On 05/11/2017 12:46 PM, Uros Bizjak wrote:
On Thu, May 11, 2017 at 8:28 PM, Jeff Law wrote:
Attached patch changes compare-elim.c order to what i386.md expects.
Thoughts?
Haven't looked at the patch itself. But I do have the necessary bits to
convert the mn103 port. It was slightly more t
On 05/11/2017 03:29 PM, Hans-Peter Nilsson wrote:
On Wed, 10 May 2017, Jakub Jelinek wrote:
On Wed, May 10, 2017 at 09:57:56PM +0200, Uros Bizjak wrote:
BTW: This patch now catches 417 cases (instead of 200+) in linux
build, including e.g.:
(parallel [
(set (reg:CCZ 17 flags)
On 12 May 2017 02:29:58 CEST, Nathan Sidwell wrote:
>On 05/11/2017 06:52 PM, Bernhard Reutner-Fischer wrote:
>> On 11 May 2017 15:10:36 CEST, Nathan Sidwell wrote:
>
>> Anyway, please also adjust the dump cleanup to handle class or I can
>take care of it sometimes later.
>
>Applied this, which se
I introduced a subtle bug in my fix for pr70828. That PR involved
implicit data mappings of subarrays which the user had explicitly mapped
inside an enclosing acc data region. E.g.
#pragma acc data copy(a[5:10])
{
#pragma acc parallel loop
for (...)
a[i] = ...
My fix for the PR
Early steppings had aarch32 support, current steppings don't, so the
aarch32 support for falkor/qdf24xx needs to be dropped. This mostly
involves removing falkor/qdf24xx references from the arm port. The
qdf24xx_extra_costs structure moves from the arm port to the aarch64
port.
This was tested w
I found a few more instances of the warning and a couple of bugs
in the fixes for it. The attached update corrects those as well.
On 05/11/2017 02:01 PM, Martin Sebor wrote:
Attached is revision 2 of the patch that addresses of the feedback
I got from Pedro (thanks again). As I explained in my
This patch by James Cowgill fixes signal aliases when libgo works out
the signal table. On MIPS, SIGABRT is defined like this:
#define SIGIOT 6
#define SIGABRT SIGIOT
This confused the mksigtab.sh script in libgo. This patch fixes the
problem. Bootstrapped and ran Go testsuite on x86_64-pc-
Since on AIX we use readdir64, we must also use opendir64 and
closedir64. This patch by Matthieu Sarter fixes the problem.
Bootstrapped and ran Go testsuite on x86_64-pc-linux-gnu. Committed
to mainline.
Ian
Index: gcc/go/gofrontend/MERGE
=
On 05/11/2017 06:52 PM, Bernhard Reutner-Fischer wrote:
On 11 May 2017 15:10:36 CEST, Nathan Sidwell wrote:
Anyway, please also adjust the dump cleanup to handle class or I can take care
of it sometimes later.
Applied this, which seems to do the trick.
nathan
--
Nathan Sidwell
2017-05-1
On 05/11/2017 06:57 PM, Rainer Orth wrote:
er, the introduction of scanlang.exp broke libatomic, libgomp, and
libitm testing. Their logfiles show
trunk/12-gcc/build/sparc-sun-solaris2.12/libatomic/testsuite/libatomic.log:ERROR:
Couldn't find library file scanlang.exp.
trunk/12-gcc/build/sparc-
The method implicit_variable_reference in the interface used between
the Go frontend and GCC proper creates a reference to a variable in a
different object file. Unfortunately, ever since it was introduced,
it has incorrectly cleared DECL_EXTERNAL and set TREE_STATIC. I'm not
entirely sure why th
On 5/11/17 3:49 PM, H.J. Lu wrote:
> This caused:
>
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=80714
This is already being tracked in https://gcc.gnu.org/PR80707
Peter
On 12 May 2017 00:57:58 CEST, Rainer Orth wrote:
>Bernhard Reutner-Fischer writes:
>
>> On 11 May 2017 15:10:36 CEST, Nathan Sidwell wrote:
>>>On 05/11/2017 08:28 AM, Rainer Orth wrote:
>>>
On the gcc-7 branch, the dump file (from -fdump-class-hierarchy)
>was
called covariant7.C.002t.c
Bernhard Reutner-Fischer writes:
> On 11 May 2017 15:10:36 CEST, Nathan Sidwell wrote:
>>On 05/11/2017 08:28 AM, Rainer Orth wrote:
>>
>>> On the gcc-7 branch, the dump file (from -fdump-class-hierarchy) was
>>> called covariant7.C.002t.class, now it's covariant7.C.002l.class.
>>But
>>> gcc/tes
On 11 May 2017 15:10:36 CEST, Nathan Sidwell wrote:
>On 05/11/2017 08:28 AM, Rainer Orth wrote:
>
>> On the gcc-7 branch, the dump file (from -fdump-class-hierarchy) was
>> called covariant7.C.002t.class, now it's covariant7.C.002l.class.
>But
>> gcc/testsuite/lib/scantree.exp (scan-tree-dump) ex
Hello world,
the memory allocation for the buffer in the library matmul
routines still has one problem: The value of 0xdeadbeef meant
as poison could end up in the calculation of the size of the
buffer for the blocked matmul.
The attached patch fixes that. Verified with regression-test,
also by
On Thu, May 11, 2017 at 02:36:26PM -0500, Will Schmidt wrote:
> On Thu, 2017-05-11 at 14:15 -0500, Segher Boessenkool wrote:
> > Hi!
> >
> > On Thu, May 11, 2017 at 10:53:33AM -0500, Will Schmidt wrote:
> > > Add handling for early expansion of vector locical operations in gimple.
> > > Specifical
On Wed, 10 May 2017, Jakub Jelinek wrote:
> On Wed, May 10, 2017 at 09:57:56PM +0200, Uros Bizjak wrote:
> > BTW: This patch now catches 417 cases (instead of 200+) in linux
> > build, including e.g.:
> >
> > (parallel [
> > (set (reg:CCZ 17 flags)
> > (compare:CCZ (lshiftrt:SI
Uros Bizjak writes:
>> I have committed a large patch to update the Go frontend and libgo to
>> the recent changes in the gofrontend repository. I had postponed
>> merging changes during the GCC 7 release process. I am now merging
>> all the changes that were pending during that period. Althou
The gc toolchain decides whether a goroutine is a system goroutine by
comparing startpc to a list of saved special PCs. In gccgo that
approach does not work as startpc is often a thunk that invokes the
real function with arguments, so the thunk address never matches the
saved special PCs.
This pa
On Tue, May 9, 2017 at 7:54 PM, Peter Bergner wrote:
> Here is the updated patch to use gimple_seq_unreachable_p() which scans the
> sequence backwards so it bails out earlier in the common case (ie, no call
> to __builtin_unreachable). As discussed in the previous thread, we remove
> case statem
Hi!
On Thu, May 11, 2017 at 10:56:48AM -0700, Carl E. Love wrote:
> * config/rs6000/rs6000-c: Add support for built-in functions
> vector unsigned char vec_popcnt (vector signed char)
> vector unsigned char vec_popcnt (vector unsigned char)
> vector unsigned short vec_popcn
On 10/05/2017 22:44, François Dumont wrote:
On 10/05/2017 11:28, Jonathan Wakely wrote:
On 10/05/17 11:15 +0200, Paolo Carlini wrote:
Hi,
On 10/05/2017 11:12, Jonathan Wakely wrote:
Looks good to me. Paolo, what do you think about bumping the versioned
namespace and SONAME to 8?
Sure, makes
Hi
When versioned namespace is active we can avoid export of
_Rb_tree_rotate_[left,right] symbols. I also took the opportunity to put
static functions in the anonymous namespace rather than using static. Is
this usage of static still planned to be deprecated ?
* src/c++98/tree.cc [_G
Hi
_Safe_container _IsCxx11AllocatorAware template allocator is only
used if C++11 Abi is not used so I simplified it.
* include/debug/safe_container.h [_GLIBCXX_USE_CXX11_ABI]
(_Safe_container<>): Remove _IsCxx11AllocatorAware template parameter.
* include/debug/string: Adapt.
Attached is revision 2 of the patch that addresses of the feedback
I got from Pedro (thanks again). As I explained in my response
(https://gcc.gnu.org/ml/gcc-patches/2017-05/msg00925.html) it's not
entirely practical to rely on the strict interpretation of the C++
definition of type triviality or
On Thu, May 11, 2017 at 10:59:41AM -0500, Bill Schmidt wrote:
> PR80695 identifies a case (similar to several others we've seen) where SLP
> vectorization is too aggressive about vectorizing stores. The problem is
> that we undervalue the cost of a vec_construct operation. vec_construct
> is the
Hi
Now that soname has been incremented I can remove several useless
debug symbols when versioned namespace is active.
* include/debug/formatter.h [_GLIBCXX_INLINE_VERSION]
(__gnu_debug::_Error_formatter::_Parameter::_M_print_field): Remove.
(__gnu_debug::_Error_formatter::_Parame
On Thu, 2017-05-11 at 14:15 -0500, Segher Boessenkool wrote:
> Hi!
>
> On Thu, May 11, 2017 at 10:53:33AM -0500, Will Schmidt wrote:
> > Add handling for early expansion of vector locical operations in gimple.
> > Specifically: vec_and, vec_andc, vec_or, vec_xor, vec_orc, vec_nand.
>
> You also h
On Thu, May 11, 2017 at 10:54:02AM -0500, Will Schmidt wrote:
> Add handling for early expansion of vector divide built-ins
>
> Bootstrapped and tested on powerpc64le-unknown-linux-gnu, and
> powerpc-unknown-linux, (p7,p8le,p8be) with no regressions.
>
> OK for trunk?
Looks good, please appl
On Thu, 2017-05-11 at 09:39 -0500, Segher Boessenkool wrote:
> On Wed, May 10, 2017 at 12:59:28PM -0500, Steven Munroe wrote:
> > > That is just for the testsuite; I meant what happens if a user tries
> > > to use it with an older target (or BE, or 32-bit)? Is there a useful,
> > > obvious error m
On 05/11/2017 02:54 PM, Christophe Lyon wrote:
On 11 May 2017 at 20:49, Nathan Sidwell wrote:
oh gee. I committed the ChangeLog but not the changed file. just done.
I have now ...
nathan
--
Nathan Sidwell
On 05/11/2017 01:50 PM, Rainer Orth wrote:
Hi Nathan,
On 05/08/2017 05:34 PM, Joseph Myers wrote:
^~
make: *** [Makefile:1102: objcp/objcp-lang.o] Error 1
Having learnt how to spell --enable-languages properly today, here's the
fix, whi
Hi!
On Thu, May 11, 2017 at 10:53:33AM -0500, Will Schmidt wrote:
> Add handling for early expansion of vector locical operations in gimple.
> Specifically: vec_and, vec_andc, vec_or, vec_xor, vec_orc, vec_nand.
You also handle nor (except in the changelog). But what about eqv?
Segher
Hi!
On Thu, May 11, 2017 at 10:51:54AM -0500, Will Schmidt wrote:
> Add support for some of the vector multiply (sp,dp) operations in
> the gimple folding. I missed these during an earlier pass that
> added the other (mule, mulo) multiply operations. This change is
> covered by the existing test
On 11 May 2017 at 20:49, Nathan Sidwell wrote:
> On 05/11/2017 02:45 PM, Christophe Lyon wrote:
>
>>
>> `/tmp/3713244_6.tmpdir/aci-gcc-fsf/builds/gcc-fsf-gccsrc/obj-aarch64-none-linux-gnu/gcc1/libcc1'
>>
>> I suspect it will be an obvious fix?
>
>
> oh gee. I committed the ChangeLog but not the c
On 05/11/2017 02:45 PM, Christophe Lyon wrote:
`/tmp/3713244_6.tmpdir/aci-gcc-fsf/builds/gcc-fsf-gccsrc/obj-aarch64-none-linux-gnu/gcc1/libcc1'
I suspect it will be an obvious fix?
oh gee. I committed the ChangeLog but not the changed file. just
done. Hopefully soon the churn will be res
On Thu, May 11, 2017 at 8:28 PM, Jeff Law wrote:
>> Attached patch changes compare-elim.c order to what i386.md expects.
>>
>> Thoughts?
>
> Haven't looked at the patch itself. But I do have the necessary bits to
> convert the mn103 port. It was slightly more than just fixing the md file,
> but
Hi Nathan
On 11 May 2017 at 19:40, Nathan Sidwell wrote:
> This patch reduces the variants of pushdecl by use of default arguments.
> gone are the _maybe_friend variants.
>
Since this commit (r247918), GCC builds fail at least for aarch64/arm.
I've noticed this error message:
libtool: compile:
On Thu, May 11, 2017 at 12:28:56PM -0600, Jeff Law wrote:
> On 05/10/2017 01:05 PM, Uros Bizjak wrote:
> > On Wed, May 10, 2017 at 5:18 PM, Uros Bizjak wrote:
> > > On Wed, May 10, 2017 at 4:27 PM, Jakub Jelinek wrote:
> > > > On Tue, May 09, 2017 at 06:06:47PM +0200, Uros Bizjak wrote:
> > > > >
2017-05-11 0:05 GMT+03:00 Alexander Ivchenko :
> Didn't quite get your point. I though that the idea is to hit this code:
>
> bounds = chkp_generate_extern_var_bounds (decl);
> ... to get the builtin_sizeof call on this variable and hence to get
> the size relocation. What exactly do you mean
On 05/10/2017 01:05 PM, Uros Bizjak wrote:
On Wed, May 10, 2017 at 5:18 PM, Uros Bizjak wrote:
On Wed, May 10, 2017 at 4:27 PM, Jakub Jelinek wrote:
On Tue, May 09, 2017 at 06:06:47PM +0200, Uros Bizjak wrote:
Attached patch enables post-reload compare elimination pass by
providing expected
Hello!
Attached patch prevents unwanted peephole2 matching. With SSE, we can
expand atomic load through XMM registers using generic DImode move
patterns. It can happen that peephole2 matches unrelated insn
sequence, and this way removes random FP store to memory.
The patch introduces specialized
On 05/11/2017 01:50 PM, Rainer Orth wrote:
however, it breaks bootstrap with --enable-languages=obj-c++:
wierd, I thought --enable-languges=all enabled that (and I have seen
objc issues pop up during development). Will take another look.
nathan
--
Nathan Sidwell
GCC Maintainers:
This patch adds support for the various vec_popcnt() and vec_slo()
builtins.
The patch has been tested on powerpc64le-unknown-linux-gnu (Power 8 LE)
with no regressions.
Is the patch OK for gcc mainline?
Carl Love
-
Hi Nathan,
> On 05/08/2017 05:34 PM, Joseph Myers wrote:
>> On Mon, 8 May 2017, Nathan Sidwell wrote:
>
>>> This patch changes the C++ FE to override the pushdecl and getdecl lang
>>> hooks.
>>> In addition to simply overriding them there, I had to fixup a couple of
>>> places
>>> in c-family/c-co
Hi Nathan,
> On 05/11/2017 01:33 PM, Rainer Orth wrote:
>
>> unfortunately, this part
>>
>> * graphite-poly.c: Don't include tree-dump.h.
>>
>> broke bootstrap with graphite:
>
> oh, that has to be configured enabled? I've committed this that I think is
right: it requires an appropriate ver
On 05/11/2017 01:33 PM, Rainer Orth wrote:
unfortunately, this part
* graphite-poly.c: Don't include tree-dump.h.
broke bootstrap with graphite:
oh, that has to be configured enabled? I've committed this that I think
is obvious. looks to be another case of using tree-dumph to get
This patch reduces the variants of pushdecl by use of default arguments.
gone are the _maybe_friend variants.
nathan
--
Nathan Sidwell
2017-05-11 Nathan Sidwell
* name-lookup.h (pushdecl): Add default friend parm.
(pushdecl_maybe_friend): Delete.
(pushdecl_top_level): Add default friend
I meant to mention that I regstrapped this on both POWER8 and POWER9.
Bill
> On May 11, 2017, at 10:59 AM, Bill Schmidt
> wrote:
>
> Hi,
>
> PR80695 identifies a case (similar to several others we've seen) where SLP
> vectorization is too aggressive about vectorizing stores. The problem is
>
Hi Nathan,
> On 05/10/2017 04:49 AM, Richard Biener wrote:
>> On Tue, May 9, 2017 at 5:41 PM, Nathan Sidwell wrote:
>>> -fdump-translation-unit is an inscrutably opaque dump. It turned out that
>>> most of the uses of the tree-dump header file was to indirectly get at
>>> dumpfile.h, and the dum
On 05/10/2017 04:40 PM, Michael Meissner wrote:
As I mentioned in the mail message:
https://gcc.gnu.org/ml/gcc/2017-05/msg00060.html
I'm working on adding the target_clones attribute support to the PowerPC. I
have an implementation right now, but I want to iterate on it somewhat.
In doing the
Hi Jerry,
This patch is good for both trunk and 7-branch.
Thanks!
Paul
On 11 May 2017 at 16:35, Jerry DeLisle wrote:
> And the actual patch ...
>
> On 05/11/2017 08:30 AM, Jerry DeLisle wrote:
>>
>> Hi all,
>>
>> The attached patch fixes this issue by moving the DTIO namelist checks
>> from n
On 05/10/2017 02:27 PM, Jakub Jelinek wrote:
On Wed, May 10, 2017 at 09:57:56PM +0200, Uros Bizjak wrote:
BTW: This patch now catches 417 cases (instead of 200+) in linux
build, including e.g.:
(parallel [
(set (reg:CCZ 17 flags)
(compare:CCZ (lshiftrt:SI (reg:SI 4 si [ori
On 05/11/2017 10:34 AM, Jakub Jelinek wrote:
On Thu, May 11, 2017 at 10:23:48AM -0600, Martin Sebor wrote:
Unlike in C, the preferred way to initialize objects in C++
is to use some form of initialization (as opposed to memset).
The preferred way to copy objects is using the copy ctor or
assignm
On Thu, May 11, 2017 at 10:23:48AM -0600, Martin Sebor wrote:
> Unlike in C, the preferred way to initialize objects in C++
> is to use some form of initialization (as opposed to memset).
> The preferred way to copy objects is using the copy ctor or
> assignment operator (as opposed to memcpy). Us
On 04/30/2017 02:02 PM, Pedro Alves wrote:
Hi Martin,
Thanks much for doing this. A few comments below, in light of my
experience doing the equivalent checks in the gdb patch linked below,
using standard C++11.
Thanks for the feedback! It gave me quite a bit to think about.
The challenge wi
Hi,
PR80695 identifies a case (similar to several others we've seen) where SLP
vectorization is too aggressive about vectorizing stores. The problem is
that we undervalue the cost of a vec_construct operation. vec_construct
is the vectorizer's representation for building a vector from scalar
ele
Hi,
Add handling for early expansion of vector divide built-ins
Bootstrapped and tested on powerpc64le-unknown-linux-gnu, and
powerpc-unknown-linux, (p7,p8le,p8be) with no regressions.
OK for trunk?
Thanks,
-Will
[gcc]
2017-05-10 Will Schmidt
* config/rs6000/rs6000.c
Hi,
Add handling for early expansion of vector locical operations in gimple.
Specifically: vec_and, vec_andc, vec_or, vec_xor, vec_orc, vec_nand.
Bootstrapped and tested on powerpc64le-unknown-linux-gnu, and
powerpc-unknown-linux, (p7,p8le,p8be) with no regressions.
(This patch requires the
Hi,
I had initially posted this back in Dec, at which time it
was given an OK. Since a bunch of time has passed, I'm reposting
to refresh memories, and ensure it's still OK. :-)
I will be needing access to the create_tmp_reg_or_ssa_name()
function for some subsequent early expansion of
Hi,
Add support for some of the vector multiply (sp,dp) operations in
the gimple folding. I missed these during an earlier pass that
added the other (mule, mulo) multiply operations. This change is
covered by the existing testcases for vector folding
(gcc.target/powerpc/fold-vec-mul-*.c).
On 05/10/2017 04:49 AM, Richard Biener wrote:
On Tue, May 9, 2017 at 5:41 PM, Nathan Sidwell wrote:
-fdump-translation-unit is an inscrutably opaque dump. It turned out that
most of the uses of the tree-dump header file was to indirectly get at
dumpfile.h, and the dump_function entry point it
During untangling the name change made sense. But didn't at the final
outcome, so I've reverted it. Sorry for the churn.
nathan
--
Nathan Sidwell
2017-05-11 Nathan Sidwell
Revert pushdecl_top_level_and_finish name change.
* name-lookup.h (pushdecl_top_level_and_finish): Resurrect old nam
And the actual patch ...
On 05/11/2017 08:30 AM, Jerry DeLisle wrote:
Hi all,
The attached patch fixes this issue by moving the DTIO namelist checks from
namelist resolution to READ/WRITE statement resolution. This allows the checks
to be specific to the io_kind. The dtio_procs_present funct
Hi all,
The attached patch fixes this issue by moving the DTIO namelist checks from
namelist resolution to READ/WRITE statement resolution. This allows the checks
to be specific to the io_kind. The dtio_procs_present function is moved and
modified to accept the io_kind as an argument and chec
Included the workaround for SLP now. With it, testsuite is clean on x86
as well.
gcc/ChangeLog:
2017-05-11 Robin Dapp
* tree-vect-data-refs.c (vect_get_data_access_cost):
Workaround for SLP handling.
(vect_enhance_data_refs_alignment):
Remove check for supporta
Hi,
When ARMv8-M Baseline support was introduced no testcase was added for
support of the MOVT instruction because GCC would always prefer a load
from literal pool. With the recent work to fix -mslow-flash-data by
Andre Simoes Dias Vieira and Prakhar Bahuguna it is now possible to
test that suppo
Hi Richard,
On 10 May 2017 at 16:20, Richard Biener wrote:
>
> So this is a patch that makes skipping unreachable code when
> doing elimination possible. Previously interesting interactions
> with tail-merging made this impossible, now I seem to have
> figured a way around this.
>
> Bootstrappe
On Tue, Jan 17, 2017 at 9:48 AM, Richard Biener
wrote:
> On Tue, Jan 10, 2017 at 2:32 PM, Robin Dapp wrote:
>> Perhaps I'm still missing how some cases are handled or not handled,
>> sorry for the noise.
>>
>>> I'm not sure there is anything to "interpret" -- the operation is unsigned
>>> and ove
On 05/11/2017 11:02 AM, Bernhard Reutner-Fischer wrote:
On 11 May 2017 at 16:52, Nathan Sidwell wrote:
Bernhard's pestering finally got through. There was an early out that
failed to stop the timer.
fixed by rearranging the code.
:) thanks.
LGTM but i cannot approve it.
Oh, already commit
On 06/05/17 22:23 +0200, Stefan Schweter wrote:
Hi,
the current *libstdc++* documentation of the `mt_allocator`
implementation [1] points to a non-working link:
http://gcc.gnu.org/viewcvs/trunk/libstdc++-v3/testsuite/ext/mt_allocator/deallocate_local-6.cc?view=markup
The `gcc` prefix is missin
On Thu, May 11, 2017 at 4:02 PM, Richard Biener wrote:
>> > Uros added the testcase in 2008 -- I think if we want to have a testcase
>> > for the original issue we need a different one. Or simply remove
>> > the testcase.
>>
>> No, there is something going on in the testcase:
>>
>> .L3:
>>
On 11 May 2017 at 16:52, Nathan Sidwell wrote:
> Bernhard's pestering finally got through. There was an early out that
> failed to stop the timer.
>
> fixed by rearranging the code.
:) thanks.
LGTM but i cannot approve it.
Hi Bin,
On 10 May 2017 at 16:31, Bin.Cheng wrote:
> On Wed, May 3, 2017 at 11:09 AM, Kyrill Tkachov
> wrote:
>> Hi Bin,
>>
>>
>> On 03/05/17 11:02, Bin.Cheng wrote:
>>>
>>> On Wed, May 3, 2017 at 9:38 AM, Bin.Cheng wrote:
On Wed, May 3, 2017 at 7:17 AM, Eric Botcazou
wrote:
>>>
Bernhard's pestering finally got through. There was an early out that
failed to stop the timer.
fixed by rearranging the code.
nathan
--
Nathan Sidwell
2017-05-11 Nathan Sidwell
* name-lookup.c (pushdecl_outermost_localscope): Always
conditionally stop timer.
Index: name-lookup.c
==
On Wed, May 10, 2017 at 12:59:28PM -0500, Steven Munroe wrote:
> > That is just for the testsuite; I meant what happens if a user tries
> > to use it with an older target (or BE, or 32-bit)? Is there a useful,
> > obvious error message?
> >
> So looking at the X86 headers, their current practice
Hi Tom,
> 2017-05-01 Tom de Vries
>
> PR testsuite/65941
> * lib/target-supports.exp (check_effective_target_rdrand): New proc.
the new effective-target keyword needs documenting in sourcebuild.texi.
Thanks.
Rainer
--
-
We were frobing lamda scopes to regular classes when pushing them, but
with a few tweaks name lookup can cope directly.
nathan
--
Nathan Sidwell
2017-05-11 Nathan Sidwell
* decl.c (xref_tag_1): Don't frob ts_lambda scope here.
* name-lookup.c (pushtag_1): Deal with ts_lambda scope.
Index:
A simple thinko caused -da to start including ipa dumps. While
investigating I noticed we might also place the kind enumeration into
pflags, so defensively stripped that out.
Fixed thusly & committed.
nathan
--
Nathan Sidwell
2017-05-11 Nathan Sidwell
* dumpfile.h (TDF_FLAGS): New.
* du
std::make_shared didn't support embedding the object within the
shared_ptr bookkeeping structure when -fno-rtti was used. This was
because the trick I use for getting the address of the embedded
object relied on typeid, so without RTTI it just creating the object
separately on the heap.
This remo
On 05/11/2017 01:50 AM, Richard Biener wrote:
It actually seems to handle negation as well. Which means it
handles disjunctive normal form.
Negation is "handled" by allowing an individual predicate to be negated.
However, the predicate must still be is_neq_zero_form_p to really
participate
Bootstrapped and tested on x86_64-unknown-linux-gnu, applied.
Richard.
2017-05-11 Richard Biener
PR tree-optimization/80705
* tree-vect-data-refs.c (vect_analyze_data_refs): DECL_NONALIASED
bases are not vectorizable.
* gcc.dg/vect/bb-slp-pr80705.c: New testc
On Thu, 11 May 2017, Uros Bizjak wrote:
> On Thu, May 11, 2017 at 2:48 PM, Richard Biener wrote:
> > On Thu, 11 May 2017, Rainer Orth wrote:
> >
> >> Hi Richard,
> >>
> >> > On Mon, 24 Apr 2017, Richard Biener wrote:
> >> >>
> >> >> One issue in PR79201 is that we don't sink pure/const calls whic
tls-disable-literal-pool.c should only be run if the toolchain and target
support native thread-local storage rather than emulated TLS. This patch also
improves the matching of the error message.
testsuite/ChangeLog:
2017-05-11 Prakhar Bahuguna
* gcc.target/arm/tls-disable-literal-poo
I had an interesting time coming to grips with these two PRs. But it
essentially comes down to the fold call in save_expr. With that, we can call
fold() on an expression whose operands weren't folded yet, but that is what
code in fold_binary_loc assumes. Since fold() is not recursive, we could e
Hello!
> Currently, the TDF_foo flags serve 3 purposes:
> 1) what kind of dump
> 2) how detailed to print it
> 3) auxiliary message control
"-da" option now dumps all IPA dumps in addition to RTL dumps. Can
this behavior be reverted, so "-da" will dump only RTL dumps, as it
was before your patch?
On 05/11/2017 08:28 AM, Rainer Orth wrote:
On the gcc-7 branch, the dump file (from -fdump-class-hierarchy) was
called covariant7.C.002t.class, now it's covariant7.C.002l.class. But
gcc/testsuite/lib/scantree.exp (scan-tree-dump) expects the NNNt.
form.
fixed thusly, applied as obvious.
the
On Thu, May 11, 2017 at 2:59 PM, Koval, Julia wrote:
>> Please macroize existing rsqrt14_ pattern.
> There is no existing macro to match this pattern - it is "mask first element
> of vec_merge", instead of mask the whole vector.(SDM Vol. 2C 5-527)
> https://software.intel.com/sites/default/files
On Thu, May 11, 2017 at 2:48 PM, Richard Biener wrote:
> On Thu, 11 May 2017, Rainer Orth wrote:
>
>> Hi Richard,
>>
>> > On Mon, 24 Apr 2017, Richard Biener wrote:
>> >>
>> >> One issue in PR79201 is that we don't sink pure/const calls which is
>> >> what the following simple patch fixes.
>> >>
>
> Please macroize existing rsqrt14_ pattern.
There is no existing macro to match this pattern - it is "mask first element of
vec_merge", instead of mask the whole vector.(SDM Vol. 2C 5-527)
https://software.intel.com/sites/default/files/managed/39/c5/325462-sdm-vol-1-2abcd-3abcd.pdf
I can creat
On Thu, 11 May 2017, Rainer Orth wrote:
> Hi Richard,
>
> > On Mon, 24 Apr 2017, Richard Biener wrote:
> >>
> >> One issue in PR79201 is that we don't sink pure/const calls which is
> >> what the following simple patch fixes.
> >>
> >> Bootstrap and regtest running on x86_64-unknown-linux-gnu.
On Thu, May 11, 2017 at 2:15 PM, Richard Biener
wrote:
> On Thu, May 11, 2017 at 2:14 PM, Richard Biener
> wrote:
>> On Thu, May 11, 2017 at 1:15 PM, Robin Dapp wrote:
>>> Included the requested changes in the patches (to follow). I removed
>>> the alignment count check now altogether.
>>>
On 05/11/2017 08:28 AM, Rainer Orth wrote:
Hi Nathan,
UNRESOLVED: g++.dg/inherit/covariant7.C -std=gnu++98 scan-tree-dump class "24
.*c6::_ZTcv0_n16_v0_n12_NV2c62f2Ev"
UNRESOLVED: g++.dg/inherit/covariant7.C -std=gnu++98 scan-tree-dump class "28
.*c6::f2"
UNRESOLVED: g++.dg/inherit
Hi Nathan,
> On 05/10/2017 05:05 AM, Richard Biener wrote:
>> On Tue, May 9, 2017 at 9:00 PM, Nathan Sidwell wrote:
>
>> +
>> +#define TDF_ADDRESS(1 << 3)/* dump node addresses */
>>
>> this leaves 1 << 2 unused.
>
> Yes, that was intentional (though I suspect my note about it was hid
Hi!
Building on the other pending patches (I'll soon commit the approved
ones), we can then support the num_gangs, num_workers, vector_length
clauses for the OpenACC 2.5 kernels construct. OK for trunk?
commit a689c52cde71960bc08ae30c3f88980f66fdd0b8
Author: Thomas Schwinge
Date: Thu May 11 1
Hi!
OK for trunk?
commit 0ba48b4faf85420fbe12971afdd6e0afe70778bb
Author: Thomas Schwinge
Date: Fri May 5 16:41:59 2017 +0200
Runtime checking of OpenACC parallelism dimensions clauses
libgomp/
* testsuite/libgomp.oacc-c-c++-common/parallel-dims.c: Rewrite.
On 05/06/2017 09:20 AM, Uros Bizjak wrote:
On Tue, Apr 18, 2017 at 8:30 PM, Denys Vlasenko wrote:
2017-04-18 Denys Vlasenko
* config/i386/i386-common.c (ix86_handle_option): Remove support
for obsolete -malign-loops, -malign-jumps and -malign-functions
options.
* config/i386
Hi Richard,
> On Mon, 24 Apr 2017, Richard Biener wrote:
>>
>> One issue in PR79201 is that we don't sink pure/const calls which is
>> what the following simple patch fixes.
>>
>> Bootstrap and regtest running on x86_64-unknown-linux-gnu.
>
> Needed some gimple_assign_lhs -> gimple_get_lhs adjus
On Thu, May 11, 2017 at 2:14 PM, Richard Biener
wrote:
> On Thu, May 11, 2017 at 1:15 PM, Robin Dapp wrote:
>> Included the requested changes in the patches (to follow). I removed
>> the alignment count check now altogether.
>>
>>> I'm not sure why you test for unlimited_cost_model here as I sai
1 - 100 of 130 matches
Mail list logo
