Richard,
I've benchmarked your patch on Skylake with SPEC CPU 20[06|17][fp|int]rate
and another smaller benchmark suites. I found that it doesn't regress
any benchmark off-noise but improves 525.x264 by 1.8%, 526.blender by 1.9% and
465.tonto by 3.2%.
I think this is a good reason to merge the p
Hi!
I'd like to ping this patch from Steve.
On Fri, Jan 05, 2018 at 02:14:26PM -0800, Steve Ellcey wrote:
> This is a fix for PR target/83335. We are asserting in
> aarch64_print_address_internal because we have a non Pmode
> address coming from an asm instruction. My fix is to
> just allow th
Ping!
On Friday 09 February 2018 01:02 PM, Siddhesh Poyarekar wrote:
> Hi,
>
> Here's v3 of the patch to disable register offset addressing mode for
> stores of 128-bit values on Falkor because they're very costly.
> Following Kyrill's suggestion, I compared the codegen for a53 and
> found that t
On Wed, 14 Feb 2018, Martin Sebor wrote:
> I was excited to find out about the recent enhancement to
> the LTO test harness to support the new dg-lto-warning and
> dg-lto-message directives (thanks, David).
>
> To make them easier to find and use (there is a C++ LTO test
> that uses them but no C
On Thu, 15 Feb 2018, Jakub Jelinek wrote:
> Hi!
>
> The function calls get_addr_base_and_unit_offset on 2 trees, but
> that can return NULL if the unit offset is not constant.
> The conditional tests just one of them for non-NULL and operand_equal_p
> ICEs if one argument is NULL, so depending on
On Wed, 14 Feb 2018, Jakub Jelinek wrote:
> On Tue, Feb 13, 2018 at 07:04:09PM +0100, Richard Biener wrote:
> > On February 13, 2018 6:51:29 PM GMT+01:00, Jakub Jelinek
> > wrote:
> > >On the following testcase, we recurse infinitely, because
> > >we have float re-association enabled, but also r
On 14/02/18 11:41, Jakub Jelinek wrote:
Hi!
As mentioned in detail in the PR, PowerPC64 ELFv1 function symbols
point to function descriptors in .opd section rather than actual
code, and one needs to read the code address from the .opd section
in order to associate symbols with .text addresses.
On Thu, Feb 15, 2018 at 01:12:08AM -0500, Jason Merrill wrote:
> > This is just a partial fix, one thing this patch doesn't change is that
> > the standard says that __VA_OPT__ ( contents ) should be treated as
> > parameter, which means that #__VA_OPT__ ( contents ) should stringify it,
> > which
On 05/01/18 22:14, Steve Ellcey wrote:
> This is a fix for PR target/83335. We are asserting in
> aarch64_print_address_internal because we have a non Pmode
> address coming from an asm instruction. My fix is to
> just allow this by checking this_is_asm_operands. This is
> what it was doing bef
On Wed, 2018-02-14 at 01:06 +0900, Oleg Endo wrote:
>
> Do you happen to have any other numbers on the resulting code
> size/speed? Looking at the new costs that the patch introduces, I'd
> expect there to be some more changes than just the 1/x...
>
I've checked your proposed patch with the CSi
Hi,
this patch requires effective target indirect_jumps for test-case pr84136.c.
Committed.
Thanks,
- Tom
[testsuite] Require et indirect_jumps for pr84136.c
2018-02-15 Tom de Vries
* gcc.c-torture/compile/pr84136.c: Require effective target
indirect_jumps.
---
gcc/testsuite/gcc.c-tort
Hi,
this patch requires effective target alloca for test-case
Wstringop-overflow-3.c.
Committed.
Thanks,
- Tom
[testsuite] Require et alloca for Wstringop-overflow-3.c
2018-02-15 Tom de Vries
* gcc.dg/Wstringop-overflow-3.c: Require effective target alloca.
---
gcc/testsuite/gcc.dg/Ws
Hi Siddhesh,
I still don't like the idea of disabling a whole class of instructions in the
md file.
It seems much better to adjust the costs here so that you get most of the
improvement now, and fine tune it once we can differentiate between
loads and stores.
Taking your example, adding -funroll
Hi,
we have been accumulating quite a few bugs under the [concepts]
meta-bug, most of which of course aren't regressions. This one is a low
hanging fruit, an error recovery issue where, after some meaningful
diagnostic, tsubst_constraint doesn't know how to handle an
error_mark_node. I believ
On Wed, Feb 14, 2018 at 6:10 PM, Jason Merrill wrote:
> This testcase involves a fastcall-qualified function type. During
> mangling, we use build_type_attribute_qual_variant to look up an
> attribute-unqualified version of that type.
> build_type_attribute_qual_variant calls type_hash_canon and
On 13/02/18 13:43, Alexandre Oliva wrote:
On Feb 12, 2018, Alexandre Oliva wrote:
This patch supersedes the previous one. Testing underway... Ok if it
succeeds?
I failed to update the patch I posted after making a correct to symbol
poisoning, that had caused builds to fail right away, sorr
Igor
> -Original Message-
> From: Joseph Myers [mailto:jos...@codesourcery.com]
> Sent: Thursday, February 15, 2018 1:24 AM
> To: Tsimbalist, Igor V
> Cc: Sandra Loosemore ; gcc-
> patc...@gcc.gnu.org; Uros Bizjak
> Subject: RE: PR84239, Reimplement CET intrinsics for rdssp/incssp ins
On Thu, Feb 15, 2018 at 4:27 AM, Jakub Jelinek wrote:
> On Thu, Feb 15, 2018 at 01:12:08AM -0500, Jason Merrill wrote:
>> > This is just a partial fix, one thing this patch doesn't change is that
>> > the standard says that __VA_OPT__ ( contents ) should be treated as
>> > parameter, which means t
On 02/13/2018 11:14 PM, Jeff Law wrote:
On 02/01/2018 04:45 PM, Martin Sebor wrote:
The previous patch didn't resolve all the false positives
in the Linux kernel. The attached is an update that fixes
the remaining one having to do with multidimensional array
members:
struct S { char a[2][4];
OK.
On Thu, Feb 15, 2018 at 9:29 AM, Paolo Carlini wrote:
> Hi,
>
> we have been accumulating quite a few bugs under the [concepts] meta-bug,
> most of which of course aren't regressions. This one is a low hanging fruit,
> an error recovery issue where, after some meaningful diagnostic,
> tsubst_
Thanks. I've seen these fixes before but I still keep forgetting
to add the directive. It would be nice if the test harness could
detect this automatically for us and remind us to add the directive
even on targets where alloca is supported.
Some sort of a script (or even DejaGnu itself) would d
I started adding non-pack locals to PACK_EXPANSION_EXTRA_ARGS, but was
still only looking up specializations for the packs...
Tested x86_64-pc-linux-gnu, applying to trunk.
commit c61b45ad2d0886dc53dde7ba95382efa29288b2d
Author: Jason Merrill
Date: Thu Feb 15 12:56:47 2018 -0500
PR
I implemented P0512 back in July, but forgot to update the value of
the feature-test macro.
commit 1d440f3731f7350426a019be7bbb636d1db47808
Author: Jason Merrill
Date: Thu Feb 15 13:11:52 2018 -0500
* c-cppbuiltin.c (c_cpp_builtins): Update __cpp_deduction_guides.
diff --git a/gcc/
On 13/02/2018 13:13, Szabolcs Nagy wrote:
> On 07/02/18 18:07, Adhemerval Zanella wrote:
> 5. The TCB support on GLIBC is meant to be included in version 2.28.
>>
> ...
>> +/* -fsplit-stack uses a TCB field available on glibc-2.27. GLIBC also
>> + exports symbol, __tcb_private_ss, to signal i
Hi,
I just committed the test case below as obvious, after testing.
Just to make sure that, when we convert all test cases from
CALL ABORT to "STOP 1" (or whatever), we actually get a failure
when we have a regression.
Regards
Thomas
2018-02-15 Thomas Koenig
PR fortran/8438
On Thu, Feb 8, 2018 at 9:09 PM, Alexandre Oliva wrote:
> + /* If it was supposed to be an rvalue but it's not, adjust
> +one of the operands so that any overload resolution
> +taking this COND_EXPR as an operand makes the correct
> +decisions. See c++/8
OK.
On Fri, Dec 15, 2017 at 4:06 AM, Tom de Vries wrote:
> [ was: Re: [PATCH, PR81844] Fix condition folding in c_parser_omp_for_loop ]
>
> On 09/14/2017 09:38 PM, Jakub Jelinek wrote:
>>
>> On Thu, Sep 14, 2017 at 07:34:14PM +, de Vries, Tom wrote:
>>
>>> --- a/libgomp/testsuite/libgomp.c++/
Hi Thomas,
apparently your commit only includes the ChangeLog addition, but not
the new test case itself. I guess you missed an "svn add"? :)
Cheers,
Janus
2018-02-15 19:47 GMT+01:00 Thomas Koenig :
> Hi,
>
> I just committed the test case below as obvious, after testing.
> Just to make sure th
Here we are asked deduction_guide_p about a TEMPLATE_ID_EXPR; give the
right answer.
Tested x86_64-pc-linux-gnu, applying to trunk.
commit 6ef5c80e09542b55e5dc4e66db87c78adfc0a266
Author: Jason Merrill
Date: Thu Feb 15 13:37:58 2018 -0500
PR c++/84376 - ICE with omitted template ar
Hi Janus,
apparently your commit only includes the ChangeLog addition, but not
the new test case itself. I guess you missed an "svn add"? :)
Actually, I messed up the ChangeLog entry in r257702 and corrected that
one in r257703 :-)
The test case is there.
Regards
Thomas
2018-02-15 20:53 GMT+01:00 Thomas Koenig :
> Hi Janus,
>
>> apparently your commit only includes the ChangeLog addition, but not
>> the new test case itself. I guess you missed an "svn add"? :)
>
>
> Actually, I messed up the ChangeLog entry in r257702 and corrected that
> one in r257703 :-)
>
> Th
Hi all,
attached is another simple patch for an accepts-invalid problem (this
time concerning DTIO), also curing an invalid test case.
Regtests cleanly on x86_64-linux-gnu. Ok for trunk?
Cheers,
Janus
2018-02-15 Janus Weil
PR fortran/84409
* interface.c (check_dtio_arg_TKR_intent)
Hi all,
This is an updated version of patch for PR 81535. The patch was last
discussed in November. The new version addresses Segher's comments in
https://gcc.gnu.org/ml/gcc-patches/2017-08/msg00558.html
The patch was tested on
x86_64-pc-linux-gnu
powerpc64-unknown-linux-gnu
powerpc64le-u
On Thu, Feb 15, 2018 at 09:03:55PM +0100, Janus Weil wrote:
>
> Regtests cleanly on x86_64-linux-gnu. Ok for trunk?
>
Looks good to me with a question below.
> Index: gcc/fortran/interface.c
> ===
> --- gcc/fortran/interface.c (r
On Thu, 2018-02-15 at 10:54 +0100, Richard Biener wrote:
> On Wed, 14 Feb 2018, Martin Sebor wrote:
>
> > I was excited to find out about the recent enhancement to
> > the LTO test harness to support the new dg-lto-warning and
> > dg-lto-message directives (thanks, David).
> >
> > To make them ea
Hi,
On 14 February 2018 at 09:47, Kugan Vivekanandarajah
wrote:
> Hi Kyrill,
>
> On 13 February 2018 at 20:47, Kyrill Tkachov
> wrote:
>> Hi Kugan,
>>
>> On 12/02/18 23:58, Kugan Vivekanandarajah wrote:
>>>
>>> Implements a machine reorg pass for aarch64/Falkor to handle
>>> prefetcher tag coll
2018-02-15 21:16 GMT+01:00 Steve Kargl :
> On Thu, Feb 15, 2018 at 09:03:55PM +0100, Janus Weil wrote:
>>
>> Regtests cleanly on x86_64-linux-gnu. Ok for trunk?
>>
>
> Looks good to me with a question below.
Thanks for the feedback, Steve.
>> Index: gcc/fortran/interface.c
>> ===
On Thu, Feb 15, 2018 at 09:55:58PM +0100, Janus Weil wrote:
> 2018-02-15 21:16 GMT+01:00 Steve Kargl :
> >
> > If the above is a numbered constraint in the Standard,
> > can you add a comment of the form /* F03:C. */
> > above the if-statement.
>
> Well, the if statement that I'm adding is no
On Fri, 2018-02-09 at 13:01 +, Nick Clifton wrote:
> Hi David, Hi Martin,
>
> OK, take 3 of the patch is attached. In this version:
>
> * The escaping is handled by a new class.
> * Self-tests have been added (and they helped find a bug - yay!)
> * The documentation has been extended
This adds the special .NOTPARALLEL target to include/Makefile when
building on darwin, to work around an apparent bug in the APFS
filesystem.
PR libstdc++/81797
* configure.ac (INCLUDE_DIR_NOTPARALLEL): Define.
* configure: Regenerate.
* include/Makefile.am (INCLUDE_DIR_NOTPARALLE
2018-02-15 22:03 GMT+01:00 Steve Kargl :
> On Thu, Feb 15, 2018 at 09:55:58PM +0100, Janus Weil wrote:
>> 2018-02-15 21:16 GMT+01:00 Steve Kargl :
>> >
>> > If the above is a numbered constraint in the Standard,
>> > can you add a comment of the form /* F03:C. */
>> > above the if-statement.
>
On Thu, 15 Feb 2018, Tsimbalist, Igor V wrote:
> I haven't managed to run it through
> ./glibc/glibc.sourceware/scripts/build-many-glibcs.py. I did bootstrap
> and CET tests.
>
> Ok for trunk?
OK.
--
Joseph S. Myers
jos...@codesourcery.com
Ping, this patch was approved for binutils by Cary:
https://sourceware.org/ml/binutils/2017-12/msg00023.html
Is it ok to apply this to GCC include/plugin-api.h now? It is a
pretty small change. Patch attached.
* plugin-api.h (ld_plugin_get_wrap_symbols): New
plugin interface.
Thanks
Sri
On
Hi,
This fixes the scan-assembler errors as seen on power9. While verifying
that fix and ensuring coverage, cleaned up some nearby bits as well.
builtins-3.c: Add/update options and skip-if stanzas, this test specifically
targets vsx/power6.
builtins-3-p9.c: Update stanzas to require -mcpu=po
This bug is a throwback to 50852 and such, where we treated a typedef
variant from dependent scope as equivalent to the underlying type,
which leads to crashes when we later try to instantiate that typedef
outside its scope. To deal with that, I introduced strip_typedefs to
remove such offending t
On Fri, 2018-02-09 at 12:02 +0100, Richard Biener wrote:
> On Thu, Feb 8, 2018 at 11:23 PM, David Malcolm
> wrote:
> > PR tree-optimization/84178 reports a couple of source files that
> > ICE inside
> > ifcvt when compiled with -03 -fno-tree-forwprop (trunk and gcc 7).
> >
> > Both cases involve
Hi Wilko,
Thanks for your comments.
On 14 February 2018 at 00:05, Wilco Dijkstra wrote:
> Hi Kugan,
>
>> Based on the previous discussions, I tried to implement a tree loop
>> unroller for partial unrolling. I would like to queue this RFC patches
>> for next stage1 review.
>
> This is a great pl
Hi,
this issue is very easy to explain and the testcase minimal:
check_special_function_return_type ICEs on a TEMPLATE_TEMPLATE_PARM as
optype, because it tries to use CLASSTYPE_TI_TEMPLATE on it. Today,
triangulating with other compilers too, I came to believe that here
essentially we only h
I wrote:
OK for trunk?
Following a remark by Steve, here is an additional test case which
checks if abort actually works. I'll add that to the patch
when it is committed.
Regards
Thomas
! { dg-do run }
! { dg-shouldfail "Program aborted." }
program main
call abort
end program mai
On 02/14/2018 04:01 PM, Jakub Jelinek wrote:
Hi!
Unlike normal insns where SUBREGs must properly validate, in
DEBUG_INSNs we allow arbitrary SUBREGs, either the dwarf2out code
will be able to use it, or it will just punt. The reason for it is
among other things that during analysis we usually
A failed template deduction in template member of a template
triggers an ICE with -std=c++17 due to what seems like
a missing handling of invalid input. Replacing
the gcc_unreachable() call that causes the ICE with a return
statement indicating the deduction failure eliminates the ICE
and restore
On Fri, Feb 16, 2018 at 12:25:11AM +0100, Thomas Koenig wrote:
> I wrote:
>
> > OK for trunk?
>
> Following a remark by Steve, here is an additional test case which
> checks if abort actually works. I'll add that to the patch
> when it is committed.
>
The patch is ok from my viewpoint. You m
On Mon, 12 Feb 2018 15:18:39 PST (-0800), Jim Wilson wrote:
On 02/12/2018 03:23 AM, Andreas Schwab wrote:
On Feb 06 2017, Palmer Dabbelt wrote:
+/* Because RISC-V only has word-sized atomics, it requries libatomic where
+ others do not. So link libatomic by default, as needed. */
+#undef
The code to avoid doing an extra copy of a list-initialized temporary
was incorrectly being used even when we actually want a
derived-to-base conversion.
Tested x86_64-pc-linux-gnu, applying to trunk.
commit b2413566649273c0b7e172213ebf41c6269c947c
Author: Jason Merrill
Date: Thu Feb 15 17:22:3
On Thursday 15 February 2018 07:50 PM, Wilco Dijkstra wrote:
> So it seems to me using existing cost mechanisms is always preferable, even
> if you
> currently can't differentiate between loads and stores.
Luis is working on address cost adjustments among other things, so I
guess the path of leas
> Oddly, the failing test in associate_35.f90 is the only one that works
> in 7-branch. I have left the PR open and changed the title
> accordingly.
The attached patch fixes this. OK for trunk?
Paul
2018-02-16 Paul Thomas
PR fortran/84115
* resolve.c (resolve_assoc_var): If a non-con
The following restores behavior of GCC 6 for volatile accesses of
automatic vars that do not have their address taken.
Bootstrapped and tested on x86_64-unknown-linux-gnu, applied to trunk
and branch.
Richard.
2018-02-16 Richard Biener
PR tree-optimization/84190
* tree-ssa.c
57 matches
Mail list logo
