[SH] PR 51244 - Add nott insn

Hello, The attached patch adds the 'nott' insn that is supported by SH2A. While the 'nott' pattern itself is usually combined away entirely (CSiBE shows no diffs for SH2A), on non-SH2A the new patterns seem to catch a few border cases, where unnecessary movt sequences can be eliminated. Tested ag

[patch] PR53881

Hello, I was relying on the assumption that all case labels in a switch vector that jump to the same target block would also have the same CASE_LABEL. This isn't so for the test case of PR53881, where DCE removes a dead statement but preserves a user label. Fortunately, using find_edge in emit_cas

[RFT] Remove -fno-tree-dominator-opts from libgcc/config/t-darwin

Hello, PR26840 was fixed long ago so this work-around shouldn't be necessary anymore: Index: ../libgcc/config/t-darwin === --- ../libgcc/config/t-darwin (revision 189348) +++ ../libgcc/config/t-darwin (working copy) @@ -1,7 +1,5

Re: [patch][RFC] Remove tree/gimple/rtl dependencies from core CFG code

On Fri, Jul 6, 2012 at 8:29 AM, Steven Bosscher wrote: > On Fri, Jul 6, 2012 at 1:06 AM, Joseph S. Myers > wrote: >> My only comment is that you add a coretypes.h include to tree-flow.h and >> remove one from tree-ssa-alias.h - but since the GCC convention is that >> all .c files start with incl

Re: [RFT] Remove -fno-tree-dominator-opts from libgcc/config/t-darwin

On Sat, Jul 7, 2012 at 7:48 AM, Steven Bosscher wrote: > Hello, > > > PR26840 was fixed long ago so this work-around shouldn't be necessary anymore: > > Index: ../libgcc/config/t-darwin > === > --- ../libgcc/config/t-darwin (revisio

Re: [RFT] Remove -fno-tree-dominator-opts from libgcc/config/t-darwin

On Sat, Jul 7, 2012 at 5:06 PM, H.J. Lu wrote: > Are you sure this the right patch? -fno-tree-dominator-opts is still here. I am sure it is not the right patch :-) Thanks! Index: libgcc/config/t-darwin === --- libgcc/config/t-darwin

Re: [patch][RFC] Remove tree/gimple/rtl dependencies from core CFG code

On Sat, 7 Jul 2012, Steven Bosscher wrote: > BTW, why do so many libgcc files include coretypes.h? It might be needed for the tm.h includes, if tm.h uses types from coretypes.h - there are still a fair number of host-side tm.h macros used in libgcc, that need to transition to libgcc-only target

[patch] Define GENERATOR_FILE explicitly for the host gengtype objects

Hello, Since r177358, a host executable gengtype is built so that plugins can use the GTY markers. Even for the host, gengtype is a GENERATOR_FILE, but since the build-% rule doesn't apply to host objects, gengtype was being built without it. So when I removed coretypes.h from double-int.h, my bu

[patch] Never include coretypes.h in another header file

Hello, As suggested by Joseph. Subject says all. This turned out to be easier than I expected: Almost all Makefile rules already include coretypes.h, I only had to adjust something for realmpfr.o. Bootstrapped on x86_64-unknown-linux-gnu. OK for trunk? Ciao! Steven no_coretypes_in_headers.diff

Re: [v3] PR 53270 fix hppa-linux bootstrap regression

On 14 June 2012 23:23, Jonathan Wakely wrote: > We've known for ages that it's not portable to do: > > __gthread_mutex_t tmp = __GTHREAD_MUTEX_INIT; > _M_mutex = __tmp; > > As PR 53270 shows, the copy assignment now actually fails in C++11 > mode on platforms using LinuxThreads, because the mutex h

Re: User directed Function Multiversioning via Function Overloading (issue5752064)

On Fri, Jul 6, 2012 at 11:05 PM, Jason Merrill wrote: > On 06/14/2012 04:13 PM, Sriraman Tallam wrote: >> >> C++ Frontend maintainers, Could you please take a look at the >> front-end part when you find the time? > > > It seems to me that what you have here are target-specific attributes that

[patch][fortran] Include coretypes.h in .c files, not in gfortran.h

Hello, Currently, gcc/flags.h and fortran/gfortran.h both include coretypes.h. In many fortran/*.c files, flags.h is included before gfortran.h, so when flags.h no longer includes coretypes.h (xf. http://gcc.gnu.org/ml/gcc-patches/2012-07/msg00275.html) gfortran no longer builds without some chang

Fix gcc.dg/lower-subreg-1.c failure, revisited.

> From: Richard Sandiford > Date: Wed, 9 May 2012 11:14:49 +0200 > Hans-Peter Nilsson writes: > >> From: Richard Sandiford > >> Date: Tue, 1 May 2012 16:46:38 +0200 > > > >> To repeat: as things stand, very few targets define proper rtx costs > >> for SET. > > > > IMHO it's wrong to start blami

Re: [patch][fortran] Include coretypes.h in .c files, not in gfortran.h

On Sat, Jul 07, 2012 at 11:28:25PM +0200, Steven Bosscher wrote: > > Currently, gcc/flags.h and fortran/gfortran.h both include > coretypes.h. In many fortran/*.c files, flags.h is included before > gfortran.h, so when flags.h no longer includes coretypes.h (xf. > http://gcc.gnu.org/ml/gcc-patches

Ping again: [RFA:] Caveat for ARM in gcc-4.7/changes.html: unaligned accesses, take 2

> From: Hans-Peter Nilsson > Date: Fri, 29 Jun 2012 15:06:44 +0200 > > From: Hans-Peter Nilsson > > Date: Fri, 22 Jun 2012 04:24:01 +0200 > > > > From: Hans-Peter Nilsson > > > Date: Fri, 15 Jun 2012 04:07:23 +0200 > > > > A ping. > > And another ping, now CCing ARM maintainers, >

ping again - Show hash table stats when -fmem-report

Hi, This patch adds many nice stats about hash tables when gcc is run with -fmem-report. Attached patch tested on x86, no regressions. Also attached is sample output of -fmem-report when compiling reload.c with -O2 -g. Especially interesting is the extreme htab usage in var-tracking (ofcours

PR #53525 - track-macro-expansion performance regression

With the attached patches I introduce four new obstacks in struct cpp_reader to substitute malloc's/realloc's when expanding macros. Numbers have been posted in the PR, but to summarize: before: 0.785 s or 2201 M instr after: 0.760 s or 2108 M instr Memory overhead is some tens kilobytes wo