Re: [PATCH, committed] Change default for powerpc -msave-toc-indirect

On Mon, Aug 29, 2011 at 11:32 PM, Michael Meissner wrote: > David asked me to reverse the default for the -msave-toc-indirect option, > since > it is buggy in a few cases, and in other places causes slowdowns if the > function has an early exit before calling the indrect function.  I also > docum

Re: [PATCH] Change vcond to vcond

On Mon, Aug 29, 2011 at 9:44 PM, Richard Guenther wrote: >> > This patch makes a conversion optab from the direct optabs vcond >> > and vcondu.  This allows to specify different modes for the >> > actual comparison and the value that is selected. >> > >> > All targets but i386 are trivially conve

Re: [PATCH, MELT, minor] add a primitive read_strv

Hello, this is a ping for http://gcc.gnu.org/ml/gcc-patches/2011-07/msg01503.html. I thought this was integrated into MELT but it does not look to be in the current trunk. It just add a primitive read_strv allowing to use meltgc_read_from_val to get a list of s-expression from a boxed C string of

Re: [PATCH 01/20] Make CLooG-ISL the only supported CLooG version.

On 08/29/2011 04:01 PM, Joseph S. Myers wrote: I looked at the ISL sources and saw that they generally had license notices of the form: * Use of this software is governed by the GNU LGPLv2.1 license Hi Joseph, at least according to a license compatibility chart[1] published by the FSF, it i

Re: [testsuite, i386] Fix for PR50185

On Tue, Aug 30, 2011 at 8:39 AM, Kirill Yukhin wrote: > Hi, > This is a ping. > > Is the patch ok for trunk? OK. Thanks, Uros.

RE: [PATCH, testsuite] Avoid architecture options conflict for case pr42894.c

Ping. BR, Terry > -Original Message- > From: Terry Guo [mailto:terry@arm.com] > Sent: Thursday, August 25, 2011 7:46 PM > To: 'gcc-patches@gcc.gnu.org' > Subject: [PATCH, testsuite] Avoid architecture options conflict for > case pr42894.c > > Hello, > > I think it is useful to run t

Re: [Ada] Reimplement Ada.Numerics.Generic_Real_Arrays in pure Ada

Dear Arnaud, Arnaud Charlet wrote: > > +-- This version of Generic_Real_Arrays avoids the use of BLAS and > LAPACK. One > +-- reason for this is new Ada 2012 requirements that prohibit algorithms > such > +-- as Strassen's algorithm, which may be used by some BLAS > implementations. In > M

Re: [PATCH] Change vcond to vcond

On Tue, 30 Aug 2011, Uros Bizjak wrote: > On Mon, Aug 29, 2011 at 9:44 PM, Richard Guenther wrote: > > >> > This patch makes a conversion optab from the direct optabs vcond > >> > and vcondu.  This allows to specify different modes for the > >> > actual comparison and the value that is selected.

[PATCH, MELT] add primitive isnull_tree

Hello, This is a small patch adding primitive isnull_tree (as there is already a primitive isnull_basicblock). Pierre Vittet Index: gcc/melt/xtramelt-ana-base.melt === --- gcc/melt/xtramelt-ana-base.melt (révision 178282) +++ gc

Re: [patch libstdc++]: Add some missing errno-constants for mingw-targets

... I committed the below because what we have now cannot possibly be right, can easily break the build. Please the target maintainers double check the whole thing, test it. Paolo. // 2011-08-30 Paolo Carlini * config/os/mingw32/error_constants.h: Fix commas

Re: [PATCH] Change vcond to vcond

On Tue, Aug 30, 2011 at 10:50 AM, Richard Guenther wrote: >> > Hm, ok - that was the first variant I tried (well, but with modeless >> > operands 1 and 2, to keep 4 and 5 selcting int vs. fp compare).  But >> > modeless operands get you that annoying warning from the gen* programs. >> >> Only fo

Re: [PATCH] Change vcond to vcond

On Tue, Aug 30, 2011 at 11:02:02AM +0200, Uros Bizjak wrote: > > Hmm.  But then I'd have to try emit an insn, right?  Currently > > the vectorizer simply looks for an optab handler ... the > > operands are not readily available (but their mode is known). > > So I'd create some fake regs, setup oper

Re: [PATCH] Change vcond to vcond

On Tue, 30 Aug 2011, Jakub Jelinek wrote: > On Tue, Aug 30, 2011 at 11:02:02AM +0200, Uros Bizjak wrote: > > > Hmm.  But then I'd have to try emit an insn, right?  Currently > > > the vectorizer simply looks for an optab handler ... the > > > operands are not readily available (but their mode is k

Re: [PATCH] Change vcond to vcond

On Tue, Aug 30, 2011 at 11:07 AM, Jakub Jelinek wrote: > On Tue, Aug 30, 2011 at 11:02:02AM +0200, Uros Bizjak wrote: >> > Hmm.  But then I'd have to try emit an insn, right?  Currently >> > the vectorizer simply looks for an optab handler ... the >> > operands are not readily available (but their

Re: [PATCH] Change vcond to vcond

On Tue, 30 Aug 2011, Uros Bizjak wrote: > On Tue, Aug 30, 2011 at 11:07 AM, Jakub Jelinek wrote: > > On Tue, Aug 30, 2011 at 11:02:02AM +0200, Uros Bizjak wrote: > >> > Hmm.  But then I'd have to try emit an insn, right?  Currently > >> > the vectorizer simply looks for an optab handler ... the >

Re: [patch libstdc++]: Add some missing errno-constants for mingw-targets

Hi Pedro, The update of copyright date is ok. The rest of the patch doesn't look right. Why you have here a stray comma? Why you remove here leading spaces for comments? Kai 2011/8/30 Paolo Carlini : > ... I committed the below because what we have now cannot possibly be right, > can easily br

Re: [PATCH] Change vcond to vcond

On Tue, Aug 30, 2011 at 11:15 AM, Richard Guenther wrote: >> >> > Hmm.  But then I'd have to try emit an insn, right?  Currently >> >> > the vectorizer simply looks for an optab handler ... the >> >> > operands are not readily available (but their mode is known). >> >> > So I'd create some fake r

Re: [patch libstdc++]: Add some missing errno-constants for mingw-targets

On 08/30/2011 11:18 AM, Kai Tietz wrote: Hi Pedro, Paolo The update of copyright date is ok. The rest of the patch doesn't look right. Why you have here a stray comma? Actually, *you* had wrong commas. Look closer. Paolo.

Re: [patch libstdc++]: Add some missing errno-constants for mingw-targets

2011/8/30 Paolo Carlini : > On 08/30/2011 11:18 AM, Kai Tietz wrote: >> >> Hi Pedro, > > Paolo >> >> The update of copyright date is ok. >> The rest of the patch doesn't look right.  Why you have here a stray >> comma? > > Actually, *you* had wrong commas. Look closer. > > Paolo. Yes, saw it now.

Re: [testsuite, i386] Fix for PR50185

Thank you! Guys who is able to write, could you please check-in my changes? K On Tue, Aug 30, 2011 at 12:12 PM, Uros Bizjak wrote: > On Tue, Aug 30, 2011 at 8:39 AM, Kirill Yukhin > wrote: >> Hi, >> This is a ping. >> >> Is the patch ok for trunk? > > OK. > > Thanks, > Uros. >

Re: [PATCH] Change vcond to vcond

On Tue, 30 Aug 2011, Uros Bizjak wrote: > On Tue, Aug 30, 2011 at 11:15 AM, Richard Guenther wrote: > > >> >> > Hmm.  But then I'd have to try emit an insn, right?  Currently > >> >> > the vectorizer simply looks for an optab handler ... the > >> >> > operands are not readily available (but thei

Re: [PATCH] Change vcond to vcond

On Tue, 30 Aug 2011, Richard Guenther wrote: > On Tue, 30 Aug 2011, Uros Bizjak wrote: > > > On Tue, Aug 30, 2011 at 11:15 AM, Richard Guenther > > wrote: > > > > >> >> > Hmm.  But then I'd have to try emit an insn, right?  Currently > > >> >> > the vectorizer simply looks for an optab handler

Re: [PATCH][ARM] -m{cpu,tune,arch}=native

On 29/08/11 04:29, Michael Hope wrote: > On Sat, Aug 27, 2011 at 3:19 AM, Andrew Stubbs wrote: >> Hi all, >> >> This patch adds support for -mcpu=native, -mtune=native, and -march=native >> for ARM Linux hosts. >> >> So far, it only recognises Cortex-A8 and Cortex-A9, so I really need to find >> o

Re: [PING][PATCH PR43513, 1/3] Replace vla with array - Implementation.

On Wed, Aug 24, 2011 at 10:47 AM, Tom de Vries wrote: > Hi Richard, > > On 07/30/2011 09:21 AM, Tom de Vries wrote: >> This is an updated version of the patch. I have 2 new patches and an updated >> testcase which I will sent out individually. >> >> Patch set was bootstrapped and reg-tested on x86

Re: [PATCH 01/20] Make CLooG-ISL the only supported CLooG version.

On Tue, 30 Aug 2011, Tobias Grosser wrote: > On 08/29/2011 04:01 PM, Joseph S. Myers wrote: > > I looked at the ISL sources and saw that they generally had license > > notices of the form: > > > > * Use of this software is governed by the GNU LGPLv2.1 license > > Hi Joseph, > > at least accor

[0/4] Make SMS schedule register moves

I'm seeing several cases in which SMS's register move handling is causing it to generate worse code than the normal schedulers on ARM Cortex-A8. The problem is that we first schedule the loop without taking the moves into account, then emit the required moves immediately before the initial definit

[1/4] SMS: remove register undo list

This patch removes the (unused) undo_replace_buff_elem list. Verges on the obvious, but still. Patch 3 splits the generation of moves into two: one function to record what needs to happen, and another function to actually do it. It's then easy to bail out if we decide that we don't want the moves

[PATCH] Fix PR48571, remove (most) array-ref re-construction

I've run into PR48571 again, and after having fun with out data-dep analysis code I indeed believe that we should not try to re-construct ARRAY_REFs from code involving pointer arithmetic. ARRAY_REFs have constraints imposed on them that are relied on by data dependence analysis and that are inde

[2/4] SMS: Use ids to represent ps_insns

Instructions in a partial schedule are currently represented as a ddg node. This patch uses a more abstract id instead. At the moment, the ids map directly to ddg nodes, but the next patch will add register moves to the end. One slight advantage of using ids is that we can leave the ASAP value o

[3/4] SMS: Record moves in the partial schedule

This patch adds infrastructure that will be used by the final patch. Specifically: - it splits the generation of register moves into two: schedule_reg_moves records moves in the partial schedule, while apply_reg_moves makes the register substitutions. This patch doesn't actually sch

Re: [PATCH Atom][PR middle-end/44382] Tree reassociation improvement

On Wed, Aug 10, 2011 at 4:49 PM, Ilya Enkovich wrote: > Hello, > > Here is a new version of the patch. Changes from the previous version > (http://gcc.gnu.org/ml/gcc-patches/2011-07/msg02240.html): >  - updated to trunk >  - TODO_remove_unused_locals flag was removed from todo_flags_finish > of re

[4/4] Make SMS schedule register moves

This is the move-scheduling patch itself. It should be fairly self-explanatory. Let me know if it isn't, and I'll try to improve the commentary. One potentially controversial change is to the way we handle moves in the prologue and epilogue. The current code uses a conservative check to decide

Re: [4/4] Make SMS schedule register moves

On Tue, Aug 30, 2011 at 2:29 PM, Richard Sandiford wrote: > This is the move-scheduling patch itself.  It should be fairly > self-explanatory.  Let me know if it isn't, and I'll try to improve > the commentary. Can you add some testcases? Thanks, Richard. > One potentially controversial change

[PATCH] [MELT] Fix versions comparisons

Comparison of MELT version were not consistent accross the codebase: - MELT .c meltdesc were generated with melt_versionstr using MELT_VERSION_STRING define - comparison at several places in the source code was done against MELT_VERSION_STRING - comparision at other places was done between

Re: [4/4] Make SMS schedule register moves

Richard Guenther writes: > On Tue, Aug 30, 2011 at 2:29 PM, Richard Sandiford > wrote: >> This is the move-scheduling patch itself.  It should be fairly >> self-explanatory.  Let me know if it isn't, and I'll try to improve >> the commentary. > > Can you add some testcases? I don't think it's an

Re: Vector shuffling

On Tue, Aug 30, 2011 at 4:31 AM, Artem Shinkarov wrote: > Hi > > This is a patch for the explicit vector shuffling we have discussed a > long time ago here: > http://gcc.gnu.org/ml/gcc-patches/2010-08/msg01092.html > > The new patch introduces the new tree code, as we agreed, and expands > this co

Re: [4/4] Make SMS schedule register moves

On Tue, Aug 30, 2011 at 2:43 PM, Richard Sandiford wrote: > Richard Guenther writes: >> On Tue, Aug 30, 2011 at 2:29 PM, Richard Sandiford >> wrote: >>> This is the move-scheduling patch itself.  It should be fairly >>> self-explanatory.  Let me know if it isn't, and I'll try to improve >>> the

[Ada] Prefixed calls as generic actuals

Generic actuals are pre-analyzed in order to capture external names, but are only fully analyzed within the instance. If an actual is a prefixed call to a dispatching operation, it may appear syntactically as a call without actuals. It is necessary to parse it as an object.operation again in order

[Ada] Ada 2012 anonymous access conversions and membership tests

This set of changes implements support for the Ada 2012 features of implicit conversions for anonymous access types and membership tests with a test expression of an anonymous access type (see AI05-0149). Expressions of an anonymous access type are now generally allowed in contexts where the expec

Re: [4/4] Make SMS schedule register moves

Richard Guenther writes: > On Tue, Aug 30, 2011 at 2:43 PM, Richard Sandiford > wrote: >> Richard Guenther writes: >>> On Tue, Aug 30, 2011 at 2:29 PM, Richard Sandiford >>> wrote: This is the move-scheduling patch itself.  It should be fairly self-explanatory.  Let me know if it isn'

[Ada] Ada 2012 accessibility of anonymous access stand-alone objects

This set of changes implements support for AI05-0148, the Ada2012 rules concerting the accessibility of a stand-alone object of an anonymous access type (hereafter, a "saooaaat"). The Extra_Accessibility attribute was previously only used for a formal parameter of an anonymous access type. The use

Re: [PATCH, i386, testsuite] FMA intrinsics

On Tue, Aug 30, 2011 at 2:42 AM, Ilya Tocar wrote: > Fixed. > > 2011/8/26 H.J. Lu : >> On Fri, Aug 26, 2011 at 8:47 AM, H.J. Lu wrote: >>> On Fri, Aug 26, 2011 at 8:06 AM, Ilya Tocar wrote: Done. Also fixed  changelog:  2011-08-26  Ilya Tocar        

[Ada] Define a light expansion mode for formal verification

Instead of the full expansion targetting code generation, a light expansion is now used in formal verification mode. Light expansion has three main objectives: 1. Perform limited expansion to explicit some Ada rules and constructs 2. Facilitate treatment for the formal verification back-end 3. Avo

[Ada] Add support to detect conflicting overriding primitive

This patch incorporates support in the compiler to detect conflicts in declarations of primitives of concurrent tagged types (see full documentation in AI05-0090-1). In addition the patch also performs a minor code cleanup in order to factorize the code which reports an error on wrong formal of pro

[Ada] Refine expansion in Alfa mode

Expansion in Alfa mode should not be the full expansion for code generation. Refine tests that control such expansion until a specific "light" expansion is defined. Tested on x86_64-pc-linux-gnu, committed on trunk 2011-08-30 Yannick Moy * exp_ch9.adb, exp_disp.adb, sem_ch9.adb, sem_r

[Ada] Access all type not considered access-to-object in generic

This patch ensures that the full view of a designated type is available in the body of an instance when the related access type acts as an actual in another instance. Tested on x86_64-pc-linux-gnu, committed on trunk 2011-08-30 Hristian Kirtchev * sem_ch12.adb (Check_Private_View): Ex

Re: [PATCH] Fix PR48571, remove (most) array-ref re-construction

On Tue, 30 Aug 2011, Richard Guenther wrote: > > I've run into PR48571 again, and after having fun with out data-dep > analysis code I indeed believe that we should not try to re-construct > ARRAY_REFs from code involving pointer arithmetic. ARRAY_REFs > have constraints imposed on them that are

[Ada] Do not catch exception if _UA_FORCE_UNWIND flag is set.

Fix a non-compliance with the unwind ABI. No pure Ada testcase. Tested on x86_64-pc-linux-gnu, committed on trunk 2011-08-30 Tristan Gingold * raise-gcc.c: Never catch exception if _UA_FORCE_UNWIND flag is set, to be compliant with the ABI. Index: raise-gcc.c

Re: [PATCH, i386, testsuite] FMA intrinsics

On Tue, Aug 30, 2011 at 6:55 AM, Ilya Tocar wrote: > Little fix. > > 2011/8/30 Ilya Tocar : >> Sorry. Regenerated. >> >> 2011/8/30 H.J. Lu : >>> On Tue, Aug 30, 2011 at 2:42 AM, Ilya Tocar wrote: Fixed. 2011/8/26 H.J. Lu : > On Fri, Aug 26, 2011 at 8:47 AM, H.J. Lu wrote:

[Ada] Define full expansion "flag" to distinguish from reduced Alfa expansion

Minor refactoring to clarify where full expansion applies wrt Alfa mode. Tested on x86_64-pc-linux-gnu, committed on trunk 2011-08-30 Yannick Moy * opt.adb, opt.ads (Full_Expander_Active): New function defines a common shorthand for (Expander_Active and not ALFA_Mode) that can

[Ada] Renames System.Parameter.Sec_Stack_Ratio

As this constant was a percentage, it is now renamed to Sec_Stack_Percentage. No functional change. Tested on x86_64-pc-linux-gnu, committed on trunk 2011-08-30 Tristan Gingold * s-parame-vms-alpha.ads, s-parame-hpux.ads, s-tassta.adb, s-tarest.adb, s-parame-vms-ia64.ads, s-so

[Ada] Define a light expansion mode for formal verification (2)

Following up on the previous change, we now move all the formal verification expansion in a separate exp_alfa.ad? file. Previous exp_*_light change is "reverted/canceled" at the same time (not shown here for convenience). Instead of the full expansion targetting code generation, a light expansion

Re: [PATCH][ARM] Thumb2 replicated constants

On 26/08/11 11:03, Andrew Stubbs wrote: There was a bug I found in final testing, so this has been delayed somewhat. I've just committed this version. There are a few minor changes to the way negative/inverted constants are generated. Bernd found another bug whist testing for arm. Apparently t

Re: [C++0x] contiguous bitfields race implementation

[I'm going to respond to this piece-meal, to make sure I don't drop anything. My apologies for the long thread, but I'm pretty sure it's in everybody's kill file by now.] + /* Be as conservative as possible on variable offsets. */ + if (TREE_OPERAND (exp, 2) +&& !host_integerp (TREE_OPERA

Re: [PATCH][ARM] Thumb2 replicated constants

On 30/08/11 15:32, Andrew Stubbs wrote: On 26/08/11 11:03, Andrew Stubbs wrote: There was a bug I found in final testing, so this has been delayed somewhat. I've just committed this version. There are a few minor changes to the way negative/inverted constants are generated. Bernd found anothe

C++ PATCH for c++/50220 (ICE with capture by copy of reference)

In this testcase, we weren't instantiating Foobar until fairly late, so the field and proxy variable didn't have a size, leading to madness. Tested x86_64-pc-linux-gnu, applying to trunk and 4.6. commit 7082532a480501ab8f8f99bc62589a9045da303a Author: Jason Merrill Date: Tue Aug 30 10:38:52

C++ PATCH for c++/50234 (ICE with constexpr and empty init-list)

This is basically the same as c++/49924, but for classes instead of arrays: if there is no explicit initializer written in the init-list, the member is value-initialized. Tested x86_64-pc-linux-gnu, applying to trunk and 4.6. commit eccc1f7f9987a394dcf9856d43244a95c51bc880 Author: Jason Merrill

Re: [PATCH, PR fortran/45170] -- properly translates substring reference

On Sun, Aug 28, 2011 at 11:36:42PM +0200, Mikael Morin wrote: > On Saturday 27 August 2011 02:34:24 Steve Kargl wrote: > > The attached patch allows gfortran to compile the > > attached testcase. The resulting executable runs > > as expected. > > > > Short story: > > > > character(len=20) :: s

[v3] Implement DR 2061 (Ready in Bloomington)

Hi, tested x86_64-linux, committed. Paolo. /// 2011-08-30 Paolo Carlini * include/bits/stl_iterator.h (make_move_iterator): Implement DR2061. * testsuite/24_iterators/move_iterator/dr2061.cc: New. * testsuite/20_util/weak_ptr/comparison/cmp_neg.cc

Re: options generation and language count

On Mon, 29 Aug 2011, Gary Funck wrote: > 1) The sanity check should probably read: >gcc_assert ((1U << cl_lang_count) <= CL_MIN_OPTION_CLASS); > In other words, it is off-by-one. Please send a patch. > 2) The fixed shift counts starting at CL_PARAMS probably need to be > adjusted upwards to

Re: options generation and language count

On 08/30/11 15:52:12, Joseph S. Myers wrote: > Please send a patch. [... on points 1, 2, and 3]0 OK, will do. > GF: Also, the use of fixed masks is problematic. Perhaps the AWK script > GF: could be changed to also generate values for CL_PARAMS, etc., ensuring > GF: that will not conflict with t

Re: [C++0x] contiguous bitfields race implementation

*bit_offset is supposed to be relative to *byte_offset then it should be easy to calculate it without another get_inner_reference. Since, as you suggested, we will terminate early on variable length offsets, we can assume both DECL_FIELD_OFFSET and DECL_FIELD_BIT_OFFSET will be constants by

[Ada] Fix ICE on initialization with discriminated aggregate

The compiler aborts on the initialization of an object with an aggregate, if its nominal subtype is a discriminated record type with a variant part for which the variants all have the same size and one of the variants contains a component whose type is tagged or controlled. Tested on x86_64-sus

Re: [PATCH, committed] Change default for powerpc -msave-toc-indirect

On Tue, Aug 30, 2011 at 09:16:40AM +0200, Richard Guenther wrote: > On Mon, Aug 29, 2011 at 11:32 PM, Michael Meissner > wrote: > > David asked me to reverse the default for the -msave-toc-indirect option, > > since > > it is buggy in a few cases, and in other places causes slowdowns if the > > f

[Ada] Clean up and document use of IN_GCC outside the compiler

This cleans up and documents the use of IN_GCC in the GNAT build system. No functional change. Tested on x86_64-suse-linux, applied on the mainline. 2011-08-30 Eric Botcazou * raise-gcc.c: Do not include coretypes.h and tm.h. (setup_to_install): Remove test for compiler macr

[PATCH PING] Remove obsolete alias check in cgraph_redirect_edge_call_stmt_to_callee

Hi, this is a ping of a patch that I have originally posted almost two months ago. Since (same body) aliases have their own cgraph_nodes, the check for them in cgraph_redirect_edge_call_stmt_to_callee is now unnecessary because e->callee is now the alias, not the function node. The following pa

[Ada] Housekeeping work in gigi (32/n)

Tested on x86_64-suse-linux, applied on the mainline. 2011-08-30 Eric Botcazou * gcc-interface/ada-tree.h (TYPE_EXTRA_SUBTYPE_P): Add internal check. * gcc-interface/utils2.c (build_binary_op): Tighten condition. (build_unary_op): Likewise. -- Eric Botcazou Index: g

Re: [Ada] Remove GCC_ZCX_Support in system-*.ads

> This constant was in fact useless as ZCX_By_Default implied it. > > No change in behavior. > > Tested on x86_64-pc-linux-gnu, committed on trunk A few missing entries. Tested on x86_64-suse-linux, applied on the mainline. 2011-08-30 Eric Botcazou * system-irix-n64.ads, system-linu

Re: [PATCH, PR 49886] Prevent fnsplit from changing signature when there are type attributes

Ping. Re-bootstrapped and re-tested yesterday on x86_64-linux. THanks, Martin On Fri, Jul 29, 2011 at 10:55:31PM +0200, Martin Jambor wrote: > Hi, > > On Thu, Jul 28, 2011 at 06:52:05PM +0200, Martin Jambor wrote: > > pass_split_functions is happy to split functions which have type > > attrib

Re: Vector shuffling

On Tue, 30 Aug 2011, Richard Guenther wrote: > oh, hum - now I remember ;) Eventually the C frontend should handle > this not via the function call mechanism but similar to how Joseph > added __builtin_complex support with > > 2011-08-19 Joseph Myers > > * c-parser.c (c_parser_postfi

Re: options generation and language count

On Tue, 30 Aug 2011, Gary Funck wrote: > On 08/30/11 15:52:12, Joseph S. Myers wrote: > > Please send a patch. [... on points 1, 2, and 3]0 > > OK, will do. > > > GF: Also, the use of fixed masks is problematic. Perhaps the AWK script > > GF: could be changed to also generate values for CL_PARA

Re: Vector shuffling

On Tue, Aug 30, 2011 at 2:03 PM, Richard Guenther wrote: > On Tue, Aug 30, 2011 at 4:31 AM, Artem Shinkarov > wrote: >> Hi >> >> This is a patch for the explicit vector shuffling we have discussed a >> long time ago here: >> http://gcc.gnu.org/ml/gcc-patches/2010-08/msg01092.html >> >> The new pa

Re: [Ada] Clean up and document use of IN_GCC outside the compiler

On Tue, 30 Aug 2011, Eric Botcazou wrote: > This cleans up and documents the use of IN_GCC in the GNAT build system. The mess of IN_GCC isn't at all GNAT-specific. See what I said in . Beyond that, include/ansidecl.h tests IN_GCC to avo

[PATCH, i386]: Fix FMA fallout

Hello! The gcc.target/i386/fma-compile.c test did not escape [ and ] correctly. Unfortunately, this problem terminated dejagnu testing early, hiding the problem with -mfma target attribute handling. Attached patch fixes both problems. 2011-08-30 Uros Bizjak * config/i386/i386.c (ix8

Re: [PATCH, i386]: Fix FMA fallout

On Tue, Aug 30, 2011 at 10:46 AM, Uros Bizjak wrote: > Hello! > > The gcc.target/i386/fma-compile.c test did not escape [ and ] > correctly.  Unfortunately, this problem terminated dejagnu testing > early, hiding the problem with -mfma target attribute handling. > > Attached patch fixes both probl

Re: [Patch][Cilkplus Branch] Updates and Bug fixes

On Mon, Aug 29, 2011 at 7:41 PM, Andi Kleen wrote: > "H.J. Lu" writes: > >> On Mon, Aug 29, 2011 at 12:44 PM, Iyer, Balaji V >> wrote: >>> Hello Everyone, >>> >>> Here is a link for a new patch (http://software.intel.com/file/38290).   >>> This patch is for the "cilkplus" branch and includes the

Re: [C++0x] contiguous bitfields race implementation

Btw, *byte_offset is still not relative to the containing object as documented, but relative to the base object of the exp reference tree (thus, to a in a.i.j.k.l instead of to a.i.j.k). If it were supposed to be relative to a.i.j.k get_inner_reference would be not needed either. Can you clari

Re: [PATCH][ARM] -m{cpu,tune,arch}=native

On Tue, Aug 30, 2011 at 10:47 PM, Stubbs, Andrew wrote: > On 29/08/11 04:29, Michael Hope wrote: >> On Sat, Aug 27, 2011 at 3:19 AM, Andrew Stubbs  wrote: >>> Hi all, >>> >>> This patch adds support for -mcpu=native, -mtune=native, and -march=native >>> for ARM Linux hosts. >>> >>> So far, it only

Re: [Patch][Cilkplus Branch] Updates and Bug fixes

> Please cleanup the formatting issues on cilkplus branch and please > follow GCC coding standard. I'm not sure if following the coding standard is required for the library. After all it's essentially a third party library and I assume shared with other cilkplus implementations. It may make more s

Re: [Patch][Cilkplus Branch] Updates and Bug fixes

On Tue, Aug 30, 2011 at 12:59 PM, Andi Kleen wrote: >> Please cleanup the formatting issues on cilkplus branch and please >> follow GCC coding standard. > > I'm not sure if following the coding standard is required for the library. > After all it's essentially a third party library and I assume sh

Re: [DOC] CP_TYPE_QUALS -> cp_type_quals

Hi, apparently the doc wasn't updated 10 years ago when the macro CP_TYPE_QUALS was replaced with the function cp_type_quals. I assume the behavior didn't change, so the documentation change should be minor. Or are functions documented differently? 2011-08-08 Marc Glisse * doc/generic.

[Patch, Fortran] PR 45044 - Named common: Different size diagnostics

Named common blocks are required to have the same size in all files and scopes. gfortran was warning before when the size was extended, but not when a smaller common block followed a larger common block with the same name. This patch adds now a non-equal check and also mentions the byte size.

C++ PATCH for c++/50084 (ICE with decltype and unnamed local class)

In this testcase the compiler was incorrectly treating the typedef as giving a name to the class for linkage purposes, and then getting confused by that. While looking at this I noticed that nothing was actually looking at the user_defined_type_p bitfield, so I repurposed it. Tested x86_64-pc

C++ PATCH for c++/50089 (ICE with qualified-id call in lambda)

Just needed to correct current_class_type to current_nonlambda_class type here. Tested x86_64-pc-linux-gnu, applying to trunk and 4.6. commit 159563fe81e80fd23ff6252e17c86012a20b169b Author: Jason Merrill Date: Tue Aug 30 15:46:31 2011 -0400 PR c++/50089 * semantics.c (finish_id_ex

Re: C++ PATCH for c++/50207 (ICE on bogus decimal::decimal32)

On Tue, 2011-08-30 at 00:27 -0400, Jason Merrill wrote: > The DFP classes are treated as transparent aggregates, so they get > mangled and passed like their first field. In the testcase it is > defined to have a base class, so the first field is an artificial base > field, which doesn't make an

C++ PATCH for c++/50114 (ICE with for-loop and lambda)

We play funny games with binding levels to support backward compatibility with the ARM-era rules for for-loop scoping, and we play funny games with binding levels to support lambdas, and the games didn't play well with each other. It seems unlikely that people will still be relying on ancient

Re: [Patch, Fortran] PR 45044 - Named common: Different size diagnostics

On Tue, Aug 30, 2011 at 11:31:38PM +0200, Tobias Burnus wrote: > Named common blocks are required to have the same size in all files and > scopes. gfortran was warning before when the size was extended, but not > when a smaller common block followed a larger common block with the same > name. >

Re: Vector shuffling

On Aug 30, 2011, at 10:01 AM, Artem Shinkarov wrote: >>> The patch at the moment lacks of some examples, but mainly it works >>> fine for me. It would be nice if i386 gurus could look into the way I >>> am doing the expansion. >>> >>> Middle-end parts seems to be more or less fine, they have not c

Re: tree hash maps and 'discards qualifiers' warnings?

On 08/17/11 08:38:44, Gary Funck wrote: > > I have been looking at changing UPC's method of > recording the blocking factor so that it uses less space > in the tree type node. The suggested method for > achieving this space reduction is to use a hash table to > map pointers to type nodes into UP

Re: C++ PATCH for c++/50207 (ICE on bogus decimal::decimal32)

Sure, the patch is safe enough. Done. Peter Bergner wrote: On Tue, 2011-08-30 at 00:27 -0400, Jason Merrill wrote: > The DFP classes are treated as transparent aggregates, so they get > mangled and passed like their first field. In the testcase it is > defined to have a base class, so the fir

Re: [PATCH] Fix ICEs in vect_finish_stmt_generation (PR tree-optimization/50133)

Jakub Jelinek wrote on 22/08/2011 05:22:59 PM: > Hi! > > The following testcase ICEs, because gsi_end_p (*gsi) and thus > there is no stmt after it from which to copy over the location. > As can be seen in the PR, we could do ugly hacks to retrieve locus > from previous stmt (non-debug of cours