Commit Graph

105928 Commits

Author SHA1 Message Date
Ian Lance Taylor
9a270ad3c0 Add cgo support routines corresponding to the ones in 6g/8g.
From-SVN: r168679
2011-01-11 18:59:30 +00:00
Richard Henderson
72d6e3c594 mn10300: Clean up costing.
Address, register, memory and rtx costs bore little relation to reality.

From-SVN: r168678
2011-01-11 10:52:43 -08:00
Richard Henderson
62d3f9b2ef Oops, forgot predicates.md line in last Changelog.
From-SVN: r168677
2011-01-11 10:40:19 -08:00
Richard Henderson
df37c0236d mn10300: Remove bset/bclr patterns.
These instructions do not use normal addressing modes and are
incorrectly implemented for that.  Corrections to legitimate
addresses expose problems here.

Delete them for now.  To be re-instated later in the form of
atomic operation builtins.

From-SVN: r168676
2011-01-11 10:38:48 -08:00
Richard Henderson
85a337b58a mn10300: Define the A and D constraints.
This will allow combining am33 and mn103 alternatives
without having to resort to the enabled attribute.

The existing 'A' constraint renamed to 'c'.  Thankfully
this existing accumulator constraint doesn't appear in
either newlib or eglibc sources.

From-SVN: r168675
2011-01-11 10:28:59 -08:00
Richard Henderson
a5c727e3ba mn10300: Add attribute enabled.
This will allow merging am33 and mn103 patterns for which the
set of alternatives can't be merged via constraint letters.

From-SVN: r168674
2011-01-11 10:26:16 -08:00
Richard Henderson
0d9e23f4ee mn10300: fp insn cleanup
Delete integer-mode abssf2, negsf2; these will be handled
by the middle-end now.  Delete unnecessary expanders.

From-SVN: r168673
2011-01-11 10:24:46 -08:00
Richard Henderson
c157b3f0d2 mn10300: Fix debug offsets into the stack frame
We were using debugging hooks to semi-correct a mistake
in the lack of ARG_POINTER_CFA_OFFSET.

From-SVN: r168672
2011-01-11 10:20:37 -08:00
Richard Henderson
cc909bba64 mn10300: Emit the movm stores in the correct order.
From-SVN: r168671
2011-01-11 10:17:25 -08:00
Richard Henderson
e3b5c2f3ea mn10300: Delete ASM_PN_FORMAT.
Is there any real reason for this?  It causes testsuite failures
because the pattern doesn't match many of the dumps.

From-SVN: r168670
2011-01-11 10:13:47 -08:00
Richard Henderson
e10ede2cf3 mn10300: Disable test tree-ssa/vrp47.c
From-SVN: r168669
2011-01-11 10:12:40 -08:00
Richard Henderson
885fe07c5c mn10300: Better definition of INCOMING_RETURN_ADDR_RTX.
The new definition is more useful to the unwinder.

From-SVN: r168668
2011-01-11 10:11:07 -08:00
Jan Hubicka
7bc6263756 Fix incorrectly applied patch.
From-SVN: r168667
2011-01-11 17:33:52 +00:00
Jan Hubicka
6e70182270 re PR lto/45721 (ICE: in function_and_variable_visibility, at ipa.c:673 with -flto)
PR lto/45721
	PR lto/45375
	* tree.h (symbol_alias_set_t): Move typedef here from varasm.c
	(symbol_alias_set_destroy, symbol_alias_set_contains,
	propagate_aliases_backward): Declare.
	* lto-streamer-out.c (struct sets): New sturcture.
	(trivally_defined_alias): New function.
	(output_alias_pair_p): Rewrite.
	(output_unreferenced_globals): Fix output of alias pairs.
	(produce_symtab): Likewise.
	* ipa.c (function_and_variable_visibility): Set weak alias destination
	as needed in lto.
	* varasm.c (symbol_alias_set_t): Remove.
	(symbol_alias_set_destroy): Export.
	(propagate_aliases_forward, propagate_aliases_backward): New functions
	based on ...
	(compute_visible_aliases): ... this one; remove.
	(trivially_visible_alias): New
	(trivially_defined_alias): New.
	(remove_unreachable_alias_pairs): Rewrite.
	(finish_aliases_1): Reorganize code checking if alias is defined.
	* passes.c (rest_of_decl_compilation): Do not call assemble_alias when
	in LTO mode.

	* lto.c (partition_cgraph_node_p, partition_varpool_node_p): Weakrefs are
	not partitioned.

	* testsuite/gcc.dg/lto/pr45721_1.c: New file.
	* testsuite/gcc.dg/lto/pr45721_0.c: New file.

From-SVN: r168666
2011-01-11 17:29:52 +00:00
Richard Guenther
1b24890758 re PR tree-optimization/46076 (constant propagation and compile-time math no longer happening versus 4.4 and 4.5)
2011-01-11  Richard Guenther  <rguenther@suse.de>

	PR tree-optimization/46076
	* tree-ssa.c (useless_type_conversion_p): Conversions from
	unprototyped to empty argument list function types are useless.

	* gcc.dg/tree-ssa/pr46076.c: New testcase.

From-SVN: r168665
2011-01-11 17:01:37 +00:00
Paolo Carlini
403c753b6f aclocal.m4: Regenerate.
2011-01-11  Paolo Carlini  <paolo.carlini@oracle.com>

	* aclocal.m4: Regenerate.

From-SVN: r168664
2011-01-11 16:43:25 +00:00
Richard Guenther
b1923f0a36 re PR rtl-optimization/45235 (const volatile read moved out of order)
2011-01-11  Richard Guenther  <rguenther@suse.de>

	PR middle-end/45235
	* emit-rtl.c (set_mem_attributes_minus_bitpos): Do not mark
	volatile MEMs as MEM_READONLY_P.

From-SVN: r168663
2011-01-11 15:41:17 +00:00
Kai Tietz
d504eedc67 re PR bootstrap/47215 (Failed to bootstrap)
2011-01-11  Kai Tietz  <kai.tietz@onevision.com>

        PR bootstrap/47215
        * decl.c (java_init_decl_processing): Initialize
        long_integer_type_node.

From-SVN: r168662
2011-01-11 15:51:07 +01:00
Richard Guenther
e98965af89 re PR tree-optimization/47239 ((int)&func & 3 is always optimized to 0 on some targets)
2011-01-11  Richard Guenther  <rguenther@suse.de>

	PR tree-optimization/47239
	* tree-ssa-ccp.c (get_value_from_alignment): Punt for FUNCTION_DECLs.

From-SVN: r168661
2011-01-11 14:23:20 +00:00
Jeff Law
9f9ca914cc re PR tree-optimization/47086 (ICE: verify_flow_info failed: BB 3 can not throw but has an EH edge with -O -fexceptions -fnon-call-exceptions -ftrapv)
* PR tree-optimization/47086
	* tree-ssa-loop-ivopts.c (find_givs_in_stmt_scev): Do not record
	IVs from statements that might throw.

	* PR tree-optimization/47086
	* gcc.dg/pr47086.c: New test.

From-SVN: r168659
2011-01-11 07:10:54 -07:00
Jason Merrill
a67e7daab5 re PR c++/46658 (internal compiler error on cp/init.c)
PR c++/46658
	* init.c (build_new_1): Handle value-init in templates differently.

From-SVN: r168655
2011-01-11 06:30:07 -05:00
Jason Merrill
41d04a8dfe re PR c++/45520 ([C++0x] compiler segmentation fault on decltype in lambda-declarator)
PR c++/45520
	* tree.c (maybe_dummy_object): Check current_class_ref against
	context, not current_class_type.

From-SVN: r168654
2011-01-11 06:29:58 -05:00
Iain Sandoe
5826c8d29e next-mapping.h: Add copyright header.
gcc/testsuite:

	* objc-obj-c++-shared/next-mapping.h: Add copyright header.
	Split type definitions and NSConstantString class into...
	* objc-obj-c++-shared/objc-test-suite-types.h: New.
	* objc-obj-c++-shared/nsconstantstring-class.h: New.
	* objc-obj-c++-shared/nsconstantstring-class-impl.h: New.
	* objc-obj-c++-shared/nsconstantstring-class-impl.m: New.
	* objc-obj-c++-shared/nsconstantstring-class-impl.mm: New.
	* objc-obj-c++-shared/next-abi.h: Add copyright header, minor update
	to comments.
	* objc-obj-c++-shared/objc-test-suite-next-encode-assist.h: Adjust
	encode defines for newer NeXT system headers.
	* obj-c++.dg/template-4.mm: Adjust to use nsconstantstring headers.
	* obj-c++.dg/torture/strings/string1.mm: Likewise.
	* objc.dg/foreach-5.m: Likewise.
	* objc.dg/foreach-2.m: Likewise.
	* objc.dg/foreach-4.m: Likewise.
	* objc.dg/torture/strings/string1.m: Likewise.
	* objc.dg/torture/strings/string2.m: Likewise.
	* objc.dg/torture/strings/string3.m: Likewise.
	* objc.dg/torture/strings/string4.m: Likewise.
	* obj-c++.dg/torture/strings/const-str-3.mm: Adjust to use type
	header and stand alone.
	* obj-c++.dg/strings/strings-2.mm: Likewise.
	* objc.dg/torture/strings/const-str-3.m: Likewise.
	* objc.dg/strings/strings-2.m: Likewise.
	* objc.dg/strings/const-str-12b.m: Likewise.

From-SVN: r168653
2011-01-11 10:14:48 +00:00
Richard Guenther
2aa9c6ae42 re PR c/32511 (GCC rejects inline+weak function)
2011-01-12  Richard Guenther  <rguenther@suse.de>

	PR middle-end/32511
	* c-common.c (handle_weak_attribute): Warn instead of error
	on declaring an inline function weak.

	* gcc.dg/attr-weak-1.c: Adjust.

From-SVN: r168652
2011-01-11 10:09:15 +00:00
Paul Thomas
93c3bf479d re PR fortran/47051 (Wrong reallocate)
2011-01-11  Paul Thomas  <pault@gcc.gnu.org>

	PR fortran/47051
	* trans-array.c (gfc_alloc_allocatable_for_assignment): Change
	to be standard compliant by testing for shape rather than size
	before skipping reallocation. Improve comments.

2011-01-11  Paul Thomas  <pault@gcc.gnu.org>

	PR fortran/47051
	* gfortran.dg/realloc_on_assign_2.f03 : Modify 'test1' to be
	standard compliant and comment.

From-SVN: r168650
2011-01-11 05:19:20 +00:00
GCC Administrator
b7e945c8e7 Daily bump.
From-SVN: r168649
2011-01-11 00:18:20 +00:00
Jan Hubicka
3e76bf2c82 bootstrap-lto.mk: -fuse-linker-plugin is default now; pass -fno-lto to STAGEprofile.
* config/bootstrap-lto.mk: -fuse-linker-plugin is default now;
	pass -fno-lto to STAGEprofile.

From-SVN: r168645
2011-01-10 23:43:54 +00:00
Jan Hubicka
345fdcb710 re PR lto/45375 ([meta-bug] Issues with building Mozilla (i.e. Firefox) with LTO)
PR lto/45375
	* lto-cgraph.c (input_profile_summary): Remove overactive sanity check.

From-SVN: r168644
2011-01-10 23:37:45 +00:00
Jan Hubicka
c459c97b48 re PR lto/45375 ([meta-bug] Issues with building Mozilla (i.e. Firefox) with LTO)
PR lto/45375
	* profile.c (read_profile_edge_counts): Ignore profile inconistency
	when correcting profile.

From-SVN: r168643
2011-01-10 23:37:11 +00:00
Jan Hubicka
06c9eb5136 re PR lto/46083 (gcc.dg/initpri1.c FAILs with -flto/-fwhopr (attribute constructor/destructor doesn't work))
PR lto/46083
	* lto-streamer-out.c (pack_ts_function_decl_value_fields): Store
	DECL_FINI_PRIORITY.
	* lto-streamer-in.c (unpack_ts_function_decl_value_fields):
	Restore DECL_FINI_PRIORITY.
	* gcc.dg/initpri3.c: New testcase.

From-SVN: r168642
2011-01-10 21:54:33 +00:00
H.J. Lu
b88e4ef16c Add __attribute__((externally_visible)).
2011-01-10  H.J. Lu  <hongjiu.lu@intel.com>

	PR lto/47222
	* g++.dg/torture/stackalign/test-unwind.h (g_edi): Mark it
	externally visible.
	(g_esi): Likewise.
	(g_ebx): Likewise.
	(g_ebp): Likewise.
	(g_esp): Likewise.
	(g_ebp_save): Likewise.
	(g_esp_save): Likewise.

From-SVN: r168640
2011-01-10 11:58:57 -08:00
Mike Stump
75b023da28 re PR target/45258 (linkage on -ldl, -lm and -lpthread should be purged from darwin build)
2010-08-17  Jack Howarth  <howarth@bromo.med.uc.edu>
 
	PR target/45258
 	* gcc.c (spec_function): Add remove-outfile.
 	(remove_outfile_spec_function): New function.
 	* config/darwin.h (LINK_SPEC): Add removal of -ldl, -lm and -lpthread.
	* invoke.texi (replace-outfile): Document.

From-SVN: r168639
2011-01-10 19:41:39 +00:00
Rainer Orth
91a22316db * g++.dg/ipa/pr46984.C: Add dg-require-effective-target lto.
From-SVN: r168638
2011-01-10 19:02:28 +00:00
Ralf Wildenhues
ae9fd815db doc: fix deftypefn markup in gccint manual.
gcc/:
	* doc/gimple.texi: Fix quoting of multi-word return values in
	@deftypefn statements.  Ensure presence of return value.  Wrap
	overlong @deftypefn lines.
	(is_gimple_operand, is_gimple_min_invariant_address): Remove
	descriptions of removed functions.
	* doc/hostconfig.texi (Host Common): Wrap long line, fix quoting
	of multi-word return value in @deftypefn statement.

From-SVN: r168637
2011-01-10 19:00:10 +00:00
Ralf Wildenhues
0ecb4a7c5e doc: avoid extra indentation in gimple.texi
gcc/:
	* doc/gimple.texi (Temporaries, Operands, Compound Lvalues)
	(Conditional Expressions, Logical Operators)
	(Statement and operand traversals): Do not indent smallexample
	code.  Fix duplicate function argument in example.

From-SVN: r168636
2011-01-10 18:59:56 +00:00
Rainer Orth
3c2e63dea1 re PR tree-optimization/46021 (3 tree-ssa tests XPASS almost everywhere)
PR tree-optimization/46021
	* gcc.dg/tree-ssa/20040204-1.c: Don't XFAIL on alpha*-*-*,
	i?86-*-*, x86_64-*-*.

From-SVN: r168635
2011-01-10 18:19:45 +00:00
Jeff Law
2e5e346da2 re PR tree-optimization/47141 (segfault)
* PR tree-optimization/47141
	* ipa-split.c (split_function): Handle case where we are returning a
	value and the return block has a virtual operand phi.

	* gcc.c-torture/compile/pr47141.c: New test.

Approved by richie in IRC

From-SVN: r168634
2011-01-10 09:48:42 -07:00
Jan Hubicka
cf9712ccc0 re PR tree-optimization/47234 (ipa-split is executed before profile feedback is read)
PR tree-optimization/47234 
	* tree-pass.h (TODO_rebuild_cgraph_edges): New TODO.
	(pass_feedback_split_functions): Declare.
	* passes.c (init_optimization_passes): Add ipa-split as subpass of
	tree-profile.
	* ipa-split.c (gate_split_functions): Update comments; disable
	split-functions for profile_arc_flag and branch_probabilities.
	(gate_feedback_split_functions): New function.
	(execute_feedback_split_functions): New function.
	(pass_feedback_split_functions): New global var.

From-SVN: r168632
2011-01-10 14:33:04 +00:00
H.J. Lu
94cd932cae Check e->call_stmt before calling gimple_call_set_cannot_inline.
2011-01-10  H.J. Lu  <hongjiu.lu@intel.com>

	PR lto/46760
	* tree-inline.c (tree_can_inline_p): Check e->call_stmt before
	calling gimple_call_set_cannot_inline.

From-SVN: r168631
2011-01-10 05:24:27 -08:00
Joseph Myers
18c9fa1720 * ja.po: Update.
From-SVN: r168630
2011-01-10 13:19:40 +00:00
Eric Botcazou
112578854e re PR testsuite/46230 (several vectorization tests FAIL or XPASS on SPARC)
PR testsuite/46230
	* gcc.dg/vect/vect-vfa-slp.c: Remove XFAIL.
	* gcc.dg/vect/slp-23.c: Remove XFAIL for SLP.
	* gcc.dg/vect/slp-35.c: Remove XFAILs.
	* gcc.dg/vect/no-tree-pre-slp-29.c: Likewise.

From-SVN: r168629
2011-01-10 12:51:00 +00:00
Eric Botcazou
81d8de64f3 re PR testsuite/46230 (several vectorization tests FAIL or XPASS on SPARC)
PR testsuite/46230
	* gcc.dg/vect/vect-peel-1.c (main): Prevent initialization loop from
	being vectorized.  Adjust dg-final pattern.  Remove XFAIL.
	* gcc.dg/vect/vect-peel-2.c (main): Likewise.
	* gcc.dg/vect/vect-peel-4.c (main): Prevent initialization loop from
	being vectorized.  Adjust dg-final pattern.

From-SVN: r168628
2011-01-10 12:41:40 +00:00
Eric Botcazou
af08e73446 re PR testsuite/46230 (several vectorization tests FAIL or XPASS on SPARC)
PR testsuite/46230
	* gcc.dg/vect/pr33804.c: XFAIL only for ilp32.
	* gcc.dg/vect/slp-24.c: Likewise.

From-SVN: r168627
2011-01-10 12:11:59 +00:00
Iain Sandoe
fbbfcaf185 darwin-sections.def: Remove unused section.
gcc:

	* config/darwin-sections.def: Remove unused section.

From-SVN: r168626
2011-01-10 08:41:58 +00:00
Dave Korn
d06865bf07 re PR c++/47218 (C++ multiple definitions of non-virtual thunk problem)
gcc/ChangeLog:

	PR c++/47218
	* cgraphunit.c (assemble_thunk): Call resolve_unique_section.

gcc/testsuite/ChangeLog:

	PR c++/47218
	* g++.dg/other/pr47218-1.C: New test file.
	* g++.dg/other/pr47218.C: Likewise.
	* g++.dg/other/pr47218.h: New supporting header.

From-SVN: r168624
2011-01-10 00:33:32 +00:00
GCC Administrator
637ebedd38 Daily bump.
From-SVN: r168623
2011-01-10 00:18:33 +00:00
Nicola Pero
96bbfbacda In gcc/: 2011-01-09 Nicola Pero <nicola.pero@meta-innovation.com>
In gcc/:
2011-01-09  Nicola Pero  <nicola.pero@meta-innovation.com>

	PR objc/47232
	* c-parser.c (c_parser_declaration_or_fndef): Improved
	error message.

In gcc/testsuite/:
2011-01-09  Nicola Pero  <nicola.pero@meta-innovation.com>

	PR objc/47232
	* objc.dg/attributes/invalid-attribute-1.m: New.
	* obj-c++.dg/attributes/invalid-attribute-1.mm: New.

From-SVN: r168619
2011-01-09 22:47:22 +00:00
Janus Weil
8977b4bb0d re PR fortran/47224 ([F03] ICE with procedure pointer component)
2011-01-09  Janus Weil  <janus@gcc.gnu.org>

	PR fortran/47224
	* resolve.c (resolve_actual_arglist): Remove unneeded and buggy piece
	of code.


2011-01-09  Janus Weil  <janus@gcc.gnu.org>

	PR fortran/47224
	* gfortran.dg/proc_ptr_comp_28.f90: New.

From-SVN: r168618
2011-01-09 22:01:38 +01:00
Iain Sandoe
20ad868865 gnu-api-2-class.mm: Skip for Darwin < 9.
gcc/testsuite:

	* obj-c++.dg/gnu-api-2-class.mm: Skip for Darwin < 9.
	* obj-c++.dg/gnu-api-2-ivar.mm: Likewise.
	* obj-c++.dg/gnu-api-2-method.mm: Likewise.
	* obj-c++.dg/gnu-api-2-objc.mm: Likewise.
	* obj-c++.dg/gnu-api-2-object.mm: Likewise.
	* obj-c++.dg/gnu-api-2-property.mm: Likewise.
	* obj-c++.dg/gnu-api-2-protocol.mm: Likewise.
	* obj-c++.dg/gnu-api-2-sel.mm: Likewise.
	* objc.dg/gnu-api-2-class.m: Likewise.
	* objc.dg/gnu-api-2-ivar.m: Likewise.
	* objc.dg/gnu-api-2-method.m: Likewise.
	* objc.dg/gnu-api-2-objc.m: Likewise.
	* objc.dg/gnu-api-2-object.m: Likewise.
	* objc.dg/gnu-api-2-property.m: Likewise.
	* objc.dg/gnu-api-2-protocol.m: Likewise.
	* objc.dg/gnu-api-2-sel.m: Likewise.

From-SVN: r168617
2011-01-09 16:59:00 +00:00
Iain Sandoe
5471aef02d foreach-1.m: Skip for Darwin < 9.
gcc/testsuite:

	* objc.dg/foreach-1.m: Skip for Darwin < 9.
	* objc.dg/objc-foreach-4.m: Likewise.
	* objc.dg/objc-foreach-5.m: Skip for Darwin < 9, adjust headers.
	* objc.dg/foreach-2.m: Skip for Darwin < 9, return self from init.
	* objc.dg/foreach-4.m: Likewise.
	* objc.dg/foreach-5.m: Likewise.
	* objc.dg/foreach-3.m: Return self from init.

From-SVN: r168616
2011-01-09 16:24:46 +00:00