"Kewen.Lin" writes:
> Hi,
>
> When I am working on IFNs for vector with length, I noticed that the
> current optab support query for mask_load/mask_store looks unexpected.
> The mask_load/mask_store requires two modes for convert_optab query,
> but the macros direct_mask_{load,store}_optab_support
Hi.
When expanding a VEC_COND_EXPR it happens that first argument (a SSA_NAME)
that can be no longer used. When that happens we need to remove the SSA_NAME,
otherwise we end up expanding it and for targets like s390x, there's no optab
expansion. We need to remove them at both places as -O0 does n
For string types we must emit more complex code than simple pointer comparison.
It's pretty much the same what we emit for cl_optimization_print_diff.
I'm going to install it as obvious.
Patch can bootstrap on x86_64-linux-gnu and survives regression tests.
Martin
gcc/ChangeLog:
* optc
On 6/18/20 5:40 PM, Martin Liška wrote:
This is bogus as these are 2 strings that are equal. Let me fix it.
Hey.
Just for the report, this is fixed on master right now.
Martin
On Tue, Jun 23, 2020 at 5:31 PM Sunil K Pandey via Gcc-patches
wrote:
>
> From: Sunil K Pandey
>
> Default for this hook is NOP. For x86, in 32 bit mode, this hook
> sets alignment of long long on stack to 32 bits if preferred stack
> boundary is 32 bits.
>
> - This patch fixes
> gcc.tar
[CC list removed, it is not shown in archives]
> In i386-builtins.c, arch_names_table is used to to map architecture name
> string to internal model. A switch statement is used to map internal
> processor name to architecture name string and internal priority.
>
> model and priority are added to
On Tue, Jun 23, 2020 at 10:37 PM Ilya Leoshkevich via Gcc-patches
wrote:
>
> Hello,
>
> I needed to test [1] on AIX and OpenBSD and noticed
> download_prerequisites doesn't work there. The attached patch fixes
> it.
>
> OK for master?
OK if David acks the AIX part.
Thanks,
Richard.
> Best regar
Jim Wilson writes:
> On Tue, Jun 23, 2020 at 5:21 AM Richard Sandiford
> wrote:
>> MVE and Power both set inactive lanes to zero. But I'm not sure about RVV.
>> AIUI, for RVV the approach instead would be to reduce the effective vector
>> length for the final iteration of the vector loop, and I'
On Wed, Jun 24, 2020 at 9:13 AM Richard Sandiford
wrote:
>
> "Kewen.Lin" writes:
> > Hi,
> >
> > When I am working on IFNs for vector with length, I noticed that the
> > current optab support query for mask_load/mask_store looks unexpected.
> > The mask_load/mask_store requires two modes for conv
Hey.
The patch is about addition of some exceptions for arc target that
address:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=92860#c26
It's again another example where optimization options influence target
options.
Ready for master?
Martin
gcc/ChangeLog:
PR tree-optimization/92860
On Wed, Jun 24, 2020 at 9:21 AM Martin Liška wrote:
>
> Hi.
>
> When expanding a VEC_COND_EXPR it happens that first argument (a SSA_NAME)
> that can be no longer used. When that happens we need to remove the SSA_NAME,
> otherwise we end up expanding it and for targets like s390x, there's no optab
On 6/18/20 5:36 PM, Martin Liška wrote:
I'm going to add this to exception list.
Done here:
https://gcc.gnu.org/pipermail/gcc-patches/2020-June/548800.html
Martin
Please find attached a fix for PR95446. Patch originally posted to the
PR by Steve Kargl.
OK to commit to master and backport?
Commit message:
Fortran : False positive for optional arguments PR95446
Check that there is non-optional argument of the same rank in the
list of actual arguments.
Please find a fix for PR95829. Original patch by Steve Kargl posted to PR.
Commit to master and backport?
Commit message:
Fortran : Bogus error with additional blanks in type(*) PR95829
Checking for "* ) " instead of "*)" clears the bogus error.
2020-06-24 Steven G. Kargl
gcc/fortran/
Hi!
We folded A <= 0 ? A : -A into -ABS (A), which is for signed integral types
incorrect - can invoke on INT_MIN UB twice, once on ABS and once on its
negation.
The following patch fixes it by instead folding it to (type)-ABSU (A).
Bootstrapped/regtested on x86_64-linux and i686-linux, ok for t
Hi!
Bootstrapped/regtested on x86_64-linux and i686-linux, committed to trunk.
2020-06-24 Jakub Jelinek
* omp-low.c (lower_omp_for): Fix two pastos.
--- gcc/omp-low.c.jj2020-06-16 16:30:43.751400100 +0200
+++ gcc/omp-low.c 2020-06-23 14:14:54.972834539 +0200
@@ -10579,7 +10
On Wed, 24 Jun 2020, Jakub Jelinek wrote:
> Hi!
>
> We folded A <= 0 ? A : -A into -ABS (A), which is for signed integral types
> incorrect - can invoke on INT_MIN UB twice, once on ABS and once on its
> negation.
>
> The following patch fixes it by instead folding it to (type)-ABSU (A).
>
> Bo
Richard Biener writes:
> On Wed, Jun 24, 2020 at 9:13 AM Richard Sandiford
> wrote:
>>
>> "Kewen.Lin" writes:
>> > Hi,
>> >
>> > When I am working on IFNs for vector with length, I noticed that the
>> > current optab support query for mask_load/mask_store looks unexpected.
>> > The mask_load/mas
On Wed, Jun 24, 2020 at 09:35:42AM +0100, Richard Sandiford wrote:
> >> I keep meaning to experiment with dropping the second mode from these
> >> optabs,
> >> since it should be uniquely determined by the first. That would make
> >> things
> >> slightly simpler and more consistent with the othe
On 6/24/20 9:43 AM, Richard Biener wrote:
Hmm, can you instead use simple_dce_from_worklist and simply
record all SSA_NAMEs you end up "forwarding" as possibly dead
in a bitmap? At least that hashmap traversal looks dangerous
with respect to address-space randomization and gsi_remove
inserting d
On Tue, Jun 23, 2020 at 09:32PM +0200, Martin Liška wrote:
> On 6/23/20 1:27 PM, Marco Elver wrote:
> > Is this one good to go, or any objections?
> Thanks for it, it's fine. Please install it.
Done, thanks!
-- Marco
PING^2
On 6/10/20 10:12 AM, Martin Liška wrote:
PING^1
On 5/25/20 3:10 PM, Martin Liška wrote:
On 5/21/20 4:53 PM, Martin Sebor wrote:
On 5/21/20 5:28 AM, Martin Liška wrote:
On 5/18/20 10:37 PM, Martin Sebor wrote:
I know there are some somewhat complex cases the attribute exclusion
mechan
On Wed, Jun 24, 2020 at 10:49 AM Martin Liška wrote:
>
> On 6/24/20 9:43 AM, Richard Biener wrote:
> > Hmm, can you instead use simple_dce_from_worklist and simply
> > record all SSA_NAMEs you end up "forwarding" as possibly dead
> > in a bitmap? At least that hashmap traversal looks dangerous
>
Am 24.06.20 um 10:12 schrieb Mark Eggleston:
> Please find a fix for PR95829. Original patch by Steve Kargl posted to PR.
>
> Commit to master and backport?
>
> Commit message:
>
> Fortran : Bogus error with additional blanks in type(*) PR95829
>
> Checking for "* ) " instead of "*)" clears the b
Hi,
This is a fix for pr95854.
Only add a judgement to make sure operand1 and operand2 are both INTEGER_CST.
Bootstrap and tested on aarch64 Linux platform. No new regression witnessed.
Is it ok to be merged?
Thanks,
Kaipeng Zhou
0001-store-merging-ICE-in-find_bswap_or_nop_1-PR95854.patch
De
Jakub Jelinek writes:
> On Wed, Jun 24, 2020 at 09:35:42AM +0100, Richard Sandiford wrote:
>> >> I keep meaning to experiment with dropping the second mode from these
>> >> optabs,
>> >> since it should be uniquely determined by the first. That would make
>> >> things
>> >> slightly simpler and
On Tue, Jun 23, 2020 at 07:25:22PM -0400, y2s1982 via Gcc-patches wrote:
> This patch adds some unit tests for omp-tools.h header. It also adds some
> simple
> functions related to OMPD API versions. It also partially defines the OMPD
> initialization function.
>
> 2020-06-23 Tony Sim
>
> lib
On 6/24/20 11:09 AM, Richard Biener wrote:
On Wed, Jun 24, 2020 at 10:49 AM Martin Liška wrote:
On 6/24/20 9:43 AM, Richard Biener wrote:
Hmm, can you instead use simple_dce_from_worklist and simply
record all SSA_NAMEs you end up "forwarding" as possibly dead
in a bitmap? At least that hash
On Wed, Jun 24, 2020 at 09:15:57AM +, zhoukaipeng (A) wrote:
> Hi,
>
> This is a fix for pr95854.
>
> Only add a judgement to make sure operand1 and operand2 are both INTEGER_CST.
So what are those two operands then when not INTEGER_CSTs? POLY_INT_CSTs?
The documentation says:
operand 1
Oh, and the testcase is wrong too, you can't use AArch64 specific
compilation options in generic tests, so either the test needs to be moved
into gcc.target/aarch64/ and match in style how such -march option is
specified in other tests there, or should not have that -m* in dg-options,
but have /* {
Hi!
Given that nobody is available to review/approve this patch, and it
cannot really cause any harm, will my (old) review/"non-formal approval"
be sufficient for Frederik to push this? Or, in other words: Frederik,
please push if nobody objects within the next week.
Grüße
Thomas
On 2020-06-
Andrea Corallo writes:
> Hi all,
>
> here the latest version of the patch to enable Armv8.1-M Mainline
> LOB (low overhead branch) extension low overhead loops (LOL) feature
> using the 'loop-doloop' pass.
>
> I posted a previous version of it during stage 4.
>
> Given the following function:
>
>
Bootstrapped and regtested on x86_64-redhat-linux, ppc64le-redhat-linux
and s390x-redhat-linux.
Ok for master?
---
Bootstrap with musl libc fails with numerous "missing sentinel in
function call" errors. This is because musl defines NULL as 0L for C++,
but gcc requires sentinel value to be a po
The following adjusts vect_stmt_dominates_stmt_p to honor out-of-region
stmts we run into which have UID -1u.
Bootstrapped / tested on x86_64-unknown-linux-gnu, applied.
Richard.
2020-06-24 Richard Biener
PR tree-optimization/95856
* tree-vectorizer.c (vect_stmt_dominates_stm
This makes sure to emit SLP vectorized loads where the first scalar
load is. This makes SLP dependence checking more powerful because
hoisting loads can use TBAA and it increases the freedom for
vector placement when there are constraints from live lanes.
Vectorized shifts block inserting vectori
Hi Nathan,
Nathan Sidwell wrote:
> On 6/11/20 3:53 PM, Iain Sandoe wrote:
>> Hi
>> We had omitted the copying of function attributes (including the
>> 'used' status). Mark the outlined functions as artificial, since
>> they are; some diagnostic processing tests this.
>
> Do we do the right thi
On Tue, Jun 23, 2020 at 07:25:22PM -0400, y2s1982 via Gcc-patches wrote:
> --- /dev/null
> +++ b/libgomp/libgompd.c
> @@ -0,0 +1,46 @@
Oh, and one more thing, this file should start with the standard comment
with license etc., like you've added it to omp-tools.h.
> --- /dev/null
> +++ b/libgomp/l
On 23/06/2020 5:44 pm, Thomas Schwinge wrote:
Hi!
On 2020-06-15T21:28:12+0100, Kwok Cheung Yeung wrote:
This patch adds support on nvptx for __sync_val_compare_and_swap operations on
1- and 2-byte values.
Is this a thorough review that these are the only functions missing, or
did you just im
On Wed, Jun 24, 2020 at 12:31 AM Uros Bizjak wrote:
>
> [CC list removed, it is not shown in archives]
>
> > In i386-builtins.c, arch_names_table is used to to map architecture name
> > string to internal model. A switch statement is used to map internal
> > processor name to architecture name st
On Wed, 24 Jun 2020, zhoukaipeng (A) wrote:
> Hi,
>
> This is a fix for pr95854.
>
> Only add a judgement to make sure operand1 and operand2 are both INTEGER_CST.
>
> Bootstrap and tested on aarch64 Linux platform. No new regression witnessed.
>
> Is it ok to be merged?
Please use tree_fits_u
Hi Alexandre,
> On Jun 8, 2020, Alexandre Oliva wrote:
>
>> * outputs.exp (skip_lto): Set when missing the linker plugin.
>
> I withdraw the above work-around patch in favor of the fix proper below.
>
> It's also supposed to fix the FAILs caused by .dSYM directories on
> platforms that crea
On Mon, Jun 22, 2020 at 04:25:46PM -0700, H.J. Lu wrote:
> On Sun, Jun 21, 2020 at 10:22 AM H.J. Lu wrote:
> >
> > On Sun, Jun 21, 2020 at 10:18 AM Uros Bizjak wrote:
> > >
> > > On Sat, Jun 20, 2020 at 3:40 PM H.J. Lu wrote:
> > >
> > > > > > > >> 2) can we automatically deduce option name:
> >
The __detail::__to_chars_2 function assumes it won't be called with zero
values. However, when the output buffer is empty the caller doesn't
handle zero values correctly, and calls __to_chars_2 with a zero value,
resulting in an overflow of the empty buffer.
The __detail::__to_chars_i function sho
libstdc++-v3/ChangeLog:
* include/bits/stl_algobase.h (__find_if): Add FALLTHRU markers.
* include/std/charconv (__detail::__to_chars): Avoid
-Wsign-compare warning.
Tested powerpc64le-linux, committed to master.
commit 25920dd18ad12ea501309b1487366e22f35db631
Author: Jo
The parser for binary numbers returned an error if the entire string
contains more digits than the result type. Leading zeros should be
ignored.
libstdc++-v3/ChangeLog:
* include/std/charconv (__from_chars_binary): Ignore leading zeros.
* testsuite/20_util/from_chars/1.cc: Check "
On Wed, Jun 24, 2020 at 2:27 PM H.J. Lu wrote:
>
> On Mon, Jun 22, 2020 at 04:25:46PM -0700, H.J. Lu wrote:
> > On Sun, Jun 21, 2020 at 10:22 AM H.J. Lu wrote:
> > >
> > > On Sun, Jun 21, 2020 at 10:18 AM Uros Bizjak wrote:
> > > >
> > > > On Sat, Jun 20, 2020 at 3:40 PM H.J. Lu wrote:
> > > >
On Wed, Jun 24, 2020 at 02:43:43PM +0200, Uros Bizjak wrote:
> >
> > Here is the updated patch for x86 backend and libgcc. driver-i386.c
> > is unchanged.
>
> Thanks. We should change driver-i386.c very carefully and in an
> independent way from this patch. It is a complex and interwoven web of
>
All Sky Lake family processors have the same CPUID model number, 0x55.
The differences are Cascade Lake has AVX512VNNI and Cooper Lake has
AVX512VNNI + AVX512BF16. Check AVX512BF16 for Cooper Lake.
PR target/95774
* common/config/i386/cpuinfo.h (get_intel_cpu): Add Cooper Lake
On 6/24/20 7:00 AM, Iain Sandoe wrote:
Hi Nathan,
Nathan Sidwell wrote:
On 6/11/20 3:53 PM, Iain Sandoe wrote:
Hi
We had omitted the copying of function attributes (including the
'used' status). Mark the outlined functions as artificial, since
they are; some diagnostic processing tests this
On 6/23/20 9:36 PM, Thomas Schwinge wrote:
(*) Thomas mentioned that this is supposed to work also in more
complex cases than the one I outlined, although, that is probably
currently the most common one.
Static linking is another such case that we've seen in the wild
What I meant was: single
On Wed, Jun 24, 2020 at 3:06 PM H.J. Lu wrote:
>
> On Wed, Jun 24, 2020 at 02:43:43PM +0200, Uros Bizjak wrote:
> > >
> > > Here is the updated patch for x86 backend and libgcc. driver-i386.c
> > > is unchanged.
> >
> > Thanks. We should change driver-i386.c very carefully and in an
> > independe
On Wed, Jun 24, 2020 at 3:37 PM Uros Bizjak wrote:
>
> On Wed, Jun 24, 2020 at 3:06 PM H.J. Lu wrote:
> >
> > On Wed, Jun 24, 2020 at 02:43:43PM +0200, Uros Bizjak wrote:
> > > >
> > > > Here is the updated patch for x86 backend and libgcc. driver-i386.c
> > > > is unchanged.
> > >
> > > Thanks.
On Wed, Jun 24, 2020 at 03:36:25PM +0200, Tobias Burnus wrote:
> > What about 'gcc/config/i386/intelmic-mkoffload.c'? I see that one
> > unconditionally passes '-fPIC' to some things -- is that doing the right
> > thing for your case, too?
>
> I have not setup intelmic to test, but it uses -fPIC
This avoids vectorizing SLP subgraphs that just compute uniform
operations on all-same operands. That fixes the less interesting
(but most embarrasing) part of the testcase in the PR. On the
way it also fixed a missing matches[0] reset in the last
refactoring touching that place.
Bootstrap / reg
On 24/06/20 12:31 +0200, Ilya Leoshkevich wrote:
Bootstrapped and regtested on x86_64-redhat-linux, ppc64le-redhat-linux
and s390x-redhat-linux.
Ok for master?
---
Bootstrap with musl libc fails with numerous "missing sentinel in
function call" errors. This is because musl defines NULL as 0L
On Wed, Jun 24, 2020 at 11:27 AM Martin Liška wrote:
>
> On 6/24/20 11:09 AM, Richard Biener wrote:
> > On Wed, Jun 24, 2020 at 10:49 AM Martin Liška wrote:
> >>
> >> On 6/24/20 9:43 AM, Richard Biener wrote:
> >>> Hmm, can you instead use simple_dce_from_worklist and simply
> >>> record all SSA_
On Wed, Jun 24, 2020 at 4:14 PM Jonathan Wakely via Gcc-patches
wrote:
>
> On 24/06/20 12:31 +0200, Ilya Leoshkevich wrote:
> >Bootstrapped and regtested on x86_64-redhat-linux, ppc64le-redhat-linux
> >and s390x-redhat-linux.
> >
> >Ok for master?
> >
> >---
> >
> >Bootstrap with musl libc fails w
On Wed, Jun 24, 2020 at 3:33 AM Richard Biener
wrote:
>
> On Tue, Jun 23, 2020 at 10:37 PM Ilya Leoshkevich via Gcc-patches
> wrote:
> >
> > Hello,
> >
> > I needed to test [1] on AIX and OpenBSD and noticed
> > download_prerequisites doesn't work there. The attached patch fixes
> > it.
> >
> > O
On Wed, Jun 24, 2020 at 03:46:25PM +0200, Uros Bizjak wrote:
> On Wed, Jun 24, 2020 at 3:37 PM Uros Bizjak wrote:
> >
> > On Wed, Jun 24, 2020 at 3:06 PM H.J. Lu wrote:
> > >
> > > On Wed, Jun 24, 2020 at 02:43:43PM +0200, Uros Bizjak wrote:
> > > > >
> > > > > Here is the updated patch for x86 b
On Wed, Jun 24, 2020 at 4:38 PM H.J. Lu wrote:
>
> On Wed, Jun 24, 2020 at 03:46:25PM +0200, Uros Bizjak wrote:
> > On Wed, Jun 24, 2020 at 3:37 PM Uros Bizjak wrote:
> > >
> > > On Wed, Jun 24, 2020 at 3:06 PM H.J. Lu wrote:
> > > >
> > > > On Wed, Jun 24, 2020 at 02:43:43PM +0200, Uros Bizjak
On Wed, Jun 24, 2020 at 3:12 PM H.J. Lu wrote:
>
> All Sky Lake family processors have the same CPUID model number, 0x55.
> The differences are Cascade Lake has AVX512VNNI and Cooper Lake has
> AVX512VNNI + AVX512BF16. Check AVX512BF16 for Cooper Lake.
>
> PR target/95774
> * comm
Brand ID was a feature that briefly existed in some Pentium III and
Pentium 4 CPUs. The CPUs that had non-zero brand ID still have had
valid family/model. Brand ID just gives a marketing name for the CPU.
Remove the extra code for brand ID check.
gcc/
PR target/95660
* common/co
On Wed, Jun 24, 2020 at 5:01 PM H.J. Lu wrote:
>
> Brand ID was a feature that briefly existed in some Pentium III and
> Pentium 4 CPUs. The CPUs that had non-zero brand ID still have had
> valid family/model. Brand ID just gives a marketing name for the CPU.
> Remove the extra code for brand ID
Hello,
Thank you for the tips. I have some follow up questions. :)
On Wed, Jun 24, 2020 at 5:25 AM Jakub Jelinek wrote:
> > libgomp/ChangeLog:
> >
> > * Makefile.am (toolexeclib_LTLIBRARIES and related): Add
> libgompd.la.
>
> Please spell out all the changes. I.e.
> * Makefile.a
Hi,
Darwin has signed chars and the fields in the insn_data
struct are const char, which leads to a bootstrap fail with
"error: comparison of integer expressions of different signedness: 'unsigned
int' and 'const char' [-Werror=sign-compare]”
OK for master?
Iain
gcc/ChangeLog:
* confi
On Wed, Jun 24, 2020 at 11:11:47AM -0400, y2s1982 . wrote:
> > And introduce libgompd.map version script which will have the exported
> > symbols in for now OMPD_5.0 symbol version.
> > So
> > OMPD_5.0 {
> > global:
> > ...;
> >
> > local:
> > *;
> > };
> >
>
> Oh I wasn't sure
On 6/24/20 5:26 PM, Jakub Jelinek via Gcc-patches wrote:
Just build the whole gcc tree.
If you are at the toplevel of gcc tree (i.e. directory that contains libgomp
subdirectory), do
mkdir obj
cd obj
../configure --disable-bootstrap --enable-languages=c,c++,fortran
make -j16 # or how many cores
CLWB isn't supported on Ice Lake client. But Ice Lake server and Tiger
Lake support it. Move PTA_CLWB to PTA_ICELAKE_SERVER and PTA_TIGERLAKE.
PR target/95874
* config/i386/i386.h (PTA_ICELAKE_CLIENT): Remove PTA_CLWB.
(PTA_ICELAKE_SERVER): Add PTA_CLWB.
(PTA_TIGE
Hello
There appears to be a bug in the handling of the 'if' clause (without a
directive name modifier) for combined OpenMP constructs in the Fortran front-end:
static void
gfc_split_omp_clauses (gfc_code *code,
gfc_omp_clauses clausesa[GFC_OMP_SPLIT_NUM])
{
...
if (code
On Wed, Jun 24, 2020 at 06:03:26PM +0200, Tobias Burnus wrote:
> On 6/24/20 5:26 PM, Jakub Jelinek via Gcc-patches wrote:
>
> > Just build the whole gcc tree.
> > If you are at the toplevel of gcc tree (i.e. directory that contains libgomp
> > subdirectory), do
> > mkdir obj
> > cd obj
> > ../conf
On Wed, Jun 24, 2020 at 05:47:06PM +0100, Kwok Cheung Yeung wrote:
> There appears to be a bug in the handling of the 'if' clause (without a
> directive name modifier) for combined OpenMP constructs in the Fortran
> front-end:
>
> static void
> gfc_split_omp_clauses (gfc_code *code,
>
Hi Harald,
here's another fix for a buffer overflow with long symbols.
Is the memory you allocate freed anywhere? We have a large
number of memory leaks in the front end, but maybe we shouldn't
add another one :-)
Best regards
Thomas
Hi Haraldd,
here's another case with a buffer that did overflow.
Regtested on x86_64-pc-linux-gnu.
OK for master / backports?
OK. Thanks for the patch!
Best regards
Thomas
Hi Thomas,
> Is the memory you allocate freed anywhere? We have a large
> number of memory leaks in the front end, but maybe we shouldn't
> add another one :-)
my reading of the definition of XALLOCAVEC
include/libiberty.h:#define XALLOCAVEC(T, N)((T *) alloca (sizeof (T) *
(N)))
and allo
Hi Kwok,
the TODO is fixed by the attached patch; I would be happy if you could handle
this patch,
e.g. together with your patch – or as follow up.
(Lightly tested only, i.e. it fixes the ICE but I did not
do a full testsuite run. But I regard it as obvious.)
Tobias
On 6/24/20 6:47 PM, Kwok C
(OpenMP 5 extends this a lot, but this is about OpenMP 4.5.
It touches code which is also used by OpenACC's attach/detach.)
@OpenACC/Julian: I think the character attach/detach for
deferred-length strings does not work properly with OpenACC;
I did not touch this code – but I think it needs some l
Hi!
On Wed, Jun 24, 2020 at 04:15:53PM +0100, Iain Sandoe wrote:
> Darwin has signed chars and the fields in the insn_data
> struct are const char, which leads to a bootstrap fail with
>
> "error: comparison of integer expressions of different signedness: 'unsigned
> int' and 'const char' [-Werr
On 11/06/20 08:32 +0200, François Dumont via Libstdc++ wrote:
As we are on patching algos we still have this old one.
   From the original patch I only kept the memory optimization part
as the new performance test was not showing good result for the other
part to change pivot value. I als
Hi Tobias,
OK for the trunk?
I just checked, and this gets a segfault for
program main
character (len=3), parameter :: x = 'abc'
print *, x(2:2)
end program main
+ if (ts)
+e->kind = ts->kind;
+ else if (e->symtree->n.sym->ts.type == BT_CHARACTER)
+e->kind = ts->kind;
There a
Hi Thomas,
could you review the second patch instead? I have sent the wrong patch
(early draft) and corrected it half an hour later!
https://gcc.gnu.org/pipermail/gcc-patches/2020-June/548779.html
Tobias
On 6/24/20 8:01 PM, Thomas Koenig via Fortran wrote:
Hi Tobias,
OK for the trunk?
I
On 6/18/20 3:42 PM, Peter Bergner wrote:
> POWER ISA 3.1 added new Matrix-Multiply Assist (MMA) instructions.
> The following patch set adds support for generating these instructions
> through built-in functions which are enabled with the -mmma option.
>
> The patch1 and patch1+patch2+patch3 have
From: Roger Sayle
2020-06-24 Roger Sayle
* simplify-rtx.c (simplify_unary_operation_1): Simplify
(parity (parity x)) as (parity x), i.e. PARITY is idempotent.
---
gcc/simplify-rtx.c | 4
1 file changed, 4 insertions(+)
diff --git a/gcc/simplify-rtx.c b/gcc/simplify-rtx.
From: Roger Sayle
2020-06-24 Roger Sayle
Segher Boessenkool
* simplify-rtx.c (simplify_unary_operation_1): Simplify rotates by 0.
---
gcc/simplify-rtx.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/gcc/simplify-rtx.c b/gcc/simplify-rtx.c
index 65008ea..3e913b5
Hi!
On Wed, Jun 24, 2020 at 02:28:00PM -0500, Peter Bergner via Gcc-patches wrote:
> On 6/18/20 3:42 PM, Peter Bergner wrote:
> > POWER ISA 3.1 added new Matrix-Multiply Assist (MMA) instructions.
> > The following patch set adds support for generating these instructions
> > through built-in funct
On 6/23/20 6:58 PM, Marek Polacek wrote:
95568 complains that CTAD for aggregates doesn't work within
requires-clause and it turned out that it doesn't work when we try
the deduction in a template. The reason is that maybe_aggr_guide
creates a guide that can look like this
template X(decltyp
Looking in the vtable for cand->conversion_path doesn't work because
it's the base where we found the function, not the base where the
function is defined. If those are different, look farther. Applying to
trunk and 10.
The second patch uses that same base lookup for the actual conversion as
Hi Tobias,
could you review the second patch instead? I have sent the wrong patch
(early draft) and corrected it half an hour later!
Sorry, I missed that. Here's the review of the real patch :-)
So, the first part is
+ if (ts)
+e->ts.kind = ts->kind;
Ok, I unerstand that - ts has been
I'm applying your patch with some tweaks.
First, I shortened your subject line; we try to keep them under 50 chars
so they look good in git log --oneline. Going over isn't a big deal,
but that's the goal.
On 6/23/20 3:47 PM, Nicholas Krause wrote:
From: Nicholas Krause
This fixes the PR95
Hi,
Just an improvement to test-coverage,
tested on x86_64-darwin,linux powerpc64-linux
applied to master,
thanks
Iain
=
We updated the handling of the errors for cases when the
ramp return cannot be constructed from the user's provided
get-return-object method. This updates the testcases
On 24/06/20 7:39 pm, Jonathan Wakely wrote:
On 11/06/20 08:32 +0200, François Dumont via Libstdc++ wrote:
As we are on patching algos we still have this old one.
   From the original patch I only kept the memory optimization
part as the new performance test was not showing good result for
Hi Martin,
Asher Gordon writes:
> Martin Sebor writes:
>
>> Unfortunately, attributes don't have locations (yet).
>
> Hmm, well maybe I could implement that. I'm not very familiar with the
> GCC source (this is my first patch), but I'll see if I can figure it
> out. It would probably be useful
Another case of buffer overflow, this time coming in pairs.
Regtested on x86_64-pc-linux-gnu.
OK for master?
Thanks,
Harald
PR fortran/95828 - Buffer overflows with SELECT RANK
With SELECT RANK, name mangling results in long internal symbols that
overflows internal buffers. Fix that.
gcc/fo
This is an ICE-on-invalid but I've been seeing it when reducing
various testcases, so it's more important for me than usually.
splice_late_return_type now checks that if we've seen a late return
type, the function return type was auto. That's a fair assumption
but grokdeclarator/cdk_function wasn
On 6/24/20 2:15 PM, Asher Gordon wrote:
Hi Martin,
Asher Gordon writes:
Martin Sebor writes:
Unfortunately, attributes don't have locations (yet).
Hmm, well maybe I could implement that. I'm not very familiar with the
GCC source (this is my first patch), but I'll see if I can figure it
o
Hi!
On Tue, Jun 23, 2020 at 01:20:53PM +0100, Richard Sandiford wrote:
> SVE supports integer division btw. :-)
So does Power (ISA 3.1, power10).
> In summary, I'm not saying we should never define the inactive values
> to be zero. I just think that we should leave it until it matters.
> And I
Hi!
On Thu, Jun 18, 2020 at 03:20:01PM -0700, Carl Love wrote:
> 2020-06-18 Carl Love
>
> * config/rs6000/altivec.md: (UNSPEC_EXTRACTL, UNSPEC_EXTRACTR)
> (vextractl, vextractr)
> (vextractl_internal, vextractr_internal)
Please say what the iterator for the is? Just adding
Hi!
On Thu, Jun 18, 2020 at 03:20:05PM -0700, Carl Love wrote:
> * config/rs6000/altivec.h (vec_insertl, vec_inserth): New defines.
> * config/rs6000/rs6000-builtin.def (VINSERTGPRBL, VINSERTGPRHL,
> VINSERTGPRWL, VINSERTGPRDL, VINSERTVPRBL, VINSERTVPRHL, VINSERTVPRWL,
> VI
Hi!
On Thu, Jun 18, 2020 at 03:20:10PM -0700, Carl Love wrote:
> +The programmer is responsible for understanding the endianness issues
> involved
> +with the first argument and the result.
:-)
> --- /dev/null
> +++ b/gcc/testsuite/gcc.target/powerpc/vec-replace-word-runnable.c
> @@ -0,0 +1,289
Martin Sebor writes:
> I have no experience with changing tree nodes but I wouldn't be
> surprised if there were assumptions baked into code that made it
> a non-trivial exercise.
>
> There's also lots of sharing of data in GCC so I'm not sure it
> makes sense for an identifier to have an associa
On Wed, Jun 24, 2020 at 12:30 AM Richard Biener
wrote:
>
> On Tue, Jun 23, 2020 at 5:31 PM Sunil K Pandey via Gcc-patches
> wrote:
> >
> > From: Sunil K Pandey
> >
> > Default for this hook is NOP. For x86, in 32 bit mode, this hook
> > sets alignment of long long on stack to 32 bits if preferre
1 - 100 of 107 matches
Mail list logo
