On Sat, 17 Dec 2011, Martin Jambor wrote:
> Hi,
>
> when I was fixing PR 50622, somehow I lost a hunk in the patch along
> the way. I punished myself by making up an ICEing testcase which is
> included alongside the fix below. The fix is needed for trunk and
> also in the 4.6 branch.
>
> I hav
> What do you mean no tests require it? For instance, all of the ones
> that currently pass with with vect_perm?
Current implementation of vect_perm doesn't check for SSSE3 - so any
x86 target is supposed to support permutation.
> Just leave vect_perm alone for now. That may not be absolutely
>
On Fri, Dec 16, 2011 at 6:45 PM, Jakub Jelinek wrote:
> Hi!
>
> This testcase apparently fails on 32-bit hppa (both hpux and linux),
> because the struct there is passed by callee copied reference
> (unlike most other targets or even 64-bit hppa).
> So "Will be eliminated" is printed 3 times inste
On Sun, Dec 18, 2011 at 1:24 PM, Andreas Schwab wrote:
> Since gcc silently accepts any -Wno-* flag on purpose, it is ineffective
> to check for support of a warning flag using the negative form. Instead
> always use the positive form when running the check (but keep the
> original spelling for t
The cas insns are among the few insns from the 68020 isa that are not
supported by cpu32.
Andreas.
2011-12-19 Andreas Schwab
PR target/51532
* config/m68k/m68k.h (FL_CAS, TARGET_CAS): Define.
* config/m68k/m68k.c (FL_FOR_isa_20): Add FL_CAS.
* config/m68k/sync.
Joseph S. Myers wrote:
> On Sun, 18 Dec 2011, Georg-Johann Lay wrote:
>
>>> This new file needs to have the standard copyright and license notices.
>>> It's desirable to generate such notices in the output files as well.
>> What is the right copyright for the generated files?
>
> See other exampl
This fixes another case of PR51572 - we need to properly stream
TYPE_DECLs in TYPE_FIELDS.
LTO Boostrap and regtest running on x86_64-unknown-linux-gnu,
SPEC 2k6 build scheduled.
Richard.
2011-12-19 Richard Guenther
PR lto/51572
* tree.c (free_lang_data_in_type): Do not unli
On 12/14/11 13:02, Henderson, Stuart wrote:
>> bfin - REGRESSION - ICE -
>> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51003
>
> This looks like a known, general issue which Bernd had a fix for, but it
> doesn't appear to have been checked in yet:
> http://gcc.gnu.org/ml/gcc-patches/2011-11/ms
On Wed, Dec 14, 2011 at 05:14:18PM -0500, Jason Merrill wrote:
> --- a/gcc/c-family/c-common.c
> +++ b/gcc/c-family/c-common.c
> @@ -6286,13 +6286,21 @@ handle_transparent_union_attribute (tree *node, tree
> name,
>
>if (TREE_CODE (type) == UNION_TYPE)
> {
> - /* When IN_PLACE is s
Hi!
When I've switched STMT_VINFO_PATTERN_DEF_STMT into
a gimple_seq STMT_VINFO_PATTERN_DEF_SEQ, I've been using
gimplify_seq_add_stmt and/or gimple_seq_alloc_with_stmt
to populate the sequences. While the former is right
for this case and doesn't mark the stmt being added for updating,
the latte
Hi!
This patch renames gimplify_seq_add_stmt to
gimple_seq_add_stmt_without_update so that it is more clear what
it does, keeping the old name because it is shorter inside
of gimplify.c to avoid having to reformat it.
Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk?
2011-12-1
Hi!
cxx_eval_vec_init_1 starts with:
if (TREE_CODE (elttype) == ARRAY_TYPE)
/* We only do this at the lowest level. */;
else if (value_init)
{
init = something;
}
else if (!init)
{
init = something_else;
}
For the ARRAY_TYPE elttype (i.e. multidimensional ar
Hi!
We were ignoring maybe_clean_or_replace_eh_stmt return value,
which tells us if stmt previously could throw but doesn't any longer.
I was a little bit worried that gimple_purge_dead_eh_edges from within
cfg cleanup wouldn't work well, but at least bootstrap/regtested
on x86_64-linux and i686-l
This fixes one old and annyoing bug wrt LTO and debug-information.
We pre-seed the LTO tree streamer cache with nodes from global_trees[]
and friends, but this causes the type variant chain to become corrupt
(in the sense that we do not merge TYPE_DECLs, so a get_qualified_type
will fail to find a
On Mon, 19 Dec 2011, Jakub Jelinek wrote:
> Hi!
>
> When I've switched STMT_VINFO_PATTERN_DEF_STMT into
> a gimple_seq STMT_VINFO_PATTERN_DEF_SEQ, I've been using
> gimplify_seq_add_stmt and/or gimple_seq_alloc_with_stmt
> to populate the sequences. While the former is right
> for this case and
On Mon, 19 Dec 2011, Jakub Jelinek wrote:
> Hi!
>
> This patch renames gimplify_seq_add_stmt to
> gimple_seq_add_stmt_without_update so that it is more clear what
> it does, keeping the old name because it is shorter inside
> of gimplify.c to avoid having to reformat it.
>
> Bootstrapped/regtest
On Mon, 19 Dec 2011, Jakub Jelinek wrote:
> Hi!
>
> We were ignoring maybe_clean_or_replace_eh_stmt return value,
> which tells us if stmt previously could throw but doesn't any longer.
> I was a little bit worried that gimple_purge_dead_eh_edges from within
> cfg cleanup wouldn't work well, but
Hi Revital,
Revital Eres writes:
> The attached patch is a resubmission following comments made by Ayal
> and Richard.
>
> Tested and bootstrap with the other patches in the series on
> ppc64-redhat-linux, enabling SMS on loops with SC 1.
Looks really good to me. I'll leave any approval to Ayal
This patch addresses three issues:
a) For SELECT TYPE: If the selector has the pointer attribute, the
associate name is a nonpointer, but it gets the target attribute.
(Rejects-valid issue; was accepted [for the wrong reasons] before PR
48887 got fixed.)
b) The example "one" is invalid, but
Dear Tobias,
> Build and regtested on x86-64-linux.
> OK for the trunk?
>
OK
Thanks for the remarkably rapid turnround!
Paul
OK.
Jason
OK.
Jason
On 2011/12/19 上午 03:18, Richard Henderson wrote:
> On 12/17/2011 10:36 PM, Chung-Lin Tang wrote:
>> I don't think it's that kind of problem; the powerpc backend uses
>> unlikely_text_section_p(), which compares the passed in argument section
>> and the value of function_section_1(current_function_d
I've just noticed that the superfluous references to
__gcc_personality_v0 aren't created any longer and thus my hack
http://gcc.gnu.org/ml/gcc-patches/2011-11/msg02113.html
isn't needed anymore. When I tried a gas/gld bootstrap on Solaris
11/x86, I still ran into a considerable number of
The following two patches are necessary to allow a i?86-linux
--enable-targets=all bootstrap to complete, but have remained unreviewed
for a week:
[ada] Support 64-bit libgnat multilib on i?86-linux
http://gcc.gnu.org/ml/gcc-patches/2011-12/msg01009.html
[libffi] Build 64-
Hi,
currently libiberty will compile strncmp while building a cross compiler. This
is due to the fact that the test for strncmp is a TRY_RUN test.
My issue is that VMS doesn't like a redefinition of a standard function,
because it always uses a shared library for C and doesn't support symbol
On 12/19/2011 12:47 AM, Michael Zolotukhin wrote:
> Ok, those were just attempts to adjust dg-scans in slp-perm-9.c, in
> which one more loop was vectorized when compiled with -mavx2. In fact,
> just SSSE3 isn't enough for vectorization of this loop - it seems that
> vector size also matters, so I
The attached patch makes -finit-= generate default
initialization for automatic arrays.
It was OK for the trunk - is it also OK for the 4.6 branch ?
Strictly speaking, it doesn't fix a regression, it is a fix for a
(non-default) debugging option.
2011-12-19 Toon Moene
PR fortran/
Tristan Gingold writes:
> 2011-12-19 Tristan Gingold
>
> * aclocal.m4: Assume strncmp works in cross case.
> * configure: Regenerate.
This is OK.
Thanks.
Ian
Thanks, Richard!
Is anyone else's approve needed for commit?
Michael
On 19 December 2011 20:55, Richard Henderson wrote:
> On 12/19/2011 12:47 AM, Michael Zolotukhin wrote:
>> Ok, those were just attempts to adjust dg-scans in slp-perm-9.c, in
>> which one more loop was vectorized when compiled
Hi,
per the audit trail, treat BIT_NOT_EXPR specially, do not try to
"adjust" it with make_typename_type. Tested x86_64-linux.
Thanks,
Paolo.
///
/cp
2011-12-19 Paolo Carlini
PR c++/51328
* pt.c (convert_template_argument): Early error out and retur
On Mon, Dec 12, 2011 at 3:00 PM, Uros Bizjak wrote:
> Hello!
>
>> I'm testing the patch in
>> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48743#c4 against trunk
>
>
> --cut here--
> --- gcc-4.7-20110521/gcc/config/i386/driver-i386.c.~1~ 2011-01-06
> 23:59:46.0 +0100
> +++ gcc-4.7-20110521
On 12/16/2011 03:54 AM, Richard Guenther wrote:
On Thu, 15 Dec 2011, Patrick Marlier wrote:
In PR51280, LTO does ICE because the object file uses TM builtin but the TM is
not enabled.
In the patch, it displays a error message if the builtin is not defined and
due to TM.
I moved is_tm_builtin()
On Sat, 2011-12-17 at 13:32 -0800, Richard Henderson wrote:
> The other thing that should be remembered is that inside a { }
> block you don't need to double-quote all the \'s.
>
Before:
"*
{
return \"\\
mova\\t1f, r0\\n\\
\\t.align\\t2\\n\\
\\tmov\\tr15, r1\\n\\
\\tmov\\t#(0f-1f), r15\\n\\
> 2011-12-19 Jakub Jelinek
>
> PR middle-end/51590
> PR tree-optimization/51606
> * tree-vect-patterns.c (append_pattern_def_seq, new_pattern_def_seq):
> New inline functions.
> (vect_recog_over_widening_pattern,
> vect_recog_vector_vector_shift_pattern,
>
> 2011-12-13 Rainer Orth
>
> * gcc-interface/Makefile.in (%86 linux%):
> (LIBGNAT_TARGET_PAIRS_32): Split off from LIBGNAT_TARGET_PAIRS.
> (LIBGNAT_TARGET_PAIRS_64): New.
> (LIBGNAT_TARGET_PAIRS): Add either depending on multilib.
OK, thanks. You can also backport it to
Richard, Jason, are you ok with just unsetting DECL_WEAK?
For now, yes.
I will come up with a suitable testcase for Patrick's case.
Once there's a testcase, we can figure out why the other patch didn't work.
Jason
Hi Jason.
I have made a testcase from Patrick's, and am committing the p
Hello Everyone,
This patch is for the C Compiler in Cilkplus branch. It is an
extension of the following patch:
http://gcc.gnu.org/ml/gcc-patches/2011-12/msg00691.html. This patch will
implement an N-dimension array notation for assignment expressions.
Thanking You,
Yours since
Hello Everyone,
This patch is for the C-Compiler in Cilkplus GCC branch. It is an
extension of the patch given in this submission:
http://gcc.gnu.org/ml/gcc-patches/2011-12/msg01397.html. it implements a
polynomial of N-dimension array notation.
Thanking You,
Yours Sincerely,
Balaji V.
This small patch fixes a build regression on GNU/kFreeBSD.
--
Robert Millan
2011-12-19 Robert Millan
Fix build regression on GNU/kFreeBSD.
* config/kfreebsd-gnu.h (GNU_USER_DYNAMIC_LINKERX32): New macro.
--- a/src/gcc/config/kfreebsd-gnu.h~ 2011-07-21 17:31:44.0 +0200
+++ b/src/gc
The code I put in back when I implemented list-initialization rejects
additional user-defined conversions when considering the copy parm of a
copy/move constructor, in order to avoid ambiguity when a nested
braced-init-list could initialize either the copy parm or the parm of
another constructo
I have committed the attached patch as obvious (Rev. 182497).
gfortran merged the symbol of a local label like:
label_name: block
end block label_name
with a later-declared procedure with the same name ("subroutine
label_name()") which lead to an ICE and interesting code paths.
Built, regt
On 12/18/2011 05:21 AM, Eric Botcazou wrote:
OK everywhere without the cap and with the same test on ARG2:
If a testcase doesn't need fancy options, it must go in gcc.c-torture/compile.
OK. I've checked in this version, which addresses both those issues.
-Sandra
2011-12-19 Sandra Loosemor
For a constexpr delegating constructor, we should be able to just use
the delegated initializer.
Tested x86_64-pc-linux-gnu, applying to trunk.
commit 4443d34fb34484de797d087066dfe87fb55afc1f
Author: Jason Merrill
Date: Mon Dec 19 15:50:09 2011 -0500
PR c++/51526
* semantics.c (bui
Andrew Pinski writes:
> Hi,
> The problem here is mips64-linux-gnu produces similar output like
> IRIX does except there might be some .cfi_startproc there. This patch
> fixes it by allow an optional .cfi_startproc and by always using that
> format for mips*-*-* .
>
> OK? Tested on mips64-linu
Eric Botcazou writes:
>> This endian difference is the cause of the mips-sde-elf build failure
>> that Maciej reported earlier in the week. Tested on that target and
>> on x86_64-linux-gnu. OK to install?
>>
>> Richard
>>
>>
>> gcc/
>> * lower-subreg.c (can_decompose_p): Check every word of
On Dec 19, 2011, at 10:15 AM, Michael Zolotukhin wrote:
> Thanks, Richard!
> Is anyone else's approve needed for commit?
>
> Michael
>
> On 19 December 2011 20:55, Richard Henderson wrote:
>> On 12/19/2011 12:47 AM, Michael Zolotukhin wrote:
>>> Ok, those were just attempts to adjust dg-scans in
The recent(ish) improvements to widening multiplication support have
disabled madd and msub for fixed-point types. The problem is that the
optab is now chosen based on:
optype = build_nonstandard_integer_type (from_mode, from_unsigned1);
which is specific to integer types.
The only time optyp
As the subject says. Not sure how this slipped through; all the other
ext-* tests have it.
Tested on mips-sde-elf and applied.
Richard
gcc/testsuite/
* gcc.target/mips/ext-2.c: Add NOMIPS16.
Index: gcc/testsuite/gcc.target/mips/ext-2.c
=
As per the subject. This test now relies on the tree-level widening
multiplication optimisation, which only runs at -O2 and above.
(It used to happen in expand instead.)
Tested on mips-sde-elf and applied.
Richard
gcc/testsuite/
* gcc.target/mips/mult-1.c: Require -O2.
Index: gcc/test
forward_propagate_subreg has code to optimise sequences like:
(set (reg:DI X) (*_extend:DI (reg:SI Y)))
(... (subreg:SI (reg:DI X)) ...)
by replacing (subreg:SI (reg:DI X)) with (reg:SI Y). However, there is
a special case to stop this happening if Y is loaded from memory and the
extensi
Just missing a case.
Tested x86_64-pc-linux-gnu, applying to trunk.
commit cd8d1cb9501d6151febfba17c240f8bbba2d25af
Author: Jason Merrill
Date: Mon Dec 19 16:28:19 2011 -0500
PR c++/51530
* pt.c (unify): Handle NULLPTR_TYPE.
diff --git a/gcc/cp/pt.c b/gcc/cp/pt.c
index 66d4c3f..769
Reviewers: xur, jingyu, shenhan, bjanakiraman_google.com,
Message:
This merges xur's patch to our mobile release branch. LGTM and I will
submit it.
Please review this at http://codereview.appspot.com/5491078/
Affected files:
Mgcc-4_6-mobile
M gcc-4_6-mobile/gcc/ChangeLog.google-4
On Mon, Dec 19, 2011 at 2:06 PM, Richard Sandiford
wrote:
> forward_propagate_subreg has code to optimise sequences like:
>
> (set (reg:DI X) (*_extend:DI (reg:SI Y)))
> (... (subreg:SI (reg:DI X)) ...)
>
> by replacing (subreg:SI (reg:DI X)) with (reg:SI Y). However, there is
> a special c
Hello!
This problem was uncovered by a libgo testsuite, where SIGSEGV handler
in html/template test resulted in
Program received signal SIGSEGV, Segmentation fault.
libgo_html.template.Error.pN25_libgo_html.template.Error (e=0x0) at error.go:185
185 if e.Line != 0 {
While x86_64 reco
On 12/19/2011 11:07 PM, Jason Merrill wrote:
Just missing a case.
Indeed. Thanks!
Paolo.
On 12/19/2011 02:14 PM, Uros Bizjak wrote:
> (gdb) bt
> #0 _Unwind_RaiseException (exc=0xf840453300) at
> ../../../gcc-svn/trunk/libgcc/unwind.inc:135
> #1 0x02559a8c in __go_unwind_stack () at
> ../../../gcc-svn/trunk/libgo/runtime/go-unwind.c:175
> #2 0x02556d44 in __go_panic (
Andrew Pinski writes:
> On Mon, Dec 19, 2011 at 2:06 PM, Richard Sandiford
> wrote:
>> forward_propagate_subreg has code to optimise sequences like:
>>
>> (set (reg:DI X) (*_extend:DI (reg:SI Y)))
>> (... (subreg:SI (reg:DI X)) ...)
>>
>> by replacing (subreg:SI (reg:DI X)) with (reg:SI Y).
OK.
Jason
somehow I missed these fails when I backported the __promote changes
to the 4.6 branch:
* testsuite/ext/type_traits/remove_unsigned_integer_neg.cc: Adjust
dg-error line numbers.
* testsuite/ext/type_traits/add_unsigned_floating_neg.cc: Likewise.
* testsuite/ext/type
On Mon, Dec 19, 2011 at 11:21 PM, Richard Henderson wrote:
> On 12/19/2011 02:14 PM, Uros Bizjak wrote:
>> (gdb) bt
>> #0 _Unwind_RaiseException (exc=0xf840453300) at
>> ../../../gcc-svn/trunk/libgcc/unwind.inc:135
>> #1 0x02559a8c in __go_unwind_stack () at
>> ../../../gcc-svn/trunk/lib
We don't have allocator_traits in 4.6 so this patch adds a test for
whether we can call allocator::construct(p) with a single argument and
calls _Construct(p) if not.
PR libstdc++/51626
* include/bits/stl_uninitialized.h (_Construct_default_a): Define
overloaded functions t
On 2011/12/19 22:09:18, asharif1 wrote:
This merges xur's patch to our mobile release branch. LGTM and I will
submit it.
OK
http://codereview.appspot.com/5491078/
On Sun, 18 Dec 2011, Eric Botcazou wrote:
>> Here g++ is invoked without an absolute path, so if you have an older
>> version of g++ in your path that does not grok -Wno-narrowing yet that
>> will fail, won't it? I assume that's why only some like you and me are
>> seeing this, but not all of us.
Committed to both google/gcc-4_6-google and google/gcc-4_6-mobile
(mobile release branch).
Diego,
I just realize we need this patch for google/gcc-main, since it is a
local patch. OK?
Thanks,
Jing
On Thu, Dec 15, 2011 at 4:42 AM, Diego Novillo wrote:
> On 11-12-14 13:43 , Jing Yu wrote:
>
>> I
Jason Merrill writes:
> I think it might be simpler to have cxx_maybe_build_cleanup avoid
> setting EXPR_LOCATION on the call in the first place.
I have bootstrapped and tested the below on x86_64-unknown-linux-gnu
against trunk.
gcc/cp/
PR debug/49951
* decl.c (cxx_maybe_build
On Mon, Dec 19, 2011 at 11:54:36PM +0100, Dodji Seketeli wrote:
> PR debug/49951
> * g++.dg/gcov/gcov-2.C: New test.
This change doesn't look like New test, just an adjustment to it...
> diff --git a/gcc/testsuite/g++.dg/gcov/gcov-2.C
> b/gcc/testsuite/g++.dg/gcov/gcov-2.C
> index 6d
On Tue, 2011-12-13 at 11:17 -0800, Richard Henderson wrote:
> On 11/30/2011 05:13 PM, Richard Henderson wrote:
> > The library is written in C++, so in theory we can use the real atomic<>
> > templates, etc. Except that we have the same horrid problem finding the
> > C++ headers as did for , so
Jakub Jelinek writes:
> On Mon, Dec 19, 2011 at 11:54:36PM +0100, Dodji Seketeli wrote:
>> PR debug/49951
>> * g++.dg/gcov/gcov-2.C: New test.
>
> This change doesn't look like New test, just an adjustment to it...
Correct. Below is the adjusted patch.
gcc/cp/
PR debug/49951
On 12/19/2011 02:58 PM, Torvald Riegel wrote:
> In the particular case (the validated loads technique used in
> method-gl.cc, load(), store(), and validate()), we actually do not need
> to have loads or stores to be really atomic, but need the compiler to
> treat them as if they were atomics wrt. t
Reviewers: Diego Novillo, shenhan, jingyu, bjanakiraman_google.com,
Message:
Please review this on behalf of shenhan@.
Please review this at http://codereview.appspot.com/5496077/
Affected files:
M gcc-4_6-mobile/gcc/ChangeLog.google-4_6
M gcc-4_6-mobile/gcc/cfgexpand.c
M gc
Remove use of _GLIBCXX__ATOMIC_BUILTINS_4 as it's no longer ever
defined...
Also, random fixup for "make check-performance."
tested x86/linux
-benjamin2011-12-19 Benjamin Kosnik
* testsuite/performance/25_algorithms/search_n.cc: Disambiguate
local variable.
2011-12-19 Benj
On Mon, 2011-12-19 at 15:17 -0800, Richard Henderson wrote:
> On 12/19/2011 02:58 PM, Torvald Riegel wrote:
> > In the particular case (the validated loads technique used in
> > method-gl.cc, load(), store(), and validate()), we actually do not need
> > to have loads or stores to be really atomic,
Hi,
thus, as explained by Jason in the audit trail, the reason why we are
ICE-ing here and failing to provide satisfactory diagnostics is that we
are not checking for virtual base classes, which are explicitly
forbidden in C++11 for constexpr constructors.
Thus I'm simply doing the below, te
On Dec 19, 2011, at 2:25 PM, Richard Sandiford wrote:
> I've just added a PR reference to the ChangeLog entry. I don't think it's
> the kind of thing that should be backported to branches, but I'm never
> too sure whether that means the bug should be closed as fixed or not.
Closed, fixed, known t
OK.
Jason
Correct. Below is the adjusted patch.
OK.
Jason
77 matches
Mail list logo