On Wed, Jan 24, 2018 at 12:15 AM, Jeff Law wrote:
>
> pr83994 is a code generation bug in the stack-clash support that affects
> openssl (we've turned on stack-clash-protection by default for the F28
> builds).
>
> The core problem is stack-clash (like stack-check)
The loop_align.c test has been broken for some time, since I put in patches to
eliminate some debug hooks (-mno-upper-regs-{df,di,sf}) that we deemed to no
longer be needed.
As Segher and I were discussing over private IRC, the root cause of this bug is
the compiler no long generates the BDNZ
On Dec 21, 2017, Jeff Law wrote:
> On 12/11/2017 07:54 PM, Alexandre Oliva wrote:
>> + ASM_GENERATE_INTERNAL_LABEL (label, "LVU", ied->view);
>> + /* FIXME: this will resolve to a small number. Could we
>> + possibly emit smaller data? Ideally
These two patches fix PR81611.
The first one improves forwprop so that we avoid adding SSA conflicting
by forwpropping the iv increment, which may cause both the incremented
and the original value to be live, even when the iv is copied between
the PHI node and the increment. We already handled
On 01/14/2018 12:16 PM, Mike Gulick wrote:
> On 01/12/2018 06:16 PM, David Malcolm wrote:
[snip]
>>
>> I was going to suggest renaming your new test to e.g.
>> location-overflow-test-pr83173.c
>> so that it matches the glob in those comments, but given that you refer
>> to the testname in
On Tue, Jan 23, 2018 at 5:24 PM, Matthias Klose wrote:
> On 17.01.2018 15:20, Ian Lance Taylor wrote:
>> This patch updates libgo to the Go1.10beta2 release. The complete
>> patch is too large to include in this e-mail message, mainly due to
>> some test changes. Bootstrapped
On 17.01.2018 15:20, Ian Lance Taylor wrote:
> This patch updates libgo to the Go1.10beta2 release. The complete
> patch is too large to include in this e-mail message, mainly due to
> some test changes. Bootstrapped and ran Go testsuite on
> x86_64-pc-linux-gnu. Committed to mainline.
gotools
pr83994 is a code generation bug in the stack-clash support that affects
openssl (we've turned on stack-clash-protection by default for the F28
builds).
The core problem is stack-clash (like stack-check) will emit a probing
loop if the prologue allocates enough stack space. When emitting a loop
This adds a -mpreferred-stack-boundary option patterned after the one supported
by the x86 port. This allows one to reduce the size of stack frames to reduce
memory usage.
This was tested with a rv32i/ilp32 target using -mpreferred-stack-boundary=3
to get 8-byte alignment instead of the default
On Tue, Jan 23, 2018 at 1:07 PM, augustine.sterl...@gmail.com
wrote:
> On Tue, Jan 23, 2018 at 9:55 AM, Max Filippov wrote:
>> libgcc/
>> 2018-01-22 Max Filippov
>>
>> * config/xtensa/ieee754-df.S (__addsf3,
* Claudiu Zissulescu [2017-11-02 13:30:33
+0100]:
> From: claziss
>
> Delegitimize address is used to undo the obfuscating effect of PIC
> addresses, returning the address in a way which is understood by the
> compiler.
>
> gcc/
>
In this testcase, we were deducing a type for g from the function f
which has not yet been deduced itself. Fixed by recognizing the case
of an undeduced initializer.
The change to undeduced_auto_decl is necessary because auto parameters
are a different story; they act more like normal template
On Tue, Jan 23, 2018 at 9:55 AM, Max Filippov wrote:
> libgcc/
> 2018-01-22 Max Filippov
>
> * config/xtensa/ieee754-df.S (__addsf3, __subsf3, __mulsf3)
> (__divsf3): Make NaN return value quiet.
> * config/xtensa/ieee754-sf.S
On Tue, Jan 23, 2018 at 1:06 PM, David Malcolm wrote:
> On Tue, 2018-01-23 at 09:54 -0500, Jason Merrill wrote:
>> On Tue, Jan 23, 2018 at 9:27 AM, David Malcolm
>> wrote:
>> > PR c++/83974 reports an ICE within fold_for_warn when calling
>> >
OK, thanks.
On Tue, Jan 23, 2018 at 2:37 PM, Paolo Carlini wrote:
> Hi,
>
> On 23/01/2018 18:38, Jason Merrill wrote:
>>
>> On 01/22/2018 05:13 PM, Paolo Carlini wrote:
>>>
>>> Hi again,
>>>
>>> On 22/01/2018 22:50, Paolo Carlini wrote:
Ok. The below passes
This is a combinatorial explosion in the combiner for the num_sign_bit_copies
mechanism present on the 7 branch and mainline during the bootstrap on MIPS
n32, i.e. a 64-bit WORD_REGISTER_OPERATIONS SIGN_EXTEND target. While I'm
still exploring various approaches to fixing it on the mainline,
Hi!
DECL_OMP_PRIVATIZED_MEMBER vars are artificial, DECL_IGNORED_P and
useful only during OpenMP lowering, their DECL_VALUE_EXPR isn't really
useful afterwards (just shouldn't be cleared because then we could try
to expand those vars if we don't optimize them as unused).
The sanitizer can add
On Tue, Jan 23, 2018 at 07:30:49PM +, Paul Richard Thomas wrote:
> Janne, Thanks.
>
> Jakub, is this OK with you?
It is indeed quite late for such large ABI changes, some distributions are
about to start using the compiler by now. How much was it tested (on which
targets)? Has the debug
Hi,
On 23/01/2018 18:38, Jason Merrill wrote:
On 01/22/2018 05:13 PM, Paolo Carlini wrote:
Hi again,
On 22/01/2018 22:50, Paolo Carlini wrote:
Ok. The below passes the C++ testsuite and I'm finishing testing it.
Therefore, as you already hinted to, we can now say that what was
*really*
Janne, Thanks.
Jakub, is this OK with you?
Paul
On 23 January 2018 at 19:09, Janne Blomqvist wrote:
> On Mon, Jan 22, 2018 at 9:12 PM, Paul Richard Thomas
> wrote:
>> This patch has been triggered by Thomas's recent message to the
On Tue, Jan 23, 2018 at 6:33 PM, Thomas Koenig wrote:
> Hi Janne,
>
>> When associating a variable of type character, if the length of the
>> target isn't known, set it to zero rather than leaving it unset. This
>> is not a complete fix for making associate of characters
Committed as 'obvious' in revision 256995.
Closing PR.
Paul
2018-23-01 Paul Thomas
PR fortran/83866
* decl.c (gfc_match_derived_decl): If eos not matched, recover
and emit error about garbage after declaration.
2018-23-01 Paul Thomas
Commit as 'obvious' in revision 256994.
I will attend to 6- and 7-branches in a little while.
Paul
2018-23-01 Paul Thomas
PR fortran/83898
* trans-stmt.c (trans_associate_var): Do not set cst_array_ctor
for characters.
2018-23-01 Paul Thomas
On Mon, Jan 22, 2018 at 9:12 PM, Paul Richard Thomas
wrote:
> This patch has been triggered by Thomas's recent message to the list.
> Not only did I start work late relative to stage 3 but debugging took
> somewhat longer than anticipated. Therefore, to get this
Hi Paul,
Could somebody please review the patch?
I'd say the patch is OK for trunk. I have also tested this on
a big-endian system, and things worked.
I will look at the resulting fallout of the GFC_DTYPE_TYPE_SIZE
stuff.
Thanks a lot for your work!
Regards
Thomas
On 01/23/2018 01:27 PM, Mike Stump wrote:
On Jan 23, 2018, at 4:18 AM, Nathan Sidwell wrote:
As discussed (https://gcc.gnu.org/ml/gcc-patches/2018-01/msg01778.html) this
patch deprecates the ARM-era for scope.
The code gives:
if you use %<-fpermissive%> G++ will accept
Hi Jerry,
Do you mean adding something like this:
diff --git a/libgfortran/libgfortran.h b/libgfortran/libgfortran.h
index 4c643b7e17b..c86e0b45e1d 100644
--- a/libgfortran/libgfortran.h
+++ b/libgfortran/libgfortran.h
@@ -600,6 +600,7 @@ typedef struct st_parameter_common
GFC_INTEGER_4
On Jan 23, 2018, at 4:18 AM, Nathan Sidwell wrote:
>
> As discussed (https://gcc.gnu.org/ml/gcc-patches/2018-01/msg01778.html) this
> patch deprecates the ARM-era for scope.
The code gives:
if you use %<-fpermissive%> G++ will accept your code
I think we should no longer
On Jan 23, 2018, at 2:31 AM, Tamar Christina wrote:
>
> This patch makes dg-cmp-results.sh reject the use of log files in the
> comparison.
No please. We like to run that script on log files from time to time for
various reasons. I'd rather fix anything that
On Tue, 2018-01-23 at 09:54 -0500, Jason Merrill wrote:
> On Tue, Jan 23, 2018 at 9:27 AM, David Malcolm
> wrote:
> > PR c++/83974 reports an ICE within fold_for_warn when calling
> > cxx_eval_constant_expression on a CAST_EXPR.
> >
> > This comes from a
libgcc/
2018-01-22 Max Filippov
* config/xtensa/ieee754-df.S (__addsf3, __subsf3, __mulsf3)
(__divsf3): Make NaN return value quiet.
* config/xtensa/ieee754-sf.S (__adddf3, __subdf3, __muldf3)
(__divdf3): Make NaN return value quiet.
---
On 01/22/2018 05:13 PM, Paolo Carlini wrote:
Hi again,
On 22/01/2018 22:50, Paolo Carlini wrote:
Ok. The below passes the C++ testsuite and I'm finishing testing it.
Therefore, as you already hinted to, we can now say that what was
*really* missing from potential_constant_expression_1 was the
Could somebody please review the patch?
Thanks
Paul
On 23 January 2018 at 06:13, Dominique d'Humières
wrote:
> Dear Paul,
>
> The test suite passed without new regression with both -m32 and -m64.
>
> Thanks for the work,
>
> Dominique
--
"If you can't explain
Hi Janne,
When associating a variable of type character, if the length of the
target isn't known, set it to zero rather than leaving it unset. This
is not a complete fix for making associate of characters work
properly, but papers over an ICE in the middle-end. See PR 83344 for
more details.
On Tue, Jan 23, 2018 at 04:46:02PM +0100, Matthias Kretz wrote:
> I just hit a compile error on AVX512 code. The fix is trivial enough that I
> didn't bother writing a PR and just fixed it. Acceptable?
>
> I hope this doesn't require the paperwork, though my employer is willing to
> do
> it
I just hit a compile error on AVX512 code. The fix is trivial enough that I
didn't bother writing a PR and just fixed it. Acceptable?
I hope this doesn't require the paperwork, though my employer is willing to do
it anyway. :-)
Cheers,
Matthias
2018-01-23 Matthias Kretz
Hi,
Here's v2 of the patch to disable register offset addressing mode for
stores of 128-bit values on Falkor because they're very costly.
Differences from the last version:
- Incorporated changes Jim made to his patch earlier that I missed,
i.e. adding an extra tuning parameter called
On 23 January 2018 at 15:57, Kyrill Tkachov
wrote:
> Hi Christophe,
>
> On 23/01/18 13:09, Christophe Lyon wrote:
>>
>> Hi Kyrill,
>>
>> On 22 January 2018 at 11:48, Kyrill Tkachov
>> wrote:
>>>
>>> Hi all,
>>>
>>> This test has
Hi Christophe,
On 23/01/18 13:09, Christophe Lyon wrote:
Hi Kyrill,
On 22 January 2018 at 11:48, Kyrill Tkachov
wrote:
Hi all,
This test has needlessly restrictive requirements. It tries to force a
soft-float target and tries to run.
This makes it unsupportable
On Tue, Jan 23, 2018 at 9:27 AM, David Malcolm wrote:
> PR c++/83974 reports an ICE within fold_for_warn when calling
> cxx_eval_constant_expression on a CAST_EXPR.
>
> This comes from a pointer-to-member-function. The result of
> build_ptrmemfunc (within
Hi all,
This patch fixes the testsuite failures gcc.target/aarch64/subs_compare_1.c and
subs_compare_2.c
The tests check that we combine a sequence like:
sub w2, w0, w1
cmp w0, w1
into
subsw2, w0, w1
This is done by a couple of peepholes in aarch64.md.
On Tue, 23 Jan 2018, Chung-Ju Wu wrote:
> In previous patch we add new options for NDS32 port:
> https://gcc.gnu.org/ml/gcc-patches/2018-01/msg01397.html
>
> So I think I need to update GCC 8 release notes as well.
Yes, please. :-)
> +New command-line options -mext-perf -mext-perf2
PR c++/83974 reports an ICE within fold_for_warn when calling
cxx_eval_constant_expression on a CAST_EXPR.
This comes from a pointer-to-member-function. The result of
build_ptrmemfunc (within cp_convert_to_pointer for a null ptr) is
a CONSTRUCTOR containing, amongst other things a CAST_EXPR of a
Hi!
These 2 PRs were fixed with r256964 and stay fixed even with r256965.
So I've just committed the testcase to the trunk and am going to close them
as fixed.
2018-01-23 Jakub Jelinek
PR c++/82882
PR c++/83978
* g++.dg/cpp0x/pr82882.C: New test.
I added an assert when recently fixing baselink substitution, but the
assert is incorrect as this testcase shows.
Fixing thusly.
nathan
--
Nathan Sidwell
2018-01-23 Nathan Sidwell
PR c++/83988
* pt.c (tsubst_baselink): Remove optype assert.
* ptree.c (cxx_print_xnode):
> Hi.
>
> This removes false positive warning when having trailing array at the end of
> a struct.
>
> Patch can bootstrap on ppc64le-redhat-linux and survives regression tests.
>
> Ready to be installed?
> Martin
>
> gcc/lto/ChangeLog:
>
> 2018-01-23 Martin Liska
>
>
> On 01/23/2018 11:02 AM, Jan Hubicka wrote:
> >> On 01/19/2018 12:57 PM, Martin Liška wrote:
> >>> Yes, there's a huge difference in between CPU 2006 and 2017. Former has
> >>> 63% w/ dominant edges,
> >>> and later one only 11%. It's caused by these 2 benchmarks with a high
> >>> coverage:
>
On 01/22/2018 06:06 PM, Jakub Jelinek wrote:
Hi!
I've recently added the complete_type call in case the structured binding
is a reference to a type that needs to be instantiated.
This testcase shows a problem where it can't be instantiated and we ICE
because we don't expect an incomplete type.
As part of the change to larger character lengths, the string copy
algorithm was temporarily pessimized to get around some spurious
-Wstringop-overflow warnings. Having tried a number of variations of
this algorithm I have managed to get it down to one spurious warning,
only with -O1
This fixes a GRAPHITE code-generation issue by eliding ISL AST plus
for large power-of-two values that don't affect the result.
I intentionally didn't extend this to other values with the same
property as I'd like to see testcases.
Bootstrap and regtest running on x86_64-unknown-linux-gnu.
Hi,
On 01/23/2018 07:40 AM, Kyrill Tkachov wrote:
Hi Luis,
On 22/01/18 13:46, Luis Machado wrote:
The following patch adds an option to control software prefetching of
memory
references with non-constant/unknown strides.
Currently we prefetch these references if the pass thinks there is
Hi Kyrill,
On 01/23/2018 07:32 AM, Kyrill Tkachov wrote:
Hi Luis,
On 22/01/18 13:46, Luis Machado wrote:
This patch adds a new option to control the minimum stride, for a memory
reference, after which the loop prefetch pass may issue software prefetch
hints for. There are two motivations:
*
On Tue, Jan 23, 2018 at 1:54 PM, Uros Bizjak wrote:
> On Tue, Jan 23, 2018 at 11:38 AM, Uros Bizjak wrote:
>> On Tue, Jan 23, 2018 at 8:32 AM, Uros Bizjak wrote:
>>> On Tue, Jan 23, 2018 at 5:49 AM, Ian Lance Taylor
Hi Kyrill,
On 22 January 2018 at 11:48, Kyrill Tkachov
wrote:
> Hi all,
>
> This test has needlessly restrictive requirements. It tries to force a
> soft-float target and tries to run.
> This makes it unsupportable for any non-soft-float variant.
> In fact, the test
On Tue, Jan 23, 2018 at 11:38 AM, Uros Bizjak wrote:
> On Tue, Jan 23, 2018 at 8:32 AM, Uros Bizjak wrote:
>> On Tue, Jan 23, 2018 at 5:49 AM, Ian Lance Taylor wrote:
>>> On Sun, Jan 21, 2018 at 3:13 PM, Uros Bizjak
Hi.
This removes false positive warning when having trailing array at the end of a
struct.
Patch can bootstrap on ppc64le-redhat-linux and survives regression tests.
Ready to be installed?
Martin
gcc/lto/ChangeLog:
2018-01-23 Martin Liska
PR lto/81440
*
On 01/23/2018 10:43 AM, Jan Hubicka wrote:
>> Hi.
>>
>> I'm aware in which development stage we are. However the patch is small and
>> makes
>> dump files readable. Hope such patch can be accepted even now?
>>
>> Patch can bootstrap on ppc64le-redhat-linux and survives regression tests.
>>
>>
As discussed (https://gcc.gnu.org/ml/gcc-patches/2018-01/msg01778.html)
this patch deprecates the ARM-era for scope.
a) in c++98 mode with -fpermissive, there's now a deprecation note when
we fix up something like
for (int i = ...) {}
... i ... // out of scope use of i
b)
On 01/23/2018 11:02 AM, Jan Hubicka wrote:
>> On 01/19/2018 12:57 PM, Martin Liška wrote:
>>> Yes, there's a huge difference in between CPU 2006 and 2017. Former has 63%
>>> w/ dominant edges,
>>> and later one only 11%. It's caused by these 2 benchmarks with a high
>>> coverage:
>>>
>>
>> Hi.
The 01/23/2018 11:30, Bernhard Reutner-Fischer wrote:
> On 23 January 2018 11:31:27 CET, Tamar Christina
> wrote:
> >Hi All,
> >
> >This patch makes dg-cmp-results.sh reject the use of log files in the
> >comparison.
> >Often when given a log file dg-cmp-results will
The failures in this PR were from forcing { dg-do run } even when
vect.exp chooses options that are incompatible with the runtime.
The default vect.exp behaviour is to execute when possible, so there's
no need for a dg-do at all.
The patch removes other unconditional { dg-do run }s too. Many of
On 23 January 2018 11:31:27 CET, Tamar Christina
wrote:
>Hi All,
>
>This patch makes dg-cmp-results.sh reject the use of log files in the
>comparison.
>Often when given a log file dg-cmp-results will give incomplete/wrong
>output and
>using log instead of sum is one
r255913 changed some constant_boolean_node calls to boolean_true_node
and boolean_false_node, which meant that the returned tree didn't
always have the right type.
Tested on aarch64-linux-gnu. Probably bordering on obvious, but just
in case: OK to install?
Richard
2018-01-23 Richard
On Sat, Jan 20, 2018 at 4:10 PM, Bin.Cheng wrote:
> On Fri, Jan 19, 2018 at 5:42 PM, Bin Cheng wrote:
>> Hi,
>> This patch is supposed to fix regression caused by loop distribution when
>> ftree-parallelize-loops. The reason is distributed memset call
Richard Biener writes:
> On Mon, Jan 22, 2018 at 6:55 PM, Richard Sandiford
> wrote:
>> vect_float is true for arm*-*-* targets, but the support is only
>> available when -funsafe-math-optimizations is on. This caused
>> failures in two
On Tue, Jan 23, 2018 at 8:32 AM, Uros Bizjak wrote:
> On Tue, Jan 23, 2018 at 5:49 AM, Ian Lance Taylor wrote:
>> On Sun, Jan 21, 2018 at 3:13 PM, Uros Bizjak wrote:
>>>
>>> The default "go build" compile options over-optimize the auxiliary
Hi All,
This patch makes dg-cmp-results.sh reject the use of log files in the
comparison.
Often when given a log file dg-cmp-results will give incomplete/wrong output and
using log instead of sum is one autocomplete tab away.
Instead have the tool guard against such mistakes.
Ok for trunk?
On Mon, Jan 22, 2018 at 9:05 PM, David Malcolm wrote:
> On Fri, 2018-01-19 at 09:45 +0100, Richard Biener wrote:
>> On Fri, Jan 19, 2018 at 12:36 AM, David Malcolm
>> wrote:
>> > PR tree-optimization/83510 reports that r255649 (for
>> > PR
On Mon, Jan 22, 2018 at 8:51 PM, David Malcolm wrote:
> "make selftest-valgrind" shows a few leaks in sbitmap.c's selftests;
> this patch fixes them.
>
> Successfully bootstrapped on x86_64-pc-linux-gnu.
> OK for trunk?
Ok.
> gcc/ChangeLog:
> * sbitmap.c
On Mon, Jan 22, 2018 at 8:52 PM, David Malcolm wrote:
> This patch adds a few extra assertions to selftest::test_location_wrappers.
>
> Successfully bootstrapped on x86_64-pc-linux-gnu.
> OK for trunk?
Ok.
> gcc/ChangeLog:
> * tree.c
On Mon, Jan 22, 2018 at 6:55 PM, Richard Sandiford
wrote:
> vect_float is true for arm*-*-* targets, but the support is only
> available when -funsafe-math-optimizations is on. This caused
> failures in two tests that disable fast-math.
>
> The easiest fix seemed to
On Mon, Jan 22, 2018 at 6:53 PM, Richard Sandiford
wrote:
> In this PR we recognised a PLUS_EXPR as a fold-left reduction,
> then applied pattern matching to convert it to a WIDEN_SUM_EXPR.
> We need to keep the original code in this case since we implement
> the
On Mon, 22 Jan 2018, Jeff Law wrote:
> On 01/22/2018 01:57 AM, Jakub Jelinek wrote:
> > On Mon, Jan 22, 2018 at 09:37:05AM +0100, Richard Biener wrote:
> >>
> >> Currently gcc_release builds GCC (for generating in-tree generated
> >> files) serially - that's prohibitly expensive and takges ages
Hi,
this is second part of fix for wrong partitining decisions. It makes
probabily_never_executed to ignore scaled profiles because they may be
low for invalid reasons. This does not seem to affect size of cold section
significantly (as it is predocminantly occupied by blocks with 0 count that
Hi,
this is first patch of two which aim to fix function partitioning issues where
we put basic blocks to cold sections while we should not.
This patch fixes issues where we think we have precise profile information but
we don't because we scaled it by some of named probability values. It is
> On 01/19/2018 12:57 PM, Martin Liška wrote:
> > Yes, there's a huge difference in between CPU 2006 and 2017. Former has 63%
> > w/ dominant edges,
> > and later one only 11%. It's caused by these 2 benchmarks with a high
> > coverage:
> >
>
> Hi.
>
> I'm sending details about the 2 edges
.. testing completed OK.
Just in case it wasn't obvious in my previous messages, when we are
calling check_for_uninitialized_const_var from
potential_constant_expression_1 we can't use CP_TYPE_CONST_P as gate for
emitting diagnostic - as just discovered - neither we can use
> Hi.
>
> I'm aware in which development stage we are. However the patch is small and
> makes
> dump files readable. Hope such patch can be accepted even now?
>
> Patch can bootstrap on ppc64le-redhat-linux and survives regression tests.
>
> Martin
>
> gcc/ChangeLog:
>
> 2018-01-22 Martin
Hi Luis,
On 22/01/18 13:46, Luis Machado wrote:
The following patch adds an option to control software prefetching of memory
references with non-constant/unknown strides.
Currently we prefetch these references if the pass thinks there is benefit to
doing so. But, since this is all based on
Hi Luis,
On 22/01/18 13:46, Luis Machado wrote:
This patch adds a new option to control the minimum stride, for a memory
reference, after which the loop prefetch pass may issue software prefetch
hints for. There are two motivations:
* Make the pass less aggressive, only issuing prefetch hints
When associating a variable of type character, if the length of the
target isn't known, set it to zero rather than leaving it unset. This
is not a complete fix for making associate of characters work
properly, but papers over an ICE in the middle-end. See PR 83344 for
more details.
Regtested on
Hi.
I'm aware in which development stage we are. However the patch is small and
makes
dump files readable. Hope such patch can be accepted even now?
Patch can bootstrap on ppc64le-redhat-linux and survives regression tests.
Martin
gcc/ChangeLog:
2018-01-22 Martin Liska
Hi, Gerald,
In previous patch we add new options for NDS32 port:
https://gcc.gnu.org/ml/gcc-patches/2018-01/msg01397.html
So I think I need to update GCC 8 release notes as well.
Index: htdocs/gcc-8/changes.html
===
RCS file:
83 matches
Mail list logo