Richard Guenther [Thu, 15 Nov 2007 11:15:08 +0000 (11:15 +0000)]
tree-ssa-alias.c (create_overlap_variables_for): Make sure to only create SFTs if we also create variable infos for PTA.
2007-11-15 Richard Guenther <rguenther@suse.de>
* tree-ssa-alias.c (create_overlap_variables_for): Make sure
to only create SFTs if we also create variable infos for PTA.
* tree-ssa-structalias.c (set_uids_in_ptset): Add all overlapping
SFTs.
Richard Guenther [Thu, 15 Nov 2007 09:06:58 +0000 (09:06 +0000)]
tree-ssa-structalias.c (set_uids_in_ptset): Use the pointed-to type for TBAA pruning.
2007-11-15 Richard Guenther <rguenther@suse.de>
* tree-ssa-structalias.c (set_uids_in_ptset): Use the pointed-to
type for TBAA pruning.
(find_what_p_points_to): Call set_uids_in_ptset with the original
pointer, not the decl of the solution we collapsed to.
Eric Botcazou [Wed, 14 Nov 2007 19:06:10 +0000 (20:06 +0100)]
re PR tree-optimization/34046 (verify_flow_info failed)
PR tree-optimization/34046
* cfg.c (update_bb_profile_for_threading): Avoid the division for the
scaling if the old probability is greater than the new base.
Eric Botcazou [Wed, 14 Nov 2007 18:31:57 +0000 (18:31 +0000)]
trans.c (call_to_gnu): Always set the source location on the call expression.
* trans.c (call_to_gnu): Always set the source location on the call
expression. If the function returns-by-target, also set it on the
address expression.
Johannes Singler [Wed, 14 Nov 2007 17:37:38 +0000 (17:37 +0000)]
multiway_merge.h: More robust finding of an arbitrary existing element inside the input sequences.
2007-11-14 Johannes Singler <singler@ira.uka.de>
* include/parallel/multiway_merge.h: More robust finding of an
arbitrary existing element inside the input sequences.
* include/bits/stl_algo.h: Fix typo to actually call appropriate
sequential version.
Jerry DeLisle [Wed, 14 Nov 2007 01:06:13 +0000 (01:06 +0000)]
re PR fortran/33162 (INTRINSIC functions as ACTUAL argument)
2007-11-13 Jerry DeLisle <jvdelisle@gcc.gnu.org>
PR fortran/33162
*gfortran.dg/proc_decl_1.f90: Update.
*gfortran.dg/proc_decl_7.f90: New test.
*gfortran.dg/proc_decl_8.f90: New test.
*gfortran.dg/proc_decl_9.f90: New test.
*gfortran.dg/proc_decl_10.f90: New test.
Jerry DeLisle [Wed, 14 Nov 2007 00:59:09 +0000 (00:59 +0000)]
re PR fortran/33162 (INTRINSIC functions as ACTUAL argument)
2007-11-11 Jerry DeLisle <jvdelisle@gcc.gnu.org>
PR fortran/33162
* decl.c (match_procedure_decl): Remove TODO and allow intrinsics in
PROCEDURE declarations. Set attr.untyped to allow the interface to be
resolved later where the symbol type will be set.
* interface.c (compare_intr_interfaces): Remove static from pointer
declarations. Add type and kind checks for dummy function arguments.
(compare_actual_formal_intr): New function to compare an actual
argument with an intrinsic function. (gfc_procedures_use): Add check for
interface that points to an intrinsic function, use the new function.
* resolve.c (resolve_specific_f0): Resolve the intrinsic interface.
(resolve_specific_s0): Ditto.
Samuel Tardieu [Tue, 13 Nov 2007 19:12:57 +0000 (19:12 +0000)]
sem_prag.adb (Process_Convention): Move the test for the entity on which the Convention pragma applies down to...
gcc/ada/
* sem_prag.adb (Process_Convention): Move the test for the
entity on which the Convention pragma applies down to also
forbid pragma Convention on enumeration literals reached
through renamings.
Jakub Jelinek [Tue, 13 Nov 2007 18:27:09 +0000 (19:27 +0100)]
re PR c++/34054 (ICE with parameter pack in return type)
PR c++/34054
PR c++/34056
PR c++/34057
PR c++/34058
PR c++/34060
* pt.c (find_parameter_packs_r): If ppd->set_packs_to_error,
set to error_mark_node the outermost POINTER_TYPE to the pack if
it is seen in a POINTER_TYPE.
(push_template_decl_real): If check_for_bare_parameter_packs
fails for function return type, set the return type to
integer_type_node. If check_for_bare_parameter_packs failed
for non-function, return error_mark_node.
* g++.dg/parse/crash36.C: Add another dg-error.
* g++.dg/cpp0x/pr34054.C: New test.
* g++.dg/cpp0x/pr34056.C: New test.
* g++.dg/cpp0x/pr34057.C: New test.
* g++.dg/cpp0x/pr34058.C: New test.
* g++.dg/cpp0x/pr34060.C: New test.
Diego Novillo [Tue, 13 Nov 2007 15:20:40 +0000 (10:20 -0500)]
pr 33870
pr 33870
* tree.h (strcut tree_memory_tag): add field unpartitionable.
remove field in_nested_struct.
(struct tree_struct_field_tag): add field nesting_level.
(sft_in_nested_struct): remove.
(sft_nesting_level): define.
(sft_unpartitionable_p): define.
* tree-ssa-alias.c (mem_sym_score): if mp->var is not
partitionable, return long_max.
(compute_memory_partitions): do not partition sfts marked
unpartitionable.
(create_sft): add argument nesting_level. set
sft_nesting_level with it. update all users.
(create_overlap_variables_for): show nesting level.
* tree-dfa.c (dump_subvars_for): likewise.
(dump_variable): likewise.
show whether the sft is partitionable or not.
* tree-flow.h (struct fieldoff): remove field
in_nested_struct.
add field nesting_level.
* tree-ssa-structalias.c (struct variable_info): remove
field in_nested_struct.
(push_fields_onto_fieldstack): add argument
nesting_level. update all users.
update documentation.
update pair->nesting_level with nesting_level.
make recursive calls with nesting_level + 1.
(set_uids_in_ptset): if an sft is added to the points-to
set, mark it as unpartitionable.
* tree-ssa-operands.c (ref_nesting_level): new.
(add_vars_for_offset): call it.
add argument full_ref. update
callers.
if var is inside a nested structure and the nesting level
of full_ref is lower than the nesting level of var,
adjust offset by the offset of var.
testsuite/ChangeLog
PR 33870
* gcc.c-torture/execute/pr33870-1.c: New test.
* gcc.dg/tree-ssa/alias-16.c: New test.
Eric Botcazou [Mon, 12 Nov 2007 21:21:37 +0000 (22:21 +0100)]
re PR rtl-optimization/34035 (ICE in calc_dfs_tree with -O2 -fnon-call-exceptions -ffast-math -fno-gcse)
PR rtl-optimization/34035
* cse.c (cse_cfg_altered): New global variable.
(cse_jumps_altered): Make boolean.
(recorded_label_ref): Likewise.
(cse_insn): Adjust for above changes.
(cse_extended_basic_block): Likewise. Set cse_cfg_altered
if dead edges have been purged.
(cse_main): Change return value specification and adjust code.
(rest_of_handle_cse): Adjust for above change.
(rest_of_handle_cse2): Likewise.
* gcse.c (rest_of_handle_gcse): Likewise.
i386.md (sse5_setcc<mode>): Use <ssemodefsuffix> to get the appropriate suffix for the coms* instruction.
2007-11-12 Dwarakanath Rajagopal <dwarak.rajagopal@amd.com>
Michael Meissner <michael.meissner@amd.com>
* config/i386/i386.md (sse5_setcc<mode>): Use <ssemodefsuffix>
to get the appropriate suffix for the coms* instruction.
(sse5_pcmov_<mode>): Restrict operands of pcmov
for scalar case to be only xmm registers and not memory.
* config/i386/sse.md (sse5_pcmov_<mode>): Correct the operand
constraints to follow the mnemonics for the pcmov instruction
Co-Authored-By: Michael Meissner <michael.meissner@amd.com>
From-SVN: r130120
re PR target/34042 (Segfault in mips_cannot_change_mode_class)
gcc/
PR target/34042
* config/mips/mips.c (mips_return_fpr_single): New function.
(mips_function_value): Use it when returning single-field
aggregates in FPRs.
(mips_expand_call): Handle the PARALLELs created by
mips_return_fpr_single.
Richard Guenther [Mon, 12 Nov 2007 14:16:05 +0000 (14:16 +0000)]
re PR middle-end/34070 (Wrong code for (int)x%4)
2007-11-12 Richard Guenther <rguenther@suse.de>
PR middle-end/34070
* fold-const.c (fold_binary): If testing for non-negative
operands with tree_expr_nonnegative_warnv_p make sure to
use op0 which has all (sign) conversions retained.
* gcc.c-torture/execute/pr34070-1.c: New testcase.
* gcc.c-torture/execute/pr34070-2.c: Likewise.
Ira Rosen [Mon, 12 Nov 2007 09:32:13 +0000 (09:32 +0000)]
re PR tree-optimization/33953 (internal compiler error: vector VEC(tree,base) index domain error, in vectorizable_operation at tree-vect-transform.c:4017)
PR tree-optimization/33953
* tree-vect-transform.c (vectorizable_operation): In case of SLP,
allocate vec_oprnds1 according to the number of created vector
statements. In case of shift with scalar argument, store scalar operand
for every vector statement to be created for the SLP node. Fix a
comment.
re PR c++/8570 (line number in error message is badly placed with illegal template friend declaration)
2007-11-12 Manuel Lopez-Ibanez <manu@gcc.gnu.org>
PR c++/8570
cp/
* pt.c (redeclare_class_template): Update error message. Use a
note to show the previous declaration.
(tsubst_friend_class): Use the location of the friend template as
the input location before calling redeclare_class_template.
testsuite/
* g++.old-deja/g++.ns/template13.C: Update expected output.
* g++.old-deja/g++.pt/friend23.C: Likewise.
* g++.dg/warn/pr8570.C: New.
Jakub Jelinek [Sat, 10 Nov 2007 07:46:31 +0000 (08:46 +0100)]
re PR middle-end/34018 (ICE: verify_stmts failed)
PR middle-end/34018
* tree-inline.h (copy_body_data): Add regimplify field.
* tree-inline.c (copy_body_r): Set id->regimplify to true
if an TREE_INVARIANT ADDR_EXPR is no longer invariant after
substitutions.
(copy_bb): Clear id->regimplify before walk_tree, if it is
set afterwards, regimplify the whole statement.
Jakub Jelinek [Sat, 10 Nov 2007 07:36:09 +0000 (08:36 +0100)]
re PR c++/32241 (ICE trying to call x.~X(); in a template)
PR c++/32241
* pt.c (tsubst_copy_and_build) <case COMPONENT_REF>: If object_type
is not scalar type, let finish_class_member_access_expr handle
diagnostics. Pass BIT_NOT_EXPR argument to
finish_pseudo_destructor_expr. Handle SCOPE_REF properly.
* gcc/sched-deps.c (sched_analyze_insn): Use MOVE_BARRIER
instead of TRUE_BARRIER for jumps. Add register dependencies
even when reg_pending_barrier is set.
haifa-sched.c (haifa_classify_insn): Rename to ...
* gcc/haifa-sched.c (haifa_classify_insn): Rename to ...
(haifa_classify_rtx): ... this. Improve handling of COND_EXECs,
handle PARALLELs by recursing. Use it ...
(haifa_classify_insn): ... here. Reimplement.
* bitmap.h (bitmap_single_bit_set_p): Declare.
* bitmap.c (bitmap_single_bit_set_p): New function.
* tree-ssa-alias.c (add_may_alias_for_new_tag): Use it.
(maybe_create_global_var): Use bitmap_empty_p.
Eric Botcazou [Fri, 9 Nov 2007 12:49:55 +0000 (13:49 +0100)]
re PR rtl-optimization/33732 (gcc.c-torture/execute/longlong.c execution at -O3)
PR rtl-optimization/33732
* reload.c (push_reload): Check that the REG_DEAD note was referring
to a hardreg or to a pseudo that has been assigned exactly one hardreg
before considering it in order to select the reload register.
(combine_reloads): Likewise.
(find_dummy_reload): Likewise.
re PR target/33604 (Revision 119502 causes significantly slower results with 4.3/4.4 compared to 4.2)
2007-11-09 Richard Guenther <rguenther@suse.de>
PR tree-optimization/33604
* tree-ssa-forwprop.c (tree_ssa_forward_propagate_single_use_vars):
Disregard changes in CV qualifiers of pointed to types for
forward propagating ADDR_EXPRs.
* tree-ssa-ccp.c (fold_stmt_r): Preserve volatileness of the original
expression.
* g++.dg/tree-ssa/pr33604.C: New testcase.
* gcc.dg/pr32721.c: Adjust pattern.
dse.c (find_shift_sequence): Always choose an integer mode for new_mode.
gcc/
* dse.c (find_shift_sequence): Always choose an integer mode for
new_mode.
(replace_read): Require both the read and store mode to be
integer ones. Remove a then-redundant FLOAT_P check.
tree-dfa.c (remove_referenced_var): If removing a variable which has subvars...
2007-11-08 Richard Guenther <rguenther@suse.de>
* tree-dfa.c (remove_referenced_var): If removing a
variable which has subvars, also remove those from
the referenced vars. Do not create a variable annotation.
Bob Wilson [Thu, 8 Nov 2007 22:08:02 +0000 (22:08 +0000)]
xtensa.c (xtensa_expand_prologue): Remove first argument for gen_entry calls.
* config/xtensa/xtensa.c (xtensa_expand_prologue): Remove first
argument for gen_entry calls.
* config/xtensa/xtensa.md: Add new "entry" value to "type" attribute.
(entry): Use the new attribute value. Remove unused first operand.
PR middle-end/33826
* ipa-pure-const (static_execute): Added code to keep recursive
functions from being marked as pure or const.
* ipa-utils (searchc): Fixed comment.
2007-11-08 Kenneth Zadeck <zadeck@naturalbridge.com>
PR middle-end/33826
* gcc.dg/pr33826.c: New.
* gcc.dg/tree-ssa/20030714-1.c: Removed two tests that depend on
recursive functions being marked pure or const.