I am currently bootstrapping and regtesting the following for backport
to the 4.8 branch and the testcase for 4.9 to trunk.
Richard.
2015-06-18 Richard Biener rguent...@suse.de
Backport from mainline
2014-12-04 Marek Polacek pola...@redhat.com
PR middle-end/56917
On Thu, 18 Jun 2015, Jakub Jelinek wrote:
On Thu, May 21, 2015 at 05:28:14PM +0200, Jakub Jelinek wrote:
For 4.9/4.8 a similar patch will be needed, but to
fold-const.c/tree-ssa-forwprop.c instead of match.pd.
Here is 4.9/4.8 backport of the patch.
Bootstrapped/regtested on x86_64-linux
On Wed, Jun 17, 2015 at 4:44 PM, Martin Liška mli...@suse.cz wrote:
On 06/17/2015 01:29 PM, Richard Biener wrote:
On Wed, Jun 17, 2015 at 1:22 PM, Jakub Jelinek ja...@redhat.com wrote:
On Wed, Jun 17, 2015 at 11:13:58AM +0200, Martin Jambor wrote:
Do you mean Richard following changes:
Ping.
On Wed, Jun 10, 2015 at 08:02:01PM +0200, Marek Polacek wrote:
Linus's kind words here https://lkml.org/lkml/2015/5/27/941 prodded
me to improving the -Wswitch-bool warning. In particular, this patch
makes the warning not warn when the case values aren't outside bool
range. But it
On 18/06/15 12:48, Richard Biener wrote:
On Thu, 18 Jun 2015, Tom de Vries wrote:
Hi,
I ran into a problem with fortran loops in oacc kernels regions not being
parallelized, after introducting transform_to_exit_first_loop_alt.
For gfortran.dg/goacc/kernels-loop.f95, we get:
...
#pragma omp
On Thu, 18 Jun 2015, Tom de Vries wrote:
On 18/06/15 12:48, Richard Biener wrote:
On Thu, 18 Jun 2015, Tom de Vries wrote:
Hi,
I ran into a problem with fortran loops in oacc kernels regions not being
parallelized, after introducting transform_to_exit_first_loop_alt.
For
* config/aarch64/aarch64-builtins.c: Builtins for rsqrt and rsqrtf.
* config/aarch64/aarch64-protos.h: Declare.
* config/aarch64/aarch64-simd.md: Matching expressions for frsqrte and
frsqrts.
* config/aarch64/aarch64.c: New functions. Emit rsqrt estimation code in
arch64 offers the instructions frsqrte and frsqrts, for rsqrt estimation and
a Newton-Raphson step, respectively.
There are ARMv8 implementations where this is faster than using fdiv and rsqrt.
It runs three steps for double and two steps for float to achieve the needed
precision.
There is one
Hi,
is there a plan to support -mrecip=rsqrt for Aarch64?
Regards,
Venkat.
-Original Message-
From: gcc-patches-ow...@gcc.gnu.org [mailto:gcc-patches-
ow...@gcc.gnu.org] On Behalf Of Benedikt Huber
Sent: Thursday, June 18, 2015 5:34 PM
To: gcc-patches@gcc.gnu.org
Cc:
On Thu, Jun 18, 2015 at 01:18:18PM +0200, Mikael Morin wrote:
I'm proposing here a fix for an OpenMP ICE regression introduced by me
at http://gcc.gnu.org/r221586
That revision changed the order in which procedures are resolved, making
it possible for a procedure to be resolved from within
On Wed, 17 Jun 2015, Matthew Fortune wrote:
Here is my prequel patch. How does this look?
Steve Ellcey
sell...@imgtec.com
2015-06-17 Steve Ellcey sell...@imgtec.com
* config/mips/mips.c (mips_rtx_costs): Remove HONOR_NAN check.
* config/mips/mips.md (*madd4mode):
On Thu, Jun 18, 2015 at 12:14 PM, Bin Cheng bin.ch...@arm.com wrote:
And the patch...
Thanks,
bin
-Original Message-
From: gcc-patches-ow...@gcc.gnu.org [mailto:gcc-patches-
ow...@gcc.gnu.org] On Behalf Of Bin Cheng
Sent: Thursday, June 18, 2015 6:13 PM
To:
Hi!
This patch adds support for linear clause on OpenMP 4.1 worksharing loops.
2015-06-18 Jakub Jelinek ja...@redhat.com
* gimplify.c (gimplify_scan_omp_clauses): For linear clause
on worksharing loop combined with parallel add shared clause
on the parallel.
*
On Wed, Jun 17, 2015 at 5:32 PM, Michael Matz m...@suse.de wrote:
Hi,
this implements support for strided grouped stores in the non-SLP case
(the SLP case existed already). Before we were ignoring all but the last
store in a group. That led to a miscompile of GemsFDTD, the testcase
Hello,
I'm proposing here a fix for an OpenMP ICE regression introduced by me
at http://gcc.gnu.org/r221586
That revision changed the order in which procedures are resolved, making
it possible for a procedure to be resolved from within an OpenMP
construct body. As the OpenMP constructs set some
On Thu, Jun 18, 2015 at 9:25 AM, Tom de Vries tom_devr...@mentor.com wrote:
Hi,
transform_to_exit_first_loop contains the following comment:
...
/* Make sure that we have phi nodes on exit for all loop header phis
(create_parallel_loop requires that). */
...
I ran into a problem
On Wed, 17 Jun 2015, Richard Sandiford wrote:
and the one mips-ps-3d.md:
(define_expand mips_abs_ps
[(set (match_operand:V2SF 0 register_operand)
(unspec:V2SF [(match_operand:V2SF 1 register_operand)]
UNSPEC_ABS_PS))]
TARGET_HARD_FLOAT
On Thu, May 21, 2015 at 05:28:14PM +0200, Jakub Jelinek wrote:
For 4.9/4.8 a similar patch will be needed, but to
fold-const.c/tree-ssa-forwprop.c instead of match.pd.
Here is 4.9/4.8 backport of the patch.
Bootstrapped/regtested on x86_64-linux and i686-linux, ok for 4.9/4.8?
2015-06-18
On Sat, Jun 13, 2015 at 12:05:42PM +0200, Marc Glisse wrote:
They look fine to me.
Thanks!
+/* (x y) + (x ^ y) - x | y */
This also works if you replace + with | or ^.
Done.
+/* (x | y) - (x ^ y) - x y */
For symmetry, you probably also want:
(x | y) - (x y) - x ^ y
Done.
On Tue, Jun 9, 2015 at 1:31 PM, Patrick Palka patr...@parcs.ath.cx wrote:
This patch refactors the entry point of -Wmisleading-indentation from:
void
warn_for_misleading_indentation (location_t guard_loc,
location_t body_loc,
On Thu, 2015-06-18 at 13:04 +0100, Maciej W. Rozycki wrote:
This change looks good to me, I have no objections. Thanks.
Maciej
OK, I checked in the prequel patch and here is a new copy of the
original patch based off of that (and with no HONOR_NAN checks in the
fma/madd instructions).
We need to handle TRY_BLOCK and TRY_FINALLY_EXPR in the constexpr code.
Tested x86_64-pc-linux-gnu, applying to trunk and 5.
commit 8f851dc2420532bda205410a99bcd672090cfe6a
Author: Jason Merrill ja...@redhat.com
Date: Wed Jun 17 18:31:02 2015 -0400
PR c++/66001
* constexpr.c
Richard,
Here is updated patch which does not include your proposal related to
the target hook deletion.
You wrote:
I still don't understand why you need the new target hook. If we have a
masked
load/store then the mask is computed by an assignment with a VEC_COND_EXPR
(in your example) and
Should be fixed, thanks.
Jason
Hi!
On Wed, 17 Jun 2015 15:15:15 +0100, Julian Brown jul...@codesourcery.com
wrote:
This is a set of tests for OpenACC private variable/state propagation
support in GCC. The associated functionality is a work-in-progress: as
such, many of these tests do not pass yet (causing incorrect
Hi!
This patch adds priority clause parsing to #pragma omp taskloop,
and passes it through to GOMP_task (new argument, in hopefully backward
compatible way) and GOMP_taskloop (we don't care about ABI here for now).
So, for priorities hopefully the only task left is write support in the
library
The macro WORD_REGISTER_OPERATIONS, if defined, means that all reg-reg
operations on data smaller than words are performed on the full word.
For TARGET_POWERPC64 words are 64 bits; but many operations on SImode
do not behave as if on DImode. So rs6000 should not define the macro.
Bootstrappped
On Thu, Jun 18, 2015 at 12:57 PM, Patrick Palka patr...@parcs.ath.cx wrote:
On Thu, Jun 18, 2015 at 12:31 PM, David Malcolm dmalc...@redhat.com wrote:
On Tue, 2015-06-09 at 13:31 -0400, Patrick Palka wrote:
This patch improves the heuristics of the warning in a number of ways.
The improvements
Hi!
The attached patch adds additional tests for the
OpenACC default clause.
Committed to gomp-4_0-branch
Jim
diff --git a/gcc/testsuite/c-c++-common/goacc/default-1.c b/gcc/testsuite/c-c++-common/goacc/default-1.c
new file mode 100644
index 000..41757f1
--- /dev/null
+++
Are there any concerns with or suggestions for changes to
the following patch?
https://gcc.gnu.org/ml/gcc-patches/2015-06/msg00886.html
Thanks
Martin
On Tue, Jun 16, 2015 at 03:35:15PM +0200, Richard Biener wrote:
We already have
/* (x y) ^ (x | y) - x ^ y */
(simplify
(bit_xor:c (bit_and @0 @1) (bit_ior @0 @1))
(bit_xor @0 @1))
but of course with minus it doesn't commutate so it's hard to
merge.
Yeah :(.
+/* (x y) + (x |
On Thu, Jun 18, 2015 at 12:39 PM, David Malcolm dmalc...@redhat.com wrote:
On Thu, 2015-06-18 at 11:41 -0400, Patrick Palka wrote:
On Tue, Jun 9, 2015 at 1:31 PM, Patrick Palka patr...@parcs.ath.cx wrote:
This patch refactors the entry point of -Wmisleading-indentation from:
void
Hi Thomas!
On 06/17/2015 04:59 AM, Thomas Schwinge wrote:
Hi Jim!
I had mentioned that the Fortran front end changes cause regressions in a
few libgomp execution tests, if configured for Intel MIC (emulation)
offloading. I have now located *where* this is coming from, but would
you please
On Tue, 2015-06-09 at 13:31 -0400, Patrick Palka wrote:
This patch improves the heuristics of the warning in a number of ways.
The improvements are hopefully adequately documented in the code
comments.
The additions to the test case also highlight the improvements.
I tested an earlier
On Thu, 2015-06-18 at 11:41 -0400, Patrick Palka wrote:
On Tue, Jun 9, 2015 at 1:31 PM, Patrick Palka patr...@parcs.ath.cx wrote:
This patch refactors the entry point of -Wmisleading-indentation from:
void
warn_for_misleading_indentation (location_t guard_loc,
On 18/06/15 13:38, Richard Biener wrote:
On Thu, Jun 18, 2015 at 9:25 AM, Tom de Vries tom_devr...@mentor.com wrote:
Hi,
transform_to_exit_first_loop contains the following comment:
...
/* Make sure that we have phi nodes on exit for all loop header phis
(create_parallel_loop requires
On Thu, Jun 18, 2015 at 12:31 PM, David Malcolm dmalc...@redhat.com wrote:
On Tue, 2015-06-09 at 13:31 -0400, Patrick Palka wrote:
This patch improves the heuristics of the warning in a number of ways.
The improvements are hopefully adequately documented in the code
comments.
The additions
On 06/18/2015 02:37 AM, Uros Bizjak wrote:
Attached patch eliminates:
gcc/doc//invoke.texi:307: warning: unlikely character , in @var.
gcc/doc//invoke.texi:307: warning: unlikely character , in @var.
gcc/doc//invoke.texi:5894: warning: unlikely character , in @var.
gcc/doc//invoke.texi:5894:
ping?
Thanks,
Kugan
On 02/06/15 09:14, Kugan wrote:
On 08/05/15 22:48, Richard Biener wrote:
You compute which promotions are unsafe, like sources/sinks of memory
(I think you miss call arguments/return values and also asm operands here).
But instead of simply marking those SSA names as
The attached patch teaches regrename to validate insns affected by each
register renaming before making the change. I can see at least two
other ways to handle this -- earlier, by rejecting renamings that result
in invalid instructions when it's searching for the best renaming; or
later, by
On Thu, Jun 18, 2015 at 08:26:19PM +0200, Mikael Morin wrote:
Hmm, for gfc_derived_types at least, I would say no.
gfc_derived_types seems to be a per root namespace global, in other
words shared between contained procedures, so it should not be
saved/restored upon resolution of sibling or
Le 18/06/2015 13:35, Jakub Jelinek a écrit :
On Thu, Jun 18, 2015 at 01:18:18PM +0200, Mikael Morin wrote:
I'm proposing here a fix for an OpenMP ICE regression introduced by me
at http://gcc.gnu.org/r221586
That revision changed the order in which procedures are resolved, making
it possible
illegal - invalid or not permitted
The ISO standard for Ada contains this (1.1.2 Structure):
Legality Rules
27 Rules that are enforced at compile time. A construct is legal if it obeys
all of the Legality Rules.
and illegal is documented in the Index.
But the standard for C doesn't, so I'll
On Jun 18, 2015, at 3:10 AM, Richard Biener rguent...@suse.de wrote:
Ok for trunk?
Ok.
Please mark this pattern with (set_attr type multiple).
Done.
While I suspect that stack probing is done before any insns with invalid
constants in the function, it would be better to model the length of
this insn so that the minipool logic is not confused later in terms of
placement of
A very cursory glance at the patch set but should we restrict this
attribute to apply only to scalar types rather than allow them on vector
extensions. I haven't had enough coffee this morning but a first
thought is that if it does apply on vector type extensions, clarifying
the lane
Hi,
I backported the fix for PR62308 to 4.9-branch. The original patch
applies cleanly.
Bootstrap + make check OK on x86_64 and aarch64.
I'm also adding the testcase I've just sent separately for trunk.
OK?
Christophe.
2015-06-18 Christophe Lyon christophe.l...@linaro.org
Backport
Hi,
While backporting the fix for PR62308 from trunk to 4.9-branch, it
appeared that a testcase would be useful.
Here it is. I'll send another email to request the backport of the fix
+ testcase to the 4.9-branch.
Christophe.
2015-06-18 Christophe Lyon christophe.l...@linaro.org
*
It appears that all the C tests check the specific syntactic form
'struct __attr_sso__ foo {...};'
What is the impact of changing the location of __attr_sso__? Eg:
struct foo {...} __attr_sso__;
None. What impact could it have?
(Note that alignment attributes can have significant
On 18 June 2015 at 21:17, Christophe Lyon christophe.l...@linaro.org wrote:
Hi,
While backporting the fix for PR62308 from trunk to 4.9-branch, it
appeared that a testcase would be useful.
Here it is. I'll send another email to request the backport of the fix
+ testcase to the 4.9-branch.
On Thu, Jun 18, 2015 at 1:08 PM, Segher Boessenkool
seg...@kernel.crashing.org wrote:
The macro WORD_REGISTER_OPERATIONS, if defined, means that all reg-reg
operations on data smaller than words are performed on the full word.
For TARGET_POWERPC64 words are 64 bits; but many operations on
I was making tags in libiberty and noticed it just didn’t work. :-(
This is one easy way to fix it.
Ok?
diff --git a/libiberty/Makefile.in b/libiberty/Makefile.in
index f06cc69..876fab2 100644
--- a/libiberty/Makefile.in
+++ b/libiberty/Makefile.in
@@ -409,8 +409,9 @@ stamp-noasandir:
etags
Mike Stump mikest...@comcast.net writes:
+ cd $(srcdir) etags `for i in $(CFILES); do echo $(srcdir)/$$i ;
done`
This doesn't work with a relative $(srcdir).
Andreas.
--
Andreas Schwab, sch...@linux-m68k.org
GPG Key fingerprint = 58CA 54C7 6D53 942B 1756 01D3 44D5 214B 8276 4ED5
On Jun 18, 2015, at 1:35 PM, Andreas Schwab sch...@linux-m68k.org wrote:
Mike Stump mikest...@comcast.net writes:
+ cd $(srcdir) etags `for i in $(CFILES); do echo $(srcdir)/$$i ;
done`
This doesn't work with a relative $(srcdir).
Sure it does. I had tested with:
srcdir =
On Tue, 2015-06-16 at 00:08 -0400, DJ Delorie wrote:
This is the first in a series of patches to make a build with an in-tree
GNU libiconv work as designed.
This patch fixes dependencies for parallel make, and avoids failures
with make targets not supported by GNU libiconv.
This is
Attached patch implements zero-extended cmove patterns. Also, the
patch merges a couple of peephole patterns into one.
2015-06-18 Uros Bizjak ubiz...@gmail.com
* config/i386/i386.md (*movsicc_noc_zext): New insn.
(zero-extended cmove with mem peephole2): New pattern.
(cmove with
On 13/06/15 16:24, Tom de Vries wrote:
Hi,
this patch allows try_transform_to_exit_first_loop_alt to succeed when
handling cases where the expression representing the number of
iterations contains a cast.
Currently, transform_to_exit_first_loop_alt testcase
On 06/16/2015 07:40 PM, Ulrich Weigand wrote:
However, I'm not clear why this should be a 4.8 only patch ... the
same problem seems to be still there on mainline, right?
Yes. But 4.8 appears to be the last where reload was really stable on these
targets. When disabling
LRA on mainline x86_64
On Thu, Jun 18, 2015 at 11:51:10AM +0200, Paolo Carlini wrote:
Nit: I think you inadvertently committed the new gcc/testsuite/ChangeLog
entry as gcc/cp/ChangeLog entry too.
Indeed, only on the trunk though. Fixed, thanks for catching it.
Jakub
Hi,
In early_inliner we do recompute inline summaries for edges after
optimize_inline_calls, but check this summary exists in case new edges appear.
But then it calls inline_update_overall_summary which also going through edges
inline summaries but with no check this time causing segfault.
Hi,
Currently instrumentation clone may be requested for a function which cannot be
cloned and a thunk is created in such case. Thunk creation is bogus and cause
original node and a clone to share arguments list (including bounds args). I
don't think such thunks are actually useful and
Currently when doing
make check-gcc RUNTESTFLAGS=TORTURE_OPTIONS=\\\{ -O3 } { -O2 }\\\
dg-torture.exp
you get -O3 and -O2 but also the two LTO torture option combinations.
That's undesired (those are the most expensive anyway). The following
patch avoids this by setting LTO_TORTURE_OPTIONS
Hi,
This patch fixes ICE in PR66449 by using POINTER_PLUS_EXPR for pointers.
Bootstrap and test on x86_64. Is it OK trunk?
Thanks,
bin
2015-06-18 Bin Cheng bin.ch...@arm.com
PR tree-optimization/66449
* tree-ssa-loop-niter.c (loop_exits_before_overflow): Use
And the patch...
Thanks,
bin
-Original Message-
From: gcc-patches-ow...@gcc.gnu.org [mailto:gcc-patches-
ow...@gcc.gnu.org] On Behalf Of Bin Cheng
Sent: Thursday, June 18, 2015 6:13 PM
To: gcc-patches@gcc.gnu.org
Subject: [PATCH PR66449]Fix ICE by using POINTER_PLUS_EXPR for
Hi,
Currently return statement with no bounds in instrumented function causes ICE.
This patch uses zero bounds in such cases instead. Bootstrapped and regtested
for x86_64-unknown-linux-gnu. Applied to trunk. Is it OK for gcc-5-branch?
Thanks,
Ilya
--
gcc/
2015-06-18 Ilya Enkovich
Hi,
this patch changes an order of bounds params expand for vararg function. This
is neede for vararg function which have bounds passed for named args but these
bounds are nor passed i nregisters. In this case we try to load them from
bounds table but all registers are used for vararg bounds
Hi,
transform_to_exit_first_loop contains the following comment:
...
/* Make sure that we have phi nodes on exit for all loop header phis
(create_parallel_loop requires that). */
...
I ran into a problem where after transform_to_exit_first_loop_alt this
property does not hold for
Attached patch eliminates:
gcc/doc//invoke.texi:307: warning: unlikely character , in @var.
gcc/doc//invoke.texi:307: warning: unlikely character , in @var.
gcc/doc//invoke.texi:5894: warning: unlikely character , in @var.
gcc/doc//invoke.texi:5894: warning: unlikely character , in @var.
Hi,
This came up in a User question last night and reminded me that I had a patch
for it in my Q.
Usually g++ driver support for -static-libstdc++ is provided by -Bstatic
-lstdc++ -Bdynamic and is currently disabled for targets without that linker
support. However, actually, there is
On Wed, Jun 17, 2015 at 11:04:24AM +0100, Uros Bizjak wrote:
On Tue, Jun 16, 2015 at 5:13 PM, Uros Bizjak ubiz...@gmail.com wrote:
Hello!
Following patch fixes:
cp_lto_pr65276_1.o: In function `std2::runtime_error::~runtime_error()':^M
On Thu, Jun 18, 2015 at 10:59 AM, James Greenhalgh
james.greenha...@arm.com wrote:
Hello!
Following patch fixes:
cp_lto_pr65276_1.o: In function `std2::runtime_error::~runtime_error()':^M
pr65276_1.C:(.text._ZN4std213runtime_errorD2Ev[_ZN4std213runtime_errorD5Ev]+0x8):
undefined
The testcase shows a series of interesting issues in the vectorizer.
First of all there is a disconnect between the number of expected
vector statements generated for the load and the permutation result,
second, the vectorizer alignment code appearantly didn't expect a
non-vector-size-multiple
On Thu, Jun 18, 2015 at 11:05 AM, Uros Bizjak ubiz...@gmail.com wrote:
This patch causes failures in arm-none-linux-gnueabihf testing:
PASS-FAIL: g++.dg/lto/pr65276 cp_lto_pr65276_0.o-cp_lto_pr65276_1.o link,
-flto -O0 -std=c++11
Hi!
I've committed (yesterday already) the following patch to trunk and 5 branch
to fix an issue where tsubst_expr on OMP_CLAUSE_DECL, when a decl had
non-reference dependent type, got a reference type after instantiation
and tsubst_expr added convert_from_reference around it.
Hi!
I've merged trunk into gomp-4_1-branch up to r224573.
Unfortunately, the PR66571 patch doesn't really work on the branch where we
support privatization of non-static data members. Therefore, I've committed
together with the merge following change plus another testcase that
covers the new
Hi Jakub,
On 06/18/2015 11:43 AM, Jakub Jelinek wrote:
Hi!
I've committed (yesterday already) the following patch to trunk and 5 branch
to fix an issue where tsubst_expr on OMP_CLAUSE_DECL, when a decl had
non-reference dependent type, got a reference type after instantiation
and tsubst_expr
Hi,
I ran into a problem with fortran loops in oacc kernels regions not
being parallelized, after introducting transform_to_exit_first_loop_alt.
For gfortran.dg/goacc/kernels-loop.f95, we get:
...
#pragma omp target oacc_parallel num_gangs(1)
...
instead of the desired num_gangs (32).
On Thu, 18 Jun 2015, Tom de Vries wrote:
Hi,
I ran into a problem with fortran loops in oacc kernels regions not being
parallelized, after introducting transform_to_exit_first_loop_alt.
For gfortran.dg/goacc/kernels-loop.f95, we get:
...
#pragma omp target oacc_parallel num_gangs(1)
spawn /daten/aranym/gcc/gcc-20150618/Build/gcc/testsuite/g++/../../xg++
-B/daten/aranym/gcc/gcc-20150618/Build/gcc/testsuite/g++/../../
/daten/aranym/gcc/gcc-20150618/gcc/testsuite/g++.dg/cpp0x/initlist96.C
-fno-diagnostics-show-caret -fdiagnostics-color=never -nostdinc++
-I/daten/aranym/gcc
Hi Uros
After bootstrap testing and regression testing on bdver4 machine. I committed
the back port on GCC 5 branch at
https://gcc.gnu.org/viewcvs/gcc?view=revisionrevision=224603
Checked with Richard as well before proceeding.
Regards,
Venkat.
-Original Message-
From: Uros
80 matches
Mail list logo