Fix PR ada/50842

We now need to link the gnattools with libiconv on Darwin 9, since we link with libcpp.a and other libraries. The patch also gets rid of EXTRA_GNATTOOLS_OBJS which doesn't make much sense any more. Tested by Dave and Dominique, applied on the mainline. 2011-10-28 Eric Botcazou PR

Re: [PATCH] Account for devirtualization opportunities in inliner

Jan, Attached is the updated patch. The only major change is the addition of indirect_call_cost to size and time weights. I've set the size cost of indirect call to 3, which is what I remember calculating when I looked into costs couple of months ago: one call instruction for the call itself,

Re: [PATCH i386] PR47698 no CMOV for volatile mem

On Thu, 27 Oct 2011, Uros Bizjak wrote: > Hello! > > > Here's a patch for PR47698, which is about CMOV should not be > > generated for memory address marked as volatile. > > Successfully bootstrapped and passed make check on x86_64-unknown-linux-gnu. > > > PR rtl-optimization/47698 >

Re: [PATCH][PING] Vectorize conversions directly

Here is the patch updated according to recent comments. 2011-10-28 Dmitry Plotnikov gcc/ * tree-cfg.c (verify_gimple_assign_unary): Allow vector conversions. * optabs.c (supportable_convert_operation): New function. * optabs.h (supportable_convert_operation): New prototype. * t

Re: [PATCH] Don't ICE on long long shifts in vectorizable_shift

On Thu, 27 Oct 2011, Jakub Jelinek wrote: > Hi! > > With the patch I'm going to post momentarily which adds vlshrv{4,2}di and > vashlv{4,2}di patterns for -mavx2 vectorizable_shift ICEs, because the > frontends for long_long_var1 << long_long_var2 emit long_long_var1 << (int) > long_long_var2 >

Re: [RFC PATCH] update to libtool-2.4.2 and regenerate

Markus Trippelsdorf writes: > By popular demand, I've prepared a patch that updates the in-tree > libtool to version 2.4.2. It is needed for lto-bootstrap with > -fno-fat-lto-objects and FreeBSD10.x versions. I see that your patch doesn't deal with libgo/config, where a private copy of libtool

Re: [PATCH] Miscompilation of __attribute__((constructor)) functions.

On Thu, Oct 27, 2011 at 7:24 PM, Paul Brook wrote: > Patch below fixes a miscompilation observed whem building uclibc libpthread on > a mips-linux system. > > The story start with the ipa-split optimization, which turns: > > void fn() > { >  if (cond) { >    DO_STUFF; >  } > } > > into: > > static

Re: [PATCH] Pass through jump functions for addressable (scalar) parameters

On Thu, Oct 27, 2011 at 9:11 PM, Martin Jambor wrote: > Hi, > > On Thu, Oct 27, 2011 at 03:07:10PM +0200, Richard Guenther wrote: >> On Wed, Oct 26, 2011 at 8:25 PM, Martin Jambor wrote: >> > Hi, >> > >> > >> > >> > 2011-10-26  Martin Jambor   >> > >> >        * ipa-prop.c (mark_modified): Moved

Re: [PATCH] Don't ICE on long long shifts in vectorizable_shift

On Fri, Oct 28, 2011 at 10:22:15AM +0200, Richard Guenther wrote: > Hm, but you are testing vector modes in the path that is supposed to > handle shifts by a scalar. That looks odd. Also it should be easy No, I'm testing it in the path that is supposed to handle shifts by a vector. That block s

[PATCH] Cleanup AVX2 vector/vector shifts (take 2)

On Thu, Oct 27, 2011 at 10:07:13PM +0200, Uros Bizjak wrote: > Please use expressive RTX forms for expanders, similar to the above > define_insn RTX. You can avoid calling gen_avx2_lshrv at the end > of c code. Also, expanders can have nonimmediate_operand as operand 2 > and conditionally move it t

Re: [PATCH] Add capability to run several iterations of early optimizations

On Thu, Oct 27, 2011 at 11:53 PM, Matt wrote: Then you'd have to analyze the compile-time impact of the IPA splitting on its own when not iterating. ?Then you should look at what actually was the optimizations that were performed that lead to the improvement (I can see some ind

[trans-mem] Explicitly go irrevocable even if transaction will always go irrevocable.

The ABI does not require a TM runtime library to immediately start a transaction in irrevocable mode if the beginTransaction flag "doesGoIrrevocable" is set. This patch fixes this by inserting a call to changeTransactionMode(0) into the entry block of all transactions that always go irrevocable eve

Re: [PATCH] Don't ICE on long long shifts in vectorizable_shift

On Fri, 28 Oct 2011, Jakub Jelinek wrote: > On Fri, Oct 28, 2011 at 10:22:15AM +0200, Richard Guenther wrote: > > Hm, but you are testing vector modes in the path that is supposed to > > handle shifts by a scalar. That looks odd. Also it should be easy > > No, I'm testing it in the path that is

Re: [PATCH, devirtualization] Detect the new type in type change detection

On Thu, Oct 27, 2011 at 9:54 PM, Martin Jambor wrote: > Hi, > > On Thu, Oct 27, 2011 at 11:06:02AM +0200, Richard Guenther wrote: >> On Thu, Oct 27, 2011 at 1:22 AM, Martin Jambor wrote: >> > Hi, >> > >> > I've been asked by Maxim Kuvyrkov to revive the following patch which >> > has not made it

Re: [PATCH] Cleanup AVX2 vector/vector shifts (take 2)

On Fri, Oct 28, 2011 at 10:57 AM, Jakub Jelinek wrote: > On Thu, Oct 27, 2011 at 10:07:13PM +0200, Uros Bizjak wrote: >> Please use expressive RTX forms for expanders, similar to the above >> define_insn RTX. You can avoid calling gen_avx2_lshrv at the end >> of c code. Also, expanders can have no

Re: [RFC PATCH] Gather vectorization (PR tree-optimization/50789)

Hi Jacob, this looks really cool. I have a liitle question, since I do not understand vectorizer as good. Say, we have a snippet: int *p; int idx[N]; int arr[M]; for (...) { p[i%4] += arr[idx[I]]; } As far as I understand, we cannot do gather we, since p may point to somewere in arr, and, idx ma

Re: [Patch]: PR49868: Named address space support for AVR

Georg-Johann Lay schrieb: > This patch adds named address space support to read data from flash (aka. > progmem) to target AVR. > > The patch has two parts: > > The first part is a repost of Ulrich's work from >http://gcc.gnu.org/ml/gcc/2011-08/msg00131.html > with the needed changes to ./gc

Re: [Patch Darwin/PPC] implement out-of-line FPR/GPR saves/restores.

On 14 Oct 2011, at 10:29, Mike Stump wrote: On Oct 14, 2011, at 2:05 AM, Iain Sandoe wrote: This implements their use and also the GPRs - the latter makes an appreciable reduction in code size, OK for trunk? Ok. Watch for problems with async stack walking (hitting sample in Activity M

Re: [RFC PATCH] Gather vectorization (PR tree-optimization/50789)

On Fri, Oct 28, 2011 at 02:01:36PM +0400, Kirill Yukhin wrote: > this looks really cool. I have a liitle question, since I do not > understand vectorizer as good. > > Say, we have a snippet: > int *p; > int idx[N]; > int arr[M]; > for (...) > { > p[i%4] += arr[idx[I]]; > } > As far as I understa

Re: [PATCH] Add capability to run several iterations of early optimizations

On Fri, Oct 28, 2011 at 1:05 AM, Maxim Kuvyrkov wrote: > Richard, > > Just as Matt posted his findings about the effect of iterating early > optimizations, I've got the new patch ready.  This patch is essentially a > complete rewrite and addresses the comments you made. > > On 18/10/2011, at 9:5

Re: [PATCH] Fix computed gotos on m68k

On Tue, 25 Oct 2011 14:49:09 +0200 Eric Botcazou wrote: > These labels are on the nonlocal_goto_handler_labels chain. You > presumably just need to apply the same treatment to them in > set_initial_label_offsets as the one applied to forced labels. > OK for the adjusted patch if it works, mainl

Re: [RFC PATCH] Gather vectorization (PR tree-optimization/50789)

On Fri, 28 Oct 2011, Jakub Jelinek wrote: > On Fri, Oct 28, 2011 at 02:01:36PM +0400, Kirill Yukhin wrote: > > this looks really cool. I have a liitle question, since I do not > > understand vectorizer as good. > > > > Say, we have a snippet: > > int *p; > > int idx[N]; > > int arr[M]; > > for (.

Re: [Patch]: PR49868: Named address space support for AVR

> The patch requires Ulrich's works which is still in review Now the correct link: http://gcc.gnu.org/ml/gcc-patches/2011-10/msg01874.html

Re: [cxx-mem-model][PATCH 0/9] Convert i386 to new atomic optabs.

On Thu, Oct 27, 2011 at 09:07:29PM -0700, Richard Henderson wrote: > Jakub, in the seventh patch, is there any good reason why OMP is > making the decision of whether or not to generate a compare-and-swap > loop? Why shouldn't we simply always generate the __sync_fetch_op > builtin and let optabs.

Re: [Patch]: PR49868: Named address space support for AVR

2011/10/28 Georg-Johann Lay : > Georg-Johann Lay schrieb: > >> This patch adds named address space support to read data from flash (aka. >> progmem) to target AVR. >> >> The patch has two parts: >> >> The first part is a repost of Ulrich's work from >>    http://gcc.gnu.org/ml/gcc/2011-08/msg00131.

Re: [PATCH] Add gcc-ar/nm/ranlib wrappers for slim LTO v2

On Fri, Oct 21, 2011 at 1:55 AM, Andi Kleen wrote: > From: Andi Kleen > > Slim LTO requires running ar/nm/ranlib with the LTO plugin. The most > convenient way to get this into existing Makefiles is using small > wrappers that pass the plugin. This matches how other compilers > (LLVM, icc) do thi

Re: Another ggc anti fragmentation patchkit

On Sat, Oct 22, 2011 at 7:54 AM, Andi Kleen wrote: > This version addresses all earlier review comments. Passes bootstrap > and testing on x86-64. Ok? Ok. Thanks, Richard. > -Andi > >

Re: [RFC PATCH] Gather vectorization (PR tree-optimization/50789)

On Fri, Oct 28, 2011 at 12:59:48PM +0200, Richard Guenther wrote: > It is also because of re-use of memory via memcpy (yes, some dubious > TBAA case from C, but essentially we don't want to break that). Thus > we can't use TBAA on anonymous memory. No, IMHO we always use a ref_all mem access in t

Re: [RFC PATCH] Gather vectorization (PR tree-optimization/50789)

On Fri, 28 Oct 2011, Jakub Jelinek wrote: > On Fri, Oct 28, 2011 at 12:59:48PM +0200, Richard Guenther wrote: > > It is also because of re-use of memory via memcpy (yes, some dubious > > TBAA case from C, but essentially we don't want to break that). Thus > > we can't use TBAA on anonymous memory

Re: [RFC PATCH] Gather vectorization (PR tree-optimization/50789)

On Fri, 28 Oct 2011, Richard Guenther wrote: > On Fri, 28 Oct 2011, Jakub Jelinek wrote: > > > On Fri, Oct 28, 2011 at 12:59:48PM +0200, Richard Guenther wrote: > > > It is also because of re-use of memory via memcpy (yes, some dubious > > > TBAA case from C, but essentially we don't want to brea

[PATCH][2/n] LTO option handling/merging rewrite

This moves the existing processing of user options from lto1 to the lto driver (lto-wrapper). It also changes the way we stream user options from some custom binary format over to simply streaming the original command-line as passed to the compiler by the driver as a COLLECT_GCC_OPTIONS-like stri

[v3] Add some basic tests for associative/unordered::count()

Hi, checked x86_64-linux, committed to mainline. Thanks, Paolo. // 2011-10-28 Paolo Carlini * testsuite/23_containers/unordered_map/operations/count.cc: New. * testsuite/23_containers/multimap/operations/count.cc: Likewise. * testsuite/23_con

Re: [Patch Darwin/Ada] work around PR target/50678

On 18 Oct 2011, at 13:31, Arnaud Charlet wrote: It's broken in all Libc versions that are in the wild (AFAICT from looking at the released sources). We will need to deal with configury/__ENVIRONMENT_MAC_OS_X_VERSION_MIN_REQUIRED__ stuff once there is a fixed Libc. OK, would be good to foll

Re: Use of vector instructions in memmov/memset expanding

Hi Jan! Thanks for the review, you could find my answers to some of your remarks below. I'll send a corrected patch soon with answers to the rest of your remarks. > - {{rep_prefix_1_byte, {{-1, rep_prefix_1_byte}}}, > + {{{rep_prefix_1_byte, {{-1, rep_prefix_1_byte}}}, >{rep_prefix_1_byte, {

Re: [PATCH i386] PR47698 no CMOV for volatile mem

On Fri, Oct 28, 2011 at 12:16 PM, Richard Guenther wrote: > On Thu, 27 Oct 2011, Uros Bizjak wrote: > >> Hello! >> >> > Here's a patch for PR47698, which is about CMOV should not be >> > generated for memory address marked as volatile. >> > Successfully bootstrapped and passed make check on >> >

Re: [PATCH i386] PR47698 no CMOV for volatile mem

On Fri, 28 Oct 2011, Sergey Ostanevich wrote: > On Fri, Oct 28, 2011 at 12:16 PM, Richard Guenther wrote: > > On Thu, 27 Oct 2011, Uros Bizjak wrote: > > > >> Hello! > >> > >> > Here's a patch for PR47698, which is about CMOV should not be > >> > generated for memory address marked as volatile. >

Re: [trans-mem] Explicitly go irrevocable even if transaction will always go irrevocable.

diff --git a/gcc/testsuite/gcc.dg/tm/memopt-1.c b/gcc/testsuite/gcc.dg/tm/memopt-1.c index 06d4f64..9a48dcb 100644 --- a/gcc/testsuite/gcc.dg/tm/memopt-1.c +++ b/gcc/testsuite/gcc.dg/tm/memopt-1.c @@ -2,8 +2,8 @@ /* { dg-options "-fgnu-tm -O -fdump-tree-tmmemopt" } */ long g, xxx, yyy; -exte

[PATCH] Fix early inliner inlining uninlinable functions

We fail to keep the cannot-inline flag up-to-date when turning indirect to direct calls. The following patch arranges to do this during statement folding (which should always be called when that happens). It also makes sure to copy the updated flag to the edge when iterating early inlining. Boo

Re: [RFC PATCH] update to libtool-2.4.2 and regenerate

Rainer Orth writes: > Markus Trippelsdorf writes: > >> By popular demand, I've prepared a patch that updates the in-tree >> libtool to version 2.4.2. It is needed for lto-bootstrap with >> -fno-fat-lto-objects and FreeBSD10.x versions. > > I see that your patch doesn't deal with libgo/config, w

[google] ThreadSanitizer instrumentation pass (issue5303083)

The patch is for google/main branch. ThreadSanitizer is a data race detector for C/C++ programs. http://code.google.com/p/data-race-test/wiki/ThreadSanitizer The tool consists of two parts: instrumentation module (this file) and a run-time library. The instrumentation module mainintains shadow cal

[Patch,AVR]: Tweak 8-bit parity expansion

This is minor tweak to support 8-bit parity. Otherwise, the input operand of 8-bit values will be extended before parity computation. The final representation as libgcc call is not generated in split1 and no more in expand. Notice that - combine is not allowed to propagate hard regs into zero-ex

Re: [Patch, fortran] [00/66] PR fortran/43829 Inline sum and product (AKA scalarization of reductions)

Mikael, The complete patch bootstraps current FSF gcc trunk on x86_64-apple-darwin11 and the resulting gfortran compiler can compile the Polyhedron 2005 benchmarks using... Compile Command : gfortran-fsf-4.7 -O3 -ffast-math -funroll-loops -flto -fwhole-program %n.f90 -o %n without runtime r

Re: [PATCH, rs6000] Update Power7 scheduling

On Thu, Oct 27, 2011 at 6:14 PM, Pat Haugen wrote: > The following patch fixes some issues with the Power7 scheduling > description. The patch is neutral on cpu2006 (was actually hoping to see > some improvements, but it's still the right thing to do since it more > accurately describes the hardwa

Re: [Patch,AVR]: Tweak 8-bit parity expansion

2011/10/28 Georg-Johann Lay : > This is minor tweak to support 8-bit parity. > > Otherwise, the input operand of 8-bit values will be extended before parity > computation. > > The final representation as libgcc call is not generated in split1 and no more > in expand. Notice that > > - combine is no

Re: [trans-mem] Explicitly go irrevocable even if transaction will always go irrevocable.

On 10/28/2011 08:53 AM, Aldy Hernandez wrote: If the original test was not wrong, you need to add a new test (and bonus points for finding out why this test is currently failing :)). long g, xxx, yyy; /* { dg-final { scan-tree-dump-times "transforming: .*_ITM_RaWU8 \\(&g\\);" 1 "tmmemopt" } }

ping: [RFA:] testsuite infrastructure for options implied by dg-final methods

Ping. Subject changed from '[RFA:] fix breakage with "Update testsuite to run with slim LTO"' except it doesn't fix *all* breakage introduced by that patch, only the one I observed and intended to fix. > Date: Fri, 21 Oct 2011 04:29:20 +0200 > From: Hans-Peter Nilsson > > Date: Fri, 21 Oct 2011

[v3] Trivial formatting changes to a recently added testcase

HI, committed to mainline. Thanks, Paolo. /// 2011-10-28 Paolo Carlini * testsuite/30_threads/condition_variable_any/50862.cc: Trivial formatting fixes. Index: testsuite/30_threads/condition_variable_any/50862.cc

Re: [PATCH][RFC] Re-write LTO option merging

On 11-10-27 01:46 , Richard Guenther wrote: On Wed, 26 Oct 2011, Richard Guenther wrote: This completely rewrites LTO option merging. At compile (uselessly now at WPA?) time we now stream a COLLECT_GCC_OPTIONS like string as it comes from argv of the compiler binary. Those options are read i

Re: [PATCH][2/n] LTO option handling/merging rewrite

Isn't this the same patch as http://gcc.gnu.org/ml/gcc-patches/2011-10/msg02348.html? Diego.

[trans-mem] fix C++ transaction_wrap attribute

The C++ front-end gives us a DECL for transaction_wrap attribute's argument. The C front-end OTOH gives us an IDENTIFIER_NODE. I have no idea why this change after the merge, but I have fixed the attribute handler to work with both front-ends. Also, for this case, distilled from testsuite/c-

Re: [cxx-mem-model][PATCH 0/9] Convert i386 to new atomic optabs.

On 10/28/2011 04:06 AM, Jakub Jelinek wrote: > It just wants a guarantee that the builtin will actually be implemented > in hw. I guess if __sync_fetch_op (new/old) isn't supported but > __sync_compare_and_swap_* is, we could just use the former and let > optabs.c deal with that. But we have to h

Re: [PATCH][PING] Vectorize conversions directly

On 10/28/2011 01:22 AM, Dmitry Plotnikov wrote: > gcc/ > * tree-cfg.c (verify_gimple_assign_unary): Allow vector conversions. > * optabs.c (supportable_convert_operation): New function. > * optabs.h (supportable_convert_operation): New prototype. > * tree-vect-stmts.c (vectorizable_

[trans-mem] Fix outer transactions to be considered abortable too.

Atomic transactions marked as outer(-atomic) transactions can abort too if they are calling functions whose type has the may_cancel_outer attribute. Given that outer transactions are probably rare, this patch just assumes that all outer transactions might abort irrespective of whether they are actu

Re: [PATCH, rs6000] Preserve link stack for 476 cpus

On 10/27/2011 06:43 PM, Peter Bergner wrote: > Ok, here's a patch to implement that, and it passes bootstrap and > regtesting. Richard, is this what you had in mind? I'll note that > I disabled rs6000_code_end for TARGET_POWERPC64, since I was running > into linker errors when building libgcc. T

Re: [trans-mem] fix C++ transaction_wrap attribute

On 10/28/2011 07:50 AM, Aldy Hernandez wrote: > * c-family/c-common.c (handle_tm_wrap_attribute): Handle decl > argument. > * testsuite/c-c++-common/tm/wrap-2.c: Move... > * testsuite/gcc.dg/tm/wrap-2.c: ...here. > * testsuite/g++.dg/tm/wrap-2.C: New. Ok. r~

Re: [trans-mem] Fix outer transactions to be considered abortable too.

On 10/28/2011 08:08 AM, Torvald Riegel wrote: > Fix outer transactions to be considered abortable too. > > * trans-mem.c (lower_transaction): Also add an "over" laber for outer > transactions. > (expand_transactions): Do not set hasNoAbort for outer transactions. >

Re: [PATCH i386] PR47698 no CMOV for volatile mem

On Fri, Oct 28, 2011 at 4:52 PM, Richard Guenther wrote: > On Fri, 28 Oct 2011, Sergey Ostanevich wrote: > >> On Fri, Oct 28, 2011 at 12:16 PM, Richard Guenther wrote: >> > On Thu, 27 Oct 2011, Uros Bizjak wrote: >> > >> >> Hello! >> >> >> >> > Here's a patch for PR47698, which is about CMOV shou

Re: [RFC PATCH] update to libtool-2.4.2 and regenerate

On Fri, 28 Oct 2011, Markus Trippelsdorf wrote: > > What about the issues with libtool's notion of sysroots, as mentioned in > > - have those > > been resolved, and if so, how? And what about the directories only > > present in the src

PING 2 : [Patch Darwin/PR49992 1/2] remove ranlib special-casing from the darwin port.

Since this is approved by Mike, if there is no further comment by Monday, I plan to apply it. On 22 Oct 2011, at 08:36, Iain Sandoe wrote: On 14 Oct 2011, at 10:36, Iain Sandoe wrote: As per the PR audit trail, there is no reason to retain this special-casing for Darwin. (given that curre

PING 2 : [Patch Darwin/PR49992 2/2] remove ranlib special-casing from the darwin port.

This is unreviewed for 2 weeks. I am sure that this issue will be affecting Ada on Darwin10/11 with the latest toolchains. It might be subtle without LTO - OTOH when LTO is engaged it breaks things completely. On 22 Oct 2011, at 08:37, Iain Sandoe wrote: On 14 Oct 2011, at 10:37, Iain

Re: [PATCH][2/n] LTO option handling/merging rewrite

On Fri, 28 Oct 2011, Richard Guenther wrote: > + /* Fallthru. */ > + case OPT_fPIC: > + case OPT_fpic: > + case OPT_fpie: > + case OPT_fcommon: > + case OPT_fexceptions: > + append_option (decoded_options, decoded_options_count, foption); > + break; No doubt

Re: [PATCH][2/n] LTO option handling/merging rewrite

On Fri, Oct 28, 2011 at 01:43:07PM +0200, Richard Guenther wrote: > > This moves the existing processing of user options from lto1 to > the lto driver (lto-wrapper). It also changes the way we stream > user options from some custom binary format over to simply > streaming the original command-lin

Re: [PATCH i386] PR47698 no CMOV for volatile mem

On 10/28/2011 05:49 AM, Sergey Ostanevich wrote: > - the code for CMOV is under #ifdef in the mddle-end, which is > explicitly marked as "have to be removed" (ifcvt.c:1446) > - I have no clear evidence all platforms that support conditional move > have the same semantics that lead to the PR We're

Re: [RFC PATCH] update to libtool-2.4.2 and regenerate

On 2011.10.28 at 15:34 +, Joseph S. Myers wrote: > On Fri, 28 Oct 2011, Markus Trippelsdorf wrote: > > > > What about the issues with libtool's notion of sysroots, as mentioned in > > > - have those > > > been resolved, and if so, how

Re: Use of vector instructions in memmov/memset expanding

On 10/28/2011 05:41 AM, Michael Zolotukhin wrote: >> > +/* Target hook. Returns rtx of mode MODE with promoted value VAL, that is >> > + supposed to represent one byte. MODE could be a vector mode. >> > + Example: >> > + 1) VAL = const_int (0xAB), mode = SImode, >> > + the result is const

[PING] Pass address space to REGNO_MODE_CODE_OK_FOR_BASE_P

The following patch still needs maintainer review: http://gcc.gnu.org/ml/gcc-patches/2011-10/msg01874.html Thanks, Ulrich -- Dr. Ulrich Weigand GNU Toolchain for Linux on System z and Cell BE ulrich.weig...@de.ibm.com

Re: [C++ Patch / RFC] PR 50870

On 10/27/2011 08:22 PM, Paolo Carlini wrote: I'm trying to figure out where the very different args argument is coming from. Earlier than that things are different: in mainline the same args, as arglist, comes from fixup_template_parm, and earlier we have fixup_template_parms which creates the

Re: [C++ Patch / RFC] PR 50870

On 10/28/2011 06:07 PM, Jason Merrill wrote: On 10/27/2011 08:22 PM, Paolo Carlini wrote: I'm trying to figure out where the very different args argument is coming from. Earlier than that things are different: in mainline the same args, as arglist, comes from fixup_template_parm, and earlier

Re: [Patch, fortran] [00/66] PR fortran/43829 Inline sum and product (AKA scalarization of reductions)

On Friday 28 October 2011 15:56:36 Jack Howarth wrote: > Mikael, > The complete patch bootstraps current FSF gcc trunk on > x86_64-apple-darwin11 and the resulting gfortran compiler can compile the > Polyhedron 2005 benchmarks using... > > Compile Command : gfortran-fsf-4.7 -O3 -ffast-math -fu

Re: [PATCH, rs6000] Preserve link stack for 476 cpus

On Fri, 2011-10-28 at 08:20 -0700, Richard Henderson wrote: > On 10/27/2011 06:43 PM, Peter Bergner wrote: > > Ok, here's a patch to implement that, and it passes bootstrap and > > regtesting. Richard, is this what you had in mind? I'll note that > > I disabled rs6000_code_end for TARGET_POWERPC6

Re: [PATCH, rs6000] Preserve link stack for 476 cpus

On 10/28/2011 09:36 AM, Peter Bergner wrote: > Oops, you're write. Had I compiled with -m64 -mcpu=power7 -mtune=476fp, > I would have caught that. I guess (supposing we don't want to support > 64-bit) I should have the following hunks instead, correct? > > + /* If not explicitly specified via o

Re: [RFC PATCH] update to libtool-2.4.2 and regenerate

On 28.10.11 01:35, Markus Trippelsdorf wrote: By popular demand, I've prepared a patch that updates the in-tree libtool to version 2.4.2. It is needed for lto-bootstrap with -fno-fat-lto-objects and FreeBSD10.x versions. It's a pretty big update as you can see by the following diffstat. I cannot

Re: Go patch committed: Implement new syscall package

Ian, >> I committed this patch which should fix this problem. Bootstrapped and >> ran Go testsuite on x86_64-unknown-linux-gnu. > > thanks, but this is not enough: > > nawk: syntax error at source line 173 > context is > ([^ >>> ]*)$", <<< cparam) == 0) { > nawk: illegal statement at

Re: [RFC PATCH] update to libtool-2.4.2 and regenerate

On 10/28/2011 10:33 AM, Rainer Orth wrote: > Markus Trippelsdorf writes: > >> By popular demand, I've prepared a patch that updates the in-tree >> libtool to version 2.4.2. It is needed for lto-bootstrap with >> -fno-fat-lto-objects and FreeBSD10.x versions. > > I see that your patch doesn't de

[C++ Patch] PR 50864

Hi, as per the recent discussion. This also changes c++/50870 from ice-on-valid to reject-valid, a *tad* better I think (but I mean to continue working on it, for a while). Tested x86_64-linux. Ok for mainline? Thanks, Paolo. /cp 2011-10-28 Paolo Carlini PR

[pph] Various Merging Fixes (issue5330048)

Add namespace merging. This change generalizes pph_out_merge_keys on the global namespace to all namspaces. Preallocate decl_lang_specific for namespaces with streaming in their keys. Stream out namespace members in declaration order. This change mysteriously fixes mysterious bugs. Add initial

[Patch Ada RFA] make sure that multilibs are built with correct s-oscons.ads

The sizes of items represented in s-oscons.ads can (and do) change with the multi-lib on targets that support libada as a multi-lib. At present, s-oscons.ads is only built once (in gcc/ada) and sym- linked to rts*/ This is causing a bunch of failures on i686-darwin9 where the m64 multi-li

Re: [C++ Patch] PR 50864

OK. Jason

[PATCH, i386]: Remove lshift_insn and lshift code attributes

Hello! We can extend existing code attributes. Also, the patch includes some stylistic changes in XOP shift patterns. No functional changes. 2011-10-28 Uros Bizjak * config/i386/i386.md (shift_insn): Rename code attribute from shiftrt_insn. Also handle ashift RTX. (

Re: scalar vector shift expansion problem on 64-bit

On Fri, Oct 28, 2011 at 06:50:49PM +0200, Jakub Jelinek wrote: > On Fri, Oct 28, 2011 at 09:07:31AM -0700, Richard Henderson wrote: > > I think this is the same problem as Jakub is attacking here: > > > > http://gcc.gnu.org/ml/gcc-patches/2011-10/msg02503.html > > It has been checked in already

Re: [PATCH, rs6000] Preserve link stack for 476 cpus

On Fri, 2011-10-28 at 09:44 -0700, Richard Henderson wrote: > Not quite. You can't allow the user to set TARGET_LINK_STACK either, > for 64-bit. Because it won't work without further fixups. More like > > if (TARGET_POWERPC64) > SET_TARGET_LINK_STACK (0); > if (TARGET_LINK_STACK == -1)

Re: scalar vector shift expansion problem on 64-bit

On 10/28/2011 11:41 AM, Jakub Jelinek wrote: > On Fri, Oct 28, 2011 at 06:50:49PM +0200, Jakub Jelinek wrote: >> On Fri, Oct 28, 2011 at 09:07:31AM -0700, Richard Henderson wrote: >>> I think this is the same problem as Jakub is attacking here: >>> >>> http://gcc.gnu.org/ml/gcc-patches/2011-10/ms

[PATCH] Pattern recognize shifts with different rhs1/rhs2 types

Hi! This patch implements what I've talked about, with this we can now with -mavx2 as well as -mxop vectorize long long/unsigned long long shifts by int or long long/unsigned long long shifts by long long (where the FE casts it to int first). Already covered by the *vshift-* testcases I've commit

Re: [Patch Darwin/PPC] implement out-of-line FPR/GPR saves/restores.

On Oct 28, 2011, at 3:40 AM, Iain Sandoe wrote: > To test what you suggested I built some code that dropped down a few stack > levels (with saves of FPR/GPR) and then either aborts or spins on a sleep. Uhm, that's not enough, for async, you need to spawn threads that do the interesting stuff in

Re: PING 2 : [Patch Darwin/PR49992 2/2] remove ranlib special-casing from the darwin port.

On Oct 28, 2011, at 8:41 AM, Iain Sandoe wrote: > This is unreviewed for 2 weeks. Odd, usually the Ada people are fairly responsive. If they want me to weigh in, I approve of the concept behind the work.

Re: [PATCH, rs6000] Preserve link stack for 476 cpus

On 10/28/2011 11:35 AM, Peter Bergner wrote: > On Fri, 2011-10-28 at 09:44 -0700, Richard Henderson wrote: >> Not quite. You can't allow the user to set TARGET_LINK_STACK either, >> for 64-bit. Because it won't work without further fixups. More like >> >> if (TARGET_POWERPC64) >> SET_TARGE

Re: [PATCH, i386]: Remove lshift_insn and lshift code attributes

On Fri, Oct 28, 2011 at 8:21 PM, Uros Bizjak wrote: > Hello! > > We can extend existing code attributes.  Also, the patch includes some > stylistic changes in XOP shift patterns. > > No functional changes. Eh, the patch was the old one, added is additional patch with updated ChangeLog: 2011-10-2

[C++ Patch] Avoid uninitialized warning in pt.c

Hi, I think we have to do this, the warning I'm seeing doesn't seem bogus: in principle comp_template_args_with_info may leave bad_old_arg and bad_new_arg uninitialized. Ok? Thanks, Paolo. 2011-10-28 Paolo Carlini * pt.c (unify_pack_expansion): Initialize bad_ol

Re: [C++ Patch] Avoid uninitialized warning in pt.c

OK. Jason

Re: [PATCH, rs6000] Preserve link stack for 476 cpus

On Fri, Oct 28, 2011 at 12:36 PM, Peter Bergner wrote: > So David, do we even want to bother trying to support this on -m64 > given the only cpu that needs this is a 32-bit only cpu?  If so, I > can try and work with Alan to figure out how we can merge the > function descriptors for the thunk rou

Re: Require canonical type comparison for typedefs again.

On Fri, Dec 17, 2010 at 12:11 PM, H.J. Lu wrote: > On Wed, Oct 27, 2010 at 7:24 AM, Dodji Seketeli wrote: >> Hello, >> >> So I forgot to remove the wrong "optimization" on the parms of >> template template parameters. In the patch at the end of this message >> this hunks fixes that: >> > ... >> >

Re: [patch][google] Allow static const floats unless -pedantic is passed. (issue 5306071)

Thanks Diego, Here's a new version of the patch with fixes for your comments. I'll submit it in a couple hours unless I hear objections. On Fri, Oct 28, 2011 at 7:57 AM, Diego Novillo wrote: > On Thu, Oct 27, 2011 at 09:27,   wrote: >> Reviewers: Diego Novillo, >> >> Message: >> This patch is in

Re: [PATCH i386] PR47698 no CMOV for volatile mem

On Fri, Oct 28, 2011 at 7:25 PM, Sergey Ostanevich wrote: > On Fri, Oct 28, 2011 at 4:52 PM, Richard Guenther wrote: >> On Fri, 28 Oct 2011, Sergey Ostanevich wrote: >> >>> On Fri, Oct 28, 2011 at 12:16 PM, Richard Guenther >>> wrote: >>> > On Thu, 27 Oct 2011, Uros Bizjak wrote: >>> > >>> >> H

Re: scalar vector shift expansion problem on 64-bit

On Fri, Oct 28, 2011 at 11:44:17AM -0700, Richard Henderson wrote: > > A wild guess, though untested, because I don't have a reproducer: > > > > 2011-10-28 Jakub Jelinek > > > > * tree-vect-stmts.c (vectorizable_shift): If op1 is vect_external_def > > and has different type from op0, c

[PR50869] don't attempt to expand CFA within cselib

An assertion check meant to verify that var loc expansions that didn't involve VALUEs (say constants, REGs, etc) didn't push values onto the dependency stack failed in an expansion of the argp reg, because equivalences for it are preserved at cselib table resets, and cselib later tries to expand it

[wwwdocs] Update home page with reference to release announcement

Committed. Gerald Index: index.html === RCS file: /cvs/gcc/wwwdocs/htdocs/index.html,v retrieving revision 1.818 diff -u -r1.818 index.html --- index.html 27 Oct 2011 12:07:48 - 1.818 +++ index.html 28 Oct 2011 20:45:33 -0

Re: [Patch, libfortran, 1/3] Simplify handling of special files

On Tuesday 18 October 2011 16:42:45 Janne Blomqvist wrote: > Hi, > > in a few places in libgfortran we have some code for handling special > and/or non-seekable files differently. The problem is that special > files don't all have some nice consistent behavior. E.g. wrt. seeking, > some allow seek

Re: [PATCH 3/6] Implement interleave via permutation.

On Mon, 24 Oct 2011, Richard Henderson wrote: > From: Richard Henderson > + /* Certain vector operations can be implemented with vector permutation. > */ > + if (VECTOR_MODE_P (mode)) > +{ > + enum tree_code tcode = ERROR_MARK; > + rtx sel; > + > + if (binoptab == vec_inte

Patch committed: Fix -fsplit-stack unwind

The CFI code for the -fsplit-stack support was incomplete. It did not correctly unwind throughout the function. This patch fixes that by adding the CFI information at the start. This is probably still not precisely correct at all times, but it is better. In 32-bit mode, the exception cleanup ro

Go patch committed: Add rune type

The Go language has a new type, rune, which is currently an alias for int. In the future it will be an alias for int32. This type represents a Unicode character. Bootstrapped and ran Go testsuite on x86_64-unknown-linux-gnu. Committed to mainline. Ian diff -r b1a5c9a0b9ff go/gogo.cc --- a/go/

  1   2   >