On 12 September 2014 05:05, Jason Merrill wrote:
> Basically all we need to do here is accept it, like check_classfn; all the
> necessary checking was already done in check_explicit_specialization.
>
> Tested x86_64-pc-linux-gnu, applying to trunk.
I'd expect you want to remove the dg-bogus direc
>>> >For PR56408 we need some fix.
>> BTW, is there anything special about Fortran ? There are at least 180 test
>> files that contain 'dg-additional-sources' >some in a very non-local way:
>The current scheme comes at its limits in that case. . See the files listed in
>the PR for issues.
So, w
On Wed, Sep 10, 2014 at 6:21 PM, Bin.Cheng wrote:
> On Wed, Sep 10, 2014 at 4:57 PM, Kyrill Tkachov
> wrote:
>>
>> On 10/09/14 09:40, Christophe Lyon wrote:
>>>
>>> Hi,
>>
>> Hi Christophe,
>>
>>>
>>> On 9 September 2014 13:02, Ramana Radhakrishnan
>>> wrote:
On Tue, Aug 19, 2014 at 4
Basically all we need to do here is accept it, like check_classfn; all
the necessary checking was already done in check_explicit_specialization.
Tested x86_64-pc-linux-gnu, applying to trunk.
commit f0ef26fb06fe5798cd7bb071a932a21cf5e5a878
Author: Jason Merrill
Date: Thu Sep 11 16:45:47 2014
This patch removes the LIBGCC2_HAS_{SF,DF,XF,TF}_MODE target macros,
replacing them by predefines with -fbuilding-libgcc, together with a
target hook that can influence those predefines when needed.
The new default is that a floating-point mode is supported in libgcc
if (a) it passes the scalar_mo
On Thu, Sep 11, 2014 at 08:06:02PM -0400, David Malcolm wrote:
> On Thu, 2014-09-11 at 18:49 -0400, tsaund...@mozilla.com wrote:
> > From: Trevor Saunders
> >
> > Hi,
> >
> > This changes all callers of INSN_DELETED_P to use one of insn->deleted ()
> > insn->set_deleted () or insn->set_undelete
On 09/09/14 02:16, Richard Biener wrote:
On Tue, Sep 9, 2014 at 2:00 AM, Aldy Hernandez wrote:
On 09/05/14 02:00, Richard Biener wrote:
What I have in mind is:
1. Move the FE specific things that come before the call to
finalize_compilation_unit currently in each LANG_HOOKS_WRITE_GLOBALS, i
Hi David,
As discussed, I have updated the to use SYMBOL-$ syntax for
PC-relative addressing and to use a new symbol, __gcc_unwind_dbase,
rather than overloading __dso_handle.
Regards,
Andrew
2014-09-09 Andrew Dixie
Move exception tables to read-only memory on AIX.
* collect2.c
On Thu, 2014-09-11 at 18:49 -0400, tsaund...@mozilla.com wrote:
> From: Trevor Saunders
>
> Hi,
>
> This changes all callers of INSN_DELETED_P to use one of insn->deleted ()
> insn->set_deleted () or insn->set_undeleted () depending on what they're
> doing (set_deleted / set_undeleted seem som
This patch is for the google/gcc-4_9 branch only.
Use one of the reserved bits in the flag byte to tell gold which
symbols in the pubnames table can be entered just once in the
.gdb_index (rather than listing all CUs in which a given symbol
appears). We set this bit for class/struct types and name
On Sep 11, 2014, at 3:15 PM, Jakub Jelinek wrote:
> That is news to me, but given the amount of test -a/-o uses e.g. in
> gcc/configure and hundreds of places, I'd say what we care is what is more
> portable to old shells.
No, we can’t care about that. If that were true, the _ && _ in the compil
From: Trevor Saunders
Hi,
This changes all callers of INSN_DELETED_P to use one of insn->deleted ()
insn->set_deleted () or insn->set_undeleted () depending on what they're doing
(set_deleted / set_undeleted seem somewhat clearer to me than = 0 / 1).
bootstrapped + regtested on x86_64-unknown
From: Trevor Saunders
Hi,
pretty trivial, but not quiet just changing the types of variables.
bootstrapped + regtested on x86_64-unknown-linux-gnu, and run through
config-list.mk with a couple other patches. ok?
Trev
gcc/ChangeLog:
2014-09-11 Trevor Saunders
* config/mn10300/mn103
On Thu, Sep 11, 2014 at 11:24:08PM +0200, Bernhard Reutner-Fischer wrote:
> On 11 September 2014 20:19:31 Jakub Jelinek wrote:
>
> >On Thu, Sep 11, 2014 at 07:26:37PM +0200, Jakub Jelinek wrote:
> >> right now. The patch below intends to serialize the content of the
> >> problematic *.exp tests
On 9/11/2014 4:18 PM, Jeff Law wrote:
On 09/07/14 11:52, John David Anglin wrote:
The attached change fixes an ICE in dbxout_type on hppa2.0w-hp-hpux11.11
compiling auto-fn24.C.
Tested on hppa2.0w-hp-hpux11.11 and hppa64-hp-hpux11.11.
OK for trunk and 4.9?
Is dbxout.c allowed to look at langua
On 11 September 2014 20:19:31 Jakub Jelinek wrote:
On Thu, Sep 11, 2014 at 07:26:37PM +0200, Jakub Jelinek wrote:
> right now. The patch below intends to serialize the content of the
> problematic *.exp tests (the first runtest to reach one of those will simply
> run all the tests from that *.
On 08/15/14 09:08, Ilya Enkovich wrote:
On 17 Jul 03:36, Jeff Law wrote:
On 04/16/14 05:52, Ilya Enkovich wrote:
diff --git a/gcc/doc/tm.texi b/gcc/doc/tm.texi
index b8ca17e..d868129 100644
--- a/gcc/doc/tm.texi
+++ b/gcc/doc/tm.texi
@@ -4333,6 +4333,13 @@ This hook returns the va_list type of
On Thu, Sep 11, 2014 at 02:55:43PM -0600, Jeff Law wrote:
> On 09/05/14 16:38, Trevor Saunders wrote:
> >
> >So, in this case it seems to me you basically have two options
> >
> >A. change the macro to an inline function, and fix up all the callers to
> >pass the right type. Then rebase that into
On 09/04/14 22:04, Kugan wrote:
Hi All,
For the bswap built-in, there are unnecessary uxts generated as reported
in pr43550. Can we rely on the argument being unsigned and set the
SUBREG promoted accordingly.
At least in ARM ABI, arguments are supposed to be properly zero/sign
extended.
Any th
2014-09-11 21:39 GMT+01:00 Jeff Law :
>
> On 09/08/14 07:57, Jiong Wang wrote:
>>>
>>> Conceptually OK. Some questions/concerns about the implementation.
>>
>> Hi Jeff,
>>
>>thanks very much for your review.
>>>
>>> It seems to me that what you're trying to describe on the RHS is
>>> REG_P ||
On 09/05/14 15:57, David Malcolm wrote:
One other aspect of my approach is that (believe it or not) I'm trying
to minimize the size of the changes, to avoid introducing pain when
backporting bugfixes from trunk to the branches.
Right. I believe and know you're trying to avoid unnecessary pain :
On 09/05/14 16:38, Trevor Saunders wrote:
So, in this case it seems to me you basically have two options
A. change the macro to an inline function, and fix up all the callers to
pass the right type. Then rebase that into some sort of reasonable
patch series.
b. add a function with a different
On 09/08/14 07:57, Jiong Wang wrote:
Conceptually OK. Some questions/concerns about the implementation.
Hi Jeff,
thanks very much for your review.
It seems to me that what you're trying to describe on the RHS is
REG_P || CONSTANT_P
yes.
and actually I am trying to detect all rtx wh
On 09/08/14 09:26, Richard Sandiford wrote:
This patch adds a destructor to target_ira_int, so that the data structures
it points to are freed when the parent target_globals is freed. It fixes
a memory leak with non-default subtargets.
Tested on x86_64-linux-gnu. OK to install?
Thanks,
Richar
On Thu, 11 Sep 2014, Marc Glisse wrote:
> On Thu, 11 Sep 2014, Joseph S. Myers wrote:
>
> > On Thu, 11 Sep 2014, Marc Glisse wrote:
> >
> > > I don't know what kind of test you have in mind, so I added a runtime
> > > test. I
> > > am just guessing that it probably fails on alpha because of PR 5
On Thu, Sep 11, 2014 at 9:52 PM, H.J. Lu wrote:
> ix86_option_override_internal turns off OPTION_MASK_ABI_X32 for -m64
> when TARGET_BI_ARCH == 2. But it fails to do so for -m16. This patch
> also turns off OPTION_MASK_ABI_X32 for -m16. OK for trunk and 4.9
> branch after testing on Linux/x86-6
On 09/07/14 11:52, John David Anglin wrote:
The attached change fixes an ICE in dbxout_type on hppa2.0w-hp-hpux11.11
compiling auto-fn24.C.
Tested on hppa2.0w-hp-hpux11.11 and hppa64-hp-hpux11.11.
OK for trunk and 4.9?
Is dbxout.c allowed to look at language specific nodes? Isn't
TEMPLATE_TYP
On 09/08/14 11:21, Wilco Dijkstra wrote:
Thanks! Jakub noticed a potential problem in this area a while back,
but I never came up with any code to trigger and have kept that issue on
my todo list ever since.
Rather than ensuring the inserted copy write a single register, it seems
to me we're be
On 09/11/14 13:01, Segher Boessenkool wrote:
Ping?
Not forgotten. Still waiting to hear back from Bin on my recommendation
that we drop the bogus note on the floor and avoid combining pseudos
with multiple sets like that.
Jeff
On 09/09/14 10:43, Vladimir Makarov wrote:
I've investigated the wrong code generation. I did a mistake in my last
patch excluding pic pseudo from live-range analysis when risky
transformations are on.
Here is the right version of all IRA/LRA changes relative to trunk. I
managed to compile an
ix86_option_override_internal turns off OPTION_MASK_ABI_X32 for -m64
when TARGET_BI_ARCH == 2. But it fails to do so for -m16. This patch
also turns off OPTION_MASK_ABI_X32 for -m16. OK for trunk and 4.9
branch after testing on Linux/x86-64 with -mx32?
Thanks.
H.J.
--
2014-09-11 H.J. Lu
> And thanks for the review, FX. Do you want to undo your Fortran-maintainer →
> mere-contributor status, given that you are now again a bit more involved in
> the GCC development?
Yeah, why not. I promise I'll be careful and only review things in my comfort
zone (which isn't so large).
I'll w
On 11.09.2014 20:33, VandeVondele Joost wrote:
>For PR56408 we need some fix.
BTW, is there anything special about Fortran ? There are at least 180 test
files that contain 'dg-additional-sources' some in a very non-local way:
Well, the question is what you want to do with the different files.
ok.
David
On Thu, Sep 11, 2014 at 11:49 AM, Teresa Johnson wrote:
> On Thu, Sep 11, 2014 at 10:17 AM, Xinliang David Li
> wrote:
>> Yes, that is what I meant.
>>
>> David
>>
>> On Thu, Sep 11, 2014 at 10:09 AM, Teresa Johnson
>> wrote:
>>> On Wed, Sep 10, 2014 at 3:31 PM, Xinliang David Li
On Thu, Sep 11, 2014 at 2:54 PM, Segher Boessenkool
wrote:
> The rs6000_get_timebase expander expands to rs6000_mftb_di for
> TARGET_POWERPC64, so the define_insn should allow it too.
>
> This fixes gcc.target/powerpc/ppc-get-timebase.c for -m32 -mpowerpc64.
>
> Okay for mainline?
>
>
> Segher
>
>
On Thu, Sep 11, 2014 at 06:33:27PM +, VandeVondele Joost wrote:
> > And these Fortran inter-test dependencies, which Tobias told me is
> > PR56408.
> > For PR56408 we need some fix.
>
> BTW, is there anything special about Fortran ? There are at least 180 test
> files that contain 'dg-additi
Ping?
On Wed, Sep 03, 2014 at 06:51:44AM -0700, Segher Boessenkool wrote:
> This came up when investigating PR62151. In that PR combine messes up a
> four-insn combination. It should really have done the combination of the
> first three insns in that. The second of those instructions sets a reg
The rs6000_get_timebase expander expands to rs6000_mftb_di for
TARGET_POWERPC64, so the define_insn should allow it too.
This fixes gcc.target/powerpc/ppc-get-timebase.c for -m32 -mpowerpc64.
Okay for mainline?
Segher
2014-11-11 Segher Boessenkool
* config/rs6000/rs6000.md (rs6000
On Thu, Sep 11, 2014 at 10:17 AM, Xinliang David Li wrote:
> Yes, that is what I meant.
>
> David
>
> On Thu, Sep 11, 2014 at 10:09 AM, Teresa Johnson wrote:
>> On Wed, Sep 10, 2014 at 3:31 PM, Xinliang David Li
>> wrote:
>>> Can you share the buildinfo reader code with the merger by defininig
On 11.09.2014 18:30, Bernd Schmidt wrote:
So it looks like the following patch would be the right thing? I'm
afraid I failed to construct a compileable Fortran testcase for scalbn.
On 11.09.2014 18:32, FX wrote:
I would think so.
Looks also good to me. Thanks for the patch, Bernd. And thanks
> And these Fortran inter-test dependencies, which Tobias told me is
> PR56408.
> For PR56408 we need some fix.
BTW, is there anything special about Fortran ? There are at least 180 test
files that contain 'dg-additional-sources' some in a very non-local way:
./objc.dg/foreach-2.m: /* {
On Thu, Sep 11, 2014 at 11:10 AM, Uros Bizjak wrote:
> On Thu, Sep 11, 2014 at 3:00 PM, Kirill Yukhin
> wrote:
>
>> Patch in the bottom extends integer max/min patterns.
>> Also, it seems, like rounding variant was generated
>> for maxmin patterns. Bug fixed.
>>
>> Bootstrapped.
>> AVX-512* test
On Thu, Sep 11, 2014 at 07:26:37PM +0200, Jakub Jelinek wrote:
> right now. The patch below intends to serialize the content of the
> problematic *.exp tests (the first runtest to reach one of those will simply
> run all the tests from that *.exp file, others will skip it).
Forgotten patch below.
On Thu, Sep 11, 2014 at 3:16 PM, Kirill Yukhin wrote:
> Patch in the bottom extends patterns for rotate, ternlog and align.
>
> Bootstrapped.
> AVX-512* tests on top of patch-set all pass
> under simulator.
>
> Is it ok for trunk?
>
> gcc/
> * config/i386/sse.md
> (define_mode_ite
On Thu, Sep 11, 2014 at 3:00 PM, Kirill Yukhin wrote:
> Patch in the bottom extends integer max/min patterns.
> Also, it seems, like rounding variant was generated
> for maxmin patterns. Bug fixed.
>
> Bootstrapped.
> AVX-512* tests on top of patch-set all pass
> under simulator.
>
> Is it ok for
On Thu, 11 Sep 2014, Joseph S. Myers wrote:
On Thu, 11 Sep 2014, Marc Glisse wrote:
I don't know what kind of test you have in mind, so I added a runtime test. I
am just guessing that it probably fails on alpha because of PR 58757, I can't
test. Computing d+d may be even more likely to trigger
Sorry I let this slide.
On Mon, Jul 14, 2014 at 01:54:13PM +0200, Jakub Jelinek wrote:
> On Sun, Jul 13, 2014 at 07:55:44PM +0200, Marek Polacek wrote:
> > 2014-07-13 Marek Polacek
> >
> > * ubsan.h (struct ubsan_mismatch_data):
>
> Missing description.
Fixed.
> > + gcc_assert (TREE_C
On Thu, Sep 11, 2014 at 05:04:56PM +, VandeVondele Joost wrote:
> > Here is a patch I'm testing now:
>
> I also tested your patch to compare timings vs a newer patch (v8) I'll send
> soon
>
> == patch v8 == make -j32 -k ==
> check-fortran 4m58.178s
> check-c++ ~10m
> check-c ~10
Yes, that is what I meant.
David
On Thu, Sep 11, 2014 at 10:09 AM, Teresa Johnson wrote:
> On Wed, Sep 10, 2014 at 3:31 PM, Xinliang David Li wrote:
>> Can you share the buildinfo reader code with the merger by defininig
>> some hooks for different callbacks?
>
> Do you mean the two blobs of co
On Wed, Sep 10, 2014 at 3:31 PM, Xinliang David Li wrote:
> Can you share the buildinfo reader code with the merger by defininig
> some hooks for different callbacks?
Do you mean the two blobs of code guarded by 'if (tag ==
GCOV_TAG_BUILD_INFO)' that I added here and the existing one in
gcov_exit
> Here is a patch I'm testing now:
Hi Jakub,
I also tested your patch to compare timings vs a newer patch (v8) I'll send soon
== patch v8 == make -j32 -k ==
check-fortran 4m58.178s
check-c++ ~10m
check-c ~10m
check 15m29.873s
== patch Jakub
check-c++ ~20m
check-fortran
> On 09/11/2014 03:06 AM, Jan Hubicka wrote:
> >http://kam.mff.cuni.cz/~hubicka/odr-warnings-firefox.txt
>
> >/aux/hubicka/firefox4/content/media/fmp4/ffmpeg/libav53/include/libavcodec/avcodec.h:997:0:
> > note: the first difference of corresponding definitions is field ‘data’
> > uint8_t *dat
> On Thu, 11 Sep 2014, Jan Hubicka wrote:
>
> > Hi,
> > this patch adds computation and streaming of mangled type names. As
> > suggested by Jason,
> > it simple calls DECL_ASSEMBLER_NAME on all names types and lets C++ supply
> > them.
> > This makes it possible to stablish precise ODR type eq
OK, thanks.
Jason
On 09/11/2014 03:06 AM, Jan Hubicka wrote:
http://kam.mff.cuni.cz/~hubicka/odr-warnings-firefox.txt
/aux/hubicka/firefox4/content/media/fmp4/ffmpeg/libav53/include/libavcodec/avcodec.h:997:0:
note: the first difference of corresponding definitions is field ‘data’
uint8_t *data[AV_NUM_DAT
Hi,
On 09/11/2014 05:06 PM, Jason Merrill wrote:
Do we need a documentation update?
I agree. Something like the below would do?
Thanks,
Paolo.
2014-09-11 Paolo Carlini
PR c++/61489
* doc/invoke.texi ([-Wmissing-field-initializers]): Update.
/cp
2014-09-11
On Thu, 11 Sep 2014, Marc Glisse wrote:
> I don't know what kind of test you have in mind, so I added a runtime test. I
> am just guessing that it probably fails on alpha because of PR 58757, I can't
> test. Computing d+d may be even more likely to trigger potential issues, if
> that's the goal.
> "Jakub" == Jakub Jelinek writes:
Jakub> I fear that is going to be too expensive, because e.g. all the
Jakub> caching that dejagnu and our tcl stuff does would be gone, all
Jakub> the tests for lp64 etc. would need to be repeated for each test.
In gdb I arranged to have this stuff saved i
On 09/11/2014 06:34 PM, Steven Bosscher wrote:
On Thu, Sep 11, 2014 at 6:19 PM, Bernd Schmidt wrote:
What do you expect that function to do different? It returns the correct
value.
No different. Just that if you want to check whether DECL is a global
variable then we have a predicate for it.
On Thu, Sep 11, 2014 at 6:19 PM, Bernd Schmidt wrote:
>>> What do you expect that function to do different? It returns the correct
>>> value.
>>>
>>
>> No different. Just that if you want to check whether DECL is a global
>> variable then we have a predicate for it. So why use TREE_STATIC
>> instea
On Thu, Sep 11, 2014 at 9:18 AM, Maxim Ostapenko
wrote:
>
> Working on ICE debugging patch, I've noted that libiberty interface doesn't
> allow to append stdout and stderr to existing files.
>
> This small patch provides two new flags for pex_run and extends open_write
> interface to handle the is
> So it looks like the following patch would be the right thing?
I would think so.
FX
On 09/11/2014 12:37 PM, FX wrote:
Changing the fntype[2] looks wrong to me, as it is also used for
powi(double, int) , where the argument order matches the current
version:
Ah, sorry. I only looked at mathbuiltins.def and didn't spot the other use.
It looks like fntype[5] is actually what you
Hi, Joseph,
Thanks for your review! I've added comments for new functions and
replaced POSIX subprocess interfaces with libiberty's ones.
In general, when cc1 or cc1plus ICE-es, we try to reproduce the bug by
running compiler 3 times and comparing stderr and stdout on each attempt
with respe
On 09/11/2014 06:15 PM, Steven Bosscher wrote:
On Thu, Sep 11, 2014 at 6:06 PM, Bernd Schmidt wrote:
On 09/11/2014 05:55 PM, Steven Bosscher wrote:
On Thu, Sep 11, 2014 at 3:26 PM, Bernd Schmidt wrote:
nvptx will be the first port to use BImode and have STORE_FLAG_VALUE==-1.
That has exposed
Working on ICE debugging patch, I've noted that libiberty interface
doesn't allow to append stdout and stderr to existing files.
This small patch provides two new flags for pex_run and extends
open_write interface to handle the issue.
Does this patch look sane?
-Maxim
libiberty/ChangeLog:
2
On Thu, Sep 11, 2014 at 6:06 PM, Bernd Schmidt wrote:
> On 09/11/2014 05:55 PM, Steven Bosscher wrote:
>>
>> On Thu, Sep 11, 2014 at 3:26 PM, Bernd Schmidt wrote:
>>>
>>> nvptx will be the first port to use BImode and have STORE_FLAG_VALUE==-1.
>>> That has exposed a bug in combine where we can end
On 09/11/2014 05:55 PM, Steven Bosscher wrote:
On Thu, Sep 11, 2014 at 3:26 PM, Bernd Schmidt wrote:
nvptx will be the first port to use BImode and have STORE_FLAG_VALUE==-1.
That has exposed a bug in combine where we can end up calling
num_sign_bit_copies for a BImode value. However, the return
On 09/11/2014 02:10 PM, Yury Gribov wrote:
On 09/11/2014 01:18 PM, Richard Biener wrote:
On Thu, Sep 11, 2014 at 11:06 AM, Richard Biener
wrote:
>On Wed, Sep 10, 2014 at 10:09 AM, Yury Gribov
wrote:
>>Hi all,
>>
>>This is a second version of patch which adds a Vim config
(.local.vimrc)
>>to r
Committed to branch dmalcolm/jit:
ChangeLog.jit:
* MAINTAINERS (Various Maintainers): Add myself as jit maintainer.
gcc/jit/ChangeLog.jit:
* TODO.rst (Initial Release): Update for addition of myself as
maintainer.
---
ChangeLog.jit | 4
MAINTAINERS
On Thu, Sep 11, 2014 at 3:27 PM, Bernd Schmidt wrote:
> It turns out that we're calling eliminate_regs for global variables which
> can't possibly have eliminable regs in their decl. At that point,
> reg_eliminate can be NULL. This patch avoids unnecessary work, and allows us
> to add an assert to
On Thu, Sep 11, 2014 at 3:26 PM, Bernd Schmidt wrote:
> nvptx will be the first port to use BImode and have STORE_FLAG_VALUE==-1.
> That has exposed a bug in combine where we can end up calling
> num_sign_bit_copies for a BImode value. However, the return value is always
> 1 in that case, so it doe
On 11/09/14 14:55, Jiong Wang wrote:
On 11/09/14 14:43, Christophe Lyon wrote:
Hi Jiong,
On 9 September 2014 12:59, Ramana Radhakrishnan
wrote:
On Mon, Aug 18, 2014 at 11:31 AM, Jiong Wang wrote:
this patch enable auto-vectorization for copysignf by using vector
bit selection instruction o
On Thu, Sep 11, 2014 at 3:25 PM, Bernd Schmidt wrote:
> Bootstrapped and tested on x86_64-linux, together with the other patches.
> Ok?
This is OK.
Ciao!
Steven
Jan Hubicka writes:
> Hi,
> Libreoffice fails to build because ltrans tries to fetch in variable
> constructor
> that is not shipped there. Fixed thus.
>
> Bootstrapped/regtested x86_64-linux.
>
> Honza
>
> * varpool.c (varpool_node::ctor_useable_for_folding_p): Do not try
> to acc
Gerald Pfeifer writes:
> Hi Igor,
>
> On Wed, 10 Sep 2014, Zamyatin, Igor wrote:
>> + Complete support for http://cilk.org";>Cilk Plus
>> features was added to GCC
>> + [2014-09-02]
features *were* added, plural.
>> + Contributed by Jakub Jelinek, Iyer Balaji and Igor
>> Zamyatin.
Wh
I've kept the integer move costs at 1 - patch attached for commit as I don't
have write access.
ChangeLog:
2014-09-11 Wilco Dijkstra
* gcc/config/aarch64/aarch64.c:
(cortexa57_regmove_cost): New cost table for A57.
(cortexa53_regmove_cost): New cost table for A53.
OK, I'll skip this patch for now as HAVE_DESIGNATED_INITIALIZERS should
always be false, so there is no point in cleaning it up.
> -Original Message-
> From: Marcus Shawcroft [mailto:marcus.shawcr...@gmail.com]
> Sent: 04 September 2014 16:42
> To: Wilco Dijkstra
> Cc: gcc-patches@gcc.gnu.
Patch attached for commit as I don't have write access.
ChangeLog:
2014-09-11 Wilco Dijkstra
* gcc/config/aarch64/aarch64.c (aarch64_register_move_cost):
Fix Q register move handling. (generic_regmove_cost): Undo raised
FP2FP move cost as Q register moves are now hand
Patch attached for commit as I don't have write access.
> -Original Message-
> From: Marcus Shawcroft [mailto:marcus.shawcr...@gmail.com]
> Sent: 04 September 2014 16:23
> To: Wilco Dijkstra
> Cc: gcc-patches@gcc.gnu.org
> Subject: Re: [PATCH 1/4] AArch64: Fix register_move_cost
>
> On 4
On Thu, Sep 11, 2014 at 03:26:49PM +0100, Marcus Shawcroft wrote:
> On 11 September 2014 13:15, James Greenhalgh wrote:
> >
> > Hi,
> >
> > I'd been putting this patch off in the hope that I might find
> > time to move these intrinsics to a C/builtin implementation, but it
> > is probably better t
Ha OK, I had misunderstood your first email, and thought you had the
test also pass in big endian.
Thanks for the clarification.
On 11 September 2014 15:56, Alan Lawrence wrote:
> Yes, I had seen that, and the failure is expected. AFAICT the test is
> correct; but the implementation of vld[234]
Do we need a documentation update?
Jason
On Thu, Sep 11, 2014 at 10:06:40AM +0200, Jakub Jelinek wrote:
> There is an option to touch say *-parallel/finished file once any of the
> check-parallel-gcc-{1,2,...} goals is done (because when it finishes, it
> means all the tests for the particular check-$lang that are parallelizable
> have ei
On 11 September 2014 15:45, VandeVondele Joost
wrote:
>
>>> could it be that the pattern in normal1 should have been '[ab]*/ de*/
>>> [ep]*/*' ?
>>
>>Yes, we are running these tests multiple times:
>>
>>PASS: 23_containers/map/modifiers/erase/abi_tag.cc (test for excess errors)
>>PASS: 23_contai
>> could it be that the pattern in normal1 should have been '[ab]*/ de*/
>> [ep]*/*' ?
>
>Yes, we are running these tests multiple times:
>
>PASS: 23_containers/map/modifiers/erase/abi_tag.cc (test for excess errors)
>PASS: 23_containers/multimap/modifiers/erase/abi_tag.cc (test for excess
>erro
On 11 September 2014 13:15, James Greenhalgh wrote:
>
> Hi,
>
> I'd been putting this patch off in the hope that I might find
> time to move these intrinsics to a C/builtin implementation, but it
> is probably better to get them right for now and come back to improving
> them later.
>
> All four o
Original patch applied cleanly to 4.9 HEAD as r215175.
Marcus Shawcroft wrote:
On 20 August 2014 10:25, Alan Lawrence wrote:
The SIMD-register variant is miscategorized as "alu_reg" despite not using
any ALU registers, and should be "neon_add" for e.g. scheduling.
Tested with check-gcc and ch
Yes, I had seen that, and the failure is expected. AFAICT the test is correct;
but the implementation of vld[234] is incorrect on bigendian, because of
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=59810 .
HTH, Alan.
Christophe Lyon wrote:
On 9 September 2014 12:19, Marcus Shawcroft wrote:
On 11/09/14 14:43, Christophe Lyon wrote:
Hi Jiong,
On 9 September 2014 12:59, Ramana Radhakrishnan
wrote:
On Mon, Aug 18, 2014 at 11:31 AM, Jiong Wang wrote:
this patch enable auto-vectorization for copysignf by using vector
bit selection instruction on arm32 when neon available.
I've no
On 9 September 2014 12:19, Marcus Shawcroft wrote:
> On 8 September 2014 11:35, Alan Lawrence wrote:
>> This adds a test of all the variants of vld2, vld2q, vld3, vld3q, vld4, and
>> vld4q. These all use typexNxM structs and the OI/CI/XImode mechanism, so the
>> test cross-checks this against pla
This testcase was breaking because multiple substitutions of ContainerEndA> into T... would lose the original arguments when we pull
the pattern out of the intermediate expansion. So the added assert
guards against that happening again, and we avoid the situation by
simplifying the handling of
Hi Jiong,
On 9 September 2014 12:59, Ramana Radhakrishnan
wrote:
> On Mon, Aug 18, 2014 at 11:31 AM, Jiong Wang wrote:
>> this patch enable auto-vectorization for copysignf by using vector
>> bit selection instruction on arm32 when neon available.
>>
I've noticed that your new testcase fails (t
It turns out that we're calling eliminate_regs for global variables
which can't possibly have eliminable regs in their decl. At that point,
reg_eliminate can be NULL. This patch avoids unnecessary work, and
allows us to add an assert to eliminate_regs later.
Bootstrapped and tested on x86_64-l
nvptx will be the first port to use BImode and have
STORE_FLAG_VALUE==-1. That has exposed a bug in combine where we can end
up calling num_sign_bit_copies for a BImode value. However, the return
value is always 1 in that case, so it doesn't tell us anything and is
going to be misinterpreted by
The nvptx backend is somewhat unusual in that call insns set a pseudo.
The combiner is surprised by this and allows combining them into other
insns, which remain as INSN rather than CALL_INSN. Aborts ensue.
Bootstrapped and tested on x86_64-linux, together with the other
patches. Ok?
Bernd
There's some code in get_uncond_jump_length to emit and then delete a
label and a jump. On ptx we skip register allocation and reload, and
this fails a "reload_completed || bb != NULL" assert in df_insn_delete.
Fixed by instead emitting the two insns into a sequence which we then
just discard.
Hello,
Patch in the bottom extends patterns for rotate, ternlog and align.
Bootstrapped.
AVX-512* tests on top of patch-set all pass
under simulator.
Is it ok for trunk?
gcc/
* config/i386/sse.md
(define_mode_iterator VI48_AVX512VL): New.
(define_expand "_vternlog_maskz")
Hello,
Patch in the bottom extends integer max/min patterns.
Also, it seems, like rounding variant was generated
for maxmin patterns. Bug fixed.
Bootstrapped.
AVX-512* tests on top of patch-set all pass
under simulator.
Is it ok for trunk?
gcc/
* config/i386/sse.md (VI128_256): Delete.
Hi all
we have merged the gcc-4_9-branch into linaro/gcc-4_9-branch up to
revision 214896 as r215060. We have also backported this set of revisions:
r211717 as r214313 : [AArch32] TARGET_ATOMIC_ASSIGN_EXPAND_FENV hook
r212927 as r214314 : [AArch32] Enable arm target in
ira-shrinkwrap-prep* testc
1 - 100 of 140 matches
Mail list logo
