Alan Modra via Gcc-patches wrote:
On 2021-05-04 8:42 a.m., Nick Clifton wrote:
Hi Guys,
On 4/30/21 7:36 PM, Simon Marchi wrote:
I think this fix is obvious enough, I encourage you to push it,
OK - I have pushed the patch to the mainline branches of both
the gcc and binutils-gdb repositorie
On Tue, May 4, 2021 at 7:43 PM Robin Dapp via Gcc-patches
wrote:
>
> Hi,
>
> on s390 we vectorize 4 statements using SLP. Add s390*-*-* to the
> appropriate dg-finals.
>
> Is that OK?
sure
> Regards
> Robin
>
> gcc/testsuite/ChangeLog:
>
> * gcc.dg/vect/slp-21.c: Add s390.
On Wed, May 05, 2021 at 08:05:29AM +0100, Iain Sandoe wrote:
> Alan Modra via Gcc-patches wrote:
>
> > On 2021-05-04 8:42 a.m., Nick Clifton wrote:
> > > Hi Guys,
> > >
> > > On 4/30/21 7:36 PM, Simon Marchi wrote:
> > > > I think this fix is obvious enough, I encourage you to push it,
> > >
>
Alan Modra wrote:
On Wed, May 05, 2021 at 08:05:29AM +0100, Iain Sandoe wrote:
Alan Modra via Gcc-patches wrote:
On 2021-05-04 8:42 a.m., Nick Clifton wrote:
Hi Guys,
On 4/30/21 7:36 PM, Simon Marchi wrote:
I think this fix is obvious enough, I encourage you to push it,
OK - I have pus
For move2add_valid_value_p we also have to ask the target whether a
register can be accessed in a different mode than it was set before.
gcc/ChangeLog:
PR rtl-optimization/100263
* postreload.c (move2add_valid_value_p): Ensure register can
change mode.
Bootstrapped and re
On Wed, May 5, 2021 at 1:32 AM Martin Sebor via Gcc-patches
wrote:
>
> With no optimization, -Wformat-overflow and -Wformat-truncation
> runs early to detect a subset of simple bugs. But as it turns out,
> the pass runs just a tad too early, before SSA. That causes it to
> miss a class of proble
On Wed, May 5, 2021 at 4:20 AM Martin Sebor via Gcc-patches
wrote:
>
> Even when explicitly enabled, -Walloca-larger-than doesn't run
> unless optimization is enabled as well. This prevents diagnosing
> alloca calls with constant arguments in excess of the limit that
> could otherwise be flagged
On Mon, May 3, 2021 at 11:31 AM Ivan Sorokin via Gcc-patches
wrote:
>
> Prior to this commit GCC -O2 generated quite bad code for this
> function:
>
> bool f()
> {
> return __builtin_cpu_supports("popcnt")
> && __builtin_cpu_supports("ssse3");
> }
>
> f:
> movl__cpu_model+1
Hi Stafford, All,
> Il giorno 5 mag 2021, alle ore 08:30, Stafford Horne ha
> scritto:
>
> On Mon, May 03, 2021 at 12:34:22PM +0200, Giulio Benetti wrote:
>> Hi Stafford, All,
>>
>> I've backported this patchset for Buildroot to versions:
>> - 9.3.0
>> - 10.3.0
>>
>> Does it make sense to se
This makes sure to follow SSA edges when folding eliminated stmts.
This reaps the same benefit as forwprop folding all stmts, not
waiting for one to produce copysign in the new testcase.
Bootstrapped on x86_64-unknown-linux-gnu, testing in progress.
2021-05-04 Richard Biener
PR tree-o
Committed as r12-477-gaf4e4d35f0b84d7c2f57a7b682a09116e9911142
Was missing in commit r12-438-g1580fc764423bf89e9b853aaa8c65999e37ccb8b
OpenMP: Support complex/float in && and || reduction
Tobias
-
Mentor Graphics (Deutschland) GmbH, Arnulfstrasse 201, 80634 München
Registergeric
In Eval_Arithmetic_Op we set Raises_Constraint_Error flag on integer
division by zero, but we didn't set it for real division by zero.
This only affects GNATprove, because GNAT sets this flag inside
Apply_Compile_Time_Constraint_Error (which has an early return guarded
by GNATprove_Mode). Possibly
Add legality checks for indexed aggregates with component associations.
RN 4.3.5 (28/5 - 30/5) specifies validity rules for indexed aggregates,
intended to simplify implementation as well as the computation of the
expected size of an indexed aggregate, such for an aggregaate of a
vector type.
Tes
This changes the implementation of output operations for floating-point
types from using Long_Long_Float for all floating-point types to using
a base type tailored to the type being operated on.
This comprises adjusting Ada.Text_IO.Float_IO and Ada.Text_IO.Complex_IO
to the new approach, as well a
In routine Apply_Compile_Time_Constraint_Error the Rep parameter was
disabling rewriting into N_Raise_Constraint_Error node. It is only
referenced when this routine is called from GNATprove, but in
GNATprove_Mode rewriting is disabled anyway.
It seems cleaner to remove this parameter and keep this
Code cleanup (and minuscule performance improvement); semantics is
unaffected.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* pprint.adb (Expr_Name): Introduce local constants to make the
code more readable and avoid repeated calls to Next to reach the
ELSE
Relational operators for universal integers come in three variants, e.g.:
function "=" (Left : Uint; Right : Uint) return Boolean renames UI_Eq;
function "=" (Left : Int; Right : Uint) return Boolean renames UI_Eq;
function "=" (Left : Uint; Right : Int) return Boolean renames UI_Eq;
p
This moves the declaration of the 4 tables to inside the function itself
for reasons explained in the code.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* exp_imgv.ads (Build_Enumeration_Image_Tables): Adjust comment.
* exp_imgv.adb (Build_Enumeration_Image_Tables):
This reduces the number of bootstrap dependencies for the compiler.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* debug.adb (d_x): Document new usage.
* exp_imgv.adb (Expand_User_Defined_Enumeration_Image): Add Typ
parameter and use it throughout the proces
In Ada2020 the expression for a postcondition can be a
Declare_Expression containing object declarations with defaults. These
defaults may generate local subtypes, but as for other default
expressions their full expansion must be deferred, typically to the
freeze point of the object being declared,
This speeds up the implementation of the Value attribute for enumeration
types with more than 3 values by using a perfect hash function to do the
lookup instead of a linear search. This also generizes the implementation
of the associated support routines in the runtime, as well as that of the
supp
GNAT currently emits style errors for declare expressions in conditions
even though their parens are required. This commit fixes that.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* par-ch5.adb (P_Condition): Check if expression is declare
expression.diff --git a/gc
When an access-to-subprogram type is qualified as not null, then the
compiler is likely to generate a more efficient code. This patch adds
such qualifiers to the internal runtime bodies, so no clients of those
runtime units are affected.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
This moves the inline expansion for user-defined enumeration types back
into the normal flow of control, moves the declarations of local objects
to where they are needed and removes an explicit check for private types
in the enumeration case, which is now superfluous.
No functional changes.
Teste
This avoids a call to the runtime and a string copy.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* debug.adb (d_x): Document extended usage.
* exp_imgv.adb (Expand_Standard_Boolean_Image): New procedure.
(Expand_Image_Attribute): Call it to expand in line t
This function really had no business being called during analysis.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* exp_ch7.adb (Expand_N_Package_Body): Add CUDA init call.
* sem_ch7.adb (Analyze_Package_Body_Helper): Remove CUDA init
call.diff --git a/gcc/ada
An assignment to an entity E triggers an improper warning that an
immediately preceding assignment to E is useless, when the right-hand
side of the new assignment includes a target name, that is to say the
Ada2020 use of "@" to designate the left-hand side of the assignment.
Tested on x86_64-pc-li
Unchecked union components were detected only when declared with
unqualified type names.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* exp_ch4.adb (Component_Is_Unconstrained_UU): Detect both
qualified and unqualified names of unchecked union components.diff --git
Fix s-os_lib.adb so vectorizing compilation works. Without this, the
compiler crashes while compiling s-os_lib.adb.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* libgnat/s-os_lib.adb (Missed_Drive_Letter): Simplify the code.diff --git a/gcc/ada/libgnat/s-os_lib.adb b/gcc/
When iterating over list elements with First/Next there is no need to
detect empty or non-existing list, because First intentionally returns
Empty in those cases.
Cleanup of code related to Unchecked_Unions; behaviour is unaffected.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
This prevents the generation of the new perfect hash function for the
Value attribute of local enumeration types when the target still uses
trampolines to implement pointers to local subprograms.
The generation of trampolines would be unexpected in these circumstances
and maybe even forbidden by t
Call to Scope returns an Entity_Id, not just Node_Id. This patch refines
the types in two occurrences of a call to Scope as an initial value.
Code cleanup; semantics is unaffected.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* exp_ch4.adb (User_Defined_Primitive_Equality_
The comment indicates that "The whole section including this comment
can be removed later ???", and 10 years qualifies as "later".
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* exp_util.adb (Is_Possibly_Unaligned_Object): Remove commented
code.diff --git a/gcc/ada/
This patch fixes an error in the compiler whereby dispatching calls to
subprograms featuring anonymous access formals may get incorrectly
expanded in such a way so as to lead to spurious runtime accessibility
check failures within the callee.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/
Routine Has_Unconstrained_UU_Component didn't traverse into nested
variant parts of a record type definition. Now it precisely follows the
Ada RM grammar to make sure that the traversal doesn't miss anything.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* exp_ch4.adb (Has_U
Expansion of attributes Input/Output for unchecked unions didn't work,
because it looked for a No_Elist where the AST has an empty list. Now we
look for a default value of a first discriminant (because the default
values must be given either for all or for none of the discrminants).
Also, it was w
Remove excessive defensive check from Has_Defaulted_Discriminants and
reuse it where possible. Cleanup only; semantics is unaffected.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* exp_attr.adb, exp_ch9.adb, sem_ch3.adb: Reuse
Has_Defaulted_Discriminants.
*
From: Florian Weimer
> Sent: 04 May 2021 20:46
>
> * Alejandro Colomar:
>
> > The thing is, in all of those threads, the only reasons to avoid
> > types in the kernel (at least, the only explicitly
> > mentioned ones) are (a bit simplified, but this is the general idea of
> > those threads):
> >
On Fri, 5 Feb 2021 at 15:42, Kyrylo Tkachov wrote:
>
> Hi Prathamesh,
>
> > -Original Message-
> > From: Prathamesh Kulkarni
> > Sent: 05 February 2021 09:53
> > To: gcc Patches ; Kyrylo Tkachov
> >
> > Subject: [PR97903][ARM] Missed optimization in lowering to vtst
> >
> > Hi,
> > For t
Hi!
If a header doesn't end with a new-line, with -fdirectives-only we right now
preprocess it as
int i = 1;# 2 "pr100392.c" 2
i.e. the line directive isn't on the next line, which means we fail to parse
it when compiling.
GCC 10 and earlier libcpp/directives-only.c had for this:
if (!pfile->st
Hi Prathamesh,
> -Original Message-
> From: Prathamesh Kulkarni
> Sent: 05 May 2021 09:35
> To: Kyrylo Tkachov
> Cc: gcc Patches
> Subject: Re: [PR97903][ARM] Missed optimization in lowering to vtst
>
> On Fri, 5 Feb 2021 at 15:42, Kyrylo Tkachov
> wrote:
> >
> > Hi Prathamesh,
> >
>
On Tue, Mar 16, 2021 at 6:13 PM Stefan Schulze Frielinghaus
wrote:
>
> [snip]
>
> Please find attached a new version of the patch. A major change compared to
> the previous patch is that I created a separate pass which hopefully makes
> reviewing also easier since it is almost self-contained. Af
On 5/3/21 1:09 PM, Ilya Leoshkevich wrote:
> On Fri, 2021-04-30 at 08:49 +0200, Andreas Krebbel wrote:
>> On 4/28/21 3:48 AM, Ilya Leoshkevich wrote:
>>> Bootstrapped and regtested on s390x-redhat-linux. Tested with
>>> valgrind
>>> too (PR 100278 is now fixed). Ok for master?
>>>
>>> v1:
>>> htt
On 04/05/21 21:42 -0400, Patrick Palka via Libstdc++ wrote:
This rewrites ranges::minmax and ranges::minmax_element so that it
performs at most 3*N/2 many comparisons, as required by the standard.
In passing, this also fixes PR100387 by avoiding a premature std::move
in ranges::minmax and in std:
On 04/05/21 21:42 -0400, Patrick Palka via Libstdc++ wrote:
These constraints are already present on the template we're partially
specilalizing for.
[ This was recently fixed editorially in
https://github.com/cplusplus/draft/pull/4519 ]
Tested on x86_64-pc-linux-gnu, does this look OK for trunk
On 04/05/21 21:42 -0400, Patrick Palka via Libstdc++ wrote:
Tested on x86_64-pc-linux-gnu, does this look OK for trunk and perhaps
10/11?
Yes, for all branches.
libstdc++-v3/ChangeLog:
* include/std/ranges (transform_view::_Iterator::iter_swap):
Remove as per LWG 3520.
On 05/05/21 10:39 +0100, Jonathan Wakely wrote:
On 04/05/21 21:42 -0400, Patrick Palka via Libstdc++ wrote:
This rewrites ranges::minmax and ranges::minmax_element so that it
performs at most 3*N/2 many comparisons, as required by the standard.
In passing, this also fixes PR100387 by avoiding a
Found with amdgcn – silences a build warning there.
Committed as r12-509-gdee371fdd4ae25f837b9b2ded7789d07ed739c9e
Tobias
-
Mentor Graphics (Deutschland) GmbH, Arnulfstrasse 201, 80634 München
Registergericht München HRB 106955, Geschäftsführer: Thomas Heurung, Frank
Thürauf
c
Hi All,
This patch removes several duplicated intrinsic definitions from
arm_mve.h mentioned in PR100419 and also fixes the wrong arguments
in few of intrinsics polymorphic variants.
Regression tested and found no issues.
Ok for master ? GCC-11 and GCC-10 branch backports?
gcc/ChangeLog:
2021-0
On 04/05/21 08:03 +0200, François Dumont wrote:
On 03/05/21 11:06 pm, Jonathan Wakely wrote:
On 03/05/21 22:17 +0200, François Dumont via Libstdc++ wrote:
Is it too early to consider this patch ? Or just lack of time ?
I haven't had time to review it yet, but my general feeling hasn't
changed
On Wed, May 5, 2021 at 11:36 AM Richard Biener
wrote:
>
> On Tue, Mar 16, 2021 at 6:13 PM Stefan Schulze Frielinghaus
> wrote:
> >
> > [snip]
> >
> > Please find attached a new version of the patch. A major change compared to
> > the previous patch is that I created a separate pass which hopeful
> OK.
Thanks.
> I suppose there's no easy way to build a gnat.dg "guality" testcase for
> this?
On all the platforms I test, I have a bunch of guality failures so I'm not
really thrilled by the idea... This should be covered by the GDB testsuite
run with -fgnat-encodings=minimal but it's actu
On 05/05/2021 10:56, Srinath Parvathaneni via Gcc-patches wrote:
Hi All,
This patch removes several duplicated intrinsic definitions from
arm_mve.h mentioned in PR100419 and also fixes the wrong arguments
in few of intrinsics polymorphic variants.
Regression tested and found no issues.
Ok f
> For move2add_valid_value_p we also have to ask the target whether a
> register can be accessed in a different mode than it was set before.
>
> gcc/ChangeLog:
>
> PR rtl-optimization/100263
> * postreload.c (move2add_valid_value_p): Ensure register can
> change mode.
>
> Boots
Hi,
this is the bootstrap failure of GCC 11 on MinGW64 configured with --enable-
tune=nocona. The bottom line is that SEH does not support CFI for epilogues
but the x86 back-end nevertheless attaches it to instructions, so we have to
filter it out and this is done by detecting the end of the pr
On Wed, May 05, 2021 at 01:00:35PM +0200, Eric Botcazou wrote:
> 2021-05-05 Eric Botcazou
>
> PR rtl-optimization/100411
> * cfgcleanup.c (try_crossjump_to_edge): Also skip end of prologue
> and beginning of function markers.
>
> --
> Eric Botcazou
> diff --git a/gcc/cfgcle
> I mean, can't we have just one while loop that skips over all debug insns,
> NOTE_INSN_BASIC_BLOCK_P, NOTE_INSN_PROLOGUE_END and NOTE_INSN_FUNCTION_BEG
> and stops on anything else, or, if we want to skip at most one of some or
> all of those note kinds, do some tracking if we've already skipped
On 24/04/21 15:46 +0200, François Dumont via Libstdc++ wrote:
Hi
Here is the patch to add backtrace generation on _GLIBCXX_DEBUG
assertions thanks to libbacktrace.
In addition to this integration I am also improving the generation
of the assertion message thanks to the "%.*s" printf
Hi Richi,
So I'm trying to look at what IVOPTs does right now and how it might be
able to help us. Looking at these two code examples:
#include
#if 0
int foo(short * a, short * b, unsigned int n)
{
int sum = 0;
for (unsigned int i = 0; i < n; ++i)
sum += a[i] + b[i];
retur
On Wed, May 05, 2021 at 01:21:20PM +0200, Eric Botcazou wrote:
> > I mean, can't we have just one while loop that skips over all debug insns,
> > NOTE_INSN_BASIC_BLOCK_P, NOTE_INSN_PROLOGUE_END and NOTE_INSN_FUNCTION_BEG
> > and stops on anything else, or, if we want to skip at most one of some or
On Tue, 4 May 2021, Jakub Jelinek wrote:
> Hi!
>
> genericize_spaceship genericizes i <=> j to approximately
> ({ int c; if (i == j) c = 0; else if (i < j) c = -1; else c = 1; c; })
> for strong ordering and
> ({ int c; if (i == j) c = 0; else if (i < j) c = -1; else if (i > j) c = 1;
> else c =
On Tue, 4 May 2021, Jakub Jelinek via Gcc-patches wrote:
2) the pr94589-2.C testcase should be matching just 12 times each, but runs
into operator>=(strong_ordering, unspecified) being defined as
(_M_value&1)==_M_value
rather than _M_value>=0. When not honoring NaNs, the 2 case should be
unreac
Passing plain char to isdigit is undefined if the value is negative.
libstdc++-v3/ChangeLog:
* include/std/charconv (__from_chars_alnum): Pass unsigned
char to std::isdigit.
Tested powerpc64le-linux. Committed to trunk.
commit d0d6ca019717305df0ef41e3fe1da48f7f561fac
Author: Jon
libstdc++-v3/ChangeLog:
* testsuite/20_util/function_objects/invoke/3.cc: Check feature
test macro.
* testsuite/20_util/function_objects/invoke/version.cc: New test.
Tested powerpc64le-linux. Committed to trunk.
commit 29745bf06276b9628d08ef1c9e28890cc56df4aa
Author: Jona
> At least for NOTE_INSN_BASIC_BLOCK skipping more than one might
> be problematic, because that would mean we've skipped into a different basic
> block and it wouldn't surprise me if split_block in that case crashed or
> did something weird (if the first argument is not BLOCK_FOR_INSN of the
> sec
Dear Prospective Panelist,
You have been selected as a Customer Satisfaction Analyst to participate in our
short exclusive 10-15 minutes paid project study on Walmart,Walgreens & CVS
stores within your area.
REWARDS: You will earn $500.00 for each completed survey project.
SURVEY [#53779-1001
On Wed, 5 May 2021, Andre Vieira (lists) wrote:
> Hi Richi,
>
> So I'm trying to look at what IVOPTs does right now and how it might be able
> to help us. Looking at these two code examples:
> #include
> #if 0
> int foo(short * a, short * b, unsigned int n)
> {
> int sum = 0;
> for (unsi
On Wed, May 05, 2021 at 02:19:27PM +0200, Eric Botcazou wrote:
> > At least for NOTE_INSN_BASIC_BLOCK skipping more than one might
> > be problematic, because that would mean we've skipped into a different basic
> > block and it wouldn't surprise me if split_block in that case crashed or
> > did so
Hi Richard,
> -Original Message-
> From: Richard Earnshaw
> Sent: 05 May 2021 11:15
> To: Srinath Parvathaneni ; gcc-
> patc...@gcc.gnu.org
> Cc: Richard Earnshaw
> Subject: Re: [GCC][PATCH] arm: Remove duplicate definitions from
> arm_mve.h (pr100419).
>
>
>
> On 05/05/2021 10:56, Sr
Hi~
On Tue, May 04, 2021 at 04:08:22PM +0100, Richard Earnshaw wrote:
> On 03/05/2021 23:55, Segher Boessenkool wrote:
> >CC_STATUS_INIT is suggested in final.c to also be useful for ports that
> >are not CC0, and at least arm seems to use it for something. So I am
> >leaving that alone, but most
Committed.
Segher
---
htdocs/backends.html | 1 -
1 file changed, 1 deletion(-)
diff --git a/htdocs/backends.html b/htdocs/backends.html
index 8034a5776360..f80378b90170 100644
--- a/htdocs/backends.html
+++ b/htdocs/backends.html
@@ -103,7 +103,6 @@ nios2 | C ia
nvptx
On Tue, May 04, 2021 at 06:30:04PM +0200, Eric Botcazou wrote:
> > Pushed. What is next? :-)
>
> You can finally remove powerpcspe. :-)
Done, thanks!
Segher
On Tue, Apr 27, 2021 at 5:29 PM Martin Jambor wrote:
>
> Hi,
>
> IPA-SRA fails to produce (very simple) edge summaries when a caller
> cannot be cloned or its signature cannot be changed which makes it
> less powerful for no good reason. This patch fixes that problem.
>
> Bootstrapped, LTO-bootst
On Tue, Mar 23, 2021 at 5:29 PM Markus Böck via Gcc-patches
wrote:
>
> GCC at the moment uses COMDAT groups for things like virtual thunks,
> even on targets that do not support COMDAT groups. This has not been a
> problem as on platforms not supporting these (such as PE COFF on
> Windows), the ba
Improve GOT addressing by emitting the instructions as a pair. This reduces
register pressure and improves code quality. With -fPIC codesize improves by
0.65% and SPECINT2017 improves by 0.25%.
Passes bootstrap and regress. OK for commit?
ChangeLog:
2021-05-05 Wilco Dijkstra
* confi
On Wed, May 05, 2021 at 01:39:32PM +0200, Richard Biener wrote:
> Can you in the above IL snippets mark COND_BB and MIDDLE_BB?
...
Thanks.
Here is an updated patch (attached) and interdiff from previous patch.
Ok for trunk if it passes bootstrap/regtest?
--- gcc/tree-ssa-phiopt.c 2021-05-03
Implement integer vector compares for 64bit vectors for TARGET_MMX_WITH_SSE.
2021-05-05 Uroš Bizjak
gcc/
PR target/98218
* config/i386/i386-expand.c (ix86_expand_int_sse_cmp):
Handle V8QI, V4HI and V2SI modes.
* config/i386/i386.c (ix86_build_const_vector): Handle V2SImode.
Hi,
This is a backport to gcc-10, cleanly applied on the branch.
This patch changes the test to use the effective-target machinery disables the
error message "ARMv8-M Security Extensions incompatible with selected FPU" when
-mfloat-abi=soft.
Further changes 'asm' to '__asm__' to avoid failures wi
Hi,
This is a backport to gcc-10, cleanly applied on the branch.
As reported in bugzilla when the -mcmse option is used while compiling for size
(-Os) with a thumb-1 target the generated code will clear the registers r7-r10.
These however are callee saved and should be preserved accross ABI bound
Hi Eric, hi all,
currently, gcn (amdgcn-amdhsa) bootstrapping fails as Alexandre's
patch to __builtin_memset (applied yesterday) now does more expansions.
The problem is [→ PR100418]
(set(reg:DI)(plus:DI(reg:DI)(const_int))) [= "adddi3"]
This fails with gcn as gcn has two clobbers for "adddi3
cr16 failed to build due to remnants of CC0 support (NOTICE_UPDATE_CC).
This removes the macro and obvious bits necessary to support it.
Committed to the trunk.
Jeff
commit 14cf6aab8578132ec89ccb46e69899ae6008ff63
Author: Jeff Law
Date: Wed May 5 07:49:28 2021 -0600
Remove NOTICE
On Tue, 4 May 2021 at 15:41, Christophe Lyon wrote:
>
> On Tue, 4 May 2021 at 13:29, Andre Vieira (lists)
> wrote:
> >
> > Hi Christophe,
> >
> > On 30/04/2021 15:09, Christophe Lyon via Gcc-patches wrote:
> > > Since MVE has a different set of vector comparison operators from
> > > Neon, we have
On Tue, 4 May 2021 at 19:03, Christophe Lyon wrote:
>
> On Tue, 4 May 2021 at 15:43, Christophe Lyon
> wrote:
> >
> > On Tue, 4 May 2021 at 13:48, Andre Vieira (lists)
> > wrote:
> > >
> > > It would be good to also add tests for NEON as you also enable auto-vec
> > > for it. I checked and I do
On Wed, 5 May 2021, Jakub Jelinek wrote:
> On Wed, May 05, 2021 at 01:39:32PM +0200, Richard Biener wrote:
> > Can you in the above IL snippets mark COND_BB and MIDDLE_BB?
> ...
>
> Thanks.
> Here is an updated patch (attached) and interdiff from previous patch.
> Ok for trunk if it passes bootst
Hi,
pr56541.c implicitly converts a float vector to an int (bool) vector:
rMin = (rMax>0) ? rMin : rBig;
It fails on some s390 targets because the do not support converting from
vector float to int. Is adding a vect_floatint_cvt as in the attached
patch the OK thing to do?
Or better an xf
Hi,
this patch adds some s390 checks for vect_*_cvts. Is it OK?
Regards
Robin
gcc/testsuite/ChangeLog:
* lib/target-supports.exp: Add s390 checks for vect conversions.
>From 959251d5d2684a9ffebec1b341a4413c2f2328db Mon Sep 17 00:00:00 2001
From: Robin Dapp
Date: Thu, 22 Apr 2021 09:3
Hi, Richard,
During the change for the 2nd version based on your previous comments, I have
the following questions need your help:
>
>> + sra_stats.subtree_deferred_init++;
>> +}
>> + else if (access->grp_to_be_debug_replaced)
>> +{
>> + /* FIXME, this part might have so
On Wed, May 05, 2021 at 09:36:16AM +0200, Richard Biener wrote:
> On Mon, May 3, 2021 at 11:31 AM Ivan Sorokin via Gcc-patches
> wrote:
> >
> > Prior to this commit GCC -O2 generated quite bad code for this
> > function:
> >
> > bool f()
> > {
> > return __builtin_cpu_supports("popcnt")
> >
This patch implements relaxing the requirements when a map with the implicit
attribute encounters
an overlapping existing map. As the OpenMP 5.0 spec describes on page 320,
lines 18-27 (and 5.1 spec,
page 352, lines 13-22):
"If a single contiguous part of the original storage of a list item wit
Hi,
this patch adds vcond_mask patterns with mixed mode for the
condition/mask and source, target so e.g. boolean conditions become
possible:
vtarget = bool_cond ? vsource1 : vsource2.
Is it OK for trunk?
Regards
Robin
gcc/ChangeLog:
* config/s390/vector.md
(vcond_mask_): Add
Removes references to CC_STATUS_INIT from the avr port, which should get
it to the point of building again.
Committed to the trunk.
Jeff
commit b927ffdd6cecd0eeda6ef77df2623519870b1e75
Author: Jeff Law
Date: Wed May 5 09:15:42 2021 -0600
Remove cc0 remnants from avr port
g
On Wed, 5 May 2021, Jonathan Wakely wrote:
> On 04/05/21 21:42 -0400, Patrick Palka via Libstdc++ wrote:
> > This rewrites ranges::minmax and ranges::minmax_element so that it
> > performs at most 3*N/2 many comparisons, as required by the standard.
> > In passing, this also fixes PR100387 by avoi
On 5/4/21 1:44 AM, Jakub Jelinek via Gcc-patches wrote:
Hi!
genericize_spaceship genericizes i <=> j to approximately
({ int c; if (i == j) c = 0; else if (i < j) c = -1; else c = 1; c; })
for strong ordering and
({ int c; if (i == j) c = 0; else if (i < j) c = -1; else if (i > j) c = 1;
else c
On Wed, 5 May 2021, Patrick Palka wrote:
> On Wed, 5 May 2021, Jonathan Wakely wrote:
>
> > On 04/05/21 21:42 -0400, Patrick Palka via Libstdc++ wrote:
> > > This rewrites ranges::minmax and ranges::minmax_element so that it
> > > performs at most 3*N/2 many comparisons, as required by the standa
Hi all,
looking for some feedback on this, one thing I would like to draw attention to
is the fact that this pattern requires 2 separate dependent reductions in the
epilogue.
The accumulator vector for the maximum/minimum elements can be reduced to a
scalar result trivially with a min/max, but
On Wed, May 05, 2021 at 01:45:29PM +0200, Marc Glisse wrote:
> On Tue, 4 May 2021, Jakub Jelinek via Gcc-patches wrote:
>
> > 2) the pr94589-2.C testcase should be matching just 12 times each, but runs
> > into operator>=(strong_ordering, unspecified) being defined as
> > (_M_value&1)==_M_value
>
On 05/05/2021 13:34, Richard Biener wrote:
On Wed, 5 May 2021, Andre Vieira (lists) wrote:
I tried to see what IVOPTs would make of this and it is able to analyze the
IVs but it doesn't realize (not even sure it tries) that one IV's end (loop 1)
could be used as the base for the other (loop 2
On 5/5/2021 8:39 AM, Robin Dapp via Gcc-patches wrote:
Hi,
this patch adds some s390 checks for vect_*_cvts. Is it OK?
Regards
Robin
gcc/testsuite/ChangeLog:
* lib/target-supports.exp: Add s390 checks for vect conversions.
OK
jeff
On 5/5/2021 8:31 AM, Robin Dapp via Gcc-patches wrote:
Hi,
pr56541.c implicitly converts a float vector to an int (bool) vector:
rMin = (rMax>0) ? rMin : rBig;
It fails on some s390 targets because the do not support converting
from vector float to int. Is adding a vect_floatint_cvt as in
>
>>
>>> @@ -11950,6 +12088,72 @@ lower_bound_in_type (tree outer, tree inner)
>>>}
>>> }
>>>
>>> +/* Returns true when the given TYPE has padding inside it.
>>> + return false otherwise. */
>>> +bool
>>> +type_has_padding (tree type)
>>
>> Would it be possible to reuse __builtin_clear_p
1 - 100 of 147 matches
Mail list logo
