On 2012.09.04 at 11:22 +0200, Richard Guenther wrote:
>
> This fixes PR54458 where DOM jump threading turns a loop into
> one with multiple latches but does not mark it so.
>
> Bootstrapped and tested on x86_64-unknown-linux-gnu, applied to trunk.
>
> Richard.
>
> 2012-09-04 Richard Guenther
The x86 cpuid instruction returns a processor ID and the
__get_cpuid_max function even explicitly makes the %ebx value directly
available. But users of that function have to use a cryptic constant.
How about adding a few macros to make this more transparent?
Index: gcc/config/i386/cpuid.h
=
The prologue code for functions with large static stack space
requirements was really ugly on moxie. This patch cleans things up
and I am committing this change.
2012-09-07 Anthony Green
* config/moxie/moxie.c (moxie_expand_prologue): Optimize prologue
for functions with large
Hi,
This small change:
http://gcc.gnu.org/ml/gcc-patches/2012-05/msg00044.html
still awaits review. Please help. Thanks.
Maciej
Hi,
This almost obviously correct proposal:
http://gcc.gnu.org/ml/gcc-patches/2012-08/msg00619.html
received no feedback whatsoever and still awaits review. Would you please
have a look at it? Thanks.
Maciej
On 07.09.2012 21:38, Jason Merrill wrote:
> On 09/07/2012 12:33 PM, Matthias Klose wrote:
>> thanks for the fix. looked at backporting this for 4.7. Is it really
>> necessary
>> to use C++ only syntax for this kind of patches, which are a candidate for
>> 4.7?
>
> It's not necessary to use this
Jim,
It is unfortunate that you did not discuss your plans to implement VLE
with me during the design phase.
This patch contains a lot of unnecessary, gratuitous changes in
addition to being very invasive. It was not edited and cleaned
sufficiently before posting. It has too much of a negative
Hi,
I'm curious about the status of this patch:
http://gcc.gnu.org/ml/gcc-patches/2012-06/msg00974.html
Chrome OS uses this, and the Ubuntu Security Team has expressed
interest in it as well. What's needed to land this in gcc?
Thanks!
-Kees
--
Kees Cook
Chrome OS Security
On Fri, Sep 7, 2012 at 3:44 PM, Cary Coutant wrote:
>> 2012-09-07 Sterling Augustine
>>
>> * gcc/dwarf2out.c (output_pubnames): Call comp_unit_die.
>
> OK for google/gcc-4_7.
>
> -cary
Committed as r191084.
> 2012-09-07 Sterling Augustine
>
> * gcc/dwarf2out.c (output_pubnames): Call comp_unit_die.
OK for google/gcc-4_7.
-cary
On Fri, 2012-09-07 at 07:52 +0100, Richard Sandiford wrote:
> MIPS_ABI_DEFAULT and MIPS_ISA_DEFAULT are better set in config.gcc.
> That also allows you to handle (say) mipsisa32-mti-linux-gnu vs.
> mipsisa64-mti-linux-gnu.
>
> I think in general we want more specific header files to come after
>
The enclosed patch for Google 4.7 fixes a bug where a pubtype
without a skeleton die would use an offset of zero--but a pubtype
with zero offset doubles as an end of list marker, preventing
gold from including all subsequent pubtypes in gdb index.
This patch fixes it by using an offset of the comp
Hi,
On 09/07/2012 11:36 PM, Jason Merrill wrote:
On 09/07/2012 05:07 PM, Paolo Carlini wrote:
the regression of template/crash83.C because of excess errors. The
latter happens because we have code in cp_parser_class_head which
tries to improve error recovery for cases of missing 'template <>'
On Wed, 2012-09-05 at 08:15 +0200, Jakub Jelinek wrote:
> On Fri, Aug 31, 2012 at 10:58:51AM -0700, Steve Ellcey wrote:
> > Here is my patch to fix the bootstrap comparision failure (PR 54128) on
> > MIPS. The reason for the comparision failure was a difference in
> > register usage and I tracked
On 09/07/2012 05:07 PM, Paolo Carlini wrote:
the regression of template/crash83.C because of excess errors. The latter happens
because we have code in cp_parser_class_head which tries to improve error recovery
for cases of missing 'template <>' in explicit specializations, and, for that
rather
On 09/07/2012 09:44 PM, Jason Merrill wrote:
On 09/06/2012 06:29 PM, Paolo Carlini wrote:
Ok, I did that in the below, also passes testing.
This patch is OK.
Great, applied.
In the meanwhile I looked a bit more into C++/24314 itself, and figured
out something, I'm attaching a draft-draft pa
Here the problem is that get_base_address() can return NULL_TREE and
this later leads to a segfault. Fix by checking that the return value is
valid.
gcc-4.6 and 4.7 are also affected.
Please commit if this looks OK.
Thanks.
Tested on x86_64-pc-linux-gnu
2012-09-07 Markus Trippelsdorf
>-Original Message-
>From: Jakub Jelinek [mailto:ja...@redhat.com]
>Sent: Friday, September 07, 2012 4:07 PM
>To: Andi Kleen
>Cc: Iyer, Balaji V; Richard Guenther; gcc-patches@gcc.gnu.org; Gabriel Dos
>Reis;
>Aldy Hernandez (al...@redhat.com); Jeff Law; r...@redhat.com
>Subject: Re: [PATCH
On 09/07/2012 02:59 AM, Richard Sandiford wrote:
ira.c:setup_pressure_classes treats "leaf" classes as pressure classes
even if moves between them are more expensive than moves to or from memory:
if (ira_class_hard_regs_num[cl] != 1
/* A register class without subclasses may con
Jakub Jelinek writes:
>
> Nope, -O2 -ftree-vectorize works just fine. Vectorization only needs
> if-conversion, but that is enabled by default if -ftree-vectorize
> (unless explicitly disabled).
How about the tree unrolling? I remember that being enabled for the
vectorizer (and then annoying me
Hi,
On 09/07/2012 04:02 PM, Jack Howarth wrote:
The attached patch eliminates the bootstrap failures in libstdc++-v3
of PR 54419 by adding a define to acinclude.m4, GLIBCXX_CHECK_X86_RDRAND, that
checks for assembler support of the new rdrnd opcode, using this new define
in configure.ac and
On Fri, Sep 07, 2012 at 12:59:26PM -0700, Andi Kleen wrote:
> "Iyer, Balaji V" writes:
> >
> > Yes, I really like this idea and that is kind of what I want. But, how do I
> > turn on vectorization on a function by function basis? I tried to set
> > flag_tree_vectorize=1 but that doesn't seem to
"Iyer, Balaji V" writes:
>
> Yes, I really like this idea and that is kind of what I want. But, how do I
> turn on vectorization on a function by function basis? I tried to set
> flag_tree_vectorize=1 but that doesn't seem to do the trick.
AFAIK vectorization needs a range of passes to work, so
On 09/06/2012 06:29 PM, Paolo Carlini wrote:
Ok, I did that in the below, also passes testing.
This patch is OK.
Jason
On 09/07/2012 12:33 PM, Matthias Klose wrote:
thanks for the fix. looked at backporting this for 4.7. Is it really necessary
to use C++ only syntax for this kind of patches, which are a candidate for 4.7?
It's not necessary to use this syntax, but even code that uses the VEC
macros would need
Hello Richard,
Please see my response below:
>-Original Message-
>From: Richard Guenther [mailto:richard.guent...@gmail.com]
>Sent: Friday, September 07, 2012 5:15 AM
>To: Richard Henderson
>Cc: gcc-patches@gcc.gnu.org; Gabriel Dos Reis; Iyer, Balaji V; Aldy Hernandez
>(al...@redha
On Fri, 7 Sep 2012, Thomas Quinot wrote:
> * c-family/c.opt (-fada-spec-parent): Define new command line switch.
> * c-family/c-ada-spec.c (get_ada_package): When -fada-spec-parent
> is specified, generate binding spec as a child of the specified unit.
c-family has its own
On Fri, 7 Sep 2012, Magnus Granberg wrote:
> * Makefile.in Add
> -fno-stack-protector when
> needed for espf.
Toplevel Makefile.in is a generated file. You need to patch Makefile.def
or Makefile.tpl and regenerate Makefile.in.
I'm surpr
On Fri, 7 Sep 2012, Magnus Granberg wrote:
> Patch: config.in.patch
> Add ENABLE_ESPF*
You don't patch config.in manually. You regenerate it with autoheader
after editing configure.ac.
--
Joseph S. Myers
jos...@codesourcery.com
On Fri, 7 Sep 2012, Magnus Granberg wrote:
> --- a/gcc/config/linux.h 2011-07-07 17:38:34.0 +0200
> +++ b/gcc/config/linux.h 2012-07-09 14:24:08.599281404 +0200
I see nothing related specifically to Linux rather than other targets that
may use GNU userspace, so I think all this
On Fri, 7 Sep 2012, Magnus Granberg wrote:
> +NOTE: With configure --enable-espf=@r{[}all@r{|}ssp@r{|}pie@r{]}is
@emph{Note:} (existing style). @option{--enable-espf}.
> +this option enabled by default for C, C++, ObjC, ObjC++.
> +To disable, use @option{-Wformat=0}.
-Wno-format rather th
>-Original Message-
>From: Andi Kleen [mailto:a...@firstfloor.org]
>Sent: Friday, September 07, 2012 1:05 PM
>To: Iyer, Balaji V
>Cc: Richard Guenther; gcc-patches@gcc.gnu.org; Gabriel Dos Reis; Aldy
>Hernandez (al...@redhat.com); Jeff Law; r...@redhat.com
>Subject: Re: [PATCH] Merging Ci
This is a bit different, in that we don't currently have an
infrastructure to test transactional memory code within the
simulate-thread framework. Luckily for this test, the LIM pass has an
actual dump message when it fails to hoist a value due to its presence
in a transaction.
Eventually/id
OK, thanks.
Jason
This is the same thing as gcc.dg/pr52558-1.c, but in this case I had to
tweak the testcase a bit because optimization passes after LIM are smart
enough to remove the condition altogether, thus never triggering the
test. Interestingly, GCC can figure out what's going on when the
condition is "l
On 2012-09-07 10:55 , Dehao Chen wrote:
gcc/ChangeLog:
2012-08-01 Dehao Chen
* toplev.c (general_init): Init block_locations.
* tree.c (tree_set_block): New.
(tree_block): Change to use LOCATION_BLOCK.
* tree.h (TREE_SET_BLOCK): New.
* final.c (reemit_
"Iyer, Balaji V" writes:
>>
>>The answer is that you should not enable the vectorizer.
>
> OK. I will fix that.
It still seems like useful functionality. Otherwise you have to compile
the whole program with -O3, just to vectoriz a few marked functions or
add additional annotations for all of them
On 09/07/2012 04:29 PM, Mikael Morin wrote:
Regression tested on x86_64-unknown-linux-gnu. OK for 4.8/4.7/4.6?
Looks fine. Thanks!
Tobias
On 06.09.2012 17:37, Jason Merrill wrote:
> Vtables were causing several different problems for constexpr:
>
> 1) Value-initializing a nearly-empty class (that has a vptr but no data) meant
> two initializers for a single base. Fixed by not bothering to zero out a type
> with no data before calli
The following change (r190962):
2012-09-04 Jason Merrill
PR c++/54441
* decl.c (reshape_init_class): Handle invalid initializer for
0-length array member.
introduces a poinential dereference of d->end. I hit this when
bootstrapping on OpenBSD/i386. Looks Like this ch
On 07/09/12 12:45, Christophe Lyon wrote:
> On 6 September 2012 18:42, Richard Earnshaw wrote:
>> On 06/09/12 17:07, Christophe Lyon wrote:
>>>
>>> But why are the thumb1_XXX patterns still necessary?
>>> I tried removing them, but compiling the testcase with -march=armv6
>>> -mthumb makes the com
* Steven Bosscher, 2012-08-24 :
> Why here instead of in c-family/c.opt?
Good point. Here is a new version of the patch:
* c-family/c.opt (-fada-spec-parent): Define new command line switch.
* c-family/c-ada-spec.c (get_ada_package): When -fada-spec-parent
is specified, g
Hi
This new configure option will add some preprocessor, compiler or link command
options as default. The default options will be -D_FORTIFY_SOURCE, -Wformat, -
Wformat-security, -fPIE -pie and -fstack-protector. Depending on what is
passed to --enable-espf=, ssp, pie, no, or all of them, will b
Patch: config_all.patch
This patch define ESPF_GCC_PIE_SPEC, ESPF_GCC_SSP_SPEC,
ESPF_CPP_UNIQUE_OPTIONS_SPEC and ESPF_EXTRA_SPECS
can be used on most linux targets.
Patch_i386.patch
Define DRIVER_SELF_SPECS and add ESPF_EXTRA_SPECS to
SUBTARGET_EXTRA_SPECS
Gentoo Hardened Project
Magnus Granberg
Patch: config.in.patch
Add ENABLE_ESPF*
Patch: gcc.c.patch
Add espf_cpp_unique_options to *cpp_unique_options
Gentoo Hardened Project
Magnus Granberg
--- a/gcc/config.in 2011-04-24 00:36:52.0 +0200
+++ b/gcc/config.in 2011-06-24 14:17:05.166236726 +0200
@@ -77,6 +77,29 @@
#endif
+/*
Patch: testsuite-espf-format.patch
Add -Wno-format if check_effective_target_espf to some testfiles.
Patch: testsuite-espf-fortify.patch
Add check_effective_target_espf.
New files for some of the testfiles so we can disable FORTIFY_SOURCE.
Patch: testsuite-espf-piessp.patch
Add -fno-PIE or -fno-s
Patch: gcc_doc.patch
Add notes to the command line options and info about the new configure option.
Gentoo Hardened Project
Magnus Granberg
--- a/gcc/doc/invoke.texi 2012-03-01 10:57:59.0 +0100
+++ b/gcc/doc/invoke.texi 2012-07-30 00:57:03.766847851 +0200
@@ -3216,6 +3216,10 @@ aspects of
Patch testsuite-espf-pr39537.patch
Is a update of the patch posten on the 39537 bug
We fix some of the format strings and casts.
Gentoo Hardened Project
Magnus Granberg
--- a/gcc/testsuite/g++.dg/ext/align1.C 2002-02-06 17:18:33.0 +0100
+++ b/gcc/testsuite/g++.dg/ext/align1.C 2012-08-11 02
Patch: Wformat.patch
Enable -Wformat and -Wformat_security
Gentoo Hardened Project
Magnus Granberg
--- a/gcc/c-family/c-common.c 2012-02-13 21:12:54.0 +0100
+++ b/gcc/c-family/c-common.c 2012-07-28 00:04:05.351725091 +0200
@@ -202,7 +202,11 @@ int warn_unknown_pragmas; /* Tri state v
/* W
Hello,
here comes a low hanging regression(4.6/4.7/4.8) fix.
In the test provided:
integer, parameter :: n=2
integer, dimension(1-min(n,2)/2:n) :: arr
integer, parameter :: i=ubound(arr,1)
gfc_match_init_expr checks right away (at parse time) that the
definition for `i' is a constant (so c
Hi Guys,
I am applying the patch below to tidy up the v850 backend a little and
fix a few minor bugs.
Cheers
Nick
gcc/ChangeLog
2012-09-07 Nick Clifton
* config/v850/v850.h (DBX_DEBUGGING_INFO): Define.
(ASM_GENERATE_INTERNAL_LABEL): Define if not already provided.
>-Original Message-
>From: Richard Guenther [mailto:richard.guent...@gmail.com]
>Sent: Friday, September 07, 2012 5:07 AM
>To: Iyer, Balaji V
>Cc: gcc-patches@gcc.gnu.org; Gabriel Dos Reis; Aldy Hernandez
>(al...@redhat.com); Jeff Law; r...@redhat.com
>Subject: Re: [PATCH] Merging Cilk Pl
The attached patch eliminates the bootstrap failures in libstdc++-v3
of PR 54419 by adding a define to acinclude.m4, GLIBCXX_CHECK_X86_RDRAND, that
checks for assembler support of the new rdrnd opcode, using this new define
in configure.ac and also defining _GLIBCXX_X86_RDRAND in config.h if sup
The attached patch eliminates the bootstrap failures in libstdc++-v3
of PR 54419 by adding a define to acinclude.m4, GLIBCXX_CHECK_X86_RDRAND, that
checks for assembler support of the new rdrnd opcode, using this new define
in configure.ac and also defining _GLIBCXX_X86_RDRAND in config.h if sup
On 09/06/2012 11:44 AM, Aldy Hernandez wrote:
The current test is failing on some architectures because the
underlying gimple has changed. I believe the best way to test the
speculative store data race is with the simulate-thread/ harness. We
already have a speculative store test in the harnes
It was presumably papering over the real problem and resulted in huge
performance regressions for some loops.
Tested on x86-64/Linux, applied on mainline and 4.7 branch.
2012-09-07 Eric Botcazou
* gcc-interface/trans.c (Loop_Statement_to_gnu): Revert to using
size_type_node
On 6 September 2012 18:42, Richard Earnshaw wrote:
> On 06/09/12 17:07, Christophe Lyon wrote:
>>
>> But why are the thumb1_XXX patterns still necessary?
>> I tried removing them, but compiling the testcase with -march=armv6
>> -mthumb makes the compiler fail (internal compiler error:
>> output_op
On 2012-09-06 05:46 , Richard Guenther wrote:
On Wed, Aug 22, 2012 at 1:54 AM, Dehao Chen wrote:
On Tue, Aug 21, 2012 at 6:25 AM, Richard Guenther
wrote:
On Mon, Aug 20, 2012 at 3:18 AM, Dehao Chen wrote:
ping
Conceptually I like the change. Can a libcpp maintainer please have a 2nd
> -Original Message-
> From: gcc-patches-ow...@gcc.gnu.org [mailto:gcc-patches-ow...@gcc.gnu.org]
On
> Behalf Of Bin Cheng
> Sent: Wednesday, September 05, 2012 7:01 PM
> To: Ramana Radhakrishnan
> Cc: gcc-patches@gcc.gnu.org
> Subject: RE: [PATCH] PR45070: Fix wrong epilogue code for cor
On 5 September 2012 16:01, Ian Lance Taylor wrote:
>...
>
> I don't know if it's a strong reason, but the problem seems to be one
> that is characteristic of a specific invocation of a compiler, rather
> than characteristic of the compiler in general. The same compiler may
> be invoked in multipl
> This fixes part of the issue described in PR gcov-profile/54487 where
> there were warnings about mismatches due to slight differences in the
> merged histograms in different object files. This can happen due to
> the truncating integer division in the merge routine, which could result
> in sligh
Applied on trunk and will wait for a few days before backporting to 4.7
branch.
Applied now to 4.7 branch .
regards,
ramana
This fixes a bug in how we treat EAF_NOESCAPE function arguments
in points-to analysis. We forgot to mark them clobbered (ugh).
To retain optimization we need to transitively close arguments
separately after this patch, thus this is what is done.
Bootstrapped on x86_64-unknown-linux-gnu, testing
There are 3 global deletions instead of 2 on SPARC64 and 0 on Alpha since the
structure is returned in memory.
Tested on x86-64/Linux and SPARC64/Linux, applied on mainline and 4.7 branch.
2012-09-07 Eric Botcazou
* gcc.dg/pr44194-1.c: Skip on Alpha and adjust regexp for SPARC64.
On Fri, Sep 7, 2012 at 10:35 AM, Richard Earnshaw wrote:
> On 20/08/12 12:36, Richard Guenther wrote:
>> On Fri, Aug 17, 2012 at 7:05 PM, Richard Earnshaw wrote:
>>> On 17/08/12 16:20, Richard Earnshaw wrote:
Ok, in which case we have to give is_widening_mult_rhs_p enough smarts
to not
On Fri, Sep 7, 2012 at 7:01 AM, Andi Kleen wrote:
> From: Andi Kleen
>
> For parallel LTO builds setting -fmem-report does not work very well
> because all the LTRANS phases dump it in parallel and typically interleave
> it to unreadability.
>
> Since usually the memory bottleneck is WPA add a fl
On Thu, Sep 6, 2012 at 5:51 PM, Richard Henderson wrote:
> On 09/06/2012 02:37 AM, Richard Guenther wrote:
>> In all this seems unrelated to CILK+ work (even if you make use of this
>> from within CILK+).
>
> While true, we also asked him to split up the work. And this piece,
> done correctly, se
On Thu, Sep 6, 2012 at 8:25 PM, Iyer, Balaji V wrote:
> Hello Richard,
> I forgot to answer one of questions. Please see it below:
>
> Thanks,
>
> Balaji V. Iyer.
>
>
>>+static tree
>>+handle_vector_attribute (tree *node, tree name ATTRIBUTE_UNUSED,
>>+ tree args ATTR
On 20/08/12 12:36, Richard Guenther wrote:
> On Fri, Aug 17, 2012 at 7:05 PM, Richard Earnshaw wrote:
>> On 17/08/12 16:20, Richard Earnshaw wrote:
>>> Ok, in which case we have to give is_widening_mult_rhs_p enough smarts
>>> to not strip
>>>
>>> (s32)u32
>>>
>>> and return u32.
>>>
>>> I'l
On Thu, 6 Sep 2012, Aldy Hernandez wrote:
> The current test is failing on some architectures because the underlying
> gimple has changed. I believe the best way to test the speculative store data
> race is with the simulate-thread/ harness. We already have a speculative
> store test in the harn
70 matches
Mail list logo
