*Ping*
I intent to incorporate Sandra's suggestions, except for the beginning of line
spacing - that's needed to avoid exceeding the 80 character line limit. I did
not include an updated patch as just pinging is easier on a mobile during
vacation :-)
Thanks,
Tobias
Loosemore, Sandra wrote:
On Thu, Jul 8, 2021 at 7:44 AM Segher Boessenkool
wrote:
>
> On Wed, Jul 07, 2021 at 11:23:48PM +0800, Hongtao Liu wrote:
> > On Wed, Jul 7, 2021 at 10:54 PM Segher Boessenkool
> > wrote:
>
> [ snip some old stuff ]
>
> > > Yeah. This stuff needs a rethink.
> > >
> > > What is wrong with just
On 7/7/21 7:48 PM, Marek Polacek wrote:
On Wed, Jul 07, 2021 at 02:38:11PM -0600, Martin Sebor via Gcc-patches wrote:
On 7/7/21 1:38 AM, Richard Biener wrote:
On Tue, Jul 6, 2021 at 5:47 PM Martin Sebor via Gcc-patches
wrote:
Ping:
On Wed, Jul 07, 2021 at 02:38:11PM -0600, Martin Sebor via Gcc-patches wrote:
> On 7/7/21 1:38 AM, Richard Biener wrote:
> > On Tue, Jul 6, 2021 at 5:47 PM Martin Sebor via Gcc-patches
> > wrote:
> > >
> > > Ping: https://gcc.gnu.org/pipermail/gcc-patches/2021-June/573349.html
> >
> > + if
It sounds plausible that this assert
int f();
static_assert(noexcept(sizeof(f(;
should pass: sizeof produces a std::size_t and its operand is not
evaluated, so it can't throw. noexcept should only evaluate to
false for potentially evaluated operands. Therefore I think that
On 7/7/21 4:23 AM, Matthias Kretz wrote:
On Tuesday, 22 June 2021 21:52:16 CEST Jason Merrill wrote:
2. About the namespace aliases: IIUC an attribute would currently be
rejected because of the C++ grammar. Do you want to make it valid before
WG21 officially decides how to proceed? And if you
On Wed, 07 Jul 2021 16:26:25 PDT (-0700), i...@maskray.me wrote:
On 2021-07-02, Kito Cheng wrote:
It was undocument before, but already used in linux kernel, so LLVM
community suggest we should document that, so that make it become
supported/documented/non-internal machine constraints.
Hi Segher,
on 2021/7/8 上午1:10, Segher Boessenkool wrote:
> Hi!
>
> On Wed, Jul 07, 2021 at 05:03:23PM +0800, Kewen.Lin wrote:
>> This patch is to make Power10 newly introduced vector
>> modulo instructions exploited in vectorized loops, it
>> just simply renames existing define_insns as standard
On Wed, Jul 07, 2021 at 11:32:59PM +0800, Hongtao Liu wrote:
> On Wed, Jul 7, 2021 at 10:54 PM Segher Boessenkool
> wrote:
> > So, a "FAKE_CALL" is very much a *real* call, on the RTL level, which is
> > where we are here. But you want it to be treated differently because it
> > will eventually
On Wed, Jul 07, 2021 at 11:23:48PM +0800, Hongtao Liu wrote:
> On Wed, Jul 7, 2021 at 10:54 PM Segher Boessenkool
> wrote:
[ snip some old stuff ]
> > Yeah. This stuff needs a rethink.
> >
> > What is wrong with just using an unspec and clobbers?
> >
> It's partial and **potential clobber**,
I'm working on reimplementing -Wanalyzer-use-of-uninitialized-value, but
I ran into issues with
region_model::add_any_constraints_from_ssa_def_stmt.
This function is from the initial commit of the analyzer and walks the
SSA names finding conditions that were missed due to the GCC 10 era
On 2021-07-02, Kito Cheng wrote:
It was undocument before, but already used in linux kernel, so LLVM
community suggest we should document that, so that make it become
supported/documented/non-internal machine constraints.
gcc/ChangeLog:
PR target/101275
* doc/md.text (Machine
On Wed, Jul 07, 2021 at 03:55:44PM -0400, Michael Meissner wrote:
> This patch adds support for the VDIVSQ, VDIVUQ, VMODSQ, and VMODUQ
> instructions to do 128-bit arithmetic.
> gcc/
> PR target/100809
> * config/rs6000/rs6000.md (udivti3): New insn.
> (divti3): New insn.
>
On 7/5/21 10:18 AM, Matthias Kretz wrote:
On Thursday, 1 July 2021 17:18:26 CEST Jason Merrill wrote:
You probably want to adjust is_late_template_attribute to change that.
Right, I hacked is_late_template_attribute but now I only see a TYPE_DECL
passed to my attribute handler
Hello, I have added myself for Write After Approval and DCO.
Thanks Gaius
=
/
* MAINTAINERS: Add myself for write after approval and DCO.
2021-07-06 Gaius Mulley
Signed-off-by: Gaius Mulley
diff --git a/MAINTAINERS b/MAINTAINERS
index
On Wed, 2021-07-07 at 21:03 +0200, Ilya Leoshkevich wrote:
> Bootstrapped and regtested on s390x-redhat-linux. Ok for master?
>
> v1: https://gcc.gnu.org/pipermail/gcc-patches/2021-June/573614.html
> v1 -> v2: Do not use UNSPEC_PLT in 64-bit code and rename it to
> UNSPEC_PLT31
On 7/7/21 11:29 AM, Patrick Palka wrote:
r12-1989 fixed the testcase in the PR, but unfortunately the fix is
buggy:
1. It breaks the case where the common template between the
TEMPLATE_DECL t and ctx_parms is the innermost template (as in
concepts-memtmpl5.C below). This can be
On 7/7/21 1:38 AM, Richard Biener wrote:
On Tue, Jul 6, 2021 at 5:47 PM Martin Sebor via Gcc-patches
wrote:
Ping: https://gcc.gnu.org/pipermail/gcc-patches/2021-June/573349.html
+ if (TREE_CODE (axstype) != UNION_TYPE)
what about QUAL_UNION_TYPE? (why constrain union type accesses
here -
[PATCH] PR 100168: Fix call test on power10.
Fix a test that was checking for 64-bit TOC calls, to also allow for
PC-relative calls.
I have verified that this test passes when run on a power10 system configured
with --with-cpu=power10 and it continues to pass on power9 little endian and
power8
[PATCH] PR 100170: Fix eq/ne tests on power10.
This patch updates eq/ne tests in the testsuite to adjust the test if
power10 code generation is used.
I have verified that these tests run on a power10 system using the
--with-cpu=power10 configuration option, and they continue to run on power9
[PATCH] PR 100167: Fix vector long long multiply/divide tests on power10.
This patch updates the vector long long multiply and divide tests to
supply the correct code information if power10 code generation is used.
2021-07-07 Michael Meissner
gcc/testsuite/
PR testsuite/100167
[PATCH] Deal with prefixed loads/stores in tests, PR testsuite/100166
This patch updates the various tests in the testsuite to treat plxv
and pstxv as being vector loads/stores. This shows up if you run the
testsuite with a compiler configured with the option: --with-cpu=power10.
I have
[PATCH] Fix vec-splati-runnable.c test.
I noticed that the vec-splati-runnable.c did not have an abort after one
of the tests. If the test was run with optimization, the optimizer could
delete some of the tests and throw off the count. However, due to the
fact that the value being loaded in
[PATCH] Change rs6000_const_f32_to_i32 return type.
The function rs6000_const_f32_to_i32 called REAL_VALUE_TO_TARGET_SINGLE
with a long long type and returns it. This patch changes the type to long
which is the proper type for REAL_VALUE_TO_TARGET_SINGLE.
2021-07-07 Michael Meissner
gcc/
[PATCH] Fix long double tests when default long double is not IBM.
This patch adds 3 more selections to target-supports.exp to see if we can force
the compiler to use a particular long double format (IEEE 128-bit, IBM extended
double, 64-bit), and the library support will track the changes for
[PATCH] Generate 128-bit int divide/modulus on power10.
This patch adds support for the VDIVSQ, VDIVUQ, VMODSQ, and VMODUQ
instructions to do 128-bit arithmetic.
Ideally this patch can be approved in time to be back ported to GCC 11.2.
Can I check this into the master branch, and eventually
On 7 Jul 2021, Nick Clifton told this:
> Hi Nick,
>
>> Ping?
>
> Oops.
I sent a bunch of pings out at the same time, to a bunch of different
projects. You are the only person to respond, so thank you!
>>> PR libctf/27482
>>> * libtool.m4 (LT_PATH_NM): Try BSDization flags with a
Bootstrapped and regtested on s390x-redhat-linux. Ok for master?
v1: https://gcc.gnu.org/pipermail/gcc-patches/2021-June/573614.html
v1 -> v2: Do not use UNSPEC_PLT in 64-bit code and rename it to
UNSPEC_PLT31 (Ulrich, Andreas). Do not append @PLT only to
weak symbols in
On 7/7/21 11:55 AM, Peter Bergner wrote:
> On 7/6/21 5:05 PM, Segher Boessenkool wrote:
>> On Tue, Jul 06, 2021 at 04:13:06PM -0500, Peter Bergner wrote:
>>> On 7/6/21 3:08 PM, Bill Schmidt wrote:
On 7/6/21 2:29 PM, Peter Bergner wrote:
> These are RS6000_BTC_GIMPLE, so I think they
This patch adds device-modifiers to the device clause:
#pragma omp target device ([ device-modifier :] integer-expression)
where device-modifier is either 'ancestor' or 'device_num'.
The 'device_num' case
#pragma omp target device (device_num : integer-expression)
is treated in the
On 7/7/2021 8:55 AM, Segher Boessenkool wrote:
On Mon, Jul 05, 2021 at 06:03:21PM -0600, Jeff Law wrote:
It reminds me a bit of millicode calls on the PA or calls to special
routines in libgcc. They're calls to functions, but those functions do
not follow the standard ABI.
Something with
Hi!
On Wed, Jul 07, 2021 at 05:03:23PM +0800, Kewen.Lin wrote:
> This patch is to make Power10 newly introduced vector
> modulo instructions exploited in vectorized loops, it
> just simply renames existing define_insns as standard
> pattern names.
>
> Is it ok for trunk?
>
> BR,
> Kewen
>
On 7/6/21 5:05 PM, Segher Boessenkool wrote:
> On Tue, Jul 06, 2021 at 04:13:06PM -0500, Peter Bergner wrote:
>> On 7/6/21 3:08 PM, Bill Schmidt wrote:
>>> On 7/6/21 2:29 PM, Peter Bergner wrote:
These are RS6000_BTC_GIMPLE, so I think they should be handled within
the "if (gimple_func)
On Wed, Jul 07, 2021 at 05:03:07PM +0800, Kewen.Lin wrote:
> This patch is to add one test case to check if vectorizer
> can exploit vector division instrutions newly introduced
> by Power10.
Okay, great, thanks!
Segher
__MINWGW32__ is typo and was not working, but anyway the MSG_WAITALL
is defined on Windows.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* s-oscons-tmplt.c (MSG_WAITALL): Remove wrong #ifdef
__MINWGW32__.diff --git a/gcc/ada/s-oscons-tmplt.c
When expansion is disabled (such as with -gnatc), there are cases where
uses of concurrent types can lead to an Assertion_Failure when
extensions are enabled (by use of -gnatX, or due to instantiation of a
predefined library generic, such as Unchecked_Conversion), because
Primitive_Operations can
Ensure that Checking_For_Potentially_Static_Expression flag is
manipulated in a stack-like manner (with stack depth 1 at the most). The
previous assertion didn't prevent us from setting the flag to True or to
False twice in a row.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
A recent change exposed a latent bug where the Is_Intrinsic_Subprogram
flag was not propagated properly, leading to errors from the front-end
inlining of the form:
cannot inline "xxx" (nested function instantiation)
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
*
Windows headers in GCC could miss some available constants. Hardcode
TCP_KEEPCNT, TCP_KEEPIDLE, and TCP_KEEPINTVL constants for such case.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* s-oscons-tmplt.c (TCP_KEEPCNT TCP_KEEPIDLE, TCP_KEEPINTVL):
Hardcode on Windows
The body of every primitive subprogram contains an elaboration check, in
case a dispatching call is made. These checks happen even in the static
model. This patch removes the checks if pragma Pure or Preelaborate is
present, because they cannot fail in that case.
Tested on x86_64-pc-linux-gnu,
Code cleanup; semantics is unaffected.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* sem_prag.adb (Analyze_Pragma): Simplify processing of pragma
CPP_Constructor.diff --git a/gcc/ada/sem_prag.adb b/gcc/ada/sem_prag.adb
--- a/gcc/ada/sem_prag.adb
+++
To help codepeer analysis.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* libgnat/g-debpoo.adb (Code_Address_For_Allocate_End): Default
Initialize.diff --git a/gcc/ada/libgnat/g-debpoo.adb b/gcc/ada/libgnat/g-debpoo.adb
--- a/gcc/ada/libgnat/g-debpoo.adb
+++
In particular, now that we are using Atomic_Unsigned which is marked
Atomic, we no longer need to mark Atomic_Counter.Value explicitly
atomic.
We can also get rid of all uses of 'Unrestricted_Access
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* libgnat/s-atocou.ads,
Ada 2022, in AI12-0412, specifies that certain uses of primitives of an
abstract type that have Pre'Class or Post'Class aspect are illegal when
an aspect is given with a nonstatic expression. Specifically, if the
primitive is nonabstract and has such aspects, it's illegal to make a
nondispatching
Ada 2022 errors about volatile compatibility between generic actual and
formal types were emitted on type declaration; now they are emitted at
the actual type within the generic instance.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* sem_ch12.adb
Code cleanup; semantics is unaffected.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* uname.adb (Get_Unit_Name): Simplify with a bounded string
buffer; also, this addresses a ??? comment about the max length
being exceeded.diff --git a/gcc/ada/uname.adb
Code cleanup related to loading of compilation units; behaviour is
unaffected.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* uname.adb (Get_Body_Name, Get_Parent_Body_Name,
Get_Parent_Spec_Name, Get_Spec_Name, Is_Child_Name,
Is_Body_Name, Is_Spec_Name,
Size and Value_Size clauses are documented to be the same, except that
Value_Size is allowed for nonfirst subtypes, and Size is allowed for
objects. This was far from true, which caused bugs such as ignoring
Value_Size for access types, in cases where a Size clause would trigger
the use of thin
Back out of an overly aggressive workaround for compilation problems
associated with a Put_Image routine for a tagged type in a Remote_Types
package and try a different (hopefully better) approach that is more
consistent with how other predefined primitives are treated.
Tested on
The Set_Socket_Option shifts timeout for -500ms on old Windows versions,
but Get_Socket_Option did +500ms for timeouts on all Windows versions.
This commit fixes it and +500ms on Get_Socket_Option only for old
Windows versions.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
A previous change to Unchecked_Convert_To removed the setting of the
Parent of the new node, because it was thought to be unnecessary.
However, in rare cases, it is necessary because for example
Remove_Side_Effects is called on the new node before attaching it
to the tree.
Tested on
In cases where it is known statically that an exception will be raised,
the frontend inserts a node of kind N_Raise_xxx_Error. This node kind
was previously not supported in GNATprove, which required special
handling for these cases in GNATprove mode. This is not needed anymore.
Tested on
Code cleanup; semantics is unaffected.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* sem_ch6.adb (Check_For_Primitive_Subprogram): Move
declarations of local variables after nested subprogram bodies.diff --git a/gcc/ada/sem_ch6.adb b/gcc/ada/sem_ch6.adb
---
Code cleanup; semantics is unaffected.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* exp_disp.adb (CPP_Num_Prims): Reuse List_Length.diff --git a/gcc/ada/exp_disp.adb b/gcc/ada/exp_disp.adb
--- a/gcc/ada/exp_disp.adb
+++ b/gcc/ada/exp_disp.adb
@@ -588,19 +588,7 @@ package
Code cleanup; semantics is unaffected.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* exp_ch3.adb, exp_ch6.adb, sem_ch6.adb: Replace Ekind
membership test in Private_Kind with a call to Is_Private_Type.diff --git a/gcc/ada/exp_ch3.adb b/gcc/ada/exp_ch3.adb
---
Remove Linker_Section_Pragma field from Record_Field_Kind.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* gen_il-gen-gen_entities.adb: Remove Linker_Section_Pragma
field from Record_Field_Kind. Minor comment improvement.diff --git
The precondition of Cot in Ada.Numerics.Generic_Elementary_Functions is
not meant for execution (as enforced by the Assertion_Policy at the top
of the file) but for analysis only. A conjunct in the precondition of
Cot applied to two arguments (with a Cycle value) was incorrect, now
fixed.
Tested
Avoid local string copy with renaming, which both make the code shorter
to read and should be marginally faster to execute. Code cleanup only
related to loading of compilation units; semantics is unaffected.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* par.adb (Par): A
There are three kinds of units that require code generation: the main
unit, its corresponding spec and generic instances needed by the main
unit.
Previously the main unit and its corresponding spec were flagged as
requiring code generation just before calling the backend, while
instance units
This adds socket options that are needed to control the keepalive status
of TCP connections. The new options are Keep_Alive_Count,
Keep_Alive_Idle, and Keep_Alive_Interval.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* libgnat/g-socket.ads (Option_Name): Add
When discovering violation of the No_Elaboration_Code restriction it is
enough to find one unit that violates it; we don't need to examine all
units.
Code cleanup; semantics is unaffected.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* lib-writ.adb (Write_ALI): Exit from
Code cleanup; semantics is unaffected.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* inline.adb (Instantiate_Bodies): Fix white in declaration.
(Remove_Dead_Instance): Change iteration from WHILE to FOR.diff --git a/gcc/ada/inline.adb b/gcc/ada/inline.adb
---
Use Tbuild.Unchecked_Convert_To instead of
Nmake.Make_Unchecked_Type_Conversion. This leads to more readable source
code in the compiler, and also more readable .dg code, because it
removes redundant unchecked conversions.
There is only one remaining call to Make_Unchecked_Type_Conversion,
which
The Terminated flag for a task was being queried without first aquiring
the task lock. It is not clear that this unsychronized access has ever
caused a problem in practice, but the thread-sanitizer tool flags it.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
*
On Wed, Jul 7, 2021 at 4:15 PM Richard Biener
wrote:
>
> On Wed, Jul 7, 2021 at 4:40 AM Hongtao Liu via Gcc-patches
> wrote:
> >
> > On Tue, Jul 6, 2021 at 9:37 AM Hongtao Liu wrote:
> > >
> > > On Tue, Jul 6, 2021 at 7:31 AM Segher Boessenkool
> > > wrote:
> > > >
> > > > Hi!
> > > >
> > > >
Hi!
On Wed, Jul 07, 2021 at 05:02:42PM +0800, Kewen.Lin wrote:
> This patch is to add test cases to check if vectorizer
> can exploit vector multiply instrutions on Power, some
> of them are supported since Power8, the other are newly
> introduced by Power10.
Okay for trunk. Thank you!
Segher
On Wed, Jul 7, 2021 at 10:54 PM Segher Boessenkool
wrote:
>
> Hi!
>
> On Wed, Jul 07, 2021 at 10:15:08AM +0200, Richard Biener wrote:
> > On Wed, Jul 7, 2021 at 4:40 AM Hongtao Liu via Gcc-patches
> > wrote:
> > > On Tue, Jul 6, 2021 at 9:37 AM Hongtao Liu wrote:
> > > > On Tue, Jul 6, 2021 at
r12-1989 fixed the testcase in the PR, but unfortunately the fix is
buggy:
1. It breaks the case where the common template between the
TEMPLATE_DECL t and ctx_parms is the innermost template (as in
concepts-memtmpl5.C below). This can be fixed by instead
passing the TREE_TYPE of
On Wed, Jul 7, 2021 at 10:54 PM Segher Boessenkool
wrote:
>
> Hi!
>
> On Wed, Jul 07, 2021 at 10:15:08AM +0200, Richard Biener wrote:
> > On Wed, Jul 7, 2021 at 4:40 AM Hongtao Liu via Gcc-patches
> > wrote:
> > > On Tue, Jul 6, 2021 at 9:37 AM Hongtao Liu wrote:
> > > > On Tue, Jul 6, 2021 at
From: liuhongt
1. Add __extendhfdf2/__extendhfsf2 to return an IEEE half converted to
IEEE double/single.
2. Add __truncdfhf2/__extendsfhf2 to truncate IEEE double/single into
IEEE half.
3. Add __eqhf2/__nehf2 to return 0 if a == b and a,b are not NAN,
otherwise return 1.
These are needed by
On Mon, Jul 05, 2021 at 06:03:21PM -0600, Jeff Law wrote:
> It reminds me a bit of millicode calls on the PA or calls to special
> routines in libgcc. They're calls to functions, but those functions do
> not follow the standard ABI.
Something with CALL_INSN_FUNCTION_USAGE? And maybe some
Hi!
On Wed, Jul 07, 2021 at 10:15:08AM +0200, Richard Biener wrote:
> On Wed, Jul 7, 2021 at 4:40 AM Hongtao Liu via Gcc-patches
> wrote:
> > On Tue, Jul 6, 2021 at 9:37 AM Hongtao Liu wrote:
> > > On Tue, Jul 6, 2021 at 7:31 AM Segher Boessenkool
> > > wrote:
> > > > I ran into this in
Hi Nick,
Ping?
Oops.
PR libctf/27482
* libtool.m4 (LT_PATH_NM): Try BSDization flags with a user-provided
Changes to libtool need to be posted to the libtool project:
https://www.gnu.org/software/libtool/
They have mailing lists for bug reports and patch submissions.
On 7/7/21 1:28 AM, Richard Biener wrote:
On Tue, Jul 6, 2021 at 5:06 PM Martin Sebor wrote:
Ping: https://gcc.gnu.org/pipermail/gcc-patches/2021-June/573968.html
Any questions/suggestions on the final patch or is it okay to commit?
I don't remember seeing one (aka saying
Hi,
Version 2 of this patch adds more code generation tests to show the
benefit of this RTL simplification as well as adding a new helper function
'rtx_vec_series_p' to reduce code duplication.
Patch tested as version 1 - ok for master?
Thanks,
Jonathan
---
gcc/ChangeLog:
2021-06-08
On Tue, Jun 29, 2021 at 11:19 AM Xionghu Luo wrote:
>
>
>
> On 2021/6/28 16:25, Richard Biener wrote:
> > On Mon, Jun 28, 2021 at 10:07 AM Xionghu Luo wrote:
> >>
> >>
> >>
> >> On 2021/6/25 18:02, Richard Biener wrote:
> >>> On Fri, Jun 25, 2021 at 11:41 AM Xionghu Luo wrote:
>
>
>
Changes since v1:
* Update assumptions for niter, add more test cases check
* Use widest_int/wide_int instead mpz to do +-/
* Move some early check for quick return
For code like:
unsigned foo(unsigned val, unsigned start)
{
unsigned cnt = 0;
for (unsigned i = start; i > val; ++i)
cnt++;
ping?
On Wed, Jun 30, 2021 at 3:58 PM Christophe LYON via Gcc-patches <
gcc-patches@gcc.gnu.org> wrote:
> I've noticed that overriding cpu/arch flags when running the testsuite
> can cause this test to fail rather than being skipped because of
> incompatible flags combination.
>
> Since the test
This improves LIM by recording aggregate copies for disambiguation
purposes instead of as UNANALYZABLE_MEM which will prevent any
invariant or store motion across it. This allows four of the six
references in the loop of the testcase to be promoted.
Bootstrapped and tested on
This adds a testcase for an old fixed PR.
Tested on x86_64-unknwon-linux-gnu, pushed.
2021-07-07 Richard Biener
PR tree-optimization/34195
* gcc.dg/vect/pr34195.c: New testcase.
---
gcc/testsuite/gcc.dg/vect/pr34195.c | 33 +
1 file changed, 33
On Jul 07 2021, Christophe Lyon wrote:
> On Wed, Jul 7, 2021 at 11:38 AM Andreas Schwab
> wrote:
>
>> This broke bootstrap on aarch64.
>>
>>
> This is https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101363
Nope. This is a *bootstap* failure.
Andreas.
--
Andreas Schwab, sch...@linux-m68k.org
On Wed, Jul 7, 2021 at 11:38 AM Andreas Schwab
wrote:
> This broke bootstrap on aarch64.
>
>
This is https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101363
Christophe
> Andreas.
>
> --
> Andreas Schwab, sch...@linux-m68k.org
> GPG Key fingerprint = 7578 EB47 D4E5 4D69 2510 2552 DF73 E780 A9DA
This broke bootstrap on aarch64.
Andreas.
--
Andreas Schwab, sch...@linux-m68k.org
GPG Key fingerprint = 7578 EB47 D4E5 4D69 2510 2552 DF73 E780 A9DA AEC1
"And now for something completely different."
Hi,
This patch is to make Power10 newly introduced vector
modulo instructions exploited in vectorized loops, it
just simply renames existing define_insns as standard
pattern names.
Is it ok for trunk?
BR,
Kewen
-
gcc/ChangeLog:
* config/rs6000/rs6000-builtin.def (MODS_V2DI,
Hi,
This patch is to add one test case to check if vectorizer
can exploit vector division instrutions newly introduced
by Power10.
Is it ok for trunk?
BR,
Kewen
-
gcc/testsuite/ChangeLog:
* gcc.target/powerpc/div-vectorize-1.c: New test.
---
Hi,
This patch is to add test cases to check if vectorizer
can exploit vector multiply instrutions on Power, some
of them are supported since Power8, the other are newly
introduced by Power10.
Is it ok for trunk?
BR,
Kewen
-
gcc/testsuite/ChangeLog:
*
On Tuesday, 22 June 2021 21:52:16 CEST Jason Merrill wrote:
> > 2. About the namespace aliases: IIUC an attribute would currently be
> > rejected because of the C++ grammar. Do you want to make it valid before
> > WG21 officially decides how to proceed? And if you have a pointer for me
> > where
On Wed, Jul 7, 2021 at 4:40 AM Hongtao Liu via Gcc-patches
wrote:
>
> On Tue, Jul 6, 2021 at 9:37 AM Hongtao Liu wrote:
> >
> > On Tue, Jul 6, 2021 at 7:31 AM Segher Boessenkool
> > wrote:
> > >
> > > Hi!
> > >
> > > I ran into this in shrink-wrap.c today.
> > >
> > > On Thu, Jun 03, 2021 at
On Wed, 7 Jul 2021, Hongtao Liu wrote:
> > > > > and I have no easy way to test things there. Handling AVX512
> > > > > should be easy as followup though.
>
> Here's the patch adding avx512f tests for FMADDSUB/FMSUBADD slp patterns.
> Pushed to the trunk.
Thanks!
Richard.
On Tue, Jul 6, 2021 at 9:01 PM Roger Sayle wrote:
>
>
> All of the optimizations/transformations mentioned in bugzilla for
> PR tree-optimization/40210 are already implemented in mainline GCC,
> with one exception. In comment #5, there's a suggestion that
> (bswap64(x)>>56)&0xff can be
OK? (I want to use the macros in libstdc++.)
On Wednesday, 30 June 2021 10:59:28 CEST Matthias Kretz wrote:
> Library code, especially in headers, sometimes needs to know how the
> compiler interprets / optimizes floating-point types and operations.
> This information can be used for additional
On Tue, Jul 6, 2021 at 5:47 PM Martin Sebor via Gcc-patches
wrote:
>
> Ping: https://gcc.gnu.org/pipermail/gcc-patches/2021-June/573349.html
+ if (TREE_CODE (axstype) != UNION_TYPE)
what about QUAL_UNION_TYPE? (why constrain union type accesses
here - note you don't seem to constrain accesses
On Tue, Jul 6, 2021 at 5:06 PM Martin Sebor wrote:
>
> Ping: https://gcc.gnu.org/pipermail/gcc-patches/2021-June/573968.html
>
> Any questions/suggestions on the final patch or is it okay to commit?
I don't remember seeing one (aka saying "bootstrapped/tested, OK to commit?"
or so) - and the
> > > > and I have no easy way to test things there. Handling AVX512
> > > > should be easy as followup though.
Here's the patch adding avx512f tests for FMADDSUB/FMSUBADD slp patterns.
Pushed to the trunk.
--
BR,
Hongtao
From 2dc666974cca3a62686f4d7135ca36c25d61a802 Mon Sep 17 00:00:00 2001
94 matches
Mail list logo
