On 12 August 2015 at 23:54, Dijk, J. van wrote:
Thanks for the clarification. I hope the attached patch against current
wwwdocs will be of use (only the name of the feature test macro is missing
in the new file cxx1z.html; did not manage to find that.)
Thanks, I've committed the attached
Hi Steve,
Overall, I don't think these optimizations are ready to include. In principle
the idea looks good but it is done at the wrong point in the compiler in my
opinion.
The biggest concern I have is that the analysis should be possible at (or
prior to) the point where the prologue/epilogue
On Thu, 13 Aug 2015, Andreas Schwab wrote:
On m68k:
FAIL: gcc.dg/tree-ssa/vrp33.c scan-tree-dump vrp1 Folding predicate.*== 1 to
0
$ gcc/xgcc -B gcc/ ../gcc/testsuite/gcc.dg/tree-ssa/vrp33.c -O2
-fdump-tree-vrp1 -S
$ grep -c Folding *.vrp1
0
I suppose for
This handles calls and conds.
Bootstrapped and tested on x86_64-unknown-linux-gnu, applied.
Richard.
2015-08-13 Richard Biener rguent...@suse.de
* gimple.h (gcall::code_): New constant static member.
(gcond::code_): Likewise.
* gimple.c (gcall::code_): Define.
Given there is now PR67167 I am going forward with the earlier posted
patch to switch SCCVN to PHI elimination in favor of another PHI
(to remove IVs) rather than in favor of its only executable edge value.
I still see no way to capture both cases without detecting the choice
and re-numbering
Ping.
-Original Message-
From: gcc-patches-ow...@gcc.gnu.org [mailto:gcc-patches-ow...@gcc.gnu.org] On
Behalf Of Robert Suchanek
Sent: 05 August 2015 09:31
To: catherine_mo...@mentor.com; Matthew Fortune; gcc-patches@gcc.gnu.org
Subject: [PATCH, MIPS] Enable load/store bonding for
Hi,
gcc/
* config/mips/mips.c (mips_rtx_cost_data): Remove costs for W32 and
W64
pseudo-processors.
* config/mips/mips.md (processor): Remove w32 and w64.
OK, thanks.
Matthew
Committed as r226851.
Regards,
Robert
On m68k:
FAIL: gcc.dg/tree-ssa/vrp33.c scan-tree-dump vrp1 Folding predicate.*== 1 to 0
$ gcc/xgcc -B gcc/ ../gcc/testsuite/gcc.dg/tree-ssa/vrp33.c -O2
-fdump-tree-vrp1 -S
$ grep -c Folding *.vrp1
0
Andreas.
--
Andreas Schwab, SUSE Labs, sch...@suse.de
GPG Key fingerprint = 0196 BAD8 1CE9
Hi Nathan!
On Wed, 12 Aug 2015 09:30:10 -0400, Nathan Sidwell nat...@acm.org wrote:
I've committed this patch to gomp4. [...]
Also put in the change I mentioned earlier this morning about the default
validate dims hook setting the dimensions to 1 on accelerators to. I'll
revert
Ping.
Thanks,
bin
On Tue, Jul 28, 2015 at 5:36 PM, Bin Cheng bin.ch...@arm.com wrote:
Hi,
Loop niter computes inaccurate bound information for different loops. This
patch is to improve it by using loop initial condition in
determine_value_range. Generally, loop niter is computed by
Ping.
Thanks,
bin
On Tue, Jul 28, 2015 at 5:38 PM, Bin Cheng bin.ch...@arm.com wrote:
Hi,
For now, SCEV may compute iv base in the form of (signed T)((unsigned
T)base + step)). This complicates other optimizations/analysis depending
on SCEV because it's hard to dive into type conversions.
On ia64 and arm64:
FAIL: g++.dg/tree-ssa/pr61034.C -std=gnu++11 scan-tree-dump-times fre2 free
10
$ gcc/xg++ -Bgcc/ ../gcc/testsuite/g++.dg/tree-ssa/pr61034.C -nostdinc++
-Iia64-suse-linux/libstdc++-v3/include/ia64-suse-linux
-Iia64-suse-linux/libstdc++-v3/include
Hi,
Sorry to bother people again. Is this OK to go now?
Thanks!
David.
On Mon, 29 Jun 2015, David Sherwood wrote:
Hi,
I have added new STRICT_MAX_EXPR and STRICT_MIN_EXPR expressions to
support the
IEEE versions of fmin and fmax. This is done by recognising the math
Le 13/08/2015 12:56, Mikael Morin a écrit :
Still, implementation-defined behavior is a progress over undefined
behaviour.
Even if it's not set in stone, the good thing about
implementation-defined behaviour is that it's known to be non-random.
So it can be checked, either with autoconf, or with
On Thu, Aug 13, 2015 at 1:08 PM, Markus Trippelsdorf
mar...@trippelsdorf.de wrote:
On 2015.08.13 at 12:56 +0200, Mikael Morin wrote:
Le 12/08/2015 22:07, Richard Sandiford a écrit :
Jeff Law l...@redhat.com writes:
On 08/12/2015 12:32 PM, Richard Biener wrote:
On August 12, 2015 8:07:13 PM
Hi Richard,
Did you have a chance to look at updated patch?
Thanks.
Yuri.
2015-08-06 14:07 GMT+03:00 Yuri Rumyantsev ysrum...@gmail.com:
HI All,
Here is updated patch which implements Richard proposal to use vector
comparison with boolean result instead of target hook. Support for it
was
Hi,
gcc/
* config/mips/mips.h (ENABLE_LD_ST_PAIRS): Enable load/store pairs for
I6400.
Sorry, I missed this one. OK to commit.
Thanks,
Matthew
Committed as r226860.
Regards,
Robert
On 2015.08.13 at 12:56 +0200, Mikael Morin wrote:
Le 12/08/2015 22:07, Richard Sandiford a écrit :
Jeff Law l...@redhat.com writes:
On 08/12/2015 12:32 PM, Richard Biener wrote:
On August 12, 2015 8:07:13 PM GMT+02:00, Jeff Law l...@redhat.com wrote:
On 08/12/2015 11:12 AM, Richard Biener
Hi All,
Here is patch for adding march/mtune options for Skylake.
Bootstrap and regression testing did not show any new failures.
Is it OK for trunk?
ChangeLog:
2015-08-13 Yuri Rumyantsev ysrum...@gmail.com
* config/i386/driver-i386.c (host_detect_local_cpu): Add support
for skylake.
*
On Wed, Aug 12, 2015 at 6:47 PM, Jeff Law l...@redhat.com wrote:
On 08/12/2015 10:24 AM, Ian Lance Taylor wrote:
On Wed, Aug 12, 2015 at 9:21 AM, Tom Tromey t...@tromey.com wrote:
Jeff In the past this has stalled on issues like how will
asynch-exceptions
Jeff be tested and the like.
It
On Wed, Aug 12, 2015 at 4:30 PM, Trevor Saunders tbsau...@tbsaunde.org wrote:
On Wed, Aug 12, 2015 at 03:31:35PM -0400, David Edelsohn wrote:
This patch has caused a bootstrap failure on AIX.
reverted, though my bet is this is some sort of generic devirt bug on
aix.
There have been a lot of
gcc/
* config/mips/mips.h (ENABLE_LD_ST_PAIRS): Enable load/store pairs for
I6400.
Sorry, I missed this one. OK to commit.
Thanks,
Matthew
On Thu, Aug 13, 2015 at 12:11 PM, David Sherwood david.sherw...@arm.com wrote:
Hi,
Sorry to bother people again. Is this OK to go now?
Hmm, why don't you go the vectorized function call path for this,
implementing the builtin_vectorized_function target hook?
Richard.
Thanks!
David.
Here's the patch for the rest of the C++17 support.
On 13 August 2015 at 09:58, Jonathan Wakely jwakely@gmail.com wrote:
On 12 August 2015 at 23:54, Dijk, J. van wrote:
Thanks for the clarification. I hope the attached patch against current
wwwdocs will be of use (only the name of the
Just remove the assert which I now know how it can trigger without
causing harm.
Bootstrapped and tested on x86_64-unknown-linux-gnu, applied.
Richard.
2015-08-13 Richard Biener rguent...@suse.de
PR tree-optimization/67191
* tree-ssa-sccvn.c
On Wed, Aug 12, 2015 at 10:52 PM, Mike Stump mikest...@comcast.net wrote:
On Aug 12, 2015, at 1:07 PM, Richard Sandiford rdsandif...@googlemail.com
wrote:
I don't think the right shifts are an issue at all.
Well, they're implementation-defined, at least in C.
The C11 wording for E1 E2 is
Richard Biener rguent...@suse.de writes:
On Thu, 13 Aug 2015, Andreas Schwab wrote:
On m68k:
FAIL: gcc.dg/tree-ssa/vrp33.c scan-tree-dump vrp1 Folding predicate.*== 1
to 0
$ gcc/xgcc -B gcc/ ../gcc/testsuite/gcc.dg/tree-ssa/vrp33.c -O2
-fdump-tree-vrp1 -S
$ grep -c Folding *.vrp1
0
On Thu, 13 Aug 2015, Andreas Schwab wrote:
On ia64 and arm64:
FAIL: g++.dg/tree-ssa/pr61034.C -std=gnu++11 scan-tree-dump-times fre2
free 10
$ gcc/xg++ -Bgcc/ ../gcc/testsuite/g++.dg/tree-ssa/pr61034.C -nostdinc++
-Iia64-suse-linux/libstdc++-v3/include/ia64-suse-linux
On Wed, Aug 12, 2015 at 10:41 PM, Tobias Grosser tob...@grosser.es wrote:
On 08/12/2015 10:33 PM, Aditya Kumar wrote:
Passes bootstrap, no regressions.
With this patch gcc bootstraps with graphite.
make BOOT_CFLAGS=-g -O2 -fgraphite-identity -floop-interchange
-floop-block
LGTM, but
Hi,
This patch follows on from
[PATCH][1/N] Change GET_MODE_INNER to always return a non-void mode
It is another tidy up, replacing the pattern
GET_MODE_BITSIZE (GET_MODE_INNER (m)) with GET_MODE_UNIT_BITSIZE (m). Also
replaces any calls to GET_MODE_PRECISION (GET_MODE_INNER (m)) with
Le 12/08/2015 22:07, Richard Sandiford a écrit :
Jeff Law l...@redhat.com writes:
On 08/12/2015 12:32 PM, Richard Biener wrote:
On August 12, 2015 8:07:13 PM GMT+02:00, Jeff Law l...@redhat.com wrote:
On 08/12/2015 11:12 AM, Richard Biener wrote:
Prec is almost never a constant and is
Hi,
It was discovered that with the attached test case compiled with -O2
-funroll-loops,
the regrename pass renamed one of the registers ($2) to $8 that was not
saved by the prologue.
The attached patch fixes it by defining macro HARD_REGNO_RENAME_OK that returns
zero iff the current function
I'd like to give Catherine chance to review this, I notice a couple
of formatting nits in the test case:
Robert Suchanek robert.sucha...@imgtec.com writes:
a/gcc/testsuite/gcc.target/mips/interrupt_handler-bug-1.c
b/gcc/testsuite/gcc.target/mips/interrupt_handler-bug-1.c
new file mode 100644
On Wed, Aug 12, 2015 at 11:10 PM, Segher Boessenkool
seg...@kernel.crashing.org wrote:
On Wed, Aug 12, 2015 at 03:31:48AM -0500, Segher Boessenkool wrote:
* This is the only platform for which I bothered collecting any sort
of performance data:
As best I can tell, there is a 9%
On Thu, Aug 13, 2015 at 1:32 PM, Yuri Rumyantsev ysrum...@gmail.com wrote:
Hi Richard,
Did you have a chance to look at updated patch?
Having a quick look now. Btw, you didn't try the simpler alternative of
tree type = type_for_mode (int_mode_for_mode (TYPE_MODE (vectype)));
build2
On 7/31/2015 6:06 PM, John Marino wrote:
So far, DragonFly only handles the C/POSIX locale through the generic
locale support files, similar to FreeBSD and Darwin. The following
patchset enables nearly complete support for named locales. The except
is for std:messages which requires gettext
Hi,
in this ICE on invalid, during error recovery resolve_typename_type gets
a TYPE which has a DECLTYPE_TYPE as TYPE_CONTEXT and
CLASSTYPE_PRIMARY_TEMPLATE_TYPE ICEs on the latter. Checking
CLASS_TYPE_P and thus returning the TYPE as is avoids the ICE and passes
testing on x86_64-linux.
Simply register an != 0 expression for them. This should solve Honzas
request to optimize away != 0 checks for this pointers during early opts.
Bootstrapped and tested on x86_64-unknown-linux-gnu, applied.
Richard.
2015-08-13 Richard Biener rguent...@suse.de
* tree.c
On 03/08/15 14:01, James Greenhalgh wrote:
On Fri, Jul 24, 2015 at 11:55:33AM +0100, Kyrill Tkachov wrote:
Hi all,
This patch implements an aarch64-specific expansion of the signed modulo by a
power of 2.
The proposed sequence makes use of the conditional negate instruction CSNEG.
For a
Due to PR67205, the deeply nested instantiations require trampolines,
which in turn requires an executable stack for the GNAT tools on
architectures such as x86_64.
Bootstrapped on x86_64-redhat-linux-gnu, and make check-ada
reports no unexpected failures.
Okay for trunk?
--
Florian Weimer /
Subject says it all.
Tested ppc64le-linux, committing to trunk.
commit d7f997a59a1383b7f08618fbf1d7855edb5fa012
Author: Jonathan Wakely jwak...@redhat.com
Date: Thu Aug 13 10:27:59 2015 +0100
* include/bits/alloc_traits.h: Add feature-test macro.
* include/bits/allocator.h:
Greetings,
Now that most systems have uchar.h for conversion between char16_t* and
char32_t* to multibyte (analogue of wchar.h and cwchar) i think we
should check off this last little crumb of C++11.
I submitted this a while back but Jonathan suggested improvements and it
took me a hot
On 2015.08.13 at 15:40 +0200, Marek Polacek wrote:
On Thu, Aug 13, 2015 at 01:11:53PM +0200, Richard Biener wrote:
Rather ubsan should not complain about implementation defined behavior (or
should separate those cases out with a different switch compared to
undefined
behavior).
I
This commits the patch from
https://gcc.gnu.org/ml/gcc-patches/2015-08/msg00663.html and adds
tests for all the mutex types.
Eelis, thanks for the patch. Please CC the libstdc++ list for all
libstdc++ patches, I don't subscribe to gcc-patches and only saw your
mail because Ville brought it to my
On Thu, Aug 13, 2015 at 01:11:53PM +0200, Richard Biener wrote:
Rather ubsan should not complain about implementation defined behavior (or
should separate those cases out with a different switch compared to undefined
behavior).
I think ubsan doesn't complain about implementation-defined
On Thu, Aug 13, 2015 at 03:52:00PM +0200, Markus Trippelsdorf wrote:
On 2015.08.13 at 15:40 +0200, Marek Polacek wrote:
On Thu, Aug 13, 2015 at 01:11:53PM +0200, Richard Biener wrote:
Rather ubsan should not complain about implementation defined behavior (or
should separate those cases
And a small correction.
On 13 August 2015 at 10:24, Jonathan Wakely jwakely@gmail.com wrote:
Here's the patch for the rest of the C++17 support.
On 13 August 2015 at 09:58, Jonathan Wakely jwakely@gmail.com wrote:
On 12 August 2015 at 23:54, Dijk, J. van wrote:
Thanks for the
On Wed, Aug 12, 2015 at 4:30 PM, Trevor Saunders tbsau...@tbsaunde.org wrote:
On Wed, Aug 12, 2015 at 03:31:35PM -0400, David Edelsohn wrote:
This patch has caused a bootstrap failure on AIX.
reverted, though my bet is this is some sort of generic devirt bug on
aix.
bootstrap with GCC 5 does
On 12/08/15 22:26, Joseph Myers wrote:
On Tue, 11 Aug 2015, Matthew Wahab wrote:
PR target/67143
* gcc.target/aarch64/pr67143.c: New
What's architecture-specific about this test? That is, why doesn't it
just go in gcc.c-torture/compile (no dg- directives needed,
On 13/08/15 10:37 -0400, Ed Smith-Rowland wrote:
Greetings,
Now that most systems have uchar.h for conversion between char16_t*
and char32_t* to multibyte (analogue of wchar.h and cwchar) i think we
should check off this last little crumb of C++11.
Ooh, I forgot about this, thanks!
AIX stdio.h header includes code specific for C++ that looks like:
extern C {
#ifdef __cplusplus
#ifdef ferror
#undef ferror
inline int ferror(FILE * _p)
{
return ((_p)-_flag _IOERR);
}
#endif /*ferror*/
which generates code that makes the AIX linker and loader rather upset
because the
On Aug 13, 2015, at 3:05 AM, Richard Biener richard.guent...@gmail.com wrote:
Ok, then guard the with __GCC__ and do the expensive bit stuff
otherwise. Just to cater for other host compilers doing sth unsensibly
implementation defined.
Ick. The guard should be specific to the
Date: Thu, 13 Aug 2015 12:02:43 +0200
Subject: Re: [PATCH] [graphite] Constrain only on INTEGER_TYPE
From: richard.guent...@gmail.com
To: tob...@grosser.es
CC: hiradi...@msn.com; gcc-patches@gcc.gnu.org; s@samsung.com;
seb...@gmail.com
On Wed,
On 08/12/2015 11:36 PM, Trevor Saunders wrote:
In many places gcc puts classes in the anon namespace so the compiler can tell
they do not get inheritted from to enable better devirtualization.
C++ does not allow class members to have static linkage, so there is no way to
tell the compiler that a
-Original Message-
From: Matthew Fortune [mailto:matthew.fort...@imgtec.com]
Sent: Thursday, August 13, 2015 7:20 AM
To: Robert Suchanek; Moore, Catherine
Cc: gcc-patches@gcc.gnu.org
Subject: RE: [PATCH][MIPS] Fix register renaming in the interrupt handlers
I'd like to give
There was a recent change which causes ira-int.h to depend on recog.h..
but doesnt include it. Every file which needs it included it before I
think, it just required some shuffling of order.
By including recog.h in ira-int.h ordering issues are fixed, and we can
simply remove recog.h form
On Aug 13, 2015, at 12:54 PM, Richard Sandiford rdsandif...@googlemail.com
wrote:
It was discovered that with the attached test case compiled with -O2
-funroll-loops, the regrename pass renamed one of the registers ($2)
to $8 that was not saved by the prologue.
The attached patch fixes it
On Thu, Aug 13, 2015 at 8:50 AM, FX fxcoud...@gmail.com wrote:
And yet, that patch has absolutely nothing to do with libbacktrace.
Hmmm. Let's try this one.
Works perfectly with the patch:
Patch tested and committed with this ChangeLog entry.
2015-08-13 Ian Lance Taylor i...@google.com
*
On Aug 12, 2015, at 7:43 PM, Segher Boessenkool seg...@kernel.crashing.org
wrote:
Yes. And there are much worse problems, like many things not working
right if your HOST_WIDE_INT would happen to be more than 64 bits; we
cannot really shake those out because there is no actual system to
test
I'm currently working on migrating debugging information for Ada from GNAT
encodings to standard DWARF. At the moment, I have worked on two topics that
I believe are not (completely) supported in standard DWARF:
- fixed point types with arbitrary scale factors;
- scalar types with biased
On Thu, Aug 13, 2015 at 11:31 AM, Yuri Rumyantsev ysrum...@gmail.com wrote:
Hi All,
Here is patch for adding march/mtune options for Skylake.
Bootstrap and regression testing did not show any new failures.
Is it OK for trunk?
OK.
Thanks,
Uros.
ChangeLog:
2015-08-13 Yuri Rumyantsev
Hi all.
These two patches are refactoring of dominator-related code.
The comment in dominance.c says: We work in a poor-mans object oriented
fashion, and carry an instance of this structure through all our 'methods'. So,
the first patch converts the mentioned structure (dom_info) into a class
On 07/28/2015 03:38 AM, Bin Cheng wrote:
Hi,
For now, SCEV may compute iv base in the form of (signed T)((unsigned
T)base + step)). This complicates other optimizations/analysis depending
on SCEV because it's hard to dive into type conversions. For many cases,
such type conversions can be
On 08/05/2015 01:44 AM, Andreas Schwab wrote:
FAIL: gcc.dg/pr66314.c (test for excess errors)
Excess errors:
/usr/local/gcc/gcc-20150805/gcc/testsuite/gcc.dg/pr66314.c:1:0: warning:
-fsanitize=address and -fsanitize=kernel-address are not supported for this
target
Fixing this is
Ughh... what was supposed to be a short hack turned out to be 2 days of
pure torture.
So...I'm implementing task priorities as per the 4.1 standard, and
decided to use a priority queue. Since we already have splay trees in
libgomp, I figured these priority queues could be built on top of
On 07/28/2015 03:36 AM, Bin Cheng wrote:
Hi,
Loop niter computes inaccurate bound information for different loops. This
patch is to improve it by using loop initial condition in
determine_value_range. Generally, loop niter is computed by subtracting
start var from end var in loop exit
Jeff Law l...@redhat.com writes:
On 08/05/2015 08:18 AM, Richard Sandiford wrote:
Building some targets results in a warning about orig_dup[i] potentially
being used uninitialised. I think the warning is fair, since it isn't
obvious that the reog_data-based loop bound remains unchanged
On 08/13/2015 04:12 AM, David Sherwood wrote:
Hi,
This patch follows on from
[PATCH][1/N] Change GET_MODE_INNER to always return a non-void mode
It is another tidy up, replacing the pattern
GET_MODE_BITSIZE (GET_MODE_INNER (m)) with GET_MODE_UNIT_BITSIZE (m). Also
replaces any calls to
On Thu, 13 Aug 2015, Mike Stump wrote:
On Aug 12, 2015, at 7:43 PM, Segher Boessenkool seg...@kernel.crashing.org
wrote:
Yes. And there are much worse problems, like many things not working
right if your HOST_WIDE_INT would happen to be more than 64 bits; we
cannot really shake those
On 08/13/2015 12:09 PM, Andrew MacLeod wrote:
There was a recent change which causes ira-int.h to depend on recog.h..
but doesnt include it. Every file which needs it included it before I
think, it just required some shuffling of order.
By including recog.h in ira-int.h ordering issues are
On 08/13/2015 04:00 AM, Richard Biener wrote:
On Wed, Aug 12, 2015 at 6:47 PM, Jeff Law l...@redhat.com wrote:
On 08/12/2015 10:24 AM, Ian Lance Taylor wrote:
On Wed, Aug 12, 2015 at 9:21 AM, Tom Tromey t...@tromey.com wrote:
Jeff In the past this has stalled on issues like how will
Looks good to me...
On Thu, Aug 13, 2015 at 10:33 AM, David Edelsohn dje@gmail.com wrote:
AIX stdio.h header includes code specific for C++ that looks like:
extern C {
#ifdef __cplusplus
#ifdef ferror
#undef ferror
inline int ferror(FILE * _p)
{
return ((_p)-_flag _IOERR);
On 08/13/2015 02:29 PM, Richard Sandiford wrote:
Jeff Law l...@redhat.com writes:
On 08/05/2015 08:18 AM, Richard Sandiford wrote:
Building some targets results in a warning about orig_dup[i] potentially
being used uninitialised. I think the warning is fair, since it isn't
obvious that the
Trevor Saunders tbsau...@tbsaunde.org writes:
On Wed, Aug 12, 2015 at 01:03:36PM -0600, Jeff Law wrote:
On 08/12/2015 12:57 PM, Richard Sandiford wrote:
Jeff Law l...@redhat.com writes:
On 08/10/2015 06:05 AM, tbsaunde+...@tbsaunde.org wrote:
From: Trevor Saunders tbsaunde+...@tbsaunde.org
On Fri, Aug 14, 2015 at 6:10 AM, Jeff Law l...@redhat.com wrote:
On 07/28/2015 03:38 AM, Bin Cheng wrote:
Hi,
For now, SCEV may compute iv base in the form of (signed T)((unsigned
T)base + step)). This complicates other optimizations/analysis depending
on SCEV because it's hard to dive into
On Fri, Aug 14, 2015 at 6:08 AM, Jeff Law l...@redhat.com wrote:
On 07/28/2015 03:36 AM, Bin Cheng wrote:
Hi,
Loop niter computes inaccurate bound information for different loops.
This
patch is to improve it by using loop initial condition in
determine_value_range. Generally, loop niter is
Reported as fixing the library on OS X Yosemite:
https://gcc.gnu.org/ml/jit/2015-q3/msg00147.html
Committed to trunk as r226882.
Committed to gcc-5-branch as r226883.
gcc/jit/ChangeLog:
* jit-playback.c (invoke_driver): On OS X, add
-Wl,-undefined,dynamic_lookup to the driver
Robert Suchanek robert.sucha...@imgtec.com writes:
Hi,
It was discovered that with the attached test case compiled with -O2
-funroll-loops, the regrename pass renamed one of the registers ($2)
to $8 that was not saved by the prologue.
The attached patch fixes it by defining macro
On Fri, 17 Jul 2015, Marek Polacek wrote:
On Fri, Jul 17, 2015 at 02:23:50PM +0200, Richard Biener wrote:
+/* Fold A - (A B) into ~B A. */
+(simplify
+ (minus (convert? @0) (convert? (bit_and:c @0 @1)))
I think you want :s on the bit_and and its convert?
Ok, I keep forgetting about this
79 matches
Mail list logo
