On Fri, Oct 8, 2021 at 1:31 PM Stefan Schulze Frielinghaus via
Gcc-patches wrote:
>
> gcc/ChangeLog:
>
> * regcprop.c (maybe_mode_change): Determine offset relative to
> high or low part depending on endianness.
>
> Bootstrapped and regtested on IBM Z. Ok for mainline and gcc-{11,1
On 10/8/21 12:23, Martin Liška wrote:
Patch can bootstrap on x86_64-linux-gnu and survives regression tests.
I've spoken to Honza and he approves the patch.
I'm going to install it.
Martin
On Mon, Oct 11, 2021 at 8:26 AM liuhongt wrote:
>
> After providing expanders for reduc_umin/umax/smin/smax_scal_v4qi,
> perfomance are a little bit faster than before for reduce operations
> w/ options -O2 -march=haswell, -O2 -march=skylake-avx512
> and -Ofast -march=skylake-avx512.
>
> gcc/Cha
This patch contains two SUBREG-related optimization enabling tweaks to
the x86 backend.
The first change, to ix86_expand_vector_extract, cures the strange
-march=cascadelake related non-determinism that affected my new test
cases last week. Extracting a QImode or HImode element from an SSE
vecto
On Sat, Oct 09, 2021 at 12:55:15PM +0800, liuhongt via Gcc-patches wrote:
> Pushed to trunk.
>
> libgomp/ChangeLog:
>
> * testsuite/libgomp.c++/scan-10.C: Add option -fvect-cost-model=cheap.
> * testsuite/libgomp.c++/scan-11.C: Ditto.
> * testsuite/libgomp.c++/scan-12.C: Ditto.
Hi,
The variable omp_atv_sequential was replaced by omp_atv_serialized in
OpenMP 5.1. This was already implemented by Jakub (C/C++, commit
ea82325afec) and Tobias (Fortran, commit fff15bad1ab).
This patch adds two tests to check if omp_atv_serialized is available
(one test for C/C++ and one for
On Mon, Oct 11, 2021 at 11:40:54AM +0200, Marcel Vollweiler wrote:
> libgomp: Add tests for omp_atv_serialized and deprecate omp_atv_sequential.
>
> The variable omp_atv_sequential was replaced by omp_atv_serialized in OpenMP
> 5.1. This was already implemented by Jakub (C/C++, commit ea82325afec)
Hi!
OpenMP 5.1 adds env vars and functions to set and query new ICVs used
as fallback if thread_limit or num_teams clauses aren't specified on
teams construct.
The following patch implements those, though further work will be needed:
1) OpenMP 5.1 also changed the num_teams clause, so that it can
MIPS release 6 requires the lw/ld/sw/sd can work with
unaligned address, while it can be implemented by
full hardware or trap&emulate.
Since it doesn't have to be fully done by hardware, we add a
pair of options -m(no-)unaligned-access. Kernels may need them.
gcc/ChangeLog:
* config/mips
The patch is about using OPTION_SET_P instead of a default
option value set in common.opt.
Patch can bootstrap on x86_64-linux-gnu and survives regression tests.
Ready to be installed?
Thanks,
Martin
gcc/ChangeLog:
* common.opt: Remove Init(2) for some options.
* toplev.c (proc
Similar patch, let's rely on OPTION_SET_P and not a default
options value.
Patch can bootstrap on x86_64-linux-gnu and survives regression tests.
Ready to be installed?
Thanks,
Martin
gcc/ChangeLog:
* common.opt: Remove usage of IRA_REGION_AUTODETECT.
* flag-types.h (enum ira_r
On Fri, 8 Oct 2021 at 21:19, Richard Sandiford
wrote:
>
> Thanks for looking at this.
>
> Prathamesh Kulkarni writes:
> > Hi,
> > As mentioned in PR, for the following test-case:
> >
> > typedef unsigned char uint8_t;
> >
> > static inline uint8_t
> > x264_clip_uint8(uint8_t x)
> > {
> > uint8_
After the recent change in Optimize attribute handling, we need
finish_option function properly auto-detecting variable tracking options.
Patch can bootstrap on x86_64-linux-gnu and survives regression tests.
Ready to be installed?
Thanks,
Martin
PR debug/102585
gcc/ChangeLog:
On 10/4/21 23:02, Andrew Pinski wrote:
It might be useful to skip tabs for the same reason as spaces really.
Sure, be my guest.
MartinFrom b66d7be2c1b3ac286257e3df4d9796e391751bef Mon Sep 17 00:00:00 2001
From: Martin Liska
Date: Mon, 4 Oct 2021 14:06:14 +0200
Subject: [PATCH] target: support
Hi!
My recent push_local_extern_decl_alias change broke error-recovery,
do_pushdecl can return error_mark_node and set_decl_tls_model can't be
called on that. There are other code paths that store error_mark_node
into DECL_LOCAL_DECL_ALIAS, with the intent to differentiate the cases
where we have
On Mon, Oct 11, 2021 at 4:55 PM Roger Sayle wrote:
>
>
> This patch contains two SUBREG-related optimization enabling tweaks to
> the x86 backend.
>
> The first change, to ix86_expand_vector_extract, cures the strange
> -march=cascadelake related non-determinism that affected my new test
> cases l
Hi Jakub,
Am 11.10.2021 um 11:49 schrieb Jakub Jelinek:
On Mon, Oct 11, 2021 at 11:40:54AM +0200, Marcel Vollweiler wrote:
libgomp: Add tests for omp_atv_serialized and deprecate omp_atv_sequential.
The variable omp_atv_sequential was replaced by omp_atv_serialized in OpenMP
5.1. This was alre
On Mon, Oct 11, 2021 at 01:28:11PM +0200, Marcel Vollweiler wrote:
> > > libgomp/ChangeLog:
> > >
> > > * allocator.c (omp_init_allocator): Replace omp_atv_sequential with
> > > omp_atv_serialized.
> > > * omp.h.in: Add deprecated flag for omp_atv_sequential.
> > > * omp_lib.f9
Hi Harald,
On 10.10.21 21:27, Harald Anlauf via Fortran wrote:
just some random remarks from initially browsing your patch.
Thanks for browsing the patch :-)
- leftover from debugging?
Yes.
- code that could be shortened/made slightly more readable:
...
Is there a reason to not use strcmp (c
Hi all,
Here's a new version of the patch.
> >>> " If an exceptional condition occurs during the evaluation of an
> >>> expression
> >> (that is, if the result is not mathematically defined or not in the
> >> range of representable values for its type), the behavior is undefined."
> >>>
> >>> So
On Sun, Oct 10, 2021 at 10:19 PM liuhongt wrote:
>
> gcc/testsuite/ChangeLog:
>
> PR middle-end/102669
> * gnat.dg/unroll1.adb: Add -fno-tree-vectorize.
> ---
> gcc/testsuite/gnat.dg/unroll1.adb | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/gcc/testsuite
On Mon, Oct 11, 2021 at 09:38:36AM +0200, Richard Biener wrote:
> On Fri, Oct 8, 2021 at 1:31 PM Stefan Schulze Frielinghaus via
> Gcc-patches wrote:
> >
> > gcc/ChangeLog:
> >
> > * regcprop.c (maybe_mode_change): Determine offset relative to
> > high or low part depending on endi
On Mon, Oct 11, 2021 at 11:28:55AM +0200, Jakub Jelinek via Gcc-patches wrote:
> I don't think this is the right thing to do.
> This just means that at some point between 2013 when -fsimd-cost-model has
> been introduced and now -fsimd-cost-model= option at least partially stopped
> working properl
On Mon, 11 Oct 2021, Jakub Jelinek wrote:
> On Mon, Oct 11, 2021 at 11:28:55AM +0200, Jakub Jelinek via Gcc-patches wrote:
> > I don't think this is the right thing to do.
> > This just means that at some point between 2013 when -fsimd-cost-model has
> > been introduced and now -fsimd-cost-model=
On Monday, October 11, 2021, Jakub Jelinek via Gcc-patches <
gcc-patches@gcc.gnu.org> wrote:
> On Mon, Oct 11, 2021 at 11:28:55AM +0200, Jakub Jelinek via Gcc-patches
> wrote:
> > I don't think this is the right thing to do.
> > This just means that at some point between 2013 when -fsimd-cost-mode
On 08/10/2021 13:17, Tejas Belagod via Gcc-patches wrote:
Hi,
This patch adds the -march feature +pacbti to Armv8.1-M Mainline.
This feature enables pointer signing and authentication instructions
on M-class architectures.
Tested on arm-none-eabi. OK for trunk?
2021-10-04 Tejas Belagod
gcc
On 10/8/21 13:12, Jakub Jelinek wrote:
I think changing the default is ok, but dropping --full is not,
it should stay and behave the way it did before (i.e. print
r12-4245-gdb3d7270b42fe27fb05664c4fdf524ab7ad13a75
same thing as the new default except for full hash instead of
first 14 chars from i
On 28/09/2021 15:30, Christophe LYON via Gcc-patches wrote:
On 28/09/2021 13:14, Kyrylo Tkachov wrote:
-Original Message-
From: Gcc-patches On Behalf Of Christophe
Lyon via Gcc-patches
Sent: 07 September 2021 10:15
To: gcc-patches@gcc.gnu.org
Subject: [PATCH 03/13] arm: Add test for
On 28/09/2021 15:32, Christophe LYON via Gcc-patches wrote:
On 28/09/2021 13:18, Kyrylo Tkachov wrote:
Hi Christophe,
-Original Message-
From: Gcc-patches On Behalf Of Christophe
LYON via Gcc-patches
Sent: 08 September 2021 08:49
To: Richard Earnshaw ; gcc-
patc...@gcc.gnu.org
Subje
On Mon, Oct 11, 2021 at 12:58 PM Martin Liška wrote:
>
> The patch is about using OPTION_SET_P instead of a default
> option value set in common.opt.
>
> Patch can bootstrap on x86_64-linux-gnu and survives regression tests.
>
> Ready to be installed?
OK, but...
> Thanks,
> Martin
>
> gcc/Change
On 08/10/2021 13:17, Tejas Belagod via Gcc-patches wrote:
Hi,
Add -mbranch-protection option and its associated parsing routines.
This option enables the code-generation of pointer signing and
authentication instructions in function prologues and epilogues.
Tested on arm-none-eabi. OK for trunk
On Mon, Oct 11, 2021 at 12:58 PM Martin Liška wrote:
>
> Similar patch, let's rely on OPTION_SET_P and not a default
> options value.
>
> Patch can bootstrap on x86_64-linux-gnu and survives regression tests.
>
> Ready to be installed?
OK.
> Thanks,
> Martin
>
> gcc/ChangeLog:
>
> * comm
On Mon, Oct 11, 2021 at 1:02 PM Martin Liška wrote:
>
> After the recent change in Optimize attribute handling, we need
> finish_option function properly auto-detecting variable tracking options.
>
> Patch can bootstrap on x86_64-linux-gnu and survives regression tests.
>
> Ready to be installed?
Hello,
On Sat, 9 Oct 2021, apinski--- via Gcc-patches wrote:
> + (lshift (convert (convert:boolean_type_node @0)) { shift; })))
> +/* a ? -1 : 0 -> -a. No need to check the TYPE_PRECISION not being 1
> + here as the powerof2cst case above will handle that case correctly.
> */
W
Sorry for the very long delay in reviewing this. Things have been
a bit hectic recently.
Christophe Lyon via Gcc-patches writes:
> VPR_REG is the only register in its class, so it should be handled by
> TARGET_CLASS_LIKELY_SPILLED_P. No test fails without this patch, but
> it seems it should be
On 10/11/21 15:05, Richard Biener wrote:
+ if (!opts_set->x_flag_var_tracking)
+opts->x_flag_var_tracking = optimize >= 1;
That's still not equivalent to the old code for -fvar-tracking-uninit which
sets opts->x_flag_var_tracking to 1 and the old code checked that
for AUTOINIT_VALUE but you
On 08/10/2021 13:17, Tejas Belagod via Gcc-patches wrote:
Hi,
Add targeting-checking entities for PACBTI in testsuite
framework.
Tested on arm-none-eabi. OK for trunk?
2021-10-04 Tejas Belagod
gcc/ChangeLog:
* testsuite/lib/target-supports.exp
(check_effective_target_arm_p
On 11/10/2021 14:36, Richard Earnshaw via Gcc-patches wrote:
On 08/10/2021 13:17, Tejas Belagod via Gcc-patches wrote:
Hi,
Add targeting-checking entities for PACBTI in testsuite
framework.
Tested on arm-none-eabi. OK for trunk?
2021-10-04 Tejas Belagod
gcc/ChangeLog:
* testsuite/lib
Set the default size of time_t_bits to match the standard runtimes. The
size must match that which is used in the VSB, since the same VSB is
used to build all of the runtimes.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* libgnat/s-parame__ae653.ads (time_t_bits): Change t
It is simpler to access the type of first formal using semantic instead
of syntactic query. Behaviour is unaffected.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* sem_util.adb (Compile_Time_Constraint_Error): Simplify getting
the type of the first formal parameter.
When complaining about a compile-time constraint error within a default
initialization procedure we assumed that this procedure initializes a
record object. However, it can initialize an array object too.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* exp_util.adb (Inside_I
During Expand, prevent the clearing of the Analyzed flag if it has
already been set by Fold_Ureal.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* expander.adb (Expand): Skip clearing of Analyzed flag if
already set for N_Real_Literal.diff --git a/gcc/ada/expander.ad
Remove the RTEMS specific version of System.Stack_Checking.Operations as
the internal RTEMS API it uses can only detect stack overflow after the
event, whereas the stack checking emulation is meant to detect the stack
overflow before it occurs. Use the standard
System.Stack_Checking.Operations pack
Concurrent record types are either task record types or protected record
types. Now we detect them collectively (when looking for any of them) or
exclusively (when looking for one or the other).
Cleanup code related to fixes in expansion of boxes in record
aggregates. Behaviour is unaffected.
Tes
There is no need to explicitly guard against an empty list where the
subsequent iteration with First/Next/Present works fine.
Cleanup related to expansion of aggregates in GNATprove mode; behaviour
is unaffected.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* exp_aggr.adb
Rewriting of boxes in record aggregates into the corresponding default
values was done in resolution, where we special-cased access types and
scalar types with a Default_Value aspect.
However, this rewriting rather belong to expansion. Also, the
special-casing didn't take Normalize_Scalars nor Ini
When detecting record components with default initialization we did two
iteration over the component list; now we do only one.
Also, there was no need to explicitly guard against an empty list where
the subsequent iteration with First/Next/Present works fine.
Cleanup related to expansion of aggre
Replace IF with a single RETURN statement. Cleanup related to expansion
of aggregates in GNATprove mode; semantics is unaffected.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* exp_aggr.adb (Is_Delayed_Aggregate): Simplify.diff --git a/gcc/ada/exp_aggr.adb b/gcc/ada/exp_agg
When building a derived class-wide pre- or postcondition we are mapping
references to inherited formals and subprogram.
Originally we only did it for simple names; recently we fixed this
mapping to also work for operator symbols; with this patch we also do
this for extended names.
Tested on x86_6
Required by the style guide and by future changes in this function.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* sem_ch4.adb (Analyze_Membership_Op): Reorder subprogram spec
and bodies in alphabetical order.diff --git a/gcc/ada/sem_ch4.adb b/gcc/ada/sem_ch4.adb
--
The various __gl_XYZ binder globals prevent some link-time optimizations
when imported as mutable. Work around this by turning them into
constants.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* libgnarl/s-intman__android.adb, libgnarl/s-intman__lynxos.adb,
libgnarl
RTEMS supports attaching interrupt handlers to hardware interrupt
vectors, which is superior to the current approach of attaching handlers
to signals. Direct attachment of handlers removes the execution
overhead of converting hardware interrupts to signals and their
subsequent propagation to the i
This fixes a couple of long-standing oversights in the fixed-point multiply
implementation that were recently copied into the divide implementation and
thus made more visible: when computing the operand size for compile-time
known values, the negative case must be taken into account and comparisons
When resolving
type Color is (Blue, Orange);
function Get_Color return Color
is begin return Blue; end Get_Color;
function Get_Color return String
is begin return "Blue"; end Get_Color;
Test : Boolean := Get_Color in Blue;
we did not try all the possible interpretations of
The following avoids generating
(insn 6 5 7 2 (set (subreg:OI (concatn/v:TI [
(reg:DI 92 [ buffer ])
(reg:DI 93 [ buffer+8 ])
]) 0)
(subreg:OI (reg/v:V8SI 85 [ __x ]) 0)) "t.ii":76:21 74
{*movoi_internal_avx}
(nil))
via store_b
All these arguments were identified programmatically as being always
used with the same value (often the default one). As such, they can be
omitted.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* ali.adb (Get_Name): Ignore_Spaces is always False.
* bindo-graphs.adb
Use collective N_Generic_Declaration subtype instead of its members
N_Generic_Subprogram_Declaration and N_Generic_Package_Declaration where
reasonable. Code cleanup related to handling of Global contracts in
generic units; semantics is unaffected.
Tested on x86_64-pc-linux-gnu, committed on trunk
Use _POSIX_Threads_Minimum_stack_size instead of
ada_pthread_minimum_stack_size so the runtime does not require the RTEMS
kernel to be configured to have Ada support.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* libgnat/s-parame__rtems.adb: use
_POSIX_Threads_Mini
In determining at run time whether a statically-known discrete value
satisifies the predicate of a subtype where both
- a Dynamic_Predicate aspect specification applies (directly or
indirectly) to a subtype; and
- at least one other predicate aspect specification (that is, either
We already had a warning for unchecked conversions that involve the
private type Ada.Calendar.Time, whose representation might differ
between releases and targets of the compiler. Now this warning is
extended to Ada.Real_Time.Time and Ada.Real_Time.Time_Span, which is
similarly non-portable.
Previ
For users, 'Old attribute references are only allowed within
postcondition expressions. Internally, the FE may build trees that
transiently (before some subsequent transformation) violate these rules;
this is ok, but these violations were being incorrectly flagged in some
cases. Fix this problem.
Rewrite the former System.Threads implementation for AE653 to work on
the new Light runtime for VxWworks7r2Cert.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* libgnat/s-thread.ads: Fix comments. Remove unused package
imports.
(Thread_Body_Exception_Exit):
Routines Make_Predefined_Primitive_Specs and
Predefined_Primitive_Bodies, which create predefined primitives for
derived tagged types, are only called when restriction
No_Dispatching_Calls is inactive. There is no need to recheck this
restriction when creating individual primitive operations relate
The reason is that the automatic alignment promotion is not yet performed
in the case where the nominal subtype is of variable size.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* gcc-interface/decl.c (promote_object_alignment): Add GNU_SIZE
parameter and use it for
This prevents the warning from being given when there may still be regular
exception handlers in the code, although some of them have been turned into
local raises, by querying the predicate that determines whether such regular
handlers are removed or not in the front-end.
Tested on x86_64-pc-linu
This implements the support for most type-generic GCC atomic builtins.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* gcc-interface/gigi.h (resolve_atomic_size): Declare.
(list_third): New inline function.
* gcc-interface/decl.c (type_for_atomic_builtin_p):
This patch fixes the baseptrs-3.C test case introduced in the patch:
https://gcc.gnu.org/pipermail/gcc-patches/2021-October/580729.html
The problematic case concerns OpenMP mapping clauses containing struct
members of reference type, e.g. "mystruct.myref.myptr[:N]". To be able
to access the ar
I noticed that the test in question now compiles properly, and in fact
runs properly too. Thus it's more useful as a runtime test than a
passing compilation test that otherwise doesn't do much. This patch
moves it to libgomp.
Tested with offloading to NVPTX and bootstrapped. (I can probably
self
Christophe Lyon via Gcc-patches writes:
> The vmvnq_n* intrinsics and have [u]int[16|32]_t arguments, so use
> iterator instead of HI in mve_vmvnq_n_.
>
> 2021-09-03 Christophe Lyon
>
> gcc/
> * config/arm/mve.md (mve_vmvnq_n_): Use V_elem mode
> for operand 1.
>
> diff --git
On Mon, Oct 11, 2021 at 3:21 PM Martin Liška wrote:
>
> On 10/11/21 15:05, Richard Biener wrote:
> >> + if (!opts_set->x_flag_var_tracking)
> >> +opts->x_flag_var_tracking = optimize >= 1;
> > That's still not equivalent to the old code for -fvar-tracking-uninit which
> > sets opts->x_flag_va
On Fri, Oct 08, 2021 at 05:31:11PM -0500, Segher Boessenkool wrote:
> On Fri, Oct 08, 2021 at 02:27:28PM -0500, Paul A. Clarke wrote:
> > On Fri, Oct 08, 2021 at 12:39:15PM -0500, Segher Boessenkool wrote:
> > I see. Thanks for the reference. If I understand correctly, volatile
> > prevents some op
On 10/5/21 12:04, Richard Biener wrote:
On Mon, Oct 4, 2021 at 1:32 PM Martin Liška wrote:
On 10/4/21 13:16, Richard Biener wrote:
I meant in merge_one_data do not check ->stamp or ->checksum but instead rely
on the counter merging code to detect mismatches (there's read_mismatch and
read_err
Christophe Lyon via Gcc-patches writes:
> This patch implements support for vectors of booleans to support MVE
> predicates, instead of HImode. Since the ABI mandates pred16_t (aka
> uint16_t) to represent predicates in intrinsics prototypes, we
> introduce a new "predicate" type qualifier so tha
On 08/10/2021 13:18, Tejas Belagod via Gcc-patches wrote:
Hi,
This patch implements target feature macros when PACBTI is
enabled through the -march option or -mbranch-protection.
Tested on arm-none-eabi. OK for trunk?
2021-10-04 Tejas Belagod
gcc/ChangeLog:
* config/arm/arm-c.c (a
Christophe Lyon via Gcc-patches writes:
> We make use of qualifier_predicate to describe MVE builtins
> prototypes, restricting to auto-vectorizable vcmp* and vpsel builtins,
> as they are exercised by the tests added earlier in the series.
>
> Special handling is needed for mve_vpselq because it
On 10/11/21 15:49, Martin Liška wrote:
Let me update the documentation bits.
There's the updated patch.
May I install the patch now?
Thanks,
MartinFrom fdeb81a960faa19f75316e279a79c231da212f99 Mon Sep 17 00:00:00 2001
From: Martin Liska
Date: Thu, 9 Sep 2021 13:02:24 +0200
Subject: [PATCH] g
Christophe Lyon via Gcc-patches writes:
> From: Christophe Lyon
>
> The problem in this PR is that we call VPSEL with a mask of vector
> type instead of HImode. This happens because operand 3 in vcond_mask
> is the pre-computed vector comparison and has vector type.
>
> This patch fixes it by imp
Christophe Lyon via Gcc-patches writes:
> This is mostly a mechanical change, only tested by the intrinsics
> expansion tests.
>
> 2021-09-02 Christophe Lyon
>
> gcc/
> PR target/100757
> PR target/101325
> * config/arm/arm-builtins.c (BINOP_UNONE_NONE_NONE_QUALIFIERS):
The following fixes the issue of ignoring side-effects on memory
from overloaded global new/delete operators by not marking them
as effectively 'const' apart from other explicitely specified
side-effects.
This will cause
FAIL: g++.dg/warn/Warray-bounds-16.C -std=gnu++1? (test for excess errors)
Christophe Lyon via Gcc-patches writes:
> This patch covers a few builtins where we do not use the
> iterator and thus we cannot use .
>
> However this introduces a problem for the v2di instructions, because
> there is not predicate for this case. For instance, changing
> STRSBS_P_QUALIFIERS bre
Christophe Lyon via Gcc-patches writes:
> This patch covers a few non-load/store builtins where we do not use
> the iterator and thus we cannot use .
>
> We need to update the expected code in cde-mve-full-assembly.c because
> we now use mve_movv16qi instead of movhi to generate the vmsr
> instru
This avoids using an integer type for which we don't have an
approprate mode when expanding .DEFERRED_INIT to a non-memory
entity.
Bootstrapped on x86_64-unknown-linux-gnu, testing in progress.
Will push after it finished.
Richard.
2021-10-11 Richard Biener
PR middle-end/102683
*ping*
From: Paul Keir
Sent: 03 September 2021 11:31
To: gcc-patches@gcc.gnu.org
Cc: libstd...@gcc.gnu.org
Subject: Re: [PATCH] libstdc++: Fix compare_three_way for constexpr and Clang
*ping*
From: Paul Keir
Sent
Hi!
As reported, the alloc-9.c test and alloc-{1,2,3}.F* and alloc-11.f90
tests fail on powerpc64-linux with -m32.
The reason why it fails just there is that malloc doesn't guarantee there
128-bit alignment (historically glibc guaranteed 2 * sizeof (void *)
alignment from malloc).
There are two s
On Fri, 1 Oct 2021 10:09:05 -0700
Julian Brown wrote:
> libgomp/
> * testsuite/libgomp.c++/baseptrs-3.C: Add test (XFAILed for
> now).
This XFAILed test is addressed in the followup patch:
https://gcc.gnu.org/pipermail/gcc-patches/2021-October/581342.html
Cheers,
Julian
Richard Sandiford via Gcc-patches writes:
> Christophe Lyon via Gcc-patches writes:
>> This patch covers a few builtins where we do not use the
>> iterator and thus we cannot use .
>>
>> However this introduces a problem for the v2di instructions, because
>> there is not predicate for this case.
Prathamesh Kulkarni writes:
> On Fri, 8 Oct 2021 at 21:19, Richard Sandiford
> wrote:
>>
>> Thanks for looking at this.
>>
>> Prathamesh Kulkarni writes:
>> > Hi,
>> > As mentioned in PR, for the following test-case:
>> >
>> > typedef unsigned char uint8_t;
>> >
>> > static inline uint8_t
>> > x
Ping.
On Tue, Sep 28, 2021 at 04:20:46PM -0400, Marek Polacek wrote:
> On Thu, Sep 23, 2021 at 02:25:16PM -0400, Jason Merrill wrote:
> > On 9/20/21 18:59, Marek Polacek via Gcc-patches wrote:
> > > +void
> > > +handle_ignored_attributes_option (vec *v)
> > > +{
> > > + if (v == nullptr)
> > > +
Any thoughts?
On Thu, Sep 23, 2021 at 12:16:36PM -0400, Marek Polacek via Gcc-patches wrote:
> Clang implements something we don't have:
>
> __attribute__((deprecated("message", "replacement")));
>
> which seems pretty neat so I wrote this patch to add it to gcc.
>
> It doesn't allow the option
Hi!
On Mon, Oct 11, 2021 at 10:47:00AM +0800, Kewen.Lin wrote:
> As PR102658 shows, commit r12-4240 enables vectorization at O2,
> some cases need to be adjusted accordingly for rs6000 port.
>
> - For target specific test cases, this adds -fno-tree-vectorize
> to retain original test points, othe
Hi!
On Thu, Aug 26, 2021 at 09:19:30AM +0800, HAO CHEN GUI wrote:
> gcc/
> ?? * config/rs6000/rs6000-call.c (rs6000_gimple_fold_builtin):
> ?? Modify the VSX_BUILTIN_XVMINDP, ALTIVEC_BUILTIN_VMINFP,
> ?? VSX_BUILTIN_XVMAXDP, ALTIVEC_BUILTIN_VMAXFP expansions.
Something mangles your em
On Fri, Sep 17, 2021 at 10:08:27AM +0200, Richard Biener wrote:
> On Mon, Sep 13, 2021 at 4:53 PM Stefan Schulze Frielinghaus
> wrote:
> >
> > On Mon, Sep 06, 2021 at 11:56:21AM +0200, Richard Biener wrote:
> > > On Fri, Sep 3, 2021 at 10:01 AM Stefan Schulze Frielinghaus
> > > wrote:
> > > >
> >
On Mon, Oct 11, 2021 at 10:55:36AM -0500, Segher Boessenkool wrote:
> On Thu, Aug 26, 2021 at 09:19:30AM +0800, HAO CHEN GUI wrote:
> > gcc/
> > ?? * config/rs6000/rs6000-call.c (rs6000_gimple_fold_builtin):
> > ?? Modify the VSX_BUILTIN_XVMINDP, ALTIVEC_BUILTIN_VMINFP,
> > ?? VSX_BUILT
Diane Meirowitz via Gcc-patches writes:
> Please review my patch. It is tiny. Thank you.
Thanks for the patch and sorry for the very slow response.
I've now pushed this to master and all active branches.
Thanks,
Richard
> Diane
>
> On 9/15/21, 5:02 PM, "Diane Meirowitz" wrote:
>
>
> d
On 10/11/21 9:30 AM, Segher Boessenkool wrote:
Hi!
On Mon, Oct 11, 2021 at 10:47:00AM +0800, Kewen.Lin wrote:
As PR102658 shows, commit r12-4240 enables vectorization at O2,
some cases need to be adjusted accordingly for rs6000 port.
- For target specific test cases, this adds -fno-tree-vector
Tamar Christina writes:
> Hi,
>
> Sending a new version of the patch because I noticed the pattern was
> overriding the nor pattern.
>
> A second pattern is needed to capture the nor case as combine will match the
> longest sequence first. So without this pattern we end up de-optimizing nor
> an
Richard,
Thank you!
Diane
> On Oct 11, 2021, at 12:15 PM, Richard Sandiford
> wrote:
>
> Diane Meirowitz via Gcc-patches writes:
>> Please review my patch. It is tiny. Thank you.
>
> Thanks for the patch and sorry for the very slow response.
> I've now pushed this to master and all active b
On Mon, Oct 11, 2021 at 08:46:17AM -0500, Paul A. Clarke wrote:
> On Fri, Oct 08, 2021 at 05:31:11PM -0500, Segher Boessenkool wrote:
> > "asm volatile" means there is a side effect in the asm. This means that
> > it has to be executed on the real machine the same as on the abstract
> > machine, w
Tamar Christina writes:
>> > Note: This patch series is working incrementally towards generating the
>> most
>> > efficient code for this and other loops in small steps.
>>
>> It looks like this could be done in the vectoriser via an extension of the
>> scalar_cond_masked_set mechanism. We
This enables lightweight checks for the __glibcxx_requires_valid_range
and __glibcxx_requires_string_len macros when _GLIBCXX_ASSERTIONS is
defined. By using __builtin_object_size we can check whether the end of
the range is part of the same object as the start of the range, and
detect problems l
1 - 100 of 137 matches
Mail list logo
