Another maybe_record_trace_start issue found by my tester. This time
the rx port.
We failed to add suitable notes for the CFI machinery. There's still
paths (particularly those which save/restore the accumulators) that may
need further adjustment. However, this patch is enough to get
libgcc/ne
On Fri, Feb 16, 2018 at 06:05:13PM -0700, Jeff Law wrote:
> > From: Segher Boessenkool
> > Date: Thu, 8 Feb 2018 16:58:33 +
> > Subject: [PATCH] Do not mess with rs6000_{single,double}_float (PR68028)
> >
> > For e500 family cores we do some questionable things with those flags,
> > which doe
On 02/09/2018 05:14 PM, Segher Boessenkool wrote:
> Hi!
>
> On Fri, Feb 09, 2018 at 09:37:28AM +, Nick Clifton wrote:
>>> I thought you were going to do a patch like the following, to make the
>>> e500 cores less special (they are not):
>>
>> Sorry - my bad. I defer to your patch then. Whate
On Thu, 2018-02-15 at 14:01 +, Richard Earnshaw (lists) wrote:
>
> Wouldn't it be better to call output_operand_lossage() with a suitable
> diagnostic message? If the operand isn't in Pmode assembly will
> (should) fail anyway.
>
> R.
How about this patch? In addtion to the code change I u
On 02/14/2018 04:57 AM, Jakub Jelinek wrote:
> On Wed, Feb 14, 2018 at 12:52:45PM +0100, Richard Biener wrote:
>> On Tue, 13 Feb 2018, Jeff Law wrote:
>>
>>> On 01/30/2018 02:59 AM, Richard Biener wrote:
This patch tries to deal with the "easy" part of a function ABI,
the return valu
On 02/14/2018 06:49 AM, Tsimbalist, Igor V wrote:
> ENDBR32 and RDSSPD are multi-byte NOPs on x86-64 processors and
> newer x86 processors, starting Pentium Pro. They are UD on older 32-bit
> processors. Detect this at configure time and adjust the default value
> for enable_cet. GCC will enable C
On 02/16/2018 05:48 PM, Jason Merrill wrote:
We hit the bug in this testcase because build_over_call replaces the
constructor with a dispatcher function, and then build_aggr_init_expr
doesn't recognize it as a constructor.
Messing with the DECL_NAME of the dispatcher seems messy, but copying
the
On 02/15/2018 10:47 AM, Martin Sebor wrote:
> On 02/13/2018 11:14 PM, Jeff Law wrote:
>> On 02/01/2018 04:45 PM, Martin Sebor wrote:
>>> The previous patch didn't resolve all the false positives
>>> in the Linux kernel. The attached is an update that fixes
>>> the remaining one having to do with m
On 02/16/2018 03:13 AM, Koval, Julia wrote:
> Hi,
> This is small fix for documentation - it adds CLWB to skylake-avx512 and
> removes it from cannonlake.
>
> gcc/
> * doc/invoke.texi (Skylake Server): Add CLWB.
> (Cannonlake): Remove CLWB.
OK.
jeff
The gnat.dg testsuite contains a dozen of spurious failures on 64-bit Windows
because of a couple of issues: the target is P64 and some patterns in filename
give rise to warning at link time.
Fixed thusly, applied to all active branches.
2018-02-16 Eric Botcazou
PR ada/84277
On 02/16/2018 03:49 PM, Jakub Jelinek wrote:
> Hi!
>
> When debugging a wrong-code in miniruby on aarch64, I've noticed a hang
> in inline_small_functions when adding __attribute__((optimize(0)))
> to one of the functions.
>
> I don't know what exactly the code is doing, but it feels like a typo
On Fri, Feb 16, 2018 at 02:14:09PM -0700, Martin Sebor wrote:
> On 02/16/2018 01:44 PM, Jakub Jelinek wrote:
> > On Fri, Feb 16, 2018 at 09:25:30PM +0100, Richard Biener wrote:
> > > But the broken compilers will overwrite the memset data with possibly
> > > uninitialized fields of a TARGET_EXPR.
Hi!
When debugging a wrong-code in miniruby on aarch64, I've noticed a hang
in inline_small_functions when adding __attribute__((optimize(0)))
to one of the functions.
I don't know what exactly the code is doing, but it feels like a typo for
me, at least I'd expect the intent is for n2 to walk th
We hit the bug in this testcase because build_over_call replaces the
constructor with a dispatcher function, and then build_aggr_init_expr
doesn't recognize it as a constructor.
Messing with the DECL_NAME of the dispatcher seems messy, but copying
the DECL_CXX_CONSTRUCTOR_P flag is easy. So we ca
Hi!
For ref linear modifier, we don't want to set OMP_CLAUSE_LINEAR_STEP
to the last_step constant converted to the scalar type, but to
a sizetype last_step times the size of the referenced type.
I'm trying to resolve on omp-lang what to do about arrays with descriptors
or scalar allocatables/poin
Hi,
here we ICE during error recovery when, after emitting a correct error
from grokdeclarator, we go on, we only clear friendp, and grokfield
proceeds to call cp_finish_decl where 'gcc_assert
(CLASS_PLACEHOLDER_TEMPLATE (auto_node));' triggers. We could imagine
solving the problem in various
On 02/16/2018 07:04 AM, Jason Merrill wrote:
On Thu, Feb 15, 2018 at 6:36 PM, Martin Sebor wrote:
A failed template deduction in template member of a template
triggers an ICE with -std=c++17 due to what seems like
a missing handling of invalid input. Replacing
the gcc_unreachable() call that c
On 02/16/2018 01:44 PM, Jakub Jelinek wrote:
On Fri, Feb 16, 2018 at 09:25:30PM +0100, Richard Biener wrote:
But the broken compilers will overwrite the memset data with possibly
uninitialized fields of a TARGET_EXPR.
Perhaps for hash-table.h we could use:
#ifndef BROKEN_VALUE_INITIALIZATION
On Mon, Feb 5, 2018 at 1:45 PM, Jason Merrill wrote:
> On Mon, Feb 5, 2018 at 8:37 AM, Marek Polacek wrote:
>> On Fri, Feb 02, 2018 at 02:11:27PM -0500, Jason Merrill wrote:
>>> On Thu, Jan 25, 2018 at 4:16 PM, Marek Polacek wrote:
>>> > This is a similar problem to 83116: we'd cached a constexp
Now that structured bindings are another use of DECL_VALUE_EXPR,
is_capture_proxy needs to recognize them.
Tested x86_64-pc-linux-gnu, applying to trunk.
commit 275feb77e153ff34289b1fc381d4da52bb68dd8e
Author: Jason Merrill
Date: Fri Feb 16 15:09:56 2018 -0500
PR c++/84420 - ICE wi
We don't yet support constexpr evaluation of pre-generic statements,
though that's a goal; in the mean time, it's trivial to ignore
USING_STMT.
Tested x86_64-pc-linux-gnu, applying to trunk.
commit 21cab302d01fddc26b26f6ad6286f235322728e1
Author: Jason Merrill
Date: Fri Feb 16 15:19:25 2018 -05
In an expression like a.f(), if f is a static member function, a is
still evaluated. But if a is a volatile lvalue, we don't want to do
the discarded-value load; what's really evaluated is the 'this'
argument, which is the address of a.
Tested x86_64-pc-linux-gnu, applying to trunk and 7.
commit
On Fri, Feb 16, 2018 at 01:54:15PM -0500, Jason Merrill wrote:
> My dejagnu doesn't like "cc1plus" as a line number:
>
> ERROR: g++.old-deja/g++.jason/scoping15.C -std=gnu++98: expected
> integer but got "cc1plus:" for " dg-warning 25 "ffriend-injection.* is
> deprecated" "" { target *-*-* } cc1p
On Fri, Feb 16, 2018 at 09:25:30PM +0100, Richard Biener wrote:
> But the broken compilers will overwrite the memset data with possibly
> uninitialized fields of a TARGET_EXPR.
Perhaps for hash-table.h we could use:
#ifndef BROKEN_VALUE_INITIALIZATION
for ( ; size; ++entries, --size)
On Fri, Feb 16, 2018 at 09:25:30PM +0100, Richard Biener wrote:
> But the broken compilers will overwrite the memset data with possibly
> uninitialized fields of a TARGET_EXPR.
For the vec.h case no, there is no temporary involved, it is constructed
directly into the memory.
For the hash-table.
On February 16, 2018 8:52:18 PM GMT+01:00, Jakub Jelinek
wrote:
>On Fri, Feb 16, 2018 at 12:47:31PM -0700, Martin Sebor wrote:
>> hoping it can be revisited in stage 1. Ultimately, we need
>> to decide if vec really needs to be a POD (or trivial or
>
>Yes, definitely. This has been discussed se
On February 16, 2018 7:05:10 PM GMT+01:00, Jakub Jelinek
wrote:
>On Fri, Feb 16, 2018 at 10:36:27AM -0700, Martin Sebor wrote:
>> The deeper problem is that vec wants to be a POD but GCC
>> instantiates the template on non-POD types. For example
>> genrecog.c instantiates it on struct parameter
The C++17 path in build_special_member_call converts to the desired
type and then builds an INIT_EXPR to initialize the object. Since
this does not involve actually creating a temporary, we should set
TARGET_EXPR_DIRECT_INIT_P.
Tested x86_64-pc-linux-gnu, applying to trunk.
commit f2cf58eab564e32
GCC maintainers:
The following patch fixes a GCC internal compiler error on the test
files:
gcc/testsuite/gcc.target/powerpc/fold-vec-neg-longlong.p8.c
gcc/testsuite/gcc.target/powerpc/fold-vec-neg-longlong.p9.c
when compiled with -mcpu=power7.
See bugzilla https://gcc.gnu.org/bugzilla/sho
On Fri, Feb 16, 2018 at 12:47:31PM -0700, Martin Sebor wrote:
> hoping it can be revisited in stage 1. Ultimately, we need
> to decide if vec really needs to be a POD (or trivial or
Yes, definitely. This has been discussed several times.
Jakub
On Fri, Feb 16, 2018 at 10:36:27AM -0700, Martin Sebor wrote:
> The deeper problem is that vec wants to be a POD but GCC
> instantiates the template on non-POD types. For example
> genrecog.c instantiates it on struct parameter that has
> a default ctor.
>
> A number of ipa-*.c files instantiate
On 02/16/2018 11:05 AM, Jakub Jelinek wrote:
On Fri, Feb 16, 2018 at 10:36:27AM -0700, Martin Sebor wrote:
The deeper problem is that vec wants to be a POD but GCC
instantiates the template on non-POD types. For example
genrecog.c instantiates it on struct parameter that has
a default ctor.
A
On 02/16/2018 01:54 PM, Jason Merrill wrote:
My dejagnu doesn't like "cc1plus" as a line number:
ERROR: g++.old-deja/g++.jason/scoping15.C -std=gnu++98: expected
integer but got "cc1plus:" for " dg-warning 25 "ffriend-injection.* is
deprecated" "" { target *-*-* } cc1plus: "
I think I got dis
84015 seems likely to be related.
On Fri, Feb 16, 2018 at 9:04 AM, Jason Merrill wrote:
> On Thu, Feb 15, 2018 at 6:36 PM, Martin Sebor wrote:
>> A failed template deduction in template member of a template
>> triggers an ICE with -std=c++17 due to what seems like
>> a missing handling of invali
On 02/16/2018 01:52 PM, Sandra Loosemore wrote:
The GCC documentation conventions say to use American spellings, not
British so in the patch hunk please
s/behaviour/behavior/g
American imperialist :)
nathan
--
Nathan Sidwell
2018-02-16 Nathan Sidwell
* doc/extend.texi (Backwards C
The use_partial_inst_tmpl code in lookup_template_class didn't work as
intended because when converting the ref-to-fn parameter to the same
type we wrapped it in a couple of extra NOP_EXPR.
Tested x86_64-pc-linux-gnu, applying to trunk and 7.
commit ac1c39476b98ab4b6b52952df5baa07cf923d4e0
Author:
My dejagnu doesn't like "cc1plus" as a line number:
ERROR: g++.old-deja/g++.jason/scoping15.C -std=gnu++98: expected
integer but got "cc1plus:" for " dg-warning 25 "ffriend-injection.* is
deprecated" "" { target *-*-* } cc1plus: "
On Fri, Feb 16, 2018 at 10:44 AM, Nathan Sidwell wrote:
> This p
On 02/16/2018 08:44 AM, Nathan Sidwell wrote:
Index: doc/extend.texi
===
--- doc/extend.texi (revision 257739)
+++ doc/extend.texi (working copy)
@@ -23881,11 +23881,23 @@ deprecated. @xref{Deprecated Features}
@table @c
On Fri, Feb 16, 2018 at 08:52:10AM -0500, Jason Merrill wrote:
> On Fri, Feb 16, 2018 at 3:28 AM, Jakub Jelinek wrote:
> > pop_stmt_list, if there is just a single stmt inside statement expression
> > moves the stmt out of the STATEMENT_LIST (and I think cp_fold does too).
> > We only initialize j
This is OK. In the future, please include the Changelog entry as a
separate text, not part of the patch, as it will rarely apply cleanly.
"Sebastian Perta" writes:
>
> * config/rl78/rl78.md (movdf): New define expand.
On Fri, 2018-02-16 at 10:21 -0800, Carl Love wrote:
>
> >
> > Do you have a new testcase, too? Or is this all covered by
> > existing
> > tests (which?)
>
> The issue was found when compiling the existing test case file with
> -mcpu=power7. No additional tests are needed for this issue.
>
So
On Fri, 2018-02-16 at 11:56 -0600, Segher Boessenkool wrote:
> Hi!
Segher:
Per our discussion on IRC. I created the bugzilla:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=84422
for this issue so it is visible to people outside IBM.
I have included this test file int the above bugzilla along w
The following patch fixes
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70023
The patch was successfully bootstrapped and tested on x86-64.
Committed as rev. 257751.
Index: ChangeLog
===
--- ChangeLog (revision 257750)
+++ ChangeL
Patch committed as obvious at revision r257750.
Dominique
--- trunk/gcc/fortran/decl.c2018/02/16 17:42:08 257749
+++ trunk/gcc/fortran/decl.c2018/02/16 18:01:02 257750
@@ -3302,7 +3302,7 @@
kind_expr = gfc_copy_expr (c1->initializer);
else if (!(actua
Hi!
On Fri, Feb 16, 2018 at 08:58:49AM -0800, Carl Love wrote:
> This patch changes the expansion macros to the equivalent Power 8
> macros. The test now cleanly exits with the message
>
> error: builtin function ‘__builtin_vsx_float2_v2di’ requires the ‘-
> mpower8-vector’ option
>
>
On 15/02/18 07:25, Uros Bizjak wrote:
On Wed, Feb 14, 2018 at 7:04 PM, Kyrill Tkachov
wrote:
On 13/02/18 16:45, Jeff Law wrote:
On 02/09/2018 07:50 AM, Kyrill Tkachov wrote:
Hi Uros,
On 08/02/18 22:54, Uros Bizjak wrote:
On Thu, Feb 8, 2018 at 6:11 PM, Kyrill Tkachov
wrote:
Hi all,
Th
On 02/16/2018 01:38 AM, Jakub Jelinek wrote:
Hi!
As the system.h comment says, it took us years to get value initialization
right, and unfortunately we've started hitting it in GCC codebase now.
There are various related PRs, 4.2.[0-3] are most broken in this regard,
value initialization of non-
From: Vladimir Mezentsev
Ramana Radhakrishnan made a workaround in gcc/config/aarch64/aarch64.c to
resolve
bootstrap comparison failure (2015-11-10, commit
bc443a71dafa2e707bae4b2fa74f83b05dea37ab).
The real bug is in gcc/varasm.c.
hash_section() returns an unstable value.
As result, two blocks
On 02/16/2018 01:14 AM, Tom de Vries wrote:
On 02/15/2018 07:01 PM, Martin Sebor wrote:
Thanks. I've seen these fixes before but I still keep forgetting
to add the directive.
Hi,
I also forget about these directives when I do an x86_64 build-test-commit.
The alloca effective target is one o
GCC maintainers:
The following patch fixes a GCC internal compiler error.
Several of the builtins use the wrong enable conditions as mentioned in
Bill's issue data base (https://github.ibm.com/wschmidt/power-gcc/issue
s/282). The implementation for the vec_float2() builtin, as
implemented, uses
We can't judge whether the condition is a constant-expression until
we've converted it to bool; if it's type-dependent, let's leave that
checking for later.
Tested x86_64-pc-linux-gnu, applying to trunk.
commit b1fa979536b793b6bb1b9262e65ba6c753d5802b
Author: Jason Merrill
Date: Fri Feb 16 10:5
output_constructor_field doesn't like to see two fields at offset zero
with non-zero size. Fixed by giving the empty base field size zero.
Tested x86_64-pc-linux-gnu, applying to trunk.
commit c5e03948968ee268d4bf9f1f7ceb8613a5b0f2a4
Author: Jason Merrill
Date: Fri Feb 16 11:23:57 2018 -0500
The gofrontend code only preempts goroutines in calls into the
runtime. This patch adds some more checks for preemption, to catch
some more cases of loops that would otherwise run without
interruption, possibly hanging a program if they are blocking some
other goroutine that could change the exit
On 02/16/2018 05:43 AM, Nick Clifton wrote:
Hi David,
Attached is a revised version of the patch which I hope addresses all
of your (very helpful) comments on the v3 patch.
OK to apply once the sources are back on stage 1 ?
Thanks (also) for adding the documentation!
I have just one mi
On Fri, Feb 16, 2018 at 11:58:07AM +0100, Martin Liška wrote:
> Btw. do we want such patch to be backported?
No.
Jakub
Hi!
On Thu, Feb 15, 2018 at 03:56:10PM -0600, Will Schmidt wrote:
>This fixes the scan-assembler errors as seen on power9. While verifying
> that fix and ensuring coverage, cleaned up some nearby bits as well.
> 2018-02-15 Will Schmidt
>
> target/pr84371
> * gcc.target/powerpc
This patch deprecates the -ffriend-injection option.
1) issue a deprecated warning, if the option is set.
2) if we do inject a visible friend, issue a warning. This warning is
not inhibitable, like the for-scope behaviour. (Notice we never
injected visible friend classes, only functions).
3)
On 22/01/18 15:23, Richard Biener wrote:
On Mon, Jan 22, 2018 at 4:10 PM, Sudakshina Das wrote:
Hi
This is a patch to backport r256526 and r256941 (Fix case fix) of trunk to
fix emit_store_flag_force () function to fix the ICE. The original
discussion is at https://gcc.gnu.org/ml/gcc-patches/2
On Fri, 2018-02-16 at 07:14 -0500, Trevor Saunders wrote:
> On Thu, Feb 15, 2018 at 04:04:45PM -0500, David Malcolm wrote:
> > On Fri, 2018-02-09 at 13:01 +, Nick Clifton wrote:
> > > +class escaped_string
> > > +{
> > > + public:
> > > + escaped_string () { m_owned = false; m_str = NULL; };
>
On Fri, 2018-02-16 at 12:43 +, Nick Clifton wrote:
> Hi David,
>
> Attached is a revised version of the patch which I hope addresses
> all
> of your (very helpful) comments on the v3 patch.
>
> OK to apply once the sources are back on stage 1 ?
Yes.
Thanks!
Dave
> Cheers
> Nick
>
Richard Biener wrote:
> With Ooo CPUs speculatively executing the next iterations I very much doubt
> that.
OoO execution is like really dumb loop unrolling, you still have all the
dependencies
between iterations, all the branches, all the pointer increments etc.
Optimizing those
reduces instr
Hi Yuri,
On Thu, Feb 15, 2018 at 08:13:38PM +, Yury Gribov wrote:
> This is an updated version of patch for PR 81535. The patch was last
> discussed in November. The new version addresses Segher's comments in
> https://gcc.gnu.org/ml/gcc-patches/2017-08/msg00558.html
> FTR here's an excerpt
On Fri, Feb 16, 2018 at 07:07:53AM +, Paul Richard Thomas wrote:
> > Oddly, the failing test in associate_35.f90 is the only one that works
> > in 7-branch. I have left the PR open and changed the title
> > accordingly.
>
> The attached patch fixes this. OK for trunk?
>
Ok.
--
steve
OK.
On Fri, Feb 16, 2018 at 10:19 AM, Jakub Jelinek wrote:
> On Fri, Feb 16, 2018 at 08:52:10AM -0500, Jason Merrill wrote:
>> On Fri, Feb 16, 2018 at 3:28 AM, Jakub Jelinek wrote:
>> > pop_stmt_list, if there is just a single stmt inside statement expression
>> > moves the stmt out of the STATE
On February 16, 2018 3:22:22 PM GMT+01:00, Wilco Dijkstra
wrote:
>Richard Biener wrote:
>>> This is a great plan - GCC urgently requires a good unroller!
>>
>> How so?
>
>I thought it is well-known for many years that the rtl unroller doesn't
>work properly.
>In practically all cases where LLVM b
Hi!
> This is the second patch that removes the non-ABI vec_insert4b and
> vec_extract4b builtin support. It also removes the two existing test
> files for the non-ABI builtin instances. A runnable test file for the
> ABI specified builtins was added by the first patch.
Looks good too; okay for
Hi all,
This patch adds Richard's march and mcpu option work to changes.html.
It is a helpful extension of the existing functionality and is certainly
user-visible
so we want to advertise it.
Checked that it looks ok in Firefox.
Note that the link to the documentation will need to be updated o
This fixes an ICE after attempted friend declaration in a local class.
We were going onto push it into the local binding, but that's a binding
in the class being defined, because . Just bail out immediately.
nathan
--
Nathan Sidwell
2018-02-16 Nathan Sidwell
PR c++/84375
* name-lookup.c
Hi!
On Wed, Feb 14, 2018 at 12:08:27PM -0800, Carl Love wrote:
> Per Segher's comments on the first version of the patch. I split the
> patch into two.
Thanks, much easier to read.
> 2018-02-13 Carl Love
>
> * config/rs6000/altivec.h: Add builtin names vec_extract4b
> ve
Richard Biener wrote:
>> This is a great plan - GCC urgently requires a good unroller!
>
> How so?
I thought it is well-known for many years that the rtl unroller doesn't work
properly.
In practically all cases where LLVM beats GCC, it is due to unrolling small
loops.
You may have noticed how p
Hello,
The following patch defines a new expand movdf, it is similar to the movdi
expand which I implemented a while ago.
The usefulness of the patch can be seen easily even on the following test
case where it reduces the code size from 199 bytes to 109 bytes:
long double long_double_add(long dou
OK.
On Fri, Jan 26, 2018 at 8:12 AM, Jonathan Wakely wrote:
> This is a backport of r251426 which incidentally fixed PR 81860 and
> its dup. The bug was closed as fixed, but as a regression it should be
> fixed for 7.x too.
>
> The patch applied cleanly to the branch except for a minor conflict i
On Thu, Feb 15, 2018 at 6:36 PM, Martin Sebor wrote:
> A failed template deduction in template member of a template
> triggers an ICE with -std=c++17 due to what seems like
> a missing handling of invalid input. Replacing
> the gcc_unreachable() call that causes the ICE with a return
> statement
On Wed, 2018-02-14 at 21:34 +0900, Oleg Endo wrote:
> > Approved - please apply - and thanks very much for doing this!
> Thanks! Committed as r257655.
>
Testing another patch
https://gcc.gnu.org/ml/gcc-patches/2018-02/msg00903.html
revealed a bug.
I've committed the attached somewhat obvious
On Thu, Feb 15, 2018 at 5:56 PM, Paolo Carlini wrote:
> this issue is very easy to explain and the testcase minimal:
> check_special_function_return_type ICEs on a TEMPLATE_TEMPLATE_PARM as
> optype, because it tries to use CLASSTYPE_TI_TEMPLATE on it. Today,
> triangulating with other compilers t
As Jakub found, after my recent combine patch at least on x86 problems
show up with RTL checking enabled. This is because the I2 generated
by a successful instruction combination can write not only a register
but it can also write a paradoxical subreg of one.
This fixes it.
Tested on powerpc64-l
On Fri, Feb 16, 2018 at 3:28 AM, Jakub Jelinek wrote:
> pop_stmt_list, if there is just a single stmt inside statement expression
> moves the stmt out of the STATEMENT_LIST (and I think cp_fold does too).
> We only initialize jump_target to non-NULL in cxx_eval_statement_list
> or for calls, so be
Hi David,
Attached is a revised version of the patch which I hope addresses all
of your (very helpful) comments on the v3 patch.
OK to apply once the sources are back on stage 1 ?
Cheers
Nick
gcc/ChangeLog
2018-02-09 Nick Clifton
PR 84195
* tree.c (escaped_string):
On Thu, Feb 15, 2018 at 04:04:45PM -0500, David Malcolm wrote:
> On Fri, 2018-02-09 at 13:01 +, Nick Clifton wrote:
> > +class escaped_string
> > +{
> > + public:
> > + escaped_string () { m_owned = false; m_str = NULL; };
> > + ~escaped_string () { if (m_owned) free (m_str); }
> > + operato
The following adds some more checks to the MEM_REF -> vector BIT_FIELD_REF
rewriting and adjusts another copy with the poly-int variant.
Bootstrapped and tested on x86_64-unknown-linux-gnu, applied to trunk.
Richard.
2018-02-16 Richard Biener
PR tree-optimization/84417
* tre
On Thu, Feb 15, 2018 at 11:30 PM, Kugan Vivekanandarajah
wrote:
> Hi Wilko,
>
> Thanks for your comments.
>
> On 14 February 2018 at 00:05, Wilco Dijkstra wrote:
>> Hi Kugan,
>>
>>> Based on the previous discussions, I tried to implement a tree loop
>>> unroller for partial unrolling. I would lik
On Fri, Feb 16, 2018 at 11:31:15AM +0100, Richard Biener wrote:
>
> After Jakubs fixes we can restore bootstrap with GCC 4.2.1 as provided
> by older Darwin hosts with the following patch.
>
> Bootstrap using a GCC 4.2.1 host compiler on x86_64-unknown-linux-gnu
> is now in stage2.
>
> Ok for tr
On Thu, Feb 15, 2018 at 11:07 PM, David Malcolm wrote:
> On Fri, 2018-02-09 at 12:02 +0100, Richard Biener wrote:
>> On Thu, Feb 8, 2018 at 11:23 PM, David Malcolm
>> wrote:
>> > PR tree-optimization/84178 reports a couple of source files that
>> > ICE inside
>> > ifcvt when compiled with -03 -fn
On 02/16/2018 10:55 AM, Jakub Jelinek wrote:
> On Fri, Feb 16, 2018 at 10:53:58AM +0100, Martin Liška wrote:
>> Hi.
>>
>> I'm sending patch that does what Jakub suggested.
>> Patch can bootstrap on ppc64le-redhat-linux and survives regression tests.
>> I consider the previous email as approval and
PING
On 15/12/17 11:57, Sudakshina Das wrote:
Hi
This patch fixes the inconsistent behavior observed at -O3 for the
unordered comparisons. According to the online docs
(https://gcc.gnu.org/onlinedocs/gcc-7.2.0/gccint/Unary-and-Binary-Expressions.html),
all of the following should not raise a
On Fri, Feb 16, 2018 at 09:48:25AM +0100, Richard Biener wrote:
> Ok. I'll see if fixing libcpp makes GCC 4.2.[0-3] usable as well.
Looking through r249234, dump_file_info is ok, because it is a POD,
ipcp_value/ipa_edge_args too, because it has a user provided ctor that
initializes everything, wi
On Fri, Feb 16, 2018 at 10:53:58AM +0100, Martin Liška wrote:
> Hi.
>
> I'm sending patch that does what Jakub suggested.
> Patch can bootstrap on ppc64le-redhat-linux and survives regression tests.
> I consider the previous email as approval and I'm going to install the patch.
Ok, thanks.
> >Fr
After Jakubs fixes we can restore bootstrap with GCC 4.2.1 as provided
by older Darwin hosts with the following patch.
Bootstrap using a GCC 4.2.1 host compiler on x86_64-unknown-linux-gnu
is now in stage2.
Ok for trunk?
Thanks,
Richard.
2018-02-16 Richard Biener
PR bootstrap/82939
Hi Kyrill,
Thanks for the review. Here's my second attempt. I've fixed the typo you raised
and added some tests (wasn't aware of those). I've also reworked the mapping
somewhat, it's explain in the description below.
Due to there being no multilib mapping for Armv8-R, default multilib
builts
Hi,
This is small fix for documentation - it adds CLWB to skylake-avx512 and
removes it from cannonlake.
gcc/
* doc/invoke.texi (Skylake Server): Add CLWB.
(Cannonlake): Remove CLWB.
Thanks,
Julia
doc_patch_16.2.18
Description: doc_patch_16.2.18
Hi.
I'm sending patch that does what Jakub suggested.
Patch can bootstrap on ppc64le-redhat-linux and survives regression tests.
I consider the previous email as approval and I'm going to install the patch.
Thanks,
Martin
>From b306ad522ed0f8a010647d0eb1d9c36b102e92c3 Mon Sep 17 00:00:00 2001
Fro
Hi!
pop_stmt_list, if there is just a single stmt inside statement expression
moves the stmt out of the STATEMENT_LIST (and I think cp_fold does too).
We only initialize jump_target to non-NULL in cxx_eval_statement_list
or for calls, so before we have a chance to diagnose the error of using
an ex
Hi!
I'd like to ping another patch:
PR84272 - P1 - fix cortex-a57-fma-steering.c
- http://gcc.gnu.org/ml/gcc-patches/2018-02/msg00539.html
Thanks
Jakub
On Fri, Feb 16, 2018 at 10:19 AM, Koval, Julia wrote:
> In latest ISE extensions document
> https://software.intel.com/sites/default/files/managed/c5/15/architecture-instruction-set-extensions-programming-reference.pdf
> CLWB was removed from Cannonlake. This patch removes it from GCC. Ok for
>
Sorry for the top-post, mobile client as I am out of office.
OK,
Thanks,
James
From: Jakub Jelinek
Sent: 09 February 2018 14:39:27
To: Richard Earnshaw; James Greenhalgh; Marcus Shawcroft
Cc: gcc-patches@gcc.gnu.org
Subject: [PATCH] cortex-a57-fma-steeri
In latest ISE extensions document
https://software.intel.com/sites/default/files/managed/c5/15/architecture-instruction-set-extensions-programming-reference.pdf
CLWB was removed from Cannonlake. This patch removes it from GCC. Ok for trunk?
Thanks,
Julia
gcc/
* config/i386/i386.c (ix86_
On Thu, Feb 15, 2018 at 8:13 PM, Yury Gribov
wrote:
> Hi all,
>
> This is an updated version of patch for PR 81535. The patch was last
> discussed in November. The new version addresses Segher's comments in
> https://gcc.gnu.org/ml/gcc-patches/2017-08/msg00558.html
Forgot to mention: I've added
Hi!
As the system.h comment says, it took us years to get value initialization
right, and unfortunately we've started hitting it in GCC codebase now.
There are various related PRs, 4.2.[0-3] are most broken in this regard,
value initialization of non-PODs doesn't actually initialize there anything
On Thu, Feb 15, 2018 at 11:08:20AM +, Szabolcs Nagy wrote:
> > --- libbacktrace/elf.c.jj 2018-02-08 20:46:10.671242369 +
> > +++ libbacktrace/elf.c 2018-02-14 08:39:06.674088951 +
> ...
> > @@ -2857,6 +2889,23 @@ elf_add (struct backtrace_state *state,
> > debuglink_crc
On Fri, 16 Feb 2018, Jakub Jelinek wrote:
> Hi!
>
> As the system.h comment says, it took us years to get value initialization
> right, and unfortunately we've started hitting it in GCC codebase now.
> There are various related PRs, 4.2.[0-3] are most broken in this regard,
> value initialization
1 - 100 of 104 matches
Mail list logo
