Hi there,
When compile below simple code:
terguo01@terry-pc01:mtpcs-frame$ cat test.c
int main(void)
{
return 0;
}
I got ICE with option -mtpcs-leaf-frame (no error if remove this option).
terguo01@terry-pc01:mtpcs-frame$
2014-12-07 12:51 GMT+03:00 Richard Sandiford rdsandif...@googlemail.com:
Ilya Enkovich enkovich@gmail.com writes:
diff --git a/gcc/config/i386/i386.md b/gcc/config/i386/i386.md
index 88435d6..9019ed8 100644
--- a/gcc/config/i386/i386.md
+++ b/gcc/config/i386/i386.md
@@ -10958,6 +10958,24
On Thu, 4 Dec 2014, Richard Biener wrote:
This is the last patch, moving stpcpy folding. There are still
string function foldings left but those are exclusively GENERIC
now with no chance of advertedly recursing from GIMPLE folding
via GENERIC folding back to GIMPLE folding. I'll deal
On Thu, Dec 4, 2014 at 8:35 PM, Ilya Verbin iver...@gmail.com wrote:
Hi,
On 30 Sep 18:53, Ilya Verbin wrote:
This patch creates 2 vectors with decls: offload_funcs and offload_vars.
libgomp will use addresses from these arrays to look up offloaded code.
During the compilation they are
On Thu, Dec 4, 2014 at 9:06 PM, Jan Hubicka hubi...@ucw.cz wrote:
On December 4, 2014 7:00:12 PM CET, Jan Hubicka hubi...@ucw.cz wrote:
Hi,
this patch fixes implementation of shift in sreal.h that ICEs on 0
value.
Also the comment is copied from shift_right and does not make sense for
2014-12-08 12:18 GMT+03:00 Richard Biener richard.guent...@gmail.com:
On Thu, Dec 4, 2014 at 8:35 PM, Ilya Verbin iver...@gmail.com wrote:
Hi,
On 30 Sep 18:53, Ilya Verbin wrote:
This patch creates 2 vectors with decls: offload_funcs and offload_vars.
libgomp will use addresses from these
On Mon, Dec 08, 2014 at 01:28:10PM +0400, Ilya Verbin wrote:
2014-12-08 12:18 GMT+03:00 Richard Biener richard.guent...@gmail.com:
On Thu, Dec 4, 2014 at 8:35 PM, Ilya Verbin iver...@gmail.com wrote:
Hi,
On 30 Sep 18:53, Ilya Verbin wrote:
This patch creates 2 vectors with decls:
2014-12-06 1:55 GMT+03:00 Jeff Law l...@redhat.com:
On 12/02/14 06:40, Ilya Enkovich wrote:
Hi,
For strchr calls bounds of the first argument are considered as returned
which is wrong because NULL may be returned. This patch fixes that.
Bootstrapped and tested on x86_64-unknown-linux-gnu.
On 11/11/14 10:38, Alex Velenko wrote:
From 98bb6d7323ce79e28be8ef892b919391ed857e1f Mon Sep 17 00:00:00 2001
From: Alex Velenko alex.vele...@arm.com
Date: Fri, 31 Oct 2014 18:43:32 +
Subject: [PATCH] [AArch64, RTL] Bics instruction generation for aarch64
Hi,
This patch adds rtl
2014-12-08 Richard Biener rguent...@suse.de
Merge from trunk r218262 through r218477.
Hello,
this surprisingly simple patch fixes PR ipa/64049. The root cause seems to be
that in this test case we try to devirtualize a method call on a return value.
Boot-Strapped and regression-tested on X86_64-linux-gnu.
OK for trunk?
Thanks
Bernd.
Hello,
this surprisingly simple patch fixes PR ipa/64049. The root cause seems to be
that in this test case we try to devirtualize a method call on a return value.
Boot-Strapped and regression-tested on X86_64-linux-gnu.
OK for trunk?
Thanks
Bernd.
Again with changelog. Sorry.
Ping
2014-11-26 15:35 GMT+03:00 Ilya Enkovich enkovich@gmail.com:
On 25 Nov 15:03, Ilya Enkovich wrote:
2014-11-25 14:11 GMT+03:00 Richard Biener richard.guent...@gmail.com:
On Tue, Nov 25, 2014 at 11:19 AM, Ilya Enkovich enkovich@gmail.com
wrote:
Ok, then it's get_for_asmname
On 08 Dec 11:30, Ilya Enkovich wrote:
2014-12-07 12:51 GMT+03:00 Richard Sandiford rdsandif...@googlemail.com:
Ilya Enkovich enkovich@gmail.com writes:
diff --git a/gcc/config/i386/i386.md b/gcc/config/i386/i386.md
index 88435d6..9019ed8 100644
--- a/gcc/config/i386/i386.md
+++
Hi all,
As the subject says, this just fixes a typo in the fprofile-generate
option name and rewords the text in the next sentence a bit.
Ok to commit?
Thanks,
KyrillIndex: htdocs/gcc-5/changes.html
===
RCS file:
Hi Kyrill,
Hi all,
As the subject says, this just fixes a typo in the fprofile-generate
option name and rewords the text in the next sentence a bit.
Ok to commit?
Thanks,
Kyrill
I think this kind of change does not really need approval.
But, maybe you should also look at fixing the
On Mon, Dec 8, 2014 at 11:24 AM, Bernd Edlinger
bernd.edlin...@hotmail.de wrote:
Hi Kyrill,
Hi all,
As the subject says, this just fixes a typo in the fprofile-generate
option name and rewords the text in the next sentence a bit.
Ok to commit?
Thanks,
Kyrill
I think this kind of
I have installed the following patch on the 4.8 branch after verifying
it still builds ok and enables graphite with system ISL 0.14 and
system ISL 0.12.2 (and the corresponding cloog).
Richard.
2014-12-08 Richard Biener rguent...@suse.de
Backport from 4.9 branch
2014-12-04
On 08/12/14 08:24, Terry Guo wrote:
Hi there,
When compile below simple code:
terguo01@terry-pc01:mtpcs-frame$ cat test.c
int main(void)
{
return 0;
}
I got ICE with option -mtpcs-leaf-frame (no error if remove this option).
terguo01@terry-pc01:mtpcs-frame$
On 08/12/14 11:26, Ramana Radhakrishnan wrote:
On Mon, Dec 8, 2014 at 11:24 AM, Bernd Edlinger
bernd.edlin...@hotmail.de wrote:
Hi Kyrill,
Hi all,
As the subject says, this just fixes a typo in the fprofile-generate
option name and rewords the text in the next sentence a bit.
Ok to commit?
On Thu, 4 Dec 2014, Joseph Myers wrote:
On Thu, 4 Dec 2014, Richard Biener wrote:
OTOH this also means the user cannot provide a conforming
implementation on his own and get that used by GCC without editing
system headers or including a header with -isystem or similar
tricks.
Well -
On Mon, Dec 08, 2014 at 01:24:12PM +0100, Richard Biener wrote:
I'm not pushing this further for stage3, but for stage1 I'd like
to eventually address this by splitting up builtin_info_type's
'implicit_p' into a flags array providing implicit_p, declared_p,
used_p and maybe
On Mon, Dec 8, 2014 at 11:57 AM, Bernd Edlinger
bernd.edlin...@hotmail.de wrote:
Hello,
this surprisingly simple patch fixes PR ipa/64049. The root cause seems to
be
that in this test case we try to devirtualize a method call on a return
value.
Boot-Strapped and regression-tested
On Mon, 8 Dec 2014, Jakub Jelinek wrote:
On Mon, Dec 08, 2014 at 01:24:12PM +0100, Richard Biener wrote:
I'm not pushing this further for stage3, but for stage1 I'd like
to eventually address this by splitting up builtin_info_type's
'implicit_p' into a flags array providing implicit_p,
On Mon, Dec 08, 2014 at 01:54:21PM +0100, Richard Biener wrote:
On Mon, 8 Dec 2014, Jakub Jelinek wrote:
On Mon, Dec 08, 2014 at 01:24:12PM +0100, Richard Biener wrote:
I'm not pushing this further for stage3, but for stage1 I'd like
to eventually address this by splitting up
Jeff Law l...@redhat.com writes:
On 12/04/14 15:42, Rainer Orth wrote:
David Malcolm dmalc...@redhat.com writes:
dejagnu.h assumed -fgnu89-inline until a recent upstream fix;
see http://lists.gnu.org/archive/html/dejagnu/2014-10/msg00011.html
Remove the workaround from jit.exp that used
Hi guys,
Could you please take a look at this issue? This fix is still urgent
for Android.
2014-12-01 18:25 GMT+03:00 Varvara Rainchik varvara.s.rainc...@gmail.com:
Hi Jakub,
Do you think this patch is ok for upstream:
2014-12-01 Varvara Rainchik varvara.rainc...@intel.com
*
On 3 December 2014 at 10:30, Alan Lawrence alan.lawre...@arm.com wrote:
On Wed, Nov 26, 2014 at 04:35:50PM +, James Greenhalgh wrote:
Why do we want to turn off folding for the V4SF/V2SF/V2DF modes of these
intrinsics? There should be no difference between the mid-end definition
and the
On Mon, Dec 08, 2014 at 04:30:46PM +0300, Varvara Rainchik wrote:
Hi guys,
Could you please take a look at this issue? This fix is still urgent
for Android.
I'm afraid it will break those targets where emutls is not on, but the C
library doesn't support TLS. I think it is acceptable not to
A lot of old user code clobbers the carry bit without telling the compiler
about it. This used to just work, because the compiler never used the bit
outside of a single RTL instruction. But that will change. Let's clobber
the carry bit in every asm; this isn't very expensive.
2014-12-08
These splitters are only used when for some reason we generated a bdnz
loop but we cannot actually use one (CTR is already used for something
else, for example). This shouldn't happen often, so making the split
sequence one insn longer isn't a big deal.
2014-12-08 Segher Boessenkool
This series makes the PowerPC port deal with the {add,subf}{c,ic,e,ze,me}
instructions directly, not as bigger primitives (which used to be emitted as
multiple machine instructions).
It does this by expanding multi-precision add/sub/neg as well as the unsigned
wordlength sCOND to separate machine
This means we do not allow GPR0 as base address of those anymore. The
alternative is to not allow the carry bit to be live over any lo_sum,
which is more expensive.
2014-12-08 Segher Boessenkool seg...@kernel.crashing.org
gcc/
PR target/64180
* config/rs6000/darwin.md
On Mon, Dec 8, 2014 at 9:18 AM, Segher Boessenkool
seg...@kernel.crashing.org wrote:
A lot of old user code clobbers the carry bit without telling the compiler
about it. This used to just work, because the compiler never used the bit
outside of a single RTL instruction. But that will change.
On Mon, Dec 8, 2014 at 9:18 AM, Segher Boessenkool
seg...@kernel.crashing.org wrote:
This means we do not allow GPR0 as base address of those anymore. The
alternative is to not allow the carry bit to be live over any lo_sum,
which is more expensive.
2014-12-08 Segher Boessenkool
On Mon, Dec 8, 2014 at 9:18 AM, Segher Boessenkool
seg...@kernel.crashing.org wrote:
These splitters are only used when for some reason we generated a bdnz
loop but we cannot actually use one (CTR is already used for something
else, for example). This shouldn't happen often, so making the
This implements the DW_TAG_atomic_type for C11 _Atomic proposal as adopted
in the latest DWARF5 draft. http://dwarfstd.org/ShowIssue.php?issue=131112.1
This is much simpler than my previous patch. Thanks to Andreas cleanups
for PR 63300 adding new qualifier support to dwarf2out.c is almost
OK.
Jason
On Mon, Dec 8, 2014 at 9:18 AM, Segher Boessenkool
seg...@kernel.crashing.org wrote:
This means we can no longer add GPR0+imm. Register allocation will have
to use a different register.
2014-12-08 Segher Boessenkool seg...@kernel.crashing.org
gcc/
PR target/64180
*
On Tue, 18 Nov 2014, Richard Biener wrote:
I'll try to replace some more TYPE_MODE during stage3...
Btw, a convenience would be to be able to write
HONOR_NANS (type)
thus effectively make HONOR_* inline functions with a machine_mode
and a type overload (and the type overload properly
Peter tells me it was an artifact of old versions of the DFP code. The
condition can never be false; delete it.
2014-12-08 Segher Boessenkool seg...@kernel.crashing.org
gcc/
* config/rs6000/rs6000.md (*addmode3): Remove condition.
---
gcc/config/rs6000/rs6000.md | 2 +-
1 file
This splits the addic patterns from the add/addi patterns, and merges the
insns and splitters.
Also change type compare to add for addic. (the dot form).
2014-12-08 Segher Boessenkool seg...@kernel.crashing.org
gcc/
PR target/64180
* config/rs6000/rs6000.md
These now are the only remaining patterns that have type compare.
And they shouldn't: two is a better type for them.
2014-12-08 Segher Boessenkool seg...@kernel.crashing.org
gcc/
* config/rs6000/rs6000.md (*anddi3_2rld_dot, *anddi3_rld_dot2):
Change type from compare to two.
This means we can no longer add GPR0+imm. Register allocation will have
to use a different register.
2014-12-08 Segher Boessenkool seg...@kernel.crashing.org
gcc/
PR target/64180
* config/rs6000/rs6000.md (*addmode3_internal1): Remove addic
alternative.
---
This implements addc, addic, adde, addze, addme, subfc, subfic, subfe,
subfze, subfme. Some of those in multiple forms: canonical RTL for the
immediate insns has four forms; subfX has a special form for subtracting
a register from itself.
All these new insns get type add for now.
2014-12-08
2014-12-08 Segher Boessenkool seg...@kernel.crashing.org
gcc/
* config/rs6000/40x.md (ppc403-compare): Remove compare.
config/rs6000/440.md (ppc440-compare): Remove compare.
config/rs6000/476.md (ppc476-compare): Remove compare.
config/rs6000/601.md
Now that everything is in place for letting GCC use the carry-using and
carry-producing machine insns as separate RTL insns, switch over all
remaining patterns that clobber CA without telling the compiler. This
are the multiple-precision add/sub/neg patterns, and the various eq/ne/
On Mon, Dec 08, 2014 at 09:44:44AM -0500, David Edelsohn wrote:
-;; Discourage ai/addic because of carry but provide it in an alternative
-;; allowing register zero as source.
(define_insn *addmode3_internal1
- [(set (match_operand:GPR 0 gpc_reg_operand =r,r,?r,r)
- (plus:GPR
On Mon, Dec 8, 2014 at 10:01 AM, Segher Boessenkool
seg...@kernel.crashing.org wrote:
On Mon, Dec 08, 2014 at 09:44:44AM -0500, David Edelsohn wrote:
-;; Discourage ai/addic because of carry but provide it in an alternative
-;; allowing register zero as source.
(define_insn
On Mon, Dec 8, 2014 at 9:18 AM, Segher Boessenkool
seg...@kernel.crashing.org wrote:
This splits the addic patterns from the add/addi patterns, and merges the
insns and splitters.
Also change type compare to add for addic. (the dot form).
2014-12-08 Segher Boessenkool
Vladimir Makarov vmaka...@redhat.com writes:
The following patch fixes
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64157
After calling target_reinit from save_target_globals for switchable
targets (as ppc), a lot of ira data (register sets, classes etc) become
undefined. After
On Mon, Dec 8, 2014 at 9:18 AM, Segher Boessenkool
seg...@kernel.crashing.org wrote:
This implements addc, addic, adde, addze, addme, subfc, subfic, subfe,
subfze, subfme. Some of those in multiple forms: canonical RTL for the
immediate insns has four forms; subfX has a special form for
On Mon, Dec 8, 2014 at 10:41 AM, David Edelsohn dje@gmail.com wrote:
On Mon, Dec 8, 2014 at 9:18 AM, Segher Boessenkool
seg...@kernel.crashing.org wrote:
This implements addc, addic, adde, addze, addme, subfc, subfic, subfe,
subfze, subfme. Some of those in multiple forms: canonical RTL
Is it ok to add GCC_CHECK_EMUTLS test in libgomp/configure.ac or
should I add in tls.m3 a similair test that would be used only in
libgomp?
2014-12-08 17:03 GMT+03:00 Jakub Jelinek ja...@redhat.com:
On Mon, Dec 08, 2014 at 04:30:46PM +0300, Varvara Rainchik wrote:
Hi guys,
Could you please
On Fri, 2014-11-21 08:45:11 -0500, Andrew MacLeod amacl...@redhat.com wrote:
During the flattening of optabs.h, I updated all the config/* files
which were affected. I've been getting spurious failures with
config-list.mk where my changes would disappear and tracked down
why.
I was
On Mon, Dec 08, 2014 at 07:01:09PM +0300, Varvara Rainchik wrote:
Is it ok to add GCC_CHECK_EMUTLS test in libgomp/configure.ac or
should I add in tls.m3 a similair test that would be used only in
libgomp?
I think config/tls.m4 would be a better place, but doing it in some way
where the users
On Sat, 2014-12-06 at 15:29 -0500, Ulrich Drepper wrote:
This patch broken out of one I sent earlier with some extensions. It
contains only little cleanups to the libgccjit code.
When creating the linker command line the code now uses an auto_vec
instead of the fixed size array.
The
On 11/28/2014 10:32 AM, Richard Biener wrote:
On Thu, Nov 27, 2014 at 6:08 PM, Martin Liška mli...@suse.cz wrote:
On 11/21/2014 04:21 PM, Martin Liška wrote:
On 11/21/2014 04:02 PM, Richard Biener wrote:
On Fri, Nov 21, 2014 at 3:39 PM, Martin Liška mli...@suse.cz wrote:
Hello.
Ok, this
On 11/14/2014 04:23 PM, Martin Liška wrote:
On 11/14/2014 03:05 PM, Martin Liška wrote:
On 11/11/2014 01:26 PM, mliska wrote:
gcc/ChangeLog:
2014-11-12 Martin Liska mli...@suse.cz
* auto-profile.c: Include cgraph_summary.h.
* cgraph.c: Likewise.
* cgraphbuild.c: Likewise.
*
On 11/14/2014 04:24 PM, Martin Liška wrote:
On 11/14/2014 03:09 PM, Martin Liška wrote:
On 11/13/2014 05:04 PM, Jan Hubicka wrote:
+ if (!inline_summary_summary)
+inline_summary_summary = (inline_summary_cgraph_summary *)
inline_summary_cgraph_summary::create_ggc (symtab);
Hehe, this
On 11/14/2014 04:23 PM, Martin Liška wrote:
On 11/14/2014 03:04 PM, Martin Liška wrote:
On 11/13/2014 04:50 PM, Jan Hubicka wrote:
gcc/ChangeLog:
2014-11-12 Martin Liska mli...@suse.cz
* Makefile.in: New object file is added.
* cgraph.h (symbol_table::allocate_cgraph_symbol):
Hi all,
please find attached a more elaborate patch for pr60255. I totally agree that
my first attempt was just scratching the surface of the work needed.
This patch also is *not* complete, but because I am really new to gfortran
patching, I don't want to present a final patch only to learn
On 2014-12-08 10:41 AM, Richard Sandiford wrote:
Vladimir Makarov vmaka...@redhat.com writes:
The following patch fixes
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64157
After calling target_reinit from save_target_globals for switchable
targets (as ppc), a lot of ira data (register
On Sat, 2014-12-06 at 15:56 -0500, Ulrich Drepper wrote:
This patch supercedes the patch I sent earlier this week to add
dependencies to the linker command line. The implementation is
different.
First, based on Dave's comment that he wants to keep the interface
simple, to enable the linker
This patch from Chris Manghane fixes a crash-on-invalid in the Go
frontend. This is GCC PR 64198. Bootstrapped and ran Go testsuite on
x86_64-unknown-linux-gnu. Committed to mainline.
Ian
diff -r 2a85649c19e1 go/parse.cc
--- a/go/parse.cc Sun Nov 30 17:28:00 2014 -0800
+++ b/go/parse.cc
On 08/12/14 10:33, Richard Earnshaw wrote:
On 11/11/14 10:38, Alex Velenko wrote:
From 98bb6d7323ce79e28be8ef892b919391ed857e1f Mon Sep 17 00:00:00 2001
From: Alex Velenko alex.vele...@arm.com
Date: Fri, 31 Oct 2014 18:43:32 +
Subject: [PATCH] [AArch64, RTL] Bics instruction generation
The high-level overview diagram in jit/notes.txt showed the
release of the JIT mutex in where I'd like it to be, rather
than where it currently is.
Fix the diagram to reflect reality, and clarify various other things.
Committed to trunk as r218488.
gcc/jit/ChangeLog:
* notes.txt: Show
On 12/06/14 06:47, Sebastian Pop wrote:
Jeff Law wrote:
OK to commit. Thanks for your patience.
Can you follow-up with a change which throttles this optimization
when -Os is in effect. You can check optimize_function_for_size_p
(cfun) and simply avoid the backward traversal or you could
Committed to trunk as r218489.
gcc/jit/ChangeLog:
* libgccjit++.h: Indent the forward declarations of the classes to
show the inheritance hierarchy.
---
gcc/jit/libgccjit++.h | 21 -
1 file changed, 12 insertions(+), 9 deletions(-)
diff --git
On 12/05/14 17:16, Segher Boessenkool wrote:
On Fri, Dec 05, 2014 at 03:31:54PM -0700, Jeff Law wrote:
Combine does not consider combining 9 into 7 because there is no LOG_LINK
between them (the link for r88 is between 8 and 7 already).
OK, yea, that's a long standing design decision. We
On 12/05/14 17:58, Segher Boessenkool wrote:
On Fri, Dec 05, 2014 at 01:40:56PM -0700, Jeff Law wrote:
My first thought would be to allow both and have combine swap the order
in the vector if recog doesn't recognize the pattern. One could argue
we could go through a full permutation of
On 12/8/2014 11:23 AM, Jan-Benedict Glaw wrote:
On Fri, 2014-11-21 08:45:11 -0500, Andrew MacLeod amacl...@redhat.com wrote:
During the flattening of optabs.h, I updated all the config/* files
which were affected. I've been getting spurious failures with
config-list.mk where my changes would
On 12/08/14 04:00, Ilya Enkovich wrote:
2014-11-26 Ilya Enkovich ilya.enkov...@intel.com
PR bootstrap/63995
* tree-chkp.c (chkp_make_static_bounds): Share bounds var
between nodes sharing assembler name.
gcc/testsuite
2014-11-26 Ilya Enkovich
On 12/05/14 15:41, Caroline Tice wrote:
When hot/cold function splitting occurs, a symbol is generated for the
cold partition, and gets output in the assembly debug info, but the
symbol currently gets a size of 0 and a type of NOTYPE, as in this
example (on x86_64-linux) from the
2014-12-09 0:00 GMT+03:00 Jeff Law l...@redhat.com:
On 12/08/14 04:00, Ilya Enkovich wrote:
2014-11-26 Ilya Enkovich ilya.enkov...@intel.com
PR bootstrap/63995
* tree-chkp.c (chkp_make_static_bounds): Share bounds var
between nodes sharing assembler name.
On 12/01/14 11:44, John David Anglin wrote:
To do unconditionally set frame_read? Or if we don't want to get
that drastic,
if (reload_completed || SIBLING_CALL_P (insn))
insn_info-frame_read = true;
Will test but I, if I read the code correctly, setting
insn_info-frame_read = true results in
On 12/04/14 08:50, John David Anglin wrote:
On 12/1/2014 11:57 AM, Jeff Law wrote:
Prior to reload (ie, in DSE1) there's a bit of magic in that we do
not set frame_read on call insns. That may in fact be wrong and
possibly the source of the problem.
/* This field is only used for the
On 12/04/14 01:43, Zhenqiang Chen wrote:
Part of PR rtl-optimization/61225
* combine.c (refer_same_reg_p): New function.
(combine_instructions): Handle I1 - I2 - I3; I2 - insn.
(try_combine): Add one more parameter TO_COMBINED_INSN, which
is
On 12/02/14 07:18, Ilya Enkovich wrote:
Hi,
Currently symbol names privatization doesn't work for instrumentation clones
because clones assembler name is transparent alias and therefore alias target
should be privatized instead. This patch does it. Bootstrapped and tested on
On 12/08/14 14:12, Ilya Enkovich wrote:
2014-12-09 0:00 GMT+03:00 Jeff Law l...@redhat.com:
On 12/08/14 04:00, Ilya Enkovich wrote:
2014-11-26 Ilya Enkovich ilya.enkov...@intel.com
PR bootstrap/63995
* tree-chkp.c (chkp_make_static_bounds): Share bounds var
Hi!
This patch attempts to fix
(set (reg:V*mode) (vec_duplicate:V*mode (reg/mem:mode)))
patterns. One issue is that there were separate patterns for
broadcast from gpr and separate patterns for broadcast from memory
(and vector reg), that isn't a good idea for reload, which can't then
freely
Hi!
I've noticed two new spots which use GET_MODE_SIZE (MODEmode)
instead of the more efficient MODE_SIZE.
Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk?
2014-12-08 Jakub Jelinek ja...@redhat.com
* config/i386/sse.md (*movmode_internal, *avx512f_gatherdimode_2):
On 12/02/14 06:54, Ilya Enkovich wrote:
Hi,
Currently cgraph nodes merge may break instrumented_version references. It
depends on order nodes are read and merged and therefore problem is not nicely
reproducible. I couldn't make a small reproducer. This patch fixes problem
and 253.perlbmk
On 11/28/14 10:26, Yury Gribov wrote:
commit 3f560e9dd16a5e914b6f2ba82edffe13dfde944c
Author: Yury Gribovy.gri...@samsung.com
Date: Thu Oct 2 15:50:52 2014 +0400
2014-10-02 Laurynas Biveinislaurynas.bivei...@gmail.com
Yury Gribovy.gri...@samsung.com
Vim config with
On Tue, 2014-11-25 at 20:34 -0500, David Malcolm wrote:
Add command-line option-parsing to the testcases, so that we can
manipulate them without needing a recompile (e.g. varying
optimization levels etc).
This uses getopt_long, which is a GNU extension to libc. Is that
acceptable?
Ping.
On 12/8/2014 3:49 PM, Jeff Law wrote:
I think the terminology and variable names certainly makes this
tougher to follow than it should.
I certainly agree. When I first looked at the code, I thought it
was completely backwards.
Dave
--
John David Anglindave.ang...@bell.net
On Sat, 2014-12-06 at 19:21 +, Sebastian Pop wrote:
I think it does not make sense to duplicate paths at -Os: I disabled the FSM
jump-threading when optimizing for size like this.
diff --git a/gcc/tree-ssa-threadedge.c b/gcc/tree-ssa-threadedge.c
index 29b20c8..ce70311 100644
---
On Dec 8, 2014, at 12:44 PM, David Malcolm dmalc...@redhat.com wrote:
On Tue, 2014-11-25 at 20:34 -0500, David Malcolm wrote:
Add command-line option-parsing to the testcases, so that we can
manipulate them without needing a recompile (e.g. varying
optimization levels etc).
This uses
On 11/15/14 20:59, David Wohlferd wrote:
On 9/15/2014 2:51 PM, Jeff Law wrote:
Let's go with your original inputs + outputs + labels change and punt
the clobbers stuff for now.
jeff
I have also added the test code you requested.
I have a release on file with the FSF, but don't have SVN
On Mon, Dec 8, 2014 at 10:43 PM, Jakub Jelinek ja...@redhat.com wrote:
Hi!
I've noticed two new spots which use GET_MODE_SIZE (MODEmode)
instead of the more efficient MODE_SIZE.
Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk?
2014-12-08 Jakub Jelinek ja...@redhat.com
On 11/20/14 11:09, Jakub Jelinek wrote:
Hi!
This patch fixes ICEs if a non-lvalue vector (say cast of one vector
to another vector type) was subscripted and used as lhs.
The following patch, if *vecp is not lvalue, will copy it to a temporary
variable which can be made addressable for the
On 12/8/2014 3:01 PM, Jeff Law wrote:
The above is wrong for sibcalls. Sibcall arguments are relative to
the incoming argument pointer. Is this always the frame pointer?
I don't think it's always the frame pointer. Don't we use an argument
pointer for the PA64 runtime? If I recall, it was
On 11/24/14 06:51, Mantas Mikaitis wrote:
check_GNU_style.sh error Lines should not exceed 80 characters does
not return the file name and line number where error is present,
only the line of code. Whereas other kind of errors return full
information.
This small patch will fix this and make
Hello!
On Tue, Nov 25, 2014 at 5:05 PM, H.J. Lu hjl.to...@gmail.com wrote:
On Tue, Nov 25, 2014 at 7:01 AM, Richard Biener
richard.guent...@gmail.com wrote:
On Tue, Nov 25, 2014 at 1:57 PM, H.J. Lu hongjiu...@intel.com wrote:
Hi,
The enclosed testcase fails on x86 when compiled with -Os
This patch requires that the gfc_error patch has been applied,
https://gcc.gnu.org/ml/gcc-patches/2014-12/msg00607.html
The patch does some missing '%s' to %qs and '...' to %...% for
gfc_error, does likewise for gfc_notify_std and converts the latter into
calls to gfc_error and gfc_warning.
On Mon, 8 Dec 2014, Richard Biener wrote:
I'm not pushing this further for stage3, but for stage1 I'd like
to eventually address this by splitting up builtin_info_type's
'implicit_p' into a flags array providing implicit_p, declared_p,
used_p and maybe declared_in_system_header_p. Would you
On 12/08/14 15:15, John David Anglin wrote:
On 12/8/2014 3:01 PM, Jeff Law wrote:
The above is wrong for sibcalls. Sibcall arguments are relative
to the incoming argument pointer. Is this always the frame
pointer?
I don't think it's always the frame pointer. Don't we use an
argument pointer
On Mon, 8 Dec 2014, Richard Biener wrote:
The alternative is to decide used in the middle-end at one point,
for example at the end of all_lowering_passes where hopefully
we have constant folded and removed dead code enough. We can also
compute an overall uses libm flag to fix the testcase I
On Mon, Dec 8, 2014 at 11:57 AM, Bernd Edlinger
bernd.edlin...@hotmail.de wrote:
Hello,
this surprisingly simple patch fixes PR ipa/64049. The root cause seems
to be
that in this test case we try to devirtualize a method call on a return
value.
Boot-Strapped and
Hi
After having installed all necessary locales on my system I end up
with 4 failures. Here is a patch to fix them all.
For numpunct test I consider that checking value of numpunct::grouping()
result was not correct. libstdc++ doesn't control this value. I prefer
to just check that
1 - 100 of 113 matches
Mail list logo
