Hi!
On Thu, 20 Jun 2013 15:08:16 -0700, Ian Lance Taylor wrote:
> Thomas Schwinge writes:
> > On Sun, 05 May 2013 18:55:09 -0700, Ian Lance Taylor wrote:
> >> Thomas Schwinge writes:
> >>
> >> >> libgcc/
> >> >> * fp-bit.c (unpack_d, pack_d): Properly preserve and restore a
> >> >>
A number of places in the rs6000 backend assume the value for a vec
splat can be found at element nunits-1 of a vector constant, which is
wrong for little-endian. This patch fixes them and the ICE found
when running altivec-consts.c on powerpc64le.
I've also updated the testcase so that it passes
At revision 200282.
Andrew --
there was a slight conflict with the new usage of is_binary_trait.
I think I resolved it properly, but double check.
-- Gaby
And this patch fixes the last patch to *really* only make function
descriptors when they are needed. Sigh. Bootstrapped and ran Go
testsuite on x86_64-unknown-linux-gnu. Committed to mainline and 4.8
branch.
Ian
diff -r 3081c3c4d620 go/expressions.cc
--- a/go/expressions.cc Thu Jun 20 17:08:43
On 06/20/2013 08:26 PM, Jason Merrill wrote:
Instantiation of a VLA capture in a template wasn't working properly;
this fixes it.
Along with this, which fixes the type of an instantiation of the MINUS_EXPR.
commit c37ab236406b9acf4423e80554055bf72d7be5fa
Author: Jason Merrill
Date: Thu Jun
Instantiation of a VLA capture in a template wasn't working properly;
this fixes it.
Tested x86_64-pc-linux-gnu, applying to trunk.
commit 9e88ab6e258122ce7d4a709919df56f7c1514c06
Author: Jason Merrill
Date: Thu Jun 13 09:29:11 2013 -0400
PR c++/55149
* decl.c (compute_array_index_
This is a follow up to the last big patch, adding function descriptors
to the Go frontend. This patch adds another compilation pass to only
create the function descriptors if they are needed: if the function is
exported, or if it is referenced in some way other than calling it.
Bootstrapped and ra
Thomas Schwinge writes:
> Hi!
>
> On Sun, 05 May 2013 18:55:09 -0700, Ian Lance Taylor wrote:
>> Thomas Schwinge writes:
>>
>> >> libgcc/
>> >> * fp-bit.c (unpack_d, pack_d): Properly preserve and restore a
>> >> NaN's payload.
>> >> * fp-bit.h [FLOAT] (QUIET_NAN): Correct value.
>>
>>
Hi!
On Sun, 05 May 2013 18:55:09 -0700, Ian Lance Taylor wrote:
> Thomas Schwinge writes:
>
> >> libgcc/
> >>* fp-bit.c (unpack_d, pack_d): Properly preserve and restore a
> >>NaN's payload.
> >>* fp-bit.h [FLOAT] (QUIET_NAN): Correct value.
>
> This is OK.
Joseph suggested these
Hi,
I have just backported r198683 from trunk to linaro/gcc-4_8-branch (as
rev 200267).
(It's the support of Address Sanitizer for ARM)
Thanks,
Christophe.
On Thu, Jun 20, 2013 at 02:26:12PM +0200, Matthias Klose wrote:
> Am 13.06.2013 11:42, schrieb Richard Sandiford:
> > "Bernhard Reutner-Fischer" writes:
> >> On 12 June 2013 20:20:50 Richard Sandiford
> >> wrote:
> >>> Matthias Klose writes:
> Index: config/mips/t-linux64
> ==
On 06/20/2013 09:09 PM, Roland Lutz wrote:
> Signed-off-by: Roland Lutz
> ---
> libjava/contrib/aot-compile.in |2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/libjava/contrib/aot-compile.in b/libjava/contrib/aot-compile.in
> index 91cfc67..2ee6739 100644
> --- a/lib
Signed-off-by: Roland Lutz
---
libjava/contrib/aot-compile.in |2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/libjava/contrib/aot-compile.in b/libjava/contrib/aot-compile.in
index 91cfc67..2ee6739 100644
--- a/libjava/contrib/aot-compile.in
+++ b/libjava/contrib/aot-compil
Since we poison "malloc" and friends in system.h, any C++ code that
includes a standard library header such as , which in turn
includes , will get poisoning errors due to lines like
#undef malloc
using ::malloc;
The solution is to include before poisoning those identifiers.
Oleg posted a
Thanks Vlad.
Vladimir Makarov writes:
> Index: lra.c
> ===
> --- lra.c (revision 200174)
> +++ lra.c (working copy)
> @@ -242,6 +242,42 @@ lra_delete_dead_insn (rtx insn)
>lra_set_insn_deleted (insn);
> }
>
> +/* Emit
It has always seemed odd that we didn't build a TEMPLATE_DECL for a
partial specialization, and with the ongoing concepts work it has become
more important to have one so that we can attach any requirements to it.
Note that there is still no way to get from the type of a partial
specialization
On 06/20/2013 01:23 PM, Jason Merrill wrote:
Since Gaby prefers SVN, let's keep using the SVN branch; it really isn't
much less convenient than a git-only branch. The main difference is
'git svn rebase'/'git svn dcommit' instead of 'git pull'/'git push'.
The one caveat is that git-svn historic
On 06/20/2013 11:14 AM, Andi Kleen wrote:
>>> +It should be only used for operands of type bool or atomic_flag. For
>>> +other types only part of the value may be set.
>>
>> @code{bool}. We have no definition for atomic_flag. Perhaps just
>> @code{char}
>> for now?
>>
>>> +It should be only use
Set $ac_aux_dir before use in libdecnumber/configure.
libdecnumber/configure uses $ac_aux_dir before it is set, causing incorrect
MISSING value. Fix with explicit AC_CONFIG_AUX_DIR.
Bootstrapped for c/c++. Okay for trunk?
libdecnumber/ChangeLog
2013-06-20 Simon Baldwin
* configure
> > +It should be only used for operands of type bool or atomic_flag. For
> > +other types only part of the value may be set.
>
> @code{bool}. We have no definition for atomic_flag. Perhaps just @code{char}
> for now?
>
> > +It should be only used for operands of type bool or atomic_flag and
Hello!
x86_64 ABI mandates long double returns in x87 register, so it is
invalid to use --mno-fp-ret-in-387. It just happened that vzeroupper
insertion mode-switching pass tripped on invalid register usage.
2013-06-20 Uros Bizjak
PR target/57655
* config/i386/i386.c (construct_contain
On 06/20/2013 06:20 AM, Andi Kleen wrote:
> gcc/:
> 2013-06-13 Andi Kleen
>
> * doc/extend.texi: Dont use __atomic_clear in HLE
> example. Fix typo.
Ok.
r~
On 06/20/2013 06:20 AM, Andi Kleen wrote:
> From: Andi Kleen
>
> Document that __atomic_clear and __atomic_test_and_set should
> only be used with bool.
>
> gcc/:
> 2013-06-13 Andi Kleen
>
> * doc/extend.texi: Document that __atomic_clear and
> __atomic_test_and_set should only
> +/* Returns true if there is a length mismatch among exprssions that are at
> the
> + same dimension and one the same side of the equal sign. The Array
> notation
> + lengths (LIST) is passed in as a 2D vector of trees. */
> +
> +static bool
> +cp_length_mismatch_in_expr_p (location_t loc
Looks good.
thanks,
David
On Thu, Jun 20, 2013 at 10:38 AM, Easwaran Raman wrote:
> In lipo mode, this patch updates the overall unit size only when the
> eventual function to which the callee is inlined is in primary module.
> This is to avoid the situation where the module growth budget is us
In lipo mode, this patch updates the overall unit size only when the
eventual function to which the callee is inlined is in primary module.
This is to avoid the situation where the module growth budget is used
up by inlines into auxiliary module functions that never get inlined
into some primary mo
On 06/20/2013 11:50 AM, Andrew Sutton wrote:
The c++-concepts SVN branch won't come in with a clone; you need to add it
to the fetch list with
git config --add remote.origin.fetch
refs/remotes/c++-concepts:refs/remotes/origin/c++-concepts
Then you can check out a local branch based on it.
Alr
This test is failing on targets with low branch costs. For those
targets we expand the code radically differently and the test really
doesn't make much sense -- much like vrp87.c.
This patch copies the list of targets where the test is meaningless from
vrp87.c.
Installed onto the trunk.
di
On 20/06/13 05:13, Peter Bergner wrote:
> On Thu, 2013-06-20 at 00:51 +0200, Torvald Riegel wrote:
>> On Wed, 2013-06-19 at 14:43 -0500, Peter Bergner wrote:
> I'd still like to hear from Andreas, whether the reentrant.c test case
> with both patches, now works on S390.
The patches fix the testca
It seems that one of the tests needed a small fix.
Attached is a corrected version.
On 20 June 2013 17:16, Michael Zolotukhin
wrote:
> Hi,
> I added two tests to verify we generate vector instructions when
> vector_loop is used.
> Is the patch ok with that change?
>
> Thanks, Michael
>
> On 5 Jun
On May 30, 2013, at 12:59 PM, Cesar Philippidis wrote:
> Here is a patch from our backlog at Mentor Graphics that addresses a
> libgomp issue where setting ENABLE_LTO=1 in site.exp causes the following
> error with dejagnu
> Is it OK for trunk?
Ok.
Committed revision 200253.
> 2013-05-30 Ia
On May 30, 2013, at 1:02 PM, Cesar Philippidis wrote:
> Here is a patch from our backlog that addresses a libitm issue where
> setting ENABLE_LTO=1
> Is this OK for trunk?
Ok.
Committed revision 200252.
> Cesar Philippidis
>
>
> 2013-05-30 Iain Sandoe
> Cesar Philippidis
>
>
On 06/20/2013 02:49 AM, Torvald Riegel wrote:
> commit c8352d4d9fa5cfa3453a61c581956835de9753e5
> Author: Torvald Riegel
> Date: Thu Jun 20 00:46:59 2013 +0200
>
> libitm: Handle HTM fastpath in status query functions.
Ok.
r~
On 06/19/2013 07:57 AM, Torvald Riegel wrote:
> commit 185af84e365e1bae31aea5afd6e67e81f3c32c72
> Author: Torvald Riegel
> Date: Wed Jun 19 16:42:24 2013 +0200
>
> libitm: Fix handling of reentrancy in the HTM fastpath.
>
> PR libitm/57643
Ok.
r~
The newly added forwprop-28.c test was missing a dg-final clause to
remove the debugging dump is scans. This trivial patch adds the missing
dg-final.
Installed as obvious.
diff --git a/gcc/testsuite/ChangeLog b/gcc/testsuite/ChangeLog
index 6dca973..5adbc01 100644
--- a/gcc/testsuite/ChangeLo
> The c++-concepts SVN branch won't come in with a clone; you need to add it
> to the fetch list with
>
> git config --add remote.origin.fetch
> refs/remotes/c++-concepts:refs/remotes/origin/c++-concepts
>
> Then you can check out a local branch based on it.
Already did that. I probably need to do
sHi,
On Thu, Jun 20, 2013 at 03:47:11PM +0100, Marcus Shawcroft wrote:
> Hi, I've been looking at an issue in mysql compilation which appears
> to be due to this patch.
>
> On 10 May 2013 18:27, Martin Jambor wrote:
> > Hi,
> >
> > as we discover targets of previously indirect calls in ipa-inli
Jason Merrill writes:
| On 06/20/2013 11:22 AM, Gabriel Dos Reis wrote:
| > Jason Merrill writes:
| >
| > | On 06/20/2013 10:17 AM, Andrew Sutton wrote:
| > | > It looks like I should be able to switch directly to the c++-concepts
| > | > branch. Or is there some configuration that has to happen
On 06/20/2013 11:22 AM, Gabriel Dos Reis wrote:
Jason Merrill writes:
| On 06/20/2013 10:17 AM, Andrew Sutton wrote:
| > It looks like I should be able to switch directly to the c++-concepts
| > branch. Or is there some configuration that has to happen on the
| > remote site?
|
| The c++-concep
Jason Merrill writes:
| On 06/20/2013 10:17 AM, Andrew Sutton wrote:
| > It looks like I should be able to switch directly to the c++-concepts
| > branch. Or is there some configuration that has to happen on the
| > remote site?
|
| The c++-concepts SVN branch won't come in with a clone; you nee
Andrew Sutton writes:
| That works. I think the current patch addresses all of Jason's comments.
OK, that sounds good.
| I'll also create a github version of this branch, so can avoid email patches.
Well, actually I prefer the email patches because I can read them them
right away when I have o
On 06/20/2013 10:17 AM, Andrew Sutton wrote:
It looks like I should be able to switch directly to the c++-concepts
branch. Or is there some configuration that has to happen on the
remote site?
The c++-concepts SVN branch won't come in with a clone; you need to add
it to the fetch list with
g
Hello Everyone,
One of the struct field used by array notations (struct inv_list) was
accidentaly set to the type enum rid, but the correct value should be enum
tree_code. This was changed in the following patch. This was causing a warning.
The field is only used by C++ (which patch is s
Hi, I've been looking at an issue in mysql compilation which appears
to be due to this patch.
On 10 May 2013 18:27, Martin Jambor wrote:
> Hi,
>
> as we discover targets of previously indirect calls in ipa-inline and
> ipa-cp, we sometimes figure out that the targets are not a function.
> One ty
Ping.
Cesar
On 6/7/13 9:50 AM, Cesar Philippidis wrote:
> On 6/6/13 9:00 AM, Richard Earnshaw wrote:
>> The pipeline offset is 4 for Thumb2 as well. So at the very least you
>> need to explain why your change doesn't apply then as well.
>
> Yes some context is lost in that comment. Thunks are
Index: ChangeLog
===
--- ChangeLog (revision 200246)
+++ ChangeLog (working copy)
@@ -1,3 +1,7 @@
+2013-06-20 Jan Hubicka
+
+ * lto-cgraph.c (input_symtab): Do not set cgraph state.
+
2013-06-20 Joern Rennecke
I didn't know it existed! Even better. Except that I'm not a git
expert. I'm reading through the docs on that site while I clone the
repo.
It looks like I should be able to switch directly to the c++-concepts
branch. Or is there some configuration that has to happen on the
remote site?
How will t
Ping.
Cesar
On 5/30/13 1:02 PM, Cesar Philippidis wrote:
> Here is a patch from our backlog that addresses a libitm issue where
> setting ENABLE_LTO=1 in site.exp causes the following error with dejagnu:
>
> ERROR: (DejaGnu) proc "libitm_target_compile linker_plugin19344.c
> linker_plugin1934
Ping.
Cesar
On 5/30/13 12:59 PM, Cesar Philippidis wrote:
> Here is a patch from our backlog at Mentor Graphics that addresses a
> libgomp issue where setting ENABLE_LTO=1 in site.exp causes the following
> error with dejagnu:
>
> ERROR: (DejaGnu) proc "libgomp_target_compile linker_plugin926
On 06/20/2013 09:18 AM, Andrew Sutton wrote:
I'll also create a github version of this branch, so can avoid email patches.
Why there rather than in the gcc.gnu.org git repository?
http://gcc.gnu.org/wiki/GitMirror
Jason
Hi all,
This patch adjusts the andsi3, iorsi3, xorsi3 patterns for -mrestrict-it. It
is done by adding 16-bit alternatives that are fit be cond_exec'd when
arm_restrict_it is on.
Tested arm-none-eabi on model and qemu with ARMv7 and ARMv8. Also tested as
part of the series with bootstrap on a Cor
On Thu, 2013-06-20 at 11:49 +0200, Torvald Riegel wrote:
> You're right, that was missing for x86 as well. Please see the updated
> second patch that is attached. It additionally checks htm_fastpath to
> see whether we are actually using the HTM. This variable is initialized
> to the value that
Hi all,
This patch adjusts the mov* patterns in the arm backend to generate code
appropriate for -mrestrict-it. The rules are:
moves between any two registers are allowed to be in IT blocks. mov immediate
are allowed if the immediate is 8-wide and the register is a low register.
The splitters th
From: Andi Kleen
The HLE example in the manual only commits when using bool
for the flag, because __atomic_clear only writes bool, and
HLE requires the acquire and release to match.
So when the example is copied with e.g. an int variable it
does not commit and causes slower than expected perform
From: Andi Kleen
Document that __atomic_clear and __atomic_test_and_set should
only be used with bool.
gcc/:
2013-06-13 Andi Kleen
* doc/extend.texi: Document that __atomic_clear and
__atomic_test_and_set should only be used with bool.
---
gcc/doc/extend.texi | 6 ++
1
That works. I think the current patch addresses all of Jason's comments.
I'll also create a github version of this branch, so can avoid email patches.
On Thu, Jun 20, 2013 at 8:09 AM, Gabriel Dos Reis wrote:
> Jason Merrill writes:
>
> | On 06/20/2013 01:30 AM, Gabriel Dos Reis wrote:
> | > As
Hi,
I added two tests to verify we generate vector instructions when
vector_loop is used.
Is the patch ok with that change?
Thanks, Michael
On 5 June 2013 18:10, Michael Zolotukhin wrote:
> I'll prepare some tests shortly,
> What about the rest questions?
>
> Thanks, Michael
>
> On 15 May 2013 1
Jason Merrill writes:
| On 06/20/2013 01:30 AM, Gabriel Dos Reis wrote:
| > As I discussed
| > with Andrew a couple of weeks ago, I have been holding back the
| > merge from trunk because he has these patch series in the queue.
|
| Incidentally, since the code is going onto a branch, we don't re
On 06/20/2013 01:30 AM, Gabriel Dos Reis wrote:
As I discussed
with Andrew a couple of weeks ago, I have been holding back the
merge from trunk because he has these patch series in the queue.
Incidentally, since the code is going onto a branch, we don't really
need to delay checkins based on c
On 06/20/2013 04:49 AM, Andreas Schwab wrote:
Jeff Law writes:
+/* { dg-final { scan-tree-dump-times "Replaced" 8 "forwprop1"} } */
$ grep -c Replaced forwprop-28.c.022t.forwprop1
16
;; Function test (test, funcdef_no=0, decl_uid=1388, symbol_order=0)
Replaced 'rotate_7 == 0' with '_6 =
Hi,
On Thu, Jun 20, 2013 at 01:32:38PM +0200, Jan Hubicka wrote:
> >
> > 2013-06-11 Martin Jambor
> >
> > PR tree-optimization/57358
> > * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
> > (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
> >
Am 13.06.2013 11:42, schrieb Richard Sandiford:
> "Bernhard Reutner-Fischer" writes:
>> On 12 June 2013 20:20:50 Richard Sandiford
>> wrote:
>>> Matthias Klose writes:
Index: config/mips/t-linux64
===
--- config/mips
>
> 2013-06-11 Martin Jambor
>
> PR tree-optimization/57358
> * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
> (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
> (ipa_analyze_params_uses): Generate pessimistic info when true.
>
>
Hi,
for Android:
FAIL: g++.dg/tls/thread_local4.C -std=gnu++11 (test for excess errors)
Excess errors:
<>/gcc/testsuite/g++.dg/tls/thread_local4.C:31:26: error:
'pthread_testcancel' was not declared in this scope
<>/gcc/testsuite/g++.dg/tls/thread_local4.C:40:24: error:
'pthread_cancel' was not d
This speeds up streamer_read_uhwi (top in mozilla LTO profile) by
delaying the section overrun check and inlining streamer_read_uchar
manually, performing CSE and optimizing the 1-byte case.
LTO bootstrapped on x86_64-unknown-linux-gnu, applied.
Richard.
2013-06-20 Richard Biener
*
> Hi,
>
> this is the simplest fix for the PR which happens because there is no
> VDEF on a stmt if a particular function is not optimized. I'd like to
> fix the bug with it on the branch. Bootstrapped and tested on
> x86_64-linux. OK?
>
> Thanks,
>
> Martin
>
>
> 2013-06-11 Martin Jambor
> Hi,
>
> PR 57539 revealed two problems with remapping reference descriptors
> during cloning of trees of inlined call graph nodes. First, when
> indirect inlining is involved, we happily remove the reference
> descriptor itself by calling ipa_free_edge_args_substructures in
> ipa_propagate_indi
Jeff Law writes:
> +/* { dg-final { scan-tree-dump-times "Replaced" 8 "forwprop1"} } */
$ grep -c Replaced forwprop-28.c.022t.forwprop1
16
;; Function test (test, funcdef_no=0, decl_uid=1388, symbol_order=0)
Replaced 'rotate_7 == 0' with '_6 == 0'
Replaced '_6 == 0' with 'code_5(D) != 22'
On Tue, 18 Jun 2013 17:10:37 +0100
Richard Earnshaw wrote:
> On 18/06/13 16:42, Julian Brown wrote:
> > Hi,
> >
> > The following patch removed pool_range/neg_pool_range attributes
> > from several instructions as a cleanup, which I believe to have been
> > incorrect:
> >
> > http://gcc.gnu.org/m
On Thu, Jun 20, 2013 at 12:11 PM, Tobias Burnus wrote:
> gfortran failed to correctly read the file
>line1,1,
>line2
> with DOS (\r\n) line endings. As the code already set EOL for "\r",
> finish_list_read didn't call eat_line. Result: The attempt to read "line2"
> actually accessed the la
We have to be careful to not end up inserting expressions with
abnormal SSA names. Thus the following makes niter analysis
more careful what it puts into number-of-iteration expressions.
Bootstrapped and tested on x86_64-unknown-linux-gnu, applied to trunk.
Richard.
2013-06-20 Richard Biener
On Wed, 2013-06-19 at 22:13 -0500, Peter Bergner wrote:
> On Thu, 2013-06-20 at 00:51 +0200, Torvald Riegel wrote:
> > On Wed, 2013-06-19 at 14:43 -0500, Peter Bergner wrote:
> > >> I'm having trouble seeing why/when _ITM_inTransaction() is
> > >> returning something other than inIrrevocableTransa
Follow up to http://gcc.gnu.org/ml/fortran/2013-06/msg00027.html - for
some reason this part of the patch got lost.
Committed as Rev. 200234.
Tobias
Index: gcc/fortran/ChangeLog
===
--- gcc/fortran/ChangeLog (Revision 200233)
+++ g
gfortran failed to correctly read the file
line1,1,
line2
with DOS (\r\n) line endings. As the code already set EOL for "\r",
finish_list_read didn't call eat_line. Result: The attempt to read
"line2" actually accessed the last byte of line one, namely "\n", which
it regarded as zero-size
On Wed, Jun 19, 2013 at 9:58 PM, Joern Rennecke
wrote:
> Quoting Michael Matz :
>
>> That's not good. You now have different order of parameters between
>> anti_dependence and canon_anti_dependence. That will be mightily
>> confusing, please instead change the caller. Currently these predicates
Ping again?
On 13/6/11 5:20 PM, Bernhard Reutner-Fischer wrote:
> ping, CCing middle-end maintainers for review.
>
> On 31 May 2013 10:13, Chung-Lin Tang wrote:
>> On 13/5/15 8:12 PM, Richard Sandiford wrote:
>>> Chung-Lin Tang writes:
On 13/5/10 6:37 PM, Richard Sandiford wrote:
> Chu
76 matches
Mail list logo
