1 2009-01-09 Steven Bosscher <steven@gcc.gnu.org>
3 * regrename.c (regrename_optimize): Fix dumping.
4 (find_oldest_value_reg): Preserve REG_POINTER.
5 (copy_hardreg_forward_1): Likewise.
7 2009-01-09 Diego Novillo <dnovillo@google.com>
9 * gimple.h (struct gimple_statement_base)<uid>: Document
10 the restrictions on its use.
12 (gimple_set_uid): Tidy.
14 2009-01-09 Jakub Jelinek <jakub@redhat.com>
16 * config/i386/i386.c (ix86_expand_movmem, ix86_expand_setmem): Add
17 zero guard even if align_bytes != 0 and count is smaller than
20 2008-01-09 Vladimir Makarov <vmakarov@redhat.com>
22 PR rtl-optimization/38495
23 * ira-emit.c (print_move_list, ira_debug_move_list): New
25 (add_range_and_copies_from_move_list): Print all added ranges.
26 Add ranges to memory optimized destination.
28 2009-01-09 Jakub Jelinek <jakub@redhat.com>
32 * config/i386/i386.c (override_options): Reject
33 -mstringop-strategy=rep_8byte with -m32.
34 (ix86_expand_movmem): For size_needed == 1 set epilogue_size_needed
35 to 1. Do count comparison against epilogue_size_needed at compile
36 time even when count_exp was constant forced into register. For
37 size_needed don't jump to epilogue, instead just avoid aligning
38 and invoke the body algorithm. If need_zero_guard, add zero guard
39 even if count is non-zero, but smaller than size_needed + number of
40 bytes that could be stored for alignment.
41 (ix86_expand_setmem): For size_needed == 1 set epilogue_size_needed
42 to 1. If need_zero_guard, add zero guard even if count is non-zero,
43 but smaller than size_needed + number of bytes that could be stored
44 for alignment. Compare size_needed with epilogue_size_needed instead
45 of desired_align - align, don't adjust size_needed, pass
46 epilogue_size_needed to the epilogue expanders.
49 * c-pretty-print.c (pp_c_expression): Handle GOTO_EXPR like
52 2009-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
54 * pa.c (last_address): Change to unsigned.
55 (update_total_code_bytes): Change argument to unsigned. Don't
56 check if insn addresses are set.
57 (pa_output_function_epilogue): Set last_address to UINT_MAX if insn
58 addresses are not set.
59 (pa_asm_output_mi_thunk): Handle wrap when updating last_address.
61 2009-01-09 Nick Clifton <nickc@redhat.com>
63 * config/sh/symbian.c: Replace uses of DECL_INLINE with
64 DECL_DECLARED_INLINE_P.
66 2009-01-09 Jakub Jelinek <jakub@redhat.com>
69 * dojump.c (do_jump_by_parts_zero_rtx): Use mode instead of
70 GET_MODE (op0) in operand_subword_force calls.
73 * fold-const.c (fold_unary): For COMPOUND_EXPR and COND_EXPR,
74 fold_convert arg0 operands to TREE_TYPE (op0) first.
76 2009-01-08 Vladimir Makarov <vmakarov@redhat.com>
78 * params.def (ira-max-conflict-table-size): Decrease default value
81 2009-01-08 Jakub Jelinek <jakub@redhat.com>
83 PR tree-optimization/37031
84 * lambda-code.c (lambda_collect_parameters): Call pointer_set_destroy
86 (build_access_matrix): Reserve correct size for AM_MATRIX vector,
87 allocate it using gc instead of heap, use VEC_quick_push instead of
89 * graphite.c (build_access_matrix): Allocate AM_MATRIX vector using gc
90 instead of heap, use VEC_quick_push instead of VEC_safe_push.
91 * tree-data-ref.h (struct access_matrix): Change matrix to gc
92 allocated vector from heap allocated.
93 * lambda.h: Add DEF_VEC_ALLOC_P for gc allocated lambda_vector.
94 * tree-loop-linear.c (linear_transform_loops): Allocate nest
95 vector only after perfect_loop_nest_depth call.
97 2009-01-08 Sebastian Pop <sebastian.pop@amd.com>
98 Jan Sjodin <jan.sjodin@amd.com>
100 PR tree-optimization/38559
101 * graphite.c (debug_value, copy_constraint,
102 swap_constraint_variables, scale_constraint_variable, ): New.
103 (get_lower_bound, get_upper_bound): Removed.
104 (graphite_trans_bb_strip_mine): Clean up this code that works
105 only for constant number of iterations. Fully copy upper and
106 lower bound constraints, not only the constant part of them.
107 * graphite.h (debug_value): Declared.
109 2009-01-08 Ira Rosen <irar@il.ibm.com>
111 PR tree-optimization/37194
112 * tree-vect-transform.c (vect_estimate_min_profitable_iters):
113 Don't add the cost of cost model guard in prologue to scalar
114 outside cost in case of known number of iterations.
116 2009-01-07 Nathan Froyd <froydnj@codesourcery.com>
117 Alan Modra <amodra@bigpond.net.au>
119 * config/rs6000/rs6000.c (rs6000_legitimize_address): Check for
120 non-word-aligned REG+CONST addressing.
122 2009-01-07 Uros Bizjak <ubizjak@gmail.com>
125 * config/alpha/alpha.c (alpha_end_function): For TARGET_ABI_OSF, call
126 free_after_compilation when outputting a thunk.
127 (alpha_output_mi_thunk_osf): Assert that we are processing a thunk.
128 Do not call free_after_compilation here.
130 2009-01-07 Uros Bizjak <ubizjak@gmail.com>
132 * config/i386/i386.c (ix86_target_string): Use ARRAY_SIZE.
133 (ix86_valid_target_attribute_inner_p): Ditto.
135 2009-01-07 Jan Sjodin <jan.sjodin@amd.com>
137 PR tree-optimization/38492
138 PR tree-optimization/38498
139 * tree-check.c (operator_is_linear, scev_is_linear_expression): New.
140 * tree-chrec.h (scev_is_linear_expression): Declared.
141 * graphite.c (graphite_cannot_represent_loop_niter): New.
142 (scopdet_basic_block_info): Call graphite_cannot_represent_loop_niter.
143 (graphite_loop_normal_form): Use gcc_assert.
144 (scan_tree_for_params): Use CASE_CONVERT.
145 (phi_node_is_iv, bb_contains_non_iv_scalar_phi_nodes): New.
146 (build_scop_conditions_1): Call bb_contains_non_iv_scalar_phi_nodes.
147 Use gcc_assert. Discard scops that contain unhandled cases.
148 (build_scop_conditions): Return a boolean status for unhandled cases.
149 (strip_mine_profitable_p): Print the loop number, not its depth.
150 (is_interchange_valid): Pass the depth of the loop nest, don't
151 recompute it wrongly.
152 (graphite_trans_bb_block): Same.
153 (graphite_trans_bb_block): Print tentative of loop blocking.
154 (graphite_trans_scop_block): Do not print that the loop has been
156 (graphite_transform_loops): Do not handle scops that contain condition
159 2009-01-07 H.J. Lu <hongjiu.lu@intel.com>
161 AVX Programming Reference (December, 2008)
162 * config/i386/avxintrin.h (_mm256_stream_si256): New.
163 (_mm256_stream_pd): Likewise.
164 (_mm256_stream_ps): Likewise.
166 * config/i386/i386.c (ix86_builtins): Add IX86_BUILTIN_MOVNTDQ256,
167 IX86_BUILTIN_MOVNTPD256 and IX86_BUILTIN_MOVNTPS256.
168 (ix86_special_builtin_type): Add VOID_FTYPE_PV4DI_V4DI.
169 (bdesc_special_args): Add __builtin_ia32_movntdq256,
170 __builtin_ia32_movntpd256 and __builtin_ia32_movntps256.
171 (ix86_init_mmx_sse_builtins): Handle VOID_FTYPE_PV4DI_V4DI.
172 (ix86_expand_special_args_builtin): Likewise.
174 * config/i386/sse.md (AVXMODEDI): New.
175 (avx_movnt<mode>): Likewise.
176 (avx_movnt<mode>): Likewise.
177 (<sse>_movnt<mode>): Remove AVX support.
178 (sse2_movntv2di): Likewise.
180 2009-01-07 Richard Guenther <rguenther@suse.de>
183 * fold-const.c (extract_muldiv): Remove obsolete comment.
184 (fold_plusminus_mult_expr): Undo MINUS_EXPR
185 to PLUS_EXPR canonicalization for the canonicalization.
187 2009-01-07 Gerald Pfeifer <gerald@pfeifer.com>
189 * doc/install.texi (alpha*-dec-osf*): Remove note on 32-bit
190 hosted cross-compilers generating less efficient code.
192 2009-01-06 Richard Sandiford <rdsandiford@googlemail.com>
194 * function.h (rtl_data): Add a dbr_scheduled_p field.
195 * reorg.c (dbr_schedule): Set it.
196 (gate_handle_delay_slots): Check it.
197 * config/mips/mips.c (mips_base_delayed_branch): Delete.
198 (mips_reorg): Check flag_delayed_branch instead of
199 mips_base_delayed_branch.
200 (mips_override_options): Don't set mips_base_delayed_branch
201 or flag_delayed_branch.
203 2009-01-06 Richard Sandiford <rdsandiford@googlemail.com>
205 PR rtl-optimization/38426.
206 * ira.c (ira): Set current_function_is_leaf earlier.
208 2009-01-06 Jakub Jelinek <jakub@redhat.com>
210 PR rtl-optimization/38722
211 * combine.c (try_combine): Don't modify PATTERN (i3) and notes
212 too early, only set a flag and modify after last possible
215 2009-01-06 Janis Johnson <janis187@us.ibm.com>
218 * ginclude/float.h: Rename DECnn_DEN to DECnn_SUBNORMAL_MIN.
219 * real.c (decimal_single_format): Correct values of emin and emax.
220 (decimal_double_format): Ditto.
221 (decimal_quad_format): Ditto.
222 * c-cppbuiltin.c (builtin_define_decimal_float_constants): Adjust
223 computation of DECnn_MIN and DECnn_MAX for corrected values of
224 emin and emax. Define __DECnn_SUBNORMAL_MIN__ instead of
225 __DECnn_MIN__, and adjust its computation for the corrected value
228 2009-01-06 Jan Hubicka <jh@suse.cz>
231 * i386.c (ix86_expand_call): Use ARRAY_SIZE.
233 2009-01-06 Gerald Pfeifer <gerald@pfeifer.com>
235 * doc/contrib.texi (Contributors): Slightly adjust the end note.
236 Add Robert Clark to the list of testers.
238 2009-01-06 Jan Hubicka <jh@suse.cz>
239 Kai Tietz <kai.tietz@onevision.com>
241 * i386.md (*msabi_syvabi): Add SSE regs clobbers.
242 * i386.c (ix86_expand_call): Add clobbers.
244 2009-01-06 Jan Hubicka <jh@suse.cz>
245 Kai Tietz <kai.tietz@onevision.com>
247 * i386.h (CONDITIONAL_CALL_USAGE): SSE regs are not used for w64 ABI.
248 * i386.c (struct ix86_frame): Add padding0 and nsseregs.
249 (ix86_nsaved_regs): Count only general purpose regs.
250 (ix86_nsaved_sseregs): New.
251 (ix86_compute_frame_layout): Update nsseregs; set preferred alignment
252 to 16 for w64; compute padding and size of sse reg save area.
253 (ix86_emit_save_regs, ix86_emit_save_regs_using_mov): Save only
254 general purpose regs.
255 (ix86_emit_save_sse_regs_using_mov): New.
256 (ix86_expand_prologue): Save SSE regs if needed.
257 (ix86_emit_restore_regs_using_mov): Use only general purpose regs.
258 (ix86_emit_restore_sse_regs_using_mov): New.
259 (ix86_expand_epilogue): Save SSE regs if needed.
261 2009-01-06 Jan Hubicka <jh@suse.cz>
262 Kai Tietz <kai.tietz@onevision.com>
264 * i386.h (ACCUMULATE_OUTGOING_ARGS): Enable for MSABI
265 * i386.c (init_cumulative_args): Disallow calls of MSABI functions
266 when accumulate outgoing args is off.
268 2009-01-06 H.J. Lu <hongjiu.lu@intel.com>
271 * ira-color.c (ira_reuse_stack_slot): Check ENABLE_IRA_CHECKING
272 before using pseudos_have_intersected_live_ranges_p.
274 * ira-int.h (ira_assert): Always define.
276 2009-01-06 H.J. Lu <hongjiu.lu@intel.com>
278 AVX Programming Reference (December, 2008)
279 * config/i386/avxintrin.h (_mm_permute2_pd): Removed.
280 (_mm256_permute2_pd): Likewise.
281 (_mm_permute2_ps): Likewise.
282 (_mm256_permute2_ps): Likewise.
283 * config/i386/i386.md (UNSPEC_VPERMIL2): Likewise.
284 * config/i386/sse.md (avx_vpermil2<mode>3): Likewise.
286 * config/i386/i386.c (ix86_builtins): Remove
287 IX86_BUILTIN_VPERMIL2PD, IX86_BUILTIN_VPERMIL2PS,
288 IX86_BUILTIN_VPERMIL2PD256 and IX86_BUILTIN_VPERMIL2PS256.
289 (ix86_builtin_type): Remove V8SF_FTYPE_V8SF_V8SF_V8SI_INT,
290 V4DF_FTYPE_V4DF_V4DF_V4DI_INT, V4SF_FTYPE_V4SF_V4SF_V4SI_INT
291 and V2DF_FTYPE_V2DF_V2DF_V2DI_INT.
292 (bdesc_args): Remove __builtin_ia32_vpermil2pd,
293 __builtin_ia32_vpermil2ps, __builtin_ia32_vpermil2pd256 and
294 __builtin_ia32_vpermil2ps256.
295 (ix86_init_mmx_sse_builtins): Updated.
296 (ix86_expand_args_builtin): Likewise.
298 2009-01-05 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
300 * pa.c (output_call): Relocate non-jump insns in the delay slot of
301 long absolute calls when generating PA 2.0 code.
303 2009-01-05 Vladimir Makarov <vmakarov@redhat.com>
305 PR rtl-optimization/38583
306 * params.h (IRA_MAX_CONFLICT_TABLE_SIZE): New macro.
308 * params.def (ira-max-conflict-table-size): New.
310 * doc/invoke.texi (ira-max-conflict-table-size): Decribe.
312 * ira.h (ira_conflicts_p): New external definition.
314 * ira-conflicts.c (build_conflict_bit_table): Do not build too big
315 table. Report this. Return result of building.
316 (ira_build_conflicts): Use ira_conflicts_p. Check result of
317 building conflict table.
319 * ira-color.c (fast_allocation): Use num instead of ira_allocnos_num.
320 (ira_color): Use ira_conflicts_p.
322 * global.c: Include ira.h.
323 (pseudo_for_reload_consideration_p, build_insn_chain): Use
326 * Makefile.in (global.o): Add ira.h.
328 * ira-build.c (mark_all_loops_for_removal,
329 propagate_some_info_from_allocno): New.
330 (remove_unnecessary_allocnos): Call
331 propagate_some_info_from_allocno.
332 (remove_low_level_allocnos): New.
333 (remove_unnecessary_regions): Add parameter. Call
334 mark_all_loops_for_removal and remove_low_level_allocnos. Pass
335 parameter to remove_unnecessary_regions.
336 (ira_build): Remove all regions but root if the conflict table was
337 not built. Update conflict hard regs for allocnos crossing calls.
339 * ira.c (ira_conflicts_p): New global.
340 (ira): Define and use ira_conflicts_p.
342 * reload1.c (compute_use_by_pseudos, reload, count_pseudo,
343 count_spilled_pseudo, find_reg, alter_reg, finish_spills,
344 emit_input_reload_insns, delete_output_reload): Use ira_conflicts_p.
346 2009-01-06 Ben Elliston <bje@au.ibm.com>
348 * gengtype-lex.l (YY_NO_INPUT): Define.
350 2009-01-05 Andrew Pinski <andrew_pinski@playstation.sony.com>
353 * c-common.c (handle_vector_size_attribute): Also reject
356 2009-01-05 Sebastian Pop <sebastian.pop@amd.com>
358 PR tree-optimization/38492
359 * graphite.c (rename_map_elt, debug_rename_elt,
360 debug_rename_map_1, debug_rename_map, new_rename_map_elt,
361 rename_map_elt_info, eq_rename_map_elts,
362 get_new_name_from_old_name, bb_in_sese_p): Moved around.
363 (sese_find_uses_to_rename_use): Renamed sese_build_livein_liveouts_use.
364 (sese_find_uses_to_rename_bb): Renamed sese_build_livein_liveouts_bb.
365 (sese_build_livein_liveouts): New.
366 (new_sese, free_sese): New.
367 (new_scop): Call new_sese.
368 (free_scop): Call free_sese.
369 (rename_variables_from_edge, rename_phis_end_scop): Removed.
370 (register_old_new_names): Renamed register_old_and_new_names.
371 (register_scop_liveout_renames, add_loop_exit_phis,
372 insert_loop_close_phis, struct igp,
373 default_liveout_before_guard, add_guard_exit_phis,
374 insert_guard_phis, copy_renames): New.
375 (translate_clast): Call insert_loop_close_phis and insert_guard_phis.
376 (sese_add_exit_phis_edge): Renamed scop_add_exit_phis_edge.
377 (rewrite_into_sese_closed_ssa): Renamed scop_insert_phis_for_liveouts.
378 (scop_adjust_phis_for_liveouts): New.
379 (gloog): Call scop_adjust_phis_for_liveouts.
381 * graphite.h (struct sese): Documented. Added fields liveout,
383 (SESE_LIVEOUT, SESE_LIVEIN, SESE_LIVEIN_VER, SESE_NUM_VER): New.
384 (new_sese, free_sese, sese_build_livein_liveouts): Declared.
385 (struct scop): Added field liveout_renames.
386 (SCOP_LIVEOUT_RENAMES): New.
388 2009-01-05 Harsha Jagasia <harsha.jagasia@amd.com>
390 PR tree-optimization/38510
391 * graphite.c (recompute_all_dominators): Call mark_irreducible_loops.
392 (translate_clast): Call recompute_all_dominators before
394 (gloog): Call recompute_all_dominators before graphite_verify.
396 2009-01-05 Harsha Jagasia <harsha.jagasia@amd.com>
397 Jan Sjodin <jan.sjodin@amd.com>
399 PR tree-optimization/38500
400 * graphite.c (create_sese_edges): Call fix_loop_structure after
403 2009-01-05 Joel Sherrill <joel.sherrill@oarcorp.com>
405 * config.gcc: Add m32r*-*-rtems*.
406 * config/m32r/rtems.h: New file.
408 2009-01-05 Ben Elliston <bje@au.ibm.com>
410 * Makefile.in (.po.gmo): Use mkinstalldirs, not test -d || mkdir.
412 (po/gcc.pot): Likewise.
414 2009-01-04 David S. Miller <davem@davemloft.net>
416 * config/sparc/sparc.h (SECONDARY_MEMORY_NEEDED_RTX): Delete.
417 (STARTING_FRAME_OFFSET): Always set to zero.
419 2009-01-04 Richard Sandiford <rdsandiford@googlemail.com>
421 * tree.def (LSHIFT_EXPR, RSHIFT_EXPR): Add commentary.
422 * tree-cfg.c (verify_gimple_assign_binary): Allow shifts of
423 fixed-point types, and vectors of the same.
425 2009-01-04 Richard Sandiford <rdsandiford@googlemail.com>
427 * config/mips/sync.md (*mb_barrier): Rename to...
428 (*memory_barrier): ...this.
430 2009-01-04 Jonathan Wakely <jwakely.gcc@gmail.com>
432 * doc/extend.texi (Function Attributes): Move @cindex after @item
433 for 'artificial' and 'flatten'. Fix grammar for 'externally_visible'
434 and put in alphabetical order. Fix 'target' name and put in order.
435 * doc/invoke.texi (-Wstrict-null-sentinel, -fipa-matrix-reorg): Fix
438 2009-01-04 Uros Bizjak <ubizjak@gmail.com>
440 * config/s390/s390.md (UNSPEC_MB): Rename from UNSPECV_MB.
441 (memory_barrier): Expand as unspec instead of unspec_volatile.
442 Remove mem:BLK from insn operands. Use Pmode scratch register.
443 (*memory_barrier): Define as unspec instead of unspec_volatile.
444 Use (match_dup 0) as input operand.
446 * config/sparc/sparc.md (UNSPEC_MEMBAR): Rename from UNSPECV_MEMBAR.
447 * config/sparc/sync.md (memory_barrier): Expand as unspec instead of
448 unspec_volatile. Remove mem:BLK from insn operands. Use Pmode
449 scratch register. Remove operand 1.
450 (*stbar): Define as unspec instead of unspec_volatile.
451 Use (match_dup 0) as input operand, remove (const_int 8).
452 (*membar): Define as unspec instead of unspec_volatile.
453 Use (match_dup 0) as input operand, remove input operand 2.
455 * config/xtensa/xtensa.md (UNSPEC_MEMW): Rename from UNSPECV_MEMW.
456 (memory_barrier): Expand as unspec instead of unspec_volatile.
457 Remove mem:BLK from insn operands. Use Pmode scratch register.
458 (*memory_barrier): Define as unspec instead of unspec_volatile.
459 Use (match_dup 0) as input operand.
461 * config/ia64/sync.md (memory_barrier): Redefine as expander pattern.
462 Remove mem:BLK from insn operands. Use Pmode scratch register.
463 Set volatile flag on operand 0.
464 (*memory_barrier): New insn pattern.
466 * config/rs6000/sync.md (memory_barrier): Remove mem:BLK from
468 (*memory_barrier): Use (match_dup 0) as input operand.
470 * config/mips/sync.md (memory_barrier): Redefine as expander pattern.
471 Remove mem:BLK from insn operands. Use Pmode scratch register.
472 Set volatile flag on operand 0.
473 (*mb_internal): New insn pattern.
475 * config/alpha/sync.md (*memory_barrier): Rename from *mb_internal.
477 2009-01-04 Steven Bosscher <steven@gcc.gnu.org>
480 * function.c (struct temp_slot): Move to the section of the file
481 that deals with temp slots. Remove field 'address'.
482 (temp_slot_address_table): New hash table of address -> temp slot.
483 (struct temp_slot_address_entry): New struct, items for the table.
484 (temp_slot_address_compute_hash, temp_slot_address_hash,
485 temp_slot_address_eq, insert_temp_slot_address): Support functions
487 (find_temp_slot_from_address): Rewrite to use the new hash table.
488 (remove_unused_temp_slot_addresses): Remove addresses of temp
489 slots that have been made available.
490 (remove_unused_temp_slot_addresses_1): Call-back for htab_traverse,
491 worker function for remove_unused_temp_slot_addresses.
492 (assign_stack_temp_for_type): Don't clear the temp slot address list.
493 Add the temp slot address to the address -> temp slot map.
494 (update_temp_slot_address): Update via insert_temp_slot_address.
495 (free_temp_slots): Call remove_unused_temp_slot_addresses.
496 (pop_temp_slots): Likewise.
497 (init_temp_slots): Allocate the address -> temp slot map, or empty
498 the map if it is already allocated.
499 (prepare_function_start): Initialize temp slot processing.
501 2009-01-04 Steven Bosscher <steven@gcc.gnu.org>
504 * cfgexpand.c (estimate_stack_frame_size): Simplify the estimate:
505 Calculate the size of all stack vars assuming no packing of stack
506 vars will happen, replacing a quadratic algorithm with a linear one.
508 2009-01-03 Jakub Jelinek <jakub@redhat.com>
511 * expmed.c (store_bit_field_1): Don't modify op0 if movstrict insn
514 2009-01-03 Diego Novillo <dnovillo@google.com>
516 * doc/contrib.texi: Update contributions.
518 2009-01-03 Jakub Jelinek <jakub@redhat.com>
521 * builtins.c (fold_builtin_memory_op): Give up if either operand
522 is volatile. Set srctype or desttype to non-qualified version
526 * builtins.c (fold_builtin_expect): Only check DECL_WEAK for VAR_DECLs
529 2009-01-02 Kenneth Zadeck <zadeck@naturalbridge.com>
531 PR rtl-optimization/35805
532 * df-problems.c (df_lr_finalize): Add recursive call to resolve lr
533 problem if fast dce is able to remove any instructions.
534 * dce.c (dce_process_block): Fix dump message.
536 2009-01-02 Mark Mitchell <mark@codesourcery.com>
539 * tree-ssa-pre.c (compute_antic): Correct loop bounds.
541 2009-01-02 Jakub Jelinek <jakub@redhat.com>
544 * tree-flow.h (op_code_prio, op_prio): New prototypes.
545 * tree-pretty-print.c (op_code_prio): New function.
546 (op_prio): No longer static. Use op_code_prio.
547 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs):
548 Use op_prio and op_code_prio to determine if () should be
549 printed around operand(s) or not.
551 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
552 dump_gimple_call, dump_gimple_switch, dump_gimple_cond,
553 dump_gimple_label, dump_gimple_try, dump_symbols, dump_gimple_phi,
554 dump_gimple_mem_ops, dump_bb_header, dump_bb_end, pp_cfg_jump): Use
555 pp_character instead of pp_string for single letter printing.
557 2009-01-02 Richard Sandiford <rdsandiford@googlemail.com>
559 * doc/extend.texi: Fix '#pragma GCC option' typo.
561 2009-01-02 Richard Guenther <rguenther@suse.de>
563 * doc/install.texi (--enable-checking): Mention different
565 (--enable-stage1-checking): Document.
567 2009-01-01 Andrew Pinski <pinskia@gmail.com>
570 * tree-cfg.c (verify_expr): Add INDIRECT_REF case. Change MODIFY_EXPR
573 2009-01-02 Ben Elliston <bje@au.ibm.com>
575 * config/fp-bit.h (pack_d): Constify argument.
576 * config/fp-bit.c (makenan): Constify return type. Remove casts.
577 (isnan): Constify argument.
581 (_fpadd_parts): Constify return type.
582 (_fpmul_parts): Likewise.
583 (_fpdiv_parts): Likewise.
585 2009-01-01 Jakub Jelinek <jakub@redhat.com>
588 * c-typeck.c (add_pending_init): Add IMPLICIT argument. Only
589 warn about overwriting initializer with side-effects or
590 -Woverride-init if !IMPLICIT.
591 (output_init_element): Likewise. Pass IMPLICIT down to
593 (process_init_element): Add IMPLICIT argument. Pass it down
594 to output_init_element.
595 (push_init_element, pop_init_level, set_designator): Adjust
596 process_init_element callers.
597 (set_nonincremental_init, set_nonincremental_init_from_string):
598 Adjust add_pending_init callers.
599 (output_pending_init_elements): Adjust output_init_element callers.
600 * c-tree.h (process_init_element): Adjust prototype.
601 * c-parser.c (c_parser_initelt, c_parser_initval): Adjust
602 process_init_element callers.