On Mon, Jun 8, 2020 at 11:13 PM Jeff Law via Gcc-patches
wrote:
>
>
> We're seeing some failures due to the local-alignment pass. For example on
> sh4:
>
> Tests that now fail, but worked before (16 tests):
>
> gcc.dg/pr48335-1.c (test for excess errors)
> gcc.dg/pr48335-2.c (test for excess err
Hi Alexandre
Alexandre Oliva wrote:
When the linker plugin is not available, dump outputs in lto runs are
different from those outputs.exp tests expect, so skip them for now.
Targets (like Darwin too) that don’t have the linker plugin (presumably)
use fat LTO objects and drive the LTO proce
The patch fixes tree-prof.exp tests on solaris11 and i686-linux-gnu,
problem was that sizeof of a pointer is different from sizeof gcov_type.
I'm going to install it if there are no objections.
Thanks,
Martin
libgcc/ChangeLog:
PR libgcov/95494
* libgcov-driver.c (write_top_count
On Mon, Jun 8, 2020 at 11:59 PM Martin Sebor via Gcc-patches
wrote:
>
> PR 95581 reports an ICE in a call to gimple_call_arg() with
> an out-of-bounds argument number (2). The attached patch avoids
> the ICE but I'm not sure it's the right fix. Other than verifying
> the ICE is gone with a power
On Tue, Jun 9, 2020 at 1:47 AM Alexandre Oliva wrote:
>
>
> An asm operand with a "VIEW_CONVERT_EXPR [...]
> }>" will output the definition of the struct as asm code. Oops.
>
> I hoped setting print_rtx_head would affect print_mem_expr, just
> because it's in print-rtl rather than tree-pretty-p
Add support to optionally emit different instrumentation for accesses to
volatile variables. While the default TSAN runtime likely will never
require this feature, other runtimes for different environments that
have subtly different memory models or assumptions may require
distinguishing volatiles.
The patch fixes a wrong ChangeLog regex, pushed to master.
Martin
contrib/ChangeLog:
* gcc-changelog/git_commit.py: Fix ChangeLog regex in order to
match the top-level ChangeLog.
* gcc-changelog/test_email.py: Add test.
* gcc-changelog/test_patches.txt: Likewise.
On 6/8/20 4:53 PM, Martin Liška wrote:
Hi.
Thank you for the report. It's caused by fact that LLVM switch in 4d474e078ac7
to c++14. So that I suggest to use gnu++14.
Patch can bootstrap on x86_64-linux-gnu and survives regression tests.
I also verified that abidiff is equal for all libsanitizer
An access type whose designated type is an incomplete or class-wide type
may end up designating a task type or a type with a task component. It
is then necessary to associate the access type with the Master of the
tasks that may be generated on an allocator for that access type. This
is done by cre
This removes an obsolete bypass present for long in the predicate
Is_Visible_Component, which is responsible for determining whether
a component present in a private tagged record type is visible in
its extensions. This bypass would essentially always return true
in bodies of instances, but there
This AI clarifies that predicate checks apply at object declaration for
types that contain components with a Default_Value or
Default_Component_Value.
Tested on x86_64-pc-linux-gnu, committed on trunk
2020-06-09 Arnaud Charlet
gcc/ada/
* sem_util.ads, sem_util.adb (Is_Partially_Initi
We were accepting conversion from anon-access-T'Class to access-all-T
which is incorrect.
Tested on x86_64-pc-linux-gnu, committed on trunk
2020-06-09 Arnaud Charlet
gcc/ada/
* sem_type.adb (Covers): Fix implementation of AI05-0149.
* sem_res.adb: Fix typo.--- gcc/ada/sem_res
This patch refines the checks and the corresponding error messsages on
the legality of instantiations where some formal objects and their
corresponding actuals may carry Atomic, Atonic_Components, Volatile,
Volatile_Components, Independent, or Independent_Components aspect
specifications. The lega
Synchronous barriers are currently not supported by GNATprove (i.e. it
doesn't recognize that they are initialized by synchronous). They are
specifically detected and rejected as not-in-SPARK, but it is more
elegant to annotate the Ada.Synchronous runtime library unit with
SPARK_Mode => Off.
Compi
This patch corrects an issue whereby the compiler would incorrectly
generate checks for overlapping formals (-gnateV) leading to spurious
runtime errors when a zero-sized array actual gets passed alongside
another formal declared in the same region.
Tested on x86_64-pc-linux-gnu, committed on trun
This patch corrects an issue whereby the compiler would crash if it
encountered an exit statement featuring a loop identifier where such
referenced loop iterates through a predicated type.
Tested on x86_64-pc-linux-gnu, committed on trunk
2020-06-09 Justin Squirek
gcc/ada/
* exp_ch5.
Make each invocation graph point to the corresponding library graph. Use
this information in Write_Invocation_Graph_Vertex to print out the name
of the library item containing this vertex.
For procedures that take both the library graph and invocation graph as
parameters, remove the library graph
When raising Program_Error due to an incorrect Kind transition, print a
more informative debugging message.
Tested on x86_64-pc-linux-gnu, committed on trunk
2020-06-09 Bob Duff
gcc/ada/
* bindo-graphs.adb (Add_Edge_Kind_Check): Add the Image of the
old and new Kinds to the r
GNAT ignores the null exclusion property of the target access type in a
membership test (e.g. Ptr in Null_Excluding_Ptr_Type), this is fixed
here.
Tested on x86_64-pc-linux-gnu, committed on trunk
2020-06-09 Arnaud Charlet
gcc/ada/
* exp_ch4.adb (Expand_N_In): Fix handling of null ex
This patch corrects the assertions in Add_Edge_Kind_Check. In
particular, a spec-->body edge can come from an Invocation_Edge or a
Forced_Edge in case of cycles. Other edges are added in a certain
order.
Tested on x86_64-pc-linux-gnu, committed on trunk
2020-06-09 Bob Duff
gcc/ada/
This patch disables a sometimes-failing assertion. The assertion is
failing because there is an invocation edge spec-->body, and then we
later add a Body_Before_Spec_Edge for the SCC computation. This is a
temporary fix; we need to investigate why the spec-->body invocation
edge exists.
Tested on
This is a useful optimization to avoid contention around the shared
empty string.
Tested on x86_64-pc-linux-gnu, committed on trunk
2020-06-09 Arnaud Charlet
gcc/ada/
* libgnat/a-strunb__shared.ads, libgnat/a-strunb__shared.adb
(Reference, Unreference): No-op for Empty_Shared
Side_Effect_Free was always assuming an aggregate isn't side effects
free, but we can do better by taking advantage of the
Compile_Time_Known_Aggregate flag.
Tested on x86_64-pc-linux-gnu, committed on trunk
2020-06-09 Arnaud Charlet
gcc/ada/
* exp_util.adb (Side_Effect_Free): Improv
This implements the support for the new C_Variadic_n conventions,
n ranging from 0 to 16, in all versions of the language since the AI is
a binding interpretation.
These new conventions are meant to be used in conjunction with the
Import aspect/pragma to call variadic C functions with the special
Now that there is no artifical limit on Convert_To_Positional, we use
the same heuristics than in Aggr_Size_OK, to share the code and to
generate static aggregates in more cases.
Tested on x86_64-pc-linux-gnu, committed on trunk
2020-06-09 Arnaud Charlet
gcc/ada/
* exp_aggr.adb (Max_
This moves around the declarations of functions that are not in the
right category, adds pragma Inline for others and fixes alphabetization
issues.
This also clearly splits the functions subject to pragma Inline from the
procedures subject to the pragma, the former being read by XEINFO but
not the
We identify all the places where error messages are issued in the
expander and should be issued in semantic analysis instead, for future
clean ups.
Tested on x86_64-pc-linux-gnu, committed on trunk
2020-06-09 Arnaud Charlet
gcc/ada/
* exp_ch3.adb, exp_ch4.adb, exp_ch6.adb, exp_ch9.ad
When a dispatching operation overrides an inherited subprogram, it must
be subtype conformant with the inherited subprogram. Such check is not
performed by the frontend on overriding primitives of private types
which are declared before the full type declaration of the private type.
Tested on x86
This change is a step towards propagating more consistently the DIC,
Invariant and Predicate attributes between different views of the
same type, in particular to the newly built underlying full views.
It also cleans up the handling of the base types for the DIC and
Invariant attributes, which loo
Before this commit, GNAT would use the type of the expression rather
than that of the subtype mark in order to decide whether it should
generate range checks or not. This means that in some cases, GNAT would
decide that no range checks were needed, which was wrong. This issue
went mostly unnoticed
This patch further refines the checks and the corresponding error
messsages on the legality of instantiations where some formal types and
their corresponding actuals may carry Atomic, Volatile, etc. aspect
specifications. The legality rules stated in RN 6.1 (12/5) are in fact
backward-incompatibl
This is a code clean up as part of removing all calls to Error_Msg* in
the expander.
Tested on x86_64-pc-linux-gnu, committed on trunk
2020-06-09 Arnaud Charlet
gcc/ada/
* exp_ch5.adb (Expand_N_Assignment): Remove kludge for
AI05-0087.
* sem_ch12.adb (Validate_Derived
On Mon, Jun 8, 2020 at 10:37 AM Eric Botcazou wrote:
>
> > OK, but all I was saying is that looking at the pointed-to type of
> > the address argument to the memcpy looks fragile to me. For the
> > case cited above it would be better to look at the actual
> > reference we are looking up and not a
On Fri, Jun 5, 2020 at 11:08 AM Jan Hubicka wrote:
>
> Hi,
> this patch fixes ICE while bulding Firefox on assert in
> record_component_aliases which checks that the alias set did not change
> while peeling off vector and array types.
> This is done so we can translate notice references to pointer
Hi C++ folks,
There are a number of places in the coroutine codegen where we need to
build calls to promise methods.
Such methods can, in several cases, also be constexpr or static (or both)
which leads to the PR,
Now the mechanism I use is this;
1/ lookup the method in the relevant scop
On 6/9/20 9:48 AM, Marco Elver wrote:
v2:
* Add Optimization keyword to -param=tsan-distinguish-volatile= as the
parameter can be different per TU.
* Add tree-dump check to test.
Hello.
I support the patch.
@Jakub: Can you please approve the patch?
Thanks,
Martin
Hi
Commit eb72dc663e9070b281be83a80f6f838a3a878822 introduces a ICE on AArch64.
See the discussion on the bug for details.
Bug: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95569
Trivial patch fixing the ICE attached.
Bootstrap and tested on aarch64 platform. No new regression witnessed
Ping?
Maybe I could mention that LLVM emits a warning in this case
(https://reviews.llvm.org/D28820).
Thanks,
Christophe
On Wed, 3 Jun 2020 at 15:23, Christophe Lyon wrote:
>
> Ping?
> https://gcc.gnu.org/pipermail/gcc-patches/2020-May/545747.html
>
> On Wed, 27 May 2020 at 13:52, Christophe
On Tue, Jun 09, 2020 at 09:48:34AM +0200, Marco Elver wrote:
> gcc/
> * params.opt: Define --param=tsan-distinguish-volatile=[0,1].
> * sanitizer.def (BUILT_IN_TSAN_VOLATILE_READ1): Define new
> builtin for volatile instrumentation of reads/writes.
> (BUILT_IN_TSAN_VOLATILE_
On Tue, 9 Jun 2020 at 11:50, Jakub Jelinek wrote:
>
> On Tue, Jun 09, 2020 at 09:48:34AM +0200, Marco Elver wrote:
> > gcc/
> > * params.opt: Define --param=tsan-distinguish-volatile=[0,1].
> > * sanitizer.def (BUILT_IN_TSAN_VOLATILE_READ1): Define new
> > builtin for volatile in
Just wanted to change this one, and noticed this
> > +DEF_SANITIZER_BUILTIN(BUILT_IN_TSAN_VOLATILE_WRITE16,
> > "__tsan_volatile_write16",
> > + BT_FN_VOID_PTR, ATTR_NOTHROW_LEAF_LIST)
is precisely 80 characters. So if I read the style guide right, it's
<= 80 chars (and not < 8
On Tue, Jun 09, 2020 at 12:08:07PM +0200, Marco Elver wrote:
> Just wanted to change this one, and noticed this
>
> > > +DEF_SANITIZER_BUILTIN(BUILT_IN_TSAN_VOLATILE_WRITE16,
> > > "__tsan_volatile_write16",
> > > + BT_FN_VOID_PTR, ATTR_NOTHROW_LEAF_LIST)
>
> is precisely 80 ch
> I'm probably completely misunderstanding how TYPE_REVERSE_STORAGE_ORDER
> works and what the constraints are. But if the memcpy is the storage order
> barrier then what cases are OK?
>
> - both source and destination objects are [only] accessed with the same
>storage order
> - both source
On 08/06/20 19:20 +0100, Jonathan Wakely wrote:
On 05/06/20 22:24 +0200, François Dumont via Libstdc++ wrote:
Hi
   Here is the last of my algo patches this time to extend the
memcmp optimization to std::deque iterators and _GLIBCXX_DEBUG mode.
   To do so I had to return int in imp
Hi Julian!
On 2020-06-05T21:31:08+0100, Julian Brown wrote:
> On Fri, 5 Jun 2020 13:17:09 +0200
> Thomas Schwinge wrote:
>> On 2019-12-17T21:03:47-0800, Julian Brown
>> wrote:
>> > This part contains the libgomp runtime support for the
>> > GOMP_MAP_ATTACH and GOMP_MAP_DETACH mapping kinds
>>
>
On Tue, Jun 09, 2020 at 12:01:24PM +0200, Marco Elver wrote:
> > Do we need/want Optimization here? Optimization means the option is
> > per-function, but to me whether you want to distinguish volatiles or not
> > seems to be a global decision for the whole project.
>
> Adding Optimization here w
On 6/9/20 9:33 AM, Martin Liška wrote:
I'm going to install it if there are no objections.
Pushed as 862b9b225fb.
Martin
On 6/8/20 1:07 PM, Martin Liška wrote:
Can you please test the following patch candidate:
I've just pushed 862b9b225fb which should fix that.
Can you please test the current master?
Thanks,
Martin
Pushed to master.
gcc/ChangeLog:
PR gcov-profile/95365
* doc/gcov.texi: Compile and link one example in 2 steps.
---
gcc/doc/gcov.texi | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/gcc/doc/gcov.texi b/gcc/doc/gcov.texi
index 61250c9407e..22e42da2ea6 10064
Hi Richard,
The 06/08/2020 16:42, Richard Sandiford wrote:
> Tamar Christina writes:
> > diff --git a/gcc/config/aarch64/aarch64.c b/gcc/config/aarch64/aarch64.c
> > index
> > 97da60762390db81df9cffaf316b909cd1609130..9cc8da338125afa01bc9fb645f4112d2d7ef548c
> > 100644
> > --- a/gcc/config/aar
Tamar Christina writes:
> Hi Richard,
> The 06/08/2020 16:42, Richard Sandiford wrote:
>> Tamar Christina writes:
>> > diff --git a/gcc/config/aarch64/aarch64.c b/gcc/config/aarch64/aarch64.c
>> > index
>> > 97da60762390db81df9cffaf316b909cd1609130..9cc8da338125afa01bc9fb645f4112d2d7ef548c
>> >
Hello.
The patch is follow up of the kernel discussion:
https://lore.kernel.org/lkml/canpmjnnrz5ovkb6pe7k6gjfogbht_zhypkng9ad+kjndzk7...@mail.gmail.com/
The patch changes inliner in the following way:
1) do not compare caller and callee sanitizer attributes
for always_inline functions - that
On Tue, Jun 09, 2020 at 01:53:38PM +0200, Martin Liška wrote:
> + const sanitize_code codes[] =
> +{
> + SANITIZE_ADDRESS,
> + SANITIZE_THREAD,
> + SANITIZE_LEAK,
Why leak? That is really a link time option only, doesn't affect code
generation in any way.
On the other side, d
On 6/9/20 2:01 PM, Jakub Jelinek wrote:
On Tue, Jun 09, 2020 at 01:53:38PM +0200, Martin Liška wrote:
+ const sanitize_code codes[] =
+{
+ SANITIZE_ADDRESS,
+ SANITIZE_THREAD,
+ SANITIZE_LEAK,
Why leak?
I was too eager ;)
That is really a link time option only, doesn't
On Tue, Jun 09, 2020 at 02:09:06PM +0200, Martin Liška wrote:
> - return ((sanitize_flags_p (SANITIZE_ADDRESS, caller)
> -== sanitize_flags_p (SANITIZE_ADDRESS, callee))
> - && (sanitize_flags_p (SANITIZE_POINTER_COMPARE, caller)
> - == sanitize_flags_p (SANITIZE_POINTER_CO
Hi Richard,
The 06/09/2020 12:44, Richard Sandiford wrote:
> Tamar Christina writes:
> > Hi Richard,
> > The 06/08/2020 16:42, Richard Sandiford wrote:
> >> Tamar Christina writes:
> >> > diff --git a/gcc/config/aarch64/aarch64.c b/gcc/config/aarch64/aarch64.c
> >> > index
> >> > 97da60762390db
On Tue, 9 Jun 2020 12:41:21 +0200
Thomas Schwinge wrote:
> Hi Julian!
>
> On 2020-06-05T21:31:08+0100, Julian Brown
> wrote:
> >> For the OpenACC runtime API 'acc_attach' etc. routines they don't,
> >> so what's the conceptual reason that for the corresponding OpenACC
> >> directive variants, '
Hi!
On 2020-05-26T04:08:44-0300, Alexandre Oliva wrote:
> Thanks, here's the combined patch I'm checking in.
>
> revamp dump and aux output names
For BRIG (HSAIL) front end testing, I'm see a lot of failures like:
Running [...]/source-gcc/gcc/testsuite/brig.dg/dg.exp ...
PASS: brig.dg/t
On 6/9/20 2:15 PM, Jakub Jelinek wrote:
Sorry for not writing everything at once, but are the
SANITIZER_POINTER_{COMPARE,SUBTRACT} differences unimportant?
They are. I got confused that they are not
part of SANITIZE_UNDEFINED or SANITIZE_UNDEFINED_NONDEFAULT.
Anyway, good point!
Martin
>From
On Tue, Jun 09, 2020 at 02:32:36PM +0200, Martin Liška wrote:
> >From 5fe0671ad79d14d1c9d0fead1a471875a4416fac Mon Sep 17 00:00:00 2001
> From: Martin Liska
> Date: Tue, 9 Jun 2020 13:03:55 +0200
> Subject: [PATCH] sanitizer: do not inline no-sanitize into sanitizer fn
>
> gcc/ChangeLog:
>
>
Pushed to master.
gcc/ChangeLog:
* gcov-dump.c (print_usage): Fix spacing for --raw option
in --help.
---
gcc/gcov-dump.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/gcc/gcov-dump.c b/gcc/gcov-dump.c
index 90cbd1ace52..ffb71ca4b68 100644
--- a/gcc/gcov-du
> Yes, but the problem is that remap_decl isn't getting called.
Right, I can get it to be called by adding a pushdecl to grokdeclarator...
> Attaching it to the BIND_EXPR doesn't help walk_tree_1 do the right
> thing with the DECL_EXPR.
... but, indeed, this still ICEs. So the key is that the D
Hi,
On Tue, Jun 09 2020, Thomas Schwinge wrote:
> Hi!
>
> On 2020-05-26T04:08:44-0300, Alexandre Oliva wrote:
>> Thanks, here's the combined patch I'm checking in.
>>
>> revamp dump and aux output names
>
> For BRIG (HSAIL) front end testing, I'm see a lot of failures like:
>
> Running [...]/
On Tue, Jun 9, 2020 at 11:47 AM qianchao wrote:
>
> Hi
>
> Commit eb72dc663e9070b281be83a80f6f838a3a878822 introduces a ICE on AArch64.
> See the discussion on the bug for details.
> Bug: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95569
>
> Trivial patch fixing the ICE attached.
>
> Bootstrap a
On Tue, Jun 9, 2020 at 12:19 PM Eric Botcazou wrote:
>
> > I'm probably completely misunderstanding how TYPE_REVERSE_STORAGE_ORDER
> > works and what the constraints are. But if the memcpy is the storage order
> > barrier then what cases are OK?
> >
> > - both source and destination objects are
Hi!
On 2020-05-26T04:08:44-0300, Alexandre Oliva wrote:
> Thanks, here's the combined patch I'm checking in.
>
> revamp dump and aux output names
For libgomp offloading testing, I'm seeing a number of failures like:
UNSUPPORTED:
libgomp.oacc-c/../libgomp.oacc-c-c++-common/nvptx-merged-loop
Add support to optionally emit different instrumentation for accesses to
volatile variables. While the default TSAN runtime likely will never
require this feature, other runtimes for different environments that
have subtly different memory models or assumptions may require
distinguishing volatiles.
On Tue, Jun 09, 2020 at 03:15:39PM +0200, Marco Elver wrote:
> gcc/
> * params.opt: Define --param=tsan-distinguish-volatile=[0,1].
> * sanitizer.def (BUILT_IN_TSAN_VOLATILE_READ1): Define new
> builtin for volatile instrumentation of reads/writes.
> (BUILT_IN_TSAN_VOLATILE_
> -Original Message-
> From: Gcc-patches On Behalf Of Richard
> Biener
> Sent: 05 May 2020 14:49
> To: gcc-patches@gcc.gnu.org
> Subject: [PATCH] move permutation validity check
>
>
> This delays the SLP permutation check to vectorizable_load and optimizes
> permutations only after all S
On Tue, Jun 9, 2020 at 3:22 PM Alex Coplan wrote:
>
> > -Original Message-
> > From: Gcc-patches On Behalf Of Richard
> > Biener
> > Sent: 05 May 2020 14:49
> > To: gcc-patches@gcc.gnu.org
> > Subject: [PATCH] move permutation validity check
> >
> >
> > This delays the SLP permutation che
> -Original Message-
> From: Richard Biener
> Sent: 09 June 2020 14:24
> To: Alex Coplan
> Cc: Richard Biener ; gcc-patches@gcc.gnu.org
> Subject: Re: [PATCH] move permutation validity check
>
> On Tue, Jun 9, 2020 at 3:22 PM Alex Coplan wrote:
> >
> > > -Original Message-
> > >
On Mon, Jun 8, 2020 at 1:04 PM Martin Liška wrote:
>
> Hello.
>
> Thank you for the approval. There's the patch that defines 4 new
> DEF_INTERNAL_OPTAB_FN.
>
> Patch can bootstrap on x86_64-linux-gnu and survives regression tests.
> It also builds on ppc64le-linux-gnu.
>
> Ready to be installed?
On Tue, 9 Jun 2020 at 15:22, Jakub Jelinek wrote:
>
> On Tue, Jun 09, 2020 at 03:15:39PM +0200, Marco Elver wrote:
> > gcc/
> > * params.opt: Define --param=tsan-distinguish-volatile=[0,1].
> > * sanitizer.def (BUILT_IN_TSAN_VOLATILE_READ1): Define new
> > builtin for volatile in
This removes dead code that was left over from the reduction
vectorization refactoring last year.
Bootstrapped / tested on x86_64-unknown-linux-gnu, applied.
Richard.
2020-06-09 Richard Biener
* tree-vect-stmts.c (vect_transform_stmt): Remove dead code.
---
gcc/tree-vect-stmts.c | 4
It turned out that this patch fails with LTO and partitions,
causing fails at runtime such as
libgomp: Duplicate node
via libgomp/splay-tree.c's splay_tree_insert.
In the test case, the problem occurred for functions - namely
main._omp_fn.* on the host.
If the code is run in LTO context, the fi
On Tue, Jun 09, 2020 at 04:02:19PM +0200, Tobias Burnus wrote:
> It turned out that this patch fails with LTO and partitions,
> causing fails at runtime such as
> libgomp: Duplicate node
> via libgomp/splay-tree.c's splay_tree_insert.
>
> In the test case, the problem occurred for functions - na
Hi,
So this is my rework of the code you sent me, I have not included the
'permute' code you included as I can't figure out what it is meant to be
doing. Maybe something to look at later.
I have also included three tests that show it working for some simple
cases and even a nested one.
Unf
Hi Alexandre,
> When the linker plugin is not available, dump outputs in lto runs are
> different from those outputs.exp tests expect, so skip them for now.
>
> Regstrapped on x86_64-linux-gnu. Ok to install?
>
>
> for gcc/testsuite/ChangeLog
>
> * outputs.exp (skip_lto): Set when missing
On Tue, 2020-06-09 at 01:00 +, Segher Boessenkool wrote:
> Some testcases failed (esp. with --with-cpu=power9) after my change
> to
> prefer xxperm over vperm when all else is equal. Fix that. (This
> also
> tightens the relevant REs somewhat).
>
> Tested on way too many configurations. Com
The 'you' here is Richi, which Richi is probably aware but maybe not the
rest of the list :')
On 09/06/2020 15:29, Andre Vieira (lists) wrote:
Hi,
So this is my rework of the code you sent me, I have not included the
'permute' code you included as I can't figure out what it is meant to
be do
On Tue, 2020-06-09 at 09:17 +0200, Richard Biener wrote:
> On Mon, Jun 8, 2020 at 11:13 PM Jeff Law via Gcc-patches
> wrote:
> >
> > We're seeing some failures due to the local-alignment pass. For example on
> > sh4:
> >
> > Tests that now fail, but worked before (16 tests):
> >
> > gcc.dg/pr
On Tue, Jun 09, 2020 at 08:54:47AM -0600, Jeff Law via Gcc-patches wrote:
> > While technically OK the issue is that it does not solve the x86 issue
> > which with incoming stack alignment < 8 bytes does not perform
> > dynamic re-alignment to align 'long long' variables appropriately.
> > Currentl
On Tue, 2020-06-09 at 16:59 +0200, Jakub Jelinek wrote:
> On Tue, Jun 09, 2020 at 08:54:47AM -0600, Jeff Law via Gcc-patches wrote:
> > > While technically OK the issue is that it does not solve the x86 issue
> > > which with incoming stack alignment < 8 bytes does not perform
> > > dynamic re-alig
On Tue, Jun 09, 2020 at 09:18:25AM -0600, Jeff Law wrote:
> On Tue, 2020-06-09 at 16:59 +0200, Jakub Jelinek wrote:
> > On Tue, Jun 09, 2020 at 08:54:47AM -0600, Jeff Law via Gcc-patches wrote:
> > > > While technically OK the issue is that it does not solve the x86 issue
> > > > which with incomin
On Tue, 2020-06-09 at 17:26 +0200, Jakub Jelinek wrote:
> On Tue, Jun 09, 2020 at 09:18:25AM -0600, Jeff Law wrote:
> > On Tue, 2020-06-09 at 16:59 +0200, Jakub Jelinek wrote:
> > > On Tue, Jun 09, 2020 at 08:54:47AM -0600, Jeff Law via Gcc-patches wrote:
> > > > > While technically OK the issue is
On 6/9/20 1:43 AM, Richard Biener wrote:
On Mon, Jun 8, 2020 at 11:59 PM Martin Sebor via Gcc-patches
wrote:
PR 95581 reports an ICE in a call to gimple_call_arg() with
an out-of-bounds argument number (2). The attached patch avoids
the ICE but I'm not sure it's the right fix. Other than ver
On Tue, Jun 9, 2020 at 5:51 PM Martin Sebor wrote:
>
> On 6/9/20 1:43 AM, Richard Biener wrote:
> > On Mon, Jun 8, 2020 at 11:59 PM Martin Sebor via Gcc-patches
> > wrote:
> >>
> >> PR 95581 reports an ICE in a call to gimple_call_arg() with
> >> an out-of-bounds argument number (2). The attache
We only need to predeclare a VLA type if it's wrapped in a pointer type;
otherwise gimplify_type_sizes will handle it.
Tested x86_64-pc-linux-gnu, applying to trunk.
gcc/cp/ChangeLog:
PR c++/95552
* cp-gimplify.c (predeclare_vla): Only predeclare a VLA if it's
wrapped in
On 6/9/20 8:41 AM, Eric Botcazou wrote:
Yes, but the problem is that remap_decl isn't getting called.
Right, I can get it to be called by adding a pushdecl to grokdeclarator...
Attaching it to the BIND_EXPR doesn't help walk_tree_1 do the right
thing with the DECL_EXPR.
... but, indeed, thi
On 09/06/20 00:02 +0100, Jonathan Wakely wrote:
On 08/06/20 22:08 +0100, Jonathan Wakely wrote:
On 08/06/20 19:20 +0100, Jonathan Wakely wrote:
On 05/06/20 22:24 +0200, François Dumont via Libstdc++ wrote:
Hi
   Here is the last of my algo patches this time to extend the
memcmp optimiza
On 09/06/20 17:11 +0100, Jonathan Wakely wrote:
On 09/06/20 00:02 +0100, Jonathan Wakely wrote:
On 08/06/20 22:08 +0100, Jonathan Wakely wrote:
On 08/06/20 19:20 +0100, Jonathan Wakely wrote:
On 05/06/20 22:24 +0200, François Dumont via Libstdc++ wrote:
Hi
   Here is the last of my algo
This removes dead code left over from the reduction vectorization
refactoring last year.
Bootstrapped on x86_64-unknown-linux-gnu, testing in progress.
Richard.
2020-06-09 Richard Biener
* tree-vect-loop.c (vectorizable_induction): Remove dead code.
---
gcc/tree-vect-loop.c | 42 ---
On Tue, Jun 9, 2020 at 10:09 AM Martin Liška wrote:
>
> On 6/8/20 4:53 PM, Martin Liška wrote:
> > Hi.
> >
> > Thank you for the report. It's caused by fact that LLVM switch in
> > 4d474e078ac7
> > to c++14. So that I suggest to use gnu++14.
> >
> > Patch can bootstrap on x86_64-linux-gnu and sur
On Sun, May 31, 2020 at 4:10 PM H.J. Lu wrote:
>
> We used to expand memcmp to "repz cmpsb" via cmpstrnsi. It was changed
> by
>
> commit 9b0f6f5e511ca512e4faeabc81d2fd3abad9b02f
> Author: Nick Clifton
> Date: Fri Aug 12 16:26:11 2011 +
>
> builtins.c (expand_builtin_memcmp): Do not us
On Fri, May 22, 2020 at 4:22 AM H.J. Lu wrote:
>
> On Sat, May 2, 2020 at 4:55 AM H.J. Lu wrote:
> >
> > Currently patchable area is at the wrong place. It is placed immediately
> > after function label, before both .cfi_startproc and ENDBR. This patch
> > adds UNSPECV_PATCHABLE_AREA for pseudo
On Tue, May 26, 2020 at 6:27 AM Martin Liška wrote:
>
> On 5/26/20 1:59 PM, H.J. Lu wrote:
> > On Tue, May 26, 2020 at 2:30 AM Martin Liška wrote:
> >>
> >> On 5/25/20 7:42 PM, H.J. Lu wrote:
> >>> Here is the updated patch. OK for master?
> >>
> >> Thank you for the updated patch.
> >>
> >> I h
May I please ping this patch?
https://gcc.gnu.org/pipermail/gcc-patches/2020-May/545426.html
Thanks!
-Lewis
On Fri, May 08, 2020 at 03:35:25PM -0400, Lewis Hyatt wrote:
> On Fri, Jan 31, 2020 at 03:31:59PM -0500, David Malcolm wrote:
> > On Fri, 2020-01-31 at 14:31 -0500, Lewis Hyatt wrote:
> >
Hi,
This patch merges the D front-end implementation with upstream dmd
13d67c575. Contains many small API changes, most contained within the
front-end, but some spill out into the codegen.
Bootstrapped and regression tested on x86_64-linux-gnu, and committed to
mainline.
Regards
Iain.
gcc/d/C
Hi Bernhard,
For Fortran identifiers, that would be really good. Would you do that?
Is it really needed?
We do not enter internal helper names (class names, anything
artificial, anything leading underscore or leading uppercase) to the
suggestions proposals so should never actually recommend a
> Like this?
Fine with me, thanks!
--
Eric Botcazou
1 - 100 of 153 matches
Mail list logo
