Hi!
The bzhi patterns are quite complicated because they need to accurately
describe the behavior of the instruction for all input values.
The following patterns are simple and make bzhi recognizable even for
cases where not all input values are valid, because the user used
a shift, in which case
On Sat, Jan 18, 2020 at 8:58 PM Mike Stump wrote:
> On Jan 17, 2020, at 2:07 PM, David Malcolm wrote:
> >
> > I ran into difficulties with the Graphviz format changing from under
> > me during an upgrade, where the new version of "dot" would reject .dot
> > files generated by the analyzer.
> >
>
This libgo patch changes the testsuite script to explicitly skip type
descriptors. Type descriptors are normally weak and GNU nm will
report them as V, so we will skip them when collecting the list of
symbols. But when not using GNU nm, they may be reported as D, so
also skip them in symstogo. T
On Jan 22, 2020, Andreas Schwab wrote:
> On Dez 25 2019, Alexandre Oliva wrote:
>> +dnl # Different versions and execution modes implement different
>> +dnl # subsets of these functions. Instead of hard-coding, test for C
>> +dnl # declarations in headers. The C primitives could be defined as
>
On Jan 22, 2020, Jonathan Wakely wrote:
> Isn't allowing arithmetic on function pointers a GNU extension?
Does that matter? This test is only supposed to be compiled by GCC.
> I think just adding the #undef to what you had originally is the best
> version.
'k, thanks, will adjust, test, post
Successfully bootstrapped & regrtested on x86_64-pc-linux-gnu.
OK for master? I'm working on various followup bugfixes that could
use this for test coverage.
gcc/testsuite/ChangeLog:
* gcc.dg/analyzer/data-model-3.c: Remove hardcoded "-O2" and move
to torture/conftest-1.c.
The ICE in PR analyzer/93382 is a validation error.
The global variable "idx" acquires a "tainted" state from local array
n1[0]. When the frame is popped, the svalue for n1[0] is purged, but
the "taint" sm_state_map's entry for "idx" has a svalue_id referencing
the now-purged svalue. This is cau
When adding namespaces to the analyzer in
r10-6151-g75038aa6aa5b562e6358108619d66ef2ccab9a53
I messed up the nesting of the #endif for #if CHECKING_P
and the closing of namespace ana.
This patch fixes it.
Successfully bootstrapped & regrtested on x86_64-pc-linux-gnu;
verified stage 1 build with -
From: Andrew Pinski
The problem here is arc looks at current_output_insn unconditional
but sometimes current_output_insn is NULL. With patchable-function-entry,
it will be. This is similar to how the nios2, handles "%.".
Committed as obvious after a simple test with -fpatchable-function-entry=1
From: Andrew Pinski
This reverts commit 9085381f1931cc3667412c8fff91878184835901 as it was
causing default dg-do to be set incorrectly on most targets.
Instead move testcases that are vect related testcase that
use "dg-require-effective-target vect_*" to the vect test area.
Committed as obvious
On Wed, Jan 22, 2020 at 4:34 PM Jakub Jelinek wrote:
>
> On Wed, Jan 22, 2020 at 04:22:54PM -0800, Andrew Pinski wrote:
> > > I see
> > > pr88497-1.c:/* { dg-require-effective-target vect_double } */
> > > pr88497-2.c:/* { dg-require-effective-target vect_float } */
> > > pr88497-3.c:/* { dg-requi
On Wed, Jan 22, 2020 at 04:22:54PM -0800, Andrew Pinski wrote:
> > I see
> > pr88497-1.c:/* { dg-require-effective-target vect_double } */
> > pr88497-2.c:/* { dg-require-effective-target vect_float } */
> > pr88497-3.c:/* { dg-require-effective-target vect_int } */
> > pr88497-4.c:/* { dg-require-
On Wed, Jan 22, 2020 at 3:39 PM Jakub Jelinek wrote:
>
> On Tue, Jan 21, 2020 at 05:56:38PM -0800, apin...@marvell.com wrote:
> > From: Andrew Pinski
> >
> > tree-ssa testcases sometimes check autovect effective target
> > but does not set it up. On MIPS, those testcases fail with
> > some TCL e
On Thu, Jan 23, 2020 at 12:39:34AM +0100, Jakub Jelinek wrote:
> I don't see how this could have been tested on x86_64-linux, it breaks there
> a lot of tests (ditto i686-linux). My wild guess is that it changes the
> default dg-do action for tests that don't have any, which in tree-ssa.exp
> used
I merged trunk revision 9e16359ce8db7180264fd5f047ca137ead8356dd to
the gccgo branch.
Ian
On Tue, Jan 21, 2020 at 05:56:38PM -0800, apin...@marvell.com wrote:
> From: Andrew Pinski
>
> tree-ssa testcases sometimes check autovect effective target
> but does not set it up. On MIPS, those testcases fail with
> some TCL error messages. This fixes the issue by calling
> check_vect_suppor
On Wed, Jan 22, 2020 at 12:18 PM Rainer Orth
wrote:
>
> > I've committed a patch to update libgo to Go 1.14beta1. As usual with
> > these updates the patch is far too large to include in this e-mail
> > message. I've included the diffs for gccgo-specific files.
> > Bootstrapped and ran Go tests
This libgo patch fixes the build on Solaris after the 1.14beta
upgrade. It adds a runtime.osinit function, and drops a duplicate
runtime.getncpu function. Bootstrapped and tested on
x86_64-pc-solaris2.11. Committed to mainline.
Ian
9e16359ce8db7180264fd5f047ca137ead8356dd
diff --git a/gcc/go/go
This patch from Andrew Pinski fixes the libgo arm64be build, by using
internal/syscall/unix/getrandom_linux_generic.go on arm64be.
Committed to mainline.
Ian
diff --git a/gcc/go/gofrontend/MERGE b/gcc/go/gofrontend/MERGE
index 61f01d739ff..544c6e66e0d 100644
--- a/gcc/go/gofrontend/MERGE
+++ b/gcc
On 1/22/20 1:20 PM, Jason Merrill wrote:
On 1/22/20 5:14 AM, Martin Liška wrote:
Hi.
Just for the record, after the change 526.blender_r fails due to:
blender/source/blender/blenlib/intern/math_color.c: In function
'rgb_float_to_uchar':
blender/source/blender/blenlib/BLI_utildefines.h:251:22:
On Wed, 2020-01-22 at 13:16 -0800, Andrew Pinski wrote:
> On Wed, Jan 22, 2020 at 12:48 PM Jeff Law wrote:
> > On Mon, 2020-01-20 at 09:42 +0100, Richard Biener wrote:
> > > On Sat, Jan 18, 2020 at 1:47 AM wrote:
> > > > From: Andrew Pinski
> > > >
> > > > On MIPS, .set noreorder/reorder needs
Documento 429kb (Visualizar)
Boa tarde, segue em anexo Guia do
pedido,Cadastramento da inclusão em CPF.
Atenciosamente,ESTOFAMENTOS PRORLEY LTDA
On Wed, 22 Jan 2020, Richard Earnshaw (lists) wrote:
> Joseph, have I got all of these right?
refs/meta in that list should be refs/meta/config, there isn't anything
else under refs/meta/ at present.
--
Joseph S. Myers
jos...@codesourcery.com
On 22.01.2020 22:27, Andreas Tobler wrote:
> Hi all,
>
> this patch adds the necessary bits to enable asan support on FreeBSD
> x86_64.
> Results will be produced over night.
>
> Ok for trunk once it is open again?
> TIA,
> Andreas
>
> gcc/
>
> * config/i386/i386.h: Define a new macro: SUBTAR
On Wed, Jan 22, 2020 at 1:16 PM Andrew Pinski wrote:
>
> On Wed, Jan 22, 2020 at 12:48 PM Jeff Law wrote:
> >
> > On Mon, 2020-01-20 at 09:42 +0100, Richard Biener wrote:
> > > On Sat, Jan 18, 2020 at 1:47 AM wrote:
> > > > From: Andrew Pinski
> > > >
> > > > On MIPS, .set noreorder/reorder nee
On 1/22/20 3:31 PM, Paolo Carlini wrote:
Hi,
On 22/01/20 17:27, Jason Merrill wrote:
On 1/22/20 10:22 AM, Paolo Carlini wrote:
Hi,
in this simple issue we either wrongly talked about variable
template-id in c++17 mode or ICEd in c++2a. I think we simply want to
handle concept-ids first, bot
Hi all,
and here the tests which should succeed now.
Ok for trunk (as usual if open...)
TIA,
Andreas
testsuite:
* gcc.dg/asan/pr87930.c: Enable on x86_64 FreeBSD.
* c-c++-common/asan/asan-interface-1.c: Likewise.
* c-c++-common/asan/clone-test-1.c: Likewise.
*
Hi all,
this patch adds the necessary bits to enable asan support on FreeBSD x86_64.
Results will be produced over night.
Ok for trunk once it is open again?
TIA,
Andreas
gcc/
* config/i386/i386.h: Define a new macro: SUBTARGET_SHADOW_OFFSET.
* config/i386/i386.c (ix86_asan_shadow_offset):
Hi all,
I'm digginig out old patches and I want to complete the libasan support
for FreeBSD x86_64. The below one was not that obvious when you have
been away for the past years.
In the last import the sanitizer_platform_limits_freebsd.cpp got
forgotten. Fix this.
Ok for trunk once it's open
On Wed, Jan 22, 2020 at 12:48 PM Jeff Law wrote:
>
> On Mon, 2020-01-20 at 09:42 +0100, Richard Biener wrote:
> > On Sat, Jan 18, 2020 at 1:47 AM wrote:
> > > From: Andrew Pinski
> > >
> > > On MIPS, .set noreorder/reorder needs to emitted around
> > > the nop. The template for the nop instruct
On 1/22/20 9:35 PM, Jeff Law wrote:
On Wed, 2020-01-22 at 15:59 +0100, Martin Sebor wrote:
The early front-end only implementation of -Wrestrict that's still
present in GCC 10 issues a false postive for %p arguments that are
the same as the destination. Bug 84919 reports an instance of this
fal
On Wed, 2020-01-22 at 10:14 +0200, Claudiu Zissulescu wrote:
> ARC processors can work with a reduced register set (i.e. registers
> r4-r9 and r16-r25 are not available). This option can be enabled
> passing -mrf16 option to the compiler, or by using -mcpu=em_mini CPU
> configuration. Using RF16 co
On Wed, 2020-01-22 at 15:39 +, Andrew Burgess wrote:
> The motivation behind this change is to make it easier for a user to
> link against static libraries on a target where dynamic libraries are
> the default library type (for example GNU/Linux).
>
> Further, my motivation is really for linki
On Mon, 2020-01-20 at 09:42 +0100, Richard Biener wrote:
> On Sat, Jan 18, 2020 at 1:47 AM wrote:
> > From: Andrew Pinski
> >
> > On MIPS, .set noreorder/reorder needs to emitted around
> > the nop. The template for the nop instruction uses %(/%) to
> > do that. But default_print_patchable_fun
On Tue, 2020-01-21 at 14:59 +, Richard Sandiford wrote:
> check-function-bodies allows individual function tests to be
> annotated with target/xfail selectors, but sometimes it's
> useful to have the same selector for all functions.
>
> Tested on aarch64-linux-gnu and x86_64-linux-gnu, and wit
On Tue, 2020-01-21 at 17:56 -0800, apin...@marvell.com wrote:
> From: Andrew Pinski
>
> tree-ssa testcases sometimes check autovect effective target
> but does not set it up. On MIPS, those testcases fail with
> some TCL error messages. This fixes the issue by calling
> check_vect_support_and_s
On Wed, 2020-01-22 at 14:56 -0500, David Malcolm wrote:
> PR analyzer/93378 reports an ICE at -O1 -g when analyzing a rewind
> via
> longjmp to a setjmp call with.
>
> The root cause is that the rewind_info_t::get_setjmp_call attempts to
> locate the setjmp GIMPLE_CALL via within the exploded_node
On Wed, 2020-01-22 at 15:59 +0100, Martin Sebor wrote:
> The early front-end only implementation of -Wrestrict that's still
> present in GCC 10 issues a false postive for %p arguments that are
> the same as the destination. Bug 84919 reports an instance of this
> false positive in the Linux kernel
Hi,
On 22/01/20 17:27, Jason Merrill wrote:
On 1/22/20 10:22 AM, Paolo Carlini wrote:
Hi,
in this simple issue we either wrongly talked about variable
template-id in c++17 mode or ICEd in c++2a. I think we simply want to
handle concept-ids first, both as represented in c++17 mode and as
rep
On Wed, 2020-01-22 at 10:14 +0200, Claudiu Zissulescu wrote:
> ARC600 when configured with mul64 instructions uses mlo and mhi
> registers to store the 64 result of the multiplication. In the ARC600
> ISA documentation we have the next register configuration when ARC600
> is configured only with mu
Hi Ian,
> I've committed a patch to update libgo to Go 1.14beta1. As usual with
> these updates the patch is far too large to include in this e-mail
> message. I've included the diffs for gccgo-specific files.
> Bootstrapped and ran Go tests on x86_64-pc-linux-gnu. Committed to
> mainline.
the
On Wed, 2020-01-22 at 10:14 +0200, Claudiu Zissulescu wrote:
> The ARC's 600 multiplication instruction can accept signed 12 bit
> instructions.
>
> gcc/
> -xx-xx Claudiu Zissulescu
>
> * config/arc/arc.c (arc_rtx_costs): Update mul64 cost.
OK
jeff
>
On Wed, 2020-01-22 at 12:11 +, Richard Sandiford wrote:
> In this PR, auto-inc-dec was trying to turn:
>
> (set (reg X) (plus (reg X) (const_int N)))
> (clobber (mem (reg X)))
>
> into:
>
> (clobber (mem (pre_modify (reg X) ...)))
>
> But bare clobber insns are just there to des
PR analyzer/93378 reports an ICE at -O1 -g when analyzing a rewind via
longjmp to a setjmp call with.
The root cause is that the rewind_info_t::get_setjmp_call attempts to
locate the setjmp GIMPLE_CALL via within the exploded_node containing
it, but the exploded_node has two stmts: a GIMPLE_DEBUG,
On Wed, Jan 22, 2020 at 02:35:13PM -0500, David Malcolm wrote:
> PR analyzer/93316 reports various testsuite failures where I
> accidentally relied on properties of x86_64-pc-linux-gnu.
>
> The following patch fixes them on sparc-sun-solaris2.11 (gcc211 in the
> GCC compile farm), and, I hope, the
On Wed, 2020-01-22 at 12:02 +, Richard Sandiford wrote:
> One consequence of r276318 was that cselib now preserves sp-based
> values across function calls. This in turn convinced cprop to
> replace the clobber in:
>
>(set (reg PSUEDO) (reg sp))
>...
>(call ...)
>...
>(clob
PR analyzer/93316 reports various testsuite failures where I
accidentally relied on properties of x86_64-pc-linux-gnu.
The following patch fixes them on sparc-sun-solaris2.11 (gcc211 in the
GCC compile farm), and, I hope, the other configurations showing
failures.
There may still be other failure
Hi!
I've backported these 15 commits from trunk to 9.x branch,
bootstrapped/regtested them on x86_64-linux and i686-linux and
committed.
Jakub
PR c++/92992
* call.c (convert_arg_to_ellipsis): For decltype(nullptr) arguments
that have side-effects use cp_build_compo
On Fri, 2020-01-17 at 16:54 -0500, David Malcolm wrote:
> PR analyzer/93307 reports that in an LTO bootstrap, there are ODR
> violations between:
> - the "region" type:
> gcc/analyzer/region-model.h:792
> vs:
> gcc/sched-int.h:1443
> - the "constraint" type:
> gcc/analyzer/constraint-
In the update of libgo to Go1.14beta1, the function runtime_nanotime
was renamed to runtime_nanotime1. I missed one reference that is not
compiled on x86 code. This patch fixes that. This fixes
https://golang.org/issue/36694. Bootstrapped and ran Go testsuite on
x86_64-pc-linux-gnu. Committed
On 1/22/20 5:14 AM, Martin Liška wrote:
Hi.
Just for the record, after the change 526.blender_r fails due to:
blender/source/blender/blenlib/intern/math_color.c: In function
'rgb_float_to_uchar':
blender/source/blender/blenlib/BLI_utildefines.h:251:22: error:
conversion from 'float' to 'unsig
On Wed, Jan 22, 2020 at 08:08:32PM +0300, Alexander Monakov wrote:
>
>
> On Wed, 22 Jan 2020, Stefan Schulze Frielinghaus wrote:
>
> > Hi David,
> >
> > In function `tree_cmp` an invariant [1] is assumed which does not
> > necessarily
> > exist. In case both input trees are finally compared vi
[updated based on v2 discussions]
This patch proposes some new (additional) rules for email subject lines
when contributing to GCC. The goal is to make sure that, as far as
possible, the subject for a patch will form a good summary when the
message is committed to the repository if applied with
Ping.
I will change the tests to use the exact input and output registers as
Richard Sandiford suggested for the AArch64 patches.
On 12/20/19 6:48 PM, Delia Burduv wrote:
> This patch adds the ARMv8.6 ACLE BFloat16 load intrinsics vld{q}_bf16
> as part of the BFloat16 extension.
> (https://deve
Ping.
I will change the tests to use the exact input and output registers as
Richard Sandiford suggested for the AArch64 patches.
On 12/20/19 6:46 PM, Delia Burduv wrote:
> This patch adds the ARMv8.6 ACLE BFloat16 store intrinsics
> vst{q}_bf16 as part of the BFloat16 extension.
> (https://dev
On Wed, Jan 22, 2020 at 12:10:12PM -0500, Jason Merrill wrote:
> On 1/22/20 11:33 AM, Marek Polacek wrote:
> > Here's what I did. Unfortunately it doesn't fix PR79620.
>
> No, I wouldn't expect it to, that looks like more of an issue with how
> is_lambda_ignored_entity skips captures in unevaluat
Ping.
I have read Richard Sandiford's comments on the AArch64 patches and I
will apply what is relevant to this patch as well. Particularly, I will
change the tests to use the exact input and output registers and I will
change the types of the rtl patterns.
On 12/20/19 6:44 PM, Delia Burduv wr
On 1/22/20 11:19 AM, Patrick Palka wrote:
On Tue, 21 Jan 2020, Jason Merrill wrote:
On 1/20/20 8:06 PM, Patrick Palka wrote:
The leak in get_mapped_args is due to auto_vec not properly supporting
destructible elements, in that auto_vec's destructor doesn't call the
destructors of its elements.
On 1/22/20 11:33 AM, Marek Polacek wrote:
Here's what I did. Unfortunately it doesn't fix PR79620.
No, I wouldn't expect it to, that looks like more of an issue with how
is_lambda_ignored_entity skips captures in unevaluated context.
+ cp_cv_quals quals;
+ if (DECL_NONSTATIC_MEMBER_FUNCT
On Wed, 22 Jan 2020, Stefan Schulze Frielinghaus wrote:
> Hi David,
>
> In function `tree_cmp` an invariant [1] is assumed which does not necessarily
> exist. In case both input trees are finally compared via `strcmp`, then
>
> tree_cmp (t1, t2) == -tree_cmp (t2, t1)
>
> does not hold in g
Hi David,
In function `tree_cmp` an invariant [1] is assumed which does not necessarily
exist. In case both input trees are finally compared via `strcmp`, then
tree_cmp (t1, t2) == -tree_cmp (t2, t1)
does not hold in general, since function `strcmp (x, y)` guarantees only that a
negative integ
On 22/01/2020 16:28, Marek Polacek wrote:
On Wed, Jan 22, 2020 at 04:05:37PM +, Richard Sandiford wrote:
"Richard Earnshaw (lists)" writes:
On 21/01/2020 17:20, Jason Merrill wrote:
On 1/21/20 10:40 AM, Richard Earnshaw (lists) wrote:
On 21/01/2020 15:39, Jakub Jelinek wrote:
On Tue, Ja
On Tue, Jan 21, 2020 at 10:40:09PM -0500, Jason Merrill wrote:
> On 1/21/20 9:08 PM, Marek Polacek wrote:
> > Here the problem is that if the noexcept specifier is used in the context
> > of a const member function, const is not considered for the member
> > variables,
> > leading to a bogus error
On Wed, Jan 22, 2020 at 04:05:37PM +, Richard Sandiford wrote:
> "Richard Earnshaw (lists)" writes:
> > On 21/01/2020 17:20, Jason Merrill wrote:
> >> On 1/21/20 10:40 AM, Richard Earnshaw (lists) wrote:
> >>> On 21/01/2020 15:39, Jakub Jelinek wrote:
> On Tue, Jan 21, 2020 at 03:33:22PM
On 1/22/20 10:22 AM, Paolo Carlini wrote:
Hi,
in this simple issue we either wrongly talked about variable template-id
in c++17 mode or ICEd in c++2a. I think we simply want to handle
concept-ids first, both as represented in c++17 mode and as represented
in c++2a mode. Tested x86_64-linux.
On Tue, 21 Jan 2020, Jason Merrill wrote:
On 1/20/20 8:06 PM, Patrick Palka wrote:
The leak in get_mapped_args is due to auto_vec not properly supporting
destructible elements, in that auto_vec's destructor doesn't call the
destructors of its elements.
Hmm, perhaps vec should static_assert __
Hi Steve,
On 1/22/20 4:54 PM, Steve Kargl wrote:
Supposedly, one can use attr.is_c_interop and attr.is_iso_c to
determine if something is interoperable (if it is consistently set)
and if it is from the ISO C Binding module.
I think this is fixable on two sides:
* For the used tree type, one
"Richard Earnshaw (lists)" writes:
> On 21/01/2020 17:20, Jason Merrill wrote:
>> On 1/21/20 10:40 AM, Richard Earnshaw (lists) wrote:
>>> On 21/01/2020 15:39, Jakub Jelinek wrote:
On Tue, Jan 21, 2020 at 03:33:22PM +, Richard Earnshaw (lists)
wrote:
>> Some examples would be us
On Wed, Jan 22, 2020 at 04:43:49PM +0100, Tobias Burnus wrote:
> Hi Steve,
>
> On 1/22/20 4:02 PM, Steve Kargl wrote:
> > On Wed, Jan 22, 2020 at 11:59:12AM +0100, Tobias Burnus wrote:
> >> And, additionally, about missing diagnostic related to (2) bind(c) and
> >> kind=4,
> > Are you sure there i
Hi Steve,
On 1/22/20 4:02 PM, Steve Kargl wrote:
On Wed, Jan 22, 2020 at 11:59:12AM +0100, Tobias Burnus wrote:
And, additionally, about missing diagnostic related to (2) bind(c) and
kind=4,
Are you sure there is a missing diagnostic? You need to
add -Wc-binding-type or -Wall to your command
The GCC git repository makes use of some non-standard refs so that by
default only the most active and useful branches and tags are fetched by
a git clone. This patch documents the additional branch and tag spaces
that exist on the server.
Joseph, have I got all of these right?
R.
diff --git
The motivation behind this change is to make it easier for a user to
link against static libraries on a target where dynamic libraries are
the default library type (for example GNU/Linux).
Further, my motivation is really for linking libraries into GDB,
however, the binutils-gdb/config/ directory
Hi,
in this simple issue we either wrongly talked about variable template-id
in c++17 mode or ICEd in c++2a. I think we simply want to handle
concept-ids first, both as represented in c++17 mode and as represented
in c++2a mode. Tested x86_64-linux.
Thanks, Paolo.
///
F
On Wed, Jan 22, 2020 at 11:59:12AM +0100, Tobias Burnus wrote:
>
> And, additionally, about missing diagnostic related to (2) bind(c) and
> kind=4,
>
Are you sure there is a missing diagnostic? You need to
add -Wc-binding-type or -Wall to your command line.
subroutine p(c) bind(c)
use iso_c
The early front-end only implementation of -Wrestrict that's still
present in GCC 10 issues a false postive for %p arguments that are
the same as the destination. Bug 84919 reports an instance of this
false positive in the Linux kernel.
That attached patch suppresses the front-end warning for th
On Wed, Jan 22, 2020 at 08:06:03AM -0600, Segher Boessenkool wrote:
> On Mon, Jan 06, 2020 at 09:24:10AM +0100, Jakub Jelinek wrote:
> > 1) always use gen_add3_insn, but if it returns NULL use the extra
> > emit_move_insn (end_addr, GEN_INT (-rounded_size));
> > and retry with end_addr instead of r
On Wed, Jan 22, 2020 at 11:59 AM Richard Sandiford
wrote:
>
> cfgexpand sorts variables by decreasing size, so when merging a later
> variable into an earlier one, there's usually no need to update the
> merged size.
>
> But for poly_int sizes, the sort function just uses a lexicographical
> compa
Hi!
On Mon, Jan 06, 2020 at 09:24:10AM +0100, Jakub Jelinek wrote:
> 1) always use gen_add3_insn, but if it returns NULL use the extra
> emit_move_insn (end_addr, GEN_INT (-rounded_size));
> and retry with end_addr instead of rs
> 2) if the first gen_add3_insn returned NULL or if it created more t
On Wed, 22 Jan 2020, Richard Earnshaw (lists) wrote:
> We're missing a statement on the main news feed about the git transition.
Lovely, thanks.
Gerald
PS: Lovely referring to you creating the patch, not the missing
announcement. ;-)
This fixes the validity check for redundant store removal by looking
at the actual stmt that represents the last (relevant) change to
the TBAA state rather than the imperfectly tracked alias set in the
expression hash table which then becomes unused (see a followup change
to get rid of that).
On
On Wed, Jan 22, 2020 at 10:00:00AM +, Richard Earnshaw (lists) wrote:
> On 21/01/2020 19:26, Segher Boessenkool wrote:
> >On Tue, Jan 21, 2020 at 02:52:00PM +, Richard Earnshaw (lists) wrote:
> >>+ A brief summary
> >
> >You could stress that this is the one thing that really matters. And
> On 1/22/20 12:09 PM, Martin Liška wrote:
> > stats for indirect_call:
> > total: 9210
> > invalid: 600
> > tracked values:
> > 0 values: 6280 times (68.19%)
> > 1 values: 1856 times (20.15%)
> > 2 values: 264 times (2.87%)
> > 3 values: 157 times (1.
On 1/22/20 12:09 PM, Martin Liška wrote:
stats for indirect_call:
total: 9210
invalid: 600
tracked values:
0 values: 6280 times (68.19%)
1 values: 1856 times (20.15%)
2 values: 264 times (2.87%)
3 values: 157 times (1.70%)
4 values: 53 tim
On 1/22/20 1:19 PM, Jan Hubicka wrote:
I think you can save an instruction in the loop if you drop "&& empty_counter ==
-1"
My motivation was to find a first empty slot.
Which means we'll have less holes, but it's probably a micro-optimization.
I'm going to install the patch without the "&& e
These tests are supposed to be testing the tlsdesc handling and
so don't apply to emultls targets.
Tested on aarch64-linux-gnu and aarch64_be-elf, applied.
Richard
2020-01-22 Richard Sandiford
gcc/testsuite/
* gcc.target/aarch64/sve/tls_preserve_1.c: Require tls_native.
* gc
A pasto in this test meant that we needed extra reverse instructions
for big-endian targets.
Tested on aarch64-linux-gnu and aarch64_be-elf, applied.
Richard
2020-01-22 Richard Sandiford
gcc/testsuite/
* gcc.target/aarch64/sve/sel_3.c (permute_vnx4sf): Take __SVFloat32_t
rat
In r279588 I'd for some reason only patched g++.dg/ext/sve-sizeless-2.C,
even though g++.dg/ext/sve-sizeless-1.C has the same problem.
Tested on aarch64-linux-gnu and aarch64_be-elf, applied.
Richard
2020-01-22 Richard Sandiford
gcc/testsuite/
* g++.dg/ext/sve-sizeless-1.C: Don't ex
> 2020-01-22 Martin Liska
>
> PR tree-optimization/92924
> * libgcov-profiler.c (__gcov_topn_values_profiler_body): First
> try to find an existing value, then find an empty slot
> if not found.
This looks good, one nit below.
> ---
> libgcc/libgcov-profiler.c | 46 +++
In this PR, auto-inc-dec was trying to turn:
(set (reg X) (plus (reg X) (const_int N)))
(clobber (mem (reg X)))
into:
(clobber (mem (pre_modify (reg X) ...)))
But bare clobber insns are just there to describe dataflow. They're
not supposed to generate any code.
This is part 2 of t
On 1/22/20 12:27 PM, Jan Hubicka wrote:
Hi.
I've updated the patch based on Honza's notes and installed it
as 5f32f9cf13f99f6295591927950aaf98aa8dba91.
Thanks, you omitted the patch, but looking at git log I think there is
still problem I commented on in the previous mail:
Whoops, sorry for t
One consequence of r276318 was that cselib now preserves sp-based
values across function calls. This in turn convinced cprop to
replace the clobber in:
(set (reg PSUEDO) (reg sp))
...
(call ...)
...
(clobber (mem:BLK (reg sp)))
with:
(clobber (mem:BLK (reg PSEUDO)))
But I dou
Hi!
On Mon, Jan 06, 2020 at 02:46:30PM -0600, Segher Boessenkool wrote:
> > > Another option would be to:
> > > 1) always use gen_add3_insn, but if it returns NULL use the extra
> > > emit_move_insn (end_addr, GEN_INT (-rounded_size));
> > > and retry with end_addr instead of rs
> > > 2) if the fi
On 1/22/20 8:32 AM, Richard Biener wrote:
On Tue, 21 Jan 2020, Alexander Monakov wrote:
On Tue, 21 Jan 2020, Richard Biener wrote:
Fourth. That PNVI (I assume it's the whole pointer-provenance stuff)
wants to get the "best" of both which can never be done since a compiler
needs to have a way
We were incorrectly assuming a merge operation is conservative enough
for not explicitely handled operations but we also need to consider
offsetting within fields when field-sensitive analysis applies.
Bootstrapped and tested on x86_64-unknown-linux-gnu, pushed.
Richard.
2020-22 Richard Biene
Hi,
Though function co_await_expander may need to be further revised, this simple
patch fixes an ICE case in co_await_expander,
Handle CO_AWAIT_EXPR in conversion in co_await_expander.
Function co_await_expander expands CO_AWAIT_EXPR and inserts expanded
code before result of c
> Hi.
>
> I've updated the patch based on Honza's notes and installed it
> as 5f32f9cf13f99f6295591927950aaf98aa8dba91.
Thanks, you omitted the patch, but looking at git log I think there is
still problem I commented on in the previous mail:
for (unsigned i = 0; i < GCOV_TOPN_VALUES; i++)
Hi.
I've updated the patch based on Honza's notes and installed it
as 5f32f9cf13f99f6295591927950aaf98aa8dba91.
Just for the record, there are stats for cc1plus for PGO bootstrap:
before:
== Stats for gcc ==
stats for indirect_call:
total: 9210
invalid: 760
tracked values:
0 values:
cfgexpand sorts variables by decreasing size, so when merging a later
variable into an earlier one, there's usually no need to update the
merged size.
But for poly_int sizes, the sort function just uses a lexicographical
comparison of the coefficients, so e.g. 2X+2 comes before 0X+32.
Which is big
Hi Thomas, hi all,
first, I have now attached a different fix for PR 85781 (= original
bug). Can you have a look?
I have the feeling (but didn't check) that your patch does not handle
the following variant of the test case: "print *, x(m:n)" (i.e. the
lower bound is not known at compile time
Jakub Jelinek writes:
> Hi!
>
> The two patterns that call aarch64_expand_subvti ensure that {low,high}_in1
> is a register, while {low,high}_in2 can be a register or immediate.
> subdi3_compare1_imm uses the aarch64_plus_immediate predicate for its last
> two operands (the value and negated value
1 - 100 of 126 matches
Mail list logo
