The Go frontend had a bug when there were multiple types with the same
name in a single function. That could happen if the types appeared in
different blocks. The compiler would use the same name for the type
descriptor for each type, leading to a multiple definition error at
assembly time. This
This patch to the libbacktrace Makefile.am adds some more -I options, so
that it can find unwind.h under more conditions. This splits the -I
options in AM_CPPFLAGS, which is where they belong anyhow. Bootstrapped
and ran libbacktrace tests on x86_64-unknown-linux-gnu. Committed to
mainline.
Ian
It turns out "make bootstrap" work now, but "make install" is still
broken on i386-unknown-freebsd10.0:
libtool: compile: gcc -DHAVE_CONFIG_H -I.
-I/scratch2/tmp/gerald/gcc-HEAD/libbacktrace -W -Wall -Wwrite-strings
-Wstrict-prototypes -Wmissing-prototypes -Wold-style-definition
-Wmissing-form
On Mon, 17 Sep 2012, Ian Lance Taylor wrote:
> While this should be fixed now, note that it should only arise if
> i386-unknown-freebsd10.0 does not use ELF. If this is an ELF based
> system, then we need to discover why it was not recognized as such.
> What is the output of
>
> awk -f SRCDIR/lib
> -Original Message-
> From: Richard Guenther [mailto:richard.guent...@gmail.com]
> Sent: Tuesday, September 18, 2012 5:50 PM
> To: Bin Cheng
> Cc: Bin.Cheng; gcc-patches@gcc.gnu.org; Richard Earnshaw
> Subject: Re: Ping^2: [PATCH]Remove duplicate check on BRANCH_COST in fold-
> const.c
>
The compiler changes are OK.
Jason
Test gcc.dg/vect/pr48765, run only for powerpc*-*-*, unconditionally
adds -mcpu=power6 to the command line. It also adds -m64, although
tests should rely on different multilib flags for that and skip the
test if not lp64 if the test can only be run with -m64.
For all of the Power multilib flags t
I am about to submit the ARCompact target port; this port needs a few
patches to target-independent code.
There is a move pattern with 20 alternatives; a few of them need a simple
function call to decide which output pattern to use. With the '@'-syntax
for multi-alternative templates, each alter
> From: John David Anglin
> Date: Sun, 16 Sep 2012 19:13:19 +0200
> This adjusts the regexp to work on hppa. Tested on hppa2.0w-hp-hpux11.11
> and hppa64-hp-hpux11.11. Committed to trunk.
> 2012-09-16 John David Anglin
>
> PR debug/54460
> * g++.dg/debug/dwarf2/nested-3.C: Add
On Mon, Sep 17, 2012 at 3:51 PM, Michael Meissner
wrote:
> This patch has support for all of the additonal cleanups I mentioned in the
> first patch that I hadn't gotten to. At this point, I am not planning any
> more
> enhancements to the patch, and I would like to check it in.
>
> On my 64-bit
On 09/19/2012 01:24 AM, Paolo Carlini wrote:
On 09/19/2012 01:12 AM, Paolo Carlini wrote:
Hi again,
On 09/18/2012 08:33 PM, Paolo Carlini wrote:
But I'm not surprised, frankly, I think the conflict is expected,
*IF* (please check) TR1 says that those three overloads, for float,
double an long
On 09/19/2012 01:12 AM, Paolo Carlini wrote:
Hi again,
On 09/18/2012 08:33 PM, Paolo Carlini wrote:
But I'm not surprised, frankly, I think the conflict is expected,
*IF* (please check) TR1 says that those three overloads, for float,
double an long double, must be declared in std::tr1 (likewis
Hi again,
On 09/18/2012 08:33 PM, Paolo Carlini wrote:
But I'm not surprised, frankly, I think the conflict is expected, *IF*
(please check) TR1 says that those three overloads, for float, double
an long double, must be declared in std::tr1 (likewise for all the
other math functions) Now, give
Which gcc version does this patch apply to?
Thanks!
Bugzilla from ja...@redhat.com wrote:
>
> Hi!
>
> This patch changes get_best_mode, so that it doesn't suggest using larger
> modes if it means it could clobber variables located after the containing
> one.
>
> On the attached testcase on x
In the recent patch for PR rtl-opt/44194 I left a case potentially ICEing
because I managed to have it covered neither on x86-64 nor on SPARC 64-bit.
This PR brings the missing coverage and it's from the gcc.c-torture/compile
testsuite on PowerPC/Darwin 64-bit.
Bootstrapped/regtested on x86-64/L
> In any case, the change in the condition you noticed was introduced by a
> recent patch by Bernd:
> http://gcc.gnu.org/ml/gcc-patches/2012-08/msg00171.html
>
> It seems that we ought to use a similar test to what Bernd introduced in
> gen_reload, that is, use the "replaced_subreg" routine on rld
On 09/18/2012 12:54 PM, Janis Johnson wrote:
> In most cases a test that requires ARM NEON should use effective target
> arm_neon, which means that flags run for all tests include NEON support.
> The result is cached the first time it is checked for a multilib.
>
> Vectorization tests, when run fo
On Tue, Sep 18, 2012 at 1:01 PM, Segher Boessenkool
wrote:
> Without these patterns, the exact same code is generated. This
> is a leftover from when we still had a "doz" instruction.
>
> Tested on powerpc64-linux --enable-languages=c,c++,fortran;
> no regressions. Also tested all these patterns
Am 18.09.2012 20:46, schrieb Thomas Koenig:
Hello world,
pretty self-explanatory. No test case because there is no change
in behavior. OK for trunk?
I forgot to add: Regression-tested.
Thomas
On 9/18/12, Richard Guenther wrote:
> On Sep 18, 2012 Lawrence Crowl wrote:
> > * cgraph.h (varpool_node): Rename to varpool_node_for_tree.
>
> Sure it should be varpool_node_for_decl, if any.
> Or varpool_node_from_decl (grep for what is more common)
Grep says _for_decl wins over _from_decl by
On 09/17/2012 07:32 PM, Ulrich Weigand wrote:
> In any case, the change in the condition you noticed was introduced by a
> recent
> patch by Bernd: http://gcc.gnu.org/ml/gcc-patches/2012-08/msg00171.html
>
> It seems that we ought to use a similar test to what Bernd introduced in
> gen_reload, th
On Tue, 18 Sep 2012, Ian Lance Taylor wrote:
> On Tue, Sep 18, 2012 at 3:58 AM, Arnaud Charlet wrote:
> >
> > Since this issue is more general, I have split my changes and introduced a
> > new
> > tentative switch called -fextra-slocs, which is the subject of this email.
>
> Sorry for picking o
Test gcc.dg/vect/fast-math-pr35982.c uses xfail in a dg-final check when
it should instead skip the check for that effective target. Tested on
arm-none-eabi for a variety of test multilibs, checked in on trunk as
obvious. I'll backport to 4.6 when the branch is open.
Janis
2012-09-18 Janis John
Seventeen tests in gcc.dg/vect that use vect_multiple_sizes have checks
similar to:
/* { dg-final { scan-tree-dump-times "can't determine dependence" 2 "vect" {
xfail vect_multiple_sizes } } } */
/* { dg-final { scan-tree-dump-times "can't determine dependence" 4 "vect" {
target vect_multiple_si
In most cases a test that requires ARM NEON should use effective target
arm_neon, which means that flags run for all tests include NEON support.
The result is cached the first time it is checked for a multilib.
Vectorization tests, when run for ARM, add flags to support NEON if it's
OK to do so, b
The infrastructure for gcc.dg/vect tests determines whether the default
is for tests to be compile-only or compile plus execute. Tests that
should not be executed use { dg-do compile }, but no test should use
{ dg-do run }. This patch removes { dg-do run} from pr52298.c.
Tested on arm-none-eabi
> >> and at that point you have lost your extra location
> >> information.
> >
> > Actually no, see the c-family/c-common.c patch, copied here, which
> > ensures that folding does preserve such information:
>
> Thanks. I think I would like some clarity on when the extra location
> information is
Hello,
There is another opportunity where SH's rotcr insn can be used.
Tested on rev 191342 with
make -k check RUNTESTFLAGS="--target_board=sh-sim
\{-m2/-ml,-m2/-mb,-m2a/-mb,-m4/-ml,-m4/-mb,-m4a/-ml,-m4a/-mb}"
and no new failures.
OK to install?
Cheers,
Oleg
gcc/ChangeLog:
PR target/5
Hello,
There is another opportunity where SH's addc insn can be used.
Tested on rev 191342 with
make -k check RUNTESTFLAGS="--target_board=sh-sim
\{-m2/-ml,-m2/-mb,-m2a/-mb,-m4/-ml,-m4/-mb,-m4a/-ml,-m4a/-mb}"
and no new failures.
OK to install?
Cheers,
Oleg
gcc/ChangeLog:
PR target/54
Hello,
Like the topic says. No functional change, just cosmetics.
Tested on rev 191342 with
make -k check RUNTESTFLAGS="--target_board=sh-sim
\{-m2/-ml,-m2/-mb,-m2a/-mb,-m4/-ml,-m4/-mb,-m4a/-ml,-m4a/-mb}"
and no new failures.
OK to install?
Cheers,
Oleg
ChangeLog:
* config/sh/sh.md (
Hello world,
pretty self-explanatory. No test case because there is no change
in behavior. OK for trunk?
Thomas
2012-09-18 Thomas König
PR fortran/54599
* dependency.c (gfc_dep_compare_expr): Clarify logic,
remove dead code.
Index: dependency.c
==
On Thu, 06 Sep 2012, Mark Kettenis wrote:
> Most bits are stolen from Linux, but there are a few subtle
> differences since our assembler is configured to be slightly more
> HP-UX-ish.
>
>
> libgcc/:
>
> 2012-09-06 Mark Kettenis
>
> * config.host (hppa-*-openbsd*): New target.
>
On 9/18/12, Eric Botcazou wrote:
>> When, the property test is embedded within a larger condition, a little
>> restructuring is required to pull out the secondary conditions. For
>> example,
>>
>> if (symtab_variable_p (node)
>> && varpool (node)->finalized)
>> varpool_analyze_node (v
Hi,
"Fabien Chêne" ha scritto:
>2012/9/18 Paolo Carlini :
>
>> I don't understand: what's wrong - exactly - with the std::tr1::pow
>(double, double) overload above? Now I can't immediately check, but do
>the EDG and CLANG front ends accept it or not?
>
>They don't. The problem is that it conflic
Sandra Loosemore writes:
> On 08/27/2012 10:36 AM, Richard Sandiford wrote:
>> Sandra Loosemore writes:
>>> On 08/19/2012 11:22 AM, Richard Sandiford wrote:
Not sure whether a peephole is the right choice here. In practice,
I'd imagine these opportunities would only come from a DI
2012/9/18 Paolo Carlini :
> I don't understand: what's wrong - exactly - with the std::tr1::pow (double,
> double) overload above? Now I can't immediately check, but do the EDG and
> CLANG front ends accept it or not?
They don't. The problem is that it conflicts with ::pow(double,double).
// i
On Tue, Sep 18, 2012 at 10:58 AM, Arnaud Charlet wrote:
>
>> and at that point you have lost your extra location
>> information.
>
> Actually no, see the c-family/c-common.c patch, copied here, which
> ensures that folding does preserve such information:
Thanks. I think I would like some clarity
> When, the property test is embedded within a larger condition, a little
> restructuring is required to pull out the secondary conditions. For
> example,
>
> if (symtab_variable_p (node)
> && varpool (node)->finalized)
> varpool_analyze_node (varpool (node));
>
> becomes
>
> if (
This patch to libbacktrace adds some support for mingw. The executable
is opened with O_BINARY. The fcntl function is not called.
Bootstrapped and ran libbacktrace testsuite on
x86_64-unknown-linux-gnu. Committed to mainline.
Ian
2012-09-18 Ian Lance Taylor
* posix.c (O_BINARY): D
Hi,
"Fabien Chêne" ha scritto:
>2012/9/11 Fabien Chêne :
>> Oops, not sure how I test that change initially, or I must be blind,
>> because it triggers an error in tr1/cmath about pow. I'll see what I
>> can do...
>
>Well, as summarized in the code below, the problem seems to be the
>redundant o
> Sorry for picking on simple stuff, but the switch name seems
No problem, and thanks for your feedback.
> meaningless, and there isn't any documentation.
Ah. I'm open for suggestion on a better name, or I can come up with a
new one.
I'll indeed add documentation as soon as there's some kind of
On Tue, Sep 18, 2012 at 3:58 AM, Arnaud Charlet wrote:
>
> Since this issue is more general, I have split my changes and introduced a new
> tentative switch called -fextra-slocs, which is the subject of this email.
Sorry for picking on simple stuff, but the switch name seems
meaningless, and ther
Hello!
g++.dg/other/vector-compare.C recently started to fail on
alphaev68-pc-linux-gnu with:
[uros@localhost other]$ ~/gcc-build-alpha/gcc/cc1plus -std=gnu++11
-quiet vector-compare.C
vector-compare.C: In function ‘int main()’:
vector-compare.C:26:5: internal compiler error: in
emit_cmp_and_jump
... And the patch.
2012/9/18 Fabien Chêne :
> 2012/9/11 Fabien Chêne :
>> Oops, not sure how I test that change initially, or I must be blind,
>> because it triggers an error in tr1/cmath about pow. I'll see what I
>> can do...
>
> Well, as summarized in the code below, the problem seems to be the
2012/9/11 Fabien Chêne :
> Oops, not sure how I test that change initially, or I must be blind,
> because it triggers an error in tr1/cmath about pow. I'll see what I
> can do...
Well, as summarized in the code below, the problem seems to be the
redundant overload of std::tr1::pow(double,double).
On Tue, Sep 18, 2012 at 4:23 AM, Richard Guenther wrote:
>
> This adds -Og as optimization level targeted at the devel-compile-debug
> cycle (formerly mostly tied to -O0 due to debug issues with even -O1).
This needs an entry in gcc-4.8/changes.html, of course.
Ian
On Tue, Sep 18, 2012 at 4:23 AM, Richard Guenther wrote:
>
> 2012-09-18 Richard Guenther
>
> PR other/53316
> * common.opt (optimize_debug): New variable.
> (Og): New optimization level.
> * doc/invoke.texi (Og): Document.
> * opts.c (maybe_default_option
Added the new avr-gcc configure option --with-avrlibc to the 4.7 release notes.
Johann
Index: gcc-4.7/changes.html
===
RCS file: /cvs/gcc/wwwdocs/htdocs/gcc-4.7/changes.html,v
retrieving revision 1.125
diff -u -p -r1.125 changes.html
On 08/27/2012 10:36 AM, Richard Sandiford wrote:
Sandra Loosemore writes:
On 08/19/2012 11:22 AM, Richard Sandiford wrote:
Not sure whether a peephole is the right choice here. In practice,
I'd imagine these opportunities would only come from a DImode move of
$0 into a doubleword register, s
This patch to libbacktrace marks the test functions with the unused
attribute. This avoids producing a compilation warning when building
the test on a system that does not support libbacktrace. Bootstrapped
and ran libbacktrace testsuite on x86_64-unknown-linux-gnu. Committed
to mainline.
Ian
Hello,
a while ago Andrew Stubbs posted a patch to use NEON registers
and instructions to perform 64-bit integer shifts:
http://gcc.gnu.org/ml/gcc-patches/2012-05/msg01645.html
As Andrew no longer works on ARM, I've now picked this up and
reworked it a bit:
- Updated for current mainline changes
Without these patterns, the exact same code is generated. This
is a leftover from when we still had a "doz" instruction.
Tested on powerpc64-linux --enable-languages=c,c++,fortran;
no regressions. Also tested all these patterns manually,
-m32 and -m64, -misel and -mno-isel.
Okay to apply?
201
On 09/18/2012 09:51 AM, Dodji Seketeli wrote:
+ VEC_safe_push (scoped_attributes, heap, attributes_table, sa);
+ result = &VEC_last (scoped_attributes, attributes_table);
Here you can set result from the return value of VEC_safe_push.
+ if ((flags & ATTR_FLAG_CXX11)
+ &
I foolishly assumed that the autoconf macro AC_CHECK_DECLS worked like
most autoconf macros, and did not define HAVE_DECL_xx when the
declaration is not available. However, it turns out that it actually
#defines it to 0. This patch fixes the test of HAVE_DECL_STRNLEN to
match that behaviour. Boo
On Tue, Sep 18, 2012 at 9:25 AM, Jakub Jelinek wrote:
> On Tue, Sep 18, 2012 at 09:03:03AM -0700, Ian Lance Taylor wrote:
>> +*/
>> +
>> +#include "config.h"
>
> Shouldn't this #include for size_t, or is config.h providing
> size_t? Or #include ? From what I can see, config.h doesn't
> always
On Tue, Sep 18, 2012 at 9:08 AM, Joseph S. Myers
wrote:
>
> I think AC_USE_SYSTEM_EXTENSIONS is preferred to defining _GNU_SOURCE in
> individual source files.
Thanks for the pointer. I have committed this patch after bootstrap
and libbacktrace test on x86_64-unknown-linux-gnu.
Ian
2012-09-18
On Tue, Sep 18, 2012 at 09:03:03AM -0700, Ian Lance Taylor wrote:
> --- strnlen.c (revision 0)
> +++ strnlen.c (revision 0)
> @@ -0,0 +1,28 @@
> +/* Portable version of strnlen.
> + This function is in the public domain. */
> +
> +/*
> +
> +@deftypefn Supplemental size_t strnlen (const char *@va
Ping.
On Fri, Aug 24, 2012 at 5:34 PM, Sriraman Tallam wrote:
> Hi Jason,
>
>I have created a new patch to use target hooks for all the
> functionality and make the front-end just call the target hooks at the
> appropriate places. This is more like what you suggested in a previous
> mail. In
On 18/09/12 14:24, Ian Bolton wrote:
> New version attached with better formatted test cases.
>
> OK for aarch64-branch and aarch64-4.7-branch?
>
> Cheers,
> Ian
>
>
> -
>
> 2012-09-18 Ian Bolton
>
> gcc/
> * config/aarch64/aarc
On Sep 18, 2012 8:43 AM, "Xinliang David Li" wrote:
>
> On Tue, Sep 18, 2012 at 1:48 AM, Sharad Singhai wrote:
> > In response to the recent comments, I have updated the patch to do the
> > following:
> >
> > - Remove pass handling from -fopt-info
> > - Support additional flags in regular dumps
>
On 09/18/2012 10:28 AM, Paolo Carlini wrote:
On 09/18/2012 03:23 AM, Ed Smith-Rowland wrote:
Here is another tweak for the Nakagami distribution.
operator() is a one liner without a local variable.
The template friend is fixed and the relevant test has been added.
Changed dates.
If this ve
On Tue, Sep 18, 2012 at 7:55 AM, Richard Guenther
wrote:
>
> Indeed - we ship binaries with stripped debug info, usually not installed.
> libbacktrace will only produce useless garbage then. So I want a way
> to disable it (at least by default) at configure time.
The library won't print useless
On Tue, Sep 18, 2012 at 7:40 AM, Markus Trippelsdorf
wrote:
>
> If you use "make install-strip" to install, then libbacktrace will have
> been build in vain. At least for this case a way to disable libbacktrace
> should be available.
Why bother? The library will fail to find anything useful to p
On Tue, Sep 18, 2012 at 1:55 AM, Richard Guenther
wrote:
> On Tue, Sep 18, 2012 at 10:54 AM, Richard Guenther
> wrote:
>> On Tue, Sep 18, 2012 at 10:32 AM, Rainer Orth
>> wrote:
>>> The libbacktrace integration broke Solaris 10 and 11 bootstrap when
>>> using gcc 4.4 (any version of gcc without
On Tue, Sep 18, 2012 at 1:32 AM, Rainer Orth
wrote:
> The libbacktrace integration broke Solaris 10 and 11 bootstrap when
> using gcc 4.4 (any version of gcc without __sync_* support actually):
The patch is fine and should fix the problem, but GCC 4.4 does have
__sync_* support. Might be worth l
On Tue, 18 Sep 2012, Uros Bizjak wrote:
> Index: dwarf.c
> ===
> --- dwarf.c (revision 191413)
> +++ dwarf.c (working copy)
> @@ -30,6 +30,8 @@
> IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
> POSSIBILI
On Tue, Sep 18, 2012 at 7:08 AM, Ian Lance Taylor wrote:
>
> I'll fix the strnlen issue some other way.
I committed a pair of patches, to libiberty and libbacktrace, that
should fix the problem.
Let me know if it is still present.
Sorry for the difficulty.
Ian
This patch to libbacktrace declares strnlen if it is not declared in a
standard header file. If necessary, the actual definition will come
from libiberty. Bootstrapped on x86_64-unknown-linux-gnu. Committed to
mainline.
Ian
2012-09-18 Ian Lance Taylor
* configure.ac: Check whether
This patch to libiberty adds support for strnlen if it is not already
present.
I rebuilt the Makefile dependencies. This revealed that maint-tool
wasn't recognizing that files that include dwarf2.h also depend on
dwarf2.def, so I fixed maint-tool.
I also rebuilt functions.texi, which had not bee
On Tue, 11 Sep 2012, David Edelsohn wrote:
> > 2012-09-10 Maciej W. Rozycki
> >
> > gcc/
> > * config/rs6000/rs6000.c (print_operand) <'c'>: Remove.
> > * config/rs6000/spe.md: Remove a leftover comment.
>
> Okay.
I have applied this change now, thanks for your review
On Tue, Sep 18, 2012 at 1:48 AM, Sharad Singhai wrote:
> In response to the recent comments, I have updated the patch to do the
> following:
>
> - Remove pass handling from -fopt-info
> - Support additional flags in regular dumps
>
> I have massaged the options so that they have the following (hop
On 2012.09.18 at 16:55 +0200, Richard Guenther wrote:
> On Tue, Sep 18, 2012 at 4:40 PM, Markus Trippelsdorf
> wrote:
> > On 2012.09.18 at 06:58 -0700, Ian Lance Taylor wrote:
> >> On Tue, Sep 18, 2012 at 1:49 AM, Richard Guenther
> >> wrote:
> >> > On Mon, Sep 17, 2012 at 7:17 PM, Ian Lance Tayl
On 09/18/2012 05:10 AM, Jakub Jelinek wrote:
> 2012-09-18 Jakub Jelinek
>
> PR target/54592
> * config/i386/i386.c (ix86_rtx_costs): Limit > UNITS_PER_WORD
> AND/IOR/XOR cost calculation to MODE_INT class modes.
>
> * gcc.target/i386/pr54592.c: New test.
Ok.
r~
On Tue, Sep 18, 2012 at 4:40 PM, Markus Trippelsdorf
wrote:
> On 2012.09.18 at 06:58 -0700, Ian Lance Taylor wrote:
>> On Tue, Sep 18, 2012 at 1:49 AM, Richard Guenther
>> wrote:
>> > On Mon, Sep 17, 2012 at 7:17 PM, Ian Lance Taylor wrote:
>> >> OK for mainline?
>> >
>> > Hm. Can you please be
Hi Tulio,
Thanks for all the cleanups!
Two quite minor things...
+(define_insn "rs6000_get_timebase_ppc64"
+ [(set (match_operand:DI 0 "gpc_reg_operand" "=r")
+(unspec_volatile:DI [(const_int 0)] UNSPECV_GETTB))]
+ "TARGET_POWERPC64"
+{
+ if (TARGET_MFCRF)
+return "mfspr %0, 268
On Tue, 18 Sep 2012, Jan Hubicka wrote:
> Hi,
> this patch reorganize lto-symtab to work across symtab's symbol table instead
> of building its own. This simplifies things a bit and with the previous
> changes it is rather straighforward - i.e. replace all uses of
> lto_symtab_entry_t by symtab_n
On 2012.09.18 at 06:58 -0700, Ian Lance Taylor wrote:
> On Tue, Sep 18, 2012 at 1:49 AM, Richard Guenther
> wrote:
> > On Mon, Sep 17, 2012 at 7:17 PM, Ian Lance Taylor wrote:
> >> OK for mainline?
> >
> > Hm. Can you please be that verbose only for ENABLE_CHECKING compilers?
>
> That would be
I've just back ported this into ARM/AARCH64-4.7.
/Marcus
2012-09-18 Marcus Shawcroft
* config/aarch64/aarch64-linux.h (MULTIARCH_TUPLE): Remove.
(STANDARD_STARTFILE_PREFIX_1): Likewise.
(STANDARD_STARTFILE_PREFIX_2): Likewise.diff --git a/gcc/config/aarch64/aarch64-lin
Ian Lance Taylor writes:
> mmapio.c uses _GNU_SOURCE so that getpagesize is available on GNU/Linux
> systems.
That should be available by default (which includes BSD things) unless
the namespace has been restricted in some other way.
There's of course also the POSIX way of using sysconf(_SC_PA
Hi,
tested x86_64-linux (with and without #include "opt_random.h" at the end
of ext/random), committed to mainline. Should be fixed now.
Thanks,
Paolo.
PS: I just noticed that in ext/random, inside namespace __gnu_cxx, we
are using, unqualified, size_t and other types. We shouldn't: it's onl
I've just committed this patch to aarch64-branch to remove the
multi-arch fudge.
/Marcus
2012-09-18 Marcus Shawcroft
* config/aarch64/aarch64-linux.h (MULTIARCH_TUPLE): Remove.
(STANDARD_STARTFILE_PREFIX_1): Likewise.
(STANDARD_STARTFILE_PREFIX_2): Likewise.diff --git
tisdag 11 september 2012 01.33.42 skrev Magnus Granberg:
> fredag 07 september 2012 18.52.11 skrev du:
> > On Fri, 7 Sep 2012, Magnus Granberg wrote:
> > > * Makefile.in Add
> > > -fno-stack-protector when
> > >
> > > needed for espf.
> >
On Tue, Sep 18, 2012 at 2:25 AM, Uros Bizjak wrote:
> On Tue, Sep 18, 2012 at 11:16 AM, Richard Guenther
> wrote:
>
>>> CentOS 5.8 uses glibc version 2.5 that needs _GNU_SOURCE defined to use
>>> strnlen.
>>
>> Hm, shouldn't libiberty contain a xstrnlen? I bet strnlen isn't available
>> everywh
On Tue, Sep 18, 2012 at 1:49 AM, Richard Guenther
wrote:
> On Mon, Sep 17, 2012 at 7:17 PM, Ian Lance Taylor wrote:
>> This patch to the diagnostic code uses the new backtrace library to
>> print a backtrace on an ICE. For example, here is the output of a test
>> case I took from a C++ PR:
>>
>>
Hi,
this patch reorganize lto-symtab to work across symtab's symbol table instead
of building its own. This simplifies things a bit and with the previous
changes it is rather straighforward - i.e. replace all uses of
lto_symtab_entry_t by symtab_node.
There are few differences in between the symt
New version attached with better formatted test cases.
OK for aarch64-branch and aarch64-4.7-branch?
Cheers,
Ian
-
2012-09-18 Ian Bolton
gcc/
* config/aarch64/aarch64.h: Define CTZ_DEFINED_VALUE_AT_ZERO.
* config/aarch
> > > diff --git a/gcc/config/aarch64/aarch64.md
> > b/gcc/config/aarch64/aarch64.md
> > > index 33815ff..5278957 100644
> > > --- a/gcc/config/aarch64/aarch64.md
> > > +++ b/gcc/config/aarch64/aarch64.md
> > > @@ -153,6 +153,8 @@
> > > (UNSPEC_CMTST 83) ; Used in aarch64-simd.md.
> >
On Tue, Sep 18, 2012 at 2:18 PM, Vladimir Yakovlev wrote:
> I tried to perform vzeroupper emitting after reload as additional pass
> of mode switching.
> I sow one problem that I don't know haw to overcome. After
> 'pro_and_epilogue', there can be no
> flow edge to exit block and pre_exit block is
On Tue, Sep 18, 2012 at 02:38:27PM +0200, Richard Guenther wrote:
> > Glibc must be compiled with optimization. Will -Og build glibc?
>
> Apply the patch and check for yourself. I suppose the answer will
> be yes (whatever "optimization" requirements glibc has).
glibc headers check __OPTIMIZE__
On Tue, Sep 18, 2012 at 2:20 PM, Jakub Jelinek wrote:
> Hi!
>
> vect_gen_perm_mask is not suitable for use outside of the vectorizer,
> it uses current vector size to determine the number of units of a vector,
> which isn't something that should be used outside of the vectorizer.
>
> The following
On Tue, 18 Sep 2012, H.J. Lu wrote:
> On Tue, Sep 18, 2012 at 4:23 AM, Richard Guenther wrote:
> >
> > This adds -Og as optimization level targeted at the devel-compile-debug
> > cycle (formerly mostly tied to -O0 due to debug issues with even -O1).
> >
> > Discussion on g...@gcc.gnu.org at least
Hi!
vect_gen_perm_mask is not suitable for use outside of the vectorizer,
it uses current vector size to determine the number of units of a vector,
which isn't something that should be used outside of the vectorizer.
The following patch just does construct the mask inline, it is not that long
cod
I tried to perform vzeroupper emitting after reload as additional pass
of mode switching.
I sow one problem that I don't know haw to overcome. After
'pro_and_epilogue', there can be no
flow edge to exit block and pre_exit block is not created in this case
(see rotine create_pre_exit).
Without that
On Wed, Jun 27, 2012 at 02:36:14PM -0700, Richard Henderson wrote:
> As noticed by Igor Zamyatin. Committed.
> PR target/53749
> * config/i386/i386.c (ix86_rtx_costs): Fix typo vs UNITS_PER_WORD
> in 2012-06-23 change. Adjust two other DImode tests as well.
This change broke cos
On Tue, Sep 18, 2012 at 4:23 AM, Richard Guenther wrote:
>
> This adds -Og as optimization level targeted at the devel-compile-debug
> cycle (formerly mostly tied to -O0 due to debug issues with even -O1).
>
> Discussion on g...@gcc.gnu.org at least shows interest in this, so this
> is a formal pa
Hello!
> This patch aims to fix all stability issues related to using the first
> scheduler in gcc
> for x86 target (there several reported issues related to this problem).
>
> Main idea of this activity is mostly to provide user a possibility to
> safely turn on first scheduler for his codes. In
Hello!
> You are right I no need the changes in mode-switchig.c at all. After I
> remove additional argument from EMIT_MODE_SET and run 'make check' I
> found no differences with make check result of previous run. So I no
> need in any changes in the middle end part.
Vladimir, can you please inv
This adds -Og as optimization level targeted at the devel-compile-debug
cycle (formerly mostly tied to -O0 due to debug issues with even -O1).
Discussion on g...@gcc.gnu.org at least shows interest in this, so this
is a formal patch submission with a request for comments on the
implementation (no
Hi All,
Forgot to attch the patch.
2012/9/18 Yuri Rumyantsev :
> Hi All,
>
> This patch aims to fix all stability issues related to using the first
> scheduler in gcc
> for x86 target (there several reported issues related to this problem).
>
> Main idea of this activity is mostly to provide user
Hi All,
This patch aims to fix all stability issues related to using the first
scheduler in gcc
for x86 target (there several reported issues related to this problem).
Main idea of this activity is mostly to provide user a possibility to
safely turn on first scheduler for his codes. In some cases
1 - 100 of 122 matches
Mail list logo