Fix i386 cmov/cstore/cbranch patterns

Hi, this patch fixes missed cmov conversion I noticed while looking into PR53346. It effectively reverts patch http://gcc.gnu.org/ml/gcc-patches/2010-07/msg02326.html I believe it is mistaken - it chnaged comparison_operator to unordered_comparison_operator on the basis that those patterns are i

Fix overflow in ipa-inline seen on profiledbootstrap

Hi, profiledbootstrap shows bug in edge_badness. With inline hints the growth is no longer limited by small constant and thus multiplying it with overall growth may overlfow. Bootstrapped/regtested & comitted. Honza PR bootstrap/55051 * ipa-inline.c (edge_badness): Improve dump

re: [Bug debug/55328] ICE: in output_addr_table_entry, at dwarf2out.c:21780 with -gsplit-dwarf

Thanks for testing this. Enclosed is a simple fix for this ICE--address_table_entries that have zero refcounts shouldn't be assigned an index. OK for trunk? Sterling 2012-11-14 Sterling Augustine * dwarf2out.c (index_address_table_entry): Check a node's refcount. patch.diff Descri

Re: [Bug debug/55328] ICE: in output_addr_table_entry, at dwarf2out.c:21780 with -gsplit-dwarf

> Enclosed is a simple fix for this ICE--address_table_entries that have > zero refcounts shouldn't be assigned an index. > > OK for trunk? > > Sterling > > 2012-11-14 Sterling Augustine > > * dwarf2out.c (index_address_table_entry): Check a node's refcount. Add PR debug/55328 to the Ch

Fix libstdc++/55320

PR libstdc++/55320 * include/std/functional (function::function(F)): Set _M_manager after operations that could throw. (_Function_base::_Ref_manager::_M_init_functor): Use addressof. * include/tr1/functional (_Function_base::_Ref_manager::_M_init_func

Re: [patch] [m68k] add multiarch definitions for m68k-linux-gnu

Am 14.11.2012 23:39, schrieb Joseph S. Myers: > On Wed, 14 Nov 2012, Matthias Klose wrote: > >> The following patch adds the multiarch definitions for m68k-linux-gnu. Tested >> using a Debian/Ubuntu package build. Ok for the trunk? >> >> Here, I'm unsure if the definition needs to be further const

Reduce complette unrolling & peeling limits

Hi, this patch reduces max-peeled-insns and max-completely-peeled-insns from 400 to 100. The reason why I am doing this is that I want to reduce code bloat caused by my cunroll work that enabled a lot more unrolling then previously causing considerable code size regression at -O3. I do not think

Re: [Bug debug/55328] ICE: in output_addr_table_entry, at dwarf2out.c:21780 with -gsplit-dwarf

On Wed, Nov 14, 2012 at 3:19 PM, Cary Coutant wrote: >> Enclosed is a simple fix for this ICE--address_table_entries that have >> zero refcounts shouldn't be assigned an index. >> >> OK for trunk? >> >> Sterling >> >> 2012-11-14 Sterling Augustine >> >> * dwarf2out.c (index_address_tabl

Re: [patch] [powerpc] add multiarch definitions for powerpc-linux-gnu

Am 14.11.2012 23:41, schrieb Joseph S. Myers: > Your t-linux setting fails to allow for the possibility of the compiler > being configured to default to soft-float (whether --with-float=soft, or > --with-cpu etc. for a CPU that implies soft-float). soft-float (only > supported for 32-bit) is AB

Re: [patch] [powerpc] add multiarch definitions for powerpc-linux-gnu

On Thu, 15 Nov 2012, Matthias Klose wrote: > Am 14.11.2012 23:41, schrieb Joseph S. Myers: > > Your t-linux setting fails to allow for the possibility of the compiler > > being configured to default to soft-float (whether --with-float=soft, or > > --with-cpu etc. for a CPU that implies soft-floa

Re: top-level configure.ac changes lost in sync

I am checking this into GCC tree for Roland. On Wed, Nov 14, 2012 at 4:13 PM, Roland McGrath wrote: > I've just noticed that this change (git mirror view): > > commit 7228bb7b806e7f0513557e3002a6dca104a768a1 > Author: H.J. Lu > Date: Sun Aug 26 14:34:38 2012 + > >

[PATCH, testsuite]: Add dg-error for unsupported floating suffix

Hello! Attached patch fixes following testsuite failure FAIL: g++.dg/cpp0x/gnu_fext-numeric-literals.C (test for excess errors) FAIL: g++.dg/cpp0x/std_fext-numeric-literals.C (test for excess errors) on targets that don't support Q and W floating suffixes. 2012-11-15 Uros Bizjak * g

Re: [patch] [powerpc] add multiarch definitions for powerpc-linux-gnu

Am 15.11.2012 00:55, schrieb Joseph S. Myers: > On Thu, 15 Nov 2012, Matthias Klose wrote: > >> Am 14.11.2012 23:41, schrieb Joseph S. Myers: >>> Your t-linux setting fails to allow for the possibility of the compiler >>> being configured to default to soft-float (whether --with-float=soft, or >

Fix bogus waring on gcov-io.c in profiledbootstrap

Hi, gcov-io.c contains bogus array bounds check that makes us to unroll one time too many and trigger bogus -Warray-bounds warning. I made testcase for PR55079 and fixed the bounds check. Bootstrapped/regtested/comitted x86_64. Honza PR bootstrap/55051 * gcov-io.c (gcov_read_summar

[PATCH] Fix up var-tracking notes emitted in between bbs (PR middle-end/43631)

Hi! Steven has been complaining for some years that var-tracking inserts NOTE_INSN_VAR_LOCATION (and NOTE_INSN_CALL_ARG_LOCATION) notes sometimes in between basic blocks, but with BLOCK_FOR_INSN set (or sometimes extends a bb ending originally with a noreturn or throwing call by a note or more aft

Re: [wwwdocs,Java] comparison of libgcj with Kaffe link is gone

- Original Message - > The regular comparison with Kaffe is gone after the restructuring/ > move to github, and I could not find any replacement URL. > > I applied this patch. If any of you has a new URL, we can resurrect. > > Gerald > > 2012-11-01 Gerald Pfeifer > > * status

Re: [PATCH] asan testsuite (take 3)

On Wed, Nov 14, 2012 at 11:12:05AM -0800, Mike Stump wrote: > On Nov 14, 2012, at 8:22 AM, Jakub Jelinek wrote: > > On Wed, Nov 14, 2012 at 12:11:13PM +0100, Jakub Jelinek wrote: > >> Anyway, once asan_symbolize actually symbolizes the output, we could use > >> something like: > >> /* { dg-output

[PATCH] Restore i?86/x86_64 4.6 and earlier cmov behavior (PR target/54073)

Hi! Before Richard Sandiford's 4.7 optab cleanups emit_conditional_move wasn't testing the predicate on the comparison, and apparently movcc patterns in i386.md can handle some comparisons that the current predicates on the expanders reject. This PR on MonteCarlo kernel shows a hardly predictable

Fix libstdc++/53841

PR libstdc++/53841 * include/std/condition_variable (condition_variable::wait_until): Handle clocks with higher resolution than __clock_t. (condition_variable::__wait_until_impl): Remove unnecessary _Clock parameter. * testsuite/30_threads/condition_v

[patch] rx: create interrupt vector aliases by attributes

This patch allows the user to specify a vector number in an interrupt attribute, allowing for more generic creation of vector tables. Ok? * doc/extend.texi (Function Attributes): Document RX extensions to "interrupt" attribute. * config/rx/rx.c (has_interrupt_vector): New

C++ PATCH for c++/55275 (abi_tag and explicit specialization)

This bug arose because I was using DECL_SOURCE_LOCATION to determine whether this is the first declaration of the class. Which may be fragile, but I don't see another way to do it. But for an explicit specialization we were treating the template declaration as the initial declaration, so the

Re: [patch] Performance patch for MIPS conditional move in expr.c

On 2012-11-14 11:02, Steve Ellcey wrote: > 2012-11-14 Steve Ellcey > > * expr.c (expand_cond_expr_using_cmove): Use promoted mode for temp. Ok. r~

Re: RFC: PATCH to add abi_tag attribute

On 11/11/2012 11:58 PM, Jason Merrill wrote: On 11/11/2012 08:01 AM, Florian Weimer wrote: I'm not sure if this sufficiently far-reaching. It seems that this doesn't allow me to implement a virtual function which takes a std::string parameter in new-ABI-mode when the base class is also used in

Re: [PATCH] Restore i?86/x86_64 4.6 and earlier cmov behavior (PR target/54073)

> Hi! > > Before Richard Sandiford's 4.7 optab cleanups emit_conditional_move > wasn't testing the predicate on the comparison, and apparently > movcc patterns in i386.md can handle some comparisons that the > current predicates on the expanders reject. This PR on MonteCarlo > kernel shows a hard

Re: C++ PATCH for c++/55275 (abi_tag and explicit specialization)

On 15 November 2012 01:51, Jason Merrill wrote: > This bug arose because I was using DECL_SOURCE_LOCATION to determine whether > this is the first declaration of the class. Which may be fragile, but I > don't see another way to do it. But for an explicit specialization we were > treating the temp

[PATCH] [DOC] [MIPS] add microMIPS option into doc

Hi all, Little change, add microMIPS option. Please checkin if it is OK. Regards, Jia 2012-11-15 Jia Liu * gcc/doc/invoke.texi: Add microMIPS option. Signed-off-by: Jia Liu --- gcc/doc/invoke.texi |7 +++ 1 file changed, 7 insertions(+) diff --git a/gcc/doc/invoke.texi

PATCH: PR other/55333: libsanitizer StackTrace::FastUnwindStack wrong x32

Hi, X32 uses 32-bit pointer in software. But its hardware pointer is 64-bit. We must use hardware pointer to unwind frames. This patch adds uhwptr for hardware pointer and uses it to unwind stack frames. Tested on Linux/x32, Linux/x86-64 and Linux/ia32. Please review it for upstream. Thanks.

Re: PATCH: PR other/55333: libsanitizer StackTrace::FastUnwindStack wrong x32

On Wed, Nov 14, 2012 at 7:51 PM, H.J. Lu wrote: > Hi, > > X32 uses 32-bit pointer in software. But its hardware pointer is > 64-bit. We must use hardware pointer to unwind frames. This patch > adds uhwptr for hardware pointer and uses it to unwind stack frames. > Tested on Linux/x32, Linux/x86-

[PATCH v4] Add support for sparc fused compare-and-branch.

This integrates the Solaris2 work from Eric Botcazou, and has some changes based upon feedback from Richard Henderson. The bootstrap comparison failure no longer happens, and this is fully regstrapped on sparc-linux-gnu w/--with-cpu=niagara4, and I also did a quick bootstrap check using --with-cp

<    1   2