On Mon, Apr 1, 2019 at 7:28 PM David Edelsohn wrote:
>
> On Mon, Apr 1, 2019 at 10:12 PM Ian Lance Taylor wrote:
> >
> > On Mon, Apr 1, 2019 at 7:06 PM David Edelsohn wrote:
> > >
> > > On Mon, Apr 1, 2019 at 9:50 PM Ian Lance Taylor wrote:
> > > >
> > > > On Mon, Apr 1, 2019 at 12:43 AM CHIGOT
Hello,
I was reading through some things and found this was missing.
As before, I don't seem to have CVS access, if its OK and someone can commit it
would be helpful.
-Stafford
Index: htdocs/gcc-9/changes.html
===
RCS file: /cvs/gc
On Tue, Apr 2, 2019 at 10:05 AM Richard Henderson wrote:
>
> On 4/1/19 8:53 PM, Sudakshina Das wrote:
> >> This could stand to use a comment, a moment's thinking about the sizes,
> >> and to
> >> use the existing asm output functions.
> >>
> >> /* PT_NOTE header: namesz, descsz, type.
> >>
On Mon, Apr 1, 2019 at 10:12 PM Ian Lance Taylor wrote:
>
> On Mon, Apr 1, 2019 at 7:06 PM David Edelsohn wrote:
> >
> > On Mon, Apr 1, 2019 at 9:50 PM Ian Lance Taylor wrote:
> > >
> > > On Mon, Apr 1, 2019 at 12:43 AM CHIGOT, CLEMENT
> > > wrote:
> > > >
> > > > Description:
> > > > * This
On Mon, Apr 1, 2019 at 7:06 PM David Edelsohn wrote:
>
> On Mon, Apr 1, 2019 at 9:50 PM Ian Lance Taylor wrote:
> >
> > On Mon, Apr 1, 2019 at 12:43 AM CHIGOT, CLEMENT
> > wrote:
> > >
> > > Description:
> > > * This patch removes -mvsx and -maltivec for go aix/ppc.
> > > These options d
On Mon, Apr 1, 2019 at 9:50 PM Ian Lance Taylor wrote:
>
> On Mon, Apr 1, 2019 at 12:43 AM CHIGOT, CLEMENT
> wrote:
> >
> > Description:
> > * This patch removes -mvsx and -maltivec for go aix/ppc.
> > These options don't seem compatible with Go stack layout.
>
> Can you give some more de
On 4/1/19 8:53 PM, Sudakshina Das wrote:
>> This could stand to use a comment, a moment's thinking about the sizes, and
>> to
>> use the existing asm output functions.
>>
>> /* PT_NOTE header: namesz, descsz, type.
>> namesz = 4 ("GNU\0")
>> descsz = 12 (see below)
> I was try
On Mon, Apr 1, 2019 at 12:43 AM CHIGOT, CLEMENT wrote:
>
> Description:
> * This patch removes -mvsx and -maltivec for go aix/ppc.
> These options don't seem compatible with Go stack layout.
Can you give some more details on the problem? gccgo should just be
using the regular stack layout
From: Andrew Pinski
Hi,
The problem here is the token->val.node is not saved over
a precompiled header for C++ operator. This can cause an
internal compiler error as we tried to print out the spelling
of the token as we assumed it was valid.
The fix is to have cpp_token_val_index return CPP_TO
>> We shouldn't include in the std::lib, the uses of assert
>> should be changed to __glibcxx_assert instead (which is enabled by
>> defining _GLIBCXX_ASSERTIONS).
>>
>
> This has to come through one of the PSTL library configuration
> macros because the "right assert" upstream won't be __glibcxx_
On Sat, 30 Mar 2019, Iain Buclaw wrote:
> Patch updates the EXCLUDES list for message strings that are now
> picked up from d/dmd after additions/removals that have happened since
> adding sources. Regenerated gcc.pot as per documentation at
> https://gcc.gnu.org/translation.html.
>
> OK for tru
On Mon, Apr 1, 2019 at 3:04 PM Jason Merrill wrote:
>
> DR 1321 clarified that two dependent names are equivalent if the names are
> the same, even if the result of name lookup is different. So template
> argument hashing should treat a lookup set like a plain identifier.
> Mangling already does.
DR 1321 clarified that two dependent names are equivalent if the names are
the same, even if the result of name lookup is different. So template
argument hashing should treat a lookup set like a plain identifier.
Mangling already does.
Tested x86_64-pc-linux-gnu, applying to trunk.
* pt.
On Mon, Apr 1, 2019 at 7:39 PM H.J. Lu wrote:
>
> check_avx2_available should check avx2_available, instead of avx_available.
> Otherwise, check_avx2_available may use result from check_avx_available.
>
> PR testsuite/89907
> * lib/target-supports.exp (check_avx2_available): Replac
Hi,
this patch by Andrey moves an assignment to reset_sched_cycles_p
after rev. 259228 placed it incorrectly.
2019-04-01 Andrey Belevantsev
PR rtl-optimization/85412
* sel-sched.c (sel_sched_region): Assign reset_sched_cycles_p before
sel_sched_region_1, not after.
check_avx2_available should check avx2_available, instead of avx_available.
Otherwise, check_avx2_available may use result from check_avx_available.
PR testsuite/89907
* lib/target-supports.exp (check_avx2_available): Replace
avx_available with avx2_available.
---
gcc/test
On 3/31/19 10:17 PM, Martin Sebor wrote:
To fix PR 89833, a P1 regression, the attached patch tries to
handle string literals as C++ 2a non-type template arguments
by treating them the same as brace enclosed initializer lists
(where the latter are handled correctly). The solution makes
sure equi
Richard Biener writes:
> On Tue, Mar 26, 2019 at 1:56 AM Richard Sandiford
> wrote:
>> Based on the "complete unrolling" view, if we number statements as
>> (i, n), where i is the outer loop iteration and n is a statement number
>> in the (completely unrolled) loop body, then the original scalar
This is a ping**3 for a patch to fix one of the test failures in PR 877763.
It fixes the gcc.target/aarch64/combine_bfi_1.c failure, but not the other
ones.
Could one of the Aarch64 maintainers take a look at it? This version of
the patch was originally submitted on February 11 after incorporatin
Richard Biener writes:
> This is an update from last years attempt to tame down vectorization
> when it runs in to ABI inefficiencies at function return. I've
> updated the patch to look for multi-reg returns instead of
> !vector ones (due to word_mode vectorization) and handle a bit
> more retur
Hi,
This patch by Andrey fixes PR 86928 by avoiding leaving basic blocks without
computed liveness info.
2019-04-01 Andrey Belevantsev
PR rtl-optimization/86928
* sel-sched-ir.c (sel_redirect_edge_and_branch_force): Invoke
compute_live if necessary.
(sel_redire
The following patch is for
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=89865
The patch was successfully tested on x86-64, aarch64, and ppc64.
The patch decreases scan matches from 82 to 8 on x86-64. I don't know
what test pr49095.c should expect. Therefore I don't change the test
itself.
Jonathan Wakely writes:
> On 29/03/19 12:12 -0700, Thomas Rodgers wrote:
>>Prevent ADL related weirdness.
>>
>> * include/pstl/algorithm_impl.h: Add namespace qualification.
>> * include/pstl/execution_defs.h: Add namespace qualification.
>> * include/pstl/execution_impl.h: Add na
Hi!
This is a fairly trivial change fixing a false negative in
-Wmaybe-uninitialized. I am pretty sure this is simply an overlooked
case (is_value_included_in() is not meant to deal with the case where
both compare codes are NE_EXPRs, neither does it need to, since their
handling is trivial).
In
On 4/1/19 12:03 AM, Yu, Mingli wrote:
> Ping.
>
> Thanks,
>
> On 2019年03月19日 16:37, mingli...@windriver.com wrote:
>> From: Mingli Yu
>>
>> Initialize st_mode member to fix the below
>> build failure when -Og included in compiler flag.
>> |
>> ./../../../../../../../../work-shared/gcc-8.3.0-r0/g
I've committed this single-liner (and added a comment) to OG8. This allows
detection of
Compute Capability 7.0 (Volta) and lets libgomp allocate a more reasonable
default gang
number for Volta GPUs.
Tested without regressions on a powerpc64le-linux system. Note that mainline has
different code
On Mon, Apr 01, 2019 at 10:15:11AM +0200, Andreas Schwab wrote:
> On Mär 28 2019, Marek Polacek wrote:
>
> > Andreas, could you please find out why we're not hitting this code in
> > digest_init_r:
> >
> > 1210 tree elt = CONSTRUCTOR_ELT (stripped_init, 0)->value;
> > 1211 if (referen
Hi,
I'm applying this patch on Andrey's behalf to resolve PR 87273 where an
overly strict assert fires; Andrey explanation in PR comment #5 provides
more detail.
Alexander
2019-04-01 Andrey Belevantsev
PR rtl-optimization/87273
* sel-sched-ir.c (merge_fences): Remove assert.
Hi James
On 29/03/2019 13:41, Sudakshina Das wrote:
> Hi James
>
> On 22/03/2019 16:25, James Greenhalgh wrote:
>> On Wed, Mar 20, 2019 at 10:17:41AM +, Sudakshina Das wrote:
>>> Hi Kyrill
>>>
>>> On 12/03/2019 12:03, Kyrill Tkachov wrote:
Hi Sudi,
On 2/22/19 10:45 AM, Sudakshi
Hi,
This patch merges the libphobos druntime library with druntime
upstream d57fa1ff.
Fixes alignment of internal core.thread locks that were found to be
the cause of abort() being called inside pthread_mutex_init(), fixing
the second part of PR d/88462.
Bootstrapped and regression tested on x86
On April 1, 2019 4:13:28 PM GMT+02:00, Richard Sandiford
wrote:
>Richard Biener writes:
>> After the fix to RPO VN in loop header copying DF via RTL invariant
>> motion takes 50% of the compile-time for the second testcase in
>> PR46590. This is caused by the DF live problem iterating over
>> a
Richard Biener writes:
> After the fix to RPO VN in loop header copying DF via RTL invariant
> motion takes 50% of the compile-time for the second testcase in
> PR46590. This is caused by the DF live problem iterating over
> all dirty blocks for the local problem which is all blocks of
> the func
Hi Richard
Thanks for the comments and pointing out the much cleaner existing asm
output functions!
On 29/03/2019 17:51, Richard Henderson wrote:
>> +#define ASM_LONG "\t.long\t"
>
> Do not replicate targetm.asm_out.aligned_op.si, or integer_asm_op, really.
>
>> +aarch64_file_end_indicate_exec
OK.
Jason
On Mon, Apr 1, 2019 at 5:10 AM bin.cheng wrote:
>
> Hi,
>
> As described in comments of PR89725, this patch fixes out-of-bound access
> during
> computing classic dist/dir vector for DDR. Basically it does two things: A)
> Handle
> relevant chrec of outer loop in multivariate access function as
On Mon, Apr 1, 2019 at 3:43 AM CHIGOT, CLEMENT wrote:
>
> Description:
> * This patch removes -mvsx and -maltivec for go aix/ppc.
> These options don't seem compatible with Go stack layout.
>
> Tests:
> * AIX 7.2: Configure/Build: SUCCESS
>
> Changelog:
> * go/gospec.c (lang_specific_d
On 29/03/19 12:12 -0700, Thomas Rodgers wrote:
Prevent ADL related weirdness.
* include/pstl/algorithm_impl.h: Add namespace qualification.
* include/pstl/execution_defs.h: Add namespace qualification.
* include/pstl/execution_impl.h: Add namespace qualification.
Currently confused by both all or all.c in the local directory.
Bootstrap / regtest running on x86_64-unknown-linux-gnu.
Richard.
2019-04-01 Richard Biener
PR lto/89896
* lto-wrapper.c (run_gcc): Avoid implicit rules making
the all target phony.
Index: gcc/lto-wrap
Hi.
Last week I was curious which warnings are disabled by default on top
of -Wall and -Wextra. Thus I used --help=warning and noticed some discrepancy
in between documentation and output of the --help option.
I created PR89885 where I explained that OPT__help_ option handling happens
early. That
On 01/04/19 12:52 +0300, Ville Voutilainen wrote:
On Mon, 1 Apr 2019 at 11:45, Ville Voutilainen
wrote:
On Sat, 30 Mar 2019 at 20:00, Ville Voutilainen
wrote:
>
> This patch makes assignments correct, because they need to
> match indices instead of types. In addition, we cut down the
> codege
On 01/04/19 11:45 +0300, Ville Voutilainen wrote:
@@ -570,45 +574,44 @@ namespace __variant
operator=(const _Copy_assign_base& __rhs)
noexcept(_Traits<_Types...>::_S_nothrow_copy_assign)
{
- __do_visit([this, &__rhs](auto&& __this_mem, auto&& __rhs_mem) mutable
-
After the fix to RPO VN in loop header copying DF via RTL invariant
motion takes 50% of the compile-time for the second testcase in
PR46590. This is caused by the DF live problem iterating over
all dirty blocks for the local problem which is all blocks of
the function because while loop-invarian
On Mon, 1 Apr 2019 at 11:45, Ville Voutilainen
wrote:
>
> On Sat, 30 Mar 2019 at 20:00, Ville Voutilainen
> wrote:
> >
> > This patch makes assignments correct, because they need to
> > match indices instead of types. In addition, we cut down the
> > codegen size. The symbols are longer than befo
Hi,
A rather long standing issue. The error-recovery problem is that, in
various circumstances, we flag as erroneous a local variable declaration
and then, when we need to come back to it later on, lookup_name doesn't
find it, keeps looking and may find instead a completely unrelated
non-vari
This fixes a quadraticness in region RPO VN which happens because
dom_walker computes whole-function RPO order in its constructor
even though in the end we do not use it. That's an artifact of
C++ (we derive from dom_walker).
The fix is to lazily do costly initialization in the ::walk
method.
On Mon, 1 Apr 2019 at 11:50, Ville Voutilainen
wrote:
> in it makes it trivial and I don't
> need to worry about the validity of the branch taken, since it's not
> instantiated. This is a new
I mean the branch *not* taken. :)
On Mon, 1 Apr 2019 at 11:45, Paolo Carlini wrote:
> >>> +__visitor_result_type(_Visitor&& __visitor, _Variants&&...
> >>> __variants)
> >>> +{
> >>> + if constexpr(__use_index)
> >>> +return __detail::__variant::__variant_idx_cookie{};
> >>> + else
> >>> + return std
On Sat, 30 Mar 2019 at 20:00, Ville Voutilainen
wrote:
>
> This patch makes assignments correct, because they need to
> match indices instead of types. In addition, we cut down the
> codegen size. The symbols are longer than before, the the amount
> of them is roughly the same, so there's no longe
Hi
On 01/04/19 10:43, Ville Voutilainen wrote:
On Mon, 1 Apr 2019 at 11:30, Paolo Carlini wrote:
Hi
On 30/03/19 19:00, Ville Voutilainen wrote:
- template
+ template
+decltype(auto)
+__visitor_result_type(_Visitor&& __visitor, _Variants&&... __variants)
+{
+ if constexpr(_
On Mon, 1 Apr 2019 at 11:30, Paolo Carlini wrote:
>
> Hi
>
> On 30/03/19 19:00, Ville Voutilainen wrote:
> > - template
> > + template
> > +decltype(auto)
> > +__visitor_result_type(_Visitor&& __visitor, _Variants&&... __variants)
> > +{
> > + if constexpr(__use_index)
> > +
Hi
On 30/03/19 19:00, Ville Voutilainen wrote:
- template
+ template
+decltype(auto)
+__visitor_result_type(_Visitor&& __visitor, _Variants&&... __variants)
+{
+ if constexpr(__use_index)
+return __detail::__variant::__variant_idx_cookie{};
+ else
+ return s
On Mär 28 2019, Marek Polacek wrote:
> Andreas, could you please find out why we're not hitting this code in
> digest_init_r:
>
> 1210 tree elt = CONSTRUCTOR_ELT (stripped_init, 0)->value;
> 1211 if (reference_related_p (type, TREE_TYPE (elt)))
This is never executed if flag_checking
Description:
* This patches disables go tests with -flto on AIX.
Tests:
* AIX 7.2: Configure/Build: SUCCESS
Changelog:
* lib/go-torture.exp: add check for lto option
Clément Chigot
ATOS Bull SAS
1 rue de Provence - 38432 Échirolles - France Index: go-torture.exp
=
Description:
* This patch removes -mvsx and -maltivec for go aix/ppc.
These options don't seem compatible with Go stack layout.
Tests:
* AIX 7.2: Configure/Build: SUCCESS
Changelog:
* go/gospec.c (lang_specific_driver): disable -mvsx and -maltivec for
aix/ppc.
Clément Chigot
54 matches
Mail list logo
