Tristan Gingold [Mon, 29 Oct 2012 10:52:28 +0000 (10:52 +0000)]
bindgen.adb (Gen_Output_File_Ada): Do not emit declaration for Is_Elaborated if not referenced.
2012-10-29 Tristan Gingold <gingold@adacore.com>
* bindgen.adb (Gen_Output_File_Ada): Do not emit declaration for
Is_Elaborated if not referenced.
2012-10-29 Tristan Gingold <gingold@adacore.com>
* exp_ch9.adb (Build_Activation_Chain_Entity): Punt in restricted
profile.
* exp_ch3.adb (Build_Initialization_Call): Do no append _Chain
parameter in restricted profile.
(Build_Init_Call_Thru): Likewise.
(Init_Formals): Likewise.
* exp_ch3.adb: Minor reformatting.
Arnaud Charlet [Mon, 29 Oct 2012 10:17:29 +0000 (11:17 +0100)]
[multiple changes]
2012-10-29 Ed Schonberg <schonberg@adacore.com>
* sem_aux.adb (Get_Rep_Item): Treat Priority and Interrupt_Priority
as equivalent, because only one of them can be specified for a
task, protected definition, or subprogram body.
* aspects.adb ((Same_Aspect): The canonical aspect of
Interrupt_Priority is Priority.
2012-10-29 Robert Dewar <dewar@adacore.com>
* sem_ch13.adb: Minor reformatting.
2012-10-29 Robert Dewar <dewar@adacore.com>
* i-cstrea.ads: Avoid redefinition of standard symbol string.
* prj-makr.adb: Add comment for OK redefinition of Stadard.
* prj.ads: Add comment for OK redefinition of Stadard.
* s-crtl.ads: Avoid redefinition of standard symbol string.
* sinfo-cn.adb (Change_Identifier_To_Defining_Identifier):
Generate warning for standard redefinition if
Warn_On_Standard_Definition set.
* usage.adb: Add lines for -gnatw.k and -gnatw.K
* warnsw.adb: Set/reset Warn_On_Standard_Redefinition
appropriately.
* warnsw.ads (Warn_On_Standard_Redefinition): New flag.
* s-stratt-xdr.adb: Avoid new warning.
2012-10-29 Ed Schonberg <schonberg@adacore.com>
* exp_dbug.ads, exp_dbug.adb (Build_Subprogram_Instance_Renamings):
in the body of a subpogram instance, introduce local renamings
for actuals of an elementary type, so that GDB can recover the
values of these actuals more directly.
Arnaud Charlet [Mon, 29 Oct 2012 10:07:33 +0000 (11:07 +0100)]
[multiple changes]
2012-10-29 Robert Dewar <dewar@adacore.com>
* sem_ch5.adb (Analyze_Loop_Statement): Add warning for identical
inner/outer ranges.
2012-10-29 Robert Dewar <dewar@adacore.com>
* einfo.ads: Change terminology "present" to "defined" in talking
about whether a given field is defined for a given entity kind.
2012-10-29 Bob Duff <duff@adacore.com>
* atree.ads: Minor comment fix.
2012-10-29 Bob Duff <duff@adacore.com>
* sem_ch13.adb (Replace_Type_Reference): Set_Comes_From_Source.
Otherwise, the node is ignored by ASIS.
* sem_ch5.adb: Minor reformatting.
2012-10-29 Thomas Quinot <quinot@adacore.com>
* exp_attr.adb, exp_dist.adb, exp_dist.ads (Build_To_Any_Call): Pass
an explicit Loc parameter to set the source location of generated
nodes.
2012-10-29 Tristan Gingold <gingold@adacore.com>
* exp_ch9.adb (Build_Task_Activation_Call): Do nothing on
restricted profile.
* bindgen.adb (System_Tasking_Restricted_Stages_Used): New variable.
(Gen_Adainit): Declare and call Activate_Tasks when the above variable
is set.
(Resolve_Binder_Options): Set the variable.
* rtsfind.ads (RE_Activate_Restricted_Tasks): Removed (now unused).
* s-tarest.adb (Tasks_Activation_Chain): New variable.
(Activate_Restricted_Tasks): Removed, and replaced by ...
(Activate_Tasks): New procedure, to activate all tasks at
the end of elaboration.
(Create_Restricted_Tasks): Chain parameter is now unreferenced. Put
the created task on the Tasks_Activation_Chain list.
* s-tarest.ads (Activate_Restricted_Tasks): Removed.
(Activate_Tasks): Added.
2012-10-29 Gary Dismukes <dismukes@adacore.com>
* sem_res.adb (Resolve_If_Expression): Compare subtype of the 'then'
expression against the subtype of the expression rather than comparing
base types, same as is already done for the 'else' expression.
Arnaud Charlet [Mon, 29 Oct 2012 10:02:16 +0000 (11:02 +0100)]
[multiple changes]
2012-10-29 Steve Baird <baird@adacore.com>
* sem_ch13.adb (Analyze_Attribute_Definition_Clause): If CodePeer_Mode
= True and we detect that an Address specification is used to overlay
another object, do not set that object's Treat_As_Volatile flag.
2012-10-29 Ed Schonberg <schonberg@adacore.com>
* sem_ch3.adb (Constant_Redeclaration): If the type of a
deferred constant has invariants, generate invariant call on
the initial value.
Arnaud Charlet [Mon, 29 Oct 2012 09:55:29 +0000 (10:55 +0100)]
[multiple changes]
2012-10-29 Yannick Moy <moy@adacore.com>
* checks.adb (Apply_Arithmetic_Overflow_Minimized_Eliminated):
Add special case for case expression alternative.
(Is_Signed_Integer_Arithmetic_Op): Remove special case for case
expression alternative.
* exp_ch4.adb Minor reformatting.
Uros Bizjak [Mon, 29 Oct 2012 08:03:47 +0000 (09:03 +0100)]
i386.c (ix86_decompose_address): Use simplify_gen_subreg to generate SImode equivalent of address...
* config/i386/i386.c (ix86_decompose_address): Use simplify_gen_subreg
to generate SImode equivalent of address, zero-extended with AND RTX.
* config/i386/i386.md (ashift to lea splitter): Split to SImode mult.
(simple lea to add/shift peephole2s): Remove peephole2s that operate
on subregs of DImode operations.
Vladimir Makarov [Mon, 29 Oct 2012 00:42:25 +0000 (00:42 +0000)]
re PR rtl-optimization/55106 (ice: Maximum number of LRA constraint passes is achieved (15))
2012-10-28 Vladimir Makarov <vmakarov@redhat.com>
PR rtl-optimization/55106
* lra-constraints.c (skip_usage_debug_insns): New function.
(check_secondary_memory_needed_p): Ditto.
(inherit_reload_reg): Use the new functions. Improve debug
output.
Vladimir Makarov [Sun, 28 Oct 2012 20:02:59 +0000 (20:02 +0000)]
rtl.h (struct rtx_def): Add a comment for member unchanging.
2012-10-28 Vladimir Makarov <vmakarov@redhat.com>
* rtl.h (struct rtx_def): Add a comment for member unchanging.
* lra-int.h (LRA_TEMP_CLOBBER_P): New macro.
(lra_hard_reg_substitution): Rename to lra_final_code_change.
* lra-constraints.c (match_reload): Mark temporary clobbers.
* lra-spill.c (lra_hard_reg_substitution): Rename to
lra_final_code_change. Remove temporary clobbers.
* lra.c (lra): Rename to lra_final_code_change.
Jonathan Wakely [Sun, 28 Oct 2012 13:20:31 +0000 (13:20 +0000)]
re PR libstdc++/55041 (prettyprinting/shared_ptr & cxx11 fails on some platforms)
PR libstdc++/55041
* python/libstdcxx/v6/printers.py (Tr1UnorderedMapPrinter): Update
to handle hashtable as member of unordered_map not base class.
(Tr1UnorderedSetPrinter): Likewise.
David S. Miller [Sat, 27 Oct 2012 02:57:50 +0000 (02:57 +0000)]
Update sparc constraint comments and documentation.
* config/sparc/constraints.md: Update unused letter list, move
"w" near other memory constraints. Remove no longer relevant
comment.
* doc/md.texi: Sync sparc constraint documentation with reality.
Ian Lance Taylor [Fri, 26 Oct 2012 20:08:29 +0000 (20:08 +0000)]
re PR other/55087 (bogus "linux-vdso.so.1: No such file or directory" caused by libbacktrace)
PR other/55087
* posix.c (backtrace_open): Add does_not_exist parameter.
* elf.c (phdr_callback): Do not warn if shared library could not
be opened.
* fileline.c (fileline_initialize): Update calls to
backtrace_open.
* internal.h (backtrace_open): Update declaration.
Jakub Jelinek [Fri, 26 Oct 2012 19:19:25 +0000 (21:19 +0200)]
re PR debug/54970 (Missing DW_OP_GNU_implicit_pointer in debuginfo)
PR debug/54970
* cfgexpand.c (expand_debug_expr): Expand &MEM_REF[&var, n]
as DEBUG_IMPLICIT_PTR + n if &var expands to DEBUG_IMPLICIT_PTR.
* tree-sra.c (create_access_replacement): Allow also MEM_REFs
with ADDR_EXPR first operand in DECL_DEBUG_EXPR expressions.
* var-tracking.c (track_expr_p): Handle MEM_REFs in DECL_DEBUG_EXPR
expressions.
* dwarf2out.c (add_var_loc_to_decl): Likewise.
PR debug/54971
* gcc.dg/guality/pr54970.c: New test.
Uros Bizjak [Fri, 26 Oct 2012 17:54:50 +0000 (19:54 +0200)]
crtfastmath.c (set_fast_math): Use __builtin_ia32_fxsave.
* config/i386/crtfastmath.c (set_fast_math): Use __builtin_ia32_fxsave.
Clear only fxsave.mxcsr_mask. Use saved mxcsr from fxsave structure
when appropriate. Correct structure element types.
* config/i386/t-crtfm (crtfastmath.o): Compile with -mfxsr, remove
-minline-all-stringops from compile flags.
Ian Lance Taylor [Fri, 26 Oct 2012 17:50:10 +0000 (17:50 +0000)]
syscall: fix creds_test to reliably close os.File
Uncovered by Uros Bizjak.
Before this patch the test would close the file descriptor but
not the os.File. When the os.File was GC'ed, the finalizer
would close the file descriptor again. That would cause
problems if the same file descriptor were returned by a later
call to open in another test.
On my system:
> GOGC=30 go test
--- FAIL: TestPassFD (0.04 seconds)
passfd_test.go:62: FileConn: dup: bad file descriptor
FAIL
Teresa Johnson [Fri, 26 Oct 2012 17:19:35 +0000 (17:19 +0000)]
Avoid conservative behavior in REE by allowing removal of redundant extends when...
Avoid conservative behavior in REE by allowing removal of redundant extends
when the def feeds another extend with a different mode. This works because
in merge_def_and_ext only calls combine_set_extension if the candidate for
removal has a wider mode than the def extend's mode, otherwise the def extend
mode is preserved. In combine_set_extension the def is modified to use the
wider candidate's mode.
2012-10-26 Teresa Johnson <tejohnson@google.com>
* ree.c (add_removable_extension): Remove unnecessary
mode check with other extension.
* testsuite/gcc.c-torture/execute/20111227-2.c: New test.
* testsuite/gcc.c-torture/execute/20111227-3.c: Ditto.
Martin Jambor [Fri, 26 Oct 2012 16:13:00 +0000 (18:13 +0200)]
re PR debug/54971 (SRA pessimizes debug info by not creating debug stmts for fields without replacements)
2012-10-26 Martin Jambor <mjambor@suse.cz>
PR debug/54971
* tree-sra.c (struct access): New flag grp_to_be_debug_replaced.
(dump_access): Dump the new flag.
(analyze_access_subtree): Set the new flag when appropriate.
(create_access_replacement): Handle debug replacements differently.
(generate_subtree_copies): Handle the grp_to_be_debug_replaced flag.
(init_subtree_with_zero): Likewise.
(sra_modify_expr): Likewise.
(load_assign_lhs_subreplacements): Likewise.
(sra_modify_assign): Likewise.
Yuri Rumyantsev [Fri, 26 Oct 2012 12:07:58 +0000 (12:07 +0000)]
i386.c (insn_is_function_arg): Add check on CALL instruction.
* config/i386/i386.c (insn_is_function_arg) : Add check on CALL
instruction.
(ix86_dependencies_evaluation_hook): Insert dependencies in all
predecessors of call block for non-trivial region avoiding creation
of loop-carried dependency to avoid cross-block motion of HW registers.
Lawrence Crowl [Thu, 25 Oct 2012 21:45:28 +0000 (21:45 +0000)]
Change hash_table to support a comparator type different from the value type stored in the hash table.
Change hash_table to support a comparator type different from the
value type stored in the hash table. The 'find' functions now may
take a different type from the value type. This requires introducing
a second typedef into the Descriptor conceptual type. Change the
Descriptor concept to use typedefs value_type and compare_type instead
of T. Change all users to match.
Add usage documentation to hash-table.h.
Tested on x86-64.
Index: gcc/ChangeLog
2012-10-25 Lawrence Crowl <crowl@google.com>
* hash-table.h: Add usage documentation.
(template struct typed_free_remove): Clarify documentation.
Rename template parameter.
(struct typed_noop_remove): Likewise.
(descriptor concept): Change typedef T to value_type.
Add typedef compare_type. Use more precise template parameter name,
Descriptor instead of Descr. Update users to match.
(struct hash_table): Change 'find' parameters to use compare_type
instead of the value type.
Jan Hubicka [Thu, 25 Oct 2012 19:23:15 +0000 (21:23 +0200)]
ipa-cp.c (ipcp_discover_new_direct_edges): If something was turned to direct call update the summary.
* ipa-cp.c (ipcp_discover_new_direct_edges): If something was turned
to direct call update the summary.
* ipa-inline-transform.c (inline_call): Sanity check that summaries
match the predicted effect; fix updating of summary after edge
redirection.
* ipa-inline-analysis.c (inline_node_duplication_hook): Do not try
to update the summary and recompute it instead.
(estimate_function_body_sizes): Fix self size estimation; double
check that it agrees with inline_update_overall_summary.
(estimate_edge_size_and_time): Handle devirtualizaiton costs.
(estimate_edge_devirt_benefit): Update to be called from
estimate_edge_size_and_time.
(estimate_calls_size_and_time): Update.
(estimate_node_size_and_time): Watch overflows.
(inline_merge_summary): Likewise.
* ipa-prob.c: Include ipa-inline.h
(ipa_make_edge_direct_to_target): After redirection update the summary.
Paolo Carlini [Thu, 25 Oct 2012 16:14:59 +0000 (16:14 +0000)]
re PR c++/53761 (ICE on incorrect transparent union (first field has floating point type))
/cp
2012-10-25 Paolo Carlini <paolo.carlini@oracle.com>
PR c++/53761
* class.c (finish_struct_1): Reject aggregates decorated with
__transparent_union__ which cannot be made transparent because
the type of the first field has a different ABI from the class
overall.
/testsuite
2012-10-25 Paolo Carlini <paolo.carlini@oracle.com>
Jason Merrill [Thu, 25 Oct 2012 15:54:00 +0000 (11:54 -0400)]
Core 1402
Core 1402
cp/
* call.c (joust): An implicitly deleted move function is
worse than any non-deleted function.
* method.c (process_subob_fn): No special rules for move.
(synthesized_method_walk, implicitly_declare_fn): Likewise.
Warn about virtual base with non-trivial move assignment.
* cp-tree.h (struct lang_decl_fn): Remove suppress_implicit_decl.
(FNDECL_SUPPRESS_IMPLICIT_DECL): Remove.
c-family/
* c.opt (Wvirtual-move-assign): New.
Jan Hubicka [Thu, 25 Oct 2012 14:52:32 +0000 (16:52 +0200)]
re PR lto/54980 (gimple check: expected gimple_cond(error_mark), have gimple_call() in gimple_cond_set_lhs, at gimple.h:2578)
PR tree-optimize/54980
* tree-ssa-loop-ivcanon.c (constant_after_peeling): Fix
obvious typo.
(loop_edge_to_cancel): Be sure that the edge is from an conditional
so we can cancel it.
Paolo Carlini [Thu, 25 Oct 2012 09:12:22 +0000 (09:12 +0000)]
re PR c++/34892 (ICE with ellipsis in default template argument)
/cp
2012-10-25 Paolo Carlini <paolo.carlini@oracle.com>
PR c++/34892
* parser.c (cp_parser_template_parameter): When
cp_parser_parameter_declaration parsed a default argument don't
see if *is_parameter_pack needs setting.
/testsuite
2012-10-25 Paolo Carlini <paolo.carlini@oracle.com>