1 2013-09-26 Richard Biener <rguenther@suse.de>
3 * alias.h (component_uses_parent_alias_set): Rename to ...
4 (component_uses_parent_alias_set_from): ... this.
5 * alias.c (component_uses_parent_alias_set): Rename to ...
6 (component_uses_parent_alias_set_from): ... this and return
8 (reference_alias_ptr_type_1): Use the result from
9 component_uses_parent_alias_set_from instead of stripping
10 components one at a time.
11 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
13 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
15 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs): Move
17 * tree-ssa-ter.h: New File. Move prototypes here.
18 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
19 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
20 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
21 from is_replaceable_p.
22 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
23 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
24 (process_replaceable): Use ter_is_replaceable_p.
25 (find_replaceable_in_bb): Use ter_is_replaceable_p.
26 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
27 newly refactored ssa_is_replaceable_p.
28 * cfgexpand.c: Include tree-outof-ssa.h.
29 * ssaexpand.h: Delete.
31 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
33 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
34 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
36 (create_gimple_tmp): Delete.
37 (get_expr_type, build_assign, build_type_cast): Move to...
38 * gimple-builder.c: New File.
39 (get_expr_type): Relocate from gimple.c.
40 (build_assign, build_type_cast): Change to only create ssanames.
41 * gimple.h: Move prototypes to...
42 * gimple-builder.h: New File. Here.
43 * tree-ssa.h: And here.
44 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
45 count_uses_and_derefs): Relocate from gimple.c.
46 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
47 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
48 * tree-ssa-math-opts (execute_cse_reciprocals): Use
49 gimple_replace_ssa_lhs.
50 * asan.c: Include gimple-builder.h.
51 * Makefile.in: Add gimple-builder.o.
53 2013-09-26 Richard Biener <rguenther@suse.de>
55 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
56 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
57 (loe_visit_block): Use gcc_checking_assert.
58 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
59 gimple_assign_ssa_name_copy_p.
60 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
63 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
65 * config/rs6000/t-rs6000 (rs6000.o): Remove.
66 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
68 2013-09-26 Richard Biener <rguenther@suse.de>
70 PR tree-optimization/58539
71 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
72 the fact that debug statements are not taking part in loop-closed
75 2013-09-26 Nick Clifton <nickc@redhat.com>
77 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
79 (msp430_print_operand_raw): Delete unused letter parameter.
80 (TARGET_PRINT_OPERAND_ADDRESS): Define.
81 (msp430_print_operand_address): New function.
82 (msp430_print_operand): Move address printing code from here to
84 * config/msp430/msp430.md (movsipsi2): Add comment in generated
86 (zero_extendpsisi2): Likewise.
87 (extendpsisi2): New pattern.
88 (andneghi3): New pattern.
90 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
92 * config/aarch64/aarch64.opt (mlra): New option.
93 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
94 (TARGET_LRA_P): Define.
96 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
98 * expr.c (expand_assignment): Remove obsolete comment.
100 2013-09-25 Jeff Law <law@redhat.com>
102 * tree-flow.h (thread_through_all_blocks): Prototype moved into
103 tree-ssa-threadupdate.h.
104 (register_jump_thread): Similarly.
105 * tree-ssa-threadupdate.h: New header file.
106 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
107 * tree-vrp.c: Likewise.
108 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
109 (thread_around_empty_blocks): Change type of path vector argument to
110 an edge,type pair from just an edge. Initialize both elements when
111 appending to a jump threading path. Tweak references to elements
113 (thread_across_edge): Similarly. Release memory for the elements
115 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
116 (dump_jump_thread_path): New function broken out from
117 register_jump_thread.
118 (register_jump_thread): Use dump_jump_thread_path. Change type of
119 path vector entries. Search the path for NULL edges and dump
120 the path if one is found. Tweak the conversion of path to 3-edge
121 form to use the block copy type information embedded in the path.
123 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
125 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
127 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
128 Vladimir Makarov <vmakarov@redhat.com>
130 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
131 from the least significant bit.
132 (strip_address_mutations): Add bitfield operations handling.
133 (must_be_index_p): Add shifting and rotate operations handling.
134 (set_address_base): Use must_be_base_p predicate.
135 (set_address_index): Use must_be_index_p predicate.
137 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
138 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
139 Sergey Lega <sergey.s.lega@intel.com>
140 Anna Tikhonova <anna.tikhonova@intel.com>
141 Ilya Tocar <ilya.tocar@intel.com>
142 Andrey Turetskiy <andrey.turetskiy@intel.com>
143 Ilya Verbin <ilya.verbin@intel.com>
144 Kirill Yukhin <kirill.yukhin@intel.com>
145 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
147 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
149 (ix86_expand_vector_move_misalign): Support new unaligned load and
150 stores and use new names.
151 (CODE_FOR_sse2_storedqu): Rename to ...
152 (CODE_FOR_sse2_storedquv16qi): ... this.
153 (CODE_FOR_sse2_loaddqu): Rename to ...
154 (CODE_FOR_sse2_loaddquv16qi): ... this.
155 (CODE_FOR_avx_loaddqu256): Rename to ...
156 (CODE_FOR_avx_loaddquv32qi): ... this.
157 (CODE_FOR_avx_storedqu256): Rename to ...
158 (CODE_FOR_avx_storedquv32qi): ... this.
159 * config/i386/i386.md (fpint_logic): New.
160 * config/i386/sse.md (VMOVE): Extend for AVX512.
164 (VI_UNALIGNED_LOADSTORE): Ditto.
165 (sse2_avx_avx512f): Ditto.
166 (sse2_avx2): Extend for AVX512.
167 (sse4_1_avx2): Ditto.
169 (sse): Extend for AVX512.
172 (avxsizesuffix): Ditto.
173 (sseintvecmode): Ditto.
175 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
176 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
177 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
178 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
179 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
180 (<sse2_avx_avx512f>_storedqu<mode): ... this.
181 (<sse>_movnt<mode>): Replace constraint "x" with "v".
182 (STORENT_MODE): Extend for AVX512.
183 (*absneg<mode>2): Replace constraint "x" with "v".
184 (*mul<mode>3): Ditto.
185 (*ieee_smin<mode>3): Ditto.
186 (*ieee_smax<mode>3): Ditto.
187 (avx_cmp<mode>3): Replace VF with VF_128_256.
188 (*<sse>_maskcmp<mode>3_comm): Ditto.
189 (<sse>_maskcmp<mode>3): Ditto.
190 (<sse>_andnot<mode>3): Extend for AVX512.
191 (<code><mode>3, anylogic): Replace VF with VF_128_256.
192 (<code><mode>3, fpint_logic): New.
193 (*<code><mode>3): Extend for AVX512.
194 (avx512flogicsuff): New.
195 (avx512f_<logic><mode>): Ditto.
196 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
198 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
199 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
200 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
201 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
202 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
203 (xop_vpermil2<mode>3): Ditto.
204 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
205 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
206 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
207 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
209 2013-09-25 Tom Tromey <tromey@redhat.com>
211 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
212 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
213 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
214 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
215 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
216 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
217 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
218 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
219 (GRAPHITE_HTAB_H): Remove.
221 2013-09-25 Tom Tromey <tromey@redhat.com>
223 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
225 2013-09-25 Tom Tromey <tromey@redhat.com>
227 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
229 2013-09-25 Tom Tromey <tromey@redhat.com>
231 * config/i386/t-i386 (i386.o): Remove.
232 (i386-c.o): Use COMPILE and POSTCOMPILE.
234 2013-09-25 Tom Tromey <tromey@redhat.com>
236 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
238 2013-09-25 Tom Tromey <tromey@redhat.com>
240 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
241 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
242 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
243 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
244 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
245 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
246 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
247 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
248 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
249 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
250 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
251 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
252 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
253 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
254 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
255 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
256 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
257 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
258 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
259 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
260 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
261 (tree-ssa-pre.o, tree-ssa-sccvn.o)
262 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
263 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
264 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
265 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
266 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
267 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
268 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
269 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
270 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
271 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
272 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
273 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
274 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
275 (graphite-clast-to-gimple.o, graphite-dependences.o)
276 (graphite-interchange.o, graphite-poly.o)
277 (graphite-scop-detection.o, graphite-sese-to-poly.o)
278 (graphite-optimize-isl.o, tree-vect-loop.o)
279 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
280 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
281 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
282 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
283 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
284 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
285 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
286 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
287 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
288 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
289 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
290 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
291 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
292 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
293 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
294 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
295 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
296 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
297 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
298 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
299 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
300 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
301 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
302 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
303 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
304 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
305 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
306 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
307 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
308 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
309 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
310 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
311 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
312 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
313 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
314 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
315 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
316 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
317 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
318 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
319 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
320 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
321 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
322 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
323 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
324 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
325 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
326 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
327 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
328 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
329 (gcov-dump.o): Remove.
330 (default-c.o): Use COMPILE and POSTCOMPILE.
331 (CFLAGS-gcc.o): New variable.
332 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
334 2013-09-25 Tom Tromey <tromey@redhat.com>
336 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
337 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
338 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
339 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
340 (c-family/c-pragma.o, c-family/c-pretty-print.o)
341 (c-family/c-semantics.o, c-family/c-ada-spec.o)
342 (c-family/array-notation-common.o, c-family/stub-objc.o)
343 (c-family/c-ubsan.o): Remove.
345 2013-09-25 Tom Tromey <tromey@redhat.com>
347 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
349 2013-09-25 Tom Tromey <tromey@redhat.com>
351 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
352 to add -DENABLE_SHARED_LIBGCC.
353 (gcc.o): Don't use subshell.
355 2013-09-25 Tom Tromey <tromey@redhat.com>
357 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
358 * configure.ac: Don't invoke AM_PROG_CC_C_O.
359 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
360 * configure, config.in: Rebuild.
362 2013-09-25 Tom Tromey <tromey@redhat.com>
364 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
365 (COMPILE, POSTCOMPILE): New variables.
366 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
367 (DEPFILES): New variable.
369 * configure.ac: Add checks for dependency checking.
370 * configure, aclocal.m4: Regenerate.
372 2013-09-25 Tom Tromey <tromey@redhat.com>
374 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
375 ($(ALL_HOST_OBJS)): Move order-only dependency to end
378 2013-09-25 Tom Tromey <tromey@redhat.com>
380 * Makefile.in (generated_files): Add options.h,
381 target-hooks-def.h, insn-opinit.h,
382 common/common-target-hooks-def.h, pass-instances.def,
383 c-family/c-target-hooks-def.h.
385 2013-09-25 Jeff Law <law@redhat.com>
387 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
388 than foo[foo.length () - 1] to access last member in a vec.
389 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
391 2013-09-25 Richard Biener <rguenther@suse.de>
394 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
396 2013-09-25 Jan Hubicka <jh@suse.cz>
398 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
401 2013-09-25 Marek Polacek <polacek@redhat.com>
404 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
405 when determining the type name.
407 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
409 * config/sh/sh.md: Fix formatting.
411 2013-09-24 Xinliang David Li <davidxl@google.com>
413 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
414 Check max peel iterations parameter.
415 * param.def: New parameter.
416 * doc/invoke.texi: Document New parameter.
418 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
420 * gimple-pretty-print.c: Various whitespace tweaks.
421 * tree-core.h: Likewise.
422 * tree-pretty-print.c: Likewise.
423 * tree-ssa-alias.c: Likewise.
424 * tree-ssa-copy.c: Likewise.
425 * tree-ssanames.c: Likewise.
426 * tree-ssanames.h: Likewise.
427 * tree-vrp.c: Likewise.
429 2013-09-24 Alan Modra <amodra@gmail.com>
433 * stmt.c (expand_asm_operands): Call expand_expr with
434 EXPAND_MEMORY for output operands that disallow regs. Don't
435 use EXPAND_WRITE on inout operands.
437 2013-09-24 Richard Biener <rguenther@suse.de>
440 * tree.c (reference_alias_ptr_type): Move ...
441 * alias.c (reference_alias_ptr_type): ... here and implement
442 in terms of the new reference_alias_ptr_type_1.
443 (ref_all_alias_ptr_type_p): New helper.
444 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
445 use ref_all_alias_ptr_type_p.
446 (get_deref_alias_set): Add flag_strict_aliasing check here.
447 (reference_alias_ptr_type_1): New function, split out from ...
448 (get_alias_set): ... here.
449 (alias_ptr_types_compatible_p): New function.
450 * alias.h (reference_alias_ptr_type): Declare.
451 (alias_ptr_types_compatible_p): Likewise.
452 * tree.h (reference_alias_ptr_type): Remove.
453 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
454 to compare MEM_REF alias types.
456 2013-09-24 Richard Biener <rguenther@suse.de>
458 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
460 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
462 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
463 reload helper function arrays into a single array reg_addr.
464 (reload_fpr_gpr): Likewise.
465 (reload_gpr_vsx): Likewise.
466 (reload_vsx_gpr): Likewise.
467 (struct rs6000_reg_addr): Likewise.
468 (reg_addr): Likewise.
469 (rs6000_debug_reg_global): Change rs6000_vector_reload,
470 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
471 (rs6000_init_hard_regno_mode_ok): Likewise.
472 (rs6000_secondary_reload_direct_move): Likewise.
473 (rs6000_secondary_reload): Likewise.
475 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
476 constraints: wu, ww, and wy. Repurpose wv constraint added during
477 power8 changes. Put wg constraint in alphabetical order.
479 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
480 for future work to add ISA 2.07 VSX single precision support.
481 (-mvsx-scalar-double): Change default from -1 to 1, update
482 documentation comment.
483 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
484 (-mupper-regs-df): New debug switch to control whether DF values
485 can go in the traditional Altivec registers.
486 (-mupper-regs-sf): New debug switch to control whether SF values
487 can go in the traditional Altivec registers.
489 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
491 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
492 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
493 Add new constraints, wu/ww/wy. Repurpose wv constraint.
494 (rs6000_debug_legitimate_address_p): Print if we are running
495 before, during, or after reload.
496 (rs6000_secondary_reload): Add a comment.
497 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
499 * config/rs6000/constraints.md (wa constraint): Sort w<x>
500 constraints. Update documentation string.
501 (wd constraint): Likewise.
502 (wf constraint): Likewise.
503 (wg constraint): Likewise.
504 (wn constraint): Likewise.
505 (ws constraint): Likewise.
506 (wt constraint): Likewise.
507 (wx constraint): Likewise.
508 (wz constraint): Likewise.
509 (wu constraint): New constraint for ISA 2.07 SFmode scalar
511 (ww constraint): Likewise.
512 (wy constraint): Likewise.
513 (wv constraint): Repurpose ISA 2.07 constraint that we not used in
514 the previous submissions.
515 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
517 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
519 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
520 (REG_BR_PROB): Say that the probability is stored in an int_list.
521 * reg-notes.def: Update commentary to mention INT_LIST.
522 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
524 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
525 * rtlanal.c (int_reg_note_p): New function.
526 (alloc_reg_note): Assert that the note does not have an int argument.
527 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
528 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
529 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
530 rather than an INSN_LIST. Handle INT_LIST.
531 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
532 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
533 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
534 Manipulate them as ints rather than rtxes.
535 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
536 * regmove.c (copy_src_to_dest): Likewise.
537 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
539 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
540 into the cases that need it.
541 * config/arm/arm.c (arm_unwind_emit): Likewise.
543 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
544 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
545 * loop-doloop.c (add_test, doloop_modify): Likewise.
546 * loop-unswitch.c (compare_and_jump_seq): Likewise.
547 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
548 * predict.c (combine_predictions_for_insn): Likewise.
549 * print-rtl.c (print_rtx): Handle INT_LIST.
550 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
551 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
552 * config/arm/arm.c (emit_unlikely_jump): Likewise.
553 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
554 (ix86_split_fp_branch, predict_jump): Likewise.
555 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
556 * config/sh/sh.c (expand_cbranchsi4): Likewise.
557 * config/spu/spu.c (ea_load_store_inline): Likewise.
559 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
560 value of a REG_BR_PROB note.
561 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
562 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
563 * emit-rtl.c (try_split): Likewise.
564 * predict.c (br_prob_note_reliable_p): Likewise.
565 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
566 * reorg.c (mostly_true_jump): Likewise.
567 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
568 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
569 * config/i386/i386.c (ix86_print_operand): Likewise.
570 * config/ia64/ia64.c (ia64_print_operand): Likewise.
571 * config/mmix/mmix.c (mmix_print_operand): Likewise.
572 * config/rs6000/rs6000.c (output_cbranch): Likewise.
573 * config/s390/s390.c (s390_expand_tbegin): Likewise.
574 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
575 * config/sparc/sparc.c (output_cbranch): Likewise.
576 * config/spu/spu.c (get_branch_target): Likewise.
577 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
578 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
580 2013-09-23 Jan Hubicka <jh@suse.cz>
582 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
584 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
585 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
587 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
588 (gimple_fold_call): Dump inconsistent devirtualizations; add
589 sanity check for type based devirtualizations.
590 * ipa-prop.c: Include ipa-utils.h
591 (ipa_intraprocedural_devirtualization): Add sanity check.
592 (try_make_edge_direct_virtual_call): Likewise.
594 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
596 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
597 assignment statements.
599 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
601 * gimple-pretty-print.c (dump_ssaname_info): New function.
602 (dump_gimple_phi): Call it.
603 (pp_gimple_stmt_1): Likewise.
604 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
605 (range_info_def): Declare.
606 * tree-pretty-print.c (pp_double_int): New function.
607 (dump_generic_node): Call it.
608 * tree-pretty-print.h (pp_double_int): Declare.
609 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
610 * tree-ssanames.h (range_info_def): New structure.
611 (value_range_type): Move definition here.
612 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
614 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
616 (set_range_info): New function.
617 (get_range_info): Likewise.
618 (duplicate_ssa_name_range_info): Likewise.
619 (duplicate_ssa_name_fn): Check pointer type and call
620 duplicate_ssa_name_range_info.
621 * tree-ssa-copy.c (fini_copy_prop): Likewise.
622 * tree-vrp.c (value_range_type): Remove definition, now in
624 (vrp_finalize): Call set_range_info to update value range of
626 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
627 (SSA_NAME_RANGE_INFO): New macro.
629 2013-09-23 Richard Biener <rguenther@suse.de>
631 PR tree-optimization/58464
632 * tree-ssa-pre.c (phi_trans_lookup): Remove.
633 (phi_trans_add): Change to add conditionally on being not
635 (phi_translate_1): Remove recursion detection here.
636 (phi_translate): Pre-seed the cache with NULL to catch
637 recursion here in a more generic way.
638 (bitmap_find_leader): Adjust comment.
639 (get_representative_for): Dump value-numbers.
640 (create_expression_by_pieces): Likewise.
641 (insert_into_preds_of_block): Likewise.
643 2013-09-23 Christian Bruel <christian.bruel@st.com>
646 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
647 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
649 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
652 2013-09-20 Renlin Li <renlin.li@arm.com>
654 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
655 (aarch64_expand_epilogue): Likewise.
656 (aarch64_legitimize_reload_address): Likewise.
658 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
660 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
661 NULL_TREE before pushing them onto the vector. Likewise for labels.
663 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
665 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
666 comparison operators when -fno-trapping-math is in effect.
667 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
668 comparison operators in TFmode and assert that unsupported operators
670 (ia64_print_operand): Likewise.
672 2013-09-21 Jan Hubicka <jh@suse.cz>
674 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
675 (sse_typeless_stores): Enable for core
676 (sse_load0_by_pxor): Likewise.
677 (four_jump_limit): Disable for core.
678 (pad_returns): Likewise.
679 (avoid_vector_decode): Likewise.
680 (fuse_cmp_and_branch): Enable for cores.
681 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
683 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
686 * config/pa/pa.c (pa_option_override): Disable auto increment and
687 decrement instructions until reload is completed.
689 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
690 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
691 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
693 2013-09-20 DJ Delorie <dj@redhat.com>
694 Nick Clifton <nickc@redhat.com>
696 * config/rl78/rl78.c: Various whitespace and comment tweaks.
697 (need_to_save): Save bank 0 on interrupts.
698 (characterize_address): Strip far address wrappers.
699 (rl78_as_legitimate_address): Likewise.
700 (transcode_memory_rtx): Likewise.
701 (rl78_peep_movhi_p): Disable this peephole after devirt.
702 (rl78_propogate_register_origins): Forget all origins when a
704 * config/rl78/rl78-virt.md: Various whitespace tweaks.
705 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
707 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
708 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
709 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
711 * config/rl78/constraints.md: Various whitespace and paren tweaks.
713 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
715 * config/pa/pa.md: In "scc" insn patterns, change output template to
716 handle const0_rtx in reg_or_0_operand operands.
718 2013-09-20 Martin Husemann <martin@NetBSD.org>
721 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
722 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
724 2013-09-20 Richard Biener <rguenther@suse.de>
727 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
728 remembering SSA name version and block index.
729 (new_scev_info_str): Adjust.
730 (hash_scev_info): Likewise. Also hash the block index.
731 (eq_scev_info): Adjust.
732 (find_var_scev_info): Likewise.
733 (struct instantiate_cache_entry): Remove.
734 (struct instantiate_cache_type): Use a htab to map name, block
736 (instantiate_cache_type::~instantiate_cache_type): Adjust.
737 (get_instantiated_value_entry): Likewise.
738 (hash_idx_scev_info, eq_idx_scev_info): New functions.
739 (instantiate_scev_name): Adjust.
741 2013-09-20 Jeff Law <law@redhat.com>
743 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
745 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
747 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
748 Call aarch64_simd_expand_args to update op[argc].
750 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
752 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
755 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
757 * gengtype.c (file_rules): Added rule for *.cc files.
758 (get_output_file_with_visibility): Give fatal message when no
761 2013-09-20 Renlin Li <renlin.li@arm.com>
763 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
764 (aarch64_expand_epilogue): Likewise.
765 (aarch64_legitimize_reload_address): Likewise.
767 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
770 * expr.c (expand_assignment): Remove misalignp code path.
772 2013-09-20 Marek Polacek <polacek@redhat.com>
775 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
776 TYPE_PRECISION. Add asserts.
778 2013-09-20 Richard Biener <rguenther@suse.de>
780 PR tree-optimization/58453
781 * tree-loop-distribution.c (distribute_loop): Apply the cost
782 model for -ftree-loop-distribute-patterns, too.
784 2013-09-20 Richard Biener <rguenther@suse.de>
787 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
788 make type comparison less strict.
790 2013-09-20 Alan Modra <amodra@gmail.com>
792 * configure: Regenerate.
793 * aclocal.m4: Regenerate.
795 2013-09-20 Marek Polacek <polacek@redhat.com>
798 * doc/extend.texi: Document that attribute used is meant to be used
799 on variables with static storage duration.
801 2013-09-19 Jakub Jelinek <jakub@redhat.com>
803 PR tree-optimization/58472
804 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
805 simd_lane_access set inv_p = false.
806 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
807 the simduid magic VAR_DECL.
809 2013-09-19 Jan Hubicka <jh@suse.cz>
811 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
813 2013-09-17 Jeff Law <law@redhat.com>
815 * tree-ssa-dom.c (record_temporary_equivalences): New function
816 split out of dom_opt_dom_walker::after_dom_children.
817 (dom_opt_dom_walker::thread_across_edge): Move common code
818 in here from dom_opt_dom_walker::after_dom_children.
819 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
821 2013-09-19 Jan Hubicka <jh@suse.cz>
823 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
824 (TARGET_GENERIC): Use PROCESOR_GENERIC
825 (enum processor_type): Unify generic32 and 64.
826 * i386.md (cpu): Likewise.
827 * x86-tune.def (use_leave): Enable for generic32.
828 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
829 * athlon.md: Change generic64 to generic in all occurences.
830 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
831 (ix86_target_macros_internal): Likewise.
832 * driver-i386.c (host_detect_local_cpu): Likewise.
833 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename to ..
834 (generic_memcpy, generic_memset, generic_cost): This one.
835 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
836 (m_GENERIC32, m_GENERIC64): Remove.
837 (m_GENERIC): Turn into one flag.
838 (processor_target): Unify generic tunnings.
839 (ix86_option_override_internal): Replace generic32/64 by generic.
840 (ix86_issue_rate): Likewise.
841 (ix86_adjust_cost): Likewise.
843 2013-09-19 Jan Hubicka <jh@suse.cz>
845 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
848 2013-09-19 Jakub Jelinek <jakub@redhat.com>
850 * omp-low.c (expand_omp_sections): Always pass len - 1 to
851 GOMP_sections_start, even if !exit_reachable.
853 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
855 * lra-constraints.c (need_for_all_save_p): Use macro
856 HARD_REGNO_CALL_PART_CLOBBERED.
857 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
858 set up pseudo conflict hard regs.
860 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
863 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
866 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
868 PR rtl-optimization/58438
869 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
870 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
871 from previous subpasses.
873 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
875 * arm.c (arm_get_frame_offsets): Validate architecture supports
876 LDRD/STRD before accepting the tuning preference.
877 (arm_expand_prologue): Likewise.
878 (arm_expand_epilogue): Likewise.
880 2013-09-18 Richard Biener <rguenther@suse.de>
882 PR tree-optimization/58417
883 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
884 have chrecs with symbols defined in the loop as operands.
885 (chrec_fold_multiply): Likewise.
886 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
887 parameters before folding binary operations.
888 (struct instantiate_cache_entry_hasher): Remove.
889 (struct instantiate_cache_type): Use a pointer-map.
890 (instantiate_cache_type::instantiate_cache_type): New function.
891 (instantiate_cache_type::get): Likewise.
892 (instantiate_cache_type::set): Likewise.
893 (instantiate_cache_type::~instantiate_cache_type): Adjust.
894 (get_instantiated_value_entry): Likewise.
895 (global_cache): New global.
896 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
897 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
898 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
899 (instantiate_scev_name): Adjust.
900 (instantiate_scev): Construct global instead of local cache.
901 (resolve_mixers): Likewise.
903 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
904 Paolo Carlini <paolo.carlini@oracle.com>
907 * doc/implement-cxx.texi: Fix references to the C++ standards.
909 2013-09-18 Jakub Jelinek <jakub@redhat.com>
911 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
912 * tree-vect-data-refs.c (vect_analyze_data_refs): For
913 simd_lane_access drs, update also DR_ALIGNED_TO.
915 2013-09-18 Marek Polacek <polacek@redhat.com>
918 * doc/extend.texi: Document no_sanitize_undefined attribute.
919 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
920 no_sanitize_undefined attribute.
922 2013-09-18 Nick Clifton <nickc@redhat.com>
924 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
925 (ASM_DECLARE_FUNCTION_NAME): Define.
927 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
929 * compare-elim.c (find_comparison_dom_walker): New class
930 (find_comparisons_in_bb): Rename to
931 find_comparison_dom_walker::before_dom_children
932 (find_comparisons): Adjust
933 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
935 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
936 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
937 (init_walk_dominator_tree): Remove declaration.
938 (fini_walk_dominator_tree): Remove declaration.
939 * fwprop.c (single_def_use_dom_walker): New class
940 (single_def_use_enter_block): Convert to
941 single_def_use_dom_walker::before_dom_children.
942 (single_def_use_leave_block): Convert to
943 single_def_use_dom_walker::after_dom_children.
944 (build_single_def_use_links): Adjust.
945 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
947 (find_candidates_in_block): Convert to
948 find_candidates_dom_walker::before_dom_children.
949 (execute_strength_reduction): Adjust.
950 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
951 (sese_dom_walker): New class.
952 (sese_dom_walker::sese_dom_walker): New constructor.
953 (sese_dom_walker::~sese_dom_walker): New destructor.
954 (build_sese_conditions_before): Convert to
955 sese_dom_walker::before_dom_children.
956 (build_sese_conditions_after): Convert to
957 sese_dom_walker::after_dom_children.
958 (build_poly_scop): Adjust
959 * tree-into-ssa.c (rewrite_dom_walker): New class
960 (rewrite_enter_block): Convert to
961 rewrite_dom_walker::before_dom_children.
962 (rewrite_leave_block): Convert to
963 rewrite_dom_walker::after_dom_children.
964 (rewrite_update_dom_walker): New class.
965 (rewrite_update_enter_block): Convert to
966 rewrite_update_dom_walker::before_dom_children.
967 (rewrite_update_leave_block): Convert to
968 rewrite_update_dom_walker::after_dom_children.
969 (rewrite_blocks, rewrite_into_ssa): Adjust.
970 (mark_def_dom_walker): New class.
971 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
972 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
973 (mark_def_sites_blocks): Convert to
974 mark_def_dom_walker::before_dom_children.
975 (mark_def_site_blocks): Remove.
976 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
977 (tree_ssa_dominator_optimize): Adjust.
978 (dom_thread_across_edge): Convert to method
979 dom_opt_dom_walker::thread_across_edge.
980 (dom_opt_enter_block): Convert to member function
981 dom_opt_dom_walker::before_dom_children.
982 (dom_opt_leave_block): Convert to member function
983 dom_opt_dom_walker::after_dom_children.
984 * tree-ssa-dse.c (dse_dom_walker): New class.
985 (dse_enter_block): Convert to member function
986 dse_dom_walker::before_dom_children.
987 (tree_ssa_dse): Adjust.
988 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
989 (determine_invariantness_stmt): Convert to method
990 invariantness_dom_walker::before_dom_children.
991 (determine_invariantness): Remove
992 (move_computations_dom_walker): New class.
993 (move_computations_stmt): Convert to method
994 move_computations_dom_walker::before_dom_children.
995 (move_computations, tree_ssa_lim): Adjust.
996 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
997 (nt_init_block): Make method
998 notrappping_dom_walker::before_dom_children.
999 (nt_fini_block): Make
1000 method nontrapping_dom_walker::after_dom_children.
1001 (get_non_trapping): Adjust.
1002 * tree-ssa-pre.c (eliminate_dom_walker): New class.
1003 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
1004 (eliminate_leave_block): Make method.
1005 eliminate_dom_walker::after_dom_children.
1007 * tree-ssa-strlen.c (strlen_dom_walker): New class.
1008 (strlen_enter_block): Make method
1009 strlen_dom_walker::before_dom_children.
1010 (strlen_leave_block): Make
1011 method strlen_dom_walker::after_dom_children.
1012 (tree_ssa_strlen): Adjust.
1013 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
1014 (tree_ssa_uncprop): Adjust.
1015 (uncprop_leave_block): Make method
1016 uncprop_dom_walker::after_dom_children.
1017 (uncprop_leave_block): Make method
1018 uncprop_dom_walker::before_dom_children.
1020 2013-09-18 Bin Cheng <bin.cheng@arm.com>
1022 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn
1023 before branch in same basic block. Check both src and dest of
1026 2013-09-17 Nick Clifton <nickc@redhat.com>
1028 * config/rl78/rl78-real.md (bf): New pattern.
1030 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
1031 (rl78_print_operand): Do not put a # before a %B.
1032 * config/rl78/rl78.opt: Tweak doc strings.
1034 2013-09-17 DJ Delorie <dj@redhat.com>
1036 * config/rl78/constraints.md (Wcv): Allow up to $r31.
1037 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
1038 (rl78_option_override): Likewise, if -mallregs.
1039 (is_virtual_register): Likewise.
1040 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
1041 (REGNO_OK_FOR_BASE_P): Likewise.
1042 * config/rl78/rl78.opt (-mallregs): New.
1044 2013-09-17 Nick Clifton <nickc@redhat.com>
1046 * config/rl78/rl78.c (need_to_save): Change return type to bool.
1047 For interrupt functions: save all call clobbered registers if the
1048 interrupt handler is not a leaf function.
1049 (rl78_expand_prologue): Always recompute the frame information.
1050 For interrupt functions: only select bank 0 if one of the bank 0
1051 registers is going to be psuhed.
1053 2013-09-17 DJ Delorie <dj@redhat.com>
1055 * config/rl78/constraints.md: For each W* constraint, rename to C*
1056 and create a W* constraint that checks for an optional ES: prefix
1058 * config/rl78/rl78.md (UNS_ES_ADDR): New.
1059 (es_addr): New. Used to wrap far addresses.
1060 * config/rl78/rl78-protos.h (rl78_es_addr): New.
1061 (rl78_es_base): New.
1062 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
1063 wrapped far addresses.
1064 (rl78_print_operand_1): Unwrap far addresses before processing.
1065 (rl78_lo16): Wrap far addresses in unspecs.
1066 (rl78_es_addr): New.
1067 (rl78_es_base): New.
1068 (insn_ok_now): Check for not-yet-wrapped far addresses.
1069 (transcode_memory_rtx): Properly re-wrap far addresses.
1071 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
1073 * config/sparc/t-rtems: Add leon3 multilibs.
1075 2013-09-17 Cong Hou <congh@google.com>
1077 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
1078 when checking the dot production pattern. The type of rhs operand
1079 of multiply is now checked correctly.
1081 2013-09-17 Jeff Law <law@redhat.com>
1083 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
1084 edge implied equivalences into successor phis.
1085 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
1086 here from tree-ssa-threadedge.c.
1087 (mark_threaded_blocks): When threading through a joiner, if both
1088 successors of the joiner's clone reach the same block, verify the
1089 PHI arguments are equal. If not, cancel the jump threading request.
1090 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
1091 tree-ssa-threadupdate.c
1092 (thread_across_edge): Don't check PHI argument equality when
1093 threading through joiner block here.
1095 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
1097 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
1098 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
1099 (warn_uninit, warn_uninitialized_vars, execute_early_warn_uninitialized,
1100 make_pass_early_warn_uninitialized): Move to tree-ssa-uninit.c.
1101 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c
1102 (has_undefined_value_p): New. Pass dependant parts of
1103 ssa_undefined_value_p.
1104 (uninit_undefined_value_p): Use has_undefined_value_p.
1105 (warn_uninit, warn_uninitialized_vars, execute_early_warn_uninitialized,
1106 make_pass_early_warn_uninitialized): Move from tree-ssa.c
1107 * tree-ssa.h: Adjust prototypes
1109 2013-09-17 Jan Hubicka <jh@suse.cz>
1112 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
1113 * ipa-inline.c (can_inline_edge_p): Do not downgrade
1114 FUNCTION_NOT_OPTIMIZED.
1115 * ipa-inline-analysis.c (compute_inline_parameters): Function
1116 not optimized is not inlinable unless it is alwaysinline.
1117 (inline_analyze_function): Force calls in not optimized
1118 function not inlinable.
1120 2013-09-17 Jan Hubicka <jh@suse.cz>
1123 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
1125 * ipa.c (function_and_variable_visibility): Likewise.
1126 * ipa-profile.c (ipa_profile): Likewise.
1128 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
1131 * cgraph.c (cgraph_function_body_availability): Check for ifunc
1132 attribute, and don't inline the resolver in this case.
1134 2013-09-17 Teresa Johnson <tejohnson@google.com>
1136 * coverage.c (get_coverage_counts): Add missing newline.
1138 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1140 PR tree-optimization/58088
1141 * fold-const.c (mask_with_trailing_zeros): New function.
1142 (fold_binary_loc): Make sure we don't recurse infinitely
1143 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
1144 Use mask_with_trailing_zeros where appropriate.
1146 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
1148 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
1149 of 'prev' var to get better distance estimation.
1151 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
1153 * tree-inline.h (struct copy_body_data): Add transform_parameter.
1154 * tree-inline.c (is_parameter_of): New predicate.
1155 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
1156 a parameter has been remapped.
1157 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
1158 (optimize_inline_calls): Initialize transform_parameter.
1159 (copy_gimple_seq_and_replace_locals): Likewise.
1160 (tree_function_versioning): Likewise.
1161 (maybe_inline_call_in_expr): Likewise.
1163 2013-09-17 Nick Clifton <nickc@redhat.com>
1165 * config/msp430/msp430-protos.h: Add prototypes for new functions.
1166 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
1168 (is_attr_func): New function.
1169 (msp430_is_interrupt_func): New function.
1170 (is_naked_func): New function.
1171 (is_reentrant_func): New function.
1172 (is_critical_func): New function.
1173 (msp430_start_function): Add annotations for function attributes.
1174 (msp430_attr): New function.
1175 (msp430_attribute_table): New.
1176 (msp430_function_section): New function.
1177 (TARGET_ASM_FUNCTION_SECTION): Define.
1178 (msp430_builtin): New enum.
1179 (msp430_init_builtins): New function.
1180 (msp430_builtin_devl): New function.
1181 (msp430_expand_builtin): New function.
1182 (TARGET_INIT_BUILTINS): Define.
1183 (TARGET_EXPAND_BUILTINS): Define.
1184 (TARGET_BUILTIN_DECL): Define.
1185 (msp430_expand_prologue): Add support for naked, interrupt,
1186 critical and reentrant functions.
1187 (msp430_expand_epilogue): Likewise.
1188 (msp430_print_operand): Handle 'O' character.
1189 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
1191 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
1192 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
1193 (pushm): Use a 'n' rather than an 'i' constraint.
1194 (msp_return): Add generation of the interrupt return instruction.
1195 (disable_interrupts): New pattern.
1196 (enable_interrupts): New pattern.
1197 (push_intr_state): New pattern.
1198 (pop_intr_state): New pattern.
1199 (bic_SR): New pattern.
1200 (bis_SR): New pattern.
1201 * doc/extend.texi: Document MSP430 function attributes and builtin
1204 2013-09-17 Richard Biener <rguenther@suse.de>
1206 PR tree-optimization/58432
1207 * tree-loop-distribution.c (tree_loop_distribution): Also
1208 scan PHIs for outside loop uses and seed a partition from them.
1210 2013-09-17 Bin Cheng <bin.cheng@arm.com>
1212 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
1213 (restructure_reference): Call backtrace_base_for_ref.
1215 2013-09-17 Alan Modra <amodra@gmail.com>
1218 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11 patch.
1220 2013-09-16 DJ Delorie <dj@redhat.com>
1222 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
1223 vregs location for RL78/G10.
1224 (rl78_expand_prologue): Avoid SEL on G10.
1225 (rl78_expand_epilogue): Likewise.
1226 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
1227 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
1228 __RL78_G10__ when appropriate.
1229 (ASM_SPEC): Pass -mg10 along to the assembler.
1230 * config/rl78/rl78.md (sel_rb): Disable for G10.
1231 * config/rl78/rl78.opt: Add -mg10 option.
1232 * config/rl78/t-rl78: Add -mg10 multilib.
1234 2013-09-16 Xinliang David Li <davidxl@google.com>
1236 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
1237 * omp-low.c (omp_max_vf): Ditto.
1238 (expand_omp_simd): Ditto.
1239 * tree-vectorizer.c (vectorize_loops): Ditto.
1240 (gate_vect_slp): Ditto.
1241 (gate_increase_alignment): Ditto.
1242 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
1243 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
1244 (gate_tree_vectorize): Name change.
1245 (tree_vectorize): Ditto.
1246 (pass_vectorize::gate): Call new function.
1247 (pass_vectorize::execute): Ditto.
1248 opts.c: O3 default setting change.
1249 (finish_options): Check new flag.
1250 * doc/invoke.texi: Document new flags.
1251 * common.opt: New flags.
1253 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
1255 * doc/tm.texi.in (Cond Exec Macros): Remove node.
1256 (Condition Code): Don't reference it.
1257 * doc/tm.texi: Regenerate.
1259 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
1262 * lra-constraints.c (undo_optional_reloads): Consider all optional
1263 reload even if it did not get a hard reg.
1265 2013-09-16 Teresa Johnson <tejohnson@google.com>
1267 * dumpfile.c (dump_loc): Remove newline emission.
1268 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
1269 emission to dump_printf_loc calls where missing.
1270 (vect_mark_for_runtime_alias_test): Ditto.
1271 (vect_analyze_data_ref_dependence): Ditto.
1272 (vect_analyze_data_ref_dependences): Ditto.
1273 (vect_slp_analyze_data_ref_dependence): Ditto.
1274 (vect_slp_analyze_data_ref_dependences): Ditto.
1275 (vect_compute_data_ref_alignment): Ditto.
1276 (vect_update_misalignment_for_peel): Ditto.
1277 (vect_verify_datarefs_alignment): Ditto.
1278 (vector_alignment_reachable_p): Ditto.
1279 (vect_get_data_access_cost): Ditto.
1280 (vect_enhance_data_refs_alignment): Ditto.
1281 (vect_find_same_alignment_drs): Ditto.
1282 (vect_analyze_data_refs_alignment): Ditto.
1283 (vect_analyze_group_access): Ditto.
1284 (vect_analyze_data_ref_access): Ditto.
1285 (vect_analyze_data_ref_accesses): Ditto.
1286 (vect_prune_runtime_alias_test_list): Ditto.
1287 (vect_analyze_data_refs): Ditto.
1288 (vect_create_addr_base_for_vector_ref): Ditto.
1289 (vect_create_data_ref_ptr): Ditto.
1290 (vect_grouped_store_supported): Ditto.
1291 (vect_grouped_load_supported): Ditto.
1292 * value-prof.c (check_counter): Ditto.
1293 (check_ic_target): Ditto.
1294 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
1295 (vect_recog_widen_mult_pattern): Ditto.
1296 (vect_recog_widen_sum_pattern): Ditto.
1297 (vect_recog_over_widening_pattern): Ditto.
1298 (vect_recog_widen_shift_pattern): Ditto.
1299 (vect_recog_rotate_pattern): Ditto.
1300 (vect_recog_vector_vector_shift_pattern): Ditto.
1301 (vect_recog_divmod_pattern): Ditto.
1302 (vect_recog_mixed_size_cond_pattern): Ditto.
1303 (vect_recog_bool_pattern): Ditto.
1304 (vect_pattern_recog_1): Ditto.
1305 (vect_pattern_recog): Ditto.
1306 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
1307 (vect_is_simple_iv_evolution): Ditto.
1308 (vect_analyze_scalar_cycles_1): Ditto.
1309 (vect_get_loop_niters): Ditto.
1310 (vect_analyze_loop_1): Ditto.
1311 (vect_analyze_loop_form): Ditto.
1312 (vect_analyze_loop_operations): Ditto.
1313 (vect_analyze_loop_2): Ditto.
1314 (vect_analyze_loop): Ditto.
1315 (report_vect_op): Ditto.
1316 (vect_is_slp_reduction): Ditto.
1317 (vect_is_simple_reduction_1): Ditto.
1318 (vect_get_known_peeling_cost): Ditto.
1319 (vect_estimate_min_profitable_iters): Ditto.
1320 (vect_model_reduction_cost): Ditto.
1321 (vect_model_induction_cost): Ditto.
1322 (get_initial_def_for_induction): Ditto.
1323 (vect_create_epilog_for_reduction): Ditto.
1324 (vectorizable_reduction): Ditto.
1325 (vectorizable_induction): Ditto.
1326 (vectorizable_live_operation): Ditto.
1327 (vect_loop_kill_debug_uses): Ditto.
1328 (vect_transform_loop): Ditto.
1329 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
1330 (vect_stmt_relevant_p): Ditto.
1331 (process_use): Ditto.
1332 (vect_mark_stmts_to_be_vectorized): Ditto.
1333 (vect_model_simple_cost): Ditto.
1334 (vect_model_promotion_demotion_cost): Ditto.
1335 (vect_model_store_cost): Ditto.
1336 (vect_get_store_cost): Ditto.
1337 (vect_model_load_cost): Ditto.
1338 (vect_get_load_cost): Ditto.
1339 (vect_init_vector_1): Ditto.
1340 (vect_get_vec_def_for_operand): Ditto.
1341 (vect_finish_stmt_generation): Ditto.
1342 (vectorizable_call): Ditto.
1343 (vectorizable_conversion): Ditto.
1344 (vectorizable_assignment): Ditto.
1345 (vectorizable_shift): Ditto.
1346 (vectorizable_operation): Ditto.
1347 (vectorizable_store): Ditto.
1348 (vectorizable_load): Ditto.
1349 (vectorizable_condition): Ditto.
1350 (vect_analyze_stmt): Ditto.
1351 (vect_transform_stmt): Ditto.
1352 (vect_is_simple_use): Ditto.
1353 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
1354 (vect_can_advance_ivs_p): Ditto.
1355 (vect_update_ivs_after_vectorizer): Ditto.
1356 (vect_do_peeling_for_loop_bound): Ditto.
1357 (vect_gen_niters_for_prolog_loop): Ditto.
1358 (vect_update_inits_of_drs): Ditto.
1359 (vect_create_cond_for_alias_checks): Ditto.
1360 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
1361 (vect_build_slp_tree_1): Ditto.
1362 (vect_supported_load_permutation_p): Ditto.
1363 (vect_analyze_slp_instance): Ditto.
1364 (vect_analyze_slp): Ditto.
1365 (vect_make_slp_decision): Ditto.
1366 (vect_detect_hybrid_slp): Ditto.
1367 (vect_bb_vectorization_profitable_p): Ditto.
1368 (vect_slp_analyze_bb_1): Ditto.
1369 (vect_update_slp_costs_according_to_vf): Ditto.
1370 (vect_get_mask_element): Ditto.
1371 (vect_transform_slp_perm_load): Ditto.
1372 (vect_schedule_slp_instance): Ditto.
1373 (vect_schedule_slp): Ditto.
1374 (vect_slp_transform_bb): Ditto.
1375 * profile.c (read_profile_edge_counts): Ditto.
1376 (compute_branch_probabilities): Ditto.
1377 * coverage.c (get_coverage_counts): Ditto.
1379 2013-09-16 Diego Novillo <dnovillo@google.com>
1381 * tree-core.h: Add missing comment lines from refactoring
1384 2013-09-16 Jan Hubicka <jh@suse.cz>
1386 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
1387 abstract functions; for static functions check the presence
1390 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
1392 * config/aarch64/aarch64-simd-builtins.def (fma): New.
1393 * config/aarch64/aarch64-simd.md
1394 (aarch64_mla_elt<mode>): New.
1395 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
1396 (aarch64_mls_elt<mode>): Likewise.
1397 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
1398 (aarch64_fma4_elt<mode>): Likewise.
1399 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
1400 (aarch64_fma4_elt_to_128v2df): Likewise.
1401 (aarch64_fma4_elt_to_64df): Likewise.
1402 (fnma<mode>4): Likewise.
1403 (aarch64_fnma4_elt<mode>): Likewise.
1404 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
1405 (aarch64_fnma4_elt_to_128v2df): Likewise.
1406 (aarch64_fnma4_elt_to_64df): Likewise.
1407 * config/aarch64/iterators.md (VDQSF): New.
1408 * config/aarch64/arm_neon.h
1409 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
1410 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
1412 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
1414 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
1415 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
1416 (aarch64_mul3_elt_to_128df): Likewise.
1417 (aarch64_mul3_elt_to_64v2df): Likewise.
1418 * config/aarch64/iterators.md (VEL): Also handle DFmode.
1420 (VMUL_CHANGE_NLANES) Likewise.
1423 * config/aarch64/arm_neon.h
1424 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
1426 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
1428 * config/aarch64/arm_neon.h
1429 (vcvtx_high_f32_f64): Fix parameters.
1431 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1432 Uros Bizjak <ubizjak@gmail.com>
1434 * config/alpha.c: Include tree-ssanames.h.
1436 2013-09-16 Richard Biener <rguenther@suse.de>
1438 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
1439 (dot_rdg_1): Handle control_dd.
1440 (create_edge_for_control_dependence): New function.
1441 (create_rdg_edges): Add control dependences if asked for.
1442 (build_rdg): Likewise.
1443 (generate_loops_for_partition): If there are not necessary
1444 control stmts remove all their dependencies.
1445 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
1446 (distribute_loop): Pass on control dependences.
1447 (tree_loop_distribution): Compute control dependences and remove
1448 restriction on number of loop nodes.
1450 2013-09-16 Jakub Jelinek <jakub@redhat.com>
1452 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
1455 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
1457 * cse.c (try_const_anchors): Punt on CC modes.
1459 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1461 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
1463 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
1466 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
1469 2013-09-14 Iain Sandoe <iain@codesourcery.com>
1472 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
1474 (darwin_objc1_section): Likewise.
1475 (darwin_file_end): Emit Image Info section when required.
1477 2013-09-14 Jan Hubicka <jh@suse.cz>
1479 * tree-into-ssa.c (gate_into_ssa): New.
1480 (pass_data_build_ssa): Use it.
1481 * cgraph.h (expand_thunk): Update prototype.
1482 * cgraphunit.c (analyze_function): Expand thunks early.
1483 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
1484 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
1485 set lowered flag; do not add new function.
1486 (assemble_thunks_and_aliases): Update.
1487 * tree-ssa.c (gate_init_datastructures): New gate.
1488 (pass_data_init_datastructures): Use it.
1490 2013-09-14 Iain Sandoe <iain@codesourcery.com>
1493 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
1494 xmm register set described in the psABI.
1496 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
1498 * dwarf2out.c (should_emit_struct_debug): Add check
1499 for type_decl variable is not NULL.
1501 2013-09-13 Jacek Caban <jacek@codeweavers.com>
1503 * config.gcc: Use new winnt-c.c target hooks
1504 * config/t-winnt: New file
1505 * config/winnt-c.c: New file
1506 * doc/tm.texi.in: Document new hook
1507 * doc/tm.texi: Regenerated
1509 2013-09-13 Jan Hubicka <jh@suse.cz>
1512 * ipa-inline.c (check_callers): New function.
1513 (check_caller_edge): Remove.
1514 (want_inline_function_to_all_callers_p): Also permit alises that are
1516 (inline_to_all_callers): Terminate the walk when devirtualization
1517 introduce new calls.
1519 2013-09-13 Jan Hubicka <jh@suse.cz>
1521 * ipa-inline-analysis.c (struct growth_data): Add node.
1522 (do_estimate_growth_1): Fix detection of recursion.
1524 2013-09-13 Jakub Jelinek <jakub@redhat.com>
1526 PR tree-optimization/58392
1527 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
1528 to avoid shadowing of outer loop variable. If
1529 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
1530 replace_by_duplicate_decl simduid of loops that have it set and
1531 set dest_cfun->has_simduid_loops and/or
1532 dest_cfun->has_force_vect_loops.
1533 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
1534 instead of maybe_lookup_decl.
1535 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
1536 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
1537 call. Copy over force_vect and copy and remap simduid. Set
1538 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
1539 (copy_cfg_body): Remove blocks_to_copy argument. Use
1540 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
1541 caller. Don't set cfun->has_simduid_loops and/or
1542 cfun->has_force_vect_loops here.
1543 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
1545 (expand_call_inline, tree_function_versioning): Adjust copy_body
1548 2013-09-13 Martin Jambor <mjambor@suse.cz>
1551 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
1552 the assert if the edge was a speculative one.
1554 2013-09-13 Richard Biener <rguenther@suse.de>
1556 * tree-data-ref.h (known_dependences_p): Move here ...
1557 * tree-loop-distribution.c (known_dependences_p): ... from here.
1558 (dump_rdg_component, debug_rdg_component): Remove.
1560 (generate_loops_for_partition): Use gimple_uid instead of
1561 relying on matching stmt visit order.
1562 (rdg_build_partitions): Take starting stmt vector.
1563 (ldist_gen): Merge into ...
1564 (distribute_loop): ... this function. Do not compute starting
1566 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
1568 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1570 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
1571 Set type attribute correctly. Set predicable_short_it attribute.
1572 (cmpsi_shiftsi): Remove %? from output template.
1574 2013-09-13 Richard Biener <rguenther@suse.de>
1576 * tree-loop-distribution.c (struct rdg_component,
1577 rdg_defs_used_in_other_loops_p, free_rdg_components,
1578 rdg_build_components): Remove.
1579 (stmts_from_loop): Do not record virtual PHIs.
1580 (generate_loops_for_partition): Skip virtual PHIs.
1581 (build_rdg_partition_for_component): Rename to ...
1582 (build_rdg_partition_for_vertex): ... this and adjust.
1583 (rdg_build_partitions): Take a vector of starting vertices
1584 instead of components. Remove unnecessary leftover handling.
1585 (ldist_gen): Do not build components or record other stores.
1586 (distribute_loop): Do not distribute loops containing stmts
1589 2013-09-13 Christian Bruel <christian.bruel@st.com>
1592 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
1594 2013-09-13 Kai Tietz <ktietz@redhat.com>
1596 * config.gcc: Separate cases for mingw and cygwin targets,
1597 and add 64-bit cygwin target case.
1599 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
1600 dll-export inline-functions.
1601 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
1604 2013-09-13 Jeff Law <law@redhat.com>
1608 2013-09-06 Jeff Law <law@redhat.com>
1610 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
1611 edge implied equivalences into successor phis.
1613 2013-09-12 DJ Delorie <dj@redhat.com>
1615 * config/rl78/rl78-virt.md: Change from | to \; for asm line
1618 2013-09-12 Brooks Moses <bmoses@google.com>
1621 * Makefile.in: Do not install driver binaries in $(target)/bin.
1623 2013-09-12 DJ Delorie <dj@redhat.com>
1625 * config/rl78/rl78.opt (mrelax): New.
1626 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
1627 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
1629 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
1631 (rl78_expand_epilogue): Likewise.
1633 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
1636 * lra-eliminations.c (remove_reg_equal_offset_note): New.
1637 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
1638 pointer elimination with using remove_reg_equal_offset_note.
1640 2013-09-12 DJ Delorie <dj@redhat.com>
1642 * config/msp430/: New port.
1643 * config.gcc (msp430): Added.
1644 * doc/invoke.texi: Document MSP430 options.
1645 * doc/install.texi: Document msp430-elf
1646 * doc/md.texi: Document msp430-elf
1647 * doc/contrib.texi: Document msp430-elf
1649 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
1652 2013-09-12 Martin Jambor <mjambor@suse.cz>
1655 * ipa-prop.c (remove_described_reference): Give up if the edge in the
1656 reference descriptor is NULL.
1657 (ipa_edge_removal_hook): If owning a reference descriptor, set its
1660 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
1662 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
1663 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
1664 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
1665 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
1666 make_temp_ssa_name): move to tree-ssanames.h
1667 * tree-ssa-alias.h: Move prototype.
1668 * tree-ssa.h: Include tree-ssanames.h.
1669 * tree-ssanames.c (FREE_SSANAMES): Move to here.
1670 * tree-ssanames.h: New. Move items from tree-flow*.h
1671 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
1673 2013-09-12 Richard Biener <rguenther@suse.de>
1675 PR tree-optimization/58404
1676 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
1677 propagate non-invariant addresses into dereferences wrapped
1678 in component references.
1680 2013-09-12 Richard Biener <rguenther@suse.de>
1682 PR tree-optimization/58402
1683 * passes.def: Move pass_late_warn_uninitialized later.
1685 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
1687 * tree-ssa.h: New. Move content from tree-flow.h and
1689 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
1690 Move prototypes belonging to tree-ssa.c.
1691 * tree-flow-inline.h (redirect_edge_var_map_def,
1692 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
1694 * gimple.h: Adjust prototypes.
1695 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
1697 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
1698 * tree.h: Move prototype to tree-ssa.h.
1699 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
1700 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
1701 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
1702 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
1703 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
1704 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
1705 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
1706 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
1707 graphite-blocking.c, graphite-clast-to-gimple.c,
1708 graphite-dependences.c, graphite-interchange.c,
1709 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
1710 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
1711 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
1712 ipa-reference.c, ipa-split.c, ipa-utils.c,
1713 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
1714 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
1715 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
1716 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
1717 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
1718 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
1719 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
1720 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
1721 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
1722 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
1723 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
1724 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
1725 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
1726 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
1727 value-prof.c, var-tracking.c,
1728 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
1730 2013-09-12 Richard Biener <rguenther@suse.de>
1732 PR tree-optimization/58396
1733 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
1734 (build_rdg): Take a loop-nest parameter, fix memleaks.
1735 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
1737 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
1739 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
1742 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
1744 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
1745 The FMA instruction names should have a 'v' prefix.
1747 2013-09-12 Richard Biener <rguenther@suse.de>
1749 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
1750 (dot_rdg): Use popen instead of system in optional code.
1751 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
1752 (already_processed_vertex_p): Adjust.
1753 (has_anti_or_output_dependence, predecessor_has_mem_write,
1754 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
1755 rdg_flag_uses): Remove.
1756 (rdg_flag_vertex): Simplify.
1757 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
1759 (build_rdg_partition_for_component): Process the first vertex
1760 of a component only.
1761 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
1763 2013-09-12 Alan Modra <amodra@gmail.com>
1765 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
1767 2013-09-11 DJ Delorie <dj@redhat.com>
1768 Nick Clifton <nickc@redhat.com>
1770 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
1771 (rl78_stack_based_mem): New.
1772 * config/rl78/constraints.md (Iv08): New.
1781 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
1783 (movsi): Change from expand to insn-and-split.
1784 (ashrsi3): Clobber AX.
1788 * config/rl78/rl78.md (CC_REG): Fix.
1789 (addsi3): Allow memory and immediate operands.
1790 (addsi3_internal): Split into...
1791 (addsi3_internal_virt): ...new, and ...
1792 (addsi3_internal_real): ...new.
1794 (subsi3_internal_virt): New.
1795 (subsi3_internal_real): New.
1796 (mulsi3): Add memory operand.
1797 (mulsi3_rl78): Likewise.
1798 (mulsi3_g13): Likewise.
1799 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
1800 (cbranchqi4_real): Add more constraint options.
1801 (cbranchhi4_real): Expand pattern.
1802 (cbranchhi4_real_signed): New.
1803 (cbranchhi4_real_inverted): New.
1804 (cbranchsi4_real_lt): New.
1805 (cbranchsi4_real_ge): New.
1806 (cbranchsi4_real_signed): New.
1807 (cbranchsi4_real): New.
1809 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
1811 (lshrsi3_virt): Likewise.
1812 (ashlsi3_virt): Likewise.
1813 (cbranchqi4_virt_signed): New.
1814 (cbranchhi4_virt_signed): New.
1815 (cbranchsi4_virt): New.
1816 * config/rl78/rl78.c: Whitespace fixes throughout.
1817 (move_elim_pass): New.
1818 (pass_data_rl78_move_elim): New.
1819 (pass_rl78_move_elim): New.
1820 (make_pass_rl78_move_elim): New.
1821 (rl78_devirt_info): Run devirt earlier.
1822 (rl78_move_elim_info): New.
1823 (rl78_asm_file_start): Register it.
1824 (rl78_split_movsi): New.
1825 (rl78_as_legitimate_address): Allow virtual base registers when
1827 (rl78_addr_space_convert): Remove spurious debug stuff.
1828 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
1829 (rl78_print_operand): More cases for not printing '#'.
1830 (rl78_expand_compare): Remove most of the logic.
1831 (content_memory): New.
1832 (clear_content_memory): New.
1833 (get_content_index): New.
1834 (get_content_name): New.
1835 (display_content_memory): New.
1836 (update_content): New.
1837 (record_content): New.
1838 (already_contains): New.
1839 (insn_ok_now): Re-recog insns with virtual registers.
1840 (add_postponed_content_update): New.
1841 (process_postponed_content_update): New.
1842 (gen_and_emit_move): New.
1843 (transcode_memory_rtx): Record new location content.
1844 Use gen_and_emit_move.
1845 (force_into_acc): New.
1846 (move_to_acc): Use gen_and_emit_move.
1847 (move_from_acc): Likewise.
1848 (move_acc_to_reg): Likewise.
1849 (move_to_x): Likewise.
1850 (move_to_hl): Likewise.
1851 (move_to_de): Likewise.
1852 (rl78_alloc_physical_registers_op1): Record location content.
1853 (has_constraint): New.
1854 (rl78_alloc_physical_registers_op2): Record location content.
1856 (rl78_alloc_physical_registers_ro1): Likewise.
1857 (rl78_alloc_physical_registers_cmp): Likewise.
1858 (rl78_alloc_physical_registers_umul): Likewise.
1859 (rl78_alloc_address_registers_macax): New.
1860 (rl78_alloc_physical_registers): Initialize and set location
1861 content memory as needed.
1862 (rl78_reorg): Make sure split2 is called.
1863 (rl78_rtx_costs): New.
1865 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
1867 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
1868 for (not (neg ...)) and (neg (not ...)) cases.
1870 2013-09-11 Richard Biener <rguenther@suse.de>
1873 * passes.def: Split critical edges before late uninit warning passes.
1874 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
1876 2013-09-11 Jakub Jelinek <jakub@redhat.com>
1878 PR tree-optimization/58385
1879 * fold-const.c (build_range_check): If both low and high are NULL,
1880 use omit_one_operand_loc to preserve exp side-effects.
1882 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1884 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
1886 2013-09-11 Richard Biener <rguenther@suse.de>
1888 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
1889 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
1890 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
1891 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
1892 stmts_from_loop, known_dependences_p, build_empty_rdg,
1893 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
1894 * tree-loop-distribution.c: ... here.
1895 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
1896 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
1897 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
1898 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
1899 * tree-loop-distribution.c: ... here.
1900 * tree-loop-distribution.c: Include gimple-pretty-print.h.
1901 (struct partition_s): Add loops member.
1902 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
1903 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
1904 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
1906 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1907 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1908 Sergey Lega <sergey.s.lega@intel.com>
1909 Anna Tikhonova <anna.tikhonova@intel.com>
1910 Ilya Tocar <ilya.tocar@intel.com>
1911 Andrey Turetskiy <andrey.turetskiy@intel.com>
1912 Ilya Verbin <ilya.verbin@intel.com>
1913 Kirill Yukhin <kirill.yukhin@intel.com>
1914 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1916 * config/i386/constraints.md (k): New.
1918 * config/i386/i386.c (const regclass_map): Add new mask registers.
1919 (dbx_register_map): Ditto.
1920 (dbx64_register_map): Ditto.
1921 (svr4_dbx_register_map): Ditto.
1922 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
1924 (ix86_preferred_reload_class): Disable constants for mask registers.
1925 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
1926 (ix86_hard_regno_mode_ok): Support new mask registers.
1927 (x86_order_regs_for_local_alloc): Ditto.
1928 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
1929 (FIXED_REGISTERS): Add new mask registers.
1930 (CALL_USED_REGISTERS): Ditto.
1931 (REG_ALLOC_ORDER): Ditto.
1932 (VALID_MASK_REG_MODE): New.
1933 (FIRST_MASK_REG): Ditto.
1934 (LAST_MASK_REG): Ditto.
1935 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
1936 (MAYBE_MASK_CLASS_P): New.
1937 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
1938 (REG_CLASS_CONTENTS): Ditto.
1939 (MASK_REGNO_P): New.
1940 (ANY_MASK_REG_P): Ditto.
1941 (HI_REGISTER_NAMES): Add new mask registers.
1942 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
1943 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
1945 (attribute "type"): Add mskmov, msklog.
1946 (attribute "length_immediate"): Support them.
1947 (attribute "memory"): Ditto.
1948 (attribute "prefix_0f"): Ditto.
1949 (*movhi_internal): Support new mask registers.
1950 (*movqi_internal): Ditto.
1951 (define_split): Split out clobber pattern is a logic
1952 insn on mask registers.
1953 (*k<logic><mode>): New.
1954 (*andhi_1): Extend to support mask regs.
1955 (*andqi_1): Extend to support mask regs.
1957 (define_split): Split and-not to and and not if operands
1959 (*<code><mode>_1): Separate HI mode to new pattern...
1960 (*<code>hi_1): This.
1961 (*<code>qi_1): Extend to support mask regs.
1963 (kortestzhi): Ditto.
1964 (kortestchi): Ditto.
1966 (*one_cmpl<mode>2_1): Remove HImode and handle it...
1967 (*one_cmplhi2_1): ...Here, now with mask registers support.
1968 (*one_cmplqi2_1): Support new mask registers.
1969 (HI/QImode arithmetics splitter): Don't split if mask registers
1971 (HI/QImode not splitter): Ditto.
1972 * config/i386/predicated.md (mask_reg_operand): New.
1973 (general_reg_operand): Ditto.
1975 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1977 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
1978 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
1980 2013-09-10 Jeff Law <law@redhat.com>
1982 PR tree-optimization/58380
1983 * tree-ssa-threadupdate.c (thread_block): Recognize another case
1984 of threading through a buried loop header.
1986 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
1987 return value for single successor case.
1989 2013-09-10 Jan Hubicka <jh@suse.cz>
1991 * ipa-devirt.c (ipa_devirt): Enable with LTO.
1993 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
1996 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
1997 support conditional execution.
1998 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
2000 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
2002 * lra.c (lra): Clear lra_optional_reload_pseudos before every
2004 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
2005 Check destination too to check move insn.
2006 (undo_optional_reloads): Add check that the original peudo did not
2007 changed its allocation and the optional reload was inherited on last
2008 inheritance pass. Break loop after deciding to keep optional reload.
2009 (lra_undo_inheritance): Add check that inherited pseudo still in
2012 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
2014 * config/aarch64/aarch64.md (generic_sched): New.
2015 * config/aarch64/aarch64-generic.md (load): Make conditional
2016 on generic_sched attribute.
2017 (nonload): Likewise.
2019 2013-09-10 Jan Hubicka <jh@suse.cz>
2021 * lto-cgraph.c: Include ipa-utils.h.
2022 (compute_ltrans_boundary): Also add possible targets into the boundary.
2024 2013-09-10 Jan Hubicka <jh@suse.cz>
2026 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
2027 VAR_DECL of vtable rather than full expression.
2029 2013-09-10 Jan Hubicka <jh@suse.cz>
2030 Paolo Carlini <paolo.carlini@oracle.com>
2032 * cgraphunit.c (analyze_functions): Save input_location, set it
2033 to UNKNOWN_LOCATION and restore it at the end.
2035 2013-09-10 Martin Jambor <mjambor@suse.cz>
2037 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
2038 represented by a thunk.
2040 2013-09-10 Jeff Law <law@redhat.com>
2042 PR tree-optimization/58343
2043 * tree-ssa-threadupdate.c (thread_block): Identify and disable
2044 jump threading requests through loop headers buried in the middle
2045 of a jump threading path.
2047 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
2048 in return value/type.
2050 2013-09-10 Jakub Jelinek <jakub@redhat.com>
2052 PR rtl-optimization/58365
2053 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
2054 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
2057 2013-09-10 Richard Biener <rguenther@suse.de>
2059 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
2060 * tree-data-ref.c (create_rdg_vertices): Collect all data
2061 references, signal failure to the caller, use data-ref API.
2062 (build_rdg): Compute data references only once. Maintain lifetime
2063 of data references and data dependences from within RDG.
2064 (free_rdg): Free dependence relations.
2065 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
2066 inventing extra dependences.
2067 (distribute_loop): Update for RDG API changes.
2069 2013-09-10 Kai Tietz <ktietz@redhat.com>
2071 * doc/invoke.texi (fms-extensions): Document changed
2072 behavior for ms-abi targets.
2073 * config/i386/i386.c (ix86_option_override_internal):
2074 Set default value of option -fms-extension for ms-abi targets.
2076 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
2078 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
2080 2013-09-10 Alan Modra <amodra@gmail.com>
2083 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
2085 2013-09-10 Alan Modra <amodra@gmail.com>
2087 * config/rs6000/predicates.md (add_cint_operand): New.
2088 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
2089 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
2090 using add_cint_operand.
2091 (largetoc_high_plus_aix): Likewise.
2093 2013-09-09 Jakub Jelinek <jakub@redhat.com>
2095 PR tree-optimization/58364
2096 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
2097 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
2098 the current range can't be an unconditional true or false.
2100 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
2102 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
2104 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
2106 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
2108 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
2111 * doc/invoke.texi (-Wdelete-incomplete): Document it.
2113 2013-09-09 Ian Bolton <ian.bolton@arm.com>
2115 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
2116 NO_REGS for immediate that can't be moved directly into FP_REGS.
2118 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2120 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
2121 comparison with negated operand.
2122 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
2125 2013-09-09 Richard Biener <rguenther@suse.de>
2128 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
2129 of a subloop record all its block as affecting loop-closed SSA form.
2131 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2133 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
2134 of an rtx/bitpos pair.
2135 (store_fixed_bit_field): Update accordingly.
2137 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2139 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
2141 * builtins.c (expand_errno_check): Likewise.
2142 * dwarf2cfi.c (init_return_column_size): Likewise.
2143 * except.c (sjlj_mark_call_sites): Likewise.
2144 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
2145 * lra-constraints.c (emit_inc): Likewise.
2146 * ree.c (combine_set_extension): Likewise.
2147 * regmove.c (fixup_match_2): Likewise.
2148 * reload1.c (inc_for_reload): Likewise.
2150 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2152 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
2153 (force_to_mode, simplify_shift_const_1, simplify_comparison):
2154 Use gen_int_mode with the mode of the associated simplify_* call.
2155 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
2156 * expmed.c (expand_shift_1): Likewise.
2157 * function.c (instantiate_virtual_regs_in_insn): Likewise.
2158 * loop-iv.c (iv_number_of_iterations): Likewise.
2159 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
2160 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
2162 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2164 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
2165 of the associated expand_* call.
2166 (asan_emit_stack_protection): Likewise.
2167 * builtins.c (round_trampoline_addr): Likewise.
2168 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
2169 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
2170 (emit_store_flag): Likewise.
2171 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
2172 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
2174 * function.c (instantiate_virtual_regs_in_insn): Likewise.
2175 * ifcvt.c (noce_try_store_flag_constants): Likewise.
2176 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
2177 * modulo-sched.c (generate_prolog_epilog): Likewise.
2178 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
2179 (expand_ctz, expand_ffs, expand_unop): Likewise.
2181 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2183 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
2184 of the associated gen_rtx_* call.
2185 * caller-save.c (init_caller_save): Likewise.
2186 * combine.c (find_split_point, make_extraction): Likewise.
2187 (make_compound_operation): Likewise.
2188 * dwarf2out.c (mem_loc_descriptor): Likewise.
2189 * explow.c (plus_constant, probe_stack_range): Likewise.
2190 * expmed.c (expand_mult_const): Likewise.
2191 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
2192 * reload1.c (init_reload): Likewise.
2193 * valtrack.c (cleanup_auto_inc_dec): Likewise.
2194 * var-tracking.c (adjust_mems): Likewise.
2195 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
2196 rather than gen_rtx_fmt_ee.
2198 2013-09-09 Jan Hubicka <jh@suse.cz>
2201 * value-prof.c (gimple_ic): Copy also abnormal edges.
2203 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
2205 * asan.c (asan_shadow_cst): Use gen_int_mode.
2207 2013-09-08 Jan Hubicka <jh@suse.cz>
2209 * ipa-profile.c: Add toplevel comment.
2210 (ipa_propagate_frequency_1): Be more conservative when profile is read.
2211 (contains_hot_call_p): New function.
2212 (ipa_propagate_frequency): Set frequencies based on counts when
2214 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
2215 profile; do not tamper with profile after inlining if it is read.
2217 2013-09-08 Jan Hubicka <jh@suse.cz>
2219 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
2222 2013-09-08 Jan Hubicka <jh@suse.cz>
2224 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
2227 2013-09-08 Jeff Law <law@redhat.com>
2230 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
2233 2013-09-08 Andi Kleen <ak@linux.intel.com>
2235 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
2237 2013-09-08 Jan Hubicka <jh@suse.cz>
2239 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
2241 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
2243 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
2244 for non-debug insns.
2245 * lra.c (new_insn_reg): Take the containing insn as a parameter.
2246 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
2247 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
2250 2013-09-08 Jan Hubicka <jh@suse.cz>
2252 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
2253 targets and devirtualize to BUILT_IN_UNREACHABLE.
2254 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
2255 * ipa.c (walk_polymorphic_call_targets): New function.
2256 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
2257 functions; use the new timevar.
2258 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
2259 was removed from the program.
2260 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
2261 not consider it in the walk when its vtable is dead.
2262 (possible_polymorphic_call_targets_1): Pass anonymous flag to
2264 (devirt_variable_node_removal_hook): New function.
2265 (possible_polymorphic_call_targets): Also register
2266 devirt_variable_node_removal_hook.
2267 (ipa_devirt): Do not do non-speculative devirtualization.
2268 (gate_ipa_devirt): One execute if devirtualizing speculatively.
2270 2013-09-08 Jan Hubicka <jh@suse.cz>
2272 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
2273 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
2274 varpool_remove_variable_insertion_hook): Declare.
2275 * varpool.c (varpool_node_hook_list): New structure.
2276 (first_varpool_node_removal_hook,
2277 first_varpool_variable_insertion_hook): New variables.
2278 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
2279 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
2280 varpool_remove_variable_insertion_hook,
2281 varpool_call_variable_insertion_hooks): New functions.
2282 (varpool_remove_node): Use it.
2284 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
2287 * diagnostic.c (diagnostic_build_prefix): When s.file is
2288 "<built-in>" don't output line and column numbers.
2290 2013-09-06 Jan Hubicka <jh@suse.cz>
2292 * cgraphunit.c (expand_thunk): Get body before touching arguments.
2293 * lto-streamer-out.c: Stream thunks, too.
2294 * lto-streamer-in.c (input_function): Pop cfun here
2295 (lto_read_body): Instead of here.
2297 2013-09-06 Caroline Tice <cmtice@google.com>
2299 * doc/install.texi: Add documentation for the --enable-vtable-verify
2300 and the --disable-libvtv configure options.
2302 2013-09-06 Jeff Law <law@redhat.com>
2304 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
2305 edge implied equivalences into successor phis.
2307 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
2309 * resource.c (mark_referenced_resources): Handle COND_EXEC.
2311 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
2313 * resource.c (mark_target_live_regs): Compute resources taking
2314 into account if a call is predicated or not.
2316 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
2318 * toplev.c (output_stack_usage): Be prepared for suffixes created by
2319 the compiler in the function names.
2321 2013-09-06 Jan Hubicka <jh@suse.cz>
2324 * ipa-inline.c (has_caller_p): New function.
2325 (want_inline_function_to_all_callers_p): Use it.
2326 (sum_callers, inline_to_all_callers): Break out from ...
2327 (ipa_inline): ... here.
2329 2013-09-06 Jan Hubicka <jh@suse.cz>
2331 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
2332 only when AVX is enabled.
2334 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2336 * config/aarch64/aarch64.md
2337 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
2338 is fpsimd_<load/store>2.
2339 (load_pair<mode>): Likewise.
2340 (store_pair<mode>): Likewise.
2342 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2344 * config/arm/types.md (type): Add "mrs" type.
2345 * config/aarch64/aarch64.md
2346 (aarch64_load_tp_hard): Make type "mrs".
2348 (load_tp_hard): Make type "mrs".
2349 * config/arm/cortex-a15.md: Update with new attributes.
2350 * config/arm/cortex-a5.md: Update with new attributes.
2351 * config/arm/cortex-a53.md: Update with new attributes.
2352 * config/arm/cortex-a7.md: Update with new attributes.
2353 * config/arm/cortex-a8.md: Update with new attributes.
2354 * config/arm/cortex-a9.md: Update with new attributes.
2355 * config/arm/cortex-m4.md: Update with new attributes.
2356 * config/arm/cortex-r4.md: Update with new attributes.
2357 * config/arm/fa526.md: Update with new attributes.
2358 * config/arm/fa606te.md: Update with new attributes.
2359 * config/arm/fa626te.md: Update with new attributes.
2360 * config/arm/fa726te.md: Update with new attributes.
2362 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2364 * config/aarch64/aarch64.md
2365 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
2366 (*movtf_aarch64): Likewise.
2368 (thumb1_movdi_insn): Use "multiple" for type where more than one
2369 instruction is used for a move.
2370 (*arm32_movhf): Likewise.
2371 (*thumb_movdf_insn): Likewise.
2373 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2375 * config/arm/types.md (type): Rename fcpys to fmov.
2377 (*arm_movsi_vfp): Rename type fcpys as fmov.
2378 (*thumb2_movsi_vfp): Likewise
2379 (*movhf_vfp_neon): Likewise
2380 (*movhf_vfp): Likewise
2381 (*movsf_vfp): Likewise
2382 (*thumb2_movsf_vfp): Likewise
2383 (*movsfcc_vfp): Likewise
2384 (*thumb2_movsfcc_vfp): Likewise
2385 * config/aarch64/aarch64-simd.md
2386 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
2387 * config/aarch64/aarch64.md
2388 (*movsi_aarch64): Replace type mov_reg with fmovs.
2389 (*movdi_aarch64): Likewise
2390 (*movsf_aarch64): Likewise
2391 (*movdf_aarch64): Likewise
2393 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
2394 * config/arm/iwmmxt.md
2395 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
2396 * config/arm/arm1020e.md: Update with new attributes.
2397 * config/arm/cortex-a15-neon.md: Update with new attributes.
2398 * config/arm/cortex-a5.md: Update with new attributes.
2399 * config/arm/cortex-a53.md: Update with new attributes.
2400 * config/arm/cortex-a7.md: Update with new attributes.
2401 * config/arm/cortex-a8-neon.md: Update with new attributes.
2402 * config/arm/cortex-a9.md: Update with new attributes.
2403 * config/arm/cortex-m4-fpu.md: Update with new attributes.
2404 * config/arm/cortex-r4f.md: Update with new attributes.
2405 * config/arm/marvell-pj4.md: Update with new attributes.
2406 * config/arm/vfp11.md: Update with new attributes.
2408 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2410 * config/aarch64/aarch64.md
2411 (*madd<mode>): Fix type attribute.
2412 (*maddsi_uxtw): Likewise.
2413 (*msub<mode>): Likewise.
2414 (*msubsi_uxtw): Likewise.
2415 (<su_optab>maddsidi4): Likewise.
2416 (<su_optab>msubsidi4): Likewise.
2418 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2420 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
2421 * config/arm/arm.md (core_cycles): Remove fdiv.
2422 * config/arm/vfp.md:
2423 (*sqrtsf2_vfp): Update for attribute changes.
2424 (*sqrtdf2_vfp): Likewise.
2425 * config/aarch64/aarch64.md:
2426 (sqrt<mode>2): Update for attribute changes.
2427 * config/arm/arm1020e.md: Update with new attributes.
2428 * config/arm/cortex-a15-neon.md: Update with new attributes.
2429 * config/arm/cortex-a5.md: Update with new attributes.
2430 * config/arm/cortex-a53.md: Update with new attributes.
2431 * config/arm/cortex-a7.md: Update with new attributes.
2432 * config/arm/cortex-a8-neon.md: Update with new attributes.
2433 * config/arm/cortex-a9.md: Update with new attributes.
2434 * config/arm/cortex-m4-fpu.md: Update with new attributes.
2435 * config/arm/cortex-r4f.md: Update with new attributes.
2436 * config/arm/marvell-pj4.md: Update with new attributes.
2437 * config/arm/vfp11.md: Update with new attributes.
2439 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2441 * config/arm/types.md
2442 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
2443 * config/aarch64/aarch64.md
2444 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
2446 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
2447 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
2448 (float<GPI:mode><GPF:mode>2): Likewise.
2450 (*truncsisf2_vfp): Update with new attributes.
2451 (*truncsidf2_vfp): Likewise.
2452 (fixuns_truncsfsi2): Likewise.
2453 (fixuns_truncdfsi2): Likewise.
2454 (*floatsisf2_vfp): Likewise.
2455 (*floatsidf2_vfp): Likewise.
2456 (floatunssisf2): Likewise.
2457 (floatunssidf2): Likewise.
2458 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
2459 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
2460 * config/arm/arm1020e.md: Update with new attributes.
2461 * config/arm/cortex-a15-neon.md: Update with new attributes.
2462 * config/arm/cortex-a5.md: Update with new attributes.
2463 * config/arm/cortex-a53.md: Update with new attributes.
2464 * config/arm/cortex-a7.md: Update with new attributes.
2465 * config/arm/cortex-a8-neon.md: Update with new attributes.
2466 * config/arm/cortex-a9.md: Update with new attributes.
2467 * config/arm/cortex-m4-fpu.md: Update with new attributes.
2468 * config/arm/cortex-r4f.md: Update with new attributes.
2469 * config/arm/marvell-pj4.md: Update with new attributes.
2470 * config/arm/vfp11.md: Update with new attributes.
2472 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2474 * config/aarch64/arm_neon.h
2475 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
2476 (vqtbx<1,2,3,4><q>_s8): Likewise.
2478 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2480 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
2481 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
2482 (add<mode>3): Add type attribute.
2483 (add<mode>3): Likewise.
2484 (usadd<mode>3): Likewise.
2485 (ssadd<mode>3): Likewise.
2486 (sub<mode>3): Likewise.
2487 (sub<mode>3): Likewise.
2488 (ussub<mode>3): Likewise.
2489 (sssub<mode>3): Likewise.
2490 (ssmulsa3): Likewise.
2491 (usmulusa3): Likewise.
2492 (arm_usatsihi): Likewise.
2494 (*movdi_vfp): Add types for all instructions.
2495 (*movdi_vfp_cortexa8): Likewise.
2496 (*movhf_vfp_neon): Likewise.
2497 (*movhf_vfp): Likewise.
2498 (*movdf_vfp): Likewise.
2499 (*thumb2_movdf_vfp): Likewise.
2500 (*thumb2_movdfcc_vfp): Likewise.
2501 * config/arm/arm.md: Add type attribute to all insn patterns.
2502 (*thumb1_adddi3): Add type attribute.
2503 (*arm_adddi3): Likewise.
2504 (*adddi_sesidi_di): Likewise.
2505 (*adddi_zesidi_di): Likewise.
2506 (*thumb1_addsi3): Likewise.
2507 (addsi3_compare0): Likewise.
2508 (*addsi3_compare0_scratch): Likewise.
2509 (*compare_negsi_si): Likewise.
2510 (cmpsi2_addneg): Likewise.
2511 (*addsi3_carryin_<optab>): Likewise.
2512 (*addsi3_carryin_alt2_<optab>): Likewise.
2513 (*addsi3_carryin_clobercc_<optab>): Likewise.
2514 (*subsi3_carryin): Likewise.
2515 (*subsi3_carryin_const): Likewise.
2516 (*subsi3_carryin_compare): Likewise.
2517 (*subsi3_carryin_compare_const): Likewise.
2518 (*arm_subdi3): Likewise.
2519 (*thumb_subdi3): Likewise.
2520 (*subdi_di_zesidi): Likewise.
2521 (*subdi_di_sesidi): Likewise.
2522 (*subdi_zesidi_di): Likewise.
2523 (*subdi_sesidi_di): Likewise.
2524 (*subdi_zesidi_ze): Likewise.
2525 (thumb1_subsi3_insn): Likewise.
2526 (*arm_subsi3_insn): Likewise.
2527 (*anddi3_insn): Likewise.
2528 (*anddi_zesidi_di): Likewise.
2529 (*anddi_sesdi_di): Likewise.
2530 (*ne_zeroextracts): Likewise.
2531 (*ne_zeroextracts): Likewise.
2532 (*ite_ne_zeroextr): Likewise.
2533 (*ite_ne_zeroextr): Likewise.
2534 (*anddi_notdi_di): Likewise.
2535 (*anddi_notzesidi): Likewise.
2536 (*anddi_notsesidi): Likewise.
2537 (andsi_notsi_si): Likewise.
2538 (thumb1_bicsi3): Likewise.
2539 (*iordi3_insn): Likewise.
2540 (*iordi_zesidi_di): Likewise.
2541 (*iordi_sesidi_di): Likewise.
2542 (*thumb1_iorsi3_insn): Likewise.
2543 (*xordi3_insn): Likewise.
2544 (*xordi_zesidi_di): Likewise.
2545 (*xordi_sesidi_di): Likewise.
2546 (*arm_xorsi3): Likewise.
2547 (*andsi_iorsi3_no): Likewise.
2548 (*smax_0): Likewise.
2549 (*smax_m1): Likewise.
2550 (*arm_smax_insn): Likewise.
2551 (*smin_0): Likewise.
2552 (*arm_smin_insn): Likewise.
2553 (*arm_umaxsi3): Likewise.
2554 (*arm_uminsi3): Likewise.
2555 (*minmax_arithsi): Likewise.
2556 (*minmax_arithsi_): Likewise.
2557 (*satsi_<SAT:code>): Likewise.
2558 (arm_ashldi3_1bit): Likewise.
2559 (arm_ashrdi3_1bit): Likewise.
2560 (arm_lshrdi3_1bit): Likewise.
2561 (*arm_negdi2): Likewise.
2562 (*thumb1_negdi2): Likewise.
2563 (*arm_negsi2): Likewise.
2564 (*thumb1_negsi2): Likewise.
2565 (*negdi_extendsid): Likewise.
2566 (*negdi_zero_extend): Likewise.
2567 (*arm_abssi2): Likewise.
2568 (*thumb1_abssi2): Likewise.
2569 (*arm_neg_abssi2): Likewise.
2570 (*thumb1_neg_abss): Likewise.
2571 (one_cmpldi2): Likewise.
2572 (extend<mode>di2): Likewise.
2573 (*compareqi_eq0): Likewise.
2574 (*arm_extendhisi2addsi): Likewise.
2575 (*arm_movdi): Likewise.
2576 (*thumb1_movdi_insn): Likewise.
2577 (*arm_movt): Likewise.
2578 (*thumb1_movsi_insn): Likewise.
2579 (pic_add_dot_plus_four): Likewise.
2580 (pic_add_dot_plus_eight): Likewise.
2581 (tls_load_dot_plus_eight): Likewise.
2582 (*thumb1_movhi_insn): Likewise.
2583 (*thumb1_movsf_insn): Likewise.
2584 (*movdf_soft_insn): Likewise.
2585 (*thumb_movdf_insn): Likewise.
2586 (cbranchsi4_insn): Likewise.
2587 (cbranchsi4_scratch): Likewise.
2588 (*negated_cbranchsi4): Likewise.
2589 (*tbit_cbranch): Likewise.
2590 (*tlobits_cbranch): Likewise.
2591 (*tstsi3_cbranch): Likewise.
2592 (*cbranchne_decr1): Likewise.
2593 (*addsi3_cbranch): Likewise.
2594 (*addsi3_cbranch_scratch): Likewise.
2595 (*arm_cmpdi_insn): Likewise.
2596 (*arm_cmpdi_unsig): Likewise.
2597 (*arm_cmpdi_zero): Likewise.
2598 (*thumb_cmpdi_zero): Likewise.
2599 (*deleted_compare): Likewise.
2600 (*mov_scc): Likewise.
2601 (*mov_negscc): Likewise.
2602 (*mov_notscc): Likewise.
2603 (*cstoresi_eq0_thumb1_insn): Likewise.
2604 (cstoresi_nltu_thumb1): Likewise.
2605 (cstoresi_ltu_thu): Likewise.
2606 (thumb1_addsi3_addgeu): Likewise.
2607 (*arm_jump): Likewise.
2608 (*thumb_jump): Likewise.
2609 (*check_arch2): Likewise.
2610 (arm_casesi_internal): Likewise.
2611 (thumb1_casesi_dispatch): Likewise.
2612 (*arm_indirect_jump): Likewise.
2613 (*thumb1_indirect_jump): Likewise.
2615 (*and_scc): Likewise.
2616 (*ior_scc): Likewise.
2617 (*compare_scc): Likewise.
2618 (*cond_move): Likewise.
2619 (*cond_arith): Likewise.
2620 (*cond_sub): Likewise.
2621 (*cmp_ite0): Likewise.
2622 (*cmp_ite1): Likewise.
2623 (*cmp_and): Likewise.
2624 (*cmp_ior): Likewise.
2625 (*ior_scc_scc): Likewise.
2626 (*ior_scc_scc_cmp): Likewise.
2627 (*and_scc_scc): Likewise.
2628 (*and_scc_scc_cmp): Likewise.
2629 (*and_scc_scc_nod): Likewise.
2630 (*negscc): Likewise.
2631 (movcond_addsi): Likewise.
2632 (movcond): Likewise.
2633 (*ifcompare_plus_move): Likewise.
2634 (*if_plus_move): Likewise.
2635 (*ifcompare_move_plus): Likewise.
2636 (*if_move_plus): Likewise.
2637 (*ifcompare_arith_arith): Likewise.
2638 (*if_arith_arith): Likewise.
2639 (*ifcompare_arith_move): Likewise.
2640 (*if_arith_move): Likewise.
2641 (*ifcompare_move_arith): Likewise.
2642 (*if_move_arith): Likewise.
2643 (*ifcompare_move_not): Likewise.
2644 (*if_move_not): Likewise.
2645 (*ifcompare_not_move): Likewise.
2646 (*if_not_move): Likewise.
2647 (*ifcompare_shift_move): Likewise.
2648 (*if_shift_move): Likewise.
2649 (*ifcompare_move_shift): Likewise.
2650 (*if_move_shift): Likewise.
2651 (*ifcompare_shift_shift): Likewise.
2652 (*ifcompare_not_arith): Likewise.
2653 (*ifcompare_arith_not): Likewise.
2654 (*if_arith_not): Likewise.
2655 (*ifcompare_neg_move): Likewise.
2656 (*if_neg_move): Likewise.
2657 (*ifcompare_move_neg): Likewise.
2658 (*if_move_neg): Likewise.
2659 (prologue_thumb1_interwork): Likewise.
2660 (*cond_move_not): Likewise.
2661 (*sign_extract_onebit): Likewise.
2662 (*not_signextract_onebit): Likewise.
2663 (stack_tie): Likewise.
2664 (align_4): Likewise.
2665 (align_8): Likewise.
2666 (consttable_end): Likewise.
2667 (consttable_1): Likewise.
2668 (consttable_2): Likewise.
2669 (consttable_4): Likewise.
2670 (consttable_8): Likewise.
2671 (consttable_16): Likewise.
2672 (*thumb1_tablejump): Likewise.
2673 (prefetch): Likewise.
2674 (force_register_use): Likewise.
2675 (thumb_eh_return): Likewise.
2676 (load_tp_hard): Likewise.
2677 (load_tp_soft): Likewise.
2678 (tlscall): Likewise.
2679 (*arm_movtas_ze): Likewise.
2680 (*arm_rev): Likewise.
2681 (*arm_revsh): Likewise.
2682 (*arm_rev16): Likewise.
2683 * config/arm/thumb2.md
2684 (*thumb2_smaxsi3): Likewise.
2685 (*thumb2_sminsi3): Likewise.
2686 (*thumb32_umaxsi3): Likewise.
2687 (*thumb2_uminsi3): Likewise.
2688 (*thumb2_negdi2): Likewise.
2689 (*thumb2_abssi2): Likewise.
2690 (*thumb2_neg_abss): Likewise.
2691 (*thumb2_movsi_insn): Likewise.
2692 (tls_load_dot_plus_four): Likewise.
2693 (*thumb2_movhi_insn): Likewise.
2694 (*thumb2_mov_scc): Likewise.
2695 (*thumb2_mov_negs): Likewise.
2696 (*thumb2_mov_negs): Likewise.
2697 (*thumb2_mov_nots): Likewise.
2698 (*thumb2_mov_nots): Likewise.
2699 (*thumb2_movsicc_): Likewise.
2700 (*thumb2_movsfcc_soft_insn): Likewise.
2701 (*thumb2_indirect_jump): Likewise.
2702 (*thumb2_and_scc): Likewise.
2703 (*thumb2_ior_scc): Likewise.
2704 (*thumb2_ior_scc_strict_it): Likewise.
2705 (*thumb2_cond_move): Likewise.
2706 (*thumb2_cond_arith): Likewise.
2707 (*thumb2_cond_ari): Likewise.
2708 (*thumb2_cond_sub): Likewise.
2709 (*thumb2_negscc): Likewise.
2710 (*thumb2_movcond): Likewise.
2711 (thumb2_casesi_internal): Likewise.
2712 (thumb2_casesi_internal_pic): Likewise.
2713 (*thumb2_alusi3_short): Likewise.
2714 (*thumb2_mov<mode>_shortim): Likewise.
2715 (*thumb2_addsi_short): Likewise.
2716 (*thumb2_subsi_short): Likewise.
2717 (thumb2_addsi3_compare0): Likewise.
2718 (*thumb2_cbz): Likewise.
2719 (*thumb2_cbnz): Likewise.
2720 (*thumb2_one_cmplsi2_short): Likewise.
2721 (*thumb2_negsi2_short): Likewise.
2722 (*orsi_notsi_si): Likewise.
2723 * config/arm/arm1020e.md: Update with new attributes.
2724 * config/arm/arm1026ejs.md: Update with new attributes.
2725 * config/arm/arm1136jfs.md: Update with new attributes.
2726 * config/arm/arm926ejs.md: Update with new attributes.
2727 * config/arm/cortex-a15.md: Update with new attributes.
2728 * config/arm/cortex-a5.md: Update with new attributes.
2729 * config/arm/cortex-a53.md: Update with new attributes.
2730 * config/arm/cortex-a7.md: Update with new attributes.
2731 * config/arm/cortex-a8.md: Update with new attributes.
2732 * config/arm/cortex-a9.md: Update with new attributes.
2733 * config/arm/cortex-m4.md: Update with new attributes.
2734 * config/arm/cortex-r4.md: Update with new attributes.
2735 * config/arm/fa526.md: Update with new attributes.
2736 * config/arm/fa606te.md: Update with new attributes.
2737 * config/arm/fa626te.md: Update with new attributes.
2738 * config/arm/fa726te.md: Update with new attributes.
2740 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
2742 * config/aarch64/aarch64-simd.md
2743 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
2744 <vwx> iterator to ensure correct register choice.
2745 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
2746 (aarch64_sqdmull_n<mode>): Likewise.
2747 (aarch64_sqdmull2_n<mode>_internal): Likewise.
2748 * config/aarch64/arm_neon.h
2749 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
2750 (vml<as><q>_n_<su>16): Likewise.
2751 (vml<as>l_high_lane<q>_<su>16): Likewise.
2752 (vml<as>l_high_n_<su>16): Likewise.
2753 (vml<as>l_lane<q>_<su>16): Likewise.
2754 (vml<as>l_n_<su>16): Likewise.
2755 (vmul<q>_lane<q>_<su>16): Likewise.
2756 (vmul<q>_n_<su>16): Likewise.
2757 (vmull_lane<q>_<su>16): Likewise.
2758 (vmull_n_<su>16): Likewise.
2759 (vmull_high_lane<q>_<su>16): Likewise.
2760 (vmull_high_n_<su>16): Likewise.
2761 (vqrdmulh<q>_n_s16): Likewise.
2763 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
2765 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
2766 have the correct lane parameter.
2768 2013-09-06 Richard Biener <rguenther@suse.de>
2770 * cfganal.c (control_dependences::~control_dependences):
2771 Properly free all of the vector.
2773 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
2776 * config/i386/i386.c (ix86_conditional_register_usage):
2777 Proper initialize extended SSE registers.
2779 2013-09-06 Jan Hubicka <jh@suse.cz>
2781 PR tree-optimization/58311
2782 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
2784 2013-09-06 Jan Hubicka <jh@suse.cz>
2786 * Makefile.in (tree-sra.o): Update dependencies.
2787 * tree-sra.c: Include ipa-utils.h
2788 (scan_function): Use recursive_call_p.
2789 (has_caller_p): New function.
2790 (cgraph_for_node_and_aliases): Count also callers of aliases.
2792 2013-09-06 Jan Hubicka <jh@suse.cz>
2795 * cgraph.h (symtab_semantically_equivalent_p): Declare.
2796 * tree-tailcall.c: Include ipa-utils.h.
2797 (find_tail_calls): Use it.
2798 * ipa-pure-const.c (check_call): Likewise.
2799 * ipa-utils.c (recursive_call_p): New function.
2800 * ipa-utils.h (recursive_call_p): Dclare.
2801 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
2802 (symtab_semantically_equivalent_p): New function.
2803 * Makefile.in (tree-tailcall.o): Update dependencies.
2805 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
2807 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
2810 2013-09-06 Richard Biener <rguenther@suse.de>
2812 * lto-streamer.h (lto_global_var_decls): Remove.
2813 * Makefile.in (OBJS): Remove lto-symtab.o.
2814 (lto-symtab.o): Remove.
2815 (GTFILES): Remove lto-symtab.c
2816 * lto-symtab.c: Move to lto/
2818 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2820 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
2821 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
2822 (UNSPEC_FPINT_RINT): New constant definitions.
2823 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
2824 definition with 2 attributes.
2825 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
2826 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
2829 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2831 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
2832 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
2833 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
2834 attribute to "z196_alone".
2835 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
2838 2013-09-06 Richard Biener <rguenther@suse.de>
2840 * basic-block.h (class control_dependences): New.
2841 * tree-ssa-dce.c (control_dependence_map): Remove.
2843 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
2844 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
2845 find_pdom, find_control_dependence, find_all_control_dependences):
2847 (mark_control_dependent_edges_necessary,
2848 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
2849 tree_dce_done, perform_tree_ssa_dce): Adjust.
2850 * cfganal.c (set_control_dependence_map_bit,
2851 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
2852 find_all_control_dependences): Move from tree-ssa-dce.c and
2853 implement as methods of control_dependences class.
2854 (control_dependences::control_dependences): New.
2855 (control_dependences::~control_dependences): Likewise.
2856 (control_dependences::get_edges_dependent_on): Likewise.
2857 (control_dependences::get_edge): Likewise.
2859 2013-09-04 Jan Hubicka <jh@suse.cz>
2861 * tree.c (types_same_for_odr): Drop overactive check.
2862 * ipa-devirt.c (hash_type_name): Likewise.
2864 2013-09-04 Jan Hubicka <jh@suse.cz>
2866 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
2867 (analyze_functions): ... here.
2869 2013-09-04 Jan Hubicka <jh@suse.cz>
2872 * cgraphunit.c (analyze_functions): Clear AUX fields
2873 after processing; initialize assembler name has.
2875 2013-09-05 Jeff Law <law@redhat.com>
2877 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
2878 from thread_around_empty_block. Record threading path into PATH.
2879 Recurse if threading through the initial block is successful.
2880 (thread_across_edge): Corresponding changes to slightly simplify.
2882 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
2884 * config/aarch64/aarch64.md
2885 (type): Remove frecpe, frecps, frecpx.
2886 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
2887 fix to be a TARGET_SIMD instruction.
2888 (aarch64_frecps): Remove.
2889 * config/aarch64/aarch64-simd.md
2890 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
2891 (aarch64_frecps<mode>): Handle all float/vector of float modes.
2893 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
2894 Sofiane Naci <sofiane.naci@arm.com>
2896 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
2897 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
2898 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
2899 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
2900 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
2901 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
2902 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
2903 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
2904 into "clz, "rbit". Rename "shift" to "shift_imm".
2905 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
2906 changes. Update for attribute changes all occurrences of arlo_* and
2908 * config/arm/arm-fixed.md: Update for attribute changes
2909 all occurrences of arlo_* types.
2910 * config/arm/thumb2.md: Update for attribute changes all occurrences
2912 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
2913 (cortexa7_older_only): Likewise.
2914 (cortexa7_younger): Likewise.
2915 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
2916 (1020alu_shift_op): Likewise.
2917 (1020alu_shift_reg_op): Likewise.
2918 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
2919 (alu_shift_op): Likewise.
2920 (alu_shift_reg_op): Likewise.
2921 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
2922 (11_alu_shift_op): Likewise.
2923 (11_alu_shift_reg_op): Likewise.
2924 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
2925 (9_alu_shift_reg_op): Likewise.
2926 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
2928 (cortex_a15_alu_shift): Likewise.
2929 (cortex_a15_alu_shift_reg): Likewise.
2930 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
2932 (cortex_a5_alu_shift): Likewise.
2933 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
2935 (cortex_a53_alu_shift): Likewise.
2936 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
2938 (cortex_a7_alu_reg): Likewise.
2939 (cortex_a7_alu_shift): Likewise.
2940 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
2942 (cortex_a8_alu_shift): Likewise.
2943 (cortex_a8_alu_shift_reg): Likewise.
2944 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
2945 (cortex_a9_dp_shift): Likewise.
2946 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
2948 * config/arm/cortex-r4.md
2949 (cortex_r4_alu): Update for attribute changes.
2950 (cortex_r4_mov): Likewise.
2951 (cortex_r4_alu_shift_reg): Likewise.
2952 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
2953 (526_alu_shift_op): Likewise.
2954 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
2955 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
2956 (626te_alu_shift_op): Likewise.
2957 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
2958 (726te_alu_shift_op): Likewise.
2959 (726te_alu_shift_reg_op): Likewise.
2960 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
2961 (mp626_alu_shift_op): Likewise.
2962 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
2963 (pj4_alu_conds): Likewise.
2964 (pj4_shift): Likewise.
2965 (pj4_shift_conds): Likewise.
2966 (pj4_alu_shift): Likewise.
2967 (pj4_alu_shift_conds): Likewise.
2968 * config/aarch64/aarch64.md: Update for attribute change
2969 all occurrences of arlo_* and shift* types.
2971 2013-09-05 Mike Stump <mikestump@comcast.net>
2973 * tree.h: Move documentation for tree_function_decl to tree-core.h
2974 with the declaration.
2976 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
2979 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
2980 looking for widest mode.
2982 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
2984 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
2986 2013-09-05 Richard Biener <rguenther@suse.de>
2988 PR tree-optimization/58137
2989 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
2990 Do not create vectors of pointers.
2991 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
2992 types for the components of the vector initializer.
2993 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
2994 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
2996 2013-09-05 Martin Jambor <mjambor@suse.cz>
2998 * ipa-prop.c (remove_described_reference): Accept missing references,
2999 return false if that hppens, otherwise return true.
3000 (cgraph_node_for_jfunc): New function.
3001 (try_decrement_rdesc_refcount): Likewise.
3002 (try_make_edge_direct_simple_call): Use them.
3003 (ipa_edge_removal_hook): Remove references from rdescs.
3004 (ipa_edge_duplication_hook): Clone rdescs and their references
3005 when the new edge has the same caller as the old one.
3006 * cgraph.c (cgraph_resolve_speculation): Remove speculative
3007 reference before removing any edges.
3009 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
3011 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
3013 * thumb2.md (thumb2_storewb_parisi): New pattern.
3015 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
3017 * config/aarch64/aarch64-option-extensions.def: Add
3018 AARCH64_OPT_EXTENSION of 'crc'.
3019 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
3020 (AARCH64_ISA_CRC): Ditto.
3021 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
3022 description of the CRC extension.
3024 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
3026 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
3027 * config/rs6000/linux.h: Ditto.
3028 * alpha/linux.h: Ditto.
3029 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
3030 no_c99_libc_has_function.
3031 * config/c6x/uclinux-elf.h: Ditto.
3032 * config/lm32/uclinux-elf.h: Ditto.
3033 * config/m68k/uclinux.h: Ditto.
3034 * config/moxie/uclinux.h: Ditto.
3035 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
3036 (crisv32-*-linux*, cris-*-linux*): Ditto.
3037 * config/bfin/bfin.c: Include "tm_p.h".
3039 2013-09-05 Richard Biener <rguenther@suse.de>
3041 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
3042 check for a definition without a basic-block.
3044 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
3045 Sofiane Naci <sofiane.naci@arm.com>
3047 * config/aarch64/aarch64.md
3048 (*movti_aarch64): Rename r_2_f and f_2_r.
3049 (*movsf_aarch64): Likewise.
3050 (*movdf_aarch64): Likewise.
3051 (*movtf_aarch64): Likewise.
3052 (aarch64_movdi_<mode>low): Likewise.
3053 (aarch64_movdi_<mode>high): Likewise.
3054 (aarch64_mov<mode>high_di): Likewise.
3055 (aarch64_mov<mode>low_di): Likewise.
3056 (aarch64_movtilow_tilow): Likewise.
3057 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
3058 values to config/arm/types.md
3059 (attribute "conds"): Update for attribute change.
3060 (anddi3_insn): Likewise.
3061 (iordi3_insn): Likewise.
3062 (xordi3_insn): Likewise.
3063 (one_cmpldi2): Likewise.
3064 * config/arm/types.md (type): Add Neon types.
3065 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
3066 use "type" attribute.
3067 (movmisalign<mode>_neon_store): Likewise.
3068 (movmisalign<mode>_neon_load): Likewise.
3069 (vec_set<mode>_internal): Likewise.
3070 (vec_setv2di_internal): Likewise.
3071 (vec_extract<mode>): Likewise.
3072 (vec_extractv2di): Likewise.
3073 (add<mode>3_neon): Likewise.
3074 (adddi3_neon): Likewise.
3075 (sub<mode>3_neon): Likewise.
3076 (subdi3_neon): Likewise.
3077 (mul<mode>3_neon): Likewise.
3078 (mul<mode>3add<mode>_neon): Likewise.
3079 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
3080 (fma<VCVTF:mode>4)): Likewise.
3081 (fma<VCVTF:mode>4_intrinsic): Likewise.
3082 (fmsub<VCVTF:mode>4)): Likewise.
3083 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
3084 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
3085 (ior<mode>3): Likewise.
3086 (and<mode>3): Likewise.
3087 (anddi3_neon): Likewise.
3088 (orn<mode>3_neon): Likewise.
3089 (orndi3_neon): Likewise.
3090 (bic<mode>3_neon): Likewise.
3091 (bicdi3_neon): Likewise.
3092 (xor<mode>3): Likewise.
3093 (one_cmpl<mode>2): Likewise.
3094 (abs<mode>2): Likewise.
3095 (neg<mode>2): Likewise.
3096 (umin<mode>3_neon): Likewise.
3097 (umax<mode>3_neon): Likewise.
3098 (smin<mode>3_neon): Likewise.
3099 (smax<mode>3_neon): Likewise.
3100 (vashl<mode>3): Likewise.
3101 (vashr<mode>3_imm): Likewise.
3102 (vlshr<mode>3_imm): Likewise.
3103 (ashl<mode>3_signed): Likewise.
3104 (ashl<mode>3_unsigned): Likewise.
3105 (neon_load_count): Likewise.
3106 (ashldi3_neon_noclobber): Likewise.
3107 (signed_shift_di3_neon): Likewise.
3108 (unsigned_shift_di3_neon): Likewise.
3109 (ashrdi3_neon_imm_noclobber): Likewise.
3110 (lshrdi3_neon_imm_noclobber): Likewise.
3111 (widen_ssum<mode>3): Likewise.
3112 (widen_usum<mode>3): Likewise.
3113 (quad_halves_<code>v4si): Likewise.
3114 (quad_halves_<code>v4sf): Likewise.
3115 (quad_halves_<code>v8hi): Likewise.
3116 (quad_halves_<code>v16qi): Likewise.
3117 (reduc_splus_v2di): Likewise.
3118 (neon_vpadd_internal<mode>): Likewise.
3119 (neon_vpsmin<mode>): Likewise.
3120 (neon_vpsmax<mode>): Likewise.
3121 (neon_vpumin<mode>): Likewise.
3122 (neon_vpumax<mode>): Likewise.
3123 (ss_add<mode>_neon): Likewise.
3124 (us_add<mode>_neon): Likewise.
3125 (ss_sub<mode>_neon): Likewise.
3126 (us_sub<mode>_neon): Likewise.
3127 (neon_vadd<mode>_unspec): Likewise.
3128 (neon_vaddl<mode>): Likewise.
3129 (neon_vaddw<mode>): Likewise.
3130 (neon_vhadd<mode>): Likewise.
3131 (neon_vqadd<mode>): Likewise.
3132 (neon_vaddhn<mode>): Likewise.
3133 (neon_vmul<mode>): Likewise.
3134 (neon_vmla<mode>): Likewise.
3135 (neon_vmlal<mode>): Likewise.
3136 (neon_vmls<mode>): Likewise.
3137 (neon_vmlsl<mode>): Likewise.
3138 (neon_vqdmulh<mode>): Likewise.
3139 (neon_vqdmlal<mode>): Likewise.
3140 (neon_vqdmlsl<mode>): Likewise.
3141 (neon_vmull<mode>): Likewise.
3142 (neon_vqdmull<mode>): Likewise.
3143 (neon_vsub<mode>_unspec): Likewise.
3144 (neon_vsubl<mode>): Likewise.
3145 (neon_vsubw<mode>): Likewise.
3146 (neon_vqsub<mode>): Likewise.
3147 (neon_vhsub<mode>): Likewise.
3148 (neon_vsubhn<mode>): Likewise.
3149 (neon_vceq<mode>): Likewise.
3150 (neon_vcge<mode>): Likewise.
3151 (neon_vcgeu<mode>): Likewise.
3152 (neon_vcgt<mode>): Likewise.
3153 (neon_vcgtu<mode>): Likewise.
3154 (neon_vcle<mode>): Likewise.
3155 (neon_vclt<mode>): Likewise.
3156 (neon_vcage<mode>): Likewise.
3157 (neon_vcagt<mode>): Likewise.
3158 (neon_vtst<mode>): Likewise.
3159 (neon_vabd<mode>): Likewise.
3160 (neon_vabdl<mode>): Likewise.
3161 (neon_vaba<mode>): Likewise.
3162 (neon_vabal<mode>): Likewise.
3163 (neon_vmax<mode>): Likewise.
3164 (neon_vmin<mode>): Likewise.
3165 (neon_vpaddl<mode>): Likewise.
3166 (neon_vpadal<mode>): Likewise.
3167 (neon_vpmax<mode>): Likewise.
3168 (neon_vpmin<mode>): Likewise.
3169 (neon_vrecps<mode>): Likewise.
3170 (neon_vrsqrts<mode>): Likewise.
3171 (neon_vqabs<mode>): Likewise.
3172 (neon_vqneg<mode>): Likewise.
3173 (neon_vcls<mode>): Likewise.
3174 (clz<mode>2): Likewise.
3175 (popcount<mode>2): Likewise.
3176 (neon_vrecpe): Likewise.
3177 (neon_vrsqrte): Likewise.
3178 (neon_vget_lane<mode>_sext_internal): Likewise.
3179 (neon_vget_lane<mode>_zext_internal): Likewise.
3180 (neon_vdup_n<mode>): Likewise.
3181 (neon_vdup_nv2di): Likewise.
3182 (neon_vdpu_lane<mode>_internal): Likewise.
3183 (neon_vswp<mode>): Likewise.
3184 (float<mode><V_cvtto>2): Likewise.
3185 (floatuns<mode><V_cvtto>2): Likewise.
3186 (fix_trunc<mode><V_cvtto>)2): Likewise
3187 (fixuns_trunc<mode><V_cvtto)2): Likewise.
3188 (neon_vcvt<mode>): Likewise.
3189 (neon_vcvtv4sfv4hf): Likewise.
3190 (neon_vcvtv4hfv4sf): Likewise.
3191 (neon_vcvt_n<mode>): Likewise.
3192 (neon_vmovn<mode>): Likewise.
3193 (neon_vqmovn<mode>): Likewise.
3194 (neon_vqmovun<mode>): Likewise.
3195 (neon_vmovl<mode>): Likewise.
3196 (neon_vmul_lane<mode>): Likewise.
3197 (neon_vmull_lane<mode>): Likewise.
3198 (neon_vqdmull_lane<mode>): Likewise.
3199 (neon_vqdmulh_lane<mode>): Likewise.
3200 (neon_vmla_lane<mode>): Likewise.
3201 (neon_vmlal_lane<mode>): Likewise.
3202 (neon_vqdmlal_lane<mode>): Likewise.
3203 (neon_vmls_lane<mode>): Likewise.
3204 (neon_vmlsl_lane<mode>): Likewise.
3205 (neon_vqdmlsl_lane<mode>): Likewise.
3206 (neon_vext<mode>): Likewise.
3207 (neon_vrev64<mode>): Likewise.
3208 (neon_vrev32<mode>): Likewise.
3209 (neon_vrev16<mode>): Likewise.
3210 (neon_vbsl<mode>_internal): Likewise.
3211 (neon_vshl<mode>): Likewise.
3212 (neon_vqshl<mode>): Likewise.
3213 (neon_vshr_n<mode>): Likewise.
3214 (neon_vshrn_n<mode>): Likewise.
3215 (neon_vqshrn_n<mode>): Likewise.
3216 (neon_vqshrun_n<mode>): Likewise.
3217 (neon_vshl_n<mode>): Likewise.
3218 (neon_vqshl_n<mode>): Likewise.
3219 (neon_vqshlu_n<mode>): Likewise.
3220 (neon_vshll_n<mode>): Likewise.
3221 (neon_vsra_n<mode>): Likewise.
3222 (neon_vsri_n<mode>): Likewise.
3223 (neon_vsli_n<mode>): Likewise.
3224 (neon_vtbl1v8qi): Likewise.
3225 (neon_vtbl2v8qi): Likewise.
3226 (neon_vtbl3v8qi): Likewise.
3227 (neon_vtbl4v8qi): Likewise.
3228 (neon_vtbx1v8qi): Likewise.
3229 (neon_vtbx2v8qi): Likewise.
3230 (neon_vtbx3v8qi): Likewise.
3231 (neon_vtbx4v8qi): Likewise.
3232 (neon_vtrn<mode>_internal): Likewise.
3233 (neon_vzip<mode>_internal): Likewise.
3234 (neon_vuzp<mode>_internal): Likewise.
3235 (neon_vld1<mode>): Likewise.
3236 (neon_vld1_lane<mode>): Likewise.
3237 (neon_vld1_dup<mode>): Likewise.
3238 (neon_vld1_dupv2di): Likewise.
3239 (neon_vst1<mode>): Likewise.
3240 (neon_vst1_lane<mode>): Likewise.
3241 (neon_vld2<mode>): Likewise.
3242 (neon_vld2_lane<mode>): Likewise.
3243 (neon_vld2_dup<mode>): Likewise.
3244 (neon_vst2<mode>): Likewise.
3245 (neon_vst2_lane<mode>): Likewise.
3246 (neon_vld3<mode>): Likewise.
3247 (neon_vld3qa<mode>): Likewise.
3248 (neon_vld3qb<mode>): Likewise.
3249 (neon_vld3_lane<mode>): Likewise.
3250 (neon_vld3_dup<mode>): Likewise.
3251 (neon_vst3<mode>): Likewise.
3252 (neon_vst3qa<mode>): Likewise.
3253 (neon_vst3qb<mode>): Likewise.
3254 (neon_vst3_lane<mode>): Likewise.
3255 (neon_vld4<mode>): Likewise.
3256 (neon_vld4qa<mode>): Likewise.
3257 (neon_vld4qb<mode>): Likewise.
3258 (neon_vld4_lane<mode>): Likewise.
3259 (neon_vld4_dup<mode>): Likewise.
3260 (neon_vst4<mode>): Likewise.
3261 (neon_vst4qa<mode>): Likewise.
3262 (neon_vst4qb<mode>): Likewise.
3263 (neon_vst4_lane<mode>): Likewise.
3264 (neon_vec_unpack<US>_lo_<mode>): Likewise.
3265 (neon_vec_unpack<US>_hi_<mode>): Likewise.
3266 (neon_vec_<US>mult_lo_<mode>): Likewise.
3267 (neon_vec_<US>mult_hi_<mode>): Likewise.
3268 (neon_vec_<US>shiftl_<mode>): Likewise.
3269 (neon_unpack<US>_<mode>): Likewise.
3270 (neon_vec_<US>mult_<mode>): Likewise.
3271 (vec_pack_trunc_<mode>): Likewise.
3272 (neon_vec_pack_trunk_<mode>): Likewise.
3273 (neon_vabd<mode>_2): Likewise.
3274 (neon_vabd<mode>_3): Likewise.
3275 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
3276 (thumb2_movsi_vfp): Likewise.
3277 (movdi_vfp): Likewise.
3278 (movdi_vfp_cortexa8): Likewise.
3279 (movhf_vfp_neon): Likewise.
3280 (movhf_vfp): Likewiwse.
3281 (movsf_vfp): Likewiwse.
3282 (thumb2_movsf_vfp): Likewiwse.
3283 (movdf_vfp): Likewise.
3284 (thumb2_movdf_vfp): Likewise.
3285 (movsfcc_vfp): Likewise.
3286 (thumb2_movsfcc_vfp): Likewise.
3287 (movdfcc_vfp): Likewise.
3288 (thumb2_movdfcc_vfp): Likewise.
3289 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
3290 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
3291 (v10_v2c): Likewise.
3292 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
3294 (cortex_a15_neon_int_2): Likewise.
3295 (cortex_a15_neon_int_3): Likewise.
3296 (cortex_a15_neon_int_4): Likewise.
3297 (cortex_a15_neon_int_5): Likewise.
3298 (cortex_a15_neon_vqneg_vqabs): Likewise.
3299 (cortex_a15_neon_vmov): Likewise.
3300 (cortex_a15_neon_vaba): Likewise.
3301 (cortex_a15_neon_vaba_qqq): Likewise.
3302 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3303 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3304 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
3305 scalar_64_32_long_scalar): Likewise.
3306 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3307 (cortex_a15_neon_mla_qqq_8_16): Likewise.
3308 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
3309 lotype_qdd_64_32_long): Likewise.
3310 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3311 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3312 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
3313 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3314 (cortex_a15_neon_shift_1): Likewise.
3315 (cortex_a15_neon_shift_2): Likewise.
3316 (cortex_a15_neon_shift_3): Likewise.
3317 (cortex_a15_neon_vshl_ddd): Likewise.
3318 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3319 (cortex_a15_neon_vsra_vrsra): Likewise.
3320 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
3321 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
3322 (cortex_a15_neon_fp_vmul_ddd): Likewise.
3323 (cortex_a15_neon_fp_vmul_qqd): Likewise.
3324 (cortex_a15_neon_fp_vmla_ddd): Likewise.
3325 (cortex_a15_neon_fp_vmla_qqq): Likewise.
3326 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
3327 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
3328 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3329 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3330 (cortex_a15_neon_bp_simple): Likewise.
3331 (cortex_a15_neon_bp_2cycle): Likewise.
3332 (cortex_a15_neon_bp_3cycle): Likewise.
3333 (cortex_a15_neon_vld1_1_2_regs): Likewise.
3334 (cortex_a15_neon_vld1_3_4_regs): Likewise.
3335 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3336 (cortex_a15_neon_vld2_4_regs): Likewise.
3337 (cortex_a15_neon_vld3_vld4): Likewise.
3338 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3339 (cortex_a15_neon_vst1_3_4_regs): Likewise.
3340 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
3341 (cortex_a15_neon_vst3_vst4): Likewise.
3342 (cortex_a15_neon_vld1_vld2_lane): Likewise.
3343 (cortex_a15_neon_vld3_vld4_lane" 10
3344 (cortex_a15_neon_vst1_vst2_lane): Likewise.
3345 (cortex_a15_neon_vst3_vst4_lane): Likewise.
3346 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
3347 (cortex_a15_neon_ldm_2): Likewise.0
3348 (cortex_a15_neon_stm_2): Likewise.
3349 (cortex_a15_neon_mcr): Likewise.
3350 (cortex_a15_neon_mcr_2_mcrr): Likewise.
3351 (cortex_a15_neon_mrc): Likewise.
3352 (cortex_a15_neon_mrrc): Likewise.
3353 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
3355 (cortex_a15_alu_shift): Likewise.
3356 (cortex_a15_alu_shift_reg): Likewise.
3357 (cortex_a15_mult32): Likewise.
3358 (cortex_a15_mult64): Likewise.
3359 (cortex_a15_block): Likewise.
3360 (cortex_a15_branch): Likewise.
3361 (cortex_a15_load1): Likewise.
3362 (cortex_a15_load3): Likewise.
3363 (cortex_a15_store1): Likewise.
3364 (cortex_a15_store3): Likewise.
3365 (cortex_a15_call): Likewise.
3366 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
3367 (cortex_a5_f2r): Likewise.
3368 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
3370 (cortex_a53_f2r): Likewise.
3371 * config/arm/cortex-a7.md
3372 (cortex_a7_branch): Update for attribute change.
3373 (cortex_a7_call): Likewise.
3374 (cortex_a7_alu_imm): Likewise.
3375 (cortex_a7_alu_reg): Likewise.
3376 (cortex_a7_alu_shift): Likewise.
3377 (cortex_a7_mul): Likewise.
3378 (cortex_a7_load1): Likewise.
3379 (cortex_a7_store1): Likewise.
3380 (cortex_a7_load2): Likewise.
3381 (cortex_a7_store2): Likewise.
3382 (cortex_a7_load3): Likewise.
3383 (cortex_a7_store3): Likewise.
3384 (cortex_a7_load4): Likewise.
3385 (cortex_a7_store4): Likewise.
3386 (cortex_a7_fpalu): Likewise.
3387 (cortex_a7_fconst): Likewise.
3388 (cortex_a7_fpmuls): Likewise.
3389 (cortex_a7_neon_mul): Likewise.
3390 (cortex_a7_fpmacs): Likewise.
3391 (cortex_a7_neon_mla: Likewise.
3392 (cortex_a7_fpmuld: Likewise.
3393 (cortex_a7_fpmacd: Likewise.
3394 (cortex_a7_fpfmad: Likewise.
3395 (cortex_a7_fdivs: Likewise.
3396 (cortex_a7_fdivd: Likewise.
3397 (cortex_a7_r2f: Likewise.
3398 (cortex_a7_f2r: Likewise.
3399 (cortex_a7_f_flags: Likewise.
3400 (cortex_a7_f_loads: Likewise.
3401 (cortex_a7_f_loadd: Likewise.
3402 (cortex_a7_f_stores: Likewise.
3403 (cortex_a7_f_stored: Likewise.
3404 (cortex_a7_neon): Likewise.
3405 * config/arm/cortex-a8-neon.md
3406 (cortex_a8_neon_mrc): Update for attribute change.
3407 (cortex_a8_neon_mrrc): Likewise.
3408 (cortex_a8_neon_int_1): Likewise.
3409 (cortex_a8_neon_int_2): Likewise.
3410 (cortex_a8_neon_int_3): Likewise.
3411 (cortex_a8_neon_int_4): Likewise.
3412 (cortex_a8_neon_int_5): Likewise.
3413 (cortex_a8_neon_vqneg_vqabs): Likewise.
3414 (cortex_a8_neon_vmov): Likewise.
3415 (cortex_a8_neon_vaba): Likewise.
3416 (cortex_a8_neon_vaba_qqq): Likewise.
3417 (cortex_a8_neon_vsma): Likewise.
3418 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3419 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3420 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
3421 long_scalar): Likewise.
3422 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3423 (cortex_a8_neon_mla_qqq_8_16): Likewise.
3424 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
3425 long_scalar_qdd_64_32_long): Likewise.
3426 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3427 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3428 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
3429 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3430 (cortex_a8_neon_shift_1): Likewise.
3431 (cortex_a8_neon_shift_2): Likewise.
3432 (cortex_a8_neon_shift_3): Likewise.
3433 (cortex_a8_neon_vshl_ddd): Likewise.
3434 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3435 (cortex_a8_neon_vsra_vrsra): Likewise.
3436 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
3437 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
3438 (cortex_a8_neon_fp_vsum): Likewise.
3439 (cortex_a8_neon_fp_vmul_ddd): Likewise.
3440 (cortex_a8_neon_fp_vmul_qqd): Likewise.
3441 (cortex_a8_neon_fp_vmla_ddd): Likewise.
3442 (cortex_a8_neon_fp_vmla_qqq): Likewise.
3443 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
3444 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
3445 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3446 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3447 (cortex_a8_neon_bp_simple): Likewise.
3448 (cortex_a8_neon_bp_2cycle): Likewise.
3449 (cortex_a8_neon_bp_3cycle): Likewise.
3450 (cortex_a8_neon_ldr): Likewise.
3451 (cortex_a8_neon_str): Likewise.
3452 (cortex_a8_neon_vld1_1_2_regs): Likewise.
3453 (cortex_a8_neon_vld1_3_4_regs): Likewise.
3454 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3455 (cortex_a8_neon_vld2_4_regs): Likewise.
3456 (cortex_a8_neon_vld3_vld4): Likewise.
3457 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3458 (cortex_a8_neon_vst1_3_4_regs): Likewise.
3459 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
3460 (cortex_a8_neon_vst3_vst4): Likewise.
3461 (cortex_a8_neon_vld1_vld2_lane): Likewise.
3462 (cortex_a8_neon_vld3_vld4_lane): Likewise.
3463 (cortex_a8_neon_vst1_vst2_lane): Likewise.
3464 (cortex_a8_neon_vst3_vst4_lane): Likewise.
3465 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
3466 (cortex_a8_neon_mcr): Likewise.
3467 (cortex_a8_neon_mcr_2_mcrr): Likewise.
3468 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
3469 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
3471 (ca9_neon_mrrc): Likewise.
3472 (cortex_a9_neon_int_1): Likewise.
3473 (cortex_a9_neon_int_2): Likewise.
3474 (cortex_a9_neon_int_3): Likewise.
3475 (cortex_a9_neon_int_4): Likewise.
3476 (cortex_a9_neon_int_5): Likewise.
3477 (cortex_a9_neon_vqneg_vqabs): Likewise.
3478 (cortex_a9_neon_vmov): Likewise.
3479 (cortex_a9_neon_vaba): Likewise.
3480 (cortex_a9_neon_vaba_qqq): Likewise.
3481 (cortex_a9_neon_vsma): Likewise.
3482 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3483 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3484 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
3485 long_scalar): Likewise.
3486 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3487 (cortex_a9_neon_mla_qqq_8_16): Likewise.
3488 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
3489 long_scalar_qdd_64_32_long): Likewise.
3490 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3491 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3492 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
3493 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3494 (cortex_a9_neon_shift_1): Likewise.
3495 (cortex_a9_neon_shift_2): Likewise.
3496 (cortex_a9_neon_shift_3): Likewise.
3497 (cortex_a9_neon_vshl_ddd): Likewise.
3498 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3499 (cortex_a9_neon_vsra_vrsra): Likewise.
3500 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
3501 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
3502 (cortex_a9_neon_fp_vsum): Likewise.
3503 (cortex_a9_neon_fp_vmul_ddd): Likewise.
3504 (cortex_a9_neon_fp_vmul_qqd): Likewise.
3505 (cortex_a9_neon_fp_vmla_ddd): Likewise.
3506 (cortex_a9_neon_fp_vmla_qqq): Likewise.
3507 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
3508 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
3509 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3510 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3511 (cortex_a9_neon_bp_simple): Likewise.
3512 (cortex_a9_neon_bp_2cycle): Likewise.
3513 (cortex_a9_neon_bp_3cycle): Likewise.
3514 (cortex_a9_neon_ldr): Likewise.
3515 (cortex_a9_neon_str): Likewise.
3516 (cortex_a9_neon_vld1_1_2_regs): Likewise.
3517 (cortex_a9_neon_vld1_3_4_regs): Likewise.
3518 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3519 (cortex_a9_neon_vld2_4_regs): Likewise.
3520 (cortex_a9_neon_vld3_vld4): Likewise.
3521 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3522 (cortex_a9_neon_vst1_3_4_regs): Likewise.
3523 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
3524 (cortex_a9_neon_vst3_vst4): Likewise.
3525 (cortex_a9_neon_vld1_vld2_lane): Likewise.
3526 (cortex_a9_neon_vld3_vld4_lane): Likewise.
3527 (cortex_a9_neon_vst1_vst2_lane): Likewise.
3528 (cortex_a9_neon_vst3_vst4_lane): Likewise.
3529 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
3530 (cortex_a9_neon_mcr): Likewise.
3531 (cortex_a9_neon_mcr_2_mcrr): Likewise.
3532 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
3533 (cortex_a9_fps): Likewise.
3534 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
3536 (cortex_m4_fmuls): Likewise.
3537 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
3539 (cortex_r4_mrc): Likewise.
3540 * config/arm/iterators.md: Update comment referring to neon_type.
3541 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
3542 (iwmmxt_movsi_insn): Likewise.
3543 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
3545 (pj4_core_to_vfp): Likewise.
3546 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
3548 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
3549 (vfp_fstore): Likewise.
3550 * doc/md.texi: Change references to neon_type to refer to type.
3552 2013-09-04 Dodji Seketeli <dodji@redhat.com>
3554 * tree.h (DECL_BUILT_IN): Fix typo in comment.
3556 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
3558 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
3561 2013-09-04 Easwaran Raman <eraman@google.com>
3564 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
3565 (build_and_add_sum): Use it.
3566 (appears_later_in_bb): Simplify code.
3568 2013-09-04 Teresa Johnson <tejohnson@google.com>
3570 * dumpfile.c (dump_finish): Don't close stderr/stdout.
3572 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
3574 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
3576 2013-09-04 Jan Hubicka <jh@suse.cz>
3578 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
3579 * ipa-devirt.c: Include diganostic.h
3580 (odr_type_d): Add types and types_set.
3581 (hash_type_name): Work for types with vtables during LTO.
3582 (odr_hasher::remove): Fix comment; destroy types_set.
3583 (add_type_duplicate): New function,
3584 (get_odr_type): Use it.
3585 (dump_type_inheritance_graph): Dump type duplicates.
3586 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
3588 * tree.c (types_same_for_odr): Give exact answers on types with
3591 2013-09-04 Dodji Seketeli <dodji@redhat.com>
3593 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
3594 explaining their differences.
3596 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
3598 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
3600 2013-09-03 Jeff Law <law@redhat.com>
3602 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
3603 when not threading through a joiner block. Pass joiner/no joiner
3604 state to register_jump_thread.
3605 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
3606 state from argument rather than implying on path length.
3607 Dump the entire jump thread path into debugging dump.
3608 * tree-flow.h (register_jump_thread): Update prototype.
3610 2013-08-29 Xinliang David Li <davidxl@google.com>
3612 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
3613 Remove a trivial gcc_assert.
3615 2013-08-29 Xinliang David Li <davidxl@google.com>
3617 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
3618 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
3619 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
3620 Delay base decl alignment adjustment.
3621 * tree-vectorizer.c (vect_destroy_datarefs): New function.
3622 * tree-vectorizer.h: New data structure.
3623 (set_dr_misalignment): New function.
3624 (dr_misalignment): Ditto.
3625 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
3626 (vectorizable_load): Ditto.
3627 (ensure_base_align): New function.
3628 (vectorize_loops): Add dbg_cnt support.
3629 (execute_vect_slp): Ditto.
3630 * dbgcnt.def: New debug counter.
3631 * Makefile: New dependency.
3633 2013-09-03 Meador Inge <meadori@codesourcery.com>
3637 2013-08-30 Meador Inge <meadori@codesourcery.com>
3639 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
3641 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
3643 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
3644 function descriptor.
3646 2013-09-03 Richard Biener <rguenther@suse.de>
3648 * tree-affine.c (add_elt_to_tree): Fix association issue,
3649 avoid useless converts and make sure to always return a
3650 properly typed result.
3652 2013-09-03 Richard Biener <rguenther@suse.de>
3655 * fold-const.c (negate_expr_p): Fix division case.
3656 (negate_expr): Likewise.
3658 2013-09-03 Richard Biener <rguenther@suse.de>
3661 * tree-streamer-out.c: Include tm.h.
3662 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
3664 2013-09-03 Jan Hubicka <jh@suse.cz>
3666 * tree-profile.c (tree_profiling): Cleanup CFG when done.
3668 2013-09-03 Alan Modra <amodra@gmail.com>
3670 * config.gcc (powerpc*-*-linux*): Add support for little-endian
3671 multilibs to big-endian target and vice versa.
3672 * config/rs6000/t-linux64: Use := assignment on all vars.
3673 (MULTILIB_EXTRA_OPTS): Remove fPIC.
3674 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
3675 * config/rs6000/t-linux64le: New file.
3676 * config/rs6000/t-linux64bele: New file.
3677 * config/rs6000/t-linux64lebe: New file.
3679 2013-09-02 Jan Hubicka <jh@suse.cz>
3681 * ipa-inline-transform.c (inline_transform): Do not
3682 optimize_inline_calls when not optimizing.
3684 2013-09-02 Jan Hubicka <jh@suse.cz>
3686 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
3687 duplicated nodes for assembler names.
3688 * symtab.c (symtab_unregister_node): Do not attempt to unlink
3689 hard registers from assembler name hash.
3691 2013-09-02 Jan Hubicka <jh@suse.cz>
3693 * ipa-split.c (execute_split_functions): Split externally visible
3694 functions called once.
3696 2013-09-02 Martin Jambor <mjambor@suse.cz>
3699 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
3700 linked list. When finding the correct duplicate, also consider also
3701 the caller in additon to its inlined_to node.
3703 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
3705 * config/aarch64/aarch64-simd-builtins.def
3706 (dup_lane_scalar): Remove.
3707 * config/aarch64/aarch64-simd.md
3708 (aarch64_simd_dup): Add 'w->w' alternative.
3709 (aarch64_dup_lane<mode>): Allow for VALL.
3710 (aarch64_dup_lane_scalar<mode>): Remove.
3711 (aarch64_dup_lane_<vswap_width_name><mode>): New.
3712 (aarch64_get_lane_signed<mode>): Add w->w altenative.
3713 (aarch64_get_lane_unsigned<mode>): Likewise.
3714 (aarch64_get_lane<mode>): Likewise.
3715 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
3716 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
3717 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
3718 (VCON): Change container of V2SF.
3719 (vswap_width_name): Likewise.
3720 * config/aarch64/arm_neon.h
3721 (__aarch64_vdup_lane_any): New.
3722 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
3723 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
3724 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
3726 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
3729 * expr.c (emit_move_complex): Do not move complex FP values as parts if
3730 the source or the destination is a single hard register.
3732 2013-09-02 Richard Biener <rguenther@suse.de>
3735 * tree-scalar-evolution.c (instantiate_scev_name): Allow
3738 2013-09-02 Richard Biener <rguenther@suse.de>
3740 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
3741 arithmetic to sizetype.
3743 2013-09-02 Bin Cheng <bin.cheng@arm.com>
3745 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
3746 Find auto-increment use both before and after candidate.
3748 2013-09-02 Marek Polacek <polacek@redhat.com>
3750 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
3752 2013-09-01 Jan Hubicka <jh@suse.cz>
3754 * Makefile.in: Add ipa-profile.o
3755 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
3756 * cgraph.c (struct cgraph_propagate_frequency_data,
3757 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
3758 ipa-profile.c; replace cgraph_ by ipa_ prefix.
3759 * cgraph.h (cgraph_propagate_frequency): Remove.
3760 * ipa-inline-analysis.c: Include ipa-utils.h;
3761 drop duplicated cfgloop.h.
3762 (inline_update_callee_summaries): Update.
3763 * ipa-profile.c: New file.
3764 * ipa-utils.h (ipa_propagate_frequency): Declare.
3765 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
3766 data-streamer.h, value-prof.h.
3767 (symtab_remove_unreachable_nodes): Update profile.
3768 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
3769 account_time_size, cmp_counts, dump_histogram,
3770 ipa_profile_generate_summary, ipa_profile_write_summary,
3771 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
3772 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
3773 Move to ipa-profile.c.
3775 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
3777 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
3779 2013-09-01 Jan Hubicka <jh@suse.cz>
3781 * common.opt (fdevirtualize-speculatively): New function.
3782 * invoke.texi (fdevirtualize-speculatively): Document.
3783 * ipa-devirt.c: Include ipa-inline.h
3784 (likely_target_p): New function.
3785 (ipa_devirt): New function.
3786 (gate_ipa_devirt): New function.
3787 (pass_data_ipa_devirt): New static var.
3788 (pass_ipa_devirt): Likewise.
3789 (make_pass_ipa_devirt): New function.
3790 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
3791 (common_handle_option): Disable devirtualization when
3792 value range profiling is available.
3793 * passes.def (pass_ipa_devirt): Add.
3794 * timever.def (TV_IPA_DEVIRT): New timevar.
3795 * tree-pass.h (make_pass_ipa_devirt):
3797 2013-09-01 Iain Sandoe <iain@codesourcery.com>
3799 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
3800 include sanitize(undefined).
3802 2013-08-31 Diego Novillo <dnovillo@google.com>
3804 * Makefile.in (TREE_CORE_H): Define.
3806 (GTFILES): Add tree-core.h.
3807 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
3809 * tree-core.h: New file.
3810 Move all data structures, enum, typedefs, global
3811 declarations and constants from ...
3814 2013-08-31 Jan Hubicka <jh@suse.cz>
3816 * bulitins.c (expand_builtin): Do not early exit for gcov
3817 instrumented functions.
3819 2013-08-31 Marek Polacek <polacek@redhat.com>
3821 * ubsan.c: Include tm_p.h.
3823 2013-08-31 Jan Hubicka <jh@suse.cz>
3825 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
3828 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
3829 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
3830 * tree-cfg.c (verify_gimple_label): ... here.
3831 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
3832 (ipa_merge_profiles): New function.
3833 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
3834 (lto_input_function_body): Likewise.
3835 * ipa-utils.h (ipa_merge_profiles): Declare.
3836 * lto-streamer.h (lto_input_function_body): Update prototype.
3837 (emit_label_in_global_context_p): Remove.
3838 * lto-symtab.c: Include ipa-utils.h
3839 (lto_cgraph_replace_node): Use ipa_merge_profiles.
3841 2013-08-31 Jan Hubicka <jh@suse.cz>
3843 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
3845 2013-08-31 Jan Hubicka <jh@suse.cz>
3847 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
3849 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
3851 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
3852 "cmp" RTX before signed_comparison_operator check to account
3855 2013-08-30 Jan Hubicka <jh@suse.cz>
3857 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
3858 (detect_type_change_1): Rename to ...
3859 (detect_type_change): ... this one; early return on non-polymorphic
3861 (detect_type_change_ssa): Add comp_type parameter; update
3862 use of detect_type_change.
3863 (compute_complex_assign_jump_func): Add param_type parameter;
3864 update use of detect_type_change_ssa.
3865 (compute_complex_ancestor_jump_func): Likewise.
3866 (ipa_get_callee_param_type): New function.
3867 (ipa_compute_jump_functions_for_edge): Compute parameter type;
3868 update calls to the jump function computation functions.
3870 2013-08-30 Teresa Johnson <tejohnson@google.com>
3871 Steven Bosscher <steven@gcc.gnu.org>
3873 * cfgrtl.c (fixup_new_cold_bb): New routine.
3874 (commit_edge_insertions): Invoke fixup_partitions.
3875 (find_partition_fixes): New routine.
3876 (fixup_partitions): Ditto.
3877 (verify_hot_cold_block_grouping): Update comments.
3878 (rtl_verify_edges): Invoke find_partition_fixes.
3879 (rtl_verify_bb_pointers): Update comments.
3880 (rtl_verify_bb_layout): Ditto.
3881 * basic-block.h (probably_never_executed_edge_p): Declare.
3882 (fixup_partitions): Ditto.
3883 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
3884 * bb-reorder.c (sanitize_hot_paths): New function.
3885 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
3887 * predict.c (probably_never_executed_edge_p): New routine.
3888 * cfg.c (check_bb_profile): Add partition insanity warnings.
3890 2013-08-30 Meador Inge <meadori@codesourcery.com>
3892 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
3894 2013-08-30 Marek Polacek <polacek@redhat.com>
3896 * Makefile.in (ubsan.o): Add.
3897 (c-family/c-ubsan.o): Add.
3898 (builtins.o): Add ubsan.h dependency.
3899 * ubsan.h: New file.
3900 * ubsan.c: New file.
3901 * common.opt: Add -fsanitize=undefined option.
3902 (flag_sanitize): Add variable.
3903 (fsanitize=): Add option. Add Driver.
3904 (fsanitize=thread): Remove option.
3905 (fsanitize=address): Likewise.
3906 (static-libubsan): New option.
3907 * doc/invoke.texi: Document the new flag and -static-libubsan.
3908 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
3909 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
3910 * builtin-attrs.def (ATTR_COLD): Define.
3911 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
3912 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
3913 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
3914 * flag-types.h (sanitize_code): New enum.
3915 * opts.c (common_handle_option): Parse command line arguments
3916 of -fsanitize=. Add -fsanitize=unreachable option.
3917 * varasm.c (get_variable_section): Adjust.
3918 (assemble_noswitch_variable): Likewise.
3919 (assemble_variable): Likewise.
3920 (output_constant_def_contents): Likewise.
3921 (categorize_decl_for_section): Likewise.
3922 (place_block_symbol): Likewise.
3923 (output_object_block): Likewise.
3924 * builtins.def: Likewise.
3925 * toplev.c (compile_file): Likewise.
3926 (process_options): Likewise.
3927 * cppbuiltin.c: Likewise.
3928 * tsan.c (tsan_pass): Likewise.
3929 (tsan_gate): Likewise.
3930 (tsan_gate_O0): Likewise.
3931 * cfgexpand.c (partition_stack_vars): Likewise.
3932 (expand_stack_vars): Likewise.
3933 (defer_stack_allocation): Likewise.
3934 (expand_used_vars): Likewise.
3935 * cfgcleanup.c (old_insns_match_p): Likewise.
3936 * asan.c (asan_finish_file): Likewise.
3937 (asan_instrument): Likewise.
3938 (gate_asan): Likewise.
3939 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
3940 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
3941 (asan_global_struct): Use pointer_sized_int_node instead
3942 calling build_nonstandard_integer_type.
3943 (initialize_sanitizer_builtins): Likewise.
3944 (asan_finish_file): Likewise.
3945 * gcc.c: Document %{%:function(args):X}.
3946 (static_spec_functions): Add sanitize.
3947 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
3948 store funcval != NULL there.
3949 (do_spec_1): Adjust handle_spec_function caller.
3950 (handle_braces): Allow %:function(args) as condition.
3951 (sanitize_spec_function): New function.
3952 (ADD_STATIC_LIBUBSAN_LIBS): Define.
3953 (LIBUBSAN_SPEC): Likewise.
3954 (LIBUBSAN_EARLY_SPEC): Likewise.
3955 (SANITIZER_SPEC): Handle libubsan.
3956 (SANITIZER_EARLY_SPEC): Likewise.
3957 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
3958 instead of fsanitize=address.
3959 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
3960 instead of fsanitize=address*.
3961 * builtins.c: Include ubsan.h.
3962 (fold_builtin_0): Instrument __builtin_unreachable.
3963 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
3964 instead of flag_asan.
3965 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
3966 (pointer_sized_int_node): Define.
3967 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
3969 2013-08-30 Mike Stump <mikestump@comcast.net>
3971 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
3974 2013-08-29 Jan Hubicka <jh@suse.cz>
3976 * cgraph.c (cgraph_function_body_availability): Handle weakref
3978 * passes.def: Remove pass_fixup_cfg.
3979 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
3980 track when we need to remove functions.
3981 (gate_ipa_inline): Execute inlining always; add comment why.
3982 (pass_data_ipa_inline): Remove TODO_remove_functions.
3983 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
3984 do not produce summaries.
3985 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
3986 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
3987 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
3988 constant pool and vtable.
3990 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
3992 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
3993 New arm_neon.h's internal macros to specify 64-bit constants.
3994 Avoid using stdint.h's macros.
3996 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
3998 * recog.c (verify_changes): Verify that changes[i].old is non-zero
3999 before applying REG_P.
4001 2013-08-30 Jakub Jelinek <jakub@redhat.com>
4003 PR tree-optimization/58277
4004 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
4005 after seeing too many stmts with vdef in between dombb and current
4006 bb, invalidate everything.
4008 2013-08-30 Richard Biener <rguenther@suse.de>
4010 * fold-const.c (fold_single_bit_test): Fix overflow test.
4012 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
4014 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
4015 and which can live in a register, always retrieve the value on entry.
4016 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
4017 passed by invisible reference specially.
4018 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
4019 (vt_add_function_parameter): Correctly deal with a parameter passed by
4020 invisible reference.
4022 2013-08-30 Jan Hubicka <jh@suse.cz>
4024 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
4026 2013-08-30 Richard Biener <rguenther@suse.de>
4028 PR tree-optimization/58228
4029 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
4030 allow invariant loads in nested loop vectorization.
4032 2013-08-30 Richard Biener <rguenther@suse.de>
4034 PR tree-optimization/58223
4035 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
4036 (has_anti_or_output_dependence): ... this and adjust to also
4037 look for output dependences.
4038 (mark_nodes_having_upstream_mem_writes): Adjust.
4039 (rdg_flag_uses): Likewise.
4041 2013-08-30 Richard Biener <rguenther@suse.de>
4043 PR tree-optimization/58010
4044 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
4045 assert that we have a loop-closed PHI.
4047 2013-08-29 Jan Hubicka <jh@suse.cz>
4049 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
4050 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
4051 * lto-section-in.c (lto_free_function_in_decl_state): New function.
4052 (lto_free_function_in_decl_state_for_node): New function.
4054 2013-08-29 Xinliang David Li <davidxl@google.com>
4056 * loop-unroll.c (report_unroll_peel): Minor message change.
4057 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
4058 Emit alignment peeling message with default -fopt-info.
4059 (vect_loop_versioning): Emit loop version info message.
4060 * tree-vectorizer.c (vectorize_loops): Minor message change.
4061 (execute_vect_slp): Ditto.
4063 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
4065 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
4066 of the clone from the DECL_NAME of the original function.
4068 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
4070 * passes.c (register_pass): Add overload.
4071 * tree-pass.h (register_pass): Forward declare it. Add comment.
4073 2013-08-29 Jan Hubicka <jh@suse.cz>
4075 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
4076 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
4077 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
4078 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
4079 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
4080 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
4081 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
4082 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
4084 2013-08-29 Teresa Johnson <tejohnson@google.com>
4086 * dumpfile.c (dump_loc): Output column number.
4087 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
4088 * doc/invoke.texi: Document optall -fopt-info flag.
4089 * profile.c (read_profile_edge_counts): Use new dump framework.
4090 (compute_branch_probabilities): Ditto.
4091 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
4092 when pass not in any opt group.
4093 * pass_manager.h (pass_manager::get_pass_profile): New method.
4094 * value-prof.c (check_counter): Use new dump framework.
4095 (check_ic_target): Ditto.
4096 * coverage.c (get_coverage_counts): Ditto.
4097 (coverage_init): Setup new dump framework.
4099 2013-08-29 Richard Biener <rguenther@suse.de>
4101 PR tree-optimization/58246
4102 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
4103 handle the dominance check inside a basic-block.
4105 2013-08-29 Richard Biener <rguenther@suse.de>
4108 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
4109 of default defs that appear in abnormal PHI nodes.
4111 2013-08-29 Richard Biener <rguenther@suse.de>
4113 PR tree-optimization/57685
4114 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
4115 single-use operands to avoid exponential complexity.
4117 2013-08-28 Dehao Chen <dehao@google.com>
4119 * ipa-inline.c (edge_badness): Fix integer underflow.
4121 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
4123 * gtm-builtins.def (_ITM_free): Declare leaf.
4125 2013-08-28 Jakub Jelinek <jakub@redhat.com>
4128 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
4129 (*tls_local_dynamic_base_64_largepic): Likewise.
4130 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
4131 Remove predicate from call operand.
4132 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
4133 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
4135 2013-08-28 Jeff Law <law@redhat.com>
4137 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
4138 checks for the number of predecessors and successors allowed.
4139 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
4140 which require copying a joiner block if there is a request which
4141 is a subpath that requires no joiner block copying.
4143 2013-08-28 Jan Hubicka <jh@suse.cz>
4145 * lto-streamer-out.c (DFS_write_tree_body): Drop
4146 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
4147 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
4148 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
4149 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
4150 Do not read DECL_ERROR_ISSUED.
4151 (unpack_ts_decl_with_vis_value_fields): Do not read
4153 (lto_input_ts_binfo_tree_pointers): Do not read
4154 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
4155 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
4156 write DECL_ERROR_ISSUED..
4157 (pack_ts_decl_with_vis_value_fields): Do not write
4159 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
4160 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
4161 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
4162 * tree.h (tree_decl_common): Update comment.
4163 (DECL_ERROR_ISSUED): Remove.
4165 2013-08-28 Jakub Jelinek <jakub@redhat.com>
4168 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
4170 2013-08-28 Jan Hubicka <jh@suse.cz>
4172 * builtins.def (free): Declare leaf.
4174 2013-08-27 David Malcolm <dmalcolm@redhat.com>
4177 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
4178 * configure: Regenerate.
4180 2013-08-27 Martin Jambor <mjambor@suse.cz>
4182 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
4183 (ipa_ancestor_jf_data): Likewise.
4184 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
4185 (ipa_get_jf_pass_through_type_preserved): New function.
4186 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
4187 (ipa_get_jf_ancestor_type_preserved): New function.
4188 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
4189 (ipa_get_jf_ancestor_result): Likewise.
4190 (propagate_vals_accross_pass_through): Use
4191 ipa_get_jf_pass_through_result to do all the value mappings.
4192 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
4193 type_preserved flag.
4194 (ipa_set_jf_cst_copy): New function.
4195 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
4196 (ipa_set_jf_arith_pass_through): Likewise.
4197 (ipa_set_ancestor_jf): Likewise.
4198 (compute_complex_assign_jump_func): Set type_preserved instead of
4200 (ipa_compute_jump_functions_for_edge): Likewise.
4201 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
4202 (update_jump_functions_after_inlining): Update type_preserved.
4203 Explicitely create jump functions when combining one with pass_through.
4204 (ipa_write_jump_function): Stream the type_preserved flags.
4205 (ipa_read_jump_function): Likewise.
4207 2013-08-27 Jakub Jelinek <jakub@redhat.com>
4208 Aldy Hernandez <aldyh@redhat.com>
4210 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
4211 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
4212 * function.h (struct function): Add has_force_vect_loops and
4214 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
4215 * gimple.c (gimple_build_omp_critical): Add KIND argument and
4217 * gimple.def: Update CLAUSES comments.
4218 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
4219 (gimple_build_omp_for): Add argument to prototype.
4220 (gimple_omp_for_kind): New.
4221 (gimple_omp_for_set_kind): New.
4222 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
4223 GOVD_DATA_SHARE_CLASS.
4224 (enum omp_region_type): Add ORT_SIMD.
4225 (gimple_add_tmp_var): Handle ORT_SIMD.
4226 (gimplify_var_or_parm_decl): Same.
4227 (is_gimple_stmt): Same.
4228 (omp_firstprivatize_variable): Same.
4229 (omp_add_variable): Only use splay_tree_insert if lookup failed.
4230 (omp_notice_variable): Handle ORT_SIMD.
4231 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
4232 (omp_check_private): Handle ORT_SIMD.
4233 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
4235 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
4236 Handle OMP_CLAUSE_LASTPRIVATE.
4237 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
4239 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
4240 (gimplify_expr): Handle OMP_SIMD.
4241 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
4242 (expand_GOMP_SIMD_VF): New.
4243 (expand_GOMP_SIMD_LAST_LANE): New.
4244 * internal-fn.def (GOMP_SIMD_LANE): New.
4245 (GOMP_SIMD_VF): New.
4246 (GOMP_SIMD_LAST_LANE): New.
4247 * omp-low.c: Include target.h.
4248 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
4250 (check_omp_nesting_restrictions): Same.
4252 (lower_rec_simd_input_clauses): New.
4253 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
4255 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
4256 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
4257 (expand_omp_build_assign): New.
4258 (expand_omp_for_init_counts): New.
4259 (expand_omp_for_init_vars): New.
4260 (extract_omp_for_update_vars): New.
4261 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
4262 and rewrite accordingly.
4263 (expand_omp_simd): New.
4264 (expand_omp_for): Use expand_omp_simd.
4265 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
4266 (lower_omp_for): Do not lower the body.
4267 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
4269 * tree-flow.h (find_omp_clause): Remove prototype.
4270 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
4271 forcing vectorization of the loop, or if flag_tree_vectorize.
4272 (gate_tree_if_conversion): Similarly.
4273 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
4274 gimple_build_omp_for.
4275 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
4276 * tree-parloops (create_parallel_loop): Pass kind argument to
4277 gimple_build_omp_for.
4278 * tree-pretty-print.c (dump_omp_clause): Add cases for
4279 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
4280 OMP_CLAUSE__SIMDUID_.
4281 (dump_generic_node): Handle OMP_SIMD.
4282 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
4283 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
4284 unroll OMP_SIMD loops here.
4285 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
4286 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
4288 (vect_analyze_data_refs): Handle simd loops.
4289 * tree-vect-loop.c (vectorizable_live_operation): Handle
4291 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
4292 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
4293 (vectorizable_load): Same.
4294 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
4295 (struct simduid_to_vf): New.
4296 (simduid_to_vf::hash): New.
4297 (simduid_to-vf::equal): New.
4298 (struct simd_array_to_simduid): New.
4299 (simd_array_to_simduid::hash): New.
4300 (simd_array_to_simduid::equal): New.
4301 (adjust_simduid_builtins): New.
4302 (struct note_simd_array_uses_struct): New.
4303 (note_simd_array_uses_cb): New.
4304 (note_simd_array_uses): New.
4305 (vectorize_loops): Handle simd hints and adjust simd builtins
4307 * tree-vectorizer.h (struct _stmt_vec_info): Add
4308 simd_lane_access_p field.
4309 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
4310 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
4311 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
4312 (omp_clause_code_name): Same.
4313 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
4314 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
4315 * tree.def (OMP_SIMD): New entry.
4316 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
4317 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
4318 (OMP_CLAUSE_DECL): Adjust range for new clauses.
4319 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
4320 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
4321 (OMP_CLAUSE_LINEAR_STEP): New.
4322 (OMP_CLAUSE_SAFELEN_EXPR): New.
4323 (OMP_CLAUSE__SIMDUID__DECL): New.
4324 (find_omp_clause): New prototype.
4326 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
4328 * config/i386/driver-i386.c (host_detect_local_cpu): Update
4329 Haswell processor detection.
4331 2013-08-27 Christian Widmer <shadow@umbrox.de>
4334 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
4335 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
4336 AVX2 capable processors.
4338 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
4340 * config/aarch64/arm_neon.h: Replace all inline asm implementations
4341 of vget_low_* with implementations in terms of other intrinsics.
4343 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
4346 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
4347 values to -1, 0 and 1.
4349 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
4351 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
4352 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
4353 (<optab><mode>3_insn): Remove.
4354 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
4355 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
4356 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
4357 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
4358 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
4359 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
4360 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
4361 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
4362 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
4363 (ror<mode>3_insn): Likewise.
4364 * config/aarch64/predicates.md (aarch64_simd_register): New.
4366 2013-08-27 Richard Biener <rguenther@suse.de>
4368 PR tree-optimization/57521
4369 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
4370 one edge is non-critical.
4371 (find_phi_replacement_condition): Make sure to use a non-critical
4372 edge. Cleanup and remove old bug workarounds.
4373 (bb_postdominates_preds): Remove.
4374 (if_convertible_loop_p_1): Do not compute post-dominators.
4375 (combine_blocks): Do not free post-dominators.
4376 (main_tree_if_conversion): Likewise.
4377 (pass_data_if_conversion): Add TODO_verify_ssa.
4379 2013-08-27 DJ Delorie <dj@redhat.com>
4381 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
4383 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
4385 * function.c (assign_parm_find_data_types): Set passed_mode and
4386 nominal_mode to the TYPE_MODE of nominal_type for the built
4387 pointer type in case of the struct-pass-by-reference.
4389 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
4391 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
4392 (UINT16_TYPE): Change default to "unsigned int".
4394 * config/avr/avr.opt (mfract-convert-truncate): New option.
4395 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
4396 is set, round negative fractional integers according to n1169
4397 when converting to integer types.
4399 2013-08-26 Jan Hubicka <jh@suse.cz>
4401 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
4402 methods can not be called indirectly when their address is not taken.
4404 2013-08-26 Jan Hubicka <jh@suse.cz>
4406 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
4409 2013-08-26 Jan Hubicka <jh@suse.cz>
4411 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
4414 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
4416 * reload.c (find_valid_class): Allow classes that do not include
4417 FIRST_PSEUDO_REGISTER - 1.
4419 2013-08-26 Jan Hubicka <jh@suse.cz>
4421 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
4422 fix edge count/frequency when speculation failed; fix type check
4423 for the direct call.
4425 2013-08-26 Jan Hubicka <jh@suse.cz>
4427 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
4429 2013-08-26 Jan Hubicka <jh@suse.cz>
4431 * ipa-inline-transform.c (inline_transform): Be ready for basic block
4432 to be changed by edge redirection.
4434 2013-08-26 Jan Hubicka <jh@suse.cz>
4436 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
4437 formating; add sanity check.
4438 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
4439 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
4440 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
4441 (resolve_noninline_speculation): Update callee keys, too.
4443 2013-08-26 Jan Hubicka <jh@suse.cz>
4445 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
4446 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
4448 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
4450 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
4453 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
4455 * config/i386/i386.c (ix86_debug_options): Remove prototype.
4456 (x86_64_elf_select_section): Ditto.
4457 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
4459 (ix86_pass_by_reference): Ditto.
4460 (output_set_got): Ditto.
4461 (ix86_unary_operator_ok): Ditto.
4462 (ix86_expand_builtin): Ditto.
4464 2013-08-23 Jan Hubicka <jh@suse.cz>
4466 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
4468 2013-08-23 Jan Hubicka <jh@suse.cz>
4470 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
4471 (tree_decl_with_vis): Add FINAL field.
4473 2013-08-23 Jeff Law <law@redhat.com>
4475 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
4476 the debugging dump when the expression is fully redundant.
4478 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
4480 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
4481 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
4482 * pretty-print.c (pp_formatted_text_data): Likewise.
4483 (pp_write_text_to_stream): Likewise.
4484 (pp_write_text_as_dot_label_to_stream): Likewise.
4485 (pp_append_r): Likewise.
4486 (pp_format): Likewise.
4487 (pp_flush): Likewise.
4488 (pp_clear_output_area): Likewise.
4489 (pp_append_text): Likewise.
4490 (pp_formatted_text): Likewise.
4491 (pp_remaining_character_count_for_line): Likewise.
4492 (pp_newline): Likewise.
4493 (pp_character): Likewise.
4494 (output_buffer::~output_buffer): Define.
4495 (pretty_printer::~pretty_printer): Destruct output buffer.
4496 * pretty-print.h (output_buffer::~output_buffer): Declare.
4497 (pretty_printer::~pretty_printer): Declare virtual.
4499 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
4502 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
4503 HOST_WIDE_INT_M1U): New macros.
4504 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
4505 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
4506 unsigned -1 for lshift.
4507 * cse.c (cse_insn): Likewise.
4508 * double-int.c (rshift_double, lshift_double): Likewise.
4509 * builtins.c (fold_builtin_bitop): Likewise.
4510 * combine.c (force_to_mode): Likewise.
4511 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
4512 * simplify-rtx.c (simplify_const_unary_operation,
4513 simplify_const_binary_operation): Likewise.
4514 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
4515 check_va_list_escapes): Likewise.
4516 * rtlanal.c (nonzero_bits1): Likewise.
4517 * expmed.c (expand_smod_pow2): Likewise.
4518 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
4520 2013-08-23 Jan Hubicka <jh@suse.cz>
4522 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
4523 as having address taken.
4525 2013-08-23 Jan Hubicka <jh@suse.cz>
4527 * ipa-utils.h (method_class_type): Declare.
4528 * ipa-devirt.c (method_class_type): Export.
4530 * cgraphunit.c (analyze_functions): Do basic devirtualization;
4531 do not walk base classes of anonymous types.
4533 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
4535 PR rtl-optimization/58220
4537 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
4538 handle SEQUENCE insns properly.
4540 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
4542 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
4544 (pp_newline_and_indent): Likewise.
4545 (pp_separate_with): Likewise.
4546 * pretty-print.c (pp_newline_and_flush): Define.
4547 (pp_newline_and_indent): Likewise.
4548 (pp_separate_with): Likewise.
4550 2013-08-23 Jakub Jelinek <jakub@redhat.com>
4553 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
4554 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
4556 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
4558 * config/i386/predicates.md (ext_sse_reg_operand): New.
4559 * config/i386/i386.md (*movti_internal): Use
4560 predicate to determine if EVEX is needed.
4561 (*movsi_internal): Ditto.
4562 (*movdf_internal): Ditto.
4563 (*movsf_internal): Ditto.
4564 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
4566 2013-08-23 Jakub Jelinek <jakub@redhat.com>
4568 PR tree-optimization/58209
4569 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
4570 (find_tail_calls): Give up for pointer result types if m is non-NULL.
4571 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
4572 emit POINTER_PLUS_EXPR.
4573 (create_tailcall_accumulator): For pointer result type accumulate in
4576 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
4578 * configure.ac: Add backslashes missing from the last change.
4579 * configure: Regenerate.
4581 2013-08-22 Jan Hubicka <jh@susue.cz>
4583 * ipa.c (function_and_variable_visibility): First remember function
4584 was global and then make it local.
4586 2013-08-22 Julian Brown <julian@codesourcery.com>
4588 * configure.ac: Add aarch64 to list of arches which use "nop" in
4590 * configure: Regenerate.
4592 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4594 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
4595 gnu_libc_has_function.
4596 * config/s390/tpf.h: Likewise.
4598 2013-08-22 Jan Hubicka <jh@susue.cz>
4600 * timevar.c (validate_phases): Add cast.
4602 2013-08-22 Jan Hubicka <jh@susue.cz>
4604 * timevar.c (validate_phases): Use size_t for memory.
4605 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
4607 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
4609 * pretty-print.h (output_buffer::output_buffer): Declare.
4610 (pretty_printer::pretty_printer): Likewise.
4611 (pp_construct): Remove.
4612 * pretty-print.c (output_buffer::output_buffer): Define.
4613 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
4614 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
4615 (print_gimple_expr): Likewise.
4616 (print_gimple_seq): Likewise.
4617 (gimple_dump_bb): Likewise.
4618 * sched-vis.c (dump_value_slim): Likewise.
4619 (dump_insn_slim): Likewise.
4620 (dump_rtl_slim): Likewise.
4621 (str_pattern_slim): Likewise.
4622 * tree-mudflap.c (mf_varname_tree): Likewise.
4623 * graph.c (print_graph_cfg): Likewise.
4624 (start_graph_dump): Likewise.
4625 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
4627 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
4628 pretty printer initialization.
4629 * coretypes.h (diagnostic_context): Remove superflous type alias
4631 (pretty_printer): Likewise. Declare directly as a class.
4632 (pretty_print_info): Remove declaration as class.
4633 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
4634 and pp_clear_output_area.
4635 (asan_add_global): Likewise.
4637 2013-08-22 Jan Hubicka <jh@suse.cz>
4639 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
4640 * ipa-utils.h (update_type_inheritance_graph): Declare.
4641 (possible_polymorphic_call_target_p): Declare.
4642 (possible_polymorphic_call_target_p): New.
4643 * ipa-devirt.c: Update toplevel comments.
4644 (cached_polymorphic_call_targets): Move up.
4645 (odr_type_d): Move ID down.
4646 (polymorphic_type_binfo_p): Update comment.
4647 (odr_hasher::remove): Likewise;
4648 (get_odr_type): Set anonymous_namespace.
4649 (dump_odr_type): Dump it.
4650 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
4651 (maybe_record_node): Record node in cached_polymorphic_call_targets.
4652 (record_binfo): Add comment.
4653 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
4655 (devirt_node_removal_hook): Do not iCE when cache is freed.
4656 (possible_polymorphic_call_target_p): New predicate.
4657 (update_type_inheritance_graph): New function.
4659 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
4660 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4661 Sergey Lega <sergey.s.lega@intel.com>
4662 Anna Tikhonova <anna.tikhonova@intel.com>
4663 Ilya Tocar <ilya.tocar@intel.com>
4664 Andrey Turetskiy <andrey.turetskiy@intel.com>
4665 Ilya Verbin <ilya.verbin@intel.com>
4666 Kirill Yukhin <kirill.yukhin@intel.com>
4667 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4669 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
4670 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
4671 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
4672 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
4673 (OPTION_MASK_ISA_AVX2_UNSET): Update.
4674 (OPTION_MASK_ISA_AVX512F_UNSET): New.
4675 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
4676 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
4677 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
4678 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
4679 OPT_mavx512pf, OPT_mavx512er cases.
4680 * config/i386/constraints.md (v): New constraint.
4681 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
4682 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
4683 (bit_AVX512CD): New.
4684 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
4685 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
4686 * config/i386/i386-c.c (ix86_target_macros_internal):
4687 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
4689 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
4690 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
4691 * config/i386/i386.c (regclass_map, dbx_register_map)
4692 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
4693 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
4694 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
4696 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
4697 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
4698 -mavx512pf options. Fix formatting.
4699 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
4700 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
4701 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
4702 -mavx512cd, -mavx512pf options.
4703 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
4704 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
4705 (ix86_preferred_output_reload_class): Replace SSE_REGS with
4707 (ix86_hard_regno_mode_ok): Support 512-bit registers.
4708 (ix86_set_reg_reg_cost): Ditto.
4709 (x86_order_regs_for_local_alloc): Ditto.
4710 (MAX_VECT_LEN): Extend to 64-byte.
4711 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
4712 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
4713 (TARGET_AVX512ER, TARGET_AVX512CD): New.
4714 (BIGGEST_ALIGNMENT): Extend to 512-bits.
4715 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
4716 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
4717 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
4718 (SSE_REG_MODE_P): Support new modes.
4719 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
4720 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
4721 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
4722 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
4723 (REG_CLASS_CONTENTS): Add new registers.
4724 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
4725 (EXT_REX_SSE_REGNO_P): New.
4726 (HI_REGISTER_NAMES): Add new registers.
4727 * config/i386/i386.md: Define constants for new registers.
4728 (mode): Add new 512-bit modes.
4729 (prefix): Support evex prefix.
4730 (isa): Support avx512f, noavx512f, fma_avx512f.
4731 (ssemodesuffix): Add new 512-bit modes.
4733 (*movxi_internal_avx512f): Ditto.
4734 (*movdi_internal): Replace constraint "x" with the new constraint "v".
4736 (*movsi_internal): Likewise.
4737 (*movdf_internal): Likewise.
4738 (*movsf_internal): Likewise.
4739 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
4740 (<code><mode>3): Likewise.
4741 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
4743 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
4744 with the new constraint "v".
4745 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
4747 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
4748 with the new constraint "v".
4749 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
4750 (<sse2>_storedqu<avxsizesuffix>): Likewise.
4751 (*<plusminus_insn><mode>3): Likewise.
4752 (<sse>_vm<plusminus_insn><mode>3): Likewise.
4753 (*mul<mode>3): Likewise.
4754 (<sse>_vmmul<mode>3): Likewise.
4755 (<sse>_div<mode>3): Likewise.
4756 (<sse>_vmdiv<mode>3): Likewise.
4757 (<sse>_sqrt<mode>2): Likewise.
4758 (<sse>_vmsqrt<mode>2): Likewise.
4759 (*<code><mode>3_finite): Likewise.
4760 (*<code><mode>3) <smaxmin>: Likewise.
4761 (<sse>_vm<code><mode>3): Likewise.
4762 (*<code><mode>3) <any_logic>: Likewise.
4763 (*fma_fmadd_<mode>): Likewise.
4764 (*fma_fmsub_<mode>): Likewise.
4765 (*fma_fnmadd_<mode>): Likewise.
4766 (*fma_fnmsub_<mode>): Likewise.
4767 (*fma_fmaddsub_<mode>): Likewise.
4768 (*fma_fmsubadd_<mode>): Likewise.
4769 (*fmai_fmadd_<mode>): Likewise.
4770 (*fmai_fmsub_<mode>): Likewise.
4771 (*fmai_fnmadd_<mode>): Likewise.
4772 (*fmai_fnmsub_<mode>): Likewise.
4773 (sse_cvtsi2ss): Likewise.
4774 (sse_cvtsi2ssq): Likewise.
4775 (sse_cvtss2si): Likewise.
4776 (sse_cvtss2si_2): Likewise.
4777 (sse_cvtss2siq): Likewise.
4778 (sse_cvtss2siq_2): Likewise.
4779 (sse_cvttss2si): Likewise.
4780 (sse_cvtss2siq_2): Likewise.
4781 (float<sseintvecmodelower><mode>2): Likewise.
4782 (sse2_cvtsd2si_2): Likewise.
4783 (sse2_cvtsd2siq_2): Likewise.
4784 (*<plusminus_insn><mode>3): Likewise.
4785 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
4786 (*<sse4_1_avx2>_mul<mode>3): Likewise.
4787 (ashr<mode>3): Likewise.
4788 (<shift_insn><mode>3): Likewise.
4789 (avx2_<code><mode>3): Likewise.
4790 (*avx2_<code><mode>3): Likewise.
4791 (*andnot<mode>3): Likewise.
4792 (*<code><mode>3) <any_logic>: Likewise.
4793 (abs<mode>2): Likewise.
4794 (avx2_permvar<mode>): Likewise.
4795 (avx2_perm<mode>_1): Likewise.
4796 (*avx_vpermilp<mode>): Likewise.
4797 (avx_vpermilvar<mode>3): Likewise.
4798 (avx2_ashrv<mode>): Likewise.
4799 (avx2_<shift_insn>v<mode>): Likewise.
4800 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
4802 * doc/rtl.texi: Document XImode.
4804 2013-08-21 Jeff Law <law@redhat.com>
4806 * tree-flow.h (register_jump_thread): Pass vector of edges
4807 instead of each important edge.
4808 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
4809 thread path into a vector and pass that to register_jump_thread.
4810 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
4811 passed in edge vector to the current 3-edge form.
4814 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
4816 * dce.c (fini_dce): Call df_analyze again just in case
4817 delete_unmarked_insns removed anything.
4819 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
4821 * reload.h (struct reg_equivs): Rename to ..
4822 (struct reg_equivs_s): .. this.
4824 2013-08-20 Martin Liska <marxin.liska@gmail.com>
4826 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
4828 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4830 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
4832 2013-08-21 Jeff Law <law@redhat.com>
4834 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
4835 simplify assignments too. If the RHS collapses to a singleton
4836 range, then return the value for the range.
4838 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
4840 * config/i386/sse.md (V16): Rename to...
4842 (mov<mode>): Update iterator name.
4843 (*mov<mode>_internal): Ditto.
4844 (push<mode>1): Ditto.
4845 (movmisalign<mode>): Ditto.
4847 2013-08-20 Jan Hubicka <jh@suse.cz>
4850 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
4851 entry for direct edges.
4852 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
4854 2013-08-20 David Malcolm <dmalcolm@redhat.com>
4856 Revert my last two changes, r201865 and r201864:
4859 2013-08-20 David Malcolm <dmalcolm@redhat.com>
4861 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
4862 instances can own GC refs.
4864 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
4865 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
4866 (gcc::context::gt_pch_nx): Likewise.
4867 (gcc::context::gt_pch_nx): Likewise.
4868 * ggc.h (gt_ggc_mx <T>): New.
4869 (gt_pch_nx_with_op <T>): New.
4870 (gt_pch_nx <T>): New.
4871 * passes.c (opt_pass::gt_ggc_mx): New.
4872 (opt_pass::gt_pch_nx): New.
4873 (opt_pass::gt_pch_nx_with_op): New.
4874 (pass_manager::gt_ggc_mx): New.
4875 (pass_manager::gt_pch_nx): New.
4876 (pass_manager::gt_pch_nx_with_op): New.
4877 (pass_manager::operator new): Use
4878 ggc_internal_cleared_alloc_stat rather than xcalloc.
4879 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
4880 (pass_manager::gt_ggc_mx): New.
4881 (pass_manager::gt_pch_nx): New.
4882 (pass_manager::gt_pch_nx_with_op): New.
4883 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
4884 (opt_pass::operator new): New.
4885 (opt_pass::gt_ggc_mx): New.
4886 (opt_pass::gt_pch_nx): New.
4887 (opt_pass::gt_pch_nx_with_op): New.
4890 2013-08-20 David Malcolm <dmalcolm@redhat.com>
4892 * Makefile.in (GTFILES): Add context.h.
4893 * context.c (gcc::context::operator new): New.
4894 (gcc::context::gt_ggc_mx): New.
4895 (gcc::context::gt_pch_nx): New.
4896 (gcc::context::gt_pch_nx): New.
4897 * context.h (gcc::context): Add GTY((user)) marking.
4898 (gcc::context::operator new): New.
4899 (gcc::context::gt_ggc_mx): New.
4900 (gcc::context::gt_pch_nx): New.
4901 (gcc::context::gt_pch_nx): New.
4902 (g): Add GTY marking.
4903 (gt_ggc_mx (gcc::context *)): New.
4904 (gt_pch_nx (gcc::context *)): New.
4905 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
4906 void *cookie)): New.
4907 * gengtype.c (open_base_files) <ifiles>: Add context.h.
4909 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
4911 * dce.c (fini_dce): Call df_analyze again just in case
4912 delete_unmarked_insns removed anything.
4914 2013-08-20 Teresa Johnson <tejohnson@google.com>
4916 PR rtl-optimizations/57451
4917 * final.c (reemit_insn_block_notes): Prevent lexical blocks
4918 from crossing split section boundaries.
4920 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
4922 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
4923 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
4924 with MULTLIB_DEFAULTS.
4926 2013-08-20 Nick Clifton <nickc@redhat.com>
4928 * target.def (narrow_volatile_bitfield): Note that the default
4929 value is false, not !TARGET_STRICT_ALIGN.
4930 * doc/tm.texi: Regenerate.
4932 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
4934 Fix LIB_SPEC for systems without libpthread.
4936 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
4937 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
4939 * config/i386/linux-common.h: Likewise.
4940 * config/mips/linux-common.h: Likewise.
4942 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
4944 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
4947 2013-08-20 David Malcolm <dmalcolm@redhat.com>
4949 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
4950 instances can own GC refs.
4952 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
4953 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
4954 (gcc::context::gt_pch_nx): Likewise.
4955 (gcc::context::gt_pch_nx): Likewise.
4956 * ggc.h (gt_ggc_mx <T>): New.
4957 (gt_pch_nx_with_op <T>): New.
4958 (gt_pch_nx <T>): New.
4959 * passes.c (opt_pass::gt_ggc_mx): New.
4960 (opt_pass::gt_pch_nx): New.
4961 (opt_pass::gt_pch_nx_with_op): New.
4962 (pass_manager::gt_ggc_mx): New.
4963 (pass_manager::gt_pch_nx): New.
4964 (pass_manager::gt_pch_nx_with_op): New.
4965 (pass_manager::operator new): Use
4966 ggc_internal_cleared_alloc_stat rather than xcalloc.
4967 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
4968 (pass_manager::gt_ggc_mx): New.
4969 (pass_manager::gt_pch_nx): New.
4970 (pass_manager::gt_pch_nx_with_op): New.
4971 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
4972 (opt_pass::operator new): New.
4973 (opt_pass::gt_ggc_mx): New.
4974 (opt_pass::gt_pch_nx): New.
4975 (opt_pass::gt_pch_nx_with_op): New.
4977 2013-08-20 David Malcolm <dmalcolm@redhat.com>
4979 * Makefile.in (GTFILES): Add context.h.
4980 * context.c (gcc::context::operator new): New.
4981 (gcc::context::gt_ggc_mx): New.
4982 (gcc::context::gt_pch_nx): New.
4983 (gcc::context::gt_pch_nx): New.
4984 * context.h (gcc::context): Add GTY((user)) marking.
4985 (gcc::context::operator new): New.
4986 (gcc::context::gt_ggc_mx): New.
4987 (gcc::context::gt_pch_nx): New.
4988 (gcc::context::gt_pch_nx): New.
4989 (g): Add GTY marking.
4990 (gt_ggc_mx (gcc::context *)): New.
4991 (gt_pch_nx (gcc::context *)): New.
4992 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
4993 void *cookie)): New.
4994 * gengtype.c (open_base_files) <ifiles>: Add context.h.
4996 2013-08-20 Alan Modra <amodra@gmail.com>
4999 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
5000 (rs6000_emit_epilogue): Likewise.
5002 2013-08-19 Dehao Chen <dehao@google.com>
5004 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
5006 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
5007 Jakub Jelinek <jakub@redhat.com>
5009 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
5010 (BUILT_IN_FABSD64): Likewise.
5011 (BUILT_IN_FABSD128): Likewise.
5012 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
5013 (fold_builtin_1): Likewise.
5014 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
5015 destination and source operands.
5016 (*abstd2_fpr): Likewise.
5017 (*nabstd2_fpr): Likewise.
5019 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
5021 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
5024 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
5026 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
5029 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
5031 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
5032 * builtins.c (default_libc_has_function): New.
5033 (gnu_libc_has_function): Ditto.
5034 (no_c99_libc_has_function): Ditto.
5035 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
5036 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
5037 (fold_builtin_sincos): Likewise.
5038 (fold_builtin_cexp): Likewise.
5039 * builtins.def (DEF_C94_BUILTIN): Likewise.
5040 (DEF_C99_BUILTIN): Likewise.
5041 (DEF_C99_C90RES_BUILTIN): Likewise.
5042 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
5043 definitions to using this define.
5044 * config/darwin-protos.h (darwin_libc_has_function): New.
5045 * config/darwin.c (darwin_libc_has_function): Ditto.
5046 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
5047 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
5048 * config/darwin.h: Ditto.
5049 * config/elfos.h: Ditto.
5050 * config/freebsd.h: Ditto.
5051 * config/i386/cygming.h: Ditto.
5052 * config/i386/djgpp.h: Ditto.
5053 * config/i386/i386-interix.h: Ditto.
5054 * config/microblaze/microblaze.h: Ditto.
5055 * config/mmix/mmix.h: Ditto.
5056 * config/gnu-user.h: Ditto.
5057 * config/ia64/hpux.h: Ditto.
5058 * config/pa/pa-hpux.h: Ditto.
5059 * config/pdp11/pdp11.h: Ditto.
5060 * config/picochip/picochip.h: Ditto.
5061 * config/linux.h: Ditto.
5062 * config/netbsd.h: Ditto.
5063 * config/openbsd.h: Ditto.
5064 * config/rs6000/aix43.h: Ditto.
5065 * config/rs6000/aix51.h: Ditto.
5066 * config/rs6000/aix52.h: Ditto.
5067 * config/rs6000/aix53.h: Ditto.
5068 * config/rs6000/aix61.h: Ditto.
5069 * config/rs6000/darwin.h: Ditto.
5070 * config/rs6000/linux.h: Ditto.
5071 * config/rs6000/linux64.h: Ditto.
5072 * config/s390/tpf.h: Ditto.
5073 * config/sol2-10.h: Ditto.
5074 * config/sol2.h: Ditto.
5075 * config/vms/vms.h: Ditto.
5076 * config/vxworks.h: Ditto.
5077 * config/linux-android.c (linux_android_libc_has_function):
5078 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
5079 * config/linux-protos.h (linux_android_libc_has_function):
5081 * config/i386/i386.c (ix86_libc_has_function): New.
5082 * config/i386/i386-protos.h
5083 (ix86_libc_has_function): New declaration.
5084 * config/i386/i386.md
5085 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
5086 ("isinf<mode>2): Likewise.
5087 * convert.c (convert_to_integer): Using new target hook
5088 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
5089 TARGET_C99_FUNCTIONS.
5090 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
5091 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
5092 * coretypes.h (function_class): New enum for different
5093 classes of functions.
5094 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
5095 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
5096 (TARGET_HAS_SINCOS): Likewise.
5097 (TARGET_LIBC_HAS_FUNCTION): New.
5098 * doc/tm.texi: Regenerated.
5099 * targhooks.h (default_libc_has_function): New declaration.
5100 (no_c99_libc_has_function): Ditto.
5101 (gnu_libc_has_function): Ditto.
5102 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
5103 and TARGET_HAS_SINCOS.
5105 2013-08-18 Jan Hubicka <jh@suse.cz>
5107 * Makeifle-in (ipa-devirt.o): New.
5108 (GTFILES): Add ipa-utils.h and ipa-devirt.c
5109 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
5110 (analyze_functions): Look into possible targets of polymorphic call.
5111 * dumpfile.c (dump_files): Add type-inheritance dump.
5112 * dumpfile.h (TDI_inheritance): New.
5113 * ipa-devirt.c: New file.
5114 * ipa-utils.h (odr_type_d): Forward declare.
5115 (odr_type): New type.
5116 (build_type_inheritance_graph): Declare.
5117 (possible_polymorphic_call_targets): Declare and introduce inline
5118 variant when only edge is pased.
5119 (dump_possible_polymorphic_call_targets): Likewise.
5120 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
5121 * tree.c (type_in_anonymous_namespace_p): Break out from ...
5122 (types_same_for_odr): ... here.
5123 * tree.h (type_in_anonymous_namespace_p): Declare.
5125 2013-08-18 Jakub Jelinek <jakub@redhat.com>
5127 PR tree-optimization/58006
5128 * tree-parloops.c (take_address_of): Don't ICE if get_name
5130 (eliminate_local_variables_stmt): Remove clobber stmts.
5132 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
5134 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
5135 error message is issued for an alias to undefined symbol.
5137 2013-08-18 Jan Hubicka <jh@suse.cz>
5139 * cgraph.c (cgraph_create_indirect_edge): Discover
5140 polymorphic calls and record basic info into indirect_info.
5141 * gimple-fold.c (gimple_fold_call): When doing BINFO based
5142 devirtualization, ignore objc function calls.
5143 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
5144 call with no parm index info.
5145 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
5146 * tree.c (virtual_method_call_p): New function.
5147 * tree.h (virtual_method_call_p): Declare.
5149 2013-08-16 Jan Hubicka <jh@suse.cz>
5152 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
5154 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
5156 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
5159 2013-08-16 David Malcolm <dmalcolm@redhat.com>
5161 * gengtype.c (type_for_name): Add special-case support for
5162 locating types within the "gcc::" namespace.
5163 (open_base_files): Emit a "using namespace gcc" directive.
5165 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
5168 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
5169 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
5171 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
5172 array instead of each individual operand as a separate argument.
5173 (emit_fusion_gpr_load): Likewise.
5174 (expand_fusion_gpr_load): Add new function declaration.
5176 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
5177 signature to have the operands passed as an array, instead of as
5178 separate arguments. Allow ZERO_EXTEND to be in the memory
5179 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
5180 depend on the register live/dead flags when peepholes are run.
5181 (expand_fusion_gpr_load): New function to be called from the
5182 peephole2 pass, to change the register that addis sets to be the
5184 (emit_fusion_gpr_load): Change the calling signature to have the
5185 operands passed as an array, instead of as separate arguments.
5186 Allow ZERO_EXTEND to be in the memory address, and also
5187 SIGN_EXTEND if -mpower8-fusion-sign.
5189 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
5191 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
5192 adjust the register addis loads up in the peephole2 pass. Do not
5193 depend on the register live/dead state when the peephole pass is done.
5195 2013-08-16 David Malcolm <dmalcolm@redhat.com>
5197 * gengtype.c (create_user_defined_type): Ensure that the kind
5198 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
5199 declaration is seen before the GTY((user)) marking.
5201 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
5204 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
5206 2013-08-16 Jan Hubicka <jh@suse.cz>
5208 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
5209 arugment expected_type.
5210 (gimple_fold_call): Use it.
5211 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
5212 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
5213 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
5214 (try_make_edge_direct_virtual_call): Likewise.
5215 * tree.c (obj_type_ref_class): New.
5216 * tree.h (obj_type_ref_class): Use it.
5218 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
5220 * sched-vis.c (rtl_slim_pp_initialized): Remove.
5221 (rtl_slim_pp): Likewise.
5222 (init_rtl_slim_pretty_print): Likewise.
5223 (dump_value_slim): Don't call it. Use local pretty printer.
5224 (dump_insn_slim): Likewise.
5225 (dump_rtl_slim): Likewise.
5226 (str_pattern_slim): Likewise.
5227 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
5230 2013-08-16 Jakub Jelinek <jakub@redhat.com>
5232 PR tree-optimization/58164
5233 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
5234 walk gimple_goto_dest of GIMPLE_GOTO.
5236 PR tree-optimization/58165
5237 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
5238 bi_call must be the last stmt in a bb, don't split_block, instead
5239 use fallthru edge from it and give up if there is none.
5240 Release conds vector when returning early.
5242 2013-08-14 Xinliang David Li <davidxl@google.com>
5244 * config/i386/i386.c (ix86_option_override_internal):
5245 Remove unused variable and field.
5247 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5250 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
5251 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
5252 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
5253 and Linux, correct BLKmode alignment when 128-bit alignment is
5254 required and compatibility flag is not set.
5255 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
5256 for zero-size arguments when compatibility flag is not set.
5258 2013-08-14 Jakub Jelinek <jakub@redhat.com>
5260 PR tree-optimization/58145
5261 * tree-sra.c (build_ref_for_offset): If prev_base has
5262 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
5264 2013-08-14 Xinliang David Li <davidxl@google.com>
5266 * config/i386/i386.c (ix86_option_override_internal):
5267 Fix uninitialized variable error.
5269 2013-08-14 Xinliang David Li <davidxl@google.com>
5271 * config/i386/i386.opt: Define two new options.
5272 * config/i386/x86-tune.def: Add arch selector field in macros.
5273 * config/i386/i386.h: Adjust macro definition.
5274 * config/i386/i386.c (ix86_option_override_internal):
5276 (parse_mtune_ctrl_str): New function.
5277 (set_ix86_tune_features): New function.
5278 (ix86_function_specific_restore): Call the new helper function.
5280 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
5282 PR rtl-optimization/57662
5283 * sel-sched.c (code_motion_process_successors): When the current insn
5284 is removed after the recursive traversal, break from the loop.
5285 Add comments and debug printouts.
5287 2013-08-14 Jakub Jelinek <jakub@redhat.com>
5288 Alexandre Oliva <aoliva@redhat.com>
5291 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
5292 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
5293 there also UNSPEC_PLTOFF.
5295 2013-08-14 Marek Polacek <polacek@redhat.com>
5297 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
5298 NUM_CONDITIONS bit positions.
5300 2013-08-13 Cary Coutant <ccoutant@google.com>
5302 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
5303 (attr_checksum): Hash vector contents instead of pointer.
5304 (attr_checksum_ordered): Likewise.
5306 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
5308 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
5309 when Pmode != word_mode. Add length_address attribute.
5310 (sse3_monitor_<mode>): Merge from sse3_monitor and
5311 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
5312 Pmode != word_mode. Update insn length attribute.
5313 * config/i386/i386.c (ix86_option_override_internal): Update
5314 ix86_gen_monitor selection for merged sse3_monitor insn.
5316 2013-08-13 Julian Brown <julian@codesourcery.com>
5318 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
5319 perform invalid legitimization on greater-than-word-size modes for
5322 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
5324 * ira.c (setup_class_translate_array): Use aclass instead of cl
5325 for classes not fully covered by allocno classes.
5327 2013-08-13 Jakub Jelinek <jakub@redhat.com>
5329 PR tree-optimization/57661
5330 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
5331 * tree-inline.c (tree_function_versioning): Initialize it.
5332 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
5333 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
5334 that is not being copied.
5337 * asan.c (instrument_strlen_call): Fix typo in comment.
5338 Use char * type even for the lhs of POINTER_PLUS_EXPR.
5340 2013-08-13 Steve Ellcey <sellcey@mips.com>
5342 * config/mips/mips.md (prefetch): Use lw instead of ld on
5343 loongson in 32bit mode.
5345 2013-08-13 Nick Clifton <nickc@redhat.com>
5347 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
5349 2013-08-13 Jan Hubicka <jh@suse.cz>
5351 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
5352 introduced edge; fix typo in sanity check.
5353 (cgraph_resolve_speculation): Export; improve diagnostic.
5354 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
5355 speculation at type mismatch.
5356 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
5357 (cgraph_resolve_speculation): Declare.
5358 (symtab_can_be_discarded): New function.
5359 * value-prof.c (gimple_ic_transform): Remove actual transform code.
5360 * ipa-inline-transform.c (speculation_removed): New global var.
5361 (clone_inlined_nodes): See if speculation can be removed.
5362 (inline_call): If speculations was removed, we growths may not match.
5363 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
5364 (speculation_useful_p): New function.
5365 (resolve_noninline_speculation): New function.
5366 (inline_small_functions): Resolve useless speculations.
5367 * ipa-inline.h (speculation_useful_p): Declare
5368 * ipa.c (can_replace_by_local_alias): Simplify.
5369 (ipa_profile): Produce speculative calls in non-lto, too;
5370 add simple cost model; produce local aliases.
5372 2013-08-13 David Malcolm <dmalcolm@redhat.com>
5374 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
5377 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
5379 * config/i386/i386.c (ix86_function_versions): Use error + inform.
5381 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
5383 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
5384 iterator instead of X87MODEF.
5386 2013-08-12 Perez Read <netfirewall@gmail.com>
5389 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
5390 operand 0 for intel asm alternative.
5391 (*movabs<mode>_2): Ditto for operand 1.
5393 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
5395 * config/aarch64/arm_none.h
5396 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
5398 2013-08-12 Nick Clifton <nickc@redhat.com>
5400 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
5402 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
5404 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
5405 expand for QI/HImode operand to produce more effictive code for
5406 unsigned char(short) --> float(double) conversion.
5408 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
5410 * doc/invoke.texi: Mention that -ftls-model does not force the final
5413 2013-08-12 Marek Polacek <polacek@redhat.com>
5414 Marc Glisse <marc.glisse@inria.fr>
5416 PR tree-optimization/57980
5417 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
5418 when creating -1 constant.
5420 2013-08-10 Jan Hubicka <jh@suse.cz>
5422 Workaround binutils PR14342.
5423 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
5424 (gimple_init_edge_profiler): Likewise.
5425 (gimple_gen_ic_func_profiler): Likewise.
5427 2013-08-09 Jan Hubicka <jh@suse.cz>
5429 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
5431 2013-08-09 Xinliang David Li <davidxl@google.com>
5433 * config/i386/stringop.def: New file.
5434 * config/i386/stringop.opt: New file.
5435 * config/i386/i386-opts.h: Include stringopt.def.
5436 * config/i386/i386.opt: Include stringopt.opt.
5437 * config/i386/i386.c (ix86_option_override_internal):
5438 Override default size based stringop inline strategies with options.
5439 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
5442 2013-08-09 Jan Hubicka <jh@suse.cz>
5444 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
5446 2013-08-09 Jan Hubicka <jh@suse.cz>
5448 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
5450 (dump_cgraph_node): Dump profile-id.
5451 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
5452 and common_target_probability.
5453 * lto-cgraph.c (lto_output_edge): Stream common targets.
5454 (lto_output_node): Stream profile ids.
5455 (input_node): Stream profile ids.
5456 (input_edge): Stream common targets.
5457 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
5458 * ipa.c: Include value-prof.h
5459 (ipa_profile_generate_summary): Turn indirect call statement histograms
5460 into common targets.
5461 (ipa_profile): Turn common targets into speculative edges.
5463 2013-08-09 Jan Hubicka <jh@suse.cz>
5465 * cgraph.h (cgraph_node): Add profile_id.
5466 * value-prof.c (cgraph_node_map): Turn into pointer_map.
5467 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
5468 (del_node_map): Update.
5469 (find_func_by_funcdef_no): Replace by ...
5470 (find_func_by_profile_id): ... this one.
5471 (gimple_ic_transform): Do not remove useful histograms when
5472 speculation is not done; dump info when indirect call removal
5474 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
5475 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
5476 (__gcov_indirect_call_profiler_v2): .. this one.
5477 * profile.h (init_node_map): Update.
5478 * coverage.c (coverage_compute_profile_id): New function.
5479 * coverage.h (coverage_compute_profile_id): Declare.
5480 * tree-profile.c (init_ic_make_global_vars): Make
5481 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
5482 (gimple_init_edge_profiler): Update prototype of
5483 __gcov_indirect_call_profiler.
5484 (gimple_gen_ic_func_profiler): Simplify.
5485 (tree_profiling): Use init_node_map
5487 2013-08-09 Jan Hubicka <jh@suse.cz>
5489 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
5490 non-speculative refs.
5491 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
5492 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
5493 (cgraph_set_call_stmt): Likewise.
5494 (cgraph_create_edge_1): Fix release checking compilatoin;
5496 (cgraph_free_edge): Free indirect info.
5497 (cgraph_turn_edge_to_speculative): New function.
5498 (cgraph_speculative_call_info): New function.
5499 (cgraph_make_edge_direct): Return direct edge; handle speculation.
5500 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
5501 (dump_cgraph_node): Dump speculation.
5502 (verify_edge_count_and_frequency): Accept speculative edges.
5503 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
5504 (verify_cgraph_node): Handle speculation.
5505 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
5506 (cgraph_set_call_stmt): Update prototype.
5507 (cgraph_make_edge_direct): Update prototype.
5508 (cgraph_speculative_call_info): Declare.
5509 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
5510 to change; update call of ipa_find_references.
5511 * ipa-ref.c (ipa_record_reference): Fix return value; clear
5512 lto_stmt_uid and speculative flags.
5513 (ipa_dump_references): Dump speculation.
5514 (ipa_clone_references): Clone speculative flag.
5515 (ipa_clone_referring): Likewise.
5516 (ipa_clone_ref): New function.
5517 (ipa_find_reference): Look into lto_stmt_uids
5518 (ipa_clear_stmts_in_references): Do not clear speculative calls.
5519 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
5520 (ipa_find_reference): Update declaration.
5521 (ipa_clone_ref): Declare.
5522 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
5523 stream speculative flag.
5524 (lto_output_ref): Stream statements uids and speculation.
5525 (input_ref): Likewise.
5526 (input_edge): Stream speuclation.
5527 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
5528 (cgraph_set_call_stmt_including_clones): Handle speculation.
5529 * ipa-inline.c (heap_edge_removal_hook): New function.
5530 (inline_small_functions): Register it.
5531 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
5532 also initialize refs.
5533 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
5535 (try_make_edge_direct_simple_call): Likewise.
5536 (try_make_edge_direct_simple_call): Likewise.
5537 (update_indirect_edges_after_inlining): Likewise.
5538 (remove_described_reference): Look proper lto_stmt_uid.
5539 (propagate_controlled_uses): Likewise.
5540 (propagate_controlled_uses): Liekwise.
5541 * tree-inline.c (copy_bb): Copy speculative edges.
5542 (redirect_all_calls): New function.
5543 (copy_cfg_body): Do redirection after loop info is updated.
5544 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
5546 2013-08-09 Jan Hubicka <jh@suse.cz>
5548 * lto-streamer-out.c (output_function): Renumber PHIs.
5549 * lto-streamer-in.c (input_function): Likewise.
5551 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
5553 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
5554 (get_lane_unsigned): Likewise.
5555 (dup_lane_scalar): Likewise.
5556 (get_lane): enable for VALL.
5557 * config/aarch64/aarch64-simd.md
5558 (aarch64_dup_lane_scalar<mode>): Remove.
5559 (aarch64_get_lane_signed<mode>): Likewise.
5560 (aarch64_get_lane_unsigned<mode>): Likewise.
5561 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
5562 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
5563 (aarch64_get_lane<mode>): Enable for all vector modes.
5564 (aarch64_get_lanedi): Remove misleading constraints.
5565 * config/aarch64/arm_neon.h
5566 (__aarch64_vget_lane_any): Define.
5567 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
5568 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
5569 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
5570 * config/aarch64/iterators.md (VDQQH): New.
5574 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
5576 * configure.ac: Add GAS check for LEON instructions on SPARC.
5577 * configure: Regenerate.
5578 * config.in: Likewise.
5579 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
5581 * config/sparc/sparc.opt (LEON, LEON3): New masks.
5582 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
5584 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
5585 (AS_LEON_FLAG): New macro.
5586 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
5587 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
5588 Deal with LEON and LEON3 for the memory model.
5589 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
5590 (atomic_compare_and_swap<mode>_1): Likewise.
5591 (*atomic_compare_and_swap<mode>_1): Likewise.
5593 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5595 * config/arm/neon.md (vcond): Fix floating-point vector
5596 comparisons against 0.
5598 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
5600 * lra-constraints.c (emit_spill_move): Remove assert.
5601 (process_alt_operands): Add more debugging
5602 output. Increase reject for spilling into memory. Decrease
5603 reject for reloading scratch.
5604 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
5606 2013-08-08 Steve Ellcey <sellcey@mips.com>
5608 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
5609 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
5610 micromips incompatible. Add nan2008.
5611 (MULTILIB_DIRNAMES): Add nan2008.
5612 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
5613 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
5614 and micromips incompatible. Add nan2008.
5615 (MULTILIB_DIRNAMES): Add nan2008.
5616 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
5618 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
5620 PR rtl-optimization/58079
5621 * combine.c (combine_simplify_rtx): Avoid using SUBST if
5622 simplify_comparison has widened a comparison with an integer.
5624 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5626 * config/arm/neon.md (movmisalign<mode>): Disable when we
5627 don't allow unaligned accesses.
5628 (*movmisalign<mode>_neon_store): Likewise.
5629 (*movmisalign<mode>_neon_load): Likewise.
5630 (*movmisalign<mode>_neon_store): Likewise.
5631 (*movmisalign<mode>_neon_load): Likewise.
5633 2013-08-08 Jan Hubicka <jh@suse.cz>
5635 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
5636 (make_pass_rebuild_cgraph_edges): Also clear references.
5637 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
5638 * ipa-inline-transform.c (inline_transform): Remove all references
5640 * cgraphunit.c (expand_function): Remove all references after
5642 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
5643 (ipa_find_reference): Rewrite to iterator.
5644 (remove_stmt_references): Likewise.
5645 (ipa_clear_stmts_in_references): New function.
5646 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
5647 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
5649 * ipa-split.c (split_function): Remove references in split function.
5651 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
5654 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
5655 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
5657 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
5660 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
5661 suggested mode for the assignment isn't compatible with the
5664 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
5667 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
5669 2013-08-07 Xinliang David Li <davidxl@google.com>
5671 * config/i386/i386.opt: New option -mtune-ctrl=.
5672 * config/i386/x86-tune.def: New file.
5673 * config/i386/i386.h: include x86-tune.def.
5674 * config/i386/i386.c (ix86_option_override_internal):
5675 Parsing -mtune-ctrl= option and set tune features.
5677 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
5680 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
5681 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
5684 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
5686 * rtl.h (update_alignments): Declare.
5687 * final.c (grow_label_align): New function extracted from...
5688 (shorten_branches): ...here. Call it.
5689 (update_alignments): New function.
5690 * reorg.c (sibling_labels): New variable.
5691 (get_label_before): Add SIBLING parameter. If it is non-zero, push
5692 the new label along with it onto the sibling_labels vector.
5693 (fill_simple_delay_slots): Adjust call to get_label_before.
5694 (fill_slots_from_thread): Likewise.
5695 (relax_delay_slots): Likewise.
5696 (make_return_insns): Likewise.
5697 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
5699 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
5701 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
5702 document its semantics.
5703 (diagnostic_report_diagnostic): Adjust accordingly.
5705 2013-08-07 David Malcolm <dmalcolm@redhat.com>
5707 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
5708 (sparc_option_override): ...and port to new C++ pass API.
5709 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
5711 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
5713 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
5715 2013-08-06 Caroline Tice <cmtice@google.com>
5717 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
5718 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
5719 * tree-pass.h: Add pass_vtable_verify.
5720 * varasm.c (assemble_variable): Add code to properly set the comdat
5721 section and name for the .vtable_map_vars section.
5722 (assemble_vtyv_preinit_initializer): New function.
5723 (default_sectin_type_flags): Make sure .vtable_map_vars section has
5725 * output.h: Add function decl for assemble_vtv_preinit_initializer.
5726 * vtable-verify.c: New file.
5727 * vtable-verify.h: New file.
5728 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
5729 initialiation levels.
5730 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
5731 * passes.def: Insert pass_vtable_verify.
5732 * aclocal.m4: Reorder includes.
5733 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
5734 -fvtv-counts options.
5735 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
5736 as appropriate, if -fvtable-verify=... is used.
5737 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
5738 -fvtable-verify=... is used.
5739 * Makefile.in (OBJS): Add vtable-verify.o to list.
5740 (vtable-verify.o): Add new build rule.
5741 (GTFILES): Add vtable-verify.c to list.
5742 * common.opt (fvtable-verify=): New flag.
5743 (vtv_priority): Values for fvtable-verify= flag.
5744 (fvtv-counts): New flag.
5745 (fvtv-debug): New flag.
5746 * tree.h (save_vtable_map_decl): New extern function decl.
5748 2013-08-07 David Malcolm <dmalcolm@redhat.com>
5750 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
5751 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
5752 (pass_data_rl78_devirt): ...new pass_data instance and...
5753 (make_pass_rl78_devirt): ...new function.
5754 (rl78_asm_file_start): Port pass registration to new C++ API.
5756 2013-08-07 David Malcolm <dmalcolm@redhat.com>
5758 * coretypes.h (rtl_opt_pass): Add.
5759 (gcc::context): Add.
5760 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
5761 (epiphany_init): Port to new C++ pass API.
5762 (epiphany_optimize_mode_switching): Likewise.
5763 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
5764 (pass_manager::get_pass_mode_switching): New.
5765 (pass_manager::get_pass_peephole2): New.
5766 * mode-switching.c (pass_mode_switching): Add clone method.
5767 * recog.c (pass_peephole2): Add clone method.
5768 (pass_split_all_insns): Add clone method.
5770 2013-08-06 David Malcolm <dmalcolm@redhat.com>
5772 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
5773 (mips_option_override): ...here, porting to new C++ API for passes.
5775 2013-08-06 Jan Hubicka <jh@suse.cz>
5777 * cgraph.c (cgraph_get_body): New function based on lto.c
5779 * cgraph.h (cgraph_get_body): Declare.
5780 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
5782 * cgraphunit.c (expand_function): Get body prior expanding.
5783 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
5784 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
5786 * passes.c (do_per_function_toporder): Get body.
5787 * tree-inline.c (expand_call_inline): Get body prior inlining it.
5788 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
5790 2013-08-06 Martin Jambor <mjambor@suse.cz>
5793 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
5794 is not re-finalized. Rename second parameter to no_collect.
5796 2013-08-06 Martin Jambor <mjambor@suse.cz>
5799 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
5800 MEM_REF has proper alignment information.
5802 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
5805 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
5807 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
5808 rtx (*) (rtx, ...) with insn_gen_fn.
5809 * genoutput.c (output_insn_data): Cast gen_? function pointers to
5810 insn_gen_fn::stored_funcptr. Add initializer braces.
5812 2013-08-05 David Malcolm <dmalcolm@redhat.com>
5814 Rewrite how instances of passes are cloned to remove assumptions
5815 about their sizes (thus allowing pass subclasses to have
5816 additional data fields, albeit non-GC-managed ones at this point).
5818 * passes.c (make_pass_instance): Now that passes have clone
5819 methods, rewrite this function to eliminate XNEW and memcpy
5820 calls that used hardcoded sizes. Since this function no longer
5821 creates pass instances, rename it to...
5822 (add_pass_instance): ...this. Document the old way that passes were
5823 numbered and flagged, and rework this function to continue using it.
5824 (next_pass_1): Add an initial_pass argument for use by
5826 (position_pass): When adding multiple instances of a pass, use
5827 the pass's clone method, rather than relying on the XNEW/memcpy
5828 within the former make_pass_instance (now add_pass_instance).
5829 (pass_manager::pass_manager): When invoking next_pass_1, also supply
5830 the initial instance of the current pass within the pass manager.
5832 2013-08-05 David Malcolm <dmalcolm@redhat.com>
5834 This is the automated part of the conversion of passes from C
5835 structs to C++ classes.
5837 Patch autogenerated by refactor_passes.py from
5838 https://github.com/davidmalcolm/gcc-refactoring-scripts
5839 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
5841 * asan.c (pass_asan): Convert from a global struct to a subclass of
5842 gimple_opt_pass along with...
5843 (pass_data_asan): ...new pass_data instance and...
5844 (make_pass_asan): ...new function.
5845 (pass_asan_O0): Convert from a global struct to a subclass of
5846 gimple_opt_pass along with...
5847 (pass_data_asan_O0): ...new pass_data instance and...
5848 (make_pass_asan_O0): ...new function.
5849 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
5850 subclass of rtl_opt_pass along with...
5851 (pass_data_inc_dec): ...new pass_data instance and...
5852 (make_pass_inc_dec): ...new function.
5853 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
5854 a subclass of rtl_opt_pass along with...
5855 (pass_data_reorder_blocks): ...new pass_data instance and...
5856 (make_pass_reorder_blocks): ...new function.
5857 (pass_duplicate_computed_gotos): Convert from a global struct to a
5858 subclass of rtl_opt_pass along with...
5859 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
5860 (make_pass_duplicate_computed_gotos): ...new function.
5861 (pass_partition_blocks): Convert from a global struct to a subclass of
5862 rtl_opt_pass along with...
5863 (pass_data_partition_blocks): ...new pass_data instance and...
5864 (make_pass_partition_blocks): ...new function.
5865 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
5866 struct to a subclass of rtl_opt_pass along with...
5867 (pass_data_branch_target_load_optimize1): ...new pass_data instance
5869 (make_pass_branch_target_load_optimize1): ...new function.
5870 (pass_branch_target_load_optimize2): Convert from a global struct to a
5871 subclass of rtl_opt_pass along with...
5872 (pass_data_branch_target_load_optimize2): ...new pass_data instance
5874 (make_pass_branch_target_load_optimize2): ...new function.
5875 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
5876 of rtl_opt_pass along with...
5877 (pass_data_jump): ...new pass_data instance and...
5878 (make_pass_jump): ...new function.
5879 (pass_jump2): Convert from a global struct to a subclass of
5880 rtl_opt_pass along with...
5881 (pass_data_jump2): ...new pass_data instance and...
5882 (make_pass_jump2): ...new function.
5883 * cfgexpand.c (pass_expand): Convert from a global struct to a
5884 subclass of rtl_opt_pass along with...
5885 (pass_data_expand): ...new pass_data instance and...
5886 (make_pass_expand): ...new function.
5887 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
5888 of rtl_opt_pass along with...
5889 (pass_data_free_cfg): ...new pass_data instance and...
5890 (make_pass_free_cfg): ...new function.
5891 (pass_into_cfg_layout_mode): Convert from a global struct to a
5892 subclass of rtl_opt_pass along with...
5893 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
5894 (make_pass_into_cfg_layout_mode): ...new function.
5895 (pass_outof_cfg_layout_mode): Convert from a global struct to a
5896 subclass of rtl_opt_pass along with...
5897 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
5898 (make_pass_outof_cfg_layout_mode): ...new function.
5899 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
5900 struct to a subclass of gimple_opt_pass along with...
5901 (pass_data_build_cgraph_edges): ...new pass_data instance and...
5902 (make_pass_build_cgraph_edges): ...new function.
5903 (pass_rebuild_cgraph_edges): Convert from a global struct to a
5904 subclass of gimple_opt_pass along with...
5905 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
5906 (make_pass_rebuild_cgraph_edges): ...new function.
5907 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
5908 subclass of gimple_opt_pass along with...
5909 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
5911 (make_pass_remove_cgraph_callee_edges): ...new function.
5912 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
5913 struct to a subclass of rtl_opt_pass along with...
5914 (pass_data_stack_adjustments): ...new pass_data instance and...
5915 (make_pass_stack_adjustments): ...new function.
5916 * combine.c (pass_combine): Convert from a global struct to a subclass
5917 of rtl_opt_pass along with...
5918 (pass_data_combine): ...new pass_data instance and...
5919 (make_pass_combine): ...new function.
5920 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
5921 global struct to a subclass of rtl_opt_pass along with...
5922 (pass_data_compare_elim_after_reload): ...new pass_data instance
5924 (make_pass_compare_elim_after_reload): ...new function.
5925 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
5926 of rtl_opt_pass along with...
5927 (pass_data_rtl_cprop): ...new pass_data instance and...
5928 (make_pass_rtl_cprop): ...new function.
5929 * cse.c (pass_cse): Convert from a global struct to a subclass of
5930 rtl_opt_pass along with...
5931 (pass_data_cse): ...new pass_data instance and...
5932 (make_pass_cse): ...new function.
5933 (pass_cse2): Convert from a global struct to a subclass of
5934 rtl_opt_pass along with...
5935 (pass_data_cse2): ...new pass_data instance and...
5936 (make_pass_cse2): ...new function.
5937 (pass_cse_after_global_opts): Convert from a global struct to a
5938 subclass of rtl_opt_pass along with...
5939 (pass_data_cse_after_global_opts): ...new pass_data instance and...
5940 (make_pass_cse_after_global_opts): ...new function.
5941 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
5942 of rtl_opt_pass along with...
5943 (pass_data_ud_rtl_dce): ...new pass_data instance and...
5944 (make_pass_ud_rtl_dce): ...new function.
5945 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
5946 rtl_opt_pass along with...
5947 (pass_data_fast_rtl_dce): ...new pass_data instance and...
5948 (make_pass_fast_rtl_dce): ...new function.
5949 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
5950 a subclass of rtl_opt_pass along with...
5951 (pass_data_df_initialize_opt): ...new pass_data instance and...
5952 (make_pass_df_initialize_opt): ...new function.
5953 (pass_df_initialize_no_opt): Convert from a global struct to a
5954 subclass of rtl_opt_pass along with...
5955 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
5956 (make_pass_df_initialize_no_opt): ...new function.
5957 (pass_df_finish): Convert from a global struct to a subclass of
5958 rtl_opt_pass along with...
5959 (pass_data_df_finish): ...new pass_data instance and...
5960 (make_pass_df_finish): ...new function.
5961 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
5962 rtl_opt_pass along with...
5963 (pass_data_rtl_dse1): ...new pass_data instance and...
5964 (make_pass_rtl_dse1): ...new function.
5965 (pass_rtl_dse2): Convert from a global struct to a subclass of
5966 rtl_opt_pass along with...
5967 (pass_data_rtl_dse2): ...new pass_data instance and...
5968 (make_pass_rtl_dse2): ...new function.
5969 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
5970 subclass of rtl_opt_pass along with...
5971 (pass_data_dwarf2_frame): ...new pass_data instance and...
5972 (make_pass_dwarf2_frame): ...new function.
5973 * except.c (pass_set_nothrow_function_flags): Convert from a global
5974 struct to a subclass of rtl_opt_pass along with...
5975 (pass_data_set_nothrow_function_flags): ...new pass_data instance
5977 (make_pass_set_nothrow_function_flags): ...new function.
5978 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
5979 subclass of rtl_opt_pass along with...
5980 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
5982 (make_pass_convert_to_eh_region_ranges): ...new function.
5983 * final.c (pass_compute_alignments): Convert from a global struct to a
5984 subclass of rtl_opt_pass along with...
5985 (pass_data_compute_alignments): ...new pass_data instance and...
5986 (make_pass_compute_alignments): ...new function.
5987 (pass_final): Convert from a global struct to a subclass of
5988 rtl_opt_pass along with...
5989 (pass_data_final): ...new pass_data instance and...
5990 (make_pass_final): ...new function.
5991 (pass_shorten_branches): Convert from a global struct to a subclass of
5992 rtl_opt_pass along with...
5993 (pass_data_shorten_branches): ...new pass_data instance and...
5994 (make_pass_shorten_branches): ...new function.
5995 (pass_clean_state): Convert from a global struct to a subclass of
5996 rtl_opt_pass along with...
5997 (pass_data_clean_state): ...new pass_data instance and...
5998 (make_pass_clean_state): ...new function.
5999 * function.c (pass_instantiate_virtual_regs): Convert from a global
6000 struct to a subclass of rtl_opt_pass along with...
6001 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
6002 (make_pass_instantiate_virtual_regs): ...new function.
6003 (pass_leaf_regs): Convert from a global struct to a subclass of
6004 rtl_opt_pass along with...
6005 (pass_data_leaf_regs): ...new pass_data instance and...
6006 (make_pass_leaf_regs): ...new function.
6007 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
6008 subclass of rtl_opt_pass along with...
6009 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
6011 (make_pass_thread_prologue_and_epilogue): ...new function.
6012 (pass_match_asm_constraints): Convert from a global struct to a
6013 subclass of rtl_opt_pass along with...
6014 (pass_data_match_asm_constraints): ...new pass_data instance and...
6015 (make_pass_match_asm_constraints): ...new function.
6016 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
6017 subclass of rtl_opt_pass along with...
6018 (pass_data_rtl_fwprop): ...new pass_data instance and...
6019 (make_pass_rtl_fwprop): ...new function.
6020 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
6021 rtl_opt_pass along with...
6022 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
6023 (make_pass_rtl_fwprop_addr): ...new function.
6024 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
6025 rtl_opt_pass along with...
6026 (pass_data_rtl_pre): ...new pass_data instance and...
6027 (make_pass_rtl_pre): ...new function.
6028 (pass_rtl_hoist): Convert from a global struct to a subclass of
6029 rtl_opt_pass along with...
6030 (pass_data_rtl_hoist): ...new pass_data instance and...
6031 (make_pass_rtl_hoist): ...new function.
6032 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
6033 subclass of gimple_opt_pass along with...
6034 (pass_data_lower_cf): ...new pass_data instance and...
6035 (make_pass_lower_cf): ...new function.
6036 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
6037 from a global struct to a subclass of gimple_opt_pass along with...
6038 (pass_data_strength_reduction): ...new pass_data instance and...
6039 (make_pass_strength_reduction): ...new function.
6040 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
6041 of rtl_opt_pass along with...
6042 (pass_data_rtl_ifcvt): ...new pass_data instance and...
6043 (make_pass_rtl_ifcvt): ...new function.
6044 (pass_if_after_combine): Convert from a global struct to a subclass of
6045 rtl_opt_pass along with...
6046 (pass_data_if_after_combine): ...new pass_data instance and...
6047 (make_pass_if_after_combine): ...new function.
6048 (pass_if_after_reload): Convert from a global struct to a subclass of
6049 rtl_opt_pass along with...
6050 (pass_data_if_after_reload): ...new pass_data instance and...
6051 (make_pass_if_after_reload): ...new function.
6052 * init-regs.c (pass_initialize_regs): Convert from a global struct to
6053 a subclass of rtl_opt_pass along with...
6054 (pass_data_initialize_regs): ...new pass_data instance and...
6055 (make_pass_initialize_regs): ...new function.
6056 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
6057 of ipa_opt_pass_d along with...
6058 (pass_data_ipa_cp): ...new pass_data instance and...
6059 (make_pass_ipa_cp): ...new function.
6060 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
6061 global struct to a subclass of gimple_opt_pass along with...
6062 (pass_data_inline_parameters): ...new pass_data instance and...
6063 (make_pass_inline_parameters): ...new function.
6064 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
6065 subclass of gimple_opt_pass along with...
6066 (pass_data_early_inline): ...new pass_data instance and...
6067 (make_pass_early_inline): ...new function.
6068 (pass_ipa_inline): Convert from a global struct to a subclass of
6069 ipa_opt_pass_d along with...
6070 (pass_data_ipa_inline): ...new pass_data instance and...
6071 (make_pass_ipa_inline): ...new function.
6072 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
6073 struct to a subclass of gimple_opt_pass along with...
6074 (pass_data_local_pure_const): ...new pass_data instance and...
6075 (make_pass_local_pure_const): ...new function.
6076 (pass_ipa_pure_const): Convert from a global struct to a subclass of
6077 ipa_opt_pass_d along with...
6078 (pass_data_ipa_pure_const): ...new pass_data instance and...
6079 (make_pass_ipa_pure_const): ...new function.
6080 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
6081 to a subclass of ipa_opt_pass_d along with...
6082 (pass_data_ipa_reference): ...new pass_data instance and...
6083 (make_pass_ipa_reference): ...new function.
6084 * ipa-split.c (pass_split_functions): Convert from a global struct to
6085 a subclass of gimple_opt_pass along with...
6086 (pass_data_split_functions): ...new pass_data instance and...
6087 (make_pass_split_functions): ...new function.
6088 (pass_feedback_split_functions): Convert from a global struct to a
6089 subclass of gimple_opt_pass along with...
6090 (pass_data_feedback_split_functions): ...new pass_data instance and...
6091 (make_pass_feedback_split_functions): ...new function.
6092 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
6093 global struct to a subclass of simple_ipa_opt_pass along with...
6094 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
6096 (make_pass_ipa_function_and_variable_visibility): ...new function.
6097 (pass_ipa_free_inline_summary): Convert from a global struct to a
6098 subclass of simple_ipa_opt_pass along with...
6099 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
6100 (make_pass_ipa_free_inline_summary): ...new function.
6101 (pass_ipa_whole_program_visibility): Convert from a global struct to a
6102 subclass of ipa_opt_pass_d along with...
6103 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
6105 (make_pass_ipa_whole_program_visibility): ...new function.
6106 (pass_ipa_profile): Convert from a global struct to a subclass of
6107 ipa_opt_pass_d along with...
6108 (pass_data_ipa_profile): ...new pass_data instance and...
6109 (make_pass_ipa_profile): ...new function.
6110 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
6111 ipa_opt_pass_d along with...
6112 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
6113 (make_pass_ipa_cdtor_merge): ...new function.
6114 * ira.c (pass_ira): Convert from a global struct to a subclass of
6115 rtl_opt_pass along with...
6116 (pass_data_ira): ...new pass_data instance and...
6117 (make_pass_ira): ...new function.
6118 (pass_reload): Convert from a global struct to a subclass of
6119 rtl_opt_pass along with...
6120 (pass_data_reload): ...new pass_data instance and...
6121 (make_pass_reload): ...new function.
6122 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
6123 subclass of rtl_opt_pass along with...
6124 (pass_data_cleanup_barriers): ...new pass_data instance and...
6125 (make_pass_cleanup_barriers): ...new function.
6126 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
6127 of rtl_opt_pass along with...
6128 (pass_data_loop2): ...new pass_data instance and...
6129 (make_pass_loop2): ...new function.
6130 (pass_rtl_loop_init): Convert from a global struct to a subclass of
6131 rtl_opt_pass along with...
6132 (pass_data_rtl_loop_init): ...new pass_data instance and...
6133 (make_pass_rtl_loop_init): ...new function.
6134 (pass_rtl_loop_done): Convert from a global struct to a subclass of
6135 rtl_opt_pass along with...
6136 (pass_data_rtl_loop_done): ...new pass_data instance and...
6137 (make_pass_rtl_loop_done): ...new function.
6138 (pass_rtl_move_loop_invariants): Convert from a global struct to a
6139 subclass of rtl_opt_pass along with...
6140 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
6141 (make_pass_rtl_move_loop_invariants): ...new function.
6142 (pass_rtl_unswitch): Convert from a global struct to a subclass of
6143 rtl_opt_pass along with...
6144 (pass_data_rtl_unswitch): ...new pass_data instance and...
6145 (make_pass_rtl_unswitch): ...new function.
6146 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
6147 subclass of rtl_opt_pass along with...
6148 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
6150 (make_pass_rtl_unroll_and_peel_loops): ...new function.
6151 (pass_rtl_doloop): Convert from a global struct to a subclass of
6152 rtl_opt_pass along with...
6153 (pass_data_rtl_doloop): ...new pass_data instance and...
6154 (make_pass_rtl_doloop): ...new function.
6155 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
6156 a subclass of rtl_opt_pass along with...
6157 (pass_data_lower_subreg): ...new pass_data instance and...
6158 (make_pass_lower_subreg): ...new function.
6159 (pass_lower_subreg2): Convert from a global struct to a subclass of
6160 rtl_opt_pass along with...
6161 (pass_data_lower_subreg2): ...new pass_data instance and...
6162 (make_pass_lower_subreg2): ...new function.
6163 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
6164 struct to a subclass of ipa_opt_pass_d along with...
6165 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
6166 (make_pass_ipa_lto_gimple_out): ...new function.
6167 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
6168 of ipa_opt_pass_d along with...
6169 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
6170 (make_pass_ipa_lto_finish_out): ...new function.
6171 * mode-switching.c (pass_mode_switching): Convert from a global struct
6172 to a subclass of rtl_opt_pass along with...
6173 (pass_data_mode_switching): ...new pass_data instance and...
6174 (make_pass_mode_switching): ...new function.
6175 * modulo-sched.c (pass_sms): Convert from a global struct to a
6176 subclass of rtl_opt_pass along with...
6177 (pass_data_sms): ...new pass_data instance and...
6178 (make_pass_sms): ...new function.
6179 * omp-low.c (pass_expand_omp): Convert from a global struct to a
6180 subclass of gimple_opt_pass along with...
6181 (pass_data_expand_omp): ...new pass_data instance and...
6182 (make_pass_expand_omp): ...new function.
6183 (pass_lower_omp): Convert from a global struct to a subclass of
6184 gimple_opt_pass along with...
6185 (pass_data_lower_omp): ...new pass_data instance and...
6186 (make_pass_lower_omp): ...new function.
6187 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
6188 of gimple_opt_pass along with...
6189 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
6190 (make_pass_diagnose_omp_blocks): ...new function.
6191 * passes.c (pass_early_local_passes): Convert from a global struct to
6192 a subclass of simple_ipa_opt_pass along with...
6193 (pass_data_early_local_passes): ...new pass_data instance and...
6194 (make_pass_early_local_passes): ...new function.
6195 (pass_all_early_optimizations): Convert from a global struct to a
6196 subclass of gimple_opt_pass along with...
6197 (pass_data_all_early_optimizations): ...new pass_data instance and...
6198 (make_pass_all_early_optimizations): ...new function.
6199 (pass_all_optimizations): Convert from a global struct to a subclass
6200 of gimple_opt_pass along with...
6201 (pass_data_all_optimizations): ...new pass_data instance and...
6202 (make_pass_all_optimizations): ...new function.
6203 (pass_all_optimizations_g): Convert from a global struct to a subclass
6204 of gimple_opt_pass along with...
6205 (pass_data_all_optimizations_g): ...new pass_data instance and...
6206 (make_pass_all_optimizations_g): ...new function.
6207 (pass_rest_of_compilation): Convert from a global struct to a subclass
6208 of rtl_opt_pass along with...
6209 (pass_data_rest_of_compilation): ...new pass_data instance and...
6210 (make_pass_rest_of_compilation): ...new function.
6211 (pass_postreload): Convert from a global struct to a subclass of
6212 rtl_opt_pass along with...
6213 (pass_data_postreload): ...new pass_data instance and...
6214 (make_pass_postreload): ...new function.
6215 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
6216 subclass of rtl_opt_pass along with...
6217 (pass_data_gcse2): ...new pass_data instance and...
6218 (make_pass_gcse2): ...new function.
6219 * postreload.c (pass_postreload_cse): Convert from a global struct to
6220 a subclass of rtl_opt_pass along with...
6221 (pass_data_postreload_cse): ...new pass_data instance and...
6222 (make_pass_postreload_cse): ...new function.
6223 * predict.c (pass_profile): Convert from a global struct to a subclass
6224 of gimple_opt_pass along with...
6225 (pass_data_profile): ...new pass_data instance and...
6226 (make_pass_profile): ...new function.
6227 (pass_strip_predict_hints): Convert from a global struct to a subclass
6228 of gimple_opt_pass along with...
6229 (pass_data_strip_predict_hints): ...new pass_data instance and...
6230 (make_pass_strip_predict_hints): ...new function.
6231 * recog.c (pass_peephole2): Convert from a global struct to a subclass
6232 of rtl_opt_pass along with...
6233 (pass_data_peephole2): ...new pass_data instance and...
6234 (make_pass_peephole2): ...new function.
6235 (pass_split_all_insns): Convert from a global struct to a subclass of
6236 rtl_opt_pass along with...
6237 (pass_data_split_all_insns): ...new pass_data instance and...
6238 (make_pass_split_all_insns): ...new function.
6239 (pass_split_after_reload): Convert from a global struct to a subclass
6240 of rtl_opt_pass along with...
6241 (pass_data_split_after_reload): ...new pass_data instance and...
6242 (make_pass_split_after_reload): ...new function.
6243 (pass_split_before_regstack): Convert from a global struct to a
6244 subclass of rtl_opt_pass along with...
6245 (pass_data_split_before_regstack): ...new pass_data instance and...
6246 (make_pass_split_before_regstack): ...new function.
6247 (pass_split_before_sched2): Convert from a global struct to a subclass
6248 of rtl_opt_pass along with...
6249 (pass_data_split_before_sched2): ...new pass_data instance and...
6250 (make_pass_split_before_sched2): ...new function.
6251 (pass_split_for_shorten_branches): Convert from a global struct to a
6252 subclass of rtl_opt_pass along with...
6253 (pass_data_split_for_shorten_branches): ...new pass_data instance
6255 (make_pass_split_for_shorten_branches): ...new function.
6256 * ree.c (pass_ree): Convert from a global struct to a subclass of
6257 rtl_opt_pass along with...
6258 (pass_data_ree): ...new pass_data instance and...
6259 (make_pass_ree): ...new function.
6260 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
6261 subclass of rtl_opt_pass along with...
6262 (pass_data_stack_regs): ...new pass_data instance and...
6263 (make_pass_stack_regs): ...new function.
6264 (pass_stack_regs_run): Convert from a global struct to a subclass of
6265 rtl_opt_pass along with...
6266 (pass_data_stack_regs_run): ...new pass_data instance and...
6267 (make_pass_stack_regs_run): ...new function.
6268 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
6269 subclass of rtl_opt_pass along with...
6270 (pass_data_cprop_hardreg): ...new pass_data instance and...
6271 (make_pass_cprop_hardreg): ...new function.
6272 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
6273 subclass of rtl_opt_pass along with...
6274 (pass_data_reginfo_init): ...new pass_data instance and...
6275 (make_pass_reginfo_init): ...new function.
6276 * regmove.c (pass_regmove): Convert from a global struct to a subclass
6277 of rtl_opt_pass along with...
6278 (pass_data_regmove): ...new pass_data instance and...
6279 (make_pass_regmove): ...new function.
6280 * regrename.c (pass_regrename): Convert from a global struct to a
6281 subclass of rtl_opt_pass along with...
6282 (pass_data_regrename): ...new pass_data instance and...
6283 (make_pass_regrename): ...new function.
6284 * reorg.c (pass_delay_slots): Convert from a global struct to a
6285 subclass of rtl_opt_pass along with...
6286 (pass_data_delay_slots): ...new pass_data instance and...
6287 (make_pass_delay_slots): ...new function.
6288 (pass_machine_reorg): Convert from a global struct to a subclass of
6289 rtl_opt_pass along with...
6290 (pass_data_machine_reorg): ...new pass_data instance and...
6291 (make_pass_machine_reorg): ...new function.
6292 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
6293 of rtl_opt_pass along with...
6294 (pass_data_sched): ...new pass_data instance and...
6295 (make_pass_sched): ...new function.
6296 (pass_sched2): Convert from a global struct to a subclass of
6297 rtl_opt_pass along with...
6298 (pass_data_sched2): ...new pass_data instance and...
6299 (make_pass_sched2): ...new function.
6300 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
6301 to a subclass of rtl_opt_pass along with...
6302 (pass_data_stack_ptr_mod): ...new pass_data instance and...
6303 (make_pass_stack_ptr_mod): ...new function.
6304 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
6305 to a subclass of rtl_opt_pass along with...
6306 (pass_data_rtl_store_motion): ...new pass_data instance and...
6307 (make_pass_rtl_store_motion): ...new function.
6308 * tracer.c (pass_tracer): Convert from a global struct to a subclass
6309 of gimple_opt_pass along with...
6310 (pass_data_tracer): ...new pass_data instance and...
6311 (make_pass_tracer): ...new function.
6312 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
6313 to a subclass of gimple_opt_pass along with...
6314 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
6315 (make_pass_diagnose_tm_blocks): ...new function.
6316 (pass_lower_tm): Convert from a global struct to a subclass of
6317 gimple_opt_pass along with...
6318 (pass_data_lower_tm): ...new pass_data instance and...
6319 (make_pass_lower_tm): ...new function.
6320 (pass_tm_init): Convert from a global struct to a subclass of
6321 gimple_opt_pass along with...
6322 (pass_data_tm_init): ...new pass_data instance and...
6323 (make_pass_tm_init): ...new function.
6324 (pass_tm_mark): Convert from a global struct to a subclass of
6325 gimple_opt_pass along with...
6326 (pass_data_tm_mark): ...new pass_data instance and...
6327 (make_pass_tm_mark): ...new function.
6328 (pass_tm_edges): Convert from a global struct to a subclass of
6329 gimple_opt_pass along with...
6330 (pass_data_tm_edges): ...new pass_data instance and...
6331 (make_pass_tm_edges): ...new function.
6332 (pass_tm_memopt): Convert from a global struct to a subclass of
6333 gimple_opt_pass along with...
6334 (pass_data_tm_memopt): ...new pass_data instance and...
6335 (make_pass_tm_memopt): ...new function.
6336 (pass_ipa_tm): Convert from a global struct to a subclass of
6337 simple_ipa_opt_pass along with...
6338 (pass_data_ipa_tm): ...new pass_data instance and...
6339 (make_pass_ipa_tm): ...new function.
6340 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
6341 subclass of gimple_opt_pass along with...
6342 (pass_data_call_cdce): ...new pass_data instance and...
6343 (make_pass_call_cdce): ...new function.
6344 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
6345 subclass of gimple_opt_pass along with...
6346 (pass_data_build_cfg): ...new pass_data instance and...
6347 (make_pass_build_cfg): ...new function.
6348 (pass_split_crit_edges): Convert from a global struct to a subclass of
6349 gimple_opt_pass along with...
6350 (pass_data_split_crit_edges): ...new pass_data instance and...
6351 (make_pass_split_crit_edges): ...new function.
6352 (pass_warn_function_return): Convert from a global struct to a
6353 subclass of gimple_opt_pass along with...
6354 (pass_data_warn_function_return): ...new pass_data instance and...
6355 (make_pass_warn_function_return): ...new function.
6356 (pass_warn_function_noreturn): Convert from a global struct to a
6357 subclass of gimple_opt_pass along with...
6358 (pass_data_warn_function_noreturn): ...new pass_data instance and...
6359 (make_pass_warn_function_noreturn): ...new function.
6360 (pass_warn_unused_result): Convert from a global struct to a subclass
6361 of gimple_opt_pass along with...
6362 (pass_data_warn_unused_result): ...new pass_data instance and...
6363 (make_pass_warn_unused_result): ...new function.
6364 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
6365 a subclass of gimple_opt_pass along with...
6366 (pass_data_merge_phi): ...new pass_data instance and...
6367 (make_pass_merge_phi): ...new function.
6368 * tree-complex.c (pass_lower_complex): Convert from a global struct to
6369 a subclass of gimple_opt_pass along with...
6370 (pass_data_lower_complex): ...new pass_data instance and...
6371 (make_pass_lower_complex): ...new function.
6372 (pass_lower_complex_O0): Convert from a global struct to a subclass of
6373 gimple_opt_pass along with...
6374 (pass_data_lower_complex_O0): ...new pass_data instance and...
6375 (make_pass_lower_complex_O0): ...new function.
6376 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
6377 subclass of gimple_opt_pass along with...
6378 (pass_data_lower_eh): ...new pass_data instance and...
6379 (make_pass_lower_eh): ...new function.
6380 (pass_refactor_eh): Convert from a global struct to a subclass of
6381 gimple_opt_pass along with...
6382 (pass_data_refactor_eh): ...new pass_data instance and...
6383 (make_pass_refactor_eh): ...new function.
6384 (pass_lower_resx): Convert from a global struct to a subclass of
6385 gimple_opt_pass along with...
6386 (pass_data_lower_resx): ...new pass_data instance and...
6387 (make_pass_lower_resx): ...new function.
6388 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
6389 of gimple_opt_pass along with...
6390 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
6391 (make_pass_lower_eh_dispatch): ...new function.
6392 (pass_cleanup_eh): Convert from a global struct to a subclass of
6393 gimple_opt_pass along with...
6394 (pass_data_cleanup_eh): ...new pass_data instance and...
6395 (make_pass_cleanup_eh): ...new function.
6396 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
6397 to a subclass of simple_ipa_opt_pass along with...
6398 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
6399 (make_pass_ipa_lower_emutls): ...new function.
6400 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
6401 a subclass of gimple_opt_pass along with...
6402 (pass_data_if_conversion): ...new pass_data instance and...
6403 (make_pass_if_conversion): ...new function.
6404 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
6405 subclass of gimple_opt_pass along with...
6406 (pass_data_build_ssa): ...new pass_data instance and...
6407 (make_pass_build_ssa): ...new function.
6408 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
6409 global struct to a subclass of gimple_opt_pass along with...
6410 (pass_data_loop_distribution): ...new pass_data instance and...
6411 (make_pass_loop_distribution): ...new function.
6412 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
6413 subclass of gimple_opt_pass along with...
6414 (pass_data_mudflap_1): ...new pass_data instance and...
6415 (make_pass_mudflap_1): ...new function.
6416 (pass_mudflap_2): Convert from a global struct to a subclass of
6417 gimple_opt_pass along with...
6418 (pass_data_mudflap_2): ...new pass_data instance and...
6419 (make_pass_mudflap_2): ...new function.
6420 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
6421 subclass of gimple_opt_pass along with...
6422 (pass_data_mudflap_1): ...new pass_data instance and...
6423 (make_pass_mudflap_1): ...new function.
6424 (pass_mudflap_2): Convert from a global struct to a subclass of
6425 gimple_opt_pass along with...
6426 (pass_data_mudflap_2): ...new pass_data instance and...
6427 (make_pass_mudflap_2): ...new function.
6428 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
6429 gimple_opt_pass along with...
6430 (pass_data_nrv): ...new pass_data instance and...
6431 (make_pass_nrv): ...new function.
6432 (pass_return_slot): Convert from a global struct to a subclass of
6433 gimple_opt_pass along with...
6434 (pass_data_return_slot): ...new pass_data instance and...
6435 (make_pass_return_slot): ...new function.
6436 * tree-object-size.c (pass_object_sizes): Convert from a global struct
6437 to a subclass of gimple_opt_pass along with...
6438 (pass_data_object_sizes): ...new pass_data instance and...
6439 (make_pass_object_sizes): ...new function.
6440 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
6441 global struct to a subclass of gimple_opt_pass along with...
6442 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
6444 (make_pass_cleanup_cfg_post_optimizing): ...new function.
6445 (pass_fixup_cfg): Convert from a global struct to a subclass of
6446 gimple_opt_pass along with...
6447 (pass_data_fixup_cfg): ...new pass_data instance and...
6448 (make_pass_fixup_cfg): ...new function.
6449 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
6450 (make_pass_mudflap_1): ...new function.
6451 (pass_mudflap_2): Replace declaration with that of...
6452 (make_pass_mudflap_2): ...new function.
6453 (pass_asan): Replace declaration with that of...
6454 (make_pass_asan): ...new function.
6455 (pass_asan_O0): Replace declaration with that of...
6456 (make_pass_asan_O0): ...new function.
6457 (pass_tsan): Replace declaration with that of...
6458 (make_pass_tsan): ...new function.
6459 (pass_tsan_O0): Replace declaration with that of...
6460 (make_pass_tsan_O0): ...new function.
6461 (pass_lower_cf): Replace declaration with that of...
6462 (make_pass_lower_cf): ...new function.
6463 (pass_refactor_eh): Replace declaration with that of...
6464 (make_pass_refactor_eh): ...new function.
6465 (pass_lower_eh): Replace declaration with that of...
6466 (make_pass_lower_eh): ...new function.
6467 (pass_lower_eh_dispatch): Replace declaration with that of...
6468 (make_pass_lower_eh_dispatch): ...new function.
6469 (pass_lower_resx): Replace declaration with that of...
6470 (make_pass_lower_resx): ...new function.
6471 (pass_build_cfg): Replace declaration with that of...
6472 (make_pass_build_cfg): ...new function.
6473 (pass_early_tree_profile): Replace declaration with that of...
6474 (make_pass_early_tree_profile): ...new function.
6475 (pass_cleanup_eh): Replace declaration with that of...
6476 (make_pass_cleanup_eh): ...new function.
6477 (pass_sra): Replace declaration with that of...
6478 (make_pass_sra): ...new function.
6479 (pass_sra_early): Replace declaration with that of...
6480 (make_pass_sra_early): ...new function.
6481 (pass_early_ipa_sra): Replace declaration with that of...
6482 (make_pass_early_ipa_sra): ...new function.
6483 (pass_tail_recursion): Replace declaration with that of...
6484 (make_pass_tail_recursion): ...new function.
6485 (pass_tail_calls): Replace declaration with that of...
6486 (make_pass_tail_calls): ...new function.
6487 (pass_tree_loop): Replace declaration with that of...
6488 (make_pass_tree_loop): ...new function.
6489 (pass_tree_loop_init): Replace declaration with that of...
6490 (make_pass_tree_loop_init): ...new function.
6491 (pass_lim): Replace declaration with that of...
6492 (make_pass_lim): ...new function.
6493 (pass_tree_unswitch): Replace declaration with that of...
6494 (make_pass_tree_unswitch): ...new function.
6495 (pass_predcom): Replace declaration with that of...
6496 (make_pass_predcom): ...new function.
6497 (pass_iv_canon): Replace declaration with that of...
6498 (make_pass_iv_canon): ...new function.
6499 (pass_scev_cprop): Replace declaration with that of...
6500 (make_pass_scev_cprop): ...new function.
6501 (pass_empty_loop): Replace declaration with that of...
6502 (make_pass_empty_loop): ...new function.
6503 (pass_record_bounds): Replace declaration with that of...
6504 (make_pass_record_bounds): ...new function.
6505 (pass_graphite): Replace declaration with that of...
6506 (make_pass_graphite): ...new function.
6507 (pass_graphite_transforms): Replace declaration with that of...
6508 (make_pass_graphite_transforms): ...new function.
6509 (pass_if_conversion): Replace declaration with that of...
6510 (make_pass_if_conversion): ...new function.
6511 (pass_loop_distribution): Replace declaration with that of...
6512 (make_pass_loop_distribution): ...new function.
6513 (pass_vectorize): Replace declaration with that of...
6514 (make_pass_vectorize): ...new function.
6515 (pass_slp_vectorize): Replace declaration with that of...
6516 (make_pass_slp_vectorize): ...new function.
6517 (pass_complete_unroll): Replace declaration with that of...
6518 (make_pass_complete_unroll): ...new function.
6519 (pass_complete_unrolli): Replace declaration with that of...
6520 (make_pass_complete_unrolli): ...new function.
6521 (pass_parallelize_loops): Replace declaration with that of...
6522 (make_pass_parallelize_loops): ...new function.
6523 (pass_loop_prefetch): Replace declaration with that of...
6524 (make_pass_loop_prefetch): ...new function.
6525 (pass_iv_optimize): Replace declaration with that of...
6526 (make_pass_iv_optimize): ...new function.
6527 (pass_tree_loop_done): Replace declaration with that of...
6528 (make_pass_tree_loop_done): ...new function.
6529 (pass_ch): Replace declaration with that of...
6530 (make_pass_ch): ...new function.
6531 (pass_ccp): Replace declaration with that of...
6532 (make_pass_ccp): ...new function.
6533 (pass_phi_only_cprop): Replace declaration with that of...
6534 (make_pass_phi_only_cprop): ...new function.
6535 (pass_build_ssa): Replace declaration with that of...
6536 (make_pass_build_ssa): ...new function.
6537 (pass_build_alias): Replace declaration with that of...
6538 (make_pass_build_alias): ...new function.
6539 (pass_build_ealias): Replace declaration with that of...
6540 (make_pass_build_ealias): ...new function.
6541 (pass_dominator): Replace declaration with that of...
6542 (make_pass_dominator): ...new function.
6543 (pass_dce): Replace declaration with that of...
6544 (make_pass_dce): ...new function.
6545 (pass_dce_loop): Replace declaration with that of...
6546 (make_pass_dce_loop): ...new function.
6547 (pass_cd_dce): Replace declaration with that of...
6548 (make_pass_cd_dce): ...new function.
6549 (pass_call_cdce): Replace declaration with that of...
6550 (make_pass_call_cdce): ...new function.
6551 (pass_merge_phi): Replace declaration with that of...
6552 (make_pass_merge_phi): ...new function.
6553 (pass_split_crit_edges): Replace declaration with that of...
6554 (make_pass_split_crit_edges): ...new function.
6555 (pass_pre): Replace declaration with that of...
6556 (make_pass_pre): ...new function.
6557 (pass_profile): Replace declaration with that of...
6558 (make_pass_profile): ...new function.
6559 (pass_strip_predict_hints): Replace declaration with that of...
6560 (make_pass_strip_predict_hints): ...new function.
6561 (pass_lower_complex_O0): Replace declaration with that of...
6562 (make_pass_lower_complex_O0): ...new function.
6563 (pass_lower_complex): Replace declaration with that of...
6564 (make_pass_lower_complex): ...new function.
6565 (pass_lower_vector): Replace declaration with that of...
6566 (make_pass_lower_vector): ...new function.
6567 (pass_lower_vector_ssa): Replace declaration with that of...
6568 (make_pass_lower_vector_ssa): ...new function.
6569 (pass_lower_omp): Replace declaration with that of...
6570 (make_pass_lower_omp): ...new function.
6571 (pass_diagnose_omp_blocks): Replace declaration with that of...
6572 (make_pass_diagnose_omp_blocks): ...new function.
6573 (pass_expand_omp): Replace declaration with that of...
6574 (make_pass_expand_omp): ...new function.
6575 (pass_expand_omp_ssa): Replace declaration with that of...
6576 (make_pass_expand_omp_ssa): ...new function.
6577 (pass_object_sizes): Replace declaration with that of...
6578 (make_pass_object_sizes): ...new function.
6579 (pass_strlen): Replace declaration with that of...
6580 (make_pass_strlen): ...new function.
6581 (pass_fold_builtins): Replace declaration with that of...
6582 (make_pass_fold_builtins): ...new function.
6583 (pass_stdarg): Replace declaration with that of...
6584 (make_pass_stdarg): ...new function.
6585 (pass_early_warn_uninitialized): Replace declaration with that of...
6586 (make_pass_early_warn_uninitialized): ...new function.
6587 (pass_late_warn_uninitialized): Replace declaration with that of...
6588 (make_pass_late_warn_uninitialized): ...new function.
6589 (pass_cse_reciprocals): Replace declaration with that of...
6590 (make_pass_cse_reciprocals): ...new function.
6591 (pass_cse_sincos): Replace declaration with that of...
6592 (make_pass_cse_sincos): ...new function.
6593 (pass_optimize_bswap): Replace declaration with that of...
6594 (make_pass_optimize_bswap): ...new function.
6595 (pass_optimize_widening_mul): Replace declaration with that of...
6596 (make_pass_optimize_widening_mul): ...new function.
6597 (pass_warn_function_return): Replace declaration with that of...
6598 (make_pass_warn_function_return): ...new function.
6599 (pass_warn_function_noreturn): Replace declaration with that of...
6600 (make_pass_warn_function_noreturn): ...new function.
6601 (pass_cselim): Replace declaration with that of...
6602 (make_pass_cselim): ...new function.
6603 (pass_phiopt): Replace declaration with that of...
6604 (make_pass_phiopt): ...new function.
6605 (pass_forwprop): Replace declaration with that of...
6606 (make_pass_forwprop): ...new function.
6607 (pass_phiprop): Replace declaration with that of...
6608 (make_pass_phiprop): ...new function.
6609 (pass_tree_ifcombine): Replace declaration with that of...
6610 (make_pass_tree_ifcombine): ...new function.
6611 (pass_dse): Replace declaration with that of...
6612 (make_pass_dse): ...new function.
6613 (pass_nrv): Replace declaration with that of...
6614 (make_pass_nrv): ...new function.
6615 (pass_rename_ssa_copies): Replace declaration with that of...
6616 (make_pass_rename_ssa_copies): ...new function.
6617 (pass_sink_code): Replace declaration with that of...
6618 (make_pass_sink_code): ...new function.
6619 (pass_fre): Replace declaration with that of...
6620 (make_pass_fre): ...new function.
6621 (pass_check_data_deps): Replace declaration with that of...
6622 (make_pass_check_data_deps): ...new function.
6623 (pass_copy_prop): Replace declaration with that of...
6624 (make_pass_copy_prop): ...new function.
6625 (pass_vrp): Replace declaration with that of...
6626 (make_pass_vrp): ...new function.
6627 (pass_uncprop): Replace declaration with that of...
6628 (make_pass_uncprop): ...new function.
6629 (pass_return_slot): Replace declaration with that of...
6630 (make_pass_return_slot): ...new function.
6631 (pass_reassoc): Replace declaration with that of...
6632 (make_pass_reassoc): ...new function.
6633 (pass_rebuild_cgraph_edges): Replace declaration with that of...
6634 (make_pass_rebuild_cgraph_edges): ...new function.
6635 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
6636 (make_pass_remove_cgraph_callee_edges): ...new function.
6637 (pass_build_cgraph_edges): Replace declaration with that of...
6638 (make_pass_build_cgraph_edges): ...new function.
6639 (pass_local_pure_const): Replace declaration with that of...
6640 (make_pass_local_pure_const): ...new function.
6641 (pass_tracer): Replace declaration with that of...
6642 (make_pass_tracer): ...new function.
6643 (pass_warn_unused_result): Replace declaration with that of...
6644 (make_pass_warn_unused_result): ...new function.
6645 (pass_diagnose_tm_blocks): Replace declaration with that of...
6646 (make_pass_diagnose_tm_blocks): ...new function.
6647 (pass_lower_tm): Replace declaration with that of...
6648 (make_pass_lower_tm): ...new function.
6649 (pass_tm_init): Replace declaration with that of...
6650 (make_pass_tm_init): ...new function.
6651 (pass_tm_mark): Replace declaration with that of...
6652 (make_pass_tm_mark): ...new function.
6653 (pass_tm_memopt): Replace declaration with that of...
6654 (make_pass_tm_memopt): ...new function.
6655 (pass_tm_edges): Replace declaration with that of...
6656 (make_pass_tm_edges): ...new function.
6657 (pass_split_functions): Replace declaration with that of...
6658 (make_pass_split_functions): ...new function.
6659 (pass_feedback_split_functions): Replace declaration with that of...
6660 (make_pass_feedback_split_functions): ...new function.
6661 (pass_strength_reduction): Replace declaration with that of...
6662 (make_pass_strength_reduction): ...new function.
6663 (pass_ipa_lower_emutls): Replace declaration with that of...
6664 (make_pass_ipa_lower_emutls): ...new function.
6665 (pass_ipa_function_and_variable_visibility): Replace declaration with
6667 (make_pass_ipa_function_and_variable_visibility): ...new function.
6668 (pass_ipa_tree_profile): Replace declaration with that of...
6669 (make_pass_ipa_tree_profile): ...new function.
6670 (pass_early_local_passes): Replace declaration with that of...
6671 (make_pass_early_local_passes): ...new function.
6672 (pass_ipa_whole_program_visibility): Replace declaration with that
6674 (make_pass_ipa_whole_program_visibility): ...new function.
6675 (pass_ipa_lto_gimple_out): Replace declaration with that of...
6676 (make_pass_ipa_lto_gimple_out): ...new function.
6677 (pass_ipa_increase_alignment): Replace declaration with that of...
6678 (make_pass_ipa_increase_alignment): ...new function.
6679 (pass_ipa_inline): Replace declaration with that of...
6680 (make_pass_ipa_inline): ...new function.
6681 (pass_ipa_free_lang_data): Replace declaration with that of...
6682 (make_pass_ipa_free_lang_data): ...new function.
6683 (pass_ipa_free_inline_summary): Replace declaration with that of...
6684 (make_pass_ipa_free_inline_summary): ...new function.
6685 (pass_ipa_cp): Replace declaration with that of...
6686 (make_pass_ipa_cp): ...new function.
6687 (pass_ipa_reference): Replace declaration with that of...
6688 (make_pass_ipa_reference): ...new function.
6689 (pass_ipa_pure_const): Replace declaration with that of...
6690 (make_pass_ipa_pure_const): ...new function.
6691 (pass_ipa_pta): Replace declaration with that of...
6692 (make_pass_ipa_pta): ...new function.
6693 (pass_ipa_lto_finish_out): Replace declaration with that of...
6694 (make_pass_ipa_lto_finish_out): ...new function.
6695 (pass_ipa_tm): Replace declaration with that of...
6696 (make_pass_ipa_tm): ...new function.
6697 (pass_ipa_profile): Replace declaration with that of...
6698 (make_pass_ipa_profile): ...new function.
6699 (pass_ipa_cdtor_merge): Replace declaration with that of...
6700 (make_pass_ipa_cdtor_merge): ...new function.
6701 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
6703 (make_pass_cleanup_cfg_post_optimizing): ...new function.
6704 (pass_init_datastructures): Replace declaration with that of...
6705 (make_pass_init_datastructures): ...new function.
6706 (pass_fixup_cfg): Replace declaration with that of...
6707 (make_pass_fixup_cfg): ...new function.
6708 (pass_expand): Replace declaration with that of...
6709 (make_pass_expand): ...new function.
6710 (pass_instantiate_virtual_regs): Replace declaration with that of...
6711 (make_pass_instantiate_virtual_regs): ...new function.
6712 (pass_rtl_fwprop): Replace declaration with that of...
6713 (make_pass_rtl_fwprop): ...new function.
6714 (pass_rtl_fwprop_addr): Replace declaration with that of...
6715 (make_pass_rtl_fwprop_addr): ...new function.
6716 (pass_jump): Replace declaration with that of...
6717 (make_pass_jump): ...new function.
6718 (pass_jump2): Replace declaration with that of...
6719 (make_pass_jump2): ...new function.
6720 (pass_lower_subreg): Replace declaration with that of...
6721 (make_pass_lower_subreg): ...new function.
6722 (pass_cse): Replace declaration with that of...
6723 (make_pass_cse): ...new function.
6724 (pass_fast_rtl_dce): Replace declaration with that of...
6725 (make_pass_fast_rtl_dce): ...new function.
6726 (pass_ud_rtl_dce): Replace declaration with that of...
6727 (make_pass_ud_rtl_dce): ...new function.
6728 (pass_rtl_dce): Replace declaration with that of...
6729 (make_pass_rtl_dce): ...new function.
6730 (pass_rtl_dse1): Replace declaration with that of...
6731 (make_pass_rtl_dse1): ...new function.
6732 (pass_rtl_dse2): Replace declaration with that of...
6733 (make_pass_rtl_dse2): ...new function.
6734 (pass_rtl_dse3): Replace declaration with that of...
6735 (make_pass_rtl_dse3): ...new function.
6736 (pass_rtl_cprop): Replace declaration with that of...
6737 (make_pass_rtl_cprop): ...new function.
6738 (pass_rtl_pre): Replace declaration with that of...
6739 (make_pass_rtl_pre): ...new function.
6740 (pass_rtl_hoist): Replace declaration with that of...
6741 (make_pass_rtl_hoist): ...new function.
6742 (pass_rtl_store_motion): Replace declaration with that of...
6743 (make_pass_rtl_store_motion): ...new function.
6744 (pass_cse_after_global_opts): Replace declaration with that of...
6745 (make_pass_cse_after_global_opts): ...new function.
6746 (pass_rtl_ifcvt): Replace declaration with that of...
6747 (make_pass_rtl_ifcvt): ...new function.
6748 (pass_into_cfg_layout_mode): Replace declaration with that of...
6749 (make_pass_into_cfg_layout_mode): ...new function.
6750 (pass_outof_cfg_layout_mode): Replace declaration with that of...
6751 (make_pass_outof_cfg_layout_mode): ...new function.
6752 (pass_loop2): Replace declaration with that of...
6753 (make_pass_loop2): ...new function.
6754 (pass_rtl_loop_init): Replace declaration with that of...
6755 (make_pass_rtl_loop_init): ...new function.
6756 (pass_rtl_move_loop_invariants): Replace declaration with that of...
6757 (make_pass_rtl_move_loop_invariants): ...new function.
6758 (pass_rtl_unswitch): Replace declaration with that of...
6759 (make_pass_rtl_unswitch): ...new function.
6760 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
6761 (make_pass_rtl_unroll_and_peel_loops): ...new function.
6762 (pass_rtl_doloop): Replace declaration with that of...
6763 (make_pass_rtl_doloop): ...new function.
6764 (pass_rtl_loop_done): Replace declaration with that of...
6765 (make_pass_rtl_loop_done): ...new function.
6766 (pass_web): Replace declaration with that of...
6767 (make_pass_web): ...new function.
6768 (pass_cse2): Replace declaration with that of...
6769 (make_pass_cse2): ...new function.
6770 (pass_df_initialize_opt): Replace declaration with that of...
6771 (make_pass_df_initialize_opt): ...new function.
6772 (pass_df_initialize_no_opt): Replace declaration with that of...
6773 (make_pass_df_initialize_no_opt): ...new function.
6774 (pass_reginfo_init): Replace declaration with that of...
6775 (make_pass_reginfo_init): ...new function.
6776 (pass_inc_dec): Replace declaration with that of...
6777 (make_pass_inc_dec): ...new function.
6778 (pass_stack_ptr_mod): Replace declaration with that of...
6779 (make_pass_stack_ptr_mod): ...new function.
6780 (pass_initialize_regs): Replace declaration with that of...
6781 (make_pass_initialize_regs): ...new function.
6782 (pass_combine): Replace declaration with that of...
6783 (make_pass_combine): ...new function.
6784 (pass_if_after_combine): Replace declaration with that of...
6785 (make_pass_if_after_combine): ...new function.
6786 (pass_ree): Replace declaration with that of...
6787 (make_pass_ree): ...new function.
6788 (pass_partition_blocks): Replace declaration with that of...
6789 (make_pass_partition_blocks): ...new function.
6790 (pass_match_asm_constraints): Replace declaration with that of...
6791 (make_pass_match_asm_constraints): ...new function.
6792 (pass_regmove): Replace declaration with that of...
6793 (make_pass_regmove): ...new function.
6794 (pass_split_all_insns): Replace declaration with that of...
6795 (make_pass_split_all_insns): ...new function.
6796 (pass_fast_rtl_byte_dce): Replace declaration with that of...
6797 (make_pass_fast_rtl_byte_dce): ...new function.
6798 (pass_lower_subreg2): Replace declaration with that of...
6799 (make_pass_lower_subreg2): ...new function.
6800 (pass_mode_switching): Replace declaration with that of...
6801 (make_pass_mode_switching): ...new function.
6802 (pass_sms): Replace declaration with that of...
6803 (make_pass_sms): ...new function.
6804 (pass_sched): Replace declaration with that of...
6805 (make_pass_sched): ...new function.
6806 (pass_ira): Replace declaration with that of...
6807 (make_pass_ira): ...new function.
6808 (pass_reload): Replace declaration with that of...
6809 (make_pass_reload): ...new function.
6810 (pass_clean_state): Replace declaration with that of...
6811 (make_pass_clean_state): ...new function.
6812 (pass_branch_prob): Replace declaration with that of...
6813 (make_pass_branch_prob): ...new function.
6814 (pass_value_profile_transformations): Replace declaration with that
6816 (make_pass_value_profile_transformations): ...new function.
6817 (pass_postreload_cse): Replace declaration with that of...
6818 (make_pass_postreload_cse): ...new function.
6819 (pass_gcse2): Replace declaration with that of...
6820 (make_pass_gcse2): ...new function.
6821 (pass_split_after_reload): Replace declaration with that of...
6822 (make_pass_split_after_reload): ...new function.
6823 (pass_branch_target_load_optimize1): Replace declaration with that
6825 (make_pass_branch_target_load_optimize1): ...new function.
6826 (pass_thread_prologue_and_epilogue): Replace declaration with that
6828 (make_pass_thread_prologue_and_epilogue): ...new function.
6829 (pass_stack_adjustments): Replace declaration with that of...
6830 (make_pass_stack_adjustments): ...new function.
6831 (pass_peephole2): Replace declaration with that of...
6832 (make_pass_peephole2): ...new function.
6833 (pass_if_after_reload): Replace declaration with that of...
6834 (make_pass_if_after_reload): ...new function.
6835 (pass_regrename): Replace declaration with that of...
6836 (make_pass_regrename): ...new function.
6837 (pass_cprop_hardreg): Replace declaration with that of...
6838 (make_pass_cprop_hardreg): ...new function.
6839 (pass_reorder_blocks): Replace declaration with that of...
6840 (make_pass_reorder_blocks): ...new function.
6841 (pass_branch_target_load_optimize2): Replace declaration with that
6843 (make_pass_branch_target_load_optimize2): ...new function.
6844 (pass_leaf_regs): Replace declaration with that of...
6845 (make_pass_leaf_regs): ...new function.
6846 (pass_split_before_sched2): Replace declaration with that of...
6847 (make_pass_split_before_sched2): ...new function.
6848 (pass_compare_elim_after_reload): Replace declaration with that of...
6849 (make_pass_compare_elim_after_reload): ...new function.
6850 (pass_sched2): Replace declaration with that of...
6851 (make_pass_sched2): ...new function.
6852 (pass_stack_regs): Replace declaration with that of...
6853 (make_pass_stack_regs): ...new function.
6854 (pass_stack_regs_run): Replace declaration with that of...
6855 (make_pass_stack_regs_run): ...new function.
6856 (pass_df_finish): Replace declaration with that of...
6857 (make_pass_df_finish): ...new function.
6858 (pass_compute_alignments): Replace declaration with that of...
6859 (make_pass_compute_alignments): ...new function.
6860 (pass_duplicate_computed_gotos): Replace declaration with that of...
6861 (make_pass_duplicate_computed_gotos): ...new function.
6862 (pass_variable_tracking): Replace declaration with that of...
6863 (make_pass_variable_tracking): ...new function.
6864 (pass_free_cfg): Replace declaration with that of...
6865 (make_pass_free_cfg): ...new function.
6866 (pass_machine_reorg): Replace declaration with that of...
6867 (make_pass_machine_reorg): ...new function.
6868 (pass_cleanup_barriers): Replace declaration with that of...
6869 (make_pass_cleanup_barriers): ...new function.
6870 (pass_delay_slots): Replace declaration with that of...
6871 (make_pass_delay_slots): ...new function.
6872 (pass_split_for_shorten_branches): Replace declaration with that of...
6873 (make_pass_split_for_shorten_branches): ...new function.
6874 (pass_split_before_regstack): Replace declaration with that of...
6875 (make_pass_split_before_regstack): ...new function.
6876 (pass_convert_to_eh_region_ranges): Replace declaration with that
6878 (make_pass_convert_to_eh_region_ranges): ...new function.
6879 (pass_shorten_branches): Replace declaration with that of...
6880 (make_pass_shorten_branches): ...new function.
6881 (pass_set_nothrow_function_flags): Replace declaration with that of...
6882 (make_pass_set_nothrow_function_flags): ...new function.
6883 (pass_dwarf2_frame): Replace declaration with that of...
6884 (make_pass_dwarf2_frame): ...new function.
6885 (pass_final): Replace declaration with that of...
6886 (make_pass_final): ...new function.
6887 (pass_rtl_seqabstr): Replace declaration with that of...
6888 (make_pass_rtl_seqabstr): ...new function.
6889 (pass_release_ssa_names): Replace declaration with that of...
6890 (make_pass_release_ssa_names): ...new function.
6891 (pass_early_inline): Replace declaration with that of...
6892 (make_pass_early_inline): ...new function.
6893 (pass_inline_parameters): Replace declaration with that of...
6894 (make_pass_inline_parameters): ...new function.
6895 (pass_update_address_taken): Replace declaration with that of...
6896 (make_pass_update_address_taken): ...new function.
6897 (pass_convert_switch): Replace declaration with that of...
6898 (make_pass_convert_switch): ...new function.
6899 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
6900 to a subclass of simple_ipa_opt_pass along with...
6901 (pass_data_ipa_tree_profile): ...new pass_data instance and...
6902 (make_pass_ipa_tree_profile): ...new function.
6903 * tree-sra.c (pass_sra_early): Convert from a global struct to a
6904 subclass of gimple_opt_pass along with...
6905 (pass_data_sra_early): ...new pass_data instance and...
6906 (make_pass_sra_early): ...new function.
6907 (pass_sra): Convert from a global struct to a subclass of
6908 gimple_opt_pass along with...
6909 (pass_data_sra): ...new pass_data instance and...
6910 (make_pass_sra): ...new function.
6911 (pass_early_ipa_sra): Convert from a global struct to a subclass of
6912 gimple_opt_pass along with...
6913 (pass_data_early_ipa_sra): ...new pass_data instance and...
6914 (make_pass_early_ipa_sra): ...new function.
6915 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
6916 subclass of gimple_opt_pass along with...
6917 (pass_data_ccp): ...new pass_data instance and...
6918 (make_pass_ccp): ...new function.
6919 (pass_fold_builtins): Convert from a global struct to a subclass of
6920 gimple_opt_pass along with...
6921 (pass_data_fold_builtins): ...new pass_data instance and...
6922 (make_pass_fold_builtins): ...new function.
6923 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
6924 subclass of gimple_opt_pass along with...
6925 (pass_data_copy_prop): ...new pass_data instance and...
6926 (make_pass_copy_prop): ...new function.
6927 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
6928 global struct to a subclass of gimple_opt_pass along with...
6929 (pass_data_rename_ssa_copies): ...new pass_data instance and...
6930 (make_pass_rename_ssa_copies): ...new function.
6931 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
6932 subclass of gimple_opt_pass along with...
6933 (pass_data_dce): ...new pass_data instance and...
6934 (make_pass_dce): ...new function.
6935 (pass_dce_loop): Convert from a global struct to a subclass of
6936 gimple_opt_pass along with...
6937 (pass_data_dce_loop): ...new pass_data instance and...
6938 (make_pass_dce_loop): ...new function.
6939 (pass_cd_dce): Convert from a global struct to a subclass of
6940 gimple_opt_pass along with...
6941 (pass_data_cd_dce): ...new pass_data instance and...
6942 (make_pass_cd_dce): ...new function.
6943 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
6944 subclass of gimple_opt_pass along with...
6945 (pass_data_dominator): ...new pass_data instance and...
6946 (make_pass_dominator): ...new function.
6947 (pass_phi_only_cprop): Convert from a global struct to a subclass of
6948 gimple_opt_pass along with...
6949 (pass_data_phi_only_cprop): ...new pass_data instance and...
6950 (make_pass_phi_only_cprop): ...new function.
6951 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
6952 subclass of gimple_opt_pass along with...
6953 (pass_data_dse): ...new pass_data instance and...
6954 (make_pass_dse): ...new function.
6955 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
6956 a subclass of gimple_opt_pass along with...
6957 (pass_data_forwprop): ...new pass_data instance and...
6958 (make_pass_forwprop): ...new function.
6959 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
6960 struct to a subclass of gimple_opt_pass along with...
6961 (pass_data_tree_ifcombine): ...new pass_data instance and...
6962 (make_pass_tree_ifcombine): ...new function.
6963 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
6964 subclass of gimple_opt_pass along with...
6965 (pass_data_ch): ...new pass_data instance and...
6966 (make_pass_ch): ...new function.
6967 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
6968 subclass of gimple_opt_pass along with...
6969 (pass_data_tree_loop): ...new pass_data instance and...
6970 (make_pass_tree_loop): ...new function.
6971 (pass_tree_loop_init): Convert from a global struct to a subclass of
6972 gimple_opt_pass along with...
6973 (pass_data_tree_loop_init): ...new pass_data instance and...
6974 (make_pass_tree_loop_init): ...new function.
6975 (pass_lim): Convert from a global struct to a subclass of
6976 gimple_opt_pass along with...
6977 (pass_data_lim): ...new pass_data instance and...
6978 (make_pass_lim): ...new function.
6979 (pass_tree_unswitch): Convert from a global struct to a subclass of
6980 gimple_opt_pass along with...
6981 (pass_data_tree_unswitch): ...new pass_data instance and...
6982 (make_pass_tree_unswitch): ...new function.
6983 (pass_predcom): Convert from a global struct to a subclass of
6984 gimple_opt_pass along with...
6985 (pass_data_predcom): ...new pass_data instance and...
6986 (make_pass_predcom): ...new function.
6987 (pass_vectorize): Convert from a global struct to a subclass of
6988 gimple_opt_pass along with...
6989 (pass_data_vectorize): ...new pass_data instance and...
6990 (make_pass_vectorize): ...new function.
6991 (pass_graphite): Convert from a global struct to a subclass of
6992 gimple_opt_pass along with...
6993 (pass_data_graphite): ...new pass_data instance and...
6994 (make_pass_graphite): ...new function.
6995 (pass_graphite_transforms): Convert from a global struct to a subclass
6996 of gimple_opt_pass along with...
6997 (pass_data_graphite_transforms): ...new pass_data instance and...
6998 (make_pass_graphite_transforms): ...new function.
6999 (pass_check_data_deps): Convert from a global struct to a subclass of
7000 gimple_opt_pass along with...
7001 (pass_data_check_data_deps): ...new pass_data instance and...
7002 (make_pass_check_data_deps): ...new function.
7003 (pass_iv_canon): Convert from a global struct to a subclass of
7004 gimple_opt_pass along with...
7005 (pass_data_iv_canon): ...new pass_data instance and...
7006 (make_pass_iv_canon): ...new function.
7007 (pass_scev_cprop): Convert from a global struct to a subclass of
7008 gimple_opt_pass along with...
7009 (pass_data_scev_cprop): ...new pass_data instance and...
7010 (make_pass_scev_cprop): ...new function.
7011 (pass_record_bounds): Convert from a global struct to a subclass of
7012 gimple_opt_pass along with...
7013 (pass_data_record_bounds): ...new pass_data instance and...
7014 (make_pass_record_bounds): ...new function.
7015 (pass_complete_unroll): Convert from a global struct to a subclass of
7016 gimple_opt_pass along with...
7017 (pass_data_complete_unroll): ...new pass_data instance and...
7018 (make_pass_complete_unroll): ...new function.
7019 (pass_complete_unrolli): Convert from a global struct to a subclass of
7020 gimple_opt_pass along with...
7021 (pass_data_complete_unrolli): ...new pass_data instance and...
7022 (make_pass_complete_unrolli): ...new function.
7023 (pass_parallelize_loops): Convert from a global struct to a subclass
7024 of gimple_opt_pass along with...
7025 (pass_data_parallelize_loops): ...new pass_data instance and...
7026 (make_pass_parallelize_loops): ...new function.
7027 (pass_loop_prefetch): Convert from a global struct to a subclass of
7028 gimple_opt_pass along with...
7029 (pass_data_loop_prefetch): ...new pass_data instance and...
7030 (make_pass_loop_prefetch): ...new function.
7031 (pass_iv_optimize): Convert from a global struct to a subclass of
7032 gimple_opt_pass along with...
7033 (pass_data_iv_optimize): ...new pass_data instance and...
7034 (make_pass_iv_optimize): ...new function.
7035 (pass_tree_loop_done): Convert from a global struct to a subclass of
7036 gimple_opt_pass along with...
7037 (pass_data_tree_loop_done): ...new pass_data instance and...
7038 (make_pass_tree_loop_done): ...new function.
7039 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
7040 struct to a subclass of gimple_opt_pass along with...
7041 (pass_data_cse_reciprocals): ...new pass_data instance and...
7042 (make_pass_cse_reciprocals): ...new function.
7043 (pass_cse_sincos): Convert from a global struct to a subclass of
7044 gimple_opt_pass along with...
7045 (pass_data_cse_sincos): ...new pass_data instance and...
7046 (make_pass_cse_sincos): ...new function.
7047 (pass_optimize_bswap): Convert from a global struct to a subclass of
7048 gimple_opt_pass along with...
7049 (pass_data_optimize_bswap): ...new pass_data instance and...
7050 (make_pass_optimize_bswap): ...new function.
7051 (pass_optimize_widening_mul): Convert from a global struct to a
7052 subclass of gimple_opt_pass along with...
7053 (pass_data_optimize_widening_mul): ...new pass_data instance and...
7054 (make_pass_optimize_widening_mul): ...new function.
7055 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
7056 subclass of gimple_opt_pass along with...
7057 (pass_data_phiopt): ...new pass_data instance and...
7058 (make_pass_phiopt): ...new function.
7059 (pass_cselim): Convert from a global struct to a subclass of
7060 gimple_opt_pass along with...
7061 (pass_data_cselim): ...new pass_data instance and...
7062 (make_pass_cselim): ...new function.
7063 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
7064 subclass of gimple_opt_pass along with...
7065 (pass_data_phiprop): ...new pass_data instance and...
7066 (make_pass_phiprop): ...new function.
7067 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
7068 subclass of gimple_opt_pass along with...
7069 (pass_data_pre): ...new pass_data instance and...
7070 (make_pass_pre): ...new function.
7071 (pass_fre): Convert from a global struct to a subclass of
7072 gimple_opt_pass along with...
7073 (pass_data_fre): ...new pass_data instance and...
7074 (make_pass_fre): ...new function.
7075 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
7076 subclass of gimple_opt_pass along with...
7077 (pass_data_reassoc): ...new pass_data instance and...
7078 (make_pass_reassoc): ...new function.
7079 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
7080 subclass of gimple_opt_pass along with...
7081 (pass_data_sink_code): ...new pass_data instance and...
7082 (make_pass_sink_code): ...new function.
7083 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
7084 subclass of gimple_opt_pass along with...
7085 (pass_data_strlen): ...new pass_data instance and...
7086 (make_pass_strlen): ...new function.
7087 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
7088 struct to a subclass of gimple_opt_pass along with...
7089 (pass_data_build_alias): ...new pass_data instance and...
7090 (make_pass_build_alias): ...new function.
7091 (pass_build_ealias): Convert from a global struct to a subclass of
7092 gimple_opt_pass along with...
7093 (pass_data_build_ealias): ...new pass_data instance and...
7094 (make_pass_build_ealias): ...new function.
7095 (pass_ipa_pta): Convert from a global struct to a subclass of
7096 simple_ipa_opt_pass along with...
7097 (pass_data_ipa_pta): ...new pass_data instance and...
7098 (make_pass_ipa_pta): ...new function.
7099 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
7100 subclass of gimple_opt_pass along with...
7101 (pass_data_uncprop): ...new pass_data instance and...
7102 (make_pass_uncprop): ...new function.
7103 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
7104 global struct to a subclass of gimple_opt_pass along with...
7105 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
7106 (make_pass_late_warn_uninitialized): ...new function.
7107 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
7108 to a subclass of gimple_opt_pass along with...
7109 (pass_data_init_datastructures): ...new pass_data instance and...
7110 (make_pass_init_datastructures): ...new function.
7111 (pass_early_warn_uninitialized): Convert from a global struct to a
7112 subclass of gimple_opt_pass along with...
7113 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
7114 (make_pass_early_warn_uninitialized): ...new function.
7115 (pass_update_address_taken): Convert from a global struct to a
7116 subclass of gimple_opt_pass along with...
7117 (pass_data_update_address_taken): ...new pass_data instance and...
7118 (make_pass_update_address_taken): ...new function.
7119 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
7120 struct to a subclass of gimple_opt_pass along with...
7121 (pass_data_release_ssa_names): ...new pass_data instance and...
7122 (make_pass_release_ssa_names): ...new function.
7123 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
7124 subclass of gimple_opt_pass along with...
7125 (pass_data_stdarg): ...new pass_data instance and...
7126 (make_pass_stdarg): ...new function.
7127 * tree-switch-conversion.c (pass_convert_switch): Convert from a
7128 global struct to a subclass of gimple_opt_pass along with...
7129 (pass_data_convert_switch): ...new pass_data instance and...
7130 (make_pass_convert_switch): ...new function.
7131 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
7132 to a subclass of gimple_opt_pass along with...
7133 (pass_data_tail_recursion): ...new pass_data instance and...
7134 (make_pass_tail_recursion): ...new function.
7135 (pass_tail_calls): Convert from a global struct to a subclass of
7136 gimple_opt_pass along with...
7137 (pass_data_tail_calls): ...new pass_data instance and...
7138 (make_pass_tail_calls): ...new function.
7139 * tree-vect-generic.c (pass_lower_vector): Convert from a global
7140 struct to a subclass of gimple_opt_pass along with...
7141 (pass_data_lower_vector): ...new pass_data instance and...
7142 (make_pass_lower_vector): ...new function.
7143 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
7144 gimple_opt_pass along with...
7145 (pass_data_lower_vector_ssa): ...new pass_data instance and...
7146 (make_pass_lower_vector_ssa): ...new function.
7147 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
7148 to a subclass of gimple_opt_pass along with...
7149 (pass_data_slp_vectorize): ...new pass_data instance and...
7150 (make_pass_slp_vectorize): ...new function.
7151 (pass_ipa_increase_alignment): Convert from a global struct to a
7152 subclass of simple_ipa_opt_pass along with...
7153 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
7154 (make_pass_ipa_increase_alignment): ...new function.
7155 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
7156 gimple_opt_pass along with...
7157 (pass_data_vrp): ...new pass_data instance and...
7158 (make_pass_vrp): ...new function.
7159 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
7160 subclass of simple_ipa_opt_pass along with...
7161 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
7162 (make_pass_ipa_free_lang_data): ...new function.
7163 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
7164 gimple_opt_pass along with...
7165 (pass_data_tsan): ...new pass_data instance and...
7166 (make_pass_tsan): ...new function.
7167 (pass_tsan_O0): Convert from a global struct to a subclass of
7168 gimple_opt_pass along with...
7169 (pass_data_tsan_O0): ...new pass_data instance and...
7170 (make_pass_tsan_O0): ...new function.
7171 * var-tracking.c (pass_variable_tracking): Convert from a global
7172 struct to a subclass of rtl_opt_pass along with...
7173 (pass_data_variable_tracking): ...new pass_data instance and...
7174 (make_pass_variable_tracking): ...new function.
7175 * web.c (pass_web): Convert from a global struct to a subclass of
7176 rtl_opt_pass along with...
7177 (pass_data_web): ...new pass_data instance and...
7178 (make_pass_web): ...new function.
7179 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
7180 declaration with that of...
7181 (make_pass_mode_switch_use): ...new function.
7182 (pass_resolve_sw_modes): Replace declaration with that of...
7183 (make_pass_resolve_sw_modes): ...new function.
7184 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
7185 from a global struct to a subclass of rtl_opt_pass along with...
7186 (pass_data_mode_switch_use): ...new pass_data instance and...
7187 (make_pass_mode_switch_use): ...new function.
7188 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
7189 from a global struct to a subclass of rtl_opt_pass along with...
7190 (pass_data_resolve_sw_modes): ...new pass_data instance and...
7191 (make_pass_resolve_sw_modes): ...new function.
7192 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
7193 struct to a subclass of rtl_opt_pass along with...
7194 (pass_data_insert_vzeroupper): ...new pass_data instance and...
7195 (make_pass_insert_vzeroupper): ...new function.
7196 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
7197 global struct to a subclass of rtl_opt_pass along with...
7198 (pass_data_work_around_errata): ...new pass_data instance and...
7199 (make_pass_work_around_errata): ...new function.
7200 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
7201 struct to a subclass of rtl_opt_pass along with...
7202 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
7203 (make_pass_mips_machine_reorg2): ...new function.
7205 2013-08-05 David Malcolm <dmalcolm@redhat.com>
7207 * passes.c (pass_manager::operator new): New.
7209 2013-08-05 David Malcolm <dmalcolm@redhat.com>
7211 Handwritten part of conversion of passes to C++ classes.
7213 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
7214 (toplev.o): Add dep on PASS_MANAGER_H.
7215 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
7216 of early local pases to reflect this moving from a global to a
7217 member of gcc::pass_manager.
7218 (cgraph_add_new_function): Likewise.
7219 * lto-cgraph.c (lto_output_node): Update for conversion of
7220 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
7221 * passes.c (opt_pass::clone): New.
7222 (opt_pass::gate): New.
7223 (opt_pass::execute): New.
7224 (opt_pass::opt_pass): New.
7225 (pass_manager::execute_early_local_passes): New.
7226 (pass_manager::execute_pass_mode_switching): new.
7227 (finish_optimization_passes): Convert to...
7228 (pass_manager::finish_optimization_passes): ...this.
7229 (finish_optimization_passes): Update for conversion of passes to
7231 (register_dump_files_1): Use has_gate since we cannot portably
7232 check a vtable entry against NULL.
7233 (dump_one_pass): Likewise.
7234 (ipa_write_summaries_2): Likewise.
7235 (ipa_write_optimization_summaries_1): Likewise.
7236 (ipa_read_summaries_1): Likewise.
7237 (ipa_read_optimization_summaries_1): Likewise.
7238 (execute_ipa_stmt_fixups): Likewise.
7239 (pass_manager::pass_manager): Rewrite pass-creation, invoking
7240 pass-creation functions rather than wiring up globals, and
7241 storing the results in fields of pass_manager generated using
7243 (pass_manager::dump_profile_report): Update for conversion of
7244 passes to C++ classes.
7245 (pass_manager::execute_ipa_summary_passes): Likewise.
7246 (execute_one_ipa_transform_pass): Likewise.
7247 (execute_one_pass): Use has_gate and has_execute since we cannot
7248 portably check a vtable entry against NULL.
7249 * pass_manager.h (pass_manager::finish_optimization_passes): New.
7250 (pass_manager): Use pass-instances.def to add fields for the
7251 various pass instances.
7252 * toplev.c (finalize): Update for move of
7253 finish_optimization_passes to a method of gcc::pass_manager.
7254 * toplev.h (finish_optimization_passes): Move to method of class
7256 * tree-pass.h (struct pass_data): New.
7257 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
7258 (opt_pass::gate): Convert to virtual function.
7259 (opt_pass::~opt_pass): New.
7260 (opt_pass::clone): New.
7261 (opt_pass::execute): Convert to virtual function.
7262 (opt_pass::opt_pass): New.
7263 (opt_pass::ctxt_): new.
7264 (gimple_opt_pass): Convert to subclass of opt_pass.
7265 (gimple_opt_pass::gimple_opt_pass): New.
7266 (rtl_opt_pass): Convert to subclass of opt_pass.
7267 (rtl_opt_pass::rtl_opt_pass): New.
7268 (ipa_opt_pass_d): Convert to subclass of opt_pass.
7269 (ipa_opt_pass_d::ipa_opt_pass_d): New.
7270 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
7271 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
7272 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
7273 invocation of pass_mode_switching to reflect this moving from a
7274 global to a member of gcc::pass_manager.
7275 (ix86_option_override): Rework how pass_insert_vzeroupper is
7276 added to the pass_manager to reflect autogenerated changes.
7277 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
7279 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
7281 PR rtl-optimization/57708
7282 * recog.c (peep2_find_free_register): Validate all regs in a
7285 2013-08-05 Jan Hubicka <jh@suse.cz>
7288 * cgraph.c (verify_cgraph_node): Accept local flags from other
7290 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
7291 (function_and_variable_visibility): Likewise.
7292 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
7294 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
7296 * graph.c (init_graph_slim_pretty_print): Remove.
7297 (print_graph_cfg): Do not call it. Use local pretty printer.
7298 (start_graph_dump): Likewise.
7300 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
7302 * gimple-pretty-print.c (buffer): Remove.
7303 (initialized): Likewise.
7304 (maybe_init_pretty_print): Likewise.
7305 (print_gimple_stmt): Do not call it. Use non-static local
7306 pretty_printer variable.
7307 (print_gimple_expr): Likewise.
7308 (print_gimple_seq): Likewise.
7309 (gimple_dump_bb): Likewise.
7311 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
7313 * asan.c (asan_pp): Remove.
7314 (asan_pp_initialized): Likewise.
7315 (asan_pp_initialize): Likewise.
7316 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
7317 (asan_emit_stack_protection): Tidy. Use local pretty printer.
7318 (asan_add_global): Likewise.
7320 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
7322 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
7323 * diagnostic.h (diagnostic_flush_buffer): Adjust.
7324 * pretty-print.c (pp_formatted_text_data): Likewise.
7325 (pp_indent): Rename from pp_base_indent.
7326 (pp_format): Rename from pp_base_format.
7327 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
7328 (pp_format_verbatim): Rename from pp_base_format_verbatim.
7329 (pp_flush): Rename from pp_base_flush.
7330 (pp_set_line_maximum_length): Rename from
7331 pp_base_set_line_maximum_length.
7332 (pp_clear_output_area): Rename from pp_base_clear_output_area.
7333 (pp_set_prefix): Rename from pp_base_set_prefix.
7334 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
7335 (pp_emit_prefix): Rename from pp_base_emit_prefix.
7336 (pp_append_text): Rename from pp_base_append_text.
7337 (pp_formatted_text): Rename from pp_base_formatted_text.
7338 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
7339 (pp_remaining_character_count_for_line): Rename from
7340 pp_base_remaining_character_count_for_line.
7341 (pp_newline): Rename from pp_base_newline.
7342 (pp_character): Rename from pp_base_character.
7343 (pp_string): Rename from pp_base_string.
7344 (pp_maybe_space): Rename from pp_base_maybe_space.
7345 * asan.c (asan_pp_string): Adjust.
7346 (asan_emit_stack_protection): Likewise.
7347 (asan_add_global): Likewise.
7348 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
7349 * tree-mudflap.c (mf_varname_tree): Likewise.
7350 * tree-pretty-print.c (pp_tree_identifier): Rename from
7351 pp_base_tree_identifier.
7352 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
7353 Declare as function.
7355 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
7357 * pretty-print.h (pp_bar_bar): New.
7358 (pp_ampersand_ampersand): Likewise.
7359 (pp_less_equal): Likewise.
7360 (pp_greater_equal): Likewise.
7361 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
7362 printer functions instead of pp_string or operators and punctuators.
7363 (dump_gimple_call): Likewise.
7364 (dump_gimple_omp_for): Likewise.
7365 (dump_gimple_transaction): Likewise.
7366 (dump_gimple_phi): Likewise.
7367 (pp_gimple_stmt_1): Likewise.
7368 * sched-vis.c (print_insn): Likewise.
7369 * tree-mudflap.c (mf_varname_tree): Likewise.
7370 * tree-pretty-print.c (dump_block_node): Likewise.
7371 (dump_generic_node): Likewise.
7373 2013-08-02 Jan Hubicka <jh@suse.cz>
7375 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
7377 * lto-streamer-out.c (tree_is_indexable): Results decls and
7378 parm decls are not indexable.
7379 (DFS_write_tree_body): Do not follow args and results.
7380 (hash_tree): Likewise.
7381 (output_functions): Rearrange so struct function is needed
7382 only when real body is output; be able to also ouptut abstract
7383 functions; output DECL_ARGUMENTS and DECL_RESULT.
7384 (lto_output): When not in WPA, ale store abstract functions.
7385 (write_symbol): Do not care about RESULT_DECL.
7386 (output_symbol_p): Handle correctly sbtract decls.
7387 * lto-streamer-in.c (input_function): Rearrange so struct
7388 function can be NULL at entry; allow streaming of
7389 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
7390 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
7391 sanity check during LTO.
7392 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
7393 RESULT_DECl and DECL_ARGUMENTS.
7394 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
7397 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
7399 * pretty-print.h (pp_underscore): New.
7401 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
7402 printer functions instead of pp_character.
7403 (dump_binary_rhs): Likewise.
7404 (dump_ternary_rhs): Likewise.
7405 (dump_gimple_call_args): Likewise.
7406 (pp_points_to_solution): Likewise.
7407 (dump_gimple_call): Likewise.
7408 (dump_gimple_switch): Likewise.
7409 (dump_gimple_cond): Likewise.
7410 (dump_gimple_bind): Likewise.
7411 (dump_gimple_try): Likewise.
7412 (dump_gimple_omp_for): Likewise.
7413 (dump_gimple_omp_continue): Likewise.
7414 (dump_gimple_omp_single): Likewise.
7415 (dump_gimple_omp_sections): Likewise.
7416 (dump_gimple_omp_block): Likewise.
7417 (dump_gimple_omp_critical): Likewise.
7418 (dump_gimple_transaction): Likewise.
7419 (dump_gimple_asm): Likewise.
7420 (dump_gimple_phi): Likewise.
7421 (dump_gimple_omp_parallel): Likewise.
7422 (dump_gimple_omp_task): Likewise.
7423 (dump_gimple_omp_atomic_load): Likewise.
7424 (dump_gimple_omp_atomic_store): Likewise.
7425 (dump_gimple_mem_ops): Likewise.
7426 (pp_gimple_stmt_1): Likewise.
7427 (pp_cfg_jump): Likewise.
7428 (dump_implicit_edges): Likewise.
7429 (gimple_dump_bb_for_graph): Likewise.
7430 * graph.c (draw_cfg_node): Likewise.
7431 * langhooks.c (lhd_print_error_function): Likewise.
7432 * sched-vis.c (print_exp): Likewise.
7433 (print_value): Likewise.
7434 (print_pattern): Likewise.
7435 (print_insn): Likewise.
7436 (rtl_dump_bb_for_graph): Likewise.
7437 * tree-pretty-print.c (dump_function_declaration): Likewise.
7438 (dump_array_domain): Likewise.
7439 (dump_omp_clause): Likewise.
7440 (dump_location): Likewise.
7441 (dump_generic_node): Likewise.
7442 (print_struct_decl): Likewise.
7443 * diagnostic.c (diagnostic_show_locus): Use pp_space.
7445 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7447 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
7448 candidate table when replacing a candidate statement.
7449 (replace_rhs_if_not_dup): Likewise.
7450 (replace_one_candidate): Likewise.
7452 2013-08-02 Jan Hubicka <jh@suse.cz>
7453 Martin Liska <marxin.liska@gmail.com>
7455 * cgraphunit.c (add_new_function): Fix logic when adding from
7457 (assemble_thunk): Rename to ...
7458 (expand_thunk); .. this one; export; get it working with
7459 general functions; make produced gimple valid.
7460 * cgraph.h (expand_thunk): Declare.
7462 2013-08-02 Jan Hubicka <jh@suse.cz>
7464 * ipa-cp.c (gather_context_independent_values): Use
7465 ipa_get_param_move_cost.
7466 (get_replacement_map): Remove PARAM; move parameter folding
7468 (create_specialized_node): Update.
7469 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
7470 assert that we have gimple body; update move_cost.
7471 (count_formal_params): Assert that we have gimple body.
7472 (ipa_dump_param): New function.
7473 (ipa_alloc_node_params): Break out from ...
7474 (ipa_initialize_node_params): ... here.
7475 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
7476 (ipa_write_node_info): Stream move costs.
7477 (ipa_read_node_info): Read move costs.
7478 (ipa_update_after_lto_read): Do not recompute node params.
7479 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
7480 (ipa_get_param): Check we are not in WPA.
7481 (ipa_get_param_move_cost): New.
7482 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
7483 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
7484 parm numbers to be present.
7486 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
7488 PR rtl-optimization/58048
7489 * lra-constraints.c (process_alt_operands): Don't check asm
7490 operand on register.
7492 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
7494 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
7495 the implied StoreLoad barrier for atomic operations if before.
7497 2013-08-02 Jan Hubicka <jh@suse.cz>
7498 Martin Liska <marxin.liska@gmail.com>
7500 * cgraph.c (cgraph_function_body_availability): Do not check
7502 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
7503 symtab_node_availability): Declare.
7504 * ipa.c (can_replace_by_local_alias): New.
7505 (function_and_variable_visibility): Use it.
7506 * symtab.c (symtab_for_node_and_aliases,
7507 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
7509 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
7511 PR rtl-optimization/57963
7512 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
7513 (lra_constraints): Use them.
7515 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
7517 * config/arm/types.md (define_attr "type"): Add "load_acq"
7519 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
7521 (cortex_a53_store1): Likewise.
7523 2013-08-01 Jan Hubicka <jh@suse.cz>
7525 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
7526 partitions are not needed.
7528 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
7530 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
7531 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
7532 MAYBE_NON_Q_CLASS_P where appropriate.
7534 2013-08-01 Jan Hubicka <jh@suse.cz>
7536 * cgraph.h (release_function_body): Declare.
7537 * tree.c (free_lang_data_in_decl): Free, parameters and return values
7538 of unused delcarations.
7540 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7542 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
7543 RTL form when subtracting a constant.
7545 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7547 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
7548 Generate canonical plus rtx with negated immediate instead of minus
7550 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
7552 2013-08-01 Jan Hubicka <jh@suse.cz>
7554 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
7555 (cgraph_release_function_body): Likewise.
7556 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
7557 * cgraph.h (cgrpah_node): Rename abstract_and_needed
7558 to used_as_abstract_origin.
7559 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
7560 symbols used as abstract origins.
7561 * cgraphunit.c (analyze_functions): Update.
7562 * ipa.c (symtab_remove_unreachable_nodes): Recompute
7563 used_as_abstract_origin.
7564 * tree-inline.c (tree_function_versioning): Update
7565 used_as_abstract_origin; be ready for DECL_RESULT and
7566 DECL_ARGUMENTS to be NULL.
7568 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
7569 for abstract functions.
7570 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
7573 2013-08-01 Jan Hubicka <jh@suse.cz>
7575 * profile.c (compute_value_histograms): Fix thinko.
7577 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
7579 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
7580 aarch-common-protos.h to extra_headers.
7581 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
7582 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
7583 * config/aarch64/t-aarch64 (aarch-common.o): Define.
7585 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
7587 * config/aarch64/aarch64.md (define_attr "type"): Delete.
7588 Include "../arm/types.md". Define "type" attribute for all patterns.
7589 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
7592 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
7594 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
7595 to support power8 load fusion.
7596 (fusion_gpr_mem_load): Likewise.
7598 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
7600 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
7601 declarations for power8 load fusion.
7602 (emit_fusion_gpr_load): Likewise.
7604 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
7605 tuning for power8, turn on fusion mode by default. Turn on sign
7606 extending fusion mode if normal fusion mode is on, and we are at
7608 (fusion_gpr_load_p): New function, return true if we can fuse an
7609 addis instruction with a dependent load to a GPR.
7610 (emit_fusion_gpr_load): Emit the instructions for power8 load
7613 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
7614 (VSX load fusion peepholes): New peepholes to fuse together an
7615 addi instruction with a VSX load instruction.
7617 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
7618 peepholes to fuse an addis instruction with a load to a GPR base
7619 register. If we are supporting sign extending fusions, convert
7620 sign extending loads to zero extending loads and add an explicit
7623 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
7625 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
7626 aarch-common-protos.h to extra_headers.
7627 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
7628 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
7629 (arm_early_store_addr_dep): Likewise.
7630 (arm_no_early_alu_shift_dep): Likewise.
7631 (arm_no_early_alu_shift_value_dep): Likewise.
7632 (arm_no_early_mul_dep): Likewise.
7633 (arm_no_early_store_addr_dep): Likewise.
7634 (arm_mac_accumulator_is_mul_result): Likewise.
7635 (arm_mac_accumulator_is_result): Likewise.
7636 * config/arm/aarch-common.c: ... here. New file.
7637 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
7639 (arm_early_store_addr_dep): Likewise.
7640 (arm_no_early_alu_shift_dep): Likewise.
7641 (arm_no_early_alu_shift_value_dep): Likewise.
7642 (arm_no_early_mul_dep): Likewise.
7643 (arm_no_early_store_addr_dep): Likewise.
7644 (arm_mac_accumulator_is_mul_result): Likewise.
7645 (arm_mac_accumulator_is_result): Likewise.
7646 * config/arm/aarch-common-protos.h: ... here. New file.
7647 * config/arm/t-arm (aarch-common.o): Define.
7649 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
7651 * config/arm/arm.md: Include new file "types.md".
7652 (define_attr "type"): Move from here to ...
7653 (define_attr "mul32"): Likewise.
7654 (define_attr "mul64"): Likewise.
7655 * config/arm/types.md: ... here. New file.
7657 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
7659 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
7660 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
7662 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7664 * gen-pass-instances.awk: Fix offset of substr().
7666 2013-07-31 David Malcolm <dmalcolm@redhat.com>
7668 * Makefile.in (pass-instances.def): New.
7669 (passes.o): Replace dependency on passes.def with one on
7672 * gen-pass-instances.awk: New.
7674 * passes.c (pass_manager::pass_manager): Use pass-instances.def
7675 rather than passes.def, updating local definition of NEXT_PASS
7676 macro to add an extra NUM parameter (currently unused).
7678 2013-07-30 David Malcolm <dmalcolm@redhat.com>
7680 * Makefile.in (PASS_MANAGER_H): New.
7681 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
7682 (passes.o): Likewise.
7683 (statistics.o): Likewise.
7684 (cgraphunit.o): Likewise.
7685 (context.o): Depend on PASS_MANAGER_H.
7687 * pass_manager.h: New.
7689 * cgraphunit.c (cgraph_add_new_function): Update for moves
7690 of globals to fields of pass_manager.
7691 (analyze_function): Likewise.
7692 (expand_function): Likewise.
7693 (ipa_passes): Likewise.
7694 (compile): Likewise.
7696 * context.c (context::context): New.
7697 * context.h (context::context): New.
7698 (context::get_passes): New.
7699 (context::passes_): New.
7701 * lto-cgraph.c (input_node): Update for moves of globals to
7702 fields of pass_manager.
7704 * passes.c (all_passes): Remove, in favor of a field of the
7705 same name within the new class pass_manager.
7706 (all_small_ipa_passes): Likewise.
7707 (all_lowering_passes): Likewise.
7708 (all_regular_ipa_passes): Likewise.
7709 (all_late_ipa_passes): Likewise.
7710 (all_lto_gen_passes): Likewise.
7711 (passes_by_id): Likewise.
7712 (passes_by_id_size): Likewise.
7713 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
7714 the new class pass_manager.
7715 (set_pass_for_id): Convert to...
7716 (pass_manager::set_pass_for_id): ...method.
7717 (get_pass_for_id): Convert to...
7718 (pass_manager::get_pass_for_id): ...method.
7719 (register_one_dump_file): Move body of implementation into...
7720 (pass_manager::register_one_dump_file): ...here.
7721 (register_dump_files_1): Convert to...
7722 (pass_manager::register_dump_files_1): ...method.
7723 (register_dump_files): Convert to...
7724 (pass_manager::register_dump_files): ...method.
7725 (create_pass_tab): Update for moves of globals to fields of
7727 (dump_passes): Move body of implementation into...
7728 (pass_manager::dump_passes): ...here.
7729 (register_pass): Move body of implementation into...
7730 (pass_manager::register_pass): ...here.
7731 (init_optimization_passes): Convert into...
7732 (pass_manager::pass_manager): ...constructor for new
7733 pass_manager class, and initialize the pass_lists array.
7734 (check_profile_consistency): Update for moves of globals to
7735 fields of pass_manager.
7736 (dump_profile_report): Move body of implementation into...
7737 (pass_manager::dump_profile_report): ...here.
7738 (ipa_write_summaries_1): Update for moves of pass lists from
7739 being globals to fields of pass_manager.
7740 (ipa_write_optimization_summaries): Likewise.
7741 (ipa_read_summaries): Likewise.
7742 (ipa_read_optimization_summaries): Likewise.
7743 (execute_all_ipa_stmt_fixups): Likewise.
7745 * statistics.c (statistics_fini): Update for moves of globals to
7746 fields of pass_manager.
7748 * toplev.c (general_init): Replace call to
7749 init_optimization_passes with construction of the pass_manager
7752 * tree-pass.h (all_passes): Remove, in favor of a field of the
7753 same name within the new class pass_manager.
7754 (all_small_ipa_passes): Likewise.
7755 (all_lowering_passes): Likewise.
7756 (all_regular_ipa_passes): Likewise.
7757 (all_lto_gen_passes): Likewise.
7758 (all_late_ipa_passes): Likewise.
7759 (passes_by_id): Likewise.
7760 (passes_by_id_size): Likewise.
7761 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
7762 the new class pass_manager.
7763 (get_pass_for_id): Remove.
7765 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
7767 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
7770 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
7772 * arm.md (mulhi3): New expand pattern.
7774 2013-07-30 Jan Hubicka <jh@suse.cz>
7775 Martin Liska <marxin.liska@gmail.com>
7777 * profile.c (compute_value_histograms): Do not ICE when
7778 there is mismatch only on some counters.
7780 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
7782 PR rtl-optimization/57637
7783 * function.c (move_insn_for_shrink_wrap): Also check the
7784 GEN set of the LIVE problem for the liveness analysis
7785 if it exists, otherwise give up.
7787 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7789 PR tree-optimization/57993
7790 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
7791 replaced statement in the candidate table.
7792 (phi_add_costs): Return infinite cost when the hidden basis does
7793 not dominate all phis on which the candidate is dependent.
7794 (replace_one_candidate): Record replaced statement in the
7797 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
7799 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
7800 (ashlv2si3): New expander.
7801 (*ashlv2si3_i): New define_insn_and_split.
7802 * predicates.md (float_operation): Allow patterns with three
7805 PR rtl-optimization/58021
7806 * mode-switching.c (create_pre_exit): Always split off preceding
7807 insns if we are not at the basic block head.
7809 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
7811 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
7812 (UCLIBC_DYNAMIC_LINKER): New macro.
7813 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
7815 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
7816 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
7818 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
7819 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
7820 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
7821 for SF and DF modes. Use ieee_quad_format for TF mode.
7822 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
7823 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
7824 (mips_option_override): Handle `-mnan=legacy'.
7825 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
7826 `-mabs=2008' and `-mnan=2008'.
7827 (OPTION_DEFAULT_SPECS): Add "nan" default.
7828 (ASM_SPEC): Handle `-mnan='.
7829 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
7830 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
7831 comment accordingly.
7832 (neg<mode>2): Likewise.
7833 * config/mips/mips.opt (mabs, mnan): New options.
7834 * doc/install.texi (Configuration): Document `--with-nan=' option.
7835 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
7837 (MIPS Options): Document them.
7838 * config.gcc <mips*-*-*>: Handle `--with-nan='.
7839 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
7840 * configure: Regenerate.
7841 * config.in: Regenerate.
7843 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
7845 * config/i386/i386.md (float post-reload splitters): Do not check
7846 for subregs of SSE registers.
7848 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
7849 H.J. Lu <hongjiu.lu@intel.com>
7853 * config/i386/i386.md (post-reload splitter
7854 to avoid partial SSE reg dependency stalls): New pattern.
7856 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
7858 * config/s390/s390.md ("movcc"): Swap load and store instructions.
7860 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
7862 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
7863 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
7865 2013-07-26 Cary Coutant <ccoutant@google.com>
7867 * dwarf2out.c (die_checksum_ordered): Don't include template
7868 instantiations in signature.
7869 (is_template_parameter): New function.
7870 (is_template_instantiation): New function.
7871 (generate_skeleton_bottom_up): Don't include template instantiations
7873 (generate_skeleton): Likewise.
7874 (break_out_comdat_types): Move recursive call to break out nested
7876 (prune_unused_types_mark_generic_parms_dies): Call
7877 is_template_parameter.
7879 2013-07-26 Ian Bolton <ian.bolton@arm.com>
7881 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
7882 uses vector registers.
7883 * config/aarch64/iterators.md: Add attributes rtn and vas.
7885 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7886 Richard Earnshaw <richard.earnshaw@arm.com>
7888 * combine.c (simplify_comparison): Re-canonicalize operands
7890 * config/arm/arm.md (movcond_addsi): New splitter.
7892 2013-07-25 Sterling Augustine <saugustine@google.com>
7894 * dwarf2out.c (size_of_pubnames): Move code to...
7895 (include_pubname_in_output): ...here. New.
7896 (want_pubnames): Rearrange.
7897 (output_pubnames): Call include_pubname_in_output. Move assertion.
7899 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
7901 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
7903 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
7906 * doc/extend.texi: Remove obsolete builtins. Fix
7907 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
7909 2013-07-25 Jan Hubicka <jh@suse.cz>
7911 * cgraph.c (release_function_body): Break out from ...
7912 (cgraph_release_function_body): ... this one; also release DECL_RESULT
7914 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
7915 old_tree in the map.
7916 (create_specialized_node): Update.
7917 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
7919 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
7920 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
7921 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
7922 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
7925 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7927 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
7928 addsi3_carryin_alt2_<optab>): Correct output template.
7930 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7932 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
7933 Adjust for arm_restrict_it.
7934 Remove trailing whitespace.
7936 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
7938 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
7939 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
7941 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
7943 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
7945 PR rtl-optimization/57960
7946 * lra-constraints.c (process_alt_operands): Use the right mode
7947 when checking strict_low.
7949 2013-07-25 Jan Hubicka <jh@suse.cz>
7951 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
7952 * cgraph.c (cgraph_remove_node): Do not release function body
7953 when in cgraph streaming.
7954 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
7955 in other partitions are not considered reachable; fix handling of
7958 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7960 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
7962 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7967 * config/arm/arm.md ("*sibcall_insn): Replace use of
7968 Ss with US. Adjust output for v5 and v4t.
7969 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
7971 * config/arm/constraints.md ("Ss"): Rename to US.
7973 2013-07-25 Terry Guo <terry.guo@arm.com>
7975 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
7976 shift_add/shift_sub0/shift_sub1 RTXs.
7978 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
7979 Anton Blanchard <anton@au1.ibm.com>
7981 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
7982 (altivec_vpks<VI_char>ss): Likewise.
7983 (altivec_vpks<VI_char>us): Likewise.
7984 (altivec_vpku<VI_char>us): Likewise.
7985 (altivec_vpku<VI_char>um): Likewise.
7987 2013-07-24 David Malcolm <dmalcolm@redhat.com>
7989 Introduce context class.
7991 * Makefile.in (CONTEXT_H): New.
7992 (OBJS): Add context.o.
7993 (toplev.o): Add CONTEXT_H to dependencies.
7996 * toplev.c (general_init): Create the singleton gcc::context instance.
8002 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
8004 PR rtl-optimization/57968
8005 * mode-switching.c (create_pre_exit): Allow instructions that
8006 don't set a return register to need a non-exit mode.
8008 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
8009 Anton Blanchard <anton@au1.ibm.com>
8011 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
8012 operands to vperm for little endian.
8013 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
8014 of lvsl to create the control mask for a vperm for little endian.
8016 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8017 Anton Blanchard <anton@au1.ibm.com>
8019 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
8020 two operands for little-endian.
8022 2013-07-23 Steve Ellcey <sellcey@mips.com>
8024 * config/mips/mips.c (mips_case_values_threshold): New.
8025 (TARGET_CASE_VALUES_THRESHOLD): Define.
8027 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8028 Anton Blanchard <anton@au1.ibm.com>
8030 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
8031 selection of field for vector splat in little endian mode.
8033 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
8035 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
8036 expanders to rs6000.md.
8037 (ior<mode>3): Likewise.
8038 (and<mode>3): Likewise.
8039 (one_cmpl<mode>2): Likewise.
8040 (nor<mode>3): Likewise.
8041 (andc<mode>3): Likewise.
8042 (eqv<mode>3): Likewise.
8043 (nand<mode>3): Likewise.
8044 (orc<mode>3): Likewise.
8046 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
8049 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
8050 to split multi-word logical operations.
8051 (rs6000_split_logical_di): Likewise.
8052 (rs6000_split_logical): Likewise.
8054 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
8055 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
8056 and allow TImode operations in 32-bit.
8057 (vsx_and<mode>3_64bit): Likewise.
8058 (vsx_ior<mode>3_32bit): Likewise.
8059 (vsx_ior<mode>3_64bit): Likewise.
8060 (vsx_xor<mode>3_32bit): Likewise.
8061 (vsx_xor<mode>3_64bit): Likewise.
8062 (vsx_one_cmpl<mode>2_32bit): Likewise.
8063 (vsx_one_cmpl<mode>2_64bit): Likewise.
8064 (vsx_nor<mode>3_32bit): Likewise.
8065 (vsx_nor<mode>3_64bit): Likewise.
8066 (vsx_andc<mode>3_32bit): Likewise.
8067 (vsx_andc<mode>3_64bit): Likewise.
8068 (vsx_eqv<mode>3_32bit): Likewise.
8069 (vsx_eqv<mode>3_64bit): Likewise.
8070 (vsx_nand<mode>3_32bit): Likewise.
8071 (vsx_nand<mode>3_64bit): Likewise.
8072 (vsx_orc<mode>3_32bit): Likewise.
8073 (vsx_orc<mode>3_64bit): Likewise.
8075 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
8076 logical types in GPRs.
8078 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
8079 logical insns to rs6000.md, and allow TImode operations in
8081 (altivec_ior<mode>3): Likewise.
8082 (altivec_xor<mode>3): Likewise.
8083 (altivec_one_cmpl<mode>2): Likewise.
8084 (altivec_nor<mode>3): Likewise.
8085 (altivec_andc<mode>3): Likewise.
8087 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
8088 attributes for moving the 128-bit logical operations into
8090 (BOOL_REGS_OUTPUT): Likewise.
8091 (BOOL_REGS_OP1): Likewise.
8092 (BOOL_REGS_OP2): Likewise.
8093 (BOOL_REGS_UNARY): Likewise.
8094 (BOOL_REGS_AND_CR0): Likewise.
8095 (one_cmpl<mode>2): Add support for DI logical operations on
8096 32-bit, splitting the operations to 32-bit.
8100 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
8101 changes to combine the 32/64-bit code, allow logical operations on
8102 TI mode in 32-bit, and to use similar match_operator patterns like
8103 scalar mode uses. Combine the Altivec and VSX code for logical
8104 operations, and move it here.
8105 (ior<mode>3, 128-bit types): Likewise.
8106 (xor<mode>3, 128-bit types): Likewise.
8107 (one_cmpl<mode>3, 128-bit types): Likewise.
8108 (nor<mode>3, 128-bit types): Likewise.
8109 (andc<mode>3, 128-bit types): Likewise.
8110 (eqv<mode>3, 128-bit types): Likewise.
8111 (nand<mode>3, 128-bit types): Likewise.
8112 (orc<mode>3, 128-bit types): Likewise.
8113 (and<mode>3_internal): Likewise.
8114 (bool<mode>3_internal): Likewise.
8115 (boolc<mode>3_internal1): Likewise.
8116 (boolc<mode>3_internal2): Likewise.
8117 (boolcc<mode>3_internal1): Likewise.
8118 (boolcc<mode>3_internal2): Likewise.
8119 (eqv<mode>3_internal1): Likewise.
8120 (eqv<mode>3_internal2): Likewise.
8121 (one_cmpl1<mode>3_internal): Likewise.
8123 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
8125 * config/microblaze/microblaze.c (microblaze_expand_prologue):
8126 Rename flag_stack_usage to flag_stack_usage_info.
8128 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
8130 * config/microblaze/sync.md: New file.
8131 * config/microblaze/microblaze.md: Include sync.md
8132 * config/microblaze/microblaze.c: Add print_operand 'y'.
8133 * config/microblaze/constraints.md: Add memory_contraint
8134 'Q' which is a single register.
8136 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
8138 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
8140 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
8142 * reload.c (find_reloads): Exit loop once we find this operand
8143 cannot be reloaded somehow for this alternative.
8145 * reload.c (find_reloads): Exit loop once we find a hard register.
8147 * rtlanal.c (computed_jump_p): Exit loop once we find label
8150 * i386.c (ix86_pad_returns): Exit loop after setting replace.
8152 * cfgloopmanip.c (remove_path): Exit loop after setting
8155 * gensupport.c (subst_dup): Avoid loop if code is not
8156 MATCH_DUP nor MATCH_OP_DUP.
8158 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
8160 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
8162 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8164 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
8165 true for SP_REGNUM if mode == ptr_mode.
8166 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
8167 with value R0_REGNUM + 31.
8169 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8171 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
8172 pad pointer-typed argument downward.
8174 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8176 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
8177 and __ILP32__ when the ILP32 model is in use.
8179 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8181 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
8182 (aarch64_load_symref_appropriately): In the case of
8183 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
8184 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
8185 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
8186 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
8187 if 'mode' doesn't equal to 'ptr_mode'.
8188 (aarch64_output_mi_thunk): Add an assertion on the alignment of
8189 'vcall_offset'; change to call aarch64_emit_move differently depending
8190 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
8191 to calculate the upper bound of 'vcall_offset'.
8192 (aarch64_cannot_force_const_mem): Change to also return true if
8194 (aarch64_legitimize_reload_address): In the case of large
8195 displacements, add new local variable 'xmode' and an assertion
8196 based on it; change to use 'xmode' to generate the new rtx and
8198 (aarch64_asm_trampoline_template): Change to generate the template
8199 differently depending on TARGET_ILP32 or not; change to use
8200 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
8201 (aarch64_trampoline_size): Removed.
8202 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
8203 and replace immediate literals with it. Change to use 'ptr_mode'
8204 instead of 'DImode' and call convert_memory_address if the mode
8205 of 'fnaddr' doesn't equal to 'ptr_mode'.
8206 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
8208 (aarch64_elf_asm_destructor): Likewise.
8209 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
8210 on TARGET_ILP32 instead of aarch64_trampoline_size.
8211 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
8212 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
8213 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
8214 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
8215 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
8216 (storewb_pair<GPI:mode>_<P:mode>): ... this.
8217 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
8218 depending on the value of 'mode'.
8219 (add_losym_<mode>): New.
8220 (ldr_got_small_<mode>): New, based on ldr_got_small.
8221 (ldr_got_small): Remove.
8222 (ldr_got_small_sidi): New.
8223 * config/aarch64/iterators.md (P): New.
8224 (PTR): Change to 'ptr_mode' in the condition.
8226 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
8228 * config.gcc (aarch64*-*-*): Support --with-abi.
8229 (aarch64*-*-elf): Support --with-multilib-list.
8230 (aarch64*-*-linux*): Likewise.
8231 (supported_defaults): Add abi to aarch64*-*-*.
8232 * configure.ac: Mention AArch64 for --with-multilib-list.
8233 * configure: Re-generated.
8234 * config/aarch64/biarchilp32.h: New file.
8235 * config/aarch64/biarchlp64.h: New file.
8236 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
8238 (MULTILIB_DEFAULTS): Ditto.
8239 (DRIVER_SELF_SPECS): Ditto.
8240 (ASM_SPEC): Update to also substitute -mabi.
8241 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
8242 file whose name depends on -mabi= and -mbig-endian.
8243 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
8245 (POINTER_SIZE): New define.
8246 (POINTERS_EXTEND_UNSIGNED): Ditto.
8247 (enum aarch64_abi_type): New enumeration tag.
8248 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
8249 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
8250 (TARGET_ILP32): New define.
8251 * config/aarch64/aarch64.opt (mabi): New.
8253 (ilp32, lp64): New values for -mabi.
8254 * config/aarch64/t-aarch64 (comma): New define.
8255 (MULTILIB_OPTIONS): Ditto.
8256 (MULTILIB_DIRNAMES): Ditto.
8257 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
8258 * doc/invoke.texi: Document -mabi for AArch64.
8260 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
8262 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
8264 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
8265 Anton Blanchard <anton@au1.ibm.com>
8267 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
8268 endianness when selecting field to splat.
8270 2013-07-22 Eric Christopher <echristo@gmail.com>
8272 * dwarf2out.c (die_odr_checksum): New function to use
8273 CHECKSUM_ macros and ULEB128 for DIE tag.
8274 (generate_type_signature): Use.
8276 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
8278 * config.gcc (sparc*-*-*): Accept leon3 processor.
8279 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
8280 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
8281 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
8282 * config/sparc/sparc.opt (enum processor_type): Add leon3.
8283 (mfix-ut699): Adjust comment.
8284 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
8285 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
8286 (CPP_CPU_SPEC): Likewise.
8287 (ASM_CPU_SPEC): Likewise.
8288 * config/sparc/sparc.c (leon3_cost): New constant.
8289 (sparc_option_override): Add leon3 support.
8290 (mem_ref): New function.
8291 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
8292 (sparc_do_work_around_errata): Look into the instruction in the delay
8293 slot and adjust accordingly. Add fix for the data cache nullify issues
8294 of the UT699. Change insertion position for the NOP.
8295 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
8296 (leon3_load): New reservation.
8297 (leon_store): Bump latency to 2.
8298 (grfpu): New automaton.
8299 (grfpu_alu): New unit.
8300 (grfpu_ds): Likewise.
8301 (leon_fp_alu): Adjust.
8302 (leon_fp_mult): Delete.
8303 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
8304 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
8305 * config/sparc/sparc.md (cpu): Add leon3.
8306 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
8308 (atomic_test_and_set): Likewise.
8311 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
8313 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
8314 default for R5900 targets.
8315 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
8316 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
8317 * config/mips/mips.c (mips_option_override): Report an error for
8318 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
8319 for -march=r5900 -mhard-float.
8321 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
8323 * df-problems.c (can_move_insns_across): Exit loop once we
8324 find a non-fixed, non-global register.
8326 * ipa-pure-const.c (propagate_nothrow): Exit loop after
8329 * omega.c (omega_eliminate_red): Break after setting red_found.
8330 (omega_problem_has_red_equations): Similarly after setting found.
8331 (omega_query_variable): Similarly after setting coupled.
8333 2013-07-22 Marek Polacek <polacek@redhat.com>
8335 * gimplify.c: Don't include gimple.h twice.
8337 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8339 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
8340 instead of TARGET_THUMB1.
8341 (Pz): New constraint.
8342 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
8344 (compare_negsi_si): Likewise.
8345 (compare_addsi2_op0): Likewise.
8346 (compare_addsi2_op1): Likewise.
8347 (addsi3_carryin_<optab>): Likewise.
8348 (addsi3_carryin_alt2_<optab>): Likewise.
8349 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
8350 for arm_restrict_it.
8351 (subsi3_carryin): Likewise.
8352 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
8353 (minmax_arithsi): Disable for arm_restrict_it.
8354 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
8355 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
8356 (satsi_<SAT:code>_shift): Likewise.
8357 (arm_shiftsi3): Add alternative for 16-bit encoding.
8358 (arm32_movhf): Disable for arm_restrict_it.
8359 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
8360 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
8362 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
8364 * config/arm/arm.md (attribute "insn"): Delete.
8365 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
8366 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
8367 (not_shiftsi): Update for attribute change.
8368 (not_shiftsi_compare0): Likewise.
8369 (not_shiftsi_compare0_scratch): Likewise.
8370 (arm_one_cmplsi2): Likewise.
8371 (thumb1_one_cmplsi2): Likewise.
8372 (notsi_compare0): Likewise.
8373 (notsi_compare0_scratch): Likewise.
8374 (thumb1_movdi_insn): Likewise.
8375 (arm_movsi_insn): Likewise.
8376 (movhi_insn_arch4): Likewise.
8377 (movhi_bytes): Likewise.
8378 (arm_movqi_insn): Likewise.
8379 (thumb1_movqi_insn): Likewise.
8380 (arm32_movhf): Likewise.
8381 (thumb1_movhf): Likewise.
8382 (arm_movsf_soft_insn): Likewise.
8383 (thumb1_movsf_insn): Likewise.
8384 (thumb_movdf_insn): Likewise.
8385 (movsicc_insn): Likewise.
8386 (movsfcc_soft_insn): Likewise.
8387 (and_scc): Likewise.
8388 (cond_move): Likewise.
8389 (if_move_not): Likewise.
8390 (if_not_move): Likewise.
8391 (if_shift_move): Likewise.
8392 (if_move_shift): Likewise.
8393 (if_shift_shift): Likewise.
8394 (if_not_arith): Likewise.
8395 (if_arith_not): Likewise.
8396 (cond_move_not): Likewise.
8397 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
8398 (neon_mov<mode>): Likewise.
8399 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
8400 (thumb2_movsi_vfp): Likewise.
8401 (movsf_vfp): Likewise.
8402 (thumb2_movsf_vfp): Likewise.
8403 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
8405 (cortexa7_older_only): Likewise.
8406 (cortexa7_younger): Likewise.
8407 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
8408 (1020alu_shift_op): Likewise.
8409 (1020alu_shift_reg_op): Likewise.
8410 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
8411 (alu_shift_op): Likewise.
8412 (alu_shift_reg_op): Likewise.
8413 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
8414 (11_alu_shift_op): Likewise.
8415 (11_alu_shift_reg_op): Likewise.
8416 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
8417 (9_alu_shift_reg_op): Likewise.
8418 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
8420 (cortex_a15_alu_shift): Likewise.
8421 (cortex_a15_alu_shift_reg): Likewise.
8422 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
8423 (cortex_a5_alu_shift): Likewise.
8424 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
8426 (cortex_a53_alu_shift): Likewise.
8427 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
8429 (cortex_a7_alu_reg): Likewise.
8430 (cortex_a7_alu_shift): Likewise.
8431 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
8432 (cortex_a8_alu_shift): Likewise.
8433 (cortex_a8_alu_shift_reg): Likewise.
8434 (cortex_a8_mov): Likewise.
8435 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
8436 (cortex_a9_dp_shift): Likewise.
8437 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
8438 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
8439 (cortex_r4_mov): Likewise.
8440 (cortex_r4_alu_shift): Likewise.
8441 (cortex_r4_alu_shift_reg): Likewise.
8442 * config/arm/fa526.md (526_alu_op): Update for attribute change.
8443 (526_alu_shift_op): Likewise.
8444 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
8445 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
8446 (626te_alu_shift_op): Likewise.
8447 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
8448 (726te_alu_op): Likewise.
8449 (726te_alu_shift_op): Likewise.
8450 (726te_alu_shift_reg_op): Likewise.
8451 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
8452 (mp626_alu_shift_op): Likewise.
8453 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
8454 (pj4_alu_e1_conds): Likewise.
8455 (pj4_alu): Likewise.
8456 (pj4_alu_conds): Likewise.
8457 (pj4_shift): Likewise.
8458 (pj4_shift_conds): Likewise.
8459 (pj4_alu_shift): Likewise.
8460 (pj4_alu_shift_conds): Likewise.
8462 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8464 * config/arm/predicates.md (shiftable_operator_strict_it):
8466 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
8467 Disable cond_exec version for arm_restrict_it.
8468 (thumb2_smaxsi3): Convert to generate cond_exec.
8469 (thumb2_sminsi3): Likewise.
8470 (thumb32_umaxsi3): Likewise.
8471 (thumb2_uminsi3): Likewise.
8472 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
8473 (thumb2_neg_abssi2): Likewise.
8474 (thumb2_mov_scc): Add alternative for 16-bit encoding.
8475 (thumb2_movsicc_insn): Adjust alternatives.
8476 (thumb2_mov_negscc): Disable for arm_restrict_it.
8477 (thumb2_mov_negscc_strict_it): New pattern.
8478 (thumb2_mov_notscc_strict_it): New pattern.
8479 (thumb2_mov_notscc): Disable for arm_restrict_it.
8480 (thumb2_ior_scc): Likewise.
8481 (thumb2_ior_scc_strict_it): New pattern.
8482 (thumb2_cond_move): Adjust for arm_restrict_it.
8483 (thumb2_cond_arith): Disable for arm_restrict_it.
8484 (thumb2_cond_arith_strict_it): New pattern.
8485 (thumb2_cond_sub): Adjust for arm_restrict_it.
8486 (thumb2_movcond): Likewise.
8487 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
8488 (thumb2_zero_extendhisi2_v6): Likewise.
8489 (thumb2_zero_extendqisi2_v6): Likewise.
8490 (orsi_notsi_si): Likewise.
8491 (orsi_not_shiftsi_si): Likewise.
8493 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
8495 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
8496 instruction sequence is 1 byte shorter.
8498 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
8500 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
8501 it is not needed after split.
8503 2013-07-20 Iain Sandoe <iain@codesourcery.com>
8506 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
8507 second label for nonlocal goto receivers. Don't output pic base labels
8508 unless we're producing PIC; mark that action unreachable().
8509 (ix86_save_reg): If the function contains a nonlocal label, save the
8511 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
8512 * config/darwin.c (emitted_pic_label_num): New GTY.
8513 (update_pic_label_number_if_needed): New.
8514 (machopic_output_function_base_name): Adjust for nonlocal receiver
8516 (machopic_should_output_picbase_label): New.
8517 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
8518 (nonlocal_goto_receiver): New insn and split.
8520 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
8522 * config/aarch64/aarch64-builtins.c
8523 (aarch64_fold_builtin): Fold abs in all modes.
8524 * config/aarch64/aarch64-simd-builtins.def
8525 (abs): Enable for all modes.
8526 * config/aarch64/arm_neon.h
8527 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
8528 (vabs_f64): Add missing intrinsic.
8530 2013-07-19 Ian Bolton <ian.bolton@arm.com>
8532 * config/aarch64/arm_neon.h (vabs_s64): New function
8534 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
8537 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
8538 * config/avr/avr.md (adjust_len): Add `round'.
8539 * config/avr/avr-protos.h (avr_out_round): New prototype.
8540 (avr_out_plus): Add `out_label' argument.
8541 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
8542 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
8543 Handle the case where `insn' is just a pattern.
8544 (avr_out_bitop): Handle the case where `insn' is just a pattern.
8545 (avr_out_round): New function.
8546 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
8548 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
8550 * config/microblaze/microblaze.c (microblaze_expand_prologue):
8551 Add check for flag_stack_usage to handle -fstack-usage support
8553 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
8555 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
8556 interaction for new Power8 flags and VSX.
8558 2013-07-18 Sriraman Tallam <tmsriram@google.com>
8561 * tree-inline.c (expand_call_inline): Emit errors during
8562 early_inlining only if optimization is not turned on.
8564 2013-07-18 David Malcolm <dmalcolm@redhat.com>
8568 * passes.c (init_optimization_passes): Move the construction of
8569 the pass hierarchy into a new passes.def file.
8571 * Makefile.in (passes.o): Add dependency on passes.def.
8573 2013-07-18 David Malcolm <dmalcolm@redhat.com>
8575 * passes.c (init_optimization_passes): Introduce macros for
8576 constructing the tree of passes (INSERT_PASSES_AFTER,
8577 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
8578 TERMINATE_PASS_LIST).
8580 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
8581 Wei Mi <wmi@google.com>
8583 PR rtl-optimization/57878
8584 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
8586 (reload_pseudo_compare_func): Check nregs first for reload
8589 2013-07-18 David Malcolm <dmalcolm@redhat.com>
8591 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
8593 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
8595 * read-rtl.c (validate_const_int): Once an invalid character is
8596 seen, quit the loop.
8598 * gengtype.c (write_roots): Similarly once we find the "deletable"
8599 or "if_marked" option.
8601 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
8603 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
8604 "xtab" and "sat". Move value "clz" from here to ...
8605 (attriubte "type"): ... here.
8606 (satsi_<SAT:code>): Delete "insn" attribute.
8607 (satsi_<SAT:code>_shift): Likewise.
8608 (arm_zero_extendqisi2addsi): Likewise.
8609 (arm_extendqisi2addsi): Likewise.
8610 (clzsi2): Update for attribute changes.
8611 (rbitsi2): Likewise.
8612 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
8614 (arm_usatsihi): Likewise.
8615 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
8617 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
8619 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
8620 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
8621 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
8622 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
8623 in alphabetical order.
8624 (attribute "core_cycles"): Update for attribute changes.
8625 (arm_addsi3): Likewise.
8626 (addsi3_compare0): Likewise.
8627 (addsi3_compare0_scratch): Likewise.
8628 (addsi3_compare_op1): Likewise.
8629 (addsi3_compare_op2): Likewise.
8630 (compare_addsi2_op0): Likewise.
8631 (compare_addsi2_op1): Likewise.
8632 (addsi3_carryin_shift_<optab>): Likewise.
8633 (subsi3_carryin_shift): Likewise.
8634 (rsbsi3_carryin_shift): Likewise.
8635 (arm_subsi3_insn): Likewise.
8636 (subsi3_compare0): Likewise.
8637 (subsi3_compare): Likewise.
8638 (arm_andsi3_insn): Likewise.
8639 (thumb1_andsi3_insn): Likewise.
8640 (andsi3_compare0): Likewise.
8641 (andsi3_compare0_scratch): Likewise.
8642 (zeroextractsi_compare0_scratch
8643 (andsi_not_shiftsi_si): Likewise.
8644 (iorsi3_insn): Likewise.
8645 (iorsi3_compare0): Likewise.
8646 (iorsi3_compare0_scratch): Likewise.
8647 (arm_xorsi3): Likewise.
8648 (thumb1_xorsi3_insn): Likewise.
8649 (xorsi3_compare0): Likewise.
8650 (xorsi3_compare0_scratch): Likewise.
8651 (satsi_<SAT:code>_shift): Likewise.
8653 (arm_shiftsi3): Likewise.
8654 (shiftsi3_compare0): Likewise.
8655 (not_shiftsi): Likewise.
8656 (not_shiftsi_compare0): Likewise.
8657 (not_shiftsi_compare0_scratch): Likewise.
8658 (arm_one_cmplsi2): Likewise.
8659 (thumb_one_complsi2): Likewise.
8660 (notsi_compare0): Likewise.
8661 (notsi_compare0_scratch): Likewise.
8662 (thumb1_zero_extendhisi2): Likewise.
8663 (arm_zero_extendhisi2): Likewise.
8664 (arm_zero_extendhisi2_v6): Likewise.
8665 (arm_zero_extendhisi2addsi): Likewise.
8666 (thumb1_zero_extendqisi2): Likewise.
8667 (thumb1_zero_extendqisi2_v6): Likewise.
8668 (arm_zero_extendqisi2): Likewise.
8669 (arm_zero_extendqisi2_v6): Likewise.
8670 (arm_zero_extendqisi2addsi): Likewise.
8671 (thumb1_extendhisi2): Likewise.
8672 (arm_extendhisi2): Likewise.
8673 (arm_extendhisi2_v6): Likewise.
8674 (arm_extendqisi): Likewise.
8675 (arm_extendqisi_v6): Likewise.
8676 (arm_extendqisi2addsi): Likewise.
8677 (thumb1_extendqisi2): Likewise.
8678 (thumb1_movdi_insn): Likewise.
8679 (arm_movsi_insn): Likewise.
8680 (movsi_compare0): Likewise.
8681 (movhi_insn_arch4): Likewise.
8682 (movhi_bytes): Likewise.
8683 (arm_movqi_insn): Likewise.
8684 (thumb1_movqi_insn): Likewise.
8685 (arm32_movhf): Likewise.
8686 (thumb1_movhf): Likewise.
8687 (arm_movsf_soft_insn): Likewise.
8688 (thumb1_movsf_insn): Likewise.
8689 (movdf_soft_insn): Likewise.
8690 (thumb_movdf_insn): Likewise.
8691 (arm_cmpsi_insn): Likewise.
8692 (cmpsi_shiftsi): Likewise.
8693 (cmpsi_shiftsi_swp): Likewise.
8694 (arm_cmpsi_negshiftsi_si): Likewise.
8695 (movsicc_insn): Likewise.
8696 (movsfcc_soft_insn): Likewise.
8697 (arith_shiftsi): Likewise.
8698 (arith_shiftsi_compare0
8699 (arith_shiftsi_compare0_scratch
8700 (sub_shiftsi): Likewise.
8701 (sub_shiftsi_compare0
8702 (sub_shiftsi_compare0_scratch
8703 (and_scc): Likewise.
8704 (cond_move): Likewise.
8705 (if_plus_move): Likewise.
8706 (if_move_plus): Likewise.
8707 (if_move_not): Likewise.
8708 (if_not_move): Likewise.
8709 (if_shift_move): Likewise.
8710 (if_move_shift): Likewise.
8711 (if_shift_shift): Likewise.
8712 (if_not_arith): Likewise.
8713 (if_arith_not): Likewise.
8714 (cond_move_not): Likewise.
8715 (thumb1_ashlsi3): Set type attribute.
8716 (thumb1_ashrsi3): Likewise.
8717 (thumb1_lshrsi3): Likewise.
8718 (thumb1_rotrsi3): Likewise.
8719 (shiftsi3_compare0_scratch): Likewise.
8720 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
8721 (neon_mov<mode>): Likewise.
8722 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
8724 (thumb2_movsi_insn): Likewise.
8725 (thumb2_cmpsi_neg_shiftsi): Likewise.
8726 (thumb2_extendqisi_v6): Likewise.
8727 (thumb2_zero_extendhisi2_v6): Likewise.
8728 (thumb2_zero_extendqisi2_v6): Likewise.
8729 (thumb2_shiftsi3_short): Likewise.
8730 (thumb2_addsi3_compare0_scratch): Likewise.
8731 (orsi_not_shiftsi_si): Likewise.
8732 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
8733 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
8735 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
8736 (1020alu_shift_op): Likewise.
8737 (1020alu_shift_reg_op): Likewise.
8738 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
8739 (alu_shift_op): Likewise.
8740 (alu_shift_reg_op): Likewise.
8741 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
8742 (11_alu_shift_op): Likewise.
8743 (11_alu_shift_reg_op): Likewise.
8744 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
8745 (9_alu_shift_reg_op): Likewise.
8746 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
8748 (cortex_a15_alu_shift): Likewise.
8749 (cortex_a15_alu_shift_reg): Likewise.
8750 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
8752 (cortex_a5_alu_shift): Likewise.
8753 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
8755 (cortex_a53_alu_shift): Likewise.
8756 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
8758 (cortex_a7_alu_reg): Likewise.
8759 (cortex_a7_alu_shift): Likewise.
8760 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
8762 (cortex_a8_alu_shift): Likewise.
8763 (cortex_a8_alu_shift_reg): Likewise.
8764 (cortex_a8_mov): Likewise.
8765 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
8766 (cortex_a9_dp_shift): Likewise.
8767 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
8769 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
8771 (cortex_r4_mov): Likewise.
8772 (cortex_r4_alu_shift): Likewise.
8773 (cortex_r4_alu_shift_reg): Likewise.
8774 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
8775 (526_alu_shift_op): Likewise.
8776 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
8777 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
8778 (626te_alu_shift_op): Likewise.
8779 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
8780 (726te_alu_op): Likewise.
8781 (726te_alu_shift_op): Likewise.
8782 (726te_alu_shift_reg_op): Likewise.
8783 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
8784 (mp626_alu_shift_op): Likewise.
8785 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
8786 (pj4_alu_e1_conds): Likewise.
8787 (pj4_alu): Likewise.
8788 (pj4_alu_conds): Likewise.
8789 (pj4_shift): Likewise.
8790 (pj4_shift_conds): Likewise.
8791 (pj4_alu_shift): Likewise.
8792 (pj4_alu_shift_conds): Likewise.
8793 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
8795 (cortexa7_older_only): Likewise.
8796 (cortexa7_younger): Likewise.
8798 2013-07-18 David Malcolm <dmalcolm@redhat.com>
8800 * ipa-pure-const.c (generate_summary): Rename to...
8801 (pure_const_generate_summary): ... this.
8803 2013-07-17 Iain Sandoe <iain@codesourcery.com>
8805 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
8807 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
8810 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
8813 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8815 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
8816 enabled without -march=zEC12.
8817 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
8820 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
8822 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
8823 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
8824 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
8825 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
8826 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
8827 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
8828 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
8830 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
8831 and ISA_HAS_NMADD3_NMSUB3.
8832 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
8833 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
8834 (nmsub4<mode>, nmsub3<mode>): Likewise.
8835 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
8837 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
8839 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
8840 TARGET_MIPS5400 checking.
8842 2013-07-16 Jakub Jelinek <jakub@redhat.com>
8843 Peter Bergner <bergner@vnet.ibm.com>
8845 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
8846 registers in the comment.
8847 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
8848 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
8849 rather than FIRST_PSEUDO_REGISTERS.
8851 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
8853 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
8854 enable extra ISA flags with TARGET_HTM.
8856 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
8858 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
8861 2013-07-15 Cong Hou <congh@google.com>
8863 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
8864 in compare function for sorting.
8866 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
8868 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
8869 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
8870 * config/rs6000/rs6000.opt: Add -mhtm option.
8871 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
8872 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
8873 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
8874 __HTM__ if the HTM instructions are available.
8875 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
8876 htm_spr_reg_operand): New define_predicates.
8877 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
8878 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
8880 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
8881 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
8882 HTM builtin functions.
8883 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
8884 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
8885 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
8886 (rs6000_builtin_mask_calculate): Likewise.
8887 (rs6000_option_override_internal): Likewise.
8888 (bdesc_htm): Add new HTM builtin support.
8889 (htm_spr_num): New function.
8890 (htm_spr_regno): Likewise.
8891 (rs6000_htm_spr_icode): Likewise.
8892 (htm_expand_builtin): Likewise.
8893 (htm_init_builtins): Likewise.
8894 (rs6000_expand_builtin): Add support for HTM builtin functions.
8895 (rs6000_init_builtins): Likewise.
8896 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
8898 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
8899 (TARGET_HTM, MASK_HTM): Define macros.
8900 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
8901 (FIXED_REGISTERS): Likewise.
8902 (CALL_USED_REGISTERS): Likewise.
8903 (CALL_REALLY_USED_REGISTERS): Likewise.
8904 (REG_ALLOC_ORDER): Likewise.
8905 (enum reg_class): Likewise.
8906 (REG_CLASS_NAMES): Likewise.
8907 (REG_CLASS_CONTENTS): Likewise.
8908 (REGISTER_NAMES): Likewise.
8909 (ADDITIONAL_REGISTER_NAMES): Likewise.
8910 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
8911 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
8912 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
8913 * config/rs6000/htm.md: New file.
8914 * config/rs6000/htmintrin.h: New file.
8915 * config/rs6000/htmxlintrin.h: New file.
8917 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
8919 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
8920 Define SYMBOL_TINY_GOT, update comment.
8921 * config/aarch64/aarch64.c
8922 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
8923 (aarch64_expand_mov_immediate): Likewise.
8924 (aarch64_print_operand): Likewise.
8925 (aarch64_classify_symbol): Likewise.
8926 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
8927 (ldr_got_tiny): Define.
8929 2013-07-13 Tobias Grosser <tobias@grosser.es>
8931 PR tree-optimization/54094
8932 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
8933 scheduling dimension for the parallelism check from the polyhedral
8934 information in the AST.
8935 * graphite-dependences.c (carries_deps): Do not assume the schedule is
8938 2013-07-13 Jason Merrill <jason@redhat.com>
8940 * print-tree.c (debug_vec_tree): Use debug_raw.
8941 (debug_raw (vec<tree, va_gc> &)): New.
8942 (debug_raw (vec<tree, va_gc> *)): New.
8943 * tree.h: Declare them.
8945 2013-07-13 Bin Cheng <bin.cheng@arm.com>
8947 * ifcvt.c (ifcvt_after_combine): New static variable.
8948 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
8950 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
8951 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
8952 rest_of_handle_if_after_reload): Pass new argument for if_convert.
8954 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
8956 * config/mips/mips.c (mips_expand_call): Remove empty statement.
8958 2013-07-12 Michael Matz <matz@suse.de>
8961 * convert.c (convert_to_real): Reject non-float inner types.
8963 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
8965 * config/aarch64/aarch64-protos.h
8966 (aarch64_simd_immediate_valid_for_move): Remove.
8967 * config/aarch64/aarch64.c (simd_immediate_info): New member.
8968 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
8970 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
8972 2013-07-11 Steve Ellcey <sellcey@mips.com>
8974 * config/mips/mips.c (mips_conditional_register_usage): Do not
8975 use t[0-7] registers in MIPS16 mode when optimizing for size.
8977 2013-07-11 Sriraman Tallam <tmsriram@google.com>
8979 * config/i386/i386.c (dispatch_function_versions): Fix array
8980 indexing of function_version_info to match actual_versions.
8982 2013-07-11 Teresa Johnson <tejohnson@google.com>
8984 * vec.h (struct va_gc): Move release out-of-line.
8985 (va_gc::release): Call ggc_free on released vec.
8987 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
8989 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
8990 Require GOT register as additional operand in UNSPEC.
8991 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
8992 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
8993 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
8994 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
8995 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
8996 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
8997 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
8999 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
9002 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
9003 name seen by assembler/linker rather if available.
9005 2013-07-11 Andreas Schwab <schwab@suse.de>
9007 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
9009 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
9011 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
9013 2013-07-10 Joseph Myers <joseph@codesourcery.com>
9015 * doc/tm.texi.in: Move hook documentation to ....
9016 * target.def: ... here.
9018 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
9020 * doc/tm.texi: Regenerate.
9022 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
9025 * doc/invoke.texi: Document -Wconditionally-supported.
9027 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
9030 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
9033 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
9036 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
9037 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
9038 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
9039 Remove duplicate devices.
9040 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
9041 * config/avr/t-multilib: Regenerate.
9042 * config/avr/avr-tables.opt: Regenerate.
9043 * doc/avr-mmcu.texi: Regenerate.
9045 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
9048 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
9050 2013-07-10 Graham Stott <graham.stott@btinternet.com>
9052 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
9053 the cost of MULT when optimizing for size.
9055 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
9057 * config/cr16/cr16-protos.h: Don't include target.h.
9059 2013-07-09 Joseph Myers <joseph@codesourcery.com>
9061 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
9062 adjust register size for TDmode and TFmode for VSX registers.
9064 2013-07-08 Kai Tietz <ktietz@redhat.com>
9067 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
9068 hook_bool_const_tree_true.
9070 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9072 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
9073 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
9074 * config/s390/s390.md: Define FPR*_REGNUM constants.
9075 Fix FPR2_REGNUM constant (18 -> 17).
9076 ("*trunc<BFP:mode><DFP_ALL:mode>2")
9077 ("*trunc<DFP_ALL:mode><BFP:mode>2")
9078 ("trunc<BFP:mode><DFP_ALL:mode>2")
9079 ("trunc<DFP_ALL:mode><BFP:mode>2")
9080 ("*extend<BFP:mode><DFP_ALL:mode>2")
9081 ("*extend<DFP_ALL:mode><BFP:mode>2")
9082 ("extend<BFP:mode><DFP_ALL:mode>2")
9083 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
9086 2013-07-08 Graham Stott <graham.stott@btinternet.com>
9088 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
9090 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9092 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
9093 and cfun_fpr_bit_p to cfun_fpr_save_p.
9094 (s390_frame_area, s390_register_info, s390_frame_info)
9095 (s390_emit_prologue, s390_emit_epilogue)
9096 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
9098 * config/s390/s390.h: Define *_REGNUM macros for floating point
9101 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
9103 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
9105 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
9107 PR rtl-optimization/57786
9108 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
9109 and break out of the loop when it is set to false.
9111 2013-07-08 Jakub Jelinek <jakub@redhat.com>
9114 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
9115 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
9116 (const_int 63)) 0)).
9117 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
9118 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
9119 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
9121 PR rtl-optimization/57829
9122 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
9123 mask bits outside of mode are just sign-extension from mode to HWI.
9125 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
9127 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
9128 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
9129 adjust_address instead of change_address to keep info about alignment.
9130 (emit_strmov): Remove.
9131 (emit_memmov): New function.
9132 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
9133 (expand_movmem_epilogue): Likewise and return updated rtx for
9135 (expand_constant_movmem_prologue): Likewise and return updated rtx for
9136 destination and source.
9137 (decide_alignment): Refactor, handle vector_loop.
9138 (ix86_expand_movmem): Likewise.
9139 (ix86_expand_setmem): Likewise.
9140 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
9142 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
9144 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
9145 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
9147 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
9149 * config/i386/sse.md (sse_movlhps): Change alternative 3
9150 of operand 2 to "m".
9152 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
9155 * config/i386/sse.md (iptr): New mode attribute.
9156 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
9157 (<sse>_vm<plusminus_insn><mode>3): Ditto.
9158 (<sse>_vmmul<mode>3): Ditto.
9159 (<sse>_vmdiv<mode>3): Ditto.
9160 (sse_vmrcpv4sf2): Ditto.
9161 (<sse>_vmsqrt<mode>2): Ditto.
9162 (sse_vmrsqrtv4sf2): Ditto.
9163 (<sse>_vm<code><mode>3): Ditto.
9164 (avx_vmcmp<mode>3): Ditto.
9165 (<sse>_vmmaskcmp<mode>3): Ditto.
9166 (<sse>_comi): Ditto.
9167 (<sse>_ucomi): Ditto.
9168 (*xop_vmfrcz_<mode>): Ditto.
9169 (*fmai_fmadd_<mode>): Ditto.
9170 (*fmai_fmsub_<mode>): Ditto.
9171 (*fmai_fnmadd_<mode>): Ditto.
9172 (*fmai_fnmsub_<mode>): Ditto.
9173 (*fma4i_vmfmadd_<mode>): Ditto.
9174 (*fma4i_vmfmsub_<mode>): Ditto.
9175 (*fma4i_vmfnmadd_<mode>): Ditto.
9176 (*fma4i_vmfnmsub_<mode>): Ditto.
9177 (*xop_vmfrcz_<mode>): Ditto.
9178 (sse_cvtps2pi): Ditto.
9179 (sse_cvttps2pi): Ditto.
9180 (sse_cvtss2si): Ditto.
9181 (sse_cvtss2si_2): Ditto.
9182 (sse_cvtss2siq_2): Ditto.
9183 (sse_cvttss2si): Ditto.
9184 (sse_cvttss2siq): Ditto.
9185 (sse_cvtsd2si): Ditto.
9186 (sse_cvtsd2si_2): Ditto.
9187 (sse_cvtsd2siq_2): Ditto.
9188 (sse_cvttsd2si): Ditto.
9189 (sse_cvttsd2siq): Ditto.
9190 (sse_cvtsd2ss): Ditto.
9191 (sse_cvtss2sd): Ditto.
9192 (avx2_pbroadcast<mode>): Ditto.
9193 (avx2_pbroadcast<mode>_1): Ditto.
9194 (*avx_vperm_broadcast_v4sf): Ditto.
9196 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
9197 (sse_movlhps): Ditto.
9198 (sse_storehps): Ditto.
9199 (sse_loadhps): Ditto.
9200 (sse_storelps): Ditto.
9201 (sse_loadlps): Ditto.
9202 (*vec_concatv4sf): Ditto.
9203 (*vec_interleave_highv2df): Ditto.
9204 (*vec_interleave_lowv2df): Ditto.
9205 (*vec_extractv2df_1_sse): Ditto.
9206 (*vec_extractv2df_0_sse): Ditto.
9207 (sse2_storelpd): Ditto.
9208 (sse2_loadlpd): Ditto.
9209 (sse2_movsd): Ditto.
9210 (*vec_concatv4si): Ditto.
9211 (vec_concatv2di): Ditto.
9213 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
9214 for Intel asm dialect.
9215 (mmx_punpcklwd): Ditto.
9216 (mmx_punpckldq): Ditto.
9218 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
9219 for intel assembler dialect.
9221 2013-07-06 Jakub Jelinek <jakub@redhat.com>
9224 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
9225 for BUILT_IN_C{LZ,LRSB}*.
9226 * tree.h (CASE_INT_FN): Add FN##IMAX case.
9227 * tree-vrp.c (extract_range_basic): Handle
9228 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
9229 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
9230 fall thru to code calling set_value*.
9231 * builtins.c (expand_builtin): Remove *IMAX cases.
9232 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
9233 if width is bigger than 2*HWI.
9235 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
9237 PR rtl-optimization/55342
9238 * lra-int.h (lra_subreg_reload_pseudos): New.
9239 * lra.c: Add undoing optional reloads to the block diagram.
9240 (lra_subreg_reload_pseudos): New.
9241 (lra_optional_reload_pseudos): Change comments.
9242 (lra): Init and clear lra_subreg_reload_pseudos. Clear
9243 lra_optional_reload_pseudos after undo transformations.
9244 * lra-assigns.c (pseudo_prefix_title): New.
9245 (lra_setup_reg_renumber): Use it.
9246 (spill_for): Ditto. Check subreg reload pseudos too.
9247 (assign_by_spills): Consider subreg reload pseudos too.
9248 * lra-constraints.c (simplify_operand_subreg): Use
9249 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
9250 (curr_insn_transform): Recognize and do optional reloads.
9251 (undo_optional_reloads): New.
9252 (lra_undo_inheritance): Call undo_optional_reloads.
9254 2013-07-05 Thomas Quinot <quinot@adacore.com>
9256 * tree-complex.c (expand_complex_operations_1): Fix typo.
9258 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
9260 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
9261 (tune_params): New member 'const vec_costs'.
9262 * config/aarch64/aarch64.c (generic_vector_cost): New.
9263 (generic_tunings): New member 'generic_vector_cost'.
9264 (aarch64_builtin_vectorization_cost): New.
9265 (aarch64_add_stmt_cost): New.
9266 (TARGET_VECTORIZE_ADD_STMT_COST): New.
9267 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
9269 2013-07-03 Jakub Jelinek <jakub@redhat.com>
9272 * config/i386/predicates.md (vsib_address_operand): Disallow
9273 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
9275 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
9278 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
9279 expand_builtin_setjmp_receiver.
9280 (expand_label): Adjust, call expand_builtin_setjmp_receiver
9281 with NULL for the label parameter.
9282 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
9283 the frame-pointer. Adjust comments.
9284 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
9285 only if LABEL is non-NULL.
9287 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
9289 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
9290 (ARM_ABI_AAPCS64): Ditto.
9292 (ARM_DEFAULT_ABI): Ditto.
9294 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
9296 * config/aarch64/aarch64-builtins.c
9297 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
9298 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
9300 * config/aarch64/aarch64-simd.md
9301 (aarch64_ld1<VALL:mode>): New.
9302 (aarch64_st1<VALL:mode>): Likewise.
9303 * config/aarch64/arm_neon.h
9304 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
9306 2013-07-02 Sriraman Tallam <tmsriram@google.com>
9308 * config/i386/i386.c (gate_insert_vzeroupper): Check if
9310 (ix86_option_override_internal):Turn on all -mavx target flags by
9311 default as they are dependent on AVX anyway.
9313 2013-07-02 Cary Coutant <ccoutant@google.com>
9315 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
9317 (loc_checksum_ordered): Likewise.
9318 (hash_loc_operands): Remove inline keyword.
9320 2013-07-02 Jakub Jelinek <jakub@redhat.com>
9322 PR tree-optimization/57741
9323 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
9324 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
9325 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
9326 Allow REAL_CST step_exprs if flag_associative_math.
9327 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
9329 2013-07-02 Ian Bolton <ian.bolton@arm.com>
9331 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
9333 2013-07-02 Ian Bolton <ian.bolton@arm.com>
9335 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
9337 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9339 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
9341 (iorsi3_insn): Likewise.
9342 (arm_xorsi3): Likewise.
9344 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
9346 * arm.md (attribute "wtype"): Delete. Move attribute values from here
9348 (attribute "type"): ... here, and prefix with "wmmx_".
9349 (attribute "core_cycles"): Update for attribute changes.
9350 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
9351 (tbcstv4hi): Likewise.
9352 (tbcstv2si): Likewise.
9353 (iwmmxt_iordi3): Likewise.
9354 (iwmmxt_xordi3): Likewise.
9355 (iwmmxt_anddi3): Likewise.
9356 (iwmmxt_nanddi3): Likewise.
9357 (iwmmxt_arm_movdi): Likewise.
9358 (iwmmxt_movsi_insn): Likewise.
9359 (mov<mode>_internal): Likewise.
9360 (and<mode>3_iwmmxt): Likewise.
9361 (ior<mode>3_iwmmxt): Likewise.
9362 (xor<mode>3_iwmmxt): Likewise.
9363 (add<mode>3_iwmmxt): Likewise.
9364 (ssaddv8qi3): Likewise.
9365 (ssaddv4hi3): Likewise.
9366 (ssaddv2si3): Likewise.
9367 (usaddv8qi3): Likewise.
9368 (usaddv4hi3): Likewise.
9369 (usaddv2si3): Likewise.
9370 (sub<mode>3_iwmmxt): Likewise.
9371 (sssubv8qi3): Likewise.
9372 (sssubv4hi3): Likewise.
9373 (sssubv2si3): Likewise.
9374 (ussubv8qi3): Likewise.
9375 (ussubv4hi3): Likewise.
9376 (ussubv2si3): Likewise.
9377 (mulv4hi3_iwmmxt): Likewise.
9378 (smulv4hi3_highpart): Likewise.
9379 (umulv4hi3_highpart): Likewise.
9380 (iwmmxt_wmacs): Likewise.
9381 (iwmmxt_wmacsz): Likewise.
9382 (iwmmxt_wmacu): Likewise.
9383 (iwmmxt_wmacuz): Likewise.
9384 (iwmmxt_clrdi): Likewise.
9385 (iwmmxt_clrv8qi): Likewise.
9386 (iwmmxt_clr4hi): Likewise.
9387 (iwmmxt_clr2si): Likewise.
9388 (iwmmxt_uavgrndv8qi3): Likewise.
9389 (iwmmxt_uavgrndv4hi3): Likewise.
9390 (iwmmxt_uavgv8qi3): Likewise.
9391 (iwmmxt_uavgv4hi3): Likewise.
9392 (iwmmxt_tinsrb): Likewise.
9393 (iwmmxt_tinsrh): Likewise.
9394 (iwmmxt_tinsrw): Likewise.
9395 (iwmmxt_textrmub): Likewise.
9396 (iwmmxt_textrmsb): Likewise.
9397 (iwmmxt_textrmuh): Likewise.
9398 (iwmmxt_textrmsh): Likewise.
9399 (iwmmxt_textrmw): Likewise.
9400 (iwmxxt_wshufh): Likewise.
9401 (eqv8qi3): Likewise.
9402 (eqv4hi3): Likewise.
9403 (eqv2si3): Likewise.
9404 (gtuv8qi3): Likewise.
9405 (gtuv4hi3): Likewise.
9406 (gtuv2si3): Likewise.
9407 (gtv8qi3): Likewise.
9408 (gtv4hi3): Likewise.
9409 (gtv2si3): Likewise.
9410 (smax<mode>3_iwmmxt): Likewise.
9411 (umax<mode>3_iwmmxt): Likewise.
9412 (smin<mode>3_iwmmxt): Likewise.
9413 (umin<mode>3_iwmmxt): Likewise.
9414 (iwmmxt_wpackhss): Likewise.
9415 (iwmmxt_wpackwss): Likewise.
9416 (iwmmxt_wpackdss): Likewise.
9417 (iwmmxt_wpackhus): Likewise.
9418 (iwmmxt_wpackwus): Likewise.
9419 (iwmmxt_wpackdus): Likewise.
9420 (iwmmxt_wunpckihb): Likewise.
9421 (iwmmxt_wunpckihh): Likewise.
9422 (iwmmxt_wunpckihw): Likewise.
9423 (iwmmxt_wunpckilb): Likewise.
9424 (iwmmxt_wunpckilh): Likewise.
9425 (iwmmxt_wunpckilw): Likewise.
9426 (iwmmxt_wunpckehub): Likewise.
9427 (iwmmxt_wunpckehuh): Likewise.
9428 (iwmmxt_wunpckehuw): Likewise.
9429 (iwmmxt_wunpckehsb): Likewise.
9430 (iwmmxt_wunpckehsh): Likewise.
9431 (iwmmxt_wunpckehsw): Likewise.
9432 (iwmmxt_wunpckelub): Likewise.
9433 (iwmmxt_wunpckeluh): Likewise.
9434 (iwmmxt_wunpckeluw): Likewise.
9435 (iwmmxt_wunpckelsb): Likewise.
9436 (iwmmxt_wunpckelsh): Likewise.
9437 (iwmmxt_wunpckelsw): Likewise.
9438 (ror<mode>3): Likewise.
9439 (ashr<mode>3_iwmmxt): Likewise.
9440 (lshr<mode>3_iwmmxt): Likewise.
9441 (ashl<mode>3_iwmmxt): Likewise.
9442 (ror<mode>3_di): Likewise.
9443 (ashr<mode>3_di): Likewise.
9444 (lshr<mode>3_di): Likewise.
9445 (ashl<mode>3_di): Likewise.
9446 (iwmmxt_wmadds): Likewise.
9447 (iwmmxt_wmaddu): Likewise.
9448 (iwmmxt_tmia): Likewise.
9449 (iwmmxt_tmiaph): Likewise.
9450 (iwmmxt_tmiabb): Likewise.
9451 (iwmmxt_tmiatb): Likewise.
9452 (iwmmxt_tmiabt): Likewise.
9453 (iwmmxt_tmiatt): Likewise.
9454 (iwmmxt_tmovmskb): Likewise.
9455 (iwmmxt_tmovmskh): Likewise.
9456 (iwmmxt_tmovmskw): Likewise.
9457 (iwmmxt_waccb): Likewise.
9458 (iwmmxt_wacch): Likewise.
9459 (iwmmxt_waccw): Likewise.
9460 (iwmmxt_waligni): Likewise.
9461 (iwmmxt_walignr): Likewise.
9462 (iwmmxt_walignr0): Likewise.
9463 (iwmmxt_walignr1): Likewise.
9464 (iwmmxt_walignr2): Likewise.
9465 (iwmmxt_walignr3): Likewise.
9466 (iwmmxt_wsadb): Likewise.
9467 (iwmmxt_wsadh): Likewise.
9468 (iwmmxt_wsadbz): Likewise.
9469 (iwmmxt_wsadhz): Likewise.
9470 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
9471 (iwmmxt_wabsdiffb): Likewise.
9472 (iwmmxt_wabsdiffh): Likewise.
9473 (iwmmxt_wabsdiffw): Likewise.
9474 (iwmmxt_waddsubhx): Likewise
9475 (iwmmxt_wsubaddhx): Likewise.
9476 (addc<mode>3): Likewise.
9477 (iwmmxt_avg4): Likewise.
9478 (iwmmxt_avg4r): Likewise.
9479 (iwmmxt_wmaddsx): Likewise.
9480 (iwmmxt_wmaddux): Likewise.
9481 (iwmmxt_wmaddsn): Likewise.
9482 (iwmmxt_wmaddun): Likewise.
9483 (iwmmxt_wmulwsm): Likewise.
9484 (iwmmxt_wmulwum): Likewise.
9485 (iwmmxt_wmulsmr): Likewise.
9486 (iwmmxt_wmulumr): Likewise.
9487 (iwmmxt_wmulwsmr): Likewise.
9488 (iwmmxt_wmulwumr): Likewise.
9489 (iwmmxt_wmulwl): Likewise.
9490 (iwmmxt_wqmulm): Likewise.
9491 (iwmmxt_wqmulwm): Likewise.
9492 (iwmmxt_wqmulmr): Likewise.
9493 (iwmmxt_wqmulwmr): Likewise.
9494 (iwmmxt_waddbhusm): Likewise.
9495 (iwmmxt_waddbhusl): Likewise.
9496 (iwmmxt_wqmiabb): Likewise.
9497 (iwmmxt_wqmiabt): Likewise.
9498 (iwmmxt_wqmiatb): Likewise.
9499 (iwmmxt_wqmiatt): Likewise.
9500 (iwmmxt_wqmiabbn): Likewise.
9501 (iwmmxt_wqmiabtn): Likewise.
9502 (iwmmxt_wqmiatbn): Likewise.
9503 (iwmmxt_wqmiattn): Likewise.
9504 (iwmmxt_wmiabb): Likewise.
9505 (iwmmxt_wmiabt): Likewise.
9506 (iwmmxt_wmiatb): Likewise.
9507 (iwmmxt_wmiatt): Likewise.
9508 (iwmmxt_wmiabbn): Likewise.
9509 (iwmmxt_wmiabtn): Likewise.
9510 (iwmmxt_wmiatbn): Likewise.
9511 (iwmmxt_wmiattn): Likewise.
9512 (iwmmxt_wmiawbb): Likewise.
9513 (iwmmxt_wmiawbt): Likewise.
9514 (iwmmxt_wmiawtb): Likewise.
9515 (iwmmxt_wmiawtt): Likewise.
9516 (iwmmxt_wmiawbbn): Likewise.
9517 (iwmmxt_wmiawbtn): Likewise.
9518 (iwmmxt_wmiawtbn): Likewise.
9519 (iwmmxt_wmiawttn): Likewise.
9520 (iwmmxt_wmerge): Likewise.
9521 (iwmmxt_tandc<mode>3): Likewise.
9522 (iwmmxt_torc<mode>3): Likewise.
9523 (iwmmxt_torvsc<mode>3): Likewise.
9524 (iwmmxt_textrc<mode>3): Likewise.
9525 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
9526 (wmmxt_pack): Likewise.
9527 (wmmxt_mult_c1): Likewise.
9528 (wmmxt_mult_c2): Likewise.
9529 (wmmxt_alu_c1): Likewise.
9530 (wmmxt_alu_c2): Likewise.
9531 (wmmxt_alu_c3): Likewise.
9532 (wmmxt_transfer_c1): Likewise.
9533 (wmmxt_transfer_c2): Likewise.
9534 (wmmxt_transfer_c3): Likewise.
9535 (marvell_f_iwmmxt_wstr): Likewise.
9536 (marvell_f_iwmmxt_wldr): Likewise.
9538 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
9540 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
9542 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
9545 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
9546 * lra-constraints.c (need_for_split_p): Check call used hard regs
9547 living through calls.
9549 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
9550 call used regs for call insn.
9552 2013-06-28 Jakub Jelinek <jakub@redhat.com>
9555 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
9556 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
9558 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
9560 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
9562 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
9564 * lra-constraints.c (need_for_split_p): Check call used hard regs
9565 living through calls.
9567 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
9570 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
9571 to tie with any other modes. Eliminate Altivec vector mode tests,
9572 since these are a subset of ALTIVEC or VSX vector modes. Simplify
9573 code, to return 0 if testing MODE2 for a condition, if we've
9574 already tested MODE1 for the same condition.
9576 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
9578 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
9581 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
9583 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
9584 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
9586 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
9588 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
9590 (aarch64_symbolic_constant_p): Remove.
9591 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
9592 static. Fix line length and white space.
9593 (aarch64_symbolic_constant_p): Remove.
9594 * config/aarch64/predicates.md (aarch64_valid_symref):
9595 Use aarch64_classify_symbol_expression.
9597 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9599 * config/arm/constraints.md (Ts): New constraint.
9600 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
9602 (compare_scc): Use "Ts" constraint for operand 0.
9603 (ior_scc_scc): Likewise.
9604 (and_scc_scc): Likewise.
9605 (and_scc_scc_nodom): Likewise.
9606 (ior_scc_scc_cmp): Likewise for operand 7.
9607 (and_scc_scc_cmp): Likewise.
9608 * config/arm/thumb2.md (thumb2_movsi_insn):
9609 Add alternatives for 16-bit encodings.
9610 (thumb2_movhi_insn): Likewise.
9611 (thumb2_movsicc_insn): Likewise.
9612 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
9613 (thumb2_negscc): Use "Ts" constraint.
9614 Move mvn instruction outside cond_exec block.
9615 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
9616 for 16-bit encodings.
9618 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9620 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
9622 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
9623 (mulsi3subsi): Likewise.
9624 (mulsidi3adddi): Likewise.
9625 (mulsidi3_v6): Likewise.
9626 (umulsidi3_v6): Likewise.
9627 (umulsidi3adddi_v6): Likewise.
9628 (smulsi3_highpart_v6): Likewise.
9629 (umulsi3_highpart_v6): Likewise.
9630 (mulhisi3tb): Likewise.
9631 (mulhisi3bt): Likewise.
9632 (mulhisi3tt): Likewise.
9633 (maddhisi4): Likewise.
9634 (maddhisi4tb): Likewise.
9635 (maddhisi4tt): Likewise.
9636 (maddhidi4): Likewise.
9637 (maddhidi4tb): Likewise.
9638 (maddhidi4tt): Likewise.
9639 (zeroextractsi_compare0_scratch): Likewise.
9640 (insv_zero): Likewise.
9641 (insv_t2): Likewise.
9642 (anddi_notzesidi_di): Likewise.
9643 (anddi_notsesidi_di): Likewise.
9644 (andsi_notsi_si): Likewise.
9645 (iordi_zesidi_di): Likewise.
9646 (xordi_zesidi_di): Likewise.
9647 (andsi_iorsi3_notsi): Likewise.
9649 (smax_m1): Likewise.
9651 (not_shiftsi): Likewise.
9652 (unaligned_loadsi): Likewise.
9653 (unaligned_loadhis): Likewise.
9654 (unaligned_loadhiu): Likewise.
9655 (unaligned_storesi): Likewise.
9656 (unaligned_storehi): Likewise.
9657 (extv_reg): Likewise.
9658 (extzv_t2): Likewise.
9660 (udivsi3): Likewise.
9661 (arm_zero_extendhisi2addsi): Likewise.
9662 (arm_zero_extendqisi2addsi): Likewise.
9663 (compareqi_eq0): Likewise.
9664 (arm_extendhisi2_v6): Likewise.
9665 (arm_extendqisi2addsi): Likewise.
9666 (arm_movt): Likewise.
9667 (thumb2_ldrd): Likewise.
9668 (thumb2_ldrd_base): Likewise.
9669 (thumb2_ldrd_base_neg): Likewise.
9670 (thumb2_strd): Likewise.
9671 (thumb2_strd_base): Likewise.
9672 (thumb2_strd_base_neg): Likewise.
9673 (arm_negsi2): Add alternative for 16-bit encoding.
9674 (arm_one_cmplsi2): Likewise.
9676 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9678 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
9679 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
9680 (movdfcc): Likewise.
9681 * config/arm/vfp.md (*thumb2_movsf_vfp):
9682 Disable predication for arm_restrict_it.
9683 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
9684 (*thumb2_movdfcc_vfp): Likewise.
9685 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
9686 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
9687 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
9688 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
9689 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
9690 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
9691 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
9692 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
9693 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
9694 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
9695 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
9696 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
9697 Disable predication for arm_restrict_it.
9699 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
9701 * config/i386/bmiintrin.h (_bextr_u32): New.
9702 (_bextr_u64): Ditto.
9704 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
9706 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
9707 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
9708 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
9709 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
9710 * config/mips/n32-elf.h: ...this new file.
9712 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
9715 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
9716 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
9718 2013-06-27 Catherine Moore <clm@codesourcery.com>
9720 * config/mips/mips-tables.opt: Regenerate.
9721 * config/mips/mips-cpus.def: Add m14ke and m14kec.
9722 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
9723 * doc/invoke.texi: Add -m14kc.
9725 2013-06-27 Jakub Jelinek <jakub@redhat.com>
9728 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
9729 constraints of operand 1 and 2.
9732 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
9733 to match RTL canonicalization. Swap predicates and
9734 constraints of operand 1 and 2.
9736 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
9738 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
9739 Process OP_INOUT regs for splitting too.
9741 2013-06-27 Jakub Jelinek <jakub@redhat.com>
9743 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
9744 decl before the loop, initialize to NULL.
9745 (vectorizable_load): Initialize ptr_incr to NULL.
9747 2013-06-27 Martin Jambor <mjambor@suse.cz>
9750 * ipa-ref.h (ipa_maybe_record_reference): Declare.
9751 * ipa-ref.c (ipa_maybe_record_reference): New function.
9752 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
9753 * ipa-cp.c (create_specialized_node): Record potential references from
9755 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
9757 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
9759 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
9760 parameter 'mode' of type 'enum machine_mode mode'; change to pass
9761 'mode' to force_reg.
9762 (aarch64_add_offset): Update calls to aarch64_force_temporary.
9763 (aarch64_expand_mov_immediate): Likewise.
9765 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
9767 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
9768 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
9770 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9772 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
9773 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
9774 (struct machine_function): Add tbegin_p.
9775 (s390_canonicalize_comparison): Fold CC mode compares to
9776 conditional jump if possible.
9777 (s390_emit_jump): Return the emitted jump.
9778 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
9779 Handle CCRAWmode compares.
9780 (s390_option_override): Default to -mhtm if available.
9781 (s390_reg_clobbered_rtx): Handle floating point regs as well.
9782 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
9783 FPRs instead of df_regs_ever_live_p.
9784 (s390_optimize_nonescaping_tx): New function.
9785 (s390_init_frame_layout): Extend clobbered_regs array to cover
9787 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
9788 (s390_expand_tbegin): New function.
9789 (enum s390_builtin): New enum definition.
9790 (code_for_builtin): New array definition.
9791 (s390_init_builtins): New function.
9792 (s390_expand_builtin): New function.
9793 (TARGET_INIT_BUILTINS): Define.
9794 (TARGET_EXPAND_BUILTIN): Define.
9795 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
9796 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
9797 (s390_alc_comparison): Likewise.
9798 * config/s390/s390-modes.def: Add CCRAWmode.
9799 * config/s390/s390.h (processor_flags): Add PF_TX.
9800 (TARGET_CPU_HTM): Define macro.
9801 (TARGET_HTM): Define macro.
9802 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
9803 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
9804 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
9805 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
9806 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
9807 (TBEGIN_MASK, TBEGINC_MASK): New constants.
9808 ("*cc_to_int"): Move up.
9809 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
9810 constants other than 0.
9811 ("*ccraw_to_int"): New insn and splitter definition.
9812 ("tbegin", "tbegin_nofloat", "tbegin_retry")
9813 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
9814 ("tx_assist"): New expander.
9815 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
9816 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
9817 * config/s390/s390.opt: Add -mhtm option.
9818 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
9819 * config/s390/htmxlintrin.h: New file.
9820 * config/s390/htmintrin.h: New file.
9821 * config/s390/s390intrin.h: New file.
9822 * doc/extend.texi: Document htm builtins.
9823 * config.gcc: Add the new header files to extra_headers.
9825 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
9827 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
9828 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
9830 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
9831 Pat Haugen <pthaugen@us.ibm.com>
9832 Peter Bergner <bergner@vnet.ibm.com>
9834 * config/rs6000/power8.md: New.
9835 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
9836 setting for power8 entry.
9837 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
9838 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
9839 test for Power4/Power5 only.
9840 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
9842 (force_new_group): Adjust comment.
9843 * config/rs6000/rs6000.md: Include power8.md.
9845 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
9847 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
9848 * config/arm/arm-protos.h (arm_max_conditional_execute): New
9850 (tune_params): Update comment.
9851 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
9852 (arm_max_conditional_execute): New function.
9853 (thumb2_final_prescan_insn): Use max_insn_skipped and
9854 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
9856 2013-06-25 Jakub Jelinek <jakub@redhat.com>
9858 PR tree-optimization/57705
9859 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
9860 SSA_NAME step, provided that it is not defined inside the loop.
9861 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
9862 (get_initial_def_for_induction): Handle SSA_NAME IV step.
9864 2013-06-25 Martin Jambor <mjambor@suse.cz>
9867 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
9868 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
9870 (ipa_note_param_call): Initialize member_ptr flag.
9871 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
9872 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
9873 (ipa_write_indirect_edge_info): Stream member_ptr flag.
9874 (ipa_read_indirect_edge_info): Likewise.
9876 2013-06-25 Richard Biener <rguenther@suse.de>
9879 * passes.c (init_optimization_passes): Move pass_fold_builtins
9880 and pass_dce earlier with -Og.
9882 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
9884 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
9885 <BIT_FIELD_REF>: Remove trailing TAB.
9886 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
9889 2013-06-24 Martin Jambor <mjambor@suse.cz>
9891 PR tree-optimization/57358
9892 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
9893 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
9894 (ipa_analyze_params_uses): Generate pessimistic info when true.
9896 2013-06-24 Martin Jambor <mjambor@suse.cz>
9898 PR tree-optimization/57539
9899 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
9900 global.inlined_to of the new node to it. All callers changed.
9901 * ipa-inline-transform.c (clone_inlined_nodes): New variable
9902 inlining_into, pass it to cgraph_clone_node.
9903 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
9904 ipa_free_edge_args_substructures.
9905 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
9906 rdesc linked list. Do not assert rdesc edges have inlined caller.
9907 Assert we have found an rdesc in the rdesc list.
9909 2013-06-24 Richard Biener <rguenther@suse.de>
9911 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
9912 (pointer_set_lookup): Declare.
9913 (class pointer_map): New template class implementing a
9914 generic pointer to T map.
9915 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
9916 pointer_map<T>::contains, pointer_map<T>::insert,
9917 pointer_map<T>::traverse): New functions.
9918 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
9919 (pointer_set_lookup): New function.
9920 (pointer_set_contains): Use pointer_set_lookup.
9921 (pointer_set_insert): Likewise.
9922 (insert_aux): Remove.
9923 (struct pointer_map_t): Embed a pointer_set_t.
9924 (pointer_map_create): Adjust.
9925 (pointer_map_destroy): Likewise.
9926 (pointer_map_contains): Likewise.
9927 (pointer_map_insert): Likewise.
9928 (pointer_map_traverse): Likewise.
9929 * tree-streamer.h (struct streamer_tree_cache_d): Use a
9930 pointer_map<unsigned> instead of a pointer_map_t.
9931 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
9932 (streamer_tree_cache_lookup): Likewise.
9933 (streamer_tree_cache_create): Likewise.
9934 (streamer_tree_cache_delete): Likewise.
9935 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
9936 pointer_map<unsigned> instead of a pointer_map_t.
9937 (lto_init_tree_ref_encoder): Adjust.
9938 (lto_destroy_tree_ref_encoder): Likewise.
9939 * lto-section-out.c (lto_output_decl_index): Likewise.
9940 (lto_record_function_out_decl_state): Likewise.
9941 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
9943 2013-06-24 Richard Biener <rguenther@suse.de>
9945 PR tree-optimization/57488
9946 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
9948 2013-06-24 Alan Modra <amodra@gmail.com>
9950 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
9951 (gen_easy_altivec_constant): Likewise.
9952 * config/rs6000/predicates.md (easy_vector_constant_add_self,
9953 easy_vector_constant_msb): Likewise.
9955 2013-06-23 Jakub Jelinek <jakub@redhat.com>
9958 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
9959 add missing return true.
9961 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
9964 * config/sh/predicates.md (general_extend_operand): Invoke
9965 general_movsrc_operand for memory operands.
9966 (general_movsrc_operand): Allow reg+reg addressing, do not use
9967 general_operand for memory operands.
9969 2013-06-23 Sriraman Tallam <tmsriram@google.com>
9971 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
9972 when current target options does not apply.
9973 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
9974 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
9975 * config/i386/bmiintrin.h: Pass appropriate target
9976 attributes to header.
9977 * config/i386/mmintrin.h: Ditto.
9978 * config/i386/nmmintrin.h: Ditto.
9979 * config/i386/avx2intrin.h: Ditto.
9980 * config/i386/fxsrintrin.h: Ditto.
9981 * config/i386/tbmintrin.h: Ditto.
9982 * config/i386/xsaveintrin.h: Ditto.
9983 * config/i386/f16cintrin.h: Ditto.
9984 * config/i386/xtestintrin.h: Ditto.
9985 * config/i386/xsaveoptintrin.h: Ditto.
9986 * config/i386/bmi2intrin.h: Ditto.
9987 * config/i386/lzcntintrin.h: Ditto.
9988 * config/i386/smmintrin.h: Ditto.
9989 * config/i386/wmmintrin.h: Ditto.
9990 * config/i386/x86intrin.h: Remove all header include guards.
9991 * config/i386/prfchwintrin.h: Ditto.
9992 * config/i386/pmmintrin.h: Ditto.
9993 * config/i386/tmmintrin.h: Ditto.
9994 * config/i386/xmmintrin.h: Ditto.
9995 * config/i386/popcntintrin.h: Ditto.
9996 * config/i386/rdseedintrin.h: Ditto.
9997 * config/i386/ammintrin.h: Ditto.
9998 * config/i386/emmintrin.h: Ditto.
9999 * config/i386/immintrin.h: Remove all header include guards.
10000 * config/i386/fma4intrin.h: Ditto.
10001 * config/i386/lwpintrin.h: Ditto.
10002 * config/i386/xopintrin.h: Ditto.
10003 * config/i386/ia32intrin.h: Ditto.
10004 * config/i386/avxintrin.h: Ditto.
10005 * config/i386/rtmintrin.h: Ditto.
10006 * config/i386/fmaintrin.h: Ditto.
10007 * config/i386/mm3dnow.h: Ditto.
10009 2013-06-22 Sriraman Tallam <tmsriram@google.com>
10011 * common/config/i386/i386-common.c: Handle LZCNT.
10013 2013-06-22 Andi Kleen <ak@linux.intel.com>
10015 * doc/extend.texi: Use __atomic_store_n instead of
10016 __atomic_store in HLE example.
10018 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
10020 * config/sh/sh.c: Remove <cstdlib> workaround.
10022 2013-06-21 Andi Kleen <ak@linux.intel.com>
10024 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
10026 2013-06-21 Andi Kleen <ak@linux.intel.com>
10028 * doc/extend.texi: Document that __atomic_clear and
10029 __atomic_test_and_set should only be used with bool.
10031 2013-06-20 Jan Hubicka <jh@suse.cz>
10033 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
10034 types_same_for_odr.
10035 * tree.c (decls_same_for_odr): New function.
10036 (same_for_edr): New function.
10037 (types_same_for_odr): New function.
10038 (get_binfo_at_offset): Use it.
10039 * tree.h (types_same_for_odr): Declare.
10041 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
10042 Jason Merrill <jason@redhat.com>
10044 * system.h: Include <cstdlib> as well as <stdlib.h>.
10046 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
10049 * config/i386/i386.c (construct_container): Report error if
10050 long double is used with disabled x87 float returns.
10052 2013-06-20 Jan Hubicka <jh@suse.cz>
10054 * lto-cgraph.c (input_symtab): Do not set cgraph state.
10056 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
10058 PR rtl-optimization/57425
10059 PR rtl-optimization/57569
10060 * alias.c (write_dependence_p): Remove parameters mem_mode and
10061 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
10062 Changed all callers.
10063 (canon_anti_dependence): Get comments and semantics in sync.
10064 Add parameter mem_canonicalized. Changed all callers.
10065 * rtl.h (canon_anti_dependence): Update prototype.
10067 2013-06-20 Richard Biener <rguenther@suse.de>
10069 * data-streamer-in.c (streamer_read_uhwi): Optimize single
10070 byte case, inline streamer_read_uchar and defer section
10073 2013-06-20 Richard Biener <rguenther@suse.de>
10075 PR tree-optimization/57584
10076 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
10077 SSA names into the expanded expression that take part in
10078 abnormal coalescing.
10080 2013-06-19 Sharad Singhai <singhai@google.com>
10082 * gcov.c (print_usage): Handle new option.
10083 (process_args): Ditto.
10084 (get_gcov_intermediate_filename): New function.
10085 (output_intermediate_file): New function.
10086 (output_gcov_file): New function
10087 (generate_results): Handle new option.
10088 (release_function): Relase demangled name.
10089 (read_graph_file): Handle demangled name.
10090 (output_lines): Ditto.
10091 * doc/gcov.texi: Document gcov intermediate format.
10093 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
10096 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
10097 (lra_emit_add): Use the functions. Add comment about Y as an
10100 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
10103 * collect2.c (collect_atexit): New.
10104 (collect_exit): Delete.
10105 (main): Register collect_atexit with atexit.
10106 (collect_wait): Change collect_exit to exit.
10108 * collect2.h (collect_exit): Delete.
10109 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
10111 2013-06-19 Wei Mi <wmi@google.com>
10113 PR rtl-optimization/57518
10114 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
10115 if regno is used in paradoxical subreg.
10116 (update_equiv_regs): Check pdx_subregs[regno] before
10117 set a reg to be equivalent with a mem.
10119 2013-06-19 Matthias Klose <doko@ubuntu.com>
10122 * file-find.h (find_a_file): Add a mode parameter.
10123 * file-find.c (find_a_file): Likewise.
10124 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
10125 with X_OK for the executables.
10126 * collect2.c (main): Call find_a_file with X_OK.
10128 2013-06-19 Steve Ellcey <sellcey@mips.com>
10131 * config/mips/mips.md (casesi_internal_mips16_<mode>):
10132 Use NEXT_INSN instead of next_real_insn.
10134 2013-06-19 Jan Hubicka <jh@suse.cz>
10136 * cgraph.h (const_value_known_p): Replace by ...
10137 (ctor_for_folding): .. this one.
10138 * cgraphunit.c (process_function_and_variable_attributes): Use it.
10139 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
10140 * expr.c (expand_expr_real_1): Likewise.
10141 (string_constant): Likewise.
10142 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
10143 * ipa.c (process_references): Likewise.
10144 (symtab_remove_unreachable_nodes): Likewise.
10145 * ipa-inline-analysis.c (param_change_prob): Likewise.
10146 * gimple-fold.c (canonicalize_constructor_val): Likewise.
10147 (get_base_constructor): Likwise.
10148 * varpool.c (varpool_remove_node): Likewise.
10149 (varpool_remove_initializer): LIkewise.
10150 (dump_varpool_node): LIkwise.
10151 (const_value_known_p): Rewrite to ...
10152 (ctor_for_folding): ... this one.
10154 2013-06-19 Jakub Jelinek <jakub@redhat.com>
10157 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
10158 PERSONALITY in $PATH derived prefixes.
10160 2013-06-19 Jeff Law <law@redhat.com>
10162 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
10165 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
10166 (simplify_bitwise_binary): Use it to simpify certain binary ops on
10169 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
10171 * config/arm/vfp.md: Move VFP instruction classification documentation
10173 * config/arm/arm.md: ... here. Update instruction classification
10176 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
10178 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
10179 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
10180 pattern. Use more efficient sequences on ARMv5 and Thumb2.
10182 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
10185 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
10186 with NEXT_INSN. Use tablejump_p to check for jump table data
10189 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
10192 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
10193 that now in C++ the value is correct per the C++ standards.
10195 2013-06-19 Richard Biener <rguenther@suse.de>
10197 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
10198 for global context.
10200 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10203 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10206 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
10207 with next_active_insn.
10209 2013-06-18 Sriraman Tallam <tmsriram@google.com>
10211 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
10212 functions are inlined during failures to flag an error.
10213 * tree-inline.c (expand_call_inline): Allow the error to be flagged
10214 in early inline pass.
10216 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
10218 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
10221 2013-06-18 Julian Brown <julian@codesourcery.com>
10223 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
10224 Permit virtual register pre-reload if !strict.
10225 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
10227 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
10229 * config/arm/neon.md (movmisalign<mode>): Use
10230 neon_perm_struct_or_reg_operand instead of
10231 neon_struct_or_register_operand.
10232 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
10233 neon_permissive_struct_operand instead of neon_struct_operand.
10234 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
10235 neon_vector_mem_operand.
10236 * config/arm/predicates.md (neon_struct_operand): Adjust call to
10237 neon_vector_mem_operand.
10238 (neon_permissive_struct_operand): New.
10239 (neon_struct_or_register_operand): Rename to...
10240 (neon_perm_struct_or_reg_operand): This. Adjust call to
10241 neon_vector_mem_operand.
10243 2013-06-18 Richard Biener <rguenther@suse.de>
10245 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
10246 * lto-streamer.h: Include pointer-set.h.
10247 (struct lto_decl_slot): Remove.
10248 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
10249 Remove next_index entry.
10250 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
10251 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
10252 (lto_init_tree_ref_encoder): Adjust.
10253 (lto_destroy_tree_ref_encoder): Likewise.
10254 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
10255 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
10256 (lto_output_decl_index): Adjust.
10257 (lto_new_out_decl_state): Likewise.
10258 (lto_record_function_out_decl_state): Likewise.
10259 * lto-streamer-out.c (copy_function): Likewise.
10261 2013-06-18 Richard Biener <rguenther@suse.de>
10263 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
10264 * cgraphunit.c: Include cfgloop.h.
10265 (init_lowered_empty_function): Initialize the loop tree.
10266 (assemble_thunk): Insert new BBs into loops.
10268 2013-06-18 Richard Biener <rguenther@suse.de>
10270 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
10271 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
10272 the map from cache entry to cache index optional.
10273 (streamer_tree_cache_replace_tree): Adjust accordingly.
10274 (streamer_tree_cache_append): Likewise.
10275 (streamer_tree_cache_delete): Likewise.
10276 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
10277 streamer cache map from cache entry to cache index.
10278 * lto-streamer-out.c (create_output_block): Adjust.
10280 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
10282 * config/arm/arm.md (attribute "insn"): Move multiplication and
10283 division attributes to...
10284 (attribute "type"): ... here. Remove mult.
10285 (attribute "mul32"): New attribute.
10286 (attribute "mul64"): Add umaal.
10287 (*arm_mulsi3): Update attributes.
10288 (*arm_mulsi3_v6): Likewise.
10289 (*thumb_mulsi3): Likewise.
10290 (*thumb_mulsi3_v6): Likewise.
10291 (*mulsi3_compare0): Likewise.
10292 (*mulsi3_compare0_v6): Likewise.
10293 (*mulsi_compare0_scratch): Likewise.
10294 (*mulsi_compare0_scratch_v6): Likewise.
10295 (*mulsi3addsi): Likewise.
10296 (*mulsi3addsi_v6): Likewise.
10297 (*mulsi3addsi_compare0): Likewise.
10298 (*mulsi3addsi_compare0_v6): Likewise.
10299 (*mulsi3addsi_compare0_scratch): Likewise.
10300 (*mulsi3addsi_compare0_scratch_v6): Likewise.
10301 (*mulsi3subsi): Likewise.
10302 (*mulsidi3adddi): Likewise.
10303 (*mulsi3addsi_v6): Likewise.
10304 (*mulsidi3adddi_v6): Likewise.
10305 (*mulsidi3_nov6): Likewise.
10306 (*mulsidi3_v6): Likewise.
10307 (*umulsidi3_nov6): Likewise.
10308 (*umulsidi3_v6): Likewise.
10309 (*umulsidi3adddi): Likewise.
10310 (*umulsidi3adddi_v6): Likewise.
10311 (*smulsi3_highpart_nov6): Likewise.
10312 (*smulsi3_highpart_v6): Likewise.
10313 (*umulsi3_highpart_nov6): Likewise.
10314 (*umulsi3_highpart_v6): Likewise.
10315 (mulhisi3): Likewise.
10316 (*mulhisi3tb): Likewise.
10317 (*mulhisi3bt): Likewise.
10318 (*mulhisi3tt): Likewise.
10319 (maddhisi4): Likewise.
10320 (*maddhisi4tb): Likewise.
10321 (*maddhisi4tt): Likewise.
10322 (maddhidi4): Likewise.
10323 (*maddhidi4tb): Likewise.
10324 (*maddhidi4tt): Likewise.
10325 (divsi3): Likewise.
10326 (udivsi3): Likewise.
10327 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
10328 (thumb2_mulsi_short_compare0): Likewise.
10329 (thumb2_mulsi_short_compare0_scratch): Likewise.
10330 * config/arm/arm1020e.md (1020mult1): Update attribute change.
10331 (1020mult2): Likewise.
10332 (1020mult3): Likewise.
10333 (1020mult4): Likewise.
10334 (1020mult5): Likewise.
10335 (1020mult6): Likewise.
10336 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
10338 (cortex_a15_mult64): Likewise.
10339 (cortex_a15_sdiv): Likewise.
10340 (cortex_a15_udiv): Likewise.
10341 * config/arm/arm1026ejs.md (mult1): Update attribute change.
10347 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
10348 (pj4_ir_div): Likewise.
10349 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
10350 (11_mult2): Likewise.
10351 (11_mult3): Likewise.
10352 (11_mult4): Likewise.
10353 (11_mult5): Likewise.
10354 (11_mult6): Likewise.
10355 (11_mult7): Likewise.
10356 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
10357 (cortex_a8_mla): Likewise.
10358 (cortex_a8_mull): Likewise.
10359 (cortex_a8_smulwy): Likewise.
10360 (cortex_a8_smlald): Likewise.
10361 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
10362 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
10363 (cortex_r4_mul_3): Likewise.
10364 (cortex_r4_mla_4): Likewise.
10365 (cortex_r4_mla_3): Likewise.
10366 (cortex_r4_smlald): Likewise.
10367 (cortex_r4_mull): Likewise.
10368 (cortex_r4_sdiv): Likewise.
10369 (cortex_r4_udiv): Likewise.
10370 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
10371 (cortex_a7_idiv): Likewise.
10372 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
10373 (9_mult2): Likewise.
10374 (9_mult3): Likewise.
10375 (9_mult4): Likewise.
10376 (9_mult5): Likewise.
10377 (9_mult6): Likewise.
10378 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
10379 (cortex_a53_sdiv): Likewise.
10380 (cortex_a53_udiv): Likewise.
10381 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
10382 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
10383 (mp626_mult2): Likewise.
10384 (mp626_mult3): Likewise.
10385 (mp626_mult4): Likewise.
10386 * config/arm/fa526.md (526_mult1): Update attribute change.
10387 (526_mult2): Likewise.
10388 * config/arm/arm-generic.md (mult): Update attribute change.
10389 (mult_ldsched_strongarm): Likewise.
10390 (mult_ldsched): Likewise.
10391 (multi_cycle): Likewise.
10392 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
10393 * config/arm/fa606te.md (606te_mult1): Update attribute change.
10394 (606te_mult2): Likewise.
10395 (606te_mult3): Likewise.
10396 (606te_mult4): Likewise.
10397 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
10398 (cortex_a9_mac16): Likewise.
10399 (cortex_a9_multiply): Likewise.
10400 (cortex_a9_mac): Likewise.
10401 (cortex_a9_multiply_long): Likewise.
10402 * config/arm/fa626te.md (626te_mult1): Update attribute change.
10403 (626te_mult2): Likewise.
10404 (626te_mult3): Likewise.
10405 (626te_mult4): Likewise.
10407 2013-06-18 Richard Biener <rguenther@suse.de>
10410 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
10412 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10415 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
10416 with next_active_insn.
10418 2013-06-18 Alan Modra <amodra@gmail.com>
10420 * config/rs6000/rs6000.h (enum data_align): New.
10421 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
10422 (DATA_ABI_ALIGNMENT): Define.
10423 (CONSTANT_ALIGNMENT): Correct comment.
10424 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
10425 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
10427 2013-06-17 David Malcolm <dmalcolm@redhat.com>
10429 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
10430 ATTRIBUTE_UNUSED marking.
10432 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
10434 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
10435 alternative and update.
10436 (aarch64_dup_lanedi): Delete.
10437 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
10438 * config/aarch64/aarch64-simd-builtins.def: Update.
10440 2013-06-17 Richard Biener <rguenther@suse.de>
10442 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
10443 (lto_input_scc): Declare.
10444 (lto_input_tree_1): Likewise.
10445 (struct lto_stats_d): Add num_tree_bodies_output and
10446 num_pickle_refs_output.
10447 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
10448 (lto_read_tree_1): Split out from ...
10449 (lto_read_tree): ... this.
10450 (lto_input_scc): New function.
10451 (lto_input_tree_1): Split out from ...
10452 (lto_input_tree): ... this. Handle LTO_tree_scc.
10453 (lto_data_in_create): Create the streamer cache without hashes.
10454 * lto-streamer-out.c (create_output_block): Create the streamer
10455 cache with hashes when not doing WPA.
10456 (lto_write_tree_1): Split out from ...
10457 (lto_write_tree): ... this.
10458 (get_symbol_initial_value): New function.
10459 (lto_output_tree_1): Split out from ...
10460 (lto_output_tree): ... this. Write trees as series of SCCs
10461 using a DFS walk via DFS_write_tree.
10462 (struct sccs, struct scc_entry): New types.
10463 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
10464 (DFS_write_tree_body): New function.
10465 (DFS_write_tree): Likewise.
10466 (hash_tree): Likewise.
10467 (scc_entry_compare): Likewise.
10468 (hash_scc): Likewise.
10469 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
10470 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
10471 TREE_CHAIN as regular reference.
10472 (streamer_read_integer_cst): Remove.
10473 (streamer_get_pickled_tree): Adjust.
10474 * tree-streamer-out.c (streamer_write_chain): Disable streaming
10475 of DECL_EXTERNALs in BLOCK_VARS for now.
10476 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
10478 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
10479 Add hash value argument and record that if hashes are recorded
10481 (streamer_tree_cache_insert_1): Adjust.
10482 (streamer_tree_cache_insert): Likewise.
10483 (streamer_tree_cache_insert_at): Rename to ...
10484 (streamer_tree_cache_replace_tree): ... this and adjust.
10485 (streamer_tree_cache_append): Adjust.
10486 (record_common_node): Likewise.
10487 (streamer_tree_cache_create): Add argument whether to
10488 record hash values together with trees.
10489 (streamer_tree_cache_delete): Adjust.
10490 * tree-streamer.h (struct streamer_tree_cache_d): Add
10492 (streamer_read_integer_cst): Remove.
10493 (streamer_tree_cache_insert): Adjust.
10494 (streamer_tree_cache_append): Likewise.
10495 (streamer_tree_cache_insert_at): Rename to ...
10496 (streamer_tree_cache_replace_tree): ... this and adjust.
10497 (streamer_tree_cache_create): Add argument whether to record hashes.
10498 (streamer_tree_cache_get): Rename to ...
10499 (streamer_tree_cache_get_tree): ... this.
10500 (streamer_tree_cache_get_hash): New function.
10501 * tree.c (cache_integer_cst): New function.
10502 * tree.h (cache_integer_cst): Declare.
10503 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
10504 * lto-symtab.c (lto_varpool_replace_node): Only release
10505 DECL_INITIAL of non-prevailing decls.
10506 * varpool.c (varpool_remove_initializer): Do not release
10507 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
10509 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
10511 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
10512 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
10513 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
10514 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
10515 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
10516 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
10517 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
10518 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
10519 instead of TARGET_64BIT.
10520 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
10521 Require ISA_HAS_<D>DIV.
10523 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
10525 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
10526 (mips*-*-linux*): Move default with_llsc setting to where other
10528 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
10530 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
10531 Likewise. Remove default with_tune setting. Move default float
10532 setting to its own block. Handle with_llsc in the same block as above.
10534 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
10536 PR rtl-optimization/57425
10537 PR rtl-optimization/57569
10538 * alias.c (write_dependence_p): Add new parameters mem_mode,
10539 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
10540 Changed all callers.
10541 (canon_anti_dependence): New function.
10542 * cse.c (check_dependence): Use canon_anti_dependence.
10543 * cselib.c (cselib_invalidate_mem): Likewise.
10544 * rtl.h (canon_anti_dependence): Declare.
10546 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
10548 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
10549 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
10550 ".set mips3" for 64-bit targets.
10552 2013-06-15 Dehao Chen <dehao@google.com>
10554 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
10555 * gimple-low.c (gimple_check_call_matching_types): Likewise.
10556 (gimple_check_call_args): Likewise.
10557 * value-prof.c (check_ic_target): Likewise.
10558 * ipa-inline.c (early_inliner): Likewise.
10559 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
10560 * cgraph.c (cgraph_create_edge_1): Likewise.
10561 (cgraph_make_edge_direct): Likewise.
10563 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
10566 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
10567 rs6000_output_move_128bit to handle emitting quad memory
10568 operations. Set attribute length to 8 bytes.
10570 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
10572 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
10574 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
10575 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
10576 (aarch64_<su>mlsl<mode>): Likewise.
10578 2013-06-14 Mike Stump <mikestump@comcast.net>
10580 * Makefile.in (TARGET_H): Add insn-codes.h.
10582 2013-06-14 Alan Modra <amodra@gmail.com>
10584 PR middle-end/57134
10585 PR middle-end/57586
10586 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
10587 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
10588 bitfield expansion when EXPAND_MEMORY.
10589 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
10591 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
10593 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
10594 test for clearing quad memory on 32-bit later.
10596 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
10598 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
10599 (fold_negate_expr): Likewise.
10600 (fold_real_zero_addition_p): Handle vectors.
10601 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
10603 2013-06-14 Alan Modra <amodra@gmail.com>
10605 * varasm.c (force_const_mem): Revert 2013-06-07 change.
10607 2013-06-13 Jan Hubicka <jh@suse.cz>
10609 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
10610 Local comdats are not externally visible.
10611 * symtab.c (dump_symtab_base): Dump externally visible.
10612 (verify_symtab_base): Verify back links in the symtab hash.
10614 2013-06-13 Bin Cheng <bin.cheng@arm.com>
10616 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
10617 CONVERT_EXPR as equal nodes.
10619 2013-06-13 Bin Cheng <bin.cheng@arm.com>
10621 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
10623 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
10625 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
10626 Generalize to complex and vector.
10627 * tree.c (build_all_ones_cst): New function.
10628 * tree.h (build_all_ones_cst): Declare it.
10630 2013-06-13 Alan Modra <amodra@gmail.com>
10632 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
10633 * config/rs6000/rs6000.md (signbittf2): New insn.
10634 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
10635 (abstf2_internal, cmptf_internal2): Likewise.
10636 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
10638 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
10639 Pat Haugen <pthaugen@us.ibm.com>
10640 Peter Bergner <bergner@vnet.ibm.com>
10642 * config/rs6000/rs6000.c (emit_load_locked): Add support for
10643 power8 byte, half-word, and quad-word atomic instructions.
10644 (emit_store_conditional): Likewise.
10645 (rs6000_expand_atomic_compare_and_swap): Likewise.
10646 (rs6000_expand_atomic_op): Likewise.
10648 * config/rs6000/sync.md (larx): Add new modes for power8.
10650 (AINT): New mode iterator to include TImode as well as normal
10651 integer modes on power8.
10652 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
10653 that VSX registers are not considered. Use AINT mode iterator
10654 instead of INT1 to allow inclusion of quad word atomic operations
10656 (load_locked<mode>): Likewise.
10657 (store_conditional<mode>): Likewise.
10658 (atomic_compare_and_swap<mode>): Likewise.
10659 (atomic_exchange<mode>): Likewise.
10660 (atomic_nand<mode>): Likewise.
10661 (atomic_fetch_<fetchop_name><mode>): Likewise.
10662 (atomic_nand_fetch<mode>): Likewise.
10663 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
10665 (ATOMIC): On power8, add QImode, HImode modes.
10666 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
10667 modes that promote to SImode.
10668 (load_lockedti): Convert TImode arguments to PTImode, so that we
10669 get a guaranteed even/odd register pair.
10670 (load_lockedpti): Likewise.
10671 (store_conditionalti): Likewise.
10672 (store_conditionalpti): Likewise.
10674 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
10675 atomic load/store instructions.
10678 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
10680 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
10682 (insn_count): New attribute, with most cases extracted from...
10683 (length): ...here. Redefine most cases in terms of insn_count.
10684 (single_insn): Delete.
10685 (can_delay): Use insn_count to check for single instructions.
10686 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
10687 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
10688 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
10689 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
10690 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
10691 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
10692 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
10693 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
10694 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
10695 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
10696 rather than "length".
10697 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
10698 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
10699 Use "insn_count" rather than "length".
10700 * config/mips/mips-dsp.md
10701 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
10702 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
10705 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
10707 PR tree-optimization/57361
10708 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
10710 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
10712 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
10714 (aarch64_simd_combine<mode>): New instruction expansion.
10715 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
10716 function prototype.
10717 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
10718 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
10720 2013-06-12 Jan Hubicka <jh@suse.cz>
10722 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
10723 decl has when in streaming stage.
10724 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
10725 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
10727 2013-06-12 Roland Stigge <stigge@antcom.de>
10730 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
10732 2013-06-12 Jakub Jelinek <jakub@redhat.com>
10734 PR tree-optimization/57537
10735 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
10736 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
10738 2013-06-12 Richard Biener <rguenther@suse.de>
10740 * data-streamer.h (streamer_write_char_stream): CSE
10741 obs->current_pointer.
10742 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
10743 streamer_write_char_stream manually and optimize the resulting loop.
10744 (streamer_write_hwi_stream): Likewise.
10746 2013-06-12 Jan Hubicka <jh@suse.cz>
10748 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
10749 * cgraph.h (varpool_create_empty_node): Declare.
10750 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
10752 * symtab.c (symtab_unregister_node): Be lax about missin entries
10754 (symtab_get_node): Update comment.
10755 * varpool.c (varpool_create_empty_node): Break out from ...
10756 (varpool_node_for_decl): ... here.
10757 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
10759 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
10761 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
10762 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
10763 part. Use straight-line flow at the end.
10764 <COMPONENT_REF>: Remove superfluous else.
10765 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
10767 2013-06-12 Jakub Jelinek <jakub@redhat.com>
10770 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
10771 target hook even for !TREE_PUBLIC decls. If no resolution info
10772 is available, return false for common and external decls.
10774 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
10776 * config/rl78/constraints.md (U): New constraint.
10777 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
10780 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
10783 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
10784 to allow returning address to AT_PLATFORM name.
10786 2013-06-11 Jan Hubicka <jh@suse.cz>
10788 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
10789 * cgraph.h (symtab_node_base): Add weakref flag.
10790 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
10791 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
10792 (output_weakrefs): Use weakref flag.
10793 * fold-const.c (simple_operand_p): Handle WEAK.
10794 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
10795 * ipa.c (varpool_externally_visible_p): Drop weakref.
10796 (function_and_variable_visibility): Update comment; fix weakref
10797 sanity checks; do not clear DECL_WEAK on them.
10798 * lto-cgraph.c (lto_output_node): update.
10799 (lto_output_varpool_node): Update.
10800 (input_overwrite_node): Update.
10801 (input_node): Update.
10802 (input_varpool_node): Update.
10803 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
10804 (lto_symtab_merge_symbols): Add sanity check.
10805 (lto_symtab_prevailing_decl): Do not special case weakrefs.
10806 * passes.c (rest_of_decl_compilation): Set static flag, too.
10807 * symtab.c (dump_symtab_base): Dump weakref.
10808 (verify_symtab_base): Sanity check weakrefs.
10809 (symtab_make_decl_local): Remove duplicated code.
10810 (symtab_alias_ultimate_target): Simplify.
10811 * varpool.c (varpool_create_variable_alias): Set weakref flag.
10813 2013-06-11 Tom de Vries <tom@codesourcery.com>
10815 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
10816 sequence_vect == NULL.
10818 2013-06-11 DJ Delorie <dj@redhat.com>
10820 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
10821 (rl78_unwind_word_mode): New.
10823 2013-06-11 David Malcolm <dmalcolm@redhat.com>
10825 * final.c (debug_prefix_maps): Make static.
10827 2013-06-11 David Malcolm <dmalcolm@redhat.com>
10829 * function.c (initial_trampoline): Remove stray copy.
10831 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
10833 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
10835 2013-06-11 Martin Jambor <mjambor@suse.cz>
10837 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
10838 within bounds at the beginning of the function.
10840 2013-06-11 Alan Modra <amodra@gmail.com>
10842 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
10844 (get_named_section): Don't NULL !DECL_P decl.
10846 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
10848 * doc/invoke.texi (core-avx2): Document.
10850 (atom): Updated with MOVBE.
10852 2013-06-11 Richard Biener <rguenther@suse.de>
10854 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
10856 2013-06-11 Anton Blanchard <anton@samba.org>
10858 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
10859 correct shift value in little-endian mode.
10861 2013-06-11 Jakub Jelinek <jakub@redhat.com>
10864 * varasm.c (get_variable_align): Move #endif to the right place.
10866 2013-06-10 Cary Coutant <ccoutant@google.com>
10868 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
10869 for hash so that hash table traversal order is deterministic.
10871 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
10872 Pat Haugen <pthaugen@us.ibm.com>
10873 Peter Bergner <bergner@vnet.ibm.com>
10875 * config/rs6000/vector.md (GPR move splitter): Do not split moves
10876 of vectors in GPRS if they are direct moves or quad word load or
10879 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
10881 (direct_move_p): Likewise.
10882 (quad_load_store_p): Likewise.
10884 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
10885 classes into bins based on the physical register type.
10886 (reg_class_to_reg_type): Likewise.
10887 (IS_STD_REG_TYPE): Likewise.
10888 (IS_FP_VECT_REG_TYPE): Likewise.
10889 (reload_fpr_gpr): Arrays to determine what insn to use if we can
10890 use direct move instructions.
10891 (reload_gpr_vsx): Likewise.
10892 (reload_vsx_gpr): Likewise.
10893 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
10894 information that is a simplification of register classes. Also
10895 precalculate direct move reload helpers.
10896 (direct_move_p): New function to return true if the operation can
10897 be done as a direct move instruciton.
10898 (quad_load_store_p): New function to return true if the operation
10899 is a quad memory operation.
10900 (rs6000_legitimize_address): If quad memory, only allow register
10901 indirect for TImode addresses.
10902 (rs6000_legitimate_address_p): Likewise.
10903 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
10904 (rs6000_reload_register_type): Likewise.
10905 (register_to_reg_type): Return register type.
10906 (rs6000_secondary_reload_simple_move): New helper function for
10907 secondary reload and secondary memory needed to identify anything
10908 that is a simple move, and does not need reloading.
10909 (rs6000_secondary_reload_direct_move): New helper function for
10910 secondary reload to identify cases that can be done with several
10911 instructions via the direct move instructions.
10912 (rs6000_secondary_reload_move): New helper function for secondary
10913 reload to identify moves between register types that can be done.
10914 (rs6000_secondary_reload): Add support for quad memory operations
10915 and for direct move.
10916 (rs6000_secondary_memory_needed): Likewise.
10917 (rs6000_debug_secondary_memory_needed): Change argument names.
10918 (rs6000_output_move_128bit): New function to return the move to
10919 use for 128-bit moves, including knowing about the various
10920 limitations of quad memory operations.
10922 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
10923 memory operations. call rs6000_output_move_128bit for the actual
10924 instruciton(s) to generate.
10925 (vsx_movti_64bit): Likewise.
10927 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
10928 (UNSPEC_P8V_MTVSRWZ): Likewise.
10929 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
10930 (UNSPEC_P8V_MTVSRD): Likewise.
10931 (UNSPEC_P8V_XXPERMDI): Likewise.
10932 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
10933 (UNSPEC_FUSION_GPR): Likewise.
10934 (FMOVE128_GPR): New iterator for direct move.
10935 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
10936 (f32_sv): Likewise.
10937 (f32_dm): Likewise.
10938 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
10939 loads and direct move instructions.
10940 (zero_extendsidi2_lfiwzx): Likewise.
10941 (extendsidi2_lfiwax): Likewise.
10942 (extendsidi2_nocell): Likewise.
10943 (floatsi<mode>2_lfiwax): Likewise.
10944 (lfiwax): Likewise.
10945 (floatunssi<mode>2_lfiwzx): Likewise.
10946 (lfiwzx): Likewise.
10947 (fix_trunc<mode>_stfiwx): Likewise.
10948 (fixuns_trunc<mode>_stfiwx): Likewise.
10949 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
10950 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
10951 (parity<mode>2_cmpb): Set length/type attr.
10952 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
10953 for 'mr.' to fast_compare.
10954 (bpermd_<mode>): Change type attr to popcnt.
10955 (p8_fmrgow_<mode>): New insns for power8 direct move support.
10956 (p8_mtvsrwz_1): Likewise.
10957 (p8_mtvsrwz_2): Likewise.
10958 (reload_fpr_from_gpr<mode>): Likewise.
10959 (p8_mtvsrd_1): Likewise.
10960 (p8_mtvsrd_2): Likewise.
10961 (p8_xxpermdi_<mode>): Likewise.
10962 (reload_vsx_from_gpr<mode>): Likewise.
10963 (reload_vsx_from_gprsf): Likewise.
10964 (p8_mfvsrd_3_<mode>): LIkewise.
10965 (reload_gpr_from_vsx<mode>): Likewise.
10966 (reload_gpr_from_vsxsf): Likewise.
10967 (p8_mfvsrd_4_disf): Likewise.
10968 (multi-word GPR splits): Do not split direct moves or quad memory
10971 2013-06-10 David Malcolm <dmalcolm@redhat.com>
10973 * tree-into-ssa.c (interesting_blocks): Make static.
10975 2013-06-10 Jakub Jelinek <jakub@redhat.com>
10978 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
10979 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
10980 Use DATA_ABI_ALIGNMENT for that case instead if defined.
10981 (get_variable_align): New function.
10982 (get_variable_section, emit_bss, emit_common,
10983 assemble_variable_contents, place_block_symbol): Use
10984 get_variable_align instead of DECL_ALIGN.
10985 (assemble_noswitch_variable): Add align argument, use it
10986 instead of DECL_ALIGN.
10987 (assemble_variable): Adjust caller. Use get_variable_align
10988 instead of DECL_ALIGN.
10989 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
10991 (DATA_ABI_ALIGNMENT): Define.
10992 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
10993 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
10994 opt is false, only return the psABI mandated alignment increase.
10995 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
10996 (DATA_ABI_ALIGNMENT): ... this.
10997 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
10998 (DATA_ABI_ALIGNMENT): ... this.
10999 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
11000 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
11001 (DATA_ABI_ALIGNMENT): ... this.
11002 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
11003 * doc/tm.texi: Regenerated.
11005 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
11007 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
11008 cmp_code to construct REG_EQUAL note.
11010 2013-06-09 Jakub Jelinek <jakub@redhat.com>
11013 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
11014 that operands[2] doesn't overlap with operands[0].
11016 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
11017 Jan Hubicka <jh@suse.cz>
11019 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
11020 hack to mark symbols as used.
11022 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
11024 PR rtl-optimization/57559
11025 * lra-constraints.c (process_alt_operands): Don't discourage
11026 memory with known offset for offsetable memory constraint.
11027 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
11029 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
11031 * varasm.c (struct oc_local_state): Reorder fields.
11032 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
11033 and adjust accordingly.
11034 (output_constructor): Reorder initialization code and adjust call to
11035 output_constructor_bitfield.
11037 2013-06-07 Jan Hubicka <jh@suse.cz>
11039 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
11041 2013-06-07 David Malcolm <dmalcolm@redhat.com>
11043 * tree-object-size.c (unknown): Make const.
11045 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11047 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
11048 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
11049 for last alternative in the cpu_facility attribute.
11051 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11054 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
11055 (xordi3): Change operand 2 constraint to arm_xordi_operand.
11056 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
11057 * config/arm/constraints.md (Dg): New constraint.
11058 * config/arm/neon.md (xordi3_neon): Remove.
11059 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
11060 * config/arm/predicates.md (arm_xordi_operand): New predicate.
11062 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11064 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
11065 Clean up alternatives.
11067 2013-06-07 Alan Modra <amodra@gmail.com>
11069 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
11072 2013-06-07 Alan Modra <amodra@gmail.com>
11074 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
11076 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11078 * config/arm/constraints.md (Df): New constraint.
11079 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
11080 Correct length attribute for last two alternatives.
11082 2013-06-07 Alan Modra <amodra@gmail.com>
11084 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
11085 override user -mfp-in-toc.
11086 (offsettable_ok_by_alignment): Consider just the current access
11087 rather than the whole object, unless BLKmode. Handle
11088 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
11089 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
11090 for -mcmodel=medium.
11091 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
11092 override user -mfp-in-toc or -msum-in-toc. Default to
11093 -mno-fp-in-toc for -mcmodel=medium.
11095 2013-06-06 DJ Delorie <dj@redhat.com>
11097 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
11098 TARGET_VALID_POINTER_MODE.
11100 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
11101 Pat Haugen <pthaugen@us.ibm.com>
11102 Peter Bergner <bergner@vnet.ibm.com>
11104 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
11105 Document new power8 builtins.
11107 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
11108 condition code register, to allow 128-bit logical operations to be
11109 done in the VSX or GPR registers.
11110 (nor<mode>3): Use the canonical form for nor.
11111 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
11112 vclz*, and vpopcnt* vector instructions.
11113 (nand<mode>3): Likewise.
11114 (orc<mode>3): Likewise.
11115 (clz<mode>2): LIkewise.
11116 (popcount<mode>2): Likewise.
11118 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
11119 that only the GPRs are recognized.
11121 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
11122 support for new power8 builtins.
11124 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
11126 (xscvdpspn): Likewise.
11132 (vpopcnt): Likewise.
11133 (vpopcntb): Likewise.
11134 (vpopcnth): Likewise.
11135 (vpopcntw): Likewise.
11136 (vpopcntd): Likewise.
11138 (vmrgew): Likewise.
11139 (vmrgow): Likewise.
11141 (eqv_v16qi3): Likewise.
11142 (eqv_v8hi3): Likewise.
11143 (eqv_v4si3): Likewise.
11144 (eqv_v2di3): Likewise.
11145 (eqv_v4sf3): Likewise.
11146 (eqv_v2df3): Likewise.
11148 (nand_v16qi3): Likewise.
11149 (nand_v8hi3): Likewise.
11150 (nand_v4si3): Likewise.
11151 (nand_v2di3): Likewise.
11152 (nand_v4sf3): Likewise.
11153 (nand_v2df3): Likewise.
11155 (orc_v16qi3): Likewise.
11156 (orc_v8hi3): Likewise.
11157 (orc_v4si3): Likewise.
11158 (orc_v2di3): Likewise.
11159 (orc_v4sf3): Likewise.
11160 (orc_v2df3): Likewise.
11162 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
11163 allow power8 quad mode in 64-bit.
11164 (rs6000_builtin_vectorized_function): Add support to vectorize
11165 ISA 2.07 count leading zeros, population count builtins.
11166 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
11167 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
11168 (builtin_function_type): Add vgbbd builtin function which takes an
11170 (altivec_expand_vec_perm_const): Add support for new power8 merge
11173 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
11174 that does not include TImdoe for use with 32-bit.
11175 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
11177 (UNSPEC_VSX_CVDPSPN): Likewise.
11178 (vsx_xscvdpspn): Likewise.
11179 (vsx_xscvspdpn): Likewise.
11180 (vsx_xscvdpspn_scalar): Likewise.
11181 (vsx_xscvspdpn_directmove): Likewise.
11182 (vsx_and<mode>3): Split logical operations into 32-bit and
11183 64-bit. Add support to do logical operations on TImode as well as
11184 VSX vector types. Allow logical operations to be done in either
11185 VSX registers or in general purpose registers in 64-bit mode. Add
11186 splitters if GPRs were used. For AND, add clobber of CCmode to
11187 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
11189 (vsx_and<mode>3_32bit): Likewise.
11190 (vsx_and<mode>3_64bit): Likewise.
11191 (vsx_ior<mode>3): Likewise.
11192 (vsx_ior<mode>3_32bit): Likewise.
11193 (vsx_ior<mode>3_64bit): Likewise.
11194 (vsx_xor<mode>3): Likewise.
11195 (vsx_xor<mode>3_32bit): Likewise.
11196 (vsx_xor<mode>3_64bit): Likewise.
11197 (vsx_one_cmpl<mode>2): Likewise.
11198 (vsx_one_cmpl<mode>2_32bit): Likewise.
11199 (vsx_one_cmpl<mode>2_64bit): Likewise.
11200 (vsx_nor<mode>3): Likewise.
11201 (vsx_nor<mode>3_32bit): Likewise.
11202 (vsx_nor<mode>3_64bit): Likewise.
11203 (vsx_andc<mode>3): Likewise.
11204 (vsx_andc<mode>3_32bit): Likewise.
11205 (vsx_andc<mode>3_64bit): Likewise.
11206 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
11207 and xxlorc instructions.
11208 (vsx_eqv<mode>3_64bit): Likewise.
11209 (vsx_nand<mode>3_32bit): Likewise.
11210 (vsx_nand<mode>3_64bit): Likewise.
11211 (vsx_orc<mode>3_32bit): Likewise.
11212 (vsx_orc<mode>3_64bit): Likewise.
11214 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
11216 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
11218 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
11219 (p8_vmrgow): Likewise.
11220 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
11221 GPRs to be split under VSX.
11222 (p8v_clz<mode>2): Add power8 count leading zero support.
11223 (p8v_popcount<mode>2): Add power8 population count support.
11224 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
11227 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
11230 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
11232 (vec_nand): Likewise.
11233 (vec_vclz): Likewise.
11234 (vec_vclzb): Likewise.
11235 (vec_vclzd): Likewise.
11236 (vec_vclzh): Likewise.
11237 (vec_vclzw): Likewise.
11238 (vec_vgbbd): Likewise.
11239 (vec_vmrgew): Likewise.
11240 (vec_vmrgow): Likewise.
11241 (vec_vpopcnt): Likewise.
11242 (vec_vpopcntb): Likewise.
11243 (vec_vpopcntd): Likewise.
11244 (vec_vpopcnth): Likewise.
11245 (vec_vpopcntw): Likewise.
11247 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
11249 PR rtl-optimization/57468
11250 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
11253 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
11255 PR rtl-optimization/57459
11256 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
11257 type when setting live regs.
11259 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
11261 * config/s390/s390.opt (mlra): New option.
11262 * config/s390/s390.c (s390_decompose_address): Check displacement
11263 for all registers for LRA.
11264 (s390_secondary_reload): Don't used secondary reloads for LRA.
11265 (s390_lra_p): New function.
11266 (TARGET_LRA_P): Define.
11267 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
11268 of attribute cpu_facility to zarch for the last alternative.
11269 (*cmpmem_short): Ditto.
11271 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
11273 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
11274 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
11275 (arm_expand_prologue): Likewise.
11277 2013-06-06 Teresa Johnson <tejohnson@google.com>
11280 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
11281 as this is now done by redirect_edge_and_branch_force.
11282 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
11283 barriers, and fix interaction with splitting.
11284 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
11285 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
11286 reflect changes made in the routine.
11287 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
11288 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
11289 since this is called in cfglayout mode, and replace partition fixup
11290 with assert as that is now done by force_nonfallthru_and_redirect.
11291 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
11292 already be marked with region crossing note.
11293 (insert_section_boundary_note): Make non-static, gate on flag
11294 has_bb_partition, rewrite to also check for multiple partitions.
11295 (rest_of_handle_reorder_blocks): Remove call to
11296 insert_section_boundary_note, now done later during free_cfg.
11297 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
11298 * bb-reorder.h (insert_section_boundary_note): Declare.
11299 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
11300 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
11301 invoke insert_section_boundary_note.
11302 (try_redirect_by_replacing_jump): Remove unnecessary
11303 check for region crossing note.
11304 (fixup_partition_crossing): New function.
11305 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
11306 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
11307 in non-cfglayout mode.
11308 (force_nonfallthru_and_redirect): Fixup partition boundaries,
11309 remove old code that tried to do this. Emit barrier correctly
11310 when we are in cfglayout mode.
11311 (last_bb_in_partition): New function.
11312 (rtl_split_edge): Correctly fixup partition boundaries.
11313 (commit_one_edge_insertion): Remove old code that tried to
11314 fixup region crossing edge since this is now handled in
11315 split_block, and set up insertion point correctly since
11316 block may now end in a jump.
11317 (verify_hot_cold_block_grouping): Guard against checking when not in
11318 linearized RTL mode.
11319 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
11321 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
11322 rtl_verify_flow_info, so not called in cfglayout mode.
11323 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
11324 (fixup_reorder_chain): Remove old code that attempted to fixup region
11325 crossing note as this is now handled in force_nonfallthru_and_redirect.
11326 (duplicate_insn_chain): Don't duplicate switch section notes.
11327 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
11329 * basic-block.h (emit_barrier_after_bb): Declare.
11331 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11333 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
11334 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
11335 arm_usatsihi): Adjust alternatives for arm_restrict_it.
11337 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11339 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
11341 * config/arm/ldmstm.md: Regenerate.
11343 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11345 * config/arm/sync.md (atomic_loaddi_1):
11346 Disable predication for arm_restrict_it.
11347 (arm_load_exclusive<mode>): Likewise.
11348 (arm_load_exclusivesi): Likewise.
11349 (arm_load_exclusivedi): Likewise.
11350 (arm_load_acquire_exclusive<mode>): Likewise.
11351 (arm_load_acquire_exclusivesi): Likewise.
11352 (arm_load_acquire_exclusivedi): Likewise.
11353 (arm_store_exclusive<mode>): Likewise.
11354 (arm_store_exclusive<mode>): Likewise.
11355 (arm_store_release_exclusivedi): Likewise.
11356 (arm_store_release_exclusive<mode>): Likewise.
11358 2013-06-06 Richard Biener <rguenther@suse.de>
11360 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
11362 (lto_tag_is_tree_code_p): Adjust.
11363 (lto_tag_is_gimple_code_p): Likewise.
11364 (lto_gimple_code_to_tag): Likewise.
11365 (lto_tag_to_gimple_code): Likewise.
11366 (lto_tree_code_to_tag): Likewise.
11367 (lto_tag_to_tree_code): Likewise.
11368 * data-streamer.h (streamer_write_hwi_in_range): Use
11369 uhwi streaming to stream the normalized range.
11370 (streamer_read_hwi_in_range): Likewise.
11372 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11374 * config/arm/arm.md (enabled_for_depr_it): New attribute.
11375 (predicable_short_it): Likewise.
11376 (predicated): Likewise.
11377 (enabled): Handle above.
11378 (define_cond_exec): Set predicated attribute to yes.
11380 2013-06-05 Mike Stump <mikestump@comcast.net>
11382 * gdbinit.in (__FUNCTION__): Add.
11384 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
11386 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
11387 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
11389 2013-06-05 Jan Hubicka <jh@suse.cz>
11391 * varasm.c (mark_decl_referenced): Revert the removal until targets
11394 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
11396 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
11397 instead of mark_decl_referenced.
11399 2013-06-05 Jan Hubicka <jh@suse.cz>
11401 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
11402 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
11403 and symtab_used_from_object_file_p.
11404 (cgraph_make_node_local_1): Clear forced_by_abi.
11405 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
11406 * cgraph.h (symtab_node_base): Add forced_by_abi.
11407 (decide_is_variable_needed): Remove.
11408 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
11409 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
11410 (decide_is_symbol_needed): ... this one; handle symbols in general;
11411 always analyze virtuals; honnor forced_by_abi.
11412 (cgraph_finalize_function): Update.
11413 (varpool_finalize_decl): Update.
11414 (symbol_defined_and_needed): Remove.
11415 (analyze_functions): Update.
11416 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
11417 output_refs, input_overwrite_node): Handle forced_by_abi.
11418 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
11419 (address_taken_from_non_vtable_p): ... this one.
11420 (comdat_can_be_unshared_p_1): New function.
11421 (cgraph_comdat_can_be_unshared_p): Rename to ...
11422 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
11423 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
11424 (function_and_variable_visibility): Clear forced_by_abi as needed.
11425 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
11426 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
11427 * symtab.c (dump_symtab_base): Dump forced_by_abi.
11428 * varpool.c (decide_is_variable_needed): Remove.
11430 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11432 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
11433 (arm_option_override): Override arm_restrict_it where appropriate.
11434 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
11435 * config/arm/arm.opt (mrestrict-it): New command-line option.
11436 * doc/invoke.texi: Document -mrestrict-it.
11438 2013-06-05 David Malcolm <dmalcolm@redhat.com>
11440 * tsan.c (tsan_atomic_table): Make const.
11442 2013-06-05 Richard Biener <rguenther@suse.de>
11444 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
11445 index associated with the tree we are supposed to replace.
11446 * tree-streamer-out.c (pack_ts_base_value_fields): Output
11447 TREE_ASM_WRITTEN as zero for everything but SSA names.
11449 2013-06-05 David Malcolm <dmalcolm@redhat.com>
11451 * tree-ssa-structalias.c (call_stmt_vars): Make static.
11453 2013-06-04 Jan Hubicka <jh@suse.cz>
11455 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
11456 (input_node, input_varpool_node): Handle correctly external same
11458 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
11459 nodes at ltrans stage.
11461 2013-06-04 Jan Hubicka <jh@suse.cz>
11463 * ipa-inline.c (update_caller_keys): Fix availability test.
11464 (update_callee_keys): Likewise.
11465 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
11466 to follow ELF standard.
11468 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
11470 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
11471 (mips64r5900el-*-elf*): New configurations.
11472 * config/mips/mips-cpus.def (r5900): New processor.
11473 * config/mips/mips-tables.opt: Regenerate.
11474 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
11475 (mips_issue_rate): Handle PROCESSOR_R5900.
11476 (mips_reorg_process_insns): Force reorder mode for the R5900.
11477 * config/mips/mips.h (TARGET_MIPS5900): Define.
11478 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
11480 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
11482 * config/mips/mips.md (processor): Add r5900.
11483 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
11485 2013-06-04 Ian Bolton <ian.bolton@arm.com>
11487 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
11488 into function to generate MOVI instruction.
11489 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
11490 (aarch64_preferred_simd_mode): Turn into wrapper.
11491 (aarch64_output_scalar_simd_mov_immediate): New function.
11492 * config/aarch64/aarch64-protos.h: Add prototype for above.
11494 2013-06-04 Ian Bolton <ian.bolton@arm.com>
11496 * config/aarch64/aarch64.c (simd_immediate_info): Remove
11497 element_char member.
11498 (sizetochar): Return signed char.
11499 (aarch64_simd_valid_immediate): Remove elchar and other
11500 unnecessary variables.
11501 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
11502 Calculate element_char as required.
11503 * config/aarch64/aarch64-protos.h: Update and move prototype
11504 for aarch64_output_simd_mov_immediate.
11505 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
11508 2013-06-04 Ian Bolton <ian.bolton@arm.com>
11510 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
11511 information completed by aarch64_simd_valid_immediate.
11512 (aarch64_legitimate_constant_p): Update arguments.
11513 (aarch64_simd_valid_immediate): Work with struct rather than many
11515 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
11516 (aarch64_simd_make_constant): Update arguments.
11517 (aarch64_output_simd_mov_immediate): Work with struct rather than
11518 many pointers. Output immediate directly rather than as operand.
11519 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
11521 * config/aarch64/constraints.md (Dn): Update arguments.
11523 2013-06-04 Ian Bolton <ian.bolton@arm.com>
11525 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
11527 (aarch64_simd_immediate_valid_for_move): Remove.
11528 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
11529 (aarch64_simd_make_constant): Update call.
11530 (aarch64_output_simd_mov_immediate): Update call.
11531 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
11533 * config/aarch64/constraints.md (Dn): Update call.
11535 2013-06-04 Ian Bolton <ian.bolton@arm.com>
11537 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
11538 return type to bool for prototype.
11539 (aarch64_legitimate_constant_p): Check for true instead of not -1.
11540 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
11541 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
11543 2013-06-04 Catherine Moore <clm@codesourcery.com>
11545 * config/mips/mips.opt (meva): New.
11546 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
11547 (ASM_SPEC): Handle -meva.
11548 * doc/invoke.texi (meva): Document.
11550 2013-06-04 Alan Modra <amodra@gmail.com>
11552 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
11555 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11557 * rtl.def: Add extra fourth optional field to define_cond_exec.
11558 * gensupport.c (process_one_cond_exec): Process attributes from
11560 * doc/md.texi: Document fourth field in define_cond_exec.
11562 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
11564 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
11565 out the processing order as in store_bit_field_1.
11567 2013-06-04 Jan Hubicka <jh@suse.cz>
11569 PR middle-end/57500
11570 * cgraphunit.c (cgraph_process_same_body_aliases): Create
11571 non-VAR_DECL node if it does not exist yet.
11573 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
11575 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
11576 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
11577 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
11578 target_cpu_default setting.
11580 2013-06-03 Teresa Johnson <tejohnson@google.com>
11582 * dumpfile.c (opt_info_switch_p): Change -fopt-info
11583 default to -fopt-info=optimized instead of all.
11584 * doc/invoke.texi: Ditto.
11585 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
11586 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
11587 (execute_vect_slp): Emit BB vectorization success under
11588 MSG_OPTIMIZED_LOCATIONS.
11589 * tree-vect-slp.c (vect_slp_transform_bb): Change
11590 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
11591 * tree-vect-loop.c (vect_transform_loop): Ditto.
11593 2013-06-03 Jason Merrill <jason@redhat.com>
11596 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
11597 Use TARGET_EXPR for C++.
11599 2013-06-03 Jakub Jelinek <jakub@redhat.com>
11601 PR rtl-optimization/57268
11602 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
11603 if DEBUG_INSN_P (insn).
11606 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
11608 PR rtl-optimization/57268
11609 * sched-deps.c (sched_analyze_2): Flush dependence lists if
11610 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
11612 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
11614 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
11615 (ix86_avoid_lea_for_addr): Likewise.
11616 (exact_dependency_1): Likewise.
11617 (ix86_adjust_cost): Likewise.
11618 (swap_top_of_ready_list): Fix formatting and !reload_completed check
11620 (do_reorder_for_imul): Fix typo, formatting and
11621 !reload_completed check removed.
11622 (ix86_sched_reorder): Fix typo and formatting.
11623 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
11626 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
11628 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
11630 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
11632 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
11633 <STRING_CST>: Likewise.
11634 <VECTOR_CST>: Likewise.
11636 2013-06-01 Janus Weil <janus@gcc.gnu.org>
11637 Mikael Morin <mikael@gcc.gnu.org>
11639 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
11640 * config.in: Regenerated.
11641 * configure: Regenerated.
11643 2013-06-01 Jan Hubicka <jh@suse.cz>
11645 PR middle-end/57366
11646 * cgraphunit.c (compile): When weakref is not supported,
11647 set up transparent aliases before final output pass.
11648 * varasm.c (assemble_alias): Do not try to do it here.
11650 2013-06-01 Jan Hubicka <jh@suse.cz>
11652 PR middle-end/57467
11653 * passes.c (for_per_function): Skip unanalyzed functions.
11655 2013-06-01 Jan Hubicka <jh@suse.cz>
11657 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
11658 (lto_symtab_merge_symbols_1): ... this one.
11659 (lto_symtab_merge_cgraph_nodes): Rename to ...
11660 (lto_symtab_merge_symbols): ... this one; simplify.
11661 * cgraph.c (same_body_aliases_done): Rename to ...
11662 (cpp_implicit_aliases_done): ... this one.
11663 (cgraph_create_function_alias): Update.
11664 (cgraph_same_body_alias): Update.
11665 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
11666 (verify_edge_corresponds_to_fndecl): Simplify.
11667 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
11668 (cgraph_node): Remove same_body_alias.
11669 (varpool_node): Remove alias_of and extra_name_alias.
11670 (same_body_aliases_done): Rename to ..
11671 (cpp_implicit_aliases_done): ... this one.
11672 (symtab_alias_ultimate_target): Add default parameter.
11673 (symtab_resolve_alias): New function.
11674 (fixup_same_cpp_alias_visibility): Declare.
11675 (cgraph_function_node): Add default parameter.
11676 (cgraph_node_asm_name): Likewise.
11677 (cgraph_function_or_thunk_node): Add default parameter; do
11678 not ICE when it is NULL.
11679 (varpool_variable_node): Likewise.
11680 * tree-emutls.c (create_emultls_var): Update.
11681 (ipa_lower_emutls): Update.
11682 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
11683 (cgraph_reset_node): Reset alias info.
11684 (cgraph_finalize_function): Update.
11685 (fixup_same_cpp_alias_visibility): Move to symtab.c.
11686 (analyze_function): Simplify.
11687 (cgraph_process_same_body_aliases): Simplify.
11688 (analyze_functions): Fixup same body aliases.
11689 (handle_alias_pairs): Simplify.
11690 (assemble_thunk): Update.
11691 (assemble_thunks_and_aliases): Update.
11692 (output_weakrefs): Rewrite.
11693 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
11694 (lto_output_varpool_node): Likewise.
11695 (compute_ltrans_boundary): Remve assert.
11696 (get_alias_symbol): New functoin.
11697 (input_node): Rewrite alias handling.
11698 (input_varpool_node): Likewise.
11699 * ipa-pure-const.c (propagate_pure_const): Fix formating.
11700 * ipa.c (process_references): Handle weakrefs correctly.
11701 (symtab_remove_unreachable_nodes): Likewise.
11702 * trans-mem.c (get_cg_data): Update.
11703 (ipa_tm_create_version_alias): Update.
11704 (ipa_tm_execute): Update.
11705 * symtab.c (dump_symtab_base): Dump aliases.
11706 (verify_symtab_base): Verify aliases.
11707 (symtab_node_availability): New function.
11708 (symtab_alias_ultimate_target): Simplify.
11709 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
11710 handle all the fixup cases.
11711 (symtab_resolve_alias): New function.
11712 * passes.c (ipa_write_summaries): Handle weakrefs.
11713 * varpool.c (varpool_analyze_node): Simplify.
11714 (assemble_aliases): Update.
11715 (varpool_create_variable_alias): Simplify.
11716 (varpool_extra_name_alias): Simplify.
11717 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
11718 (lto_symtab_merge_symbols): ... this one.
11720 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
11723 PR rtl-optimization/57268
11724 * sched-deps.c (sched_analyze_2): Flush dependence lists if
11725 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
11727 2013-06-01 Tobias Burnus <burnus@net-b.de>
11729 Partially reverted:
11730 2013-05-31 Tobias Burnus <burnus@net-b.de>
11732 PR middle-end/57073
11733 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
11736 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
11738 PR rtl-optimization/57268
11739 * sched-deps.c (sched_analyze_2): Flush dependence lists if
11740 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
11742 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
11744 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
11745 unordered comparison operators when -fno-trapping-math is in effect
11747 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
11748 and implement unordered comparison operators properly on the e500.
11750 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
11752 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
11753 for constant scalar integers.
11754 (simplify_relational_operation_1): Likewise.
11756 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
11758 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
11759 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
11762 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
11763 Igor Zamyatin <igor.zamyatin@intel.com>
11765 Silvermont (SLM) architecture performance tuning.
11766 * config/i386/i386.h (enum ix86_tune_indices): Add
11767 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
11768 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
11770 * config/i386/i386.c (initial_ix86_tune_features)
11771 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
11772 (ix86_lea_outperforms): Handle Silvermont tuning.
11773 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
11775 (ix86_use_lea_for_mov): Likewise.
11776 (ix86_avoid_lea_for_addr): Likewise.
11777 (ix86_lea_for_add_ok): Likewise.
11778 (exact_dependency_1): New function.
11779 (exact_store_load_dependency): Likewise.
11780 (ix86_adjust_cost): Handle Silvermont tuning.
11781 (do_reoder_for_imul): Likewise.
11782 (swap_top_of_ready_list): New function.
11783 (ix86_sched_reorder): Changed to handle Silvermont tuning.
11785 * config/i386/i386.md (peepholes that split memory operand in fp
11788 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
11790 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
11791 Remove un-necessary braces.
11793 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
11795 * config/aarch64/aarch64.c (aarch64_classify_symbol):
11796 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
11798 2013-05-31 Tobias Burnus <burnus@net-b.de>
11800 PR middle-end/57073
11801 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
11803 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11806 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
11807 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
11808 * config/arm/neon.md (iordi3_neon): Remove.
11809 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
11810 * config/arm/predicates.md (imm_for_neon_logic_operand):
11811 Move to earlier in the file.
11812 (neon_logic_op2): Likewise.
11813 (arm_iordi_operand_neon): New predicate.
11815 2013-05-31 Richard Biener <rguenther@suse.de>
11817 PR tree-optimization/57478
11818 PR tree-optimization/57453
11819 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
11822 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
11824 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
11825 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
11827 2013-05-30 Tobias Burnus <burnus@net-b.de>
11828 Thomas Koenig <tkoenig@gcc.gnu.org>
11830 PR middle-end/57073
11831 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
11832 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
11834 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
11836 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
11838 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
11840 * target.def (register_usage_leveling_p): New hook.
11841 * targhooks.c (default_register_usage_leveling_p): New.
11842 * targhooks.h (default_register_usage_leveling_p): New prototype.
11843 * lra-assigns.c (register_usage_leveling_p): Use the hook.
11844 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
11845 * doc/tm.texi: Update.
11846 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
11848 2013-05-30 Ian Bolton <ian.bolton@arm.com>
11850 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
11851 (*insv_reg<mode>): New define_insn.
11853 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
11855 PR rtl-optimization/57439
11856 * postreload.c (move2add_valid_value_p): Check that we have
11857 a zero subreg_regno_offset when accessing the register in
11858 the requested mode.
11860 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
11861 Igor Zamyatin <igor.zamyatin@intel.com>
11863 Silvermont (SLM) architecture pipeline model, tuning and
11865 * config.gcc: Add slm config options and target.
11867 * config/i386/slm.md: New.
11869 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
11871 * config/i386/i386-c.c (ix86_target_macros_internal): New case
11873 (ix86_target_macros_internal): Likewise.
11875 * config/i386/i386.c (slm_cost): New cost.
11876 (m_SLM): New macro flag.
11877 (initial_ix86_tune_features): Set m_SLM.
11878 (x86_accumulate_outgoing_args): Likewise.
11879 (x86_arch_always_fancy_math_387): Likewise.
11880 (processor_target_table): Add slm cost.
11881 (cpu_names): Add slm cpu name.
11882 (x86_option_override_internal): Set SLM ISA.
11883 (ix86_issue_rate): New case PROCESSOR_SLM.
11884 (ia32_multipass_dfa_lookahead): Likewise.
11885 (fold_builtin_cpu): Add slm.
11887 * config/i386/i386.h (TARGET_SLM): New target macro.
11888 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
11889 (processor_type): Add PROCESSOR_SLM.
11891 * config/i386/i386.md (cpu): Add new value "slm".
11892 (slm.md): Include slm.md.
11894 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
11895 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11897 * config/arm/arm-protos.h: Add and update function protos.
11898 * config/arm/arm.c (use_simple_return_p): New added.
11899 (thumb2_expand_return): Check simple_return flag.
11900 * config/arm/arm.md: Add simple_return and conditional simple_return.
11901 * config/arm/iterators.md: Add iterator for return and simple_return.
11903 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11905 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
11906 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
11907 (arm_emit_vfp_multi_reg_pop): Likewise.
11908 (thumb2_emit_ldrd_pop): Likewise.
11909 (arm_expand_epilogue): Add misc REG_CFA notes.
11910 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
11912 2013-05-29 Lawrence Crowl <crowl@google.com>
11914 * config/arm/t-arm: Update for below.
11916 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
11917 Change type to hash_table. Update dependent calls and types.
11919 * config/i386/t-cygming: Update for below.
11921 * config/i386/t-interix: Update for below.
11923 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
11924 Change type to hash_table. Update dependent calls and types.
11925 (i386_find_on_wrapper_list::wrappers): Likewise.
11927 * config/ia64/t-ia64: Update for below.
11929 * config/ia64/ia64.c (bundle_state_table):
11930 Change type to hash_table. Update dependent calls and types.
11932 * config/mips/mips.c (mips_reorg_process_insns::htab):
11933 Change type to hash_table. Update dependent calls and types.
11935 * config/sol2.c (solaris_comdat_htab):
11936 Change type to hash_table. Update dependent calls and types.
11938 * config/t-sol2: Update for above.
11940 2013-05-29 Teresa Johnson <tejohnson@google.com>
11942 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
11943 functions are not yet marked as defined.
11945 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
11946 Pat Haugen <pthaugen@us.ibm.com>
11947 Peter Bergner <bergner@vnet.ibm.com>
11949 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
11953 (vrotl<mode>3): Likewise.
11954 (vashl<mode>3): Likewise.
11955 (vlshr<mode>3): Likewise.
11956 (vashr<mode>3): Likewise.
11958 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
11959 support for power8 V2DI builtins.
11961 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
11962 power8 V2DI builtins.
11963 (vupkhsw): Likewise.
11964 (vupklsw): Likewise.
11965 (vaddudm): Likewise.
11966 (vminsd): Likewise.
11967 (vmaxsd): Likewise.
11968 (vminud): Likewise.
11969 (vmaxud): Likewise.
11970 (vpkudum): Likewise.
11971 (vpksdss): Likewise.
11972 (vpkudus): Likewise.
11973 (vpksdus): Likewise.
11978 (vsubudm): Likewise.
11979 (vcmpequd): Likewise.
11980 (vcmpgtsd): Likewise.
11981 (vcmpgtud): Likewise.
11982 (vcmpequd_p): Likewise.
11983 (vcmpgtsd_p): Likewise.
11984 (vcmpgtud_p): Likewise.
11985 (vupkhsw): Likewise.
11986 (vupklsw): Likewise.
11987 (vaddudm): Likewise.
11988 (vmaxsd): Likewise.
11989 (vmaxud): Likewise.
11990 (vminsd): Likewise.
11991 (vminud): Likewise.
11992 (vpksdss): Likewise.
11993 (vpksdus): Likewise.
11994 (vpkudum): Likewise.
11995 (vpkudus): Likewise.
12000 (vsubudm): Likewise.
12002 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
12003 support for power8 V2DI instructions.
12005 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
12006 power8 V2DI instructions. Combine pack and unpack insns to use an
12007 iterator for each mode. Check whether a particular mode supports
12008 Altivec instructions instead of just checking TARGET_ALTIVEC.
12009 (UNSPEC_VPKUWUM): Likewise.
12010 (UNSPEC_VPKSHSS): Likewise.
12011 (UNSPEC_VPKSWSS): Likewise.
12012 (UNSPEC_VPKUHUS): Likewise.
12013 (UNSPEC_VPKSHUS): Likewise.
12014 (UNSPEC_VPKUWUS): Likewise.
12015 (UNSPEC_VPKSWUS): Likewise.
12016 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
12017 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
12018 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
12019 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
12020 (UNSPEC_VUPKHSB): Likewise.
12021 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
12022 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
12023 (UNSPEC_VUPKHSH): Likewise.
12024 (UNSPEC_VUPKLSB): Likewise.
12025 (UNSPEC_VUPKLSH): Likewise.
12027 (VI_char): Likewise.
12028 (VI_scalar): Likewise.
12029 (VI_unit): Likewise.
12031 (VP_small): Likewise.
12032 (VP_small_lc): Likewise.
12033 (VU_char): Likewise.
12034 (add<mode>3): Likewise.
12035 (altivec_vaddcuw): Likewise.
12036 (altivec_vaddu<VI_char>s): Likewise.
12037 (altivec_vadds<VI_char>s): Likewise.
12038 (sub<mode>3): Likewise.
12039 (altivec_vsubcuw): Likewise.
12040 (altivec_vsubu<VI_char>s): Likewise.
12041 (altivec_vsubs<VI_char>s): Likewise.
12042 (altivec_vavgs<VI_char>): Likewise.
12043 (altivec_vcmpbfp): Likewise.
12044 (altivec_eq<mode>): Likewise.
12045 (altivec_gt<mode>): Likewise.
12046 (altivec_gtu<mode>): Likewise.
12047 (umax<mode>3): Likewise.
12048 (smax<mode>3): Likewise.
12049 (umin<mode>3): Likewise.
12050 (smin<mode>3): Likewise.
12051 (altivec_vpkuhum): Likewise.
12052 (altivec_vpkuwum): Likewise.
12053 (altivec_vpkshss): Likewise.
12054 (altivec_vpkswss): Likewise.
12055 (altivec_vpkuhus): Likewise.
12056 (altivec_vpkshus): Likewise.
12057 (altivec_vpkuwus): Likewise.
12058 (altivec_vpkswus): Likewise.
12059 (altivec_vpks<VI_char>ss): Likewise.
12060 (altivec_vpks<VI_char>us): Likewise.
12061 (altivec_vpku<VI_char>us): Likewise.
12062 (altivec_vpku<VI_char>um): Likewise.
12063 (altivec_vrl<VI_char>): Likewise.
12064 (altivec_vsl<VI_char>): Likewise.
12065 (altivec_vsr<VI_char>): Likewise.
12066 (altivec_vsra<VI_char>): Likewise.
12067 (altivec_vsldoi_<mode>): Likewise.
12068 (altivec_vupkhsb): Likewise.
12069 (altivec_vupkhs<VU_char>): Likewise.
12070 (altivec_vupkls<VU_char>): Likewise.
12071 (altivec_vupkhsh): Likewise.
12072 (altivec_vupklsb): Likewise.
12073 (altivec_vupklsh): Likewise.
12074 (altivec_vcmpequ<VI_char>_p): Likewise.
12075 (altivec_vcmpgts<VI_char>_p): Likewise.
12076 (altivec_vcmpgtu<VI_char>_p): Likewise.
12077 (abs<mode>2): Likewise.
12078 (vec_unpacks_hi_v16qi): Likewise.
12079 (vec_unpacks_hi_v8hi): Likewise.
12080 (vec_unpacks_lo_v16qi): Likewise.
12081 (vec_unpacks_hi_<VP_small_lc>): Likewise.
12082 (vec_unpacks_lo_v8hi): Likewise.
12083 (vec_unpacks_lo_<VP_small_lc>): Likewise.
12084 (vec_pack_trunc_v8h): Likewise.
12085 (vec_pack_trunc_v4si): Likewise.
12086 (vec_pack_trunc_<mode>): Likewise.
12088 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
12090 (vec_vmaxsd): Likewise.
12091 (vec_vmaxud): Likewise.
12092 (vec_vminsd): Likewise.
12093 (vec_vminud): Likewise.
12094 (vec_vpksdss): Likewise.
12095 (vec_vpksdus): Likewise.
12096 (vec_vpkudum): Likewise.
12097 (vec_vpkudus): Likewise.
12098 (vec_vrld): Likewise.
12099 (vec_vsld): Likewise.
12100 (vec_vsrad): Likewise.
12101 (vec_vsrd): Likewise.
12102 (vec_vsubudm): Likewise.
12103 (vec_vupkhsw): Likewise.
12104 (vec_vupklsw): Likewise.
12106 2013-05-29 Jan Hubicka <jh@suse.cz>
12108 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
12109 flags; reorder rest of fields in more consistent way.
12110 (varpool_node): Remove analyzed, finalized and alias.
12111 (cgraph_ndoe): Likewise.
12112 (symtab_alias_ultimate_target): New function.
12113 (cgraph_function_node): Move offline.
12114 (cgraph_reset_node): Declare.
12115 (cgraph_comdat_can_be_unshared_p): Remove.
12116 (varpool_remove_initializer): Declare.
12117 (varpool_first_defined_variable, varpool_next_defined_variable
12118 cgraph_first_defined_function, cgraph_next_defined_function): Update.
12119 (cgraph_function_with_gimple_body_p): Update.
12120 (varpool_all_refs_explicit_p): Update.
12121 (symtab_alias_target): New function.
12122 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
12123 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
12124 (cgraph_function_or_thunk_node): Simplify using
12125 symtab_alias_ultimate_target.
12126 (varpool_variable_node): Likewise.
12127 * cgraph.c (cgraph_create_function_alias): Update.
12128 (cgraph_add_thunk): Update.
12129 (cgraph_remove_node): Update.
12130 (dump_cgraph_node): Do not dump removed flags.
12131 (cgraph_function_body_availability): Update.
12132 (cgraph_propagate_frequency): Update.
12133 (verify_cgraph_node): Check sanity of local flag.
12134 (cgraph_function_node): Move here from cgraph.h; revamp for
12135 cgraph_function_or_thunk_node.
12136 * lto-symtab.c (lto_varpool_replace_node): Update.
12137 (lto_symtab_resolve_can_prevail_p): Update.
12138 (lto_symtab_merge_cgraph_nodes): Update.
12139 * ipa-cp.c (determine_versionability, initialize_node_lattices,
12140 propagate_constants_accross_call, devirtualization_time_bonus,
12141 ipcp_propagate_stage): Update.
12142 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
12143 * ipa-inline-transform.c (clone_inlined_nodes,
12144 preserve_function_body_p): Update.
12145 * ipa-reference.c (propagate): Update.
12146 (write_node_summary_p): Update.
12147 * toplev.c (wrapup_global_declaration_2): Update.
12148 * cgraphunit.c (cgraph_analyze_function): Rename to ...
12149 (analyze_function) ... this one.
12150 (cgraph_process_new_functions): Update.
12151 (cgraph_reset_node): Export.
12152 (cgraph_finalize_function): Update.
12153 (cgraph_add_new_function): Update.
12154 (process_function_and_variable_attributes): Update.
12155 (varpool_finalize_decl): Update.
12156 (symbol_finalized): Remove.
12157 (symbol_finalized_and_needed): Rename to ...
12158 (symbol_defined_and_needed): ... update.
12159 (cgraph_analyze_functions): Update.
12160 (handle_alias_pairs): Update.
12161 (mark_functions_to_output): Update.
12162 (assemble_thunk): Update.
12163 (output_in_order): Update.
12164 (output_weakrefs): Update.
12165 (finalize_compilation_unit): Update.
12166 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
12167 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
12168 input_node, input_varpool_node): Update.
12169 * dbxout.c (dbxout_expand_expr): Update.
12170 * cgraphclones.c (cgraph_clone_node): Update.
12171 (cgraph_copy_node_for_versioning): Update.
12172 (cgraph_materialize_clone): Update.
12173 (cgraph_materialize_all_clones): Update.
12174 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
12175 propagate_pure_const, propagate_nothrow): Update.
12176 * lto-streamer-out.c (lto_output, write_symbol): Update.
12177 * ipa-utils.c (ipa_reverse_postorder): Update.
12178 * ipa-inline.c (can_inline_edge_p): Update.
12179 (update_caller_keys, ipa_inline): Update.
12180 * dwarf2out.c (reference_to_unused,
12181 premark_types_used_by_global_vars_helper): Update.
12182 * tree-eh.c (tree_could_trap_p): Update.
12183 * ipa-split.c (consider_split, execute_split_functions): Update.
12184 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
12185 has_addr_references_p): Update; move ahead in file for better
12187 (process_references): Simplify.
12188 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
12189 bodies are removed.
12190 (cgraph_comdat_can_be_unshared_p): Make static.
12191 (cgraph_externally_visible_p): Update.
12192 (varpool_externally_visible_p): Update.
12193 (function_and_variable_visibility): Update.
12194 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
12195 ipa_tm_mark_force_output_node): Update.
12196 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
12197 estimate_edge_devirt_benefit, inline_generate_summary,
12198 inline_write_summary): Update.
12199 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
12200 * ipa-prop.c (ipa_compute_jump_functions): Update.
12201 (ipa_print_node_params, ipa_prop_read_section,
12202 ipa_update_after_lto_read, read_replacements_section): Update.
12203 * varasm.c (mark_decl_referenced): Update.
12204 (assemble_alias, dump_tm_clone_pairs): Update.
12205 * tree-inline.c (copy_bb): Update.
12206 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
12208 * symtab.c (dump_symtab_base): Print new flags.
12209 (verify_symtab_base): Verify new flags.
12210 (symtab_alias_ultimate_target): New function.
12211 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
12212 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
12214 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
12216 * i386.c (ix86_get_function_versions_dispatcher,
12217 ix86_generate_version_dispatcher_body): Update.
12218 (fold_builtin_cpu): Use varpool_add_new_variable.
12219 * varpool.c (varpool_remove_initializer): Break out from ...
12220 (varpool_remove_node): ... this one.
12221 (dump_varpool_node, varpool_node_for_asm,
12222 cgraph_variable_initializer_availability, varpool_analyze_node,
12223 varpool_assemble_decl, varpool_remove_unreferenced_decls,
12224 varpool_finalize_named_section_flags, varpool_create_variable_alias):
12227 2013-05-29 Jan Hubicka <jh@suse.cz>
12229 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
12231 2013-05-29 Easwaran Raman <eraman@google.com>
12233 PR tree-optimization/57442
12234 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
12235 when control exits the main loop.
12237 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
12239 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
12241 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
12242 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
12243 * rx/t-rx: Add rx100 under multi library matches option for nofpu
12246 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12248 PR tree-optimization/57441
12249 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
12250 Don't limit size of incr_vec to number of candidates.
12252 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
12254 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
12255 and mips16 directories.
12256 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
12257 (MULTILIB_DIRNAMES): Ditto.
12258 (MULTILIB_EXCEPTIONS): Add new exceptions.
12259 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
12260 (MULTILIB_DIRNAMES): Ditto.
12261 (MULTILIB_EXCEPTIONS): Add new exceptions.
12263 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
12264 Marcus Shawcroft <marcus.shawcroft@arm.com>
12266 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
12267 SYMBOL_TINY_ABSOLUTE.
12268 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
12269 SYMBOL_TINY_ABSOLUTE.
12270 (aarch64_expand_mov_immediate): Likewise.
12271 (aarch64_classify_symbol): Likewise.
12272 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
12273 Permit SYMBOL_TINY_ABSOLUTE.
12274 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
12276 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
12277 Marcus Shawcroft <marcus.shawcroft@arm.com>
12279 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
12280 Refactor if/switch. Replace gcc_assert with if.
12282 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
12284 * config/i386/i386.c (initial_ix86_tune_features): Enable
12285 FP Reassociation for AMD bdver1 and bdver2.
12287 2013-05-29 Martin Jambor <mjambor@suse.cz>
12289 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
12290 and IMAGPART_EXPR do not occur within other handled_components.
12292 2013-05-29 Richard Biener <rguenther@suse.de>
12294 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
12295 access on whether the use is in the BB we currently try to
12297 (vect_bb_vectorization_profitable_p): Pass the BB we currently
12298 vectorize to vect_bb_slp_scalar_cost.
12300 2013-05-29 Richard Biener <rguenther@suse.de>
12302 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
12303 computing scalar cost offsetted by stmts that are kept live
12305 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
12306 for computation of scalar cost.
12308 2013-05-28 Steve Ellcey <sellcey@mips.com>
12310 * config/mips/mips-cpus.def (mips32r2): Change processor type.
12312 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
12314 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
12315 array notation built-in reduction functions.
12316 * doc/passes.texi (Passes): Added documentation about changes done
12318 * doc/invoke.texi (C Dialect Options): Added documentation about
12319 the -fcilkplus flag.
12320 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
12321 (BUILTINS_DEF): Depend on cilkplus.def.
12322 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
12323 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
12324 * cilkplus.def: New file.
12326 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
12328 PR rtl-optimization/57439
12329 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
12331 2013-05-28 Easwaran Raman <eraman@google.com>
12333 PR tree-optimization/57337
12334 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
12335 (find_insert_point): Correctly identify the insertion point
12336 when two statements with the same UID is compared.
12338 2013-05-28 Richard Biener <rguenther@suse.de>
12340 PR tree-optimization/56787
12341 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
12342 from the list of data references.
12343 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
12345 (vect_analyze_loop_operations): Likewise.
12346 (vect_transform_loop): Remove clobbers.
12348 2013-05-28 Martin Jambor <mjambor@suse.cz>
12350 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
12351 and REALPART_EXPRs have scalar type.
12353 2013-05-28 Richard Biener <rguenther@suse.de>
12355 PR tree-optimization/57411
12356 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
12358 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
12359 virtual operand propagation.
12361 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
12363 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
12364 destination register for bmasksi_vis.
12365 (vector_init_bshuffle): Likewise.
12366 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
12368 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
12370 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
12371 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
12372 mode if the instruction isn't available in the original mode.
12373 * config/sparc/sparc.opt (mfix-ut699): New option.
12374 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
12375 (divdf3): Turn into expander.
12376 (divdf3_nofix): New insn.
12377 (divdf3_fix): Likewise.
12378 (divsf3): Disable if -mfix-ut699.
12379 (sqrtdf2): Turn into expander.
12380 (sqrtdf2_nofix): New insn.
12381 (sqrtdf2_fix): Likewise.
12382 (sqrtsf2): Disable if -mfix-ut699.
12384 2013-05-27 Richard Biener <rguenther@suse.de>
12386 PR middle-end/57412
12387 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
12388 block for the new loop.
12390 2013-05-27 Richard Biener <rguenther@suse.de>
12392 PR tree-optimization/57343
12393 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
12394 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
12395 (number_of_iterations_cond): Do not build the folded tree.
12397 2013-05-27 Richard Biener <rguenther@suse.de>
12400 PR middle-end/57381
12401 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
12402 OEP_CONSTANT_ADDRESS_OF retained.
12404 PR tree-optimization/57417
12405 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
12406 for unchanged base.
12407 (set_ssa_val_to): Compare addresses using
12408 get_addr_base_and_unit_offset.
12410 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
12412 PR rtl-optimization/56833
12413 * postreload.c (move2add_record_mode): New function.
12414 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
12415 (move2add_use_add2_insn): Use move2add_record_sym_value.
12416 (move2add_use_add3_insn): Likewise.
12417 (reload_cse_move2add): Use move2add_valid_value_p and
12418 move2add_record_mode. Invalidate call-clobbered and REG_INC
12419 affected regs by setting reg_mode to VOIDmode.
12420 (move2add_note_store): Don't pretend the inside of a SUBREG is
12421 the actual destination. Invalidate single/leading registers by
12422 setting reg_mode to VOIDmode.
12423 Use move2add_record_sym_value, move2add_valid_value_p and
12424 move2add_record_mode.
12426 2013-05-27 Richard Biener <rguenther@suse.de>
12428 PR tree-optimization/57396
12429 * tree-affine.c (double_int_constant_multiple_p): Properly
12430 return false for val == 0 and div != 0.
12432 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
12434 * config/mips/mips.h: Use #elif in preprocessor conditions.
12436 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
12439 * config/mips/constraints.md (kl): New constraint.
12440 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
12441 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
12442 constraint for operand 0. Split after CSE for MIPS16. Emit a move
12443 from LO for MIPS16.
12444 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
12446 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
12449 * config/mips/mips.c (mips_can_inline_p): New function.
12450 (TARGET_CAN_INLINE_P): Define.
12452 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
12454 * sched-int.h (ds_t, dw_t): Make unsigned int.
12455 Fix documentation that describes how all the ds_t bits are used.
12456 Reserve the last bit for delayed-branch scheduling.
12457 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
12458 (BITS_PER_DEP_WEAK): Fix definition and documentation.
12459 (gen_dep_weak_1): Remove prototype.
12460 * sched-deps.c (get_dep_weak_1): Make static.
12461 * target.def (speculate_insn, needs_block_p, gen_spec_check,
12462 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
12463 * doc/tm.texi: Regenerate.
12464 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
12466 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
12469 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
12471 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
12472 Sandra Loosemore <sandra@codesourcery.com>
12474 * config.gcc (powerpc-*): Allow native for with-cpu.
12476 2013-05-24 Jeff Law <law@redhat.com>
12478 PR tree-optimization/57124
12479 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
12480 conversion feeding a condition if the range has an overflow
12481 if -fstrict-overflow. Add warnings for when we do make the
12484 2013-05-24 Dehao Chen <dehao@google.com>
12486 * tree-cfg.c (locus_discrim_map): Fix the typo.
12487 (locus_discrim_hasher): Likewise.
12488 (locus_discrim_hasher::hash): Likewise.
12489 (locus_discrim_hasher::equal): Likewise.
12491 2013-05-24 Martin Jambor <mjambor@suse.cz>
12493 PR tree-optimization/57294
12494 * cgraph.h (ipa_record_stmt_references): Declare.
12495 * cgraphbuild.c (ipa_record_stmt_references): New function.
12496 (build_cgraph_edges): Use ipa_record_stmt_references.
12497 (rebuild_cgraph_edges): Likewise.
12498 (cgraph_rebuild_references): Likewise.
12499 * ipa-prop.c (ipa_modify_call_arguments): Discard references
12500 associated with the old statement and build references from the
12501 newly built statements.
12502 * ipa-ref.c (ipa_remove_stmt_references): New function.
12503 * ipa-ref.h (ipa_remove_stmt_references): Declare.
12505 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
12507 * lra-constraints.c (emit_spill_move): Use smaller mode for
12509 (check_and_process_move): Consider mem-reg moves for secondary
12511 (curr_insn_transform): Don't lose insns emitted before for
12512 secondary memory moves.
12513 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
12514 reg set up in the current insn.
12516 2013-05-24 Dehao Chen <dehao@google.com>
12518 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
12520 (locus_descrim_hasher::equal): Likewise.
12521 (build_gimple_cfg): New discriminator assignment algorithm.
12522 (make_edges): Likewise.
12523 (next_discriminator_for_locus): Likewise.
12524 (same_line_p): Likewise.
12525 (assign_discriminators): Likewise.
12526 (make_cond_expr_edges): Likewise.
12527 (make_gimple_switch_edges): Likewise.
12528 (make_goto_expr_edges): Likewise.
12529 (make_gimple_asm_edges): Likewise.
12531 2013-05-24 Ian Bolton <ian.bolton@arm.com>
12533 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
12534 X format specifier to only display bottom 16 bits.
12535 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
12536 immediate to match for operand 2, since it will be masked.
12538 2013-05-24 Richard Biener <rguenther@suse.de>
12540 PR tree-optimization/57287
12541 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
12542 all SSA names that occur in abnormal PHIs.
12544 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
12547 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
12548 that index is not negative.
12550 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
12552 PR rtl-optimization/55177
12553 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
12554 (simplify_byte_swapping_operation): New.
12555 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
12556 (simplify_relational_operation_1): Deal with BSWAP.
12558 2013-05-23 Richard Henderson <rth@redhat.com>
12561 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
12562 (ix86_reorg): Call it.
12564 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
12567 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
12568 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
12569 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
12571 2013-05-23 Christian Bruel <christian.bruel@st.com>
12574 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
12576 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
12577 Marcus Shawcroft <marcus.shawcroft@arm.com>
12579 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
12580 * config/aarch64/constraints.md (Usa): Remove.
12581 * doc/md.texi (AArch64 Usa): Remove.
12583 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
12584 Marcus Shawcroft <marcus.shawcroft@arm.com>
12586 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
12587 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
12588 * config/aarch64/predicates.md (aarch64_const_address): Remove.
12589 (aarch64_mov_operand): Use aarch64_mov_operand_p.
12591 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
12593 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
12594 instruction (AdvSIMD).
12595 * config/aarch64/aarch64-builtins.c
12596 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
12597 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
12599 2013-05-23 Martin Jambor <mjambor@suse.cz>
12601 PR middle-end/57347
12602 * tree.h (contains_bitfld_component_ref_p): Declare.
12603 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
12604 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
12606 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
12607 not access a bit-field. Assert all final offsets are byte-aligned.
12609 2013-05-23 Richard Biener <rguenther@suse.de>
12611 PR tree-optimization/57380
12612 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
12613 least one invariant or re-used load.
12614 * passes.c (init_optimization_passes): Move pass_phiprop before
12617 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
12619 * config/aarch64/aarch64-simd.md
12620 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
12622 2013-05-23 Richard Biener <rguenther@suse.de>
12624 PR middle-end/57381
12625 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
12626 OEP_CONSTANT_ADDRESS_OF retained.
12628 2013-05-23 Jakub Jelinek <jakub@redhat.com>
12630 PR middle-end/57344
12631 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
12632 don't lower unit. Handle unit not being always BITS_PER_WORD.
12634 2013-05-23 Richard Biener <rguenther@suse.de>
12636 PR rtl-optimization/57341
12637 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
12638 instead of true_dependence.
12640 2013-05-22 David Malcolm <dmalcolm@redhat.com>
12642 * bb-reorder.c (branch_threshold): Make const.
12643 (exec_threshold): Ditto.
12645 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
12646 Pat Haugen <pthaugen@us.ibm.com>
12647 Peter Bergner <bergner@vnet.ibm.com>
12649 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
12650 documentation for the power8 crypto builtins.
12652 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
12654 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
12655 macros for defining power8 builtin functions.
12656 (BU_P8V_AV_2): Likewise.
12657 (BU_P8V_AV_P): Likewise.
12658 (BU_P8V_VSX_1): Likewise.
12659 (BU_P8V_OVERLOAD_1): Likewise.
12660 (BU_P8V_OVERLOAD_2): Likewise.
12661 (BU_CRYPTO_1): Likewise.
12662 (BU_CRYPTO_2): Likewise.
12663 (BU_CRYPTO_3): Likewise.
12664 (BU_CRYPTO_OVERLOAD_1): Likewise.
12665 (BU_CRYPTO_OVERLOAD_2): Likewise.
12666 (XSCVSPDP): Fix typo, point to the correct instruction.
12667 (VCIPHER): Add power8 crypto builtins.
12668 (VCIPHERLAST): Likewise.
12669 (VNCIPHER): Likewise.
12670 (VNCIPHERLAST): Likewise.
12671 (VPMSUMB): Likewise.
12672 (VPMSUMH): Likewise.
12673 (VPMSUMW): Likewise.
12674 (VPERMXOR_V2DI): Likewise.
12675 (VPERMXOR_V4SI: Likewise.
12676 (VPERMXOR_V8HI: Likewise.
12677 (VPERMXOR_V16QI: Likewise.
12678 (VSHASIGMAW): Likewise.
12679 (VSHASIGMAD): Likewise.
12680 (VPMSUM): Likewise.
12681 (VPERMXOR): Likewise.
12682 (VSHASIGMA): Likewise.
12684 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
12685 __CRYPTO__ if the crypto instructions are available.
12686 (altivec_overloaded_builtins): Add support for overloaded power8
12689 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
12690 support for power8 crypto builtins.
12691 (builtin_function_type): Likewise.
12692 (altivec_init_builtins): Add support for builtins that take vector
12693 long long (V2DI) arguments.
12695 * config/rs6000/crypto.md: New file, define power8 crypto
12698 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
12699 Pat Haugen <pthaugen@us.ibm.com>
12700 Peter Bergner <bergner@vnet.ibm.com>
12702 * doc/invoke.texi (Option Summary): Add power8 options.
12703 (RS/6000 and PowerPC Options): Likewise.
12705 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
12706 constraints.md instead of rs6000.h. Reorder w* constraints. Add
12707 wm, wn, wr documentation.
12709 * config/rs6000/constraints.md (wm): New constraint for VSX
12710 registers if direct move instructions are enabled.
12711 (wn): New constraint for no registers.
12712 (wq): New constraint for quad word even GPR registers.
12713 (wr): New constraint if 64-bit instructions are enabled.
12714 (wv): New constraint if power8 vector instructions are enabled.
12715 (wQ): New constraint for quad word memory locations.
12717 * config/rs6000/predicates.md (const_0_to_15_operand): New
12718 constraint for 0..15 for crypto instructions.
12719 (gpc_reg_operand): If VSX allow registers in VSX registers as well
12720 as GPR and floating point registers.
12721 (int_reg_operand): New predicate to match only GPR registers.
12722 (base_reg_operand): New predicate to match base registers.
12723 (quad_int_reg_operand): New predicate to match even GPR registers
12724 for quad memory operations.
12725 (vsx_reg_or_cint_operand): New predicate to allow vector logical
12726 operations in both GPR and VSX registers.
12727 (quad_memory_operand): New predicate for quad memory operations.
12728 (reg_or_indexed_operand): New predicate for direct move support.
12730 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
12731 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
12732 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
12733 (POWERPC_MASKS): Add power8 options.
12734 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
12737 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
12738 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
12740 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
12741 (-mpower8-fusion): New power8 options.
12742 (-mpower8-fusion-sign): Likewise.
12743 (-mpower8-vector): Likewise.
12744 (-mcrypto): Likewise.
12745 (-mdirect-move): Likewise.
12746 (-mquad-memory): Likewise.
12748 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
12749 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
12751 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
12752 (rs6000_debug_vector_unit): Add p8_vector.
12753 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
12754 definitions. Also print fusion state.
12755 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
12756 (rs6000_builtin_mask_calculate): Add power8 builtin support.
12757 (rs6000_option_override_internal): Add support for power8.
12758 (rs6000_common_init_builtins): Add debugging for skipped builtins
12759 if -mdebug=builtin.
12760 (rs6000_adjust_cost): Add power8 support.
12761 (rs6000_issue_rate): Likewise.
12762 (insn_must_be_first_in_group): Likewise.
12763 (insn_must_be_last_in_group): Likewise.
12764 (force_new_group): Likewise.
12765 (rs6000_register_move_cost): Likewise.
12766 (rs6000_opt_masks): Likewise.
12768 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
12769 power8 capable assembler, default to power7 options.
12770 (TARGET_DIRECT_MOVE): Likewise.
12771 (TARGET_CRYPTO): Likewise.
12772 (TARGET_P8_VECTOR): Likewise.
12773 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
12774 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
12775 (VECTOR_MEM_P8_VECTOR_P): Likewise.
12776 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
12777 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
12778 (TARGET_XSCVDPSPN): Likewise.
12779 (TARGET_XSCVSPDPN): Likewsie.
12780 (TARGET_SYNC_HI_QI): Likewise.
12781 (TARGET_SYNC_TI): Likewise.
12782 (MASK_CRYPTO): Likewise.
12783 (MASK_DIRECT_MOVE): Likewise.
12784 (MASK_P8_FUSION): Likewise.
12785 (MASK_P8_VECTOR): Likewise.
12786 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
12787 temporary used by some of the direct move instructions to get two FP
12788 temporary registers does not force creation of a stack frame.
12789 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
12790 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
12791 that any VSX registers are tieable, even if they are also an
12792 Altivec vector mode.
12793 (r6000_reg_class_enum): Add wm, wr, wv constraints.
12794 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
12795 (RS6000_BTM_CRYPTO): Likewise.
12796 (RS6000_BTM_COMMON): Likewise.
12798 * config/rs6000/rs6000.md (cpu attribute): Add power8.
12799 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
12800 (enum rs6000_vector): Add power8 vector support.
12802 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
12806 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
12807 (any_sibcall_could_use_r3): this and handle indirect calls.
12808 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
12810 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12812 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
12814 2013-05-22 Richard Biener <rguenther@suse.de>
12816 PR middle-end/57349
12817 * profile.c (branch_prob): Do not split blocks that are
12818 abnormally receiving from ECF_RETURNS_TWICE functions.
12820 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
12822 * recog.c (offsettable_address_addr_space_p): Fix calculation of
12823 address mode. Move pointer mode initialization to the same place.
12825 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
12827 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
12828 while it has any effect.
12830 2013-05-21 Easwaran Raman <eraman@google.com>
12832 PR tree-optimization/57322
12833 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
12834 UID of the statement added to the BB to be 1.
12836 2013-05-21 Jakub Jelinek <jakub@redhat.com>
12838 PR tree-optimization/57331
12839 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
12840 of conversion from pointer type to integral type with integer.
12842 2013-05-21 Martin Jambor <mjambor@suse.cz>
12845 * ipa-prop.c (ipa_read_node_info): Process param_used and
12846 controlled_uses in the same order as when writing.
12848 2013-05-21 Magnus Granberg <baldrick@free.fr>
12851 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
12853 2013-05-21 Richard Biener <rguenther@suse.de>
12855 PR tree-optimization/57318
12856 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
12857 estimate stmts with side-effects as likely eliminated.
12859 2013-05-21 Richard Biener <rguenther@suse.de>
12861 PR tree-optimization/57330
12862 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
12863 preserve the call stmts fntype.
12865 2013-05-21 Richard Biener <rguenther@suse.de>
12867 PR tree-optimization/57303
12868 * tree-ssa-sink.c (statement_sink_location): Improve killing
12869 stmt detection and properly handle self-assignments.
12871 2013-05-21 Christian Bruel <christian.bruel@st.com>
12873 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
12874 spanning registers. LEAF_REG_REMAP is supported only for contiguous
12875 registers. Set register size out of the PARALLEL loop.
12877 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
12880 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
12881 (*fmasf4, *fmasf4_media): New insns.
12883 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
12885 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
12886 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
12887 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
12888 (mips_idiv_insns): Update the comments to say that the returned
12889 instruction counts are in units of BASE_INSN_LENGTH.
12890 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
12891 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
12892 using 2 rather than 4 as the length of indirect MIPS16 and
12893 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
12894 length of a NOP. Don't divide MIPS16 lengths by 2.
12895 (mips16_split_long_branches): Assume a branch is long if the
12896 length is greater than 4 rather than 8.
12897 * config/mips/mips.md (length): Give MIPS16 lengths directly,
12898 rather than multiplying them by 2. Multiply instruction counts
12899 by BASE_INSN_LENGTH rather than 4.
12900 (*jump_mips16, tls_get_tp_mips16_<mode>)
12901 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
12903 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
12905 * config/mips/mips.md (extended_mips16): Remove branch case.
12906 (length): Remove duplicated extended_mips16 test.
12908 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
12910 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
12912 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
12914 * recog.h (Recog_data): Rename to...
12915 (recog_data_d): ...this.
12916 (recog_data): Update accordingly.
12917 * recog.c (recog_data): Likewise.
12918 * reload.c (save_recog_data): Likewise.
12919 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
12920 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
12922 2013-05-17 Julian Brown <julian@codesourcery.com>
12924 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
12925 found in a REG_EQUAL note, invalidate it.
12927 2013-05-17 Easwaran Raman <eraman@google.com>
12929 * tree-ssa-reassoc.c (find_insert_point): New function.
12930 (insert_stmt_after): Likewise.
12931 (get_def_stmt): Likewise.
12932 (ensure_ops_are_available): Likewise.
12933 (not_dominated_by): Likewise.
12934 (rewrite_expr_tree): Do not move statements beyond what is
12935 necessary. Remove call to swap_ops_for_binary_stmt...
12936 (reassociate_bb): ... and move it here.
12937 (build_and_add_sum): Assign UIDs for new statements.
12938 (linearize_expr): Likewise.
12939 (do_reassoc): Renumber gimple statement UIDs.
12941 2013-05-17 Jan Hubicka <jh@suse.cz>
12943 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
12945 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
12946 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
12948 (output_weakrefs): Update.
12950 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
12951 Martin Jambor <mjambor@suse.cz>
12953 PR middle-end/57276
12954 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
12955 value that corresponds to the given aggval is found in values vector.
12957 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
12959 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
12960 sse, sse2, sse3, ssse3 and sse4a flags to options.
12962 2013-05-17 David Malcolm <dmalcolm@redhat.com>
12964 * gengtype-state.c: (s_expr_writer): New class, to handle
12965 prettifying of output layout of s-expressions.
12966 (state_writer): New class, to write out gtype.state.
12967 (state_written_type_count): Move this variable into member data of
12969 (s_expr_writer::s_expr_writer): New code: constructor for new class
12970 (state_writer::state_writer(): ditto
12971 (s_expr_writer::write_new_line): New function
12972 (s_expr_writer::write_any_indent): ditto
12973 (s_expr_writer::begin_s_expr): ditto
12974 (s_expr_writer::end_s_expr): ditto
12975 (write_state_fileloc): convert to method of state_writer...
12976 (state_writer:: write_state_fileloc): ...and use methods of
12977 s_expr_writer to write indentation into the gtype.state output file
12978 to visually represent the hierarchical structure of the list
12980 (write_state_fields): ditto, renaming to...
12981 (state_writer::write_state_fields)
12982 (write_state_a_string): ditto, renaming to...
12983 (state_writer::write_state_a_string)
12984 (write_state_string_option): ditto, renaming to...
12985 (state_writer::write_state_string_option)
12986 (write_state_type_option): ditto, renaming to...
12987 (state_writer::write_state_type_option)
12988 (write_state_nested_option): ditto, renaming to...
12989 (state_writer::write_state_nested_option)
12990 (write_state_option): ditto, renaming to...
12991 (state_writer::write_state_option)
12992 (write_state_options): ditto, renaming to...
12993 (state_writer::write_state_options)
12994 (write_state_lang_bitmap): ditto, renaming to...
12995 (state_writer::write_state_lang_bitmap)
12996 (write_state_version): ditto, renaming to...
12997 (state_writer::write_state_version)
12998 (write_state_scalar_type): ditto, renaming to...
12999 (state_writer::write_state_scalar_type)
13000 (write_state_string_type): ditto, renaming to...
13001 (state_writer::write_state_string_type)
13002 (write_state_undefined_type): ditto, renaming to...
13003 (state_writer::write_state_undefined_type)
13004 (write_state_struct_union_type): ditto, renaming to...
13005 (state_writer::write_state_struct_union_type)
13006 (write_state_struct_type): ditto, renaming to...
13007 (state_writer::write_state_struct_type)
13008 (write_state_user_struct_type): ditto, renaming to...
13009 (state_writer::write_state_user_struct_type)
13010 (write_state_lang_struct_type): ditto, renaming to...
13011 (state_writer::write_state_lang_struct_type)
13012 (write_state_param_struct_type): ditto, renaming to...
13013 (state_writer::write_state_param_struct_type)
13014 (write_state_pointer_type): ditto, renaming to...
13015 (state_writer::write_state_pointer_type)
13016 (write_state_array_type): ditto, renaming to...
13017 (state_writer::write_state_array_type)
13018 (write_state_gc_used): ditto, renaming to...
13019 (state_writer::write_state_gc_used)
13020 (write_state_common_type_content): ditto, renaming to...
13021 (state_writer::write_state_common_type_content)
13022 (write_state_type): ditto, renaming to...
13023 (state_writer::write_state_type)
13024 (write_state_pair_list): ditto, renaming to...
13025 (state_writer::write_state_pair_list)
13026 (write_state_pair): ditto, renaming to...
13027 (state_writer::write_state_pair)
13028 (write_state_typedefs): ditto, renaming to...
13029 (state_writer::write_state_typedefs)
13030 (write_state_structures): ditto, renaming to...
13031 (state_writer::write_state_structures)
13032 (write_state_param_structs): ditto, renaming to...
13033 (state_writer::write_state_param_structs)
13034 (write_state_variables): ditto, renaming to...
13035 (state_writer::write_state_variables)
13036 (write_state_srcdir): ditto, renaming to...
13037 (state_writer::write_state_srcdir)
13038 (write_state_files_list): ditto, renaming to...
13039 (state_writer::write_state_files_list)
13040 (write_state_languages): ditto, renaming to...
13041 (state_writer::write_state_languages)
13042 (write_state): create a state_writer instance and use it when
13043 writing out the state file
13045 2013-05-17 Mike Stump <mikestump@comcast.net>
13047 PR rtl-optimization/57304
13048 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
13049 accessing DF_REF_REAL_LOC.
13051 2013-05-17 Jakub Jelinek <jakub@redhat.com>
13053 PR rtl-optimization/57281
13054 PR rtl-optimization/57300
13055 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
13056 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
13057 what the other splitter did if the registers are dead.
13059 2013-05-17 Richard Biener <rguenther@suse.de>
13061 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
13064 2013-05-17 Jakub Jelinek <jakub@redhat.com>
13066 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
13069 2013-05-17 Marek Polacek <polacek@redhat.com>
13071 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
13072 length when doing non-zero store of storing '\0' to '\0'.
13074 2013-05-17 Jakub Jelinek <jakub@redhat.com>
13076 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
13077 vect_external_def oprnd1 with loop_vinfo, try to emit
13078 optional cast, negation and and stmts on the loop preheader
13079 edge instead of into the pattern def seq.
13081 PR tree-optimization/57051
13082 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
13083 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
13085 2013-05-16 Nick Clifton <nickc@redhat.com>
13087 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
13088 (rl78_is_naked_func): New function.
13089 (rl78_expand_prologue): Skip prologue generation for naked functions.
13090 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
13091 * doc/extend.texi (naked): Add RL78 to the list of processors
13092 that supports this attribute.
13094 2013-05-16 Jeff Law <law@redhat.com>
13096 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
13098 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
13100 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
13101 cache parameters using detect_caches_amd also for CYRIX,
13102 NSC and TM2 signatures.
13104 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
13105 Dzianis Kahanovich <mahatma@eu.by>
13109 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
13110 VIA/Centaur processors and determine their cache parameters
13111 using detect_caches_amd.
13113 2013-05-16 Teresa Johnson <tejohnson@google.com>
13115 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
13116 (rtl_verify_edges): New function.
13117 (rtl_verify_bb_insns): Ditto.
13118 (rtl_verify_bb_pointers): Ditto.
13119 (rtl_verify_bb_insn_chain): Ditto.
13120 (rtl_verify_fallthru): Ditto.
13121 (rtl_verify_bb_layout): Ditto.
13122 (rtl_verify_flow_info_1): Outline checks into new functions.
13123 (rtl_verify_flow_info): Ditto.
13125 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
13127 * cfghooks.c (copy_bbs): Add update_dominance argument.
13128 * cfghooks.h (copy_bbs): Update prototype.
13129 * tree-cfg.c (gimple_duplicate_sese_region):
13130 Add update_dominance argument.
13131 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
13132 * tree-ssa-loop-ch.c (copy_loop_headers): Update
13133 gimple_duplicate_sese_region call.
13134 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
13135 Update copy_bbs call.
13136 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
13137 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
13139 2013-05-16 Jakub Jelinek <jakub@redhat.com>
13141 * tree-vectorizer.h (NUM_PATTERNS): Increment.
13142 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
13143 vect_recog_rotate_pattern.
13144 (vect_recog_rotate_pattern): New function.
13146 2013-05-16 Jason Merrill <jason@redhat.com>
13148 * Makefile.in (LLINKER): New variable.
13149 (mostlyclean): Remove link mutex.
13150 * configure.ac: Handle --enable-link-mutex.
13151 * lock-and-run.sh: New script.
13153 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
13156 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
13159 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
13161 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
13163 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
13165 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
13166 * config/arm/arm.c (next_consecutive_mem): New function.
13167 (gen_movmem_ldrd_strd): Likewise.
13168 * config/arm/arm.md (movmemqi): Update condition and code.
13169 (unaligned_loaddi, unaligned_storedi): New patterns.
13171 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
13173 * config.gcc: Obsolete *-*-solaris2.9*.
13174 * doc/install.texi (Specific, *-*-solaris2*): Document it.
13176 2013-05-16 Richard Biener <rguenther@suse.de>
13178 * passes.c (init_optimization_passes): Move pass_parallelize_loops
13179 earlier, after GRAPHITE transforms and IV canonicalization.
13181 2013-05-16 Jakub Jelinek <jakub@redhat.com>
13183 * omp-low.c (extract_omp_for_data): For collapsed loops,
13184 if at least one of the loops is known at compile time to
13185 iterate zero times, set count to 0.
13186 (expand_omp_regimplify_p): New function.
13187 (expand_omp_for_generic): For collapsed loops, if at least
13188 one of the loops isn't known to iterate at least once,
13189 add runtime check with setting count to 0.
13190 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
13191 For unsigned types if it isn't known at compile time that
13192 the loop will iterate at least once, add runtime check to bypass
13193 the whole loop if initial condition isn't true.
13195 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
13197 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
13199 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
13201 PR middle-end/57286
13202 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
13203 transformations to avoid an infinite loop.
13205 2013-05-16 Marek Polacek <polacek@redhat.com>
13207 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
13209 2013-05-15 Leif Ekblad <leif@rdos.net>
13211 * config/i386/i386.c (ix86_decompose_address): Use
13212 DEFAULT_TLS_SEG_REG to access TLS segment register.
13213 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
13214 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
13215 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
13217 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
13220 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
13221 sibling calls to functions that would normally be lazily bound,
13222 unless $gp is call-clobbered.
13224 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
13226 * config/i386/i386.c (ix86_option_override_internal): Update
13227 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
13228 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
13229 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
13230 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
13232 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
13234 2013-05-15 Andreas Schwab <schwab@suse.de>
13236 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
13237 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
13240 2013-05-15 Teresa Johnson <tejohnson@google.com>
13242 * loop-unroll.c (report_unroll_peel): Check decision before
13243 emitting unroll/peel message.
13245 2013-05-15 Teresa Johnson <tejohnson@google.com>
13247 * function.h (has_bb_partition): New rtl_data flag.
13248 (bb_reorder_complete): Ditto.
13249 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
13250 instead of flag_reorder_blocks_and_partition.
13251 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
13252 with some enhancements.
13253 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
13254 * bb-reorder.c (connect_traces): Check for has_bb_partition
13255 instead of flag_reorder_blocks_and_partition.
13256 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
13257 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
13258 verify_hot_cold_block_grouping.
13259 (partition_hot_cold_basic_blocks): Set has_bb_partition.
13261 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
13264 * config/arm/predicates.md (call_insn_operand): New predicate.
13265 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
13266 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
13267 if insn is not a tail call.
13268 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
13270 * config/arm/arm.h (enum reg_class): New caller save register class.
13271 (REG_CLASS_NAMES): Likewise.
13272 (REG_CLASS_CONTENTS): Likewise.
13273 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
13276 2013-05-15 Richard Biener <rguenther@suse.de>
13278 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
13279 of MSG_OPTIMIZED_LOCATIONS.
13280 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
13281 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
13283 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
13284 of MSG_OPTIMIZED_LOCATIONS.
13285 (execute_vect_slp): Likewise.
13286 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
13287 (vect_create_cond_for_alias_checks): Likewise.
13288 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
13289 (vect_recog_widen_mult_pattern): Likewise.
13290 (vect_recog_widen_sum_pattern): Likewise.
13291 (vect_recog_over_widening_pattern): Likewise.
13292 (vect_recog_widen_shift_pattern): Likewise.
13293 (vect_recog_vector_vector_shift_pattern): Likewise.
13294 (vect_recog_divmod_pattern): Likewise.
13295 (vect_recog_mixed_size_cond_pattern): Likewise.
13296 (vect_recog_bool_pattern): Likewise.
13297 (vect_pattern_recog_1): Likewise.
13299 2013-05-15 Martin Jambor <mjambor@suse.cz>
13301 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
13302 non-functions to builtin_unreachable.
13303 * ipa-inline-transform.c (inline_call): Do not assert estimates were
13304 correct when new direct edges were discovered.
13306 2013-05-15 Martin Jambor <mjambor@suse.cz>
13308 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
13309 header, print symbol order instead of node uid, print more information
13310 about indirect edge targets.
13311 (ipa_make_edge_direct_to_target): Print symbol order instead of node
13313 (ipa_make_edge_direct_to_target): Likewise.
13314 (remove_described_reference): Likewise.
13315 (propagate_controlled_uses): Likewise.
13316 (ipa_print_node_params): Also print symbol order.
13317 (ipcp_transform_function): Print symbol order instead of node uids.
13318 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
13319 (cgraph_get_create_real_symbol_node): Likewise.
13320 * ipa-cp.c (print_lattice): Likewise.
13321 (print_all_lattices): Likewise.
13322 (determine_versionability): Likewise.
13323 (initialize_node_lattices): Likewise.
13324 (estimate_local_effects): Likewise.
13325 (update_profiling_info): Likewise.
13326 (create_specialized_node): Likewise.
13327 (perhaps_add_new_callers): Likewise.
13328 (decide_about_value): Likewise.
13329 (decide_whether_version_node): Likewise.
13330 (identify_dead_nodes): Likewise.
13331 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
13332 (dump_inline_summary): Likewise.
13333 (estimate_node_size_and_time): Likewise.
13334 (inline_analyze_function): Likewise.
13335 * ipa-inline.c (report_inline_failed_reason): Likewise.
13336 (want_early_inline_function_p): Likewise.
13337 (edge_badness): Likewise.
13338 (update_edge_key): Likewise.
13339 (inline_small_functions): Likewise. Add dumping of order to two other
13341 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
13342 instead of node uids.
13343 (propagate_pure_const): Likewise.
13344 (propagate_pure_const): Likewise.
13345 * ipa-utils.c (dump_cgraph_node_set): Likewise.
13346 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
13347 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
13349 * tree-pretty-print.c (dump_function_header): Likewise.
13350 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
13351 Print symbol order instead of node uids.
13353 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13355 * config/s390/s390.c (s390_register_move_cost): Don't impose the
13356 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
13358 2013-05-15 Richard Biener <rguenther@suse.de>
13360 PR tree-optimization/57275
13361 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
13362 return value for fail to do runtime alias checks for gather loads.
13364 2013-05-15 Jan Hubicka <jh@suse.cz>
13368 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
13369 weakrefs are not external.
13370 (lto_symtab_merge_decls): Fix thinko when dealing with
13371 non-lto_symtab decls.
13372 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
13373 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
13374 * varpool.c (dump_varpool_node): Dump more flags.
13376 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
13378 * config/i386/i386.c (processor_alias_table): Add instruction
13379 FSGSBASE for AMD bdver3 architecture.
13381 2013-05-14 Jakub Jelinek <jakub@redhat.com>
13383 * tree.c (warn_deprecated_use): Print file:line using locus color.
13384 * diagnostic.c (diagnostic_report_current_module): Print file:line
13385 and file:line:column using locus color.
13387 2013-05-14 Mike Stump <mikestump@comcast.net>
13389 * gdbinit.in: Add __null.
13391 2013-05-14 Mike Stump <mikestump@comcast.net>
13393 * recog.h: Rename struct recog_data to Recog_data.
13394 * recog.c: Likewise.
13395 * reload.c (can_reload_into): Likewise.
13396 * config/picochip/picochip.c: Likewise.
13398 2013-05-14 Mike Stump <mikestump@comcast.net>
13400 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
13402 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
13404 * resource.h (struct resources): Remove unch_memory member.
13405 (CLEAR_RESOURCE): Don't clear unch_memory.
13406 * resource.c (mark_referenced_resources): Don't set it.
13407 (mark_set_resources): Likewise.
13408 (mark_target_live_regs): Don't clear it.
13409 (init_resource_info): Likewise.
13410 * reorg.c (resource_conflicts_p): Don't compare it.
13411 (redundant_insn): Don't set it.
13413 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
13415 * emit-rtl.c (next_label): Remove unused function.
13416 (skip_consecutive_labels, link_cc0_insns): Move to ...
13417 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
13418 only place where these functions are used, and make them static.
13420 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
13422 * fold-const.c (fold_negate_expr): Handle vectors.
13423 (fold_truth_not_expr): Make it static.
13424 (fold_invert_truthvalue): New static function.
13425 (invert_truthvalue_loc): Handle vectors. Do not call
13426 fold_truth_not_expr directly.
13427 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
13428 <TRUTH_NOT_EXPR>: Do not cast to boolean.
13429 (fold_comparison): Handle vector constants.
13430 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
13431 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
13432 * tree.h (fold_truth_not_expr): Remove declaration.
13434 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
13436 * config/aarch64/aarch64-simd.md
13437 (aarch64_vcond_internal<mode>): Rename to...
13438 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
13439 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
13440 float modes. Clarify all iterator modes.
13441 (vcond<mode><mode>): Use new name for vcond expanders.
13442 (vcond<v_cmp_result><mode>): Likewise.
13443 (vcondu<mode><mode>: Likewise.
13444 * config/aarch64/iterators.md (VDQF_COND): New.
13446 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
13449 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
13450 variable for the shift amount. Check that we shift by non-negative
13453 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
13456 * config/arm/arm.h (EPILOGUE_USES): Only return true
13457 for LR_REGNUM after epilogue_completed.
13459 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
13461 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
13462 is error_mark_node.
13464 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
13467 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
13468 and Solaris 11+/x86 with gld.
13469 * configure: Regenerate.
13471 2013-05-14 Jakub Jelinek <jakub@redhat.com>
13473 * expmed.c (expand_shift_1): Canonicalize rotates by
13474 constant bitsize / 2 to bitsize - 1.
13475 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
13476 case ROTATERT>: Likewise.
13479 2013-05-10 Jakub Jelinek <jakub@redhat.com>
13481 * config/i386/i386.md (rotateinv): New code attr.
13482 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
13483 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
13484 roll $31, %eax, etc.
13486 2013-05-14 Richard Biener <rguenther@suse.de>
13488 PR middle-end/57235
13489 * tree-eh.c (sink_clobbers): Give up for successors with
13490 multiple predecessors and no virtual uses.
13492 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
13494 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
13495 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
13497 2013-05-14 Jakub Jelinek <jakub@redhat.com>
13499 PR middle-end/57251
13500 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
13501 the case when both op0 and op1 have VOIDmode.
13503 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
13505 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
13506 in multiply-accumulate mode.
13508 2013-05-13 Guozhi Wei <carrot@google.com>
13510 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
13512 2013-05-13 Kai Tietz <ktietz@redhat.com>
13515 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
13516 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
13517 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
13518 * config/i386/i386.c (ix86_option_override_internal): Likewise.
13519 (ix86_expand_prologue): Likewise.
13520 (ix86_expand_split_stack_prologue): Likewise.
13521 (legitimate_pic_address_disp_p): Likewise.
13522 (legitimize_pic_address): Likewise.
13523 (legitimize_tls_address): Likewise.
13524 (legitimize_pe_coff_symbol): Likewise.
13525 (output_pic_addr_const): Likewise.
13526 (construct_plt_address): Likewise.
13527 (ix86_expand_call): Likewise.
13528 (x86_output_mi_thunk): Likewise.
13529 (x86_function_profiler): Likewise.
13531 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
13533 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
13534 similar switch cases.
13535 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
13536 (aarch64_simd_mov_to_<mode>low): Delete.
13537 (aarch64_simd_mov_to_<mode>high): Delete.
13538 (move_lo_quad_<mode>): Add w<-r alternative.
13539 (aarch64_simd_move_hi_quad_<mode>): Likewise.
13540 (aarch64_simd_mov_from_*): Update type attribute.
13541 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
13544 2013-05-13 Jan Hubicka <jh@suse.cz>
13546 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
13547 * config/i386/i386.c (ix86_compute_frame_layout,
13548 ix86_expand_epilogue, emit_i387_cw_initialization,
13549 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
13550 ix86_local_alignment): Fix use of size/speed predicates.
13552 2013-05-13 Jakub Jelinek <jakub@redhat.com>
13554 PR tree-optimization/45216
13555 PR tree-optimization/57157
13556 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
13557 the (-Y) & (B - 1) variant if OP is |.
13558 * expmed.c (expand_shift_1): For rotations by const0_rtx just
13559 return shifted. Use (-op1) & (prec - 1) as other_amount
13560 instead of prec - op1.
13562 2013-05-13 Martin Jambor <mjambor@suse.cz>
13564 PR middle-end/42371
13565 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
13566 (ipa_constant_data): New type.
13567 (ipa_jump_func): Use ipa_constant_data to hold information about
13568 constant jump functions.
13569 (ipa_get_jf_constant): Adjust to jump function type changes.
13570 (ipa_get_jf_constant_rdesc): New function.
13571 (ipa_param_descriptor): New field controlled_uses.
13572 (ipa_get_controlled_uses): New function.
13573 (ipa_set_controlled_uses): Likewise.
13574 * ipa-ref.h (ipa_find_reference): Declare.
13575 * ipa-prop.c (ipa_cst_ref_desc): New type.
13576 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
13578 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
13579 New parameter cs. Adjust all callers.
13580 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
13581 (remove_described_reference): New function.
13582 (jfunc_rdesc_usable): Likewise.
13583 (try_make_edge_direct_simple_call): Decrement controlled use count,
13584 attempt to remove reference if it hits zero.
13585 (combine_controlled_uses_counters): New function.
13586 (propagate_controlled_uses): Likewise.
13587 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
13588 (ipa_edge_duplication_hook): Duplicate reference descriptions.
13589 (ipa_print_node_params): Print described use counter.
13590 (ipa_write_jump_function): Adjust to jump function type changes.
13591 (ipa_read_jump_function): New parameter CS, pass it to
13592 ipa_set_jf_constant. Adjust caller.
13593 (ipa_write_node_info): Stream controlled use count
13594 (ipa_read_node_info): Likewise.
13595 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
13597 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
13598 count. Remove cloning-added reference if it reaches zero.
13599 * ipa-ref.c (ipa_find_reference): New function.
13601 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
13603 * config/i386/i386.c (processor_target_table): Modified default
13604 alignment values for AMD BD and BT architectures.
13606 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
13608 * tree-vect-generic.c (uniform_vector_p): Move ...
13609 * tree.c (uniform_vector_p): ... here.
13610 * tree.h (uniform_vector_p): Declare it.
13611 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
13614 2013-05-13 Jakub Jelinek <jakub@redhat.com>
13616 PR tree-optimization/57230
13617 * tree-ssa-strlen.c (handle_char_store): Record length for
13618 array store from STRING_CST.
13620 PR tree-optimization/57230
13621 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
13624 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
13626 * config/epiphany/epiphany.c (epiphany_init): Check size of
13627 NUM_MODES_FOR_MODE_SWITCHING.
13628 (epiphany_expand_prologue):
13629 Remove CONFIG_REGNUM initial value handling code.
13630 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
13631 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
13632 (emit_set_fp_mode, epiphany_mode_after): Likewise.
13633 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
13634 Don't return 1 for FP_MODE_NONE.
13635 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
13636 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
13637 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
13638 * config/epiphany/epiphany.md (save_config): New pattern.
13640 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
13642 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
13644 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
13646 * config/i386/i386.md (memory): Handle sseishft1.
13647 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
13648 (*vec_extractv2di_1): Ditto.
13650 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
13652 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
13655 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
13657 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
13658 Add mthumb/march=armv7-a multilib.
13659 Add mthumb/march=armv7-r multilib.
13660 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
13662 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
13664 * config/v850/t-rtems: Add more multilibs.
13666 2013-05-10 Richard Biener <rguenther@suse.de>
13668 PR tree-optimization/57214
13669 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
13670 not propagate from SSA names that occur in abnormal PHI nodes.
13672 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
13674 * stor-layout.c (element_precision): New function.
13675 * machmode.h (element_precision): Declare it.
13676 * tree.c (build_minus_one_cst): New function.
13677 (element_precision): Likewise.
13678 * tree.h (build_minus_one_cst): Declare new function.
13679 (element_precision): Likewise.
13680 * fold-const.c (operand_equal_p): Use element_precision.
13681 (fold_binary_loc): Handle vector types.
13682 * convert.c (convert_to_integer): Use element_precision.
13683 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
13686 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
13688 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
13689 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
13690 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
13691 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
13692 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
13693 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
13694 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
13695 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
13696 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
13697 (Uuw8): New constraints.
13698 (Usb4): Move into alphabetical order.
13699 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
13700 (sd8_operand, ub8_operand, uw8_operand): New predicates.
13701 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
13702 previously unnamed patterns.
13703 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
13704 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
13705 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
13706 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
13707 of set_attr_alternative/if_then_else. Use extended_mips16 instead
13708 of specific lengths.
13710 2013-05-10 Jakub Jelinek <jakub@redhat.com>
13712 * config/i386/i386.md (rotateinv): New code attr.
13713 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
13714 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
13715 roll $31, %eax, etc.
13717 PR tree-optimization/45216
13718 PR tree-optimization/57157
13719 * tree-ssa-forwprop.c (simplify_rotate): New function.
13720 (ssa_forward_propagate_and_combine): Call it.
13722 2013-05-10 Richard Biener <rguenther@suse.de>
13724 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
13725 disable peeling when we version for aliasing.
13726 (vector_alignment_reachable_p): Honor explicit user alignment.
13727 (vect_supportable_dr_alignment): Likewise.
13728 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
13729 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
13730 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
13731 then peeling to arrange for the cost-model check to come first.
13733 2013-05-10 Alan Modra <amodra@gmail.com>
13735 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
13736 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
13737 * configure: Regenerate.
13739 2013-05-10 Alan Modra <amodra@gmail.com>
13742 * varasm.c (default_elf_select_section): Move !DECL_P check..
13743 (get_named_section): ..to here before calling get_section_name.
13745 (default_section_type_flags): Add DECL_P check.
13746 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
13747 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
13749 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
13751 * config/epiphany/epiphany.c (epiphany_expand_prologue):
13752 When using gen_stack_adjust_str with a register offset, add a
13753 REG_FRAME_RELATED_EXPR note.
13755 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
13757 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
13758 (*vec_extractv4si_zext_mem): Ditto.
13759 (*vec_extractv2di): Add 0->x and x->x alternatives.
13760 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
13761 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
13763 2013-05-09 Jason Merrill <jason@redhat.com>
13765 N3639 C++1y VLA support
13766 * gimplify.c (gimplify_vla_decl): Don't touch an existing
13769 * tree.c (build_constructor_va): New.
13770 * tree.h: Declare it.
13772 2013-05-09 Martin Jambor <mjambor@suse.cz>
13775 * gimple-fold.c (canonicalize_constructor_val): Call
13776 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
13778 2013-05-09 Jan Hubicka <jh@suse.cz>
13779 Richard Biener <rguenther@suse.de>
13782 * symtab.c (symtab_make_decl_local): Do not add private names.
13784 2013-05-09 Jan Hubicka <jh@suse.cz>
13787 * symtab.c (insert_to_assembler_name_hash): Handle clones.
13788 (unlink_from_assembler_name_hash): Likewise.
13789 (symtab_prevail_in_asm_name_hash, symtab_register_node,
13790 symtab_unregister_node, symtab_initialize_asm_name_hash,
13791 change_decl_assembler_name): Update.
13793 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
13795 * config/aarch64/aarch64.md: New movtf split.
13796 (*movtf_aarch64): Update.
13797 (aarch64_movdi_tilow): Handle TF modes and rename to
13798 aarch64_movdi_<mode>low.
13799 (aarch64_movdi_tihigh): Handle TF modes and rename to
13800 aarch64_movdi_<mode>high
13801 (aarch64_movtihigh_di): Handle TF modes and rename to
13802 aarch64_mov<mode>high_di
13803 (aarch64_movtilow_di): Handle TF modes and rename to
13804 aarch64_mov<mode>low_di
13805 (aarch64_movtilow_tilow): Remove spurious whitespace.
13806 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
13808 (aarch64_print_operand): Update.
13810 2013-05-09 Alan Modra <amodra@gmail.com>
13812 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
13814 * configure: Regenerate.
13816 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
13818 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
13819 splitter preparation statements.
13820 * config/i386/sse.md (*vec_extract* splitters): Ditto.
13821 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
13824 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13826 * gimple-ssa-strength-reduction.c (count_candidates): Change
13827 return value to int.
13828 (analyze_candidates_and_replace): Change type of length to int.
13830 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
13832 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
13833 (*vec_extract<mode>): Use VI12_128 mode iterator.
13834 (*vec_extract<mode>_mem): Ditto.
13835 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
13838 2013-05-08 Diego Novillo <dnovillo@google.com>
13843 2012-08-17 Diego Novillo <dnovillo@google.com>
13846 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
13847 * config.in: Regenerate.
13848 * configure: Regenerate.
13849 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
13851 2013-05-08 Jan Hubicka <jh@suse.cz>
13854 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
13855 * cgraph.h (symtab_node_base): Add unique_name.
13856 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
13857 input_overwrite_node, input_varpool_node): Stream unique_name.
13858 * cgraphclones.c (cgraph_create_virtual_clone,
13859 cgraph_function_versioning): Set unique_name.
13860 * ipa.c (function_and_variable_visibility): Set unique_name.
13862 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13864 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
13865 (alloc_cand_and_find_basis): Restrict conditional candidate
13866 processing to CAND_MULTs.
13868 2013-05-08 Jan Hubicka <jh@suse.cz>
13871 lto-symtab.c (lto_symtab_symbol_p): New function.
13872 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
13873 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
13874 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
13875 Skip static symbols.
13877 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
13879 PR tree-optimization/57200
13880 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
13881 Only call inform if the preceding warning_at returns true.
13883 2013-05-07 Han Shen <shenhan@google.com>
13885 * cfgexpand.c (record_or_union_type_has_array_p): New function.
13886 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
13887 * common.opt (fstack-protector-strong): New option.
13888 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
13889 * doc/invoke.texi (Optimization Options): Document
13890 "-fstack-protector-strong".
13891 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
13893 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
13895 * config/mips/mips.c (mips_machine_reorg2): Return 0.
13897 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
13899 * ira.c (update_equiv_regs): Add insn having equiv memory even if
13900 it is not lhs of the insn.
13901 (setup_reg_equiv): Remove insn having equiv memory which it is not
13903 * lra-constraints.c (process_address): Try to improve generation
13904 code for address base + disp.
13905 (lra_constraints): Make correct the code for checking insn setting
13906 up backward equivalence. Remove insn only if it is in the init
13908 * lra-eliminations.c (update_reg_eliminate): Change return value.
13909 (lra_eliminate): Use the result.
13911 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
13913 * config/i386/sse.md (ssescalarnummask): New mode attribute.
13914 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
13915 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
13916 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
13917 register target operands.
13918 (*vec_extractv8hi_sse2): New pattern.
13919 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
13920 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
13921 (*vec_extract<mode>_mem): New insn and split pattern.
13923 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
13925 * config/arm/arm.c (arm_asan_shadow_offset): New function.
13926 (TARGET_ASAN_SHADOW_OFFSET): Define.
13927 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
13928 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
13930 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13932 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
13933 (incr_vec_index): Return -1 if increment not found.
13934 (create_add_on_incoming_edge): Assert if increment not found.
13935 (record_increment): Limit number of increments recorded.
13936 (all_phi_incrs_profitable): Return false if an increment not found.
13937 (replace_profitable_candidates): Don't process increments that were
13939 (analyze_candidates_and_replace): Limit size of incr_vec.
13941 2013-05-07 Richard Biener <rguenther@suse.de>
13943 * calls.c (special_function_p): setjmp-like functions are leaf.
13944 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
13945 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
13947 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
13949 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
13950 (aarch64_simd_mov<mode>): New expander.
13951 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
13952 (aarch64_simd_mov_to_<mode>high): Likewise.
13953 (aarch64_simd_mov_from_<mode>low): Likewise.
13954 (aarch64_simd_mov_from_<mode>high): Likewise.
13955 (aarch64_dup_lane<mode>): Update.
13956 (aarch64_dup_lanedi): New instruction pattern.
13957 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
13958 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
13960 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13962 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
13963 (replace_mult_candidate): Remove unnecessary argument; remove
13964 unnecessary parameter from call to introduce_cast_before_cand.
13965 (replace_unconditional_candidate): Remove unnecessary parameter
13966 from call to replace_mult_candidate.
13967 (replace_conditional_candidate): Likewise.
13968 (insert_initializers): Use make_temp_ssa_name.
13969 (introduce_cast_before_cand): Remove unnecessary argument; use
13970 make_temp_ssa_name.
13971 (replace_one_candidate): Remove unnecessary argument; remove
13972 unnecessary parameter from calls to introduce_cast_before_cand.
13973 (replace_profitable_candidates): Remove unnecessary parameters
13974 from calls to replace_one_candidate.
13976 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13978 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
13979 phi def as possibly hiding a basis for a CAND_ADD whose operands
13980 have been commuted in the analysis.
13981 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
13983 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
13985 * config/aarch64/aarch64.md
13986 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
13987 shift value between 0-4.
13989 2013-05-07 Richard Biener <rguenther@suse.de>
13991 * double-int.h (rshift): New overload.
13992 * double-int.c (rshift): New function.
13993 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
13994 (create_reference_ops_from_ref): Remove.
13995 (vn_reference_insert): Use shared ops for constructing the
13996 reference and copy it.
13998 2013-05-07 Richard Biener <rguenther@suse.de>
14000 PR middle-end/57190
14001 * tree-eh.c (sink_clobbers): Properly propagate
14002 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
14004 2013-05-07 Jakub Jelinek <jakub@redhat.com>
14006 PR tree-optimization/57149
14007 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
14008 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
14009 collect_phi_def_edges, execute_late_warn_uninitialized): Use
14010 uninit_undefined_value_p instead of ssa_undefined_value_p.
14013 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
14014 for modifier == EXPAND_INITIALIZER.
14016 2013-05-07 Anton Blanchard <anton@samba.org>
14018 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
14019 for powerpc64 little endian.
14020 * configure: Regenerate.
14022 2013-05-06 Graham Stott <grahams@btinternet.com>
14024 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
14025 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
14026 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
14027 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
14029 2013-05-06 Graham Stott <grahams@btinternet.com>
14031 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
14032 codes which allow non-lvalues.
14034 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
14036 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
14037 components are all 1s.
14038 (integer_minus_onep): New function.
14039 * tree.h (integer_minus_onep): Declare it.
14040 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
14041 integer_minus_onep instead of integer_all_onesp.
14043 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
14046 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
14047 variations of these patterns.
14049 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
14051 * config/i386/i386.md (isa): Add x64_sse4 member.
14052 (enabled): Handle x64_sse4.
14053 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
14054 instruction for 64bit SSE4_1 targets. Update insn attributes.
14055 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
14056 instruction for SSE4_1 targets. Update insn attributes.
14057 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
14058 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
14059 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
14061 (*vec_extractv4si): Rename from *sse4_1_pextrd.
14062 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
14063 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
14065 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
14068 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
14070 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
14072 * final.c (do_assembler_dialects): Don't handle curly braces and
14073 vertical bar escaped by % as dialect delimiters.
14074 (output_asm_insn): Print curly braces and vertical bar if escaped
14075 by % and ASSEMBLER_DIALECT defined.
14076 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
14077 * doc/tm.texi: Regenerated.
14079 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
14081 * config/mips/mips.c: Include tree-pass.h.
14082 (mips_reorg): Split in pre- and post-dbr_schedule parts.
14083 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
14084 (pass_mips_machine_reorg2): New machine specific pass.
14085 (insert_pass_mips_machine_reorg2): New pass plugin definition.
14086 (mips_option_override): Register the new pass.
14087 * rtl.h (cleanup_barriers): Remove prototype.
14088 (dbr_schedule): Likewise.
14089 * jump.c (cleanup_barriers): Make static.
14090 * reorg.c (dbr_schedule): Likewise.
14092 2013-05-06 Richard Biener <rguenther@suse.de>
14094 PR tree-optimization/57185
14095 * tree-parloops.c (add_field_for_reduction): Handle anonymous
14096 SSA names properly.
14098 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
14101 * config/i386/i386.c (add_parameter_dependencies): Add dependence
14102 between "first_arg" and "insn", not "last" and "insn".
14104 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14106 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
14107 (find_candidates_in_block): Re-enable slsr_process_phi.
14108 (create_phi_basis): Fix double counting of candidate adjustment.
14110 2013-05-06 Richard Biener <rguenther@suse.de>
14112 PR middle-end/57147
14113 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
14114 the edge is also fallthru, preserve it and just clear the
14116 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
14117 also complex, preserve that and just clear the fallthru flag.
14118 * tree-inline.c (update_ssa_across_abnormal_edges): Also
14119 update virtual operands.
14121 2013-05-06 Alan Modra <amodra@gmail.com>
14123 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
14124 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
14125 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
14126 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
14127 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
14128 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
14130 2013-05-06 Alan Modra <amodra@gmail.com>
14132 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
14133 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
14134 (DEFAULT_ASM_ENDIAN): Define.
14135 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
14136 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
14137 Update -K PIC clause from sysv4.h.
14138 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
14139 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
14141 2013-05-06 Alan Modra <amodra@gmail.com>
14143 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
14144 twice for little-endian.
14145 (ashrdi3_no_power, ashrdi3): Support little-endian.
14147 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
14150 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
14151 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
14153 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
14154 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
14155 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
14158 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
14159 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
14161 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
14162 * config/pa/pa.opt: Make mbig-switch a no-op.
14163 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
14164 (CASE_VECTOR_MODE): Always return SImode.
14165 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
14166 for the !TARGET_BIG_SWITCH case.
14167 * config/pa/pa-linux.h: Likewise.
14168 * config/pa/pa-openbsd.h: Likewise.
14169 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
14170 * config/pa/pa.md (short_jump): Remove define_insn.
14171 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
14172 (casesi0): Remove define_insn.
14173 (type): Remove btable_branch.
14174 (pa_combine_type): Likewise.
14175 (in_nullified_branch_delay): Likewise.
14176 (in_call_delay): Likewise.
14177 (define_delay): Likewise.
14178 (define_insn_reservation "Z3"): Likewise.
14179 (define_insn_reservation "Z4"): Likewise.
14180 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
14181 (pa_adjust_insn_length): Remove adjustment for btable branches.
14182 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
14185 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
14187 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
14188 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
14189 Add m->r,x alternatives.
14190 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
14191 splitters using SWI48x mode iterator.
14192 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
14193 TARGET_64BIT. Add m->x alternative.
14194 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
14195 Add o->x alternative. Enable for TARGET_SSE.
14196 (sse_storeq): Remove expander.
14197 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
14198 with memory input operand.
14199 (*vec_extractv2di_1 splitter): New.
14200 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
14201 * config/i386/i386.md (ssevecmodelower): New mode attribute.
14203 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
14205 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
14206 (INT_LOWPART): Delete.
14207 (extract_MB): Adjust.
14208 (extract_ME): Adjust.
14209 (print_operand): Adjust.
14211 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
14213 * config/rs6000/predicates.md (reg_or_add_cint_operand,
14214 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
14215 (reg_or_logical_cint_operand, easy_fp_constant,
14216 logical_const_operand): Delete "CONST_DOUBLE" case.
14217 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
14218 "HOST_BITS_PER_WIDE_INT == 64" test.
14219 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
14220 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
14221 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
14222 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
14224 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
14225 CONST_DOUBLE DImode/VOIDmode case.
14226 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
14227 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
14228 CONST_DOUBLE VOIDmode case.
14229 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
14230 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
14231 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
14232 Delete CONST_DOUBLE case.
14233 (splitters for mov FMOVE64 const_double): Delete
14234 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
14235 "HOST_BITS_PER_WIDE_INT >= 64" test.
14236 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
14238 (mov DI const_double): Delete.
14240 2013-05-04 Jakub Jelinek <jakub@redhat.com>
14242 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
14243 on op shows all bits zero in mode of a lowpart subreg, return zero.
14245 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
14248 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
14249 to save TFmode registers and DImode to save TImode registers for
14250 caller save operations.
14251 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
14252 mark being partially clobbered since they only use the first
14255 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
14256 and TDmode only use the upper 64-bits of each VSX register.
14258 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14260 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
14261 (find_candidates_in_block): Disable slsr_process_phi.
14263 2013-05-03 Guozhi Wei <carrot@google.com>
14265 * coverage.c (coverage_obj_init): Move the construction of gcov
14267 (build_init_ctor): ... here.
14269 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14271 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
14272 (slsr_cand_d): Redefine def_phi.
14273 (stride_status, phi_adjust_status, count_phis_status): New enums.
14274 (find_phi_def): New.
14275 (find_basis_for_base_expr): New.
14276 (find_basis_for_candidate): Handle hidden bases.
14277 (alloc_cand_and_find_basis): Handle phi candidates.
14278 (slsr_process_phi): New.
14279 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
14280 (create_mul_imm_cand): Likewise.
14281 (create_add_ssa_cand): Exclude phi base candidates.
14282 (create_add_imm_cand): Likewise.
14283 (slsr_process_cast): Likewise.
14284 (slsr_process_copy): Likewise.
14285 (find_candidates_in_block): Handle phi candidates.
14286 (dump_candidate): Likewise.
14287 (unconditional_cands): Delete.
14288 (unconditional_cands_with_known_stride_p): Delete.
14289 (phi_dependent_cand_p): New.
14290 (cand_increment): Handle phi-dependent candidates.
14291 (replace_dependent): Delete.
14292 (replace_mult_candidate): New.
14293 (replace_unconditional_candidate): New.
14294 (incr_vec_index): Move to avoid forward reference.
14295 (create_add_on_incoming_edge): New.
14296 (create_phi_basis): New.
14297 (replace_dependents): Delete.
14298 (replace_conditional_candidate): New.
14299 (phi_add_costs): New.
14300 (replace_uncond_cands_and_profitable_phis): New.
14301 (record_increment): Handle phi adjustments.
14302 (record_phi_increments): New.
14303 (record_increments): Handle phi adjustments.
14304 (phi_incr_cost): New.
14305 (lowest_cost_path): Handle phis.
14306 (total_savings): Likewise.
14307 (analyze_increments): Likewise.
14308 (ncd_with_phi): New.
14309 (ncd_of_cand_and_phis): New.
14310 (nearest_common_dominator_for_cands): Handle phi increments.
14311 (all_phi_incrs_profitable): New.
14312 (replace_profitable_candidates): Handle phi-dependent candidates.
14313 (analyze_candidates_and_replace): Likewise.
14315 2013-05-03 Teresa Johnson <tejohnson@google.com>
14318 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
14319 do not exceed REG_BR_PROB_BASE.
14321 2013-05-03 Jeff Law <law@redhat.com>
14323 PR tree-optimization/57144
14324 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
14325 operand of the condition will bit into the new type when eliminating
14326 a cast feeding a condition.
14328 2013-05-03 Jakub Jelinek <jakub@redhat.com>
14330 PR rtl-optimization/57130
14331 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
14332 of COMPARE as in_code to the recursive call if needed.
14334 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
14336 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
14337 (enabled): Handle new members.
14338 * config/i386/sse.md (*vec_concatv2si): Merge from
14339 *vec_concatv2si_sse2 and vec_concatv2si_sse.
14340 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
14342 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
14344 PR tree-optimization/57027
14345 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
14346 for fnms opportunity, check we got the prerequisite kind
14347 of tree / gimple before using accessor functions.
14349 2013-05-03 Richard Biener <rguenther@suse.de>
14351 * double-int.h (lshift): New overload without precision
14352 and arith argument.
14353 (operator *=, operator +=, operator -=): Move ...
14354 * double-int.c (operator *=, operator +=, operator -=): ... here
14355 and implement more efficiently.
14356 (mul_double_with_sign): Remove.
14357 (lshift_double): Adjust to take unsinged shift argument, push
14358 dispatching code to callers.
14359 (mul_double_wide_with_sign): Add early out for callers that
14360 are not interested in high parts or overflow.
14361 (lshift): New function.
14362 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
14363 dispatch code here.
14364 (lrotate, rrotate): Use logical shifts.
14365 * expr.c (get_inner_reference): Use lshift.
14366 * fixed-value.c (do_fixed_divide): Likewise.
14367 * tree-dfa.c (get_ref_base_and_extent): Likewise.
14368 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
14369 (indirect_refs_may_alias_p): Likewise.
14370 (stmt_kills_ref_p_1): Likewise.
14372 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
14374 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
14376 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
14378 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
14379 scalar form of FABD instruction.
14381 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
14383 * lra-constraints.c (process_alt_operands): Add checking alt
14384 number to choose the best alternative.
14386 2013-05-02 Richard Biener <rguenther@suse.de>
14388 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
14389 bitmap and its handling.
14390 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
14392 2013-05-02 Richard Biener <rguenther@suse.de>
14394 PR middle-end/57140
14395 * tree-inline.c (copy_loops): Properly handle removed loops.
14396 (copy_cfg_body): Mark destination loops for fixup if source
14397 loops needed fixup.
14399 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
14402 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
14403 generating simple_return for naked functions.
14405 2013-05-02 Martin Jambor <mjambor@suse.cz>
14407 PR middle-end/56988
14408 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
14409 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
14411 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
14412 ipa_agg_replacement_value structures.
14413 (known_aggs_to_agg_replacement_list): Likewise.
14414 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
14415 (read_agg_replacement_chain): Likewise.
14416 (ipcp_transform_function): Also check that by_ref flags match.
14418 2013-05-02 Richard Biener <rguenther@suse.de>
14420 * graphds.h (struct graph): Add obstack member.
14421 * graphds.c (new_graph): Initialize obstack and allocate
14423 (add_edge): Allocate edge from the obstack.
14424 (free_graph): Free the obstack instead of all edges and vertices.
14426 2013-05-02 Teresa Johnson <tejohnson@google.com>
14428 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
14430 * cfg.c (update_bb_profile_for_threading): Ditto.
14431 * tree-inline.c (copy_bb): Ditto.
14432 (copy_edges_for_bb): Ditto.
14433 (initialize_cfun): Ditto.
14434 (copy_cfg_body): Ditto.
14435 (expand_call_inline): Ditto.
14436 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
14437 (estimate_node_size_and_time): Ditto.
14438 (inline_merge_summary): Ditto.
14439 * cgraphclones.c (cgraph_clone_edge): Ditto.
14440 (cgraph_clone_node): Ditto.
14441 * sched-rgn.c (compute_dom_prob_ps): Ditto.
14442 (compute_trg_info): Ditto.
14444 2013-05-02 Ian Bolton <ian.bolton@arm.com>
14446 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
14447 S reg when fp attribute set.
14448 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
14450 2013-05-02 Ian Bolton <ian.bolton@arm.com>
14452 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
14454 (*and_one_cmplsi3_compare0_uxtw): Likewise.
14455 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
14456 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
14458 2013-05-02 Richard Biener <rguenther@suse.de>
14460 * tree-scalar-evolution.c (scev_info_hasher): Remove.
14461 (struct instantiate_cache_entry): New type.
14462 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
14463 (struct instantiate_cache_type): New type.
14464 (set_instantiated_value, get_instantiated_value): Remove.
14465 (get_instantiated_value_entry): New function.
14466 (instantiate_scev_name): Use the new cache and adjust.
14467 (instantiate_scev_poly): Adjust.
14468 (instantiate_scev_binary): Likewise.
14469 (instantiate_array_ref): Likewise.
14470 (instantiate_scev_convert): Likewise.
14471 (instantiate_scev_not): Likewise.
14472 (instantiate_scev_3): Likewise.
14473 (instantiate_scev_2): Likewise.
14474 (instantiate_scev_r): Likewise.
14475 (instantiate_scev): Likewise.
14476 (resolve_mixers): Likewise.
14478 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
14481 * lra-constraints.c (best_small_class_operands_num): Remove.
14482 (process_alt_operands): Remove small_class_operands_num. Take
14483 small classes operands into losers and only if the operand is not
14484 matched. Modify debugging output.
14485 (curr_insn_transform): Remove best_small_class_operands_num.
14488 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14490 * config/aarch64/aarch64-builtins.c
14491 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
14492 * config/aarch64/aarch64-simd-builtins.def
14493 (reduc_splus_): Add new modes.
14494 (reduc_uplus_): New.
14495 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
14496 (reduc_uplus_v4sf): Likewise.
14497 (reduc_splus_v4sf): Likewise.
14498 (aarch64_addv<mode>): Likewise.
14499 (reduc_uplus_<mode>): Likewise.
14500 (reduc_splus_<mode>): Likewise.
14501 (aarch64_addvv2di): Likewise.
14502 (reduc_uplus_v2di): Likewise.
14503 (reduc_splus_v2di): Likewise.
14504 (aarch64_addvv2si): Likewise.
14505 (reduc_uplus_v2si): Likewise.
14506 (reduc_splus_v2si): Likewise.
14507 (reduc_<sur>plus_<mode>): New.
14508 (reduc_<sur>plus_v2di): Likewise.
14509 (reduc_<sur>plus_v2si): Likewise.
14510 (reduc_<sur>plus_v4sf): Likewise.
14511 (aarch64_addpv4sf): Likewise.
14512 * config/aarch64/arm_neon.h
14513 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
14514 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
14515 add UNSPEC_SADDV, UNSPEC_UADDV.
14517 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
14519 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14521 * config/aarch64/arm_neon.h
14522 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
14524 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14526 * config/aarch64/aarch64-builtins
14527 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
14529 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14531 * config/aarch64/aarch64-simd-builtins.def
14532 (reduc_smax_): New.
14533 (reduc_smin_): Likewise.
14534 (reduc_umax_): Likewise.
14535 (reduc_umin_): Likewise.
14536 (reduc_smax_nan_): Likewise.
14537 (reduc_smin_nan_): Likewise.
14540 (smax): Update for V2SF, V4SF and V2DF modes.
14543 (smin_nan): Likewise.
14544 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
14545 (<su><maxmin><mode>3): ...This, refactor.
14546 (s<maxmin><mode>3): New.
14547 (<maxmin_uns><mode>3): Likewise.
14548 (reduc_<maxmin_uns>_<mode>): Refactor.
14549 (reduc_<maxmin_uns>_v4sf): Likewise.
14550 (reduc_<maxmin_uns>_v2si): Likewise.
14551 (aarch64_<fmaxmin><mode>: Remove.
14552 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
14554 (vmin<q>_f<32,64>): Likewise.
14555 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
14557 (su): Add mappings for smax, smin, umax, umin.
14559 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
14560 (FMAXMIN): Rename as...
14561 (FMAXMIN_UNS): ...This.
14563 (fmaxminv): Likewise.
14564 (fmaxmin): Likewise.
14566 (maxmin_uns_op): Likewise.
14568 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14570 * config/aarch64/arm_neon.h
14571 (vac<ge, gt><sd>_f<32, 64>): Rename to...
14572 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
14573 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
14575 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14577 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
14578 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
14580 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14582 * config/aarch64/aarch64-simd.md
14583 (vcond<mode>_internal): Handle special cases for constant masks.
14584 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
14585 (vcondu<mode><mode>): Likewise.
14586 (vcond<v_cmp_result><mode>): New.
14588 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14590 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
14591 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
14592 * config/aarch64/aarch64-simd-builtins.def
14593 (cmeq): Update to BUILTIN_VALLDI.
14598 * config/aarch64/arm_neon.h
14599 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
14600 to builtins or C as appropriate.
14602 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
14604 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
14606 (cmhi): Rename to...
14608 * config/aarch64/aarch64-simd.md
14609 (simd_mode): Add SF.
14610 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
14611 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
14612 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
14613 (cstore<mode>_neg): ...This.
14614 * config/aarch64/iterators.md
14616 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
14617 (COMPARISONS): New.
14618 (UCOMPARISONS): Likewise.
14619 (optab): Add missing comparisons.
14625 (VCMP_S): Likewise.
14626 (VCMP_U): Likewise.
14627 (V_cmp_result): Add DF, SF modes.
14628 (v_cmp_result): Likewise.
14630 (vmtype): Likewise.
14631 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
14633 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
14635 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
14636 define_insn to define_insn_and_split.
14637 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
14638 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
14639 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
14640 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
14641 (thumb2_negscc): Likewise.
14643 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
14645 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
14647 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
14649 * config/arm/thumb2.md: Remove trailing whitespaces.
14651 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
14653 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
14654 Use gen_int_mode rather than GEN_INT.
14656 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
14658 * value-prof.c (stream_in_histogram_value): Remove the strayed
14661 2013-04-30 Richard Biener <rguenther@suse.de>
14663 PR middle-end/57122
14664 * cfghooks.c (split_edge): Properly check for the loop latch edge.
14666 2013-04-30 Richard Biener <rguenther@suse.de>
14668 PR middle-end/57107
14669 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
14671 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
14673 PR rtl-optimization/56957
14674 PR rtl-optimization/57105
14675 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
14676 variable. Use just INSN_UID for determining whether an insn
14677 should be only disconnected from the insn stream.
14678 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
14680 2013-04-30 Jakub Jelinek <jakub@redhat.com>
14682 PR tree-optimization/57104
14683 * tsan.c (instrument_expr): Don't instrument accesses to
14684 DECL_HARD_REGISTER VAR_DECLs.
14686 2013-04-30 Richard Biener <rguenther@suse.de>
14688 * function.h (loops_for_fn): New inline function.
14689 (set_loops_for_fn): Likewise.
14690 * cfgloop.h (place_new_loop): Add struct function parameter.
14691 (get_loop): Likewise.
14692 (get_loops): Likewise.
14693 (number_of_loops): Likewise.
14694 (fel_next): Adjust.
14695 (fel_init): Likewise.
14696 * cfg.c (get_loop_copy): Adjust.
14697 * cfgloop.c (flow_loops_dump): Likewise.
14698 (record_loop_exits): Likewise.
14699 (verify_loop_structure): Likewise.
14700 * cfgloopanal.c (mark_irreducible_loops): Likewise.
14701 (estimate_reg_pressure_cost): Likewise.
14702 (mark_loop_exit_edges): Likewise.
14703 * cfgloopmanip.c (place_new_loop): Likewise.
14704 (add_loop): Likewise.
14705 (duplicate_loop): Likewise.
14706 * graph.c (draw_cfg_nodes): Likewise.
14707 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
14708 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
14709 (extract_affine_chrec): Likewise.
14710 (build_scop_iteration_domain): Likewise.
14711 * graphite.c (graphite_initialize): Likewise.
14712 * ira-build.c (create_loop_tree_nodes): Likewise.
14713 (more_one_region_p): Likewise.
14714 (rebuild_regno_allocno_maps): Likewise.
14715 (mark_loops_for_removal): Likewise.
14716 (mark_all_loops_for_removal): Likewise.
14717 (remove_unnecessary_regions): Likewise.
14718 (ira_build): Likewise.
14719 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
14720 * loop-init.c (fix_loop_structure): Likewise.
14721 (gate_rtl_move_loop_invariants): Likewise.
14722 (gate_rtl_unswitch): Likewise.
14723 (gate_rtl_unroll_and_peel_loops): Likewise.
14724 (rtl_doloop): Likewise.
14725 * lto-streamer-in.c (input_cfg): Likewise.
14726 * lto-streamer-out.c (output_cfg): Likewise.
14727 * modulo-sched.c (sms_schedule): Likewise.
14728 * predict.c (tree_estimate_probability): Likewise.
14729 (tree_estimate_probability_driver): Likewise.
14730 (estimate_loops): Likewise.
14731 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
14732 (move_sese_region_to_fn): Likewise.
14733 (debug_loop_num): Likewise.
14734 * tree-chrec.c (chrec_evaluate): Likewise.
14735 (hide_evolution_in_other_loops_than_loop): Likewise.
14736 (chrec_component_in_loop_num): Likewise.
14737 (reset_evolution_in_loop): Likewise.
14738 (evolution_function_is_invariant_rec_p): Likewise.
14739 * tree-if-conv.c (main_tree_if_conversion): Likewise.
14740 * tree-inline.c (copy_loops): Likewise.
14741 (copy_cfg_body): Likewise.
14742 (tree_function_versioning): Likewise.
14743 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
14744 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
14746 (add_to_evolution_1): Likewise.
14747 (scev_const_prop): Likewise.
14748 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
14749 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
14750 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
14751 (tree_ssa_lim_initialize): Likewise.
14752 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
14753 (verify_loop_closed_ssa): Likewise.
14754 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
14755 (tree_ssa_loop_im): Likewise.
14756 (tree_ssa_loop_unswitch): Likewise.
14757 (tree_vectorize): Likewise.
14758 (check_data_deps): Likewise.
14759 (tree_ssa_loop_ivcanon): Likewise.
14760 (tree_ssa_loop_bounds): Likewise.
14761 (tree_complete_unroll): Likewise.
14762 (tree_complete_unroll_inner): Likewise.
14763 (tree_parallelize_loops): Likewise.
14764 (tree_ssa_loop_prefetch): Likewise.
14765 (tree_ssa_loop_ivopts): Likewise.
14766 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
14767 * tree-vectorizer.c (vectorize_loops): Likewise.
14769 2013-04-29 Mike Frysinger <vapier@gentoo.org>
14771 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
14772 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
14773 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
14774 with EABI_LINK_SPEC.
14776 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
14779 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
14782 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
14785 * lra-constraints.c (process_alt_operands): Discourage a bit more
14786 using memory for pseudos. Print cost dump for alternatives.
14787 Modify cost values for conflicts with early clobbers.
14788 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
14790 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
14793 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
14795 2013-04-29 Ian Bolton <ian.bolton@arm.com>
14797 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
14798 from/to S register.
14799 (movdi_aarch64): Support LDR/STR from/to D register.
14801 2013-04-29 Ian Bolton <ian.bolton@arm.com>
14803 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
14804 or higher by default.
14806 2013-04-29 Richard Biener <rguenther@suse.de>
14808 PR middle-end/57075
14809 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
14810 even if not adding abnormal edges for calls that can make
14813 2013-04-29 Richard Biener <rguenther@suse.de>
14815 PR middle-end/57103
14816 * tree-cfg.c (move_stmt_op): Fix condition under which to update
14818 (move_stmt_r): Remove redundant checking.
14820 2013-04-29 Teresa Johnson <tejohnson@google.com>
14823 * basic-block.h (apply_scale): New function.
14824 (apply_probability): Use apply_scale.
14825 * gimple-streamer-in.c (input_bb): Ditto.
14826 * lto-streamer-in.c (input_cfg): Ditto.
14827 * lto-cgraph.c (merge_profile_summaries): Ditto.
14828 * tree-optimize.c (execute_fixup_cfg): Ditto.
14829 * tree-inline.c (copy_bb): Update comment to use apply_scale.
14830 (copy_edges_for_bb): Ditto.
14831 (copy_cfg_body): Ditto.
14833 2013-04-29 Tom de Vries <tom@codesourcery.com>
14835 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
14836 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
14837 (tail_merge_optimize): Handle current_loops == NULL.
14839 2013-04-26 Jeff Law <law@redhat.com>
14841 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
14842 (simplify_cond_using_ranges): Generalize code to simplify
14843 COND_EXPRs where one argument is a constant and the other
14844 is an SSA_NAME created by an integral type conversion.
14846 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14848 * config/arm/arm.md (store_minmaxsi): Use only when
14849 optimize_insn_for_size_p.
14851 2013-04-29 Christian Bruel <christian.bruel@st.com>
14854 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
14856 2013-04-29 Richard Biener <rguenther@suse.de>
14858 PR middle-end/57089
14859 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
14860 loop tree make sure to schedule a fixup for the child as well.
14861 (expand_omp_for_generic): Properly add loops.
14862 (expand_omp_for_static_nochunk): Likewise.
14863 (expand_omp_for_static_chunk): Likewise.
14864 (expand_omp_for): For the degenerate case fixup loops.
14865 (expand_omp_sections): Fix default bb placement in loops.
14866 (expand_omp_atomic_pipeline): Properly add loops.
14868 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14870 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
14872 2013-04-29 Tom de Vries <tom@codesourcery.com>
14874 * tree-ssa-tail-merge.c: Update header comment.
14876 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14878 * config/aarch64/arm_neon.h
14879 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
14880 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
14881 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
14882 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
14883 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
14884 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
14885 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
14886 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
14888 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14890 * config/aarch64/aarch64-simd.md
14891 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
14892 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
14893 fix_trunc, fixuns_trunc.
14894 (ftrunc<VDQF:mode>2): New.
14895 * config/aarch64/iterators.md (optab): Add fix, fixuns.
14896 (fix_trunc_optab): New.
14898 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14900 * config/aarch64/aarch64-builtins.c
14901 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
14902 iceilf, lround, iroundf.
14904 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
14907 * config/i386/i386.h (enum ix86_tune_indices)
14908 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
14909 New, split from X86_TUNE_INTER_UNIT_MOVES.
14910 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
14911 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
14912 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
14913 (TARGET_INTER_UNIT_MOVES): Remove.
14914 * config/i386/i386.c (initial_ix86_tune_features): Update.
14915 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
14916 (ix86_expand_convert_uns_didf_sse): Use
14917 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
14918 (ix86_expand_vector_init_one_nonzero): Ditto.
14919 (ix86_expand_vector_init_interleave): Ditto.
14920 (inline_secondary_memory_needed): Return true for moves from SSE class
14921 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
14922 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
14923 * config/i386/constraints.md (Yi, Ym): Depend on
14924 TARGET_INTER_UNIT_MOVES_TO_VEC.
14925 (Yj, Yn): New constraints.
14926 * config/i386/i386.md (*movdi_internal): Change constraints of
14927 operand 1 from Yi to Yj and from Ym to Yn.
14928 (*movsi_internal): Ditto.
14929 (*movdf_internal): Ditto.
14930 (*movsf_internal): Ditto.
14931 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
14932 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
14933 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
14934 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
14935 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
14936 * config/i386/sse.md (movdi_to_sse): Ditto.
14937 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
14938 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
14939 TARGET_INTER_UNIT_MOVES.
14940 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
14941 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
14942 instead of TARGET_INTER_UNIT_MOVES.
14943 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
14944 operand 1 from Yi to Yj and from Ym to Yn.
14946 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14948 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
14949 (float_truncate_hi_): Likewise.
14950 (float_extend_lo_): Likewise.
14951 (float_truncate_lo_): Likewise.
14952 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
14953 (aarch64_float_extend_lo_v2df): Likewise.
14954 (vec_unpacks_hi_v4sf): Likewise.
14955 (aarch64_float_truncate_lo_v2sf): Likewise.
14956 (aarch64_float_truncate_hi_v4sf): Likewise.
14957 (vec_pack_trunc_v2df): Likewise.
14958 (vec_pack_trunc_df): Likewise.
14960 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14962 * config/aarch64/aarch64-builtins.c
14963 (aarch64_fold_builtin): Fold float conversions.
14964 * config/aarch64/aarch64-simd-builtins.def
14965 (floatv2si, floatv4si, floatv2di): New.
14966 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
14967 * config/aarch64/aarch64-simd.md
14968 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
14969 * config/aarch64/iterators.md (FLOATUORS): New.
14970 (optab): Add float, floatuns.
14971 (su_optab): Likewise.
14973 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
14975 * config/aarch64/aarch64-builtins.c
14976 (aarch64_builtin_vectorized_function): Use new names for
14978 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
14979 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
14980 (fcvtzu): Split as...
14981 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
14982 (fcvtas): Split as...
14983 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
14984 (fcvtau): Split as...
14985 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
14986 (fcvtps): Split as...
14987 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
14988 (fcvtpu): Split as...
14989 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
14990 (fcvtms): Split as...
14991 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
14992 (fcvtmu): Split as...
14993 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
14994 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
14995 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
14996 (lfrintnusf, lfrintnudf): Likewise.
14997 * config/aarch64/aarch64-simd.md
14998 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
15000 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
15001 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
15002 (fcvt_pattern): Likewise.
15004 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15006 * config/aarch64/aarch64-simd.md
15007 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
15008 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
15010 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15012 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
15013 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
15014 (vrnd<a,m,n,p>_f32): Implement using builtins.
15015 (vrnd<i,x><q>_f<32, 64>): New.
15017 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
15019 * config/aarch64/aarch64-builtins.c
15020 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
15021 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
15022 (frintz): Rename to...
15024 (frintp): Rename to...
15026 (frintm): Rename to...
15028 (frinti): Rename to...
15029 (nearbyint): ...this.
15030 (frintx): Rename to...
15032 (frinta): Rename to...
15034 * config/aarch64/aarch64-simd.md
15035 (aarch64_frint<frint_suffix><mode>): Delete.
15036 (<frint_pattern><mode>2): Convert to insn.
15037 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
15038 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
15039 (frint_pattern): Likewise.
15040 (frint_suffix): Likewise.
15042 2013-04-29 Richard Biener <rguenther@suse.de>
15044 PR tree-optimization/57081
15045 * loop-init.c: Include tree-flow.h.
15046 (loop_optimizer_finalize): Free number of iteration estimates.
15047 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
15049 2013-04-29 Jakub Jelinek <jakub@redhat.com>
15051 PR tree-optimization/57083
15052 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
15053 non-singleton shift count range, zero extend low_bound for uns case.
15055 * config/i386/predicates.md (general_vector_operand): New predicate.
15056 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
15057 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
15058 if they aren't nonimmediate operands. If their original values
15059 satisfy const_vector_equal_evenodd_p, don't shift them.
15060 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
15061 predicates. For the SSE4.1 case force operands[{1,2}] into registers
15062 if not nonimmediate_operand.
15063 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
15064 instead of register_operand.
15065 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
15067 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
15069 * stor-layout.c (finalize_size_functions): Allocate a structure and
15070 reset cfun before dumping the functions.
15072 2013-04-27 Jakub Jelinek <jakub@redhat.com>
15074 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
15077 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
15078 use xop_pmacsdqh if uns_p.
15079 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
15080 the immediate rotate count.
15082 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
15084 * rtl.h (struct rtx_def): Add comment for field jump.
15085 (LRA_SUBREG_P): New macro.
15086 * recog.c (register_operand): Check LRA_SUBREG_P.
15087 * lra.c (lra): Add note at the end of RTL code. Align non-empty
15089 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
15090 (lra_final_code_change): Skip subreg change for operators.
15091 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
15092 if there are no operand changes.
15093 * lra-constraints.c (curr_insn_set): New.
15094 (match_reload): Set LRA_SUBREG_P.
15095 (emit_spill_move): Ditto.
15096 (check_and_process_move): Use curr_insn_set. Process only single
15097 set insns. Don't initialize sec_mem_p and change_p.
15098 (simplify_operand_subreg): Use LRA_SUBREG_P.
15099 (reg_in_class_p): New function.
15100 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
15101 of #ifdef. Add code to remove cycling.
15102 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
15103 non-null disp. Reload inner instead of disp when base and index
15104 are null. Try to put lo_sum into register.
15105 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
15106 (check_and_process_move): Move code for move cost check to
15107 simple_move_p. Remove equiv_substitution.
15108 (simple_move_p): New function.
15109 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
15110 curr_insn_set. Call check_and_process_move only for single set
15111 insns. Use the new function. Move call of check_and_process_move
15112 after operand equiv substitution and address process.
15114 2013-04-26 Jakub Jelinek <jakub@redhat.com>
15117 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
15118 with nonlocal goto receivers or returns twice calls, ignore
15119 unininitialized values from abnormal edges to nl goto receiver
15120 or returns twice call.
15122 2013-04-26 Jakub Jelinek <jakub@redhat.com>
15124 PR tree-optimization/57051
15125 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
15126 and VEC_RSHIFT_EXPR if shift count is a multiple of element
15129 2013-04-26 Richard Biener <rguenther@suse.de>
15131 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
15132 (expand_omp_taskreg): Likewise. Mark loops for fixup.
15133 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
15134 (fixup_loop_arrays_after_move): New function.
15135 (move_sese_region_to_fn): Properly outline the loop tree parts
15136 of the SESE region.
15138 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
15140 * config/i386/i386.md (type, unit): Fix long lines.
15142 2013-04-26 Richard Biener <rguenther@suse.de>
15144 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
15145 (lto-streamer-out.o): Likewise.
15146 * cfgloop.c (init_loops_structure): Export, add struct function
15147 argument and adjust.
15148 (flow_loops_find): Adjust.
15149 * cfgloop.h (enum loop_estimation): Add EST_LAST.
15150 (init_loops_structure): Declare.
15151 * lto-streamer-in.c: Include cfgloop.h.
15152 (input_cfg): Input the loop tree.
15153 * lto-streamer-out.c: Include cfgloop.h.
15154 (output_cfg): Output the loop tree.
15155 (output_struct_function_base): Do not drop PROP_loops.
15157 2013-03-26 Richard Biener <rguenther@suse.de>
15159 * tree-cfg.c (execute_build_cfg): Build the loop tree.
15160 (pass_build_cfg): Provide PROP_loops.
15161 (move_sese_region_to_fn): Remove loops that are outlined into fn
15163 * tree-inline.c: Include cfgloop.h.
15164 (initialize_cfun): Do not drop PROP_loops.
15165 (copy_loops): New function.
15166 (copy_cfg_body): Copy loop structure.
15167 (tree_function_versioning): Initialize destination loop tree.
15168 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
15169 (pass_parallelize_loops): Do IL verification.
15170 * loop-init.c (loop_optimizer_init): Fixup loops if required.
15171 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
15172 the CFG make sure we fixup loops as well.
15173 * tree-ssa-tail-merge.c: Include cfgloop.h.
15174 (replace_block_by): When merging loop latches mark loops for fixup.
15175 * lto-streamer-out.c (output_struct_function_base): Drop
15176 PROP_loops for now.
15177 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
15178 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
15179 * ipa-split.c: Include cfgloop.h.
15180 (split_function): Add the new return block to the loop tree root.
15181 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
15182 whether we have removed the forwarder block.
15183 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
15184 * cfgloop.h (place_new_loop): Declare.
15185 * cfgloopmanip.c (place_new_loop): Export.
15186 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
15187 (tree-switch-conversion.o): Likewise.
15188 (tree-complex.o): Likewise.
15189 (tree-inline.o): Likewise.
15190 (tree-ssa-tailmerge.o): Likewise.
15191 (ipa-split.o): Likewise.
15192 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
15193 (tree-ssa-copy.o): Likewise.
15194 * tree-switch-conversion.c: Include cfgloop.h
15195 (process_switch): If we emit a bit-test cascade, schedule loops
15197 * tree-complex.c: Include cfgloop.h.
15198 (expand_complex_div_wide): Properly add new basic-blocks to loops.
15199 * asan.c: Include cfgloop.h.
15200 (create_cond_insert_point): Properly add new basic-blocks to
15201 loops, schedule loop fixup.
15202 * cfgloop.c (verify_loop_structure): Check that looks are not
15204 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
15206 (expand_omp_for_generic): Likewise.
15207 (expand_omp_sections): Likewise.
15208 (expand_omp_atomic_pipeline): Schedule loops for fixup.
15209 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
15210 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
15211 is initialized, not when loops are present.
15212 * tree-parloops.c (parallelize_loops): Remove checking here.
15213 * passes.c (init_optimization_passes): Schedule a copy-propagation
15214 pass before complete unrolling of inner loops.
15216 2013-04-26 Jakub Jelinek <jakub@redhat.com>
15218 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
15219 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
15220 (colorize_init): Add argument to _WIN32 version.
15221 * toplev.c: Include diagnostic-color.h.
15222 (process_options): Default to -fdiagnostics-color=auto if
15223 GCC_COLORS env var is in the environment.
15224 * common.opt (fdiagnostics-color=): Add Var and Init.
15225 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
15226 env var is in the environment, the default is auto rather than never.
15228 * diagnostic.h (file_name_as_prefix): Add context argument.
15229 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
15230 the string as locus.
15231 * langhooks.c (lhd_print_error_function): Adjust caller.
15233 2013-04-25 Lawrence Crowl <crowl@google.com>
15235 * var-tracking.c (shared_hash_def::htab):
15236 Change type to hash_table. Update dependent calls and types.
15238 2013-04-25 Lawrence Crowl <crowl@google.com>
15240 * Makefile.in: Update as needed below.
15242 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
15243 Move declaration to after the type's method definitons.
15245 * attribs.c (htab_t scoped_attributes::attribute_hash):
15246 Change type to hash_table. Update dependent calls and types.
15248 * bitmap.c (htab_t bitmap_desc_hash):
15249 Change type to hash_table. Update dependent calls and types.
15251 * cselib.c (htab_t cselib_hash_table):
15252 Change type to hash_table. Update dependent calls and types.
15254 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
15255 (hash_string_slot_node): Move implementation into lto-streamer.h
15256 struct string_slot_hasher.
15257 (eq_string_slot_node): Likewise.
15259 * data-streamer-out.c: Update output_block::string_hash_table
15260 dependent calls and types.
15262 * dwarf2cfi.c (htab_t trace_index):
15263 Change type to hash_table. Update dependent calls and types.
15265 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
15266 Change type to hash_table. Update dependent calls and types.
15267 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
15268 (htab_t optimize_external_refs::map): Likewise.
15269 (htab_t output_comp_unit::extern_map): Likewise.
15270 (htab_t output_comdat_type_unit::extern_map): Likewise.
15271 (htab_t output_macinfo::macinfo_htab): Likewise.
15272 (htab_t optimize_location_lists::htab): Likewise.
15273 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
15275 * except.c (htab_t ehspec_hash_type):
15276 Change type to hash_table. Update dependent calls and types.
15277 (assign_filter_values::ttypes): Likewise.
15278 (assign_filter_values::ehspec): Likewise.
15279 (sjlj_assign_call_site_values::ar_hash): Likewise.
15280 (convert_to_eh_region_ranges::ar_hash): Likewise.
15282 * gcse.c (htab_t pre_ldst_table):
15283 Change type to hash_table. Update dependent calls and types.
15285 * ggc-common.c (htab_t saving_htab):
15286 Change type to hash_table. Update dependent calls and types.
15287 (htab_t loc_hash): Likewise.
15288 (htab_t ptr_hash): Likewise.
15289 (call_count): Rename ggc_call_count.
15290 (call_alloc): Rename ggc_call_alloc.
15291 (loc_descriptor): Rename make_loc_descriptor.
15292 (add_statistics): Rename ggc_add_statistics.
15294 * ggc-common.c (saving_htab):
15295 Change type to hash_table. Update dependent calls and types.
15297 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
15298 (push_gimplify_context): Likewise.
15299 (pop_gimplify_context): Likewise.
15300 (struct gimple_temp_hash_elt): Added.
15301 (struct gimplify_hasher): Likewise.
15302 (struct gimplify_ctx.temp_htab):
15303 Change type to hash_table. Update dependent calls and types.
15305 * gimple-fold.c: Include gimplify-ctx.h.
15307 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
15308 Change type to hash_table. Update dependent calls and types.
15309 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
15310 avoid potential global name collision.
15312 * gimplify.c: Include gimplify-ctx.h.
15313 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
15314 (htab_t gimplify_ctx::temp_htab):
15315 Update dependent calls and types for new type hash_table.
15316 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
15317 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
15319 * gimplify-ctx.h: New.
15320 (struct gimple_temp_hash_elt): Move from gimplify.c.
15321 (class gimplify_hasher): New.
15322 (struct gimplify_ctx): Move from gimple.h.
15323 (htab_t gimplify_ctx::temp_htab):
15324 Change type to hash_table. Update dependent calls and types.
15326 * graphite-clast-to-gimple.c: Include graphite-htab.h.
15327 (htab_t ivs_params::newivs_index):
15328 Change type to hash_table. Update dependent calls and types.
15329 (htab_t ivs_params::params_index): Likewise.
15330 (htab_t print_generated_program::params_index): Likewise.
15331 (htab_t gloog::newivs_index): Likewise.
15332 (htab_t gloog::params_index): Likewise.
15334 * graphite.c: Include graphite-htab.h.
15335 4htab_t graphite_transform_loops::bb_pbb_mapping):
15336 Change type to hash_table. Update dependent calls and types.
15338 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
15339 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
15340 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
15342 * graphite-dependences.c: Include graphite-htab.h.
15343 (loop_is_parallel_p): Change hash table type of parameter.
15345 * graphite-htab.h: New.
15346 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
15347 (extern find_pbb_via_hash): Move from graphite-poly.h.
15348 (extern loop_is_parallel_p): Move from graphite-poly.h.
15349 (extern get_loop_body_pbbs): Move from graphite-poly.h.
15351 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
15352 (extern loop_is_parallel_p): Move to graphite-htab.h.
15353 (extern get_loop_body_pbbs): Move to graphite-htab.h.
15355 * haifa-sched.c (htab_t delay_htab):
15356 Change type to hash_table. Update dependent calls and types.
15357 (htab_t delay_htab_i2): Likewise.
15359 * ira-color.c (htab_t allocno_hard_regs_htab):
15360 Change type to hash_table. Update dependent calls and types.
15362 * ira-costs.c (htab_t cost_classes_htab):
15363 Change type to hash_table. Update dependent calls and types.
15365 * loop-invariant.c (htab_t merge_identical_invariants::eq):
15366 Change type to hash_table. Update dependent calls and types.
15368 * loop-iv.c (htab_t bivs):
15369 Change type to hash_table. Update dependent calls and types.
15371 * loop-unroll.c (htab_t opt_info::insns_to_split):
15372 Change type to hash_table. Update dependent calls and types.
15373 (htab_t opt_info::insns_with_var_to_expand): Likewise.
15375 * lto-streamer.h (struct string_slot): Move from data-streamer.h
15376 (struct string_slot_hasher): New.
15377 (htab_t output_block::string_hash_table):
15378 Change type to hash_table. Update dependent calls and types.
15380 * lto-streamer-in.c (freeing_string_slot_hasher): New.
15381 (htab_t file_name_hash_table):
15382 Change type to hash_table. Update dependent calls and types.
15384 * lto-streamer-out.c: Update output_block::string_hash_table dependent
15387 * lto-streamer.c (htab_t tree_htab):
15388 Change type to hash_table. Update dependent calls and types.
15390 * omp-low.c: Include gimplify-ctx.h.
15392 * passes.c (htab_t name_to_pass_map):
15393 Change type to hash_table. Update dependent calls and types.
15394 (pass_traverse): Rename to passes_pass_traverse.
15396 * plugin.c (htab_t event_tab):
15397 Change type to hash_table. Update dependent calls and types.
15399 * postreload-gcse.c (htab_t expr_table):
15400 Change type to hash_table. Update dependent calls and types.
15401 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
15403 * sese.c (debug_rename_map_1): Make extern.
15404 (htab_t copy_bb_and_scalar_dependences::rename_map):
15405 Change type to hash_table. Update dependent calls and types.
15407 * sese.h (extern debug_rename_map): Move to .c file.
15409 * store-motion.c (htab_t store_motion_mems_table):
15410 Change type to hash_table. Update dependent calls and types.
15412 * trans-mem.c (htab_t tm_new_mem_hash):
15413 Change type to hash_table. Update dependent calls and types.
15415 * tree-browser.c (htab_t TB_up_ht):
15416 Change type to hash_table. Update dependent calls and types.
15418 * tree-cfg.c (htab_t discriminator_per_locus):
15419 Change type to hash_table. Update dependent calls and types.
15421 * tree-complex.c: Include tree-hasher.h
15422 (htab_t complex_variable_components):
15423 Change type to hash_table. Update dependent calls and types.
15425 * tree-eh.c (htab_t finally_tree):
15426 Change type to hash_table. Update dependent calls and types.
15428 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
15429 struct int_tree_hasher.
15430 (extern int_tree_map_eq): Likewise.
15431 (uid_decl_map_hash): Removed.
15432 (extern decl_tree_map_eq): Likewise.
15434 * tree-hasher.h: New.
15435 (struct int_tree_hasher): New.
15436 (typedef int_tree_htab_type): New.
15438 * tree-inline.c: Include gimplify-ctx.h.
15440 * tree-mudflap.c: Include gimplify-ctx.h.
15442 * tree-parloops.c: Include tree-hasher.h.
15443 (htab_t eliminate_local_variables_stmt::decl_address):
15444 Change type to hash_table. Update dependent calls and types.
15445 (htab_t separate_decls_in_region::decl_copies): Likewise.
15447 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
15448 Change type to hash_table. Update dependent calls and types.
15450 * tree-sra.c (candidates):
15451 Change type to hash_table. Update dependent calls and types.
15453 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
15455 (int_tree_map_hash): Likewise.
15457 * tree-ssa-dom.c (htab_t avail_exprs):
15458 Change type to hash_table. Update dependent calls and types.
15460 * tree-ssa-live.c (var_map_base_init::tree_to_index):
15461 Change type to hash_table. Update dependent calls and types.
15463 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
15464 Change type to hash_table. Update dependent calls and types.
15466 * tree-ssa-phiopt.c (seen_ssa_names):
15467 Change type to hash_table. Update dependent calls and types.
15469 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
15470 Change type to hash_table. Update dependent calls and types.
15472 * tree-ssa-uncprop.c (equiv):
15473 Change type to hash_table. Update dependent calls and types.
15475 2013-04-25 Jakub Jelinek <jakub@redhat.com>
15477 PR rtl-optimization/57003
15478 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
15479 call note_stores with kill_clobbered_value callback again after
15480 killing regs_invalidated_by_call.
15482 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
15484 * config/aarch64/aarch64-simd.md
15485 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
15486 (aarch64_simd_bsl<mode>): Likewise.
15487 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
15489 2013-04-25 Marek Polacek <polacek@redhat.com>
15491 PR tree-optimization/57066
15492 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
15494 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
15496 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
15498 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
15500 * config/aarch64/aarch64-builtins.c
15501 (aarch64_fold_builtin): New.
15502 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
15503 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
15504 * config/aarch64/aarch64-simd-builtins.def (abs): New.
15505 * config/aarch64/arm_neon.h
15506 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
15508 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
15509 Tejas Belagod <tejas.belagod@arm.com>
15511 * config/aarch64/aarch64-builtins.c
15512 (aarch64_gimple_fold_builtin): New.
15513 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
15514 * config/aarch64/aarch64-simd-builtins.def (addv): New.
15515 * config/aarch64/aarch64-simd.md (addpv4sf): New.
15516 (addvv4sf): Update.
15517 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
15519 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
15521 * config/aarch64/aarch64.md
15522 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
15524 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
15526 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
15527 (*ngcsi_uxtw): New pattern.
15529 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15530 Julian Brown <julian@codesourcery.com>
15532 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
15533 (TB_DREG): Add T_V4HF.
15534 (v4hf_UP): New macro.
15535 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
15536 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
15537 Handle initialisation of V4HF. Adjust initialisation of reinterpret
15539 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
15540 (arm_vector_mode_supported_p): Handle V4HF.
15541 (arm_mangle_map): Handle V4HFmode.
15542 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
15543 * config/arm/arm_neon_builtins.def: Add entries for
15544 vcvtv4hfv4sf, vcvtv4sfv4hf.
15545 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
15546 (neon_vcvtv4hfv4sf): Likewise.
15547 * config/arm/neon-gen.ml: Handle half-precision floating point
15549 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
15550 * config/arm/arm_neon.h: Regenerate.
15551 * config/arm/neon.ml (type elts): Add F16.
15552 (type vectype): Add T_float16x4, T_floatHF.
15553 (type vecmode): Add V4HF.
15554 (type features): Add Requires_FP_bit feature.
15555 (elt_width): Handle F16.
15556 (elt_class): Likewise.
15557 (elt_of_class_width): Likewise.
15558 (mode_of_elt): Refactor.
15559 (type_for_elt): Handle F16, fix error messages.
15560 (vectype_size): Handle T_float16x4.
15561 (vcvt_sh): New function.
15562 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
15563 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
15564 (string_of_mode): Handle V4HF.
15565 * doc/arm-neon-intrinsics.texi: Regenerate.
15567 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
15569 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
15570 format specifier in 'X' case.
15572 2013-04-25 Alan Modra <amodra@gmail.com>
15575 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
15576 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
15577 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
15578 Repeat for many other rotate/shift and mask patterns using subregs.
15579 Name lshiftrt insns.
15580 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
15581 on WORDS_BIG_ENDIAN.
15583 2013-04-25 Alan Modra <amodra@gmail.com>
15585 * config.gcc: Support little-endian powerpc-linux targets.
15586 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
15587 (LINK_OS_LINUX_SPEC): Define.
15588 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
15589 Preserve MASK_LITTLE_ENDIAN.
15590 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
15591 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
15592 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
15593 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
15594 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
15595 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
15596 Correct fp word order for little-endian. Don't shift toc entries
15597 smaller than a word for little-endian.
15598 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
15599 (bswapdi2 splits): Correct low-part subreg for little-endian.
15600 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
15601 low/high where such is correct only for be.
15602 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
15603 little-endian for -mcall-aixdesc.
15605 2013-04-25 Alan Modra <amodra@gmail.com>
15607 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
15608 replace_equiv_address_nv.
15610 2013-04-25 Alan Modra <amodra@gmail.com>
15612 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
15614 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
15617 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
15618 * rtl.h (struct rtx_def): ...
15620 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
15622 PR rtl-optimizations/57046
15623 * lra-constraints (split_reg): Set up lra_risky_transformations_p
15624 for multi-reg splits.
15626 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
15628 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
15630 2013-04-24 Sterling Augustine <saugustine@google.com>
15632 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
15633 (comp_dir_string, debug_str_dwo_section): New.
15634 (DEBUG_STR_DWO_SECTION): Rename to ...
15635 (DEBUG_DWO_STR_SECTION): ... this.
15636 (DEBUG_NORM_STR_SECTION): Delete.
15637 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
15638 (DEBUG_STR_DWO_SECTION_FLAGS): New.
15639 (find_AT_string): Move most logic to ...
15640 (find_AT_string_in_table): ... here. New.
15641 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
15642 add_skeleton_AT_string. Delete logic.
15643 (output_skeleton_debug_sections): Remove call to
15644 add_top_level_skeleton_die_attrs.
15645 (add_comp_dir_attribute): Move logic to comp_dir_string.
15646 (dwarf2out_init): Initialize debug_str_dwo_section.
15647 (output_indirect_string): Call find_string_form.
15648 (output_indirect_strings): Rewrite.
15649 (prune_unused_types): Empty skeleton_debug_str_hash.
15650 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
15651 (dwarf2out_finish): Call output_indirect_strings.
15653 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
15655 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
15657 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
15659 * rtl.h (struct rtx_def): Add comment for field jump.
15660 (LRA_SUBREG_P): New macro.
15661 * recog.c (register_operand): Check LRA_SUBREG_P.
15662 * lra.c (lra): Add note at the end of RTL code. Align non-empty
15664 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
15665 (lra_final_code_change): Skip subreg change for operators.
15666 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
15667 if there are no operand changes.
15668 * lra-constraints.c (curr_insn_set): New.
15669 (match_reload): Set LRA_SUBREG_P.
15670 (emit_spill_move): Ditto.
15671 (check_and_process_move): Use curr_insn_set. Process only single
15672 set insns. Don't initialize sec_mem_p and change_p.
15673 (simplify_operand_subreg): Use LRA_SUBREG_P.
15674 (reg_in_class_p): New function.
15675 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
15676 of #ifdef. Add code to remove cycling.
15677 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
15678 non-null disp. Reload inner instead of disp when base and index
15679 are null. Try to put lo_sum into register.
15680 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
15681 (check_and_process_move): Move code for move cost check to
15682 simple_move_p. Remove equiv_substitution.
15683 (simple_move_p): New function.
15684 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
15685 curr_insn_set. Call check_and_process_move only for single set
15686 insns. Use the new function. Move call of check_and_process_move
15687 after operand equiv substitution and address process.
15689 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
15691 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
15692 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
15693 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
15695 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
15697 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
15699 2013-04-24 Marek Polacek <polacek@redhat.com>
15701 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
15702 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
15703 (select_loops_exit_conditions): Likewise.
15704 (number_of_iterations_for_all_loops): Likewise.
15705 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
15706 (scev_analysis): Likewise.
15708 2013-04-02 Catherine Moore <clm@codesourcery.com>
15709 Chao-ying Fu <fu@mips.com>
15711 * config/mips/micromips.md (jraddiusp): New pattern.
15712 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
15713 instruction if possible.
15715 2013-04-24 Alan Modra <amodra@gmail.com>
15717 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
15719 2013-04-24 Julian Brown <julian@codesourcery.com>
15720 Chung-Lin Tang <cltang@codesourcery.com>
15722 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
15723 dependency behavior in enumeration type DIE generation. Add TODO note
15724 to comments about future DW_FORM_sdata/udata re-work of related code.
15726 2013-04-23 Lawrence Crowl <crowl@google.com>
15728 * Makefile.in: Update as needed below.
15730 * hash-table.h (class hash_table):
15731 Correct many methods with parameter types compare_type to the correct
15732 value_type. (Correct code was unlikely to notice the change.)
15733 (hash_table::elements_with_deleted) New.
15734 (class hashtable::iterator): New.
15735 (hashtable::begin()): New.
15736 (hashtable::end()): New.
15737 (FOR_EACH_HASH_TABLE_ELEMENT): New.
15739 * statistics.c (statistics_hashes):
15740 Change type to hash_table. Update dependent calls and types.
15742 * tree-into-ssa.c (var_infos):
15743 Change type to hash_table. Update dependent calls and types.
15745 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
15746 Change type to hash_table. Update dependent calls and types.
15748 * tree-ssa-loop-im.c (struct mem_ref.refs):
15749 Change type to hash_table. Update dependent calls and types.
15751 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
15752 Change type to hash_table. Update dependent calls and types.
15754 * tree-ssa-sccvn.c (vn_tables_s::nary):
15755 Change type to hash_table. Update dependent calls and types.
15756 (vn_tables_s::phis): Likewise.
15757 (vn_tables_s::references): Likewise.
15759 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
15760 (vn_reference_eq): Update parameter and return types.
15762 * tree-ssa-structalias.c (pointer_equiv_class_table):
15763 Change type to hash_table. Update dependent calls and types.
15764 (location_equiv_class_table): Likewise.
15766 * tree-vect-data-refs.c: Consequential changes for making
15767 peeling a hash_table.
15769 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
15770 (destroy_loop_vec_info): Dependent hash_table update.
15772 * tree-vectorizer.h (peeling_htab):
15773 Change type to hash_table. Update dependent calls and types.
15775 2013-04-23 Shiva Chen <shiva0217@gmail.com>
15777 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
15778 to check the register content is equal or not.
15779 * lra-constraints.c (match_reload): Use lra_assign_reg_val
15780 to assign register content record.
15781 * lra-eliminations.c (update_reg_eliminate): Use
15782 lra_update_reg_val_offset to update register content offset.
15783 * lra-int.h (struct lra_reg): Add offset member.
15784 (lra_reg_val_equal_p): New static inline function.
15785 (lra_update_reg_val_offset): New static inline function.
15786 (lra_assign_reg_val): New static inline function.
15787 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
15788 to assign register content record.
15789 (initialize_lra_reg_info_element): Initial offset to zero.
15791 2013-04-23 Catherine Moore <clm@codesourcery.com>
15793 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
15794 operands. Record compression.
15796 2013-04-23 Xinliang David Li <davidxl@google.com>
15798 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
15800 2013-04-23 Richard Biener <rguenther@suse.de>
15802 PR middle-end/57036
15803 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
15804 parameter, only add abnormal goto edges from the copied body
15805 if the call could perform abnormal gotos.
15806 (copy_cfg_body): Adjust.
15808 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
15810 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
15812 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
15814 * coretypes.h (gimple_stmt_iterator): Add struct to make
15817 2013-04-23 Richard Biener <rguenther@suse.de>
15819 PR tree-optimization/57026
15820 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
15821 from SSA names occuring in abnormal PHI nodes.
15823 2013-04-22 Andi Kleen <ak@linux.intel.com>
15825 * lto/lto.c (print_lto_report_1): Fix LTO report names.
15827 2013-04-22 Andi Kleen <ak@linux.intel.com>
15829 * lto/lto.c (print_lto_report_1): Declare early.
15830 (read_cgraph_and_symbols): Call print_lto_report_1 early.
15832 2013-04-22 Andi Kleen <ak@linux.intel.com>
15834 * common.opt (-flto-report-wpa): Add.
15835 * doc/invoke.texi (-flto-report-wpa): Add.
15836 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
15839 2013-04-22 Xinliang David Li <davidxl@google.com>
15841 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
15842 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
15843 * Makefile.in: New dependency
15845 David Daney <ddaney.cavm@gmail.com>
15847 * configure.ac (gcc_cv_as_micromips_support): Use the
15848 --fatal-warnings option.
15849 * configure: Regenerate.
15851 2013-04-22 Marek Polacek <polacek@redhat.com>
15854 * tsan.c (instrument_expr): Don't instrument expression
15855 in case its size is zero.
15857 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
15861 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
15863 * config/alpha/alpha.c (TARGET_LRA_P): New define.
15865 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
15867 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
15868 (gimple_stmt_iterator): New typedef.
15869 * gimple.h (gimple_stmt_iterator): Rename to...
15870 (gimple_stmt_iterator_d): ... This.
15871 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
15872 trees be valid for GIMPLE and GENERIC.
15873 (TARGET_GIMPLE_FOLD_BUILTIN): New.
15874 * gimple-fold.c (gimple_fold_call): Call target hook
15875 gimple_fold_builtin.
15876 * hooks.c (hook_bool_gsiptr_false): New.
15877 * hooks.h (hook_bool_gsiptr_false): New.
15878 * target.def (fold_stmt): New.
15879 * doc/tm.texi: Regenerate.
15881 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
15884 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
15885 a set sp if no stack realignment.
15887 2013-04-22 Nick Clifton <nickc@redhat.com>
15889 * config.gcc (tilegx-linux): Extend extra_objs rather than
15891 (tilepro-linux): Likewise.
15893 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
15895 * config/aarch64/aarch64-builtins.c
15897 (CF0, CF1, CF2, CF3, CF4, CF10): New.
15898 (VAR<1-12>): Add MAP parameter.
15899 (BUILTIN_*): Likewise.
15900 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
15901 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
15902 (aarch64_ushl_n<mode>): Likewise.
15903 (aarch64_sshr_n<mode>): Likewise.
15904 (aarch64_ushr_n<mode>): Likewise.
15905 (aarch64_<maxmin><mode>): Likewise.
15906 (aarch64_sqrt<mode>): Likewise.
15907 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
15908 (vshr<q>_n_*): Likewise.
15910 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
15912 * config/aarch64/aarch64-builtins.c
15913 (aarch64_simd_builtin_type_mode): Handle SF types.
15915 (BUILTIN_GPF): Define.
15916 (aarch64_init_simd_builtins): Handle SF types.
15917 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
15918 (frecps): Likewise.
15919 (frecpx): Likewise.
15920 * config/aarch64/aarch64-simd.md
15921 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
15922 (aarch64_frecpe<mode>): New.
15923 (aarch64_frecps<mode>): Likewise.
15924 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
15925 (v8type): Add frecp<esx>.
15926 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
15927 (aarch64_frecps<mode>): Likewise.
15928 * config/aarch64/iterators.md (FRECP): New.
15929 (frecp_suffix): Likewise.
15930 * config/aarch64/arm_neon.h
15931 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
15933 2013-04-22 Christian Bruel <christian.bruel@st.com>
15936 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
15937 (REG_CLASS_NAMES): Idem.
15938 (REG_CLASS_CONTENTS): Idem.
15939 (REGCLASS_HAS_FP_REG): Idem.
15940 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
15941 (sh_conditional_register_usage): Idem.
15943 2013-04-21 Jeff Law <law@redhat.com>
15945 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
15946 (ssa_forward_propagate_and_combine): Use it.
15948 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
15950 * lra.c: Update the flow chart diagram.
15952 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
15954 PR rtl-optimization/56847
15955 * lra-constraints.c (process_alt_operands): Discourage alternative
15956 with non-matche doffsettable memory constraint fro memory with
15959 2013-04-19 Richard Biener <rguenther@suse.de>
15961 PR tree-optimization/56982
15962 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
15964 * gimplify.c (gimplify_call_expr): Notice special calls.
15965 (gimplify_modify_expr): Likewise.
15966 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
15967 abnormal control flow receivers.
15968 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
15969 in the same way as cfun->has_nonlocal_labels.
15970 (gimple_purge_dead_abnormal_call_edges): Likewise.
15971 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
15972 receivers start a basic-block.
15974 2013-04-19 Richard Biener <rguenther@suse.de>
15976 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
15978 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
15979 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
15980 (SLP_TREE_LOAD_PERMUTATION): Add.
15981 (vect_transform_slp_perm_load): Adjust prototype.
15982 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
15983 (vect_free_slp_instance): Likewise.
15984 (vect_create_new_slp_node): Likewise.
15985 (vect_supported_slp_permutation_p): Remove.
15986 (vect_slp_rearrange_stmts): Adjust.
15987 (vect_supported_load_permutation_p): Likewise. Inline
15988 vect_supported_slp_permutation_p here.
15989 (vect_analyze_slp_instance): Compute load permutations per
15990 slp node instead of per instance.
15991 (vect_get_slp_defs): Adjust.
15992 (vect_transform_slp_perm_load): Likewise.
15993 (vect_schedule_slp_instance): Remove redundant code.
15994 (vect_schedule_slp): Remove hack for PR56270, add it ...
15995 * tree-vect-stmts.c (vectorizable_load): ... here, do not
15996 CSE loads for SLP. Adjust.
15998 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
16000 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
16001 spelling in two comments.
16003 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
16006 * config/arm/arm.c (load_multiple_sequence): Require SP
16007 as base register for loads if SP is in the register list.
16009 2013-04-19 Martin Jambor <mjambor@suse.cz>
16011 PR tree-optimization/56718
16012 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
16013 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
16014 and made public. Adjusted all callers.
16015 (ipa_intraprocedural_devirtualization): New function.
16016 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
16017 (ipa_intraprocedural_devirtualization): Likewise.
16018 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
16020 2013-04-19 Richard Biener <rguenther@suse.de>
16022 PR tree-optimization/57000
16023 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
16025 2013-04-19 Terry Guo <terry.guo@arm.com>
16027 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
16029 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
16030 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
16031 (cortex_m4_fmacs): Use new reservations.
16032 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
16034 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
16036 PR rtl-optimization/56999
16037 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
16039 (lra_coalesce): Remove split_origin_bitmap and related code.
16040 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
16041 ranges if necessary.
16043 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
16045 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
16047 (ix86_expand_call): Remove clobbered_registers array and use
16048 x86_64_ms_sysv_extra_clobbered_registers instead.
16049 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
16051 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
16053 * config/i386/i386.md (*call_rex64_ms_sysv): Use
16054 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
16055 (*call_value_rex64_ms_sysv): Ditto.
16057 2013-04-18 Cary Coutant <ccoutant@google.com>
16059 * dwarf2out.c (output_pubnames): Check die_perennial_p of
16060 parent instead of die_mark.
16062 2013-04-18 Diego Novillo <dnovillo@google.com>
16064 * gimple.c (create_gimple_tmp): New.
16065 (get_expr_type): New.
16066 (build_assign): New.
16067 (build_type_cast): New.
16068 * gimple.h (enum ssa_mode): Define.
16069 (gimple_seq_set_location): New.
16070 * asan.c (build_check_stmt): Change some gimple_build_* calls
16071 to use build_assign and build_type_cast.
16073 2013-04-18 Richard Biener <rguenther@suse.de>
16075 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
16076 handle negative step. Remove redundant checks.
16077 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
16078 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
16079 for negative step and grouped loads fail to vectorize.
16081 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
16083 * emit-rtl.c (reset_insn_used_flags): New function.
16084 (reset_all_used_flags): Use it.
16085 (verify_insn_sharing): New function.
16086 (verify_rtl_sharing): Fix verification for SEQUENCEs.
16088 2013-04-18 Jakub Jelinek <jakub@redhat.com>
16090 PR tree-optimization/56984
16091 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
16092 and (x >> M) >= N don't register any assertion if N << M is the
16095 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
16097 * lower-subreg.c (resolve_simple_move): If called self-recursive,
16098 do not delete_insn insns that have not yet been emitted, only
16099 unlink them with remove_insn.
16100 * df-scan.c (df_insn_delete): Revert r197492.
16102 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
16104 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
16105 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
16107 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
16109 * config/arm/arm.md (movsicc_insn): Convert define_insn into
16110 define_insn_and_split.
16111 (and_scc,ior_scc,negscc): Likewise.
16112 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
16114 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
16116 * config/arm/arm.c (use_return_insn): Return 0 for targets that
16117 can benefit from using a sequence of LDRD instructions in epilogue
16118 instead of a single LDM instruction.
16120 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
16123 * doc/extend.texi: Fix typo.
16125 2013-04-17 Richard Biener <rguenther@suse.de>
16127 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
16128 (vect_build_slp_tree): ... here.
16129 (vect_build_slp_tree_1): Compute which stmts of the SLP group
16130 match. Remove special-casing of mismatched complex loads.
16131 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
16132 re-try the match with swapped commutative operands.
16133 (vect_supported_load_permutation_p): Remove special-casing of
16134 mismatched complex loads.
16135 (vect_analyze_slp_instance): Adjust.
16137 2013-04-17 Richard Biener <rguenther@suse.de>
16139 PR rtl-optimization/56921
16140 * cfgloop.h (struct loop): Add simple_loop_desc member.
16141 (struct niter_desc): Mark with GTY(()).
16142 (simple_loop_desc): Do not use aux field but simple_loop_desc.
16143 * loop-iv.c (get_simple_loop_desc): Likewise.
16144 (free_simple_loop_desc): Likewise.
16147 2013-04-16 Richard Biener <rguenther@suse.de>
16149 PR rtl-optimization/56921
16150 * loop-init.c (pass_rtl_move_loop_invariants): Add
16151 TODO_do_not_ggc_collect to todo_flags_finish.
16152 (pass_rtl_unswitch): Same.
16153 (pass_rtl_unroll_and_peel_loops): Same.
16154 (pass_rtl_doloop): Same.
16156 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
16158 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
16159 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
16160 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
16162 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
16163 * tree-streamer.c (record_common_node): Adjust reference in comment.
16165 2013-04-17 Terry Guo <terry.guo@arm.com>
16167 * config/arm/cortex-m4.md: Add a new bypass.
16169 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16171 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
16173 (*subs_<optab><mode>_multp2): New pattern.
16174 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
16175 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
16177 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16179 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
16180 (*subs_mul_imm_<mode>): New pattern.
16182 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
16185 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
16186 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
16187 (vsx_movti_32bit): Change j->wa to O->wa.
16189 2013-04-16 Richard Biener <rguenther@suse.de>
16191 PR rtl-optimization/56921
16192 * loop-init.c (pass_rtl_move_loop_invariants): Add
16193 TODO_do_not_ggc_collect to todo_flags_finish.
16194 (pass_rtl_unswitch): Same.
16195 (pass_rtl_unroll_and_peel_loops): Same.
16196 (pass_rtl_doloop): Same.
16198 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
16200 * config/arm/arm.c (emit_multi_reg_push): New declaration
16201 for an existing function.
16202 (arm_emit_strd_push): New function.
16203 (arm_expand_prologue): Used here.
16204 (arm_emit_ldrd_pop): New function.
16205 (arm_expand_epilogue): Used here.
16206 (arm_get_frame_offsets): Update condition.
16207 (arm_emit_multi_reg_pop): Add a special case for load of a single
16208 register with writeback.
16210 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
16212 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
16215 2013-04-16 Richard Biener <rguenther@suse.de>
16217 PR tree-optimization/56756
16218 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
16219 (first_mem_ref_loc): New.
16220 (execute_sm): Place the load temporarily before a previous
16221 access instead of in the latch edge to ensure its SSA dependencies
16222 are defined at points dominating the load.
16224 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
16226 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
16227 correct fix by moving header and footer insn to the footer of
16228 the merged basic block. Clear BB_END of the merged-away block.
16230 PR middle-end/43631
16231 * emit-rtl.c (make_note_raw): New function.
16232 (link_insn_into_chain): New static inline function.
16233 (add_insn): Use it.
16234 (add_insn_before, add_insn_after): Factor insn chain linking code...
16235 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
16236 using link_insn_into_chain.
16237 (note_outside_basic_block_p): New helper function for emit_note_after
16238 and emit_note_before.
16239 (emit_note_after): Use nobb variant of add_insn_after if the note
16240 should not be contained in a basic block.
16241 (emit_note_before): Use nobb variant of add_insn_before if the note
16242 should not be contained in a basic block.
16243 (emit_note_copy): Use make_note_raw.
16244 (emit_note): Likewise.
16245 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
16246 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
16247 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
16248 the moved barrier the tail of the basic block it follows.
16249 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
16251 2013-04-15 Jakub Jelinek <jakub@redhat.com>
16253 PR tree-optimization/56962
16254 * gimple-ssa-strength-reduction.c (record_increment): Only set
16255 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
16256 either rhs1 or rhs2 is equal to c->base_expr.
16258 2013-04-15 Richard Biener <rguenther@suse.de>
16260 PR tree-optimization/56933
16261 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
16263 (GROUP_READ_WRITE_DEPENDENCE): Remove.
16264 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
16265 * tree-vect-data-refs.c (vect_analyze_group_access): Move
16266 dependence check ...
16267 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
16269 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
16270 GROUP_READ_WRITE_DEPENDENCE.
16272 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16274 * emit-rtl.c (reset_all_used_flags): New function.
16275 (verify_rtl_sharing): Call reset_all_used_flags before and after
16276 performing the checks.
16278 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16280 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
16281 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
16282 * config/arm/constraints.md (De): New constraint.
16283 * config/arm/neon.md (anddi3_neon): Delete.
16284 (neon_vand<mode>): Expand to standard anddi3 pattern.
16285 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
16286 Move earlier in the file.
16287 (neon_inv_logic_op2): Likewise.
16288 (arm_anddi_operand_neon): New predicate.
16290 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16292 * configure.ac (gcc_cv_ld_as_needed): Set
16293 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
16294 Use -z ignore, -z record on *-*-solaris2*.
16295 (HAVE_LD_AS_NEEDED): Update comment.
16296 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
16297 * configure: Regenerate.
16298 * config.in: Regenerate.
16299 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
16300 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
16301 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
16302 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
16303 equivalents. Fix markup.
16304 * doc/tm.texi: Regenerate.
16306 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
16308 * config/i386/i386.opt: New option mstack-protector-guard=.
16309 * config/i386/i386-opts.h: Add enum stack_protector_guard.
16310 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
16311 TARGET_SSP_TLS_GUARD.
16312 * config/i386/i386.c (ix86_option_override_internal): Set
16313 ix86_stack_protector_guard.
16314 * config/i386/i386.md (stack_protect_set): Enable for
16315 TARGET_SSP_TLS_GUARD only.
16316 (stack_protect_set_<mode>): Ditto.
16317 (stack_protect_test): Ditto.
16318 (stack_protect_test_<mode>): Ditto.
16319 * doc/invoke.texi (i386 Option): Document.
16321 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
16324 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
16325 (S_MODES): Set H_MODE bit.
16326 (SF_MODES): Set only S_MODE and SF_MODE bits.
16327 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
16328 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
16329 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
16330 <MODE_FLOAT>: Likewise.
16332 2013-04-15 Joey Ye <joey.ye@arm.com>
16334 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
16336 2013-04-15 Joey Ye <joey.ye@arm.com>
16338 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
16340 (thumb_far_jump_used_p): Count instruction size and set
16343 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
16345 * reorg.c (fill_simple_delay_slots): Reindent block of code.
16346 * resource.c (mark_target_live_regs): Reformat conditional block.
16348 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
16350 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
16351 notes, they are emitted only just before final.
16352 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
16354 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
16356 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
16357 * cfgrtl.c (delete_insn): Call it here instead.
16358 * lra-spills.c (lra_final_code_change): Use delete_insn.
16359 * haifa-sched.c (sched_remove_insn): Likewise.
16360 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
16361 returning to the nop pool.
16362 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
16363 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
16365 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
16367 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
16368 * doc/tm.texi: Regenerated.
16370 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
16372 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
16375 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
16377 * df-core.c (df_find_def): Compare register numbers.
16378 (df_find_use): Likewise.
16380 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
16383 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
16384 lra_in_progress for return.
16386 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
16388 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
16389 define_insn into define_insn_and_split and emit movsicc patterns.
16391 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
16393 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
16395 2013-04-12 Richard Biener <rguenther@suse.de>
16397 * tree-pass.h (TODO_do_not_ggc_collect): New.
16398 * passes.c (execute_one_ipa_transform_pass): Honor
16399 TODO_do_not_ggc_collect.
16400 (execute_one_pass): Likewise.
16403 2013-04-10 Richard Biener <rguenther@suse.de>
16405 * passes.c (init_optimization_passes): Remove reload pass.
16406 * ira.c (do_reload): Merge into ...
16408 (rest_of_handle_reload): Remove.
16409 (pass_reload): Likewise.
16410 * config/i386/i386.c (ix86_option_override): Refer to ira instead
16411 of reload for vzeroupper pass placement.
16413 2013-04-12 Jakub Jelinek <jakub@redhat.com>
16415 PR tree-optimization/56918
16416 PR tree-optimization/56920
16417 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
16418 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
16419 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
16420 use wide_mul_with_sign method.
16422 2013-04-12 Richard Biener <rguenther@suse.de>
16424 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
16425 not be considered a gimple constant.
16427 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
16429 * fold-const.c (const_binop): Handle vector shifts by a scalar.
16430 (fold_binary_loc): Call const_binop also for mixed vector-scalar
16433 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
16434 Jakub Jelinek <jakub@redhat.com>
16436 * opts.c: Include diagnostic-color.h.
16437 (common_handle_option): Handle OPT_fdiagnostics_color_.
16438 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
16439 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
16440 (diagnostic-color.o): New.
16441 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
16442 (diagnostic_color_rule): New enum.
16443 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
16444 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
16445 the location string.
16446 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
16447 either NULL, or color kind.
16448 * diagnostic-color.c: New file.
16449 * diagnostic-color.h: New file.
16450 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
16452 * doc/invoke.texi (-fdiagnostics-color): Document.
16453 * pretty-print.h (pp_show_color): Define.
16454 (struct pretty_print_info): Add show_color field.
16455 * diagnostic.c: Include diagnostic-color.h.
16456 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
16457 macros. Colorize error:, warning: etc. strings and also the location
16459 (diagnostic_show_locus): Colorize the caret line.
16460 * pretty-print.c: Include diagnostic-color.h.
16461 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
16462 inside of %< %> quotes or quoted through q format modifier.
16464 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16466 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
16468 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16470 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
16471 code in CC_NZ mode.
16472 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
16475 2013-04-11 Marek Polacek <polacek@redhat.com>
16477 PR tree-optimization/48184
16478 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
16480 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
16482 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
16483 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
16484 (skip_simple_arithmetic): Tidy up.
16485 * tree.h (skip_simple_constant_arithmetic): Declare.
16487 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16489 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
16491 2013-04-11 Richard Biener <rguenther@suse.de>
16493 * tree-vect-loop.c (get_initial_def_for_induction): Properly
16494 generate vector constants.
16496 2013-04-11 Richard Biener <rguenther@suse.de>
16498 PR tree-optimization/56878
16499 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
16500 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
16502 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
16503 Prefer to align the DR with the most invariant base address.
16505 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
16507 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
16510 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
16512 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
16513 floating-point vector comparisons against 0.
16515 2013-04-11 Jakub Jelinek <jakub@redhat.com>
16517 PR tree-optimization/56899
16518 * fold-const.c (extract_muldiv_1): Apply distributive law
16519 only if TYPE_OVERFLOW_WRAPS (ctype).
16521 2013-04-11 Bin Cheng <bin.cheng@arm.com>
16524 * ira-costs.c (scan_one_insn): Check whether the source rtx of
16525 loading has side effect.
16527 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
16529 * config/sparc/sparc.c: Include tree-pass.h.
16530 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
16531 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
16532 head of file. Change return type. Split off gate function.
16533 (sparc_gate_work_around_errata): New function.
16534 (pass_work_around_errata): New pass definition.
16535 (insert_pass_work_around_errata) New pass insert definition to
16536 insert pass_work_around_errata just after delayed-branch scheduling.
16537 (sparc_option_override): Insert the pass.
16538 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
16540 2013-04-10 David S. Miller <davem@davemloft.net>
16542 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
16543 or -mcpu=hypersparc.
16545 * target.def (cstore_mode): New hook.
16546 * target.h: Include insn-codes.h
16547 * targhooks.c: Likewise.
16548 (default_cstore_mode): New function.
16549 * targhooks.h: Declare it.
16550 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
16551 * doc/tm.texi: Rebuild.
16552 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
16553 target hook, rather than inspecting the insn_data.
16554 * config/sparc/sparc.c (sparc_cstore_mode): New function.
16555 (TARGET_CSTORE_MODE): Redefine.
16556 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
16558 * config/sparc/predicates.md (cstore_result_operand): New special
16560 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
16561 Use it for operand 0.
16562 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
16563 (*snesi_special): Likewise.
16564 (*snesi_zero): Likewise.
16565 (*seqsi_zero): Likewise.
16566 (*sltu_insn): Likewise.
16567 (*sgeu_insn): Likewise.
16568 (*seqdi_special): Make operand 0 and comparison operation be of
16570 (*snedi_special): Likewise.
16571 (*snedi_special_vis3): Likewise.
16572 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
16573 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
16574 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
16575 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
16576 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
16577 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
16578 (*sltu_extend_sp64): Likewise.
16579 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
16580 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
16581 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
16582 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
16583 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
16585 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
16587 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
16588 (aarch64_start_file): Use the new function.
16590 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
16591 Jason Merrill <jason@redhat.com>
16593 * common.opt: Add -gdwarf.
16594 * opts.c (common_handle_option): Handle it.
16595 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
16597 2013-04-10 Richard Biener <rguenther@suse.de>
16599 * passes.c (execute_todo): Do not call ggc_collect conditional here.
16600 (execute_one_ipa_transform_pass): But unconditionally here.
16601 (execute_one_pass): And here.
16602 (init_optimization_passes): Remove reload pass.
16603 * tree-pass.h (TODO_ggc_collect): Remove.
16604 (pass_reload): Likewise.
16605 * ira.c (do_reload): Merge into ...
16607 (rest_of_handle_reload): Remove.
16608 (pass_reload): Likewise.
16609 * config/i386/i386.c (ix86_option_override): Refer to ira instead
16610 of reload for vzeroupper pass placement.
16611 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
16612 and todo_flags_finish of all passes.
16614 2013-04-10 Richard Biener <rguenther@suse.de>
16616 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
16617 first_const_oprnd field, rename first_def_type to first_op_type.
16618 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
16619 (vect_get_and_check_slp_defs): Always use the type of the
16620 operand. Allow mixed vect_external_def, vect_constant_def types.
16621 (vect_get_constant_vectors): Handle mixed vect_external_def,
16622 vect_constant_def types.
16624 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
16626 PR tree-optimization/55524
16627 * tree-ssa-math-opts.c
16628 (convert_mult_to_fma): Don't use an fms construct
16629 when we don't have an fms operation, but fnma, and it looks
16630 likely that we'll be able to use the latter.
16632 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
16634 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
16636 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
16637 inline fail caused by overwritable functions.
16639 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
16641 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
16642 unnecessary bits in the constant power of two case.
16644 2013-04-10 Richard Biener <rguenther@suse.de>
16646 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
16647 broken code swapping operands.
16648 (vect_build_slp_tree): Do not compute load permutations here.
16649 (vect_analyze_slp_instance): Compute load permutations here,
16650 after building the SLP tree.
16652 2013-04-09 Christian Bruel <christian.bruel@st.com>
16654 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
16655 of next/prev_real_insn.
16657 2013-04-09 Jan Hubicka <jh@suse.cz>
16659 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
16660 Drop aliased parameter.
16661 (function_and_variable_visibility): Do not handle alias pairs.
16662 * cgraph.c (varpool_externally_visible_p): Update prototype.
16663 * varpool.c (varpool_add_new_variable): Update.
16665 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16667 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
16669 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
16671 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
16673 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
16675 2013-04-09 Marek Polacek <polacek@redhat.com>
16677 PR tree-optimization/48762
16678 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
16680 2013-04-09 Richard Biener <rguenther@suse.de>
16682 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
16684 (vect_build_slp_tree): Likewise.
16685 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
16686 calculating the cost of a SLP instance.
16687 (vect_analyze_slp_instance): Use it from here, after building
16690 2013-04-09 Jakub Jelinek <jakub@redhat.com>
16692 PR middle-end/56883
16693 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
16694 expand_omp_for_static_chunk): Use simple_p = true in
16695 force_gimple_operand_gsi calls when assigning to addressable decls.
16697 2013-04-09 Jeff Law <law@redhat.com>
16699 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
16700 when the boolean was created by converting a wider object which
16701 had a boolean range.
16703 2013-04-09 Richard Biener <rguenther@suse.de>
16705 * tree-vectorizer.h (slp_void_p): Remove.
16706 (slp_tree): Typedef before _slp_tree declaration.
16707 (struct _slp_tree): Use a vector of slp_tree as children.
16708 (vect_get_place_in_interleaving_chain): Remove.
16709 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
16711 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
16713 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
16714 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
16715 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
16716 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
16717 Use slp_node instead of slp_void_p and adjust.
16719 2013-04-09 Richard Biener <rguenther@suse.de>
16721 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
16722 work that is not necessary.
16724 2013-04-09 Jakub Jelinek <jakub@redhat.com>
16726 PR tree-optimization/56854
16727 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
16728 forward into clobber stmts if it would change MEM_REF lhs into
16731 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
16733 * tree.c (type_hash_lookup, type_hash_add): Make static.
16734 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
16736 2013-04-09 Richard Biener <rguenther@suse.de>
16738 * tree.h (unsave_expr_now): Remove.
16739 * tree-inline.c (mark_local_for_remap_r): Remove.
16740 (unsave_expr_1): Likewise.
16741 (unsave_r): Likewise.
16742 (unsave_expr_now): Likewise.
16743 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
16744 (propagate_tree_value): Likewise.
16746 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
16748 * doc/rtl.texi (sequence): Rewrite documentation to match the
16749 current use of SEQUENCE rtl objects.
16750 * rtl.def (SEQUENCE): Likewise.
16752 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
16753 Update documentation.
16754 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
16755 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
16757 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
16759 2013-04-08 Teresa Johnson <tejohnson@google.com>
16761 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
16762 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
16764 (estimate_edge_size_and_time): Add comment to suggest using rounding
16766 (estimate_node_size_and_time): Ditto.
16767 (remap_edge_change_prob): Use helper rounding divide methods.
16768 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
16769 (gimple_mod_pow2_value_transform): Ditto.
16770 (gimple_mod_subtract_transform): Ditto.
16771 (gimple_ic_transform): Ditto.
16772 (gimple_stringops_transform): Ditto.
16773 * stmt.c (conditional_probability): Ditto.
16774 (emit_case_dispatch_table): Ditto.
16775 * lto-cgraph.c (merge_profile_summaries): Ditto.
16776 * tree-optimize.c (execute_fixup_cfg): Ditto.
16777 * cfgcleanup.c (try_forward_edges): Ditto.
16778 * cfgloopmanip.c (scale_loop_profile): Ditto.
16780 (duplicate_loop_to_header_edge): Ditto.
16781 (lv_adjust_loop_entry_edge): Ditto.
16782 * tree-vect-loop.c (vect_transform_loop): Ditto.
16783 * profile.c (compute_branch_probabilities): Ditto.
16784 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
16785 * lto-streamer-in.c (input_cfg): Ditto.
16786 * gimple-streamer-in.c (input_bb): Ditto.
16787 * ipa-cp.c (update_profiling_info): Ditto.
16788 (update_specialized_profile): Ditto.
16789 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
16790 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
16792 * sched-rgn.c (compute_dom_prob_ps): Ditto.
16793 (compute_trg_info): Ditto.
16794 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
16795 (purge_dead_edges): Ditto.
16796 * loop-unswitch.c (unswitch_loop): Ditto.
16797 * cgraphclones.c (cgraph_clone_edge): Ditto.
16798 (cgraph_clone_node): Ditto.
16799 * tree-inline.c (copy_bb): Ditto.
16800 (copy_edges_for_bb): Ditto.
16801 (initialize_cfun): Ditto.
16802 (copy_cfg_body): Ditto.
16803 (expand_call_inline): Ditto.
16805 2013-04-08 Kai Tietz <ktietz@redhat.com>
16807 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
16808 TARGET_CYGWIN64 by TARGET_64BIT.
16810 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
16812 * config/epiphany/epiphany.md (GPR_1): New constant.
16813 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
16814 * config/epiphany/epiphany.c (gen_compare_reg):
16815 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
16816 is already in place.
16817 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
16818 Don't require being called during rtl expansion; If y operlaps r0,
16820 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
16821 (epiphany_expand_epilogue): Likewise.
16823 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
16824 Don't use CC_FPmode for ORDERED / UNORDERED.
16825 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
16827 * config/epiphany/constraints.md (CnL): New constraint.
16828 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
16829 * config/epiphany/predicates.md (add_operand): Allow 1024.
16831 * config/epiphany/epiphany.md (logical_op): New code iterator.
16832 (op_mnc): New code attribute.
16833 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
16834 (mov_f+1, mov_f+2): New peephole2 patterns.
16836 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
16837 (cstoresi4): Also allow re-use of zero result when doing a NE
16838 comparison to a non-zero operand.
16839 Use (clobber (scratch)) for first insn if the gpr output is not needed.
16841 * config/epiphany/epiphany.md (<insn_opname>v2si3):
16842 Use gen_addsi3_i / gen_subsi3_i.
16844 2013-04-08 Jakub Jelinek <jakub@redhat.com>
16848 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
16849 contain anything but clobbers, at most one __builtin_stack_restore,
16850 optionally debug stmts and final resx, and if it has at least one
16851 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
16852 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
16853 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
16854 which isn't defaut definition, remove them.
16855 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
16856 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
16857 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
16858 with MEM_REF LHS with SSA_NAME address.
16860 2013-04-08 Jeff Law <law@redhat.com>
16862 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
16864 2013-04-08 Richard Biener <rguenther@suse.de>
16866 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
16868 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
16869 determined vector type.
16870 (vect_analyze_data_refs): Likewise.
16871 (vect_get_new_vect_var): Adjust.
16872 (vect_create_destination_var): Preserve SSA name versions.
16873 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
16874 not dump anything here.
16876 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
16878 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
16879 Add member lr_slot_known.
16880 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
16882 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
16883 Remove code that sets lr_slot_offset according to what a previous
16884 version of epiphany_emit_save_restore used to do.
16885 (epiphany_emit_save_restore): When doing an lr save or restore,
16886 set/verify lr_slot_known and lr_slot_offset.
16888 2013-04-08 Xinyu Qi <xyqi@marvell.com>
16891 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
16894 2013-04-08 Richard Biener <rguenther@suse.de>
16896 * alias.c (find_base_term): Fix thinko in previous change.
16898 2013-04-08 Jakub Jelinek <jakub@redhat.com>
16900 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
16901 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
16902 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
16903 if possible to compute val.
16904 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
16905 For QImode integers don't require anything about precision. Use
16906 const_with_all_bytes_same to find out if the constant doesn't have
16907 repeated bytes in it.
16909 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16911 * config/s390/s390.c (s390_expand_insv): Only accept insertions
16914 2013-04-08 Marek Polacek <polacek@redhat.com>
16916 PR rtl-optimization/48182
16917 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
16920 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
16923 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
16924 nuses, make sure we have a label.
16926 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16929 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
16930 (rs6000_emit_swdiv_low_precision): Remove.
16931 (rs6000_emit_swdiv): Rewrite to handle between one and four
16932 iterations of Newton-Raphson generally; modify required number of
16933 iterations for some cases.
16934 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
16936 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
16938 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
16939 set-but-unused variable.
16941 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
16942 basic blocks of released function bodies garbage-collectable.
16944 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
16945 (struct rtl_opt_pass): Add TODO_df_finish.
16947 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
16949 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16951 * config/arm/constraints.md (q): New constraint.
16952 * config/arm/ldrdstrd.md: New file.
16953 * config/arm/arm.md (ldrdstrd.md) New include.
16954 (arm_movdi): Use "q" instead of "r" constraint
16955 for double-word memory access.
16956 (movdf_soft_insn): Likewise.
16957 * config/arm/vfp.md (movdi_vfp): Likewise.
16958 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
16959 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
16960 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
16961 (mem_ok_for_ldrd_strd): Likewise.
16962 (output_move_double): Update assertion.
16964 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16966 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
16968 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16970 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
16971 define_insn_and_split.
16972 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
16974 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16976 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
16977 define_insn_and_split.
16978 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
16979 (shiftsi3_compare): New pattern.
16980 (rrx): New pattern.
16981 * config/arm/unspecs.md (UNSPEC_RRX): New.
16983 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16985 * config/arm/arm.md (negdi_extendsidi): New pattern.
16986 (negdi_zero_extendsidi): Likewise.
16988 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16990 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
16991 define_insn_and_split.
16992 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
16993 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
16995 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
16997 * config/arm/arm.md (arm_subdi3): Convert define_insn into
16998 define_insn_and_split.
16999 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
17000 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
17002 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17004 * config/arm/arm.md (subsi3_carryin): New pattern.
17005 (subsi3_carryin_const): Likewise.
17006 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
17007 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
17009 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17011 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
17013 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
17015 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
17016 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
17018 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17020 * config/arm/arm.c (arm_expand_builtin): Change fcode
17021 type to unsigned int.
17023 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17025 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
17027 2013-04-04 Ian Lance Taylor <iant@google.com>
17029 * doc/standards.texi (Standards): The Go frontend supports the Go 1
17032 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
17034 PR middle-end/56729
17035 * df-scan.c (df_insn_delete): Disable failing assert.
17037 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17039 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
17040 New function prototype.
17041 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
17042 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
17043 (arm_builtin_vectorized_function): New function.
17045 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17047 * config/arm/arm_neon_builtins.def: New file.
17048 * config/arm/arm.c (neon_builtin_data): Move contents to
17049 arm_neon_builtins.def.
17050 (enum arm_builtins): Include neon builtin definitions.
17051 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
17052 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
17054 2013-04-04 Marek Polacek <polacek@redhat.com>
17056 PR tree-optimization/48186
17057 * predict.c (maybe_hot_frequency_p): Return false if
17058 HOT_BB_FREQUENCY_FRACTION is 0.
17059 (cgraph_maybe_hot_edge_p): Likewise.
17061 2013-04-04 Richard Biener <rguenther@suse.de>
17063 PR tree-optimization/56826
17064 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
17067 2013-04-04 Richard Biener <rguenther@suse.de>
17069 PR tree-optimization/56213
17070 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
17071 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
17073 2013-04-04 Richard Biener <rguenther@suse.de>
17075 PR tree-optimization/56837
17076 * tree-loop-distribution.c (classify_partition): For non-zero
17077 values require that the value has the same precision as its
17078 mode to be useful as memset value.
17080 2013-04-03 Nick Clifton <nickc@redhat.com>
17082 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
17083 (fmssf4): Use fmsf.s on E3V5 architectures.
17084 (fnmasf4): Use fnmaf.s on E3V5 architectures.
17085 (fnmssf4): Use fnmsf.s on E3V5 architectures.
17087 2013-04-03 Jeff Law <law@redhat.com>
17089 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
17090 (lra-eliminations.o): Likewise.
17092 2013-04-03 Teresa Johnson <tejohnson@google.com>
17094 * gcov-io.c (compute_working_sets): Moved most of body of old
17095 compute_working_sets here from profile.c.
17096 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
17097 (gcov_working_set_t): Moved typedef here from basic-block.h
17098 (compute_working_set): Declare.
17099 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
17100 (get_working_sets): Renamed from compute_working_set,
17101 replace most of body with call to new compute_working_sets.
17102 (get_exec_counts): Replace call to compute_working_sets
17103 to get_working_sets.
17104 * profile.h (get_working_sets): Renamed from compute_working_set.
17105 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
17106 to get_working_sets.
17107 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
17108 * gcov-dump.c (dump_working_sets): New function.
17110 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
17112 * hwint.c (sext_hwi, zext_hwi): New functions.
17113 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
17114 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
17115 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
17116 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
17117 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
17118 (sext_hwi, zext_hwi): New functions.
17120 2013-04-03 Jeff Law <law@redhat.com>
17122 PR tree-optimization/56799
17123 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
17124 back test for widening conversion erroneously dropped in prior change.
17126 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17129 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
17130 instead of next_real_insn.
17132 2013-04-03 Marek Polacek <polacek@redhat.com>
17135 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
17137 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17140 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
17142 (thumb1_output_casesi): Likewise.
17143 (thumb2_output_casesi): Likewise.
17145 2013-04-03 Richard Biener <rguenther@suse.de>
17147 PR tree-optimization/56817
17148 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
17150 (tree_unroll_loops_completely_1): ... new function to manually
17151 walk the loop tree, properly defering outer loops of unrolled
17152 loops to later iterations.
17154 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
17156 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
17157 (vectorizable_load): Likewise.
17158 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
17159 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
17161 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
17163 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
17166 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
17168 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
17170 2013-04-03 Bin Cheng <bin.cheng@arm.com>
17172 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
17174 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
17176 PR tree-optimization/56790
17177 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
17180 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
17182 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
17184 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
17185 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
17188 2013-04-03 Jakub Jelinek <jakub@redhat.com>
17191 * tree.h (force_folding_builtin_constant_p): New decl.
17192 * builtins.c (force_folding_builtin_constant_p): New variable.
17193 (fold_builtin_constant_p): Fold immediately also if
17194 force_folding_builtin_constant_p.
17196 2013-04-03 Richard Biener <rguenther@suse.de>
17198 PR tree-optimization/56812
17199 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
17200 DRs of the same interleaving chain are independent.
17202 2013-04-02 Jason Merrill <jason@redhat.com>
17204 * gdbinit.in (pbb): Use debug fn.
17206 2013-04-02 Lawrence Crowl <crowl@google.com>
17208 * sese.h (struct ivtype_map_elt_s): Remove unused.
17209 (extern debug_ivtype_map): Remove unused.
17210 (extern eq_ivtype_map_elts): Remove unused.
17211 * sese.c (debug_ivtype_map): Removed unused.
17212 (debug_ivtype_map_1): Removed unused.
17213 (debug_ivtype_elt): Remove unused.
17214 (eq_ivtype_map_elts): Remove unused.
17217 2013-04-02 Kai Tietz <ktietz@redhat.com>
17220 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
17221 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
17222 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
17224 (legitimize_pe_coff_symbol): Likewise.
17225 (is_imported_p): New helper-function.
17226 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
17227 for Windows x64 targets.
17228 (ix86_expand_prologue): Optimize for pe-coff targets.
17229 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
17230 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
17231 medium/large code-model.
17232 (legitimize_pic_address): Likewise.
17233 (legitimize_tls_address): Likewise.
17234 (ix86_expand_call): Likewise.
17235 (x86_output_mi_thunk): Likewise.
17236 (get_dllimport_decl): Add new beimport argument.
17237 (construct_plt_address): Don't assert for x64 pe-coff targets.
17238 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
17240 (SYMBOL_FLAG_STUBVAR): New macro.
17241 (SYMBOL_REF_STUBVAR_P): Likewise.
17242 * config/i386/winnt.c (stub_list): New structure.
17243 (stub_head): New local variable.
17244 (i386_pe_record_stub): New function.
17245 (i386_pe_file_end): Emit refptr-stubs.
17247 2013-04-02 Jakub Jelinek <jakub@redhat.com>
17249 PR rtl-optimization/56745
17250 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
17251 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
17254 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
17255 and both of them are MEM_REFs, just compare first argument for
17256 equality and attempt to deal even with differing offsets.
17259 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
17260 of gimple_clobber_p to be MEM_REF.
17261 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
17262 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
17263 after gimplification.
17264 * asan.c (get_mem_ref_of_assignment): Don't instrument
17265 gimple_clobber_p stmts.
17266 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
17267 gimple_clobber_p stmt if they have MEM_REF lhs and
17268 are dead because of another gimple_clobber_p stmt.
17269 * tree-ssa-live.c (clear_unused_block_pointer): Treat
17270 gimple_clobber_p stmts like debug stmts.
17271 (remove_unused_locals): Remove clobbers with MEM_REF lhs
17272 that refer to unused VAR_DECLs or uninitialized values.
17273 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
17274 gimple_clobber_p stmts if they refer to removed parameters.
17275 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
17278 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
17280 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
17281 using SWI48 mode attribute.
17283 2013-04-02 Wei Mi <wmi@google.com>
17285 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
17286 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
17287 *<rotate_insn><mode>3_mask in i386.md.
17289 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
17291 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
17293 2013-04-02 Richard Biener <rguenther@suse.de>
17295 PR tree-optimization/56778
17296 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
17297 Runtime alias tests are not supported for gather loads.
17298 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
17299 stmts referenced from SSA operands before updating SSA form.
17301 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
17302 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17304 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
17305 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
17306 * config/arm/cortex-a53.md: New file.
17307 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
17308 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
17309 * config/arm/arm.c (arm_issue_rate): Likewise.
17310 * config/arm/arm-tune.md: Regenerate
17311 * config/arm/arm-tables.opt: Regenerate.
17312 * config/arm/arm-cores.def: Add cortex-a53.
17314 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
17316 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
17319 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
17321 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
17322 scalar load/store operations using B/H registers.
17323 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
17325 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
17327 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
17329 * config/aarch64/aarch64.c
17330 (aarch64_simd_scalar_immediate_valid_for_move): New.
17331 * config/aarch64/aarch64-protos.h
17332 (aarch64_simd_scalar_immediate_valid_for_move): New.
17333 * config/aarch64/constraints.md (Dh, Dq): New.
17334 * config/aarch64/iterators.md (hq): New.
17336 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
17338 * reorg.c (get_branch_condition): Deal with conditional returns.
17339 (fill_simple_delay_slots): Remove dead code dealing with jumps.
17341 2013-04-01 Wei Mi <wmi@google.com>
17343 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
17344 Truncate operand 2 using %b asm operand modifier.
17345 (*<shift_insn><mode>3_mask): Ditto.
17346 (*<rotate_insn><mode>3_mask): Ditto.
17348 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
17350 PR middle-end/56798
17351 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
17353 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
17355 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
17357 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
17359 2013-03-30 Lawrence Crowl <crowl@google.com>
17361 * dse.c (clear_alias_sets): Remove never set.
17362 (disqualified_clear_alias_sets): Remove never set.
17363 (clear_alias_mode_pool): Remove never set.
17364 (dse_step0): Remove condition that is never true.
17365 (canon_address): Remove condition that is never true.
17366 (dse_step7): Remove condition that is never true.
17367 (rest_of_handle_dse): Remove condition that is never true.
17368 (rest_of_handle_dse::did_global): Remove never read from above.
17369 (dse_step2_spill): Remove never called from above.
17370 (dse_step5_spill): Remove never called from above.
17372 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
17374 * doc/md.texi (Standard Names) <casesi>: Update documentation for
17375 JUMP_TABLE_DATA changes.
17376 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
17377 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
17378 (Insns) <jump_table_data>: New entry.
17379 * doc/tm.texi: Regenerate.
17381 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
17383 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
17384 for table jump at the end of a basic block using tablejump_p.
17385 * targhooks.c (default_invalid_within_doloop): Likewise.
17386 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
17387 target hook implementation that is identical to the default hook.
17388 (rs6000_invalid_within_doloop): Remove.
17390 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
17391 unused variable from tablejump_p call.
17393 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
17394 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
17395 (INSN_DELETED_P): Likewise.
17396 (emit_jump_table_data): New prototype.
17397 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
17398 after 4th as unused.
17399 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
17400 * sched-vis.c (print_insn): Likewise.
17401 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
17402 insn for compatibility with back ends that use next_active_insn to
17403 identify jump table data.
17404 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
17405 (remove_insn): Likewise.
17406 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
17408 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
17409 (emit_jump_table_data): New function.
17411 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
17412 basic block, a JUMP_TABLE_DATA never is.
17413 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
17414 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
17415 off from code handling real insns.
17416 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
17417 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
17418 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
17419 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
17420 is not a NONDEBUG_INSN_P.
17421 * ira-costs.c (scan_one_insn): Likewise.
17422 * jump.c (mark_all_labels): Likewise.
17423 (mark_jump_label_1): Likewise.
17424 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
17425 * lra.c (get_insn_freq): Expect all insns reaching here to be in
17427 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
17428 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
17429 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
17430 JUMP_TABLE_DATA_P insns.
17431 (calculate_elim_costs_all_insns): Likewise.
17432 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
17433 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
17434 (delete_output_reload): Code style fixups.
17435 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
17436 insn flags on this non-insn.
17437 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
17438 as scheduling barriers, for pre-change compatibility.
17439 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
17440 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
17442 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
17443 redundant JUMP_TABLE_DATA_P test.
17444 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
17445 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
17446 (frv_for_each_packet): Likewise.
17447 * config/i386/i386.c (min_insn_size): Likewise.
17448 (ix86_avoid_jump_mispredicts): Likewise.
17449 * config/m32r/m32r.c (m32r_is_insn): Likewise.
17450 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
17451 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
17452 (mips16_insn_length): Robustify.
17453 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
17454 (mips16_split_long_branches): Likewise.
17455 * config/pa/pa.c (pa_combine_instructions): Likewise.
17456 * config/rs6000/rs6000.c (get_next_active_insn): Treat
17457 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
17458 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
17459 as contributing to pool range lengths.
17460 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
17461 Remove redundant JUMP_TABLE_DATA_P test.
17462 (sh_loop_align): Likewise.
17463 (split_branches): Likewise.
17464 (sh_insn_length_adjustment): Likewise.
17465 * config/spu/spu.c (get_branch_target): Likewise.
17467 2013-03-29 Jan Hubicka <jh@suse.cz>
17469 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
17470 gcov streaming; stream hot bb threshold to ltrans.
17471 * predict.c (get_hot_bb_threshold): Break out from ....
17472 (maybe_hot_count_p): ... here.
17473 (set_hot_bb_threshold): New function.
17474 * lto-section-in.c (lto_section_name): Add profile.
17475 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
17476 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
17477 and data-streamer.h
17478 (histogram_entry): New structure.
17479 (histogram, histogram_pool): New global vars.
17480 (histogram_hash): New structure.
17481 (histogram_hash::hash): New method.
17482 (histogram_hash::equal): Likewise.
17483 (account_time_size): New function.
17484 (cmp_counts): New function.
17485 (dump_histogram): New function.
17486 (ipa_profile_generate_summary): New function.
17487 (ipa_profile_write_summary): New function.
17488 (ipa_profile_read_summary): New function.
17489 (ipa_profile): Decide on threshold.
17490 (pass_ipa_profile): Add ipa_profile_write_summary and
17491 ipa_profile_read_summary.
17492 * Makefile.in (ipa.o): Update dependencies.
17493 * lto-streamer.h (LTO_section_ipa_profile): New section.
17495 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
17497 * tree.h (VAR_P): New.
17499 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
17502 * doc/invoke.texi ([-fwhole-program]): Fix typo.
17504 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
17506 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
17507 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
17508 (control_flow_insn_p): Likewise.
17509 * cfgrtl.c (duplicate_insn_chain): Likewise.
17510 * final.c (get_attr_length_1): Likewise.
17511 (shorten_branches): Likewise.
17512 (final_scan_insn): Likewise.
17513 * function.c (instantiate_virtual_regs): Likewise.
17514 * gcse.c (insert_insn_end_basic_block): Likewise.
17515 * ira-costs.c (scan_one_insn): Likewise.
17516 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
17517 * lra.c (check_rtl): Likewise.
17518 * reload1.c (elimination_costs_in_insn): Likewise.
17519 * reorg.c (follow_jumps): Likewise.
17521 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
17522 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
17523 (thumb_far_jump_used_p): Likewise.
17524 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
17525 (workaround_speculation): Likewise.
17526 (add_sched_insns_for_speculation): Likewise.
17527 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
17528 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
17529 (frv_for_each_packet): Likewise.
17530 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
17531 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
17532 (final_emit_insn_group_barriers): Likewise.
17533 * config/m32r/m32r.c (m32r_is_insn): Likewise.
17534 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
17535 (mips16_insn_length): Likewise.
17536 * config/pa/pa.c (pa_reorg): Likewise.
17537 (pa_combine_instructions): Likewise.
17538 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
17539 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
17540 (sh_reorg): Likewise.
17541 (split_branches): Likewise.
17542 * config/spu/spu.c (get_branch_target): Likewise.
17544 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
17547 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
17549 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
17550 Fix declaration name.
17552 2013-03-28 Lawrence Crowl <crowl@google.com>
17554 * graphds.h (struct graph.indicies): Remove unused.
17555 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
17556 (SCOP_ORIGINAL_PDDRS): Remove unused.
17557 * sese.h (extern insert_loop_close_phis): Removed unused.
17558 (extern insert_guard_phis): Removed unused.
17559 (extern ivtype_map_elt_info): Removed unused.
17560 (new_ivtype_map_elt): Removed unused.
17561 * sese.c (ivtype_map_elt_info): Removed unused.
17563 2013-03-28 Lawrence Crowl <crowl@google.com>
17565 * Makefile.in: Add several missing include dependences.
17567 (test-dump.o): New. This object is not added to any executable,
17568 but is present for ad-hoc testing.
17570 (debug (const bitmap_head_def &)): New.
17571 (debug (const bitmap_head_def *)): New.
17573 (extern debug (const bitmap_head_def &)): New.
17574 (extern debug (const bitmap_head_def *)): New.
17576 (debug (edge_def &)): New.
17577 (debug (edge_def *)): New.
17579 (debug (basic_block_def &)): New.
17580 (debug (basic_block_def *)): New.
17582 (dump_node (const_tree, int, FILE *)): Correct source file.
17584 (debug (die_struct &)): New.
17585 (debug (die_struct *)): New.
17587 (extern debug (die_struct &)): New.
17588 (extern debug (die_struct *)): New.
17589 * gimple-pretty-print.c
17590 (debug (gimple_statement_d &)): New.
17591 (debug (gimple_statement_d *)): New.
17592 * gimple-pretty-print.h
17593 (extern debug (gimple_statement_d &)): New.
17594 (extern debug (gimple_statement_d *)): New.
17596 (debug (ira_allocno_copy &)): New.
17597 (debug (ira_allocno_copy *)): New.
17598 (debug (ira_allocno &)): New.
17599 (debug (ira_allocno *)): New.
17601 (extern debug (ira_allocno_copy &)): New.
17602 (extern debug (ira_allocno_copy *)): New.
17603 (extern debug (ira_allocno &)): New.
17604 (extern debug (ira_allocno *)): New.
17606 (debug (live_range &)): New.
17607 (debug (live_range *)): New.
17609 (debug (lra_live_range &)): New.
17610 (debug (lra_live_range *)): New.
17612 (debug (lra_live_range &)): New.
17613 (debug (lra_live_range *)): New.
17615 (debug (omega_pb_d &)): New.
17616 (debug (omega_pb_d *)): New.
17618 (extern debug (omega_pb_d &)): New.
17619 (extern debug (omega_pb_d *)): New.
17621 (debug (const rtx_def &)): New.
17622 (debug (const rtx_def *)): New.
17624 (debug_tree (tree): Move within file.
17625 (debug_raw (const tree_node &)): New.
17626 (debug_raw (const tree_node *)): New.
17627 (dump_tree_via_hooks (const tree_node *, int)): New.
17628 (debug (const tree_node &)): New.
17629 (debug (const tree_node *)): New.
17630 (debug_verbose (const tree_node &)): New.
17631 (debug_verbose (const tree_node *)): New.
17632 (debug_head (const tree_node &)): New.
17633 (debug_head (const tree_node *)): New.
17634 (debug_body (const tree_node &)): New.
17635 (debug_body (const tree_node *)): New.
17636 (debug_vec_tree (tree): Move and reimplement in terms of dump.
17637 (debug (vec<tree, va_gc> &)): New.
17638 (debug (vec<tree, va_gc> *)): New.
17640 (extern debug (const rtx_def &)): New.
17641 (extern debug (const rtx_def *)): New.
17643 (debug_raw (simple_bitmap_def &)): New.
17644 (debug_raw (simple_bitmap_def *)): New.
17645 (debug (simple_bitmap_def &)): New.
17646 (debug (simple_bitmap_def *)): New.
17648 (extern debug (simple_bitmap_def &)): New.
17649 (extern debug (simple_bitmap_def *)): New.
17650 (extern debug_raw (simple_bitmap_def &)): New.
17651 (extern debug_raw (simple_bitmap_def *)): New.
17653 (debug (vinsn_def &)): New.
17654 (debug (vinsn_def *)): New.
17655 (debug_verbose (vinsn_def &)): New.
17656 (debug_verbose (vinsn_def *)): New.
17657 (debug (expr_def &)): New.
17658 (debug (expr_def *)): New.
17659 (debug_verbose (expr_def &)): New.
17660 (debug_verbose (expr_def *)): New.
17661 (debug (vec<rtx> &)): New.
17662 (debug (vec<rtx> *)): New.
17664 (extern debug (vinsn_def &)): New.
17665 (extern debug (vinsn_def *)): New.
17666 (extern debug_verbose (vinsn_def &)): New.
17667 (extern debug_verbose (vinsn_def *)): New.
17668 (extern debug (expr_def &)): New.
17669 (extern debug (expr_def *)): New.
17670 (extern debug_verbose (expr_def &)): New.
17671 (extern debug_verbose (expr_def *)): New.
17672 (extern debug (vec<rtx> &)): New.
17673 (extern debug (vec<rtx> *)): New.
17675 (_list_iter_cond_expr): Make inline instead of static.
17677 (debug (sreal &)): New.
17678 (debug (sreal *)): New.
17680 (extern debug (sreal &)): New.
17681 (extern debug (sreal *)): New.
17683 (extern debug_raw (const tree_node &)): New.
17684 (extern debug_raw (const tree_node *)): New.
17685 (extern debug (const tree_node &)): New.
17686 (extern debug (const tree_node *)): New.
17687 (extern debug_verbose (const tree_node &)): New.
17688 (extern debug_verbose (const tree_node *)): New.
17689 (extern debug_head (const tree_node &)): New.
17690 (extern debug_head (const tree_node *)): New.
17691 (extern debug_body (const tree_node &)): New.
17692 (extern debug_body (const tree_node *)): New.
17693 (extern debug (vec<tree, va_gc> &)): New.
17694 (extern debug (vec<tree, va_gc> *)): New.
17696 (debug (struct loop &)): New.
17697 (debug (struct loop *)): New.
17698 (debug_verbose (struct loop &)): New.
17699 (debug_verbose (struct loop *)): New.
17700 * tree-dump.c: Add header dependence.
17702 (extern debug (struct loop &)): New.
17703 (extern debug (struct loop *)): New.
17704 (extern debug_verbose (struct loop &)): New.
17705 (extern debug_verbose (struct loop *)): New.
17707 (debug (data_reference &)): New.
17708 (debug (data_reference *)): New.
17709 (debug (vec<data_reference_p> &)): New.
17710 (debug (vec<data_reference_p> *)): New.
17711 (debug (vec<ddr_p> &)): New.
17712 (debug (vec<ddr_p> *)): New.
17714 (extern debug (data_reference &)): New.
17715 (extern debug (data_reference *)): New.
17716 (extern debug (vec<data_reference_p> &)): New.
17717 (extern debug (vec<data_reference_p> *)): New.
17718 (extern debug (vec<ddr_p> &)): New.
17719 (extern debug (vec<ddr_p> *)): New.
17721 (debug (pt_solution &)): New.
17722 (debug (pt_solution *)): New.
17724 (extern debug (pt_solution &)): New.
17725 (extern debug (pt_solution *)): New.
17727 (debug (_var_map &)): New.
17728 (debug (_var_map *)): New.
17729 (debug (tree_live_info_d &)): New.
17730 (debug (tree_live_info_d *)): New.
17732 (extern debug (_var_map &)): New.
17733 (extern debug (_var_map *)): New.
17734 (extern debug (tree_live_info_d &)): New.
17735 (extern debug (tree_live_info_d *)): New.
17737 2013-03-28 Jan Hubicka <jh@suse.cz>
17739 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
17741 2013-03-28 Ian Bolton <ian.bolton@arm.com>
17743 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
17744 record only when desired or required.
17746 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
17748 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
17749 *vec_extractv2di_1_rex64. Use x64 isa attribute.
17751 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17753 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
17754 (*andsi3_compare0_uxtw): New pattern.
17755 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
17756 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
17758 2013-03-28 Jan Hubicka <jh@suse.cz>
17760 * data-streamer-in.c (streamer_read_gcov_count): New function.
17761 * gimple-streamer-out.c: Include value-prof.h.
17762 (output_gimple_stmt): Output histogram.
17763 (output_bb): Use streamer_write_gcov_count.
17764 * value-prof.c: Include data-streamer.h
17765 (dump_histogram_value): Add HIST_TYPE_MAX.
17766 (stream_out_histogram_value): New function.
17767 (stream_in_histogram_value): New function.
17768 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
17769 (stream_out_histogram_value, stream_in_histogram_value): Declare.
17770 * data-streamer-out.c (streamer_write_gcov_count): New function.
17771 (streamer_write_gcov_count_stream): New function.
17772 * lto-cgraph.c (lto_output_edge): Update counter streaming.
17773 (lto_output_node): Likewise.
17774 (input_node, input_edge): Likewise.
17775 * lto-streamer-out.c (output_cfg): Update streaming.
17776 * lto-streamer-in.c (input_cfg): Likewise.
17777 * data-streamer.h (streamer_write_gcov_count,
17778 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
17779 * gimple-streamer-in.c: Include value-prof.h
17780 (input_gimple_stmt): Input histograms.
17781 (input_bb): Update profile streaming.
17783 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
17785 * genmodes.c (emit_max_int): New function.
17786 (emit_insn_modes_h): Added call to emit_max_function.
17787 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
17789 * machmode.def: Fixed comment.
17791 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
17793 * combine.c (try_combine): Removed useless assert.
17794 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
17796 2013-03-28 Marek Polacek <polacek@redhat.com>
17797 Richard Biener <rguenther@suse.de>
17799 PR tree-optimization/56695
17800 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
17801 build signed result of a vector comparison.
17802 * tree-cfg.c (verify_gimple_comparison): Check that a result
17803 of a vector comparison has signed type.
17805 2013-03-28 Richard Biener <rguenther@suse.de>
17807 PR tree-optimization/37021
17808 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
17809 do not restrict gaps between groups.
17810 * tree-vect-stmts.c (vectorizable_load): Properly account for
17811 a gap between groups.
17813 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
17815 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
17816 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
17819 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
17821 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
17822 * doc/extend.texi (Named Address Spaces): Ditto.
17823 (Variable Attributes): Ditto.
17825 2013-03-27 Kai Tietz <ktietz@redhat.com>
17827 * config.build: Add support for cygwin x64 target.
17828 * config.gcc: Likewise.
17829 * config.host: Likewise.
17830 * configure.ac: Likewise
17831 * configure: Regenerated.
17833 2013-03-27 Kai Tietz <ktietz@redhat.com>
17835 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
17836 * config/i386/t-cygwin-w64: New file.
17837 * config/i386/cygwin-w64.h: New file.
17838 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
17839 and add support for x64-cygwin target.
17840 (CPP_SPEC): Likewise.
17841 (CXX_WRAP_SPEC_LIST): Undefine before define.
17842 (LIBGCJ_SONAME): Use 15 as version.
17844 2013-03-27 Richard Biener <rguenther@suse.de>
17846 PR tree-optimization/56716
17847 * tree-ssa-structalias.c (perform_var_substitution): Adjust
17848 dumping for ref nodes.
17850 2013-03-27 Martin Jambor <mjambor@suse.cz>
17852 PR tree-optimization/55334
17853 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
17854 restricted pointers to arrays.
17856 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
17858 * Makefile.in (.SUFFIXES): Add .cc.
17859 (.c.o): Apply same recipe for implicit rule .cc.o.
17861 2013-03-27 Richard Biener <rguenther@suse.de>
17863 PR tree-optimization/37021
17864 * tree-vect-data-refs.c (vect_check_strided_load): Allow
17865 REALPART/IMAGPART_EXPRs around the supported refs.
17866 * tree-ssa-structalias.c (find_func_aliases): Assume that
17867 floating-point values are not used to transfer pointers.
17869 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
17871 * target.def (TARGET_HAS_IFUNC_P): New target hook.
17872 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
17873 * doc/tm.texi: Regenerate.
17874 * targhooks.h (default_has_ifunc_p): New.
17875 * targhooks.c (default_has_ifunc_p): Ditto.
17876 * config/linux-protos.h: New file.
17877 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
17878 hook for linux which disables support of indirect functions in android.
17879 * config/linux-android.c: New file.
17880 * config/t-linux-android.c: Ditto.
17881 * config.gcc: Added new object file linux-android.o.
17882 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
17883 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
17884 * varasm.c (do_assemble_alias): Likewise.
17885 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
17886 doesn't support indirect functions.
17887 * configure: Regenerate.
17889 2013-03-27 Bin Cheng <bin.cheng@arm.com>
17892 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
17893 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
17896 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17898 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
17900 2013-03-27 Terry Guo <terry.guo@arm.com>
17902 * config/arm/arm-cores.def: Added core cortex-r7.
17903 * config/arm/arm-tune.md: Regenerated.
17904 * config/arm/arm-tables.opt: Regenerated.
17905 * doc/invoke.texi: Added entry for core cortex-r7.
17907 2013-03-27 Walter Lee <walt@tilera.com>
17909 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
17910 double-decrement of next_scratch_regno.
17912 2013-03-27 Walter Lee <walt@tilera.com>
17914 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
17916 (insn_v1mulus): Ditto.
17917 (insn_v2muls): Ditto.
17919 2013-03-27 Walter Lee <walt@tilera.com>
17921 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
17922 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
17924 2013-03-27 Walter Lee <walt@tilera.com>
17926 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
17927 (*sibcall_value): Ditto.
17929 2013-03-27 Walter Lee <walt@tilera.com>
17931 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
17932 (insn_mnz_v8qi): ... this ...
17933 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
17935 (insn_v<n>mnz): Replaced by ...
17936 (insn_v1mnz): ... this ...
17937 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
17939 (insn_mz_<mode>): Replaced by ...
17940 (insn_mz_v8qi): ... this ...
17941 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
17943 (insn_v<n>mz): Replaced by ...
17944 (insn_v1mz): ... this ...
17945 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
17948 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
17950 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
17952 2013-03-26 Roland McGrath <mcgrathr@google.com>
17954 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
17955 than fprintf with a non-constant, non-format string.
17957 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
17959 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
17960 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
17961 operand 0 predicate.
17962 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
17963 attribute. Use general_x64nomem_operand as operand 1 predicate.
17964 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
17965 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
17966 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
17967 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
17968 (mov<mode>_insv_1): Remove expander. Merge insn with
17969 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
17970 Use general_x64nomem_operand as operand 1 predicate.
17971 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
17972 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
17973 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
17974 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
17975 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
17976 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
17977 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
17978 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
17979 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
17980 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
17981 (general_x64nomem_operand): Ditto.
17983 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
17985 * config/rtems.opt: Add -pthread option.
17987 2013-03-26 Richard Biener <rguenther@suse.de>
17989 * alias.c (find_base_term): Avoid redundant and not used recursion.
17990 (base_alias_check): Get the initial base term from the caller.
17991 (true_dependence_1): Compute and pass base terms to base_alias_check.
17992 (write_dependence_p): Likewise.
17993 (may_alias_p): Likewise.
17995 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
17997 * config/aarch64/aarch64.c (aarch64_classify_address): Support
17998 PC-relative load in SI modes and above only.
18000 2013-03-26 Xinyu Qi <xyqi@marvell.com>
18002 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
18003 * config/arm/iwmmxt.md (WCGR0): Update.
18004 (WCGR1, WCGR2, WCGR3): Likewise.
18006 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
18008 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
18009 Use x64 and nox64 isa attributes.
18011 2013-03-26 Richard Biener <rguenther@suse.de>
18013 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
18014 alignment computations and rely on get_object_alignment_1
18015 for the !TYPE_P case.
18016 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
18018 2013-03-26 Walter Lee <walt@tilera.com>
18020 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
18021 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
18023 2013-03-25 Jeff Law <law@redhat.com>
18025 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
18026 check for INTEGRAL_TYPE_P that was missing due to checking in
18027 wrong version of prior patch.
18029 2013-03-25 Walter Lee <walt@tilera.com>
18031 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
18032 TILEGX_INSN_SHUFFLEBYTES1.
18033 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
18035 (tilegx_builtins): Ditto.
18036 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
18038 2013-03-25 Walter Lee <walt@tilera.com>
18040 * config/tilegx/tilegx.md (floatsisf2): New pattern.
18041 (floatunssisf2): New pattern.
18042 (floatsidf2): New pattern.
18043 (floatunssidf2): New pattern.
18045 2013-03-25 Walter Lee <walt@tilera.com>
18047 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
18048 tests for constraint J, K, N, P.
18050 2013-03-25 Walter Lee <walt@tilera.com>
18052 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
18053 Use indirect/pcrel encoding.
18054 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
18057 2013-03-25 Steve Ellcey <sellcey@mips.com>
18059 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
18060 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
18061 * config/mips/mips.c (mips_option_override): Set IMADD default.
18062 * config/mips/mips.h (PTF_AVOID_IMADD): New.
18063 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
18064 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
18065 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
18066 * doc/invoke.texi (-mimadd/-mno-imadd): New.
18068 2013-03-25 Jeff Law <law@redhat.com>
18070 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
18071 slightly to avoid creating and folding useless trees. Simplify
18072 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
18074 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
18076 * config/i386/i386.md (*zero_extendsidi2): Merge with
18077 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
18078 * config/i386/predicates.md (x86_64_zext_operand): Rename from
18079 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
18080 targets. Clarify comment.
18082 2013-03-25 Martin Jambor <mjambor@suse.cz>
18084 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
18085 pass-through jump functions differently.
18086 (ipa_read_jump_function): Likewise. Also use setter functions to set
18089 2013-03-25 Martin Jambor <mjambor@suse.cz>
18091 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
18092 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
18094 (ipa_get_indirect_edge_target): New function.
18095 (devirtualization_time_bonus): New parameter known_aggs, pass it to
18096 ipa_get_indirect_edge_target. Update all callers.
18097 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
18098 ipa_get_indirect_edge_target_1 instead of calling
18099 ipa_get_indirect_edge_target.
18100 (create_specialized_node): Pass aggvlas to
18101 ipcp_discover_new_direct_edges.
18103 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18105 * config/arm/arm.md (f_sels, f_seld): New types.
18106 (*cmov<mode>): New pattern.
18107 * config/arm/predicates.md (arm_vsel_comparison_operator): New
18110 2013-03-25 Kai Tietz <ktietz@redhat.com>
18112 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
18113 POSIX-printf for mingw-hosted builds.
18115 2013-03-25 Richard Biener <rguenther@suse.de>
18117 PR middle-end/56694
18118 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
18119 must-not-throw stmt location.
18121 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18123 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
18124 Emit load-acquire versions when acq is true.
18125 (arm_emit_store_exclusive): Add rel parameter.
18126 Emit store-release versions when rel is true.
18127 (arm_split_compare_and_swap): Use acquire-release instructions
18129 of barriers when appropriate.
18130 (arm_split_atomic_op): Likewise.
18131 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
18132 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
18133 (VUNSPEC_SLX): Likewise.
18134 (VUNSPEC_LDA): Likewise.
18135 (VUNSPEC_STL): Likewise.
18136 * config/arm/sync.md (atomic_load<mode>): New pattern.
18137 (atomic_store<mode>): Likewise.
18138 (arm_load_acquire_exclusive<mode>): Likewise.
18139 (arm_load_acquire_exclusivesi): Likewise.
18140 (arm_load_acquire_exclusivedi): Likewise.
18141 (arm_store_release_exclusive<mode>): Likewise.
18143 2013-03-25 Catherine Moore <clm@codesourcery.com>
18145 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
18146 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
18147 * config/mip/predicates.md (lwsp_swsp_operand,
18148 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
18149 sb16_operand, db4_operand, db7_operand, ib3_operand,
18150 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
18151 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
18152 andi16_operand): New predicates.
18153 * config/mips/mips.md (compression): New attribute.
18154 (enabled): New attribute.
18155 (length): Consider compression in computing length.
18156 (shift_compression): New code attribute.
18157 (*add<mode>3): New operands. Record compression.
18158 (sub<mode>3): Likewise.
18159 (one_cmpl<mode>2): Likewise.
18160 (*and<mode>3): Likewise.
18161 (*ior<mode>3): Likewise.
18162 (unnamed pattern for xor): Likewise.
18163 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
18164 (*<optab><mode>3): Likewise.
18165 (*mov<mode>_internal: Likewise.
18166 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
18167 (mips_unsigned_immediate_p): New.
18168 (umips_lwsp_swsp_address_p): New.
18169 (m16_based_address_p): New.
18170 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
18171 (mips_unsigned_immediate_p): New prototype.
18172 (lwsp_swsp_address_p): New prototype.
18173 (m16_based_address_p): New prototype.
18174 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
18175 (mips_signed_immediate_p): New function.
18176 (m16_based_address_p): New function.
18177 (lwsp_swsp_address_p): New function.
18178 (mips_print_operand_punctuation): Recognize short delay slot insns
18179 for microMIPS.add<mode>3"
18181 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18184 * config/arm/iterators.md (v_cmp_result): New mode attribute.
18185 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
18187 2013-03-25 Richard Biener <rguenther@suse.de>
18189 PR tree-optimization/56689
18190 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
18193 2013-03-25 Richard Biener <rguenther@suse.de>
18195 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
18197 (memory_references): Likewise.
18198 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
18199 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
18200 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
18201 (gather_mem_refs_in_loops): Fold into ...
18202 (analyze_memory_references): ... this. Move initialization
18203 to tree_ssa_lim_initialize.
18204 (fill_always_executed_in): Rename to ...
18205 (fill_always_executed_in_1): ... this.
18206 (fill_always_executed_in): Move contains_call computation to
18207 this new function from ...
18208 (tree_ssa_lim_initialize): ... here.
18209 (tree_ssa_lim): Call fill_always_executed_in.
18211 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
18213 * postreload.c (reload_combine): Fix code detecting returns.
18215 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
18217 * function.c (emit_use_return_register_into_block): On cc0 targets,
18218 do not emit the sequence between cc0 setter and user.
18220 2013-03-25 Kai Tietz <ktietz@redhat.com>
18222 * config/i386/predicates.md (local_symbolic_operand): Interpret
18223 dll-imported symbols as none-local.
18225 2013-03-25 Richard Biener <rguenther@suse.de>
18227 * tree-ssa-loop-im.c (struct depend): Remove.
18228 (struct lim_aux_data): Make depends a vec of gimples.
18229 (free_lim_aux_data): Adjust.
18230 (add_dependency): Likewise.
18231 (set_level): Likewise.
18233 2013-03-25 Richard Biener <rguenther@suse.de>
18235 PR middle-end/56434
18236 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
18237 the pointer returned by calls with ECF_MALLOC set.
18239 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
18241 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
18243 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
18245 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
18246 using MMXMODE mode iterator.
18247 (*move<mode>_internal): Merge with *movv2sf_internal and
18248 *movv2sf_internal_rex64 using MMXMODE mode iterator.
18250 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
18252 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
18253 (record_last_mem_set_info): Likewise.
18255 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
18256 of XNEWVEC followed by memset.
18257 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
18259 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
18261 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
18262 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
18263 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
18264 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
18265 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
18266 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
18267 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
18268 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
18269 BARRIER_P instead of GET_CODE.
18271 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
18273 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
18274 inaccuracy in the probing code.
18276 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
18277 (ctrapdi4): Likewise.
18279 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
18281 * calls.c (expand_call): Add missing guard to code handling return
18282 of non-BLKmode structures in MSB.
18283 * function.c (expand_function_end): Likewise.
18285 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
18287 * combine.c (try_combine): Adjust comment. Do not add the set of
18288 insn #0 if the destination indirectly is set or dies in insn #2.
18289 Tidy up code to distribute a new note.
18291 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
18293 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
18294 also for alternatives 16 and 17.
18296 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
18298 * config/i386/sse.md (*mov<mode>_internal): Merge with
18299 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
18300 Emit insn template depending on type attribute. Use
18301 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
18302 movd instead of movq mnemonic for interunit moves. Rewrite mode
18303 attribute calculation. Remove unit attribute calculation.
18304 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
18305 Set prefix_data16 attribute for DImode ssemov types.
18306 Use Ym instead of y for SSE-MMX conversion alternatives.
18307 Reorder operand constraints.
18309 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
18311 * df.h (df_insn_delete): Adjust prototype.
18312 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
18313 and let it decide whether mark the basic block dirty.
18314 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
18315 * df-scan.c (df_insn_info_delete): New helper function, split
18316 off from df_insn_delete.
18317 (df_scan_free_bb_info): Use it.
18318 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
18320 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
18321 that the insn is actually an insn and it has a non-NULL basic block.
18322 Do not mark basic block dirty if only deleting a DEBUG_INSN.
18324 2013-03-22 Richard Biener <rguenther@suse.de>
18326 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
18328 (mem_ref_alloc): Do not allocate them.
18329 (refs_independent_p): Do not query or maintain a cache.
18331 2013-03-22 Richard Biener <rguenther@suse.de>
18333 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
18334 (gather_mem_refs_in_loops): Do not compute it.
18335 (analyze_memory_references): Do not allocate it.
18336 (tree_ssa_lim_finalize): Do not free it.
18337 (for_all_locs_in_loop): Do not query all_refs_in_loop.
18339 2013-03-22 Richard Biener <rguenther@suse.de>
18341 * is-a.h (as_a): Use gcc_checking_assert.
18343 2013-03-22 Ian Bolton <ian.bolton@arm.com>
18345 * config/aarch64/aarch64.c (aarch64_print_operand): New
18346 format specifier for printing a constant in hex.
18347 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
18348 format specifier for printing second operand.
18350 2013-03-22 Richard Biener <rguenther@suse.de>
18352 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
18354 (gather_mem_refs_in_loops): Perform store accumulation here.
18355 (create_vop_ref_mapping_loop): Remove.
18356 (create_vop_ref_mapping): Likewise.
18357 (analyze_memory_references): Initialize refs_stored_in_loop.
18358 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
18359 (record_indep_loop): Remove.
18360 (record_dep_loop): New function.
18361 (ref_indep_loop_p_1): Adjust to only walk over references
18362 in the loop, not its subloops.
18363 (ref_indep_loop_p): Rename to ...
18364 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
18365 maintaining a more fine-grained cache.
18366 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
18367 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
18369 2013-03-22 Richard Biener <rguenther@suse.de>
18371 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
18372 (struct mem_ref): Make accesses_in_loop a vec of a vec of
18373 aggregate mem_ref_loc.
18374 (free_mem_ref_locs): Inline into ...
18375 (memref_free): ... this and adjust.
18376 (mem_ref_alloc): Adjust.
18377 (mem_ref_locs_alloc): Remove.
18378 (record_mem_ref_loc): Adjust.
18379 (get_all_locs_in_loop): Rewrite into ...
18380 (for_all_locs_in_loop): ... this iterator.
18381 (rewrite_mem_ref_loc): New functor.
18382 (rewrite_mem_refs): Use for_all_locs_in_loop.
18383 (sm_set_flag_if_changed): New functor.
18384 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
18385 (ref_always_accessed): New functor.
18386 (ref_always_accessed_p): Use for_all_locs_in_loop.
18388 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
18390 * tree-pass.h (PROP_gimple_lvec): New.
18391 * passes.c (dump_properties): Handle PROP_gimple_lvec.
18392 (init_optimization_passes): Move pass_lower_vector.
18393 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
18395 (pass_lower_vector): Provide PROP_gimple_lvec.
18396 (pass_lower_vector_ssa): Likewise.
18397 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
18399 2013-03-21 Mark Wielaard <mjw@redhat.com>
18401 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
18403 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
18405 * config/i386/i386.md (*movdi_internal): Disparage slightly
18406 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
18407 conversion alternatives.
18409 2013-03-21 Jakub Jelinek <jakub@redhat.com>
18411 PR middle-end/48087
18412 * diagnostic.def (DK_WERROR): New kind.
18413 * diagnostic.h (werrorcount): Define.
18414 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
18415 promoted to DK_ERROR, increment DK_WERROR counter instead of
18417 * toplev.c (toplev_main): Call print_ignored_options even if
18418 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
18419 even if just werrorcount is non-zero.
18422 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
18424 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
18425 (string_cst_pool_decl): New function.
18426 (optimize_one_addr_into_implicit_ptr): New function.
18427 (resolve_addr_in_expr): Optimize DWARF location expression
18428 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
18429 which doesn't live in memory, but has DW_AT_location or
18430 DW_AT_const_value, or refers to a string literal, into
18431 DW_OP_GNU_implicit_pointer.
18432 (optimize_location_into_implicit_ptr): New function.
18433 (resolve_addr): If removing DW_AT_location of a variable because
18434 it was DW_OP_addr of address of the variable, but the variable doesn't
18435 live in memory, try to emit const value attribute for the initializer.
18437 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
18439 * tree.h (VECTOR_TYPE_P): New macro.
18440 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
18441 TYPE_MODE): Use it.
18442 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
18443 VEC_COND_EXPR cannot be lvalues.
18444 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
18446 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
18448 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
18449 Restrict the transformation to equal modes.
18451 2013-03-21 Richard Biener <rguenther@suse.de>
18453 PR tree-optimization/39326
18454 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
18455 (MEM_ANALYZABLE): Adjust.
18456 (record_mem_ref_loc): Move bitmap ops ...
18457 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
18458 unanalyzable refs, do not record locations for it.
18459 (analyze_memory_references): Allocate ref zero as shared
18461 (refs_independent_p): Do not test for unanalyzed mems here.
18462 (ref_indep_loop_p_1): Special-case disambiguation against
18463 the unanalyzed ref.
18464 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
18466 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
18468 * config/arm/arm-protos.h (tune_params): Add
18469 prefer_neon_for_64bits field.
18470 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
18471 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
18472 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
18473 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
18474 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
18475 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
18476 (arm_option_override): Handle -mneon-for-64bits new option.
18477 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
18478 (prefer_neon_for_64bits): Declare new variable.
18479 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
18480 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
18481 (arch_enabled): Handle new arch types. Remove support for onlya8
18483 (one_cmpldi2): Use new arch names.
18484 (zero_extend<mode>di2, extend<mode>di2): Ditto.
18485 * config/arm/arm.opt (mneon-for-64bits): Add option.
18486 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
18487 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
18488 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
18490 * doc/invoke.texi (-mneon-for-64bits): Document.
18492 2013-03-21 Richard Biener <rguenther@suse.de>
18494 PR tree-optimization/39326
18495 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
18496 (sort_bbs_in_loop_postorder_cmp): New function.
18497 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
18499 2013-03-21 Richard Biener <rguenther@suse.de>
18501 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
18502 (vect_insert_into_interleaving_chain): Likewise.
18503 (vect_drs_dependent_in_basic_block): Inline ...
18504 (vect_slp_analyze_data_ref_dependence): ... here. New function,
18506 (vect_analyze_data_ref_dependence): ... here. Simplify.
18507 (vect_check_interleaving): Simplify.
18508 (vect_analyze_data_ref_dependences): Likewise. Split out ...
18509 (vect_slp_analyze_data_ref_dependences): ... this new function.
18510 (dr_group_sort_cmp): New function.
18511 (vect_analyze_data_ref_accesses): Compute data-reference groups
18512 here instead of in vect_analyze_data_ref_dependence. Use
18513 a more efficient algorithm.
18514 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
18515 vect_slp_analyze_data_ref_dependences. Call
18516 vect_analyze_data_ref_accesses earlier.
18517 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
18518 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
18519 (vect_slp_analyze_data_ref_dependences): New prototype.
18521 2013-03-21 Richard Biener <rguenther@suse.de>
18523 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
18524 ref is stored in the loop.
18525 (find_refs_for_sm): Walk only over all stores.
18526 (store_motion_loop): Allocate from lim_bitmap_obstack.
18527 (store_motion): Likewise.
18529 2013-03-21 Richard Biener <rguenther@suse.de>
18531 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
18532 Update virtual SSA form.
18534 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
18536 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
18537 * configure: Regenerate.
18538 * config.in: Regenerate.
18539 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
18540 if !HAVE_LD_EH_FRAME_CIEV3.
18542 2013-03-21 Richard Biener <rguenther@suse.de>
18544 * tree-cfg.c (verify_expr_no_block): New function.
18545 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
18546 nor DECL_VALUE_EXPR have locations with associated blocks.
18547 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
18548 (clear_unused_block_pointer): Remove code dealing with
18549 blocks in DECL_DEBUG_EXPR locations.
18551 2013-03-21 Richard Biener <rguenther@suse.de>
18553 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
18554 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
18555 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
18556 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
18557 instead of DECL_DEBUG_EXPR_IS_FROM.
18558 * gimplify.c (gimplify_modify_expr): Likewise.
18559 * tree-cfg.c (verify_expr_location_1): Likewise.
18560 * tree-complex.c (create_one_component_var): Likewise.
18561 * tree-sra.c (create_access_replacement): Likewise.
18562 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
18563 (clear_unused_block_pointer): Likewise.
18564 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
18565 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
18566 * var-tracking.c (var_debug_decl): Likewise.
18567 (track_expr_p): Likewise.
18568 * tree-inline.c (add_local_variables): Likewise. Set
18569 DECL_HAS_DEBUG_EXPR_P after copying it.
18570 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
18571 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
18573 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
18576 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
18577 * configure: Regenerate.
18578 * config.in: Regenerate.
18579 * config/i386/i386.md (*movdf_internal): Use
18580 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
18581 movd instead of movq mnemonic for interunit moves.
18582 (*movdi_internal): Ditto.
18584 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18586 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
18587 (abd<mode>_3): New pattern.
18588 (aba<mode>_3): New pattern.
18589 (fabd<mode>_3): New pattern.
18591 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18593 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
18594 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
18595 occurrence of REGISTER_PREFIX as its empty string.
18597 2013-03-20 Jeff Law <law@redhat.com>
18599 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
18600 addititional equivalences for equality comparisons between an SSA_NAME
18601 and a constant where the SSA_NAME was set from a widening conversion.
18603 2013-03-20 Walter Lee <walt@tilera.com>
18605 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
18607 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
18609 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
18610 depending on type attribute.
18611 (*movti_internal): Ditto.
18612 (*movtf_internal): Ditto.
18613 (*movxf_internal): Ditto.
18614 (*movdf_internal): Ditto.
18615 (*movsf_internal): Ditto.
18617 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
18619 * config/i386/i386.md (*movti_internal): Set prefix attribute to
18620 maybe_vex for sselog1 and ssemov types.
18621 (*movdi_internal): Reorder operand constraints.
18622 (*movsi_internal): Ditto. Set prefix attribute to
18623 maybe_vex for sselog1 and ssemov types.
18624 (*movtf_internal): Set prefix attribute to maybe_vex
18625 for sselog1 and ssemov types.
18626 (*movdf_internal): Ditto. Set prefix_data16 attribute for
18627 DImode ssemov types. Reorder operand constraints.
18628 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
18629 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
18630 attribute for SImode ssemov types. Reorder operand constraints.
18632 2013-03-20 Martin Jambor <mjambor@suse.cz>
18634 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
18635 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
18637 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
18639 * config/rs6000/predicates.md (indexed_address, update_address_mem
18640 update_indexed_address_mem): New predicates.
18641 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
18642 attribute for load/store instructions.
18643 * config/rs6000/dfp.md (movsd_store): Likewise.
18644 (movsd_load): Likewise.
18645 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
18646 (unnamed HI->DI extend define_insn): Likewise.
18647 (unnamed SI->DI extend define_insn): Likewise.
18648 (unnamed QI->SI extend define_insn): Likewise.
18649 (unnamed QI->HI extend define_insn): Likewise.
18650 (unnamed HI->SI extend define_insn): Likewise.
18651 (unnamed HI->SI extend define_insn): Likewise.
18652 (extendsfdf2_fpr): Likewise.
18653 (movsi_internal1): Likewise.
18654 (movsi_internal1_single): Likewise.
18655 (movhi_internal): Likewise.
18656 (movqi_internal): Likewise.
18657 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
18658 attribute for load/store instructions.
18659 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
18661 (mov<mode>_softfloat): Likewise.
18662 (mov<mode>_hardfloat32): Likewise.
18663 (mov<mode>_hardfloat64): Likewise.
18664 (mov<mode>_softfloat64): Likewise.
18665 (movdi_internal32): Likewise.
18666 (movdi_internal64): Likewise.
18667 (probe_stack_<mode>): Likewise.
18669 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
18671 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
18672 floating point, and decimal floating point to reload iterator.
18674 * config/rs6000/constraints.md (wl constraint): New constraints to
18675 return FLOAT_REGS if certain options are used to reduce the number
18676 of separate patterns that exist in the file.
18677 (wx constraint): Likewise.
18678 (wz constraint): Likewise.
18680 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
18681 -mdebug=reg, print wg, wl, wx, and wz constraints.
18682 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
18683 Initialize the reload functions for 64-bit binary/decimal floating
18685 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
18686 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
18687 create the buffer on the stack to overcome not having a 32-bit
18689 (rs6000_emit_move): Likewise.
18690 (rs6000_secondary_memory_needed_rtx): Likewise.
18691 (rs6000_alloc_sdmode_stack_slot): Likewise.
18692 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
18693 via xxlxor, just like DFmode 0.0.
18695 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
18696 define as 1 if we are running on a power7 or newer.
18697 (enum r6000_reg_class_enum): Add new constraints.
18699 * config/rs6000/dfp.md (movsd): Delete, combine with binary
18700 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
18701 with other moves by using conditional constraits (wg). Use LFIWZX
18702 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
18703 (movsd splitter): Likewise.
18704 (movsd_hardfloat): Likewise.
18705 (movsd_softfloat): Likewise.
18707 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
18708 binary and decimal floating point moves.
18709 (fmove_ok): New attributes to combine binary and decimal floating
18710 point moves, and to combine power6x (mfpgpr) moves along normal
18712 (real_value_to_target): Likewise.
18713 (f32_lr): Likewise.
18714 (f32_lm): Likewise.
18715 (f32_li): Likewise.
18716 (f32_sr): Likewise.
18717 (f32_sm): Likewise.
18718 (f32_si): Likewise.
18719 (movsf): Combine binary and decimal floating point moves. Combine
18720 power6x (mfpgpr) moves with other moves by using conditional
18721 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
18722 (mov<mode> for SFmode/SDmode); Likewise.
18723 (SFmode/SDmode splitters): Likewise.
18724 (movsf_hardfloat): Likewise.
18725 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
18726 (movsf_softfloat): Likewise.
18727 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
18729 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
18730 wx and wz constraints.
18732 * config/rs6000/constraints.md (wg constraint): New constraint to
18733 return FLOAT_REGS if -mmfpgpr (power6x) was used.
18735 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
18738 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
18739 -mdebug=reg, print wg, wl, wx, and wz constraints.
18740 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
18741 Initialize the reload functions for 64-bit binary/decimal floating
18743 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
18744 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
18745 create the buffer on the stack to overcome not having a 32-bit
18747 (rs6000_emit_move): Likewise.
18748 (rs6000_secondary_memory_needed_rtx): Likewise.
18749 (rs6000_alloc_sdmode_stack_slot): Likewise.
18750 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
18751 via xxlxor, just like DFmode 0.0.
18753 * config/rs6000/dfp.md (movdd): Delete, combine with binary
18754 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
18755 with other moves by using conditional constraits (wg). Use LFIWZX
18756 and STFIWX for loading SDmode on power7.
18757 (movdd splitters): Likewise.
18758 (movdd_hardfloat32): Likewise.
18759 (movdd_softfloat32): Likewise.
18760 (movdd_hardfloat64_mfpgpr): Likewise.
18761 (movdd_hardfloat64): Likewise.
18762 (movdd_softfloat64): Likewise.
18764 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
18765 64-bit binary and decimal floating point moves.
18766 (FMOVE64X): Likewise.
18767 (movdf): Combine 64-bit binary and decimal floating point moves.
18768 Combine power6x (mfpgpr) moves with other moves by using
18769 conditional constraits (wg).
18770 (mov<mode> for DFmode/DDmode): Likewise.
18771 (DFmode/DDmode splitters): Likewise.
18772 (movdf_hardfloat32): Likewise.
18773 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
18774 (movdf_softfloat32): Likewise.
18775 (movdf_hardfloat64_mfpgpr): Likewise.
18776 (movdf_hardfloat64): Likewise.
18777 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
18778 (movdf_softfloat64): Likewise.
18779 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
18780 (reload_<mode>_load): Move to later in the file so they aren't in
18781 the middle of the floating point move insns.
18782 (reload_<mode>_store): Likewise.
18784 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
18787 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
18788 constraint if -mdebug=reg.
18789 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
18790 Enable using dd reload support if needed.
18792 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
18793 binary and decimal floating point moves in rs6000.md.
18794 (movtd_internal): Likewise.
18796 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
18797 decimal floating point moves.
18799 (movtf_internal): Likewise.
18800 (mov<mode>_internal, TDmode/TFmode): Likewise.
18801 (movtf_softfloat): Likewise.
18802 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
18804 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
18805 movdi_internal64, using wg constraint for move direct operations.
18806 (movdi_internal64): Likewise.
18808 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
18809 MODES_TIEABLE_P for selected modes. Print the numerical value of
18810 the various virtual registers. Use GPR/FPR first/last values,
18811 instead of hard coding the register numbers. Print which modes
18812 have reload functions registered.
18813 (rs6000_option_override_internal): If -mdebug=reg, trace the options
18814 settings before/after setting cpu, target and subtarget settings.
18815 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
18816 and for secondary reload failures in rs6000_secondary_reload_inner.
18817 (rs6000_secondary_reload_fail): Likewise.
18818 (rs6000_secondary_reload_inner): Likewise.
18820 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
18821 macros for first/last GPR and FPR registers.
18822 (LAST_GPR_REGNO): Likewise.
18823 (FIRST_FPR_REGNO): Likewise.
18824 (LAST_FPR_REGNO): Likewise.
18826 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
18827 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
18828 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
18829 (vcond<mode><mode>): Likewise.
18830 (vcondu<mode><mode>): Likewise.
18831 (vector_gtu<mode>): Likewise.
18832 (vector_gte<mode>): Likewise.
18833 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
18834 to prevent the compiler from converting DImode operations to TImode.
18835 (ior<mode>3): Likewise.
18836 (and<mode>3): Likewise.
18837 (one_cmpl<mode>2): Likewise.
18838 (nor<mode>3): Likewise.
18839 (andc<mode>3): Likewise.
18841 * config/rs6000/constraints.md (wt constraint): New constraint
18842 that returns VSX_REGS if TImode is allowed in VSX registers.
18844 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
18845 constant under VSX.
18847 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
18848 similar to TImode, but it is restricted to being in the GPRs.
18850 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
18851 TImode to occupy a single VSX register.
18853 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
18854 -mvsx-timode for power7/power8.
18855 (power7 cpu): Likewise.
18856 (power8 cpu): Likewise.
18858 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
18859 sure that TFmode/TDmode take up two registers if they are ever
18860 allowed in the upper VSX registers.
18861 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
18863 (rs6000_init_hard_regno_mode_ok): Likewise.
18864 (rs6000_debug_reg_global): Add debugging for PTImode and wt
18865 constraint. Print if LRA is turned on.
18866 (rs6000_option_override_internal): Give an error if -mvsx-timode
18867 and VSX is not enabled.
18868 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
18869 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
18870 to reg+offset addressing. Use PTImode when checking offset
18871 addresses for validity.
18872 (reg_offset_addressing_ok_p): Likewise.
18873 (rs6000_legitimate_offset_address_p): Likewise.
18874 (rs6000_legitimize_address): Likewise.
18875 (rs6000_legitimize_reload_address): Likewise.
18876 (rs6000_legitimate_address_p): Likewise.
18877 (rs6000_eliminate_indexed_memrefs): Likewise.
18878 (rs6000_emit_move): Likewise.
18879 (rs6000_secondary_reload): Likewise.
18880 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
18881 reloads to fpr registers to continue to use reg+offset addressing,
18882 but 64-bit reloads to altivec registers need reg+reg addressing.
18883 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
18884 it. Treat LO_SUM like a PLUS operation.
18885 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
18886 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
18887 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
18888 registers to share a register with a smaller sized type, since VSX
18889 puts scalars in the upper 64-bits.
18890 (print_operand): Add support for PTImode.
18891 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
18892 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
18893 registers, but don't have arithmetic support.
18894 (rs6000_memory_move_cost): Add test for VSX.
18895 (rs6000_opt_masks): Add -mvsx-timode.
18897 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
18900 (VSr): Use wt constraint for TImode.
18901 (VSv): Drop TImode support.
18902 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
18903 (vsx_movti_64bit): Likewise.
18904 (vsx_movti_32bit): Likewise.
18905 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
18906 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
18907 one '?' on the appropriate output constraint. Do not allow TImode
18908 logical operations on 32-bit systems.
18909 (vsx_ior<mode>3): Likewise.
18910 (vsx_xor<mode>3): Likewise.
18911 (vsx_one_cmpl<mode>2): Likewise.
18912 (vsx_nor<mode>3): Likewise.
18913 (vsx_andc<mode>3): Likewise.
18914 (vsx_concat_<mode>): Likewise.
18915 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
18917 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
18918 OPTION_MASK_VSX_TIMODE.
18919 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
18920 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
18922 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
18923 (TI2 iterator): New iterator for TImode, PTImode.
18924 (wd mode attribute): Add values for vector types.
18925 (movti_string): Replace TI move operations with operations for TImode
18926 and PTImode. Add support for TImode being allowed in VSX registers.
18927 (mov<mode>_string, TImode/PTImode): Likewise.
18928 (movti_ppc64): Likewise.
18929 (mov<mode>_ppc64, TImode/PTImode): Likewise.
18930 (TI mode splitters): Likewise.
18932 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
18935 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
18937 PR tree-optimization/56355
18938 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
18939 Also handle integers with undefined overflow.
18941 2013-03-20 Catherine Moore <clm@codesourcery.com>
18942 Maciej W. Rozycki <macro@codesourcery.com>
18943 Tom de Vries <tom@codesourcery.com>
18944 Nathan Sidwell <nathan@codesourcery.com>
18945 Iain Sandoe <iain@codesourcery.com>
18946 Nathan Froyd <froydnj@codesourcery.com>
18947 Chao-ying Fu <fu@mips.com>
18949 * doc/extend.texi: (micromips, nomicromips, nocompression):
18950 Document new function attributes.
18951 * doc/invoke.texi (minterlink-compressed, mmicromips,
18952 m14k, m14ke, m14kec): Document new options.
18953 (minterlink-mips16): Update documentation.
18954 * doc/md.texi (ZC, ZD): Document new constraints.
18955 * configure.ac (gcc_cv_as_micromips): Check if linker
18956 supports the .set micromips directive.
18957 * configure: Regenerate.
18958 * config.in: Regenerate.
18959 * config/mips/mips-tables.opt: Regenerate.
18960 * config/mips/micromips.md: New file.
18961 * constraints.md (ZC, ZD): New constraints.
18962 * config/mips/predicates.md (movep_src_register): New predicate.
18963 (movep_src_operand): New predicate.
18964 (non_volatile_mem_operand): New predicate.
18965 * config/mips/mips.md (multimem): New type.
18966 (length): Differentiate between 17-bit and 18-bit branch offsets.
18967 (MOVEP1, MOVEP2): New mode iterator.
18968 (mov_<load>l): Use ZC constraint.
18969 (mov_<load>r): Likewise.
18970 (mov_<store>l): Likewise.
18971 (mov_<store>r): Likewise.
18972 (*branch_equality<mode>_inverted): Add microMIPS support.
18973 (*branch_equality<mode>): Likewise.
18974 (*jump_absolute): Likewise.
18975 (indirect_jump_<mode>): Likewise.
18976 (tablejump_<mode>): Likewise.
18977 (<optab>_internal): Likewise.
18978 (sibcall_internal): Likewise.
18979 (sibcall_value_internal): Likewise.
18980 (prefetch): Use constraint ZD.
18981 * config/mips/mips.opt (minterlink-compressed): New option.
18982 (minterlink-mips16): Now an alias for minterlink-compressed.
18983 (mmicromips): New option.
18984 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
18985 (compare_and_swap_12): Likewise.
18986 (sync_add<mode>): Likewise.
18987 (sync_<optab>_12): Likewise.
18988 (sync_old_<optab>_12): Likewise.
18989 (sync_new_<optab>_12): Likewise.
18990 (sync_nand_12): Likewise.
18991 (sync_old_nand_12): Likewise.
18992 (sync_new_nand_12): Likewise.
18993 (sync_sub<mode>): Likewise.
18994 (sync_old_add<mode>): Likewise.
18995 (sync_old_sub<mode>): Likewise.
18996 (sync_new_add<mode>): Likewise.
18997 (sync_new_sub<mode>): Likewise.
18998 (sync_<optab><mode>): Likewise.
18999 (sync_old_<optab><mode>): Likewise.
19000 (sync_new_<optab><mode>): Likewise.
19001 (sync_nand<mode>): Likewise.
19002 (sync_old_nand<mode>): Likewise.
19003 (sync_new_nand<mode>): Likewise.
19004 (sync_lock_test_and_set<mode>): Likewise.
19005 (test_and_set_12): Likewise.
19006 (atomic_compare_and_swap<mode>): Likewise.
19007 (atomic_exchange<mode>_llsc): Likewise.
19008 (atomic_fetch_add<mode>_llsc): Likewise.
19009 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
19010 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
19011 (umips_save_restore_pattern_p): Likewise.
19012 (umips_load_store_pair_p): Likewise.
19013 (umips_output_load_store_pair): Likewise.
19014 (umips_movep_target_p): Likewise.
19015 (umips_12bit_offset_address_p): Likewise.
19016 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
19017 (mips_base_mips16): Rename this...
19018 (mips_base_compression_flags): ...to this. Update all uses.
19019 (mips_attribute_table): Add micromips, nomicromips and nocompression.
19020 (mips_mips16_decl_p): Delete.
19021 (mips_nomips16_decl_p): Delete.
19022 (mips_get_compress_on_flags): New function.
19023 (mips_get_compress_off_flags): New function.
19024 (mips_get_compress_mode): New function.
19025 (mips_get_compress_on_name): New function.
19026 (mips_get_compress_off_name): New function.
19027 (mips_insert_attributes): Support multiple compression types.
19028 (mips_merge_decl_attributes): Likewise.
19029 (umips_12bit_offset_address_p): New function.
19030 (mips_start_function_definition): Emit .set micromips directive.
19031 (mips_call_may_need_jalx_p): New function.
19032 (mips_function_ok_for_sibcall): Add microMIPS support.
19033 (mips_print_operand_punctuation): Support short delay slots and
19035 (umips_swm_mask, umips_swm_encoding): New.
19036 (umips_build_save_restore): New function.
19037 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
19038 (was_mips16_p): Remove.
19039 (old_compression_mode): New.
19040 (mips_set_compression_mode): New function.
19041 (mips_set_current_function): Add microMIPS support.
19042 (mips_option_override): Likewise.
19043 (umips_save_restore_pattern_p): New function.
19044 (umips_output_save_restore): New function.
19045 (umips_load_store_pair_p_1): New function.
19046 (umips_load_store_pair_p): New function.
19047 (umips_output_load_store_pair_1): New function.
19048 (umips_output_load_store_pair): New function.
19049 (umips_movep_target_p) New function.
19050 (mips_prepare_pch_save): Add microMIPS support.
19051 * config/mips/mips.h (TARGET_COMPRESSION): New.
19052 (TARGET_CPU_CPP_BUILTINS): Update macro
19053 to use new compression flags and to support microMIPS.
19054 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
19055 (MIPS_ARCH_FLOAT_SPEC): Likewise.
19056 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
19057 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
19058 (ASM_SPEC): Support mmicromips and mno-micromips.
19059 (M16STORE_REG_P): New macro.
19060 (MIPS_CALL): Support TARGET_MICROMIPS.
19061 (MICROMIPS_J): New macro.
19062 (mips_base_mips16): Rename this...
19063 (mips_base_compression_flags): ...to this.
19064 (UMIPS_12BIT_OFFSET_P): New macro.
19065 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
19066 (MULTILIB_DIRNAMES): Likewise.
19067 2013-03-20 Richard Biener <rguenther@suse.de>
19069 PR tree-optimization/56661
19070 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
19071 the result does not have to be distinct.
19073 2013-03-20 Richard Biener <rguenther@suse.de>
19075 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
19078 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19079 Steven Bosscher <steven@gcc.gnu.org>
19081 PR rtl-optimization/56605
19082 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
19084 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
19087 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
19088 that require movd instead of movq.
19090 2013-03-20 Richard Biener <rguenther@suse.de>
19092 * tree-ssa-structalias.c (struct variable_info): Add pointer
19093 to the first field of an aggregate with sub-vars. Make
19094 this and the pointer to the next subfield its ID.
19095 (vi_next): New function.
19096 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
19097 storedanything_id, integer_id): Increment by one.
19098 (new_var_info, get_call_vi, lookup_call_clobber_vi,
19099 get_call_clobber_vi): Adjust.
19100 (solution_set_expand): Simplify and speedup.
19101 (solution_set_add): Inline into ...
19102 (set_union_with_increment): ... this. Adjust accordingly.
19103 (do_sd_constraint): Likewise.
19104 (do_ds_constraint): Likewise.
19105 (do_complex_constraint): Simplify.
19106 (build_pred_graph): Adjust.
19107 (solve_graph): Likewise. Simplify and speedup.
19108 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
19109 get_constraint_for_component_ref, get_constraint_for_1,
19110 first_vi_for_offset, first_or_preceding_vi_for_offset,
19111 create_function_info_for, create_variable_info_for_1,
19112 create_variable_info_for, intra_create_variable_infos): Adjust.
19113 (init_base_vars): Push NULL for ID zero.
19114 (compute_points_to_sets): Adjust.
19116 2013-03-20 Richard Biener <rguenther@suse.de>
19118 * cfgloop.c (verify_loop_structure): Streamline and avoid
19119 ICEing on corrupt loop tree.
19120 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
19123 2013-03-20 Richard Biener <rguenther@suse.de>
19125 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
19126 check whether an SSA update is needed.
19128 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
19130 * config/mips/constraints.md (T): Rename to...
19134 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
19135 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
19137 2013-03-19 Ian Bolton <ian.bolton@arm.com>
19139 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
19140 (*subsi3_carryin_uxtw): Likewise.
19142 2013-03-19 Ian Bolton <ian.bolton@arm.com>
19144 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
19145 (*rorsi3_insn_uxtw): Likewise.
19147 2013-03-19 Ian Bolton <ian.bolton@arm.com>
19149 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
19150 (*extrsi5_insn_uxtw): Likewise.
19152 2013-03-19 Richard Biener <rguenther@suse.de>
19154 PR tree-optimization/56273
19155 * passes.c (init_optimization_passes): Move second VRP after DOM.
19157 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
19159 * config/i386/i386.md (*movti_internal): Merge from
19160 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
19161 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
19162 nox64 isa attributes.
19164 2013-03-18 Richard Biener <rguenther@suse.de>
19166 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
19168 (merge_node_constraints): Likewise.
19169 (build_succ_graph): Likewise.
19170 (valid_graph_edge): Inline into single caller.
19171 (unify_nodes): Likewise. Use bitmap_set_bit return value
19173 (scc_visit): Fix formatting and variable use.
19174 (do_sd_constraint): Use gcc_checking_assert.
19175 (do_ds_constraint): Likewise.
19176 (do_complex_constraint): Likewise.
19177 (condense_visit): Likewise. Cleanup.
19178 (dump_pred_graph): New function.
19179 (perform_var_substitution): Dump the pred-graph before
19180 variable substitution.
19181 (find_equivalent_node): Use gcc_checking_assert.
19182 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
19184 2013-03-18 Richard Biener <rguenther@suse.de>
19186 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
19187 Remove cond_expr_stmt_list argument and do not gimplify the
19189 (vect_loop_versioning): Adjust.
19190 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
19191 Cleanup to use less temporaries.
19192 (vect_create_data_ref_ptr): Cleanup.
19194 2013-03-18 Jakub Jelinek <jakub@redhat.com>
19196 PR tree-optimization/56635
19197 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
19198 require types_compatible_p types.
19200 2013-03-18 Nick Clifton <nickc@redhat.com>
19202 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
19203 spurious backslash.
19205 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
19206 Add missing line to comment describing function.
19208 2013-03-18 Richard Biener <rguenther@suse.de>
19210 PR tree-optimization/56210
19211 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
19212 Handle string / character search functions.
19213 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
19215 2013-03-18 Richard Biener <rguenther@suse.de>
19217 PR middle-end/56483
19218 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
19219 and implement properly.
19220 * gimple.h (gimple_cond_single_var_p): Remove.
19222 2013-03-18 Richard Biener <rguenther@suse.de>
19224 * tree-data-ref.h (find_data_references_in_loop): Declare.
19225 * tree-data-ref.c (get_references_in_stmt): Use a stack
19226 vector pre-allocated in the callers.
19227 (find_data_references_in_stmt): Adjust.
19228 (graphite_find_data_references_in_stmt): Likewise.
19229 (create_rdg_vertices): Likewise.
19230 (find_data_references_in_loop): Export.
19231 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
19232 Compute dependences here...
19233 (vect_analyze_data_refs): ...not here. When we encounter
19234 a non-vectorizable data reference in basic-block vectorization
19235 truncate the data reference vector. Do not bother to
19236 fixup data-dependence information for gather loads.
19237 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
19238 of data references, as reported.
19240 2013-03-18 Richard Biener <rguenther@suse.de>
19242 PR tree-optimization/3713
19243 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
19244 has_constants and expr.
19245 (stmt_has_constants): Properly valueize SSA names when deciding
19246 whether the stmt has constants.
19248 2013-03-18 Richard Biener <rguenther@suse.de>
19250 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
19251 whole function when there is nothing to do.
19252 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
19253 * tree-vectorizer.c (vectorize_loops): Update virtual and
19254 loop-closed SSA once.
19255 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
19257 2013-03-18 Richard Biener <rguenther@suse.de>
19259 PR middle-end/56113
19260 * domwalk.c (bb_postorder): New global static.
19261 (cmp_bb_postorder): New function.
19262 (walk_dominator_tree): Replace scheme imposing an order for
19263 visiting dominator sons by one sorting them at the time they
19264 are pushed on the stack.
19266 2013-03-18 Richard Biener <rguenther@suse.de>
19268 PR tree-optimization/39326
19269 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
19270 (struct mem_ref): Replace mem member with ao_ref typed member.
19271 (MEM_ANALYZABLE): Adjust.
19272 (memref_eq): Likewise.
19273 (mem_ref_alloc): Likewise.
19274 (gather_mem_refs_stmt): Likewise.
19275 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
19276 (execute_sm_if_changed_flag_set): Adjust.
19277 (execute_sm): Likewise.
19278 (ref_always_accessed_p): Likewise.
19279 (refs_independent_p): Likewise.
19280 (can_sm_ref_p): Likewise.
19282 2013-03-18 Jakub Jelinek <jakub@redhat.com>
19285 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
19286 return 1 even for !unsignedp.
19288 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
19290 * config/i386/i386.md (isa): Add x64 and nox64.
19291 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
19292 (*pushtf): Enable *roF alternative for x64 isa only.
19293 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
19294 mode attribute of integer alternatives to DImode for TARGET_64BIT.
19295 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
19296 (*movtf_internal): Merge from *movtf_internal_rex64 and
19297 *movtf_internal_sse. Use x64 and nox64 isa attributes.
19298 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
19299 nox64 isa attributes.
19300 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
19301 nox64 isa attributes.
19302 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
19304 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
19306 * config/alpha/alpha.c (TARGET_LRA_P): New define.
19308 2013-03-17 Jakub Jelinek <jakub@redhat.com>
19311 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
19312 class names. Remove trailing comma after "ALL_REGS".
19314 2013-03-16 Jan Hubicka <jh@suse.cz>
19316 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
19317 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
19318 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
19319 of cgraph_get_create_node.
19320 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
19322 2013-03-16 Jason Merrill <jason@redhat.com>
19325 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
19326 with DW_AT_default_value.
19328 2013-03-16 Jakub Jelinek <jakub@redhat.com>
19330 * BASE-VER: Set to 4.9.0.
19332 2013-03-14 Andi Kleen <ak@linux.intel.com>
19335 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
19336 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
19337 Document _x* TSX intrinsics.
19339 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
19340 David Holsgrove <david.holsgrove@xilinx.com>
19342 * configure.ac: Add MicroBlaze TLS support detection.
19343 * configure: Regenerate.
19344 * config/microblaze/microblaze-protos.h
19345 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
19346 symbol_mentioned_p, label_mentioned_p): Add prototypes.
19347 * config/microblaze/microblaze.c (microblaze_address_type): Add
19348 ADDRESS_TLS and tls_reloc address types.
19349 (microblaze_address_info): Add tls_reloc.
19350 (TARGET_HAVE_TLS): Define.
19351 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
19352 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
19353 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
19354 load_tls_operand, microblaze_call_tls_get_addr,
19355 microblaze_legitimize_tls_address): New functions.
19356 (microblaze_classify_unspec): Handle UNSPEC_TLS.
19357 (get_base_reg): Use microblaze_tls_symbol_p.
19358 (microblaze_classify_address): Handle TLS.
19359 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
19360 label_mentioned_p and microblaze_tls_referenced_p.
19361 (microblaze_legitimize_address): Handle TLS.
19362 (microblaze_address_insns): Handle ADDRESS_TLS.
19363 (pic_address_needs_scratch): Handle TLS.
19364 (print_operand_address): Handle TLS.
19365 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
19366 (microblaze_expand_move): Handle TLS.
19367 (microblaze_legitimate_constant_p): Check
19368 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
19369 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
19370 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
19371 (PIC_OFFSET_TABLE_REGNUM): Set.
19372 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
19373 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
19374 (addsi3, movsi_internal2, movdf_internal): Update constraints
19375 * config/microblaze/predicates.md (arith_plus_operand): Define
19376 (move_operand): Redefine as move_src_operand,
19377 check microblaze_tls_referenced_p.
19379 2013-03-14 Ian Bolton <ian.bolton@arm.com>
19381 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
19382 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
19384 2013-03-14 Ian Bolton <ian.bolton@arm.com>
19386 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
19389 2013-03-14 Jakub Jelinek <jakub@redhat.com>
19391 PR tree-optimization/53265
19392 * common.opt (Waggressive-loop-optimizations): New option.
19393 * tree-ssa-loop-niter.c: Include tree-pass.h.
19394 (do_warn_aggressive_loop_optimizations): New function.
19395 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
19396 if number_of_latch_executions returned constant.
19397 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
19398 early. If number_of_latch_executions returned constant, set
19399 nb_iterations_upper_bound back to it.
19400 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
19402 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
19403 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
19405 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
19406 (MULTILIB_OSDIRNAMES): Set.
19407 * genmultilib: If defaultosdirname doesn't start with :: , set
19408 defaultosdirname2 instead, clear it and emit two . multilib_raw
19409 entries instead of just one.
19411 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
19413 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
19414 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
19415 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
19416 (SUBTARGET_OVERRIDE_OPTIONS): New.
19418 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
19421 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
19422 (musermode): Convert to Var(TARGET_USERMODE).
19423 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
19424 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
19425 * config/sh/sh.c (sh_option_override): Use
19426 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
19427 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
19429 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
19431 (udivsi3_i4_single, divsi3_i4_single): Use
19432 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
19434 2013-03-13 Dave Korn <dave.korn.cygwin@....>
19436 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
19439 2013-03-13 Richard Biener <rguenther@suse.de>
19441 PR tree-optimization/56608
19442 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
19443 calls when vectorizing basic-blocks.
19445 2013-03-13 Jakub Jelinek <jakub@redhat.com>
19448 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
19451 2013-03-12 Jakub Jelinek <jakub@redhat.com>
19453 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
19455 2013-03-11 Jan Hubicka <jh@suse.cz>
19458 * lto-streamer-out.c (output_symbol_p): Skip references from
19459 constructors of external variables.
19461 2013-03-11 Jan Hubicka <jh@suse.cz>
19463 PR middle-end/56571
19464 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
19466 * emit-rtl.c (verify_rtx_sharing): Likewise.
19467 (copy_insn_1): Likewise.
19468 * rtl.c (copy_rtx): Likewise.
19470 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
19473 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
19474 output_operand_lossage message.
19476 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
19479 * arm.c (shift_op): Validate RTL pattern on the fly.
19480 (arm_print_operand, case 'S'): Don't use shift_operator to validate
19483 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19486 * config/pa/pa.md (call_value): Check for calls to powf and direct to
19487 new call patterns that clobber %fr12.
19488 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
19489 split and postreload patterns.
19490 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
19491 registers %fr12 and %fr12R as call used.
19493 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
19495 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
19496 (canon_address, record_store, replace_read, check_mem_read_rtx,
19497 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
19498 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
19499 rest_of_handle_dse): Likewise.
19501 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
19503 PR middle-end/56524
19504 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
19506 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
19507 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
19508 (save_optabs_if_changed): Replace with...
19509 (init_tree_optimization_optabs): ...this.
19510 * optabs.c (save_optabs_if_changed): Rename to...
19511 (init_tree_optimization_optabs): ...this. Take the optimization node
19512 as argument. Do nothing if the base optabs are already correct.
19513 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
19514 to recompute optabs.
19515 * function.h (function): Remove optabs field.
19516 * function.c (invoke_set_current_function_hook): Call
19517 init_tree_optimization_optabs. Use the result to initialize
19520 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
19522 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
19523 if GTMA_HAS_NO_INSTRUMENTATION.
19524 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
19525 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
19526 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
19527 * gimple-pretty-print.c (dump_gimple_transaction): Handle
19528 GTMA_HAS_NO_INSTRUMENTATION.
19530 2013-03-08 Jakub Jelinek <jakub@redhat.com>
19532 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
19535 2013-03-08 Marek Polacek <polacek@redhat.com>
19536 Jakub Jelinek <jakub@redhat.com>
19538 PR tree-optimization/56478
19539 * predict.c (is_comparison_with_loop_invariant_p): Change the
19540 type of loop_step to tree.
19541 (predict_loops): Adjust.
19542 (predict_iv_comparison): Perform the computations on double_ints.
19544 2013-03-08 Richard Biener <rguenther@suse.de>
19546 PR tree-optimization/56570
19547 * tree-cfg.c (verify_expr_location_1): Verify locations for
19549 * tree-sra.c (create_access_replacement): Strip locations
19550 from DECL_DEBUG_EXPRs.
19552 2013-03-08 Richard Biener <rguenther@suse.de>
19554 * tree-inline.c (expand_call_inline): Do not associate
19555 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
19556 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
19558 2013-03-08 Richard Biener <rguenther@suse.de>
19560 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
19561 or block changes with -Og. Fix for location / block encoding
19562 changes and PHI arguments with locations.
19564 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
19566 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
19568 (struct output_info): Likewise.
19569 (register_overhead): Remove bad gcc_assert.
19570 (bitmap_find_bit): If there is only a single bitmap element, do not
19571 count a miss as a search.
19572 (print_statistics): Update for counter type changes.
19573 (dump_bitmap_statistics): Likewise. Print headers such that they
19574 are properly lined up with the printed counters.
19576 2013-03-07 Jakub Jelinek <jakub@redhat.com>
19578 PR tree-optimization/56559
19579 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
19580 check that it has only a single use.
19582 2013-03-07 Richard Biener <rguenther@suse.de>
19584 * doc/invoke.texi (fwhole-program): Discourage use in combination
19587 2013-03-06 Jakub Jelinek <jakub@redhat.com>
19589 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
19591 PR tree-optimization/56539
19592 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
19593 instead of GSI_CONTINUE_LINKING as last argument to
19594 force_gimple_operand_gsi. Adjust function comment.
19596 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
19599 PR middle-end/56548
19600 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
19601 promoted mode, convert the result back to the original mode.
19603 2013-03-06 Richard Biener <rguenther@suse.de>
19605 PR middle-end/56294
19606 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
19607 (insert_updated_phi_nodes_compare_uids): New function.
19608 (update_ssa): Sort symbols_to_rename after UID before
19609 traversing it to insert PHI nodes.
19611 2013-03-06 Richard Biener <rguenther@suse.de>
19613 PR middle-end/50494
19614 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
19615 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
19618 2013-02-13 Richard Biener <rguenther@suse.de>
19621 * varasm.c (output_constant_def_1): Get the decl representing
19622 the constant as argument.
19623 (output_constant_def): Wrap output_constant_def_1.
19624 (make_decl_rtl): Use output_constant_def_1 with the decl
19625 representing the constant.
19626 (build_constant_desc): Optionally re-use a decl already
19627 representing the constant.
19628 (tree_output_constant_def): Adjust.
19630 2013-03-06 Joey Ye <joey.ye@arm.com>
19633 * gcc.c (convert_white_space): New function.
19634 (main): Handles white space in function name.
19636 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
19639 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
19640 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
19641 to SH_DIV_CALL_TABLE for TARGET_SH2.
19642 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
19644 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
19645 call-table options.
19647 2013-03-05 Sterling Augustine <saugustine@google.com>
19648 Cary Coutant <ccoutant@google.com>
19651 * dwarf2out.c (resolve_addr): Don't call
19652 remove_loc_list_addr_table_entries a second time for the same
19655 2013-03-05 Jakub Jelinek <jakub@redhat.com>
19658 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
19659 (avoid_complex_debug_insns): New function.
19660 (expand_debug_locations): Call it.
19662 PR rtl-optimization/56484
19663 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
19664 lifetimes of hard registers on small register class machines.
19666 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
19668 * config/microblaze/microblaze-protos.h: Rename
19669 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
19670 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
19672 (microblaze_fast_interrupt_function_p): New function.
19673 (microblaze_is_interrupt_handler): Rename to
19674 microblaze_is_interrupt_variant and add fast_interrupt check.
19675 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
19676 (save_restore_insns): Likewise.
19677 (compute_frame_size): Likewise.
19678 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
19679 (microblaze_globalize_label): Likewise.
19680 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
19681 * config/microblaze/microblaze.md: Use wrapper
19682 microblaze_is_interrupt_variant.
19684 2013-03-05 Kai Tietz <ktietz@redhat.com>
19686 * sdbout.c (sdbout_one_type): Switch to current function's section
19687 supporting cold/hot.
19689 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
19691 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
19694 2013-03-05 Jakub Jelinek <jakub@redhat.com>
19696 PR middle-end/56461
19697 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
19698 if VALGRIND_GET_VBITS is defined, temporarily make object
19699 memory all defined, and restore previous valgrind addressability
19700 and definability afterwards. Free this_object at the end.
19702 PR middle-end/56461
19703 * lra.c (lra): Call lra_clear_live_ranges if live_p,
19704 right before calling lra_create_live_ranges, also call it
19705 when clearing live_p. Only call lra_clear_live_ranges
19706 at the end if live_p.
19708 PR middle-end/56461
19709 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
19711 2013-03-05 Richard Biener <rguenther@suse.de>
19713 PR tree-optimization/56521
19714 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
19717 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
19720 * except.h (remove_unreachable_eh_regions): New prototype.
19721 * except.c (remove_eh_handler_splicer): New function, split out
19722 of remove_eh_handler.
19723 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
19724 warning about running it on many EH regions one at a time.
19725 (remove_unreachable_eh_regions_worker): New function, walk the
19726 EH tree in depth-first order and remove non-marked regions.
19727 (remove_unreachable_eh_regions): New function.
19728 * tree-eh.c (mark_reachable_handlers): New function, split out
19729 from remove_unreachable_handlers.
19730 (remove_unreachable_handlers): Use mark_reachable_handlers and
19731 remove_unreachable_eh_regions.
19732 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
19733 and remove_unreachable_eh_regions.
19735 2013-03-05 Richard Biener <rguenther@suse.de>
19737 PR middle-end/56525
19738 * loop-init.c (fix_loop_structure): Remove loops in two stages,
19739 not freeing them until the end.
19741 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19743 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
19745 2013-03-05 Richard Biener <rguenther@suse.de>
19747 PR tree-optimization/56270
19748 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
19749 of loads after scheduling an SLP instance.
19751 2013-03-05 Jakub Jelinek <jakub@redhat.com>
19753 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
19755 (check_gcc_parallelize): Run guality.exp as a separate job from
19756 vect.exp with unsorted.exp and $(dg_target_exps) separately from
19757 struct-layout-1.exp with stackalign.exp.
19759 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
19761 PR middle-end/56461
19762 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
19763 load_index sbitmap even if some bit in it isn't set.
19765 PR middle-end/56461
19766 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
19767 (discover_iteration_bound_by_body_walk): Change queues to
19768 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
19769 spelling in comment. Call safe_push on queues[bound_index] directly.
19770 Release queues[queue_index] in every iteration unconditionally.
19771 Release bounds vector.
19773 PR middle-end/56461
19774 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
19775 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
19776 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
19779 2013-03-05 Richard Biener <rguenther@suse.de>
19782 * tree-inline.c (remap_blocks_to_null): New function.
19783 (expand_call_inline): When expanding a call stmt without
19784 an associated block inline remap all callee blocks to NULL.
19786 2013-03-05 Jakub Jelinek <jakub@redhat.com>
19788 PR rtl-optimization/56494
19789 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
19790 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
19791 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
19793 PR middle-end/56461
19794 * sel-sched-ir.c (free_sched_pools): Release
19795 succs_info_pool.stack[succs_info_pool.max_top] vectors too
19796 if succs_info_pool.max_top isn't -1.
19799 * opts.c (opts_obstack, opts_concat): Moved to...
19800 * opts-common.c (opts_obstack, opts_concat): ... here.
19802 2013-03-04 Jakub Jelinek <jakub@redhat.com>
19804 PR middle-end/56461
19805 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
19807 2013-03-04 Martin Jambor <mjambor@suse.cz>
19809 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
19810 all appropriate places.
19812 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
19814 PR tree-optimization/56424
19815 * ipa-split.c (split_function): Do not set the RSO flag if result is
19816 not by reference and its type is a register type.
19818 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
19820 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
19821 (microblaze_legitimate_pic_operand): Likewise
19822 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
19823 new function microblaze_legitimate_pic_operand
19824 * config/microblaze/microblaze-protos.h
19825 (microblaze_legitimate_pic_operand): Declare.
19827 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19829 * config/microblaze/predicates.md (call_insn_simple_operand):
19830 New predicate for supported rtx code types.
19831 * config/microblaze/microblaze.md (call_internal1): Use
19832 call_insn_simple_operand predicate.
19834 2013-03-04 Jakub Jelinek <jakub@redhat.com>
19836 PR middle-end/56461
19837 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
19838 partitions.ordered_remove.
19840 PR middle-end/56461
19841 * tree-vect-stmts.c (vectorizable_conversion): Don't call
19842 vec_oprnds0.create (1) for modifier == NONE.
19844 PR middle-end/56461
19845 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
19846 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
19847 vec_oprnds1 right before pushing anything to it for
19850 PR middle-end/56461
19851 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
19852 set nbbs to 0 instead of having separate code path.
19853 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
19854 instead of false as last argument if returning NULL.
19856 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
19858 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
19859 the attribute is now called "target" instead of "option".
19860 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
19861 * doc/tm.texi.in (Target Attributes): Likewise document the correct
19862 attribute/pragma name for TARGET_OPTION_VALID_P and
19863 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
19864 * doc/tm.texi: Regenerated.
19866 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
19868 * config/microblaze/microblaze.c:
19869 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
19870 * config/microblaze/microblaze.h: Add -mxl-reorder to
19872 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
19873 instructions emitted if TARGET_REORDER.
19874 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
19875 or 0 for -m/-mno case, but initialises as 2 to detect default use case
19878 2013-03-01 Xinliang David Li <davidxl@google.com>
19880 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
19883 2013-03-01 Jakub Jelinek <jakub@redhat.com>
19885 PR middle-end/56461
19886 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
19887 vector even when returning true. Fix up function comment formatting.
19889 PR middle-end/56461
19890 * ira-build.c (ira_loop_nodes_count): New variable.
19891 (create_loop_tree_nodes): Initialize it.
19892 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
19894 PR middle-end/56461
19895 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
19896 method on dr_chain and result_chain.
19897 * tree-vect-stmts.c (vectorizable_store): Only call
19898 result_chain.create if j == 0.
19900 PR middle-end/56461
19901 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
19902 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
19903 before overwriting it.
19905 2013-03-01 Tobias Burnus <burnus@net-b.de>
19907 * doc/extended.texi (C Extensions): Change order in @menu
19909 (Other MIPS Built-in Functions): Move last MIPS entry before
19910 "picoChip Built-in Functions".
19911 (SH Built-in Functions): Move after RX Built-in Functions.
19912 * doc/gcc.texi (Introduction): Change order in @menu
19914 * doc/md.texi (Constraints): Ditto.
19915 * gty.texi (Type Information): Ditto.
19916 (User-provided marking routines for template types): Make
19918 * doc/invoke.texi (AArch64 Options): Move before
19919 "Adapteva Epiphany Options".
19921 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
19922 Jakub Jelinek <jakub@redhat.com>
19925 * asan.c (gate_asan): Lookup no_sanitize_address instead of
19926 no_address_safety_analysis attribute.
19927 * doc/extend.texi (no_address_safety_attribute): Rename to
19928 no_sanitize_address attribute, mention no_address_safety_analysis
19929 attribute as deprecated alias.
19931 2013-02-28 Jakub Jelinek <jakub@redhat.com>
19933 PR middle-end/56461
19934 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
19935 type to vec<vec<tree> > *.
19936 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
19937 to be vec<tree> instead of vec<tree> *, set vec_defs
19938 to vNULL and call vec_defs.create (number_of_vects), adjust other
19940 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
19941 vectorizable_condition): Adjust vect_get_slp_defs callers.
19943 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
19945 * config/aarch64/aarch64.c
19946 (aarch64_float_const_representable): Remove unused variable.
19948 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
19950 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
19952 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
19954 * config/aarch64/aarch64-builtins.c
19955 (aarch64_init_simd_builtins): Make static.
19957 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
19959 * config/aarch64/aarch64.c
19960 (aarch64_simd_make_constant): Make static.
19962 2013-02-28 Martin Jambor <mjambor@suse.cz>
19964 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
19965 with no initialization to the RHS of debug statements.
19967 2013-02-28 Martin Jambor <mjambor@suse.cz>
19969 PR tree-optimization/56294
19970 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
19972 (get_access_replacement): Do not call create_access_replacement.
19973 Assert a replacement exists.
19974 (get_repl_default_def_ssa_name): Create the replacement declaration
19977 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19979 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
19980 final_end_function.
19982 2013-02-28 Marek Polacek <polacek@redhat.com>
19984 PR rtl-optimization/56466
19985 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
19986 if we're changing a loop.
19987 (peel_loops_completely): Likewise.
19989 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
19992 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
19994 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
19997 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
19998 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
19999 INTX_FTYPE_FX, FX_FTYPE_INTX.
20000 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
20002 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
20004 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
20005 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
20006 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
20007 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
20008 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
20009 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
20010 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
20011 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
20012 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
20013 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
20014 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
20015 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
20016 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
20017 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
20018 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
20019 (avrxmega6): Increase max flash segments from 5 to 6.
20020 * config/avr/t-multilib: Regenerate.
20021 * config/avr/avr-tables.opt: Regenerate.
20022 * doc/avr-mmcu.texi: Regenerate.
20024 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
20026 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
20027 (avr_device_to_arch): Rename to avr_device_to_ld.
20028 (avr_device_to_as): New prototype.
20029 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
20030 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
20031 * config/avr/driver-avr.c (avr_device_to_as): New.
20032 (avr_device_to_arch): Rename to avr_device_to_ld.
20034 2013-02-27 Jakub Jelinek <jakub@redhat.com>
20036 PR middle-end/56461
20037 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
20038 method on dr_chain and result_chain.
20040 PR middle-end/56461
20041 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
20042 pointer_set_destroy on not_executed_last_iteration.
20044 PR middle-end/56461
20045 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
20047 PR middle-end/56461
20048 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
20049 FOR_EACH_DEFINED_FUNCTION when freeing state.
20051 PR middle-end/56461
20052 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
20054 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
20057 PR middle-end/56461
20058 * ipa-cp.c (decide_whether_version_node): Call vec_free on
20059 known_aggs[i].items and release known_aggs vector.
20061 PR middle-end/56461
20062 * ipa-reference.c (propagate): Free node_info even for alias nodes.
20064 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20066 * config/microblaze/microblaze.c (microblaze_emit_compare):
20067 Use xor for EQ/NE comparisions.
20068 * config/microblaze/microblaze.md (cstoresf4): Add constraints
20069 (cbranchsf4): Adjust operator to comparison_operator.
20071 2013-02-27 Jakub Jelinek <jakub@redhat.com>
20073 PR middle-end/56461
20074 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
20076 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
20077 vec_safe_push, always update *slot.
20078 (redirect_edge_var_map_clear): Use vec_free.
20079 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
20080 (free_var_map_entry): Use vec_free.
20081 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
20082 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
20084 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
20086 PR middle-end/45472
20087 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
20088 when the may_trap_p bit of the exprs being merged differs.
20089 Reorder tests for speculativeness in the logical and operator.
20091 2013-02-27 Jakub Jelinek <jakub@redhat.com>
20093 * incpath.c (add_standard_paths): Use reconcat instead of concat
20094 where appropriate and avoid leaking memory.
20096 * opts.h: Include obstack.h.
20097 (opts_concat): New prototype.
20098 (opts_obstack): New declaration.
20099 * opts.c (opts_concat): New function.
20100 (opts_obstack): New variable.
20101 (init_options_struct): Call gcc_init_obstack on opts_obstack.
20102 (finish_options): Use opts_concat instead of concat
20103 and XOBNEWVEC instead of XNEWVEC.
20104 * opts-common.c (generate_canonical_option, decode_cmdline_option,
20105 generate_option): Likewise.
20106 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
20107 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
20110 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
20111 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
20113 2013-02-26 Jakub Jelinek <jakub@redhat.com>
20115 PR middle-end/56461
20116 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
20118 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
20120 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
20121 (arm_block_move_unaligned_straight): Likewise.
20122 (arm_adjust_block_mem): Likewise.
20124 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
20127 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
20128 temp, cond and label.
20129 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
20132 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
20133 * config/c6x/c6x.h (dbx_register_map): Update declaration.
20136 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
20137 of prologue/epilogue functions.
20140 * config/tilegx/tilegx.c (tilegx_expand_prologue):
20141 Remove unused variable cfa_offset.
20142 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
20145 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
20146 type promotion to unsigned.
20149 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
20150 for HOST_WIDE_INT of 32 bit / same size as int.
20151 (arm_block_move_unaligned_straight): Likewise.
20152 (arm_adjust_block_mem): Likewise.
20155 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
20158 2013-02-26 Marek Polacek <polacek@redhat.com>
20160 PR tree-optimization/56426
20161 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
20163 2013-02-26 Richard Biener <rguenther@suse.de>
20166 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
20167 unused variable loops.
20169 2013-02-26 Jakub Jelinek <jakub@redhat.com>
20171 PR tree-optimization/56448
20172 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
20173 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
20174 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
20175 later operands of the references, or even first operand for
20176 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
20178 PR tree-optimization/56443
20179 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
20180 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
20181 to type_for_mode langhook.
20183 2013-02-25 Matt Turner <mattst88@gmail.com>
20185 * doc/invoke.texi: Document r4700.
20187 2013-02-25 Richard Biener <rguenther@suse.de>
20189 PR tree-optimization/56175
20190 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
20192 (simplify_bitwise_binary): ... here. Also guard the conversion
20193 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
20195 2013-02-25 Catherine Moore <clm@codesourcery.com>
20198 2013-02-24 Catherine Moore <clm@codesourcery.com>
20199 Maciej W. Rozycki <macro@codesourcery.com>
20200 Tom de Vries <tom@codesourcery.com>
20201 Nathan Sidwell <nathan@codesourcery.com>
20202 Iain Sandoe <iain@codesourcery.com>
20203 Nathan Froyd <froydnj@codesourcery.com>
20204 Chao-ying Fu <fu@mips.com>
20206 * doc/extend.texi: (micromips, nomicromips, nocompression):
20207 Document new function attributes.
20208 * doc/invoke.texi (minterlink-compressed, mmicromips,
20209 m14k, m14ke, m14kec): Document new options.
20210 (minterlink-mips16): Update documentation.
20211 * doc/md.texi (ZC, ZD): Document new constraints.
20212 * configure.ac (gcc_cv_as_micromips): Check if linker
20213 supports the .set micromips directive.
20214 * configure: Regenerate.
20215 * config.in: Regenerate.
20216 * config/mips/mips-tables.opt: Regenerate.
20217 * config/mips/micromips.md: New file.
20218 * constraints.md (ZC, AD): New constraints.
20219 * config/mips/predicates.md (movep_src_register): New predicate.
20220 (movep_src_operand): New predicate.
20221 (non_volatile_mem_operand): New predicate.
20222 * config/mips/mips.md (multimem): New type.
20223 (length): Differentiate between 17-bit and 18-bit branch offsets.
20224 (MOVEP1, MOVEP2): New mode iterator.
20225 (mov_<load>l): Use ZC constraint.
20226 (mov_<load>r): Likewise.
20227 (mov_<store>l): Likewise.
20228 (mov_<store>r): Likewise.
20229 (*branch_equality<mode>_inverted): Add microMIPS support.
20230 (*branch_equality<mode>): Likewise.
20231 (*jump_absolute): Likewise.
20232 (indirect_jump_<mode>): Likewise.
20233 (tablejump_<mode>): Likewise.
20234 (<optab>_internal): Likewise.
20235 (sibcall_internal): Likewise.
20236 (sibcall_value_internal): Likewise.
20237 (prefetch): Use constraint ZD.
20238 * config/mips/mips.opt (minterlink-compressed): New option.
20239 (minterlink-mips16): Now an alias for minterlink-compressed.
20240 (mmicromips): New option.
20241 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
20242 (compare_and_swap_12): Likewise.
20243 (sync_add<mode>): Likewise.
20244 (sync_<optab>_12): Likewise.
20245 (sync_old_<optab>_12): Likewise.
20246 (sync_new_<optab>_12): Likewise.
20247 (sync_nand_12): Likewise.
20248 (sync_old_nand_12): Likewise.
20249 (sync_new_nand_12): Likewise.
20250 (sync_sub<mode>): Likewise.
20251 (sync_old_add<mode>): Likewise.
20252 (sync_old_sub<mode>): Likewise.
20253 (sync_new_add<mode>): Likewise.
20254 (sync_new_sub<mode>): Likewise.
20255 (sync_<optab><mode>): Likewise.
20256 (sync_old_<optab><mode>): Likewise.
20257 (sync_new_<optab><mode>): Likewise.
20258 (sync_nand<mode>): Likewise.
20259 (sync_old_nand<mode>): Likewise.
20260 (sync_new_nand<mode>): Likewise.
20261 (sync_lock_test_and_set<mode>): Likewise.
20262 (test_and_set_12): Likewise.
20263 (atomic_compare_and_swap<mode>): Likewise.
20264 (atomic_exchange<mode>_llsc): Likewise.
20265 (atomic_fetch_add<mode>_llsc): Likewise.
20266 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
20267 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
20268 (umips_save_restore_pattern_p): Likewise.
20269 (umips_load_store_pair_p): Likewise.
20270 (umips_output_load_store_pair): Likewise.
20271 (umips_movep_target_p): Likewise.
20272 (umips_12bit_offset_address_p): Likewise.
20273 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
20274 (mips_base_mips16): Rename this...
20275 (mips_base_compression_flags): ...to this. Update all uses.
20276 (mips_attribute_table): Add micromips, nomicromips and nocompression.
20277 (mips_mips16_decl_p): Delete.
20278 (mips_nomips16_decl_p): Delete.
20279 (mips_get_compress_on_flags): New function.
20280 (mips_get_compress_off_flags): New function.
20281 (mips_get_compress_mode): New function.
20282 (mips_get_compress_on_name): New function.
20283 (mips_get_compress_off_name): New function.
20284 (mips_insert_attributes): Support multiple compression types.
20285 (mips_merge_decl_attributes): Likewise.
20286 (umips_12bit_offset_address_p): New function.
20287 (mips_start_function_definition): Emit .set micromips directive.
20288 (mips_call_may_need_jalx_p): New function.
20289 (mips_function_ok_for_sibcall): Add microMIPS support.
20290 (mips_print_operand_punctuation): Support short delay slots and
20292 (umips_swm_mask, umips_swm_encoding): New.
20293 (umips_build_save_restore): New function.
20294 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
20295 (was_mips16_p): Remove.
20296 (old_compression_mode): New.
20297 (mips_set_compression_mode): New function.
20298 (mips_set_current_function): Add microMIPS support.
20299 (mips_option_override): Likewise.
20300 (umips_save_restore_pattern_p): New function.
20301 (umips_output_save_restore): New function.
20302 (umips_load_store_pair_p_1): New function.
20303 (umips_load_store_pair_p): New function.
20304 (umips_output_load_store_pair_1): New function.
20305 (umips_output_load_store_pair): New function.
20306 (umips_movep_target_p) New function.
20307 (mips_prepare_pch_save): Add microMIPS support.
20308 * config/mips/mips.h (TARGET_COMPRESSION): New.
20309 (TARGET_CPU_CPP_BUILTINS): Update macro
20310 to use new compression flags and to support microMIPS.
20311 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
20312 (MIPS_ARCH_FLOAT_SPEC): Likewise.
20313 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
20314 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
20315 (ASM_SPEC): Support mmicromips and mno-micromips.
20316 (M16STORE_REG_P): New macro.
20317 (MIPS_CALL): Support TARGET_MICROMIPS.
20318 (MICROMIPS_J): New macro.
20319 (mips_base_mips16): Rename this...
20320 (mips_base_compression_flags): ...to this.
20321 (UMIPS_12BIT_OFFSET_P): New macro.
20322 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
20323 (MULTILIB_DIRNAMES): Likewise.
20325 2013-02-25 Tom de Vries <tom@codesourcery.com>
20327 PR rtl-optimization/56131
20328 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
20329 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
20330 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
20332 2013-02-25 Tobias Burnus <burnus@net-b.de>
20334 * doc/invoke.texi (-fsanitize=): Move from optimization
20335 to debugging options.
20337 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
20339 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
20341 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
20342 Alexander Monakov <amonakov@ispras.ru>
20344 PR middle-end/56077
20345 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
20346 flush pending lists also on non-jumps. Adjust comment.
20348 2013-02-24 Catherine Moore <clm@codesourcery.com>
20349 Maciej W. Rozycki <macro@codesourcery.com>
20350 Tom de Vries <tom@codesourcery.com>
20351 Nathan Sidwell <nathan@codesourcery.com>
20352 Iain Sandoe <iain@codesourcery.com>
20353 Nathan Froyd <froydnj@codesourcery.com>
20354 Chao-ying Fu <fu@mips.com>
20356 * doc/extend.texi: (micromips, nomicromips, nocompression):
20357 Document new function attributes.
20358 * doc/invoke.texi (minterlink-compressed, mmicromips,
20359 m14k, m14ke, m14kec): Document new options.
20360 (minterlink-mips16): Update documentation.
20361 * doc/md.texi (ZC, ZD): Document new constraints.
20362 * configure.ac (gcc_cv_as_micromips): Check if linker
20363 supports the .set micromips directive.
20364 * configure: Regenerate.
20365 * config.in: Regenerate.
20366 * config/mips/mips-tables.opt: Regenerate.
20367 * config/mips/micromips.md: New file.
20368 * constraints.md (ZC, AD): New constraints.
20369 * config/mips/predicates.md (movep_src_register): New predicate.
20370 (movep_src_operand): New predicate.
20371 (non_volatile_mem_operand): New predicate.
20372 * config/mips/mips.md (multimem): New type.
20373 (length): Differentiate between 17-bit and 18-bit branch offsets.
20374 (MOVEP1, MOVEP2): New mode iterator.
20375 (mov_<load>l): Use ZC constraint.
20376 (mov_<load>r): Likewise.
20377 (mov_<store>l): Likewise.
20378 (mov_<store>r): Likewise.
20379 (*branch_equality<mode>_inverted): Add microMIPS support.
20380 (*branch_equality<mode>): Likewise.
20381 (*jump_absolute): Likewise.
20382 (indirect_jump_<mode>): Likewise.
20383 (tablejump_<mode>): Likewise.
20384 (<optab>_internal): Likewise.
20385 (sibcall_internal): Likewise.
20386 (sibcall_value_internal): Likewise.
20387 (prefetch): Use constraint ZD.
20388 * config/mips/mips.opt (minterlink-compressed): New option.
20389 (minterlink-mips16): Now an alias for minterlink-compressed.
20390 (mmicromips): New option.
20391 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
20392 (compare_and_swap_12): Likewise.
20393 (sync_add<mode>): Likewise.
20394 (sync_<optab>_12): Likewise.
20395 (sync_old_<optab>_12): Likewise.
20396 (sync_new_<optab>_12): Likewise.
20397 (sync_nand_12): Likewise.
20398 (sync_old_nand_12): Likewise.
20399 (sync_new_nand_12): Likewise.
20400 (sync_sub<mode>): Likewise.
20401 (sync_old_add<mode>): Likewise.
20402 (sync_old_sub<mode>): Likewise.
20403 (sync_new_add<mode>): Likewise.
20404 (sync_new_sub<mode>): Likewise.
20405 (sync_<optab><mode>): Likewise.
20406 (sync_old_<optab><mode>): Likewise.
20407 (sync_new_<optab><mode>): Likewise.
20408 (sync_nand<mode>): Likewise.
20409 (sync_old_nand<mode>): Likewise.
20410 (sync_new_nand<mode>): Likewise.
20411 (sync_lock_test_and_set<mode>): Likewise.
20412 (test_and_set_12): Likewise.
20413 (atomic_compare_and_swap<mode>): Likewise.
20414 (atomic_exchange<mode>_llsc): Likewise.
20415 (atomic_fetch_add<mode>_llsc): Likewise.
20416 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
20417 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
20418 (umips_save_restore_pattern_p): Likewise.
20419 (umips_load_store_pair_p): Likewise.
20420 (umips_output_load_store_pair): Likewise.
20421 (umips_movep_target_p): Likewise.
20422 (umips_12bit_offset_address_p): Likewise.
20423 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
20424 (mips_base_mips16): Rename this...
20425 (mips_base_compression_flags): ...to this. Update all uses.
20426 (mips_attribute_table): Add micromips, nomicromips and nocompression.
20427 (mips_mips16_decl_p): Delete.
20428 (mips_nomips16_decl_p): Delete.
20429 (mips_get_compress_on_flags): New function.
20430 (mips_get_compress_off_flags): New function.
20431 (mips_get_compress_mode): New function.
20432 (mips_get_compress_on_name): New function.
20433 (mips_get_compress_off_name): New function.
20434 (mips_insert_attributes): Support multiple compression types.
20435 (mips_merge_decl_attributes): Likewise.
20436 (umips_12bit_offset_address_p): New function.
20437 (mips_start_function_definition): Emit .set micromips directive.
20438 (mips_call_may_need_jalx_p): New function.
20439 (mips_function_ok_for_sibcall): Add microMIPS support.
20440 (mips_print_operand_punctuation): Support short delay slots and
20442 (umips_swm_mask, umips_swm_encoding): New.
20443 (umips_build_save_restore): New function.
20444 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
20445 (was_mips16_p): Remove.
20446 (old_compression_mode): New.
20447 (mips_set_compression_mode): New function.
20448 (mips_set_current_function): Add microMIPS support.
20449 (mips_option_override): Likewise.
20450 (umips_save_restore_pattern_p): New function.
20451 (umips_output_save_restore): New function.
20452 (umips_load_store_pair_p_1): New function.
20453 (umips_load_store_pair_p): New function.
20454 (umips_output_load_store_pair_1): New function.
20455 (umips_output_load_store_pair): New function.
20456 (umips_movep_target_p) New function.
20457 (mips_prepare_pch_save): Add microMIPS support.
20458 * config/mips/mips.h (TARGET_COMPRESSION): New.
20459 (TARGET_CPU_CPP_BUILTINS): Update macro
20460 to use new compression flags and to support microMIPS.
20461 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
20462 (MIPS_ARCH_FLOAT_SPEC): Likewise.
20463 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
20464 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
20465 (ASM_SPEC): Support mmicromips and mno-micromips.
20466 (M16STORE_REG_P): New macro.
20467 (MIPS_CALL): Support TARGET_MICROMIPS.
20468 (MICROMIPS_J): New macro.
20469 (mips_base_mips16): Rename this...
20470 (mips_base_compression_flags): ...to this.
20471 (UMIPS_12BIT_OFFSET_P): New macro.
20472 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
20473 (MULTILIB_DIRNAMES): Likewise.
20475 2013-02-24 Jakub Jelinek <jakub@redhat.com>
20478 * target-globals.c (save_target_globals): For init_reg_sets and
20479 target_reinit remporarily set this_fn_optabs to this_target_optabs.
20481 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
20483 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
20484 * config/aarch64/t-aarch64
20485 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
20487 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
20489 PR inline-asm/56148
20490 * lra-constraints.c (process_alt_operands): Reload operand
20491 conflicting with earlier clobber only if no more other conflicting
20494 2013-02-22 Jakub Jelinek <jakub@redhat.com>
20497 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
20498 if not linking a shared library.
20500 2013-02-22 Seth LaForge <sethml@google.com>
20502 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
20504 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
20506 * config/arm/arm.md (split for extendsidi): Update condition.
20507 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
20508 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
20509 (qhs_zextenddi_cstr): Likewise.
20511 2013-02-21 Jakub Jelinek <jakub@redhat.com>
20513 PR middle-end/56420
20514 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
20515 avoid signed wrapping.
20516 (expand_mult): Handle properly multiplication by
20517 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
20518 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
20519 in the compiler if coeff is HOST_WIDE_INT_MIN.
20520 (expand_divmod): Don't make ext_op1 static, change it's type to
20521 uhwi. Avoid undefined behavior in -INTVAL (op1).
20523 PR rtl-optimization/50339
20524 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
20526 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
20527 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
20528 into splitting_ashiftrt field.
20529 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
20531 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
20534 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
20536 PR middle-end/56108
20537 * trans-mem.c (execute_tm_mark): Do not expand transactions that
20538 are sure to go irrevocable.
20540 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
20542 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
20543 scalars are valid operands.
20545 2013-02-21 Martin Jambor <mjambor@suse.cz>
20547 PR tree-optimization/56310
20548 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
20549 only matching indices and non-negative final offsets.
20550 (intersect_aggregates_with_edge): Pass src_idx to
20551 agg_replacements_to_vector. Pass src_idx insstead of index to
20552 intersect_with_agg_replacements.
20554 2013-02-21 Martin Jambor <mjambor@suse.cz>
20556 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
20557 instead of hard-wired defaults.
20559 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
20561 * doc/invoke.texi (MIPS Options): Update documentation of the
20562 floating-point multiply-accumulate instruction restrictions.
20564 2013-02-21 Kostya Serebryany <kcc@google.com>
20566 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
20567 asan_shadow_offset on x86_64 linux.
20569 2013-02-21 Richard Biener <rguenther@suse.de>
20571 PR tree-optimization/56415
20573 2013-02-11 Richard Biener <rguenther@suse.de>
20575 PR tree-optimization/56273
20576 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
20579 2013-02-21 Jakub Jelinek <jakub@redhat.com>
20582 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
20585 PR inline-asm/56405
20586 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
20587 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
20589 2013-02-20 Jan Hubicka <jh@suse.cz>
20591 PR tree-optimization/56265
20592 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
20593 when target is referenced for first time.
20595 2013-02-20 Richard Biener <rguenther@suse.de>
20597 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
20598 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
20599 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
20600 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
20601 not return anything.
20602 (rename_ssa_copies): Do not remove unused locals.
20603 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
20604 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
20605 * passes.c (execute_function_todo): Do not schedule unused locals
20606 removal if cleanup_tree_cfg did something.
20607 * tree-ssa-live.c (remove_unused_locals): Dump statistics
20608 about the number of removed locals.
20610 2013-02-20 Richard Biener <rguenther@suse.de>
20612 PR tree-optimization/56398
20613 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
20615 2013-02-20 Martin Jambor <mjambor@suse.cz>
20617 PR tree-optimization/55334
20618 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
20619 restricted pointers to arrays.
20621 2013-02-20 Richard Biener <rguenther@suse.de>
20622 Jakub Jelinek <jakub@redhat.com>
20624 PR tree-optimization/56396
20625 * tree-ssa-ccp.c (n_const_val): New static variable.
20626 (get_value): Return NULL for SSA names we don't have a lattice
20628 (ccp_initialize): Initialize n_const_val.
20629 * tree-ssa-copy.c (n_copy_of): New static variable.
20630 (init_copy_prop): Initialize n_copy_of.
20631 (get_value): Return NULL_TREE for SSA names we don't have a
20634 2013-02-20 Martin Jambor <mjambor@suse.cz>
20636 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
20638 2013-02-20 Richard Biener <rguenther@suse.de>
20640 * genpreds.c (write_lookup_constraint): Do not compare first
20641 letter of the constraint again.
20643 2013-02-20 Richard Biener <rguenther@suse.de>
20645 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
20647 (get_use_iv_cost): Terminate hashtable walk when coming across
20650 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
20652 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
20653 reassociation for avx2 targets.
20655 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20657 * config/microblaze/microblaze.c: microblaze_has_clz = 0
20658 Add version check for v8.10.a to enable microblaze_has_clz
20659 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
20660 version and TARGET_PATTERN_COMPARE check
20661 * config/microblaze/microblaze.md: New clzsi2 instruction
20663 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20665 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
20666 function before branching.
20668 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
20670 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
20672 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
20674 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
20676 PR middle-end/55889
20677 * sel-sched.c: Include ira.h.
20678 (implicit_clobber_conflict_p): New function.
20679 (moveup_expr): Use it.
20680 * Makefile.in (sel-sched.o): Depend on ira.h.
20682 2013-02-19 Richard Biener <rguenther@suse.de>
20684 PR tree-optimization/56384
20685 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
20686 (vn_hash_type): Split out from ...
20687 (vn_hash_constant_with_type): ... here.
20688 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
20689 (vn_phi_eq): Compare types from vn_phi_s structure.
20690 (vn_phi_lookup): Populate vn_phi_s type.
20691 (vn_phi_insert): Likewise.
20693 2013-02-19 Jakub Jelinek <jakub@redhat.com>
20695 PR tree-optimization/56350
20696 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
20697 if haven't found reduction or nested cycle operand, rather than
20698 asserting we must find it.
20700 PR tree-optimization/56381
20701 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
20704 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
20705 Jakub Jelinek <jakub@redhat.com>
20708 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
20709 (swap_optab_enable): Same.
20710 (init_all_optabs): Use argument instead of global.
20711 * tree.h (struct tree_optimization_option): New field target_optabs.
20712 * expr.h (init_all_optabs): Add argument to prototype.
20713 (TREE_OPTIMIZATION_OPTABS): New.
20714 (save_optabs_if_changed): Protoize.
20715 * optabs.h: Declare this_fn_optabs.
20716 * optabs.c (save_optabs_if_changed): New.
20717 Declare this_fn_optabs.
20718 (init_optabs): Add argument to init_all_optabs() call.
20719 * function.c (invoke_set_current_function_hook): Handle per
20721 * function.h (struct function): New field optabs.
20722 * config/mips/mips.c (mips_set_mips16_mode): Handle when
20723 optimization_current_node has changed.
20724 * target-globals.h (save_target_globals_default_opts): Protoize.
20725 * target-globals.c (save_target_globals_default_opts): New.
20727 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20730 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
20731 registers %fr12 and %fr12R as call used.
20734 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
20735 and HImode, require all displacements to be an integer multiple of
20737 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
20738 only allow QImode and HImode when reload is in progress and strict is
20739 true. Likewise for symbolic addresses. Use base14_operand to check
20740 displacements in REG+BASE addresses.
20742 2013-02-18 Richard Biener <rguenther@suse.de>
20744 PR tree-optimization/56366
20745 * tree-vect-loop.c (get_initial_def_for_induction): Properly
20746 handle sign-conversion of outer-loop initial induction value.
20748 2013-02-18 Richard Biener <rguenther@suse.de>
20750 PR middle-end/56349
20751 * cfghooks.c (merge_blocks): If we merge a latch into another
20752 block adjust references to it.
20753 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
20754 (verify_loop_structure): Verify that a recorded latch is in fact
20757 2013-02-18 Richard Biener <rguenther@suse.de>
20759 PR tree-optimization/56321
20760 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
20761 order SSA name release and virtual operand unlinking.
20763 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20765 * config/microblaze/microblaze.md (save_stack_block): Define.
20766 (restore_stack_block): Likewise.
20768 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20770 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
20771 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
20772 * config/microblaze/microblaze.c (microblaze_option_override):
20773 Bail out early for PIC modes when target does not support PIC.
20775 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20777 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
20778 Replace with a microblaze version.
20779 (microblaze_trampoline_init): Adapt for microblaze.
20780 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
20783 2013-02-16 Jakub Jelinek <jakub@redhat.com>
20784 Dodji Seketeli <dodji@redhat.com>
20787 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
20788 (instrument_mem_region_access): Do not forget to always put
20789 instrumentation of the of 'base' and 'base + len' in a "if (len !=
20790 0) statement, even for cases where either 'base' or 'base + len'
20791 are not instrumented -- because they have been previously
20792 instrumented. Simplify the logic by putting all the statements
20793 instrument 'base + len' inside a sequence, and then insert that
20794 sequence right before the current insertion point. Then, to
20795 instrument 'base + len', just get an iterator on that statement.
20796 And do not forget to update the pointer to iterator the function
20797 received as argument.
20799 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
20801 PR rtl-optimization/56348
20802 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
20804 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
20806 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
20807 (clean_graph_dump_file): Pass base to start_graph_dump.
20809 2013-02-14 Richard Henderson <rth@redhat.com>
20812 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
20814 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
20816 * collect2-aix.h: Define F_LOADONLY.
20818 2013-02-14 Richard Biener <rguenther@suse.de>
20821 * varasm.c (output_constant_def_1): Get the decl representing
20822 the constant as argument.
20823 (output_constant_def): Wrap output_constant_def_1.
20824 (make_decl_rtl): Use output_constant_def_1 with the decl
20825 representing the constant.
20826 (build_constant_desc): Optionally re-use a decl already
20827 representing the constant.
20828 (tree_output_constant_def): Adjust.
20830 2013-02-14 Dodji Seketeli <dodji@redhat.com>
20833 * asan.c (instrument_builtin_call): Really put the length of the
20834 second source argument into src1_len.
20836 2013-02-13 Jakub Jelinek <jakub@redhat.com>
20838 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
20839 argument. If it is false, don't create edge from then_bb to
20841 (insert_if_then_before_iter): Pass true to it.
20842 (build_check_stmt): Pass false to it.
20843 (transform_statements): Flush hash table only on extended basic
20844 block boundaries, rather than at the beginning of every bb.
20845 Don't flush hash table on nonfreeing_call_p calls.
20846 * tree-flow.h (nonfreeing_call_p): New prototype.
20847 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
20849 2013-02-13 David S. Miller <davem@davemloft.net>
20851 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
20853 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
20856 * ira.c (max_regno_before_ira): Move from ...
20858 (fix_reg_equiv_init): Use max_regno_before_ira instead of
20861 2013-02-13 Jakub Jelinek <jakub@redhat.com>
20863 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
20865 2013-02-13 Richard Biener <rguenther@suse.de>
20868 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
20869 globals in MEM_REFs.
20871 2013-02-13 Richard Biener <rguenther@suse.de>
20873 * loop-init.c (loop_optimizer_init): Clear loop state when
20874 re-initializing preserved loops.
20875 * loop-unswitch.c (unswitch_single_loop): Return whether
20876 we unswitched the loop. Do not verify loop state here.
20877 (unswitch_loops): When we unswitched a loop discover new loops.
20879 2013-02-13 Kostya Serebryany <kcc@google.com>
20881 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
20883 * sanitizer.def: Rename __asan_init to __asan_init_v1.
20885 2013-02-12 Dodji Seketeli <dodji@redhat.com>
20887 Avoid instrumenting duplicated memory access in the same basic block
20888 * Makefile.in (asan.o): Add new dependency on hash-table.h
20889 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
20890 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
20891 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
20892 (free_mem_ref_resources, has_mem_ref_been_instrumented)
20893 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
20894 (get_mem_ref_of_assignment): New functions.
20895 (get_mem_refs_of_builtin_call): Extract from
20896 instrument_builtin_call and tweak a little bit to make it fit with
20898 (instrument_builtin_call): Use the new
20899 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
20900 of is_gimple_builtin_call.
20901 (instrument_derefs, instrument_mem_region_access): Insert the
20902 instrumented memory reference into the hash table.
20903 (maybe_instrument_assignment): Renamed instrument_assignment into
20904 this, and change it to advance the iterator when instrumentation
20905 actually happened and return true in that case. This makes it
20906 homogeneous with maybe_instrument_assignment, and thus give a
20907 chance to callers to be more 'regular'.
20908 (transform_statements): Clear the memory reference hash table
20909 whenever we enter a new BB, when we cross a function call, or when
20910 we are done transforming statements. Use
20911 maybe_instrument_assignment instead of instrumentation. No more
20912 need to special case maybe_instrument_assignment and advance the
20913 iterator after calling it; it's now handled just like
20914 maybe_instrument_call. Update comment.
20916 2013-02-13 Richard Biener <rguenther@suse.de>
20918 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
20919 Fix loop discovery code.
20921 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
20923 PR inline-asm/56148
20924 * lra-constraints.c (process_alt_operands): Match early clobber
20925 operand with itself. Check conflicts with earlyclobber only if
20926 the operand is not reloaded. Prefer to reload conflicting operand
20927 if earlyclobber and matching operands are the same.
20929 2013-02-12 Richard Biener <rguenther@suse.de>
20932 * lto-streamer-out.c (write_symbol): Do not output symbols
20933 for hard register variables.
20935 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
20938 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
20939 (umulsidi3_insn, mulsidi3_insn): New insns.
20941 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
20943 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
20944 (struct tune_params): Add vec_costs field.
20945 * config/arm/arm.c (arm_builtin_vectorization_cost)
20946 (arm_add_stmt_cost): New functions.
20947 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
20948 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
20949 (arm_default_vec_cost): New struct of type cpu_vec_costs.
20950 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
20951 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
20952 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
20953 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
20955 2013-02-12 Richard Biener <rguenther@suse.de>
20958 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
20959 decls again if possible.
20961 2013-02-12 Richard Biener <rguenther@suse.de>
20963 PR middle-end/56288
20964 * tree-ssa.c (verify_ssa_name): Fix check, move
20965 SSA_NAME_IN_FREE_LIST check up.
20967 2013-02-12 Jakub Jelinek <jakub@redhat.com>
20968 Steven Bosscher <steven@gcc.gnu.org>
20970 PR rtl-optimization/56151
20971 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
20972 equal to op0 or op1, and last_insn pattern is CODE operation
20973 with MEM dest and one of the operands matches that MEM.
20975 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
20977 * doc/extend.texi: Document Function Multiversioning and "default"
20978 parameter string to target attribute.
20979 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
20980 target attribute parameter is "default".
20981 (ix86_compare_version_priority): Remove checks for target attribute.
20982 (ix86_mangle_function_version_assembler_name): Change error to sorry.
20983 Remove check for target attribute equal to NULL. Add assert.
20984 (ix86_generate_version_dispatcher_body): Change error to sorry.
20986 2013-02-11 Iain Sandoe <iain@codesourcery.com>
20987 Jack Howarth <howarth@bromo.med.uc.edu>
20988 Patrick Marlier <patrick.marlier@gmail.com>
20991 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
20992 define ENDFILE_SPEC as TM_DESTRUCTOR.
20993 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
20995 2013-02-11 Alexander Potapenko <glider@google.com>
20996 Jack Howarth <howarth@bromo.med.uc.edu>
20997 Jakub Jelinek <jakub@redhat.com>
21000 * config/darwin.c (cdtor_record): Rename ctor_record.
21001 (sort_cdtor_records): Rename sort_ctor_records.
21002 (finalize_dtors): New routine to sort destructors by
21003 priority before use in assemble_integer.
21004 (machopic_asm_out_destructor): Use finalize_dtors if needed.
21006 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
21008 PR rtl-optimization/56275
21009 * simplify-rtx.c (avoid_constant_pool_reference): Check that
21010 offset is non-negative and less than cmode size before
21011 calling simplify_subreg.
21013 2013-02-11 Richard Biener <rguenther@suse.de>
21015 PR tree-optimization/56264
21016 * cfgloop.h (fix_loop_structure): Adjust prototype.
21017 * loop-init.c (fix_loop_structure): Return the number of
21018 newly discovered loops.
21019 * tree-cfgcleanup.c (repair_loop_structures): When new loops
21020 are discovered, do a full loop-closed SSA rewrite.
21022 2013-02-11 Richard Biener <rguenther@suse.de>
21024 PR tree-optimization/56273
21025 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
21027 (check_array_ref): Fix missing newline in dumps.
21028 (search_for_addr_array): Likewise.
21030 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
21032 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
21034 2013-02-09 Jakub Jelinek <jakub@redhat.com>
21037 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
21039 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
21041 PR rtl-optimization/56246
21042 * lra-constraints.c (simplify_operand_subreg): Try to reuse
21044 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
21045 constraints are satisfied.
21047 2013-02-08 Jeff Law <law@redhat.com>
21050 * emit-rtl.c (reg_is_parm_p): New function.
21051 * regs.h (reg_is_parm_p): New prototype.
21052 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
21053 callee-clobbered registers.
21055 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
21058 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
21059 If there is no implicit builtin declaration, just return NULL.
21061 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
21063 * config/i386/sse.md (FMAMODEM): New mode iterator.
21064 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
21065 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
21067 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
21069 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
21070 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
21071 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
21073 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21075 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
21076 (microblaze*-*-elf): Likewise.
21077 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
21079 * config/microblaze/microblaze-c.c: Add builtin defines for
21080 _LITTLE_ENDIAN and _BIG_ENDIAN.
21081 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
21082 add to TARGET_DEFAULT flags.
21083 Expand ASM_SPEC and LINK_SPEC.
21084 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
21085 * config/microblaze/microblaze.md: Update extendsidi2 and
21086 movdi_internal instructions to use low-order / high-order reg
21088 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
21089 options and inversemask / mask of LITTLE_ENDIAN.
21090 * config/microblaze/t-microblaze: Expand multilib options to
21091 include mlittle-endian (le) and update exceptions patterns.
21093 2013-02-08 Jakub Jelinek <jakub@redhat.com>
21095 PR rtl-optimization/56195
21096 * lra-constraints.c (get_reload_reg): Don't reuse regs
21097 if they have smaller mode than requested, if they have
21098 wider mode than requested, try to return a SUBREG.
21100 PR tree-optimization/56250
21101 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
21102 if type is unsigned and code isn't MULT_EXPR.
21104 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
21106 PR tree-optimization/56064
21107 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
21108 bits according to mode.
21109 * fixed-value.h (fixed_from_double_int)
21110 (const_fixed_from_double_int): Adjust comments.
21112 2013-02-08 Richard Biener <rguenther@suse.de>
21115 * lto-streamer.h (struct data_in): Remove current_file, current_line
21116 and current_col members.
21117 * lto-streamer-out.c (lto_output_location): Stream changed bits
21118 en-block for efficiency.
21119 * lto-streamer-in.c (clear_line_info): Remove.
21120 (lto_input_location): Cache current file, line and column
21121 globally via local statics. Read changed bits en-block.
21122 (input_function): Do not call clear_line_info.
21123 (lto_read_body): Likewise.
21124 (lto_input_toplevel_asms): Likewise.
21126 2013-02-08 Michael Matz <matz@suse.de>
21128 PR tree-optimization/52448
21129 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
21130 (nt_call_phase): New static.
21131 (add_or_mark_expr): Only mark accesses with newer phase than any
21133 (nonfreeing_call_p): New.
21134 (nt_init_block): Update nt_call_phase, mark blocks as visited.
21135 (nt_fini_block): Keep blocks marked as visited.
21136 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
21138 2013-02-08 Richard Biener <rguenther@suse.de>
21140 * ira.c (ira): Free broken dominator information.
21142 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
21144 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
21146 2013-02-08 Marek Polacek <polacek@redhat.com>
21148 * cfgloop.c (verify_loop_structure): Add more checking of headers.
21150 2013-02-08 Richard Biener <rguenther@suse.de>
21152 PR middle-end/56181
21153 * cfgloop.h (flow_loops_find): Adjust.
21154 (bb_loop_header_p): Declare.
21155 * cfgloop.c (bb_loop_header_p): New function split out from ...
21156 (flow_loops_find): ... here. Adjust function signature,
21157 support incremental loop structure update.
21158 (verify_loop_structure): Cleanup. Verify a loop is a loop.
21159 * cfgloopmanip.c (fix_loop_structure): Move ...
21160 * loop-init.c (fix_loop_structure): ... here.
21161 (apply_loop_flags): Split out from ...
21162 (loop_optimizer_init): ... here.
21163 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
21164 in incremental mode, only remove dead loops here.
21166 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
21169 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
21170 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
21171 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
21172 (*round<mode>3.libgcc): New insns for fixed-modes.
21173 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
21174 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
21175 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
21176 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
21177 implementations. Define to __builtin_avr_absFX,
21178 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
21179 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
21180 __builtin_avr_countlsFX, respectively.
21181 * config/avr/avr-c.c (target.h): Include it.
21182 (enum avr_builtin_id): New enum.
21183 (avr_resolve_overloaded_builtin): New static function.
21184 (avr_register_target_pragmas): Use it to set
21185 targetm.resolve_overloaded_builtin.
21186 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
21187 tree nodes used by DEF_BUILTIN.
21188 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
21189 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
21190 <AVR_BUILTIN_xxBITS>: Same.
21192 2013-02-08 Richard Biener <rguenther@suse.de>
21194 * cfgloop.c (verify_loop_structure): Properly handle
21195 a loop exiting to another loop header.
21196 * ira-int.h (ira_loops): Remove.
21197 * ira.c (ira_loops): Remove.
21198 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
21199 (do_reload): Use loop_optimizer_finalize.
21200 * ira-build.c (create_loop_tree_nodes): Use get_loops and
21201 number_of_loops to access the loop tree.
21202 (more_one_region_p): Likewise.
21203 (finish_loop_tree_nodes): Likewise.
21204 (rebuild_regno_allocno_maps): Likewise.
21205 (mark_loops_for_removal): Likewise.
21206 (mark_all_loops_for_removal): Likewise.
21207 (remove_unnecessary_regions): Likewise.
21208 (ira_build): Likewise.
21209 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
21211 2013-02-08 Richard Biener <rguenther@suse.de>
21213 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
21214 * ipa-pure-const.c (analyze_function): Avoid calling
21215 mark_irreducible_loops twice.
21216 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
21218 2013-02-07 David S. Miller <davem@davemloft.net>
21220 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
21222 * var-tracking.c (vt_add_function_parameter): Test the presence of
21223 HAVE_window_save properly and do not remap argument registers when
21224 we have a leaf function.
21226 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
21229 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
21231 * config/i386/i386.c (ix86_print_operand): Ditto.
21233 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
21235 * lra-constraints.c (process_alt_operands): Fix recently added comment.
21237 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
21239 PR rtl-optimization/56225
21240 * lra-constraints.c (process_alt_operands): Check that reload hard
21241 reg can hold value for strict_low_part.
21243 2013-02-07 Jakub Jelinek <jakub@redhat.com>
21246 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
21247 dwarf2out_end_function.
21248 (in_first_function_p, maybe_at_text_label_p,
21249 first_loclabel_num_not_at_text_label): New variables.
21250 (dwarf2out_var_location): In the first function find out
21251 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
21252 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
21255 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
21257 PR rtl-optimization/56178
21258 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
21259 SUBREG of a register. Tidy up related block of code.
21260 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
21261 note if the source is a register or a SUBREG of a register.
21263 2013-02-07 Jakub Jelinek <jakub@redhat.com>
21266 * config/rs6000/rs6000.md (ptrm): New mode attr.
21267 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
21268 call_value_indirect_aix<pttrsize>,
21269 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
21272 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
21274 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
21275 if -bnortl. Convert to strcmp and strncmp.
21277 2013-02-07 Alan Modra <amodra@gmail.com>
21280 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
21281 addresses won't wrap when offsetting.
21282 (rs6000_secondary_reload): Provide secondary reloads needed for
21283 wrapping LO_SUM addresses.
21285 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
21287 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
21288 MACH, just __MACH__.
21290 2013-02-06 Richard Biener <rguenther@suse.de>
21292 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
21293 instead of calling fix_loop_structure.
21295 2013-02-06 Jakub Jelinek <jakub@redhat.com>
21297 PR middle-end/56217
21298 * omp-low.c (use_pointer_for_field): Return false if
21299 lower_send_shared_vars doesn't generate any copy-out code.
21301 2013-02-06 Tom de Vries <tom@codesourcery.com>
21303 PR rtl-optimization/56131
21304 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
21305 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
21306 of the label is NULL. Add comment.
21308 2013-02-05 Jakub Jelinek <jakub@redhat.com>
21310 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
21313 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
21314 (STATIC_LIBTSAN_LIBS): Likewise.
21315 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
21316 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
21317 is defined, don't add anything else beyond that.
21318 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
21319 (LINK_COMMAND_SPEC): Use them.
21321 PR tree-optimization/56205
21322 * tree-stdarg.c (check_all_va_list_escapes): Return true if
21323 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
21324 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
21326 2013-02-05 Richard Biener <rguenther@suse.de>
21328 PR tree-optimization/53342
21329 PR tree-optimization/53185
21330 * tree-vectorizer.h (vect_check_strided_load): Remove.
21331 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
21332 not disallow peeling for vectorized strided loads.
21333 (vect_check_strided_load): Make static and simplify.
21334 (vect_analyze_data_refs): Adjust.
21335 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
21336 correctly when vectorizing strided loads.
21338 2013-02-05 Richard Biener <rguenther@suse.de>
21340 * doc/install.texi: Refer to ISL, not PPL.
21342 2013-02-05 Jan Hubicka <jh@suse.cz>
21344 PR tree-optimization/55789
21345 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
21347 2013-02-05 Jan Hubicka <jh@suse.cz>
21349 PR tree-optimization/55789
21350 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
21351 the dead call anyway.
21353 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
21356 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
21358 2013-02-04 Alexander Potapenko <glider@google.com>
21359 Jack Howarth <howarth@bromo.med.uc.edu>
21360 Jakub Jelinek <jakub@redhat.com>
21363 * config/darwin.c (sort_ctor_records): Stabilized qsort
21364 on constructor priority by using original position.
21365 (finalize_ctors): New routine to sort constructors by
21366 priority before use in assemble_integer.
21367 (machopic_asm_out_constructor): Use finalize_ctors if needed.
21369 2013-02-04 Jakub Jelinek <jakub@redhat.com>
21372 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
21373 about visibility on artificial decls.
21374 * config/sol2.c (solaris_assemble_visibility): Likewise.
21376 2013-02-04 Kai Tietz <ktietz@redhat.com>
21379 * config/i386/i386.c (function_value_ms_64): Add additional valtype
21380 argument and improve checking of return-argument types for 16-byte
21382 (ix86_function_value_1): Add additional valtype argument on call
21383 of function_value_64.
21384 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
21385 handling infunction_value_64 function.
21387 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
21389 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
21391 2013-02-04 Richard Biener <rguenther@suse.de>
21393 PR tree-optimization/56188
21394 * tree-ssa-structalias.c (label_visit): Consider case with
21395 initially non-empty points-to set.
21396 (perform_var_substitution): Dump node mapping and clean up.
21398 2013-02-04 Richard Guenther <rguenther@suse.de>
21401 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
21402 node prevail as last resort.
21403 (lto_symtab_merge_decls): Remove guard on LTRANS here.
21404 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
21406 2013-02-04 Richard Biener <rguenther@suse.de>
21408 PR tree-optimization/56113
21409 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
21411 (equiv_class_lookup_or_add): ... this.
21412 (label_visit): Adjust and fix error in previous patch.
21413 (perform_var_substitution): Adjust.
21415 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
21417 * config/sh/divtab.c: Fix formatting and comments throughout the file.
21418 * config/sh/sh4-300.md: Likewise.
21419 * config/sh/sh4a.md: Likewise.
21420 * config/sh/constraints.md: Likewise.
21421 * config/sh/sh.md: Likewise.
21422 * config/sh/netbsd-elf.h: Likewise.
21423 * config/sh/predicates.md: Likewise.
21424 * config/sh/sh-protos.h: Likewise.
21425 * config/sh/ushmedia.h: Likewise.
21426 * config/sh/linux.h: Likewise.
21427 * config/sh/sh.c: Likewise.
21428 * config/sh/superh.h: Likewise.
21429 * config/sh/elf.h: Likewise.
21430 * config/sh/sh4.md: Likewise.
21431 * config/sh/sh.h: Likewise.
21433 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21435 * config/pa/constraints.md: Adjust unused letters. Change "T"
21436 constraint to match_test floating_point_store_memory_operand().
21437 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
21438 (base14_operand): New.
21439 (floating_point_store_memory_operand): New.
21440 (integer_store_memory_operand): Revise to use base14_operand and
21441 reg_plus_base_memory_operand.
21442 (move_dest_operand): Allow symbolic_memory_operands.
21443 (symbolic_memory_operand): Check for LO_SOM.
21444 (symbolic_operand): Change default case to break.
21445 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
21446 CONST_DOUBLE values to be reloaded by putting them into memory when
21447 the destination is a floating point register.
21448 (movdf): Remove code to handle CONST_DOUBLE.
21450 (reload_indf_r1): New.
21451 (reload_insf_r1): New.
21452 Consistently use "Q" and "T" constraints with integer and floating
21453 point move instructions, respectively.
21454 (movdi): Remove FAIL.
21455 Change predicate for source operand unamed DImode move from
21456 general_operand to move_src_operand.
21457 (umulsidi3): Change predicate for destination operand to
21459 Likewise for similar unamed patterns.
21460 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
21461 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
21462 (hppa_legitimize_address): Simplify mask calculation.
21463 (pa_emit_move_sequence): Revised handling of secondary reloads from
21464 REG+D addresses for floating point loads and stores. Directly handle
21465 loading CONST0_RTX (mode) to a floating point register.
21466 (pa_secondary_reload): Handle reloading DF and SFmode constant values
21467 to floating point registers. Don't restrict secondary reloads to
21468 floating point registers to integer modes. Revise some comments and
21470 (TARGET_LEGITIMATE_ADDRESS_P): Define.
21471 (pa_legitimate_address_p): New.
21472 (pa_legitimize_reload_address): New.
21473 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
21474 (STRICT_REG_OK_FOR_BASE_P): New.
21475 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
21476 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
21478 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
21479 Andrew Dixie <andrewd@gentrack.com>
21481 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
21484 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
21486 * expmed.c (extract_bit_field_1): Pass the full width of the
21487 structure to get_best_reg_extraction_insn.
21489 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
21492 * configure.ac (use_cxa_atexit): Add AIX.
21493 * configure: Regenerate.
21495 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
21497 2013-02-01 Jakub Jelinek <jakub@redhat.com>
21500 * final.c (need_profile_function): New variable.
21501 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
21502 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
21503 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
21504 notes, targetm.asm_out.function_prologue doesn't emit anything,
21505 HAVE_prologue and profiler should be emitted before prologue,
21506 set need_profile_function instead of emitting it.
21507 (final_scan_insn): If need_profile_function, emit
21508 profile_function on the first NOTE_INSN_BASIC_BLOCK or
21509 NOTE_INSN_FUNCTION_BEG note.
21511 2013-02-01 Richard Henderson <rth@redhat.com>
21513 * config/rs6000/rs6000.md (smulditi3): New.
21516 * config/alpha/alpha.md (umulditi3): New.
21518 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
21520 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
21521 (ASM_OUTPUT_ALIGNED_LOCAL): New.
21523 2013-02-01 Richard Biener <rguenther@suse.de>
21525 PR tree-optimization/56113
21526 * tree-ssa-structalias.c (label_visit): Reduce work for
21527 single-predecessor nodes.
21529 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
21531 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
21532 range isn't testing for zero.
21534 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
21536 PR middle-end/56113
21537 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
21539 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
21540 Nick Clifton <nickc@redhat.com>
21542 * config/v850/constraints.md (Q): Define as a memory constraint.
21543 * config/v850/predicates.md (label_ref_operand): New predicate.
21544 (e3v5_shift_operand): New predicate.
21545 (ior_operator): New predicate.
21546 * config/v850/t-v850: Add e3v5 multilib.
21547 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
21548 (v850_gen_movdi): Prototype.
21549 * config/v850/v850.c: Add support for e3v5 architecture.
21550 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
21552 (construct_save_jarl): Add e3v5 long JARL support.
21553 (v850_adjust_insn_length): New function. Adjust length of call
21554 insns when using e3v5 instructions.
21555 (v850_gen_movdi): New function: Generate instructions to move a
21557 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
21558 (CPP_SPEC): Define __v850e3v5__ as appropriate.
21559 (TARGET_USE_FPU): Enable for e3v5.
21560 (CONST_OK_FOR_W): New macro.
21561 (ADJUST_INSN_LENGTH): Define.
21562 * config/v850/v850.md (UNSPEC_LOOP): Define.
21563 (attr cpu): Add v850e3v5.
21564 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
21565 (movdi): New pattern.
21566 (movdi_internal): New pattern.
21567 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
21568 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
21569 (cstoresf4): Likewise.
21570 (cstoredf4): Likewise.
21571 (insv): New pattern.
21572 (rotlso3_a): New pattern.
21573 (rotlsi3_b): New pattern
21574 (rotlsi3_v850e3v5): New pattern.
21575 (doloop_begin): New pattern.
21576 (fix_loop_counter): New pattern.
21577 (doloop_end): New pattern.
21578 (branch_normal): Add e3v5 long branch support.
21579 (branch_invert): Likewise.
21580 (branch_z_normal): Likewise.
21581 (branch_z_invert): Likewise.
21582 (branch_nz_normal): Likewise.
21583 (branch_nz_invert): Likewise.
21584 (call_internal_short): Add e3v5 register-indirect JARL support.
21585 (call_internal_long): Likewise.
21586 (call_value_internal_short): Likewise.
21587 (call_value_internal_long): Likewise.
21588 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
21589 (mloop): New option.
21590 * config.gcc: Add support for configuring v840e3v5 target.
21591 * doc/invoke.texi: Document new v850 specific command line options.
21593 2013-01-31 Paul Koning <ni1d@arrl.net>
21597 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
21598 children if parent is a class.
21599 (prune_unused_types_prune): Don't add DW_AT_declaration.
21601 2013-01-31 Richard Biener <rguenther@suse.de>
21603 PR tree-optimization/56157
21604 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
21605 match up operand with SLP child.
21607 2013-01-31 Jason Merrill <jason@redhat.com>
21610 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
21611 parameters the first time.
21612 (gen_scheduled_generic_parms_dies): Check completeness here.
21614 2013-01-31 Richard Biener <rguenther@suse.de>
21616 PR middle-end/53073
21617 * common.opt (faggressive-loop-optimizations): New flag,
21618 enabled by default.
21619 * doc/invoke.texi (faggressive-loop-optimizations): Document.
21620 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
21621 infer_loop_bounds_from_undefined by it.
21623 2013-01-31 Richard Biener <rguenther@suse.de>
21625 PR tree-optimization/56150
21626 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
21627 visit virtual operands.
21628 (find_uses_to_rename_bb): Likewise.
21630 2013-01-31 Richard Biener <rguenther@suse.de>
21632 PR tree-optimization/56150
21633 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
21634 mixed store non-store stmts.
21636 2013-01-30 Jakub Jelinek <jakub@redhat.com>
21639 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
21640 LIBASAN_EARLY_SPEC is defined.
21641 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
21642 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
21644 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
21647 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
21648 invalid args instead of ICEing on it.
21649 (ix86_valid_target_attribute_tree): Return error_mark_node if
21650 ix86_valid_target_attribute_inner_p failed.
21651 (ix86_valid_target_attribute_p): Return false only if
21652 ix86_valid_target_attribute_tree returned error_mark_node. Allow
21653 target("default") attribute.
21654 (sorted_attr_string): Change argument from const char * to tree,
21655 merge in all target attribute arguments rather than just one.
21656 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
21657 instead of free. Avoid using strcat.
21658 (ix86_mangle_function_version_assembler_name): Mangle
21659 target("default") as if no target attribute is present. Adjust
21660 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
21661 instead of xmalloc and XDELETEVEC instead of free.
21662 (ix86_function_versions): Don't return true if one of the decls
21663 doesn't have target attribute. If they don't and one of the decls
21664 is DECL_FUNCTION_VERSIONED, report an error. Adjust
21665 sorted_attr_string caller. Use XDELETEVEC instead of free.
21666 (ix86_supports_function_versions): Remove.
21667 (make_name): Fix up formatting.
21668 (make_dispatcher_decl): Remove resolver_name and its initialization.
21669 Avoid leaking memory.
21670 (is_function_default_version): Return true if there is
21671 target("default") attribute rather than no target attribute at all.
21672 (make_resolver_func): Avoid leaking memory.
21673 (ix86_generate_version_dispatcher_body): Likewise.
21674 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
21675 * target.def (supports_function_versions): Remove.
21676 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
21677 * doc/tm.texi: Regenerated.
21679 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
21681 PR rtl-optimization/56144
21682 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
21683 for values with side effects.
21685 2013-01-30 Richard Biener <rguenther@suse.de>
21687 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
21688 (sparseset_pop): Likewise.
21689 * cfganal.c (compute_idf): Likewise. Increase work-stack size
21690 to be able to use quick_push in the worker loop.
21692 2013-01-30 Marek Polacek <polacek@redhat.com>
21694 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
21696 2013-01-30 Richard Biener <rguenther@suse.de>
21699 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
21701 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
21703 PR tree-optimization/56064
21704 * fixed-value.c (fixed_from_double_int): New function.
21705 * fixed-value.h (fixed_from_double_int): New prototype.
21706 (const_fixed_from_double_int): New static inline function.
21707 * fold-const.c (native_interpret_fixed): New static function.
21708 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
21709 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
21710 (native_encode_fixed): New static function.
21711 (native_encode_expr) <FIXED_CST>: Use it.
21712 (native_interpret_int): Move double_int worker code to...
21713 * double-int.c (double_int::from_buffer): ...this new static method.
21714 * double-int.h (double_int::from_buffer): Prototype it.
21716 2013-01-30 Richard Biener <rguenther@suse.de>
21718 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
21719 New pointer-map and obstack.
21720 (init_alias_vars): Allocate pointer-map and obstack.
21721 (delete_points_to_sets): Free them.
21722 (find_what_var_points_to): Cache result.
21723 (find_what_p_points_to): Adjust for changed interface of
21724 find_what_var_points_to.
21725 (compute_points_to_sets): Likewise.
21726 (ipa_pta_execute): Likewise.
21728 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
21730 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
21731 * configure: Regenerate.
21732 * config.in: Regenerate.
21733 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
21734 #nobits/#progbits if supported.
21736 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
21739 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
21740 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
21741 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
21743 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
21745 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
21746 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
21748 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
21750 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
21751 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
21753 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
21755 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
21757 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
21758 * config/arm/cortex-a7.md: New bypasses using
21759 arm_mac_accumulator_is_result.
21761 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
21763 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
21764 (cortex_a7_neon_mla): Likewise.
21765 (cortex_a7_fpfmad): New reservation.
21766 (cortex_a7_fpmacs): Use ffmas and update required units.
21767 (cortex_a7_fpmuld): Update required units and latency.
21768 (cortex_a7_fpmacd): Likewise.
21769 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
21770 (cortex_a7_neon). Likewise.
21771 (bypass) Update participating units.
21773 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
21775 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
21776 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
21778 * config/arm/vfp11.md (vfp_farith): Use ffmas.
21779 (vfp_fmul): Use ffmad.
21780 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
21781 (cortex_r4_fmacd): Use ffmad.
21782 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
21783 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
21784 (cortex_a9_fmacd): Use ffmad.
21785 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
21786 (cortex_a8_vfp_macd): Use ffmad.
21787 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
21788 (cortex_a5_fpmacd): Use ffmad.
21789 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
21790 (cortex_a15_vfp_macd): Use ffmad.
21791 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
21793 2013-01-29 Jason Merrill <jason@redhat.com>
21796 * varasm.c (default_assemble_visibility): Don't warn about
21797 visibility on artificial decls.
21799 2013-01-29 Richard Biener <rguenther@suse.de>
21801 PR tree-optimization/56113
21802 * tree-ssa-structalias.c (equiv_class_lookup): Also return
21804 (label_visit): Free duplicate bitmaps and record the leader instead.
21805 (perform_var_substitution): Adjust.
21807 2013-01-29 Richard Biener <rguenther@suse.de>
21809 PR tree-optimization/55270
21810 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
21811 the CFG, schedule loops for fixup.
21813 2013-01-29 Nick Clifton <nickc@redhat.com>
21815 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
21818 2013-01-28 Leif Ekblad <leif@rdos.net>
21820 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
21821 * config/i386/i386.h (TARGET_RDOS): New macro.
21822 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
21823 * config/i386/i386.c (ix86_option_override_internal): For 64bit
21824 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
21825 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
21826 DEFAULT_LARGE_SECTION_THRESHOLD.
21827 * config/i386/i386.md (R14_REG, R15_REG): New constants.
21828 * config/i386/rdos.h: New file.
21829 * config/i386/rdos64.h: New file.
21831 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
21834 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
21837 2013-01-28 Jakub Jelinek <jakub@redhat.com>
21839 PR rtl-optimization/56117
21840 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
21841 call cselib_lookup_from_insn on the MEM before calling
21842 add_insn_mem_dependence.
21844 2013-01-28 Richard Biener <rguenther@suse.de>
21846 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
21847 to a stmt that didn't have one.
21848 (copy_phis_for_bb): Likewise for PHI arguments.
21849 (copy_debug_stmt): Likewise for debug stmts.
21851 2013-01-28 Richard Biener <rguenther@suse.de>
21853 PR tree-optimization/56034
21854 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
21855 (partition_builtin_p): Adjust.
21856 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
21857 it is the last partition.
21858 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
21859 up the vertex for the definition.
21860 (classify_partition): Classify whether a partition is a
21861 PKIND_REDUCTION, thus has uses outside of the loop.
21862 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
21863 Merge all PKIND_REDUCTION partitions into the last partition.
21864 (tree_loop_distribution): Seed partitions from reductions as well.
21866 2013-01-28 Jakub Jelinek <jakub@redhat.com>
21868 PR tree-optimization/56125
21869 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
21870 pow(x,c) into sqrt(x) * powi(x, n/2) or
21871 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
21872 optimizing for size.
21873 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
21874 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
21877 PR tree-optimization/56094
21878 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
21879 to UNKNOWN_LOCATION while gimplifying expr.
21881 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
21884 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
21885 operand 0 in movabs insn template for -masm=intel asm alternative.
21886 (*movabs<mode>_2): Ditto for operand 1.
21888 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
21891 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
21894 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
21896 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
21899 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21901 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
21902 Undef to avoid warning.
21904 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
21906 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
21907 * configure: Regenerate.
21909 2013-01-25 Jakub Jelinek <jakub@redhat.com>
21911 PR tree-optimization/56098
21912 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
21913 for stmts with volatile ops.
21914 (cond_store_replacement): Don't optimize if assign has volatile ops.
21915 (cond_if_else_store_replacement_1): Don't optimize if either
21916 then_assign or else_assign have volatile ops.
21917 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
21920 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
21922 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
21924 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
21926 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
21927 missing ':' in asm example.
21929 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
21931 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
21932 entries into lane and laneq entries.
21933 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
21934 Remove AdvSIMD scalar modes.
21935 (aarch64_sq<r>dmulh_laneq<mode>): New.
21936 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
21938 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
21939 builtin implementations to relfect changes in RTL in aarch64-simd.md.
21940 * config/aarch64/iterators.md (VCOND): New.
21943 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
21946 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
21947 Add NULL LIBNAME argument to existing definitions.
21948 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
21949 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
21950 * config/avr/avr.c (DEF_BUILTIN): Same.
21951 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
21952 (avr_expand_builtin): Expand to a vanilla call if a libgcc
21953 implementation is available (DECL_ASSEMBLER_NAME is set).
21954 (avr_fold_absfx): New static function.
21955 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
21956 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
21957 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
21958 AVR_BUILTIN_ABSLLK.
21959 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
21960 (abshk, absk, abslk, absllk): Provide as static inline functions.
21962 2013-01-25 Marek Polacek <polacek@redhat.com>
21964 PR tree-optimization/56035
21965 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
21967 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
21969 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
21970 (*movtf_internal_rex64): Add (!o,C) alternative
21971 (*movxf_internal_rex64): Ditto.
21972 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
21974 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
21976 * doc/invoke.texi: fix typo.
21977 * doc/objc.texi: fix typo.
21979 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
21981 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
21982 for the first two alternatives.
21984 2013-01-24 Diego Novillo <dnovillo@google.com>
21986 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
21987 (ggc-zone.o): Remove.
21988 * configure.ac: Remove option --with-gc.
21989 * configure: Re-generate.
21990 * doc/install.texi: Remove documentation for --with-gc.
21991 * gengtype.c (write_enum_defn): Remove. Update all users.
21992 (write_Types_process_field): Remove generation of gt_e_* argument.
21993 (output_type_enum): Remove. Update all users.
21994 (write_enum_defn): Remove. Update all users.
21995 (enum alloc_zone): Remove. Update all users.
21996 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
21997 * ggc-common.c (ggc_splay_alloc): Remove first argument.
21998 Update all callers.
21999 (struct ptr_data): Remove field TYPE. Update all users.
22000 (gt_pch_note_object): Remove argument TYPE. Update all users.
22001 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
22003 * ggc-none.c (ggc_alloc_typed_stat): Remove.
22004 (struct alloc_zone): Remove.
22005 (ggc_internal_alloc_zone_stat): Remove.
22006 (ggc_internal_cleared_alloc_zone_stat): Remove.
22007 * ggc-page.c (ggc_alloc_typed_stat): Remove.
22008 (ggc_pch_count_object): Remove last argument. Update all users.
22009 (ggc_pch_alloc_object): Remove last argument. Update all users.
22010 (struct alloc_zone): Remove.
22011 * ggc-zone.c: Remove.
22012 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
22013 (struct alloc_zone): Remove.
22014 (ggc_alloc_typed_stat): Remove.
22015 (ggc_alloc_typed): Remove.
22016 (ggc_splay_alloc): Remove first argument.
22017 (rtl_zone): Remove. Update all users.
22018 (tree_zone): Remove. Update all users.
22019 (tree_id_zone): Remove. Update all users.
22020 (ggc_internal_zone_alloc_stat): Remove. Update all users.
22021 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
22022 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
22023 * tree-ssanames.c: Remove references to zone allocator in comments.
22025 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
22027 * config/avr/avr.c (avr_out_fract): Make register numbers that
22028 might be outside of source operand signed.
22030 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
22032 * config/i386/constraints.md (Yf): New constraint.
22033 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
22034 of f constraint to conditionaly disable x87 register preferences.
22035 (*movdf_internal): Ditto.
22036 (*movsf_internal): Ditto.
22038 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
22040 PR inline-asm/55934
22041 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
22042 that have operands with impossible constraints.
22043 Add a FIXME for a speed-up opportunity.
22044 * lra-constraints.c (process_alt_operands): Verify that a class
22045 selected from constraints on asms is valid for the operand mode.
22046 (curr_insn_transform): Remove incorrect comment.
22048 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
22050 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
22051 TOC operand is a valid symbol ref in the constant pool.
22053 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
22055 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
22057 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
22060 * config/avr/stdfix.h: New file.
22061 * t-avr (stdfix-gcc.h): New rule to build it.
22062 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
22064 2013-01-23 Kostya Serebryany <kcc@google.com>
22066 * config/darwin.h: remove dependency on
22067 CoreFoundation (asan on Mac OS).
22069 2013-01-23 Jakub Jelinek <jakub@redhat.com>
22072 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
22073 instead of cmpdi_operand for first comparison operand.
22074 Don't assert that comparison operands aren't both constants.
22076 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
22078 * doc/install.texi (Downloading the Source): Update references to
22079 downloading separate components.
22081 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
22083 * doc/extend.texi (__int128): Improve grammar.
22085 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
22088 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
22089 alternative to (o,r).
22090 (*movdi_internal_rex64): Remove (!o,n) alternative.
22091 (DImode immediate->memory splitter): Remove.
22092 (DImode immediate->memory peephole2): Remove.
22093 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
22094 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
22095 alternative to (!o,*r).
22096 (*movtf_internal_sse): New pattern.
22097 (*movxf_internal_rex64): New pattern.
22098 (*movxf_internal): Disable for TARGET_64BIT.
22099 (*movdf_internal_rex64): Remove (!o,F) alternative.
22101 2013-01-22 Jakub Jelinek <jakub@redhat.com>
22103 PR middle-end/56074
22104 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
22105 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
22106 * tree-vect-loop-manip.c (find_loop_location): Also ignore
22107 stmt locations where LOCATION_LOCUS of the stmt location is
22108 UNKNOWN_LOCATION or BUILTINS_LOCATION.
22111 * config/i386/i386.md (UNSPEC_STOS): New.
22112 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
22113 *strsetqi_1): Add UNSPEC_STOS.
22115 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
22118 * doc/invoke.texi: Remove left over -Wsynth example.
22120 2013-01-21 Jakub Jelinek <jakub@redhat.com>
22122 PR tree-optimization/56051
22123 * fold-const.c (fold_binary_loc): Don't fold
22124 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
22125 a narrowing conversion, or widening conversion from signed
22128 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
22130 PR rtl-optimization/56023
22131 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
22132 dependent on debug instruction.
22134 2013-01-21 Martin Jambor <mjambor@suse.cz>
22136 PR middle-end/56022
22137 * function.c (allocate_struct_function): Call
22138 invoke_set_current_function_hook earlier.
22140 2013-01-21 Jakub Jelinek <jakub@redhat.com>
22142 * reload1.c (init_reload): Only initialize reload_obstack
22143 during the first call.
22145 2013-01-21 Marek Polacek <polacek@redhat.com>
22147 * cfgloop.c (verify_loop_structure): Fix up grammar.
22149 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
22151 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
22152 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
22154 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
22157 * config/arm/marvell-pj4.md: Update copyright year.
22158 Fix up use of alu to alu_reg and simple_alu_imm.
22160 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
22162 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
22164 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
22167 * lra-constraints.c (curr_insn_transform): Don't reuse original
22168 insn for secondary memory move when memory mode should be different.
22170 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22172 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
22173 atomic_storedi_1): New patterns.
22175 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
22177 btver2 pipeline descriptions.
22178 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
22180 * config/i386/i386.md (btver2_decode): New type attributes.
22181 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
22183 * config/i386/btver2.md: New file describing btver2 pipelines.
22185 2013-01-19 Andrew Pinski <apinski@cavium.com>
22187 PR tree-optimization/52631
22188 * tree-ssa-sccvn (visit_use): Before looking up the original
22189 statement, try looking up the simplified expression.
22191 2013-01-19 Anthony Green <green@moxielogic.com>
22193 * config/moxie/moxie.c (moxie_expand_prologue): Set
22194 current_function_static_stack_size.
22196 2013-01-18 Jakub Jelinek <jakub@redhat.com>
22198 PR tree-optimization/56029
22199 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
22200 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
22202 2013-01-18 Sharad Singhai <singhai@google.com>
22204 PR tree-optimization/55995
22205 * dumpfile.c (dump_loc): Print location only if available.
22206 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
22208 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
22211 * lra-constraints.c (curr_insn_transform): Reuse original insn for
22212 secondary memory move.
22213 (inherit_reload_reg): Use rclass instead of cl for
22214 check_secondary_memory_needed_p.
22216 2013-01-18 Jakub Jelinek <jakub@redhat.com>
22218 PR middle-end/56015
22219 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
22220 the case where writing real complex part of target modifies op1.
22222 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
22224 * config/aarch64/aarch64-simd.md
22225 (aarch64_vcond_internal<mode>): Handle unordered cases.
22226 * config/aarch64/iterators.md (v_cmp_result): New.
22228 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
22229 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
22231 * config/arm/marvell-pj4.md: New file.
22232 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
22233 * config/arm/arm.md (generic_sched): Add marvell_pj4.
22234 (generic_vfp): Likewise.
22235 * config/arm/arm-cores.def: Add marvell-pj4.
22236 * config/arm/arm-tune.md: Regenerate.
22237 * config/arm/arm-tables.opt: Regenerate.
22238 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
22239 * doc/invoke.texi: Document marvell-pj4.
22241 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
22243 * config/aarch64/arm_neon.h: Map scalar types to standard types.
22245 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
22250 * var-tracking.c (negative_power_of_two_p): New.
22251 (global_get_addr_cache, local_get_addr_cache): New.
22252 (get_addr_from_global_cache, get_addr_from_local_cache): New.
22253 (vt_canonicalize_addr): Rewrite using the above. Adjust the
22255 (vt_stack_offset_p): Remove.
22256 (vt_canon_true_dep): Always canonicalize loc's address.
22257 (clobber_overlapping_mems): Make sure we have a MEM.
22258 (local_get_addr_clear_given_value): New.
22259 (val_reset): Clear local cached entries.
22260 (compute_bb_dataflow): Create and release the local cache.
22261 Disable duplicate MEMs clobbering.
22262 (emit_notes_in_bb): Clobber MEMs likewise.
22263 (vt_emit_notes): Create and release the local cache.
22264 (vt_initialize, vt_finalize): Create and release the global
22265 cache, respectively.
22266 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
22268 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
22270 PR libmudflap/53359
22271 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
22272 not found in the symtab.
22274 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
22277 PR rtl-optimization/55547
22278 PR rtl-optimization/53827
22281 * alias.c (offset_overlap_p): New, factored out of...
22282 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
22283 the conservative special case for symbolic constants. Don't
22284 adjust zero sizes on alignment.
22286 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
22288 PR rtl-optimization/52573
22289 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
22290 REG_UNUSED for the same register.
22292 2013-01-17 Richard Biener <rguenther@suse.de>
22293 Marek Polacek <polacek@redhat.com>
22295 PR rtl-optimization/55833
22296 * loop-unswitch.c (unswitch_loops): Move loop verification...
22297 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
22298 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
22299 Set it to true when we're removing a loop from hierarchy tree in
22300 an irreducible region.
22301 (fix_bb_placements): Adjust caller.
22302 (fix_loop_placements): Likewise.
22304 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
22306 * config/avr/builtins.def (DEF_BUILTIN): Factor out
22307 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
22308 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
22309 Remove ID. Adjust comments.
22310 * config/avr/avr-c.c (avr_builtin_name): Remove.
22311 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
22312 * config/avr/avr.c (avr_tolower): New static function.
22313 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
22314 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
22315 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
22318 2013-01-17 Jan Hubicka <jh@suse.cz>
22320 PR tree-optimization/55273
22321 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
22323 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
22326 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
22327 store through atomic_store<mode>_1.
22328 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
22330 2013-01-17 Martin Jambor <mjambor@suse.cz>
22332 PR tree-optimizations/55264
22333 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
22334 for virtual methods.
22335 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
22336 virtual methods before inlining is over.
22337 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
22339 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
22342 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
22344 PR rtl-optimization/56005
22345 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
22346 pending reads for prefetch.
22348 2013-01-16 Ian Bolton <ian.bolton@arm.com>
22350 * config/aarch64/aarch64.md
22351 (*cstoresi_neg_uxtw): New pattern.
22352 (*cmovsi_insn_uxtw): New pattern.
22353 (*<optab>si3_uxtw): New pattern.
22354 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
22355 (*<optab>si3_insn_uxtw): New pattern.
22356 (*bswapsi2_uxtw): New pattern.
22358 2013-01-16 Richard Biener <rguenther@suse.de>
22360 * tree-inline.c (tree_function_versioning): Remove set but
22361 never used variable.
22363 2013-01-16 Richard Biener <rguenther@suse.de>
22365 PR tree-optimization/55964
22366 * tree-flow.h (rename_variables_in_loop): Remove.
22367 (rename_variables_in_bb): Likewise.
22368 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
22369 (copy_loop_before): Adjust and delete update-ssa status.
22370 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
22371 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
22372 (rename_variables_in_loop): Remove.
22373 (slpeel_update_phis_for_duplicate_loop): Likewise.
22374 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
22375 use available cfg machinery instead of duplicating it.
22376 Update PHI nodes and perform poor-mans SSA update here.
22377 (slpeel_tree_peel_loop_to_edge): Adjust.
22379 2013-01-16 Richard Biener <rguenther@suse.de>
22381 PR tree-optimization/54767
22382 PR tree-optimization/53465
22383 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
22384 (vrp_visit_phi_node): For PHI arguments coming via backedges
22385 drop all symbolical range information.
22386 (execute_vrp): Compute backedges.
22388 2013-01-16 Richard Biener <rguenther@suse.de>
22390 * doc/install.texi: Update CLooG and ISL requirements to
22393 2013-01-16 Christian Bruel <christian.bruel@st.com>
22396 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
22397 (broken_move): Handle UNSPECV_SP_SWITCH_B.
22398 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
22400 2013-01-16 DJ Delorie <dj@redhat.com>
22402 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
22403 (UNSPECV_SP_SWITCH_E): New.
22404 (sp_switch_1): Change to an unspec.
22405 (sp_switch_2): Change to an unspec. Don't use post-inc when we
22408 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
22410 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
22411 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
22412 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
22413 (expand_mem_thread_fence): Ditto.
22414 (expand_mem_signal_fence): Ditto.
22415 (expand_atomic_load): Ditto.
22416 (expand_atomic_store): Ditto.
22418 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
22420 PR rtl-optimization/55547
22421 PR rtl-optimization/53827
22424 * alias.c (memrefs_conflict_p): Set sizes to negative after
22427 2013-01-15 Jakub Jelinek <jakub@redhat.com>
22430 * function.c (thread_prologue_and_epilogue_insns): Always
22431 add crtl->drap_reg to set_up_by_prologue.set, even if
22432 stack_realign_drap is false.
22434 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
22436 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
22437 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
22438 *call): Fix indention.
22440 2013-01-15 Tom de Vries <tom@codesourcery.com>
22443 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
22446 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
22448 PR rtl-optimization/55153
22449 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
22451 2013-01-15 Martin Jambor <mjambor@suse.cz>
22453 PR tree-optimization/55920
22454 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
22455 accesses as grp_to_be_debug_replaced.
22457 2013-01-15 Jakub Jelinek <jakub@redhat.com>
22459 PR tree-optimization/55920
22460 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
22461 there is non-useless type conversion needed from debug rhs to lhs,
22462 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
22464 2013-01-15 Joseph Myers <joseph@codesourcery.com>
22465 Mikael Pettersson <mikpe@it.uu.se>
22468 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
22470 (ASM_OUTPUT_CASE_LABEL): Remove.
22471 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
22472 * final.c (shorten_branches): Update alignment of labels before
22473 jump tables if CASE_VECTOR_SHORTEN_MODE.
22475 2013-01-15 Richard Biener <rguenther@suse.de>
22478 * system.h: Do not include gmp.h for building host tools.
22480 2013-01-15 Richard Biener <rguenther@suse.de>
22482 PR middle-end/55882
22483 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
22484 account for bitpos when computing alignment.
22486 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
22488 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
22489 (ix86_target_macros_internal): Likewise.
22491 * config/i386/i386.c (m_CORE2I7): Removed.
22492 (m_CORE_HASWELL): New macro.
22493 (m_CORE_ALL): Likewise.
22494 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
22495 (initial_ix86_arch_features): Likewise.
22496 (processor_target_table): Initializations for Core avx2.
22497 (cpu_names): New names "core-avx2".
22498 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
22499 PROCESSOR_CORE_HASWELL.
22500 (ix86_issue_rate): New case.
22501 (ia32_multipass_dfa_lookahead): Likewise.
22502 (ix86_sched_init_global): Likewise.
22504 * config/i386/i386.h (TARGET_HASWELL): New macro.
22505 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
22506 (processor_type): New PROCESSOR_HASWELL.
22508 2013-01-15 Jakub Jelinek <jakub@redhat.com>
22510 PR tree-optimization/55955
22511 * tree-vect-loop.c (vectorizable_reduction): Give up early on
22512 *SHIFT_EXPR and *ROTATE_EXPR codes.
22514 PR tree-optimization/48766
22515 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
22516 -ftrapv disable -fwrapv.
22518 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
22521 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
22522 etc. to 1 and not to __flash.
22523 Use LL suffix for __INT24_MAX__ with -mint8.
22524 Use ULL suffix for __UINT24_MAX__ with -mint8.
22526 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
22528 * config/avr/avr-arch.h
22529 (struct base_arch_s): Use typedef avr_arch_t instead.
22530 (struct arch_info_s): Use typedef avr_arch_info_t instead.
22531 (struct mcu_type_s): Use typedef avr_mcu_t instead.
22532 * config/avr/avr.c: Same.
22533 * config/avr/avr-devices.c: Same.
22534 * config/avr/driver-avr.c: Same.
22535 * config/avr/gen-avr-mmcu-texi.c: Same.
22536 * config/avr/avr-mcus.def: Adjust comment.
22538 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
22540 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
22541 * config/aarch64/iterators.md (VALLDI): New.
22543 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
22544 Andi Kleen <ak@linux.intel.com>
22547 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
22548 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
22551 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
22553 * config/avr/avr-stdint.h: Remove trailing blanks.
22554 * config/avr/avr-log.h: Same.
22555 * config/avr/avr-arch.h: Same.
22556 * config/avr/avr-devices.c: Same.
22557 * config/avr/avr-dimode.md: Same.
22558 * config/avr/predicates.md: Same.
22559 * config/avr/avr-c.c: Same. And fix typo.
22561 * config/avr/avr-protos.h: Same. And:
22562 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
22563 (init_cumulative_args): Rename to avr_init_cumulative_args.
22564 (expand_prologue): Rename to avr_expand_prologue.
22565 (expand_epilogue): Rename to avr_expand_epilogue.
22566 (adjust_insn_length): Rename to avr_adjust_insn_length.
22567 (notice_update_cc): Rename to avr_notice_update_cc.
22568 (final_prescan_insn): Rename to avr_final_prescan_insn.
22569 * config/avr/avr.c: Same.
22570 * config/avr/avr.h: Same.
22571 * config/avr/avr.md: Remove trailing blanks.
22572 (prologue): Use avr_expand_prologue.
22573 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
22575 2013-01-14 Richard Biener <rguenther@suse.de>
22577 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
22578 verify_location, collect_subblocks): New functions.
22579 (verify_gimple_in_cfg): Verify that locations only reference
22580 BLOCKs in the functions BLOCK tree.
22582 2013-01-14 Richard Biener <rguenther@suse.de>
22584 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
22586 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
22588 (insert_out_of_ssa_copy_on_edge): Likewise.
22589 (rewrite_close_phi_out_of_ssa): Likewise.
22590 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
22592 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
22593 propagated constants.
22594 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
22597 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
22599 * config/avr/avr-modes.def: Add GPL copyright notice.
22601 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
22603 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
22604 MEMMODEL_MASK to determine memory model.
22605 (atomic_store<mode>): Ditto from operands[2].
22606 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
22608 2013-01-13 Jakub Jelinek <jakub@redhat.com>
22611 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
22612 (fold_gimple_assign): Don't call unshare_expr here.
22613 (fold_ctor_reference): Call unshare_expr.
22615 2013-01-13 Terry Guo <terry.guo@arm.com>
22617 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
22618 * doc/fragments.texi: Document MULTILIB_REUSE.
22619 * gcc.c (multilib_reuse): New internal spec.
22620 (set_multilib_dir): Also search multilib from multilib_reuse.
22621 * genmultilib (tmpmultilib3): Refactor code.
22622 (tmpmultilib4): Ditto.
22623 (multilib_reuse): New multilib argument.
22625 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
22627 * Makefile.in: Update copyright.
22629 2013-01-12 Tom de Vries <tom@codesourcery.com>
22631 PR middle-end/55890
22632 * calls.c (expand_call): Check if arg_nr is valid.
22634 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
22636 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
22637 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
22638 documentation. Add missing '__' in front of
22639 __builtin_ia32_packssdw256.
22641 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22644 * config/s390/s390.c (s390_preferred_reload_class): Do not return
22645 NO_REGS for larl operands.
22646 (s390_reload_larl_operand): Use s390_load_address instead of
22649 2013-01-11 Richard Biener <rguenther@suse.de>
22651 * tree-cfg.c (verify_node_sharing_1): Split out from ...
22652 (verify_node_sharing): ... here.
22653 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
22655 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
22657 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
22658 Substitute TREECHECKING.
22659 * configure: Regenerate.
22660 * Makefile.in (TREECHECKING): New.
22662 2013-01-11 Richard Guenther <rguenther@suse.de>
22664 PR tree-optimization/44061
22665 * tree-vrp.c (extract_range_basic): Compute zero as
22666 value-range for __builtin_constant_p of function parameters.
22668 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
22670 Update copyright years.
22672 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
22674 PR rtl-optimization/55672
22675 * lra-eliminations.c (mark_not_eliminable): Permit addition with
22676 const to be eliminable.
22678 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
22680 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
22681 * configure: Regenerate.
22683 2013-01-10 Richard Biener <rguenther@suse.de>
22685 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
22687 2013-01-10 Richard Biener <rguenther@suse.de>
22690 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
22691 locations for virtual PHI arguments.
22692 (rewrite_update_phi_arguments): Likewise.
22694 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
22696 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
22699 2013-01-10 Jakub Jelinek <jakub@redhat.com>
22701 PR tree-optimization/55921
22702 * tree-complex.c (expand_complex_asm): New function.
22703 (expand_complex_operations_1): Call it for GIMPLE_ASM.
22705 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22708 * config/s390/s390.c (s390_symref_operand_p)
22709 (s390_loadrelative_operand_p): Merge the two functions.
22710 (s390_check_qrst_address, print_operand_address): Add parameters
22711 to s390_loadrelative_operand_p invokation.
22712 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
22713 (s390_reload_larl_operand, s390_secondary_reload): Use
22714 s390_loadrelative_operand_p instead of s390_symref_operand_p.
22715 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
22717 2013-01-09 Mike Stump <mikestump@comcast.net>
22719 * dse.c (record_store): Remove unnecessary assert.
22721 2013-01-09 Jan Hubicka <jh@suse.cz>
22723 PR tree-optimization/55569
22724 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
22725 * cfgloop.h (scale_loop_profile): Likewise.
22727 2013-01-09 Jan Hubicka <jh@suse.cz>
22730 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
22732 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
22734 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
22736 PR middle-end/55114
22737 * expr.h (maybe_emit_group_store): Declare.
22738 * expr.c (maybe_emit_group_store): New function.
22739 * builtins.c (expand_builtin_int_roundingfn): Call it.
22740 (expand_builtin_int_roundingfn_2): Likewise.
22742 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
22744 PR rtl-optimization/55829
22745 * lra-constraints.c (match_reload): Add code for absent output.
22746 (curr_insn_transform): Add code for reloads of matched inputs
22749 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
22751 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
22752 attribute of movddup insn to DF.
22753 (*vec_interleave_lowv2df): Ditto.
22754 (vec_dupv2df): Ditto.
22756 2013-01-09 Jan Hubicka <jh@suse.cz>
22758 PR tree-optimiation/55875
22759 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
22760 EVERY_ITERATION parameter.
22761 (number_of_iterations_exit): Check if exit is executed every iteration.
22762 (idx_infer_loop_bounds): Similarly here.
22763 (n_of_executions_at_most): Simplify
22764 to only test for cases where statement is dominated by the
22765 particular bound; handle correctly the "postdominance" test.
22766 (scev_probably_wraps_p): Use max loop iterations info
22767 as a global bound first.
22769 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
22770 Nick Clifton <nickc@redhat.com>
22772 * config/v850/v850.md (cbranchsf4): New pattern.
22773 (cstoresf4): New pattern.
22774 (cbranchdf4): New pattern.
22775 (cstoredf4): New pattern.
22776 (movsicc): Disallow floating point comparisons.
22777 (cmpsf_le_insn): Fix order of operators.
22778 (cmpsf_lt_insn): Likewise.
22779 (cmpsf_eq_insn): Likewise.
22780 (cmpdf_le_insn): Likewise.
22781 (cmpdf_lt_insn): Likewise.
22782 (cmpdf_eq_insn): Likewise.
22783 (cmpsf_ge_insn): Use LE comparison.
22784 (cmpdf_ge_insn): Likewise.
22785 (cmpsf_gt_insn): Use LT comparison.
22786 (cmpdf_gt_insn): Likewise.
22787 (cmpsf_ne_insn): Delete pattern.
22788 (cmpdf_ne_insn): Delete pattern.
22789 * config/v850/v850.c (v850_gen_float_compare): Use
22790 gen_cmpdf_eq_insn for NE comparison.
22791 (v850_float_z_comparison_operator)
22792 (v850_float_nz_comparison_operator): Move from here ...
22793 * config/v850/predicates.md: ... to here. Move GT and GE
22794 comparisons into v850_float_z_comparison_operator.
22795 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
22797 (v850_float_nz_comparison_operator): Likewise.
22799 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22801 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
22802 with calls to gen_insvsi/gen_insvdi.
22804 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
22806 * config/i386/i386.c (initial_ix86_tune_features): Set up
22807 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
22809 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
22810 Jakub Jelinek <jakub@redhat.com>
22812 PR tree-optimization/48189
22813 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
22814 If nitercst is 0, don't predict the exit edge.
22816 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22818 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
22819 in asm_fprintf with reg_names.
22820 (aarch64_print_operand_address): Likewise.
22821 (aarch64_return_addr): Likewise.
22822 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
22824 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22826 * config/pa/pa.h (VAL_U6_BITS_P): Define.
22827 (INT_U6_BITS): Likewise.
22828 * config/pa/predicates.md (uint6_operand): New predicate.
22829 (shift5_operand, shift6_operand): Likewise.
22830 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
22832 (lshrdi3): Use shift6_operand.
22833 (shrpsi4, shrpdi4): New insn patterns.
22834 (extzv): Delete expander.
22835 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
22836 predicates in unamed zero extract patterns. Tighten common constraint.
22837 (extv): Delete expander.
22838 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
22839 predicates in unamed sign extract patterns. Tighten common constraint.
22840 (insv): Delete expander.
22841 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
22842 predicates in unamed insert patterns. Tighten common constraint.
22843 Change uint32_operand predicate to uint6_operand predicate in unamed
22844 DImode pattern to insert constant values of type 1...1xxxx.
22846 2013-01-04 Jan Hubicka <jh@suse.cz>
22848 PR tree-optimization/55823
22849 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
22852 2013-01-08 Jakub Jelinek <jakub@redhat.com>
22853 Uros Bizjak <ubizjak@gmail.com>
22855 PR rtl-optimization/55845
22856 * df-problems.c (can_move_insns_across): Stop scanning at
22857 volatile_insn_p source instruction or give up if
22858 across_from .. across_to range contains any volatile_insn_p
22861 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
22863 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
22864 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
22866 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
22867 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
22869 2013-01-08 Jakub Jelinek <jakub@redhat.com>
22872 * asan.c (asan_clear_shadow): New function.
22873 (asan_emit_stack_protection): Use it.
22875 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
22877 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
22878 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
22879 with tab instead of space.
22881 2013-01-08 Nick Clifton <nickc@redhat.com>
22883 * config/rl78/rl78.c (rl78_expand_prologue): Always select
22884 register bank 0 at the start of an interrupt handler.
22885 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
22888 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
22890 * config/aarch64/aarch64-simd.md
22891 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
22892 (aarch64_simd_bsl): Likewise.
22893 (aarch64_vcond_internal<mode>): Likewise.
22894 (vcond<mode><mode>): Likewise.
22895 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
22896 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
22898 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
22900 * config/aarch64/aarch64-builtins.c
22901 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
22903 2013-01-08 Martin Jambor <mjambor@suse.cz>
22906 * tree-sra.c (analyze_access_subtree): Return true also after
22907 potentially creating a debug-only replacement.
22909 2013-01-08 Jakub Jelinek <jakub@redhat.com>
22911 PR middle-end/55890
22912 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
22914 PR tree-optimization/54120
22915 * tree-vrp.c (range_fits_type_p): Don't allow
22916 src_precision < precision from signed vr to unsigned_p
22917 if vr->min or vr->max is negative.
22918 (simplify_float_conversion_using_ranges): Test can_float_p
22919 against CODE_FOR_nothing.
22921 2013-01-08 Jakub Jelinek <jakub@redhat.com>
22922 Richard Biener <rguenther@suse.de>
22924 PR middle-end/55851
22925 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
22926 types instead of just INTEGER_TYPE types.
22928 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
22930 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
22933 2013-01-07 Steve Ellcey <sellcey@mips.com>
22936 * config/mips/mips.opt: Change mad to mmad to match documentation.
22938 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
22941 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
22942 .progmemx.data now.
22944 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
22947 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
22948 (avr_addrspace_t): Add .section_name field.
22949 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
22951 (avr_addrspace): Same. Initialize .section_name. Remove last
22952 NULL entry. Put __memx into .progmemx.data.
22953 (progmem_section_prefix): Remove.
22954 (avr_asm_init_sections): No need to initialize progmem_section.
22955 (avr_asm_named_section): Use avr_addrspace[].section_name to get
22956 section name prefix.
22957 (avr_asm_select_section): Ditto. And use get_unnamed_section to
22958 retrieve the progmem section.
22959 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
22960 boundary to run over avr_addrspace[].
22961 (avr_register_target_pragmas): Ditto.
22963 2013-01-06 Jakub Jelinek <jakub@redhat.com>
22965 * varasm.c (output_constant_def_contents): For asan_protect_global
22966 protected strings, adjust DECL_ALIGN if needed, before testing for
22968 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
22969 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
22971 (output_object_block): For asan protected decls, emit asan padding
22972 after their contents.
22973 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
22974 (asan_finish_file): Test it here instead.
22976 2013-01-07 Nick Clifton <nickc@redhat.com>
22977 Matthias Klose <doko@debian.org>
22978 Doug Kwan <dougkwan@google.com>
22979 H.J. Lu <hongjiu.lu@intel.com>
22982 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
22984 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
22986 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
22988 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
22990 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
22992 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
22995 * doc/install.texi (Cross-Compiler-Specific Options): Document
22998 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
23000 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
23001 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
23002 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
23003 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
23004 vqmovun_high_s64): Fix source operand number and update copyright.
23006 2013-01-07 Richard Biener <rguenther@suse.de>
23008 PR middle-end/55890
23009 * gimple.h (gimple_call_builtin_p): New overload.
23010 * gimple.c (validate_call): New function.
23011 (gimple_call_builtin_p): Likewise.
23012 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
23013 Use gimple_call_builtin_p.
23014 (find_func_clobbers): Likewise.
23015 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
23016 (strlen_optimize_stmt): Likewise.
23018 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
23020 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
23021 (vld1q_dup_*): Likewise.
23022 (vld1_*): Likewise.
23023 (vld1q_*): Likewise.
23024 (vld1_lane_*): Likewise.
23025 (vld1q_lane_*): Likewise.
23027 2013-01-07 Richard Biener <rguenther@suse.de>
23029 * lto-streamer.h (LTO_minor_version): Bump to 2.
23031 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
23033 * config/aarch64/aarch64-protos.h
23034 (aarch64_const_double_zero_rtx_p): Rename to...
23035 (aarch64_float_const_zero_rtx_p): ...this.
23036 (aarch64_float_const_representable_p): New.
23037 (aarch64_output_simd_mov_immediate): Likewise.
23038 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
23039 move immediate case.
23040 * config/aarch64/aarch64.c
23041 (aarch64_const_double_zero_rtx_p): Rename to...
23042 (aarch64_float_const_zero_rtx_p): ...this.
23043 (aarch64_print_operand): Allow printing of new constants.
23044 (aarch64_valid_floating_const): New.
23045 (aarch64_legitimate_constant_p): Check for valid floating-point
23047 (aarch64_simd_valid_immediate): Likewise.
23048 (aarch64_vect_float_const_representable_p): New.
23049 (aarch64_float_const_representable_p): Likewise.
23050 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
23051 (aarch64_output_simd_mov_immediate): New.
23052 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
23053 (*movdf_aarch64): Likewise.
23054 * config/aarch64/constraints.md (Ufc): New.
23055 (Y): call aarch64_float_const_zero_rtx.
23056 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
23058 2013-01-07 Richard Biener <rguenther@suse.de>
23060 PR tree-optimization/55888
23061 PR tree-optimization/55862
23062 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
23063 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
23064 not if it is contained therein.
23066 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
23068 * config/avr/t-avr: Typo.
23070 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
23073 * config/avr/t-avr: Don't automatically rebuild
23074 $(srcdir)/config/avr/t-multilib
23075 $(srcdir)/config/avr/avr-tables.opt
23076 $(srcdir)/doc/avr-mmcu.texi
23077 (avr-mcus): New phony target to build them on request.
23078 (s-avr-mlib, s-avr-mmcu-texi): Remove.
23079 * avr/avr-mcus.def: Adjust comments.
23081 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
23083 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
23085 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
23087 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
23089 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
23091 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
23093 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
23095 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
23096 to generate profiling.
23097 * config/rs6000/aix64.h (LIB_SPEC): Same.
23099 2013-01-04 Andrew Pinski <apinski@cavium.com>
23101 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
23103 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
23105 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
23107 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
23109 (ix86_expand_move): Ditto.
23110 (ix86_zero_extend_to_Pmode): Ditto.
23111 (ix86_expand_call): Ditto.
23112 (ix86_expand_special_args_builtin): Ditto.
23113 (ix86_expand_builtin): Ditto.
23115 2013-01-04 Richard Biener <rguenther@suse.de>
23117 PR tree-optimization/55862
23118 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
23119 translating them through PHI nodes.
23121 2013-01-04 Martin Jambor <mjambor@suse.cz>
23123 PR tree-optimization/55755
23124 * tree-sra.c (sra_modify_assign): Do not check that an access has no
23125 children when trying to avoid producing a VIEW_CONVERT_EXPR.
23127 2013-01-04 Marek Polacek <polacek@redhat.com>
23129 PR middle-end/55859
23130 * opts.c (default_options_optimization): Clarify error message.
23132 2013-01-04 Richard Biener <rguenther@suse.de>
23134 PR middle-end/55863
23135 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
23138 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23141 * config/pa/pa.md (movsi): Revert previous change.
23142 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
23145 2013-01-03 Richard Henderson <rth@redhat.com>
23147 * config/i386/i386.c (ix86_expand_move): Always assign to op1
23148 after eliminating TLS symbols.
23150 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
23153 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
23154 * graphite-poly.c (debug_gmp_value): Likewise.
23156 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
23159 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
23160 selected code model, define __code_mode_small__, __code_model_medium__,
23161 __code_model_large__, __code_model_32__ or __code_model_kernel__.
23162 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
23163 xchg temporary register with %k. Declare temporary register as
23165 [__x86_64__]: For medium and large code models, preserve %rbx register.
23167 2013-01-03 Richard Biener <rguenther@suse.de>
23169 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
23170 (dump_subscript): Adjust.
23171 (finalize_ddr_dependent): Do not dump redundant info.
23172 (analyze_siv_subscript): Adjust.
23173 (subscript_dependence_tester): Likewise.
23174 (compute_affine_dependence): Likewise.
23176 2013-01-03 Richard Biener <rguenther@suse.de>
23179 2013-01-03 Richard Biener <rguenther@suse.de>
23181 PR tree-optimization/55857
23182 * tree-vect-stmts.c (vectorizable_load): Do not setup
23183 re-alignment for invariant loads.
23185 2013-01-02 Richard Biener <rguenther@suse.de>
23187 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
23188 invariant load do not generate a vector load from the scalar location.
23190 2013-01-03 Richard Biener <rguenther@suse.de>
23192 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
23193 for not vectorizing.
23194 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
23195 not build INDIRECT_REFs, call get_name once only.
23196 (vect_create_data_ref_ptr): Likewise. Dump base object kind
23197 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
23199 2013-01-03 Richard Biener <rguenther@suse.de>
23201 PR tree-optimization/55857
23202 * tree-vect-stmts.c (vectorizable_load): Do not setup
23203 re-alignment for invariant loads.
23205 2013-01-03 Richard Biener <rguenther@suse.de>
23208 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
23209 prefer a built-in decl.
23211 2013-01-03 Jakub Jelinek <jakub@redhat.com>
23213 * gcc.c (process_command): Update copyright notice dates.
23214 * gcov.c (print_version): Likewise.
23215 * gcov-dump.c (print_version): Likewise.
23217 PR rtl-optimization/55838
23218 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
23219 iv0.step, iv1.step and step.
23221 2013-01-03 Jakub Jelinek <jakub@redhat.com>
23222 Marc Glisse <marc.glisse@inria.fr>
23224 PR tree-optimization/55832
23225 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
23226 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
23227 integer_{one,zero}_node.
23229 2013-01-03 Jakub Jelinek <jakub@redhat.com>
23232 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
23233 * var-tracking.c (reverse_op): Don't add reverse ops to
23234 VALUEs that have already
23235 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
23237 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
23239 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
23241 2013-01-02 Teresa Johnson <tejohnson@google.com>
23243 * dumpfile.c (dump_loc): Print filename with location.
23244 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
23245 new location_t parameter to emit complete unroll message with
23246 new dump framework.
23247 (canonicalize_loop_induction_variables): Compute loops location
23248 and pass to try_unroll_loop_completely.
23249 * loop-unroll.c (report_unroll_peel): New function.
23250 (peel_loops_completely): Use new dump format with location
23251 for main dumpfile message, and invoke report_unroll_peel on success.
23252 (decide_unrolling_and_peeling): Ditto.
23253 (decide_peel_once_rolling): Remove old dumpfile message subsumed
23254 by report_unroll_peel.
23255 (decide_peel_completely): Ditto.
23256 (decide_unroll_constant_iterations): Ditto.
23257 (decide_unroll_runtime_iterations): Ditto.
23258 (decide_peel_simple): Ditto.
23259 (decide_unroll_stupid): Ditto.
23260 * cfgloop.c (get_loop_location): New function.
23261 * cfgloop.h (get_loop_location): Declare.
23263 2013-01-02 Sriraman Tallam <tmsriram@google.com>
23265 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
23268 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23270 PR middle-end/55198
23271 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
23272 BLKmode objects when EXPAND_MEMORY is specified.
23274 2013-01-02 Sriraman Tallam <tmsriram@google.com>
23276 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
23278 (fold_builtin_cpu): Do not share cpu model decls across statements.
23280 2013-01-02 Jason Merrill <jason@redhat.com>
23283 * tree.c (build_array_type_1): Revert earlier change.
23285 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
23287 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
23289 * config/aarch64/aarch64-tune.md: Re-generate.
23291 2013-01-02 Richard Biener <rguenther@suse.de>
23293 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
23294 invariant load do not generate a vector load from the scalar location.
23296 2013-01-02 Richard Biener <rguenther@suse.de>
23299 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
23300 * configure: Regenerate.
23302 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
23304 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
23305 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
23306 (expand_builtin_int_roundingfn_2): Keep the original target around
23307 for the fallback case.
23309 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
23311 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
23312 to be clear for sign changes.
23314 2013-01-01 Jan Hubicka <jh@suse.cz>
23316 * ipa-inline-analysis.c: Fix formatting.
23318 2013-01-01 Jakub Jelinek <jakub@redhat.com>
23320 PR tree-optimization/55831
23321 * tree-vect-loop.c (get_initial_def_for_induction): Use
23322 gsi_after_labels instead of gsi_start_bb.
23324 Copyright (C) 2013 Free Software Foundation, Inc.
23326 Copying and distribution of this file, with or without modification,
23327 are permitted in any medium without royalty provided the copyright
23328 notice and this notice are preserved.