On Thu, Aug 14, 2014 at 8:53 PM, Konstantin Serebryany
wrote:
> In order for your work to be generally useful, I'd ask several things:
> - Update
> https://code.google.com/p/address-sanitizer/wiki/CompileTimeOptimizations
> with examples that will be handled
Done (to be honest I only plan to do
On Thu, Aug 14, 2014 at 11:29 PM, Sebastian Pop wrote:
> Bin.Cheng wrote:
>> >> The overflow check can be improved by using deeper inspection to prove the
>> >> equality. This patch deals with that by making below two improvements:
>> >> a) Handles constant cases.
>> >> b) Uses affine expansi
On Fri, Aug 8, 2014 at 4:54 PM, Richard Biener
wrote:
> On Fri, Aug 8, 2014 at 10:05 AM, Bin.Cheng wrote:
>> On Thu, Aug 7, 2014 at 8:06 PM, Richard Biener
>> wrote:
>>> On Thu, Aug 7, 2014 at 11:46 AM, Bin Cheng wrote:
Hi,
As analyzed in PR62032, this patch fixes the latent lto bug b
On 08/06/14 11:22, David Malcolm wrote:
/
* rtx-classes-status.txt: Update.
---
rtx-classes-status.txt | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/rtx-classes-status.txt b/rtx-classes-status.txt
index b22cb1e..90d6efd 100644
--- a/rtx-classes-status.txt
+++
On 08/14/14 15:36, Richard Sandiford wrote:
Right. I think they're held up on patch 40 (ironically the one that
conflicts with yours).
I think we could declare side effects in notes as invalid and add some
ENABLE_CHECKING bits to enforce that. With those in place, my concerns
around #40 from
The attached patch enables the -pthread flag for cygwin targets. While
not strictly necessary for compiling or linking with pthread or _r
functions on Cygwin, accepting it allows for better compatibility with
other platforms.
--
Yaakov Selkowitz
Associate Software Engineer, ARM
Red Hat, Inc.
Andi Kleen writes:
Ping!
> From: Andi Kleen
>
> To use gcc-{ar,ranlib} for boot strap we need to add a -B option
> to the tool. Since ar has weird and unusual argument conventions
> implement the code by hand instead of using any libraries.
>
> v2: Fix typo
>
> gcc/:
>
> 2014-08-04 Andi Kleen
Minor optimization. Committed.
* config/rl78/rl78-real.md (addqi3_real): Allow adding global
variables to the accumulator.
Index: config/rl78/rl78-real.md
===
--- config/rl78/rl78-real.md(revision 213996)
+++ co
Letting GCC think that any mem-mem alternative is OK leads to trouble
with far mem to far mem moves, so split out the moves we can make.
Committed.
* config/rl78/predicates.md (rl78_near_mem_operand): New.
* config/rl78/rl78-virt.md (movqi_virt_mm, movqi_virt)
(movhi_virt_
The G10 family doesn't support this opcode. Committed.
* config/rl78/rl78-expand.md (umulqihi3): Disable for G10.
* config/rl78/rl78-virt.md (umulhi3_shift_virt): Likewise.
(umulqihi3_virt): Likewise.
* config/rl78/rl78-real.md (umulhi3_shift_real): Likewise.
On Thu, Aug 14, 2014 at 11:18 PM, Jason Merrill wrote:
> On 08/14/2014 04:31 AM, Bin Cheng wrote:
>>
>> g++.dg/ext/arm-fp16/fp16-mangle-1.C is failed because GCC now sets
>> DECL_COMDAT on template instantiations if flag_implicit_templates is in
>> effect. Then DECL_WEAK will be set accordingly.
Ping?
Best regards,
Thomas
> -Original Message-
> From: gcc-patches-ow...@gcc.gnu.org [mailto:gcc-patches-
> ow...@gcc.gnu.org] On Behalf Of Thomas Preud'homme
> Sent: Thursday, August 07, 2014 1:57 PM
> To: gcc-patches@gcc.gnu.org
> Subject: RE: [PATCH] Fix confusion between target, hos
Minor tweak, committed.
* config/rl78/rl78-virt.md (movhi_virt): Allow const->far moves.
Index: config/rl78/rl78-virt.md
===
--- config/rl78/rl78-virt.md(revision 213994)
+++ config/rl78/rl78-virt.md(working copy)
@@
On 08/14/2014 01:06 AM, Ramana Radhakrishnan wrote:
> On Mon, Aug 11, 2014 at 11:01 PM, Janis Johnson
> wrote:
>> The check for effective target arm_v8_neon_ok passes even if __ARM_ARCH
>> is not 8 or greater, but then some tests fail because intrinsic functions
>> used in the test have not been d
Hi,
testcase in PR 62092 triggers assert I added to temporarily check that new
ipa-devirt code handles all cases as ipa-prop devirtualization. In this case
ipa-prop gives wrong answer because of bug I introduced into
walk_aliased_vdefs. The function_entry_reached is supposed to be cleared at
begg
Hi,
I get a segafult in decl_maybe_in_construction_p during function versioning.
We have following steps in clone creation (e.g. as in
create_version_clone_with_body):
1. Create function decl
2. Create clone of cgraph node
3. Copy function body
After the first step there is no body attached
On 08/13/2014 08:48 PM, Ville Voutilainen wrote:
Ok, modified patch attached. The changelog entry was attached to the previous
mail as an evil binary attachment (gmail...)
Ah, so it was.
here, for convenience:
Thanks; having it in the body of the mail is more convenient for me as
'git am'
David Malcolm writes:
> On Tue, 2014-08-12 at 15:08 -0600, Jeff Law wrote:
>> On 08/06/14 11:19, David Malcolm wrote:
>> > gcc/
>> >* rtl.h (for_each_rtx_in_insn): New function.
>> >* rtlanal.c (for_each_rtx_in_insn): Likewise.
>> OK. Note that we're moving away from for_each_rtx... I h
On Mon, 28 Jul 2014, Herman, Andrei wrote:
> Please find attached the fixed patch files for this change:
> 1. Add command line option -fforce-dwarf-lexical-blocks.
> 2. Support flag_force_dwarf_blocks in C.
The front-end parts of these patches are OK with the changes indicated
below
On Tue, 2014-08-12 at 15:08 -0600, Jeff Law wrote:
> On 08/06/14 11:19, David Malcolm wrote:
> > gcc/
> > * rtl.h (for_each_rtx_in_insn): New function.
> > * rtlanal.c (for_each_rtx_in_insn): Likewise.
> OK. Note that we're moving away from for_each_rtx... I haven't
> looked, but there'
The attached patch implements premature DLL unloading prevention in
__cxa_thread_atexit for Cygwin and MinGW-w64 targets. The mingw.org
target is welcome to do the same in their os_defines.h, but this code
does require Windows XP/2003, and they have historically catered to
older platforms.
M
This patch implements __cxa_atexit support for Cygwin targets. This
requires Cygwin 1.7.32 and binutils master.
Net difference in check-c++ results on i686-pc-cygwin:
# of unexpected failures-11
# of unexpected successes -3
# of expected failures -61
# of unsupported test
Thank you. I fixed the typo and committed.
On Thu, Aug 14, 2014 at 1:49 PM, Teresa Johnson wrote:
> On Thu, Aug 14, 2014 at 1:46 PM, Yi Yang wrote:
>> Patch v2.
>>
>> Trunk no longer set SECTION_NAME for implicit section names, so this
>> probably does not apply to trunk. It's probably not neces
On Thu, Aug 14, 2014 at 1:46 PM, Yi Yang wrote:
> Patch v2.
>
> Trunk no longer set SECTION_NAME for implicit section names, so this
> probably does not apply to trunk. It's probably not necessary for
> trunk either.
>
> Tested for Google 4.8(albeit unnecessary) and 4.9 branch.
>
> diff --git gcc/
Patch v2.
Trunk no longer set SECTION_NAME for implicit section names, so this
probably does not apply to trunk. It's probably not necessary for
trunk either.
Tested for Google 4.8(albeit unnecessary) and 4.9 branch.
diff --git gcc/bb-reorder.c gcc/bb-reorder.c
index a1b3e65..b9a829e 100644
---
On Wed, 2014-08-13 at 20:55 -0600, Jeff Law wrote:
> On 08/13/14 18:11, David Malcolm wrote:
> > On Wed, 2014-08-13 at 14:34 -0600, Jeff Law wrote:
> >> On 08/13/14 14:28, David Malcolm wrote:
> >>> Thanks. Although this function gets converted back to a macro in patch
> >>> 191, I just realized t
Hi Tobias (or any other Fortran maintainer),
Is this patch OK?
https://gcc.gnu.org/ml/gcc-patches/2014-08/msg00488.html
Thanks,
Manuel.
On Thu, Aug 14, 2014 at 11:36 AM, Xinliang David Li wrote:
> Ok.
>
> The interfaces of counter reading/getting now becomes confusing. Should it
> be better documented somewhere so that developer knows what is the right one
> to use in a certain context?
I think it is documented in libgcov.h and a
On 08/06/14 11:21, David Malcolm wrote:
gcc/
* config/bfin/bfin-protos.h (asm_conditional_branch): Strengthen
param 1 from rtx to rtx_insn *.
* config/bfin/bfin.c (expand_prologue_reg_save): Likewise for
the various locals named "insn".
(expand_epilogue_reg
OK, thanks.
Jason
On 13/08/2014 11:50, Jonathan Wakely wrote:
Yes you can, it's conforming to replace a (non-virtual) member function
with default arguments by two or more member functions. We do it all
the time.
See 17.6.5.5 [member.functions] p2.
You should have told it sooner ! But of course no-one is s
--- a/gcc/fortran/options.c
+++ b/gcc/fortran/options.c
@@ -693,6 +693,10 @@ gfc_handle_option (size_t scode, const char *arg,
int value,
gfc_option.warn_line_truncation = value;
break;
+case OPT_Wmissing_return:
+ warn_missing_return = value;
+ break;
+
case OPT_W
Steve Ellcey wrote:
> I understand the desire not to add optimizations just for benchmarks but
> we do know other compilers have added this optimization for coremark
> (See
> http://community.arm.com/groups/embedded/blog/2013/02/21/coremark-and-compiler-performance)
> and the 13 people on the CC li
Hi,
a few cases where under SFINAE we just go ahead and we don't immediately
return error_mark_node. During the work I also noticed a glitch in
cxx_sizeof_or_alignof_type where we do complain & tf_warning_or_error
but in that case complain is a bool.
Tested x86_64-linux as usual.
Thanks,
Pa
Fix a couple problems found during testing.
Backport from trunk (r212694) failed to fixup gcov_read_counter
invocations in google-specific code. Also, forward port
r211800 from google/4_8 to tolerate differences after COMDAT
fixup.
Passes manual testing, ok if passes regression tests?
Thanks,
Te
On Thu, 2014-08-14 at 10:21 -0600, Jeff Law wrote:
> On 08/14/14 10:12, David Malcolm wrote:
> > On Thu, 2014-08-14 at 09:56 -0600, Jeff Law wrote:
> >> On 08/14/14 04:32, Richard Biener wrote:
> You'll note in a separate thread Steve and I discussed this during
> Cauldron
> and it
On Wed, Aug 13, 2014 at 9:03 PM, Yi Yang wrote:
> This bug is caused by my last patch, which did not differentiate
> between explicit section names (via attributes) and implicit section
> names (via -ffunction-section).
>
> This patch fixes that.
>
> --
>
> diff --git gcc/bb-reorder.c gcc/bb-reord
On Thu, Aug 14, 2014 at 07:40:52PM +0200, Thomas Koenig wrote:
> Hello world,
>
> the attached patch fixes the regression by making sure we never
> try to create a temporary variable from a temporary variable,
> which happened in the wrong order.
>
> Regression-tested. OK for trunk and 4.9?
>
On Thu, 14 Aug 2014, Marek Polacek wrote:
> The DR#458 is about the usage of ATOMIC_*_LOCK_FREE macros defined in
> stdatomic.h in the #if directives. Proposed Technical Corrigendum of
> this DR is that these macros should expand to constant expressions
> suitable for use in #if preprocessing dir
Richard Biener wrote:
>
> This fixes wrong answer from data-dependence analysis by realizing
> that _all_ (even non-evolving) indirect accesses cannot be constrained
> to a full object size. This also gets rid of that ugly
> DR_UNCONSTRAINED_BASE hack (but effectively make it always active).
>
The DR#458 is about the usage of ATOMIC_*_LOCK_FREE macros defined in
stdatomic.h in the #if directives. Proposed Technical Corrigendum of
this DR is that these macros should expand to constant expressions
suitable for use in #if preprocessing directives. This patch does
that by mapping these mac
On Wed, 2014-08-13 at 11:52 +0200, Richard Biener wrote:
> On Wed, Aug 13, 2014 at 4:54 AM, Bin.Cheng wrote:
> > On Wed, Aug 13, 2014 at 4:40 AM, Jeff Law wrote:
> >> On 08/12/14 14:23, Richard Biener wrote:
> >>> On August 12, 2014 8:31:16 PM CEST, Jeff Law wrote:
> On 08/12/14 11:46, Stev
Hello world,
the attached patch fixes the regression by making sure we never
try to create a temporary variable from a temporary variable,
which happened in the wrong order.
Regression-tested. OK for trunk and 4.9?
2014-08-19 Thomas Koenig
PR fortran/62106
* gfortran.h (symb
On 08/13/2014 05:29 AM, Kyrill Tkachov wrote:
> Is the attached patch ok? It just moves the section as you suggested. I did a
> build of the Linux kernel with and without this patch to make sure no code-gen
> was accidentally affected.
Looks good.
> We'd need to store a mapping from constant to R
On Wed, Aug 13, 2014 at 09:03:37PM +0200, Manuel López-Ibáñez wrote:
> I don't think this is the right fix. The problem is that we are trying
> to print the caret in a column that is larger than the line_width. We
> do this because the file given by the line directive has nothing to do
> with the a
Richard Biener wrote:
>
> The following fixes missing dominator computation before fixing loops.
> Rather than doing even more such weird stuff in a pass gate function
> this puts this into a new pass scheduled before the loop passes gate.
>
Ok.
> +unsigned int
> +pass_fix_loops::execute (funct
On 12/08/2014 19:48, Joseph S. Myers wrote:
> On Mon, 11 Aug 2014, Sylvestre Ledru wrote:
>
>>> The test Wmissing-return2.c only has one of the two warnings. But as per
>>> "-Wreturn-type => Run both", and for backwards compatibility with the
>>> existing definition of -Wreturn-type, both warnin
Indeed, thanks for working on this.
We've been wanting such optimization phase from day one, but never got
to implementing it (except for a few simple ones).
https://code.google.com/p/address-sanitizer/wiki/CompileTimeOptimizations
There have been several attempts outside of our team to do such
opt
Hi!
I've committed following fix for two issues revealed e.g. by valgrind
on some of the udr*.f90 testcases.
buffer could be uninitialized, and gfc_free_omp_udr could free symbols
in the combiner_ns (or initializer_ns) when freeing those whole namespaces,
so if some symbols were queued for commit/
OK.
Jason
Hi!
Right now when -fsanitize=leak adds -llsan, it adds it late on the command
line, so e.g. -lstdc++ comes after it, which seems to be bad.
The following patch puts it early on the link command line like we do for
-lasan or -ltsan. Bootstrapped/regtested on x86_64-linux and i686-linux,
ok for tr
On Wed, 2014-08-13 at 20:13 -0400, David Malcolm wrote:
> On Wed, 2014-08-06 at 13:19 -0400, David Malcolm wrote:
> > This is the patch series I spoke about at Cauldron in the talk
> > "A proposal for typesafe RTL"; slides here:
> > http://dmalcolm.fedorapeople.org/presentations/cauldron-2014/rtl
>
On Thu, 2014-08-14 at 09:56 -0600, Jeff Law wrote:
> On 08/14/14 04:32, Richard Biener wrote:
> >> You'll note in a separate thread Steve and I discussed this during Cauldron
> >> and it was at my recommendation Steve resurrected his proof of concept
> >> plugin and started beating it into shape.
>
On 08/14/2014 05:07 AM, Richard Biener wrote:
So - can you take over this C++ frontend issue?
OK.
Jason
On 08/14/14 10:12, David Malcolm wrote:
On Thu, 2014-08-14 at 09:56 -0600, Jeff Law wrote:
On 08/14/14 04:32, Richard Biener wrote:
You'll note in a separate thread Steve and I discussed this during Cauldron
and it was at my recommendation Steve resurrected his proof of concept
plugin and start
Hi,
On 08/14/2014 04:18 PM, Jason Merrill wrote:
On 08/14/2014 07:14 AM, Paolo Carlini wrote:
+ nparms -= variadic_p ? variadic_p : default_p;
What if you have both default arguments and parameter packs?
Right. Got distracted by the minor secondary issues...
It seems to me that
Hi all
we have merged the gcc-4_9-branch into linaro/gcc-4_9-branch up to
revision 213803 as r213943. We have also backported this set of revisions:
r211140 as r213455 [AArch64] Drop ISB after FPCR write.
r211270 as r213790 [AArch64] Remove from arm_neon.h functions not
in the spec
r21
On 08/14/14 04:32, Richard Biener wrote:
You'll note in a separate thread Steve and I discussed this during Cauldron
and it was at my recommendation Steve resurrected his proof of concept
plugin and started beating it into shape.
But do we really want a pass just to help coremark?
And that's th
Hi all
we have merged the gcc-4_8-branch into linaro/gcc-4_8-branch up to
revision 213802 as r213944. We have also backported this set of revisions:
r204251 as r213841 PR sanitizer/58543
r206529 as r213842 PR target/59744
r206530 as r213842 PR target/59744 / fix changelog typo
Th
On 08/14/2014 05:50 PM, Jakub Jelinek wrote:
I hope the scheduler doesn't attempt to swap sp += 24 with flags setter
because of the sp += 16 vs. flags setter dependency and sp += 24 vs. sp +=
16 dependency, but I feel kind of uneasy with find_inc assuming the recorded
dependency is the one for th
On Thu, Aug 14, 2014 at 11:49:37AM +0200, Jakub Jelinek wrote:
> On Thu, Aug 14, 2014 at 11:34:04AM +0200, Jakub Jelinek wrote:
> > So, to set DEP_MULTIPLE even in the case where ask_depencency_caches
> > returns DEP_PRESENT, you'd need to find the old dependency anyway (isn't
> > that going to be
It does not help Silvermont, i.e. only Haswell and SandyBridge are affected.
I don't use splitter since (1) it deletes zeroing of dest reg; (2)
scheduler can hoist them up . I will try r16/r32 variants and tell you
later.
2014-08-14 19:18 GMT+04:00 H.J. Lu :
> On Thu, Aug 14, 2014 at 4:50 AM, Yuri
Bin.Cheng wrote:
> >> The overflow check can be improved by using deeper inspection to prove the
> >> equality. This patch deals with that by making below two improvements:
> >> a) Handles constant cases.
> >> b) Uses affine expansion as deeper inspection to check the equality.
Looks good to
On 08/14/2014 04:31 AM, Bin Cheng wrote:
g++.dg/ext/arm-fp16/fp16-mangle-1.C is failed because GCC now sets
DECL_COMDAT on template instantiations if flag_implicit_templates is in
effect. Then DECL_WEAK will be set accordingly. As a result, checking for
".global " would fail on this case.
On Thu, Aug 14, 2014 at 4:50 AM, Yuri Rumyantsev wrote:
> Hi All,
>
> Here is a fix for PR 62011 - remove false dependency for unary
> bit-manipulation instructions for latest BigCore chips (Sandybridge
> and Haswell) by outputting in assembly file zeroing destination
> register before bmi instruc
On Thu, Aug 14, 2014 at 4:06 PM, Ilya Enkovich wrote:
> 2014-08-14 18:00 GMT+04:00 Alexander Monakov :
>>
>>
>> On Thu, 14 Aug 2014, Yuri Rumyantsev wrote:
>>
>>> Hi All,
>>>
>>> Here is a fix for PR 62011 - remove false dependency for unary
>>> bit-manipulation instructions for latest BigCore chi
On Thu, Aug 14, 2014 at 1:08 AM, Evgeny Stupachenko wrote:
> Ping.
>
> On Thu, Jul 10, 2014 at 7:29 PM, Evgeny Stupachenko
> wrote:
>> On Mon, Jul 7, 2014 at 6:40 PM, Richard Henderson wrote:
>>> On 07/03/2014 02:53 AM, Evgeny Stupachenko wrote:
-expand_vec_perm_palignr (struct expand_vec_
For example, for the first loop for attached test-case we do not
prepend xor to popcnt because of using destination register:
.L23:
leal 1(%rdx), %ecx
popcntq (%rbx,%rax,8), %rax
leal 2(%rdx), %r8d
popcntq (%rbx,%rcx,8), %rcx
addq %rax, %rcx
leal 3(%rdx), %esi
xorq %rax, %rax
popcntq (%rbx,%r8,8),
This fixes wrong answer from data-dependence analysis by realizing
that _all_ (even non-evolving) indirect accesses cannot be constrained
to a full object size. This also gets rid of that ugly
DR_UNCONSTRAINED_BASE hack (but effectively make it always active).
Bootstrapped on x86_64-unknown-linu
This hides a part of the internal strings processing detail by
using appropriate APIs.
LTO bootstrap ongoing on x86_64-unknown-linux-gnu.
Richard.
2014-08-14 Richard Biener
* data-streamer.h (streamer_string_index, string_for_index):
Remove.
* data-streamer-out.c (st
On Sat, Aug 9, 2014 at 7:14 AM, Tom de Vries wrote:
> On 08-08-14 17:17, Tom de Vries wrote:
>>>
>>> Maybe instead of a new mem_alias_equal_p simply compare MEM_ATTRs
>>> with mem_attrs_eq_p?
>>
>>
>> I propose to fix it this way (as attached) on 4.8/4.9/trunk, and maybe do
>> a more
>> efficient
On Mon, Jul 28, 2014 at 8:08 AM, Wei Mi wrote:
>> But fact is that it is _not_ necessary to split the block because there
>> are no outgoing abnormal edges from it.
>>
>> The verifier failure is an artifact from using the same predicates during
>> CFG building and CFG verifying (usually ok, but fo
On Thu, Aug 14, 2014 at 2:20 PM, Kirill Yukhin wrote:
> Hello,
> This patch extends vcvtudq2ps to support
> AVX-512VL new insns.
>
> Bootstrapped.
> New tests on top of patch-set all pass
> under simulator.
>
> Is it ok for trunk?
>
> gcc/
> * config/i386/sse.md
> (define_mode_iter
On 08/14/2014 07:14 AM, Paolo Carlini wrote:
+ nparms -= variadic_p ? variadic_p : default_p;
What if you have both default arguments and parameter packs?
Jason
On Thu, Aug 14, 2014 at 2:12 PM, Kirill Yukhin wrote:
> Hello,
> This patch introduces vcvt[t]pd2 patterns.
>
> Bootstrapped.
> New tests on top of patch-set all pass
> under simulator.
>
> Is it ok for trunk?
>
> gcc/
> * config/i386/i386.md
> (define_code_attr ufix_bool): New.
>
2014-08-14 18:00 GMT+04:00 Alexander Monakov :
>
>
> On Thu, 14 Aug 2014, Yuri Rumyantsev wrote:
>
>> Hi All,
>>
>> Here is a fix for PR 62011 - remove false dependency for unary
>> bit-manipulation instructions for latest BigCore chips (Sandybridge
>> and Haswell) by outputting in assembly file ze
On Thu, 14 Aug 2014, Yuri Rumyantsev wrote:
> Hi All,
>
> Here is a fix for PR 62011 - remove false dependency for unary
> bit-manipulation instructions for latest BigCore chips (Sandybridge
> and Haswell) by outputting in assembly file zeroing destination
> register before bmi instruction. I c
On Thu, Aug 14, 2014 at 1:42 PM, Kirill Yukhin wrote:
> Hello,
> This patch extends moves to OI/TI mode.
>
> Bootstrapped.
> New tests on top of patch-set all pass
> under simulator.
>
> Is it ok for trunk?
>
> gcc/
> * config/i386/i386.md
> (define_insn "*movoi_internal_avx"): Add
On 08/07/2014 12:52 PM, Marat Zakirov wrote:
On 07/31/2014 04:08 PM, Marat Zakirov wrote:
On 07/24/2014 07:40 PM, Marat Zakirov wrote:
On 07/24/2014 04:27 PM, Marat Zakirov wrote:
On 07/23/2014 06:23 PM, Marat Zakirov wrote:
Hi there!
I made a patch which fixes regressions on ARM platfo
On Thu, Aug 14, 2014 at 03:09:26PM +0200, Richard Biener wrote:
>
> See $subject.
>
> Ok?
>
> Thanks,
> Richard.
>
> 2014-08-14 Richard Biener
>
> maintainer-scripts/
> * crontab: Change trunk snapshots from 4.10 to 5.
>
> gcc/
> * BASE-VER: Change to 5.0.0
LGTM.
> >I've observed SPEC2006 failure on avx512-vlbwdq branch.
> >It was caused by hardreg_cprop. In maybe_mode_change it was
> >assumed, that all values of the same register class and same mode.
> >are ok. This is not the case for i386/avx512. We need to honor
> >HARD_REGNO_MODE_OK.
> One could argu
See $subject.
Ok?
Thanks,
Richard.
2014-08-14 Richard Biener
maintainer-scripts/
* crontab: Change trunk snapshots from 4.10 to 5.
gcc/
* BASE-VER: Change to 5.0.0
Index: maintainer-scripts/crontab
===
On 14 Aug 13:45, Uros Bizjak wrote:
> Please update the above entry.
Whoops. Updated ChangeLog:
gcc/
* config/i386/i386.c
(ix86_expand_special_args_builtin): Handle avx512vl_storev8sf_mask,
avx512vl_storev8si_mask, avx512vl_storev4df_mask,
avx512vl_storev4di_mask,
a
Adding community.
-- Forwarded message --
From: Kirill Yukhin
Date: Thu, Aug 14, 2014 at 4:16 PM
Subject: [PATCH i386 AVX512] [14/n] Add convert to PS insn patterns.
To: Uros Bizjak
Hello,
This patch introduces new patterns for conversions
to PS.
Bootstrapped.
New tests on to
Hello,
This patch extends vcvtudq2ps to support
AVX-512VL new insns.
Bootstrapped.
New tests on top of patch-set all pass
under simulator.
Is it ok for trunk?
gcc/
* config/i386/sse.md
(define_mode_iterator VF1_AVX512VL): New.
(define_insn "ufloatv16siv16sf2"): Delete.
Hello,
This patch introduces vcvt[t]pd2 patterns.
Bootstrapped.
New tests on top of patch-set all pass
under simulator.
Is it ok for trunk?
gcc/
* config/i386/i386.md
(define_code_attr ufix_bool): New.
* config/i386/sse.md
(define_mode_iterator VF2_AVX512VL): New.
On Thu, Aug 14, 2014 at 1:36 PM, Kirill Yukhin wrote:
> This patch extends "vec_extract_hi_"
> pattern to support AVX-512DQ insn.
>
> Bootstrapped.
>
> Bootstrapped.
> New tests on top of patch-set all pass
> under simulator.
>
> Is it ok for trunk?
>
> gcc/
> * config/i386/i386.md
>
Hello Uroš,
On 14 Aug 13:35, Uros Bizjak wrote:
> OK (although I'd put single-use attributes - blendbits and probably
> new dbpsadbwmode - nearby their users).
Thanks! I'll move `blendbits' to its user.
As far as currently no user for `dbpsadbwmode' exists in main trunk - I'll
move it to the user i
The following fixes missing dominator computation before fixing loops.
Rather than doing even more such weird stuff in a pass gate function
this puts this into a new pass scheduled before the loop passes gate.
Bootstrap and regtest running on x86_64-unknown-linux-gnu.
Richard.
2014-08-14 Richa
Hi All,
Here is a fix for PR 62011 - remove false dependency for unary
bit-manipulation instructions for latest BigCore chips (Sandybridge
and Haswell) by outputting in assembly file zeroing destination
register before bmi instruction. I checked that performance restored
for popcnt, lzcnt and tzcn
A static checker determined that LTO def_fn_type has erratic
va_start/end combinations on error paths. The following fixes
that.
Bootstrap and regtest running on x86_64-unknown-linux-gnu.
Richard.
2014-08-14 Richard Biener
* lto-lang.c (def_fn_type): Fix error handling wrt va_end.
On Thu, Aug 14, 2014 at 1:30 PM, Kirill Yukhin wrote:
> Hello,
> This patch extends load/store insns.
> No built-ins added in this patch.
>
> Bootstrapped.
> New tests on top of patch-set all pass
> under simulator.
>
> Is it ok for trunk?
>
> gcc/
> * config/i386/i386.c
> (ix86_ex
Hello,
This patch extends moves to OI/TI mode.
Bootstrapped.
New tests on top of patch-set all pass
under simulator.
Is it ok for trunk?
gcc/
* config/i386/i386.md
(define_insn "*movoi_internal_avx"): Add EVEX version.
(define_insn "*movti_internal"): Ditto.
--
Thanks, K
Hello,
This patch extends "vec_extract_hi_"
pattern to support AVX-512DQ insn.
Bootstrapped.
Bootstrapped.
New tests on top of patch-set all pass
under simulator.
Is it ok for trunk?
gcc/
* config/i386/i386.md
(define_attr "isa"): Add avx512dq, noavx512dq.
(define_attr "
On Thu, Aug 14, 2014 at 1:20 PM, Kirill Yukhin wrote:
> This patch extends iterators and iterator modes
> to support new patterns (future patches).
>
> Bootstrapped.
>
> Is it ok trunk?
>
> * config/i386/sse.md (define_mode_attr avx512): New.
> (define_mode_attr sse2_avx_avx512f):
Hello,
This patch extends load/store insns.
No built-ins added in this patch.
Bootstrapped.
New tests on top of patch-set all pass
under simulator.
Is it ok for trunk?
gcc/
* config/i386/i386.c
(ix86_expand_special_args_builtin): Handle avx512vl_storev8sf_mask,
avx512vl_s
Hello,
This patch extends iterators and iterator modes
to support new patterns (future patches).
Bootstrapped.
Is it ok trunk?
* config/i386/sse.md (define_mode_attr avx512): New.
(define_mode_attr sse2_avx_avx512f): Allow V8HI, V16HI, V32HI, V2DI,
V4DI modes.
(de
Hi!
On Wed, 13 Aug 2014 22:41:47 +0200, Tobias Burnus wrote:
> Cesar Philippidis wrote:
> > According to section 2.6.1 in the openacc spec, fortran loop variables
> > should be implicitly private like in openmp. This patch does just so.
>
> Makes sense. Looking at the patch, I wonder whether the
Hi,
this is a diagnostic issue about the error message for a wrong number of
template arguments vs default arguments. The fix is simple but also
replacing the second error with a default, as recommended by Manuel in
the audit trail, means that we have to adjust a few existing testcases.
Note,
This speeds up type_hash_canon by avoiding a 2nd hashtable lookup
in the case no previous same type is in the hashtable.
Bootstrapped and tested on x86_64-unknown-linux-gnu, applied.
Richard.
2014-08-14 Richard Biener
* tree.c (type_hash_lookup, type_hash_add): Merge into ...
1 - 100 of 118 matches
Mail list logo
