Hi!
The following testcase is miscompiled on x86_64-linux, the bitfield store
is implemented as a RMW 64-bit operation at d+24 when the d variable has
size of only 28 bytes and scheduling moves in between the R and W part
a store to a different variable that happens to be right after the d
variabl
This makes us pass down the vector type for the two-operand
SLP node build rather than picking that from operand one which,
when constant or external, could be NULL.
Bootstrapped and tested on x86_64-unknown-linux-gnu, pushed.
2021-06-16 Richard Biener
PR tree-optimization/101083
Hi!
As the following testcase shows, libffi didn't handle properly
classify_arguments of structures at byte offsets not divisible by
UNITS_PER_WORD. The following patch adjusts it to match what
config/i386/ classify_argument does for that and also ports the
PR38781 fix there (the second chunk).
On Wed, Jun 16, 2021 at 9:49 AM Jakub Jelinek wrote:
>
> Hi!
>
> As the following testcase shows, libffi didn't handle properly
> classify_arguments of structures at byte offsets not divisible by
> UNITS_PER_WORD. The following patch adjusts it to match what
> config/i386/ classify_argument does
This removes obsolete stuff.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* libgnat/s-except.ads (ZCX_By_Default): Delete.
(Require_Body): Likewise.
* libgnat/s-except.adb: Replace body with pragma No_Body.diff --git a/gcc/ada/libgnat/s-except.adb b/gcc/ada/
The placement of check for types of an unchecked conversion being
generic was done in the middle of switching from their original to
underlying views, which was inconsistent and made the surrounding
comments confusing.
Code cleanup in preparation for a follow-up improvement; behaviour is
unaffecte
During the implementation of AI12-0172, we incorrectly allowed illegal
cases of unconstrained components.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* sem_ch3.adb (Analyze_Component_Declaration): Do not special
case raise expressions.
gcc/testsuite/
* gn
This removes obsolete stuff.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* initialize.c: Do not include vxWorks.h and fcntl.h from here.
(__gnat_initialize) [__MINGW32__]: Remove #ifdef and attribute
(__gnat_initialize) [init_float]: Delete.
(__gnat
This new version is both more straightforward to understand and to map
to the generated code at high optimization. No functional changes.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* sem_util.adb (Current_Entity_In_Scope): Reimplement.diff --git a/gcc/ada/sem_util.adb b/
For an assignment statement of the form "A.B(C).D := ...", in a loop,
the index check on C can be missing.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* ghost.adb: Add another special case where full analysis is
needed. This bug is due to quirks in the way
Completing previous patch since it introduced a regression on ACATS
c611a03 under certified runtime.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* exp_disp.adb (Build_Class_Wide_Check): Ensure that evaluation
of actuals is side effects free (since the check duplica
There is no need to walk the hierarchy rooted at a type that does not
come from source only to drop the result on the floor. No functional
changes.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* freeze.adb (Check_No_Parts_Violations): Return earlier if the
type is
The value returned by Incomplete_Or_Partial_View depends on the current
scope, which is unexpected at best.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* sem_util.adb (Incomplete_Or_Partial_View): Retrieve the scope of
the parameter and use it to find its incomplet
This patch fixes an issue in the compiler whereby a non-static
expression is allowed as an argument to aspect Interrupt_Priority
despite restriction Static_Priorities being in effect.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* sem_ch13.adb (Make_Aitem_Pragma): Check for
As First_Sloc is used to compute source location for entry guards, it
must be also used to compute the dominance marker source location in
this particular case (when a statement is dominated by an entry guard).
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* par_sco.adb (Set
This patch fixes an issue in the compiler whereby taking 'Access of a
"for of" loop parameter caused spurious compile-time accessibility
errors.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* sem_util.adb (Accessibility_Level): Take into account
renamings of loop pa
This patch fixes an issue in the compiler whereby the mixing of
positional and named entries in an enumeration representation clause was
erroneously allowed instead of rejected - as per RM rules.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* sem_ch13.adb (Analyze_Enumerati
Code cleanup; behaviour is unaffected. The extra initialization most
likely come from a initial version of the patch, before the limit of
parameters to a subsequent call to New_List was discovered.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* exp_ch3.adb (Build_Slice_Assi
Routine Is_Actual_Parameter now also detects entry parameters; this
change is harmless for GNAT and allows this routine to be reused in
GNATprove.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* sem_util.ads (Is_Actual_Parameter): Update comment.
* sem_util.adb (Is_A
The expanded code should never reference entities in the tasking runtime
(libgnarl) except when expanding tasking constructs directly.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* rtsfind.ads, libgnarl/s-taskin.ads, exp_ch3.adb, exp_ch4.adb,
exp_ch6.adb, exp_ch9.a
This patch implements the legality rules specified in RM 11.3 (2) for
the placement of Raise_Expressions. In Ada2020 aspect specifications can
appear in several new declarative contexts, and the keyword "with" can
be part of a Raise_Expression or the start of a list of aspect
specifications. To pr
The aliasing check applies when some of the formals has their passing
mechanism unspecified; RM 6.2 (12/3). Previously it only applied when
the first formal had its passing mechanism unspecified and the second
had its passing mechanism either unspecified or by-reference.
Tested on x86_64-pc-linux-
For each instance of implementation advice given in the Ada RM, the
Implementation Advice section of the GNAT RM documents whether the
advice is followed (see Ada RM M.3(1)). Such documention was missing for
the implementation advice regarding Ada.Containers and its child units.
Rectify this omissi
As part of the work on changing side-effects removal in SPARK, a special
case was introduced to generate an Itype_Reference for Itypes in slices.
This was based on a misunderstanding of existing checks for bounds when
analyzing slices. These Itype_Reference are actually not needed to get
the corres
It works neither at compile time nor at run time because of the classical
issue that -Integer'First is not a valid Integer value.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* urealp.adb (Scale): Change first paramter to Uint and adjust.
(Equivalent_Decimal_Exponen
Detection of volatile expressions, i.e. references to volatile objects
and allocators, is done in two steps: first when analysing entity names
and allocators themselves (except when they occur within actual
parameters of subprogram calls) and then after the subprogram call has
been resolved (so tha
Function declared immediately within a protected body is a not a
protected function; the exact definition is RM 9.5.1(1): "A protected
subprogram is a subprogram declared immediately within a protected
definition."
Consequently, functions declared immediately within a protected body are
not volati
When detecting references to volatile objects in expressions of the
expression functions we couldn't determine the enclosing function. This
was because we examined a copy of the expression made for preanalysis
and this copy is not properly decorated. Consequently, we wrongly
rejected valid referenc
Hi,
This patch adds tests to verify that Neon narrowing-shift instructions
clear the top half of the result vector. It is sufficient to show that a
subsequent combine with a zero-vector is optimized away - leaving
just the narrowing-shift instruction.
Ok for master?
Thanks,
Jonathan
---
gcc/te
Hi,
Version 2 of this patch adds tests to verify the benefit of this change.
Ok for master?
Thanks,
Jonathan
---
gcc/ChangeLog:
2021-06-11 Jonathan Wright
* config/aarch64/aarch64-simd.md (aarch64_xtn_insn_le):
Define - modeling zero-high-half semantics.
(aarch64_x
Hi Richi,
On Tue, Jun 15 2021, Richard Biener wrote:
> On June 15, 2021 5:09:40 PM GMT+02:00, Martin Jambor wrote:
>>Hi,
>>
>>When SRA transforms an assignment where the RHS is an aggregate decl
>>that it creates replacements for, the (least efficient) fallback method
>>of dealing with them is to
Hi,
Version 2 of the patch adds tests to verify the benefit of this change.
Ok for master?
Thanks,
Jonathan
---
gcc/ChangeLog:
2021-06-14 Jonathan Wright
* config/aarch64/aarch64-simd-builtins.def: Split generator
for aarch64_sqmovun builtins into scalar and vector variants
Hi,
Version 2 of the patch adds tests to verify the benefit of this change.
Ok for master?
Thanks,
Jonathan
---
gcc/ChangeLog:
2021-06-14 Jonathan Wright
* config/aarch64/aarch64-simd-builtins.def: Split generator
for aarch64_qmovn builtins into scalar and vector
v
Hi,
Version 2 of this patch adds tests to verify the benefit of this change.
Ok for master?
Thanks,
Jonathan
---
gcc/ChangeLog:
2021-06-14 Jonathan Wright
* config/aarch64/aarch64-simd.md (aarch64_hn):
Change to an expander that emits the correct instruction
depend
Changed the variable "gcc_cv_as_global_load_fixed" into
"gcc_cv_as_gcn_global_load_fixed" in order to have the "gcn" substring
also in the config.patch file.
Am 09.06.2021 um 16:47 schrieb Marcel Vollweiler:
This patch fixes an issue with global_load assembler functions leading
to a "invalid op
> On 15 Jun 2021, at 00:07, Andrew MacLeod via Gcc-patches
> wrote:
>
> As mentioned in the Text from the PR:
>
> "When a range is being calculated for an ssa-name, the propagation process
> often goes along back edges. These back edges sometime require other
> ssa-names which have not be p
On Wed, 16 Jun 2021, Jakub Jelinek wrote:
> Hi!
>
> The following testcase is miscompiled on x86_64-linux, the bitfield store
> is implemented as a RMW 64-bit operation at d+24 when the d variable has
> size of only 28 bytes and scheduling moves in between the R and W part
> a store to a differen
On Wed, 16 Jun 2021, Martin Jambor wrote:
> Hi Richi,
>
> On Tue, Jun 15 2021, Richard Biener wrote:
> > On June 15, 2021 5:09:40 PM GMT+02:00, Martin Jambor
> > wrote:
> >>Hi,
> >>
> >>When SRA transforms an assignment where the RHS is an aggregate decl
> >>that it creates replacements for, th
On Wed, Jun 16, 2021 at 5:18 AM Trevor Saunders wrote:
>
> - Unfortunately using_auto_storage () needs to handle m_vec being null.
> - Handle self move of an auto_vec to itself.
> - Make sure auto_vec defines the classes move constructor and assignment
> operator, as well as ones taking vec, so
When we face a sm_ord vs sm_unord for the same ref during
store sequence merging we assert that the ref is already marked
unsupported. But it can be that it will only be marked so
during the ongoing merging so instead of asserting mark it here.
Also apply some optimization to not waste resources
On Mon, 14 Jun 2021 at 16:15, Kyrylo Tkachov wrote:
>
>
>
> > -Original Message-
> > From: Prathamesh Kulkarni
> > Sent: 14 June 2021 08:58
> > To: gcc Patches ; Kyrylo Tkachov
> >
> > Subject: Re: [ARM] PR97906 - Missed lowering abs(a) >= abs(b) to vacge
> >
> > On Mon, 7 Jun 2021 at 12
On 14/06/2021 11:57, Richard Biener wrote:
On Mon, 14 Jun 2021, Richard Biener wrote:
Indeed. For example a simple
int a[1024], b[1024], c[1024];
void foo(int n)
{
for (int i = 0; i < n; ++i)
a[i+1] += c[i+i] ? b[i+1] : 0;
}
should usually see peeling for alignment (though on x86 yo
On Wed, 16 Jun 2021, Andre Vieira (lists) wrote:
>
> On 14/06/2021 11:57, Richard Biener wrote:
> > On Mon, 14 Jun 2021, Richard Biener wrote:
> >
> >> Indeed. For example a simple
> >> int a[1024], b[1024], c[1024];
> >>
> >> void foo(int n)
> >> {
> >>for (int i = 0; i < n; ++i)
> >> a
Hi!
I've noticed this test now on various arches sometimes FAILs, sometimes
PASSes (the line 12 test in particular).
The problem is that a = 0; initialization in the caller no longer happens
before the f(&a) call as what the argument points to is only used in
debug info.
Making the function noip
On Tue, Jun 15, 2021 at 06:11:27PM +0200, Richard Biener wrote:
> >--- a/gcc/tree-sra.c
> >+++ b/gcc/tree-sra.c
> >@@ -915,6 +915,12 @@ create_access (tree expr, gimple *stmt, bool
> >write)
> >if (!DECL_P (base) || !bitmap_bit_p (candidate_bitmap, DECL_UID
> >(base)))
> > return NULL;
> >
> >
Jonathan Wright writes:
> Hi,
>
> Version 2 of this patch adds tests to verify the benefit of this change.
>
> Ok for master?
>
> Thanks,
> Jonathan
>
> ---
>
> gcc/ChangeLog:
>
> 2021-06-11 Jonathan Wright
>
> * config/aarch64/aarch64-simd.md (aarch64_xtn_insn_le):
> Define - m
Jonathan Wright writes:
> Hi,
>
> Version 2 of the patch adds tests to verify the benefit of this change.
>
> Ok for master?
>
> Thanks,
> Jonathan
>
> ---
> gcc/ChangeLog:
>
> 2021-06-14 Jonathan Wright
>
> * config/aarch64/aarch64-simd-builtins.def: Split generator
> for aarch
Jonathan Wright writes:
> Hi,
>
> Version 2 of the patch adds tests to verify the benefit of this change.
>
> Ok for master?
>
> Thanks,
> Jonathan
>
> ---
>
> gcc/ChangeLog:
>
> 2021-06-14 Jonathan Wright
>
> * config/aarch64/aarch64-simd-builtins.def: Split generator
> for aar
Jonathan Wright writes:
> Hi,
>
> Version 2 of this patch adds tests to verify the benefit of this change.
>
> Ok for master?
>
> Thanks,
> Jonathan
>
> ---
>
> gcc/ChangeLog:
>
> 2021-06-14 Jonathan Wright
>
> * config/aarch64/aarch64-simd.md (aarch64_hn):
> Change to an expand
On Tue, 2021-06-15 at 19:48 -0600, Martin Sebor wrote:
Thanks for writing the patch.
> While debugging locations I noticed the semi_embedded_vec template
> in line-map.h doesn't declare a copy ctor or copy assignment, but
> is being copied in a couple of places in the C++ parser (via
> gcc_rich_l
Jonathan Wright via Gcc-patches writes:
> Hi,
>
> This patch adds tests to verify that Neon narrowing-shift instructions
> clear the top half of the result vector. It is sufficient to show that a
> subsequent combine with a zero-vector is optimized away - leaving
> just the narrowing-shift instruc
This adds a simple reduction vectorization capability to the
non-loop vectorizer. Simple meaning it lacks any of the fancy
ways to generate the reduction epilogue but only supports
those we can handle via a direct internal function reducing
a vector to a scalar. One of the main reasons is to avoi
Richard Biener via Gcc-patches writes:
> On Tue, Jun 15, 2021 at 8:02 AM Trevor Saunders wrote:
>>
>> This makes it clear the caller owns the vector, and ensures it is cleaned up.
>>
>> Signed-off-by: Trevor Saunders
>>
>> bootstrapped and regtested on x86_64-linux-gnu, ok?
>
> OK.
>
> Btw, are
Richard Biener writes:
> This adds a simple reduction vectorization capability to the
> non-loop vectorizer. Simple meaning it lacks any of the fancy
> ways to generate the reduction epilogue but only supports
> those we can handle via a direct internal function reducing
> a vector to a scalar.
On Tue, 15 Jun 2021 at 22:33, Jonathan Wakely wrote:
>
> On Tue, 15 Jun 2021 at 21:32, Tim Song wrote:
> >
> > CPOs are specified as actual semiregular function objects that can be
> > copied and constructed freely, so it seems a bit hostile to make them
> > final/non-addressable? (It's debatable w
On Tue, 15 Jun 2021 at 19:28, Jonathan Wakely wrote:
>
> By changing __cust_access::__decay_copy from a function template to a
> function object we avoid ADL. That means it's fine to call it
> unqualified (the compiler won't waste time doing ADL in associated
> namespaces, and won't try to complete
On Wed, 16 Jun 2021, Richard Sandiford wrote:
> Richard Biener writes:
> > This adds a simple reduction vectorization capability to the
> > non-loop vectorizer. Simple meaning it lacks any of the fancy
> > ways to generate the reduction epilogue but only supports
> > those we can handle via a di
On Tue, 15 Jun 2021 at 19:32, Jonathan Wakely wrote:
>
> Signed-off-by: Jonathan Wakely
>
> libstdc++-v3/ChangeLog:
>
> * include/bits/ranges_util.h (view_interface): Add noexcept to
> empty, operator bool, data and size members.
> (subrange): Add noexcept to constructors.
On Tue, 15 Jun 2021 at 20:29, Patrick Palka via Libstdc++
wrote:
>
> On Tue, 15 Jun 2021, Patrick Palka wrote:
>
> > This force-enables perfect forwarding call wrapper semantics whenever
> > the extra arguments of a partially applied range adaptor aren't all
> > trivially copyable, so as to avoid
On Mon, 14 Jun 2021 at 19:14, Paul Smith via Libstdc++
wrote:
>
> std::tuple elements are retrieved via std::get<> (template) not
> [] (array); have the generated output string match this.
Both of your patches seem to be based on the idea that the output is
supposed to correspond to how you acces
In addition to V8QI permutations, several other missing permutations are
added for 64bit vector modes for TARGET_SSSE3 and TARGET_SSE4_1 targets.
2021-06-16 Uroš Bizjak
gcc/
PR target/89021
* config/i386/i386-expand.c (expand_vec_perm_2perm_pblendv):
Handle 64bit modes for TARGET_S
On Mon, Jun 14, 2021 at 7:26 PM Stefan Schulze Frielinghaus
wrote:
>
> On Thu, May 20, 2021 at 08:37:24PM +0200, Stefan Schulze Frielinghaus wrote:
> [...]
> > > but we won't ever arrive here because of the niters condition. But
> > > yes, doing the pattern matching in the innermost loop processi
On 6/16/2021 4:00 AM, Richard Biener wrote:
On Wed, 16 Jun 2021, Martin Jambor wrote:
Hi Richi,
On Tue, Jun 15 2021, Richard Biener wrote:
On June 15, 2021 5:09:40 PM GMT+02:00, Martin Jambor wrote:
Hi,
When SRA transforms an assignment where the RHS is an aggregate decl
that it creates
On 6/16/21 6:38 AM, David Malcolm wrote:
On Tue, 2021-06-15 at 19:48 -0600, Martin Sebor wrote:
Thanks for writing the patch.
While debugging locations I noticed the semi_embedded_vec template
in line-map.h doesn't declare a copy ctor or copy assignment, but
is being copied in a couple of plac
On 6/16/21 5:48 AM, Jakub Jelinek wrote:
On Tue, Jun 15, 2021 at 06:11:27PM +0200, Richard Biener wrote:
--- a/gcc/tree-sra.c
+++ b/gcc/tree-sra.c
@@ -915,6 +915,12 @@ create_access (tree expr, gimple *stmt, bool
write)
if (!DECL_P (base) || !bitmap_bit_p (candidate_bitmap, DECL_UID
(base)))
Hi, Richard,
On Jun 16, 2021, at 1:19 AM, Richard Biener
mailto:rguent...@suse.de>> wrote:
+/* Expand the IFN_DEFERRED_INIT function according to its second
argument. */
+static void
+expand_DEFERRED_INIT (internal_fn, gcall *stmt)
+{
+ tree var = gimple_call_lhs (stmt);
+ tree init = NULL_TR
On Wed, 2021-06-16 at 08:52 -0600, Martin Sebor wrote:
> On 6/16/21 6:38 AM, David Malcolm wrote:
> > On Tue, 2021-06-15 at 19:48 -0600, Martin Sebor wrote:
> >
> > Thanks for writing the patch.
> >
> > > While debugging locations I noticed the semi_embedded_vec template
> > > in line-map.h doesn
*PING*
Forwarded Message
Subject: [Patch, fortran] PR fortran/96724 - Bogus warnings with the
repeat intrinsic and the flag -Wconversion-extra
Date: Thu, 20 Aug 2020 16:52:10 +
From: José Rui Faustino de Sousa
To: fort...@gcc.gnu.org, gcc-patches@gcc.gnu.org
Hi all!
Pr
*PING*
Forwarded Message
Subject: [Patch, fortran] PR fortran/96870 - Class name on error message
Date: Mon, 31 Aug 2020 16:09:32 +
From: José Rui Faustino de Sousa
To: fort...@gcc.gnu.org, gcc-patches@gcc.gnu.org
Hi all!
Proposed patch to PR96870 - Class name on error m
The patch for 96391 changed linemap_compare_locations to give up on
comparing locations from macro expansions if we don't have column
information. But in this testcase, the BOILERPLATE macro is multiple lines
long, so we do want to compare locations within the macro. So this patch
moves the LINE_
Hi,
On Wed, Jun 16 2021, Jakub Jelinek wrote:
> On Tue, Jun 15, 2021 at 06:11:27PM +0200, Richard Biener wrote:
>> >--- a/gcc/tree-sra.c
>> >+++ b/gcc/tree-sra.c
>> >@@ -915,6 +915,12 @@ create_access (tree expr, gimple *stmt, bool
>> >write)
>> >if (!DECL_P (base) || !bitmap_bit_p (candidate_bitm
On 6/16/21 6:46 AM, Richard Sandiford via Gcc-patches wrote:
Richard Biener via Gcc-patches writes:
On Tue, Jun 15, 2021 at 8:02 AM Trevor Saunders wrote:
This makes it clear the caller owns the vector, and ensures it is cleaned up.
Signed-off-by: Trevor Saunders
bootstrapped and regteste
At the risk of overstepping my GCN backend review remit...
On Wed, 16 Jun 2021 11:34:53 +0200
Marcel Vollweiler wrote:
> index d9fc3c2..e179ce1 100644
> --- a/gcc/configure.ac
> +++ b/gcc/configure.ac
> @@ -5357,6 +5357,30 @@ case "$target" in
> ;;
> esac
>
> +# This tests if the assembl
On 6/16/21 5:41 AM, Maxim Kuvyrkov wrote:
On 15 Jun 2021, at 00:07, Andrew MacLeod via Gcc-patches
wrote:
As mentioned in the Text from the PR:
"When a range is being calculated for an ssa-name, the propagation process often goes
along back edges. These back edges sometime require other ssa-
On 6/16/21 9:06 AM, David Malcolm wrote:
On Wed, 2021-06-16 at 08:52 -0600, Martin Sebor wrote:
On 6/16/21 6:38 AM, David Malcolm wrote:
On Tue, 2021-06-15 at 19:48 -0600, Martin Sebor wrote:
Thanks for writing the patch.
While debugging locations I noticed the semi_embedded_vec template
in
On 6/16/21 12:11 PM, Martin Sebor wrote:
On 6/16/21 9:06 AM, David Malcolm wrote:
On Wed, 2021-06-16 at 08:52 -0600, Martin Sebor wrote:
On 6/16/21 6:38 AM, David Malcolm wrote:
On Tue, 2021-06-15 at 19:48 -0600, Martin Sebor wrote:
Thanks for writing the patch.
While debugging locations I
Hi All!
Proposed patch to:
PR100097 - Unlimited polymorphic pointers and allocatables have
incorrect rank
PR100098 - Polymorphic pointers and allocatables have incorrect rank
Patch tested only on x86_64-pc-linux-gnu.
Version 2 no longer re-initializes explicit initialized variables, which
a
On 6/16/21 4:13 AM, Richard Biener via Gcc-patches wrote:
On Wed, Jun 16, 2021 at 5:18 AM Trevor Saunders wrote:
- Unfortunately using_auto_storage () needs to handle m_vec being null.
- Handle self move of an auto_vec to itself.
- Make sure auto_vec defines the classes move constructor and as
On 6/16/21 5:41 AM, Maxim Kuvyrkov wrote:
+ m_new_value_p = state;
+ return ret;
}
// Dump the caches for basic block BB to file F.
Thanks,
--
Maxim Kuvyrkov
https://www.linaro.org
Let me know if the problem is resolved.
pushed as obvious.
Andrew
commit bdfc1207bd20cf1ad81fca12
On Wed, 16 Jun 2021, Julian Brown wrote:
> > +if test x$gcc_cv_as_gcn_global_load_fixed = xyes; then
> > + AC_DEFINE(HAVE_GCN_ASM_GLOBAL_LOAD_FIXED, 1, [Define if your
> > assembler has fixed global_load functions.])
> > +else
> > + AC_DEFINE(HAVE_GCN_ASM_GLOBAL_LOAD_FIXED, 0, [D
On 6/16/21 10:35 AM, Jason Merrill wrote:
On 6/16/21 12:11 PM, Martin Sebor wrote:
On 6/16/21 9:06 AM, David Malcolm wrote:
On Wed, 2021-06-16 at 08:52 -0600, Martin Sebor wrote:
On 6/16/21 6:38 AM, David Malcolm wrote:
On Tue, 2021-06-15 at 19:48 -0600, Martin Sebor wrote:
Thanks for writin
This is an 'interim' fix. For now it forces all waiting threads to wake
on _M_release(). This isn't exactly efficient but resolves the issue
in the immediate term.
libstdc++-v3/ChangeLog:
libstdc++/PR100806
* include/bits/semaphore_base.h (__atomic_semaphore::_M_release():
Hi Richard,
Thanks for the feedback. From what you said, I can think of two possible
solutions (though I'm not sure if either is feasible/fully correct):
Option 1: Have the new X * (Y / X) --> Y - (Y % X) optimization only run in
scenarios that don't interfere with the existing X - (X / Y) * Y
On Wed, 2021-06-16 at 11:21 -0600, Martin Sebor wrote:
> On 6/16/21 10:35 AM, Jason Merrill wrote:
> > On 6/16/21 12:11 PM, Martin Sebor wrote:
> > > On 6/16/21 9:06 AM, David Malcolm wrote:
> > > > On Wed, 2021-06-16 at 08:52 -0600, Martin Sebor wrote:
> > > > > On 6/16/21 6:38 AM, David Malcolm w
The add-logical and add-add fusion patterns all have constraint
alternatives "=0,1,&r,r" for the output (3). The inputs 0 and 1
are used in the first fusion instruction and then either may be
reused as a temp for the output of the first insn which is
input to the second. However, if input 2 is the
So, the major question now is:
Is one single repeatable pattern enough for pattern initialization for all
different types of auto variables?
If YES, then the implementation for pattern initialization will be much easier
and simpler
as you pointed out. And will save me a lot of pain to imp
Same as previous version except removing the copyright notice from the
test.
libstdc++-v3/ChangeLog:
libstdc++/PR100806
* include/bits/semaphore_base.h (__atomic_semaphore::_M_release():
Force _M_release() to wake all waiting threads.
* testsuite/30_threads/semaphor
On Wed, Jun 16, 2021 at 01:51:16PM -0500, Aaron Sawdey wrote:
> The add-logical and add-add fusion patterns all have constraint
> alternatives "=0,1,&r,r" for the output (3). The inputs 0 and 1
> are used in the first fusion instruction and then either may be
> reused as a temp for the output of th
After my patch for PR91706, or before that with the qualified call,
tsubst_baselink returned a BASELINK with BASELINK_BINFO indicating a base of
a still-dependent derived class. We need to look up the relevant base binfo
in the substituted class.
Tested x86_64-pc-linux-gnu, applying to trunk.
A user pointed out to me that the manual documented the wrong
defaults for the -Wmismatched-dealloc and -Wmismatched-new-delete
options. I pushed the correction in r12-1544 (also attached).
Martin
commit 139564821dd2e4a9cbb08677ff12cf291d7d0218
Author: Martin Sebor
Date: Wed Jun 16 16:49:56 2
This patch mostly affects an internal routine in the ranger_cache. When
doing the cache rework to also be a range_query object, I implemented
range_on_edge to always return a range, but only return TRUE when the
range was actually changed on the edge. This was a well intention but
misguided a
This change adds first degree recomputation to the gori engine.
"Exports" are any ssa_name which is in the definition chain of one of
the SSA_NAMES on the condition.
a_2 = b_3 +6
if (a_2 > 10 && a_2 <20) // a_2 has the range [11, 19]
The condition tells us that the range of a_2 is [1
Ping patch. In particular, we would like to get this into the GCC 11.2
backport as it is power10 enablement.
| Date: Tue, 8 Jun 2021 20:21:25 -0400
| Subject: [PATCH 1/3] Add IEEE 128-bit min/max support on PowerPC.
| Message-ID: <20210609002125.ga18...@ibm-toto.the-meissners.org>
--
Michael Me
Ping patch. In particular, we like to get this into GCC 11.2 as part of
power10 enablement.
| Date: Tue, 8 Jun 2021 20:22:40 -0400
| Subject: [PATCH 2/3] Fix IEEE 128-bit min/max test.
| Message-ID: <20210609002240.gb18...@ibm-toto.the-meissners.org>
--
Michael Meissner, IBM
IBM, M/S 2506R, 550
Ping patch. In particular, we would like to get this patch into GCC 11.2 for
power10 enablement.
| Date: Tue, 8 Jun 2021 20:24:47 -0400
| Subject: [PATCH 3/3] Add IEEE 128-bit fp conditional move on PowerPC.
| Message-ID: <20210609002447.gc18...@ibm-toto.the-meissners.org>
--
Michael Meissner,
Ping patch. In particular, we would like to get this to GCC 11.2 because it is
power10 enablement.
| Date: Fri, 4 Jun 2021 11:10:37 -0400
| Subject: Generate 128-bit divide/modulus
| Message-ID: <20210604151037.ga27...@ibm-toto.the-meissners.org>
--
Michael Meissner, IBM
IBM, M/S 2506R, 550 Kin
The commit from this patch
(https://gcc.gnu.org/git/?p=gcc.git;a=commit;h=23eb66d1d46a34cb28c4acbdf8a1deb80a7c5a05)
changed the semantics of gcov_read_string and gcov_write_string. Before this
change the string storage was as described in gcov-io.h:
"Strings are
padded with 1 to 4 NUL bytes,
I've been unable to build ada on my fedora box since:
commit abcf5174979bcb91ac4c921eaa19a5b37f231ae4
Author: Arnaud Charlet
Date: Wed Jan 13 08:49:15 2021 -0500
[Ada] Use runtime from base compiler during stage1
gcc/ada/
* Make-generated.in: Add rule to copy runtime fil
On 2021-06-15 12:57, guojiufu via Gcc-patches wrote:
On 2021-06-14 17:16, Jan Hubicka wrote:
On 5/6/2021 8:36 PM, guojiufu via Gcc-patches wrote:
> Gentle ping.
>
> Original message:
> https://gcc.gnu.org/pipermail/gcc-patches/2020-October/555871.html
I think you need a more aggressive ping :
The case of an initializer with side effects for a zero-length array seems
extremely unlikely, but we should still return the right type in that case.
Tested x86_64-pc-linux-gnu, applying to trunk.
PR c++/101029
gcc/cp/ChangeLog:
* init.c (build_vec_init): Preserve the type of b
1 - 100 of 107 matches
Mail list logo
