Re: [PATCH] Fix ICEs with power8 fixuns_truncdi2 patterns (PR target/79197)

On Wed, Feb 01, 2017 at 11:56:34PM -0600, Segher Boessenkool wrote: > On Mon, Jan 30, 2017 at 10:27:15PM +0100, Jakub Jelinek wrote: > > Accoring to make mddump generated tmp-mddump.md, on powerpc the only pattern > > with unsigned_fix:DI where the inner operand is SF or DFmode is the > >

Re: [PATCH] use zero as the lower bound for a signed-unsigned range (PR 79327)

On Wed, Feb 01, 2017 at 04:52:35PM -0700, Martin Sebor wrote: > --- a/gcc/gimple-ssa-sprintf.c > +++ b/gcc/gimple-ssa-sprintf.c > @@ -1382,13 +1382,26 @@ format_integer (const directive , tree arg) > would reflect the largest possible precision (i.e., INT_MAX). */ >res.range.min

Re: [PATCH, pr63256] update powerpc dg options

On Thu, Jan 19, 2017 at 05:00:18PM -0600, Aaron Sawdey wrote: > SMS does process the loop in sms-8.c on powerpc now so I have updated > the options to reflect that. It passes with the --param, yes. Okay for trunk. Thanks, Segher

Re: [PATCH] Fix ICEs with power8 fixuns_truncdi2 patterns (PR target/79197)

On Tue, Jan 31, 2017 at 09:11:11AM +0100, Jakub Jelinek wrote: > What I think this patch doesn't handle properly is SFmode -> DImode > unsigned_fix. I have no idea how SFmode is represented in vector registers, > so have no idea if xscvdpuxds actually handles SFmode as well. All floating point

Re: [PATCH] Fix ICEs with power8 fixuns_truncdi2 patterns (PR target/79197)

On Mon, Jan 30, 2017 at 10:27:15PM +0100, Jakub Jelinek wrote: > Accoring to make mddump generated tmp-mddump.md, on powerpc the only pattern > with unsigned_fix:DI where the inner operand is SF or DFmode is the > *fixuns_truncdi2_fctiduz. There is an expander for that instruction, > which uses

Re: [PATCH doc] clean up -fdump-tree- options (PR 32003)

On 02/01/2017 08:26 PM, Martin Sebor wrote: On 02/01/2017 08:06 PM, Sandra Loosemore wrote: On 02/01/2017 06:57 PM, Martin Sebor wrote: As discussed in bug 32003 - Undocumented -fdump-tree options, rather than duplicating the same boiler-plate text for each of the dozens (138 by my count) of

Re: [PATCH/AARCH64] Add scheduler for Thunderx2t99

Hi James, Thanks for reviewing the patch and comments. >> I wonder whether the current modeling of: >> (define_insn_reservation "thunderx2t99_asimd_load4_elts" 6 >> Actually benefits the schedule in a meaningful way, or if it just increases Done. Removed the scheduler modeling for

Re: [PATCH] [AArch64] Enable AES and cmp_branch fusion for Thunderx2t99

Hi Kyrill, Thanks for the review and comments. >> but there are a couple of issues with the ChangeLog 2017-02-02 Naveen H.S * config/aarch64/aarch64.c (thunderx2t99_tunings): Enable AES and cmp_branch fusion. Thanks, Naveen

Re: [PATCH] [AArch64] Implement popcount pattern

Hi James and Kyrill, Thanks for the review and comments on the patch. >> On ILP32 systems this would still use the SImode patterns, >> so I suggest you use __builtin_popcountll and >> an unsigned long long return type to ensure you always exercise the 64-bit >> code. Sorry for not commenting

Re: [PATCH doc] clean up -fdump-tree- options (PR 32003)

On 02/01/2017 08:06 PM, Sandra Loosemore wrote: On 02/01/2017 06:57 PM, Martin Sebor wrote: As discussed in bug 32003 - Undocumented -fdump-tree options, rather than duplicating the same boiler-plate text for each of the dozens (138 by my count) of undocumented passes, the attached patch

Re: [PATCH doc] clean up -fdump-tree- options (PR 32003)

On 02/01/2017 06:57 PM, Martin Sebor wrote: As discussed in bug 32003 - Undocumented -fdump-tree options, rather than duplicating the same boiler-plate text for each of the dozens (138 by my count) of undocumented passes, the attached patch removes the pass-specific -fdump-tree- options

Re: [RFC] Bug lto/78140

Hi Richard, On 30/01/17 21:08, Richard Biener wrote: On Mon, Jan 30, 2017 at 12:23 AM, kugan wrote: Hi All, As suggested by Richard in the PR, I tried to implement variable size structures for VR as shown in attached patch. That is, I changed ipa-prop.h to:

Re: [PATCH] Fix PR79278, amend ADJUST_FIELD_ALIGN interface

On Wed, Feb 01, 2017 at 03:53:09PM -0600, Segher Boessenkool wrote: > On Wed, Feb 01, 2017 at 11:59:10AM +0100, Richard Biener wrote: > > Wasn't successful in making a cross to ppc64-linux build its libobjc. > > I'll do a native build. Just the patch in the first message in this > thread? And

[PATCH doc] clean up -fdump-tree- options (PR 32003)

As discussed in bug 32003 - Undocumented -fdump-tree options, rather than duplicating the same boiler-plate text for each of the dozens (138 by my count) of undocumented passes, the attached patch removes the pass-specific -fdump-tree- options replacing them with a list of generic steps to

Re: [RFC] Bug lto/78140

Hi Richard, On 30/01/17 21:08, Richard Biener wrote: On Mon, Jan 30, 2017 at 12:23 AM, kugan wrote: lto1: internal compiler error: Segmentation fault 0xdedc4b crash_signal ../../gcc/gcc/toplev.c:333 0xb46680 ipa_node_params_t::duplicate(cgraph_node*,

Re: [PATCH] relax -Wformat-overflow for precision ranges (PR 79275)

On 01/31/2017 03:33 PM, Jeff Law wrote: On 01/30/2017 02:28 PM, Martin Sebor wrote: Bug 79275 - -Wformat-overflow false positive exceeding INT_MAX in glibc sysdeps/posix/tempname.c points out a false positive found during a Glibc build and caused by the checker using the upper bound of a range

[PATCH] use zero as the lower bound for a signed-unsigned range (PR 79327)

PR tree-optimization/79327 exposes a bug in the gimple-ssa-sprintf pass in the computation of the expected size of output of an integer format directive whose argument is in a signed-unsigned range (such as [-12, 345]). The current algorithm on trunk uses the bounds of the range to compute the

Re: Go patch committed: Fixes for m68k

On Mon, Jan 23, 2017 at 10:15 AM, Ian Lance Taylor wrote: > I committed a patch to the Go frontend and libgo to fix alignment > issues on m68k. This fixes PR 79037. Bootstrapped and ran Go tests > on x86_64-pc-linux-gnu. Tested by John Paul Adrian Glaubitz on > m68k-linux-gnu.

Re: [PATCH] Add support for Fuchsia (OS)

On Tue, 17 Jan 2017, Josh Conner via gcc-patches wrote: > Attached is my recommended patch for changes to the web docs describing > Fuchsia support. Please let me know if there's anything else I can do. This looks fine (just remove the blank before "Fuchsia"); and sorry for the delay getting back

GCC 6 patch: Add aligned attribute for m68k support

This patch by John Paul Adrian Glaubitz fixes m68k libgo on the GCC 6 branch, by adding alignment required by the kernel but not otherwise imposed by the m68k backend. I bootstrapped and ran Go tests on x86_64-pc-linux-gnu, Adrian tested on m68k. Committed to GCC 6 branch. Ian @@ -, +, @@ ---

New German PO file for 'gcc' (version 7.1-b20170101)

Hello, gentle maintainer. This is a message from the Translation Project robot. A revised PO file for textual domain 'gcc' has been submitted by the German team of translators. The file is available at: http://translationproject.org/latest/gcc/de.po (This file, 'gcc-7.1-b20170101.de.po',

Re: [PATCH] Fix PR79278, amend ADJUST_FIELD_ALIGN interface

On Tue, Jan 31, 2017 at 10:01:46AM +0100, Richard Biener wrote: > * doc/tm.texi.in (ADJUST_FIELD_ALIGN): Adjust to take additional > type parameter. > * doc/tm.texi: Regenerate. You didn't include tm.texi.in in the patch, only tm.texi . Segher

Re: [PATCH] Fix PR79278, amend ADJUST_FIELD_ALIGN interface

On Wed, Feb 01, 2017 at 11:59:10AM +0100, Richard Biener wrote: > Wasn't successful in making a cross to ppc64-linux build its libobjc. I'll do a native build. Just the patch in the first message in this thread? And just running the testsuite is enough, or is there something specific you want

Re: [PATCH] Fix ada __gnat_killprocesstree (PR ada/79309)

> I believe > is right. sizeof (statfile) is 64, sizeof ("/proc//stat") is > 12 (that includes the terminating '\0'), and 52 characters long d->d_name > still fits (6 bytes /proc/, 52 bytes d->d_name, 5 bytes /stat and 1 byte > '\0') while 53 characters are too much. Equivalent of the above

Re: [PATCH] Fix ada __gnat_killprocesstree (PR ada/79309)

On Wed, Feb 01, 2017 at 09:47:40PM +0100, Eric Botcazou wrote: > > 2017-02-01 Jakub Jelinek > > > > PR ada/79309 > > * adaint.c (__gnat_killprocesstree): Don't clear statfile > > before overwriting it. If d->d_name is too long, skip trying > > to construct

Re: [PATCH] Fix ada __gnat_killprocesstree (PR ada/79309)

> As mentioned in the PR, strncat does something different from what the > code expects (the last argument is the maximum number of characters > to be copied, rather than maximum number of characters in the destination > buffer). As for the (highly unlikely, because d->d_name really should be >

[Ada] Fix PR ada/79309

This is the questionable string logic in __gnat_killprocesstree detected by the new -Wstringop-overflow= warning. The fix is Jakub's. Tested on x86_64-suse-linux, applied on the mainline. 2017-02-01 Eric Botcazou Jakub Jelinek

Re: [PATCH, pr63256] update powerpc dg options

On Thu, 2017-01-19 at 17:00 -0600, Aaron Sawdey wrote: > SMS does process the loop in sms-8.c on powerpc now so I have updated > the options to reflect that. > > Test now passes on powerpc -m64/-m32/-m32 -mpowerpc64. Ok for trunk? > > testsuite/ChangeLog > 2017-01-19  Aaron Sawdey  

[wwwdocs] gcc-3.2/c++-abi.html - adjust link text for C++ ABI

Applied, after a friendly hint off-list. At first I was wondering whether to leave the text, but then it had been updated already earlier (to mentor.com), so why not fix it again? Gerald Index: gcc-3.2/c++-abi.html === RCS file:

[PATCH] Fix ada __gnat_killprocesstree (PR ada/79309)

Hi! As mentioned in the PR, strncat does something different from what the code expects (the last argument is the maximum number of characters to be copied, rather than maximum number of characters in the destination buffer). As for the (highly unlikely, because d->d_name really should be the

Re: [PATCH, rs6000] Fix PR70012 (gcc.dg/vect/costmodel/ppc/costmodel-vect-33.c)

On Wed, Feb 01, 2017 at 10:42:31AM -0600, Bill Schmidt wrote: > https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70012 reports that the subject > test case is now failing in some circumstances. Prior to POWER8, the test > was checking for conditions that apply when data alignment is unknown; either >

Re: [PATCH, wwwdocs/ARM] Mention new rmprofile value for --with-multilib-list

On Mon, 30 Jan 2017, Thomas Preudhomme wrote: ARM backend now support a new set of multilib libraries enabled with --with-multilib-list=rmprofile [1]. This patch documents it in the changes for GCC 7. I will admit that even following this list, I had to look up what this really referred to.

Re: [patch 79279] combine/simplify_set issue

Hi Aurelien, On Wed, Feb 01, 2017 at 10:02:56AM +0100, Aurelien Buhrig wrote: > >> This patch fixes a combiner bug in simplify_set which calls > >> CANNOT_CHANGE_MODE_CLASS with wrong mode params. > >> It occurs when trying to simplify paradoxical subregs of hw regs (whose > >> natural mode is

Re: [wwwdocs] changes.html - document new warning options

On Tue, 31 Jan 2017, Martin Sebor wrote: > Thanks for the careful review (and debugging)! Thanks for taking the time to prepare all this to begin with. ;-) On Wed, 1 Feb 2017, Jakub Jelinek wrote: >>> void f (size_t n) >>> { >>> char *d = alloca (n) > Missing semicolon after alloca (n)

Re: [gomp4] enable GOMP_MAP_FIRSTPRIVATE_INT in OpenACC

On 01/30/2017 02:18 AM, Thomas Schwinge wrote: > Hi Cesar! > > On Fri, 27 Jan 2017 07:45:52 -0800, Cesar Philippidis > wrote: >> If you take a close look at lower_omp_target, you'll notice that I'm >> gave reference types special treatment. Specifically, I disabled this

Re: [gomp4] add support for derived types in ACC UPDATE

On 01/30/2017 02:08 AM, Thomas Schwinge wrote: > Hi Cesar! > > On Thu, 10 Nov 2016 09:38:33 -0800, Cesar Philippidis > wrote: >> This patch has been committed to gomp-4_0-branch. > >> --- a/gcc/fortran/openmp.c >> +++ b/gcc/fortran/openmp.c > >> @@ -242,7 +243,8 @@

Re: [PATCH] Fix __atomic to not implement atomic loads with CAS.

On Mon, 2017-01-30 at 19:54 +0100, Torvald Riegel wrote: > This patch fixes the __atomic builtins to not implement supposedly > lock-free atomic loads based on just a compare-and-swap operation. After an off-list OK by Jakub, I have committed this as r245098. Jakub will take care of the OpenMP

Re: [PATCH 0/5] OpenMP/PTX: improve correctness in SIMD regions

On Wed, Feb 01, 2017 at 08:09:27PM +0300, Alexander Monakov wrote: > > That said, I think pointers to gimple stmts in struct loop or something > > similar is problematic, you'd need to adjust those whenever something would > > remove those stmts, or e.g. duplicate the loop and stmts, handle those

Re: [PATCH 0/5] OpenMP/PTX: improve correctness in SIMD regions

On Wed, 1 Feb 2017, Jakub Jelinek wrote: > > Yes; I imagine the approach taken in patch 2/5 can be extended to achieve > > this. > > That is, instead of just storing a flag 'bool in_simtreg' in struct loop, > > store > > pointers to corresponding SIMT_ENTER/EXIT gimple statements, use a similar

[PATCH, rs6000] Fix PR70012 (gcc.dg/vect/costmodel/ppc/costmodel-vect-33.c)

Hi, https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70012 reports that the subject test case is now failing in some circumstances. Prior to POWER8, the test was checking for conditions that apply when data alignment is unknown; either peeling for alignment or versioning for alignment may be used.

Re: [PATCH] XFAIL gcc.dg/graphite/scop-dsyrk.c

Sorry for duplicates, I'm resending as plain text for the mailing list. On Wed, Feb 1, 2017 at 6:57 AM, Richard Biener wrote: > > > The following XFAILs the testcases, making them fail reliably independelty > of int/long type sizes and also providing new testcase variants that

Re: [PR63238] output alignment debug information

On Fri, Jan 27, 2017 at 04:24:58AM -0200, Alexandre Oliva wrote: > Output DWARFv5+ DW_AT_alignment for non-default alignment of > variables, fields and types. The new tests all fail on targets that default to -gstrict-dwarf because they have buggy or prehistoric linkers/debug info consumers like

Re: [PATCH] Provide opt-info for autopar and graphite

On Wed, Feb 1, 2017 at 4:43 AM, Richard Biener wrote: > * graphite.c: Include tree-vectorizer.h for find_loop_location. > (graphite_transform_loops): Provide opt-info for optimized nests. > * tree-parloop.c (parallelize_loops): Provide opt-info for >

Re: [PATCH 0/5] OpenMP/PTX: improve correctness in SIMD regions

On Wed, Feb 01, 2017 at 06:44:39PM +0300, Alexander Monakov wrote: > > That said, I understand how would you add these arguments during > > lowering, but don't understand what would you want to do during inlining, > > if you have addressable vars in inlined function, you need to avoid > >

Re: [PATCH 0/5] OpenMP/PTX: improve correctness in SIMD regions

On Wed, 1 Feb 2017, Jakub Jelinek wrote: > IFN_ASAN_POISON is treated that way too. That also means that if a > variable is previously addressable and the only spot that takes its address > is that IFN, it can be rewritten into SSA form, but the IFN has to be > adjusted to something different

Re: [PATCH][C++] Save memory/time when folding CONSTRUCTORs

On Wed, 1 Feb 2017, Nathan Sidwell wrote: > On 02/01/2017 10:18 AM, Richard Biener wrote: > > > True. Updated patch below. > > > > Richard. > > > > 2017-02-01 Richard Biener > > > > PR cp/14179 > > cp/ > > * cp-gimplify.c (cp_fold): When folding a

Re: [PATCH] Fix PR79278, amend ADJUST_FIELD_ALIGN interface

On Wed, Feb 01, 2017 at 10:27:27AM -0500, Trevor Saunders wrote: > Yeah, different targets deal with this mess in different ways, looks > like frv uses a different macro and ppc redefines macros as needed > rs6000_special_adjust_field_align_p is also defined in ppc headers. > > > > > /* @@@ A

Re: [PATCH][C++] Save memory/time when folding CONSTRUCTORs

On 02/01/2017 10:18 AM, Richard Biener wrote: True. Updated patch below. Richard. 2017-02-01 Richard Biener PR cp/14179 cp/ * cp-gimplify.c (cp_fold): When folding a CONSTRUCTOR copy it lazily on the first changed element only and copy

Re: [PATCH][C++] Save memory/time when folding CONSTRUCTORs

On Wed, Feb 01, 2017 at 04:18:49PM +0100, Richard Biener wrote: > 2017-02-01 Richard Biener > > PR cp/14179 > cp/ > * cp-gimplify.c (cp_fold): When folding a CONSTRUCTOR copy > it lazily on the first changed element only and copy it > fully

Re: [PATCH][C++] Save memory/time when folding CONSTRUCTORs

On Wed, 1 Feb 2017, Nathan Sidwell wrote: > On 02/01/2017 09:03 AM, Richard Biener wrote: > > > > Currently we copy CONSTRUCTORs we cp_fold even if no elements fold > > (we throw the copy away then). That's wasteful and we can easily > > do the copying on-demand. For simplicity the following

Re: [PATCH][C++] Improve memory use for PR12245

On Wed, 1 Feb 2017, Jakub Jelinek wrote: > On Wed, Feb 01, 2017 at 02:14:20PM +0100, Richard Biener wrote: > > > > Looks like we cache the answer to maybe_constant_value (INTEGER_CST) > > which results in (-fmem-report): > > > > cp/constexpr.c:4814 (maybe_constant_value) 67108816:100.0%

Re: [PATCH][C++] Save memory/time when folding CONSTRUCTORs

On 02/01/2017 09:03 AM, Richard Biener wrote: Currently we copy CONSTRUCTORs we cp_fold even if no elements fold (we throw the copy away then). That's wasteful and we can easily do the copying on-demand. For simplicity the following resorts to memcpy-ing the whole original vector on the first

Re: [PATCH] Fix PR79278, amend ADJUST_FIELD_ALIGN interface

On Wed, Feb 01, 2017 at 09:42:08AM +0100, Richard Biener wrote: > On Tue, 31 Jan 2017, Jeff Law wrote: > > > On 01/31/2017 02:01 AM, Richard Biener wrote: > > > > > > This amends ADJUST_FIELD_ALIGN to always get the type of the field > > > as argument but make the field itself optional. All

Re: [C++ Patch] PR 69637

On 02/01/2017 08:24 AM, Paolo Carlini wrote: Hi, I'm still working on a number of ICEs on invalid happening during error recovery... In these cases, after a meaningful diagnostic, we ICE in cxx_eval_constant_expression because it doesn't handle OVERLOADs and TEMPLATE_ID_EXPRs in the main

Re: [PATCH 0/5] OpenMP/PTX: improve correctness in SIMD regions

On Wed, Feb 01, 2017 at 04:28:14PM +0300, Alexander Monakov wrote: > Hi, > > Earlier Richard mentioned the possibility to special-case GOMP_SIMT_ENTER to > allow passing privatized variables to it by reference without making them > addressable. I now see that such special-casing is already done

Re: [PATCH][C++] Save memory/time when folding CONSTRUCTORs

On Wed, 1 Feb 2017, Jakub Jelinek wrote: > On Wed, Feb 01, 2017 at 03:35:49PM +0100, Richard Biener wrote: > > > > changed = false; > > > > break; > > > > } > > > > - changed = true; > > > > + if (! changed) > >

Re: [PATCH][C++] Save memory/time when folding CONSTRUCTORs

On Wed, Feb 01, 2017 at 03:35:49PM +0100, Richard Biener wrote: > > > changed = false; > > > break; > > > } > > > - changed = true; > > > + if (! changed) > > > + { > > > + nelts = elts->copy (); > > > > Isn't the above

Re: [PATCH][C++] Improve memory use for PR12245

On Wed, Feb 01, 2017 at 02:14:20PM +0100, Richard Biener wrote: > > Looks like we cache the answer to maybe_constant_value (INTEGER_CST) > which results in (-fmem-report): > > cp/constexpr.c:4814 (maybe_constant_value) 67108816:100.0% > 10066310417: 0.0% ggc > > this can

Re: [PATCH][C++] Save memory/time when folding CONSTRUCTORs

On Wed, 1 Feb 2017, Jakub Jelinek wrote: > On Wed, Feb 01, 2017 at 03:03:15PM +0100, Richard Biener wrote: > > 2017-02-01 Richard Biener > > > > PR cp/14179 > > cp/ > > * cp-gimplify.c (cp_fold): When folding a CONSTRUCTOR copy > > it lazily on the first

Re: [PATCH v2] aarch64: Add split-stack initial support

Hi Adhermerval, The argument code looks good now, but this isn't right: +  int ninsn = aarch64_internal_mov_immediate (reg10, GEN_INT (-allocate), + true, Pmode); +  gcc_assert (ninsn == 1 || ninsn == 2); +  if (ninsn == 1) +    { +  if (allocate >

Re: [PATCH][C++] Save memory/time when folding CONSTRUCTORs

On Wed, Feb 01, 2017 at 03:03:15PM +0100, Richard Biener wrote: > 2017-02-01 Richard Biener > > PR cp/14179 > cp/ > * cp-gimplify.c (cp_fold): When folding a CONSTRUCTOR copy > it lazily on the first changed element only and copy it > fully

Re: [PATCH] Add dg-require-alias to a ICF test (PR testsuite/79272).

On 02/01/2017 02:14 PM, John David Anglin wrote: > On 2017-02-01, at 8:03 AM, Martin Liška wrote: > >> As mentioned in the PR, hppa2.0w-hp-hpux11.11 does not support aliasing and >> thus >> the scanned pattern is invalid. Fixed in patch. > > Looks fine to me and obvious. Ok, installed as

[PATCH][C++] Save memory/time when folding CONSTRUCTORs

Currently we copy CONSTRUCTORs we cp_fold even if no elements fold (we throw the copy away then). That's wasteful and we can easily do the copying on-demand. For simplicity the following resorts to memcpy-ing the whole original vector on the first change and only overwrites changed elements in

Re: [PATCH] [AArch64] Implement popcount pattern

On Tue, Dec 13, 2016 at 11:59:36AM +, Kyrill Tkachov wrote: > Hi Naveen, > > On 13/12/16 11:51, Hurugalawadi, Naveen wrote: > >Hi Kyrill, > > > >Thanks for reviewing the patch and your useful comments. > > > >>>looks good to me if it has gone through the normal required > >>>bootstrap and

PR79286, ira combine_and_move_insns in loops

This patch cures PR79286 by restoring the REG_DEAD note test used prior to r235660, but modified to only exclude insns that may trap. I'd like to allow combine/move without a REG_DEAD note in loops because insns in loops often lack such notes, and I recall seeing quite a few cases at the time I

Re: [PATCH 0/5] OpenMP/PTX: improve correctness in SIMD regions

Hi, Earlier Richard mentioned the possibility to special-case GOMP_SIMT_ENTER to allow passing privatized variables to it by reference without making them addressable. I now see that such special-casing is already done for IFN_ATOMIC_COMPARE_EXCHANGE in tree-ssa.c: execute_update_addresses_taken

[C++ Patch] PR 69637

Hi, I'm still working on a number of ICEs on invalid happening during error recovery... In these cases, after a meaningful diagnostic, we ICE in cxx_eval_constant_expression because it doesn't handle OVERLOADs and TEMPLATE_ID_EXPRs in the main switch. I tried a number of different

Re: [PATCH] Add dg-require-alias to a ICF test (PR testsuite/79272).

On 2017-02-01, at 8:03 AM, Martin Liška wrote: > As mentioned in the PR, hppa2.0w-hp-hpux11.11 does not support aliasing and > thus > the scanned pattern is invalid. Fixed in patch. Looks fine to me and obvious. Dave -- John David Anglin dave.ang...@bell.net

[PATCH][C++] Improve memory use for PR12245

Looks like we cache the answer to maybe_constant_value (INTEGER_CST) which results in (-fmem-report): cp/constexpr.c:4814 (maybe_constant_value) 67108816:100.0% 10066310417: 0.0% ggc this can be improved trivially to cp/constexpr.c:4817 (maybe_constant_value)

[PATCH] Add dg-require-alias to a ICF test (PR testsuite/79272).

Hello. As mentioned in the PR, hppa2.0w-hp-hpux11.11 does not support aliasing and thus the scanned pattern is invalid. Fixed in patch. Ready to be installed? Martin >From 2bc48dc64eeb0dbe55ca8f9f5abe2841f78c3c80 Mon Sep 17 00:00:00 2001 From: marxin Date: Tue, 31 Jan 2017

[PATCH] PR78346 make handle stashing iterators

I didn't get an answer explaining why these function objects store a reference not the iterator, so I'm fixing the regression by storing the iterator and dereferencing it on every comparison. PR libstdc++/78346 * include/bits/predefined_ops.h (_Iter_equals_iter): Store iterator

[PATCH] XFAIL gcc.dg/graphite/scop-dsyrk.c

The following XFAILs the testcases, making them fail reliably independelty of int/long type sizes and also providing new testcase variants that succeed reliably. I don't see us fixing the underlying niter analysis issue for GCC 7. Tested on x86_64-unknown-linux-gnu with {,-m32}, applied.

Re: [PATCH] PR libstdc++/79195 fix make_array type deduction

On 23/01/17 15:55 +, Jonathan Wakely wrote: This adds indirection through a class template partial specialization so we don't instantiate common_type<_Types...>::type unless we need it. PR libstdc++/79195 * include/experimental/array (__make_array_elem): New class template

Re: [PATCH] PR libstdc++/79254 fix exception-safety in std::string::operator=

On 01/02/17 11:42 +, Jonathan Wakely wrote: On 27/01/17 16:16 +, Jonathan Wakely wrote: This implements the strong exception-safety guarantee that is required by [string.require] p2, which the new string can fail to meet when propagate_on_container_copy_assignment (POCCA) is true. The

Re: [PATCH] Fix PGO bootstrap on x390x (PR bootstrap/78985).

On Wed, Feb 01, 2017 at 12:52:16PM +0100, Martin Liška wrote: > On 02/01/2017 11:38 AM, Jakub Jelinek wrote: > > On Wed, Feb 01, 2017 at 11:34:48AM +0100, Martin Liška wrote: > >>> Presumably the issue with print_operand_address is that there are paths > >>> where s390_decompose_address can

Re: [PATCH] Fix PGO bootstrap on x390x (PR bootstrap/78985).

On 02/01/2017 11:38 AM, Jakub Jelinek wrote: > On Wed, Feb 01, 2017 at 11:34:48AM +0100, Martin Liška wrote: >>> Presumably the issue with print_operand_address is that there are paths >>> where s390_decompose_address can return without initializing AD/OUT. But >>> AFAICT those are invalid

Re: gcc.css colors

On 2017.02.01 at 12:41 +0100, Jakub Jelinek wrote: > Hi! > > On Wed, Feb 01, 2017 at 12:27:05PM +0100, Markus Trippelsdorf wrote: > > > I've tried various color settings of gnome-terminal (both white on black > > > and > > > black on white plus the different color sets) and all of them except >

Re: gcc.css colors

On Wed, Feb 01, 2017 at 12:33:30PM +0100, Markus Trippelsdorf wrote: > > E.g. in changes.html, the Asan spans with boldlime are using > > \033[1m\033[32m > > in libsanitizer, and the note color in gcc by default is > > \033[1m\033[36m > > > > I've tried various color settings of gnome-terminal

Re: [patch] Fix PR middle-end/78468

> I'd say let's not have a middle ground - this stuff is sufficiently > brain-twisting that I'd rather go back to a known working state. If > there was an error in the previous patch, let's roll it back until we > fully understand the situation. Here's a revised version along these lines, OK for

Re: [PATCH] PR libstdc++/79254 fix exception-safety in std::string::operator=

On 27/01/17 16:16 +, Jonathan Wakely wrote: This implements the strong exception-safety guarantee that is required by [string.require] p2, which the new string can fail to meet when propagate_on_container_copy_assignment (POCCA) is true. The solution is to define a helper that takes

Re: gcc.css colors

On 2017.02.01 at 12:14 +0100, Jakub Jelinek wrote: > On Wed, Feb 01, 2017 at 11:55:53AM +0100, Markus Trippelsdorf wrote: > > On 2017.02.01 at 11:48 +0100, Jakub Jelinek wrote: > > > On Wed, Feb 01, 2017 at 11:45:14AM +0100, Markus Trippelsdorf wrote: > > > > Some colors on e.g.

Re: gcc.css colors

On Wed, Feb 01, 2017 at 11:55:53AM +0100, Markus Trippelsdorf wrote: > On 2017.02.01 at 11:48 +0100, Jakub Jelinek wrote: > > On Wed, Feb 01, 2017 at 11:45:14AM +0100, Markus Trippelsdorf wrote: > > > Some colors on e.g. https://gcc.gnu.org/gcc-7/changes.html are nearly > > > unreadable. So what

Re: gcc.css colors

On Wed, 1 Feb 2017, Markus Trippelsdorf wrote: Some colors on e.g. https://gcc.gnu.org/gcc-7/changes.html are nearly unreadable. I recently noticed that gcc's website has an extremely strict Content Security Policy, which makes it harder to customize its appearance using for instance

Re: [PATCH] Fix PR79278, amend ADJUST_FIELD_ALIGN interface

On Wed, 1 Feb 2017, Iain Sandoe wrote: > > > On 1 Feb 2017, at 08:42, Richard Biener wrote: > > > > On Tue, 31 Jan 2017, Jeff Law wrote: > > > >> On 01/31/2017 02:01 AM, Richard Biener wrote: > >>> > >>> This amends ADJUST_FIELD_ALIGN to always get the type of the field >

Re: gcc.css colors

On 2017.02.01 at 11:48 +0100, Jakub Jelinek wrote: > On Wed, Feb 01, 2017 at 11:45:14AM +0100, Markus Trippelsdorf wrote: > > Some colors on e.g. https://gcc.gnu.org/gcc-7/changes.html are nearly > > unreadable. So what about the following patch? > > > > --- gcc_orig.css2017-02-01

Re: [wwwdocs] changes.html - document new warning options

On Wed, Feb 01, 2017 at 05:50:08AM -0500, Aldy Hernandez wrote: > > > Since you ask so nicely I added another example but I'm afraid it > > isn't terribly interesting: > > > > In contrast, a call to alloca that isn't bounded at all such as > > in the following function will elicit the

Re: [wwwdocs] changes.html - document new warning options

Since you ask so nicely I added another example but I'm afraid it isn't terribly interesting: In contrast, a call to alloca that isn't bounded at all such as in the following function will elicit the warning below regardless of the size argument to the option. void f (size_t n) {

Re: gcc.css colors

On Wed, Feb 01, 2017 at 11:45:14AM +0100, Markus Trippelsdorf wrote: > Some colors on e.g. https://gcc.gnu.org/gcc-7/changes.html are nearly > unreadable. So what about the following patch? > > --- gcc_orig.css 2017-02-01 11:39:17.634017498 +0100 > +++ gcc.css 2017-02-01 11:40:23.979244263

Re: [PATCH] Fix PR79278, amend ADJUST_FIELD_ALIGN interface

> On 1 Feb 2017, at 08:42, Richard Biener wrote: > > On Tue, 31 Jan 2017, Jeff Law wrote: > >> On 01/31/2017 02:01 AM, Richard Biener wrote: >>> >>> This amends ADJUST_FIELD_ALIGN to always get the type of the field >>> as argument but make the field itself optional. All

Re: [PATCH] Fix bool vs. unsigned:1 vectorization (PR tree-optimization/79284)

On Wed, 1 Feb 2017, Jakub Jelinek wrote: > On Wed, Feb 01, 2017 at 11:07:18AM +0100, Richard Biener wrote: > > > On Wed, Feb 01, 2017 at 10:58:29AM +0100, Richard Biener wrote: > > > > > > +/* Nonzero if TYPE represents a (scalar) boolean type or type > > > > > > + in the middle-end compatible

gcc.css colors

Some colors on e.g. https://gcc.gnu.org/gcc-7/changes.html are nearly unreadable. So what about the following patch? --- gcc_orig.css2017-02-01 11:39:17.634017498 +0100 +++ gcc.css 2017-02-01 11:40:23.979244263 +0100 @@ -58,8 +58,8 @@ } div.copyright p:nth-child(3) { margin-bottom:

[PATCH] Provide opt-info for autopar and graphite

The following provides opt-info for autopar and graphite. That should help to track down miscompilations by them given they usually do not apply in many places (like for hunting down PR79321). I've used the loop location of the loop containing the SCOP which is probably reasonable for an option

[PATCH] Fix PR79315

The following fixes ICEs when building SPEC 2k6 with autopar. Bootstrapped on x86_64-unknown-linux-gnu, testing in progress. Richard. 2017-02-01 Richard Biener PR middle-end/79315 * tree-cfg.c (move_stmt_op): Never set TREE_BLOCK when it was not

Re: [PATCH] Fix PGO bootstrap on x390x (PR bootstrap/78985).

On Wed, Feb 01, 2017 at 11:34:48AM +0100, Martin Liška wrote: > > Presumably the issue with print_operand_address is that there are paths > > where s390_decompose_address can return without initializing AD/OUT. But > > AFAICT those are invalid addresses that presumably shouldn't be showing up >

Re: [PATCH] Fix PGO bootstrap on x390x (PR bootstrap/78985).

On 01/31/2017 05:47 PM, Jeff Law wrote: > On 01/30/2017 06:32 AM, Martin Liška wrote: >> On 01/30/2017 12:27 PM, Martin Liška wrote: >>> Hi. >>> >>> Following patch simply fixes issues reported by -Wmaybe-unitialized. That >>> enables PGO bootstrap >>> on a s390x machine. >>> >>> Ready to be

Re: [PATCH] Fix bool vs. unsigned:1 vectorization (PR tree-optimization/79284)

On Wed, Feb 01, 2017 at 11:07:18AM +0100, Richard Biener wrote: > > On Wed, Feb 01, 2017 at 10:58:29AM +0100, Richard Biener wrote: > > > > > +/* Nonzero if TYPE represents a (scalar) boolean type or type > > > > > + in the middle-end compatible with it. */ > > > > > + > > > > > +#define

RE: [PATCH] MIPS: Fix mode mismatch error between Loongson builtin arguments and insn operands.

> From: Matthew Fortune > > +/* The third argument needs to be in SImode in order to succesfully > > match > > + the operand from the insn definition. */ > > Please refer to operand here not argument as it is the second argument > to the builtin but third operand of the instruction.

Re: [PATCH] Fix bool vs. unsigned:1 vectorization (PR tree-optimization/79284)

On Wed, 1 Feb 2017, Jakub Jelinek wrote: > On Wed, Feb 01, 2017 at 10:58:29AM +0100, Richard Biener wrote: > > > > +/* Nonzero if TYPE represents a (scalar) boolean type or type > > > > + in the middle-end compatible with it. */ > > > > + > > > > +#define INTEGRAL_BOOLEAN_TYPE_P(TYPE) \ > > >

Re: [PATCH] Fix bool vs. unsigned:1 vectorization (PR tree-optimization/79284)

On Wed, Feb 01, 2017 at 10:58:29AM +0100, Richard Biener wrote: > > > +/* Nonzero if TYPE represents a (scalar) boolean type or type > > > + in the middle-end compatible with it. */ > > > + > > > +#define INTEGRAL_BOOLEAN_TYPE_P(TYPE) \ > > > + (TREE_CODE (TYPE) == BOOLEAN_TYPE\ >

Re: [PATCH/VECT/AARCH64] Improve cost model for ThunderX2 CN99xx

On 31/01/17 22:34, Andrew Pinski wrote: > On Sat, Jan 28, 2017 at 12:34 PM, Andrew Pinski wrote: >> Hi, >> On some (most) AARCH64 cores, it is not always profitable to >> vectorize some integer loops. This patch does two things (I can split >> it into different patches if

  1   2   >