On January 4, 2019 11:49:45 PM GMT+01:00, Jakub Jelinek
wrote:
>Hi!
>
>The following patch adds an optimization, if we know certain SSA_NAME
>has two possible values and have GIMPLE_COND EQ_EXPR/NE_EXPR of that
>SSA_NAME with one of the two values and PHI which has two adjacent
>constants, we can
On January 3, 2019 11:36:00 PM GMT+01:00, Jakub Jelinek
wrote:
>Hi!
>
>The following testcase FAILs on 8.x branch, went latent later on on the
>trunk and I suppose Alex' i386.c ix86_const_not_ok_for_debug_p change
>would
>have prevented it too.
>
>The problem is in what the comment in const_ok_fo
On January 3, 2019 11:55:16 PM GMT+01:00, Jakub Jelinek
wrote:
>Hi!
>
>The following testcases ICE on x86_64 and other targets.
>
>The problem is that a function with nested functions returning VLA
>structures is inlined (or versioned) so that the variable sizes become
>constant, but we still hav
Hi Thomas,
this is the current version of the oacc-* parts of the Async Re-work patch.
I have reverted away from the earlier mentioned attempt of using lockless
techniques to manage the asyncqueues; it is really hard to do in a 100% correct
manner, unless we only use something like simple lists t
Hi Martin,
> On 4 Jan 2019, at 22:30, Mike Stump wrote:
>
> On Jan 4, 2019, at 2:03 PM, Martin Sebor wrote:
>>
>> The improved handling of attribute positional arguments added
>> in r266195 introduced a regression on Darwin where attribute
>> format with the CFString archetype accepts CFString
On 22-12-18 03:13, Tom de Vries wrote:
> If you have a test-case where this is indeed failing without the
> proposed hook implementation, then please try to remove the hardcoding
> of vector_length > 32 from the test-source and instead set it using
> -fopenacc-dim. AFAIU, the proposed hook does not
On Sat, Jan 05, 2019 at 09:51:31AM +0100, Richard Biener wrote:
> >+ if (TYPE_PRECISION (TREE_TYPE (lhs)) == TYPE_PRECISION (TREE_TYPE
> >(arg0)))
> >+{
> >+ if (TREE_CODE (TREE_TYPE (arg0)) == BOOLEAN_TYPE
> >+ || !TYPE_OVERFLOW_WRAPS (TREE_TYPE (arg0)))
> >+type = TREE_TYPE (lh
Hi,
co-incidentally my "Make strlen range computations more conservative" patch
contained a fix on the same spot, I just did not have a test case for it:
@@ -3184,7 +3146,10 @@ get_min_string_length (tree rhs, bool *f
&& TREE_READONLY (rhs))
rhs = DECL_INITIAL (rhs);
- if (rhs && TR
Hi Steve,
thanks for taking this on!
Maybe it would be better to put the checking and argument reordering
into its own routine (something like gfc_check_minloc_maxloc in
check.c) so all three arguments would be present, with an
expression possibly containing NULL, for the later routines.
This co
On Sat, Jan 05, 2019 at 12:20:24PM +, Bernd Edlinger wrote:
> co-incidentally my "Make strlen range computations more conservative" patch
> contained a fix on the same spot, I just did not have a test case for it:
>
> @@ -3184,7 +3146,10 @@ get_min_string_length (tree rhs, bool *f
>&&
On 1/5/19 1:31 PM, Jakub Jelinek wrote:
> On Sat, Jan 05, 2019 at 12:20:24PM +, Bernd Edlinger wrote:
>> co-incidentally my "Make strlen range computations more conservative" patch
>> contained a fix on the same spot, I just did not have a test case for it:
>>
>> @@ -3184,7 +3146,10 @@ get_min_
On Sat, Jan 05, 2019 at 01:17:21PM +, Bernd Edlinger wrote:
> Well, yes it works, but this can only optimize invalid code like
> strlen((char*)L"wide").
There is nothing invalid on it. Furthermore, the strlen pass doesn't
optimize just strlen, it does many optimizations that rely on knowing
Hi all,
the attached patch fixes PR 88009, an ICE-on-invalid regression caused
by one of my earlier commits. Apart from adding some extra checks to
avoid ICEs, it also uses the 'artificial' attribute to suppress bogus
errors (see comment #3) and does some minor cleanup in
resolve_fl_variable.
Reg
Hi Janus,
the attached patch fixes PR 88009, an ICE-on-invalid regression caused
by one of my earlier commits. Apart from adding some extra checks to
avoid ICEs, it also uses the 'artificial' attribute to suppress bogus
errors (see comment #3) and does some minor cleanup in
resolve_fl_variable.
Am Sa., 5. Jan. 2019 um 14:37 Uhr schrieb Thomas Koenig :
>
> Hi Janus,
>
> > the attached patch fixes PR 88009, an ICE-on-invalid regression caused
> > by one of my earlier commits. Apart from adding some extra checks to
> > avoid ICEs, it also uses the 'artificial' attribute to suppress bogus
> >
From: Paul Hua
In some older Loongson3 processors there is a LL/SC errata that can
cause the CPU to deadlock occasionally. The details are very
complicated. We find a way to work around this errata by
a) adding a sync before ll/lld instruction,
b) adding a sync before branch target that betwee
On 1/4/19 10:22 PM, Jason Merrill wrote:
> Hmm, I'm uncomfortable with starting to pass in the decl just for the sake of
> deciding whether this diagnostic should be a pedwarn or error. In general,
> because of copy elision, we can't know at this point what we're initializing,
> so I'd rather no
On Sat, Jan 05, 2019 at 01:23:14PM +0100, Thomas Koenig wrote:
>
> Maybe it would be better to put the checking and argument reordering
> into its own routine (something like gfc_check_minloc_maxloc in
> check.c) so all three arguments would be present, with an
> expression possibly containing NUL
Committed on trunk as obvious
* gcc.dg/plugin/plugindir1.c: Adjust dg-prune-output for Darwin.
* gcc.dg/plugin/plugindir2.c: Likewise.
* gcc.dg/plugin/plugindir3.c: Likewise.
* gcc.dg/plugin/plugindir4.c: Likewise.
Dominique
On 1/5/19 3:31 AM, Iain Sandoe wrote:
Hi Martin,
On 4 Jan 2019, at 22:30, Mike Stump wrote:
On Jan 4, 2019, at 2:03 PM, Martin Sebor wrote:
The improved handling of attribute positional arguments added
in r266195 introduced a regression on Darwin where attribute
format with the CFString ar
On Sat, Jan 05, 2019 at 09:22:34AM -0800, Steve Kargl wrote:
> On Sat, Jan 05, 2019 at 01:23:14PM +0100, Thomas Koenig wrote:
> >
> > Maybe it would be better to put the checking and argument reordering
> > into its own routine (something like gfc_check_minloc_maxloc in
> > check.c) so all three a
Hi,
this patch fixes the way we account time. analyze_function_body
uses frequency of BB execution and expected time to execute the stmt
to compute overall time it takes to invoke function, but it manages to
account time which is not scaled by frequency, yet.
Bootstrapped/regtested x86_64-linux,
> On 5 Jan 2019, at 17:39, Martin Sebor wrote:
>
> On 1/5/19 3:31 AM, Iain Sandoe wrote:
>> Hi Martin,
>>> On 4 Jan 2019, at 22:30, Mike Stump wrote:
>>>
>>> On Jan 4, 2019, at 2:03 PM, Martin Sebor wrote:
The improved handling of attribute positional arguments added
in r2661
Hi,
this patch adds new parameters to ipa-inline. max-inline-insns-size is
useful to increase inlining limits for programs with large abstraction
penalty.
uinlined-* should be useful for architecutures with greater function
call overhead than modern x86 chips (which is good portion of them,
especi
> Committed on trunk as obvious
Committed as revision r267599. Patch attached.
Dominique
patch-plug
Description: Binary data
On Sat, Jan 05, 2019 at 03:35:21PM +0100, Janus Weil wrote:
> Am Sa., 5. Jan. 2019 um 14:37 Uhr schrieb Thomas Koenig
> :
> >
> > Hi Janus,
> >
> > > the attached patch fixes PR 88009, an ICE-on-invalid regression caused
> > > by one of my earlier commits. Apart from adding some extra checks to
>
On Sat, Jan 05, 2019 at 11:39:08AM -0800, Steve Kargl wrote:
> On Sat, Jan 05, 2019 at 03:35:21PM +0100, Janus Weil wrote:
> > Am Sa., 5. Jan. 2019 um 14:37 Uhr schrieb Thomas Koenig
> > :
> > >
> > > Hi Janus,
> > >
> > > > the attached patch fixes PR 88009, an ICE-on-invalid regression caused
>
On Sat, Jan 05, 2019 at 11:45:42AM -0800, Steve Kargl wrote:
> Execution timeout is: 300
> spawn [open ...]
> STOP 2
> FAIL: gfortran.dg/class_alias.f90 -Os execution test
>
> STOP 2 occurs at the end of testcase, so I suspect finalization is messed up.
This is bizarre. Changing the code wher
In C++17 the clock used for filesystem::file_time_type is unspecified,
allowing it to be chrono::system_clock. The C++2a draft requires it to
be a distinct type, with additional member functions to convert to/from
other clocks (either the system clock or UTC). In order to avoid an ABI
change later
Hi Martin,
The patch on top of r267591 fixes pr88638 without regression.
Note
FAIL: c-c++-common/attributes-4.c -std=gnu++14 (test for excess errors)
FAIL: c-c++-common/attributes-4.c -std=gnu++17 (test for excess errors)
FAIL: c-c++-common/attributes-4.c -std=gnu++98 (test for excess erro
This patch adds thread-local globals to the libgccjit frontend. The
library user can mark a global as being thread-local by calling
`gcc_jit_lvalue_set_bool_thread_local'. It is implemented by calling
`set_decl_tls_model (inner, decl_default_tls_model (inner))', where
`inner' is the GENERIC tree co
> > FAIL: gfortran.dg/class_alias.f90 -Os execution test
>
> I see this failure on trunk as well, but I don't think it was
> introduced by my commit. Instead, I see I it starting with r267601:
>
> https://gcc.gnu.org/viewcvs/gcc?view=revision&revision=267601
>
> Jan, could you please have a l
On Sat, Jan 05, 2019 at 11:48:15PM +0100, Jan Hubicka wrote:
> > > FAIL: gfortran.dg/class_alias.f90 -Os execution test
> >
> > I see this failure on trunk as well, but I don't think it was
> > introduced by my commit. Instead, I see I it starting with r267601:
> >
> > https://gcc.gnu.org/view
On Sat, Jan 05, 2019 at 03:50:06PM -0800, Steve Kargl wrote:
>
> /gfortran.dg/pr79966.f90 -fno-diagnostics-show-caret
> -fno-diagnostics-show-line-numbers -fdiagnostics-color=never -O -O2
> -fpeel-loops -finline-functions -fipa-cp-clone -fdump-ipa-inline-details -S
> -o pr79966.s
> PASS: gfortr
> On Sat, Jan 05, 2019 at 03:50:06PM -0800, Steve Kargl wrote:
> >
> > /gfortran.dg/pr79966.f90 -fno-diagnostics-show-caret
> > -fno-diagnostics-show-line-numbers -fdiagnostics-color=never -O -O2
> > -fpeel-loops -finline-functions -fipa-cp-clone -fdump-ipa-inline-details -S
> > -o pr79966.s
>
On 03/01/19 22:44 +, Jonathan Wakely wrote:
On 03/01/19 22:07 +, Jonathan Wakely wrote:
On 03/01/19 20:38 +, Jonathan Wakely wrote:
There are a number of UTF-8 characters in comments which add no value
and can be replaced with ASCII equivalents, or removed entirely for the
section s
36 matches
Mail list logo
