
10637 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
10639 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
10641 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
10642 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
10644 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
10646 * config/i386/sse.md (sse_movlhps): Change alternative 3
10647 of operand 2 to "m".
10649 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
10652 * config/i386/sse.md (iptr): New mode attribute.
10653 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
10654 (<sse>_vm<plusminus_insn><mode>3): Ditto.
10655 (<sse>_vmmul<mode>3): Ditto.
10656 (<sse>_vmdiv<mode>3): Ditto.
10657 (sse_vmrcpv4sf2): Ditto.
10658 (<sse>_vmsqrt<mode>2): Ditto.
10659 (sse_vmrsqrtv4sf2): Ditto.
10660 (<sse>_vm<code><mode>3): Ditto.
10661 (avx_vmcmp<mode>3): Ditto.
10662 (<sse>_vmmaskcmp<mode>3): Ditto.
10663 (<sse>_comi): Ditto.
10664 (<sse>_ucomi): Ditto.
10665 (*xop_vmfrcz_<mode>): Ditto.
10666 (*fmai_fmadd_<mode>): Ditto.
10667 (*fmai_fmsub_<mode>): Ditto.
10668 (*fmai_fnmadd_<mode>): Ditto.
10669 (*fmai_fnmsub_<mode>): Ditto.
10670 (*fma4i_vmfmadd_<mode>): Ditto.
10671 (*fma4i_vmfmsub_<mode>): Ditto.
10672 (*fma4i_vmfnmadd_<mode>): Ditto.
10673 (*fma4i_vmfnmsub_<mode>): Ditto.
10674 (*xop_vmfrcz_<mode>): Ditto.
10675 (sse_cvtps2pi): Ditto.
10676 (sse_cvttps2pi): Ditto.
10677 (sse_cvtss2si): Ditto.
10678 (sse_cvtss2si_2): Ditto.
10679 (sse_cvtss2siq_2): Ditto.
10680 (sse_cvttss2si): Ditto.
10681 (sse_cvttss2siq): Ditto.
10682 (sse_cvtsd2si): Ditto.
10683 (sse_cvtsd2si_2): Ditto.
10684 (sse_cvtsd2siq_2): Ditto.
10685 (sse_cvttsd2si): Ditto.
10686 (sse_cvttsd2siq): Ditto.
10687 (sse_cvtsd2ss): Ditto.
10688 (sse_cvtss2sd): Ditto.
10689 (avx2_pbroadcast<mode>): Ditto.
10690 (avx2_pbroadcast<mode>_1): Ditto.
10691 (*avx_vperm_broadcast_v4sf): Ditto.
10693 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
10694 (sse_movlhps): Ditto.
10695 (sse_storehps): Ditto.
10696 (sse_loadhps): Ditto.
10697 (sse_storelps): Ditto.
10698 (sse_loadlps): Ditto.
10699 (*vec_concatv4sf): Ditto.
10700 (*vec_interleave_highv2df): Ditto.
10701 (*vec_interleave_lowv2df): Ditto.
10702 (*vec_extractv2df_1_sse): Ditto.
10703 (*vec_extractv2df_0_sse): Ditto.
10704 (sse2_storelpd): Ditto.
10705 (sse2_loadlpd): Ditto.
10706 (sse2_movsd): Ditto.
10707 (*vec_concatv4si): Ditto.
10708 (vec_concatv2di): Ditto.
10710 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
10711 for Intel asm dialect.
10712 (mmx_punpcklwd): Ditto.
10713 (mmx_punpckldq): Ditto.
10715 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
10716 for intel assembler dialect.
10718 2013-07-06 Jakub Jelinek <jakub@redhat.com>
10721 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
10722 for BUILT_IN_C{LZ,LRSB}*.
10723 * tree.h (CASE_INT_FN): Add FN##IMAX case.
10724 * tree-vrp.c (extract_range_basic): Handle
10725 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
10726 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
10727 fall thru to code calling set_value*.
10728 * builtins.c (expand_builtin): Remove *IMAX cases.
10729 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
10730 if width is bigger than 2*HWI.
10732 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
10734 PR rtl-optimization/55342
10735 * lra-int.h (lra_subreg_reload_pseudos): New.
10736 * lra.c: Add undoing optional reloads to the block diagram.
10737 (lra_subreg_reload_pseudos): New.
10738 (lra_optional_reload_pseudos): Change comments.
10739 (lra): Init and clear lra_subreg_reload_pseudos. Clear
10740 lra_optional_reload_pseudos after undo transformations.
10741 * lra-assigns.c (pseudo_prefix_title): New.
10742 (lra_setup_reg_renumber): Use it.
10743 (spill_for): Ditto. Check subreg reload pseudos too.
10744 (assign_by_spills): Consider subreg reload pseudos too.
10745 * lra-constraints.c (simplify_operand_subreg): Use
10746 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
10747 (curr_insn_transform): Recognize and do optional reloads.
10748 (undo_optional_reloads): New.
10749 (lra_undo_inheritance): Call undo_optional_reloads.
10751 2013-07-05 Thomas Quinot <quinot@adacore.com>
10753 * tree-complex.c (expand_complex_operations_1): Fix typo.
10755 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
10757 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
10758 (tune_params): New member 'const vec_costs'.
10759 * config/aarch64/aarch64.c (generic_vector_cost): New.
10760 (generic_tunings): New member 'generic_vector_cost'.
10761 (aarch64_builtin_vectorization_cost): New.
10762 (aarch64_add_stmt_cost): New.
10763 (TARGET_VECTORIZE_ADD_STMT_COST): New.
10764 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
10766 2013-07-03 Jakub Jelinek <jakub@redhat.com>
10769 * config/i386/predicates.md (vsib_address_operand): Disallow
10770 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
10772 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
10774 PR middle-end/55030
10775 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
10776 expand_builtin_setjmp_receiver.
10777 (expand_label): Adjust, call expand_builtin_setjmp_receiver
10778 with NULL for the label parameter.
10779 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
10780 the frame-pointer. Adjust comments.
10781 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
10782 only if LABEL is non-NULL.
10784 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
10786 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
10787 (ARM_ABI_AAPCS64): Ditto.
10789 (ARM_DEFAULT_ABI): Ditto.
10791 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
10793 * config/aarch64/aarch64-builtins.c
10794 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
10795 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
10797 * config/aarch64/aarch64-simd.md
10798 (aarch64_ld1<VALL:mode>): New.
10799 (aarch64_st1<VALL:mode>): Likewise.
10800 * config/aarch64/arm_neon.h
10801 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
10803 2013-07-02 Sriraman Tallam <tmsriram@google.com>
10805 * config/i386/i386.c (gate_insert_vzeroupper): Check if
10807 (ix86_option_override_internal):Turn on all -mavx target flags by
10808 default as they are dependent on AVX anyway.
10810 2013-07-02 Cary Coutant <ccoutant@google.com>
10812 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
10813 deterministic hash.
10814 (loc_checksum_ordered): Likewise.
10815 (hash_loc_operands): Remove inline keyword.
10817 2013-07-02 Jakub Jelinek <jakub@redhat.com>
10819 PR tree-optimization/57741
10820 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
10821 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
10822 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
10823 Allow REAL_CST step_exprs if flag_associative_math.
10824 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
10826 2013-07-02 Ian Bolton <ian.bolton@arm.com>
10828 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
10830 2013-07-02 Ian Bolton <ian.bolton@arm.com>
10832 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
10834 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10836 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
10838 (iorsi3_insn): Likewise.
10839 (arm_xorsi3): Likewise.
10841 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
10843 * arm.md (attribute "wtype"): Delete. Move attribute values from here
10845 (attribute "type"): ... here, and prefix with "wmmx_".
10846 (attribute "core_cycles"): Update for attribute changes.
10847 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
10848 (tbcstv4hi): Likewise.
10849 (tbcstv2si): Likewise.
10850 (iwmmxt_iordi3): Likewise.
10851 (iwmmxt_xordi3): Likewise.
10852 (iwmmxt_anddi3): Likewise.
10853 (iwmmxt_nanddi3): Likewise.
10854 (iwmmxt_arm_movdi): Likewise.
10855 (iwmmxt_movsi_insn): Likewise.
10856 (mov<mode>_internal): Likewise.
10857 (and<mode>3_iwmmxt): Likewise.
10858 (ior<mode>3_iwmmxt): Likewise.
10859 (xor<mode>3_iwmmxt): Likewise.
10860 (add<mode>3_iwmmxt): Likewise.
10861 (ssaddv8qi3): Likewise.
10862 (ssaddv4hi3): Likewise.
10863 (ssaddv2si3): Likewise.
10864 (usaddv8qi3): Likewise.
10865 (usaddv4hi3): Likewise.
10866 (usaddv2si3): Likewise.
10867 (sub<mode>3_iwmmxt): Likewise.
10868 (sssubv8qi3): Likewise.
10869 (sssubv4hi3): Likewise.
10870 (sssubv2si3): Likewise.
10871 (ussubv8qi3): Likewise.
10872 (ussubv4hi3): Likewise.
10873 (ussubv2si3): Likewise.
10874 (mulv4hi3_iwmmxt): Likewise.
10875 (smulv4hi3_highpart): Likewise.
10876 (umulv4hi3_highpart): Likewise.
10877 (iwmmxt_wmacs): Likewise.
10878 (iwmmxt_wmacsz): Likewise.
10879 (iwmmxt_wmacu): Likewise.
10880 (iwmmxt_wmacuz): Likewise.
10881 (iwmmxt_clrdi): Likewise.
10882 (iwmmxt_clrv8qi): Likewise.
10883 (iwmmxt_clr4hi): Likewise.
10884 (iwmmxt_clr2si): Likewise.
10885 (iwmmxt_uavgrndv8qi3): Likewise.
10886 (iwmmxt_uavgrndv4hi3): Likewise.
10887 (iwmmxt_uavgv8qi3): Likewise.
10888 (iwmmxt_uavgv4hi3): Likewise.
10889 (iwmmxt_tinsrb): Likewise.
10890 (iwmmxt_tinsrh): Likewise.
10891 (iwmmxt_tinsrw): Likewise.
10892 (iwmmxt_textrmub): Likewise.
10893 (iwmmxt_textrmsb): Likewise.
10894 (iwmmxt_textrmuh): Likewise.
10895 (iwmmxt_textrmsh): Likewise.
10896 (iwmmxt_textrmw): Likewise.
10897 (iwmxxt_wshufh): Likewise.
10898 (eqv8qi3): Likewise.
10899 (eqv4hi3): Likewise.
10900 (eqv2si3): Likewise.
10901 (gtuv8qi3): Likewise.
10902 (gtuv4hi3): Likewise.
10903 (gtuv2si3): Likewise.
10904 (gtv8qi3): Likewise.
10905 (gtv4hi3): Likewise.
10906 (gtv2si3): Likewise.
10907 (smax<mode>3_iwmmxt): Likewise.
10908 (umax<mode>3_iwmmxt): Likewise.
10909 (smin<mode>3_iwmmxt): Likewise.
10910 (umin<mode>3_iwmmxt): Likewise.
10911 (iwmmxt_wpackhss): Likewise.
10912 (iwmmxt_wpackwss): Likewise.
10913 (iwmmxt_wpackdss): Likewise.
10914 (iwmmxt_wpackhus): Likewise.
10915 (iwmmxt_wpackwus): Likewise.
10916 (iwmmxt_wpackdus): Likewise.
10917 (iwmmxt_wunpckihb): Likewise.
10918 (iwmmxt_wunpckihh): Likewise.
10919 (iwmmxt_wunpckihw): Likewise.
10920 (iwmmxt_wunpckilb): Likewise.
10921 (iwmmxt_wunpckilh): Likewise.
10922 (iwmmxt_wunpckilw): Likewise.
10923 (iwmmxt_wunpckehub): Likewise.
10924 (iwmmxt_wunpckehuh): Likewise.
10925 (iwmmxt_wunpckehuw): Likewise.
10926 (iwmmxt_wunpckehsb): Likewise.
10927 (iwmmxt_wunpckehsh): Likewise.
10928 (iwmmxt_wunpckehsw): Likewise.
10929 (iwmmxt_wunpckelub): Likewise.
10930 (iwmmxt_wunpckeluh): Likewise.
10931 (iwmmxt_wunpckeluw): Likewise.
10932 (iwmmxt_wunpckelsb): Likewise.
10933 (iwmmxt_wunpckelsh): Likewise.
10934 (iwmmxt_wunpckelsw): Likewise.
10935 (ror<mode>3): Likewise.
10936 (ashr<mode>3_iwmmxt): Likewise.
10937 (lshr<mode>3_iwmmxt): Likewise.
10938 (ashl<mode>3_iwmmxt): Likewise.
10939 (ror<mode>3_di): Likewise.
10940 (ashr<mode>3_di): Likewise.
10941 (lshr<mode>3_di): Likewise.
10942 (ashl<mode>3_di): Likewise.
10943 (iwmmxt_wmadds): Likewise.
10944 (iwmmxt_wmaddu): Likewise.
10945 (iwmmxt_tmia): Likewise.
10946 (iwmmxt_tmiaph): Likewise.
10947 (iwmmxt_tmiabb): Likewise.
10948 (iwmmxt_tmiatb): Likewise.
10949 (iwmmxt_tmiabt): Likewise.
10950 (iwmmxt_tmiatt): Likewise.
10951 (iwmmxt_tmovmskb): Likewise.
10952 (iwmmxt_tmovmskh): Likewise.
10953 (iwmmxt_tmovmskw): Likewise.
10954 (iwmmxt_waccb): Likewise.
10955 (iwmmxt_wacch): Likewise.
10956 (iwmmxt_waccw): Likewise.
10957 (iwmmxt_waligni): Likewise.
10958 (iwmmxt_walignr): Likewise.
10959 (iwmmxt_walignr0): Likewise.
10960 (iwmmxt_walignr1): Likewise.
10961 (iwmmxt_walignr2): Likewise.
10962 (iwmmxt_walignr3): Likewise.
10963 (iwmmxt_wsadb): Likewise.
10964 (iwmmxt_wsadh): Likewise.
10965 (iwmmxt_wsadbz): Likewise.
10966 (iwmmxt_wsadhz): Likewise.
10967 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
10968 (iwmmxt_wabsdiffb): Likewise.
10969 (iwmmxt_wabsdiffh): Likewise.
10970 (iwmmxt_wabsdiffw): Likewise.
10971 (iwmmxt_waddsubhx): Likewise
10972 (iwmmxt_wsubaddhx): Likewise.
10973 (addc<mode>3): Likewise.
10974 (iwmmxt_avg4): Likewise.
10975 (iwmmxt_avg4r): Likewise.
10976 (iwmmxt_wmaddsx): Likewise.
10977 (iwmmxt_wmaddux): Likewise.
10978 (iwmmxt_wmaddsn): Likewise.
10979 (iwmmxt_wmaddun): Likewise.
10980 (iwmmxt_wmulwsm): Likewise.
10981 (iwmmxt_wmulwum): Likewise.
10982 (iwmmxt_wmulsmr): Likewise.
10983 (iwmmxt_wmulumr): Likewise.
10984 (iwmmxt_wmulwsmr): Likewise.
10985 (iwmmxt_wmulwumr): Likewise.
10986 (iwmmxt_wmulwl): Likewise.
10987 (iwmmxt_wqmulm): Likewise.
10988 (iwmmxt_wqmulwm): Likewise.
10989 (iwmmxt_wqmulmr): Likewise.
10990 (iwmmxt_wqmulwmr): Likewise.
10991 (iwmmxt_waddbhusm): Likewise.
10992 (iwmmxt_waddbhusl): Likewise.
10993 (iwmmxt_wqmiabb): Likewise.
10994 (iwmmxt_wqmiabt): Likewise.
10995 (iwmmxt_wqmiatb): Likewise.
10996 (iwmmxt_wqmiatt): Likewise.
10997 (iwmmxt_wqmiabbn): Likewise.
10998 (iwmmxt_wqmiabtn): Likewise.
10999 (iwmmxt_wqmiatbn): Likewise.
11000 (iwmmxt_wqmiattn): Likewise.
11001 (iwmmxt_wmiabb): Likewise.
11002 (iwmmxt_wmiabt): Likewise.
11003 (iwmmxt_wmiatb): Likewise.
11004 (iwmmxt_wmiatt): Likewise.
11005 (iwmmxt_wmiabbn): Likewise.
11006 (iwmmxt_wmiabtn): Likewise.
11007 (iwmmxt_wmiatbn): Likewise.
11008 (iwmmxt_wmiattn): Likewise.
11009 (iwmmxt_wmiawbb): Likewise.
11010 (iwmmxt_wmiawbt): Likewise.
11011 (iwmmxt_wmiawtb): Likewise.
11012 (iwmmxt_wmiawtt): Likewise.
11013 (iwmmxt_wmiawbbn): Likewise.
11014 (iwmmxt_wmiawbtn): Likewise.
11015 (iwmmxt_wmiawtbn): Likewise.
11016 (iwmmxt_wmiawttn): Likewise.
11017 (iwmmxt_wmerge): Likewise.
11018 (iwmmxt_tandc<mode>3): Likewise.
11019 (iwmmxt_torc<mode>3): Likewise.
11020 (iwmmxt_torvsc<mode>3): Likewise.
11021 (iwmmxt_textrc<mode>3): Likewise.
11022 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
11023 (wmmxt_pack): Likewise.
11024 (wmmxt_mult_c1): Likewise.
11025 (wmmxt_mult_c2): Likewise.
11026 (wmmxt_alu_c1): Likewise.
11027 (wmmxt_alu_c2): Likewise.
11028 (wmmxt_alu_c3): Likewise.
11029 (wmmxt_transfer_c1): Likewise.
11030 (wmmxt_transfer_c2): Likewise.
11031 (wmmxt_transfer_c3): Likewise.
11032 (marvell_f_iwmmxt_wstr): Likewise.
11033 (marvell_f_iwmmxt_wldr): Likewise.
11035 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
11037 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
11039 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
11042 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
11043 * lra-constraints.c (need_for_split_p): Check call used hard regs
11044 living through calls.
11046 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
11047 call used regs for call insn.
11049 2013-06-28 Jakub Jelinek <jakub@redhat.com>
11052 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
11053 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
11055 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
11057 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
11059 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
11061 * lra-constraints.c (need_for_split_p): Check call used hard regs
11062 living through calls.
11064 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
11067 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
11068 to tie with any other modes. Eliminate Altivec vector mode tests,
11069 since these are a subset of ALTIVEC or VSX vector modes. Simplify
11070 code, to return 0 if testing MODE2 for a condition, if we've
11071 already tested MODE1 for the same condition.
11073 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
11075 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
11078 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
11080 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
11081 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
11083 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
11085 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
11087 (aarch64_symbolic_constant_p): Remove.
11088 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
11089 static. Fix line length and white space.
11090 (aarch64_symbolic_constant_p): Remove.
11091 * config/aarch64/predicates.md (aarch64_valid_symref):
11092 Use aarch64_classify_symbol_expression.
11094 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11096 * config/arm/constraints.md (Ts): New constraint.
11097 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
11099 (compare_scc): Use "Ts" constraint for operand 0.
11100 (ior_scc_scc): Likewise.
11101 (and_scc_scc): Likewise.
11102 (and_scc_scc_nodom): Likewise.
11103 (ior_scc_scc_cmp): Likewise for operand 7.
11104 (and_scc_scc_cmp): Likewise.
11105 * config/arm/thumb2.md (thumb2_movsi_insn):
11106 Add alternatives for 16-bit encodings.
11107 (thumb2_movhi_insn): Likewise.
11108 (thumb2_movsicc_insn): Likewise.
11109 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
11110 (thumb2_negscc): Use "Ts" constraint.
11111 Move mvn instruction outside cond_exec block.
11112 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
11113 for 16-bit encodings.
11115 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11117 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
11119 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
11120 (mulsi3subsi): Likewise.
11121 (mulsidi3adddi): Likewise.
11122 (mulsidi3_v6): Likewise.
11123 (umulsidi3_v6): Likewise.
11124 (umulsidi3adddi_v6): Likewise.
11125 (smulsi3_highpart_v6): Likewise.
11126 (umulsi3_highpart_v6): Likewise.
11127 (mulhisi3tb): Likewise.
11128 (mulhisi3bt): Likewise.
11129 (mulhisi3tt): Likewise.
11130 (maddhisi4): Likewise.
11131 (maddhisi4tb): Likewise.
11132 (maddhisi4tt): Likewise.
11133 (maddhidi4): Likewise.
11134 (maddhidi4tb): Likewise.
11135 (maddhidi4tt): Likewise.
11136 (zeroextractsi_compare0_scratch): Likewise.
11137 (insv_zero): Likewise.
11138 (insv_t2): Likewise.
11139 (anddi_notzesidi_di): Likewise.
11140 (anddi_notsesidi_di): Likewise.
11141 (andsi_notsi_si): Likewise.
11142 (iordi_zesidi_di): Likewise.
11143 (xordi_zesidi_di): Likewise.
11144 (andsi_iorsi3_notsi): Likewise.
11145 (smax_0): Likewise.
11146 (smax_m1): Likewise.
11147 (smin_0): Likewise.
11148 (not_shiftsi): Likewise.
11149 (unaligned_loadsi): Likewise.
11150 (unaligned_loadhis): Likewise.
11151 (unaligned_loadhiu): Likewise.
11152 (unaligned_storesi): Likewise.
11153 (unaligned_storehi): Likewise.
11154 (extv_reg): Likewise.
11155 (extzv_t2): Likewise.
11156 (divsi3): Likewise.
11157 (udivsi3): Likewise.
11158 (arm_zero_extendhisi2addsi): Likewise.
11159 (arm_zero_extendqisi2addsi): Likewise.
11160 (compareqi_eq0): Likewise.
11161 (arm_extendhisi2_v6): Likewise.
11162 (arm_extendqisi2addsi): Likewise.
11163 (arm_movt): Likewise.
11164 (thumb2_ldrd): Likewise.
11165 (thumb2_ldrd_base): Likewise.
11166 (thumb2_ldrd_base_neg): Likewise.
11167 (thumb2_strd): Likewise.
11168 (thumb2_strd_base): Likewise.
11169 (thumb2_strd_base_neg): Likewise.
11170 (arm_negsi2): Add alternative for 16-bit encoding.
11171 (arm_one_cmplsi2): Likewise.
11173 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11175 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
11176 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
11177 (movdfcc): Likewise.
11178 * config/arm/vfp.md (*thumb2_movsf_vfp):
11179 Disable predication for arm_restrict_it.
11180 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
11181 (*thumb2_movdfcc_vfp): Likewise.
11182 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
11183 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
11184 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
11185 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
11186 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
11187 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
11188 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
11189 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
11190 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
11191 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
11192 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
11193 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
11194 Disable predication for arm_restrict_it.
11196 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
11198 * config/i386/bmiintrin.h (_bextr_u32): New.
11199 (_bextr_u64): Ditto.
11201 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
11203 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
11204 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
11205 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
11206 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
11207 * config/mips/n32-elf.h: ...this new file.
11209 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
11212 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
11213 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
11215 2013-06-27 Catherine Moore <clm@codesourcery.com>
11217 * config/mips/mips-tables.opt: Regenerate.
11218 * config/mips/mips-cpus.def: Add m14ke and m14kec.
11219 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
11220 * doc/invoke.texi: Add -m14kc.
11222 2013-06-27 Jakub Jelinek <jakub@redhat.com>
11225 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
11226 constraints of operand 1 and 2.
11229 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
11230 to match RTL canonicalization. Swap predicates and
11231 constraints of operand 1 and 2.
11233 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
11235 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
11236 Process OP_INOUT regs for splitting too.
11238 2013-06-27 Jakub Jelinek <jakub@redhat.com>
11240 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
11241 decl before the loop, initialize to NULL.
11242 (vectorizable_load): Initialize ptr_incr to NULL.
11244 2013-06-27 Martin Jambor <mjambor@suse.cz>
11247 * ipa-ref.h (ipa_maybe_record_reference): Declare.
11248 * ipa-ref.c (ipa_maybe_record_reference): New function.
11249 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
11250 * ipa-cp.c (create_specialized_node): Record potential references from
11252 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
11254 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
11256 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
11257 parameter 'mode' of type 'enum machine_mode mode'; change to pass
11258 'mode' to force_reg.
11259 (aarch64_add_offset): Update calls to aarch64_force_temporary.
11260 (aarch64_expand_mov_immediate): Likewise.
11262 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
11264 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
11265 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
11267 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11269 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
11270 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
11271 (struct machine_function): Add tbegin_p.
11272 (s390_canonicalize_comparison): Fold CC mode compares to
11273 conditional jump if possible.
11274 (s390_emit_jump): Return the emitted jump.
11275 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
11276 Handle CCRAWmode compares.
11277 (s390_option_override): Default to -mhtm if available.
11278 (s390_reg_clobbered_rtx): Handle floating point regs as well.
11279 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
11280 FPRs instead of df_regs_ever_live_p.
11281 (s390_optimize_nonescaping_tx): New function.
11282 (s390_init_frame_layout): Extend clobbered_regs array to cover
11284 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
11285 (s390_expand_tbegin): New function.
11286 (enum s390_builtin): New enum definition.
11287 (code_for_builtin): New array definition.
11288 (s390_init_builtins): New function.
11289 (s390_expand_builtin): New function.
11290 (TARGET_INIT_BUILTINS): Define.
11291 (TARGET_EXPAND_BUILTIN): Define.
11292 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
11293 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
11294 (s390_alc_comparison): Likewise.
11295 * config/s390/s390-modes.def: Add CCRAWmode.
11296 * config/s390/s390.h (processor_flags): Add PF_TX.
11297 (TARGET_CPU_HTM): Define macro.
11298 (TARGET_HTM): Define macro.
11299 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
11300 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
11301 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
11302 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
11303 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
11304 (TBEGIN_MASK, TBEGINC_MASK): New constants.
11305 ("*cc_to_int"): Move up.
11306 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
11307 constants other than 0.
11308 ("*ccraw_to_int"): New insn and splitter definition.
11309 ("tbegin", "tbegin_nofloat", "tbegin_retry")
11310 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
11311 ("tx_assist"): New expander.
11312 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
11313 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
11314 * config/s390/s390.opt: Add -mhtm option.
11315 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
11316 * config/s390/htmxlintrin.h: New file.
11317 * config/s390/htmintrin.h: New file.
11318 * config/s390/s390intrin.h: New file.
11319 * doc/extend.texi: Document htm builtins.
11320 * config.gcc: Add the new header files to extra_headers.
11322 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
11324 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
11325 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
11327 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
11328 Pat Haugen <pthaugen@us.ibm.com>
11329 Peter Bergner <bergner@vnet.ibm.com>
11331 * config/rs6000/power8.md: New.
11332 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
11333 setting for power8 entry.
11334 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
11335 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
11336 test for Power4/Power5 only.
11337 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
11339 (force_new_group): Adjust comment.
11340 * config/rs6000/rs6000.md: Include power8.md.
11342 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
11344 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
11345 * config/arm/arm-protos.h (arm_max_conditional_execute): New
11347 (tune_params): Update comment.
11348 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
11349 (arm_max_conditional_execute): New function.
11350 (thumb2_final_prescan_insn): Use max_insn_skipped and
11351 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
11353 2013-06-25 Jakub Jelinek <jakub@redhat.com>
11355 PR tree-optimization/57705
11356 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
11357 SSA_NAME step, provided that it is not defined inside the loop.
11358 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
11359 (get_initial_def_for_induction): Handle SSA_NAME IV step.
11361 2013-06-25 Martin Jambor <mjambor@suse.cz>
11363 PR middle-end/57670
11364 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
11365 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
11367 (ipa_note_param_call): Initialize member_ptr flag.
11368 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
11369 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
11370 (ipa_write_indirect_edge_info): Stream member_ptr flag.
11371 (ipa_read_indirect_edge_info): Likewise.
11373 2013-06-25 Richard Biener <rguenther@suse.de>
11375 PR middle-end/56977
11376 * passes.c (init_optimization_passes): Move pass_fold_builtins
11377 and pass_dce earlier with -Og.
11379 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
11381 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
11382 <BIT_FIELD_REF>: Remove trailing TAB.
11383 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
11386 2013-06-24 Martin Jambor <mjambor@suse.cz>
11388 PR tree-optimization/57358
11389 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
11390 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
11391 (ipa_analyze_params_uses): Generate pessimistic info when true.
11393 2013-06-24 Martin Jambor <mjambor@suse.cz>
11395 PR tree-optimization/57539
11396 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
11397 global.inlined_to of the new node to it. All callers changed.
11398 * ipa-inline-transform.c (clone_inlined_nodes): New variable
11399 inlining_into, pass it to cgraph_clone_node.
11400 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
11401 ipa_free_edge_args_substructures.
11402 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
11403 rdesc linked list. Do not assert rdesc edges have inlined caller.
11404 Assert we have found an rdesc in the rdesc list.
11406 2013-06-24 Richard Biener <rguenther@suse.de>
11408 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
11409 (pointer_set_lookup): Declare.
11410 (class pointer_map): New template class implementing a
11411 generic pointer to T map.
11412 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
11413 pointer_map<T>::contains, pointer_map<T>::insert,
11414 pointer_map<T>::traverse): New functions.
11415 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
11416 (pointer_set_lookup): New function.
11417 (pointer_set_contains): Use pointer_set_lookup.
11418 (pointer_set_insert): Likewise.
11419 (insert_aux): Remove.
11420 (struct pointer_map_t): Embed a pointer_set_t.
11421 (pointer_map_create): Adjust.
11422 (pointer_map_destroy): Likewise.
11423 (pointer_map_contains): Likewise.
11424 (pointer_map_insert): Likewise.
11425 (pointer_map_traverse): Likewise.
11426 * tree-streamer.h (struct streamer_tree_cache_d): Use a
11427 pointer_map<unsigned> instead of a pointer_map_t.
11428 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
11429 (streamer_tree_cache_lookup): Likewise.
11430 (streamer_tree_cache_create): Likewise.
11431 (streamer_tree_cache_delete): Likewise.
11432 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
11433 pointer_map<unsigned> instead of a pointer_map_t.
11434 (lto_init_tree_ref_encoder): Adjust.
11435 (lto_destroy_tree_ref_encoder): Likewise.
11436 * lto-section-out.c (lto_output_decl_index): Likewise.
11437 (lto_record_function_out_decl_state): Likewise.
11438 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
11440 2013-06-24 Richard Biener <rguenther@suse.de>
11442 PR tree-optimization/57488
11443 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
11445 2013-06-24 Alan Modra <amodra@gmail.com>
11447 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
11448 (gen_easy_altivec_constant): Likewise.
11449 * config/rs6000/predicates.md (easy_vector_constant_add_self,
11450 easy_vector_constant_msb): Likewise.
11452 2013-06-23 Jakub Jelinek <jakub@redhat.com>
11455 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
11456 add missing return true.
11458 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
11461 * config/sh/predicates.md (general_extend_operand): Invoke
11462 general_movsrc_operand for memory operands.
11463 (general_movsrc_operand): Allow reg+reg addressing, do not use
11464 general_operand for memory operands.
11466 2013-06-23 Sriraman Tallam <tmsriram@google.com>
11468 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
11469 when current target options does not apply.
11470 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
11471 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
11472 * config/i386/bmiintrin.h: Pass appropriate target
11473 attributes to header.
11474 * config/i386/mmintrin.h: Ditto.
11475 * config/i386/nmmintrin.h: Ditto.
11476 * config/i386/avx2intrin.h: Ditto.
11477 * config/i386/fxsrintrin.h: Ditto.
11478 * config/i386/tbmintrin.h: Ditto.
11479 * config/i386/xsaveintrin.h: Ditto.
11480 * config/i386/f16cintrin.h: Ditto.
11481 * config/i386/xtestintrin.h: Ditto.
11482 * config/i386/xsaveoptintrin.h: Ditto.
11483 * config/i386/bmi2intrin.h: Ditto.
11484 * config/i386/lzcntintrin.h: Ditto.
11485 * config/i386/smmintrin.h: Ditto.
11486 * config/i386/wmmintrin.h: Ditto.
11487 * config/i386/x86intrin.h: Remove all header include guards.
11488 * config/i386/prfchwintrin.h: Ditto.
11489 * config/i386/pmmintrin.h: Ditto.
11490 * config/i386/tmmintrin.h: Ditto.
11491 * config/i386/xmmintrin.h: Ditto.
11492 * config/i386/popcntintrin.h: Ditto.
11493 * config/i386/rdseedintrin.h: Ditto.
11494 * config/i386/ammintrin.h: Ditto.
11495 * config/i386/emmintrin.h: Ditto.
11496 * config/i386/immintrin.h: Remove all header include guards.
11497 * config/i386/fma4intrin.h: Ditto.
11498 * config/i386/lwpintrin.h: Ditto.
11499 * config/i386/xopintrin.h: Ditto.
11500 * config/i386/ia32intrin.h: Ditto.
11501 * config/i386/avxintrin.h: Ditto.
11502 * config/i386/rtmintrin.h: Ditto.
11503 * config/i386/fmaintrin.h: Ditto.
11504 * config/i386/mm3dnow.h: Ditto.
11506 2013-06-22 Sriraman Tallam <tmsriram@google.com>
11508 * common/config/i386/i386-common.c: Handle LZCNT.
11510 2013-06-22 Andi Kleen <ak@linux.intel.com>
11512 * doc/extend.texi: Use __atomic_store_n instead of
11513 __atomic_store in HLE example.
11515 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
11517 * config/sh/sh.c: Remove <cstdlib> workaround.
11519 2013-06-21 Andi Kleen <ak@linux.intel.com>
11521 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
11523 2013-06-21 Andi Kleen <ak@linux.intel.com>
11525 * doc/extend.texi: Document that __atomic_clear and
11526 __atomic_test_and_set should only be used with bool.
11528 2013-06-20 Jan Hubicka <jh@suse.cz>
11530 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
11531 types_same_for_odr.
11532 * tree.c (decls_same_for_odr): New function.
11533 (same_for_edr): New function.
11534 (types_same_for_odr): New function.
11535 (get_binfo_at_offset): Use it.
11536 * tree.h (types_same_for_odr): Declare.
11538 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
11539 Jason Merrill <jason@redhat.com>
11541 * system.h: Include <cstdlib> as well as <stdlib.h>.
11543 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
11546 * config/i386/i386.c (construct_container): Report error if
11547 long double is used with disabled x87 float returns.
11549 2013-06-20 Jan Hubicka <jh@suse.cz>
11551 * lto-cgraph.c (input_symtab): Do not set cgraph state.
11553 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
11555 PR rtl-optimization/57425
11556 PR rtl-optimization/57569
11557 * alias.c (write_dependence_p): Remove parameters mem_mode and
11558 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
11559 Changed all callers.
11560 (canon_anti_dependence): Get comments and semantics in sync.
11561 Add parameter mem_canonicalized. Changed all callers.
11562 * rtl.h (canon_anti_dependence): Update prototype.
11564 2013-06-20 Richard Biener <rguenther@suse.de>
11566 * data-streamer-in.c (streamer_read_uhwi): Optimize single
11567 byte case, inline streamer_read_uchar and defer section
11570 2013-06-20 Richard Biener <rguenther@suse.de>
11572 PR tree-optimization/57584
11573 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
11574 SSA names into the expanded expression that take part in
11575 abnormal coalescing.
11577 2013-06-19 Sharad Singhai <singhai@google.com>
11579 * gcov.c (print_usage): Handle new option.
11580 (process_args): Ditto.
11581 (get_gcov_intermediate_filename): New function.
11582 (output_intermediate_file): New function.
11583 (output_gcov_file): New function
11584 (generate_results): Handle new option.
11585 (release_function): Relase demangled name.
11586 (read_graph_file): Handle demangled name.
11587 (output_lines): Ditto.
11588 * doc/gcov.texi: Document gcov intermediate format.
11590 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
11593 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
11594 (lra_emit_add): Use the functions. Add comment about Y as an
11597 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
11600 * collect2.c (collect_atexit): New.
11601 (collect_exit): Delete.
11602 (main): Register collect_atexit with atexit.
11603 (collect_wait): Change collect_exit to exit.
11605 * collect2.h (collect_exit): Delete.
11606 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
11608 2013-06-19 Wei Mi <wmi@google.com>
11610 PR rtl-optimization/57518
11611 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
11612 if regno is used in paradoxical subreg.
11613 (update_equiv_regs): Check pdx_subregs[regno] before
11614 set a reg to be equivalent with a mem.
11616 2013-06-19 Matthias Klose <doko@ubuntu.com>
11619 * file-find.h (find_a_file): Add a mode parameter.
11620 * file-find.c (find_a_file): Likewise.
11621 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
11622 with X_OK for the executables.
11623 * collect2.c (main): Call find_a_file with X_OK.
11625 2013-06-19 Steve Ellcey <sellcey@mips.com>
11628 * config/mips/mips.md (casesi_internal_mips16_<mode>):
11629 Use NEXT_INSN instead of next_real_insn.
11631 2013-06-19 Jan Hubicka <jh@suse.cz>
11633 * cgraph.h (const_value_known_p): Replace by ...
11634 (ctor_for_folding): .. this one.
11635 * cgraphunit.c (process_function_and_variable_attributes): Use it.
11636 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
11637 * expr.c (expand_expr_real_1): Likewise.
11638 (string_constant): Likewise.
11639 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
11640 * ipa.c (process_references): Likewise.
11641 (symtab_remove_unreachable_nodes): Likewise.
11642 * ipa-inline-analysis.c (param_change_prob): Likewise.
11643 * gimple-fold.c (canonicalize_constructor_val): Likewise.
11644 (get_base_constructor): Likwise.
11645 * varpool.c (varpool_remove_node): Likewise.
11646 (varpool_remove_initializer): LIkewise.
11647 (dump_varpool_node): LIkwise.
11648 (const_value_known_p): Rewrite to ...
11649 (ctor_for_folding): ... this one.
11651 2013-06-19 Jakub Jelinek <jakub@redhat.com>
11654 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
11655 PERSONALITY in $PATH derived prefixes.
11657 2013-06-19 Jeff Law <law@redhat.com>
11659 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
11662 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
11663 (simplify_bitwise_binary): Use it to simpify certain binary ops on
11666 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
11668 * config/arm/vfp.md: Move VFP instruction classification documentation
11670 * config/arm/arm.md: ... here. Update instruction classification
11673 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
11675 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
11676 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
11677 pattern. Use more efficient sequences on ARMv5 and Thumb2.
11679 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
11682 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
11683 with NEXT_INSN. Use tablejump_p to check for jump table data
11686 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
11689 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
11690 that now in C++ the value is correct per the C++ standards.
11692 2013-06-19 Richard Biener <rguenther@suse.de>
11694 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
11695 for global context.
11697 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11700 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11703 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
11704 with next_active_insn.
11706 2013-06-18 Sriraman Tallam <tmsriram@google.com>
11708 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
11709 functions are inlined during failures to flag an error.
11710 * tree-inline.c (expand_call_inline): Allow the error to be flagged
11711 in early inline pass.
11713 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
11715 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
11718 2013-06-18 Julian Brown <julian@codesourcery.com>
11720 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
11721 Permit virtual register pre-reload if !strict.
11722 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
11724 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
11726 * config/arm/neon.md (movmisalign<mode>): Use
11727 neon_perm_struct_or_reg_operand instead of
11728 neon_struct_or_register_operand.
11729 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
11730 neon_permissive_struct_operand instead of neon_struct_operand.
11731 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
11732 neon_vector_mem_operand.
11733 * config/arm/predicates.md (neon_struct_operand): Adjust call to
11734 neon_vector_mem_operand.
11735 (neon_permissive_struct_operand): New.
11736 (neon_struct_or_register_operand): Rename to...
11737 (neon_perm_struct_or_reg_operand): This. Adjust call to
11738 neon_vector_mem_operand.
11740 2013-06-18 Richard Biener <rguenther@suse.de>
11742 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
11743 * lto-streamer.h: Include pointer-set.h.
11744 (struct lto_decl_slot): Remove.
11745 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
11746 Remove next_index entry.
11747 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
11748 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
11749 (lto_init_tree_ref_encoder): Adjust.
11750 (lto_destroy_tree_ref_encoder): Likewise.
11751 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
11752 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
11753 (lto_output_decl_index): Adjust.
11754 (lto_new_out_decl_state): Likewise.
11755 (lto_record_function_out_decl_state): Likewise.
11756 * lto-streamer-out.c (copy_function): Likewise.
11758 2013-06-18 Richard Biener <rguenther@suse.de>
11760 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
11761 * cgraphunit.c: Include cfgloop.h.
11762 (init_lowered_empty_function): Initialize the loop tree.
11763 (assemble_thunk): Insert new BBs into loops.
11765 2013-06-18 Richard Biener <rguenther@suse.de>
11767 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
11768 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
11769 the map from cache entry to cache index optional.
11770 (streamer_tree_cache_replace_tree): Adjust accordingly.
11771 (streamer_tree_cache_append): Likewise.
11772 (streamer_tree_cache_delete): Likewise.
11773 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
11774 streamer cache map from cache entry to cache index.
11775 * lto-streamer-out.c (create_output_block): Adjust.
11777 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
11779 * config/arm/arm.md (attribute "insn"): Move multiplication and
11780 division attributes to...
11781 (attribute "type"): ... here. Remove mult.
11782 (attribute "mul32"): New attribute.
11783 (attribute "mul64"): Add umaal.
11784 (*arm_mulsi3): Update attributes.
11785 (*arm_mulsi3_v6): Likewise.
11786 (*thumb_mulsi3): Likewise.
11787 (*thumb_mulsi3_v6): Likewise.
11788 (*mulsi3_compare0): Likewise.
11789 (*mulsi3_compare0_v6): Likewise.
11790 (*mulsi_compare0_scratch): Likewise.
11791 (*mulsi_compare0_scratch_v6): Likewise.
11792 (*mulsi3addsi): Likewise.
11793 (*mulsi3addsi_v6): Likewise.
11794 (*mulsi3addsi_compare0): Likewise.
11795 (*mulsi3addsi_compare0_v6): Likewise.
11796 (*mulsi3addsi_compare0_scratch): Likewise.
11797 (*mulsi3addsi_compare0_scratch_v6): Likewise.
11798 (*mulsi3subsi): Likewise.
11799 (*mulsidi3adddi): Likewise.
11800 (*mulsi3addsi_v6): Likewise.
11801 (*mulsidi3adddi_v6): Likewise.
11802 (*mulsidi3_nov6): Likewise.
11803 (*mulsidi3_v6): Likewise.
11804 (*umulsidi3_nov6): Likewise.
11805 (*umulsidi3_v6): Likewise.
11806 (*umulsidi3adddi): Likewise.
11807 (*umulsidi3adddi_v6): Likewise.
11808 (*smulsi3_highpart_nov6): Likewise.
11809 (*smulsi3_highpart_v6): Likewise.
11810 (*umulsi3_highpart_nov6): Likewise.
11811 (*umulsi3_highpart_v6): Likewise.
11812 (mulhisi3): Likewise.
11813 (*mulhisi3tb): Likewise.
11814 (*mulhisi3bt): Likewise.
11815 (*mulhisi3tt): Likewise.
11816 (maddhisi4): Likewise.
11817 (*maddhisi4tb): Likewise.
11818 (*maddhisi4tt): Likewise.
11819 (maddhidi4): Likewise.
11820 (*maddhidi4tb): Likewise.
11821 (*maddhidi4tt): Likewise.
11822 (divsi3): Likewise.
11823 (udivsi3): Likewise.
11824 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
11825 (thumb2_mulsi_short_compare0): Likewise.
11826 (thumb2_mulsi_short_compare0_scratch): Likewise.
11827 * config/arm/arm1020e.md (1020mult1): Update attribute change.
11828 (1020mult2): Likewise.
11829 (1020mult3): Likewise.
11830 (1020mult4): Likewise.
11831 (1020mult5): Likewise.
11832 (1020mult6): Likewise.
11833 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
11835 (cortex_a15_mult64): Likewise.
11836 (cortex_a15_sdiv): Likewise.
11837 (cortex_a15_udiv): Likewise.
11838 * config/arm/arm1026ejs.md (mult1): Update attribute change.
11844 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
11845 (pj4_ir_div): Likewise.
11846 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
11847 (11_mult2): Likewise.
11848 (11_mult3): Likewise.
11849 (11_mult4): Likewise.
11850 (11_mult5): Likewise.
11851 (11_mult6): Likewise.
11852 (11_mult7): Likewise.
11853 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
11854 (cortex_a8_mla): Likewise.
11855 (cortex_a8_mull): Likewise.
11856 (cortex_a8_smulwy): Likewise.
11857 (cortex_a8_smlald): Likewise.
11858 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
11859 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
11860 (cortex_r4_mul_3): Likewise.
11861 (cortex_r4_mla_4): Likewise.
11862 (cortex_r4_mla_3): Likewise.
11863 (cortex_r4_smlald): Likewise.
11864 (cortex_r4_mull): Likewise.
11865 (cortex_r4_sdiv): Likewise.
11866 (cortex_r4_udiv): Likewise.
11867 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
11868 (cortex_a7_idiv): Likewise.
11869 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
11870 (9_mult2): Likewise.
11871 (9_mult3): Likewise.
11872 (9_mult4): Likewise.
11873 (9_mult5): Likewise.
11874 (9_mult6): Likewise.
11875 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
11876 (cortex_a53_sdiv): Likewise.
11877 (cortex_a53_udiv): Likewise.
11878 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
11879 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
11880 (mp626_mult2): Likewise.
11881 (mp626_mult3): Likewise.
11882 (mp626_mult4): Likewise.
11883 * config/arm/fa526.md (526_mult1): Update attribute change.
11884 (526_mult2): Likewise.
11885 * config/arm/arm-generic.md (mult): Update attribute change.
11886 (mult_ldsched_strongarm): Likewise.
11887 (mult_ldsched): Likewise.
11888 (multi_cycle): Likewise.
11889 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
11890 * config/arm/fa606te.md (606te_mult1): Update attribute change.
11891 (606te_mult2): Likewise.
11892 (606te_mult3): Likewise.
11893 (606te_mult4): Likewise.
11894 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
11895 (cortex_a9_mac16): Likewise.
11896 (cortex_a9_multiply): Likewise.
11897 (cortex_a9_mac): Likewise.
11898 (cortex_a9_multiply_long): Likewise.
11899 * config/arm/fa626te.md (626te_mult1): Update attribute change.
11900 (626te_mult2): Likewise.
11901 (626te_mult3): Likewise.
11902 (626te_mult4): Likewise.
11904 2013-06-18 Richard Biener <rguenther@suse.de>
11907 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
11909 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11912 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
11913 with next_active_insn.
11915 2013-06-18 Alan Modra <amodra@gmail.com>
11917 * config/rs6000/rs6000.h (enum data_align): New.
11918 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
11919 (DATA_ABI_ALIGNMENT): Define.
11920 (CONSTANT_ALIGNMENT): Correct comment.
11921 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
11922 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
11924 2013-06-17 David Malcolm <dmalcolm@redhat.com>
11926 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
11927 ATTRIBUTE_UNUSED marking.
11929 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
11931 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
11932 alternative and update.
11933 (aarch64_dup_lanedi): Delete.
11934 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
11935 * config/aarch64/aarch64-simd-builtins.def: Update.
11937 2013-06-17 Richard Biener <rguenther@suse.de>
11939 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
11940 (lto_input_scc): Declare.
11941 (lto_input_tree_1): Likewise.
11942 (struct lto_stats_d): Add num_tree_bodies_output and
11943 num_pickle_refs_output.
11944 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
11945 (lto_read_tree_1): Split out from ...
11946 (lto_read_tree): ... this.
11947 (lto_input_scc): New function.
11948 (lto_input_tree_1): Split out from ...
11949 (lto_input_tree): ... this. Handle LTO_tree_scc.
11950 (lto_data_in_create): Create the streamer cache without hashes.
11951 * lto-streamer-out.c (create_output_block): Create the streamer
11952 cache with hashes when not doing WPA.
11953 (lto_write_tree_1): Split out from ...
11954 (lto_write_tree): ... this.
11955 (get_symbol_initial_value): New function.
11956 (lto_output_tree_1): Split out from ...
11957 (lto_output_tree): ... this. Write trees as series of SCCs
11958 using a DFS walk via DFS_write_tree.
11959 (struct sccs, struct scc_entry): New types.
11960 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
11961 (DFS_write_tree_body): New function.
11962 (DFS_write_tree): Likewise.
11963 (hash_tree): Likewise.
11964 (scc_entry_compare): Likewise.
11965 (hash_scc): Likewise.
11966 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
11967 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
11968 TREE_CHAIN as regular reference.
11969 (streamer_read_integer_cst): Remove.
11970 (streamer_get_pickled_tree): Adjust.
11971 * tree-streamer-out.c (streamer_write_chain): Disable streaming
11972 of DECL_EXTERNALs in BLOCK_VARS for now.
11973 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
11975 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
11976 Add hash value argument and record that if hashes are recorded
11978 (streamer_tree_cache_insert_1): Adjust.
11979 (streamer_tree_cache_insert): Likewise.
11980 (streamer_tree_cache_insert_at): Rename to ...
11981 (streamer_tree_cache_replace_tree): ... this and adjust.
11982 (streamer_tree_cache_append): Adjust.
11983 (record_common_node): Likewise.
11984 (streamer_tree_cache_create): Add argument whether to
11985 record hash values together with trees.
11986 (streamer_tree_cache_delete): Adjust.
11987 * tree-streamer.h (struct streamer_tree_cache_d): Add
11989 (streamer_read_integer_cst): Remove.
11990 (streamer_tree_cache_insert): Adjust.
11991 (streamer_tree_cache_append): Likewise.
11992 (streamer_tree_cache_insert_at): Rename to ...
11993 (streamer_tree_cache_replace_tree): ... this and adjust.
11994 (streamer_tree_cache_create): Add argument whether to record hashes.
11995 (streamer_tree_cache_get): Rename to ...
11996 (streamer_tree_cache_get_tree): ... this.
11997 (streamer_tree_cache_get_hash): New function.
11998 * tree.c (cache_integer_cst): New function.
11999 * tree.h (cache_integer_cst): Declare.
12000 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
12001 * lto-symtab.c (lto_varpool_replace_node): Only release
12002 DECL_INITIAL of non-prevailing decls.
12003 * varpool.c (varpool_remove_initializer): Do not release
12004 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
12006 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
12008 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
12009 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
12010 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
12011 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
12012 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
12013 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
12014 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
12015 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
12016 instead of TARGET_64BIT.
12017 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
12018 Require ISA_HAS_<D>DIV.
12020 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
12022 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
12023 (mips*-*-linux*): Move default with_llsc setting to where other
12025 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
12027 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
12028 Likewise. Remove default with_tune setting. Move default float
12029 setting to its own block. Handle with_llsc in the same block as above.
12031 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
12033 PR rtl-optimization/57425
12034 PR rtl-optimization/57569
12035 * alias.c (write_dependence_p): Add new parameters mem_mode,
12036 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
12037 Changed all callers.
12038 (canon_anti_dependence): New function.
12039 * cse.c (check_dependence): Use canon_anti_dependence.
12040 * cselib.c (cselib_invalidate_mem): Likewise.
12041 * rtl.h (canon_anti_dependence): Declare.
12043 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
12045 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
12046 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
12047 ".set mips3" for 64-bit targets.
12049 2013-06-15 Dehao Chen <dehao@google.com>
12051 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
12052 * gimple-low.c (gimple_check_call_matching_types): Likewise.
12053 (gimple_check_call_args): Likewise.
12054 * value-prof.c (check_ic_target): Likewise.
12055 * ipa-inline.c (early_inliner): Likewise.
12056 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
12057 * cgraph.c (cgraph_create_edge_1): Likewise.
12058 (cgraph_make_edge_direct): Likewise.
12060 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
12063 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
12064 rs6000_output_move_128bit to handle emitting quad memory
12065 operations. Set attribute length to 8 bytes.
12067 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
12069 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
12071 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
12072 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
12073 (aarch64_<su>mlsl<mode>): Likewise.
12075 2013-06-14 Mike Stump <mikestump@comcast.net>
12077 * Makefile.in (TARGET_H): Add insn-codes.h.
12079 2013-06-14 Alan Modra <amodra@gmail.com>
12081 PR middle-end/57134
12082 PR middle-end/57586
12083 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
12084 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
12085 bitfield expansion when EXPAND_MEMORY.
12086 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
12088 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
12090 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
12091 test for clearing quad memory on 32-bit later.
12093 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
12095 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
12096 (fold_negate_expr): Likewise.
12097 (fold_real_zero_addition_p): Handle vectors.
12098 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
12100 2013-06-14 Alan Modra <amodra@gmail.com>
12102 * varasm.c (force_const_mem): Revert 2013-06-07 change.
12104 2013-06-13 Jan Hubicka <jh@suse.cz>
12106 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
12107 Local comdats are not externally visible.
12108 * symtab.c (dump_symtab_base): Dump externally visible.
12109 (verify_symtab_base): Verify back links in the symtab hash.
12111 2013-06-13 Bin Cheng <bin.cheng@arm.com>
12113 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
12114 CONVERT_EXPR as equal nodes.
12116 2013-06-13 Bin Cheng <bin.cheng@arm.com>
12118 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
12120 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
12122 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
12123 Generalize to complex and vector.
12124 * tree.c (build_all_ones_cst): New function.
12125 * tree.h (build_all_ones_cst): Declare it.
12127 2013-06-13 Alan Modra <amodra@gmail.com>
12129 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
12130 * config/rs6000/rs6000.md (signbittf2): New insn.
12131 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
12132 (abstf2_internal, cmptf_internal2): Likewise.
12133 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
12135 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
12136 Pat Haugen <pthaugen@us.ibm.com>
12137 Peter Bergner <bergner@vnet.ibm.com>
12139 * config/rs6000/rs6000.c (emit_load_locked): Add support for
12140 power8 byte, half-word, and quad-word atomic instructions.
12141 (emit_store_conditional): Likewise.
12142 (rs6000_expand_atomic_compare_and_swap): Likewise.
12143 (rs6000_expand_atomic_op): Likewise.
12145 * config/rs6000/sync.md (larx): Add new modes for power8.
12147 (AINT): New mode iterator to include TImode as well as normal
12148 integer modes on power8.
12149 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
12150 that VSX registers are not considered. Use AINT mode iterator
12151 instead of INT1 to allow inclusion of quad word atomic operations
12153 (load_locked<mode>): Likewise.
12154 (store_conditional<mode>): Likewise.
12155 (atomic_compare_and_swap<mode>): Likewise.
12156 (atomic_exchange<mode>): Likewise.
12157 (atomic_nand<mode>): Likewise.
12158 (atomic_fetch_<fetchop_name><mode>): Likewise.
12159 (atomic_nand_fetch<mode>): Likewise.
12160 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
12162 (ATOMIC): On power8, add QImode, HImode modes.
12163 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
12164 modes that promote to SImode.
12165 (load_lockedti): Convert TImode arguments to PTImode, so that we
12166 get a guaranteed even/odd register pair.
12167 (load_lockedpti): Likewise.
12168 (store_conditionalti): Likewise.
12169 (store_conditionalpti): Likewise.
12171 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
12172 atomic load/store instructions.
12175 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
12177 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
12179 (insn_count): New attribute, with most cases extracted from...
12180 (length): ...here. Redefine most cases in terms of insn_count.
12181 (single_insn): Delete.
12182 (can_delay): Use insn_count to check for single instructions.
12183 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
12184 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
12185 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
12186 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
12187 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
12188 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
12189 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
12190 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
12191 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
12192 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
12193 rather than "length".
12194 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
12195 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
12196 Use "insn_count" rather than "length".
12197 * config/mips/mips-dsp.md
12198 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
12199 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
12202 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
12204 PR tree-optimization/57361
12205 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
12207 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
12209 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
12211 (aarch64_simd_combine<mode>): New instruction expansion.
12212 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
12213 function prototype.
12214 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
12215 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
12217 2013-06-12 Jan Hubicka <jh@suse.cz>
12219 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
12220 decl has when in streaming stage.
12221 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
12222 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
12224 2013-06-12 Roland Stigge <stigge@antcom.de>
12227 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
12229 2013-06-12 Jakub Jelinek <jakub@redhat.com>
12231 PR tree-optimization/57537
12232 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
12233 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
12235 2013-06-12 Richard Biener <rguenther@suse.de>
12237 * data-streamer.h (streamer_write_char_stream): CSE
12238 obs->current_pointer.
12239 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
12240 streamer_write_char_stream manually and optimize the resulting loop.
12241 (streamer_write_hwi_stream): Likewise.
12243 2013-06-12 Jan Hubicka <jh@suse.cz>
12245 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
12246 * cgraph.h (varpool_create_empty_node): Declare.
12247 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
12249 * symtab.c (symtab_unregister_node): Be lax about missin entries
12251 (symtab_get_node): Update comment.
12252 * varpool.c (varpool_create_empty_node): Break out from ...
12253 (varpool_node_for_decl): ... here.
12254 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
12256 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
12258 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
12259 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
12260 part. Use straight-line flow at the end.
12261 <COMPONENT_REF>: Remove superfluous else.
12262 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
12264 2013-06-12 Jakub Jelinek <jakub@redhat.com>
12267 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
12268 target hook even for !TREE_PUBLIC decls. If no resolution info
12269 is available, return false for common and external decls.
12271 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
12273 * config/rl78/constraints.md (U): New constraint.
12274 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
12277 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
12280 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
12281 to allow returning address to AT_PLATFORM name.
12283 2013-06-11 Jan Hubicka <jh@suse.cz>
12285 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
12286 * cgraph.h (symtab_node_base): Add weakref flag.
12287 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
12288 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
12289 (output_weakrefs): Use weakref flag.
12290 * fold-const.c (simple_operand_p): Handle WEAK.
12291 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
12292 * ipa.c (varpool_externally_visible_p): Drop weakref.
12293 (function_and_variable_visibility): Update comment; fix weakref
12294 sanity checks; do not clear DECL_WEAK on them.
12295 * lto-cgraph.c (lto_output_node): update.
12296 (lto_output_varpool_node): Update.
12297 (input_overwrite_node): Update.
12298 (input_node): Update.
12299 (input_varpool_node): Update.
12300 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
12301 (lto_symtab_merge_symbols): Add sanity check.
12302 (lto_symtab_prevailing_decl): Do not special case weakrefs.
12303 * passes.c (rest_of_decl_compilation): Set static flag, too.
12304 * symtab.c (dump_symtab_base): Dump weakref.
12305 (verify_symtab_base): Sanity check weakrefs.
12306 (symtab_make_decl_local): Remove duplicated code.
12307 (symtab_alias_ultimate_target): Simplify.
12308 * varpool.c (varpool_create_variable_alias): Set weakref flag.
12310 2013-06-11 Tom de Vries <tom@codesourcery.com>
12312 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
12313 sequence_vect == NULL.
12315 2013-06-11 DJ Delorie <dj@redhat.com>
12317 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
12318 (rl78_unwind_word_mode): New.
12320 2013-06-11 David Malcolm <dmalcolm@redhat.com>
12322 * final.c (debug_prefix_maps): Make static.
12324 2013-06-11 David Malcolm <dmalcolm@redhat.com>
12326 * function.c (initial_trampoline): Remove stray copy.
12328 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
12330 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
12332 2013-06-11 Martin Jambor <mjambor@suse.cz>
12334 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
12335 within bounds at the beginning of the function.
12337 2013-06-11 Alan Modra <amodra@gmail.com>
12339 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
12341 (get_named_section): Don't NULL !DECL_P decl.
12343 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
12345 * doc/invoke.texi (core-avx2): Document.
12347 (atom): Updated with MOVBE.
12349 2013-06-11 Richard Biener <rguenther@suse.de>
12351 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
12353 2013-06-11 Anton Blanchard <anton@samba.org>
12355 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
12356 correct shift value in little-endian mode.
12358 2013-06-11 Jakub Jelinek <jakub@redhat.com>
12361 * varasm.c (get_variable_align): Move #endif to the right place.
12363 2013-06-10 Cary Coutant <ccoutant@google.com>
12365 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
12366 for hash so that hash table traversal order is deterministic.
12368 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
12369 Pat Haugen <pthaugen@us.ibm.com>
12370 Peter Bergner <bergner@vnet.ibm.com>
12372 * config/rs6000/vector.md (GPR move splitter): Do not split moves
12373 of vectors in GPRS if they are direct moves or quad word load or
12376 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
12378 (direct_move_p): Likewise.
12379 (quad_load_store_p): Likewise.
12381 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
12382 classes into bins based on the physical register type.
12383 (reg_class_to_reg_type): Likewise.
12384 (IS_STD_REG_TYPE): Likewise.
12385 (IS_FP_VECT_REG_TYPE): Likewise.
12386 (reload_fpr_gpr): Arrays to determine what insn to use if we can
12387 use direct move instructions.
12388 (reload_gpr_vsx): Likewise.
12389 (reload_vsx_gpr): Likewise.
12390 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
12391 information that is a simplification of register classes. Also
12392 precalculate direct move reload helpers.
12393 (direct_move_p): New function to return true if the operation can
12394 be done as a direct move instruciton.
12395 (quad_load_store_p): New function to return true if the operation
12396 is a quad memory operation.
12397 (rs6000_legitimize_address): If quad memory, only allow register
12398 indirect for TImode addresses.
12399 (rs6000_legitimate_address_p): Likewise.
12400 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
12401 (rs6000_reload_register_type): Likewise.
12402 (register_to_reg_type): Return register type.
12403 (rs6000_secondary_reload_simple_move): New helper function for
12404 secondary reload and secondary memory needed to identify anything
12405 that is a simple move, and does not need reloading.
12406 (rs6000_secondary_reload_direct_move): New helper function for
12407 secondary reload to identify cases that can be done with several
12408 instructions via the direct move instructions.
12409 (rs6000_secondary_reload_move): New helper function for secondary
12410 reload to identify moves between register types that can be done.
12411 (rs6000_secondary_reload): Add support for quad memory operations
12412 and for direct move.
12413 (rs6000_secondary_memory_needed): Likewise.
12414 (rs6000_debug_secondary_memory_needed): Change argument names.
12415 (rs6000_output_move_128bit): New function to return the move to
12416 use for 128-bit moves, including knowing about the various
12417 limitations of quad memory operations.
12419 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
12420 memory operations. call rs6000_output_move_128bit for the actual
12421 instruciton(s) to generate.
12422 (vsx_movti_64bit): Likewise.
12424 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
12425 (UNSPEC_P8V_MTVSRWZ): Likewise.
12426 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
12427 (UNSPEC_P8V_MTVSRD): Likewise.
12428 (UNSPEC_P8V_XXPERMDI): Likewise.
12429 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
12430 (UNSPEC_FUSION_GPR): Likewise.
12431 (FMOVE128_GPR): New iterator for direct move.
12432 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
12433 (f32_sv): Likewise.
12434 (f32_dm): Likewise.
12435 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
12436 loads and direct move instructions.
12437 (zero_extendsidi2_lfiwzx): Likewise.
12438 (extendsidi2_lfiwax): Likewise.
12439 (extendsidi2_nocell): Likewise.
12440 (floatsi<mode>2_lfiwax): Likewise.
12441 (lfiwax): Likewise.
12442 (floatunssi<mode>2_lfiwzx): Likewise.
12443 (lfiwzx): Likewise.
12444 (fix_trunc<mode>_stfiwx): Likewise.
12445 (fixuns_trunc<mode>_stfiwx): Likewise.
12446 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
12447 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
12448 (parity<mode>2_cmpb): Set length/type attr.
12449 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
12450 for 'mr.' to fast_compare.
12451 (bpermd_<mode>): Change type attr to popcnt.
12452 (p8_fmrgow_<mode>): New insns for power8 direct move support.
12453 (p8_mtvsrwz_1): Likewise.
12454 (p8_mtvsrwz_2): Likewise.
12455 (reload_fpr_from_gpr<mode>): Likewise.
12456 (p8_mtvsrd_1): Likewise.
12457 (p8_mtvsrd_2): Likewise.
12458 (p8_xxpermdi_<mode>): Likewise.
12459 (reload_vsx_from_gpr<mode>): Likewise.
12460 (reload_vsx_from_gprsf): Likewise.
12461 (p8_mfvsrd_3_<mode>): LIkewise.
12462 (reload_gpr_from_vsx<mode>): Likewise.
12463 (reload_gpr_from_vsxsf): Likewise.
12464 (p8_mfvsrd_4_disf): Likewise.
12465 (multi-word GPR splits): Do not split direct moves or quad memory
12468 2013-06-10 David Malcolm <dmalcolm@redhat.com>
12470 * tree-into-ssa.c (interesting_blocks): Make static.
12472 2013-06-10 Jakub Jelinek <jakub@redhat.com>
12475 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
12476 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
12477 Use DATA_ABI_ALIGNMENT for that case instead if defined.
12478 (get_variable_align): New function.
12479 (get_variable_section, emit_bss, emit_common,
12480 assemble_variable_contents, place_block_symbol): Use
12481 get_variable_align instead of DECL_ALIGN.
12482 (assemble_noswitch_variable): Add align argument, use it
12483 instead of DECL_ALIGN.
12484 (assemble_variable): Adjust caller. Use get_variable_align
12485 instead of DECL_ALIGN.
12486 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
12488 (DATA_ABI_ALIGNMENT): Define.
12489 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
12490 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
12491 opt is false, only return the psABI mandated alignment increase.
12492 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
12493 (DATA_ABI_ALIGNMENT): ... this.
12494 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
12495 (DATA_ABI_ALIGNMENT): ... this.
12496 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
12497 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
12498 (DATA_ABI_ALIGNMENT): ... this.
12499 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
12500 * doc/tm.texi: Regenerated.
12502 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
12504 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
12505 cmp_code to construct REG_EQUAL note.
12507 2013-06-09 Jakub Jelinek <jakub@redhat.com>
12510 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
12511 that operands[2] doesn't overlap with operands[0].
12513 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
12514 Jan Hubicka <jh@suse.cz>
12516 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
12517 hack to mark symbols as used.
12519 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
12521 PR rtl-optimization/57559
12522 * lra-constraints.c (process_alt_operands): Don't discourage
12523 memory with known offset for offsetable memory constraint.
12524 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
12526 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
12528 * varasm.c (struct oc_local_state): Reorder fields.
12529 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
12530 and adjust accordingly.
12531 (output_constructor): Reorder initialization code and adjust call to
12532 output_constructor_bitfield.
12534 2013-06-07 Jan Hubicka <jh@suse.cz>
12536 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
12538 2013-06-07 David Malcolm <dmalcolm@redhat.com>
12540 * tree-object-size.c (unknown): Make const.
12542 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12544 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
12545 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
12546 for last alternative in the cpu_facility attribute.
12548 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12551 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
12552 (xordi3): Change operand 2 constraint to arm_xordi_operand.
12553 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
12554 * config/arm/constraints.md (Dg): New constraint.
12555 * config/arm/neon.md (xordi3_neon): Remove.
12556 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
12557 * config/arm/predicates.md (arm_xordi_operand): New predicate.
12559 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12561 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
12562 Clean up alternatives.
12564 2013-06-07 Alan Modra <amodra@gmail.com>
12566 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
12569 2013-06-07 Alan Modra <amodra@gmail.com>
12571 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
12573 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12575 * config/arm/constraints.md (Df): New constraint.
12576 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
12577 Correct length attribute for last two alternatives.
12579 2013-06-07 Alan Modra <amodra@gmail.com>
12581 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
12582 override user -mfp-in-toc.
12583 (offsettable_ok_by_alignment): Consider just the current access
12584 rather than the whole object, unless BLKmode. Handle
12585 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
12586 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
12587 for -mcmodel=medium.
12588 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
12589 override user -mfp-in-toc or -msum-in-toc. Default to
12590 -mno-fp-in-toc for -mcmodel=medium.
12592 2013-06-06 DJ Delorie <dj@redhat.com>
12594 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
12595 TARGET_VALID_POINTER_MODE.
12597 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
12598 Pat Haugen <pthaugen@us.ibm.com>
12599 Peter Bergner <bergner@vnet.ibm.com>
12601 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
12602 Document new power8 builtins.
12604 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
12605 condition code register, to allow 128-bit logical operations to be
12606 done in the VSX or GPR registers.
12607 (nor<mode>3): Use the canonical form for nor.
12608 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
12609 vclz*, and vpopcnt* vector instructions.
12610 (nand<mode>3): Likewise.
12611 (orc<mode>3): Likewise.
12612 (clz<mode>2): LIkewise.
12613 (popcount<mode>2): Likewise.
12615 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
12616 that only the GPRs are recognized.
12618 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
12619 support for new power8 builtins.
12621 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
12623 (xscvdpspn): Likewise.
12629 (vpopcnt): Likewise.
12630 (vpopcntb): Likewise.
12631 (vpopcnth): Likewise.
12632 (vpopcntw): Likewise.
12633 (vpopcntd): Likewise.
12635 (vmrgew): Likewise.
12636 (vmrgow): Likewise.
12638 (eqv_v16qi3): Likewise.
12639 (eqv_v8hi3): Likewise.
12640 (eqv_v4si3): Likewise.
12641 (eqv_v2di3): Likewise.
12642 (eqv_v4sf3): Likewise.
12643 (eqv_v2df3): Likewise.
12645 (nand_v16qi3): Likewise.
12646 (nand_v8hi3): Likewise.
12647 (nand_v4si3): Likewise.
12648 (nand_v2di3): Likewise.
12649 (nand_v4sf3): Likewise.
12650 (nand_v2df3): Likewise.
12652 (orc_v16qi3): Likewise.
12653 (orc_v8hi3): Likewise.
12654 (orc_v4si3): Likewise.
12655 (orc_v2di3): Likewise.
12656 (orc_v4sf3): Likewise.
12657 (orc_v2df3): Likewise.
12659 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
12660 allow power8 quad mode in 64-bit.
12661 (rs6000_builtin_vectorized_function): Add support to vectorize
12662 ISA 2.07 count leading zeros, population count builtins.
12663 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
12664 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
12665 (builtin_function_type): Add vgbbd builtin function which takes an
12667 (altivec_expand_vec_perm_const): Add support for new power8 merge
12670 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
12671 that does not include TImdoe for use with 32-bit.
12672 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
12674 (UNSPEC_VSX_CVDPSPN): Likewise.
12675 (vsx_xscvdpspn): Likewise.
12676 (vsx_xscvspdpn): Likewise.
12677 (vsx_xscvdpspn_scalar): Likewise.
12678 (vsx_xscvspdpn_directmove): Likewise.
12679 (vsx_and<mode>3): Split logical operations into 32-bit and
12680 64-bit. Add support to do logical operations on TImode as well as
12681 VSX vector types. Allow logical operations to be done in either
12682 VSX registers or in general purpose registers in 64-bit mode. Add
12683 splitters if GPRs were used. For AND, add clobber of CCmode to
12684 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
12686 (vsx_and<mode>3_32bit): Likewise.
12687 (vsx_and<mode>3_64bit): Likewise.
12688 (vsx_ior<mode>3): Likewise.
12689 (vsx_ior<mode>3_32bit): Likewise.
12690 (vsx_ior<mode>3_64bit): Likewise.
12691 (vsx_xor<mode>3): Likewise.
12692 (vsx_xor<mode>3_32bit): Likewise.
12693 (vsx_xor<mode>3_64bit): Likewise.
12694 (vsx_one_cmpl<mode>2): Likewise.
12695 (vsx_one_cmpl<mode>2_32bit): Likewise.
12696 (vsx_one_cmpl<mode>2_64bit): Likewise.
12697 (vsx_nor<mode>3): Likewise.
12698 (vsx_nor<mode>3_32bit): Likewise.
12699 (vsx_nor<mode>3_64bit): Likewise.
12700 (vsx_andc<mode>3): Likewise.
12701 (vsx_andc<mode>3_32bit): Likewise.
12702 (vsx_andc<mode>3_64bit): Likewise.
12703 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
12704 and xxlorc instructions.
12705 (vsx_eqv<mode>3_64bit): Likewise.
12706 (vsx_nand<mode>3_32bit): Likewise.
12707 (vsx_nand<mode>3_64bit): Likewise.
12708 (vsx_orc<mode>3_32bit): Likewise.
12709 (vsx_orc<mode>3_64bit): Likewise.
12711 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
12713 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
12715 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
12716 (p8_vmrgow): Likewise.
12717 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
12718 GPRs to be split under VSX.
12719 (p8v_clz<mode>2): Add power8 count leading zero support.
12720 (p8v_popcount<mode>2): Add power8 population count support.
12721 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
12724 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
12727 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
12729 (vec_nand): Likewise.
12730 (vec_vclz): Likewise.
12731 (vec_vclzb): Likewise.
12732 (vec_vclzd): Likewise.
12733 (vec_vclzh): Likewise.
12734 (vec_vclzw): Likewise.
12735 (vec_vgbbd): Likewise.
12736 (vec_vmrgew): Likewise.
12737 (vec_vmrgow): Likewise.
12738 (vec_vpopcnt): Likewise.
12739 (vec_vpopcntb): Likewise.
12740 (vec_vpopcntd): Likewise.
12741 (vec_vpopcnth): Likewise.
12742 (vec_vpopcntw): Likewise.
12744 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
12746 PR rtl-optimization/57468
12747 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
12750 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
12752 PR rtl-optimization/57459
12753 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
12754 type when setting live regs.
12756 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
12758 * config/s390/s390.opt (mlra): New option.
12759 * config/s390/s390.c (s390_decompose_address): Check displacement
12760 for all registers for LRA.
12761 (s390_secondary_reload): Don't used secondary reloads for LRA.
12762 (s390_lra_p): New function.
12763 (TARGET_LRA_P): Define.
12764 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
12765 of attribute cpu_facility to zarch for the last alternative.
12766 (*cmpmem_short): Ditto.
12768 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
12770 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
12771 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
12772 (arm_expand_prologue): Likewise.
12774 2013-06-06 Teresa Johnson <tejohnson@google.com>
12777 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
12778 as this is now done by redirect_edge_and_branch_force.
12779 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
12780 barriers, and fix interaction with splitting.
12781 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
12782 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
12783 reflect changes made in the routine.
12784 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
12785 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
12786 since this is called in cfglayout mode, and replace partition fixup
12787 with assert as that is now done by force_nonfallthru_and_redirect.
12788 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
12789 already be marked with region crossing note.
12790 (insert_section_boundary_note): Make non-static, gate on flag
12791 has_bb_partition, rewrite to also check for multiple partitions.
12792 (rest_of_handle_reorder_blocks): Remove call to
12793 insert_section_boundary_note, now done later during free_cfg.
12794 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
12795 * bb-reorder.h (insert_section_boundary_note): Declare.
12796 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
12797 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
12798 invoke insert_section_boundary_note.
12799 (try_redirect_by_replacing_jump): Remove unnecessary
12800 check for region crossing note.
12801 (fixup_partition_crossing): New function.
12802 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
12803 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
12804 in non-cfglayout mode.
12805 (force_nonfallthru_and_redirect): Fixup partition boundaries,
12806 remove old code that tried to do this. Emit barrier correctly
12807 when we are in cfglayout mode.
12808 (last_bb_in_partition): New function.
12809 (rtl_split_edge): Correctly fixup partition boundaries.
12810 (commit_one_edge_insertion): Remove old code that tried to
12811 fixup region crossing edge since this is now handled in
12812 split_block, and set up insertion point correctly since
12813 block may now end in a jump.
12814 (verify_hot_cold_block_grouping): Guard against checking when not in
12815 linearized RTL mode.
12816 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
12818 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
12819 rtl_verify_flow_info, so not called in cfglayout mode.
12820 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
12821 (fixup_reorder_chain): Remove old code that attempted to fixup region
12822 crossing note as this is now handled in force_nonfallthru_and_redirect.
12823 (duplicate_insn_chain): Don't duplicate switch section notes.
12824 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
12826 * basic-block.h (emit_barrier_after_bb): Declare.
12828 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12830 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
12831 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
12832 arm_usatsihi): Adjust alternatives for arm_restrict_it.
12834 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12836 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
12838 * config/arm/ldmstm.md: Regenerate.
12840 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12842 * config/arm/sync.md (atomic_loaddi_1):
12843 Disable predication for arm_restrict_it.
12844 (arm_load_exclusive<mode>): Likewise.
12845 (arm_load_exclusivesi): Likewise.
12846 (arm_load_exclusivedi): Likewise.
12847 (arm_load_acquire_exclusive<mode>): Likewise.
12848 (arm_load_acquire_exclusivesi): Likewise.
12849 (arm_load_acquire_exclusivedi): Likewise.
12850 (arm_store_exclusive<mode>): Likewise.
12851 (arm_store_exclusive<mode>): Likewise.
12852 (arm_store_release_exclusivedi): Likewise.
12853 (arm_store_release_exclusive<mode>): Likewise.
12855 2013-06-06 Richard Biener <rguenther@suse.de>
12857 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
12859 (lto_tag_is_tree_code_p): Adjust.
12860 (lto_tag_is_gimple_code_p): Likewise.
12861 (lto_gimple_code_to_tag): Likewise.
12862 (lto_tag_to_gimple_code): Likewise.
12863 (lto_tree_code_to_tag): Likewise.
12864 (lto_tag_to_tree_code): Likewise.
12865 * data-streamer.h (streamer_write_hwi_in_range): Use
12866 uhwi streaming to stream the normalized range.
12867 (streamer_read_hwi_in_range): Likewise.
12869 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12871 * config/arm/arm.md (enabled_for_depr_it): New attribute.
12872 (predicable_short_it): Likewise.
12873 (predicated): Likewise.
12874 (enabled): Handle above.
12875 (define_cond_exec): Set predicated attribute to yes.
12877 2013-06-05 Mike Stump <mikestump@comcast.net>
12879 * gdbinit.in (__FUNCTION__): Add.
12881 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
12883 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
12884 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
12886 2013-06-05 Jan Hubicka <jh@suse.cz>
12888 * varasm.c (mark_decl_referenced): Revert the removal until targets
12891 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
12893 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
12894 instead of mark_decl_referenced.
12896 2013-06-05 Jan Hubicka <jh@suse.cz>
12898 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
12899 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
12900 and symtab_used_from_object_file_p.
12901 (cgraph_make_node_local_1): Clear forced_by_abi.
12902 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
12903 * cgraph.h (symtab_node_base): Add forced_by_abi.
12904 (decide_is_variable_needed): Remove.
12905 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
12906 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
12907 (decide_is_symbol_needed): ... this one; handle symbols in general;
12908 always analyze virtuals; honnor forced_by_abi.
12909 (cgraph_finalize_function): Update.
12910 (varpool_finalize_decl): Update.
12911 (symbol_defined_and_needed): Remove.
12912 (analyze_functions): Update.
12913 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
12914 output_refs, input_overwrite_node): Handle forced_by_abi.
12915 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
12916 (address_taken_from_non_vtable_p): ... this one.
12917 (comdat_can_be_unshared_p_1): New function.
12918 (cgraph_comdat_can_be_unshared_p): Rename to ...
12919 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
12920 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
12921 (function_and_variable_visibility): Clear forced_by_abi as needed.
12922 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
12923 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
12924 * symtab.c (dump_symtab_base): Dump forced_by_abi.
12925 * varpool.c (decide_is_variable_needed): Remove.
12927 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12929 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
12930 (arm_option_override): Override arm_restrict_it where appropriate.
12931 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
12932 * config/arm/arm.opt (mrestrict-it): New command-line option.
12933 * doc/invoke.texi: Document -mrestrict-it.
12935 2013-06-05 David Malcolm <dmalcolm@redhat.com>
12937 * tsan.c (tsan_atomic_table): Make const.
12939 2013-06-05 Richard Biener <rguenther@suse.de>
12941 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
12942 index associated with the tree we are supposed to replace.
12943 * tree-streamer-out.c (pack_ts_base_value_fields): Output
12944 TREE_ASM_WRITTEN as zero for everything but SSA names.
12946 2013-06-05 David Malcolm <dmalcolm@redhat.com>
12948 * tree-ssa-structalias.c (call_stmt_vars): Make static.
12950 2013-06-04 Jan Hubicka <jh@suse.cz>
12952 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
12953 (input_node, input_varpool_node): Handle correctly external same
12955 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
12956 nodes at ltrans stage.
12958 2013-06-04 Jan Hubicka <jh@suse.cz>
12960 * ipa-inline.c (update_caller_keys): Fix availability test.
12961 (update_callee_keys): Likewise.
12962 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
12963 to follow ELF standard.
12965 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
12967 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
12968 (mips64r5900el-*-elf*): New configurations.
12969 * config/mips/mips-cpus.def (r5900): New processor.
12970 * config/mips/mips-tables.opt: Regenerate.
12971 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
12972 (mips_issue_rate): Handle PROCESSOR_R5900.
12973 (mips_reorg_process_insns): Force reorder mode for the R5900.
12974 * config/mips/mips.h (TARGET_MIPS5900): Define.
12975 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
12977 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
12979 * config/mips/mips.md (processor): Add r5900.
12980 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
12982 2013-06-04 Ian Bolton <ian.bolton@arm.com>
12984 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
12985 into function to generate MOVI instruction.
12986 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
12987 (aarch64_preferred_simd_mode): Turn into wrapper.
12988 (aarch64_output_scalar_simd_mov_immediate): New function.
12989 * config/aarch64/aarch64-protos.h: Add prototype for above.
12991 2013-06-04 Ian Bolton <ian.bolton@arm.com>
12993 * config/aarch64/aarch64.c (simd_immediate_info): Remove
12994 element_char member.
12995 (sizetochar): Return signed char.
12996 (aarch64_simd_valid_immediate): Remove elchar and other
12997 unnecessary variables.
12998 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
12999 Calculate element_char as required.
13000 * config/aarch64/aarch64-protos.h: Update and move prototype
13001 for aarch64_output_simd_mov_immediate.
13002 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
13005 2013-06-04 Ian Bolton <ian.bolton@arm.com>
13007 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
13008 information completed by aarch64_simd_valid_immediate.
13009 (aarch64_legitimate_constant_p): Update arguments.
13010 (aarch64_simd_valid_immediate): Work with struct rather than many
13012 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
13013 (aarch64_simd_make_constant): Update arguments.
13014 (aarch64_output_simd_mov_immediate): Work with struct rather than
13015 many pointers. Output immediate directly rather than as operand.
13016 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
13018 * config/aarch64/constraints.md (Dn): Update arguments.
13020 2013-06-04 Ian Bolton <ian.bolton@arm.com>
13022 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
13024 (aarch64_simd_immediate_valid_for_move): Remove.
13025 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
13026 (aarch64_simd_make_constant): Update call.
13027 (aarch64_output_simd_mov_immediate): Update call.
13028 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
13030 * config/aarch64/constraints.md (Dn): Update call.
13032 2013-06-04 Ian Bolton <ian.bolton@arm.com>
13034 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
13035 return type to bool for prototype.
13036 (aarch64_legitimate_constant_p): Check for true instead of not -1.
13037 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
13038 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
13040 2013-06-04 Catherine Moore <clm@codesourcery.com>
13042 * config/mips/mips.opt (meva): New.
13043 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
13044 (ASM_SPEC): Handle -meva.
13045 * doc/invoke.texi (meva): Document.
13047 2013-06-04 Alan Modra <amodra@gmail.com>
13049 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
13052 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13054 * rtl.def: Add extra fourth optional field to define_cond_exec.
13055 * gensupport.c (process_one_cond_exec): Process attributes from
13057 * doc/md.texi: Document fourth field in define_cond_exec.
13059 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
13061 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
13062 out the processing order as in store_bit_field_1.
13064 2013-06-04 Jan Hubicka <jh@suse.cz>
13066 PR middle-end/57500
13067 * cgraphunit.c (cgraph_process_same_body_aliases): Create
13068 non-VAR_DECL node if it does not exist yet.
13070 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
13072 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
13073 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
13074 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
13075 target_cpu_default setting.
13077 2013-06-03 Teresa Johnson <tejohnson@google.com>
13079 * dumpfile.c (opt_info_switch_p): Change -fopt-info
13080 default to -fopt-info=optimized instead of all.
13081 * doc/invoke.texi: Ditto.
13082 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
13083 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
13084 (execute_vect_slp): Emit BB vectorization success under
13085 MSG_OPTIMIZED_LOCATIONS.
13086 * tree-vect-slp.c (vect_slp_transform_bb): Change
13087 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
13088 * tree-vect-loop.c (vect_transform_loop): Ditto.
13090 2013-06-03 Jason Merrill <jason@redhat.com>
13093 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
13094 Use TARGET_EXPR for C++.
13096 2013-06-03 Jakub Jelinek <jakub@redhat.com>
13098 PR rtl-optimization/57268
13099 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
13100 if DEBUG_INSN_P (insn).
13103 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
13105 PR rtl-optimization/57268
13106 * sched-deps.c (sched_analyze_2): Flush dependence lists if
13107 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
13109 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
13111 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
13112 (ix86_avoid_lea_for_addr): Likewise.
13113 (exact_dependency_1): Likewise.
13114 (ix86_adjust_cost): Likewise.
13115 (swap_top_of_ready_list): Fix formatting and !reload_completed check
13117 (do_reorder_for_imul): Fix typo, formatting and
13118 !reload_completed check removed.
13119 (ix86_sched_reorder): Fix typo and formatting.
13120 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
13123 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
13125 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
13127 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
13129 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
13130 <STRING_CST>: Likewise.
13131 <VECTOR_CST>: Likewise.
13133 2013-06-01 Janus Weil <janus@gcc.gnu.org>
13134 Mikael Morin <mikael@gcc.gnu.org>
13136 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
13137 * config.in: Regenerated.
13138 * configure: Regenerated.
13140 2013-06-01 Jan Hubicka <jh@suse.cz>
13142 PR middle-end/57366
13143 * cgraphunit.c (compile): When weakref is not supported,
13144 set up transparent aliases before final output pass.
13145 * varasm.c (assemble_alias): Do not try to do it here.
13147 2013-06-01 Jan Hubicka <jh@suse.cz>
13149 PR middle-end/57467
13150 * passes.c (for_per_function): Skip unanalyzed functions.
13152 2013-06-01 Jan Hubicka <jh@suse.cz>
13154 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
13155 (lto_symtab_merge_symbols_1): ... this one.
13156 (lto_symtab_merge_cgraph_nodes): Rename to ...
13157 (lto_symtab_merge_symbols): ... this one; simplify.
13158 * cgraph.c (same_body_aliases_done): Rename to ...
13159 (cpp_implicit_aliases_done): ... this one.
13160 (cgraph_create_function_alias): Update.
13161 (cgraph_same_body_alias): Update.
13162 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
13163 (verify_edge_corresponds_to_fndecl): Simplify.
13164 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
13165 (cgraph_node): Remove same_body_alias.
13166 (varpool_node): Remove alias_of and extra_name_alias.
13167 (same_body_aliases_done): Rename to ..
13168 (cpp_implicit_aliases_done): ... this one.
13169 (symtab_alias_ultimate_target): Add default parameter.
13170 (symtab_resolve_alias): New function.
13171 (fixup_same_cpp_alias_visibility): Declare.
13172 (cgraph_function_node): Add default parameter.
13173 (cgraph_node_asm_name): Likewise.
13174 (cgraph_function_or_thunk_node): Add default parameter; do
13175 not ICE when it is NULL.
13176 (varpool_variable_node): Likewise.
13177 * tree-emutls.c (create_emultls_var): Update.
13178 (ipa_lower_emutls): Update.
13179 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
13180 (cgraph_reset_node): Reset alias info.
13181 (cgraph_finalize_function): Update.
13182 (fixup_same_cpp_alias_visibility): Move to symtab.c.
13183 (analyze_function): Simplify.
13184 (cgraph_process_same_body_aliases): Simplify.
13185 (analyze_functions): Fixup same body aliases.
13186 (handle_alias_pairs): Simplify.
13187 (assemble_thunk): Update.
13188 (assemble_thunks_and_aliases): Update.
13189 (output_weakrefs): Rewrite.
13190 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
13191 (lto_output_varpool_node): Likewise.
13192 (compute_ltrans_boundary): Remve assert.
13193 (get_alias_symbol): New functoin.
13194 (input_node): Rewrite alias handling.
13195 (input_varpool_node): Likewise.
13196 * ipa-pure-const.c (propagate_pure_const): Fix formating.
13197 * ipa.c (process_references): Handle weakrefs correctly.
13198 (symtab_remove_unreachable_nodes): Likewise.
13199 * trans-mem.c (get_cg_data): Update.
13200 (ipa_tm_create_version_alias): Update.
13201 (ipa_tm_execute): Update.
13202 * symtab.c (dump_symtab_base): Dump aliases.
13203 (verify_symtab_base): Verify aliases.
13204 (symtab_node_availability): New function.
13205 (symtab_alias_ultimate_target): Simplify.
13206 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
13207 handle all the fixup cases.
13208 (symtab_resolve_alias): New function.
13209 * passes.c (ipa_write_summaries): Handle weakrefs.
13210 * varpool.c (varpool_analyze_node): Simplify.
13211 (assemble_aliases): Update.
13212 (varpool_create_variable_alias): Simplify.
13213 (varpool_extra_name_alias): Simplify.
13214 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
13215 (lto_symtab_merge_symbols): ... this one.
13217 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
13220 PR rtl-optimization/57268
13221 * sched-deps.c (sched_analyze_2): Flush dependence lists if
13222 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
13224 2013-06-01 Tobias Burnus <burnus@net-b.de>
13226 Partially reverted:
13227 2013-05-31 Tobias Burnus <burnus@net-b.de>
13229 PR middle-end/57073
13230 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
13233 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
13235 PR rtl-optimization/57268
13236 * sched-deps.c (sched_analyze_2): Flush dependence lists if
13237 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
13239 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
13241 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
13242 unordered comparison operators when -fno-trapping-math is in effect
13244 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
13245 and implement unordered comparison operators properly on the e500.
13247 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
13249 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
13250 for constant scalar integers.
13251 (simplify_relational_operation_1): Likewise.
13253 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
13255 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
13256 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
13259 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
13260 Igor Zamyatin <igor.zamyatin@intel.com>
13262 Silvermont (SLM) architecture performance tuning.
13263 * config/i386/i386.h (enum ix86_tune_indices): Add
13264 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
13265 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
13267 * config/i386/i386.c (initial_ix86_tune_features)
13268 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
13269 (ix86_lea_outperforms): Handle Silvermont tuning.
13270 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
13272 (ix86_use_lea_for_mov): Likewise.
13273 (ix86_avoid_lea_for_addr): Likewise.
13274 (ix86_lea_for_add_ok): Likewise.
13275 (exact_dependency_1): New function.
13276 (exact_store_load_dependency): Likewise.
13277 (ix86_adjust_cost): Handle Silvermont tuning.
13278 (do_reoder_for_imul): Likewise.
13279 (swap_top_of_ready_list): New function.
13280 (ix86_sched_reorder): Changed to handle Silvermont tuning.
13282 * config/i386/i386.md (peepholes that split memory operand in fp
13285 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
13287 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
13288 Remove un-necessary braces.
13290 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
13292 * config/aarch64/aarch64.c (aarch64_classify_symbol):
13293 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
13295 2013-05-31 Tobias Burnus <burnus@net-b.de>
13297 PR middle-end/57073
13298 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
13300 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13303 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
13304 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
13305 * config/arm/neon.md (iordi3_neon): Remove.
13306 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
13307 * config/arm/predicates.md (imm_for_neon_logic_operand):
13308 Move to earlier in the file.
13309 (neon_logic_op2): Likewise.
13310 (arm_iordi_operand_neon): New predicate.
13312 2013-05-31 Richard Biener <rguenther@suse.de>
13314 PR tree-optimization/57478
13315 PR tree-optimization/57453
13316 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
13319 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
13321 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
13322 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
13324 2013-05-30 Tobias Burnus <burnus@net-b.de>
13325 Thomas Koenig <tkoenig@gcc.gnu.org>
13327 PR middle-end/57073
13328 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
13329 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
13331 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
13333 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
13335 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
13337 * target.def (register_usage_leveling_p): New hook.
13338 * targhooks.c (default_register_usage_leveling_p): New.
13339 * targhooks.h (default_register_usage_leveling_p): New prototype.
13340 * lra-assigns.c (register_usage_leveling_p): Use the hook.
13341 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
13342 * doc/tm.texi: Update.
13343 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
13345 2013-05-30 Ian Bolton <ian.bolton@arm.com>
13347 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
13348 (*insv_reg<mode>): New define_insn.
13350 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
13352 PR rtl-optimization/57439
13353 * postreload.c (move2add_valid_value_p): Check that we have
13354 a zero subreg_regno_offset when accessing the register in
13355 the requested mode.
13357 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
13358 Igor Zamyatin <igor.zamyatin@intel.com>
13360 Silvermont (SLM) architecture pipeline model, tuning and
13362 * config.gcc: Add slm config options and target.
13364 * config/i386/slm.md: New.
13366 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
13368 * config/i386/i386-c.c (ix86_target_macros_internal): New case
13370 (ix86_target_macros_internal): Likewise.
13372 * config/i386/i386.c (slm_cost): New cost.
13373 (m_SLM): New macro flag.
13374 (initial_ix86_tune_features): Set m_SLM.
13375 (x86_accumulate_outgoing_args): Likewise.
13376 (x86_arch_always_fancy_math_387): Likewise.
13377 (processor_target_table): Add slm cost.
13378 (cpu_names): Add slm cpu name.
13379 (x86_option_override_internal): Set SLM ISA.
13380 (ix86_issue_rate): New case PROCESSOR_SLM.
13381 (ia32_multipass_dfa_lookahead): Likewise.
13382 (fold_builtin_cpu): Add slm.
13384 * config/i386/i386.h (TARGET_SLM): New target macro.
13385 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
13386 (processor_type): Add PROCESSOR_SLM.
13388 * config/i386/i386.md (cpu): Add new value "slm".
13389 (slm.md): Include slm.md.
13391 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
13392 Zhenqiang Chen <zhenqiang.chen@linaro.org>
13394 * config/arm/arm-protos.h: Add and update function protos.
13395 * config/arm/arm.c (use_simple_return_p): New added.
13396 (thumb2_expand_return): Check simple_return flag.
13397 * config/arm/arm.md: Add simple_return and conditional simple_return.
13398 * config/arm/iterators.md: Add iterator for return and simple_return.
13400 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
13402 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
13403 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
13404 (arm_emit_vfp_multi_reg_pop): Likewise.
13405 (thumb2_emit_ldrd_pop): Likewise.
13406 (arm_expand_epilogue): Add misc REG_CFA notes.
13407 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
13409 2013-05-29 Lawrence Crowl <crowl@google.com>
13411 * config/arm/t-arm: Update for below.
13413 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
13414 Change type to hash_table. Update dependent calls and types.
13416 * config/i386/t-cygming: Update for below.
13418 * config/i386/t-interix: Update for below.
13420 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
13421 Change type to hash_table. Update dependent calls and types.
13422 (i386_find_on_wrapper_list::wrappers): Likewise.
13424 * config/ia64/t-ia64: Update for below.
13426 * config/ia64/ia64.c (bundle_state_table):
13427 Change type to hash_table. Update dependent calls and types.
13429 * config/mips/mips.c (mips_reorg_process_insns::htab):
13430 Change type to hash_table. Update dependent calls and types.
13432 * config/sol2.c (solaris_comdat_htab):
13433 Change type to hash_table. Update dependent calls and types.
13435 * config/t-sol2: Update for above.
13437 2013-05-29 Teresa Johnson <tejohnson@google.com>
13439 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
13440 functions are not yet marked as defined.
13442 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
13443 Pat Haugen <pthaugen@us.ibm.com>
13444 Peter Bergner <bergner@vnet.ibm.com>
13446 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
13450 (vrotl<mode>3): Likewise.
13451 (vashl<mode>3): Likewise.
13452 (vlshr<mode>3): Likewise.
13453 (vashr<mode>3): Likewise.
13455 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
13456 support for power8 V2DI builtins.
13458 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
13459 power8 V2DI builtins.
13460 (vupkhsw): Likewise.
13461 (vupklsw): Likewise.
13462 (vaddudm): Likewise.
13463 (vminsd): Likewise.
13464 (vmaxsd): Likewise.
13465 (vminud): Likewise.
13466 (vmaxud): Likewise.
13467 (vpkudum): Likewise.
13468 (vpksdss): Likewise.
13469 (vpkudus): Likewise.
13470 (vpksdus): Likewise.
13475 (vsubudm): Likewise.
13476 (vcmpequd): Likewise.
13477 (vcmpgtsd): Likewise.
13478 (vcmpgtud): Likewise.
13479 (vcmpequd_p): Likewise.
13480 (vcmpgtsd_p): Likewise.
13481 (vcmpgtud_p): Likewise.
13482 (vupkhsw): Likewise.
13483 (vupklsw): Likewise.
13484 (vaddudm): Likewise.
13485 (vmaxsd): Likewise.
13486 (vmaxud): Likewise.
13487 (vminsd): Likewise.
13488 (vminud): Likewise.
13489 (vpksdss): Likewise.
13490 (vpksdus): Likewise.
13491 (vpkudum): Likewise.
13492 (vpkudus): Likewise.
13497 (vsubudm): Likewise.
13499 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
13500 support for power8 V2DI instructions.
13502 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
13503 power8 V2DI instructions. Combine pack and unpack insns to use an
13504 iterator for each mode. Check whether a particular mode supports
13505 Altivec instructions instead of just checking TARGET_ALTIVEC.
13506 (UNSPEC_VPKUWUM): Likewise.
13507 (UNSPEC_VPKSHSS): Likewise.
13508 (UNSPEC_VPKSWSS): Likewise.
13509 (UNSPEC_VPKUHUS): Likewise.
13510 (UNSPEC_VPKSHUS): Likewise.
13511 (UNSPEC_VPKUWUS): Likewise.
13512 (UNSPEC_VPKSWUS): Likewise.
13513 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
13514 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
13515 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
13516 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
13517 (UNSPEC_VUPKHSB): Likewise.
13518 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
13519 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
13520 (UNSPEC_VUPKHSH): Likewise.
13521 (UNSPEC_VUPKLSB): Likewise.
13522 (UNSPEC_VUPKLSH): Likewise.
13524 (VI_char): Likewise.
13525 (VI_scalar): Likewise.
13526 (VI_unit): Likewise.
13528 (VP_small): Likewise.
13529 (VP_small_lc): Likewise.
13530 (VU_char): Likewise.
13531 (add<mode>3): Likewise.
13532 (altivec_vaddcuw): Likewise.
13533 (altivec_vaddu<VI_char>s): Likewise.
13534 (altivec_vadds<VI_char>s): Likewise.
13535 (sub<mode>3): Likewise.
13536 (altivec_vsubcuw): Likewise.
13537 (altivec_vsubu<VI_char>s): Likewise.
13538 (altivec_vsubs<VI_char>s): Likewise.
13539 (altivec_vavgs<VI_char>): Likewise.
13540 (altivec_vcmpbfp): Likewise.
13541 (altivec_eq<mode>): Likewise.
13542 (altivec_gt<mode>): Likewise.
13543 (altivec_gtu<mode>): Likewise.
13544 (umax<mode>3): Likewise.
13545 (smax<mode>3): Likewise.
13546 (umin<mode>3): Likewise.
13547 (smin<mode>3): Likewise.
13548 (altivec_vpkuhum): Likewise.
13549 (altivec_vpkuwum): Likewise.
13550 (altivec_vpkshss): Likewise.
13551 (altivec_vpkswss): Likewise.
13552 (altivec_vpkuhus): Likewise.
13553 (altivec_vpkshus): Likewise.
13554 (altivec_vpkuwus): Likewise.
13555 (altivec_vpkswus): Likewise.
13556 (altivec_vpks<VI_char>ss): Likewise.
13557 (altivec_vpks<VI_char>us): Likewise.
13558 (altivec_vpku<VI_char>us): Likewise.
13559 (altivec_vpku<VI_char>um): Likewise.
13560 (altivec_vrl<VI_char>): Likewise.
13561 (altivec_vsl<VI_char>): Likewise.
13562 (altivec_vsr<VI_char>): Likewise.
13563 (altivec_vsra<VI_char>): Likewise.
13564 (altivec_vsldoi_<mode>): Likewise.
13565 (altivec_vupkhsb): Likewise.
13566 (altivec_vupkhs<VU_char>): Likewise.
13567 (altivec_vupkls<VU_char>): Likewise.
13568 (altivec_vupkhsh): Likewise.
13569 (altivec_vupklsb): Likewise.
13570 (altivec_vupklsh): Likewise.
13571 (altivec_vcmpequ<VI_char>_p): Likewise.
13572 (altivec_vcmpgts<VI_char>_p): Likewise.
13573 (altivec_vcmpgtu<VI_char>_p): Likewise.
13574 (abs<mode>2): Likewise.
13575 (vec_unpacks_hi_v16qi): Likewise.
13576 (vec_unpacks_hi_v8hi): Likewise.
13577 (vec_unpacks_lo_v16qi): Likewise.
13578 (vec_unpacks_hi_<VP_small_lc>): Likewise.
13579 (vec_unpacks_lo_v8hi): Likewise.
13580 (vec_unpacks_lo_<VP_small_lc>): Likewise.
13581 (vec_pack_trunc_v8h): Likewise.
13582 (vec_pack_trunc_v4si): Likewise.
13583 (vec_pack_trunc_<mode>): Likewise.
13585 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
13587 (vec_vmaxsd): Likewise.
13588 (vec_vmaxud): Likewise.
13589 (vec_vminsd): Likewise.
13590 (vec_vminud): Likewise.
13591 (vec_vpksdss): Likewise.
13592 (vec_vpksdus): Likewise.
13593 (vec_vpkudum): Likewise.
13594 (vec_vpkudus): Likewise.
13595 (vec_vrld): Likewise.
13596 (vec_vsld): Likewise.
13597 (vec_vsrad): Likewise.
13598 (vec_vsrd): Likewise.
13599 (vec_vsubudm): Likewise.
13600 (vec_vupkhsw): Likewise.
13601 (vec_vupklsw): Likewise.
13603 2013-05-29 Jan Hubicka <jh@suse.cz>
13605 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
13606 flags; reorder rest of fields in more consistent way.
13607 (varpool_node): Remove analyzed, finalized and alias.
13608 (cgraph_ndoe): Likewise.
13609 (symtab_alias_ultimate_target): New function.
13610 (cgraph_function_node): Move offline.
13611 (cgraph_reset_node): Declare.
13612 (cgraph_comdat_can_be_unshared_p): Remove.
13613 (varpool_remove_initializer): Declare.
13614 (varpool_first_defined_variable, varpool_next_defined_variable
13615 cgraph_first_defined_function, cgraph_next_defined_function): Update.
13616 (cgraph_function_with_gimple_body_p): Update.
13617 (varpool_all_refs_explicit_p): Update.
13618 (symtab_alias_target): New function.
13619 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
13620 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
13621 (cgraph_function_or_thunk_node): Simplify using
13622 symtab_alias_ultimate_target.
13623 (varpool_variable_node): Likewise.
13624 * cgraph.c (cgraph_create_function_alias): Update.
13625 (cgraph_add_thunk): Update.
13626 (cgraph_remove_node): Update.
13627 (dump_cgraph_node): Do not dump removed flags.
13628 (cgraph_function_body_availability): Update.
13629 (cgraph_propagate_frequency): Update.
13630 (verify_cgraph_node): Check sanity of local flag.
13631 (cgraph_function_node): Move here from cgraph.h; revamp for
13632 cgraph_function_or_thunk_node.
13633 * lto-symtab.c (lto_varpool_replace_node): Update.
13634 (lto_symtab_resolve_can_prevail_p): Update.
13635 (lto_symtab_merge_cgraph_nodes): Update.
13636 * ipa-cp.c (determine_versionability, initialize_node_lattices,
13637 propagate_constants_accross_call, devirtualization_time_bonus,
13638 ipcp_propagate_stage): Update.
13639 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
13640 * ipa-inline-transform.c (clone_inlined_nodes,
13641 preserve_function_body_p): Update.
13642 * ipa-reference.c (propagate): Update.
13643 (write_node_summary_p): Update.
13644 * toplev.c (wrapup_global_declaration_2): Update.
13645 * cgraphunit.c (cgraph_analyze_function): Rename to ...
13646 (analyze_function) ... this one.
13647 (cgraph_process_new_functions): Update.
13648 (cgraph_reset_node): Export.
13649 (cgraph_finalize_function): Update.
13650 (cgraph_add_new_function): Update.
13651 (process_function_and_variable_attributes): Update.
13652 (varpool_finalize_decl): Update.
13653 (symbol_finalized): Remove.
13654 (symbol_finalized_and_needed): Rename to ...
13655 (symbol_defined_and_needed): ... update.
13656 (cgraph_analyze_functions): Update.
13657 (handle_alias_pairs): Update.
13658 (mark_functions_to_output): Update.
13659 (assemble_thunk): Update.
13660 (output_in_order): Update.
13661 (output_weakrefs): Update.
13662 (finalize_compilation_unit): Update.
13663 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
13664 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
13665 input_node, input_varpool_node): Update.
13666 * dbxout.c (dbxout_expand_expr): Update.
13667 * cgraphclones.c (cgraph_clone_node): Update.
13668 (cgraph_copy_node_for_versioning): Update.
13669 (cgraph_materialize_clone): Update.
13670 (cgraph_materialize_all_clones): Update.
13671 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
13672 propagate_pure_const, propagate_nothrow): Update.
13673 * lto-streamer-out.c (lto_output, write_symbol): Update.
13674 * ipa-utils.c (ipa_reverse_postorder): Update.
13675 * ipa-inline.c (can_inline_edge_p): Update.
13676 (update_caller_keys, ipa_inline): Update.
13677 * dwarf2out.c (reference_to_unused,
13678 premark_types_used_by_global_vars_helper): Update.
13679 * tree-eh.c (tree_could_trap_p): Update.
13680 * ipa-split.c (consider_split, execute_split_functions): Update.
13681 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
13682 has_addr_references_p): Update; move ahead in file for better
13684 (process_references): Simplify.
13685 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
13686 bodies are removed.
13687 (cgraph_comdat_can_be_unshared_p): Make static.
13688 (cgraph_externally_visible_p): Update.
13689 (varpool_externally_visible_p): Update.
13690 (function_and_variable_visibility): Update.
13691 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
13692 ipa_tm_mark_force_output_node): Update.
13693 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
13694 estimate_edge_devirt_benefit, inline_generate_summary,
13695 inline_write_summary): Update.
13696 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
13697 * ipa-prop.c (ipa_compute_jump_functions): Update.
13698 (ipa_print_node_params, ipa_prop_read_section,
13699 ipa_update_after_lto_read, read_replacements_section): Update.
13700 * varasm.c (mark_decl_referenced): Update.
13701 (assemble_alias, dump_tm_clone_pairs): Update.
13702 * tree-inline.c (copy_bb): Update.
13703 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
13705 * symtab.c (dump_symtab_base): Print new flags.
13706 (verify_symtab_base): Verify new flags.
13707 (symtab_alias_ultimate_target): New function.
13708 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
13709 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
13711 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
13713 * i386.c (ix86_get_function_versions_dispatcher,
13714 ix86_generate_version_dispatcher_body): Update.
13715 (fold_builtin_cpu): Use varpool_add_new_variable.
13716 * varpool.c (varpool_remove_initializer): Break out from ...
13717 (varpool_remove_node): ... this one.
13718 (dump_varpool_node, varpool_node_for_asm,
13719 cgraph_variable_initializer_availability, varpool_analyze_node,
13720 varpool_assemble_decl, varpool_remove_unreferenced_decls,
13721 varpool_finalize_named_section_flags, varpool_create_variable_alias):
13724 2013-05-29 Jan Hubicka <jh@suse.cz>
13726 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
13728 2013-05-29 Easwaran Raman <eraman@google.com>
13730 PR tree-optimization/57442
13731 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
13732 when control exits the main loop.
13734 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
13736 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
13738 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
13739 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
13740 * rx/t-rx: Add rx100 under multi library matches option for nofpu
13743 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13745 PR tree-optimization/57441
13746 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
13747 Don't limit size of incr_vec to number of candidates.
13749 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
13751 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
13752 and mips16 directories.
13753 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
13754 (MULTILIB_DIRNAMES): Ditto.
13755 (MULTILIB_EXCEPTIONS): Add new exceptions.
13756 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
13757 (MULTILIB_DIRNAMES): Ditto.
13758 (MULTILIB_EXCEPTIONS): Add new exceptions.
13760 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
13761 Marcus Shawcroft <marcus.shawcroft@arm.com>
13763 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
13764 SYMBOL_TINY_ABSOLUTE.
13765 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
13766 SYMBOL_TINY_ABSOLUTE.
13767 (aarch64_expand_mov_immediate): Likewise.
13768 (aarch64_classify_symbol): Likewise.
13769 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
13770 Permit SYMBOL_TINY_ABSOLUTE.
13771 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
13773 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
13774 Marcus Shawcroft <marcus.shawcroft@arm.com>
13776 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
13777 Refactor if/switch. Replace gcc_assert with if.
13779 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
13781 * config/i386/i386.c (initial_ix86_tune_features): Enable
13782 FP Reassociation for AMD bdver1 and bdver2.
13784 2013-05-29 Martin Jambor <mjambor@suse.cz>
13786 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
13787 and IMAGPART_EXPR do not occur within other handled_components.
13789 2013-05-29 Richard Biener <rguenther@suse.de>
13791 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
13792 access on whether the use is in the BB we currently try to
13794 (vect_bb_vectorization_profitable_p): Pass the BB we currently
13795 vectorize to vect_bb_slp_scalar_cost.
13797 2013-05-29 Richard Biener <rguenther@suse.de>
13799 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
13800 computing scalar cost offsetted by stmts that are kept live
13802 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
13803 for computation of scalar cost.
13805 2013-05-28 Steve Ellcey <sellcey@mips.com>
13807 * config/mips/mips-cpus.def (mips32r2): Change processor type.
13809 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
13811 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
13812 array notation built-in reduction functions.
13813 * doc/passes.texi (Passes): Added documentation about changes done
13815 * doc/invoke.texi (C Dialect Options): Added documentation about
13816 the -fcilkplus flag.
13817 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
13818 (BUILTINS_DEF): Depend on cilkplus.def.
13819 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
13820 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
13821 * cilkplus.def: New file.
13823 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
13825 PR rtl-optimization/57439
13826 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
13828 2013-05-28 Easwaran Raman <eraman@google.com>
13830 PR tree-optimization/57337
13831 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
13832 (find_insert_point): Correctly identify the insertion point
13833 when two statements with the same UID is compared.
13835 2013-05-28 Richard Biener <rguenther@suse.de>
13837 PR tree-optimization/56787
13838 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
13839 from the list of data references.
13840 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
13842 (vect_analyze_loop_operations): Likewise.
13843 (vect_transform_loop): Remove clobbers.
13845 2013-05-28 Martin Jambor <mjambor@suse.cz>
13847 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
13848 and REALPART_EXPRs have scalar type.
13850 2013-05-28 Richard Biener <rguenther@suse.de>
13852 PR tree-optimization/57411
13853 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
13855 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
13856 virtual operand propagation.
13858 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
13860 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
13861 destination register for bmasksi_vis.
13862 (vector_init_bshuffle): Likewise.
13863 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
13865 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
13867 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
13868 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
13869 mode if the instruction isn't available in the original mode.
13870 * config/sparc/sparc.opt (mfix-ut699): New option.
13871 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
13872 (divdf3): Turn into expander.
13873 (divdf3_nofix): New insn.
13874 (divdf3_fix): Likewise.
13875 (divsf3): Disable if -mfix-ut699.
13876 (sqrtdf2): Turn into expander.
13877 (sqrtdf2_nofix): New insn.
13878 (sqrtdf2_fix): Likewise.
13879 (sqrtsf2): Disable if -mfix-ut699.
13881 2013-05-27 Richard Biener <rguenther@suse.de>
13883 PR middle-end/57412
13884 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
13885 block for the new loop.
13887 2013-05-27 Richard Biener <rguenther@suse.de>
13889 PR tree-optimization/57343
13890 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
13891 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
13892 (number_of_iterations_cond): Do not build the folded tree.
13894 2013-05-27 Richard Biener <rguenther@suse.de>
13897 PR middle-end/57381
13898 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
13899 OEP_CONSTANT_ADDRESS_OF retained.
13901 PR tree-optimization/57417
13902 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
13903 for unchanged base.
13904 (set_ssa_val_to): Compare addresses using
13905 get_addr_base_and_unit_offset.
13907 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
13909 PR rtl-optimization/56833
13910 * postreload.c (move2add_record_mode): New function.
13911 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
13912 (move2add_use_add2_insn): Use move2add_record_sym_value.
13913 (move2add_use_add3_insn): Likewise.
13914 (reload_cse_move2add): Use move2add_valid_value_p and
13915 move2add_record_mode. Invalidate call-clobbered and REG_INC
13916 affected regs by setting reg_mode to VOIDmode.
13917 (move2add_note_store): Don't pretend the inside of a SUBREG is
13918 the actual destination. Invalidate single/leading registers by
13919 setting reg_mode to VOIDmode.
13920 Use move2add_record_sym_value, move2add_valid_value_p and
13921 move2add_record_mode.
13923 2013-05-27 Richard Biener <rguenther@suse.de>
13925 PR tree-optimization/57396
13926 * tree-affine.c (double_int_constant_multiple_p): Properly
13927 return false for val == 0 and div != 0.
13929 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
13931 * config/mips/mips.h: Use #elif in preprocessor conditions.
13933 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
13936 * config/mips/constraints.md (kl): New constraint.
13937 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
13938 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
13939 constraint for operand 0. Split after CSE for MIPS16. Emit a move
13940 from LO for MIPS16.
13941 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
13943 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
13946 * config/mips/mips.c (mips_can_inline_p): New function.
13947 (TARGET_CAN_INLINE_P): Define.
13949 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
13951 * sched-int.h (ds_t, dw_t): Make unsigned int.
13952 Fix documentation that describes how all the ds_t bits are used.
13953 Reserve the last bit for delayed-branch scheduling.
13954 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
13955 (BITS_PER_DEP_WEAK): Fix definition and documentation.
13956 (gen_dep_weak_1): Remove prototype.
13957 * sched-deps.c (get_dep_weak_1): Make static.
13958 * target.def (speculate_insn, needs_block_p, gen_spec_check,
13959 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
13960 * doc/tm.texi: Regenerate.
13961 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
13963 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
13966 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
13968 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
13969 Sandra Loosemore <sandra@codesourcery.com>
13971 * config.gcc (powerpc-*): Allow native for with-cpu.
13973 2013-05-24 Jeff Law <law@redhat.com>
13975 PR tree-optimization/57124
13976 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
13977 conversion feeding a condition if the range has an overflow
13978 if -fstrict-overflow. Add warnings for when we do make the
13981 2013-05-24 Dehao Chen <dehao@google.com>
13983 * tree-cfg.c (locus_discrim_map): Fix the typo.
13984 (locus_discrim_hasher): Likewise.
13985 (locus_discrim_hasher::hash): Likewise.
13986 (locus_discrim_hasher::equal): Likewise.
13988 2013-05-24 Martin Jambor <mjambor@suse.cz>
13990 PR tree-optimization/57294
13991 * cgraph.h (ipa_record_stmt_references): Declare.
13992 * cgraphbuild.c (ipa_record_stmt_references): New function.
13993 (build_cgraph_edges): Use ipa_record_stmt_references.
13994 (rebuild_cgraph_edges): Likewise.
13995 (cgraph_rebuild_references): Likewise.
13996 * ipa-prop.c (ipa_modify_call_arguments): Discard references
13997 associated with the old statement and build references from the
13998 newly built statements.
13999 * ipa-ref.c (ipa_remove_stmt_references): New function.
14000 * ipa-ref.h (ipa_remove_stmt_references): Declare.
14002 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
14004 * lra-constraints.c (emit_spill_move): Use smaller mode for
14006 (check_and_process_move): Consider mem-reg moves for secondary
14008 (curr_insn_transform): Don't lose insns emitted before for
14009 secondary memory moves.
14010 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
14011 reg set up in the current insn.
14013 2013-05-24 Dehao Chen <dehao@google.com>
14015 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
14017 (locus_descrim_hasher::equal): Likewise.
14018 (build_gimple_cfg): New discriminator assignment algorithm.
14019 (make_edges): Likewise.
14020 (next_discriminator_for_locus): Likewise.
14021 (same_line_p): Likewise.
14022 (assign_discriminators): Likewise.
14023 (make_cond_expr_edges): Likewise.
14024 (make_gimple_switch_edges): Likewise.
14025 (make_goto_expr_edges): Likewise.
14026 (make_gimple_asm_edges): Likewise.
14028 2013-05-24 Ian Bolton <ian.bolton@arm.com>
14030 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
14031 X format specifier to only display bottom 16 bits.
14032 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
14033 immediate to match for operand 2, since it will be masked.
14035 2013-05-24 Richard Biener <rguenther@suse.de>
14037 PR tree-optimization/57287
14038 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
14039 all SSA names that occur in abnormal PHIs.
14041 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
14044 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
14045 that index is not negative.
14047 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
14049 PR rtl-optimization/55177
14050 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
14051 (simplify_byte_swapping_operation): New.
14052 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
14053 (simplify_relational_operation_1): Deal with BSWAP.
14055 2013-05-23 Richard Henderson <rth@redhat.com>
14058 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
14059 (ix86_reorg): Call it.
14061 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
14064 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
14065 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
14066 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
14068 2013-05-23 Christian Bruel <christian.bruel@st.com>
14071 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
14073 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
14074 Marcus Shawcroft <marcus.shawcroft@arm.com>
14076 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
14077 * config/aarch64/constraints.md (Usa): Remove.
14078 * doc/md.texi (AArch64 Usa): Remove.
14080 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
14081 Marcus Shawcroft <marcus.shawcroft@arm.com>
14083 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
14084 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
14085 * config/aarch64/predicates.md (aarch64_const_address): Remove.
14086 (aarch64_mov_operand): Use aarch64_mov_operand_p.
14088 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
14090 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
14091 instruction (AdvSIMD).
14092 * config/aarch64/aarch64-builtins.c
14093 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
14094 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
14096 2013-05-23 Martin Jambor <mjambor@suse.cz>
14098 PR middle-end/57347
14099 * tree.h (contains_bitfld_component_ref_p): Declare.
14100 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
14101 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
14103 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
14104 not access a bit-field. Assert all final offsets are byte-aligned.
14106 2013-05-23 Richard Biener <rguenther@suse.de>
14108 PR tree-optimization/57380
14109 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
14110 least one invariant or re-used load.
14111 * passes.c (init_optimization_passes): Move pass_phiprop before
14114 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
14116 * config/aarch64/aarch64-simd.md
14117 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
14119 2013-05-23 Richard Biener <rguenther@suse.de>
14121 PR middle-end/57381
14122 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
14123 OEP_CONSTANT_ADDRESS_OF retained.
14125 2013-05-23 Jakub Jelinek <jakub@redhat.com>
14127 PR middle-end/57344
14128 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
14129 don't lower unit. Handle unit not being always BITS_PER_WORD.
14131 2013-05-23 Richard Biener <rguenther@suse.de>
14133 PR rtl-optimization/57341
14134 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
14135 instead of true_dependence.
14137 2013-05-22 David Malcolm <dmalcolm@redhat.com>
14139 * bb-reorder.c (branch_threshold): Make const.
14140 (exec_threshold): Ditto.
14142 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
14143 Pat Haugen <pthaugen@us.ibm.com>
14144 Peter Bergner <bergner@vnet.ibm.com>
14146 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
14147 documentation for the power8 crypto builtins.
14149 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
14151 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
14152 macros for defining power8 builtin functions.
14153 (BU_P8V_AV_2): Likewise.
14154 (BU_P8V_AV_P): Likewise.
14155 (BU_P8V_VSX_1): Likewise.
14156 (BU_P8V_OVERLOAD_1): Likewise.
14157 (BU_P8V_OVERLOAD_2): Likewise.
14158 (BU_CRYPTO_1): Likewise.
14159 (BU_CRYPTO_2): Likewise.
14160 (BU_CRYPTO_3): Likewise.
14161 (BU_CRYPTO_OVERLOAD_1): Likewise.
14162 (BU_CRYPTO_OVERLOAD_2): Likewise.
14163 (XSCVSPDP): Fix typo, point to the correct instruction.
14164 (VCIPHER): Add power8 crypto builtins.
14165 (VCIPHERLAST): Likewise.
14166 (VNCIPHER): Likewise.
14167 (VNCIPHERLAST): Likewise.
14168 (VPMSUMB): Likewise.
14169 (VPMSUMH): Likewise.
14170 (VPMSUMW): Likewise.
14171 (VPERMXOR_V2DI): Likewise.
14172 (VPERMXOR_V4SI: Likewise.
14173 (VPERMXOR_V8HI: Likewise.
14174 (VPERMXOR_V16QI: Likewise.
14175 (VSHASIGMAW): Likewise.
14176 (VSHASIGMAD): Likewise.
14177 (VPMSUM): Likewise.
14178 (VPERMXOR): Likewise.
14179 (VSHASIGMA): Likewise.
14181 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
14182 __CRYPTO__ if the crypto instructions are available.
14183 (altivec_overloaded_builtins): Add support for overloaded power8
14186 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
14187 support for power8 crypto builtins.
14188 (builtin_function_type): Likewise.
14189 (altivec_init_builtins): Add support for builtins that take vector
14190 long long (V2DI) arguments.
14192 * config/rs6000/crypto.md: New file, define power8 crypto
14195 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
14196 Pat Haugen <pthaugen@us.ibm.com>
14197 Peter Bergner <bergner@vnet.ibm.com>
14199 * doc/invoke.texi (Option Summary): Add power8 options.
14200 (RS/6000 and PowerPC Options): Likewise.
14202 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
14203 constraints.md instead of rs6000.h. Reorder w* constraints. Add
14204 wm, wn, wr documentation.
14206 * config/rs6000/constraints.md (wm): New constraint for VSX
14207 registers if direct move instructions are enabled.
14208 (wn): New constraint for no registers.
14209 (wq): New constraint for quad word even GPR registers.
14210 (wr): New constraint if 64-bit instructions are enabled.
14211 (wv): New constraint if power8 vector instructions are enabled.
14212 (wQ): New constraint for quad word memory locations.
14214 * config/rs6000/predicates.md (const_0_to_15_operand): New
14215 constraint for 0..15 for crypto instructions.
14216 (gpc_reg_operand): If VSX allow registers in VSX registers as well
14217 as GPR and floating point registers.
14218 (int_reg_operand): New predicate to match only GPR registers.
14219 (base_reg_operand): New predicate to match base registers.
14220 (quad_int_reg_operand): New predicate to match even GPR registers
14221 for quad memory operations.
14222 (vsx_reg_or_cint_operand): New predicate to allow vector logical
14223 operations in both GPR and VSX registers.
14224 (quad_memory_operand): New predicate for quad memory operations.
14225 (reg_or_indexed_operand): New predicate for direct move support.
14227 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
14228 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
14229 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
14230 (POWERPC_MASKS): Add power8 options.
14231 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
14234 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
14235 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
14237 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
14238 (-mpower8-fusion): New power8 options.
14239 (-mpower8-fusion-sign): Likewise.
14240 (-mpower8-vector): Likewise.
14241 (-mcrypto): Likewise.
14242 (-mdirect-move): Likewise.
14243 (-mquad-memory): Likewise.
14245 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
14246 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
14248 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
14249 (rs6000_debug_vector_unit): Add p8_vector.
14250 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
14251 definitions. Also print fusion state.
14252 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
14253 (rs6000_builtin_mask_calculate): Add power8 builtin support.
14254 (rs6000_option_override_internal): Add support for power8.
14255 (rs6000_common_init_builtins): Add debugging for skipped builtins
14256 if -mdebug=builtin.
14257 (rs6000_adjust_cost): Add power8 support.
14258 (rs6000_issue_rate): Likewise.
14259 (insn_must_be_first_in_group): Likewise.
14260 (insn_must_be_last_in_group): Likewise.
14261 (force_new_group): Likewise.
14262 (rs6000_register_move_cost): Likewise.
14263 (rs6000_opt_masks): Likewise.
14265 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
14266 power8 capable assembler, default to power7 options.
14267 (TARGET_DIRECT_MOVE): Likewise.
14268 (TARGET_CRYPTO): Likewise.
14269 (TARGET_P8_VECTOR): Likewise.
14270 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
14271 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
14272 (VECTOR_MEM_P8_VECTOR_P): Likewise.
14273 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
14274 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
14275 (TARGET_XSCVDPSPN): Likewise.
14276 (TARGET_XSCVSPDPN): Likewsie.
14277 (TARGET_SYNC_HI_QI): Likewise.
14278 (TARGET_SYNC_TI): Likewise.
14279 (MASK_CRYPTO): Likewise.
14280 (MASK_DIRECT_MOVE): Likewise.
14281 (MASK_P8_FUSION): Likewise.
14282 (MASK_P8_VECTOR): Likewise.
14283 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
14284 temporary used by some of the direct move instructions to get two FP
14285 temporary registers does not force creation of a stack frame.
14286 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
14287 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
14288 that any VSX registers are tieable, even if they are also an
14289 Altivec vector mode.
14290 (r6000_reg_class_enum): Add wm, wr, wv constraints.
14291 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
14292 (RS6000_BTM_CRYPTO): Likewise.
14293 (RS6000_BTM_COMMON): Likewise.
14295 * config/rs6000/rs6000.md (cpu attribute): Add power8.
14296 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
14297 (enum rs6000_vector): Add power8 vector support.
14299 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14303 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
14304 (any_sibcall_could_use_r3): this and handle indirect calls.
14305 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
14307 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14309 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
14311 2013-05-22 Richard Biener <rguenther@suse.de>
14313 PR middle-end/57349
14314 * profile.c (branch_prob): Do not split blocks that are
14315 abnormally receiving from ECF_RETURNS_TWICE functions.
14317 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
14319 * recog.c (offsettable_address_addr_space_p): Fix calculation of
14320 address mode. Move pointer mode initialization to the same place.
14322 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
14324 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
14325 while it has any effect.
14327 2013-05-21 Easwaran Raman <eraman@google.com>
14329 PR tree-optimization/57322
14330 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
14331 UID of the statement added to the BB to be 1.
14333 2013-05-21 Jakub Jelinek <jakub@redhat.com>
14335 PR tree-optimization/57331
14336 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
14337 of conversion from pointer type to integral type with integer.
14339 2013-05-21 Martin Jambor <mjambor@suse.cz>
14342 * ipa-prop.c (ipa_read_node_info): Process param_used and
14343 controlled_uses in the same order as when writing.
14345 2013-05-21 Magnus Granberg <baldrick@free.fr>
14348 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
14350 2013-05-21 Richard Biener <rguenther@suse.de>
14352 PR tree-optimization/57318
14353 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
14354 estimate stmts with side-effects as likely eliminated.
14356 2013-05-21 Richard Biener <rguenther@suse.de>
14358 PR tree-optimization/57330
14359 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
14360 preserve the call stmts fntype.
14362 2013-05-21 Richard Biener <rguenther@suse.de>
14364 PR tree-optimization/57303
14365 * tree-ssa-sink.c (statement_sink_location): Improve killing
14366 stmt detection and properly handle self-assignments.
14368 2013-05-21 Christian Bruel <christian.bruel@st.com>
14370 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
14371 spanning registers. LEAF_REG_REMAP is supported only for contiguous
14372 registers. Set register size out of the PARALLEL loop.
14374 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
14377 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
14378 (*fmasf4, *fmasf4_media): New insns.
14380 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
14382 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
14383 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
14384 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
14385 (mips_idiv_insns): Update the comments to say that the returned
14386 instruction counts are in units of BASE_INSN_LENGTH.
14387 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
14388 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
14389 using 2 rather than 4 as the length of indirect MIPS16 and
14390 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
14391 length of a NOP. Don't divide MIPS16 lengths by 2.
14392 (mips16_split_long_branches): Assume a branch is long if the
14393 length is greater than 4 rather than 8.
14394 * config/mips/mips.md (length): Give MIPS16 lengths directly,
14395 rather than multiplying them by 2. Multiply instruction counts
14396 by BASE_INSN_LENGTH rather than 4.
14397 (*jump_mips16, tls_get_tp_mips16_<mode>)
14398 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
14400 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
14402 * config/mips/mips.md (extended_mips16): Remove branch case.
14403 (length): Remove duplicated extended_mips16 test.
14405 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
14407 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
14409 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
14411 * recog.h (Recog_data): Rename to...
14412 (recog_data_d): ...this.
14413 (recog_data): Update accordingly.
14414 * recog.c (recog_data): Likewise.
14415 * reload.c (save_recog_data): Likewise.
14416 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
14417 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
14419 2013-05-17 Julian Brown <julian@codesourcery.com>
14421 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
14422 found in a REG_EQUAL note, invalidate it.
14424 2013-05-17 Easwaran Raman <eraman@google.com>
14426 * tree-ssa-reassoc.c (find_insert_point): New function.
14427 (insert_stmt_after): Likewise.
14428 (get_def_stmt): Likewise.
14429 (ensure_ops_are_available): Likewise.
14430 (not_dominated_by): Likewise.
14431 (rewrite_expr_tree): Do not move statements beyond what is
14432 necessary. Remove call to swap_ops_for_binary_stmt...
14433 (reassociate_bb): ... and move it here.
14434 (build_and_add_sum): Assign UIDs for new statements.
14435 (linearize_expr): Likewise.
14436 (do_reassoc): Renumber gimple statement UIDs.
14438 2013-05-17 Jan Hubicka <jh@suse.cz>
14440 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
14442 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
14443 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
14445 (output_weakrefs): Update.
14447 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
14448 Martin Jambor <mjambor@suse.cz>
14450 PR middle-end/57276
14451 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
14452 value that corresponds to the given aggval is found in values vector.
14454 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
14456 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
14457 sse, sse2, sse3, ssse3 and sse4a flags to options.
14459 2013-05-17 David Malcolm <dmalcolm@redhat.com>
14461 * gengtype-state.c: (s_expr_writer): New class, to handle
14462 prettifying of output layout of s-expressions.
14463 (state_writer): New class, to write out gtype.state.
14464 (state_written_type_count): Move this variable into member data of
14466 (s_expr_writer::s_expr_writer): New code: constructor for new class
14467 (state_writer::state_writer(): ditto
14468 (s_expr_writer::write_new_line): New function
14469 (s_expr_writer::write_any_indent): ditto
14470 (s_expr_writer::begin_s_expr): ditto
14471 (s_expr_writer::end_s_expr): ditto
14472 (write_state_fileloc): convert to method of state_writer...
14473 (state_writer:: write_state_fileloc): ...and use methods of
14474 s_expr_writer to write indentation into the gtype.state output file
14475 to visually represent the hierarchical structure of the list
14477 (write_state_fields): ditto, renaming to...
14478 (state_writer::write_state_fields)
14479 (write_state_a_string): ditto, renaming to...
14480 (state_writer::write_state_a_string)
14481 (write_state_string_option): ditto, renaming to...
14482 (state_writer::write_state_string_option)
14483 (write_state_type_option): ditto, renaming to...
14484 (state_writer::write_state_type_option)
14485 (write_state_nested_option): ditto, renaming to...
14486 (state_writer::write_state_nested_option)
14487 (write_state_option): ditto, renaming to...
14488 (state_writer::write_state_option)
14489 (write_state_options): ditto, renaming to...
14490 (state_writer::write_state_options)
14491 (write_state_lang_bitmap): ditto, renaming to...
14492 (state_writer::write_state_lang_bitmap)
14493 (write_state_version): ditto, renaming to...
14494 (state_writer::write_state_version)
14495 (write_state_scalar_type): ditto, renaming to...
14496 (state_writer::write_state_scalar_type)
14497 (write_state_string_type): ditto, renaming to...
14498 (state_writer::write_state_string_type)
14499 (write_state_undefined_type): ditto, renaming to...
14500 (state_writer::write_state_undefined_type)
14501 (write_state_struct_union_type): ditto, renaming to...
14502 (state_writer::write_state_struct_union_type)
14503 (write_state_struct_type): ditto, renaming to...
14504 (state_writer::write_state_struct_type)
14505 (write_state_user_struct_type): ditto, renaming to...
14506 (state_writer::write_state_user_struct_type)
14507 (write_state_lang_struct_type): ditto, renaming to...
14508 (state_writer::write_state_lang_struct_type)
14509 (write_state_param_struct_type): ditto, renaming to...
14510 (state_writer::write_state_param_struct_type)
14511 (write_state_pointer_type): ditto, renaming to...
14512 (state_writer::write_state_pointer_type)
14513 (write_state_array_type): ditto, renaming to...
14514 (state_writer::write_state_array_type)
14515 (write_state_gc_used): ditto, renaming to...
14516 (state_writer::write_state_gc_used)
14517 (write_state_common_type_content): ditto, renaming to...
14518 (state_writer::write_state_common_type_content)
14519 (write_state_type): ditto, renaming to...
14520 (state_writer::write_state_type)
14521 (write_state_pair_list): ditto, renaming to...
14522 (state_writer::write_state_pair_list)
14523 (write_state_pair): ditto, renaming to...
14524 (state_writer::write_state_pair)
14525 (write_state_typedefs): ditto, renaming to...
14526 (state_writer::write_state_typedefs)
14527 (write_state_structures): ditto, renaming to...
14528 (state_writer::write_state_structures)
14529 (write_state_param_structs): ditto, renaming to...
14530 (state_writer::write_state_param_structs)
14531 (write_state_variables): ditto, renaming to...
14532 (state_writer::write_state_variables)
14533 (write_state_srcdir): ditto, renaming to...
14534 (state_writer::write_state_srcdir)
14535 (write_state_files_list): ditto, renaming to...
14536 (state_writer::write_state_files_list)
14537 (write_state_languages): ditto, renaming to...
14538 (state_writer::write_state_languages)
14539 (write_state): create a state_writer instance and use it when
14540 writing out the state file
14542 2013-05-17 Mike Stump <mikestump@comcast.net>
14544 PR rtl-optimization/57304
14545 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
14546 accessing DF_REF_REAL_LOC.
14548 2013-05-17 Jakub Jelinek <jakub@redhat.com>
14550 PR rtl-optimization/57281
14551 PR rtl-optimization/57300
14552 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
14553 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
14554 what the other splitter did if the registers are dead.
14556 2013-05-17 Richard Biener <rguenther@suse.de>
14558 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
14561 2013-05-17 Jakub Jelinek <jakub@redhat.com>
14563 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
14566 2013-05-17 Marek Polacek <polacek@redhat.com>
14568 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
14569 length when doing non-zero store of storing '\0' to '\0'.
14571 2013-05-17 Jakub Jelinek <jakub@redhat.com>
14573 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
14574 vect_external_def oprnd1 with loop_vinfo, try to emit
14575 optional cast, negation and and stmts on the loop preheader
14576 edge instead of into the pattern def seq.
14578 PR tree-optimization/57051
14579 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
14580 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
14582 2013-05-16 Nick Clifton <nickc@redhat.com>
14584 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
14585 (rl78_is_naked_func): New function.
14586 (rl78_expand_prologue): Skip prologue generation for naked functions.
14587 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
14588 * doc/extend.texi (naked): Add RL78 to the list of processors
14589 that supports this attribute.
14591 2013-05-16 Jeff Law <law@redhat.com>
14593 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
14595 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
14597 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
14598 cache parameters using detect_caches_amd also for CYRIX,
14599 NSC and TM2 signatures.
14601 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
14602 Dzianis Kahanovich <mahatma@eu.by>
14606 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
14607 VIA/Centaur processors and determine their cache parameters
14608 using detect_caches_amd.
14610 2013-05-16 Teresa Johnson <tejohnson@google.com>
14612 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
14613 (rtl_verify_edges): New function.
14614 (rtl_verify_bb_insns): Ditto.
14615 (rtl_verify_bb_pointers): Ditto.
14616 (rtl_verify_bb_insn_chain): Ditto.
14617 (rtl_verify_fallthru): Ditto.
14618 (rtl_verify_bb_layout): Ditto.
14619 (rtl_verify_flow_info_1): Outline checks into new functions.
14620 (rtl_verify_flow_info): Ditto.
14622 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
14624 * cfghooks.c (copy_bbs): Add update_dominance argument.
14625 * cfghooks.h (copy_bbs): Update prototype.
14626 * tree-cfg.c (gimple_duplicate_sese_region):
14627 Add update_dominance argument.
14628 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
14629 * tree-ssa-loop-ch.c (copy_loop_headers): Update
14630 gimple_duplicate_sese_region call.
14631 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
14632 Update copy_bbs call.
14633 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
14634 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
14636 2013-05-16 Jakub Jelinek <jakub@redhat.com>
14638 * tree-vectorizer.h (NUM_PATTERNS): Increment.
14639 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
14640 vect_recog_rotate_pattern.
14641 (vect_recog_rotate_pattern): New function.
14643 2013-05-16 Jason Merrill <jason@redhat.com>
14645 * Makefile.in (LLINKER): New variable.
14646 (mostlyclean): Remove link mutex.
14647 * configure.ac: Handle --enable-link-mutex.
14648 * lock-and-run.sh: New script.
14650 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14653 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
14656 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
14658 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
14660 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
14662 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
14663 * config/arm/arm.c (next_consecutive_mem): New function.
14664 (gen_movmem_ldrd_strd): Likewise.
14665 * config/arm/arm.md (movmemqi): Update condition and code.
14666 (unaligned_loaddi, unaligned_storedi): New patterns.
14668 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
14670 * config.gcc: Obsolete *-*-solaris2.9*.
14671 * doc/install.texi (Specific, *-*-solaris2*): Document it.
14673 2013-05-16 Richard Biener <rguenther@suse.de>
14675 * passes.c (init_optimization_passes): Move pass_parallelize_loops
14676 earlier, after GRAPHITE transforms and IV canonicalization.
14678 2013-05-16 Jakub Jelinek <jakub@redhat.com>
14680 * omp-low.c (extract_omp_for_data): For collapsed loops,
14681 if at least one of the loops is known at compile time to
14682 iterate zero times, set count to 0.
14683 (expand_omp_regimplify_p): New function.
14684 (expand_omp_for_generic): For collapsed loops, if at least
14685 one of the loops isn't known to iterate at least once,
14686 add runtime check with setting count to 0.
14687 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
14688 For unsigned types if it isn't known at compile time that
14689 the loop will iterate at least once, add runtime check to bypass
14690 the whole loop if initial condition isn't true.
14692 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
14694 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
14696 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
14698 PR middle-end/57286
14699 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
14700 transformations to avoid an infinite loop.
14702 2013-05-16 Marek Polacek <polacek@redhat.com>
14704 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
14706 2013-05-15 Leif Ekblad <leif@rdos.net>
14708 * config/i386/i386.c (ix86_decompose_address): Use
14709 DEFAULT_TLS_SEG_REG to access TLS segment register.
14710 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
14711 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
14712 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
14714 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
14717 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
14718 sibling calls to functions that would normally be lazily bound,
14719 unless $gp is call-clobbered.
14721 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
14723 * config/i386/i386.c (ix86_option_override_internal): Update
14724 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
14725 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
14726 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
14727 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
14729 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
14731 2013-05-15 Andreas Schwab <schwab@suse.de>
14733 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
14734 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
14737 2013-05-15 Teresa Johnson <tejohnson@google.com>
14739 * loop-unroll.c (report_unroll_peel): Check decision before
14740 emitting unroll/peel message.
14742 2013-05-15 Teresa Johnson <tejohnson@google.com>
14744 * function.h (has_bb_partition): New rtl_data flag.
14745 (bb_reorder_complete): Ditto.
14746 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
14747 instead of flag_reorder_blocks_and_partition.
14748 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
14749 with some enhancements.
14750 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
14751 * bb-reorder.c (connect_traces): Check for has_bb_partition
14752 instead of flag_reorder_blocks_and_partition.
14753 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
14754 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
14755 verify_hot_cold_block_grouping.
14756 (partition_hot_cold_basic_blocks): Set has_bb_partition.
14758 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14761 * config/arm/predicates.md (call_insn_operand): New predicate.
14762 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
14763 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
14764 if insn is not a tail call.
14765 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
14767 * config/arm/arm.h (enum reg_class): New caller save register class.
14768 (REG_CLASS_NAMES): Likewise.
14769 (REG_CLASS_CONTENTS): Likewise.
14770 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
14773 2013-05-15 Richard Biener <rguenther@suse.de>
14775 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
14776 of MSG_OPTIMIZED_LOCATIONS.
14777 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
14778 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
14780 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
14781 of MSG_OPTIMIZED_LOCATIONS.
14782 (execute_vect_slp): Likewise.
14783 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
14784 (vect_create_cond_for_alias_checks): Likewise.
14785 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
14786 (vect_recog_widen_mult_pattern): Likewise.
14787 (vect_recog_widen_sum_pattern): Likewise.
14788 (vect_recog_over_widening_pattern): Likewise.
14789 (vect_recog_widen_shift_pattern): Likewise.
14790 (vect_recog_vector_vector_shift_pattern): Likewise.
14791 (vect_recog_divmod_pattern): Likewise.
14792 (vect_recog_mixed_size_cond_pattern): Likewise.
14793 (vect_recog_bool_pattern): Likewise.
14794 (vect_pattern_recog_1): Likewise.
14796 2013-05-15 Martin Jambor <mjambor@suse.cz>
14798 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
14799 non-functions to builtin_unreachable.
14800 * ipa-inline-transform.c (inline_call): Do not assert estimates were
14801 correct when new direct edges were discovered.
14803 2013-05-15 Martin Jambor <mjambor@suse.cz>
14805 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
14806 header, print symbol order instead of node uid, print more information
14807 about indirect edge targets.
14808 (ipa_make_edge_direct_to_target): Print symbol order instead of node
14810 (ipa_make_edge_direct_to_target): Likewise.
14811 (remove_described_reference): Likewise.
14812 (propagate_controlled_uses): Likewise.
14813 (ipa_print_node_params): Also print symbol order.
14814 (ipcp_transform_function): Print symbol order instead of node uids.
14815 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
14816 (cgraph_get_create_real_symbol_node): Likewise.
14817 * ipa-cp.c (print_lattice): Likewise.
14818 (print_all_lattices): Likewise.
14819 (determine_versionability): Likewise.
14820 (initialize_node_lattices): Likewise.
14821 (estimate_local_effects): Likewise.
14822 (update_profiling_info): Likewise.
14823 (create_specialized_node): Likewise.
14824 (perhaps_add_new_callers): Likewise.
14825 (decide_about_value): Likewise.
14826 (decide_whether_version_node): Likewise.
14827 (identify_dead_nodes): Likewise.
14828 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
14829 (dump_inline_summary): Likewise.
14830 (estimate_node_size_and_time): Likewise.
14831 (inline_analyze_function): Likewise.
14832 * ipa-inline.c (report_inline_failed_reason): Likewise.
14833 (want_early_inline_function_p): Likewise.
14834 (edge_badness): Likewise.
14835 (update_edge_key): Likewise.
14836 (inline_small_functions): Likewise. Add dumping of order to two other
14838 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
14839 instead of node uids.
14840 (propagate_pure_const): Likewise.
14841 (propagate_pure_const): Likewise.
14842 * ipa-utils.c (dump_cgraph_node_set): Likewise.
14843 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
14844 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
14846 * tree-pretty-print.c (dump_function_header): Likewise.
14847 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
14848 Print symbol order instead of node uids.
14850 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14852 * config/s390/s390.c (s390_register_move_cost): Don't impose the
14853 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
14855 2013-05-15 Richard Biener <rguenther@suse.de>
14857 PR tree-optimization/57275
14858 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
14859 return value for fail to do runtime alias checks for gather loads.
14861 2013-05-15 Jan Hubicka <jh@suse.cz>
14865 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
14866 weakrefs are not external.
14867 (lto_symtab_merge_decls): Fix thinko when dealing with
14868 non-lto_symtab decls.
14869 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
14870 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
14871 * varpool.c (dump_varpool_node): Dump more flags.
14873 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
14875 * config/i386/i386.c (processor_alias_table): Add instruction
14876 FSGSBASE for AMD bdver3 architecture.
14878 2013-05-14 Jakub Jelinek <jakub@redhat.com>
14880 * tree.c (warn_deprecated_use): Print file:line using locus color.
14881 * diagnostic.c (diagnostic_report_current_module): Print file:line
14882 and file:line:column using locus color.
14884 2013-05-14 Mike Stump <mikestump@comcast.net>
14886 * gdbinit.in: Add __null.
14888 2013-05-14 Mike Stump <mikestump@comcast.net>
14890 * recog.h: Rename struct recog_data to Recog_data.
14891 * recog.c: Likewise.
14892 * reload.c (can_reload_into): Likewise.
14893 * config/picochip/picochip.c: Likewise.
14895 2013-05-14 Mike Stump <mikestump@comcast.net>
14897 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
14899 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
14901 * resource.h (struct resources): Remove unch_memory member.
14902 (CLEAR_RESOURCE): Don't clear unch_memory.
14903 * resource.c (mark_referenced_resources): Don't set it.
14904 (mark_set_resources): Likewise.
14905 (mark_target_live_regs): Don't clear it.
14906 (init_resource_info): Likewise.
14907 * reorg.c (resource_conflicts_p): Don't compare it.
14908 (redundant_insn): Don't set it.
14910 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
14912 * emit-rtl.c (next_label): Remove unused function.
14913 (skip_consecutive_labels, link_cc0_insns): Move to ...
14914 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
14915 only place where these functions are used, and make them static.
14917 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
14919 * fold-const.c (fold_negate_expr): Handle vectors.
14920 (fold_truth_not_expr): Make it static.
14921 (fold_invert_truthvalue): New static function.
14922 (invert_truthvalue_loc): Handle vectors. Do not call
14923 fold_truth_not_expr directly.
14924 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
14925 <TRUTH_NOT_EXPR>: Do not cast to boolean.
14926 (fold_comparison): Handle vector constants.
14927 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
14928 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
14929 * tree.h (fold_truth_not_expr): Remove declaration.
14931 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
14933 * config/aarch64/aarch64-simd.md
14934 (aarch64_vcond_internal<mode>): Rename to...
14935 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
14936 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
14937 float modes. Clarify all iterator modes.
14938 (vcond<mode><mode>): Use new name for vcond expanders.
14939 (vcond<v_cmp_result><mode>): Likewise.
14940 (vcondu<mode><mode>: Likewise.
14941 * config/aarch64/iterators.md (VDQF_COND): New.
14943 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
14946 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
14947 variable for the shift amount. Check that we shift by non-negative
14950 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
14953 * config/arm/arm.h (EPILOGUE_USES): Only return true
14954 for LR_REGNUM after epilogue_completed.
14956 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
14958 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
14959 is error_mark_node.
14961 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
14964 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
14965 and Solaris 11+/x86 with gld.
14966 * configure: Regenerate.
14968 2013-05-14 Jakub Jelinek <jakub@redhat.com>
14970 * expmed.c (expand_shift_1): Canonicalize rotates by
14971 constant bitsize / 2 to bitsize - 1.
14972 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
14973 case ROTATERT>: Likewise.
14976 2013-05-10 Jakub Jelinek <jakub@redhat.com>
14978 * config/i386/i386.md (rotateinv): New code attr.
14979 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
14980 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
14981 roll $31, %eax, etc.
14983 2013-05-14 Richard Biener <rguenther@suse.de>
14985 PR middle-end/57235
14986 * tree-eh.c (sink_clobbers): Give up for successors with
14987 multiple predecessors and no virtual uses.
14989 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
14991 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
14992 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
14994 2013-05-14 Jakub Jelinek <jakub@redhat.com>
14996 PR middle-end/57251
14997 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
14998 the case when both op0 and op1 have VOIDmode.
15000 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
15002 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
15003 in multiply-accumulate mode.
15005 2013-05-13 Guozhi Wei <carrot@google.com>
15007 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
15009 2013-05-13 Kai Tietz <ktietz@redhat.com>
15012 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
15013 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
15014 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
15015 * config/i386/i386.c (ix86_option_override_internal): Likewise.
15016 (ix86_expand_prologue): Likewise.
15017 (ix86_expand_split_stack_prologue): Likewise.
15018 (legitimate_pic_address_disp_p): Likewise.
15019 (legitimize_pic_address): Likewise.
15020 (legitimize_tls_address): Likewise.
15021 (legitimize_pe_coff_symbol): Likewise.
15022 (output_pic_addr_const): Likewise.
15023 (construct_plt_address): Likewise.
15024 (ix86_expand_call): Likewise.
15025 (x86_output_mi_thunk): Likewise.
15026 (x86_function_profiler): Likewise.
15028 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
15030 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
15031 similar switch cases.
15032 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
15033 (aarch64_simd_mov_to_<mode>low): Delete.
15034 (aarch64_simd_mov_to_<mode>high): Delete.
15035 (move_lo_quad_<mode>): Add w<-r alternative.
15036 (aarch64_simd_move_hi_quad_<mode>): Likewise.
15037 (aarch64_simd_mov_from_*): Update type attribute.
15038 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
15041 2013-05-13 Jan Hubicka <jh@suse.cz>
15043 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
15044 * config/i386/i386.c (ix86_compute_frame_layout,
15045 ix86_expand_epilogue, emit_i387_cw_initialization,
15046 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
15047 ix86_local_alignment): Fix use of size/speed predicates.
15049 2013-05-13 Jakub Jelinek <jakub@redhat.com>
15051 PR tree-optimization/45216
15052 PR tree-optimization/57157
15053 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
15054 the (-Y) & (B - 1) variant if OP is |.
15055 * expmed.c (expand_shift_1): For rotations by const0_rtx just
15056 return shifted. Use (-op1) & (prec - 1) as other_amount
15057 instead of prec - op1.
15059 2013-05-13 Martin Jambor <mjambor@suse.cz>
15061 PR middle-end/42371
15062 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
15063 (ipa_constant_data): New type.
15064 (ipa_jump_func): Use ipa_constant_data to hold information about
15065 constant jump functions.
15066 (ipa_get_jf_constant): Adjust to jump function type changes.
15067 (ipa_get_jf_constant_rdesc): New function.
15068 (ipa_param_descriptor): New field controlled_uses.
15069 (ipa_get_controlled_uses): New function.
15070 (ipa_set_controlled_uses): Likewise.
15071 * ipa-ref.h (ipa_find_reference): Declare.
15072 * ipa-prop.c (ipa_cst_ref_desc): New type.
15073 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
15075 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
15076 New parameter cs. Adjust all callers.
15077 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
15078 (remove_described_reference): New function.
15079 (jfunc_rdesc_usable): Likewise.
15080 (try_make_edge_direct_simple_call): Decrement controlled use count,
15081 attempt to remove reference if it hits zero.
15082 (combine_controlled_uses_counters): New function.
15083 (propagate_controlled_uses): Likewise.
15084 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
15085 (ipa_edge_duplication_hook): Duplicate reference descriptions.
15086 (ipa_print_node_params): Print described use counter.
15087 (ipa_write_jump_function): Adjust to jump function type changes.
15088 (ipa_read_jump_function): New parameter CS, pass it to
15089 ipa_set_jf_constant. Adjust caller.
15090 (ipa_write_node_info): Stream controlled use count
15091 (ipa_read_node_info): Likewise.
15092 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
15094 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
15095 count. Remove cloning-added reference if it reaches zero.
15096 * ipa-ref.c (ipa_find_reference): New function.
15098 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
15100 * config/i386/i386.c (processor_target_table): Modified default
15101 alignment values for AMD BD and BT architectures.
15103 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
15105 * tree-vect-generic.c (uniform_vector_p): Move ...
15106 * tree.c (uniform_vector_p): ... here.
15107 * tree.h (uniform_vector_p): Declare it.
15108 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
15111 2013-05-13 Jakub Jelinek <jakub@redhat.com>
15113 PR tree-optimization/57230
15114 * tree-ssa-strlen.c (handle_char_store): Record length for
15115 array store from STRING_CST.
15117 PR tree-optimization/57230
15118 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
15121 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
15123 * config/epiphany/epiphany.c (epiphany_init): Check size of
15124 NUM_MODES_FOR_MODE_SWITCHING.
15125 (epiphany_expand_prologue):
15126 Remove CONFIG_REGNUM initial value handling code.
15127 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
15128 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
15129 (emit_set_fp_mode, epiphany_mode_after): Likewise.
15130 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
15131 Don't return 1 for FP_MODE_NONE.
15132 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
15133 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
15134 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
15135 * config/epiphany/epiphany.md (save_config): New pattern.
15137 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
15139 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
15141 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
15143 * config/i386/i386.md (memory): Handle sseishft1.
15144 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
15145 (*vec_extractv2di_1): Ditto.
15147 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
15149 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
15152 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
15154 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
15155 Add mthumb/march=armv7-a multilib.
15156 Add mthumb/march=armv7-r multilib.
15157 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
15159 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
15161 * config/v850/t-rtems: Add more multilibs.
15163 2013-05-10 Richard Biener <rguenther@suse.de>
15165 PR tree-optimization/57214
15166 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
15167 not propagate from SSA names that occur in abnormal PHI nodes.
15169 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
15171 * stor-layout.c (element_precision): New function.
15172 * machmode.h (element_precision): Declare it.
15173 * tree.c (build_minus_one_cst): New function.
15174 (element_precision): Likewise.
15175 * tree.h (build_minus_one_cst): Declare new function.
15176 (element_precision): Likewise.
15177 * fold-const.c (operand_equal_p): Use element_precision.
15178 (fold_binary_loc): Handle vector types.
15179 * convert.c (convert_to_integer): Use element_precision.
15180 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
15183 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
15185 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
15186 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
15187 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
15188 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
15189 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
15190 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
15191 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
15192 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
15193 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
15194 (Uuw8): New constraints.
15195 (Usb4): Move into alphabetical order.
15196 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
15197 (sd8_operand, ub8_operand, uw8_operand): New predicates.
15198 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
15199 previously unnamed patterns.
15200 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
15201 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
15202 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
15203 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
15204 of set_attr_alternative/if_then_else. Use extended_mips16 instead
15205 of specific lengths.
15207 2013-05-10 Jakub Jelinek <jakub@redhat.com>
15209 * config/i386/i386.md (rotateinv): New code attr.
15210 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
15211 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
15212 roll $31, %eax, etc.
15214 PR tree-optimization/45216
15215 PR tree-optimization/57157
15216 * tree-ssa-forwprop.c (simplify_rotate): New function.
15217 (ssa_forward_propagate_and_combine): Call it.
15219 2013-05-10 Richard Biener <rguenther@suse.de>
15221 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
15222 disable peeling when we version for aliasing.
15223 (vector_alignment_reachable_p): Honor explicit user alignment.
15224 (vect_supportable_dr_alignment): Likewise.
15225 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
15226 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
15227 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
15228 then peeling to arrange for the cost-model check to come first.
15230 2013-05-10 Alan Modra <amodra@gmail.com>
15232 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
15233 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
15234 * configure: Regenerate.
15236 2013-05-10 Alan Modra <amodra@gmail.com>
15239 * varasm.c (default_elf_select_section): Move !DECL_P check..
15240 (get_named_section): ..to here before calling get_section_name.
15242 (default_section_type_flags): Add DECL_P check.
15243 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
15244 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
15246 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
15248 * config/epiphany/epiphany.c (epiphany_expand_prologue):
15249 When using gen_stack_adjust_str with a register offset, add a
15250 REG_FRAME_RELATED_EXPR note.
15252 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
15254 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
15255 (*vec_extractv4si_zext_mem): Ditto.
15256 (*vec_extractv2di): Add 0->x and x->x alternatives.
15257 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
15258 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
15260 2013-05-09 Jason Merrill <jason@redhat.com>
15262 N3639 C++1y VLA support
15263 * gimplify.c (gimplify_vla_decl): Don't touch an existing
15266 * tree.c (build_constructor_va): New.
15267 * tree.h: Declare it.
15269 2013-05-09 Martin Jambor <mjambor@suse.cz>
15272 * gimple-fold.c (canonicalize_constructor_val): Call
15273 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
15275 2013-05-09 Jan Hubicka <jh@suse.cz>
15276 Richard Biener <rguenther@suse.de>
15279 * symtab.c (symtab_make_decl_local): Do not add private names.
15281 2013-05-09 Jan Hubicka <jh@suse.cz>
15284 * symtab.c (insert_to_assembler_name_hash): Handle clones.
15285 (unlink_from_assembler_name_hash): Likewise.
15286 (symtab_prevail_in_asm_name_hash, symtab_register_node,
15287 symtab_unregister_node, symtab_initialize_asm_name_hash,
15288 change_decl_assembler_name): Update.
15290 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
15292 * config/aarch64/aarch64.md: New movtf split.
15293 (*movtf_aarch64): Update.
15294 (aarch64_movdi_tilow): Handle TF modes and rename to
15295 aarch64_movdi_<mode>low.
15296 (aarch64_movdi_tihigh): Handle TF modes and rename to
15297 aarch64_movdi_<mode>high
15298 (aarch64_movtihigh_di): Handle TF modes and rename to
15299 aarch64_mov<mode>high_di
15300 (aarch64_movtilow_di): Handle TF modes and rename to
15301 aarch64_mov<mode>low_di
15302 (aarch64_movtilow_tilow): Remove spurious whitespace.
15303 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
15305 (aarch64_print_operand): Update.
15307 2013-05-09 Alan Modra <amodra@gmail.com>
15309 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
15311 * configure: Regenerate.
15313 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
15315 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
15316 splitter preparation statements.
15317 * config/i386/sse.md (*vec_extract* splitters): Ditto.
15318 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
15321 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15323 * gimple-ssa-strength-reduction.c (count_candidates): Change
15324 return value to int.
15325 (analyze_candidates_and_replace): Change type of length to int.
15327 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
15329 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
15330 (*vec_extract<mode>): Use VI12_128 mode iterator.
15331 (*vec_extract<mode>_mem): Ditto.
15332 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
15335 2013-05-08 Diego Novillo <dnovillo@google.com>
15340 2012-08-17 Diego Novillo <dnovillo@google.com>
15343 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
15344 * config.in: Regenerate.
15345 * configure: Regenerate.
15346 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
15348 2013-05-08 Jan Hubicka <jh@suse.cz>
15351 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
15352 * cgraph.h (symtab_node_base): Add unique_name.
15353 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
15354 input_overwrite_node, input_varpool_node): Stream unique_name.
15355 * cgraphclones.c (cgraph_create_virtual_clone,
15356 cgraph_function_versioning): Set unique_name.
15357 * ipa.c (function_and_variable_visibility): Set unique_name.
15359 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15361 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
15362 (alloc_cand_and_find_basis): Restrict conditional candidate
15363 processing to CAND_MULTs.
15365 2013-05-08 Jan Hubicka <jh@suse.cz>
15368 lto-symtab.c (lto_symtab_symbol_p): New function.
15369 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
15370 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
15371 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
15372 Skip static symbols.
15374 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
15376 PR tree-optimization/57200
15377 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
15378 Only call inform if the preceding warning_at returns true.
15380 2013-05-07 Han Shen <shenhan@google.com>
15382 * cfgexpand.c (record_or_union_type_has_array_p): New function.
15383 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
15384 * common.opt (fstack-protector-strong): New option.
15385 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
15386 * doc/invoke.texi (Optimization Options): Document
15387 "-fstack-protector-strong".
15388 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
15390 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
15392 * config/mips/mips.c (mips_machine_reorg2): Return 0.
15394 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
15396 * ira.c (update_equiv_regs): Add insn having equiv memory even if
15397 it is not lhs of the insn.
15398 (setup_reg_equiv): Remove insn having equiv memory which it is not
15400 * lra-constraints.c (process_address): Try to improve generation
15401 code for address base + disp.
15402 (lra_constraints): Make correct the code for checking insn setting
15403 up backward equivalence. Remove insn only if it is in the init
15405 * lra-eliminations.c (update_reg_eliminate): Change return value.
15406 (lra_eliminate): Use the result.
15408 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
15410 * config/i386/sse.md (ssescalarnummask): New mode attribute.
15411 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
15412 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
15413 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
15414 register target operands.
15415 (*vec_extractv8hi_sse2): New pattern.
15416 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
15417 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
15418 (*vec_extract<mode>_mem): New insn and split pattern.
15420 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
15422 * config/arm/arm.c (arm_asan_shadow_offset): New function.
15423 (TARGET_ASAN_SHADOW_OFFSET): Define.
15424 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
15425 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
15427 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15429 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
15430 (incr_vec_index): Return -1 if increment not found.
15431 (create_add_on_incoming_edge): Assert if increment not found.
15432 (record_increment): Limit number of increments recorded.
15433 (all_phi_incrs_profitable): Return false if an increment not found.
15434 (replace_profitable_candidates): Don't process increments that were
15436 (analyze_candidates_and_replace): Limit size of incr_vec.
15438 2013-05-07 Richard Biener <rguenther@suse.de>
15440 * calls.c (special_function_p): setjmp-like functions are leaf.
15441 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
15442 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
15444 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
15446 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
15447 (aarch64_simd_mov<mode>): New expander.
15448 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
15449 (aarch64_simd_mov_to_<mode>high): Likewise.
15450 (aarch64_simd_mov_from_<mode>low): Likewise.
15451 (aarch64_simd_mov_from_<mode>high): Likewise.
15452 (aarch64_dup_lane<mode>): Update.
15453 (aarch64_dup_lanedi): New instruction pattern.
15454 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
15455 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
15457 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15459 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
15460 (replace_mult_candidate): Remove unnecessary argument; remove
15461 unnecessary parameter from call to introduce_cast_before_cand.
15462 (replace_unconditional_candidate): Remove unnecessary parameter
15463 from call to replace_mult_candidate.
15464 (replace_conditional_candidate): Likewise.
15465 (insert_initializers): Use make_temp_ssa_name.
15466 (introduce_cast_before_cand): Remove unnecessary argument; use
15467 make_temp_ssa_name.
15468 (replace_one_candidate): Remove unnecessary argument; remove
15469 unnecessary parameter from calls to introduce_cast_before_cand.
15470 (replace_profitable_candidates): Remove unnecessary parameters
15471 from calls to replace_one_candidate.
15473 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15475 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
15476 phi def as possibly hiding a basis for a CAND_ADD whose operands
15477 have been commuted in the analysis.
15478 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
15480 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
15482 * config/aarch64/aarch64.md
15483 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
15484 shift value between 0-4.
15486 2013-05-07 Richard Biener <rguenther@suse.de>
15488 * double-int.h (rshift): New overload.
15489 * double-int.c (rshift): New function.
15490 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
15491 (create_reference_ops_from_ref): Remove.
15492 (vn_reference_insert): Use shared ops for constructing the
15493 reference and copy it.
15495 2013-05-07 Richard Biener <rguenther@suse.de>
15497 PR middle-end/57190
15498 * tree-eh.c (sink_clobbers): Properly propagate
15499 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
15501 2013-05-07 Jakub Jelinek <jakub@redhat.com>
15503 PR tree-optimization/57149
15504 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
15505 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
15506 collect_phi_def_edges, execute_late_warn_uninitialized): Use
15507 uninit_undefined_value_p instead of ssa_undefined_value_p.
15510 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
15511 for modifier == EXPAND_INITIALIZER.
15513 2013-05-07 Anton Blanchard <anton@samba.org>
15515 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
15516 for powerpc64 little endian.
15517 * configure: Regenerate.
15519 2013-05-06 Graham Stott <grahams@btinternet.com>
15521 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
15522 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
15523 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
15524 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
15526 2013-05-06 Graham Stott <grahams@btinternet.com>
15528 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
15529 codes which allow non-lvalues.
15531 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
15533 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
15534 components are all 1s.
15535 (integer_minus_onep): New function.
15536 * tree.h (integer_minus_onep): Declare it.
15537 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
15538 integer_minus_onep instead of integer_all_onesp.
15540 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
15543 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
15544 variations of these patterns.
15546 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
15548 * config/i386/i386.md (isa): Add x64_sse4 member.
15549 (enabled): Handle x64_sse4.
15550 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
15551 instruction for 64bit SSE4_1 targets. Update insn attributes.
15552 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
15553 instruction for SSE4_1 targets. Update insn attributes.
15554 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
15555 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
15556 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
15558 (*vec_extractv4si): Rename from *sse4_1_pextrd.
15559 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
15560 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
15562 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
15565 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
15567 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
15569 * final.c (do_assembler_dialects): Don't handle curly braces and
15570 vertical bar escaped by % as dialect delimiters.
15571 (output_asm_insn): Print curly braces and vertical bar if escaped
15572 by % and ASSEMBLER_DIALECT defined.
15573 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
15574 * doc/tm.texi: Regenerated.
15576 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
15578 * config/mips/mips.c: Include tree-pass.h.
15579 (mips_reorg): Split in pre- and post-dbr_schedule parts.
15580 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
15581 (pass_mips_machine_reorg2): New machine specific pass.
15582 (insert_pass_mips_machine_reorg2): New pass plugin definition.
15583 (mips_option_override): Register the new pass.
15584 * rtl.h (cleanup_barriers): Remove prototype.
15585 (dbr_schedule): Likewise.
15586 * jump.c (cleanup_barriers): Make static.
15587 * reorg.c (dbr_schedule): Likewise.
15589 2013-05-06 Richard Biener <rguenther@suse.de>
15591 PR tree-optimization/57185
15592 * tree-parloops.c (add_field_for_reduction): Handle anonymous
15593 SSA names properly.
15595 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
15598 * config/i386/i386.c (add_parameter_dependencies): Add dependence
15599 between "first_arg" and "insn", not "last" and "insn".
15601 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15603 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
15604 (find_candidates_in_block): Re-enable slsr_process_phi.
15605 (create_phi_basis): Fix double counting of candidate adjustment.
15607 2013-05-06 Richard Biener <rguenther@suse.de>
15609 PR middle-end/57147
15610 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
15611 the edge is also fallthru, preserve it and just clear the
15613 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
15614 also complex, preserve that and just clear the fallthru flag.
15615 * tree-inline.c (update_ssa_across_abnormal_edges): Also
15616 update virtual operands.
15618 2013-05-06 Alan Modra <amodra@gmail.com>
15620 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
15621 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
15622 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
15623 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
15624 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
15625 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
15627 2013-05-06 Alan Modra <amodra@gmail.com>
15629 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
15630 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
15631 (DEFAULT_ASM_ENDIAN): Define.
15632 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
15633 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
15634 Update -K PIC clause from sysv4.h.
15635 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
15636 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
15638 2013-05-06 Alan Modra <amodra@gmail.com>
15640 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
15641 twice for little-endian.
15642 (ashrdi3_no_power, ashrdi3): Support little-endian.
15644 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
15647 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
15648 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
15650 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
15651 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
15652 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
15655 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
15656 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
15658 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
15659 * config/pa/pa.opt: Make mbig-switch a no-op.
15660 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
15661 (CASE_VECTOR_MODE): Always return SImode.
15662 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
15663 for the !TARGET_BIG_SWITCH case.
15664 * config/pa/pa-linux.h: Likewise.
15665 * config/pa/pa-openbsd.h: Likewise.
15666 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
15667 * config/pa/pa.md (short_jump): Remove define_insn.
15668 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
15669 (casesi0): Remove define_insn.
15670 (type): Remove btable_branch.
15671 (pa_combine_type): Likewise.
15672 (in_nullified_branch_delay): Likewise.
15673 (in_call_delay): Likewise.
15674 (define_delay): Likewise.
15675 (define_insn_reservation "Z3"): Likewise.
15676 (define_insn_reservation "Z4"): Likewise.
15677 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
15678 (pa_adjust_insn_length): Remove adjustment for btable branches.
15679 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
15682 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
15684 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
15685 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
15686 Add m->r,x alternatives.
15687 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
15688 splitters using SWI48x mode iterator.
15689 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
15690 TARGET_64BIT. Add m->x alternative.
15691 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
15692 Add o->x alternative. Enable for TARGET_SSE.
15693 (sse_storeq): Remove expander.
15694 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
15695 with memory input operand.
15696 (*vec_extractv2di_1 splitter): New.
15697 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
15698 * config/i386/i386.md (ssevecmodelower): New mode attribute.
15700 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
15702 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
15703 (INT_LOWPART): Delete.
15704 (extract_MB): Adjust.
15705 (extract_ME): Adjust.
15706 (print_operand): Adjust.
15708 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
15710 * config/rs6000/predicates.md (reg_or_add_cint_operand,
15711 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
15712 (reg_or_logical_cint_operand, easy_fp_constant,
15713 logical_const_operand): Delete "CONST_DOUBLE" case.
15714 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
15715 "HOST_BITS_PER_WIDE_INT == 64" test.
15716 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
15717 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
15718 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
15719 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
15721 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
15722 CONST_DOUBLE DImode/VOIDmode case.
15723 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
15724 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
15725 CONST_DOUBLE VOIDmode case.
15726 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
15727 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
15728 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
15729 Delete CONST_DOUBLE case.
15730 (splitters for mov FMOVE64 const_double): Delete
15731 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
15732 "HOST_BITS_PER_WIDE_INT >= 64" test.
15733 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
15735 (mov DI const_double): Delete.
15737 2013-05-04 Jakub Jelinek <jakub@redhat.com>
15739 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
15740 on op shows all bits zero in mode of a lowpart subreg, return zero.
15742 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
15745 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
15746 to save TFmode registers and DImode to save TImode registers for
15747 caller save operations.
15748 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
15749 mark being partially clobbered since they only use the first
15752 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
15753 and TDmode only use the upper 64-bits of each VSX register.
15755 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15757 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
15758 (find_candidates_in_block): Disable slsr_process_phi.
15760 2013-05-03 Guozhi Wei <carrot@google.com>
15762 * coverage.c (coverage_obj_init): Move the construction of gcov
15764 (build_init_ctor): ... here.
15766 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15768 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
15769 (slsr_cand_d): Redefine def_phi.
15770 (stride_status, phi_adjust_status, count_phis_status): New enums.
15771 (find_phi_def): New.
15772 (find_basis_for_base_expr): New.
15773 (find_basis_for_candidate): Handle hidden bases.
15774 (alloc_cand_and_find_basis): Handle phi candidates.
15775 (slsr_process_phi): New.
15776 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
15777 (create_mul_imm_cand): Likewise.
15778 (create_add_ssa_cand): Exclude phi base candidates.
15779 (create_add_imm_cand): Likewise.
15780 (slsr_process_cast): Likewise.
15781 (slsr_process_copy): Likewise.
15782 (find_candidates_in_block): Handle phi candidates.
15783 (dump_candidate): Likewise.
15784 (unconditional_cands): Delete.
15785 (unconditional_cands_with_known_stride_p): Delete.
15786 (phi_dependent_cand_p): New.
15787 (cand_increment): Handle phi-dependent candidates.
15788 (replace_dependent): Delete.
15789 (replace_mult_candidate): New.
15790 (replace_unconditional_candidate): New.
15791 (incr_vec_index): Move to avoid forward reference.
15792 (create_add_on_incoming_edge): New.
15793 (create_phi_basis): New.
15794 (replace_dependents): Delete.
15795 (replace_conditional_candidate): New.
15796 (phi_add_costs): New.
15797 (replace_uncond_cands_and_profitable_phis): New.
15798 (record_increment): Handle phi adjustments.
15799 (record_phi_increments): New.
15800 (record_increments): Handle phi adjustments.
15801 (phi_incr_cost): New.
15802 (lowest_cost_path): Handle phis.
15803 (total_savings): Likewise.
15804 (analyze_increments): Likewise.
15805 (ncd_with_phi): New.
15806 (ncd_of_cand_and_phis): New.
15807 (nearest_common_dominator_for_cands): Handle phi increments.
15808 (all_phi_incrs_profitable): New.
15809 (replace_profitable_candidates): Handle phi-dependent candidates.
15810 (analyze_candidates_and_replace): Likewise.
15812 2013-05-03 Teresa Johnson <tejohnson@google.com>
15815 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
15816 do not exceed REG_BR_PROB_BASE.
15818 2013-05-03 Jeff Law <law@redhat.com>
15820 PR tree-optimization/57144
15821 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
15822 operand of the condition will bit into the new type when eliminating
15823 a cast feeding a condition.
15825 2013-05-03 Jakub Jelinek <jakub@redhat.com>
15827 PR rtl-optimization/57130
15828 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
15829 of COMPARE as in_code to the recursive call if needed.
15831 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
15833 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
15834 (enabled): Handle new members.
15835 * config/i386/sse.md (*vec_concatv2si): Merge from
15836 *vec_concatv2si_sse2 and vec_concatv2si_sse.
15837 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
15839 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
15841 PR tree-optimization/57027
15842 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
15843 for fnms opportunity, check we got the prerequisite kind
15844 of tree / gimple before using accessor functions.
15846 2013-05-03 Richard Biener <rguenther@suse.de>
15848 * double-int.h (lshift): New overload without precision
15849 and arith argument.
15850 (operator *=, operator +=, operator -=): Move ...
15851 * double-int.c (operator *=, operator +=, operator -=): ... here
15852 and implement more efficiently.
15853 (mul_double_with_sign): Remove.
15854 (lshift_double): Adjust to take unsinged shift argument, push
15855 dispatching code to callers.
15856 (mul_double_wide_with_sign): Add early out for callers that
15857 are not interested in high parts or overflow.
15858 (lshift): New function.
15859 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
15860 dispatch code here.
15861 (lrotate, rrotate): Use logical shifts.
15862 * expr.c (get_inner_reference): Use lshift.
15863 * fixed-value.c (do_fixed_divide): Likewise.
15864 * tree-dfa.c (get_ref_base_and_extent): Likewise.
15865 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
15866 (indirect_refs_may_alias_p): Likewise.
15867 (stmt_kills_ref_p_1): Likewise.
15869 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
15871 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
15873 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
15875 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
15876 scalar form of FABD instruction.
15878 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
15880 * lra-constraints.c (process_alt_operands): Add checking alt
15881 number to choose the best alternative.
15883 2013-05-02 Richard Biener <rguenther@suse.de>
15885 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
15886 bitmap and its handling.
15887 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
15889 2013-05-02 Richard Biener <rguenther@suse.de>
15891 PR middle-end/57140
15892 * tree-inline.c (copy_loops): Properly handle removed loops.
15893 (copy_cfg_body): Mark destination loops for fixup if source
15894 loops needed fixup.
15896 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
15899 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
15900 generating simple_return for naked functions.
15902 2013-05-02 Martin Jambor <mjambor@suse.cz>
15904 PR middle-end/56988
15905 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
15906 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
15908 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
15909 ipa_agg_replacement_value structures.
15910 (known_aggs_to_agg_replacement_list): Likewise.
15911 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
15912 (read_agg_replacement_chain): Likewise.
15913 (ipcp_transform_function): Also check that by_ref flags match.
15915 2013-05-02 Richard Biener <rguenther@suse.de>
15917 * graphds.h (struct graph): Add obstack member.
15918 * graphds.c (new_graph): Initialize obstack and allocate
15920 (add_edge): Allocate edge from the obstack.
15921 (free_graph): Free the obstack instead of all edges and vertices.
15923 2013-05-02 Teresa Johnson <tejohnson@google.com>
15925 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
15927 * cfg.c (update_bb_profile_for_threading): Ditto.
15928 * tree-inline.c (copy_bb): Ditto.
15929 (copy_edges_for_bb): Ditto.
15930 (initialize_cfun): Ditto.
15931 (copy_cfg_body): Ditto.
15932 (expand_call_inline): Ditto.
15933 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
15934 (estimate_node_size_and_time): Ditto.
15935 (inline_merge_summary): Ditto.
15936 * cgraphclones.c (cgraph_clone_edge): Ditto.
15937 (cgraph_clone_node): Ditto.
15938 * sched-rgn.c (compute_dom_prob_ps): Ditto.
15939 (compute_trg_info): Ditto.
15941 2013-05-02 Ian Bolton <ian.bolton@arm.com>
15943 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
15944 S reg when fp attribute set.
15945 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
15947 2013-05-02 Ian Bolton <ian.bolton@arm.com>
15949 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
15951 (*and_one_cmplsi3_compare0_uxtw): Likewise.
15952 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
15953 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
15955 2013-05-02 Richard Biener <rguenther@suse.de>
15957 * tree-scalar-evolution.c (scev_info_hasher): Remove.
15958 (struct instantiate_cache_entry): New type.
15959 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
15960 (struct instantiate_cache_type): New type.
15961 (set_instantiated_value, get_instantiated_value): Remove.
15962 (get_instantiated_value_entry): New function.
15963 (instantiate_scev_name): Use the new cache and adjust.
15964 (instantiate_scev_poly): Adjust.
15965 (instantiate_scev_binary): Likewise.
15966 (instantiate_array_ref): Likewise.
15967 (instantiate_scev_convert): Likewise.
15968 (instantiate_scev_not): Likewise.
15969 (instantiate_scev_3): Likewise.
15970 (instantiate_scev_2): Likewise.
15971 (instantiate_scev_r): Likewise.
15972 (instantiate_scev): Likewise.
15973 (resolve_mixers): Likewise.
15975 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
15978 * lra-constraints.c (best_small_class_operands_num): Remove.
15979 (process_alt_operands): Remove small_class_operands_num. Take
15980 small classes operands into losers and only if the operand is not
15981 matched. Modify debugging output.
15982 (curr_insn_transform): Remove best_small_class_operands_num.
15985 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
15987 * config/aarch64/aarch64-builtins.c
15988 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
15989 * config/aarch64/aarch64-simd-builtins.def
15990 (reduc_splus_): Add new modes.
15991 (reduc_uplus_): New.
15992 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
15993 (reduc_uplus_v4sf): Likewise.
15994 (reduc_splus_v4sf): Likewise.
15995 (aarch64_addv<mode>): Likewise.
15996 (reduc_uplus_<mode>): Likewise.
15997 (reduc_splus_<mode>): Likewise.
15998 (aarch64_addvv2di): Likewise.
15999 (reduc_uplus_v2di): Likewise.
16000 (reduc_splus_v2di): Likewise.
16001 (aarch64_addvv2si): Likewise.
16002 (reduc_uplus_v2si): Likewise.
16003 (reduc_splus_v2si): Likewise.
16004 (reduc_<sur>plus_<mode>): New.
16005 (reduc_<sur>plus_v2di): Likewise.
16006 (reduc_<sur>plus_v2si): Likewise.
16007 (reduc_<sur>plus_v4sf): Likewise.
16008 (aarch64_addpv4sf): Likewise.
16009 * config/aarch64/arm_neon.h
16010 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
16011 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
16012 add UNSPEC_SADDV, UNSPEC_UADDV.
16014 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
16016 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
16018 * config/aarch64/arm_neon.h
16019 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
16021 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
16023 * config/aarch64/aarch64-builtins
16024 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
16026 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
16028 * config/aarch64/aarch64-simd-builtins.def
16029 (reduc_smax_): New.
16030 (reduc_smin_): Likewise.
16031 (reduc_umax_): Likewise.
16032 (reduc_umin_): Likewise.
16033 (reduc_smax_nan_): Likewise.
16034 (reduc_smin_nan_): Likewise.
16037 (smax): Update for V2SF, V4SF and V2DF modes.
16040 (smin_nan): Likewise.
16041 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
16042 (<su><maxmin><mode>3): ...This, refactor.
16043 (s<maxmin><mode>3): New.
16044 (<maxmin_uns><mode>3): Likewise.
16045 (reduc_<maxmin_uns>_<mode>): Refactor.
16046 (reduc_<maxmin_uns>_v4sf): Likewise.
16047 (reduc_<maxmin_uns>_v2si): Likewise.
16048 (aarch64_<fmaxmin><mode>: Remove.
16049 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
16051 (vmin<q>_f<32,64>): Likewise.
16052 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
16054 (su): Add mappings for smax, smin, umax, umin.
16056 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
16057 (FMAXMIN): Rename as...
16058 (FMAXMIN_UNS): ...This.
16060 (fmaxminv): Likewise.
16061 (fmaxmin): Likewise.
16063 (maxmin_uns_op): Likewise.
16065 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
16067 * config/aarch64/arm_neon.h
16068 (vac<ge, gt><sd>_f<32, 64>): Rename to...
16069 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
16070 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
16072 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
16074 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
16075 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
16077 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
16079 * config/aarch64/aarch64-simd.md
16080 (vcond<mode>_internal): Handle special cases for constant masks.
16081 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
16082 (vcondu<mode><mode>): Likewise.
16083 (vcond<v_cmp_result><mode>): New.
16085 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
16087 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
16088 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
16089 * config/aarch64/aarch64-simd-builtins.def
16090 (cmeq): Update to BUILTIN_VALLDI.
16095 * config/aarch64/arm_neon.h
16096 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
16097 to builtins or C as appropriate.
16099 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
16101 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
16103 (cmhi): Rename to...
16105 * config/aarch64/aarch64-simd.md
16106 (simd_mode): Add SF.
16107 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
16108 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
16109 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
16110 (cstore<mode>_neg): ...This.
16111 * config/aarch64/iterators.md
16113 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
16114 (COMPARISONS): New.
16115 (UCOMPARISONS): Likewise.
16116 (optab): Add missing comparisons.
16122 (VCMP_S): Likewise.
16123 (VCMP_U): Likewise.
16124 (V_cmp_result): Add DF, SF modes.
16125 (v_cmp_result): Likewise.
16127 (vmtype): Likewise.
16128 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
16130 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
16132 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
16133 define_insn to define_insn_and_split.
16134 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
16135 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
16136 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
16137 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
16138 (thumb2_negscc): Likewise.
16140 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
16142 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
16144 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
16146 * config/arm/thumb2.md: Remove trailing whitespaces.
16148 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
16150 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
16151 Use gen_int_mode rather than GEN_INT.
16153 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
16155 * value-prof.c (stream_in_histogram_value): Remove the strayed
16158 2013-04-30 Richard Biener <rguenther@suse.de>
16160 PR middle-end/57122
16161 * cfghooks.c (split_edge): Properly check for the loop latch edge.
16163 2013-04-30 Richard Biener <rguenther@suse.de>
16165 PR middle-end/57107
16166 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
16168 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
16170 PR rtl-optimization/56957
16171 PR rtl-optimization/57105
16172 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
16173 variable. Use just INSN_UID for determining whether an insn
16174 should be only disconnected from the insn stream.
16175 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
16177 2013-04-30 Jakub Jelinek <jakub@redhat.com>
16179 PR tree-optimization/57104
16180 * tsan.c (instrument_expr): Don't instrument accesses to
16181 DECL_HARD_REGISTER VAR_DECLs.
16183 2013-04-30 Richard Biener <rguenther@suse.de>
16185 * function.h (loops_for_fn): New inline function.
16186 (set_loops_for_fn): Likewise.
16187 * cfgloop.h (place_new_loop): Add struct function parameter.
16188 (get_loop): Likewise.
16189 (get_loops): Likewise.
16190 (number_of_loops): Likewise.
16191 (fel_next): Adjust.
16192 (fel_init): Likewise.
16193 * cfg.c (get_loop_copy): Adjust.
16194 * cfgloop.c (flow_loops_dump): Likewise.
16195 (record_loop_exits): Likewise.
16196 (verify_loop_structure): Likewise.
16197 * cfgloopanal.c (mark_irreducible_loops): Likewise.
16198 (estimate_reg_pressure_cost): Likewise.
16199 (mark_loop_exit_edges): Likewise.
16200 * cfgloopmanip.c (place_new_loop): Likewise.
16201 (add_loop): Likewise.
16202 (duplicate_loop): Likewise.
16203 * graph.c (draw_cfg_nodes): Likewise.
16204 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
16205 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
16206 (extract_affine_chrec): Likewise.
16207 (build_scop_iteration_domain): Likewise.
16208 * graphite.c (graphite_initialize): Likewise.
16209 * ira-build.c (create_loop_tree_nodes): Likewise.
16210 (more_one_region_p): Likewise.
16211 (rebuild_regno_allocno_maps): Likewise.
16212 (mark_loops_for_removal): Likewise.
16213 (mark_all_loops_for_removal): Likewise.
16214 (remove_unnecessary_regions): Likewise.
16215 (ira_build): Likewise.
16216 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
16217 * loop-init.c (fix_loop_structure): Likewise.
16218 (gate_rtl_move_loop_invariants): Likewise.
16219 (gate_rtl_unswitch): Likewise.
16220 (gate_rtl_unroll_and_peel_loops): Likewise.
16221 (rtl_doloop): Likewise.
16222 * lto-streamer-in.c (input_cfg): Likewise.
16223 * lto-streamer-out.c (output_cfg): Likewise.
16224 * modulo-sched.c (sms_schedule): Likewise.
16225 * predict.c (tree_estimate_probability): Likewise.
16226 (tree_estimate_probability_driver): Likewise.
16227 (estimate_loops): Likewise.
16228 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
16229 (move_sese_region_to_fn): Likewise.
16230 (debug_loop_num): Likewise.
16231 * tree-chrec.c (chrec_evaluate): Likewise.
16232 (hide_evolution_in_other_loops_than_loop): Likewise.
16233 (chrec_component_in_loop_num): Likewise.
16234 (reset_evolution_in_loop): Likewise.
16235 (evolution_function_is_invariant_rec_p): Likewise.
16236 * tree-if-conv.c (main_tree_if_conversion): Likewise.
16237 * tree-inline.c (copy_loops): Likewise.
16238 (copy_cfg_body): Likewise.
16239 (tree_function_versioning): Likewise.
16240 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
16241 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
16243 (add_to_evolution_1): Likewise.
16244 (scev_const_prop): Likewise.
16245 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
16246 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
16247 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
16248 (tree_ssa_lim_initialize): Likewise.
16249 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
16250 (verify_loop_closed_ssa): Likewise.
16251 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
16252 (tree_ssa_loop_im): Likewise.
16253 (tree_ssa_loop_unswitch): Likewise.
16254 (tree_vectorize): Likewise.
16255 (check_data_deps): Likewise.
16256 (tree_ssa_loop_ivcanon): Likewise.
16257 (tree_ssa_loop_bounds): Likewise.
16258 (tree_complete_unroll): Likewise.
16259 (tree_complete_unroll_inner): Likewise.
16260 (tree_parallelize_loops): Likewise.
16261 (tree_ssa_loop_prefetch): Likewise.
16262 (tree_ssa_loop_ivopts): Likewise.
16263 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
16264 * tree-vectorizer.c (vectorize_loops): Likewise.
16266 2013-04-29 Mike Frysinger <vapier@gentoo.org>
16268 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
16269 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
16270 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
16271 with EABI_LINK_SPEC.
16273 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
16276 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
16279 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
16282 * lra-constraints.c (process_alt_operands): Discourage a bit more
16283 using memory for pseudos. Print cost dump for alternatives.
16284 Modify cost values for conflicts with early clobbers.
16285 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
16287 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
16290 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
16292 2013-04-29 Ian Bolton <ian.bolton@arm.com>
16294 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
16295 from/to S register.
16296 (movdi_aarch64): Support LDR/STR from/to D register.
16298 2013-04-29 Ian Bolton <ian.bolton@arm.com>
16300 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
16301 or higher by default.
16303 2013-04-29 Richard Biener <rguenther@suse.de>
16305 PR middle-end/57075
16306 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
16307 even if not adding abnormal edges for calls that can make
16310 2013-04-29 Richard Biener <rguenther@suse.de>
16312 PR middle-end/57103
16313 * tree-cfg.c (move_stmt_op): Fix condition under which to update
16315 (move_stmt_r): Remove redundant checking.
16317 2013-04-29 Teresa Johnson <tejohnson@google.com>
16320 * basic-block.h (apply_scale): New function.
16321 (apply_probability): Use apply_scale.
16322 * gimple-streamer-in.c (input_bb): Ditto.
16323 * lto-streamer-in.c (input_cfg): Ditto.
16324 * lto-cgraph.c (merge_profile_summaries): Ditto.
16325 * tree-optimize.c (execute_fixup_cfg): Ditto.
16326 * tree-inline.c (copy_bb): Update comment to use apply_scale.
16327 (copy_edges_for_bb): Ditto.
16328 (copy_cfg_body): Ditto.
16330 2013-04-29 Tom de Vries <tom@codesourcery.com>
16332 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
16333 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
16334 (tail_merge_optimize): Handle current_loops == NULL.
16336 2013-04-26 Jeff Law <law@redhat.com>
16338 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
16339 (simplify_cond_using_ranges): Generalize code to simplify
16340 COND_EXPRs where one argument is a constant and the other
16341 is an SSA_NAME created by an integral type conversion.
16343 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16345 * config/arm/arm.md (store_minmaxsi): Use only when
16346 optimize_insn_for_size_p.
16348 2013-04-29 Christian Bruel <christian.bruel@st.com>
16351 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
16353 2013-04-29 Richard Biener <rguenther@suse.de>
16355 PR middle-end/57089
16356 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
16357 loop tree make sure to schedule a fixup for the child as well.
16358 (expand_omp_for_generic): Properly add loops.
16359 (expand_omp_for_static_nochunk): Likewise.
16360 (expand_omp_for_static_chunk): Likewise.
16361 (expand_omp_for): For the degenerate case fixup loops.
16362 (expand_omp_sections): Fix default bb placement in loops.
16363 (expand_omp_atomic_pipeline): Properly add loops.
16365 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16367 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
16369 2013-04-29 Tom de Vries <tom@codesourcery.com>
16371 * tree-ssa-tail-merge.c: Update header comment.
16373 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
16375 * config/aarch64/arm_neon.h
16376 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
16377 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
16378 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
16379 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
16380 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
16381 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
16382 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
16383 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
16385 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
16387 * config/aarch64/aarch64-simd.md
16388 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
16389 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
16390 fix_trunc, fixuns_trunc.
16391 (ftrunc<VDQF:mode>2): New.
16392 * config/aarch64/iterators.md (optab): Add fix, fixuns.
16393 (fix_trunc_optab): New.
16395 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
16397 * config/aarch64/aarch64-builtins.c
16398 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
16399 iceilf, lround, iroundf.
16401 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
16404 * config/i386/i386.h (enum ix86_tune_indices)
16405 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
16406 New, split from X86_TUNE_INTER_UNIT_MOVES.
16407 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
16408 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
16409 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
16410 (TARGET_INTER_UNIT_MOVES): Remove.
16411 * config/i386/i386.c (initial_ix86_tune_features): Update.
16412 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
16413 (ix86_expand_convert_uns_didf_sse): Use
16414 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
16415 (ix86_expand_vector_init_one_nonzero): Ditto.
16416 (ix86_expand_vector_init_interleave): Ditto.
16417 (inline_secondary_memory_needed): Return true for moves from SSE class
16418 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
16419 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
16420 * config/i386/constraints.md (Yi, Ym): Depend on
16421 TARGET_INTER_UNIT_MOVES_TO_VEC.
16422 (Yj, Yn): New constraints.
16423 * config/i386/i386.md (*movdi_internal): Change constraints of
16424 operand 1 from Yi to Yj and from Ym to Yn.
16425 (*movsi_internal): Ditto.
16426 (*movdf_internal): Ditto.
16427 (*movsf_internal): Ditto.
16428 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
16429 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
16430 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
16431 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
16432 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
16433 * config/i386/sse.md (movdi_to_sse): Ditto.
16434 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
16435 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
16436 TARGET_INTER_UNIT_MOVES.
16437 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
16438 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
16439 instead of TARGET_INTER_UNIT_MOVES.
16440 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
16441 operand 1 from Yi to Yj and from Ym to Yn.
16443 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
16445 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
16446 (float_truncate_hi_): Likewise.
16447 (float_extend_lo_): Likewise.
16448 (float_truncate_lo_): Likewise.
16449 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
16450 (aarch64_float_extend_lo_v2df): Likewise.
16451 (vec_unpacks_hi_v4sf): Likewise.
16452 (aarch64_float_truncate_lo_v2sf): Likewise.
16453 (aarch64_float_truncate_hi_v4sf): Likewise.
16454 (vec_pack_trunc_v2df): Likewise.
16455 (vec_pack_trunc_df): Likewise.
16457 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
16459 * config/aarch64/aarch64-builtins.c
16460 (aarch64_fold_builtin): Fold float conversions.
16461 * config/aarch64/aarch64-simd-builtins.def
16462 (floatv2si, floatv4si, floatv2di): New.
16463 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
16464 * config/aarch64/aarch64-simd.md
16465 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
16466 * config/aarch64/iterators.md (FLOATUORS): New.
16467 (optab): Add float, floatuns.
16468 (su_optab): Likewise.
16470 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
16472 * config/aarch64/aarch64-builtins.c
16473 (aarch64_builtin_vectorized_function): Use new names for
16475 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
16476 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
16477 (fcvtzu): Split as...
16478 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
16479 (fcvtas): Split as...
16480 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
16481 (fcvtau): Split as...
16482 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
16483 (fcvtps): Split as...
16484 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
16485 (fcvtpu): Split as...
16486 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
16487 (fcvtms): Split as...
16488 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
16489 (fcvtmu): Split as...
16490 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
16491 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
16492 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
16493 (lfrintnusf, lfrintnudf): Likewise.
16494 * config/aarch64/aarch64-simd.md
16495 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
16497 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
16498 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
16499 (fcvt_pattern): Likewise.
16501 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
16503 * config/aarch64/aarch64-simd.md
16504 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
16505 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
16507 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
16509 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
16510 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
16511 (vrnd<a,m,n,p>_f32): Implement using builtins.
16512 (vrnd<i,x><q>_f<32, 64>): New.
16514 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
16516 * config/aarch64/aarch64-builtins.c
16517 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
16518 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
16519 (frintz): Rename to...
16521 (frintp): Rename to...
16523 (frintm): Rename to...
16525 (frinti): Rename to...
16526 (nearbyint): ...this.
16527 (frintx): Rename to...
16529 (frinta): Rename to...
16531 * config/aarch64/aarch64-simd.md
16532 (aarch64_frint<frint_suffix><mode>): Delete.
16533 (<frint_pattern><mode>2): Convert to insn.
16534 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
16535 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
16536 (frint_pattern): Likewise.
16537 (frint_suffix): Likewise.
16539 2013-04-29 Richard Biener <rguenther@suse.de>
16541 PR tree-optimization/57081
16542 * loop-init.c: Include tree-flow.h.
16543 (loop_optimizer_finalize): Free number of iteration estimates.
16544 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
16546 2013-04-29 Jakub Jelinek <jakub@redhat.com>
16548 PR tree-optimization/57083
16549 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
16550 non-singleton shift count range, zero extend low_bound for uns case.
16552 * config/i386/predicates.md (general_vector_operand): New predicate.
16553 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
16554 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
16555 if they aren't nonimmediate operands. If their original values
16556 satisfy const_vector_equal_evenodd_p, don't shift them.
16557 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
16558 predicates. For the SSE4.1 case force operands[{1,2}] into registers
16559 if not nonimmediate_operand.
16560 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
16561 instead of register_operand.
16562 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
16564 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
16566 * stor-layout.c (finalize_size_functions): Allocate a structure and
16567 reset cfun before dumping the functions.
16569 2013-04-27 Jakub Jelinek <jakub@redhat.com>
16571 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
16574 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
16575 use xop_pmacsdqh if uns_p.
16576 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
16577 the immediate rotate count.
16579 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
16581 * rtl.h (struct rtx_def): Add comment for field jump.
16582 (LRA_SUBREG_P): New macro.
16583 * recog.c (register_operand): Check LRA_SUBREG_P.
16584 * lra.c (lra): Add note at the end of RTL code. Align non-empty
16586 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
16587 (lra_final_code_change): Skip subreg change for operators.
16588 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
16589 if there are no operand changes.
16590 * lra-constraints.c (curr_insn_set): New.
16591 (match_reload): Set LRA_SUBREG_P.
16592 (emit_spill_move): Ditto.
16593 (check_and_process_move): Use curr_insn_set. Process only single
16594 set insns. Don't initialize sec_mem_p and change_p.
16595 (simplify_operand_subreg): Use LRA_SUBREG_P.
16596 (reg_in_class_p): New function.
16597 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
16598 of #ifdef. Add code to remove cycling.
16599 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
16600 non-null disp. Reload inner instead of disp when base and index
16601 are null. Try to put lo_sum into register.
16602 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
16603 (check_and_process_move): Move code for move cost check to
16604 simple_move_p. Remove equiv_substitution.
16605 (simple_move_p): New function.
16606 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
16607 curr_insn_set. Call check_and_process_move only for single set
16608 insns. Use the new function. Move call of check_and_process_move
16609 after operand equiv substitution and address process.
16611 2013-04-26 Jakub Jelinek <jakub@redhat.com>
16614 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
16615 with nonlocal goto receivers or returns twice calls, ignore
16616 unininitialized values from abnormal edges to nl goto receiver
16617 or returns twice call.
16619 2013-04-26 Jakub Jelinek <jakub@redhat.com>
16621 PR tree-optimization/57051
16622 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
16623 and VEC_RSHIFT_EXPR if shift count is a multiple of element
16626 2013-04-26 Richard Biener <rguenther@suse.de>
16628 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
16629 (expand_omp_taskreg): Likewise. Mark loops for fixup.
16630 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
16631 (fixup_loop_arrays_after_move): New function.
16632 (move_sese_region_to_fn): Properly outline the loop tree parts
16633 of the SESE region.
16635 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
16637 * config/i386/i386.md (type, unit): Fix long lines.
16639 2013-04-26 Richard Biener <rguenther@suse.de>
16641 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
16642 (lto-streamer-out.o): Likewise.
16643 * cfgloop.c (init_loops_structure): Export, add struct function
16644 argument and adjust.
16645 (flow_loops_find): Adjust.
16646 * cfgloop.h (enum loop_estimation): Add EST_LAST.
16647 (init_loops_structure): Declare.
16648 * lto-streamer-in.c: Include cfgloop.h.
16649 (input_cfg): Input the loop tree.
16650 * lto-streamer-out.c: Include cfgloop.h.
16651 (output_cfg): Output the loop tree.
16652 (output_struct_function_base): Do not drop PROP_loops.
16654 2013-03-26 Richard Biener <rguenther@suse.de>
16656 * tree-cfg.c (execute_build_cfg): Build the loop tree.
16657 (pass_build_cfg): Provide PROP_loops.
16658 (move_sese_region_to_fn): Remove loops that are outlined into fn
16660 * tree-inline.c: Include cfgloop.h.
16661 (initialize_cfun): Do not drop PROP_loops.
16662 (copy_loops): New function.
16663 (copy_cfg_body): Copy loop structure.
16664 (tree_function_versioning): Initialize destination loop tree.
16665 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
16666 (pass_parallelize_loops): Do IL verification.
16667 * loop-init.c (loop_optimizer_init): Fixup loops if required.
16668 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
16669 the CFG make sure we fixup loops as well.
16670 * tree-ssa-tail-merge.c: Include cfgloop.h.
16671 (replace_block_by): When merging loop latches mark loops for fixup.
16672 * lto-streamer-out.c (output_struct_function_base): Drop
16673 PROP_loops for now.
16674 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
16675 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
16676 * ipa-split.c: Include cfgloop.h.
16677 (split_function): Add the new return block to the loop tree root.
16678 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
16679 whether we have removed the forwarder block.
16680 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
16681 * cfgloop.h (place_new_loop): Declare.
16682 * cfgloopmanip.c (place_new_loop): Export.
16683 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
16684 (tree-switch-conversion.o): Likewise.
16685 (tree-complex.o): Likewise.
16686 (tree-inline.o): Likewise.
16687 (tree-ssa-tailmerge.o): Likewise.
16688 (ipa-split.o): Likewise.
16689 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
16690 (tree-ssa-copy.o): Likewise.
16691 * tree-switch-conversion.c: Include cfgloop.h
16692 (process_switch): If we emit a bit-test cascade, schedule loops
16694 * tree-complex.c: Include cfgloop.h.
16695 (expand_complex_div_wide): Properly add new basic-blocks to loops.
16696 * asan.c: Include cfgloop.h.
16697 (create_cond_insert_point): Properly add new basic-blocks to
16698 loops, schedule loop fixup.
16699 * cfgloop.c (verify_loop_structure): Check that looks are not
16701 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
16703 (expand_omp_for_generic): Likewise.
16704 (expand_omp_sections): Likewise.
16705 (expand_omp_atomic_pipeline): Schedule loops for fixup.
16706 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
16707 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
16708 is initialized, not when loops are present.
16709 * tree-parloops.c (parallelize_loops): Remove checking here.
16710 * passes.c (init_optimization_passes): Schedule a copy-propagation
16711 pass before complete unrolling of inner loops.
16713 2013-04-26 Jakub Jelinek <jakub@redhat.com>
16715 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
16716 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
16717 (colorize_init): Add argument to _WIN32 version.
16718 * toplev.c: Include diagnostic-color.h.
16719 (process_options): Default to -fdiagnostics-color=auto if
16720 GCC_COLORS env var is in the environment.
16721 * common.opt (fdiagnostics-color=): Add Var and Init.
16722 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
16723 env var is in the environment, the default is auto rather than never.
16725 * diagnostic.h (file_name_as_prefix): Add context argument.
16726 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
16727 the string as locus.
16728 * langhooks.c (lhd_print_error_function): Adjust caller.
16730 2013-04-25 Lawrence Crowl <crowl@google.com>
16732 * var-tracking.c (shared_hash_def::htab):
16733 Change type to hash_table. Update dependent calls and types.
16735 2013-04-25 Lawrence Crowl <crowl@google.com>
16737 * Makefile.in: Update as needed below.
16739 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
16740 Move declaration to after the type's method definitons.
16742 * attribs.c (htab_t scoped_attributes::attribute_hash):
16743 Change type to hash_table. Update dependent calls and types.
16745 * bitmap.c (htab_t bitmap_desc_hash):
16746 Change type to hash_table. Update dependent calls and types.
16748 * cselib.c (htab_t cselib_hash_table):
16749 Change type to hash_table. Update dependent calls and types.
16751 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
16752 (hash_string_slot_node): Move implementation into lto-streamer.h
16753 struct string_slot_hasher.
16754 (eq_string_slot_node): Likewise.
16756 * data-streamer-out.c: Update output_block::string_hash_table
16757 dependent calls and types.
16759 * dwarf2cfi.c (htab_t trace_index):
16760 Change type to hash_table. Update dependent calls and types.
16762 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
16763 Change type to hash_table. Update dependent calls and types.
16764 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
16765 (htab_t optimize_external_refs::map): Likewise.
16766 (htab_t output_comp_unit::extern_map): Likewise.
16767 (htab_t output_comdat_type_unit::extern_map): Likewise.
16768 (htab_t output_macinfo::macinfo_htab): Likewise.
16769 (htab_t optimize_location_lists::htab): Likewise.
16770 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
16772 * except.c (htab_t ehspec_hash_type):
16773 Change type to hash_table. Update dependent calls and types.
16774 (assign_filter_values::ttypes): Likewise.
16775 (assign_filter_values::ehspec): Likewise.
16776 (sjlj_assign_call_site_values::ar_hash): Likewise.
16777 (convert_to_eh_region_ranges::ar_hash): Likewise.
16779 * gcse.c (htab_t pre_ldst_table):
16780 Change type to hash_table. Update dependent calls and types.
16782 * ggc-common.c (htab_t saving_htab):
16783 Change type to hash_table. Update dependent calls and types.
16784 (htab_t loc_hash): Likewise.
16785 (htab_t ptr_hash): Likewise.
16786 (call_count): Rename ggc_call_count.
16787 (call_alloc): Rename ggc_call_alloc.
16788 (loc_descriptor): Rename make_loc_descriptor.
16789 (add_statistics): Rename ggc_add_statistics.
16791 * ggc-common.c (saving_htab):
16792 Change type to hash_table. Update dependent calls and types.
16794 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
16795 (push_gimplify_context): Likewise.
16796 (pop_gimplify_context): Likewise.
16797 (struct gimple_temp_hash_elt): Added.
16798 (struct gimplify_hasher): Likewise.
16799 (struct gimplify_ctx.temp_htab):
16800 Change type to hash_table. Update dependent calls and types.
16802 * gimple-fold.c: Include gimplify-ctx.h.
16804 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
16805 Change type to hash_table. Update dependent calls and types.
16806 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
16807 avoid potential global name collision.
16809 * gimplify.c: Include gimplify-ctx.h.
16810 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
16811 (htab_t gimplify_ctx::temp_htab):
16812 Update dependent calls and types for new type hash_table.
16813 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
16814 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
16816 * gimplify-ctx.h: New.
16817 (struct gimple_temp_hash_elt): Move from gimplify.c.
16818 (class gimplify_hasher): New.
16819 (struct gimplify_ctx): Move from gimple.h.
16820 (htab_t gimplify_ctx::temp_htab):
16821 Change type to hash_table. Update dependent calls and types.
16823 * graphite-clast-to-gimple.c: Include graphite-htab.h.
16824 (htab_t ivs_params::newivs_index):
16825 Change type to hash_table. Update dependent calls and types.
16826 (htab_t ivs_params::params_index): Likewise.
16827 (htab_t print_generated_program::params_index): Likewise.
16828 (htab_t gloog::newivs_index): Likewise.
16829 (htab_t gloog::params_index): Likewise.
16831 * graphite.c: Include graphite-htab.h.
16832 4htab_t graphite_transform_loops::bb_pbb_mapping):
16833 Change type to hash_table. Update dependent calls and types.
16835 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
16836 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
16837 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
16839 * graphite-dependences.c: Include graphite-htab.h.
16840 (loop_is_parallel_p): Change hash table type of parameter.
16842 * graphite-htab.h: New.
16843 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
16844 (extern find_pbb_via_hash): Move from graphite-poly.h.
16845 (extern loop_is_parallel_p): Move from graphite-poly.h.
16846 (extern get_loop_body_pbbs): Move from graphite-poly.h.
16848 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
16849 (extern loop_is_parallel_p): Move to graphite-htab.h.
16850 (extern get_loop_body_pbbs): Move to graphite-htab.h.
16852 * haifa-sched.c (htab_t delay_htab):
16853 Change type to hash_table. Update dependent calls and types.
16854 (htab_t delay_htab_i2): Likewise.
16856 * ira-color.c (htab_t allocno_hard_regs_htab):
16857 Change type to hash_table. Update dependent calls and types.
16859 * ira-costs.c (htab_t cost_classes_htab):
16860 Change type to hash_table. Update dependent calls and types.
16862 * loop-invariant.c (htab_t merge_identical_invariants::eq):
16863 Change type to hash_table. Update dependent calls and types.
16865 * loop-iv.c (htab_t bivs):
16866 Change type to hash_table. Update dependent calls and types.
16868 * loop-unroll.c (htab_t opt_info::insns_to_split):
16869 Change type to hash_table. Update dependent calls and types.
16870 (htab_t opt_info::insns_with_var_to_expand): Likewise.
16872 * lto-streamer.h (struct string_slot): Move from data-streamer.h
16873 (struct string_slot_hasher): New.
16874 (htab_t output_block::string_hash_table):
16875 Change type to hash_table. Update dependent calls and types.
16877 * lto-streamer-in.c (freeing_string_slot_hasher): New.
16878 (htab_t file_name_hash_table):
16879 Change type to hash_table. Update dependent calls and types.
16881 * lto-streamer-out.c: Update output_block::string_hash_table dependent
16884 * lto-streamer.c (htab_t tree_htab):
16885 Change type to hash_table. Update dependent calls and types.
16887 * omp-low.c: Include gimplify-ctx.h.
16889 * passes.c (htab_t name_to_pass_map):
16890 Change type to hash_table. Update dependent calls and types.
16891 (pass_traverse): Rename to passes_pass_traverse.
16893 * plugin.c (htab_t event_tab):
16894 Change type to hash_table. Update dependent calls and types.
16896 * postreload-gcse.c (htab_t expr_table):
16897 Change type to hash_table. Update dependent calls and types.
16898 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
16900 * sese.c (debug_rename_map_1): Make extern.
16901 (htab_t copy_bb_and_scalar_dependences::rename_map):
16902 Change type to hash_table. Update dependent calls and types.
16904 * sese.h (extern debug_rename_map): Move to .c file.
16906 * store-motion.c (htab_t store_motion_mems_table):
16907 Change type to hash_table. Update dependent calls and types.
16909 * trans-mem.c (htab_t tm_new_mem_hash):
16910 Change type to hash_table. Update dependent calls and types.
16912 * tree-browser.c (htab_t TB_up_ht):
16913 Change type to hash_table. Update dependent calls and types.
16915 * tree-cfg.c (htab_t discriminator_per_locus):
16916 Change type to hash_table. Update dependent calls and types.
16918 * tree-complex.c: Include tree-hasher.h
16919 (htab_t complex_variable_components):
16920 Change type to hash_table. Update dependent calls and types.
16922 * tree-eh.c (htab_t finally_tree):
16923 Change type to hash_table. Update dependent calls and types.
16925 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
16926 struct int_tree_hasher.
16927 (extern int_tree_map_eq): Likewise.
16928 (uid_decl_map_hash): Removed.
16929 (extern decl_tree_map_eq): Likewise.
16931 * tree-hasher.h: New.
16932 (struct int_tree_hasher): New.
16933 (typedef int_tree_htab_type): New.
16935 * tree-inline.c: Include gimplify-ctx.h.
16937 * tree-mudflap.c: Include gimplify-ctx.h.
16939 * tree-parloops.c: Include tree-hasher.h.
16940 (htab_t eliminate_local_variables_stmt::decl_address):
16941 Change type to hash_table. Update dependent calls and types.
16942 (htab_t separate_decls_in_region::decl_copies): Likewise.
16944 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
16945 Change type to hash_table. Update dependent calls and types.
16947 * tree-sra.c (candidates):
16948 Change type to hash_table. Update dependent calls and types.
16950 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
16952 (int_tree_map_hash): Likewise.
16954 * tree-ssa-dom.c (htab_t avail_exprs):
16955 Change type to hash_table. Update dependent calls and types.
16957 * tree-ssa-live.c (var_map_base_init::tree_to_index):
16958 Change type to hash_table. Update dependent calls and types.
16960 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
16961 Change type to hash_table. Update dependent calls and types.
16963 * tree-ssa-phiopt.c (seen_ssa_names):
16964 Change type to hash_table. Update dependent calls and types.
16966 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
16967 Change type to hash_table. Update dependent calls and types.
16969 * tree-ssa-uncprop.c (equiv):
16970 Change type to hash_table. Update dependent calls and types.
16972 2013-04-25 Jakub Jelinek <jakub@redhat.com>
16974 PR rtl-optimization/57003
16975 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
16976 call note_stores with kill_clobbered_value callback again after
16977 killing regs_invalidated_by_call.
16979 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
16981 * config/aarch64/aarch64-simd.md
16982 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
16983 (aarch64_simd_bsl<mode>): Likewise.
16984 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
16986 2013-04-25 Marek Polacek <polacek@redhat.com>
16988 PR tree-optimization/57066
16989 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
16991 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
16993 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
16995 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
16997 * config/aarch64/aarch64-builtins.c
16998 (aarch64_fold_builtin): New.
16999 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
17000 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
17001 * config/aarch64/aarch64-simd-builtins.def (abs): New.
17002 * config/aarch64/arm_neon.h
17003 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
17005 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
17006 Tejas Belagod <tejas.belagod@arm.com>
17008 * config/aarch64/aarch64-builtins.c
17009 (aarch64_gimple_fold_builtin): New.
17010 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
17011 * config/aarch64/aarch64-simd-builtins.def (addv): New.
17012 * config/aarch64/aarch64-simd.md (addpv4sf): New.
17013 (addvv4sf): Update.
17014 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
17016 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17018 * config/aarch64/aarch64.md
17019 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
17021 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17023 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
17024 (*ngcsi_uxtw): New pattern.
17026 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17027 Julian Brown <julian@codesourcery.com>
17029 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
17030 (TB_DREG): Add T_V4HF.
17031 (v4hf_UP): New macro.
17032 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
17033 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
17034 Handle initialisation of V4HF. Adjust initialisation of reinterpret
17036 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
17037 (arm_vector_mode_supported_p): Handle V4HF.
17038 (arm_mangle_map): Handle V4HFmode.
17039 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
17040 * config/arm/arm_neon_builtins.def: Add entries for
17041 vcvtv4hfv4sf, vcvtv4sfv4hf.
17042 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
17043 (neon_vcvtv4hfv4sf): Likewise.
17044 * config/arm/neon-gen.ml: Handle half-precision floating point
17046 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
17047 * config/arm/arm_neon.h: Regenerate.
17048 * config/arm/neon.ml (type elts): Add F16.
17049 (type vectype): Add T_float16x4, T_floatHF.
17050 (type vecmode): Add V4HF.
17051 (type features): Add Requires_FP_bit feature.
17052 (elt_width): Handle F16.
17053 (elt_class): Likewise.
17054 (elt_of_class_width): Likewise.
17055 (mode_of_elt): Refactor.
17056 (type_for_elt): Handle F16, fix error messages.
17057 (vectype_size): Handle T_float16x4.
17058 (vcvt_sh): New function.
17059 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
17060 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
17061 (string_of_mode): Handle V4HF.
17062 * doc/arm-neon-intrinsics.texi: Regenerate.
17064 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
17066 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
17067 format specifier in 'X' case.
17069 2013-04-25 Alan Modra <amodra@gmail.com>
17072 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
17073 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
17074 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
17075 Repeat for many other rotate/shift and mask patterns using subregs.
17076 Name lshiftrt insns.
17077 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
17078 on WORDS_BIG_ENDIAN.
17080 2013-04-25 Alan Modra <amodra@gmail.com>
17082 * config.gcc: Support little-endian powerpc-linux targets.
17083 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
17084 (LINK_OS_LINUX_SPEC): Define.
17085 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
17086 Preserve MASK_LITTLE_ENDIAN.
17087 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
17088 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
17089 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
17090 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
17091 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
17092 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
17093 Correct fp word order for little-endian. Don't shift toc entries
17094 smaller than a word for little-endian.
17095 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
17096 (bswapdi2 splits): Correct low-part subreg for little-endian.
17097 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
17098 low/high where such is correct only for be.
17099 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
17100 little-endian for -mcall-aixdesc.
17102 2013-04-25 Alan Modra <amodra@gmail.com>
17104 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
17105 replace_equiv_address_nv.
17107 2013-04-25 Alan Modra <amodra@gmail.com>
17109 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
17111 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
17114 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
17115 * rtl.h (struct rtx_def): ...
17117 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
17119 PR rtl-optimizations/57046
17120 * lra-constraints (split_reg): Set up lra_risky_transformations_p
17121 for multi-reg splits.
17123 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
17125 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
17127 2013-04-24 Sterling Augustine <saugustine@google.com>
17129 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
17130 (comp_dir_string, debug_str_dwo_section): New.
17131 (DEBUG_STR_DWO_SECTION): Rename to ...
17132 (DEBUG_DWO_STR_SECTION): ... this.
17133 (DEBUG_NORM_STR_SECTION): Delete.
17134 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
17135 (DEBUG_STR_DWO_SECTION_FLAGS): New.
17136 (find_AT_string): Move most logic to ...
17137 (find_AT_string_in_table): ... here. New.
17138 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
17139 add_skeleton_AT_string. Delete logic.
17140 (output_skeleton_debug_sections): Remove call to
17141 add_top_level_skeleton_die_attrs.
17142 (add_comp_dir_attribute): Move logic to comp_dir_string.
17143 (dwarf2out_init): Initialize debug_str_dwo_section.
17144 (output_indirect_string): Call find_string_form.
17145 (output_indirect_strings): Rewrite.
17146 (prune_unused_types): Empty skeleton_debug_str_hash.
17147 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
17148 (dwarf2out_finish): Call output_indirect_strings.
17150 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
17152 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
17154 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
17156 * rtl.h (struct rtx_def): Add comment for field jump.
17157 (LRA_SUBREG_P): New macro.
17158 * recog.c (register_operand): Check LRA_SUBREG_P.
17159 * lra.c (lra): Add note at the end of RTL code. Align non-empty
17161 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
17162 (lra_final_code_change): Skip subreg change for operators.
17163 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
17164 if there are no operand changes.
17165 * lra-constraints.c (curr_insn_set): New.
17166 (match_reload): Set LRA_SUBREG_P.
17167 (emit_spill_move): Ditto.
17168 (check_and_process_move): Use curr_insn_set. Process only single
17169 set insns. Don't initialize sec_mem_p and change_p.
17170 (simplify_operand_subreg): Use LRA_SUBREG_P.
17171 (reg_in_class_p): New function.
17172 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
17173 of #ifdef. Add code to remove cycling.
17174 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
17175 non-null disp. Reload inner instead of disp when base and index
17176 are null. Try to put lo_sum into register.
17177 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
17178 (check_and_process_move): Move code for move cost check to
17179 simple_move_p. Remove equiv_substitution.
17180 (simple_move_p): New function.
17181 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
17182 curr_insn_set. Call check_and_process_move only for single set
17183 insns. Use the new function. Move call of check_and_process_move
17184 after operand equiv substitution and address process.
17186 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
17188 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
17189 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
17190 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
17192 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
17194 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
17196 2013-04-24 Marek Polacek <polacek@redhat.com>
17198 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
17199 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
17200 (select_loops_exit_conditions): Likewise.
17201 (number_of_iterations_for_all_loops): Likewise.
17202 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
17203 (scev_analysis): Likewise.
17205 2013-04-02 Catherine Moore <clm@codesourcery.com>
17206 Chao-ying Fu <fu@mips.com>
17208 * config/mips/micromips.md (jraddiusp): New pattern.
17209 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
17210 instruction if possible.
17212 2013-04-24 Alan Modra <amodra@gmail.com>
17214 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
17216 2013-04-24 Julian Brown <julian@codesourcery.com>
17217 Chung-Lin Tang <cltang@codesourcery.com>
17219 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
17220 dependency behavior in enumeration type DIE generation. Add TODO note
17221 to comments about future DW_FORM_sdata/udata re-work of related code.
17223 2013-04-23 Lawrence Crowl <crowl@google.com>
17225 * Makefile.in: Update as needed below.
17227 * hash-table.h (class hash_table):
17228 Correct many methods with parameter types compare_type to the correct
17229 value_type. (Correct code was unlikely to notice the change.)
17230 (hash_table::elements_with_deleted) New.
17231 (class hashtable::iterator): New.
17232 (hashtable::begin()): New.
17233 (hashtable::end()): New.
17234 (FOR_EACH_HASH_TABLE_ELEMENT): New.
17236 * statistics.c (statistics_hashes):
17237 Change type to hash_table. Update dependent calls and types.
17239 * tree-into-ssa.c (var_infos):
17240 Change type to hash_table. Update dependent calls and types.
17242 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
17243 Change type to hash_table. Update dependent calls and types.
17245 * tree-ssa-loop-im.c (struct mem_ref.refs):
17246 Change type to hash_table. Update dependent calls and types.
17248 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
17249 Change type to hash_table. Update dependent calls and types.
17251 * tree-ssa-sccvn.c (vn_tables_s::nary):
17252 Change type to hash_table. Update dependent calls and types.
17253 (vn_tables_s::phis): Likewise.
17254 (vn_tables_s::references): Likewise.
17256 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
17257 (vn_reference_eq): Update parameter and return types.
17259 * tree-ssa-structalias.c (pointer_equiv_class_table):
17260 Change type to hash_table. Update dependent calls and types.
17261 (location_equiv_class_table): Likewise.
17263 * tree-vect-data-refs.c: Consequential changes for making
17264 peeling a hash_table.
17266 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
17267 (destroy_loop_vec_info): Dependent hash_table update.
17269 * tree-vectorizer.h (peeling_htab):
17270 Change type to hash_table. Update dependent calls and types.
17272 2013-04-23 Shiva Chen <shiva0217@gmail.com>
17274 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
17275 to check the register content is equal or not.
17276 * lra-constraints.c (match_reload): Use lra_assign_reg_val
17277 to assign register content record.
17278 * lra-eliminations.c (update_reg_eliminate): Use
17279 lra_update_reg_val_offset to update register content offset.
17280 * lra-int.h (struct lra_reg): Add offset member.
17281 (lra_reg_val_equal_p): New static inline function.
17282 (lra_update_reg_val_offset): New static inline function.
17283 (lra_assign_reg_val): New static inline function.
17284 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
17285 to assign register content record.
17286 (initialize_lra_reg_info_element): Initial offset to zero.
17288 2013-04-23 Catherine Moore <clm@codesourcery.com>
17290 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
17291 operands. Record compression.
17293 2013-04-23 Xinliang David Li <davidxl@google.com>
17295 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
17297 2013-04-23 Richard Biener <rguenther@suse.de>
17299 PR middle-end/57036
17300 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
17301 parameter, only add abnormal goto edges from the copied body
17302 if the call could perform abnormal gotos.
17303 (copy_cfg_body): Adjust.
17305 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
17307 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
17309 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
17311 * coretypes.h (gimple_stmt_iterator): Add struct to make
17314 2013-04-23 Richard Biener <rguenther@suse.de>
17316 PR tree-optimization/57026
17317 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
17318 from SSA names occuring in abnormal PHI nodes.
17320 2013-04-22 Andi Kleen <ak@linux.intel.com>
17322 * lto/lto.c (print_lto_report_1): Fix LTO report names.
17324 2013-04-22 Andi Kleen <ak@linux.intel.com>
17326 * lto/lto.c (print_lto_report_1): Declare early.
17327 (read_cgraph_and_symbols): Call print_lto_report_1 early.
17329 2013-04-22 Andi Kleen <ak@linux.intel.com>
17331 * common.opt (-flto-report-wpa): Add.
17332 * doc/invoke.texi (-flto-report-wpa): Add.
17333 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
17336 2013-04-22 Xinliang David Li <davidxl@google.com>
17338 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
17339 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
17340 * Makefile.in: New dependency
17342 David Daney <ddaney.cavm@gmail.com>
17344 * configure.ac (gcc_cv_as_micromips_support): Use the
17345 --fatal-warnings option.
17346 * configure: Regenerate.
17348 2013-04-22 Marek Polacek <polacek@redhat.com>
17351 * tsan.c (instrument_expr): Don't instrument expression
17352 in case its size is zero.
17354 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
17358 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
17360 * config/alpha/alpha.c (TARGET_LRA_P): New define.
17362 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
17364 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
17365 (gimple_stmt_iterator): New typedef.
17366 * gimple.h (gimple_stmt_iterator): Rename to...
17367 (gimple_stmt_iterator_d): ... This.
17368 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
17369 trees be valid for GIMPLE and GENERIC.
17370 (TARGET_GIMPLE_FOLD_BUILTIN): New.
17371 * gimple-fold.c (gimple_fold_call): Call target hook
17372 gimple_fold_builtin.
17373 * hooks.c (hook_bool_gsiptr_false): New.
17374 * hooks.h (hook_bool_gsiptr_false): New.
17375 * target.def (fold_stmt): New.
17376 * doc/tm.texi: Regenerate.
17378 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
17381 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
17382 a set sp if no stack realignment.
17384 2013-04-22 Nick Clifton <nickc@redhat.com>
17386 * config.gcc (tilegx-linux): Extend extra_objs rather than
17388 (tilepro-linux): Likewise.
17390 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
17392 * config/aarch64/aarch64-builtins.c
17394 (CF0, CF1, CF2, CF3, CF4, CF10): New.
17395 (VAR<1-12>): Add MAP parameter.
17396 (BUILTIN_*): Likewise.
17397 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
17398 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
17399 (aarch64_ushl_n<mode>): Likewise.
17400 (aarch64_sshr_n<mode>): Likewise.
17401 (aarch64_ushr_n<mode>): Likewise.
17402 (aarch64_<maxmin><mode>): Likewise.
17403 (aarch64_sqrt<mode>): Likewise.
17404 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
17405 (vshr<q>_n_*): Likewise.
17407 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
17409 * config/aarch64/aarch64-builtins.c
17410 (aarch64_simd_builtin_type_mode): Handle SF types.
17412 (BUILTIN_GPF): Define.
17413 (aarch64_init_simd_builtins): Handle SF types.
17414 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
17415 (frecps): Likewise.
17416 (frecpx): Likewise.
17417 * config/aarch64/aarch64-simd.md
17418 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
17419 (aarch64_frecpe<mode>): New.
17420 (aarch64_frecps<mode>): Likewise.
17421 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
17422 (v8type): Add frecp<esx>.
17423 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
17424 (aarch64_frecps<mode>): Likewise.
17425 * config/aarch64/iterators.md (FRECP): New.
17426 (frecp_suffix): Likewise.
17427 * config/aarch64/arm_neon.h
17428 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
17430 2013-04-22 Christian Bruel <christian.bruel@st.com>
17433 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
17434 (REG_CLASS_NAMES): Idem.
17435 (REG_CLASS_CONTENTS): Idem.
17436 (REGCLASS_HAS_FP_REG): Idem.
17437 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
17438 (sh_conditional_register_usage): Idem.
17440 2013-04-21 Jeff Law <law@redhat.com>
17442 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
17443 (ssa_forward_propagate_and_combine): Use it.
17445 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
17447 * lra.c: Update the flow chart diagram.
17449 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
17451 PR rtl-optimization/56847
17452 * lra-constraints.c (process_alt_operands): Discourage alternative
17453 with non-matche doffsettable memory constraint fro memory with
17456 2013-04-19 Richard Biener <rguenther@suse.de>
17458 PR tree-optimization/56982
17459 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
17461 * gimplify.c (gimplify_call_expr): Notice special calls.
17462 (gimplify_modify_expr): Likewise.
17463 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
17464 abnormal control flow receivers.
17465 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
17466 in the same way as cfun->has_nonlocal_labels.
17467 (gimple_purge_dead_abnormal_call_edges): Likewise.
17468 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
17469 receivers start a basic-block.
17471 2013-04-19 Richard Biener <rguenther@suse.de>
17473 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
17475 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
17476 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
17477 (SLP_TREE_LOAD_PERMUTATION): Add.
17478 (vect_transform_slp_perm_load): Adjust prototype.
17479 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
17480 (vect_free_slp_instance): Likewise.
17481 (vect_create_new_slp_node): Likewise.
17482 (vect_supported_slp_permutation_p): Remove.
17483 (vect_slp_rearrange_stmts): Adjust.
17484 (vect_supported_load_permutation_p): Likewise. Inline
17485 vect_supported_slp_permutation_p here.
17486 (vect_analyze_slp_instance): Compute load permutations per
17487 slp node instead of per instance.
17488 (vect_get_slp_defs): Adjust.
17489 (vect_transform_slp_perm_load): Likewise.
17490 (vect_schedule_slp_instance): Remove redundant code.
17491 (vect_schedule_slp): Remove hack for PR56270, add it ...
17492 * tree-vect-stmts.c (vectorizable_load): ... here, do not
17493 CSE loads for SLP. Adjust.
17495 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
17497 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
17498 spelling in two comments.
17500 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
17503 * config/arm/arm.c (load_multiple_sequence): Require SP
17504 as base register for loads if SP is in the register list.
17506 2013-04-19 Martin Jambor <mjambor@suse.cz>
17508 PR tree-optimization/56718
17509 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
17510 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
17511 and made public. Adjusted all callers.
17512 (ipa_intraprocedural_devirtualization): New function.
17513 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
17514 (ipa_intraprocedural_devirtualization): Likewise.
17515 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
17517 2013-04-19 Richard Biener <rguenther@suse.de>
17519 PR tree-optimization/57000
17520 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
17522 2013-04-19 Terry Guo <terry.guo@arm.com>
17524 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
17526 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
17527 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
17528 (cortex_m4_fmacs): Use new reservations.
17529 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
17531 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
17533 PR rtl-optimization/56999
17534 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
17536 (lra_coalesce): Remove split_origin_bitmap and related code.
17537 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
17538 ranges if necessary.
17540 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
17542 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
17544 (ix86_expand_call): Remove clobbered_registers array and use
17545 x86_64_ms_sysv_extra_clobbered_registers instead.
17546 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
17548 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
17550 * config/i386/i386.md (*call_rex64_ms_sysv): Use
17551 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
17552 (*call_value_rex64_ms_sysv): Ditto.
17554 2013-04-18 Cary Coutant <ccoutant@google.com>
17556 * dwarf2out.c (output_pubnames): Check die_perennial_p of
17557 parent instead of die_mark.
17559 2013-04-18 Diego Novillo <dnovillo@google.com>
17561 * gimple.c (create_gimple_tmp): New.
17562 (get_expr_type): New.
17563 (build_assign): New.
17564 (build_type_cast): New.
17565 * gimple.h (enum ssa_mode): Define.
17566 (gimple_seq_set_location): New.
17567 * asan.c (build_check_stmt): Change some gimple_build_* calls
17568 to use build_assign and build_type_cast.
17570 2013-04-18 Richard Biener <rguenther@suse.de>
17572 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
17573 handle negative step. Remove redundant checks.
17574 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
17575 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
17576 for negative step and grouped loads fail to vectorize.
17578 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
17580 * emit-rtl.c (reset_insn_used_flags): New function.
17581 (reset_all_used_flags): Use it.
17582 (verify_insn_sharing): New function.
17583 (verify_rtl_sharing): Fix verification for SEQUENCEs.
17585 2013-04-18 Jakub Jelinek <jakub@redhat.com>
17587 PR tree-optimization/56984
17588 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
17589 and (x >> M) >= N don't register any assertion if N << M is the
17592 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
17594 * lower-subreg.c (resolve_simple_move): If called self-recursive,
17595 do not delete_insn insns that have not yet been emitted, only
17596 unlink them with remove_insn.
17597 * df-scan.c (df_insn_delete): Revert r197492.
17599 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
17601 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
17602 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
17604 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
17606 * config/arm/arm.md (movsicc_insn): Convert define_insn into
17607 define_insn_and_split.
17608 (and_scc,ior_scc,negscc): Likewise.
17609 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
17611 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
17613 * config/arm/arm.c (use_return_insn): Return 0 for targets that
17614 can benefit from using a sequence of LDRD instructions in epilogue
17615 instead of a single LDM instruction.
17617 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
17620 * doc/extend.texi: Fix typo.
17622 2013-04-17 Richard Biener <rguenther@suse.de>
17624 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
17625 (vect_build_slp_tree): ... here.
17626 (vect_build_slp_tree_1): Compute which stmts of the SLP group
17627 match. Remove special-casing of mismatched complex loads.
17628 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
17629 re-try the match with swapped commutative operands.
17630 (vect_supported_load_permutation_p): Remove special-casing of
17631 mismatched complex loads.
17632 (vect_analyze_slp_instance): Adjust.
17634 2013-04-17 Richard Biener <rguenther@suse.de>
17636 PR rtl-optimization/56921
17637 * cfgloop.h (struct loop): Add simple_loop_desc member.
17638 (struct niter_desc): Mark with GTY(()).
17639 (simple_loop_desc): Do not use aux field but simple_loop_desc.
17640 * loop-iv.c (get_simple_loop_desc): Likewise.
17641 (free_simple_loop_desc): Likewise.
17644 2013-04-16 Richard Biener <rguenther@suse.de>
17646 PR rtl-optimization/56921
17647 * loop-init.c (pass_rtl_move_loop_invariants): Add
17648 TODO_do_not_ggc_collect to todo_flags_finish.
17649 (pass_rtl_unswitch): Same.
17650 (pass_rtl_unroll_and_peel_loops): Same.
17651 (pass_rtl_doloop): Same.
17653 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
17655 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
17656 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
17657 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
17659 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
17660 * tree-streamer.c (record_common_node): Adjust reference in comment.
17662 2013-04-17 Terry Guo <terry.guo@arm.com>
17664 * config/arm/cortex-m4.md: Add a new bypass.
17666 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17668 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
17670 (*subs_<optab><mode>_multp2): New pattern.
17671 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
17672 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
17674 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17676 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
17677 (*subs_mul_imm_<mode>): New pattern.
17679 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
17682 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
17683 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
17684 (vsx_movti_32bit): Change j->wa to O->wa.
17686 2013-04-16 Richard Biener <rguenther@suse.de>
17688 PR rtl-optimization/56921
17689 * loop-init.c (pass_rtl_move_loop_invariants): Add
17690 TODO_do_not_ggc_collect to todo_flags_finish.
17691 (pass_rtl_unswitch): Same.
17692 (pass_rtl_unroll_and_peel_loops): Same.
17693 (pass_rtl_doloop): Same.
17695 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
17697 * config/arm/arm.c (emit_multi_reg_push): New declaration
17698 for an existing function.
17699 (arm_emit_strd_push): New function.
17700 (arm_expand_prologue): Used here.
17701 (arm_emit_ldrd_pop): New function.
17702 (arm_expand_epilogue): Used here.
17703 (arm_get_frame_offsets): Update condition.
17704 (arm_emit_multi_reg_pop): Add a special case for load of a single
17705 register with writeback.
17707 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
17709 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
17712 2013-04-16 Richard Biener <rguenther@suse.de>
17714 PR tree-optimization/56756
17715 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
17716 (first_mem_ref_loc): New.
17717 (execute_sm): Place the load temporarily before a previous
17718 access instead of in the latch edge to ensure its SSA dependencies
17719 are defined at points dominating the load.
17721 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
17723 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
17724 correct fix by moving header and footer insn to the footer of
17725 the merged basic block. Clear BB_END of the merged-away block.
17727 PR middle-end/43631
17728 * emit-rtl.c (make_note_raw): New function.
17729 (link_insn_into_chain): New static inline function.
17730 (add_insn): Use it.
17731 (add_insn_before, add_insn_after): Factor insn chain linking code...
17732 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
17733 using link_insn_into_chain.
17734 (note_outside_basic_block_p): New helper function for emit_note_after
17735 and emit_note_before.
17736 (emit_note_after): Use nobb variant of add_insn_after if the note
17737 should not be contained in a basic block.
17738 (emit_note_before): Use nobb variant of add_insn_before if the note
17739 should not be contained in a basic block.
17740 (emit_note_copy): Use make_note_raw.
17741 (emit_note): Likewise.
17742 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
17743 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
17744 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
17745 the moved barrier the tail of the basic block it follows.
17746 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
17748 2013-04-15 Jakub Jelinek <jakub@redhat.com>
17750 PR tree-optimization/56962
17751 * gimple-ssa-strength-reduction.c (record_increment): Only set
17752 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
17753 either rhs1 or rhs2 is equal to c->base_expr.
17755 2013-04-15 Richard Biener <rguenther@suse.de>
17757 PR tree-optimization/56933
17758 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
17760 (GROUP_READ_WRITE_DEPENDENCE): Remove.
17761 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
17762 * tree-vect-data-refs.c (vect_analyze_group_access): Move
17763 dependence check ...
17764 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
17766 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
17767 GROUP_READ_WRITE_DEPENDENCE.
17769 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17771 * emit-rtl.c (reset_all_used_flags): New function.
17772 (verify_rtl_sharing): Call reset_all_used_flags before and after
17773 performing the checks.
17775 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17777 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
17778 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
17779 * config/arm/constraints.md (De): New constraint.
17780 * config/arm/neon.md (anddi3_neon): Delete.
17781 (neon_vand<mode>): Expand to standard anddi3 pattern.
17782 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
17783 Move earlier in the file.
17784 (neon_inv_logic_op2): Likewise.
17785 (arm_anddi_operand_neon): New predicate.
17787 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17789 * configure.ac (gcc_cv_ld_as_needed): Set
17790 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
17791 Use -z ignore, -z record on *-*-solaris2*.
17792 (HAVE_LD_AS_NEEDED): Update comment.
17793 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
17794 * configure: Regenerate.
17795 * config.in: Regenerate.
17796 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
17797 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
17798 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
17799 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
17800 equivalents. Fix markup.
17801 * doc/tm.texi: Regenerate.
17803 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
17805 * config/i386/i386.opt: New option mstack-protector-guard=.
17806 * config/i386/i386-opts.h: Add enum stack_protector_guard.
17807 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
17808 TARGET_SSP_TLS_GUARD.
17809 * config/i386/i386.c (ix86_option_override_internal): Set
17810 ix86_stack_protector_guard.
17811 * config/i386/i386.md (stack_protect_set): Enable for
17812 TARGET_SSP_TLS_GUARD only.
17813 (stack_protect_set_<mode>): Ditto.
17814 (stack_protect_test): Ditto.
17815 (stack_protect_test_<mode>): Ditto.
17816 * doc/invoke.texi (i386 Option): Document.
17818 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
17821 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
17822 (S_MODES): Set H_MODE bit.
17823 (SF_MODES): Set only S_MODE and SF_MODE bits.
17824 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
17825 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
17826 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
17827 <MODE_FLOAT>: Likewise.
17829 2013-04-15 Joey Ye <joey.ye@arm.com>
17831 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
17833 2013-04-15 Joey Ye <joey.ye@arm.com>
17835 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
17837 (thumb_far_jump_used_p): Count instruction size and set
17840 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
17842 * reorg.c (fill_simple_delay_slots): Reindent block of code.
17843 * resource.c (mark_target_live_regs): Reformat conditional block.
17845 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
17847 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
17848 notes, they are emitted only just before final.
17849 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
17851 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
17853 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
17854 * cfgrtl.c (delete_insn): Call it here instead.
17855 * lra-spills.c (lra_final_code_change): Use delete_insn.
17856 * haifa-sched.c (sched_remove_insn): Likewise.
17857 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
17858 returning to the nop pool.
17859 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
17860 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
17862 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
17864 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
17865 * doc/tm.texi: Regenerated.
17867 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
17869 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
17872 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
17874 * df-core.c (df_find_def): Compare register numbers.
17875 (df_find_use): Likewise.
17877 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
17880 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
17881 lra_in_progress for return.
17883 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
17885 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
17886 define_insn into define_insn_and_split and emit movsicc patterns.
17888 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
17890 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
17892 2013-04-12 Richard Biener <rguenther@suse.de>
17894 * tree-pass.h (TODO_do_not_ggc_collect): New.
17895 * passes.c (execute_one_ipa_transform_pass): Honor
17896 TODO_do_not_ggc_collect.
17897 (execute_one_pass): Likewise.
17900 2013-04-10 Richard Biener <rguenther@suse.de>
17902 * passes.c (init_optimization_passes): Remove reload pass.
17903 * ira.c (do_reload): Merge into ...
17905 (rest_of_handle_reload): Remove.
17906 (pass_reload): Likewise.
17907 * config/i386/i386.c (ix86_option_override): Refer to ira instead
17908 of reload for vzeroupper pass placement.
17910 2013-04-12 Jakub Jelinek <jakub@redhat.com>
17912 PR tree-optimization/56918
17913 PR tree-optimization/56920
17914 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
17915 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
17916 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
17917 use wide_mul_with_sign method.
17919 2013-04-12 Richard Biener <rguenther@suse.de>
17921 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
17922 not be considered a gimple constant.
17924 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
17926 * fold-const.c (const_binop): Handle vector shifts by a scalar.
17927 (fold_binary_loc): Call const_binop also for mixed vector-scalar
17930 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
17931 Jakub Jelinek <jakub@redhat.com>
17933 * opts.c: Include diagnostic-color.h.
17934 (common_handle_option): Handle OPT_fdiagnostics_color_.
17935 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
17936 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
17937 (diagnostic-color.o): New.
17938 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
17939 (diagnostic_color_rule): New enum.
17940 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
17941 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
17942 the location string.
17943 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
17944 either NULL, or color kind.
17945 * diagnostic-color.c: New file.
17946 * diagnostic-color.h: New file.
17947 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
17949 * doc/invoke.texi (-fdiagnostics-color): Document.
17950 * pretty-print.h (pp_show_color): Define.
17951 (struct pretty_print_info): Add show_color field.
17952 * diagnostic.c: Include diagnostic-color.h.
17953 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
17954 macros. Colorize error:, warning: etc. strings and also the location
17956 (diagnostic_show_locus): Colorize the caret line.
17957 * pretty-print.c: Include diagnostic-color.h.
17958 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
17959 inside of %< %> quotes or quoted through q format modifier.
17961 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17963 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
17965 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17967 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
17968 code in CC_NZ mode.
17969 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
17972 2013-04-11 Marek Polacek <polacek@redhat.com>
17974 PR tree-optimization/48184
17975 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
17977 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
17979 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
17980 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
17981 (skip_simple_arithmetic): Tidy up.
17982 * tree.h (skip_simple_constant_arithmetic): Declare.
17984 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17986 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
17988 2013-04-11 Richard Biener <rguenther@suse.de>
17990 * tree-vect-loop.c (get_initial_def_for_induction): Properly
17991 generate vector constants.
17993 2013-04-11 Richard Biener <rguenther@suse.de>
17995 PR tree-optimization/56878
17996 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
17997 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
17999 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
18000 Prefer to align the DR with the most invariant base address.
18002 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
18004 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
18007 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
18009 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
18010 floating-point vector comparisons against 0.
18012 2013-04-11 Jakub Jelinek <jakub@redhat.com>
18014 PR tree-optimization/56899
18015 * fold-const.c (extract_muldiv_1): Apply distributive law
18016 only if TYPE_OVERFLOW_WRAPS (ctype).
18018 2013-04-11 Bin Cheng <bin.cheng@arm.com>
18021 * ira-costs.c (scan_one_insn): Check whether the source rtx of
18022 loading has side effect.
18024 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
18026 * config/sparc/sparc.c: Include tree-pass.h.
18027 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
18028 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
18029 head of file. Change return type. Split off gate function.
18030 (sparc_gate_work_around_errata): New function.
18031 (pass_work_around_errata): New pass definition.
18032 (insert_pass_work_around_errata) New pass insert definition to
18033 insert pass_work_around_errata just after delayed-branch scheduling.
18034 (sparc_option_override): Insert the pass.
18035 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
18037 2013-04-10 David S. Miller <davem@davemloft.net>
18039 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
18040 or -mcpu=hypersparc.
18042 * target.def (cstore_mode): New hook.
18043 * target.h: Include insn-codes.h
18044 * targhooks.c: Likewise.
18045 (default_cstore_mode): New function.
18046 * targhooks.h: Declare it.
18047 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
18048 * doc/tm.texi: Rebuild.
18049 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
18050 target hook, rather than inspecting the insn_data.
18051 * config/sparc/sparc.c (sparc_cstore_mode): New function.
18052 (TARGET_CSTORE_MODE): Redefine.
18053 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
18055 * config/sparc/predicates.md (cstore_result_operand): New special
18057 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
18058 Use it for operand 0.
18059 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
18060 (*snesi_special): Likewise.
18061 (*snesi_zero): Likewise.
18062 (*seqsi_zero): Likewise.
18063 (*sltu_insn): Likewise.
18064 (*sgeu_insn): Likewise.
18065 (*seqdi_special): Make operand 0 and comparison operation be of
18067 (*snedi_special): Likewise.
18068 (*snedi_special_vis3): Likewise.
18069 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
18070 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
18071 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
18072 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
18073 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
18074 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
18075 (*sltu_extend_sp64): Likewise.
18076 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
18077 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
18078 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
18079 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
18080 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
18082 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
18084 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
18085 (aarch64_start_file): Use the new function.
18087 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
18088 Jason Merrill <jason@redhat.com>
18090 * common.opt: Add -gdwarf.
18091 * opts.c (common_handle_option): Handle it.
18092 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
18094 2013-04-10 Richard Biener <rguenther@suse.de>
18096 * passes.c (execute_todo): Do not call ggc_collect conditional here.
18097 (execute_one_ipa_transform_pass): But unconditionally here.
18098 (execute_one_pass): And here.
18099 (init_optimization_passes): Remove reload pass.
18100 * tree-pass.h (TODO_ggc_collect): Remove.
18101 (pass_reload): Likewise.
18102 * ira.c (do_reload): Merge into ...
18104 (rest_of_handle_reload): Remove.
18105 (pass_reload): Likewise.
18106 * config/i386/i386.c (ix86_option_override): Refer to ira instead
18107 of reload for vzeroupper pass placement.
18108 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
18109 and todo_flags_finish of all passes.
18111 2013-04-10 Richard Biener <rguenther@suse.de>
18113 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
18114 first_const_oprnd field, rename first_def_type to first_op_type.
18115 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
18116 (vect_get_and_check_slp_defs): Always use the type of the
18117 operand. Allow mixed vect_external_def, vect_constant_def types.
18118 (vect_get_constant_vectors): Handle mixed vect_external_def,
18119 vect_constant_def types.
18121 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
18123 PR tree-optimization/55524
18124 * tree-ssa-math-opts.c
18125 (convert_mult_to_fma): Don't use an fms construct
18126 when we don't have an fms operation, but fnma, and it looks
18127 likely that we'll be able to use the latter.
18129 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
18131 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
18133 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
18134 inline fail caused by overwritable functions.
18136 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
18138 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
18139 unnecessary bits in the constant power of two case.
18141 2013-04-10 Richard Biener <rguenther@suse.de>
18143 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
18144 broken code swapping operands.
18145 (vect_build_slp_tree): Do not compute load permutations here.
18146 (vect_analyze_slp_instance): Compute load permutations here,
18147 after building the SLP tree.
18149 2013-04-09 Christian Bruel <christian.bruel@st.com>
18151 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
18152 of next/prev_real_insn.
18154 2013-04-09 Jan Hubicka <jh@suse.cz>
18156 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
18157 Drop aliased parameter.
18158 (function_and_variable_visibility): Do not handle alias pairs.
18159 * cgraph.c (varpool_externally_visible_p): Update prototype.
18160 * varpool.c (varpool_add_new_variable): Update.
18162 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18164 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
18166 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
18168 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
18170 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
18172 2013-04-09 Marek Polacek <polacek@redhat.com>
18174 PR tree-optimization/48762
18175 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
18177 2013-04-09 Richard Biener <rguenther@suse.de>
18179 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
18181 (vect_build_slp_tree): Likewise.
18182 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
18183 calculating the cost of a SLP instance.
18184 (vect_analyze_slp_instance): Use it from here, after building
18187 2013-04-09 Jakub Jelinek <jakub@redhat.com>
18189 PR middle-end/56883
18190 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
18191 expand_omp_for_static_chunk): Use simple_p = true in
18192 force_gimple_operand_gsi calls when assigning to addressable decls.
18194 2013-04-09 Jeff Law <law@redhat.com>
18196 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
18197 when the boolean was created by converting a wider object which
18198 had a boolean range.
18200 2013-04-09 Richard Biener <rguenther@suse.de>
18202 * tree-vectorizer.h (slp_void_p): Remove.
18203 (slp_tree): Typedef before _slp_tree declaration.
18204 (struct _slp_tree): Use a vector of slp_tree as children.
18205 (vect_get_place_in_interleaving_chain): Remove.
18206 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
18208 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
18210 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
18211 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
18212 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
18213 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
18214 Use slp_node instead of slp_void_p and adjust.
18216 2013-04-09 Richard Biener <rguenther@suse.de>
18218 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
18219 work that is not necessary.
18221 2013-04-09 Jakub Jelinek <jakub@redhat.com>
18223 PR tree-optimization/56854
18224 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
18225 forward into clobber stmts if it would change MEM_REF lhs into
18228 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
18230 * tree.c (type_hash_lookup, type_hash_add): Make static.
18231 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
18233 2013-04-09 Richard Biener <rguenther@suse.de>
18235 * tree.h (unsave_expr_now): Remove.
18236 * tree-inline.c (mark_local_for_remap_r): Remove.
18237 (unsave_expr_1): Likewise.
18238 (unsave_r): Likewise.
18239 (unsave_expr_now): Likewise.
18240 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
18241 (propagate_tree_value): Likewise.
18243 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
18245 * doc/rtl.texi (sequence): Rewrite documentation to match the
18246 current use of SEQUENCE rtl objects.
18247 * rtl.def (SEQUENCE): Likewise.
18249 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
18250 Update documentation.
18251 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
18252 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
18254 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
18256 2013-04-08 Teresa Johnson <tejohnson@google.com>
18258 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
18259 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
18261 (estimate_edge_size_and_time): Add comment to suggest using rounding
18263 (estimate_node_size_and_time): Ditto.
18264 (remap_edge_change_prob): Use helper rounding divide methods.
18265 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
18266 (gimple_mod_pow2_value_transform): Ditto.
18267 (gimple_mod_subtract_transform): Ditto.
18268 (gimple_ic_transform): Ditto.
18269 (gimple_stringops_transform): Ditto.
18270 * stmt.c (conditional_probability): Ditto.
18271 (emit_case_dispatch_table): Ditto.
18272 * lto-cgraph.c (merge_profile_summaries): Ditto.
18273 * tree-optimize.c (execute_fixup_cfg): Ditto.
18274 * cfgcleanup.c (try_forward_edges): Ditto.
18275 * cfgloopmanip.c (scale_loop_profile): Ditto.
18277 (duplicate_loop_to_header_edge): Ditto.
18278 (lv_adjust_loop_entry_edge): Ditto.
18279 * tree-vect-loop.c (vect_transform_loop): Ditto.
18280 * profile.c (compute_branch_probabilities): Ditto.
18281 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
18282 * lto-streamer-in.c (input_cfg): Ditto.
18283 * gimple-streamer-in.c (input_bb): Ditto.
18284 * ipa-cp.c (update_profiling_info): Ditto.
18285 (update_specialized_profile): Ditto.
18286 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
18287 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
18289 * sched-rgn.c (compute_dom_prob_ps): Ditto.
18290 (compute_trg_info): Ditto.
18291 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
18292 (purge_dead_edges): Ditto.
18293 * loop-unswitch.c (unswitch_loop): Ditto.
18294 * cgraphclones.c (cgraph_clone_edge): Ditto.
18295 (cgraph_clone_node): Ditto.
18296 * tree-inline.c (copy_bb): Ditto.
18297 (copy_edges_for_bb): Ditto.
18298 (initialize_cfun): Ditto.
18299 (copy_cfg_body): Ditto.
18300 (expand_call_inline): Ditto.
18302 2013-04-08 Kai Tietz <ktietz@redhat.com>
18304 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
18305 TARGET_CYGWIN64 by TARGET_64BIT.
18307 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
18309 * config/epiphany/epiphany.md (GPR_1): New constant.
18310 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
18311 * config/epiphany/epiphany.c (gen_compare_reg):
18312 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
18313 is already in place.
18314 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
18315 Don't require being called during rtl expansion; If y operlaps r0,
18317 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
18318 (epiphany_expand_epilogue): Likewise.
18320 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
18321 Don't use CC_FPmode for ORDERED / UNORDERED.
18322 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
18324 * config/epiphany/constraints.md (CnL): New constraint.
18325 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
18326 * config/epiphany/predicates.md (add_operand): Allow 1024.
18328 * config/epiphany/epiphany.md (logical_op): New code iterator.
18329 (op_mnc): New code attribute.
18330 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
18331 (mov_f+1, mov_f+2): New peephole2 patterns.
18333 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
18334 (cstoresi4): Also allow re-use of zero result when doing a NE
18335 comparison to a non-zero operand.
18336 Use (clobber (scratch)) for first insn if the gpr output is not needed.
18338 * config/epiphany/epiphany.md (<insn_opname>v2si3):
18339 Use gen_addsi3_i / gen_subsi3_i.
18341 2013-04-08 Jakub Jelinek <jakub@redhat.com>
18345 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
18346 contain anything but clobbers, at most one __builtin_stack_restore,
18347 optionally debug stmts and final resx, and if it has at least one
18348 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
18349 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
18350 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
18351 which isn't defaut definition, remove them.
18352 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
18353 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
18354 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
18355 with MEM_REF LHS with SSA_NAME address.
18357 2013-04-08 Jeff Law <law@redhat.com>
18359 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
18361 2013-04-08 Richard Biener <rguenther@suse.de>
18363 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
18365 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
18366 determined vector type.
18367 (vect_analyze_data_refs): Likewise.
18368 (vect_get_new_vect_var): Adjust.
18369 (vect_create_destination_var): Preserve SSA name versions.
18370 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
18371 not dump anything here.
18373 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
18375 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
18376 Add member lr_slot_known.
18377 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
18379 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
18380 Remove code that sets lr_slot_offset according to what a previous
18381 version of epiphany_emit_save_restore used to do.
18382 (epiphany_emit_save_restore): When doing an lr save or restore,
18383 set/verify lr_slot_known and lr_slot_offset.
18385 2013-04-08 Xinyu Qi <xyqi@marvell.com>
18388 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
18391 2013-04-08 Richard Biener <rguenther@suse.de>
18393 * alias.c (find_base_term): Fix thinko in previous change.
18395 2013-04-08 Jakub Jelinek <jakub@redhat.com>
18397 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
18398 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
18399 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
18400 if possible to compute val.
18401 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
18402 For QImode integers don't require anything about precision. Use
18403 const_with_all_bytes_same to find out if the constant doesn't have
18404 repeated bytes in it.
18406 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18408 * config/s390/s390.c (s390_expand_insv): Only accept insertions
18411 2013-04-08 Marek Polacek <polacek@redhat.com>
18413 PR rtl-optimization/48182
18414 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
18417 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18420 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
18421 nuses, make sure we have a label.
18423 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18426 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
18427 (rs6000_emit_swdiv_low_precision): Remove.
18428 (rs6000_emit_swdiv): Rewrite to handle between one and four
18429 iterations of Newton-Raphson generally; modify required number of
18430 iterations for some cases.
18431 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
18433 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
18435 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
18436 set-but-unused variable.
18438 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
18439 basic blocks of released function bodies garbage-collectable.
18441 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
18442 (struct rtl_opt_pass): Add TODO_df_finish.
18444 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
18446 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
18448 * config/arm/constraints.md (q): New constraint.
18449 * config/arm/ldrdstrd.md: New file.
18450 * config/arm/arm.md (ldrdstrd.md) New include.
18451 (arm_movdi): Use "q" instead of "r" constraint
18452 for double-word memory access.
18453 (movdf_soft_insn): Likewise.
18454 * config/arm/vfp.md (movdi_vfp): Likewise.
18455 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
18456 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
18457 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
18458 (mem_ok_for_ldrd_strd): Likewise.
18459 (output_move_double): Update assertion.
18461 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
18463 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
18465 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
18467 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
18468 define_insn_and_split.
18469 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
18471 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
18473 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
18474 define_insn_and_split.
18475 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
18476 (shiftsi3_compare): New pattern.
18477 (rrx): New pattern.
18478 * config/arm/unspecs.md (UNSPEC_RRX): New.
18480 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
18482 * config/arm/arm.md (negdi_extendsidi): New pattern.
18483 (negdi_zero_extendsidi): Likewise.
18485 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
18487 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
18488 define_insn_and_split.
18489 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
18490 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
18492 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
18494 * config/arm/arm.md (arm_subdi3): Convert define_insn into
18495 define_insn_and_split.
18496 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
18497 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
18499 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
18501 * config/arm/arm.md (subsi3_carryin): New pattern.
18502 (subsi3_carryin_const): Likewise.
18503 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
18504 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
18506 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
18508 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
18510 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
18512 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
18513 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
18515 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18517 * config/arm/arm.c (arm_expand_builtin): Change fcode
18518 type to unsigned int.
18520 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
18522 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
18524 2013-04-04 Ian Lance Taylor <iant@google.com>
18526 * doc/standards.texi (Standards): The Go frontend supports the Go 1
18529 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
18531 PR middle-end/56729
18532 * df-scan.c (df_insn_delete): Disable failing assert.
18534 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18536 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
18537 New function prototype.
18538 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
18539 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
18540 (arm_builtin_vectorized_function): New function.
18542 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18544 * config/arm/arm_neon_builtins.def: New file.
18545 * config/arm/arm.c (neon_builtin_data): Move contents to
18546 arm_neon_builtins.def.
18547 (enum arm_builtins): Include neon builtin definitions.
18548 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
18549 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
18551 2013-04-04 Marek Polacek <polacek@redhat.com>
18553 PR tree-optimization/48186
18554 * predict.c (maybe_hot_frequency_p): Return false if
18555 HOT_BB_FREQUENCY_FRACTION is 0.
18556 (cgraph_maybe_hot_edge_p): Likewise.
18558 2013-04-04 Richard Biener <rguenther@suse.de>
18560 PR tree-optimization/56826
18561 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
18564 2013-04-04 Richard Biener <rguenther@suse.de>
18566 PR tree-optimization/56213
18567 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
18568 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
18570 2013-04-04 Richard Biener <rguenther@suse.de>
18572 PR tree-optimization/56837
18573 * tree-loop-distribution.c (classify_partition): For non-zero
18574 values require that the value has the same precision as its
18575 mode to be useful as memset value.
18577 2013-04-03 Nick Clifton <nickc@redhat.com>
18579 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
18580 (fmssf4): Use fmsf.s on E3V5 architectures.
18581 (fnmasf4): Use fnmaf.s on E3V5 architectures.
18582 (fnmssf4): Use fnmsf.s on E3V5 architectures.
18584 2013-04-03 Jeff Law <law@redhat.com>
18586 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
18587 (lra-eliminations.o): Likewise.
18589 2013-04-03 Teresa Johnson <tejohnson@google.com>
18591 * gcov-io.c (compute_working_sets): Moved most of body of old
18592 compute_working_sets here from profile.c.
18593 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
18594 (gcov_working_set_t): Moved typedef here from basic-block.h
18595 (compute_working_set): Declare.
18596 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
18597 (get_working_sets): Renamed from compute_working_set,
18598 replace most of body with call to new compute_working_sets.
18599 (get_exec_counts): Replace call to compute_working_sets
18600 to get_working_sets.
18601 * profile.h (get_working_sets): Renamed from compute_working_set.
18602 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
18603 to get_working_sets.
18604 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
18605 * gcov-dump.c (dump_working_sets): New function.
18607 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
18609 * hwint.c (sext_hwi, zext_hwi): New functions.
18610 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
18611 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
18612 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
18613 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
18614 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
18615 (sext_hwi, zext_hwi): New functions.
18617 2013-04-03 Jeff Law <law@redhat.com>
18619 PR tree-optimization/56799
18620 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
18621 back test for widening conversion erroneously dropped in prior change.
18623 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18626 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
18627 instead of next_real_insn.
18629 2013-04-03 Marek Polacek <polacek@redhat.com>
18632 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
18634 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18637 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
18639 (thumb1_output_casesi): Likewise.
18640 (thumb2_output_casesi): Likewise.
18642 2013-04-03 Richard Biener <rguenther@suse.de>
18644 PR tree-optimization/56817
18645 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
18647 (tree_unroll_loops_completely_1): ... new function to manually
18648 walk the loop tree, properly defering outer loops of unrolled
18649 loops to later iterations.
18651 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
18653 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
18654 (vectorizable_load): Likewise.
18655 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
18656 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
18658 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
18660 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
18663 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
18665 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
18667 2013-04-03 Bin Cheng <bin.cheng@arm.com>
18669 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
18671 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
18673 PR tree-optimization/56790
18674 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
18677 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
18679 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
18681 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
18682 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
18685 2013-04-03 Jakub Jelinek <jakub@redhat.com>
18688 * tree.h (force_folding_builtin_constant_p): New decl.
18689 * builtins.c (force_folding_builtin_constant_p): New variable.
18690 (fold_builtin_constant_p): Fold immediately also if
18691 force_folding_builtin_constant_p.
18693 2013-04-03 Richard Biener <rguenther@suse.de>
18695 PR tree-optimization/56812
18696 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
18697 DRs of the same interleaving chain are independent.
18699 2013-04-02 Jason Merrill <jason@redhat.com>
18701 * gdbinit.in (pbb): Use debug fn.
18703 2013-04-02 Lawrence Crowl <crowl@google.com>
18705 * sese.h (struct ivtype_map_elt_s): Remove unused.
18706 (extern debug_ivtype_map): Remove unused.
18707 (extern eq_ivtype_map_elts): Remove unused.
18708 * sese.c (debug_ivtype_map): Removed unused.
18709 (debug_ivtype_map_1): Removed unused.
18710 (debug_ivtype_elt): Remove unused.
18711 (eq_ivtype_map_elts): Remove unused.
18714 2013-04-02 Kai Tietz <ktietz@redhat.com>
18717 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
18718 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
18719 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
18721 (legitimize_pe_coff_symbol): Likewise.
18722 (is_imported_p): New helper-function.
18723 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
18724 for Windows x64 targets.
18725 (ix86_expand_prologue): Optimize for pe-coff targets.
18726 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
18727 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
18728 medium/large code-model.
18729 (legitimize_pic_address): Likewise.
18730 (legitimize_tls_address): Likewise.
18731 (ix86_expand_call): Likewise.
18732 (x86_output_mi_thunk): Likewise.
18733 (get_dllimport_decl): Add new beimport argument.
18734 (construct_plt_address): Don't assert for x64 pe-coff targets.
18735 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
18737 (SYMBOL_FLAG_STUBVAR): New macro.
18738 (SYMBOL_REF_STUBVAR_P): Likewise.
18739 * config/i386/winnt.c (stub_list): New structure.
18740 (stub_head): New local variable.
18741 (i386_pe_record_stub): New function.
18742 (i386_pe_file_end): Emit refptr-stubs.
18744 2013-04-02 Jakub Jelinek <jakub@redhat.com>
18746 PR rtl-optimization/56745
18747 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
18748 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
18751 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
18752 and both of them are MEM_REFs, just compare first argument for
18753 equality and attempt to deal even with differing offsets.
18756 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
18757 of gimple_clobber_p to be MEM_REF.
18758 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
18759 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
18760 after gimplification.
18761 * asan.c (get_mem_ref_of_assignment): Don't instrument
18762 gimple_clobber_p stmts.
18763 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
18764 gimple_clobber_p stmt if they have MEM_REF lhs and
18765 are dead because of another gimple_clobber_p stmt.
18766 * tree-ssa-live.c (clear_unused_block_pointer): Treat
18767 gimple_clobber_p stmts like debug stmts.
18768 (remove_unused_locals): Remove clobbers with MEM_REF lhs
18769 that refer to unused VAR_DECLs or uninitialized values.
18770 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
18771 gimple_clobber_p stmts if they refer to removed parameters.
18772 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
18775 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
18777 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
18778 using SWI48 mode attribute.
18780 2013-04-02 Wei Mi <wmi@google.com>
18782 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
18783 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
18784 *<rotate_insn><mode>3_mask in i386.md.
18786 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
18788 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
18790 2013-04-02 Richard Biener <rguenther@suse.de>
18792 PR tree-optimization/56778
18793 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
18794 Runtime alias tests are not supported for gather loads.
18795 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
18796 stmts referenced from SSA operands before updating SSA form.
18798 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
18799 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
18801 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
18802 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
18803 * config/arm/cortex-a53.md: New file.
18804 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
18805 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
18806 * config/arm/arm.c (arm_issue_rate): Likewise.
18807 * config/arm/arm-tune.md: Regenerate
18808 * config/arm/arm-tables.opt: Regenerate.
18809 * config/arm/arm-cores.def: Add cortex-a53.
18811 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
18813 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
18816 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
18818 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
18819 scalar load/store operations using B/H registers.
18820 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
18822 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
18824 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
18826 * config/aarch64/aarch64.c
18827 (aarch64_simd_scalar_immediate_valid_for_move): New.
18828 * config/aarch64/aarch64-protos.h
18829 (aarch64_simd_scalar_immediate_valid_for_move): New.
18830 * config/aarch64/constraints.md (Dh, Dq): New.
18831 * config/aarch64/iterators.md (hq): New.
18833 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
18835 * reorg.c (get_branch_condition): Deal with conditional returns.
18836 (fill_simple_delay_slots): Remove dead code dealing with jumps.
18838 2013-04-01 Wei Mi <wmi@google.com>
18840 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
18841 Truncate operand 2 using %b asm operand modifier.
18842 (*<shift_insn><mode>3_mask): Ditto.
18843 (*<rotate_insn><mode>3_mask): Ditto.
18845 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
18847 PR middle-end/56798
18848 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
18850 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
18852 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
18854 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
18856 2013-03-30 Lawrence Crowl <crowl@google.com>
18858 * dse.c (clear_alias_sets): Remove never set.
18859 (disqualified_clear_alias_sets): Remove never set.
18860 (clear_alias_mode_pool): Remove never set.
18861 (dse_step0): Remove condition that is never true.
18862 (canon_address): Remove condition that is never true.
18863 (dse_step7): Remove condition that is never true.
18864 (rest_of_handle_dse): Remove condition that is never true.
18865 (rest_of_handle_dse::did_global): Remove never read from above.
18866 (dse_step2_spill): Remove never called from above.
18867 (dse_step5_spill): Remove never called from above.
18869 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
18871 * doc/md.texi (Standard Names) <casesi>: Update documentation for
18872 JUMP_TABLE_DATA changes.
18873 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
18874 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
18875 (Insns) <jump_table_data>: New entry.
18876 * doc/tm.texi: Regenerate.
18878 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
18880 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
18881 for table jump at the end of a basic block using tablejump_p.
18882 * targhooks.c (default_invalid_within_doloop): Likewise.
18883 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
18884 target hook implementation that is identical to the default hook.
18885 (rs6000_invalid_within_doloop): Remove.
18887 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
18888 unused variable from tablejump_p call.
18890 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
18891 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
18892 (INSN_DELETED_P): Likewise.
18893 (emit_jump_table_data): New prototype.
18894 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
18895 after 4th as unused.
18896 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
18897 * sched-vis.c (print_insn): Likewise.
18898 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
18899 insn for compatibility with back ends that use next_active_insn to
18900 identify jump table data.
18901 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
18902 (remove_insn): Likewise.
18903 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
18905 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
18906 (emit_jump_table_data): New function.
18908 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
18909 basic block, a JUMP_TABLE_DATA never is.
18910 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
18911 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
18912 off from code handling real insns.
18913 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
18914 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
18915 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
18916 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
18917 is not a NONDEBUG_INSN_P.
18918 * ira-costs.c (scan_one_insn): Likewise.
18919 * jump.c (mark_all_labels): Likewise.
18920 (mark_jump_label_1): Likewise.
18921 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
18922 * lra.c (get_insn_freq): Expect all insns reaching here to be in
18924 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
18925 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
18926 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
18927 JUMP_TABLE_DATA_P insns.
18928 (calculate_elim_costs_all_insns): Likewise.
18929 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
18930 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
18931 (delete_output_reload): Code style fixups.
18932 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
18933 insn flags on this non-insn.
18934 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
18935 as scheduling barriers, for pre-change compatibility.
18936 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
18937 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
18939 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
18940 redundant JUMP_TABLE_DATA_P test.
18941 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
18942 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
18943 (frv_for_each_packet): Likewise.
18944 * config/i386/i386.c (min_insn_size): Likewise.
18945 (ix86_avoid_jump_mispredicts): Likewise.
18946 * config/m32r/m32r.c (m32r_is_insn): Likewise.
18947 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
18948 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
18949 (mips16_insn_length): Robustify.
18950 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
18951 (mips16_split_long_branches): Likewise.
18952 * config/pa/pa.c (pa_combine_instructions): Likewise.
18953 * config/rs6000/rs6000.c (get_next_active_insn): Treat
18954 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
18955 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
18956 as contributing to pool range lengths.
18957 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
18958 Remove redundant JUMP_TABLE_DATA_P test.
18959 (sh_loop_align): Likewise.
18960 (split_branches): Likewise.
18961 (sh_insn_length_adjustment): Likewise.
18962 * config/spu/spu.c (get_branch_target): Likewise.
18964 2013-03-29 Jan Hubicka <jh@suse.cz>
18966 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
18967 gcov streaming; stream hot bb threshold to ltrans.
18968 * predict.c (get_hot_bb_threshold): Break out from ....
18969 (maybe_hot_count_p): ... here.
18970 (set_hot_bb_threshold): New function.
18971 * lto-section-in.c (lto_section_name): Add profile.
18972 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
18973 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
18974 and data-streamer.h
18975 (histogram_entry): New structure.
18976 (histogram, histogram_pool): New global vars.
18977 (histogram_hash): New structure.
18978 (histogram_hash::hash): New method.
18979 (histogram_hash::equal): Likewise.
18980 (account_time_size): New function.
18981 (cmp_counts): New function.
18982 (dump_histogram): New function.
18983 (ipa_profile_generate_summary): New function.
18984 (ipa_profile_write_summary): New function.
18985 (ipa_profile_read_summary): New function.
18986 (ipa_profile): Decide on threshold.
18987 (pass_ipa_profile): Add ipa_profile_write_summary and
18988 ipa_profile_read_summary.
18989 * Makefile.in (ipa.o): Update dependencies.
18990 * lto-streamer.h (LTO_section_ipa_profile): New section.
18992 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
18994 * tree.h (VAR_P): New.
18996 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
18999 * doc/invoke.texi ([-fwhole-program]): Fix typo.
19001 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
19003 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
19004 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
19005 (control_flow_insn_p): Likewise.
19006 * cfgrtl.c (duplicate_insn_chain): Likewise.
19007 * final.c (get_attr_length_1): Likewise.
19008 (shorten_branches): Likewise.
19009 (final_scan_insn): Likewise.
19010 * function.c (instantiate_virtual_regs): Likewise.
19011 * gcse.c (insert_insn_end_basic_block): Likewise.
19012 * ira-costs.c (scan_one_insn): Likewise.
19013 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
19014 * lra.c (check_rtl): Likewise.
19015 * reload1.c (elimination_costs_in_insn): Likewise.
19016 * reorg.c (follow_jumps): Likewise.
19018 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
19019 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
19020 (thumb_far_jump_used_p): Likewise.
19021 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
19022 (workaround_speculation): Likewise.
19023 (add_sched_insns_for_speculation): Likewise.
19024 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
19025 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
19026 (frv_for_each_packet): Likewise.
19027 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
19028 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
19029 (final_emit_insn_group_barriers): Likewise.
19030 * config/m32r/m32r.c (m32r_is_insn): Likewise.
19031 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
19032 (mips16_insn_length): Likewise.
19033 * config/pa/pa.c (pa_reorg): Likewise.
19034 (pa_combine_instructions): Likewise.
19035 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
19036 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
19037 (sh_reorg): Likewise.
19038 (split_branches): Likewise.
19039 * config/spu/spu.c (get_branch_target): Likewise.
19041 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
19044 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
19046 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
19047 Fix declaration name.
19049 2013-03-28 Lawrence Crowl <crowl@google.com>
19051 * graphds.h (struct graph.indicies): Remove unused.
19052 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
19053 (SCOP_ORIGINAL_PDDRS): Remove unused.
19054 * sese.h (extern insert_loop_close_phis): Removed unused.
19055 (extern insert_guard_phis): Removed unused.
19056 (extern ivtype_map_elt_info): Removed unused.
19057 (new_ivtype_map_elt): Removed unused.
19058 * sese.c (ivtype_map_elt_info): Removed unused.
19060 2013-03-28 Lawrence Crowl <crowl@google.com>
19062 * Makefile.in: Add several missing include dependences.
19064 (test-dump.o): New. This object is not added to any executable,
19065 but is present for ad-hoc testing.
19067 (debug (const bitmap_head_def &)): New.
19068 (debug (const bitmap_head_def *)): New.
19070 (extern debug (const bitmap_head_def &)): New.
19071 (extern debug (const bitmap_head_def *)): New.
19073 (debug (edge_def &)): New.
19074 (debug (edge_def *)): New.
19076 (debug (basic_block_def &)): New.
19077 (debug (basic_block_def *)): New.
19079 (dump_node (const_tree, int, FILE *)): Correct source file.
19081 (debug (die_struct &)): New.
19082 (debug (die_struct *)): New.
19084 (extern debug (die_struct &)): New.
19085 (extern debug (die_struct *)): New.
19086 * gimple-pretty-print.c
19087 (debug (gimple_statement_d &)): New.
19088 (debug (gimple_statement_d *)): New.
19089 * gimple-pretty-print.h
19090 (extern debug (gimple_statement_d &)): New.
19091 (extern debug (gimple_statement_d *)): New.
19093 (debug (ira_allocno_copy &)): New.
19094 (debug (ira_allocno_copy *)): New.
19095 (debug (ira_allocno &)): New.
19096 (debug (ira_allocno *)): New.
19098 (extern debug (ira_allocno_copy &)): New.
19099 (extern debug (ira_allocno_copy *)): New.
19100 (extern debug (ira_allocno &)): New.
19101 (extern debug (ira_allocno *)): New.
19103 (debug (live_range &)): New.
19104 (debug (live_range *)): New.
19106 (debug (lra_live_range &)): New.
19107 (debug (lra_live_range *)): New.
19109 (debug (lra_live_range &)): New.
19110 (debug (lra_live_range *)): New.
19112 (debug (omega_pb_d &)): New.
19113 (debug (omega_pb_d *)): New.
19115 (extern debug (omega_pb_d &)): New.
19116 (extern debug (omega_pb_d *)): New.
19118 (debug (const rtx_def &)): New.
19119 (debug (const rtx_def *)): New.
19121 (debug_tree (tree): Move within file.
19122 (debug_raw (const tree_node &)): New.
19123 (debug_raw (const tree_node *)): New.
19124 (dump_tree_via_hooks (const tree_node *, int)): New.
19125 (debug (const tree_node &)): New.
19126 (debug (const tree_node *)): New.
19127 (debug_verbose (const tree_node &)): New.
19128 (debug_verbose (const tree_node *)): New.
19129 (debug_head (const tree_node &)): New.
19130 (debug_head (const tree_node *)): New.
19131 (debug_body (const tree_node &)): New.
19132 (debug_body (const tree_node *)): New.
19133 (debug_vec_tree (tree): Move and reimplement in terms of dump.
19134 (debug (vec<tree, va_gc> &)): New.
19135 (debug (vec<tree, va_gc> *)): New.
19137 (extern debug (const rtx_def &)): New.
19138 (extern debug (const rtx_def *)): New.
19140 (debug_raw (simple_bitmap_def &)): New.
19141 (debug_raw (simple_bitmap_def *)): New.
19142 (debug (simple_bitmap_def &)): New.
19143 (debug (simple_bitmap_def *)): New.
19145 (extern debug (simple_bitmap_def &)): New.
19146 (extern debug (simple_bitmap_def *)): New.
19147 (extern debug_raw (simple_bitmap_def &)): New.
19148 (extern debug_raw (simple_bitmap_def *)): New.
19150 (debug (vinsn_def &)): New.
19151 (debug (vinsn_def *)): New.
19152 (debug_verbose (vinsn_def &)): New.
19153 (debug_verbose (vinsn_def *)): New.
19154 (debug (expr_def &)): New.
19155 (debug (expr_def *)): New.
19156 (debug_verbose (expr_def &)): New.
19157 (debug_verbose (expr_def *)): New.
19158 (debug (vec<rtx> &)): New.
19159 (debug (vec<rtx> *)): New.
19161 (extern debug (vinsn_def &)): New.
19162 (extern debug (vinsn_def *)): New.
19163 (extern debug_verbose (vinsn_def &)): New.
19164 (extern debug_verbose (vinsn_def *)): New.
19165 (extern debug (expr_def &)): New.
19166 (extern debug (expr_def *)): New.
19167 (extern debug_verbose (expr_def &)): New.
19168 (extern debug_verbose (expr_def *)): New.
19169 (extern debug (vec<rtx> &)): New.
19170 (extern debug (vec<rtx> *)): New.
19172 (_list_iter_cond_expr): Make inline instead of static.
19174 (debug (sreal &)): New.
19175 (debug (sreal *)): New.
19177 (extern debug (sreal &)): New.
19178 (extern debug (sreal *)): New.
19180 (extern debug_raw (const tree_node &)): New.
19181 (extern debug_raw (const tree_node *)): New.
19182 (extern debug (const tree_node &)): New.
19183 (extern debug (const tree_node *)): New.
19184 (extern debug_verbose (const tree_node &)): New.
19185 (extern debug_verbose (const tree_node *)): New.
19186 (extern debug_head (const tree_node &)): New.
19187 (extern debug_head (const tree_node *)): New.
19188 (extern debug_body (const tree_node &)): New.
19189 (extern debug_body (const tree_node *)): New.
19190 (extern debug (vec<tree, va_gc> &)): New.
19191 (extern debug (vec<tree, va_gc> *)): New.
19193 (debug (struct loop &)): New.
19194 (debug (struct loop *)): New.
19195 (debug_verbose (struct loop &)): New.
19196 (debug_verbose (struct loop *)): New.
19197 * tree-dump.c: Add header dependence.
19199 (extern debug (struct loop &)): New.
19200 (extern debug (struct loop *)): New.
19201 (extern debug_verbose (struct loop &)): New.
19202 (extern debug_verbose (struct loop *)): New.
19204 (debug (data_reference &)): New.
19205 (debug (data_reference *)): New.
19206 (debug (vec<data_reference_p> &)): New.
19207 (debug (vec<data_reference_p> *)): New.
19208 (debug (vec<ddr_p> &)): New.
19209 (debug (vec<ddr_p> *)): New.
19211 (extern debug (data_reference &)): New.
19212 (extern debug (data_reference *)): New.
19213 (extern debug (vec<data_reference_p> &)): New.
19214 (extern debug (vec<data_reference_p> *)): New.
19215 (extern debug (vec<ddr_p> &)): New.
19216 (extern debug (vec<ddr_p> *)): New.
19218 (debug (pt_solution &)): New.
19219 (debug (pt_solution *)): New.
19221 (extern debug (pt_solution &)): New.
19222 (extern debug (pt_solution *)): New.
19224 (debug (_var_map &)): New.
19225 (debug (_var_map *)): New.
19226 (debug (tree_live_info_d &)): New.
19227 (debug (tree_live_info_d *)): New.
19229 (extern debug (_var_map &)): New.
19230 (extern debug (_var_map *)): New.
19231 (extern debug (tree_live_info_d &)): New.
19232 (extern debug (tree_live_info_d *)): New.
19234 2013-03-28 Jan Hubicka <jh@suse.cz>
19236 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
19238 2013-03-28 Ian Bolton <ian.bolton@arm.com>
19240 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
19241 record only when desired or required.
19243 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
19245 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
19246 *vec_extractv2di_1_rex64. Use x64 isa attribute.
19248 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19250 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
19251 (*andsi3_compare0_uxtw): New pattern.
19252 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
19253 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
19255 2013-03-28 Jan Hubicka <jh@suse.cz>
19257 * data-streamer-in.c (streamer_read_gcov_count): New function.
19258 * gimple-streamer-out.c: Include value-prof.h.
19259 (output_gimple_stmt): Output histogram.
19260 (output_bb): Use streamer_write_gcov_count.
19261 * value-prof.c: Include data-streamer.h
19262 (dump_histogram_value): Add HIST_TYPE_MAX.
19263 (stream_out_histogram_value): New function.
19264 (stream_in_histogram_value): New function.
19265 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
19266 (stream_out_histogram_value, stream_in_histogram_value): Declare.
19267 * data-streamer-out.c (streamer_write_gcov_count): New function.
19268 (streamer_write_gcov_count_stream): New function.
19269 * lto-cgraph.c (lto_output_edge): Update counter streaming.
19270 (lto_output_node): Likewise.
19271 (input_node, input_edge): Likewise.
19272 * lto-streamer-out.c (output_cfg): Update streaming.
19273 * lto-streamer-in.c (input_cfg): Likewise.
19274 * data-streamer.h (streamer_write_gcov_count,
19275 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
19276 * gimple-streamer-in.c: Include value-prof.h
19277 (input_gimple_stmt): Input histograms.
19278 (input_bb): Update profile streaming.
19280 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
19282 * genmodes.c (emit_max_int): New function.
19283 (emit_insn_modes_h): Added call to emit_max_function.
19284 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
19286 * machmode.def: Fixed comment.
19288 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
19290 * combine.c (try_combine): Removed useless assert.
19291 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
19293 2013-03-28 Marek Polacek <polacek@redhat.com>
19294 Richard Biener <rguenther@suse.de>
19296 PR tree-optimization/56695
19297 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
19298 build signed result of a vector comparison.
19299 * tree-cfg.c (verify_gimple_comparison): Check that a result
19300 of a vector comparison has signed type.
19302 2013-03-28 Richard Biener <rguenther@suse.de>
19304 PR tree-optimization/37021
19305 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
19306 do not restrict gaps between groups.
19307 * tree-vect-stmts.c (vectorizable_load): Properly account for
19308 a gap between groups.
19310 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
19312 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
19313 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
19316 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
19318 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
19319 * doc/extend.texi (Named Address Spaces): Ditto.
19320 (Variable Attributes): Ditto.
19322 2013-03-27 Kai Tietz <ktietz@redhat.com>
19324 * config.build: Add support for cygwin x64 target.
19325 * config.gcc: Likewise.
19326 * config.host: Likewise.
19327 * configure.ac: Likewise
19328 * configure: Regenerated.
19330 2013-03-27 Kai Tietz <ktietz@redhat.com>
19332 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
19333 * config/i386/t-cygwin-w64: New file.
19334 * config/i386/cygwin-w64.h: New file.
19335 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
19336 and add support for x64-cygwin target.
19337 (CPP_SPEC): Likewise.
19338 (CXX_WRAP_SPEC_LIST): Undefine before define.
19339 (LIBGCJ_SONAME): Use 15 as version.
19341 2013-03-27 Richard Biener <rguenther@suse.de>
19343 PR tree-optimization/56716
19344 * tree-ssa-structalias.c (perform_var_substitution): Adjust
19345 dumping for ref nodes.
19347 2013-03-27 Martin Jambor <mjambor@suse.cz>
19349 PR tree-optimization/55334
19350 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
19351 restricted pointers to arrays.
19353 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
19355 * Makefile.in (.SUFFIXES): Add .cc.
19356 (.c.o): Apply same recipe for implicit rule .cc.o.
19358 2013-03-27 Richard Biener <rguenther@suse.de>
19360 PR tree-optimization/37021
19361 * tree-vect-data-refs.c (vect_check_strided_load): Allow
19362 REALPART/IMAGPART_EXPRs around the supported refs.
19363 * tree-ssa-structalias.c (find_func_aliases): Assume that
19364 floating-point values are not used to transfer pointers.
19366 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
19368 * target.def (TARGET_HAS_IFUNC_P): New target hook.
19369 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
19370 * doc/tm.texi: Regenerate.
19371 * targhooks.h (default_has_ifunc_p): New.
19372 * targhooks.c (default_has_ifunc_p): Ditto.
19373 * config/linux-protos.h: New file.
19374 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
19375 hook for linux which disables support of indirect functions in android.
19376 * config/linux-android.c: New file.
19377 * config/t-linux-android.c: Ditto.
19378 * config.gcc: Added new object file linux-android.o.
19379 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
19380 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
19381 * varasm.c (do_assemble_alias): Likewise.
19382 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
19383 doesn't support indirect functions.
19384 * configure: Regenerate.
19386 2013-03-27 Bin Cheng <bin.cheng@arm.com>
19389 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
19390 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
19393 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19395 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
19397 2013-03-27 Terry Guo <terry.guo@arm.com>
19399 * config/arm/arm-cores.def: Added core cortex-r7.
19400 * config/arm/arm-tune.md: Regenerated.
19401 * config/arm/arm-tables.opt: Regenerated.
19402 * doc/invoke.texi: Added entry for core cortex-r7.
19404 2013-03-27 Walter Lee <walt@tilera.com>
19406 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
19407 double-decrement of next_scratch_regno.
19409 2013-03-27 Walter Lee <walt@tilera.com>
19411 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
19413 (insn_v1mulus): Ditto.
19414 (insn_v2muls): Ditto.
19416 2013-03-27 Walter Lee <walt@tilera.com>
19418 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
19419 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
19421 2013-03-27 Walter Lee <walt@tilera.com>
19423 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
19424 (*sibcall_value): Ditto.
19426 2013-03-27 Walter Lee <walt@tilera.com>
19428 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
19429 (insn_mnz_v8qi): ... this ...
19430 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
19432 (insn_v<n>mnz): Replaced by ...
19433 (insn_v1mnz): ... this ...
19434 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
19436 (insn_mz_<mode>): Replaced by ...
19437 (insn_mz_v8qi): ... this ...
19438 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
19440 (insn_v<n>mz): Replaced by ...
19441 (insn_v1mz): ... this ...
19442 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
19445 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
19447 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
19449 2013-03-26 Roland McGrath <mcgrathr@google.com>
19451 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
19452 than fprintf with a non-constant, non-format string.
19454 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
19456 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
19457 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
19458 operand 0 predicate.
19459 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
19460 attribute. Use general_x64nomem_operand as operand 1 predicate.
19461 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
19462 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
19463 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
19464 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
19465 (mov<mode>_insv_1): Remove expander. Merge insn with
19466 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
19467 Use general_x64nomem_operand as operand 1 predicate.
19468 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
19469 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
19470 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
19471 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
19472 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
19473 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
19474 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
19475 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
19476 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
19477 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
19478 (general_x64nomem_operand): Ditto.
19480 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
19482 * config/rtems.opt: Add -pthread option.
19484 2013-03-26 Richard Biener <rguenther@suse.de>
19486 * alias.c (find_base_term): Avoid redundant and not used recursion.
19487 (base_alias_check): Get the initial base term from the caller.
19488 (true_dependence_1): Compute and pass base terms to base_alias_check.
19489 (write_dependence_p): Likewise.
19490 (may_alias_p): Likewise.
19492 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
19494 * config/aarch64/aarch64.c (aarch64_classify_address): Support
19495 PC-relative load in SI modes and above only.
19497 2013-03-26 Xinyu Qi <xyqi@marvell.com>
19499 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
19500 * config/arm/iwmmxt.md (WCGR0): Update.
19501 (WCGR1, WCGR2, WCGR3): Likewise.
19503 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
19505 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
19506 Use x64 and nox64 isa attributes.
19508 2013-03-26 Richard Biener <rguenther@suse.de>
19510 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
19511 alignment computations and rely on get_object_alignment_1
19512 for the !TYPE_P case.
19513 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
19515 2013-03-26 Walter Lee <walt@tilera.com>
19517 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
19518 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
19520 2013-03-25 Jeff Law <law@redhat.com>
19522 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
19523 check for INTEGRAL_TYPE_P that was missing due to checking in
19524 wrong version of prior patch.
19526 2013-03-25 Walter Lee <walt@tilera.com>
19528 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
19529 TILEGX_INSN_SHUFFLEBYTES1.
19530 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
19532 (tilegx_builtins): Ditto.
19533 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
19535 2013-03-25 Walter Lee <walt@tilera.com>
19537 * config/tilegx/tilegx.md (floatsisf2): New pattern.
19538 (floatunssisf2): New pattern.
19539 (floatsidf2): New pattern.
19540 (floatunssidf2): New pattern.
19542 2013-03-25 Walter Lee <walt@tilera.com>
19544 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
19545 tests for constraint J, K, N, P.
19547 2013-03-25 Walter Lee <walt@tilera.com>
19549 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
19550 Use indirect/pcrel encoding.
19551 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
19554 2013-03-25 Steve Ellcey <sellcey@mips.com>
19556 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
19557 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
19558 * config/mips/mips.c (mips_option_override): Set IMADD default.
19559 * config/mips/mips.h (PTF_AVOID_IMADD): New.
19560 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
19561 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
19562 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
19563 * doc/invoke.texi (-mimadd/-mno-imadd): New.
19565 2013-03-25 Jeff Law <law@redhat.com>
19567 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
19568 slightly to avoid creating and folding useless trees. Simplify
19569 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
19571 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
19573 * config/i386/i386.md (*zero_extendsidi2): Merge with
19574 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
19575 * config/i386/predicates.md (x86_64_zext_operand): Rename from
19576 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
19577 targets. Clarify comment.
19579 2013-03-25 Martin Jambor <mjambor@suse.cz>
19581 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
19582 pass-through jump functions differently.
19583 (ipa_read_jump_function): Likewise. Also use setter functions to set
19586 2013-03-25 Martin Jambor <mjambor@suse.cz>
19588 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
19589 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
19591 (ipa_get_indirect_edge_target): New function.
19592 (devirtualization_time_bonus): New parameter known_aggs, pass it to
19593 ipa_get_indirect_edge_target. Update all callers.
19594 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
19595 ipa_get_indirect_edge_target_1 instead of calling
19596 ipa_get_indirect_edge_target.
19597 (create_specialized_node): Pass aggvlas to
19598 ipcp_discover_new_direct_edges.
19600 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19602 * config/arm/arm.md (f_sels, f_seld): New types.
19603 (*cmov<mode>): New pattern.
19604 * config/arm/predicates.md (arm_vsel_comparison_operator): New
19607 2013-03-25 Kai Tietz <ktietz@redhat.com>
19609 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
19610 POSIX-printf for mingw-hosted builds.
19612 2013-03-25 Richard Biener <rguenther@suse.de>
19614 PR middle-end/56694
19615 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
19616 must-not-throw stmt location.
19618 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19620 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
19621 Emit load-acquire versions when acq is true.
19622 (arm_emit_store_exclusive): Add rel parameter.
19623 Emit store-release versions when rel is true.
19624 (arm_split_compare_and_swap): Use acquire-release instructions
19626 of barriers when appropriate.
19627 (arm_split_atomic_op): Likewise.
19628 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
19629 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
19630 (VUNSPEC_SLX): Likewise.
19631 (VUNSPEC_LDA): Likewise.
19632 (VUNSPEC_STL): Likewise.
19633 * config/arm/sync.md (atomic_load<mode>): New pattern.
19634 (atomic_store<mode>): Likewise.
19635 (arm_load_acquire_exclusive<mode>): Likewise.
19636 (arm_load_acquire_exclusivesi): Likewise.
19637 (arm_load_acquire_exclusivedi): Likewise.
19638 (arm_store_release_exclusive<mode>): Likewise.
19640 2013-03-25 Catherine Moore <clm@codesourcery.com>
19642 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
19643 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
19644 * config/mip/predicates.md (lwsp_swsp_operand,
19645 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
19646 sb16_operand, db4_operand, db7_operand, ib3_operand,
19647 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
19648 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
19649 andi16_operand): New predicates.
19650 * config/mips/mips.md (compression): New attribute.
19651 (enabled): New attribute.
19652 (length): Consider compression in computing length.
19653 (shift_compression): New code attribute.
19654 (*add<mode>3): New operands. Record compression.
19655 (sub<mode>3): Likewise.
19656 (one_cmpl<mode>2): Likewise.
19657 (*and<mode>3): Likewise.
19658 (*ior<mode>3): Likewise.
19659 (unnamed pattern for xor): Likewise.
19660 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
19661 (*<optab><mode>3): Likewise.
19662 (*mov<mode>_internal: Likewise.
19663 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
19664 (mips_unsigned_immediate_p): New.
19665 (umips_lwsp_swsp_address_p): New.
19666 (m16_based_address_p): New.
19667 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
19668 (mips_unsigned_immediate_p): New prototype.
19669 (lwsp_swsp_address_p): New prototype.
19670 (m16_based_address_p): New prototype.
19671 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
19672 (mips_signed_immediate_p): New function.
19673 (m16_based_address_p): New function.
19674 (lwsp_swsp_address_p): New function.
19675 (mips_print_operand_punctuation): Recognize short delay slot insns
19676 for microMIPS.add<mode>3"
19678 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19681 * config/arm/iterators.md (v_cmp_result): New mode attribute.
19682 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
19684 2013-03-25 Richard Biener <rguenther@suse.de>
19686 PR tree-optimization/56689
19687 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
19690 2013-03-25 Richard Biener <rguenther@suse.de>
19692 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
19694 (memory_references): Likewise.
19695 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
19696 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
19697 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
19698 (gather_mem_refs_in_loops): Fold into ...
19699 (analyze_memory_references): ... this. Move initialization
19700 to tree_ssa_lim_initialize.
19701 (fill_always_executed_in): Rename to ...
19702 (fill_always_executed_in_1): ... this.
19703 (fill_always_executed_in): Move contains_call computation to
19704 this new function from ...
19705 (tree_ssa_lim_initialize): ... here.
19706 (tree_ssa_lim): Call fill_always_executed_in.
19708 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
19710 * postreload.c (reload_combine): Fix code detecting returns.
19712 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
19714 * function.c (emit_use_return_register_into_block): On cc0 targets,
19715 do not emit the sequence between cc0 setter and user.
19717 2013-03-25 Kai Tietz <ktietz@redhat.com>
19719 * config/i386/predicates.md (local_symbolic_operand): Interpret
19720 dll-imported symbols as none-local.
19722 2013-03-25 Richard Biener <rguenther@suse.de>
19724 * tree-ssa-loop-im.c (struct depend): Remove.
19725 (struct lim_aux_data): Make depends a vec of gimples.
19726 (free_lim_aux_data): Adjust.
19727 (add_dependency): Likewise.
19728 (set_level): Likewise.
19730 2013-03-25 Richard Biener <rguenther@suse.de>
19732 PR middle-end/56434
19733 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
19734 the pointer returned by calls with ECF_MALLOC set.
19736 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
19738 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
19740 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
19742 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
19743 using MMXMODE mode iterator.
19744 (*move<mode>_internal): Merge with *movv2sf_internal and
19745 *movv2sf_internal_rex64 using MMXMODE mode iterator.
19747 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
19749 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
19750 (record_last_mem_set_info): Likewise.
19752 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
19753 of XNEWVEC followed by memset.
19754 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
19756 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
19758 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
19759 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
19760 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
19761 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
19762 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
19763 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
19764 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
19765 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
19766 BARRIER_P instead of GET_CODE.
19768 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
19770 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
19771 inaccuracy in the probing code.
19773 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
19774 (ctrapdi4): Likewise.
19776 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
19778 * calls.c (expand_call): Add missing guard to code handling return
19779 of non-BLKmode structures in MSB.
19780 * function.c (expand_function_end): Likewise.
19782 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
19784 * combine.c (try_combine): Adjust comment. Do not add the set of
19785 insn #0 if the destination indirectly is set or dies in insn #2.
19786 Tidy up code to distribute a new note.
19788 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
19790 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
19791 also for alternatives 16 and 17.
19793 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
19795 * config/i386/sse.md (*mov<mode>_internal): Merge with
19796 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
19797 Emit insn template depending on type attribute. Use
19798 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
19799 movd instead of movq mnemonic for interunit moves. Rewrite mode
19800 attribute calculation. Remove unit attribute calculation.
19801 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
19802 Set prefix_data16 attribute for DImode ssemov types.
19803 Use Ym instead of y for SSE-MMX conversion alternatives.
19804 Reorder operand constraints.
19806 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
19808 * df.h (df_insn_delete): Adjust prototype.
19809 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
19810 and let it decide whether mark the basic block dirty.
19811 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
19812 * df-scan.c (df_insn_info_delete): New helper function, split
19813 off from df_insn_delete.
19814 (df_scan_free_bb_info): Use it.
19815 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
19817 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
19818 that the insn is actually an insn and it has a non-NULL basic block.
19819 Do not mark basic block dirty if only deleting a DEBUG_INSN.
19821 2013-03-22 Richard Biener <rguenther@suse.de>
19823 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
19825 (mem_ref_alloc): Do not allocate them.
19826 (refs_independent_p): Do not query or maintain a cache.
19828 2013-03-22 Richard Biener <rguenther@suse.de>
19830 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
19831 (gather_mem_refs_in_loops): Do not compute it.
19832 (analyze_memory_references): Do not allocate it.
19833 (tree_ssa_lim_finalize): Do not free it.
19834 (for_all_locs_in_loop): Do not query all_refs_in_loop.
19836 2013-03-22 Richard Biener <rguenther@suse.de>
19838 * is-a.h (as_a): Use gcc_checking_assert.
19840 2013-03-22 Ian Bolton <ian.bolton@arm.com>
19842 * config/aarch64/aarch64.c (aarch64_print_operand): New
19843 format specifier for printing a constant in hex.
19844 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
19845 format specifier for printing second operand.
19847 2013-03-22 Richard Biener <rguenther@suse.de>
19849 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
19851 (gather_mem_refs_in_loops): Perform store accumulation here.
19852 (create_vop_ref_mapping_loop): Remove.
19853 (create_vop_ref_mapping): Likewise.
19854 (analyze_memory_references): Initialize refs_stored_in_loop.
19855 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
19856 (record_indep_loop): Remove.
19857 (record_dep_loop): New function.
19858 (ref_indep_loop_p_1): Adjust to only walk over references
19859 in the loop, not its subloops.
19860 (ref_indep_loop_p): Rename to ...
19861 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
19862 maintaining a more fine-grained cache.
19863 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
19864 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
19866 2013-03-22 Richard Biener <rguenther@suse.de>
19868 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
19869 (struct mem_ref): Make accesses_in_loop a vec of a vec of
19870 aggregate mem_ref_loc.
19871 (free_mem_ref_locs): Inline into ...
19872 (memref_free): ... this and adjust.
19873 (mem_ref_alloc): Adjust.
19874 (mem_ref_locs_alloc): Remove.
19875 (record_mem_ref_loc): Adjust.
19876 (get_all_locs_in_loop): Rewrite into ...
19877 (for_all_locs_in_loop): ... this iterator.
19878 (rewrite_mem_ref_loc): New functor.
19879 (rewrite_mem_refs): Use for_all_locs_in_loop.
19880 (sm_set_flag_if_changed): New functor.
19881 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
19882 (ref_always_accessed): New functor.
19883 (ref_always_accessed_p): Use for_all_locs_in_loop.
19885 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
19887 * tree-pass.h (PROP_gimple_lvec): New.
19888 * passes.c (dump_properties): Handle PROP_gimple_lvec.
19889 (init_optimization_passes): Move pass_lower_vector.
19890 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
19892 (pass_lower_vector): Provide PROP_gimple_lvec.
19893 (pass_lower_vector_ssa): Likewise.
19894 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
19896 2013-03-21 Mark Wielaard <mjw@redhat.com>
19898 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
19900 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
19902 * config/i386/i386.md (*movdi_internal): Disparage slightly
19903 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
19904 conversion alternatives.
19906 2013-03-21 Jakub Jelinek <jakub@redhat.com>
19908 PR middle-end/48087
19909 * diagnostic.def (DK_WERROR): New kind.
19910 * diagnostic.h (werrorcount): Define.
19911 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
19912 promoted to DK_ERROR, increment DK_WERROR counter instead of
19914 * toplev.c (toplev_main): Call print_ignored_options even if
19915 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
19916 even if just werrorcount is non-zero.
19919 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
19921 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
19922 (string_cst_pool_decl): New function.
19923 (optimize_one_addr_into_implicit_ptr): New function.
19924 (resolve_addr_in_expr): Optimize DWARF location expression
19925 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
19926 which doesn't live in memory, but has DW_AT_location or
19927 DW_AT_const_value, or refers to a string literal, into
19928 DW_OP_GNU_implicit_pointer.
19929 (optimize_location_into_implicit_ptr): New function.
19930 (resolve_addr): If removing DW_AT_location of a variable because
19931 it was DW_OP_addr of address of the variable, but the variable doesn't
19932 live in memory, try to emit const value attribute for the initializer.
19934 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
19936 * tree.h (VECTOR_TYPE_P): New macro.
19937 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
19938 TYPE_MODE): Use it.
19939 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
19940 VEC_COND_EXPR cannot be lvalues.
19941 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
19943 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
19945 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
19946 Restrict the transformation to equal modes.
19948 2013-03-21 Richard Biener <rguenther@suse.de>
19950 PR tree-optimization/39326
19951 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
19952 (MEM_ANALYZABLE): Adjust.
19953 (record_mem_ref_loc): Move bitmap ops ...
19954 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
19955 unanalyzable refs, do not record locations for it.
19956 (analyze_memory_references): Allocate ref zero as shared
19958 (refs_independent_p): Do not test for unanalyzed mems here.
19959 (ref_indep_loop_p_1): Special-case disambiguation against
19960 the unanalyzed ref.
19961 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
19963 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
19965 * config/arm/arm-protos.h (tune_params): Add
19966 prefer_neon_for_64bits field.
19967 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
19968 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
19969 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
19970 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
19971 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
19972 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
19973 (arm_option_override): Handle -mneon-for-64bits new option.
19974 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
19975 (prefer_neon_for_64bits): Declare new variable.
19976 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
19977 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
19978 (arch_enabled): Handle new arch types. Remove support for onlya8
19980 (one_cmpldi2): Use new arch names.
19981 (zero_extend<mode>di2, extend<mode>di2): Ditto.
19982 * config/arm/arm.opt (mneon-for-64bits): Add option.
19983 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
19984 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
19985 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
19987 * doc/invoke.texi (-mneon-for-64bits): Document.
19989 2013-03-21 Richard Biener <rguenther@suse.de>
19991 PR tree-optimization/39326
19992 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
19993 (sort_bbs_in_loop_postorder_cmp): New function.
19994 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
19996 2013-03-21 Richard Biener <rguenther@suse.de>
19998 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
19999 (vect_insert_into_interleaving_chain): Likewise.
20000 (vect_drs_dependent_in_basic_block): Inline ...
20001 (vect_slp_analyze_data_ref_dependence): ... here. New function,
20003 (vect_analyze_data_ref_dependence): ... here. Simplify.
20004 (vect_check_interleaving): Simplify.
20005 (vect_analyze_data_ref_dependences): Likewise. Split out ...
20006 (vect_slp_analyze_data_ref_dependences): ... this new function.
20007 (dr_group_sort_cmp): New function.
20008 (vect_analyze_data_ref_accesses): Compute data-reference groups
20009 here instead of in vect_analyze_data_ref_dependence. Use
20010 a more efficient algorithm.
20011 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
20012 vect_slp_analyze_data_ref_dependences. Call
20013 vect_analyze_data_ref_accesses earlier.
20014 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
20015 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
20016 (vect_slp_analyze_data_ref_dependences): New prototype.
20018 2013-03-21 Richard Biener <rguenther@suse.de>
20020 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
20021 ref is stored in the loop.
20022 (find_refs_for_sm): Walk only over all stores.
20023 (store_motion_loop): Allocate from lim_bitmap_obstack.
20024 (store_motion): Likewise.
20026 2013-03-21 Richard Biener <rguenther@suse.de>
20028 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
20029 Update virtual SSA form.
20031 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20033 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
20034 * configure: Regenerate.
20035 * config.in: Regenerate.
20036 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
20037 if !HAVE_LD_EH_FRAME_CIEV3.
20039 2013-03-21 Richard Biener <rguenther@suse.de>
20041 * tree-cfg.c (verify_expr_no_block): New function.
20042 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
20043 nor DECL_VALUE_EXPR have locations with associated blocks.
20044 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
20045 (clear_unused_block_pointer): Remove code dealing with
20046 blocks in DECL_DEBUG_EXPR locations.
20048 2013-03-21 Richard Biener <rguenther@suse.de>
20050 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
20051 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
20052 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
20053 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
20054 instead of DECL_DEBUG_EXPR_IS_FROM.
20055 * gimplify.c (gimplify_modify_expr): Likewise.
20056 * tree-cfg.c (verify_expr_location_1): Likewise.
20057 * tree-complex.c (create_one_component_var): Likewise.
20058 * tree-sra.c (create_access_replacement): Likewise.
20059 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
20060 (clear_unused_block_pointer): Likewise.
20061 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
20062 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
20063 * var-tracking.c (var_debug_decl): Likewise.
20064 (track_expr_p): Likewise.
20065 * tree-inline.c (add_local_variables): Likewise. Set
20066 DECL_HAS_DEBUG_EXPR_P after copying it.
20067 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
20068 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
20070 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
20073 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
20074 * configure: Regenerate.
20075 * config.in: Regenerate.
20076 * config/i386/i386.md (*movdf_internal): Use
20077 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
20078 movd instead of movq mnemonic for interunit moves.
20079 (*movdi_internal): Ditto.
20081 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20083 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
20084 (abd<mode>_3): New pattern.
20085 (aba<mode>_3): New pattern.
20086 (fabd<mode>_3): New pattern.
20088 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20090 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
20091 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
20092 occurrence of REGISTER_PREFIX as its empty string.
20094 2013-03-20 Jeff Law <law@redhat.com>
20096 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
20097 addititional equivalences for equality comparisons between an SSA_NAME
20098 and a constant where the SSA_NAME was set from a widening conversion.
20100 2013-03-20 Walter Lee <walt@tilera.com>
20102 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
20104 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
20106 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
20107 depending on type attribute.
20108 (*movti_internal): Ditto.
20109 (*movtf_internal): Ditto.
20110 (*movxf_internal): Ditto.
20111 (*movdf_internal): Ditto.
20112 (*movsf_internal): Ditto.
20114 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
20116 * config/i386/i386.md (*movti_internal): Set prefix attribute to
20117 maybe_vex for sselog1 and ssemov types.
20118 (*movdi_internal): Reorder operand constraints.
20119 (*movsi_internal): Ditto. Set prefix attribute to
20120 maybe_vex for sselog1 and ssemov types.
20121 (*movtf_internal): Set prefix attribute to maybe_vex
20122 for sselog1 and ssemov types.
20123 (*movdf_internal): Ditto. Set prefix_data16 attribute for
20124 DImode ssemov types. Reorder operand constraints.
20125 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
20126 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
20127 attribute for SImode ssemov types. Reorder operand constraints.
20129 2013-03-20 Martin Jambor <mjambor@suse.cz>
20131 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
20132 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
20134 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
20136 * config/rs6000/predicates.md (indexed_address, update_address_mem
20137 update_indexed_address_mem): New predicates.
20138 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
20139 attribute for load/store instructions.
20140 * config/rs6000/dfp.md (movsd_store): Likewise.
20141 (movsd_load): Likewise.
20142 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
20143 (unnamed HI->DI extend define_insn): Likewise.
20144 (unnamed SI->DI extend define_insn): Likewise.
20145 (unnamed QI->SI extend define_insn): Likewise.
20146 (unnamed QI->HI extend define_insn): Likewise.
20147 (unnamed HI->SI extend define_insn): Likewise.
20148 (unnamed HI->SI extend define_insn): Likewise.
20149 (extendsfdf2_fpr): Likewise.
20150 (movsi_internal1): Likewise.
20151 (movsi_internal1_single): Likewise.
20152 (movhi_internal): Likewise.
20153 (movqi_internal): Likewise.
20154 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
20155 attribute for load/store instructions.
20156 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
20158 (mov<mode>_softfloat): Likewise.
20159 (mov<mode>_hardfloat32): Likewise.
20160 (mov<mode>_hardfloat64): Likewise.
20161 (mov<mode>_softfloat64): Likewise.
20162 (movdi_internal32): Likewise.
20163 (movdi_internal64): Likewise.
20164 (probe_stack_<mode>): Likewise.
20166 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
20168 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
20169 floating point, and decimal floating point to reload iterator.
20171 * config/rs6000/constraints.md (wl constraint): New constraints to
20172 return FLOAT_REGS if certain options are used to reduce the number
20173 of separate patterns that exist in the file.
20174 (wx constraint): Likewise.
20175 (wz constraint): Likewise.
20177 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
20178 -mdebug=reg, print wg, wl, wx, and wz constraints.
20179 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
20180 Initialize the reload functions for 64-bit binary/decimal floating
20182 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
20183 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
20184 create the buffer on the stack to overcome not having a 32-bit
20186 (rs6000_emit_move): Likewise.
20187 (rs6000_secondary_memory_needed_rtx): Likewise.
20188 (rs6000_alloc_sdmode_stack_slot): Likewise.
20189 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
20190 via xxlxor, just like DFmode 0.0.
20192 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
20193 define as 1 if we are running on a power7 or newer.
20194 (enum r6000_reg_class_enum): Add new constraints.
20196 * config/rs6000/dfp.md (movsd): Delete, combine with binary
20197 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
20198 with other moves by using conditional constraits (wg). Use LFIWZX
20199 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
20200 (movsd splitter): Likewise.
20201 (movsd_hardfloat): Likewise.
20202 (movsd_softfloat): Likewise.
20204 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
20205 binary and decimal floating point moves.
20206 (fmove_ok): New attributes to combine binary and decimal floating
20207 point moves, and to combine power6x (mfpgpr) moves along normal
20209 (real_value_to_target): Likewise.
20210 (f32_lr): Likewise.
20211 (f32_lm): Likewise.
20212 (f32_li): Likewise.
20213 (f32_sr): Likewise.
20214 (f32_sm): Likewise.
20215 (f32_si): Likewise.
20216 (movsf): Combine binary and decimal floating point moves. Combine
20217 power6x (mfpgpr) moves with other moves by using conditional
20218 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
20219 (mov<mode> for SFmode/SDmode); Likewise.
20220 (SFmode/SDmode splitters): Likewise.
20221 (movsf_hardfloat): Likewise.
20222 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
20223 (movsf_softfloat): Likewise.
20224 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
20226 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
20227 wx and wz constraints.
20229 * config/rs6000/constraints.md (wg constraint): New constraint to
20230 return FLOAT_REGS if -mmfpgpr (power6x) was used.
20232 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
20235 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
20236 -mdebug=reg, print wg, wl, wx, and wz constraints.
20237 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
20238 Initialize the reload functions for 64-bit binary/decimal floating
20240 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
20241 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
20242 create the buffer on the stack to overcome not having a 32-bit
20244 (rs6000_emit_move): Likewise.
20245 (rs6000_secondary_memory_needed_rtx): Likewise.
20246 (rs6000_alloc_sdmode_stack_slot): Likewise.
20247 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
20248 via xxlxor, just like DFmode 0.0.
20250 * config/rs6000/dfp.md (movdd): Delete, combine with binary
20251 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
20252 with other moves by using conditional constraits (wg). Use LFIWZX
20253 and STFIWX for loading SDmode on power7.
20254 (movdd splitters): Likewise.
20255 (movdd_hardfloat32): Likewise.
20256 (movdd_softfloat32): Likewise.
20257 (movdd_hardfloat64_mfpgpr): Likewise.
20258 (movdd_hardfloat64): Likewise.
20259 (movdd_softfloat64): Likewise.
20261 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
20262 64-bit binary and decimal floating point moves.
20263 (FMOVE64X): Likewise.
20264 (movdf): Combine 64-bit binary and decimal floating point moves.
20265 Combine power6x (mfpgpr) moves with other moves by using
20266 conditional constraits (wg).
20267 (mov<mode> for DFmode/DDmode): Likewise.
20268 (DFmode/DDmode splitters): Likewise.
20269 (movdf_hardfloat32): Likewise.
20270 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
20271 (movdf_softfloat32): Likewise.
20272 (movdf_hardfloat64_mfpgpr): Likewise.
20273 (movdf_hardfloat64): Likewise.
20274 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
20275 (movdf_softfloat64): Likewise.
20276 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
20277 (reload_<mode>_load): Move to later in the file so they aren't in
20278 the middle of the floating point move insns.
20279 (reload_<mode>_store): Likewise.
20281 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
20284 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
20285 constraint if -mdebug=reg.
20286 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
20287 Enable using dd reload support if needed.
20289 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
20290 binary and decimal floating point moves in rs6000.md.
20291 (movtd_internal): Likewise.
20293 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
20294 decimal floating point moves.
20296 (movtf_internal): Likewise.
20297 (mov<mode>_internal, TDmode/TFmode): Likewise.
20298 (movtf_softfloat): Likewise.
20299 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
20301 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
20302 movdi_internal64, using wg constraint for move direct operations.
20303 (movdi_internal64): Likewise.
20305 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
20306 MODES_TIEABLE_P for selected modes. Print the numerical value of
20307 the various virtual registers. Use GPR/FPR first/last values,
20308 instead of hard coding the register numbers. Print which modes
20309 have reload functions registered.
20310 (rs6000_option_override_internal): If -mdebug=reg, trace the options
20311 settings before/after setting cpu, target and subtarget settings.
20312 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
20313 and for secondary reload failures in rs6000_secondary_reload_inner.
20314 (rs6000_secondary_reload_fail): Likewise.
20315 (rs6000_secondary_reload_inner): Likewise.
20317 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
20318 macros for first/last GPR and FPR registers.
20319 (LAST_GPR_REGNO): Likewise.
20320 (FIRST_FPR_REGNO): Likewise.
20321 (LAST_FPR_REGNO): Likewise.
20323 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
20324 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
20325 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
20326 (vcond<mode><mode>): Likewise.
20327 (vcondu<mode><mode>): Likewise.
20328 (vector_gtu<mode>): Likewise.
20329 (vector_gte<mode>): Likewise.
20330 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
20331 to prevent the compiler from converting DImode operations to TImode.
20332 (ior<mode>3): Likewise.
20333 (and<mode>3): Likewise.
20334 (one_cmpl<mode>2): Likewise.
20335 (nor<mode>3): Likewise.
20336 (andc<mode>3): Likewise.
20338 * config/rs6000/constraints.md (wt constraint): New constraint
20339 that returns VSX_REGS if TImode is allowed in VSX registers.
20341 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
20342 constant under VSX.
20344 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
20345 similar to TImode, but it is restricted to being in the GPRs.
20347 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
20348 TImode to occupy a single VSX register.
20350 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
20351 -mvsx-timode for power7/power8.
20352 (power7 cpu): Likewise.
20353 (power8 cpu): Likewise.
20355 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
20356 sure that TFmode/TDmode take up two registers if they are ever
20357 allowed in the upper VSX registers.
20358 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
20360 (rs6000_init_hard_regno_mode_ok): Likewise.
20361 (rs6000_debug_reg_global): Add debugging for PTImode and wt
20362 constraint. Print if LRA is turned on.
20363 (rs6000_option_override_internal): Give an error if -mvsx-timode
20364 and VSX is not enabled.
20365 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
20366 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
20367 to reg+offset addressing. Use PTImode when checking offset
20368 addresses for validity.
20369 (reg_offset_addressing_ok_p): Likewise.
20370 (rs6000_legitimate_offset_address_p): Likewise.
20371 (rs6000_legitimize_address): Likewise.
20372 (rs6000_legitimize_reload_address): Likewise.
20373 (rs6000_legitimate_address_p): Likewise.
20374 (rs6000_eliminate_indexed_memrefs): Likewise.
20375 (rs6000_emit_move): Likewise.
20376 (rs6000_secondary_reload): Likewise.
20377 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
20378 reloads to fpr registers to continue to use reg+offset addressing,
20379 but 64-bit reloads to altivec registers need reg+reg addressing.
20380 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
20381 it. Treat LO_SUM like a PLUS operation.
20382 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
20383 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
20384 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
20385 registers to share a register with a smaller sized type, since VSX
20386 puts scalars in the upper 64-bits.
20387 (print_operand): Add support for PTImode.
20388 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
20389 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
20390 registers, but don't have arithmetic support.
20391 (rs6000_memory_move_cost): Add test for VSX.
20392 (rs6000_opt_masks): Add -mvsx-timode.
20394 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
20397 (VSr): Use wt constraint for TImode.
20398 (VSv): Drop TImode support.
20399 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
20400 (vsx_movti_64bit): Likewise.
20401 (vsx_movti_32bit): Likewise.
20402 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
20403 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
20404 one '?' on the appropriate output constraint. Do not allow TImode
20405 logical operations on 32-bit systems.
20406 (vsx_ior<mode>3): Likewise.
20407 (vsx_xor<mode>3): Likewise.
20408 (vsx_one_cmpl<mode>2): Likewise.
20409 (vsx_nor<mode>3): Likewise.
20410 (vsx_andc<mode>3): Likewise.
20411 (vsx_concat_<mode>): Likewise.
20412 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
20414 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
20415 OPTION_MASK_VSX_TIMODE.
20416 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
20417 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
20419 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
20420 (TI2 iterator): New iterator for TImode, PTImode.
20421 (wd mode attribute): Add values for vector types.
20422 (movti_string): Replace TI move operations with operations for TImode
20423 and PTImode. Add support for TImode being allowed in VSX registers.
20424 (mov<mode>_string, TImode/PTImode): Likewise.
20425 (movti_ppc64): Likewise.
20426 (mov<mode>_ppc64, TImode/PTImode): Likewise.
20427 (TI mode splitters): Likewise.
20429 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
20432 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
20434 PR tree-optimization/56355
20435 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
20436 Also handle integers with undefined overflow.
20438 2013-03-20 Catherine Moore <clm@codesourcery.com>
20439 Maciej W. Rozycki <macro@codesourcery.com>
20440 Tom de Vries <tom@codesourcery.com>
20441 Nathan Sidwell <nathan@codesourcery.com>
20442 Iain Sandoe <iain@codesourcery.com>
20443 Nathan Froyd <froydnj@codesourcery.com>
20444 Chao-ying Fu <fu@mips.com>
20446 * doc/extend.texi: (micromips, nomicromips, nocompression):
20447 Document new function attributes.
20448 * doc/invoke.texi (minterlink-compressed, mmicromips,
20449 m14k, m14ke, m14kec): Document new options.
20450 (minterlink-mips16): Update documentation.
20451 * doc/md.texi (ZC, ZD): Document new constraints.
20452 * configure.ac (gcc_cv_as_micromips): Check if linker
20453 supports the .set micromips directive.
20454 * configure: Regenerate.
20455 * config.in: Regenerate.
20456 * config/mips/mips-tables.opt: Regenerate.
20457 * config/mips/micromips.md: New file.
20458 * constraints.md (ZC, ZD): New constraints.
20459 * config/mips/predicates.md (movep_src_register): New predicate.
20460 (movep_src_operand): New predicate.
20461 (non_volatile_mem_operand): New predicate.
20462 * config/mips/mips.md (multimem): New type.
20463 (length): Differentiate between 17-bit and 18-bit branch offsets.
20464 (MOVEP1, MOVEP2): New mode iterator.
20465 (mov_<load>l): Use ZC constraint.
20466 (mov_<load>r): Likewise.
20467 (mov_<store>l): Likewise.
20468 (mov_<store>r): Likewise.
20469 (*branch_equality<mode>_inverted): Add microMIPS support.
20470 (*branch_equality<mode>): Likewise.
20471 (*jump_absolute): Likewise.
20472 (indirect_jump_<mode>): Likewise.
20473 (tablejump_<mode>): Likewise.
20474 (<optab>_internal): Likewise.
20475 (sibcall_internal): Likewise.
20476 (sibcall_value_internal): Likewise.
20477 (prefetch): Use constraint ZD.
20478 * config/mips/mips.opt (minterlink-compressed): New option.
20479 (minterlink-mips16): Now an alias for minterlink-compressed.
20480 (mmicromips): New option.
20481 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
20482 (compare_and_swap_12): Likewise.
20483 (sync_add<mode>): Likewise.
20484 (sync_<optab>_12): Likewise.
20485 (sync_old_<optab>_12): Likewise.
20486 (sync_new_<optab>_12): Likewise.
20487 (sync_nand_12): Likewise.
20488 (sync_old_nand_12): Likewise.
20489 (sync_new_nand_12): Likewise.
20490 (sync_sub<mode>): Likewise.
20491 (sync_old_add<mode>): Likewise.
20492 (sync_old_sub<mode>): Likewise.
20493 (sync_new_add<mode>): Likewise.
20494 (sync_new_sub<mode>): Likewise.
20495 (sync_<optab><mode>): Likewise.
20496 (sync_old_<optab><mode>): Likewise.
20497 (sync_new_<optab><mode>): Likewise.
20498 (sync_nand<mode>): Likewise.
20499 (sync_old_nand<mode>): Likewise.
20500 (sync_new_nand<mode>): Likewise.
20501 (sync_lock_test_and_set<mode>): Likewise.
20502 (test_and_set_12): Likewise.
20503 (atomic_compare_and_swap<mode>): Likewise.
20504 (atomic_exchange<mode>_llsc): Likewise.
20505 (atomic_fetch_add<mode>_llsc): Likewise.
20506 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
20507 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
20508 (umips_save_restore_pattern_p): Likewise.
20509 (umips_load_store_pair_p): Likewise.
20510 (umips_output_load_store_pair): Likewise.
20511 (umips_movep_target_p): Likewise.
20512 (umips_12bit_offset_address_p): Likewise.
20513 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
20514 (mips_base_mips16): Rename this...
20515 (mips_base_compression_flags): ...to this. Update all uses.
20516 (mips_attribute_table): Add micromips, nomicromips and nocompression.
20517 (mips_mips16_decl_p): Delete.
20518 (mips_nomips16_decl_p): Delete.
20519 (mips_get_compress_on_flags): New function.
20520 (mips_get_compress_off_flags): New function.
20521 (mips_get_compress_mode): New function.
20522 (mips_get_compress_on_name): New function.
20523 (mips_get_compress_off_name): New function.
20524 (mips_insert_attributes): Support multiple compression types.
20525 (mips_merge_decl_attributes): Likewise.
20526 (umips_12bit_offset_address_p): New function.
20527 (mips_start_function_definition): Emit .set micromips directive.
20528 (mips_call_may_need_jalx_p): New function.
20529 (mips_function_ok_for_sibcall): Add microMIPS support.
20530 (mips_print_operand_punctuation): Support short delay slots and
20532 (umips_swm_mask, umips_swm_encoding): New.
20533 (umips_build_save_restore): New function.
20534 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
20535 (was_mips16_p): Remove.
20536 (old_compression_mode): New.
20537 (mips_set_compression_mode): New function.
20538 (mips_set_current_function): Add microMIPS support.
20539 (mips_option_override): Likewise.
20540 (umips_save_restore_pattern_p): New function.
20541 (umips_output_save_restore): New function.
20542 (umips_load_store_pair_p_1): New function.
20543 (umips_load_store_pair_p): New function.
20544 (umips_output_load_store_pair_1): New function.
20545 (umips_output_load_store_pair): New function.
20546 (umips_movep_target_p) New function.
20547 (mips_prepare_pch_save): Add microMIPS support.
20548 * config/mips/mips.h (TARGET_COMPRESSION): New.
20549 (TARGET_CPU_CPP_BUILTINS): Update macro
20550 to use new compression flags and to support microMIPS.
20551 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
20552 (MIPS_ARCH_FLOAT_SPEC): Likewise.
20553 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
20554 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
20555 (ASM_SPEC): Support mmicromips and mno-micromips.
20556 (M16STORE_REG_P): New macro.
20557 (MIPS_CALL): Support TARGET_MICROMIPS.
20558 (MICROMIPS_J): New macro.
20559 (mips_base_mips16): Rename this...
20560 (mips_base_compression_flags): ...to this.
20561 (UMIPS_12BIT_OFFSET_P): New macro.
20562 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
20563 (MULTILIB_DIRNAMES): Likewise.
20564 2013-03-20 Richard Biener <rguenther@suse.de>
20566 PR tree-optimization/56661
20567 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
20568 the result does not have to be distinct.
20570 2013-03-20 Richard Biener <rguenther@suse.de>
20572 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
20575 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20576 Steven Bosscher <steven@gcc.gnu.org>
20578 PR rtl-optimization/56605
20579 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
20581 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
20584 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
20585 that require movd instead of movq.
20587 2013-03-20 Richard Biener <rguenther@suse.de>
20589 * tree-ssa-structalias.c (struct variable_info): Add pointer
20590 to the first field of an aggregate with sub-vars. Make
20591 this and the pointer to the next subfield its ID.
20592 (vi_next): New function.
20593 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
20594 storedanything_id, integer_id): Increment by one.
20595 (new_var_info, get_call_vi, lookup_call_clobber_vi,
20596 get_call_clobber_vi): Adjust.
20597 (solution_set_expand): Simplify and speedup.
20598 (solution_set_add): Inline into ...
20599 (set_union_with_increment): ... this. Adjust accordingly.
20600 (do_sd_constraint): Likewise.
20601 (do_ds_constraint): Likewise.
20602 (do_complex_constraint): Simplify.
20603 (build_pred_graph): Adjust.
20604 (solve_graph): Likewise. Simplify and speedup.
20605 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
20606 get_constraint_for_component_ref, get_constraint_for_1,
20607 first_vi_for_offset, first_or_preceding_vi_for_offset,
20608 create_function_info_for, create_variable_info_for_1,
20609 create_variable_info_for, intra_create_variable_infos): Adjust.
20610 (init_base_vars): Push NULL for ID zero.
20611 (compute_points_to_sets): Adjust.
20613 2013-03-20 Richard Biener <rguenther@suse.de>
20615 * cfgloop.c (verify_loop_structure): Streamline and avoid
20616 ICEing on corrupt loop tree.
20617 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
20620 2013-03-20 Richard Biener <rguenther@suse.de>
20622 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
20623 check whether an SSA update is needed.
20625 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
20627 * config/mips/constraints.md (T): Rename to...
20631 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
20632 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
20634 2013-03-19 Ian Bolton <ian.bolton@arm.com>
20636 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
20637 (*subsi3_carryin_uxtw): Likewise.
20639 2013-03-19 Ian Bolton <ian.bolton@arm.com>
20641 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
20642 (*rorsi3_insn_uxtw): Likewise.
20644 2013-03-19 Ian Bolton <ian.bolton@arm.com>
20646 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
20647 (*extrsi5_insn_uxtw): Likewise.
20649 2013-03-19 Richard Biener <rguenther@suse.de>
20651 PR tree-optimization/56273
20652 * passes.c (init_optimization_passes): Move second VRP after DOM.
20654 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
20656 * config/i386/i386.md (*movti_internal): Merge from
20657 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
20658 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
20659 nox64 isa attributes.
20661 2013-03-18 Richard Biener <rguenther@suse.de>
20663 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
20665 (merge_node_constraints): Likewise.
20666 (build_succ_graph): Likewise.
20667 (valid_graph_edge): Inline into single caller.
20668 (unify_nodes): Likewise. Use bitmap_set_bit return value
20670 (scc_visit): Fix formatting and variable use.
20671 (do_sd_constraint): Use gcc_checking_assert.
20672 (do_ds_constraint): Likewise.
20673 (do_complex_constraint): Likewise.
20674 (condense_visit): Likewise. Cleanup.
20675 (dump_pred_graph): New function.
20676 (perform_var_substitution): Dump the pred-graph before
20677 variable substitution.
20678 (find_equivalent_node): Use gcc_checking_assert.
20679 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
20681 2013-03-18 Richard Biener <rguenther@suse.de>
20683 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
20684 Remove cond_expr_stmt_list argument and do not gimplify the
20686 (vect_loop_versioning): Adjust.
20687 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
20688 Cleanup to use less temporaries.
20689 (vect_create_data_ref_ptr): Cleanup.
20691 2013-03-18 Jakub Jelinek <jakub@redhat.com>
20693 PR tree-optimization/56635
20694 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
20695 require types_compatible_p types.
20697 2013-03-18 Nick Clifton <nickc@redhat.com>
20699 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
20700 spurious backslash.
20702 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
20703 Add missing line to comment describing function.
20705 2013-03-18 Richard Biener <rguenther@suse.de>
20707 PR tree-optimization/56210
20708 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
20709 Handle string / character search functions.
20710 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
20712 2013-03-18 Richard Biener <rguenther@suse.de>
20714 PR middle-end/56483
20715 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
20716 and implement properly.
20717 * gimple.h (gimple_cond_single_var_p): Remove.
20719 2013-03-18 Richard Biener <rguenther@suse.de>
20721 * tree-data-ref.h (find_data_references_in_loop): Declare.
20722 * tree-data-ref.c (get_references_in_stmt): Use a stack
20723 vector pre-allocated in the callers.
20724 (find_data_references_in_stmt): Adjust.
20725 (graphite_find_data_references_in_stmt): Likewise.
20726 (create_rdg_vertices): Likewise.
20727 (find_data_references_in_loop): Export.
20728 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
20729 Compute dependences here...
20730 (vect_analyze_data_refs): ...not here. When we encounter
20731 a non-vectorizable data reference in basic-block vectorization
20732 truncate the data reference vector. Do not bother to
20733 fixup data-dependence information for gather loads.
20734 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
20735 of data references, as reported.
20737 2013-03-18 Richard Biener <rguenther@suse.de>
20739 PR tree-optimization/3713
20740 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
20741 has_constants and expr.
20742 (stmt_has_constants): Properly valueize SSA names when deciding
20743 whether the stmt has constants.
20745 2013-03-18 Richard Biener <rguenther@suse.de>
20747 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
20748 whole function when there is nothing to do.
20749 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
20750 * tree-vectorizer.c (vectorize_loops): Update virtual and
20751 loop-closed SSA once.
20752 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
20754 2013-03-18 Richard Biener <rguenther@suse.de>
20756 PR middle-end/56113
20757 * domwalk.c (bb_postorder): New global static.
20758 (cmp_bb_postorder): New function.
20759 (walk_dominator_tree): Replace scheme imposing an order for
20760 visiting dominator sons by one sorting them at the time they
20761 are pushed on the stack.
20763 2013-03-18 Richard Biener <rguenther@suse.de>
20765 PR tree-optimization/39326
20766 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
20767 (struct mem_ref): Replace mem member with ao_ref typed member.
20768 (MEM_ANALYZABLE): Adjust.
20769 (memref_eq): Likewise.
20770 (mem_ref_alloc): Likewise.
20771 (gather_mem_refs_stmt): Likewise.
20772 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
20773 (execute_sm_if_changed_flag_set): Adjust.
20774 (execute_sm): Likewise.
20775 (ref_always_accessed_p): Likewise.
20776 (refs_independent_p): Likewise.
20777 (can_sm_ref_p): Likewise.
20779 2013-03-18 Jakub Jelinek <jakub@redhat.com>
20782 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
20783 return 1 even for !unsignedp.
20785 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
20787 * config/i386/i386.md (isa): Add x64 and nox64.
20788 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
20789 (*pushtf): Enable *roF alternative for x64 isa only.
20790 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
20791 mode attribute of integer alternatives to DImode for TARGET_64BIT.
20792 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
20793 (*movtf_internal): Merge from *movtf_internal_rex64 and
20794 *movtf_internal_sse. Use x64 and nox64 isa attributes.
20795 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
20796 nox64 isa attributes.
20797 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
20798 nox64 isa attributes.
20799 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
20801 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
20803 * config/alpha/alpha.c (TARGET_LRA_P): New define.
20805 2013-03-17 Jakub Jelinek <jakub@redhat.com>
20808 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
20809 class names. Remove trailing comma after "ALL_REGS".
20811 2013-03-16 Jan Hubicka <jh@suse.cz>
20813 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
20814 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
20815 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
20816 of cgraph_get_create_node.
20817 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
20819 2013-03-16 Jason Merrill <jason@redhat.com>
20822 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
20823 with DW_AT_default_value.
20825 2013-03-16 Jakub Jelinek <jakub@redhat.com>
20827 * BASE-VER: Set to 4.9.0.
20829 2013-03-14 Andi Kleen <ak@linux.intel.com>
20832 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
20833 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
20834 Document _x* TSX intrinsics.
20836 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
20837 David Holsgrove <david.holsgrove@xilinx.com>
20839 * configure.ac: Add MicroBlaze TLS support detection.
20840 * configure: Regenerate.
20841 * config/microblaze/microblaze-protos.h
20842 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
20843 symbol_mentioned_p, label_mentioned_p): Add prototypes.
20844 * config/microblaze/microblaze.c (microblaze_address_type): Add
20845 ADDRESS_TLS and tls_reloc address types.
20846 (microblaze_address_info): Add tls_reloc.
20847 (TARGET_HAVE_TLS): Define.
20848 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
20849 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
20850 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
20851 load_tls_operand, microblaze_call_tls_get_addr,
20852 microblaze_legitimize_tls_address): New functions.
20853 (microblaze_classify_unspec): Handle UNSPEC_TLS.
20854 (get_base_reg): Use microblaze_tls_symbol_p.
20855 (microblaze_classify_address): Handle TLS.
20856 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
20857 label_mentioned_p and microblaze_tls_referenced_p.
20858 (microblaze_legitimize_address): Handle TLS.
20859 (microblaze_address_insns): Handle ADDRESS_TLS.
20860 (pic_address_needs_scratch): Handle TLS.
20861 (print_operand_address): Handle TLS.
20862 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
20863 (microblaze_expand_move): Handle TLS.
20864 (microblaze_legitimate_constant_p): Check
20865 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
20866 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
20867 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
20868 (PIC_OFFSET_TABLE_REGNUM): Set.
20869 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
20870 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
20871 (addsi3, movsi_internal2, movdf_internal): Update constraints
20872 * config/microblaze/predicates.md (arith_plus_operand): Define
20873 (move_operand): Redefine as move_src_operand,
20874 check microblaze_tls_referenced_p.
20876 2013-03-14 Ian Bolton <ian.bolton@arm.com>
20878 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
20879 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
20881 2013-03-14 Ian Bolton <ian.bolton@arm.com>
20883 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
20886 2013-03-14 Jakub Jelinek <jakub@redhat.com>
20888 PR tree-optimization/53265
20889 * common.opt (Waggressive-loop-optimizations): New option.
20890 * tree-ssa-loop-niter.c: Include tree-pass.h.
20891 (do_warn_aggressive_loop_optimizations): New function.
20892 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
20893 if number_of_latch_executions returned constant.
20894 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
20895 early. If number_of_latch_executions returned constant, set
20896 nb_iterations_upper_bound back to it.
20897 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
20899 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
20900 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
20902 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
20903 (MULTILIB_OSDIRNAMES): Set.
20904 * genmultilib: If defaultosdirname doesn't start with :: , set
20905 defaultosdirname2 instead, clear it and emit two . multilib_raw
20906 entries instead of just one.
20908 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
20910 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
20911 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
20912 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
20913 (SUBTARGET_OVERRIDE_OPTIONS): New.
20915 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
20918 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
20919 (musermode): Convert to Var(TARGET_USERMODE).
20920 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
20921 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
20922 * config/sh/sh.c (sh_option_override): Use
20923 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
20924 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
20926 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
20928 (udivsi3_i4_single, divsi3_i4_single): Use
20929 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
20931 2013-03-13 Dave Korn <dave.korn.cygwin@....>
20933 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
20936 2013-03-13 Richard Biener <rguenther@suse.de>
20938 PR tree-optimization/56608
20939 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
20940 calls when vectorizing basic-blocks.
20942 2013-03-13 Jakub Jelinek <jakub@redhat.com>
20945 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
20948 2013-03-12 Jakub Jelinek <jakub@redhat.com>
20950 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
20952 2013-03-11 Jan Hubicka <jh@suse.cz>
20955 * lto-streamer-out.c (output_symbol_p): Skip references from
20956 constructors of external variables.
20958 2013-03-11 Jan Hubicka <jh@suse.cz>
20960 PR middle-end/56571
20961 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
20963 * emit-rtl.c (verify_rtx_sharing): Likewise.
20964 (copy_insn_1): Likewise.
20965 * rtl.c (copy_rtx): Likewise.
20967 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
20970 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
20971 output_operand_lossage message.
20973 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
20976 * arm.c (shift_op): Validate RTL pattern on the fly.
20977 (arm_print_operand, case 'S'): Don't use shift_operator to validate
20980 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20983 * config/pa/pa.md (call_value): Check for calls to powf and direct to
20984 new call patterns that clobber %fr12.
20985 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
20986 split and postreload patterns.
20987 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
20988 registers %fr12 and %fr12R as call used.
20990 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
20992 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
20993 (canon_address, record_store, replace_read, check_mem_read_rtx,
20994 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
20995 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
20996 rest_of_handle_dse): Likewise.
20998 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
21000 PR middle-end/56524
21001 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
21003 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
21004 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
21005 (save_optabs_if_changed): Replace with...
21006 (init_tree_optimization_optabs): ...this.
21007 * optabs.c (save_optabs_if_changed): Rename to...
21008 (init_tree_optimization_optabs): ...this. Take the optimization node
21009 as argument. Do nothing if the base optabs are already correct.
21010 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
21011 to recompute optabs.
21012 * function.h (function): Remove optabs field.
21013 * function.c (invoke_set_current_function_hook): Call
21014 init_tree_optimization_optabs. Use the result to initialize
21017 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
21019 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
21020 if GTMA_HAS_NO_INSTRUMENTATION.
21021 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
21022 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
21023 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
21024 * gimple-pretty-print.c (dump_gimple_transaction): Handle
21025 GTMA_HAS_NO_INSTRUMENTATION.
21027 2013-03-08 Jakub Jelinek <jakub@redhat.com>
21029 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
21032 2013-03-08 Marek Polacek <polacek@redhat.com>
21033 Jakub Jelinek <jakub@redhat.com>
21035 PR tree-optimization/56478
21036 * predict.c (is_comparison_with_loop_invariant_p): Change the
21037 type of loop_step to tree.
21038 (predict_loops): Adjust.
21039 (predict_iv_comparison): Perform the computations on double_ints.
21041 2013-03-08 Richard Biener <rguenther@suse.de>
21043 PR tree-optimization/56570
21044 * tree-cfg.c (verify_expr_location_1): Verify locations for
21046 * tree-sra.c (create_access_replacement): Strip locations
21047 from DECL_DEBUG_EXPRs.
21049 2013-03-08 Richard Biener <rguenther@suse.de>
21051 * tree-inline.c (expand_call_inline): Do not associate
21052 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
21053 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
21055 2013-03-08 Richard Biener <rguenther@suse.de>
21057 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
21058 or block changes with -Og. Fix for location / block encoding
21059 changes and PHI arguments with locations.
21061 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
21063 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
21065 (struct output_info): Likewise.
21066 (register_overhead): Remove bad gcc_assert.
21067 (bitmap_find_bit): If there is only a single bitmap element, do not
21068 count a miss as a search.
21069 (print_statistics): Update for counter type changes.
21070 (dump_bitmap_statistics): Likewise. Print headers such that they
21071 are properly lined up with the printed counters.
21073 2013-03-07 Jakub Jelinek <jakub@redhat.com>
21075 PR tree-optimization/56559
21076 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
21077 check that it has only a single use.
21079 2013-03-07 Richard Biener <rguenther@suse.de>
21081 * doc/invoke.texi (fwhole-program): Discourage use in combination
21084 2013-03-06 Jakub Jelinek <jakub@redhat.com>
21086 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
21088 PR tree-optimization/56539
21089 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
21090 instead of GSI_CONTINUE_LINKING as last argument to
21091 force_gimple_operand_gsi. Adjust function comment.
21093 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
21096 PR middle-end/56548
21097 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
21098 promoted mode, convert the result back to the original mode.
21100 2013-03-06 Richard Biener <rguenther@suse.de>
21102 PR middle-end/56294
21103 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
21104 (insert_updated_phi_nodes_compare_uids): New function.
21105 (update_ssa): Sort symbols_to_rename after UID before
21106 traversing it to insert PHI nodes.
21108 2013-03-06 Richard Biener <rguenther@suse.de>
21110 PR middle-end/50494
21111 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
21112 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
21115 2013-02-13 Richard Biener <rguenther@suse.de>
21118 * varasm.c (output_constant_def_1): Get the decl representing
21119 the constant as argument.
21120 (output_constant_def): Wrap output_constant_def_1.
21121 (make_decl_rtl): Use output_constant_def_1 with the decl
21122 representing the constant.
21123 (build_constant_desc): Optionally re-use a decl already
21124 representing the constant.
21125 (tree_output_constant_def): Adjust.
21127 2013-03-06 Joey Ye <joey.ye@arm.com>
21130 * gcc.c (convert_white_space): New function.
21131 (main): Handles white space in function name.
21133 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
21136 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
21137 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
21138 to SH_DIV_CALL_TABLE for TARGET_SH2.
21139 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
21141 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
21142 call-table options.
21144 2013-03-05 Sterling Augustine <saugustine@google.com>
21145 Cary Coutant <ccoutant@google.com>
21148 * dwarf2out.c (resolve_addr): Don't call
21149 remove_loc_list_addr_table_entries a second time for the same
21152 2013-03-05 Jakub Jelinek <jakub@redhat.com>
21155 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
21156 (avoid_complex_debug_insns): New function.
21157 (expand_debug_locations): Call it.
21159 PR rtl-optimization/56484
21160 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
21161 lifetimes of hard registers on small register class machines.
21163 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
21165 * config/microblaze/microblaze-protos.h: Rename
21166 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
21167 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
21169 (microblaze_fast_interrupt_function_p): New function.
21170 (microblaze_is_interrupt_handler): Rename to
21171 microblaze_is_interrupt_variant and add fast_interrupt check.
21172 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
21173 (save_restore_insns): Likewise.
21174 (compute_frame_size): Likewise.
21175 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
21176 (microblaze_globalize_label): Likewise.
21177 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
21178 * config/microblaze/microblaze.md: Use wrapper
21179 microblaze_is_interrupt_variant.
21181 2013-03-05 Kai Tietz <ktietz@redhat.com>
21183 * sdbout.c (sdbout_one_type): Switch to current function's section
21184 supporting cold/hot.
21186 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
21188 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
21191 2013-03-05 Jakub Jelinek <jakub@redhat.com>
21193 PR middle-end/56461
21194 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
21195 if VALGRIND_GET_VBITS is defined, temporarily make object
21196 memory all defined, and restore previous valgrind addressability
21197 and definability afterwards. Free this_object at the end.
21199 PR middle-end/56461
21200 * lra.c (lra): Call lra_clear_live_ranges if live_p,
21201 right before calling lra_create_live_ranges, also call it
21202 when clearing live_p. Only call lra_clear_live_ranges
21203 at the end if live_p.
21205 PR middle-end/56461
21206 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
21208 2013-03-05 Richard Biener <rguenther@suse.de>
21210 PR tree-optimization/56521
21211 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
21214 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
21217 * except.h (remove_unreachable_eh_regions): New prototype.
21218 * except.c (remove_eh_handler_splicer): New function, split out
21219 of remove_eh_handler.
21220 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
21221 warning about running it on many EH regions one at a time.
21222 (remove_unreachable_eh_regions_worker): New function, walk the
21223 EH tree in depth-first order and remove non-marked regions.
21224 (remove_unreachable_eh_regions): New function.
21225 * tree-eh.c (mark_reachable_handlers): New function, split out
21226 from remove_unreachable_handlers.
21227 (remove_unreachable_handlers): Use mark_reachable_handlers and
21228 remove_unreachable_eh_regions.
21229 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
21230 and remove_unreachable_eh_regions.
21232 2013-03-05 Richard Biener <rguenther@suse.de>
21234 PR middle-end/56525
21235 * loop-init.c (fix_loop_structure): Remove loops in two stages,
21236 not freeing them until the end.
21238 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21240 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
21242 2013-03-05 Richard Biener <rguenther@suse.de>
21244 PR tree-optimization/56270
21245 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
21246 of loads after scheduling an SLP instance.
21248 2013-03-05 Jakub Jelinek <jakub@redhat.com>
21250 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
21252 (check_gcc_parallelize): Run guality.exp as a separate job from
21253 vect.exp with unsorted.exp and $(dg_target_exps) separately from
21254 struct-layout-1.exp with stackalign.exp.
21256 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
21258 PR middle-end/56461
21259 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
21260 load_index sbitmap even if some bit in it isn't set.
21262 PR middle-end/56461
21263 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
21264 (discover_iteration_bound_by_body_walk): Change queues to
21265 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
21266 spelling in comment. Call safe_push on queues[bound_index] directly.
21267 Release queues[queue_index] in every iteration unconditionally.
21268 Release bounds vector.
21270 PR middle-end/56461
21271 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
21272 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
21273 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
21276 2013-03-05 Richard Biener <rguenther@suse.de>
21279 * tree-inline.c (remap_blocks_to_null): New function.
21280 (expand_call_inline): When expanding a call stmt without
21281 an associated block inline remap all callee blocks to NULL.
21283 2013-03-05 Jakub Jelinek <jakub@redhat.com>
21285 PR rtl-optimization/56494
21286 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
21287 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
21288 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
21290 PR middle-end/56461
21291 * sel-sched-ir.c (free_sched_pools): Release
21292 succs_info_pool.stack[succs_info_pool.max_top] vectors too
21293 if succs_info_pool.max_top isn't -1.
21296 * opts.c (opts_obstack, opts_concat): Moved to...
21297 * opts-common.c (opts_obstack, opts_concat): ... here.
21299 2013-03-04 Jakub Jelinek <jakub@redhat.com>
21301 PR middle-end/56461
21302 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
21304 2013-03-04 Martin Jambor <mjambor@suse.cz>
21306 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
21307 all appropriate places.
21309 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
21311 PR tree-optimization/56424
21312 * ipa-split.c (split_function): Do not set the RSO flag if result is
21313 not by reference and its type is a register type.
21315 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
21317 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
21318 (microblaze_legitimate_pic_operand): Likewise
21319 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
21320 new function microblaze_legitimate_pic_operand
21321 * config/microblaze/microblaze-protos.h
21322 (microblaze_legitimate_pic_operand): Declare.
21324 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21326 * config/microblaze/predicates.md (call_insn_simple_operand):
21327 New predicate for supported rtx code types.
21328 * config/microblaze/microblaze.md (call_internal1): Use
21329 call_insn_simple_operand predicate.
21331 2013-03-04 Jakub Jelinek <jakub@redhat.com>
21333 PR middle-end/56461
21334 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
21335 partitions.ordered_remove.
21337 PR middle-end/56461
21338 * tree-vect-stmts.c (vectorizable_conversion): Don't call
21339 vec_oprnds0.create (1) for modifier == NONE.
21341 PR middle-end/56461
21342 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
21343 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
21344 vec_oprnds1 right before pushing anything to it for
21347 PR middle-end/56461
21348 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
21349 set nbbs to 0 instead of having separate code path.
21350 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
21351 instead of false as last argument if returning NULL.
21353 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
21355 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
21356 the attribute is now called "target" instead of "option".
21357 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
21358 * doc/tm.texi.in (Target Attributes): Likewise document the correct
21359 attribute/pragma name for TARGET_OPTION_VALID_P and
21360 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
21361 * doc/tm.texi: Regenerated.
21363 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
21365 * config/microblaze/microblaze.c:
21366 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
21367 * config/microblaze/microblaze.h: Add -mxl-reorder to
21369 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
21370 instructions emitted if TARGET_REORDER.
21371 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
21372 or 0 for -m/-mno case, but initialises as 2 to detect default use case
21375 2013-03-01 Xinliang David Li <davidxl@google.com>
21377 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
21380 2013-03-01 Jakub Jelinek <jakub@redhat.com>
21382 PR middle-end/56461
21383 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
21384 vector even when returning true. Fix up function comment formatting.
21386 PR middle-end/56461
21387 * ira-build.c (ira_loop_nodes_count): New variable.
21388 (create_loop_tree_nodes): Initialize it.
21389 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
21391 PR middle-end/56461
21392 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
21393 method on dr_chain and result_chain.
21394 * tree-vect-stmts.c (vectorizable_store): Only call
21395 result_chain.create if j == 0.
21397 PR middle-end/56461
21398 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
21399 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
21400 before overwriting it.
21402 2013-03-01 Tobias Burnus <burnus@net-b.de>
21404 * doc/extended.texi (C Extensions): Change order in @menu
21406 (Other MIPS Built-in Functions): Move last MIPS entry before
21407 "picoChip Built-in Functions".
21408 (SH Built-in Functions): Move after RX Built-in Functions.
21409 * doc/gcc.texi (Introduction): Change order in @menu
21411 * doc/md.texi (Constraints): Ditto.
21412 * gty.texi (Type Information): Ditto.
21413 (User-provided marking routines for template types): Make
21415 * doc/invoke.texi (AArch64 Options): Move before
21416 "Adapteva Epiphany Options".
21418 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
21419 Jakub Jelinek <jakub@redhat.com>
21422 * asan.c (gate_asan): Lookup no_sanitize_address instead of
21423 no_address_safety_analysis attribute.
21424 * doc/extend.texi (no_address_safety_attribute): Rename to
21425 no_sanitize_address attribute, mention no_address_safety_analysis
21426 attribute as deprecated alias.
21428 2013-02-28 Jakub Jelinek <jakub@redhat.com>
21430 PR middle-end/56461
21431 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
21432 type to vec<vec<tree> > *.
21433 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
21434 to be vec<tree> instead of vec<tree> *, set vec_defs
21435 to vNULL and call vec_defs.create (number_of_vects), adjust other
21437 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
21438 vectorizable_condition): Adjust vect_get_slp_defs callers.
21440 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
21442 * config/aarch64/aarch64.c
21443 (aarch64_float_const_representable): Remove unused variable.
21445 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
21447 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
21449 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
21451 * config/aarch64/aarch64-builtins.c
21452 (aarch64_init_simd_builtins): Make static.
21454 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
21456 * config/aarch64/aarch64.c
21457 (aarch64_simd_make_constant): Make static.
21459 2013-02-28 Martin Jambor <mjambor@suse.cz>
21461 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
21462 with no initialization to the RHS of debug statements.
21464 2013-02-28 Martin Jambor <mjambor@suse.cz>
21466 PR tree-optimization/56294
21467 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
21469 (get_access_replacement): Do not call create_access_replacement.
21470 Assert a replacement exists.
21471 (get_repl_default_def_ssa_name): Create the replacement declaration
21474 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21476 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
21477 final_end_function.
21479 2013-02-28 Marek Polacek <polacek@redhat.com>
21481 PR rtl-optimization/56466
21482 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
21483 if we're changing a loop.
21484 (peel_loops_completely): Likewise.
21486 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
21489 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
21491 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
21494 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
21495 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
21496 INTX_FTYPE_FX, FX_FTYPE_INTX.
21497 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
21499 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
21501 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
21502 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
21503 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
21504 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
21505 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
21506 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
21507 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
21508 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
21509 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
21510 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
21511 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
21512 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
21513 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
21514 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
21515 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
21516 (avrxmega6): Increase max flash segments from 5 to 6.
21517 * config/avr/t-multilib: Regenerate.
21518 * config/avr/avr-tables.opt: Regenerate.
21519 * doc/avr-mmcu.texi: Regenerate.
21521 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
21523 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
21524 (avr_device_to_arch): Rename to avr_device_to_ld.
21525 (avr_device_to_as): New prototype.
21526 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
21527 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
21528 * config/avr/driver-avr.c (avr_device_to_as): New.
21529 (avr_device_to_arch): Rename to avr_device_to_ld.
21531 2013-02-27 Jakub Jelinek <jakub@redhat.com>
21533 PR middle-end/56461
21534 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
21535 method on dr_chain and result_chain.
21537 PR middle-end/56461
21538 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
21539 pointer_set_destroy on not_executed_last_iteration.
21541 PR middle-end/56461
21542 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
21544 PR middle-end/56461
21545 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
21546 FOR_EACH_DEFINED_FUNCTION when freeing state.
21548 PR middle-end/56461
21549 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
21551 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
21554 PR middle-end/56461
21555 * ipa-cp.c (decide_whether_version_node): Call vec_free on
21556 known_aggs[i].items and release known_aggs vector.
21558 PR middle-end/56461
21559 * ipa-reference.c (propagate): Free node_info even for alias nodes.
21561 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21563 * config/microblaze/microblaze.c (microblaze_emit_compare):
21564 Use xor for EQ/NE comparisions.
21565 * config/microblaze/microblaze.md (cstoresf4): Add constraints
21566 (cbranchsf4): Adjust operator to comparison_operator.
21568 2013-02-27 Jakub Jelinek <jakub@redhat.com>
21570 PR middle-end/56461
21571 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
21573 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
21574 vec_safe_push, always update *slot.
21575 (redirect_edge_var_map_clear): Use vec_free.
21576 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
21577 (free_var_map_entry): Use vec_free.
21578 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
21579 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
21581 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
21583 PR middle-end/45472
21584 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
21585 when the may_trap_p bit of the exprs being merged differs.
21586 Reorder tests for speculativeness in the logical and operator.
21588 2013-02-27 Jakub Jelinek <jakub@redhat.com>
21590 * incpath.c (add_standard_paths): Use reconcat instead of concat
21591 where appropriate and avoid leaking memory.
21593 * opts.h: Include obstack.h.
21594 (opts_concat): New prototype.
21595 (opts_obstack): New declaration.
21596 * opts.c (opts_concat): New function.
21597 (opts_obstack): New variable.
21598 (init_options_struct): Call gcc_init_obstack on opts_obstack.
21599 (finish_options): Use opts_concat instead of concat
21600 and XOBNEWVEC instead of XNEWVEC.
21601 * opts-common.c (generate_canonical_option, decode_cmdline_option,
21602 generate_option): Likewise.
21603 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
21604 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
21607 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
21608 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
21610 2013-02-26 Jakub Jelinek <jakub@redhat.com>
21612 PR middle-end/56461
21613 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
21615 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
21617 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
21618 (arm_block_move_unaligned_straight): Likewise.
21619 (arm_adjust_block_mem): Likewise.
21621 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
21624 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
21625 temp, cond and label.
21626 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
21629 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
21630 * config/c6x/c6x.h (dbx_register_map): Update declaration.
21633 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
21634 of prologue/epilogue functions.
21637 * config/tilegx/tilegx.c (tilegx_expand_prologue):
21638 Remove unused variable cfa_offset.
21639 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
21642 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
21643 type promotion to unsigned.
21646 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
21647 for HOST_WIDE_INT of 32 bit / same size as int.
21648 (arm_block_move_unaligned_straight): Likewise.
21649 (arm_adjust_block_mem): Likewise.
21652 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
21655 2013-02-26 Marek Polacek <polacek@redhat.com>
21657 PR tree-optimization/56426
21658 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
21660 2013-02-26 Richard Biener <rguenther@suse.de>
21663 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
21664 unused variable loops.
21666 2013-02-26 Jakub Jelinek <jakub@redhat.com>
21668 PR tree-optimization/56448
21669 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
21670 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
21671 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
21672 later operands of the references, or even first operand for
21673 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
21675 PR tree-optimization/56443
21676 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
21677 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
21678 to type_for_mode langhook.
21680 2013-02-25 Matt Turner <mattst88@gmail.com>
21682 * doc/invoke.texi: Document r4700.
21684 2013-02-25 Richard Biener <rguenther@suse.de>
21686 PR tree-optimization/56175
21687 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
21689 (simplify_bitwise_binary): ... here. Also guard the conversion
21690 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
21692 2013-02-25 Catherine Moore <clm@codesourcery.com>
21695 2013-02-24 Catherine Moore <clm@codesourcery.com>
21696 Maciej W. Rozycki <macro@codesourcery.com>
21697 Tom de Vries <tom@codesourcery.com>
21698 Nathan Sidwell <nathan@codesourcery.com>
21699 Iain Sandoe <iain@codesourcery.com>
21700 Nathan Froyd <froydnj@codesourcery.com>
21701 Chao-ying Fu <fu@mips.com>
21703 * doc/extend.texi: (micromips, nomicromips, nocompression):
21704 Document new function attributes.
21705 * doc/invoke.texi (minterlink-compressed, mmicromips,
21706 m14k, m14ke, m14kec): Document new options.
21707 (minterlink-mips16): Update documentation.
21708 * doc/md.texi (ZC, ZD): Document new constraints.
21709 * configure.ac (gcc_cv_as_micromips): Check if linker
21710 supports the .set micromips directive.
21711 * configure: Regenerate.
21712 * config.in: Regenerate.
21713 * config/mips/mips-tables.opt: Regenerate.
21714 * config/mips/micromips.md: New file.
21715 * constraints.md (ZC, AD): New constraints.
21716 * config/mips/predicates.md (movep_src_register): New predicate.
21717 (movep_src_operand): New predicate.
21718 (non_volatile_mem_operand): New predicate.
21719 * config/mips/mips.md (multimem): New type.
21720 (length): Differentiate between 17-bit and 18-bit branch offsets.
21721 (MOVEP1, MOVEP2): New mode iterator.
21722 (mov_<load>l): Use ZC constraint.
21723 (mov_<load>r): Likewise.
21724 (mov_<store>l): Likewise.
21725 (mov_<store>r): Likewise.
21726 (*branch_equality<mode>_inverted): Add microMIPS support.
21727 (*branch_equality<mode>): Likewise.
21728 (*jump_absolute): Likewise.
21729 (indirect_jump_<mode>): Likewise.
21730 (tablejump_<mode>): Likewise.
21731 (<optab>_internal): Likewise.
21732 (sibcall_internal): Likewise.
21733 (sibcall_value_internal): Likewise.
21734 (prefetch): Use constraint ZD.
21735 * config/mips/mips.opt (minterlink-compressed): New option.
21736 (minterlink-mips16): Now an alias for minterlink-compressed.
21737 (mmicromips): New option.
21738 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
21739 (compare_and_swap_12): Likewise.
21740 (sync_add<mode>): Likewise.
21741 (sync_<optab>_12): Likewise.
21742 (sync_old_<optab>_12): Likewise.
21743 (sync_new_<optab>_12): Likewise.
21744 (sync_nand_12): Likewise.
21745 (sync_old_nand_12): Likewise.
21746 (sync_new_nand_12): Likewise.
21747 (sync_sub<mode>): Likewise.
21748 (sync_old_add<mode>): Likewise.
21749 (sync_old_sub<mode>): Likewise.
21750 (sync_new_add<mode>): Likewise.
21751 (sync_new_sub<mode>): Likewise.
21752 (sync_<optab><mode>): Likewise.
21753 (sync_old_<optab><mode>): Likewise.
21754 (sync_new_<optab><mode>): Likewise.
21755 (sync_nand<mode>): Likewise.
21756 (sync_old_nand<mode>): Likewise.
21757 (sync_new_nand<mode>): Likewise.
21758 (sync_lock_test_and_set<mode>): Likewise.
21759 (test_and_set_12): Likewise.
21760 (atomic_compare_and_swap<mode>): Likewise.
21761 (atomic_exchange<mode>_llsc): Likewise.
21762 (atomic_fetch_add<mode>_llsc): Likewise.
21763 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
21764 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
21765 (umips_save_restore_pattern_p): Likewise.
21766 (umips_load_store_pair_p): Likewise.
21767 (umips_output_load_store_pair): Likewise.
21768 (umips_movep_target_p): Likewise.
21769 (umips_12bit_offset_address_p): Likewise.
21770 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
21771 (mips_base_mips16): Rename this...
21772 (mips_base_compression_flags): ...to this. Update all uses.
21773 (mips_attribute_table): Add micromips, nomicromips and nocompression.
21774 (mips_mips16_decl_p): Delete.
21775 (mips_nomips16_decl_p): Delete.
21776 (mips_get_compress_on_flags): New function.
21777 (mips_get_compress_off_flags): New function.
21778 (mips_get_compress_mode): New function.
21779 (mips_get_compress_on_name): New function.
21780 (mips_get_compress_off_name): New function.
21781 (mips_insert_attributes): Support multiple compression types.
21782 (mips_merge_decl_attributes): Likewise.
21783 (umips_12bit_offset_address_p): New function.
21784 (mips_start_function_definition): Emit .set micromips directive.
21785 (mips_call_may_need_jalx_p): New function.
21786 (mips_function_ok_for_sibcall): Add microMIPS support.
21787 (mips_print_operand_punctuation): Support short delay slots and
21789 (umips_swm_mask, umips_swm_encoding): New.
21790 (umips_build_save_restore): New function.
21791 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
21792 (was_mips16_p): Remove.
21793 (old_compression_mode): New.
21794 (mips_set_compression_mode): New function.
21795 (mips_set_current_function): Add microMIPS support.
21796 (mips_option_override): Likewise.
21797 (umips_save_restore_pattern_p): New function.
21798 (umips_output_save_restore): New function.
21799 (umips_load_store_pair_p_1): New function.
21800 (umips_load_store_pair_p): New function.
21801 (umips_output_load_store_pair_1): New function.
21802 (umips_output_load_store_pair): New function.
21803 (umips_movep_target_p) New function.
21804 (mips_prepare_pch_save): Add microMIPS support.
21805 * config/mips/mips.h (TARGET_COMPRESSION): New.
21806 (TARGET_CPU_CPP_BUILTINS): Update macro
21807 to use new compression flags and to support microMIPS.
21808 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
21809 (MIPS_ARCH_FLOAT_SPEC): Likewise.
21810 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
21811 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
21812 (ASM_SPEC): Support mmicromips and mno-micromips.
21813 (M16STORE_REG_P): New macro.
21814 (MIPS_CALL): Support TARGET_MICROMIPS.
21815 (MICROMIPS_J): New macro.
21816 (mips_base_mips16): Rename this...
21817 (mips_base_compression_flags): ...to this.
21818 (UMIPS_12BIT_OFFSET_P): New macro.
21819 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
21820 (MULTILIB_DIRNAMES): Likewise.
21822 2013-02-25 Tom de Vries <tom@codesourcery.com>
21824 PR rtl-optimization/56131
21825 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
21826 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
21827 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
21829 2013-02-25 Tobias Burnus <burnus@net-b.de>
21831 * doc/invoke.texi (-fsanitize=): Move from optimization
21832 to debugging options.
21834 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
21836 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
21838 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
21839 Alexander Monakov <amonakov@ispras.ru>
21841 PR middle-end/56077
21842 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
21843 flush pending lists also on non-jumps. Adjust comment.
21845 2013-02-24 Catherine Moore <clm@codesourcery.com>
21846 Maciej W. Rozycki <macro@codesourcery.com>
21847 Tom de Vries <tom@codesourcery.com>
21848 Nathan Sidwell <nathan@codesourcery.com>
21849 Iain Sandoe <iain@codesourcery.com>
21850 Nathan Froyd <froydnj@codesourcery.com>
21851 Chao-ying Fu <fu@mips.com>
21853 * doc/extend.texi: (micromips, nomicromips, nocompression):
21854 Document new function attributes.
21855 * doc/invoke.texi (minterlink-compressed, mmicromips,
21856 m14k, m14ke, m14kec): Document new options.
21857 (minterlink-mips16): Update documentation.
21858 * doc/md.texi (ZC, ZD): Document new constraints.
21859 * configure.ac (gcc_cv_as_micromips): Check if linker
21860 supports the .set micromips directive.
21861 * configure: Regenerate.
21862 * config.in: Regenerate.
21863 * config/mips/mips-tables.opt: Regenerate.
21864 * config/mips/micromips.md: New file.
21865 * constraints.md (ZC, AD): New constraints.
21866 * config/mips/predicates.md (movep_src_register): New predicate.
21867 (movep_src_operand): New predicate.
21868 (non_volatile_mem_operand): New predicate.
21869 * config/mips/mips.md (multimem): New type.
21870 (length): Differentiate between 17-bit and 18-bit branch offsets.
21871 (MOVEP1, MOVEP2): New mode iterator.
21872 (mov_<load>l): Use ZC constraint.
21873 (mov_<load>r): Likewise.
21874 (mov_<store>l): Likewise.
21875 (mov_<store>r): Likewise.
21876 (*branch_equality<mode>_inverted): Add microMIPS support.
21877 (*branch_equality<mode>): Likewise.
21878 (*jump_absolute): Likewise.
21879 (indirect_jump_<mode>): Likewise.
21880 (tablejump_<mode>): Likewise.
21881 (<optab>_internal): Likewise.
21882 (sibcall_internal): Likewise.
21883 (sibcall_value_internal): Likewise.
21884 (prefetch): Use constraint ZD.
21885 * config/mips/mips.opt (minterlink-compressed): New option.
21886 (minterlink-mips16): Now an alias for minterlink-compressed.
21887 (mmicromips): New option.
21888 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
21889 (compare_and_swap_12): Likewise.
21890 (sync_add<mode>): Likewise.
21891 (sync_<optab>_12): Likewise.
21892 (sync_old_<optab>_12): Likewise.
21893 (sync_new_<optab>_12): Likewise.
21894 (sync_nand_12): Likewise.
21895 (sync_old_nand_12): Likewise.
21896 (sync_new_nand_12): Likewise.
21897 (sync_sub<mode>): Likewise.
21898 (sync_old_add<mode>): Likewise.
21899 (sync_old_sub<mode>): Likewise.
21900 (sync_new_add<mode>): Likewise.
21901 (sync_new_sub<mode>): Likewise.
21902 (sync_<optab><mode>): Likewise.
21903 (sync_old_<optab><mode>): Likewise.
21904 (sync_new_<optab><mode>): Likewise.
21905 (sync_nand<mode>): Likewise.
21906 (sync_old_nand<mode>): Likewise.
21907 (sync_new_nand<mode>): Likewise.
21908 (sync_lock_test_and_set<mode>): Likewise.
21909 (test_and_set_12): Likewise.
21910 (atomic_compare_and_swap<mode>): Likewise.
21911 (atomic_exchange<mode>_llsc): Likewise.
21912 (atomic_fetch_add<mode>_llsc): Likewise.
21913 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
21914 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
21915 (umips_save_restore_pattern_p): Likewise.
21916 (umips_load_store_pair_p): Likewise.
21917 (umips_output_load_store_pair): Likewise.
21918 (umips_movep_target_p): Likewise.
21919 (umips_12bit_offset_address_p): Likewise.
21920 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
21921 (mips_base_mips16): Rename this...
21922 (mips_base_compression_flags): ...to this. Update all uses.
21923 (mips_attribute_table): Add micromips, nomicromips and nocompression.
21924 (mips_mips16_decl_p): Delete.
21925 (mips_nomips16_decl_p): Delete.
21926 (mips_get_compress_on_flags): New function.
21927 (mips_get_compress_off_flags): New function.
21928 (mips_get_compress_mode): New function.
21929 (mips_get_compress_on_name): New function.
21930 (mips_get_compress_off_name): New function.
21931 (mips_insert_attributes): Support multiple compression types.
21932 (mips_merge_decl_attributes): Likewise.
21933 (umips_12bit_offset_address_p): New function.
21934 (mips_start_function_definition): Emit .set micromips directive.
21935 (mips_call_may_need_jalx_p): New function.
21936 (mips_function_ok_for_sibcall): Add microMIPS support.
21937 (mips_print_operand_punctuation): Support short delay slots and
21939 (umips_swm_mask, umips_swm_encoding): New.
21940 (umips_build_save_restore): New function.
21941 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
21942 (was_mips16_p): Remove.
21943 (old_compression_mode): New.
21944 (mips_set_compression_mode): New function.
21945 (mips_set_current_function): Add microMIPS support.
21946 (mips_option_override): Likewise.
21947 (umips_save_restore_pattern_p): New function.
21948 (umips_output_save_restore): New function.
21949 (umips_load_store_pair_p_1): New function.
21950 (umips_load_store_pair_p): New function.
21951 (umips_output_load_store_pair_1): New function.
21952 (umips_output_load_store_pair): New function.
21953 (umips_movep_target_p) New function.
21954 (mips_prepare_pch_save): Add microMIPS support.
21955 * config/mips/mips.h (TARGET_COMPRESSION): New.
21956 (TARGET_CPU_CPP_BUILTINS): Update macro
21957 to use new compression flags and to support microMIPS.
21958 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
21959 (MIPS_ARCH_FLOAT_SPEC): Likewise.
21960 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
21961 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
21962 (ASM_SPEC): Support mmicromips and mno-micromips.
21963 (M16STORE_REG_P): New macro.
21964 (MIPS_CALL): Support TARGET_MICROMIPS.
21965 (MICROMIPS_J): New macro.
21966 (mips_base_mips16): Rename this...
21967 (mips_base_compression_flags): ...to this.
21968 (UMIPS_12BIT_OFFSET_P): New macro.
21969 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
21970 (MULTILIB_DIRNAMES): Likewise.
21972 2013-02-24 Jakub Jelinek <jakub@redhat.com>
21975 * target-globals.c (save_target_globals): For init_reg_sets and
21976 target_reinit remporarily set this_fn_optabs to this_target_optabs.
21978 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
21980 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
21981 * config/aarch64/t-aarch64
21982 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
21984 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
21986 PR inline-asm/56148
21987 * lra-constraints.c (process_alt_operands): Reload operand
21988 conflicting with earlier clobber only if no more other conflicting
21991 2013-02-22 Jakub Jelinek <jakub@redhat.com>
21994 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
21995 if not linking a shared library.
21997 2013-02-22 Seth LaForge <sethml@google.com>
21999 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
22001 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
22003 * config/arm/arm.md (split for extendsidi): Update condition.
22004 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
22005 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
22006 (qhs_zextenddi_cstr): Likewise.
22008 2013-02-21 Jakub Jelinek <jakub@redhat.com>
22010 PR middle-end/56420
22011 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
22012 avoid signed wrapping.
22013 (expand_mult): Handle properly multiplication by
22014 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
22015 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
22016 in the compiler if coeff is HOST_WIDE_INT_MIN.
22017 (expand_divmod): Don't make ext_op1 static, change it's type to
22018 uhwi. Avoid undefined behavior in -INTVAL (op1).
22020 PR rtl-optimization/50339
22021 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
22023 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
22024 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
22025 into splitting_ashiftrt field.
22026 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
22028 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
22031 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
22033 PR middle-end/56108
22034 * trans-mem.c (execute_tm_mark): Do not expand transactions that
22035 are sure to go irrevocable.
22037 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
22039 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
22040 scalars are valid operands.
22042 2013-02-21 Martin Jambor <mjambor@suse.cz>
22044 PR tree-optimization/56310
22045 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
22046 only matching indices and non-negative final offsets.
22047 (intersect_aggregates_with_edge): Pass src_idx to
22048 agg_replacements_to_vector. Pass src_idx insstead of index to
22049 intersect_with_agg_replacements.
22051 2013-02-21 Martin Jambor <mjambor@suse.cz>
22053 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
22054 instead of hard-wired defaults.
22056 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
22058 * doc/invoke.texi (MIPS Options): Update documentation of the
22059 floating-point multiply-accumulate instruction restrictions.
22061 2013-02-21 Kostya Serebryany <kcc@google.com>
22063 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
22064 asan_shadow_offset on x86_64 linux.
22066 2013-02-21 Richard Biener <rguenther@suse.de>
22068 PR tree-optimization/56415
22070 2013-02-11 Richard Biener <rguenther@suse.de>
22072 PR tree-optimization/56273
22073 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
22076 2013-02-21 Jakub Jelinek <jakub@redhat.com>
22079 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
22082 PR inline-asm/56405
22083 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
22084 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
22086 2013-02-20 Jan Hubicka <jh@suse.cz>
22088 PR tree-optimization/56265
22089 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
22090 when target is referenced for first time.
22092 2013-02-20 Richard Biener <rguenther@suse.de>
22094 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
22095 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
22096 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
22097 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
22098 not return anything.
22099 (rename_ssa_copies): Do not remove unused locals.
22100 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
22101 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
22102 * passes.c (execute_function_todo): Do not schedule unused locals
22103 removal if cleanup_tree_cfg did something.
22104 * tree-ssa-live.c (remove_unused_locals): Dump statistics
22105 about the number of removed locals.
22107 2013-02-20 Richard Biener <rguenther@suse.de>
22109 PR tree-optimization/56398
22110 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
22112 2013-02-20 Martin Jambor <mjambor@suse.cz>
22114 PR tree-optimization/55334
22115 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
22116 restricted pointers to arrays.
22118 2013-02-20 Richard Biener <rguenther@suse.de>
22119 Jakub Jelinek <jakub@redhat.com>
22121 PR tree-optimization/56396
22122 * tree-ssa-ccp.c (n_const_val): New static variable.
22123 (get_value): Return NULL for SSA names we don't have a lattice
22125 (ccp_initialize): Initialize n_const_val.
22126 * tree-ssa-copy.c (n_copy_of): New static variable.
22127 (init_copy_prop): Initialize n_copy_of.
22128 (get_value): Return NULL_TREE for SSA names we don't have a
22131 2013-02-20 Martin Jambor <mjambor@suse.cz>
22133 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
22135 2013-02-20 Richard Biener <rguenther@suse.de>
22137 * genpreds.c (write_lookup_constraint): Do not compare first
22138 letter of the constraint again.
22140 2013-02-20 Richard Biener <rguenther@suse.de>
22142 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
22144 (get_use_iv_cost): Terminate hashtable walk when coming across
22147 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
22149 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
22150 reassociation for avx2 targets.
22152 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
22154 * config/microblaze/microblaze.c: microblaze_has_clz = 0
22155 Add version check for v8.10.a to enable microblaze_has_clz
22156 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
22157 version and TARGET_PATTERN_COMPARE check
22158 * config/microblaze/microblaze.md: New clzsi2 instruction
22160 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
22162 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
22163 function before branching.
22165 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
22167 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
22169 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
22171 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
22173 PR middle-end/55889
22174 * sel-sched.c: Include ira.h.
22175 (implicit_clobber_conflict_p): New function.
22176 (moveup_expr): Use it.
22177 * Makefile.in (sel-sched.o): Depend on ira.h.
22179 2013-02-19 Richard Biener <rguenther@suse.de>
22181 PR tree-optimization/56384
22182 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
22183 (vn_hash_type): Split out from ...
22184 (vn_hash_constant_with_type): ... here.
22185 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
22186 (vn_phi_eq): Compare types from vn_phi_s structure.
22187 (vn_phi_lookup): Populate vn_phi_s type.
22188 (vn_phi_insert): Likewise.
22190 2013-02-19 Jakub Jelinek <jakub@redhat.com>
22192 PR tree-optimization/56350
22193 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
22194 if haven't found reduction or nested cycle operand, rather than
22195 asserting we must find it.
22197 PR tree-optimization/56381
22198 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
22201 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
22202 Jakub Jelinek <jakub@redhat.com>
22205 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
22206 (swap_optab_enable): Same.
22207 (init_all_optabs): Use argument instead of global.
22208 * tree.h (struct tree_optimization_option): New field target_optabs.
22209 * expr.h (init_all_optabs): Add argument to prototype.
22210 (TREE_OPTIMIZATION_OPTABS): New.
22211 (save_optabs_if_changed): Protoize.
22212 * optabs.h: Declare this_fn_optabs.
22213 * optabs.c (save_optabs_if_changed): New.
22214 Declare this_fn_optabs.
22215 (init_optabs): Add argument to init_all_optabs() call.
22216 * function.c (invoke_set_current_function_hook): Handle per
22218 * function.h (struct function): New field optabs.
22219 * config/mips/mips.c (mips_set_mips16_mode): Handle when
22220 optimization_current_node has changed.
22221 * target-globals.h (save_target_globals_default_opts): Protoize.
22222 * target-globals.c (save_target_globals_default_opts): New.
22224 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22227 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
22228 registers %fr12 and %fr12R as call used.
22231 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
22232 and HImode, require all displacements to be an integer multiple of
22234 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
22235 only allow QImode and HImode when reload is in progress and strict is
22236 true. Likewise for symbolic addresses. Use base14_operand to check
22237 displacements in REG+BASE addresses.
22239 2013-02-18 Richard Biener <rguenther@suse.de>
22241 PR tree-optimization/56366
22242 * tree-vect-loop.c (get_initial_def_for_induction): Properly
22243 handle sign-conversion of outer-loop initial induction value.
22245 2013-02-18 Richard Biener <rguenther@suse.de>
22247 PR middle-end/56349
22248 * cfghooks.c (merge_blocks): If we merge a latch into another
22249 block adjust references to it.
22250 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
22251 (verify_loop_structure): Verify that a recorded latch is in fact
22254 2013-02-18 Richard Biener <rguenther@suse.de>
22256 PR tree-optimization/56321
22257 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
22258 order SSA name release and virtual operand unlinking.
22260 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
22262 * config/microblaze/microblaze.md (save_stack_block): Define.
22263 (restore_stack_block): Likewise.
22265 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
22267 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
22268 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
22269 * config/microblaze/microblaze.c (microblaze_option_override):
22270 Bail out early for PIC modes when target does not support PIC.
22272 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
22274 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
22275 Replace with a microblaze version.
22276 (microblaze_trampoline_init): Adapt for microblaze.
22277 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
22280 2013-02-16 Jakub Jelinek <jakub@redhat.com>
22281 Dodji Seketeli <dodji@redhat.com>
22284 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
22285 (instrument_mem_region_access): Do not forget to always put
22286 instrumentation of the of 'base' and 'base + len' in a "if (len !=
22287 0) statement, even for cases where either 'base' or 'base + len'
22288 are not instrumented -- because they have been previously
22289 instrumented. Simplify the logic by putting all the statements
22290 instrument 'base + len' inside a sequence, and then insert that
22291 sequence right before the current insertion point. Then, to
22292 instrument 'base + len', just get an iterator on that statement.
22293 And do not forget to update the pointer to iterator the function
22294 received as argument.
22296 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
22298 PR rtl-optimization/56348
22299 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
22301 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
22303 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
22304 (clean_graph_dump_file): Pass base to start_graph_dump.
22306 2013-02-14 Richard Henderson <rth@redhat.com>
22309 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
22311 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
22313 * collect2-aix.h: Define F_LOADONLY.
22315 2013-02-14 Richard Biener <rguenther@suse.de>
22318 * varasm.c (output_constant_def_1): Get the decl representing
22319 the constant as argument.
22320 (output_constant_def): Wrap output_constant_def_1.
22321 (make_decl_rtl): Use output_constant_def_1 with the decl
22322 representing the constant.
22323 (build_constant_desc): Optionally re-use a decl already
22324 representing the constant.
22325 (tree_output_constant_def): Adjust.
22327 2013-02-14 Dodji Seketeli <dodji@redhat.com>
22330 * asan.c (instrument_builtin_call): Really put the length of the
22331 second source argument into src1_len.
22333 2013-02-13 Jakub Jelinek <jakub@redhat.com>
22335 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
22336 argument. If it is false, don't create edge from then_bb to
22338 (insert_if_then_before_iter): Pass true to it.
22339 (build_check_stmt): Pass false to it.
22340 (transform_statements): Flush hash table only on extended basic
22341 block boundaries, rather than at the beginning of every bb.
22342 Don't flush hash table on nonfreeing_call_p calls.
22343 * tree-flow.h (nonfreeing_call_p): New prototype.
22344 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
22346 2013-02-13 David S. Miller <davem@davemloft.net>
22348 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
22350 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
22353 * ira.c (max_regno_before_ira): Move from ...
22355 (fix_reg_equiv_init): Use max_regno_before_ira instead of
22358 2013-02-13 Jakub Jelinek <jakub@redhat.com>
22360 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
22362 2013-02-13 Richard Biener <rguenther@suse.de>
22365 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
22366 globals in MEM_REFs.
22368 2013-02-13 Richard Biener <rguenther@suse.de>
22370 * loop-init.c (loop_optimizer_init): Clear loop state when
22371 re-initializing preserved loops.
22372 * loop-unswitch.c (unswitch_single_loop): Return whether
22373 we unswitched the loop. Do not verify loop state here.
22374 (unswitch_loops): When we unswitched a loop discover new loops.
22376 2013-02-13 Kostya Serebryany <kcc@google.com>
22378 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
22380 * sanitizer.def: Rename __asan_init to __asan_init_v1.
22382 2013-02-12 Dodji Seketeli <dodji@redhat.com>
22384 Avoid instrumenting duplicated memory access in the same basic block
22385 * Makefile.in (asan.o): Add new dependency on hash-table.h
22386 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
22387 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
22388 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
22389 (free_mem_ref_resources, has_mem_ref_been_instrumented)
22390 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
22391 (get_mem_ref_of_assignment): New functions.
22392 (get_mem_refs_of_builtin_call): Extract from
22393 instrument_builtin_call and tweak a little bit to make it fit with
22395 (instrument_builtin_call): Use the new
22396 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
22397 of is_gimple_builtin_call.
22398 (instrument_derefs, instrument_mem_region_access): Insert the
22399 instrumented memory reference into the hash table.
22400 (maybe_instrument_assignment): Renamed instrument_assignment into
22401 this, and change it to advance the iterator when instrumentation
22402 actually happened and return true in that case. This makes it
22403 homogeneous with maybe_instrument_assignment, and thus give a
22404 chance to callers to be more 'regular'.
22405 (transform_statements): Clear the memory reference hash table
22406 whenever we enter a new BB, when we cross a function call, or when
22407 we are done transforming statements. Use
22408 maybe_instrument_assignment instead of instrumentation. No more
22409 need to special case maybe_instrument_assignment and advance the
22410 iterator after calling it; it's now handled just like
22411 maybe_instrument_call. Update comment.
22413 2013-02-13 Richard Biener <rguenther@suse.de>
22415 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
22416 Fix loop discovery code.
22418 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
22420 PR inline-asm/56148
22421 * lra-constraints.c (process_alt_operands): Match early clobber
22422 operand with itself. Check conflicts with earlyclobber only if
22423 the operand is not reloaded. Prefer to reload conflicting operand
22424 if earlyclobber and matching operands are the same.
22426 2013-02-12 Richard Biener <rguenther@suse.de>
22429 * lto-streamer-out.c (write_symbol): Do not output symbols
22430 for hard register variables.
22432 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
22435 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
22436 (umulsidi3_insn, mulsidi3_insn): New insns.
22438 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
22440 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
22441 (struct tune_params): Add vec_costs field.
22442 * config/arm/arm.c (arm_builtin_vectorization_cost)
22443 (arm_add_stmt_cost): New functions.
22444 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
22445 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
22446 (arm_default_vec_cost): New struct of type cpu_vec_costs.
22447 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
22448 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
22449 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
22450 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
22452 2013-02-12 Richard Biener <rguenther@suse.de>
22455 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
22456 decls again if possible.
22458 2013-02-12 Richard Biener <rguenther@suse.de>
22460 PR middle-end/56288
22461 * tree-ssa.c (verify_ssa_name): Fix check, move
22462 SSA_NAME_IN_FREE_LIST check up.
22464 2013-02-12 Jakub Jelinek <jakub@redhat.com>
22465 Steven Bosscher <steven@gcc.gnu.org>
22467 PR rtl-optimization/56151
22468 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
22469 equal to op0 or op1, and last_insn pattern is CODE operation
22470 with MEM dest and one of the operands matches that MEM.
22472 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
22474 * doc/extend.texi: Document Function Multiversioning and "default"
22475 parameter string to target attribute.
22476 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
22477 target attribute parameter is "default".
22478 (ix86_compare_version_priority): Remove checks for target attribute.
22479 (ix86_mangle_function_version_assembler_name): Change error to sorry.
22480 Remove check for target attribute equal to NULL. Add assert.
22481 (ix86_generate_version_dispatcher_body): Change error to sorry.
22483 2013-02-11 Iain Sandoe <iain@codesourcery.com>
22484 Jack Howarth <howarth@bromo.med.uc.edu>
22485 Patrick Marlier <patrick.marlier@gmail.com>
22488 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
22489 define ENDFILE_SPEC as TM_DESTRUCTOR.
22490 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
22492 2013-02-11 Alexander Potapenko <glider@google.com>
22493 Jack Howarth <howarth@bromo.med.uc.edu>
22494 Jakub Jelinek <jakub@redhat.com>
22497 * config/darwin.c (cdtor_record): Rename ctor_record.
22498 (sort_cdtor_records): Rename sort_ctor_records.
22499 (finalize_dtors): New routine to sort destructors by
22500 priority before use in assemble_integer.
22501 (machopic_asm_out_destructor): Use finalize_dtors if needed.
22503 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
22505 PR rtl-optimization/56275
22506 * simplify-rtx.c (avoid_constant_pool_reference): Check that
22507 offset is non-negative and less than cmode size before
22508 calling simplify_subreg.
22510 2013-02-11 Richard Biener <rguenther@suse.de>
22512 PR tree-optimization/56264
22513 * cfgloop.h (fix_loop_structure): Adjust prototype.
22514 * loop-init.c (fix_loop_structure): Return the number of
22515 newly discovered loops.
22516 * tree-cfgcleanup.c (repair_loop_structures): When new loops
22517 are discovered, do a full loop-closed SSA rewrite.
22519 2013-02-11 Richard Biener <rguenther@suse.de>
22521 PR tree-optimization/56273
22522 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
22524 (check_array_ref): Fix missing newline in dumps.
22525 (search_for_addr_array): Likewise.
22527 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
22529 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
22531 2013-02-09 Jakub Jelinek <jakub@redhat.com>
22534 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
22536 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
22538 PR rtl-optimization/56246
22539 * lra-constraints.c (simplify_operand_subreg): Try to reuse
22541 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
22542 constraints are satisfied.
22544 2013-02-08 Jeff Law <law@redhat.com>
22547 * emit-rtl.c (reg_is_parm_p): New function.
22548 * regs.h (reg_is_parm_p): New prototype.
22549 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
22550 callee-clobbered registers.
22552 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
22555 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
22556 If there is no implicit builtin declaration, just return NULL.
22558 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
22560 * config/i386/sse.md (FMAMODEM): New mode iterator.
22561 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
22562 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
22564 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
22566 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
22567 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
22568 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
22570 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
22572 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
22573 (microblaze*-*-elf): Likewise.
22574 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
22576 * config/microblaze/microblaze-c.c: Add builtin defines for
22577 _LITTLE_ENDIAN and _BIG_ENDIAN.
22578 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
22579 add to TARGET_DEFAULT flags.
22580 Expand ASM_SPEC and LINK_SPEC.
22581 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
22582 * config/microblaze/microblaze.md: Update extendsidi2 and
22583 movdi_internal instructions to use low-order / high-order reg
22585 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
22586 options and inversemask / mask of LITTLE_ENDIAN.
22587 * config/microblaze/t-microblaze: Expand multilib options to
22588 include mlittle-endian (le) and update exceptions patterns.
22590 2013-02-08 Jakub Jelinek <jakub@redhat.com>
22592 PR rtl-optimization/56195
22593 * lra-constraints.c (get_reload_reg): Don't reuse regs
22594 if they have smaller mode than requested, if they have
22595 wider mode than requested, try to return a SUBREG.
22597 PR tree-optimization/56250
22598 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
22599 if type is unsigned and code isn't MULT_EXPR.
22601 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
22603 PR tree-optimization/56064
22604 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
22605 bits according to mode.
22606 * fixed-value.h (fixed_from_double_int)
22607 (const_fixed_from_double_int): Adjust comments.
22609 2013-02-08 Richard Biener <rguenther@suse.de>
22612 * lto-streamer.h (struct data_in): Remove current_file, current_line
22613 and current_col members.
22614 * lto-streamer-out.c (lto_output_location): Stream changed bits
22615 en-block for efficiency.
22616 * lto-streamer-in.c (clear_line_info): Remove.
22617 (lto_input_location): Cache current file, line and column
22618 globally via local statics. Read changed bits en-block.
22619 (input_function): Do not call clear_line_info.
22620 (lto_read_body): Likewise.
22621 (lto_input_toplevel_asms): Likewise.
22623 2013-02-08 Michael Matz <matz@suse.de>
22625 PR tree-optimization/52448
22626 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
22627 (nt_call_phase): New static.
22628 (add_or_mark_expr): Only mark accesses with newer phase than any
22630 (nonfreeing_call_p): New.
22631 (nt_init_block): Update nt_call_phase, mark blocks as visited.
22632 (nt_fini_block): Keep blocks marked as visited.
22633 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
22635 2013-02-08 Richard Biener <rguenther@suse.de>
22637 * ira.c (ira): Free broken dominator information.
22639 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
22641 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
22643 2013-02-08 Marek Polacek <polacek@redhat.com>
22645 * cfgloop.c (verify_loop_structure): Add more checking of headers.
22647 2013-02-08 Richard Biener <rguenther@suse.de>
22649 PR middle-end/56181
22650 * cfgloop.h (flow_loops_find): Adjust.
22651 (bb_loop_header_p): Declare.
22652 * cfgloop.c (bb_loop_header_p): New function split out from ...
22653 (flow_loops_find): ... here. Adjust function signature,
22654 support incremental loop structure update.
22655 (verify_loop_structure): Cleanup. Verify a loop is a loop.
22656 * cfgloopmanip.c (fix_loop_structure): Move ...
22657 * loop-init.c (fix_loop_structure): ... here.
22658 (apply_loop_flags): Split out from ...
22659 (loop_optimizer_init): ... here.
22660 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
22661 in incremental mode, only remove dead loops here.
22663 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
22666 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
22667 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
22668 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
22669 (*round<mode>3.libgcc): New insns for fixed-modes.
22670 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
22671 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
22672 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
22673 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
22674 implementations. Define to __builtin_avr_absFX,
22675 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
22676 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
22677 __builtin_avr_countlsFX, respectively.
22678 * config/avr/avr-c.c (target.h): Include it.
22679 (enum avr_builtin_id): New enum.
22680 (avr_resolve_overloaded_builtin): New static function.
22681 (avr_register_target_pragmas): Use it to set
22682 targetm.resolve_overloaded_builtin.
22683 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
22684 tree nodes used by DEF_BUILTIN.
22685 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
22686 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
22687 <AVR_BUILTIN_xxBITS>: Same.
22689 2013-02-08 Richard Biener <rguenther@suse.de>
22691 * cfgloop.c (verify_loop_structure): Properly handle
22692 a loop exiting to another loop header.
22693 * ira-int.h (ira_loops): Remove.
22694 * ira.c (ira_loops): Remove.
22695 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
22696 (do_reload): Use loop_optimizer_finalize.
22697 * ira-build.c (create_loop_tree_nodes): Use get_loops and
22698 number_of_loops to access the loop tree.
22699 (more_one_region_p): Likewise.
22700 (finish_loop_tree_nodes): Likewise.
22701 (rebuild_regno_allocno_maps): Likewise.
22702 (mark_loops_for_removal): Likewise.
22703 (mark_all_loops_for_removal): Likewise.
22704 (remove_unnecessary_regions): Likewise.
22705 (ira_build): Likewise.
22706 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
22708 2013-02-08 Richard Biener <rguenther@suse.de>
22710 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
22711 * ipa-pure-const.c (analyze_function): Avoid calling
22712 mark_irreducible_loops twice.
22713 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
22715 2013-02-07 David S. Miller <davem@davemloft.net>
22717 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
22719 * var-tracking.c (vt_add_function_parameter): Test the presence of
22720 HAVE_window_save properly and do not remap argument registers when
22721 we have a leaf function.
22723 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
22726 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
22728 * config/i386/i386.c (ix86_print_operand): Ditto.
22730 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
22732 * lra-constraints.c (process_alt_operands): Fix recently added comment.
22734 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
22736 PR rtl-optimization/56225
22737 * lra-constraints.c (process_alt_operands): Check that reload hard
22738 reg can hold value for strict_low_part.
22740 2013-02-07 Jakub Jelinek <jakub@redhat.com>
22743 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
22744 dwarf2out_end_function.
22745 (in_first_function_p, maybe_at_text_label_p,
22746 first_loclabel_num_not_at_text_label): New variables.
22747 (dwarf2out_var_location): In the first function find out
22748 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
22749 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
22752 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
22754 PR rtl-optimization/56178
22755 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
22756 SUBREG of a register. Tidy up related block of code.
22757 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
22758 note if the source is a register or a SUBREG of a register.
22760 2013-02-07 Jakub Jelinek <jakub@redhat.com>
22763 * config/rs6000/rs6000.md (ptrm): New mode attr.
22764 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
22765 call_value_indirect_aix<pttrsize>,
22766 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
22769 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
22771 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
22772 if -bnortl. Convert to strcmp and strncmp.
22774 2013-02-07 Alan Modra <amodra@gmail.com>
22777 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
22778 addresses won't wrap when offsetting.
22779 (rs6000_secondary_reload): Provide secondary reloads needed for
22780 wrapping LO_SUM addresses.
22782 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
22784 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
22785 MACH, just __MACH__.
22787 2013-02-06 Richard Biener <rguenther@suse.de>
22789 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
22790 instead of calling fix_loop_structure.
22792 2013-02-06 Jakub Jelinek <jakub@redhat.com>
22794 PR middle-end/56217
22795 * omp-low.c (use_pointer_for_field): Return false if
22796 lower_send_shared_vars doesn't generate any copy-out code.
22798 2013-02-06 Tom de Vries <tom@codesourcery.com>
22800 PR rtl-optimization/56131
22801 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
22802 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
22803 of the label is NULL. Add comment.
22805 2013-02-05 Jakub Jelinek <jakub@redhat.com>
22807 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
22810 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
22811 (STATIC_LIBTSAN_LIBS): Likewise.
22812 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
22813 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
22814 is defined, don't add anything else beyond that.
22815 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
22816 (LINK_COMMAND_SPEC): Use them.
22818 PR tree-optimization/56205
22819 * tree-stdarg.c (check_all_va_list_escapes): Return true if
22820 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
22821 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
22823 2013-02-05 Richard Biener <rguenther@suse.de>
22825 PR tree-optimization/53342
22826 PR tree-optimization/53185
22827 * tree-vectorizer.h (vect_check_strided_load): Remove.
22828 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
22829 not disallow peeling for vectorized strided loads.
22830 (vect_check_strided_load): Make static and simplify.
22831 (vect_analyze_data_refs): Adjust.
22832 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
22833 correctly when vectorizing strided loads.
22835 2013-02-05 Richard Biener <rguenther@suse.de>
22837 * doc/install.texi: Refer to ISL, not PPL.
22839 2013-02-05 Jan Hubicka <jh@suse.cz>
22841 PR tree-optimization/55789
22842 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
22844 2013-02-05 Jan Hubicka <jh@suse.cz>
22846 PR tree-optimization/55789
22847 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
22848 the dead call anyway.
22850 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
22853 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
22855 2013-02-04 Alexander Potapenko <glider@google.com>
22856 Jack Howarth <howarth@bromo.med.uc.edu>
22857 Jakub Jelinek <jakub@redhat.com>
22860 * config/darwin.c (sort_ctor_records): Stabilized qsort
22861 on constructor priority by using original position.
22862 (finalize_ctors): New routine to sort constructors by
22863 priority before use in assemble_integer.
22864 (machopic_asm_out_constructor): Use finalize_ctors if needed.
22866 2013-02-04 Jakub Jelinek <jakub@redhat.com>
22869 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
22870 about visibility on artificial decls.
22871 * config/sol2.c (solaris_assemble_visibility): Likewise.
22873 2013-02-04 Kai Tietz <ktietz@redhat.com>
22876 * config/i386/i386.c (function_value_ms_64): Add additional valtype
22877 argument and improve checking of return-argument types for 16-byte
22879 (ix86_function_value_1): Add additional valtype argument on call
22880 of function_value_64.
22881 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
22882 handling infunction_value_64 function.
22884 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
22886 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
22888 2013-02-04 Richard Biener <rguenther@suse.de>
22890 PR tree-optimization/56188
22891 * tree-ssa-structalias.c (label_visit): Consider case with
22892 initially non-empty points-to set.
22893 (perform_var_substitution): Dump node mapping and clean up.
22895 2013-02-04 Richard Guenther <rguenther@suse.de>
22898 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
22899 node prevail as last resort.
22900 (lto_symtab_merge_decls): Remove guard on LTRANS here.
22901 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
22903 2013-02-04 Richard Biener <rguenther@suse.de>
22905 PR tree-optimization/56113
22906 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
22908 (equiv_class_lookup_or_add): ... this.
22909 (label_visit): Adjust and fix error in previous patch.
22910 (perform_var_substitution): Adjust.
22912 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
22914 * config/sh/divtab.c: Fix formatting and comments throughout the file.
22915 * config/sh/sh4-300.md: Likewise.
22916 * config/sh/sh4a.md: Likewise.
22917 * config/sh/constraints.md: Likewise.
22918 * config/sh/sh.md: Likewise.
22919 * config/sh/netbsd-elf.h: Likewise.
22920 * config/sh/predicates.md: Likewise.
22921 * config/sh/sh-protos.h: Likewise.
22922 * config/sh/ushmedia.h: Likewise.
22923 * config/sh/linux.h: Likewise.
22924 * config/sh/sh.c: Likewise.
22925 * config/sh/superh.h: Likewise.
22926 * config/sh/elf.h: Likewise.
22927 * config/sh/sh4.md: Likewise.
22928 * config/sh/sh.h: Likewise.
22930 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22932 * config/pa/constraints.md: Adjust unused letters. Change "T"
22933 constraint to match_test floating_point_store_memory_operand().
22934 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
22935 (base14_operand): New.
22936 (floating_point_store_memory_operand): New.
22937 (integer_store_memory_operand): Revise to use base14_operand and
22938 reg_plus_base_memory_operand.
22939 (move_dest_operand): Allow symbolic_memory_operands.
22940 (symbolic_memory_operand): Check for LO_SOM.
22941 (symbolic_operand): Change default case to break.
22942 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
22943 CONST_DOUBLE values to be reloaded by putting them into memory when
22944 the destination is a floating point register.
22945 (movdf): Remove code to handle CONST_DOUBLE.
22947 (reload_indf_r1): New.
22948 (reload_insf_r1): New.
22949 Consistently use "Q" and "T" constraints with integer and floating
22950 point move instructions, respectively.
22951 (movdi): Remove FAIL.
22952 Change predicate for source operand unamed DImode move from
22953 general_operand to move_src_operand.
22954 (umulsidi3): Change predicate for destination operand to
22956 Likewise for similar unamed patterns.
22957 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
22958 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
22959 (hppa_legitimize_address): Simplify mask calculation.
22960 (pa_emit_move_sequence): Revised handling of secondary reloads from
22961 REG+D addresses for floating point loads and stores. Directly handle
22962 loading CONST0_RTX (mode) to a floating point register.
22963 (pa_secondary_reload): Handle reloading DF and SFmode constant values
22964 to floating point registers. Don't restrict secondary reloads to
22965 floating point registers to integer modes. Revise some comments and
22967 (TARGET_LEGITIMATE_ADDRESS_P): Define.
22968 (pa_legitimate_address_p): New.
22969 (pa_legitimize_reload_address): New.
22970 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
22971 (STRICT_REG_OK_FOR_BASE_P): New.
22972 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
22973 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
22975 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
22976 Andrew Dixie <andrewd@gentrack.com>
22978 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
22981 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
22983 * expmed.c (extract_bit_field_1): Pass the full width of the
22984 structure to get_best_reg_extraction_insn.
22986 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
22989 * configure.ac (use_cxa_atexit): Add AIX.
22990 * configure: Regenerate.
22992 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
22994 2013-02-01 Jakub Jelinek <jakub@redhat.com>
22997 * final.c (need_profile_function): New variable.
22998 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
22999 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
23000 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
23001 notes, targetm.asm_out.function_prologue doesn't emit anything,
23002 HAVE_prologue and profiler should be emitted before prologue,
23003 set need_profile_function instead of emitting it.
23004 (final_scan_insn): If need_profile_function, emit
23005 profile_function on the first NOTE_INSN_BASIC_BLOCK or
23006 NOTE_INSN_FUNCTION_BEG note.
23008 2013-02-01 Richard Henderson <rth@redhat.com>
23010 * config/rs6000/rs6000.md (smulditi3): New.
23013 * config/alpha/alpha.md (umulditi3): New.
23015 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
23017 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
23018 (ASM_OUTPUT_ALIGNED_LOCAL): New.
23020 2013-02-01 Richard Biener <rguenther@suse.de>
23022 PR tree-optimization/56113
23023 * tree-ssa-structalias.c (label_visit): Reduce work for
23024 single-predecessor nodes.
23026 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
23028 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
23029 range isn't testing for zero.
23031 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
23033 PR middle-end/56113
23034 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
23036 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
23037 Nick Clifton <nickc@redhat.com>
23039 * config/v850/constraints.md (Q): Define as a memory constraint.
23040 * config/v850/predicates.md (label_ref_operand): New predicate.
23041 (e3v5_shift_operand): New predicate.
23042 (ior_operator): New predicate.
23043 * config/v850/t-v850: Add e3v5 multilib.
23044 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
23045 (v850_gen_movdi): Prototype.
23046 * config/v850/v850.c: Add support for e3v5 architecture.
23047 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
23049 (construct_save_jarl): Add e3v5 long JARL support.
23050 (v850_adjust_insn_length): New function. Adjust length of call
23051 insns when using e3v5 instructions.
23052 (v850_gen_movdi): New function: Generate instructions to move a
23054 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
23055 (CPP_SPEC): Define __v850e3v5__ as appropriate.
23056 (TARGET_USE_FPU): Enable for e3v5.
23057 (CONST_OK_FOR_W): New macro.
23058 (ADJUST_INSN_LENGTH): Define.
23059 * config/v850/v850.md (UNSPEC_LOOP): Define.
23060 (attr cpu): Add v850e3v5.
23061 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
23062 (movdi): New pattern.
23063 (movdi_internal): New pattern.
23064 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
23065 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
23066 (cstoresf4): Likewise.
23067 (cstoredf4): Likewise.
23068 (insv): New pattern.
23069 (rotlso3_a): New pattern.
23070 (rotlsi3_b): New pattern
23071 (rotlsi3_v850e3v5): New pattern.
23072 (doloop_begin): New pattern.
23073 (fix_loop_counter): New pattern.
23074 (doloop_end): New pattern.
23075 (branch_normal): Add e3v5 long branch support.
23076 (branch_invert): Likewise.
23077 (branch_z_normal): Likewise.
23078 (branch_z_invert): Likewise.
23079 (branch_nz_normal): Likewise.
23080 (branch_nz_invert): Likewise.
23081 (call_internal_short): Add e3v5 register-indirect JARL support.
23082 (call_internal_long): Likewise.
23083 (call_value_internal_short): Likewise.
23084 (call_value_internal_long): Likewise.
23085 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
23086 (mloop): New option.
23087 * config.gcc: Add support for configuring v840e3v5 target.
23088 * doc/invoke.texi: Document new v850 specific command line options.
23090 2013-01-31 Paul Koning <ni1d@arrl.net>
23094 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
23095 children if parent is a class.
23096 (prune_unused_types_prune): Don't add DW_AT_declaration.
23098 2013-01-31 Richard Biener <rguenther@suse.de>
23100 PR tree-optimization/56157
23101 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
23102 match up operand with SLP child.
23104 2013-01-31 Jason Merrill <jason@redhat.com>
23107 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
23108 parameters the first time.
23109 (gen_scheduled_generic_parms_dies): Check completeness here.
23111 2013-01-31 Richard Biener <rguenther@suse.de>
23113 PR middle-end/53073
23114 * common.opt (faggressive-loop-optimizations): New flag,
23115 enabled by default.
23116 * doc/invoke.texi (faggressive-loop-optimizations): Document.
23117 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
23118 infer_loop_bounds_from_undefined by it.
23120 2013-01-31 Richard Biener <rguenther@suse.de>
23122 PR tree-optimization/56150
23123 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
23124 visit virtual operands.
23125 (find_uses_to_rename_bb): Likewise.
23127 2013-01-31 Richard Biener <rguenther@suse.de>
23129 PR tree-optimization/56150
23130 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
23131 mixed store non-store stmts.
23133 2013-01-30 Jakub Jelinek <jakub@redhat.com>
23136 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
23137 LIBASAN_EARLY_SPEC is defined.
23138 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
23139 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
23141 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
23144 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
23145 invalid args instead of ICEing on it.
23146 (ix86_valid_target_attribute_tree): Return error_mark_node if
23147 ix86_valid_target_attribute_inner_p failed.
23148 (ix86_valid_target_attribute_p): Return false only if
23149 ix86_valid_target_attribute_tree returned error_mark_node. Allow
23150 target("default") attribute.
23151 (sorted_attr_string): Change argument from const char * to tree,
23152 merge in all target attribute arguments rather than just one.
23153 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
23154 instead of free. Avoid using strcat.
23155 (ix86_mangle_function_version_assembler_name): Mangle
23156 target("default") as if no target attribute is present. Adjust
23157 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
23158 instead of xmalloc and XDELETEVEC instead of free.
23159 (ix86_function_versions): Don't return true if one of the decls
23160 doesn't have target attribute. If they don't and one of the decls
23161 is DECL_FUNCTION_VERSIONED, report an error. Adjust
23162 sorted_attr_string caller. Use XDELETEVEC instead of free.
23163 (ix86_supports_function_versions): Remove.
23164 (make_name): Fix up formatting.
23165 (make_dispatcher_decl): Remove resolver_name and its initialization.
23166 Avoid leaking memory.
23167 (is_function_default_version): Return true if there is
23168 target("default") attribute rather than no target attribute at all.
23169 (make_resolver_func): Avoid leaking memory.
23170 (ix86_generate_version_dispatcher_body): Likewise.
23171 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
23172 * target.def (supports_function_versions): Remove.
23173 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
23174 * doc/tm.texi: Regenerated.
23176 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
23178 PR rtl-optimization/56144
23179 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
23180 for values with side effects.
23182 2013-01-30 Richard Biener <rguenther@suse.de>
23184 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
23185 (sparseset_pop): Likewise.
23186 * cfganal.c (compute_idf): Likewise. Increase work-stack size
23187 to be able to use quick_push in the worker loop.
23189 2013-01-30 Marek Polacek <polacek@redhat.com>
23191 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
23193 2013-01-30 Richard Biener <rguenther@suse.de>
23196 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
23198 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
23200 PR tree-optimization/56064
23201 * fixed-value.c (fixed_from_double_int): New function.
23202 * fixed-value.h (fixed_from_double_int): New prototype.
23203 (const_fixed_from_double_int): New static inline function.
23204 * fold-const.c (native_interpret_fixed): New static function.
23205 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
23206 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
23207 (native_encode_fixed): New static function.
23208 (native_encode_expr) <FIXED_CST>: Use it.
23209 (native_interpret_int): Move double_int worker code to...
23210 * double-int.c (double_int::from_buffer): ...this new static method.
23211 * double-int.h (double_int::from_buffer): Prototype it.
23213 2013-01-30 Richard Biener <rguenther@suse.de>
23215 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
23216 New pointer-map and obstack.
23217 (init_alias_vars): Allocate pointer-map and obstack.
23218 (delete_points_to_sets): Free them.
23219 (find_what_var_points_to): Cache result.
23220 (find_what_p_points_to): Adjust for changed interface of
23221 find_what_var_points_to.
23222 (compute_points_to_sets): Likewise.
23223 (ipa_pta_execute): Likewise.
23225 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
23227 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
23228 * configure: Regenerate.
23229 * config.in: Regenerate.
23230 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
23231 #nobits/#progbits if supported.
23233 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
23236 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
23237 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
23238 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
23240 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
23242 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
23243 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
23245 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
23247 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
23248 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
23250 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
23252 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
23254 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
23255 * config/arm/cortex-a7.md: New bypasses using
23256 arm_mac_accumulator_is_result.
23258 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
23260 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
23261 (cortex_a7_neon_mla): Likewise.
23262 (cortex_a7_fpfmad): New reservation.
23263 (cortex_a7_fpmacs): Use ffmas and update required units.
23264 (cortex_a7_fpmuld): Update required units and latency.
23265 (cortex_a7_fpmacd): Likewise.
23266 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
23267 (cortex_a7_neon). Likewise.
23268 (bypass) Update participating units.
23270 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
23272 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
23273 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
23275 * config/arm/vfp11.md (vfp_farith): Use ffmas.
23276 (vfp_fmul): Use ffmad.
23277 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
23278 (cortex_r4_fmacd): Use ffmad.
23279 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
23280 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
23281 (cortex_a9_fmacd): Use ffmad.
23282 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
23283 (cortex_a8_vfp_macd): Use ffmad.
23284 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
23285 (cortex_a5_fpmacd): Use ffmad.
23286 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
23287 (cortex_a15_vfp_macd): Use ffmad.
23288 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
23290 2013-01-29 Jason Merrill <jason@redhat.com>
23293 * varasm.c (default_assemble_visibility): Don't warn about
23294 visibility on artificial decls.
23296 2013-01-29 Richard Biener <rguenther@suse.de>
23298 PR tree-optimization/56113
23299 * tree-ssa-structalias.c (equiv_class_lookup): Also return
23301 (label_visit): Free duplicate bitmaps and record the leader instead.
23302 (perform_var_substitution): Adjust.
23304 2013-01-29 Richard Biener <rguenther@suse.de>
23306 PR tree-optimization/55270
23307 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
23308 the CFG, schedule loops for fixup.
23310 2013-01-29 Nick Clifton <nickc@redhat.com>
23312 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
23315 2013-01-28 Leif Ekblad <leif@rdos.net>
23317 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
23318 * config/i386/i386.h (TARGET_RDOS): New macro.
23319 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
23320 * config/i386/i386.c (ix86_option_override_internal): For 64bit
23321 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
23322 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
23323 DEFAULT_LARGE_SECTION_THRESHOLD.
23324 * config/i386/i386.md (R14_REG, R15_REG): New constants.
23325 * config/i386/rdos.h: New file.
23326 * config/i386/rdos64.h: New file.
23328 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
23331 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
23334 2013-01-28 Jakub Jelinek <jakub@redhat.com>
23336 PR rtl-optimization/56117
23337 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
23338 call cselib_lookup_from_insn on the MEM before calling
23339 add_insn_mem_dependence.
23341 2013-01-28 Richard Biener <rguenther@suse.de>
23343 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
23344 to a stmt that didn't have one.
23345 (copy_phis_for_bb): Likewise for PHI arguments.
23346 (copy_debug_stmt): Likewise for debug stmts.
23348 2013-01-28 Richard Biener <rguenther@suse.de>
23350 PR tree-optimization/56034
23351 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
23352 (partition_builtin_p): Adjust.
23353 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
23354 it is the last partition.
23355 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
23356 up the vertex for the definition.
23357 (classify_partition): Classify whether a partition is a
23358 PKIND_REDUCTION, thus has uses outside of the loop.
23359 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
23360 Merge all PKIND_REDUCTION partitions into the last partition.
23361 (tree_loop_distribution): Seed partitions from reductions as well.
23363 2013-01-28 Jakub Jelinek <jakub@redhat.com>
23365 PR tree-optimization/56125
23366 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
23367 pow(x,c) into sqrt(x) * powi(x, n/2) or
23368 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
23369 optimizing for size.
23370 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
23371 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
23374 PR tree-optimization/56094
23375 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
23376 to UNKNOWN_LOCATION while gimplifying expr.
23378 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
23381 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
23382 operand 0 in movabs insn template for -masm=intel asm alternative.
23383 (*movabs<mode>_2): Ditto for operand 1.
23385 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
23388 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
23391 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23393 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
23396 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23398 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
23399 Undef to avoid warning.
23401 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
23403 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
23404 * configure: Regenerate.
23406 2013-01-25 Jakub Jelinek <jakub@redhat.com>
23408 PR tree-optimization/56098
23409 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
23410 for stmts with volatile ops.
23411 (cond_store_replacement): Don't optimize if assign has volatile ops.
23412 (cond_if_else_store_replacement_1): Don't optimize if either
23413 then_assign or else_assign have volatile ops.
23414 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
23417 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
23419 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
23421 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
23423 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
23424 missing ':' in asm example.
23426 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
23428 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
23429 entries into lane and laneq entries.
23430 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
23431 Remove AdvSIMD scalar modes.
23432 (aarch64_sq<r>dmulh_laneq<mode>): New.
23433 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
23435 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
23436 builtin implementations to relfect changes in RTL in aarch64-simd.md.
23437 * config/aarch64/iterators.md (VCOND): New.
23440 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
23443 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
23444 Add NULL LIBNAME argument to existing definitions.
23445 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
23446 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
23447 * config/avr/avr.c (DEF_BUILTIN): Same.
23448 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
23449 (avr_expand_builtin): Expand to a vanilla call if a libgcc
23450 implementation is available (DECL_ASSEMBLER_NAME is set).
23451 (avr_fold_absfx): New static function.
23452 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
23453 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
23454 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
23455 AVR_BUILTIN_ABSLLK.
23456 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
23457 (abshk, absk, abslk, absllk): Provide as static inline functions.
23459 2013-01-25 Marek Polacek <polacek@redhat.com>
23461 PR tree-optimization/56035
23462 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
23464 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
23466 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
23467 (*movtf_internal_rex64): Add (!o,C) alternative
23468 (*movxf_internal_rex64): Ditto.
23469 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
23471 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
23473 * doc/invoke.texi: fix typo.
23474 * doc/objc.texi: fix typo.
23476 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
23478 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
23479 for the first two alternatives.
23481 2013-01-24 Diego Novillo <dnovillo@google.com>
23483 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
23484 (ggc-zone.o): Remove.
23485 * configure.ac: Remove option --with-gc.
23486 * configure: Re-generate.
23487 * doc/install.texi: Remove documentation for --with-gc.
23488 * gengtype.c (write_enum_defn): Remove. Update all users.
23489 (write_Types_process_field): Remove generation of gt_e_* argument.
23490 (output_type_enum): Remove. Update all users.
23491 (write_enum_defn): Remove. Update all users.
23492 (enum alloc_zone): Remove. Update all users.
23493 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
23494 * ggc-common.c (ggc_splay_alloc): Remove first argument.
23495 Update all callers.
23496 (struct ptr_data): Remove field TYPE. Update all users.
23497 (gt_pch_note_object): Remove argument TYPE. Update all users.
23498 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
23500 * ggc-none.c (ggc_alloc_typed_stat): Remove.
23501 (struct alloc_zone): Remove.
23502 (ggc_internal_alloc_zone_stat): Remove.
23503 (ggc_internal_cleared_alloc_zone_stat): Remove.
23504 * ggc-page.c (ggc_alloc_typed_stat): Remove.
23505 (ggc_pch_count_object): Remove last argument. Update all users.
23506 (ggc_pch_alloc_object): Remove last argument. Update all users.
23507 (struct alloc_zone): Remove.
23508 * ggc-zone.c: Remove.
23509 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
23510 (struct alloc_zone): Remove.
23511 (ggc_alloc_typed_stat): Remove.
23512 (ggc_alloc_typed): Remove.
23513 (ggc_splay_alloc): Remove first argument.
23514 (rtl_zone): Remove. Update all users.
23515 (tree_zone): Remove. Update all users.
23516 (tree_id_zone): Remove. Update all users.
23517 (ggc_internal_zone_alloc_stat): Remove. Update all users.
23518 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
23519 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
23520 * tree-ssanames.c: Remove references to zone allocator in comments.
23522 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
23524 * config/avr/avr.c (avr_out_fract): Make register numbers that
23525 might be outside of source operand signed.
23527 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
23529 * config/i386/constraints.md (Yf): New constraint.
23530 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
23531 of f constraint to conditionaly disable x87 register preferences.
23532 (*movdf_internal): Ditto.
23533 (*movsf_internal): Ditto.
23535 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
23537 PR inline-asm/55934
23538 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
23539 that have operands with impossible constraints.
23540 Add a FIXME for a speed-up opportunity.
23541 * lra-constraints.c (process_alt_operands): Verify that a class
23542 selected from constraints on asms is valid for the operand mode.
23543 (curr_insn_transform): Remove incorrect comment.
23545 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
23547 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
23548 TOC operand is a valid symbol ref in the constant pool.
23550 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23552 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
23554 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
23557 * config/avr/stdfix.h: New file.
23558 * t-avr (stdfix-gcc.h): New rule to build it.
23559 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
23561 2013-01-23 Kostya Serebryany <kcc@google.com>
23563 * config/darwin.h: remove dependency on
23564 CoreFoundation (asan on Mac OS).
23566 2013-01-23 Jakub Jelinek <jakub@redhat.com>
23569 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
23570 instead of cmpdi_operand for first comparison operand.
23571 Don't assert that comparison operands aren't both constants.
23573 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
23575 * doc/install.texi (Downloading the Source): Update references to
23576 downloading separate components.
23578 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
23580 * doc/extend.texi (__int128): Improve grammar.
23582 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
23585 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
23586 alternative to (o,r).
23587 (*movdi_internal_rex64): Remove (!o,n) alternative.
23588 (DImode immediate->memory splitter): Remove.
23589 (DImode immediate->memory peephole2): Remove.
23590 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
23591 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
23592 alternative to (!o,*r).
23593 (*movtf_internal_sse): New pattern.
23594 (*movxf_internal_rex64): New pattern.
23595 (*movxf_internal): Disable for TARGET_64BIT.
23596 (*movdf_internal_rex64): Remove (!o,F) alternative.
23598 2013-01-22 Jakub Jelinek <jakub@redhat.com>
23600 PR middle-end/56074
23601 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
23602 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
23603 * tree-vect-loop-manip.c (find_loop_location): Also ignore
23604 stmt locations where LOCATION_LOCUS of the stmt location is
23605 UNKNOWN_LOCATION or BUILTINS_LOCATION.
23608 * config/i386/i386.md (UNSPEC_STOS): New.
23609 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
23610 *strsetqi_1): Add UNSPEC_STOS.
23612 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
23615 * doc/invoke.texi: Remove left over -Wsynth example.
23617 2013-01-21 Jakub Jelinek <jakub@redhat.com>
23619 PR tree-optimization/56051
23620 * fold-const.c (fold_binary_loc): Don't fold
23621 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
23622 a narrowing conversion, or widening conversion from signed
23625 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
23627 PR rtl-optimization/56023
23628 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
23629 dependent on debug instruction.
23631 2013-01-21 Martin Jambor <mjambor@suse.cz>
23633 PR middle-end/56022
23634 * function.c (allocate_struct_function): Call
23635 invoke_set_current_function_hook earlier.
23637 2013-01-21 Jakub Jelinek <jakub@redhat.com>
23639 * reload1.c (init_reload): Only initialize reload_obstack
23640 during the first call.
23642 2013-01-21 Marek Polacek <polacek@redhat.com>
23644 * cfgloop.c (verify_loop_structure): Fix up grammar.
23646 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
23648 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
23649 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
23651 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23654 * config/arm/marvell-pj4.md: Update copyright year.
23655 Fix up use of alu to alu_reg and simple_alu_imm.
23657 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
23659 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
23661 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
23664 * lra-constraints.c (curr_insn_transform): Don't reuse original
23665 insn for secondary memory move when memory mode should be different.
23667 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23669 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
23670 atomic_storedi_1): New patterns.
23672 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
23674 btver2 pipeline descriptions.
23675 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
23677 * config/i386/i386.md (btver2_decode): New type attributes.
23678 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
23680 * config/i386/btver2.md: New file describing btver2 pipelines.
23682 2013-01-19 Andrew Pinski <apinski@cavium.com>
23684 PR tree-optimization/52631
23685 * tree-ssa-sccvn (visit_use): Before looking up the original
23686 statement, try looking up the simplified expression.
23688 2013-01-19 Anthony Green <green@moxielogic.com>
23690 * config/moxie/moxie.c (moxie_expand_prologue): Set
23691 current_function_static_stack_size.
23693 2013-01-18 Jakub Jelinek <jakub@redhat.com>
23695 PR tree-optimization/56029
23696 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
23697 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
23699 2013-01-18 Sharad Singhai <singhai@google.com>
23701 PR tree-optimization/55995
23702 * dumpfile.c (dump_loc): Print location only if available.
23703 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
23705 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
23708 * lra-constraints.c (curr_insn_transform): Reuse original insn for
23709 secondary memory move.
23710 (inherit_reload_reg): Use rclass instead of cl for
23711 check_secondary_memory_needed_p.
23713 2013-01-18 Jakub Jelinek <jakub@redhat.com>
23715 PR middle-end/56015
23716 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
23717 the case where writing real complex part of target modifies op1.
23719 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
23721 * config/aarch64/aarch64-simd.md
23722 (aarch64_vcond_internal<mode>): Handle unordered cases.
23723 * config/aarch64/iterators.md (v_cmp_result): New.
23725 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
23726 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23728 * config/arm/marvell-pj4.md: New file.
23729 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
23730 * config/arm/arm.md (generic_sched): Add marvell_pj4.
23731 (generic_vfp): Likewise.
23732 * config/arm/arm-cores.def: Add marvell-pj4.
23733 * config/arm/arm-tune.md: Regenerate.
23734 * config/arm/arm-tables.opt: Regenerate.
23735 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
23736 * doc/invoke.texi: Document marvell-pj4.
23738 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
23740 * config/aarch64/arm_neon.h: Map scalar types to standard types.
23742 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
23747 * var-tracking.c (negative_power_of_two_p): New.
23748 (global_get_addr_cache, local_get_addr_cache): New.
23749 (get_addr_from_global_cache, get_addr_from_local_cache): New.
23750 (vt_canonicalize_addr): Rewrite using the above. Adjust the
23752 (vt_stack_offset_p): Remove.
23753 (vt_canon_true_dep): Always canonicalize loc's address.
23754 (clobber_overlapping_mems): Make sure we have a MEM.
23755 (local_get_addr_clear_given_value): New.
23756 (val_reset): Clear local cached entries.
23757 (compute_bb_dataflow): Create and release the local cache.
23758 Disable duplicate MEMs clobbering.
23759 (emit_notes_in_bb): Clobber MEMs likewise.
23760 (vt_emit_notes): Create and release the local cache.
23761 (vt_initialize, vt_finalize): Create and release the global
23762 cache, respectively.
23763 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
23765 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
23767 PR libmudflap/53359
23768 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
23769 not found in the symtab.
23771 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
23774 PR rtl-optimization/55547
23775 PR rtl-optimization/53827
23778 * alias.c (offset_overlap_p): New, factored out of...
23779 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
23780 the conservative special case for symbolic constants. Don't
23781 adjust zero sizes on alignment.
23783 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
23785 PR rtl-optimization/52573
23786 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
23787 REG_UNUSED for the same register.
23789 2013-01-17 Richard Biener <rguenther@suse.de>
23790 Marek Polacek <polacek@redhat.com>
23792 PR rtl-optimization/55833
23793 * loop-unswitch.c (unswitch_loops): Move loop verification...
23794 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
23795 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
23796 Set it to true when we're removing a loop from hierarchy tree in
23797 an irreducible region.
23798 (fix_bb_placements): Adjust caller.
23799 (fix_loop_placements): Likewise.
23801 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
23803 * config/avr/builtins.def (DEF_BUILTIN): Factor out
23804 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
23805 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
23806 Remove ID. Adjust comments.
23807 * config/avr/avr-c.c (avr_builtin_name): Remove.
23808 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
23809 * config/avr/avr.c (avr_tolower): New static function.
23810 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
23811 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
23812 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
23815 2013-01-17 Jan Hubicka <jh@suse.cz>
23817 PR tree-optimization/55273
23818 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
23820 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
23823 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
23824 store through atomic_store<mode>_1.
23825 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
23827 2013-01-17 Martin Jambor <mjambor@suse.cz>
23829 PR tree-optimizations/55264
23830 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
23831 for virtual methods.
23832 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
23833 virtual methods before inlining is over.
23834 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
23836 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
23839 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
23841 PR rtl-optimization/56005
23842 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
23843 pending reads for prefetch.
23845 2013-01-16 Ian Bolton <ian.bolton@arm.com>
23847 * config/aarch64/aarch64.md
23848 (*cstoresi_neg_uxtw): New pattern.
23849 (*cmovsi_insn_uxtw): New pattern.
23850 (*<optab>si3_uxtw): New pattern.
23851 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
23852 (*<optab>si3_insn_uxtw): New pattern.
23853 (*bswapsi2_uxtw): New pattern.
23855 2013-01-16 Richard Biener <rguenther@suse.de>
23857 * tree-inline.c (tree_function_versioning): Remove set but
23858 never used variable.
23860 2013-01-16 Richard Biener <rguenther@suse.de>
23862 PR tree-optimization/55964
23863 * tree-flow.h (rename_variables_in_loop): Remove.
23864 (rename_variables_in_bb): Likewise.
23865 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
23866 (copy_loop_before): Adjust and delete update-ssa status.
23867 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
23868 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
23869 (rename_variables_in_loop): Remove.
23870 (slpeel_update_phis_for_duplicate_loop): Likewise.
23871 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
23872 use available cfg machinery instead of duplicating it.
23873 Update PHI nodes and perform poor-mans SSA update here.
23874 (slpeel_tree_peel_loop_to_edge): Adjust.
23876 2013-01-16 Richard Biener <rguenther@suse.de>
23878 PR tree-optimization/54767
23879 PR tree-optimization/53465
23880 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
23881 (vrp_visit_phi_node): For PHI arguments coming via backedges
23882 drop all symbolical range information.
23883 (execute_vrp): Compute backedges.
23885 2013-01-16 Richard Biener <rguenther@suse.de>
23887 * doc/install.texi: Update CLooG and ISL requirements to
23890 2013-01-16 Christian Bruel <christian.bruel@st.com>
23893 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
23894 (broken_move): Handle UNSPECV_SP_SWITCH_B.
23895 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
23897 2013-01-16 DJ Delorie <dj@redhat.com>
23899 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
23900 (UNSPECV_SP_SWITCH_E): New.
23901 (sp_switch_1): Change to an unspec.
23902 (sp_switch_2): Change to an unspec. Don't use post-inc when we
23905 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
23907 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
23908 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
23909 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
23910 (expand_mem_thread_fence): Ditto.
23911 (expand_mem_signal_fence): Ditto.
23912 (expand_atomic_load): Ditto.
23913 (expand_atomic_store): Ditto.
23915 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
23917 PR rtl-optimization/55547
23918 PR rtl-optimization/53827
23921 * alias.c (memrefs_conflict_p): Set sizes to negative after
23924 2013-01-15 Jakub Jelinek <jakub@redhat.com>
23927 * function.c (thread_prologue_and_epilogue_insns): Always
23928 add crtl->drap_reg to set_up_by_prologue.set, even if
23929 stack_realign_drap is false.
23931 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
23933 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
23934 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
23935 *call): Fix indention.
23937 2013-01-15 Tom de Vries <tom@codesourcery.com>
23940 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
23943 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
23945 PR rtl-optimization/55153
23946 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
23948 2013-01-15 Martin Jambor <mjambor@suse.cz>
23950 PR tree-optimization/55920
23951 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
23952 accesses as grp_to_be_debug_replaced.
23954 2013-01-15 Jakub Jelinek <jakub@redhat.com>
23956 PR tree-optimization/55920
23957 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
23958 there is non-useless type conversion needed from debug rhs to lhs,
23959 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
23961 2013-01-15 Joseph Myers <joseph@codesourcery.com>
23962 Mikael Pettersson <mikpe@it.uu.se>
23965 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
23967 (ASM_OUTPUT_CASE_LABEL): Remove.
23968 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
23969 * final.c (shorten_branches): Update alignment of labels before
23970 jump tables if CASE_VECTOR_SHORTEN_MODE.
23972 2013-01-15 Richard Biener <rguenther@suse.de>
23975 * system.h: Do not include gmp.h for building host tools.
23977 2013-01-15 Richard Biener <rguenther@suse.de>
23979 PR middle-end/55882
23980 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
23981 account for bitpos when computing alignment.
23983 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
23985 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
23986 (ix86_target_macros_internal): Likewise.
23988 * config/i386/i386.c (m_CORE2I7): Removed.
23989 (m_CORE_HASWELL): New macro.
23990 (m_CORE_ALL): Likewise.
23991 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
23992 (initial_ix86_arch_features): Likewise.
23993 (processor_target_table): Initializations for Core avx2.
23994 (cpu_names): New names "core-avx2".
23995 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
23996 PROCESSOR_CORE_HASWELL.
23997 (ix86_issue_rate): New case.
23998 (ia32_multipass_dfa_lookahead): Likewise.
23999 (ix86_sched_init_global): Likewise.
24001 * config/i386/i386.h (TARGET_HASWELL): New macro.
24002 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
24003 (processor_type): New PROCESSOR_HASWELL.
24005 2013-01-15 Jakub Jelinek <jakub@redhat.com>
24007 PR tree-optimization/55955
24008 * tree-vect-loop.c (vectorizable_reduction): Give up early on
24009 *SHIFT_EXPR and *ROTATE_EXPR codes.
24011 PR tree-optimization/48766
24012 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
24013 -ftrapv disable -fwrapv.
24015 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
24018 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
24019 etc. to 1 and not to __flash.
24020 Use LL suffix for __INT24_MAX__ with -mint8.
24021 Use ULL suffix for __UINT24_MAX__ with -mint8.
24023 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
24025 * config/avr/avr-arch.h
24026 (struct base_arch_s): Use typedef avr_arch_t instead.
24027 (struct arch_info_s): Use typedef avr_arch_info_t instead.
24028 (struct mcu_type_s): Use typedef avr_mcu_t instead.
24029 * config/avr/avr.c: Same.
24030 * config/avr/avr-devices.c: Same.
24031 * config/avr/driver-avr.c: Same.
24032 * config/avr/gen-avr-mmcu-texi.c: Same.
24033 * config/avr/avr-mcus.def: Adjust comment.
24035 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
24037 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
24038 * config/aarch64/iterators.md (VALLDI): New.
24040 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
24041 Andi Kleen <ak@linux.intel.com>
24044 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
24045 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
24048 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
24050 * config/avr/avr-stdint.h: Remove trailing blanks.
24051 * config/avr/avr-log.h: Same.
24052 * config/avr/avr-arch.h: Same.
24053 * config/avr/avr-devices.c: Same.
24054 * config/avr/avr-dimode.md: Same.
24055 * config/avr/predicates.md: Same.
24056 * config/avr/avr-c.c: Same. And fix typo.
24058 * config/avr/avr-protos.h: Same. And:
24059 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
24060 (init_cumulative_args): Rename to avr_init_cumulative_args.
24061 (expand_prologue): Rename to avr_expand_prologue.
24062 (expand_epilogue): Rename to avr_expand_epilogue.
24063 (adjust_insn_length): Rename to avr_adjust_insn_length.
24064 (notice_update_cc): Rename to avr_notice_update_cc.
24065 (final_prescan_insn): Rename to avr_final_prescan_insn.
24066 * config/avr/avr.c: Same.
24067 * config/avr/avr.h: Same.
24068 * config/avr/avr.md: Remove trailing blanks.
24069 (prologue): Use avr_expand_prologue.
24070 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
24072 2013-01-14 Richard Biener <rguenther@suse.de>
24074 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
24075 verify_location, collect_subblocks): New functions.
24076 (verify_gimple_in_cfg): Verify that locations only reference
24077 BLOCKs in the functions BLOCK tree.
24079 2013-01-14 Richard Biener <rguenther@suse.de>
24081 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
24083 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
24085 (insert_out_of_ssa_copy_on_edge): Likewise.
24086 (rewrite_close_phi_out_of_ssa): Likewise.
24087 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
24089 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
24090 propagated constants.
24091 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
24094 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
24096 * config/avr/avr-modes.def: Add GPL copyright notice.
24098 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
24100 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
24101 MEMMODEL_MASK to determine memory model.
24102 (atomic_store<mode>): Ditto from operands[2].
24103 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
24105 2013-01-13 Jakub Jelinek <jakub@redhat.com>
24108 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
24109 (fold_gimple_assign): Don't call unshare_expr here.
24110 (fold_ctor_reference): Call unshare_expr.
24112 2013-01-13 Terry Guo <terry.guo@arm.com>
24114 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
24115 * doc/fragments.texi: Document MULTILIB_REUSE.
24116 * gcc.c (multilib_reuse): New internal spec.
24117 (set_multilib_dir): Also search multilib from multilib_reuse.
24118 * genmultilib (tmpmultilib3): Refactor code.
24119 (tmpmultilib4): Ditto.
24120 (multilib_reuse): New multilib argument.
24122 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
24124 * Makefile.in: Update copyright.
24126 2013-01-12 Tom de Vries <tom@codesourcery.com>
24128 PR middle-end/55890
24129 * calls.c (expand_call): Check if arg_nr is valid.
24131 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
24133 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
24134 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
24135 documentation. Add missing '__' in front of
24136 __builtin_ia32_packssdw256.
24138 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
24141 * config/s390/s390.c (s390_preferred_reload_class): Do not return
24142 NO_REGS for larl operands.
24143 (s390_reload_larl_operand): Use s390_load_address instead of
24146 2013-01-11 Richard Biener <rguenther@suse.de>
24148 * tree-cfg.c (verify_node_sharing_1): Split out from ...
24149 (verify_node_sharing): ... here.
24150 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
24152 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
24154 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
24155 Substitute TREECHECKING.
24156 * configure: Regenerate.
24157 * Makefile.in (TREECHECKING): New.
24159 2013-01-11 Richard Guenther <rguenther@suse.de>
24161 PR tree-optimization/44061
24162 * tree-vrp.c (extract_range_basic): Compute zero as
24163 value-range for __builtin_constant_p of function parameters.
24165 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
24167 Update copyright years.
24169 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
24171 PR rtl-optimization/55672
24172 * lra-eliminations.c (mark_not_eliminable): Permit addition with
24173 const to be eliminable.
24175 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
24177 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
24178 * configure: Regenerate.
24180 2013-01-10 Richard Biener <rguenther@suse.de>
24182 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
24184 2013-01-10 Richard Biener <rguenther@suse.de>
24187 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
24188 locations for virtual PHI arguments.
24189 (rewrite_update_phi_arguments): Likewise.
24191 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
24193 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
24196 2013-01-10 Jakub Jelinek <jakub@redhat.com>
24198 PR tree-optimization/55921
24199 * tree-complex.c (expand_complex_asm): New function.
24200 (expand_complex_operations_1): Call it for GIMPLE_ASM.
24202 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
24205 * config/s390/s390.c (s390_symref_operand_p)
24206 (s390_loadrelative_operand_p): Merge the two functions.
24207 (s390_check_qrst_address, print_operand_address): Add parameters
24208 to s390_loadrelative_operand_p invokation.
24209 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
24210 (s390_reload_larl_operand, s390_secondary_reload): Use
24211 s390_loadrelative_operand_p instead of s390_symref_operand_p.
24212 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
24214 2013-01-09 Mike Stump <mikestump@comcast.net>
24216 * dse.c (record_store): Remove unnecessary assert.
24218 2013-01-09 Jan Hubicka <jh@suse.cz>
24220 PR tree-optimization/55569
24221 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
24222 * cfgloop.h (scale_loop_profile): Likewise.
24224 2013-01-09 Jan Hubicka <jh@suse.cz>
24227 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
24229 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
24231 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
24233 PR middle-end/55114
24234 * expr.h (maybe_emit_group_store): Declare.
24235 * expr.c (maybe_emit_group_store): New function.
24236 * builtins.c (expand_builtin_int_roundingfn): Call it.
24237 (expand_builtin_int_roundingfn_2): Likewise.
24239 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
24241 PR rtl-optimization/55829
24242 * lra-constraints.c (match_reload): Add code for absent output.
24243 (curr_insn_transform): Add code for reloads of matched inputs
24246 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
24248 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
24249 attribute of movddup insn to DF.
24250 (*vec_interleave_lowv2df): Ditto.
24251 (vec_dupv2df): Ditto.
24253 2013-01-09 Jan Hubicka <jh@suse.cz>
24255 PR tree-optimiation/55875
24256 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
24257 EVERY_ITERATION parameter.
24258 (number_of_iterations_exit): Check if exit is executed every iteration.
24259 (idx_infer_loop_bounds): Similarly here.
24260 (n_of_executions_at_most): Simplify
24261 to only test for cases where statement is dominated by the
24262 particular bound; handle correctly the "postdominance" test.
24263 (scev_probably_wraps_p): Use max loop iterations info
24264 as a global bound first.
24266 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
24267 Nick Clifton <nickc@redhat.com>
24269 * config/v850/v850.md (cbranchsf4): New pattern.
24270 (cstoresf4): New pattern.
24271 (cbranchdf4): New pattern.
24272 (cstoredf4): New pattern.
24273 (movsicc): Disallow floating point comparisons.
24274 (cmpsf_le_insn): Fix order of operators.
24275 (cmpsf_lt_insn): Likewise.
24276 (cmpsf_eq_insn): Likewise.
24277 (cmpdf_le_insn): Likewise.
24278 (cmpdf_lt_insn): Likewise.
24279 (cmpdf_eq_insn): Likewise.
24280 (cmpsf_ge_insn): Use LE comparison.
24281 (cmpdf_ge_insn): Likewise.
24282 (cmpsf_gt_insn): Use LT comparison.
24283 (cmpdf_gt_insn): Likewise.
24284 (cmpsf_ne_insn): Delete pattern.
24285 (cmpdf_ne_insn): Delete pattern.
24286 * config/v850/v850.c (v850_gen_float_compare): Use
24287 gen_cmpdf_eq_insn for NE comparison.
24288 (v850_float_z_comparison_operator)
24289 (v850_float_nz_comparison_operator): Move from here ...
24290 * config/v850/predicates.md: ... to here. Move GT and GE
24291 comparisons into v850_float_z_comparison_operator.
24292 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
24294 (v850_float_nz_comparison_operator): Likewise.
24296 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24298 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
24299 with calls to gen_insvsi/gen_insvdi.
24301 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
24303 * config/i386/i386.c (initial_ix86_tune_features): Set up
24304 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
24306 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
24307 Jakub Jelinek <jakub@redhat.com>
24309 PR tree-optimization/48189
24310 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
24311 If nitercst is 0, don't predict the exit edge.
24313 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24315 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
24316 in asm_fprintf with reg_names.
24317 (aarch64_print_operand_address): Likewise.
24318 (aarch64_return_addr): Likewise.
24319 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
24321 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24323 * config/pa/pa.h (VAL_U6_BITS_P): Define.
24324 (INT_U6_BITS): Likewise.
24325 * config/pa/predicates.md (uint6_operand): New predicate.
24326 (shift5_operand, shift6_operand): Likewise.
24327 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
24329 (lshrdi3): Use shift6_operand.
24330 (shrpsi4, shrpdi4): New insn patterns.
24331 (extzv): Delete expander.
24332 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
24333 predicates in unamed zero extract patterns. Tighten common constraint.
24334 (extv): Delete expander.
24335 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
24336 predicates in unamed sign extract patterns. Tighten common constraint.
24337 (insv): Delete expander.
24338 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
24339 predicates in unamed insert patterns. Tighten common constraint.
24340 Change uint32_operand predicate to uint6_operand predicate in unamed
24341 DImode pattern to insert constant values of type 1...1xxxx.
24343 2013-01-04 Jan Hubicka <jh@suse.cz>
24345 PR tree-optimization/55823
24346 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
24349 2013-01-08 Jakub Jelinek <jakub@redhat.com>
24350 Uros Bizjak <ubizjak@gmail.com>
24352 PR rtl-optimization/55845
24353 * df-problems.c (can_move_insns_across): Stop scanning at
24354 volatile_insn_p source instruction or give up if
24355 across_from .. across_to range contains any volatile_insn_p
24358 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
24360 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
24361 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
24363 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
24364 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
24366 2013-01-08 Jakub Jelinek <jakub@redhat.com>
24369 * asan.c (asan_clear_shadow): New function.
24370 (asan_emit_stack_protection): Use it.
24372 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
24374 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
24375 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
24376 with tab instead of space.
24378 2013-01-08 Nick Clifton <nickc@redhat.com>
24380 * config/rl78/rl78.c (rl78_expand_prologue): Always select
24381 register bank 0 at the start of an interrupt handler.
24382 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
24385 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
24387 * config/aarch64/aarch64-simd.md
24388 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
24389 (aarch64_simd_bsl): Likewise.
24390 (aarch64_vcond_internal<mode>): Likewise.
24391 (vcond<mode><mode>): Likewise.
24392 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
24393 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
24395 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
24397 * config/aarch64/aarch64-builtins.c
24398 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
24400 2013-01-08 Martin Jambor <mjambor@suse.cz>
24403 * tree-sra.c (analyze_access_subtree): Return true also after
24404 potentially creating a debug-only replacement.
24406 2013-01-08 Jakub Jelinek <jakub@redhat.com>
24408 PR middle-end/55890
24409 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
24411 PR tree-optimization/54120
24412 * tree-vrp.c (range_fits_type_p): Don't allow
24413 src_precision < precision from signed vr to unsigned_p
24414 if vr->min or vr->max is negative.
24415 (simplify_float_conversion_using_ranges): Test can_float_p
24416 against CODE_FOR_nothing.
24418 2013-01-08 Jakub Jelinek <jakub@redhat.com>
24419 Richard Biener <rguenther@suse.de>
24421 PR middle-end/55851
24422 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
24423 types instead of just INTEGER_TYPE types.
24425 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
24427 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
24430 2013-01-07 Steve Ellcey <sellcey@mips.com>
24433 * config/mips/mips.opt: Change mad to mmad to match documentation.
24435 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
24438 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
24439 .progmemx.data now.
24441 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
24444 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
24445 (avr_addrspace_t): Add .section_name field.
24446 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
24448 (avr_addrspace): Same. Initialize .section_name. Remove last
24449 NULL entry. Put __memx into .progmemx.data.
24450 (progmem_section_prefix): Remove.
24451 (avr_asm_init_sections): No need to initialize progmem_section.
24452 (avr_asm_named_section): Use avr_addrspace[].section_name to get
24453 section name prefix.
24454 (avr_asm_select_section): Ditto. And use get_unnamed_section to
24455 retrieve the progmem section.
24456 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
24457 boundary to run over avr_addrspace[].
24458 (avr_register_target_pragmas): Ditto.
24460 2013-01-06 Jakub Jelinek <jakub@redhat.com>
24462 * varasm.c (output_constant_def_contents): For asan_protect_global
24463 protected strings, adjust DECL_ALIGN if needed, before testing for
24465 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
24466 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
24468 (output_object_block): For asan protected decls, emit asan padding
24469 after their contents.
24470 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
24471 (asan_finish_file): Test it here instead.
24473 2013-01-07 Nick Clifton <nickc@redhat.com>
24474 Matthias Klose <doko@debian.org>
24475 Doug Kwan <dougkwan@google.com>
24476 H.J. Lu <hongjiu.lu@intel.com>
24479 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
24481 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
24483 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
24485 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
24487 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
24489 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
24492 * doc/install.texi (Cross-Compiler-Specific Options): Document
24495 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
24497 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
24498 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
24499 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
24500 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
24501 vqmovun_high_s64): Fix source operand number and update copyright.
24503 2013-01-07 Richard Biener <rguenther@suse.de>
24505 PR middle-end/55890
24506 * gimple.h (gimple_call_builtin_p): New overload.
24507 * gimple.c (validate_call): New function.
24508 (gimple_call_builtin_p): Likewise.
24509 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
24510 Use gimple_call_builtin_p.
24511 (find_func_clobbers): Likewise.
24512 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
24513 (strlen_optimize_stmt): Likewise.
24515 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
24517 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
24518 (vld1q_dup_*): Likewise.
24519 (vld1_*): Likewise.
24520 (vld1q_*): Likewise.
24521 (vld1_lane_*): Likewise.
24522 (vld1q_lane_*): Likewise.
24524 2013-01-07 Richard Biener <rguenther@suse.de>
24526 * lto-streamer.h (LTO_minor_version): Bump to 2.
24528 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
24530 * config/aarch64/aarch64-protos.h
24531 (aarch64_const_double_zero_rtx_p): Rename to...
24532 (aarch64_float_const_zero_rtx_p): ...this.
24533 (aarch64_float_const_representable_p): New.
24534 (aarch64_output_simd_mov_immediate): Likewise.
24535 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
24536 move immediate case.
24537 * config/aarch64/aarch64.c
24538 (aarch64_const_double_zero_rtx_p): Rename to...
24539 (aarch64_float_const_zero_rtx_p): ...this.
24540 (aarch64_print_operand): Allow printing of new constants.
24541 (aarch64_valid_floating_const): New.
24542 (aarch64_legitimate_constant_p): Check for valid floating-point
24544 (aarch64_simd_valid_immediate): Likewise.
24545 (aarch64_vect_float_const_representable_p): New.
24546 (aarch64_float_const_representable_p): Likewise.
24547 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
24548 (aarch64_output_simd_mov_immediate): New.
24549 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
24550 (*movdf_aarch64): Likewise.
24551 * config/aarch64/constraints.md (Ufc): New.
24552 (Y): call aarch64_float_const_zero_rtx.
24553 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
24555 2013-01-07 Richard Biener <rguenther@suse.de>
24557 PR tree-optimization/55888
24558 PR tree-optimization/55862
24559 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
24560 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
24561 not if it is contained therein.
24563 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
24565 * config/avr/t-avr: Typo.
24567 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
24570 * config/avr/t-avr: Don't automatically rebuild
24571 $(srcdir)/config/avr/t-multilib
24572 $(srcdir)/config/avr/avr-tables.opt
24573 $(srcdir)/doc/avr-mmcu.texi
24574 (avr-mcus): New phony target to build them on request.
24575 (s-avr-mlib, s-avr-mmcu-texi): Remove.
24576 * avr/avr-mcus.def: Adjust comments.
24578 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
24580 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
24582 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
24584 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
24586 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
24588 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
24590 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
24592 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
24593 to generate profiling.
24594 * config/rs6000/aix64.h (LIB_SPEC): Same.
24596 2013-01-04 Andrew Pinski <apinski@cavium.com>
24598 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
24600 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
24602 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
24604 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
24606 (ix86_expand_move): Ditto.
24607 (ix86_zero_extend_to_Pmode): Ditto.
24608 (ix86_expand_call): Ditto.
24609 (ix86_expand_special_args_builtin): Ditto.
24610 (ix86_expand_builtin): Ditto.
24612 2013-01-04 Richard Biener <rguenther@suse.de>
24614 PR tree-optimization/55862
24615 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
24616 translating them through PHI nodes.
24618 2013-01-04 Martin Jambor <mjambor@suse.cz>
24620 PR tree-optimization/55755
24621 * tree-sra.c (sra_modify_assign): Do not check that an access has no
24622 children when trying to avoid producing a VIEW_CONVERT_EXPR.
24624 2013-01-04 Marek Polacek <polacek@redhat.com>
24626 PR middle-end/55859
24627 * opts.c (default_options_optimization): Clarify error message.
24629 2013-01-04 Richard Biener <rguenther@suse.de>
24631 PR middle-end/55863
24632 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
24635 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24638 * config/pa/pa.md (movsi): Revert previous change.
24639 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
24642 2013-01-03 Richard Henderson <rth@redhat.com>
24644 * config/i386/i386.c (ix86_expand_move): Always assign to op1
24645 after eliminating TLS symbols.
24647 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
24650 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
24651 * graphite-poly.c (debug_gmp_value): Likewise.
24653 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
24656 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
24657 selected code model, define __code_mode_small__, __code_model_medium__,
24658 __code_model_large__, __code_model_32__ or __code_model_kernel__.
24659 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
24660 xchg temporary register with %k. Declare temporary register as
24662 [__x86_64__]: For medium and large code models, preserve %rbx register.
24664 2013-01-03 Richard Biener <rguenther@suse.de>
24666 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
24667 (dump_subscript): Adjust.
24668 (finalize_ddr_dependent): Do not dump redundant info.
24669 (analyze_siv_subscript): Adjust.
24670 (subscript_dependence_tester): Likewise.
24671 (compute_affine_dependence): Likewise.
24673 2013-01-03 Richard Biener <rguenther@suse.de>
24676 2013-01-03 Richard Biener <rguenther@suse.de>
24678 PR tree-optimization/55857
24679 * tree-vect-stmts.c (vectorizable_load): Do not setup
24680 re-alignment for invariant loads.
24682 2013-01-02 Richard Biener <rguenther@suse.de>
24684 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
24685 invariant load do not generate a vector load from the scalar location.
24687 2013-01-03 Richard Biener <rguenther@suse.de>
24689 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
24690 for not vectorizing.
24691 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
24692 not build INDIRECT_REFs, call get_name once only.
24693 (vect_create_data_ref_ptr): Likewise. Dump base object kind
24694 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
24696 2013-01-03 Richard Biener <rguenther@suse.de>
24698 PR tree-optimization/55857
24699 * tree-vect-stmts.c (vectorizable_load): Do not setup
24700 re-alignment for invariant loads.
24702 2013-01-03 Richard Biener <rguenther@suse.de>
24705 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
24706 prefer a built-in decl.
24708 2013-01-03 Jakub Jelinek <jakub@redhat.com>
24710 * gcc.c (process_command): Update copyright notice dates.
24711 * gcov.c (print_version): Likewise.
24712 * gcov-dump.c (print_version): Likewise.
24714 PR rtl-optimization/55838
24715 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
24716 iv0.step, iv1.step and step.
24718 2013-01-03 Jakub Jelinek <jakub@redhat.com>
24719 Marc Glisse <marc.glisse@inria.fr>
24721 PR tree-optimization/55832
24722 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
24723 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
24724 integer_{one,zero}_node.
24726 2013-01-03 Jakub Jelinek <jakub@redhat.com>
24729 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
24730 * var-tracking.c (reverse_op): Don't add reverse ops to
24731 VALUEs that have already
24732 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
24734 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
24736 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
24738 2013-01-02 Teresa Johnson <tejohnson@google.com>
24740 * dumpfile.c (dump_loc): Print filename with location.
24741 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
24742 new location_t parameter to emit complete unroll message with
24743 new dump framework.
24744 (canonicalize_loop_induction_variables): Compute loops location
24745 and pass to try_unroll_loop_completely.
24746 * loop-unroll.c (report_unroll_peel): New function.
24747 (peel_loops_completely): Use new dump format with location
24748 for main dumpfile message, and invoke report_unroll_peel on success.
24749 (decide_unrolling_and_peeling): Ditto.
24750 (decide_peel_once_rolling): Remove old dumpfile message subsumed
24751 by report_unroll_peel.
24752 (decide_peel_completely): Ditto.
24753 (decide_unroll_constant_iterations): Ditto.
24754 (decide_unroll_runtime_iterations): Ditto.
24755 (decide_peel_simple): Ditto.
24756 (decide_unroll_stupid): Ditto.
24757 * cfgloop.c (get_loop_location): New function.
24758 * cfgloop.h (get_loop_location): Declare.
24760 2013-01-02 Sriraman Tallam <tmsriram@google.com>
24762 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
24765 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24767 PR middle-end/55198
24768 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
24769 BLKmode objects when EXPAND_MEMORY is specified.
24771 2013-01-02 Sriraman Tallam <tmsriram@google.com>
24773 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
24775 (fold_builtin_cpu): Do not share cpu model decls across statements.
24777 2013-01-02 Jason Merrill <jason@redhat.com>
24780 * tree.c (build_array_type_1): Revert earlier change.
24782 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
24784 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
24786 * config/aarch64/aarch64-tune.md: Re-generate.
24788 2013-01-02 Richard Biener <rguenther@suse.de>
24790 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
24791 invariant load do not generate a vector load from the scalar location.
24793 2013-01-02 Richard Biener <rguenther@suse.de>
24796 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
24797 * configure: Regenerate.
24799 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
24801 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
24802 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
24803 (expand_builtin_int_roundingfn_2): Keep the original target around
24804 for the fallback case.
24806 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
24808 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
24809 to be clear for sign changes.
24811 2013-01-01 Jan Hubicka <jh@suse.cz>
24813 * ipa-inline-analysis.c: Fix formatting.
24815 2013-01-01 Jakub Jelinek <jakub@redhat.com>
24817 PR tree-optimization/55831
24818 * tree-vect-loop.c (get_initial_def_for_induction): Use
24819 gsi_after_labels instead of gsi_start_bb.
24821 Copyright (C) 2013 Free Software Foundation, Inc.
24823 Copying and distribution of this file, with or without modification,
24824 are permitted in any medium without royalty provided the copyright
24825 notice and this notice are preserved.