[patch] Fix libstdc++/55043 - issue with nesting unordered_map containing unique_ptr into vector

This fixes a regression caused by supporting the C++11 allocator requirements in std::vector, and the fact that the unordered containers don't have noexcept move constructors. Fixed by specializing is_copy_constructible for the unordered containers so vector doesn't try to copy them when their ele

Re: [PATCH] Fix PR55882

On Wed, 16 Jan 2013, Eric Botcazou wrote: > > Not necessarily. The following is a 4.7 variant of the patch > > (on trunk get_object_alignment_1 got one extra output which > > moved the align return value to by-reference). > > OK, I obviously didn't try very hard here... > > > Can you give it te

Re: [PR libmudflap/53359] don't register symbols not emitted

On Jan 7, 2013, Richard Biener wrote: > On Sun, Jan 6, 2013 at 8:47 PM, Alexandre Oliva wrote: >> On Jan 2, 2013, Richard Biener wrote: >> >>> On Sun, Dec 30, 2012 at 1:22 AM, Alexandre Oliva wrote: On Dec 21, 2012, Richard Biener wrote: > On Fri, Dec 21, 2012 at 6:33 AM, Al

Re: [PATCH] Allow new ISL/CLooG versions

On Tue, 15 Jan 2013, Jack Howarth wrote: > On Tue, Jan 15, 2013 at 11:05:51AM +0100, Richard Biener wrote: > > On Tue, 15 Jan 2013, Richard Biener wrote: > > > > > On Mon, 14 Jan 2013, Jack Howarth wrote: > > > > > > > On Mon, Jan 14, 2013 at 08:27:12PM +0100, Dominique Dhumieres wrote: > > > >

Re: [PATCH] Allow new ISL/CLooG versions

On Wed, 16 Jan 2013, Richard Biener wrote: > On Tue, 15 Jan 2013, Jack Howarth wrote: > > > On Tue, Jan 15, 2013 at 11:05:51AM +0100, Richard Biener wrote: > > > On Tue, 15 Jan 2013, Richard Biener wrote: > > > > > > > On Mon, 14 Jan 2013, Jack Howarth wrote: > > > > > > > > > On Mon, Jan 14, 2

Re: [patch] Fix libstdc++/55043 - issue with nesting unordered_map containing unique_ptr into vector

On 16 January 2013 09:25, Jonathan Wakely wrote: > This fixes a regression caused by supporting the C++11 allocator > requirements in std::vector, and the fact that the unordered > containers don't have noexcept move constructors. Fixed by > specializing is_copy_constructible for the unordered con

[PATCH, PR 55264] Do not remove as unreachable any virtual methods before inlining

Hi, PR 55264 is caused by cgraph machinery thinking it knows all calls to a virtual method when that is actually not true. As discussed with Honza, prior to inlining, we should not assume some virtual functions (namely those that are neither DECL_COMDAT nor DECL_EXTERNAL) are not reachable. DECL

Re: [PR libmudflap/53359] don't register symbols not emitted

> No, the bug is that we're registering something because at an earlier > point (when we emitted checks) there were references to it. The > references were all optimized away, we correctly decided (elsewhere) the > object was not referenced, and we removed it from the symbol table, but > mudflap s

fix for PR49888 var-tracking compile-time regression

PR49888 introduced clobber_overlapping_mems to detach VALUEs (and variables bound to them) from MEMs as the MEMs are modified. This turned out to be quite expensive, particularly the computation of canonical addresses passed to alias dependency. This patch introduces caching of the canonical addr

Re: [PATCH, PR 55264] Do not remove as unreachable any virtual methods before inlining

On Wed, Jan 16, 2013 at 11:01 AM, Martin Jambor wrote: > Hi, > > PR 55264 is caused by cgraph machinery thinking it knows all calls to > a virtual method when that is actually not true. As discussed with > Honza, prior to inlining, we should not assume some virtual functions > (namely those that

Re: fix for PR49888 var-tracking compile-time regression

On Wed, Jan 16, 2013 at 08:28:59AM -0200, Alexandre Oliva wrote: > PR49888 introduced clobber_overlapping_mems to detach VALUEs (and > variables bound to them) from MEMs as the MEMs are modified. This > turned out to be quite expensive, particularly the computation of > canonical addresses passed

Re: [PATCH, PR 55264] Do not remove as unreachable any virtual methods before inlining

> On Wed, Jan 16, 2013 at 11:01 AM, Martin Jambor wrote: > > Hi, > > > > PR 55264 is caused by cgraph machinery thinking it knows all calls to > > a virtual method when that is actually not true. As discussed with > > Honza, prior to inlining, we should not assume some virtual functions > > (name

Re: [PATCH, PR 55264] Do not remove as unreachable any virtual methods before inlining

> Perhaps could you first change cgraph_non_local_node_p_1 and try to check > some code > if codegen differs significantly? It should not at all. > ipa-cp is the sole user of this flag in IPA passes, so you should know what > it does. Thinking deeper of ipa-cp and local virtuals, I think this is

Re: [PR libmudflap/53359] don't register symbols not emitted

On Jan 16, 2013, Jan Hubicka wrote: > 2) has the obvious advantage that unused vars are not going to be output just > for sake of checking code. For 2) the symtab_get_node test seems resonable to > me. That's what I first implemented, and I still firmly believe symtab_get_node is the correct te

Re: fix for PR49888 var-tracking compile-time regression

On Jan 16, 2013, Jakub Jelinek wrote: > On Wed, Jan 16, 2013 at 08:28:59AM -0200, Alexandre Oliva wrote: >> PR49888 introduced clobber_overlapping_mems to detach VALUEs (and >> variables bound to them) from MEMs as the MEMs are modified. This >> turned out to be quite expensive, particularly the

Re: PING: gcc.target/arm: skip 5 tests for flag conflicts

Hi Janis, > Back in September I submitted a patch to fix five ARM tests in > . > You responded in < http://gcc.gnu.org/ml/gcc-patches/2012-10/msg00972.html> > and I answered your questions in a reply. I believe that Richard's main point wa

Re: [testsuite] remove ARM big-endian from some vect effective targets

Hi Janis, > 2013-01-15 Janis Johnson > > PR testsuite/54622 > > * lib/target-supports.exp (check_effective_target_vect_perm_byte, > check_effective_target_vect_perm_short, > check_effective_target_vect_widen_mult_qi_to_hi_pattern, > check_effective_target_vect64)

Re: [testsuite] add option to LTO flags for c-torture/execute/builtins tests

Hi Janis, > 2013-01-15 Janis Johnson > > PR testsuite/55994 > * gcc.c-torture/execute/builtins/builtins.exp: Add > -Wl,--allow-multiple-definition for eabi and elf targets. Approved - please apply. Cheers Nick

[PATCH] Change DO loop translation, avoid undefined overflow and repeated step sign tests

The following patch fixes a few things I noticed when looking at PR42108 again. First of all the current setup of /* Calculate the loop count. to-from can overflow, so we cast to unsigned. */ doesn't work as advertised (well, it does, for the result and -fwrapv) - it does not

Re: [testsuite] fix ARM test gcc.target/arm/neon-vld1_dupQ.c

Hi Janis, > 2013-01-14 Janis Johnson > > * gcc.target/arm/neon-vld1_dupQ.c: Use types that match function > prototypes. Approved - please apply. Cheers Nick

Re: [ARM] Turning off 64bits ops in Neon and gfortran/modulo-scheduling problem

Ping^2 ? On 8 January 2013 17:24, Christophe Lyon wrote: > Ping? > http://gcc.gnu.org/ml/gcc-patches/2012-12/msg01197.html > > Thanks, > > Christophe > > On 19 December 2012 16:59, Christophe Lyon wrote: >> On 17 December 2012 16:12, Richard Earnshaw wrote: >>> On 29/11/12 17:16, Christophe Ly

Re: [PR libmudflap/53359] don't register symbols not emitted

On Wed, Jan 16, 2013 at 2:17 PM, Alexandre Oliva wrote: > On Jan 16, 2013, Jan Hubicka wrote: > >> 2) has the obvious advantage that unused vars are not going to be output just >> for sake of checking code. For 2) the symtab_get_node test seems resonable >> to >> me. > > That's what I first imp

Re: [testsuite] replace gcc.target/arm/ftest-*.c with compile-only tests

Hi Janis, > The gcc.target/arm/ftest-*.c tests check various macros that are set for > ARM targets by setting flags based on preprocessor directives that check > those macros. The tests are skipped if the current test platform > doesn't support executing programs for the architecture for which fl

Re: [PATCH][RFC] Fix PR55964

On Tue, 15 Jan 2013, Jakub Jelinek wrote: > On Mon, Jan 14, 2013 at 04:58:09PM +0100, Richard Biener wrote: > > I happen to have a patch for PR55964 around - preparatory work > > to make loop distribution (and vectorization) handle nested loops. > > It mostly kills the broken way loop distribution

[Patch, Fortran] PR 55983: [4.7/4.8 Regression] ICE in find_typebound_proc_uop, at fortran/class.c:2711

Hi all, here is a close-to-obvious patch for an ICE-on-invalid regression. It removes as assert, which is reasonable for valid code but can fail under error conditions (as the PR shows), and replaces it with an equivalent IF clause. Regtested on x86_64-unknown-linux-gnu. Ok for trunk and 4.7? Ch

Re: fix for PR49888 var-tracking compile-time regression

On Wed, Jan 16, 2013 at 11:25:46AM -0200, Alexandre Oliva wrote: > > Can you safely cache the canon addresses already during vt_initialize > > (when cselib_* is still processing new insns, cselib VALUEs contain > > REGs and MEMs that are flushed at the end of processing the current bb > > in vt_ini

Re: [PATCH] Change DO loop translation, avoid undefined overflow and repeated step sign tests

Richard Biener wrote: if (D.1912 < 0) { if (D.1911 > D.1910) { goto L.6; }

[PATCH] Remove dead code

Bootstrapped and tested on x86_64-unknown-linux-gnu, applied. Richard. 2013-01-16 Richard Biener * tree-inline.c (tree_function_versioning): Remove set but never used variable. Index: gcc/tree-inline.c === --- g

[PATCH] Reorder step != +-1 do code to allow empty latch (PR fortran/52865)

Hi! As discussed in the PR, this patch performs the decrement of countm1 before the condition, so that the loop can have empty latch block. The testcase from the PR then can be vectorized. Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk? 2013-01-16 Jakub Jelinek P

[patch] libmudflap: check for NULL path in dlopen wrapper

>From the bug report: If mudflap is used to instrument a program using dlopen, and the program (assuming it is compiled with -rdynamic) loads itself by passing NULL for the path to dlopen, the program will crash unconditionally; that is, regardless of the options passed to mudflap, so long as inst

Re: [patch] libmudflap: check for NULL path in dlopen wrapper

On Wed, Jan 16, 2013 at 04:32:25PM +0100, Matthias Klose wrote: > PR mudflap/24619 > * mf-hooks2.c (dlopen wrapper): Check for NULL path. > > Index: b/src/libmudflap/mf-hooks2.c > === > --- a/libmudflap/mf-hooks2.c > +++ b

[PATCH] gcc/go/gospec.c: fix static linking of Go programs (issue 7130047)

LGTM https://codereview.appspot.com/7130047/diff/1/gcc/go/gospec.c File gcc/go/gospec.c (right): https://codereview.appspot.com/7130047/diff/1/gcc/go/gospec.c#newcode230 gcc/go/gospec.c:230: num_args = argc + need_math + shared_libgcc + (library > 0) * 5 + 5; I wonder if we should change + 5 to

[PATCH] gcc/go/gospec.c: fix static linking of Go programs (issue 7130047)

LGTM https://codereview.appspot.com/7130047/diff/1/gcc/go/gospec.c File gcc/go/gospec.c (right): https://codereview.appspot.com/7130047/diff/1/gcc/go/gospec.c#newcode230 gcc/go/gospec.c:230: num_args = argc + need_math + shared_libgcc + (library > 0) * 5 + 5; I wonder if we should change + 5 to

Re: [PATCH] Reorder step != +-1 do code to allow empty latch (PR fortran/52865)

Jakub Jelinek wrote: As discussed in the PR, this patch performs the decrement of countm1 before the condition, so that the loop can have empty latch block. The testcase from the PR then can be vectorized. Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk? OK from the FE side

Re: [PATCH] Reorder step != +-1 do code to allow empty latch (PR fortran/52865)

On Wed, 16 Jan 2013, Tobias Burnus wrote: > Jakub Jelinek wrote: > > As discussed in the PR, this patch performs the decrement of countm1 > > before the condition, so that the loop can have empty latch block. > > The testcase from the PR then can be vectorized. > > > > Bootstrapped/regtested on x

[PATCH] Fix PR3713

This fixes PR3713 by properly propagating ->has_constants in SCCVN. With that we are able to simplify (unsigned) Bar & 1 properly. Only copyprop later turns the call into a direct one though, so I'm testing the important fact - that Bar is inlined and eliminated by IPA inlining. Bootstrapped on x

Re: [PATCH] gcc/go/gospec.c: fix static linking of Go programs (issue 7130047)

Reviewers: iant, Message: please take another look. Description: 2013-01-16 Shenghou Ma * gospec.c: pass -u pthread_create to linker when static linking. Please review this at https://codereview.appspot.com/7130047/ Affected files: M gcc/go/gospec.c Index: gcc/go/gospec.c ==

Fix loop-iv.c ICE

Hi, this patch fixes ICE seen in PR51083 on PPC. Here the flags ^= 0x8000 expression is translated as PLUS. This makes us to consider flags to be IV and work out that the loop do not really iterate. It is a missed optimization that we do not work out this on all targets and do not unloop

Re: [PATCH] Fix shrink-wrapping with vDRAP (PR target/55940)

On Tue, Jan 15, 2013 at 2:38 PM, Jakub Jelinek wrote: > Hi! > > As the following testcase shows, even when stack_realign_drap we might need > to prevent crtl->drap_reg accesses in the bbs considered for > shrink-wrapping, even if reload decides stack realignment isn't needed, the > vDRAP (in the t

Re: [PATCH] Fix shrink-wrapping with vDRAP (PR target/55940)

On Wed, Jan 16, 2013 at 08:28:54AM -0800, H.J. Lu wrote: > > --- gcc/function.c.jj 2013-01-11 09:02:55.0 +0100 > > +++ gcc/function.c 2013-01-15 19:23:20.648826011 +0100 > > @@ -6029,7 +6029,7 @@ thread_prologue_and_epilogue_insns (void > >if (pic_offset_table_rtx) > >

Re: [PATCH] gcc/go/gospec.c: fix static linking of Go programs (issue 7130047)

Approved and applied. Thanks! https://codereview.appspot.com/7130047/

Re: fix for PR49888 var-tracking compile-time regression

On Wed, Jan 16, 2013 at 11:25:46AM -0200, Alexandre Oliva wrote: > > Also, what effects (if any) does the patch have on the > > .debug_info/.debug_loc size and coverage? > > It shouldn't have any, since it's just caching results that would have > been recomputed over and over. However, there's a

[PATCH, i386] Fix PR55981, atomic store is split in two smaller stores

Hello! Using plain movdi pattern is not guaranteed to be atomic for all operands. For x86_64, when storing DImode immediate outside SImode range, the compiler splits the move into two separate SImode moves, violating atomic assumptions. Attached patch generates atomic store for all supported inpu

Re: [PATCH, i386] Fix PR55981, atomic store is split in two smaller stores

On 01/16/2013 09:26 AM, Uros Bizjak wrote: 2012-01-16 Uros Bizjak PR target/55981 * config/i386/sync.md (atomic_store): Always generate SWImode store through atomic_store_1. (atomic_store_1): Macroize insn using SWI mode iterator. testsuite/ChangeLog: 2012-01-1

Re: [testsuite] replace gcc.target/arm/ftest-*.c with compile-only tests

On 01/16/2013 05:53 AM, Nick Clifton wrote: > Hi Janis, > >> The gcc.target/arm/ftest-*.c tests check various macros that are set for >> ARM targets by setting flags based on preprocessor directives that check >> those macros. The tests are skipped if the current test platform >> doesn't support

Re: [google integration, gcc-4_7] Rename __google_stl_debug_string_dangling -> __google_stl_debug_dangling_string

On Wed, Jan 16, 2013 at 12:14 PM, Paul Pluzhnikov wrote: > [Resending to correct patches address.] > > Diego, > > This harmonizes with __google_stl_debug_dangling_vector > > Ok for google/integration and google/gcc-4_7 branches? OK. Diego.

Re: [Patch, Fortran] PR 55983: [4.7/4.8 Regression] ICE in find_typebound_proc_uop, at fortran/class.c:2711

Dear Janus, As you say, this is close to being obvious - OK for trunk and for 4.7. Thanks for the patch. Cheers Paul On 16 January 2013 15:08, Janus Weil wrote: > Hi all, > > here is a close-to-obvious patch for an ICE-on-invalid regression. It > removes as assert, which is reasonable for val

patch to fix PR56005

The following patch fixes PR56005. The details are on http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56005 2013-01-16 Vladimir Makarov PR rtl-optimization/56005 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding pending reads for prefetch. The patch was

Re: [RFC, middlend] Fix for PR54218

On Mon, Jan 14, 2013 at 12:50 PM, Richard Biener wrote: > On Fri, Jan 11, 2013 at 6:37 PM, George Thomas > wrote: >> On Fri, Jan 11, 2013 at 9:53 PM, Andrew Pinski wrote: >>> On Fri, Jan 11, 2013 at 8:17 AM, George Thomas >>> wrote: Hi, I am sending a patch which solves the debug

Re: [PATCH, PR 55264] Do not remove as unreachable any virtual methods before inlining

On Wed, Jan 16, 2013 at 01:44:20PM +0100, Jan Hubicka wrote: > > Perhaps could you first change cgraph_non_local_node_p_1 and try to check > > some code > > if codegen differs significantly? It should not at all. > > ipa-cp is the sole user of this flag in IPA passes, so you should know what > >

Re: [PR55547] fix alias regression on alpha on misaligned symbols (was: Re: do you have time to review this alpha P1 patch?)

On Wed, Jan 16, 2013 at 8:33 AM, Uros Bizjak wrote: >>> On 01/15/2013 08:24 AM, Aldy Hernandez wrote: Ok, it's really an alias.c bug, but it is Alpha, and aoliva has already provided an unreviewed patch... >> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55547 >> >>> The patch in #C4

Re: [PR55547] fix alias regression on alpha on misaligned symbols

On 01/15/2013 08:29 PM, Alexandre Oliva wrote: if (rtx_equal_for_memref_p (x, y)) { - if (xsize <= 0 || ysize <= 0) + if (xsize == 0 || ysize == 0) return 1; - if (c >= 0 && xsize > c) + if (c >= 0 && abs (xsize) - c > 0) return 1; - if (c < 0 &&

Re: [google gcc-4_7] Inlining and devirtualization tests

On 16/01/2013, at 6:41 PM, Xinliang David Li wrote: > Looks fine. Why adding tests that are expected to fail? Are these > tests passing with trunk? They are expected to fail only due to optimization analysis not being up to snuff yet, not because it's a bad idea to optimize a particular testcas

Re: [Patch, Fortran] PR 55983: [4.7/4.8 Regression] ICE in find_typebound_proc_uop, at fortran/class.c:2711

> As you say, this is close to being obvious - OK for trunk and for 4.7. Thanks, guys. Committed to trunk as 195251. Will do 4.7 soon ... Cheers, Janus > On 16 January 2013 15:08, Janus Weil wrote: >> Hi all, >> >> here is a close-to-obvious patch for an ICE-on-invalid regression. It >> remov

[wwwdocs,avr,committed]: Mention avr specific improvments

http://gcc.gnu.org/ml/gcc-cvs-wwwdocs/2013/msg00015.html http://gcc.gnu.org/gcc-4.8/changes.html#avr Index: changes.html === RCS file: /cvs/gcc/wwwdocs/htdocs/gcc-4.8/changes.html,v retrieving revision 1.85 diff -u -p -r1.85 changes.h

Re: [PATCH, PR 55264] Do not remove as unreachable any virtual methods before inlining

> On Wed, Jan 16, 2013 at 01:44:20PM +0100, Jan Hubicka wrote: > > > Perhaps could you first change cgraph_non_local_node_p_1 and try to check > > > some code > > > if codegen differs significantly? It should not at all. > > > ipa-cp is the sole user of this flag in IPA passes, so you should know

[google 4.7] Fix for PR 8013197 (issue7140044)

commit 257910a4dd56ac0c95ad79053a4364a3ff34a674 Author: Sterling Augustine Date: Wed Jan 16 13:22:59 2013 -0800 Fix for PR 8013197. Address table entries will already have been removed at this point in resolve_addr, so no need to call again. M gcc/dwarf2out.c Tested:

Re: PING: gcc.target/arm: skip 5 tests for flag conflicts

On 01/16/2013 05:31 AM, Nick Clifton wrote: > Hi Janis, > >> Back in September I submitted a patch to fix five ARM tests in >> . >> You responded in < http://gcc.gnu.org/ml/gcc-patches/2012-10/msg00972.html> >> and I answered your questions

Re: [patch] Fix libstdc++/55043 - issue with nesting unordered_map containing unique_ptr into vector

Here's another attempt to fix this regression, I hope this time it doesn't cause more problems than it solves. Instead of specializing is_copy_constructible for the unordered containers this causes their copy constructors to be deleted if the value_type is not CopyInsertable into the container. T

[patch] fix libstdc++/56012 - narrowing conversion in std::atomic_flag

This fixes a regression since 4.6 when -Wsystem-headers is used. The initialization of the __atomic_flag_base base class has a narrowing conversion from int (the macro) to either bool or unsigned char. The patch fixes it by calling a constexpr function which implicitly converts the value to the r

Re: [google 4.7] Fix for PR 8013197 (issue7140044)

> 2013-01-16 Sterling Augustine > > * gcc/dwarf2out.c (resolve_addr): Delete call to > remove_addr_table_entry. OK for google/gcc-4_7. The commit log entry should say "Google ref: b/8013197" instead of "PR ...". Thanks! -cary

libbacktrace patch committed: Handle missing line number entry

It turns out that it is possible to construct debug info in which a compilation unit has a low_pc and a high_pc, and a line number table, but the line number table does not cover PC values from low_pc up to some value. For example, this will happen in an assembler file if you use a .cfi_startproc

[patch] libstdc++/52887 - fix AIX bootstrap

Add required instantiations for AIX. PR libstdc++/52887 * src/c++11/regex.cc: Add instantiations for AIX. Committed to the 4.7 branch only. commit df31b423330bab88fee84c8f32376dce7ca9242b Author: Jonathan Wakely Date: Thu Jan 17 01:36:42 2013 + PR libstdc++/52887

Go patch committed: Provide line number for init function

Related to the libbacktrace patch I submitted earlier (http://gcc.gnu.org/ml/gcc-patches/2013-01/msg00889.html), this patch ensures that the initialization function, created by the Go frontend to handle package imports and run init functions, has line number information. Previously it did not. Th

[PATCH] [testsuite] [arm] Test thumb1 far jump

Test cases for previous patch "no lr save for non-far branches in leaf function". * gcc.target/arm/thumb1-far-jump-1.c: New. * gcc.target/arm/thumb1-far-jump-2.c: New. Index: gcc/testsuite/gcc.target/arm/thumb1-far-jump-2.c =

Re: [PATCH] [testsuite] [arm] Test thumb1 far jump

On 01/16/2013 06:05 PM, Joey Ye wrote: > Test cases for previous patch "no lr save for non-far branches in leaf > function". > > * gcc.target/arm/thumb1-far-jump-1.c: New. > * gcc.target/arm/thumb1-far-jump-2.c: New. > > Index: gcc/testsuite/gcc.target/arm/thumb1-far-jump-2.c > ==

Re: [PR55547] fix alias regression on alpha on misaligned symbols

On Jan 16, 2013, Richard Henderson wrote: > I notice that these expressions (including the first hunk that uses > ifs) are now all the same. *nod* > It would seem extremely prudent to pull > this out to a function so that they stay the same. ack, will do. > That said, I question the change of

RE: [PATCH] [testsuite] [arm] Test thumb1 far jump

> -Original Message- > From: Janis Johnson [mailto:janis_john...@mentor.com] > Sent: Thursday, January 17, 2013 10:41 > To: Joey Ye > Cc: gcc-patches@gcc.gnu.org > Subject: Re: [PATCH] [testsuite] [arm] Test thumb1 far jump > > On 01/16/2013 06:05 PM, Joey Ye wrote: > > Test cases for pr

Re: [PR55547] fix alias regression on alpha on misaligned symbols

On Jan 16, 2013, Richard Henderson wrote: > I notice that these expressions (including the first hunk that uses > ifs) are now all the same. It would seem extremely prudent to pull > this out to a function so that they stay the same. Here's a revised patch that makes that change, making the over

Re: [Patch i386]: btver2 pipeline descriptions.

On Thu, Jan 17, 2013 at 8:17 AM, Kumar, Venkataramanan wrote: > Hi Maintainers, > > This patch adds pipeline descriptions for -march=btver2. > > Completed bootstrap and gcc regression test passes with r194705 > > Is this ok for trunk? > > > Change log > -- > 2013-01-17 Venkataramana

Re: [patch] fix libstdc++/56012 - narrowing conversion in std::atomic_flag

2013/1/17 Jonathan Wakely : > This fixes a regression since 4.6 when -Wsystem-headers is used. The > initialization of the __atomic_flag_base base class has a narrowing > conversion from int (the macro) to either bool or unsigned char. The > patch fixes it by calling a constexpr function which im