Hi Steve,
I have a couple of questions before I have to hurry off to work:
First, why is
@@ -2253,22 +2253,19 @@ gfc_simplify_dim (gfc_expr *x, gfc_expr *y)
gfc_expr*
gfc_simplify_dot_product (gfc_expr *vector_a, gfc_expr *vector_b)
{
+ /* If vector_a is a zero-sized array, the result
On Wed, 24 Jan 2018 21:34:48 +0100
François Dumont wrote:
> On 24/01/2018 18:53, Petr Ovtchenkov wrote:
> > On Wed, 24 Jan 2018 17:39:59 +0100
> > François Dumont wrote:
> >
> >> Hi
> >>
> >> I'd like to propose this new debug check. Comparing
This libbacktrace patch fixes the setting of str_size on PE/COFF to
not leave some bytes uninitialized on a 64-bit host. Committed to
mainline.
Ian
2018-01-24 Ian Lance Taylor
* pecoff.c (coff_add): Use coff_read4, not memcpy.
Index: pecoff.c
PR 68239 points out that libbacktrace can sometimes take a long time
scanning the list of free memory blocks looking for one that is large
enough. Since the libbacktrace memory allocator does not have to be
perfect in practice, only keep the 16 largest entries on the free
list. Bootstrapped and
All,
The attach patch fixes a regression with dot_product and
zero-sized arrays. I bootstrapped and regression tested
the patch on x86_64-*-freebsd. OK to commit?
2018-01-23 Steven G. Kargl
PR fortran/83998
* simplify.c (gfc_simplify_dot_product): Deal
On Wed, Jan 24, 2018 at 05:00:39PM -0500, Michael Meissner wrote:
> Replacing 'int' with 'unsigned long' allows the test to succeed once again. I
> have checked this on a big endian power8 (both 32-bit and 64-bit) and on a
> little endian power8 (64-bit only), and it passes in all three
These testcases jit-compile functions that return char, but
were erroneously calling them as if they returned int.
This led to errors for certain target configurations (e.g.
reading from %eax (32-bit) in the harness when only %al (8-bit)
had been written to in the jit-compiled function).
On Wed, Jan 24, 2018 at 03:19:00PM -0500, Michael Meissner wrote:
> On Wed, Jan 24, 2018 at 12:35:38PM -0600, Segher Boessenkool wrote:
> > Although, hrm, in your patch you also change "int i" to "long i"; that
> > alone seems to be enough to fix everything? Could you check that please?
>
>
This patch to the Go frontend rationalizes the external symbol names
that appear in assembler code. It changes from the ad hoc mechanisms
used to date to produce a set of names that are at least somewhat more
coherent. They are also more readable, after applying a simple
demangling algorithms
Hi!
cxx_eval_outermost_constant_expr clears TREE_CONSTANT on ADDR_EXPRs that
aren't considered by C++ constant expressions, but that breaks middle-end
which relies on TREE_CONSTANT being set on ADDR_EXPR where the address
is constant.
The following patch just special cases ADDR_EXPR not to clear
Hi!
In constexpr evaluation of array references for arrays with unknown bounds,
we need to diagnose out of bounds accesses, but really don't know the bounds
at compile time, right now GCC will see nelts as error_mark_node + 1 and
will not consider them a constant expression at all.
>From the
On 01/24/2018 03:24 PM, Joseph Myers wrote:
> PR target/68467 is libgcc failing to build for m68k-linux-gnu
> configured for ColdFire.
>
> Jeff has an analysis in the PR identifying the problem as resulting
> from the callers of libcalls with 1-byte or 2-byte arguments wanting
> to push just 1 or
PR target/68467 is libgcc failing to build for m68k-linux-gnu
configured for ColdFire.
Jeff has an analysis in the PR identifying the problem as resulting
from the callers of libcalls with 1-byte or 2-byte arguments wanting
to push just 1 or 2 bytes on the stack, while the libcall
implementations
On Wed, Jan 24, 2018 at 08:19:58PM +, Paul Richard Thomas wrote:
> (Jakub, This is all hidden behind the -fcoarray option. To my mind
> this is safe for release.)
Ok from RM POV.
Jakub
On 01/24/2018 12:11 AM, Uros Bizjak wrote:
> On Wed, Jan 24, 2018 at 12:15 AM, Jeff Law wrote:
>>
>> pr83994 is a code generation bug in the stack-clash support that affects
>> openssl (we've turned on stack-clash-protection by default for the F28
>> builds).
>>
>> The
On Wed, Jan 24, 2018 at 12:35:38PM -0600, Segher Boessenkool wrote:
> Hi!
>
> On Wed, Jan 24, 2018 at 12:27:55AM -0500, Michael Meissner wrote:
> >
> > As Segher and I were discussing over private IRC, the root cause of this
> > bug is
> > the compiler no long generates the BDNZ instruction for
Hi All,
Here is a patch for popcount builtin detection similar to LLVM. I
would like to queue this for review for next stage 1.
1. This is done part of loop-distribution and effective for -O3 and above.
2. This does not distribute loop to detect popcount (like
memcpy/memmove). I dont think that
The following patch fixes both PR56010 and PR83743. PR56010 is fixed by
adding an extra altname field to the RS6000_CPU table which matches the
cases where the Linux kernel's AT_PLATFORM name differs from the name GCC
expects. If we match on the altname, then we return the canonical name.
On January 24, 2018 at 1:29:12 PM, Steve Kargl
(s...@troutmask.apl.washington.edu) wrote:
Yes, thanks, Paul. Unfortunately, I've run out of time.
Damian, GCC is in stage 3, we need to wait for approval
from the release manager (aka Jakub) before committing
the patch.
Will do.
On Wed, Jan 24, 2018 at 01:25:51PM -0800, Damian Rouson wrote:
> Thank you, Paul. I think Alessandro has commit rights.
> If so, then I’ll ask him to make the requested edits and commit it.
>
> Damian
>
Yes, thanks, Paul. Unfortunately, I've run out of time.
Damian, GCC is in stage 3, we need
Thank you, Paul. I think Alessandro has commit rights. If so, then I’ll ask
him to make the requested edits and commit it.
Damian
On January 24, 2018 at 12:19:58 PM, Paul Richard Thomas
(paul.richard.tho...@gmail.com) wrote:
Hi All,
Given the delay relative to the start of stage 3, I
On 24/01/2018 18:53, Petr Ovtchenkov wrote:
On Wed, 24 Jan 2018 17:39:59 +0100
François Dumont wrote:
Hi
I'd like to propose this new debug check. Comparing with non-eos
istreambuf_iterator sounds like an obvious coding mistake.
I propose it despite the
Jeff Law writes:
> On 11/22/2017 11:12 AM, Richard Sandiford wrote:
>> Richard Sandiford writes:
>>> This patch adds support for the SVE bitwise reduction instructions
>>> (ANDV, ORV and EORV). It's a fairly mechanical extension of existing
>>>
Hi All,
Given the delay relative to the start of stage 3, I thought that I had
better deal with this asap:
+ /* TODO: this works on any derived type when
+ it should only work with team_type. */
+ if (team->ts.type != BT_DERIVED)
Why don't you give the team_type derived type
On Wed, Jan 24, 2018 at 12:35:38PM -0600, Segher Boessenkool wrote:
> Hi!
>
> On Wed, Jan 24, 2018 at 12:27:55AM -0500, Michael Meissner wrote:
> >
> > As Segher and I were discussing over private IRC, the root cause of this
> > bug is
> > the compiler no long generates the BDNZ instruction for
Szabolcs Nagy writes:
> Fix test failures with -mcmodel=tiny when adr is generated instead of adrp.
>
> FAIL: gcc.target/aarch64/sve/peel_ind_1.c -march=armv8.2-a+sve
> scan-assembler \\tadrp\\tx[0-9]+, x\\n
> FAIL: gcc.target/aarch64/sve/peel_ind_2.c -march=armv8.2-a+sve
>
The following patch fixes
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=84014
The patch was tested on powerpc64 and bootstrapped on x86-64.
Committed as rev. 257029.
Index: ChangeLog
===
--- ChangeLog (revision 257028)
+++
Hi!
On Wed, Jan 24, 2018 at 12:27:55AM -0500, Michael Meissner wrote:
>
> As Segher and I were discussing over private IRC, the root cause of this bug
> is
> the compiler no long generates the BDNZ instruction for a count down loop,
> instead it decrements the index in a GPR and does a
On January 24, 2018 6:51:54 PM GMT+01:00, Richard Biener
wrote:
>On January 24, 2018 6:40:25 PM GMT+01:00, Jakub Jelinek
> wrote:
>>On Wed, Jan 24, 2018 at 06:36:02PM +0100, Martin Jambor wrote:
>>> > I think there's already a set of attributes that prevent
On Wed, 24 Jan 2018 17:39:59 +0100
François Dumont wrote:
> Hi
>
> I'd like to propose this new debug check. Comparing with non-eos
> istreambuf_iterator sounds like an obvious coding mistake.
>
> I propose it despite the stage 1 as it is just a new debug
On January 24, 2018 6:40:25 PM GMT+01:00, Jakub Jelinek
wrote:
>On Wed, Jan 24, 2018 at 06:36:02PM +0100, Martin Jambor wrote:
>> > I think there's already a set of attributes that prevent cloning
>and
>> > or are adjusted by the IPA param machinery. The Martins or Honza
>> >
Prompted by PR target/83838 (Many gcc.target/i386/indirect-thunk*.c
tests FAIL), which is caused by this snippet in i386/sol2.h
/* Only recent versions of Solaris 11 ld properly support hidden .gnu.linkonce
sections, so don't use them. */
#ifndef USE_GLD
#define USE_HIDDEN_LINKONCE 0
#endif
On Wed, Jan 24, 2018 at 06:36:02PM +0100, Martin Jambor wrote:
> > I think there's already a set of attributes that prevent cloning and
> > or are adjusted by the IPA param machinery. The Martins or Honza
> > should know better.
>
> I am not sure I understand the problem but if
>
On Wed, Jan 24 2018, Richard Biener wrote:
> On January 24, 2018 5:16:45 PM GMT+01:00, Jakub Jelinek
> wrote:
>>On Wed, Jan 24, 2018 at 05:08:10PM +0100, Richard Biener wrote:
>>> >The "omp declare simd" attribute refers to argument numbers of the
>>> >functions, so trying to
On Tue, Dec 12, 2017 at 12:54:13AM -0200, Alexandre Oliva wrote:
> +/* Check whether BLOCK, a lexical block, is nested within OUTER, or is
> + OUTER itself. If BOTHWAYS, check not only that BLOCK can reach
> + OUTER through BLOCK_SUPERCONTEXT links, but also that there is a
> + path from
On Tue, Dec 12, 2017 at 12:52:18AM -0200, Alexandre Oliva wrote:
> --- a/include/dwarf2.h
> +++ b/include/dwarf2.h
> @@ -298,6 +298,14 @@ enum dwarf_location_list_entry_type
> DW_LLE_start_end = 0x07,
> DW_LLE_start_length = 0x08,
>
> +/*
>
On Wednesday 24 January 2018 06:29 PM, Siddhesh Poyarekar wrote:
>>> + /* Avoid register indexing for 128-bit stores when the
>>> + AARCH64_EXTRA_TUNE_SLOW_REGOFFSET_QUADWORD_STORE option is set. */
>>> + if (!optimize_size
>>> + && type == ADDR_QUERY_STR
>>> + &&
Hi,
Assorted updates for the vector intrinsics / gimple folding vec_cmp()
testcases to handle codegen variations as seen between P8,P9 targets.
This breaks apart the testcases into a #included header that contains the
testcase contents, and fold-vec-cmp-*.c tests that contain the target specific
Hi,
Assorted testcase updates to handle codegen variations between P7,p8,p9.
This breaks out the tests into p7,p8,p9 -specific versions of the same.
Sniff-tested on multiple systems, this clears up multiple errors
currently seen on P9.
OK for trunk?
Thanks
-Will
[testsuite]
Hi,
Update the vec-neg-longlong folding tests to handle codegen variations
as seen between p8 and p9 targets.
This breaks out the tests into p7,p8,p9 versions of the same, while moving the
common testcase content into a #included header.
sniff-tested across power systems (P6,P8,P9)
This
Hi
I'd like to propose this new debug check. Comparing with non-eos
istreambuf_iterator sounds like an obvious coding mistake.
I propose it despite the stage 1 as it is just a new debug check,
it doesn't impact the lib in normal mode.
Tested under Linux x86_64, ok to commit ?
On January 24, 2018 5:16:45 PM GMT+01:00, Jakub Jelinek
wrote:
>On Wed, Jan 24, 2018 at 05:08:10PM +0100, Richard Biener wrote:
>> >The "omp declare simd" attribute refers to argument numbers of the
>> >functions, so trying to apply it on versioned functions that can
>>
Fix test failures with -mcmodel=tiny when adr is generated instead of adrp.
FAIL: gcc.target/aarch64/sve/peel_ind_1.c -march=armv8.2-a+sve scan-assembler
\\tadrp\\tx[0-9]+, x\\n
FAIL: gcc.target/aarch64/sve/peel_ind_2.c -march=armv8.2-a+sve scan-assembler
\\tadrp\\tx[0-9]+, x\\n
FAIL:
Hi,
this patch fixes another issue found by Martin Liska's patch to trap in unlikely
section (last one which I need to bootstrap).
Here we confused local and global counts, which was misupdate at a time I
introudced them. Bootstrapped/regtested x86_64-linux.
Honza
* ipa-profile.c
On Wed, Jan 24, 2018 at 05:08:10PM +0100, Richard Biener wrote:
> >The "omp declare simd" attribute refers to argument numbers of the
> >functions, so trying to apply it on versioned functions that can
> >perhaps
> >have different number and types of arguments results in ICEs or
> >wrong-code.
>
On January 24, 2018 4:47:06 PM GMT+01:00, Jakub Jelinek
wrote:
>Hi!
>
>The "omp declare simd" attribute refers to argument numbers of the
>functions, so trying to apply it on versioned functions that can
>perhaps
>have different number and types of arguments results in ICEs or
Hi,
I'm looking into this rather mild regression, which should be relatively
easy to fix. In short, Jason's fix for c++/54325 moved an
abstract_virtuals_error_sfinae check from build_aggr_init_expr to
build_cplus_new therefore the testcase in this new bug isn't rejected
anymore because a
Hi!
The "omp declare simd" attribute refers to argument numbers of the
functions, so trying to apply it on versioned functions that can perhaps
have different number and types of arguments results in ICEs or wrong-code.
Unfortunately, if simd attribute or #pragma omp declare simd is used
on C++
On 01/24/2018 03:07 PM, Jakub Jelinek wrote:
On Wed, Jan 24, 2018 at 02:56:28PM +0100, Tom de Vries wrote:
+#if WORKAROUND_PTXJIT_BUG_2
+/* Variant of pc_set that only requires JUMP_P (INSN) if STRICT. This variant
+ is needed in the nvptx target because the branches generated for
+
Hi all,
This test fails on arm hardfloat targets because it sets an explicit
-mfloat-abi=softfp.
The usual approach to setting the NEON options is to use dg-add-options
arm_neon.
But in the lto tests we don't have that framework, we can only set them
explicitly with dg-lto-options.
The
On Tue, Jan 23, 2018 at 12:25 PM, Richard Sandiford
wrote:
> r255913 changed some constant_boolean_node calls to boolean_true_node
> and boolean_false_node, which meant that the returned tree didn't
> always have the right type.
>
> Tested on aarch64-linux-gnu.
On Wed, Jan 24, 2018 at 4:42 AM, Alexandre Oliva wrote:
> These two patches fix PR81611.
>
> The first one improves forwprop so that we avoid adding SSA conflicting
> by forwpropping the iv increment, which may cause both the incremented
> and the original value to be live,
On Tue, Jan 23, 2018 at 12:32 PM, Richard Sandiford
wrote:
> The failures in this PR were from forcing { dg-do run } even when
> vect.exp chooses options that are incompatible with the runtime.
> The default vect.exp behaviour is to execute when possible, so there's
On Wed, Jan 24, 2018 at 02:56:28PM +0100, Tom de Vries wrote:
> +#if WORKAROUND_PTXJIT_BUG_2
> +/* Variant of pc_set that only requires JUMP_P (INSN) if STRICT. This
> variant
> + is needed in the nvptx target because the branches generated for
> + parititioning are NONJUMP_INSN_P, not
tsubst_pack_expansion already knows how to deal with partial
instantiation of a pack expansion, where we end up with arguments for
some packs but not others, but my recent lambda work made this come up
in a new situation: within a function, where we need to deal with
function parameter packs as
Hi,
Fixed it. Ok for trunk?
gcc/
* config/i386/avx512bitalgintrin.h (_mm512_bitshuffle_epi64_mask,
_mm512_mask_bitshuffle_epi64_mask, _mm256_bitshuffle_epi64_mask,
_mm256_mask_bitshuffle_epi64_mask, _mm_bitshuffle_epi64_mask,
_mm_mask_bitshuffle_epi64_mask): Fix
On 01/24/2018 12:03 PM, Jakub Jelinek wrote:
On Wed, Jan 24, 2018 at 11:41:45AM +0100, Tom de Vries wrote:
+/* Insert a dummy ptx insn when encountering a branch to a label with no ptx
+ insn inbetween the branch and the label. This works around a JIT bug
+ observed at driver version
On 01/24/2018 12:53 PM, Richard Biener wrote:
On Wed, 24 Jan 2018, Tom de Vries wrote:
I'll commit this shortly for stage4. Strictly speaking, this is not an 8
regression, but a wrong code bug. But I think that the code generation error
seems fundamental enough, and the fix simple and localized
On Wednesday 24 January 2018 05:50 PM, Kyrill Tkachov wrote:
> I would tend towards making the costs usage more intelligent and
> differentiating
> between loads and stores but I agree that is definitely GCC 9 material.
> Whether this approach is an acceptable stopgap for GCC 8 is up to the
>
Hi Siddhesh,
On 23/01/18 15:41, Siddhesh Poyarekar wrote:
Hi,
Here's v2 of the patch to disable register offset addressing mode for
stores of 128-bit values on Falkor because they're very costly.
Differences from the last version:
- Incorporated changes Jim made to his patch earlier that I
On Wed, 24 Jan 2018, Tom de Vries wrote:
> Hi,
>
> atm the test-case contained in this patch hangs.
>
> For the test-case we generate:
> ...
> @ %r79 bra $L18;
> {
> call _gfortran_abort;
> trap;
> exit;
> }
> $L18:
> ...
>
> which results in SASS code (at
Ping.
Any comments on this?
The one line summary is that using self sets instead of clobber high would
result in a
patch roughly the same, but with different condition checks.
It depends if people think it really is useful for self sets to not be live.
Given that we are at stage 4 now, and this
Hi all,
This test fails to optimise away the PLUS reduction in the loop on arm targets
when vectorisation
is not enabled due to absence of SIMD instructions.
From reading the logs and the PR I gather that the presence or absence of SIMD
affects the passing of this test
on other targets as
Hi,
atm the test-case contained in this patch hangs.
For the test-case we generate:
...
@ %r79 bra $L18;
{
call _gfortran_abort;
trap;
exit;
}
$L18:
...
which results in SASS code (at GOMP_NVPTX_JIT=-O4):
...
/*05d8*/ @P0 BRA `(.L_18);
/*05e8*/
On Wed, 24 Jan 2018, Koval, Julia wrote:
> I think we may want to extend it to more than 2 ints someday, when we run out
> of bits again. It won't break the existing functionality if 3rd int will be
> zero by default. That's why I tried to avoid "two" in the name.
>
> Julia
>
> >
I think we may want to extend it to more than 2 ints someday, when we run out
of bits again. It won't break the existing functionality if 3rd int will be
zero by default. That's why I tried to avoid "two" in the name.
Julia
> -Original Message-
> From: Jakub Jelinek
On Wed, Jan 24, 2018 at 12:00:26PM +0100, Uros Bizjak wrote:
> On Mon, Jan 22, 2018 at 3:44 PM, Koval, Julia wrote:
> > Yes, you are right, any() is not required. Here is the patch.
>
> Please also attach ChangeLog.
>
> The patch is OK for x86 target, it needs global
On Wed, Jan 24, 2018 at 11:41:45AM +0100, Tom de Vries wrote:
> +/* Insert a dummy ptx insn when encountering a branch to a label with no ptx
> + insn inbetween the branch and the label. This works around a JIT bug
> + observed at driver version 384.111, at -O0 for sm_50. */
> +
> +static
On Mon, Jan 22, 2018 at 3:44 PM, Koval, Julia wrote:
> Yes, you are right, any() is not required. Here is the patch.
Please also attach ChangeLog.
The patch is OK for x86 target, it needs global reviewer approval
(Maybe Jakub, as the patch touches OMP part).
Uros.
>
On Wed, 24 Jan 2018, Tom de Vries wrote:
> Hi,
>
> this patch adds a workaround for the nvptx target JIT bug PR83589 - "[nvptx]
> mode-transitions.c and private-variables.{c,f90} execution FAILs at
> GOMP_NVPTX_JIT=-O0".
>
>
> When compiling a branch-around-nothing (where the branch is warp
Hi,
this patch adds a workaround for the nvptx target JIT bug PR83589 -
"[nvptx] mode-transitions.c and private-variables.{c,f90} execution
FAILs at GOMP_NVPTX_JIT=-O0".
When compiling a branch-around-nothing (where the branch is warp
neutering, so it's a divergent branch):
...
.reg
On Wed, Jan 24, 2018 at 10:58:57AM +0100, Tom de Vries wrote:
> I realize this is not really a stage4 patch, so: OK for stage1?
Ok.
Jakub
Hi,
The nvptx target PR83589 - "[nvptx] mode-transitions.c and
private-variables.{c,f90} execution FAILs at GOMP_NVPTX_JIT=-O0" is a
JIT bug.
I've written a workaround for this JIT bug (the second patch in this
series).
I've only managed to reproduce the JIT bug at JIT optimization level
The following enhances chrec_fold_plus_1 to handle a case to avoid
ICEing during GRAPHITE code-gen. It teaches us to handle
(signed T) { (T) x, +, y } + (signed T) z as
(signed T) ( { (T) x, +, y } + (T) (singed T) z ) instead of
failing with chrec_not_known.
Bootstrapped and tested on
Hi Jakub,
The lateness is indeed embarrassing but couldn't be helped.
> Given above my preference would be to keep version an int, and
> change rank and type to signed char and attribute to signed short.
> That way there will be no padding on either 32-bit or 64-bit targets,
> and the structure
Committed as obvious.
Richard.
2018-01-24 Richard Biener
PR middle-end/84000
* tree-cfg.c (replace_loop_annotate): Handle annot_expr_parallel_kind.
Index: gcc/tree-cfg.c
===
---
On Wed, Jan 24, 2018 at 10:30:41AM +0200, Janne Blomqvist wrote:
> On Tue, Jan 23, 2018 at 10:30 PM, Jakub Jelinek wrote:
> > On Tue, Jan 23, 2018 at 07:30:49PM +, Paul Richard Thomas wrote:
> >> Janne, Thanks.
> >>
> >> Jakub, is this OK with you?
> >
> > It is indeed quite
On Tue, Jan 23, 2018 at 10:30 PM, Jakub Jelinek wrote:
> On Tue, Jan 23, 2018 at 07:30:49PM +, Paul Richard Thomas wrote:
>> Janne, Thanks.
>>
>> Jakub, is this OK with you?
>
> It is indeed quite late for such large ABI changes, some distributions are
> about to start using
Hi,
this patch fixes another issue where basic block is incorrectly marked as
unlikely which is caught by Martin's hack to bb-reorder to insert trap
to all blocks in cold partition.
The problem solved here is that I have missed logic to set probabilities
to adjusted when doing basic arithmetic on
79 matches
Mail list logo