Re: [patch, ARM] Make 128 bits the default vector size for NEON

On 6 April 2011 16:07, Hans-Peter Nilsson wrote: >> Date: Thu, 31 Mar 2011 13:39:05 +0200 >> From: Ira Rosen > >> This patch changes NEON's default vector size from 64 to 128 bits. > > I'm wondering, are there NEON-specific measurements to support > this change? > > A colleague of mine implemente

[gc-improv] Permanent vs function RTL obstack fix

Fixes a bunch of C testsuite failures. Committed to gc-improv. 2011-04-07 Laurynas Biveinis * stmt.c (label_rtx): Allocate RTX in permanent RTL memory. Index: stmt.c === --- stmt.c (revision 170593) +++ stmt.c (working cop

[gc-improv] Missing prototype in the last patch

Committed to gc-improv. 2011-04-07 Laurynas Biveinis * rtl.h (gen_rtx_CONST): New. Index: rtl.h === --- rtl.h (revision 171651) +++ rtl.h (working copy) @@ -2188,6 +2188,7 @@ generation functions included above do the

Re: Cloog ISL - and linking of libisl

* Jack Howarth wrote on Thu, Apr 07, 2011 at 12:20:00AM CEST: > On Tue, Feb 15, 2011 at 08:07:43PM +0100, Tobias Burnus wrote: > > --- config/cloog.m4 (revision 166641) > > +++ config/cloog.m4 (working copy) > > @@ -143,7 +143,7 @@ > >;; > > "ISL") > >clooginc="${clooginc} ${_c

Re: PATCH [trunk] gengtype should generate ggc_alloc macros in plugin mode.

Hello Basile - The patch is correct in general. But a few nitpicks: > 2011-04-04 Basile Starynkevitch >* gengtype.c (write_typed_alloc_def): Gets extra outf_p argument >and use it. ...: New argument f. Use it instead of header_file. >(write_typed_struct_alloc_def, wr

Re: [patch] Import iota and is_sorted from std to __gnu_cxx in C++0x mode. (issue4384043)

Hi, > I think this is a candidate to backport to libstdc++-4.6.1. Ok mainline and 4_6-branch if testing passes. Thanks, Paolo

[patch] Import iota and is_sorted from std to __gnu_cxx in C++0x mode. (issue4384043)

In C++0x mode, import iota and is_sorted from namespace std into namespace __gnu_cxx to avoid ambiguous call errors. copy_n would be a good candidate for this too, except it was standardized with a different return type. I think this is a candidate to backport to libstdc++-4.6.1. Tested: `make

Re: [Ping #2][Patch, testsuite]: Don't xfail sibcalls on AVR

On Wed, 6 Apr 2011, Georg-Johann Lay wrote: > I signed a fsf copyright assignment. AFAIK there is no automatism > that grants write access when copyright assignment is approved. Yes, this is correct. > I would have to be maintainer or something like that. Nope. All you need is a maintainer (Eri

[PATCH 4/7] Removal of cgraph_node function

This patch contains the rest of the changes required to remove cgraph_node function. First, new functions cgraph_create_node and cgraph_get_create_node are introduced. The former creates a new node for decl (and aborts compilation if th node already exists) and the latter creates it if it does no

[PATCH 1/7] Simple cgraph_node -> cgraph_get_node conversions

Hi, this patch changes most of current calls to cgraph_node to calls to cgraph_get_node. The function should never return NULL in he contexts of the callers below and either the probability is so low that it does not warrant an assert (like cgraph_get_node (current_function_decl) or the result is

[PATCH 6/7] A tweak to fortran -> call graph interface

Hi, it seems to me that fortran can call cgraph_create_node directly without checking for its existence first. Bootstrapped and tested on x86_64-linux without any problems, tests on i686 in progress. Thanks, Martin 2011-03-18 Martin Jambor * trans-decl.c (gfc_generate_function_cod

[PATCH 7/7] Tweaks to objc -> call graph interface

Hi, this is really only based on successful testing and not much analyzis of the context but it seems that we don't need lazy node construction here. It would be nice not to have it after the big patch gets in. Bootstrapped and tested on x86_64-linux without any problems, tests on i686 in progre

[PATCH 5/7] Tweaks to C++ -> call graph interface

I concede that my understanding of the C++ front-end inner workings are quite narrow and so the folling is basically a suggestion. But it seems to me that at a few places where C++ queries the call graph for a node, the lazy node creation is not necessary. If a maintainer can verify and approve (

[PATCH 2/7] cgraph_node -> cgraph_get_node conversions accepting NULL results

Hi, this patch converts a number of calls to cgraph_node to calls to cgraph_get_node and provides means to deal with returned NULL value. These are essentially the places where lazy node creation was happening for no good reason. Bootstrapped and tested separately on x86_64-linux without any prob

[PATCH 3/7] cgraph_node -> cgraph_get_node with asserts

Hi, the patch below changes a number of calls to cgraph_node to calls to cgraph_get_node. These calls should never return NULL but because the callers do not immediately dereference the pointer they get, a NULL value would result in a segmentation fault at some later point, obscuring the real cau

[PATCH 0/7] Change of call graph interface - cgraph_node function removal

Hi, I'm sending this email describing the set of patches also to the gcc mailing list because quite a few people do not follow the gcc-patches mailing list (where I am sending the individual patches). The patch set changes the interface of call graph and tries to avoid lazy call graph node creati

Re: [C++0x] Range-based for statements and ADL

Hi! It took some time but I finally re-wrote the patch and added a few testsuites. A few comments as usual: I've moved the array stuff into cp_parser_perform_range_for_lookup. I think that it belongs there: it's just a special lookup case . I finally agreed with Jason about calling lookup_member

[PATCH] change TDI_generic into TDI_gimple

Hey, while searching through the code related to debug dumps, I realized that that the TDI_generic enum value is misleading since it is used to dump out GIMPLE code instead of GENERIC tree code. This is already indicated by the used dump file name extension (.gimple). Changelog: * d

Go patch committed: Use backend interface for switch statements

This patch to the Go frontend uses the backend interface to produce switch statements for which all the cases are integer constants. Bootstrapped and ran Go testsuite on x86_64-unknown-linux-gnu. Committed to mainline. Ian 2011-04-06 Ian Lance Taylor * go-gcc.cc (if_statement): Use b

Re: [wwdocs] add libstdc++ docs for 4.6.0

> I ended up fiddling a bit with permissions and such, but everything > should be in place now and I also committed the web patch on your > behalf. Thanks! This pleases me a great deal. > Testing the links, there is one thing does does not work, and where > I was not sure how to address it: Th

Re: [wwdocs] add libstdc++ docs for 4.6.0

On Tue, 5 Apr 2011, Benjamin Kosnik wrote: > I've gone ahead and generated libstdc++ documentation for the 4.6.0 > release. Please find enclosed a patch for the main gcc documentation > page (http://gcc.gnu.org/onlinedocs) that adds these links. Nice! > Of course, these links are predicated on yo

Re: Cloog ISL - and linking of libisl

On Tue, Feb 15, 2011 at 08:07:43PM +0100, Tobias Burnus wrote: > Dear Sebastian and Tobias, hello all, > > I think GCC should automatically link libisl ("-lisl") if one configures > GCC to use cloog-isl. That's the first patch at > http://gcc.gnu.org/ml/gcc-patches/2010-11/msg01272.html , which

Convert legacy m68k options to .opt aliases

Similar to , this patch converts legacy m68k options for non-ColdFire CPUs into aliases for the corresponding -mcpu= options. (Note that -mcpu32 is an alias for -mcpu=68332 rather than -mcpu=cpu32, to match the old code in m68k_handle_option

[MIPS] Remove REG_OK_FOR_BASE_P and REG_OK_FOR_INDEX_P macros

Hello. This patch remove unused REG_OK_FOR_BASE_P and REG_OK_FOR_INDEX_P macros from the MIPS back end. Bootstrapped and regression tested on mips64el-unknown-linux-gnu. OK to install? * config/mips/mips.h (REG_MODE_OK_FOR_BASE_P, REG_OK_FOR_INDEX_P): Remove macros. Ind

Re: [RFC] Remove kludge in commit_edge_insertions

-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On 04/06/11 12:39, Steven Bosscher wrote: > On Wed, Apr 6, 2011 at 8:37 PM, Steven Bosscher wrote: >> On Wed, Apr 6, 2011 at 8:15 PM, Jeff Law wrote: >> AFAIU the patch doesn't change behavior for the GIMPLE CFG. It only affects RTL. >>> B

[PATCH] Remove ASM_OUTPUT_BSS macro.

Hi. No one back end does not use ASM_OUTPUT_BSS macro now, this patch remove it. The GCC have more flexible ASM_OUTPUT_ALIGNED_BSS macro. The patch has been bootstrapped on and regression tested on x86_64-unknown-linux-gnu for c. OK to install? * doc/tm.texi.in (ASM_OUTPUT_BSS): R

Re: [RFC] Remove kludge in commit_edge_insertions

-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On 04/06/11 12:37, Steven Bosscher wrote: > On Wed, Apr 6, 2011 at 8:15 PM, Jeff Law wrote: > >>> AFAIU the patch doesn't change behavior for the GIMPLE CFG. It only affects >>> RTL. >> But the elimination of the PHI results in creating RTL that is

Re: [Patch][AVR]: Initial builtins support

Added Anatoly and Eric to changelog entry Johann 2011-04-06 Anatoly Sokolov Eric Weddington Georg-Johann Lay * config/avr/avr.c: ("insn-codes.h", "optabs.h", "langhooks.h"): New Includes (avr_init_builtins, avr_expand_builtin, avr_ex

Re: [RFC] Remove kludge in commit_edge_insertions

> What about when we have a PHI, which we eliminate by inserting insns on > edges and those insns actually form a loop? You can see an example of > this in PR48389. Interesting case, thanks. I think it's simply a matter of defining the API: does commit_edge_insertions need to fix up the CFG or

Re: [RFC] Remove kludge in commit_edge_insertions

On Wed, Apr 6, 2011 at 8:37 PM, Steven Bosscher wrote: > On Wed, Apr 6, 2011 at 8:15 PM, Jeff Law wrote: > >>> AFAIU the patch doesn't change behavior for the GIMPLE CFG. It only affects >>> RTL. >> But the elimination of the PHI results in creating RTL that is inserted >> on a CFG edge. > > Yes

Re: [RFC] Remove kludge in commit_edge_insertions

On Wed, Apr 6, 2011 at 8:15 PM, Jeff Law wrote: >> AFAIU the patch doesn't change behavior for the GIMPLE CFG. It only affects >> RTL. > But the elimination of the PHI results in creating RTL that is inserted > on a CFG edge. Yes, but gimple_expand_cfg() calls find_many_sub_basic_blocks(), and

[Patch, Fortran, committed] Some coarray fixes (PR 18918 and 48477)

I have committed the following three rather obvious patches (separately): 1. libgfortran/caf/mpi.c: I am now finally convinced that one indeed need to make use of MPI_Initialized's return value when calling MPI_Finalize. Thanks to all who pointed it out several times. 2. gcc/fortran/trans-int

Re: [PATCH] Cleanup, use add_to_hard_reg_set instead of SET_HARD_REG_BIT loops.

This looks like a nice cleanup. One comment: On Wed, Apr 06, 2011 at 10:21:12PM +0400, Anatoly Sokolov wrote: > @@ -4958,60 +4958,54 @@ > mark_reload_reg_in_use (unsigned int regno, int opnum, enum reload_type type, > enum machine_mode mode) > { > - unsigned int nregs =

Re: [patch] Preserve return statements at -O0

> Isn't one of the effects of this patch to generate debug info for the > artificial label decl? No, DECL_IGNORED_P is preserved on the label decl. > When I first saw your patch, I thought "FORCED_LABEL()", perhaps I'm > not sure that it's any better than your solution... Have you tried > that?

[PATCH] Cleanup, use add_to_hard_reg_set instead of SET_HARD_REG_BIT loops.

Hi. This patch converts loops of SET_HARD_REG_BIT in to add_to_hard_reg_set functions call. The patch has been bootstrapped on and regression tested on x86_64-unknown-linux-gnu for c. OK to install? * expr.c (expand_expr_real_1): Use add_to_hard_reg_set function instead o

Re: [RFC] Remove kludge in commit_edge_insertions

-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On 04/06/11 12:01, Steven Bosscher wrote: > On Wed, Apr 6, 2011 at 6:52 PM, Jeff Law wrote: >> What about when we have a PHI, which we eliminate by inserting insns on >> edges and those insns actually form a loop? You can see an example of >> this in

Re: [patch] Preserve return statements at -O0

On Wed, Apr 6, 2011 at 8:57 AM, Eric Botcazou wrote: > this patchlet ensures that return statements are preserved at -O0 so that you > can put a breakpoint on them in the debugger. Isn't one of the effects of this patch to generate debug info for the artificial label decl? When I first saw your

Re: [RFC] Remove kludge in commit_edge_insertions

On Wed, Apr 6, 2011 at 6:52 PM, Jeff Law wrote: > What about when we have a PHI, which we eliminate by inserting insns on > edges and those insns actually form a loop?  You can see an example of > this in PR48389. AFAIU the patch doesn't change behavior for the GIMPLE CFG. It only affects RTL. C

Re: Fix powerpc rs6000_stack_info ICE

On Wed, Apr 6, 2011 at 8:18 AM, Alan Modra wrote: > The ENABLE_CHECKING tests I added at the end of rs6000_stack_info > are now asserting on gcc.c-torture/execute/20041011-1.c and other > testcases.  The cause is the late running DSE pass removing a set of > LR, which means lr_save_p changes.  Why

Re: [PATCH, PR43920, 6/9] Cross-jumping - Use reg-notes.

-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On 04/01/11 08:54, Tom de Vries wrote: > On 03/31/2011 11:16 PM, Tom de Vries wrote: >> On 03/31/2011 08:52 PM, Jeff Law wrote: >> >>> On 03/31/11 12:42, Tom de Vries wrote: Uses regnotes to analyze whether we can replace insn a by insn b, even >>

[PATCH 2/n, i386]: Merge SSE and AVX patterns using "enable" attribute.

Hello! This patch continues merging of SSE and AVX patterns in "Parallel floating point logical operations", "Parallel single-precision floating point conversion operations" and "Parallel double-precision floating point conversion operations" sections of sse.md. Additionally, it introduces handli

Re: [cxx-mem-model] bitfield tests

On 04/06/11 10:29, Michael Matz wrote: Hi, On Mon, 4 Apr 2011, Aldy Hernandez wrote: (5) Do we agree that all such cpus use a byte-granular modification mask? Now, as of (0) I might agree to disregard the original Alpha, but as the embedded world moves to SMP I'm not sure we can disregard

Re: [PATCH, ARM] Switch to EABI version 5 for RTEMS

On 06/04/11 18:24, Ralf Corsepius wrote: > On 04/06/2011 05:20 PM, Sebastian Huber wrote: > >> there were several requests for ARM Cortex-M support on RTEMS >> recently. The >> first step towards this is a suitable ARM tool chain. I want to use >> this event >> to clean up the multilibs and switc

Re: PATCH: PR middle-end/48440: [4.7 Regression] FAIL: gcc.c-torture/compile/labels-3.c

On Tue, Apr 5, 2011 at 3:29 AM, Richard Guenther wrote: > On Tue, Apr 5, 2011 at 8:44 AM, Paolo Bonzini wrote: > Index: cgraphbuild.c > === > --- cgraphbuild.c.orig  2011-04-03 11:28:45.0 +0200 > +++ cgraphbui

Rename LINUX_TARGET_OS_CPP_BUILTINS and LINUX_DYNAMIC_LINKER*

This patch, relative to a tree with (pending review) applied, further cleans up the handling of GNU-userspace targets by renaming LINUX_TARGET_OS_CPP_BUILTINS (a macro also defined in headers such as kfreebsd-gnu.h) to GNU_USER_TARGET_OS_CPP

Re: [RFC] Remove kludge in commit_edge_insertions

-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On 04/04/11 15:07, Eric Botcazou wrote: > commit_edge_insertions contains this kludge: > > /* In the old rtl CFG API, it was OK to insert control flow on an > edge, apparently? In cfglayout mode, this will *not* work, and > the caller is

Re: [PATCH 1/n, i386]: Merge SSE and AVX patterns using "enable" attribute.

On Wed, Apr 6, 2011 at 1:12 PM, Kirill Yukhin wrote: > Nice patch. But I have two question: > - What will be with future extensions? E.g. what if `avxXX` will appear, do > you think that common parts with previous extentions should be shrinked in > that way? The patch in fact mixes two somehow o

Re: [PATCH, ARM] Switch to EABI version 5 for RTEMS

On 04/06/2011 05:20 PM, Sebastian Huber wrote: there were several requests for ARM Cortex-M support on RTEMS recently. The first step towards this is a suitable ARM tool chain. I want to use this event to clean up the multilibs and switch to the EABI version 5. The benefit of EABI version 5 i

Re: License on gcov-io.c

RMS approved relicensing gcov-io.c, so I've applied this patch to trunk, 4.6 and 4.5 branch to add the license exception. Perhaps someone with a 4.4 branch checkout could apply it there as well? 2011-04-06 Joseph Myers * gcov-io.c: Use GCC Runtime Library Exception. Index: gcov-io.

Re: RFA: MN10300: Handle MDR register in movqi/movhi pattern.

-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On 04/06/11 03:38, Nick Clifton wrote: > Hi Jeff, Hi Alex, Hi Richard, > > The MN10300 gcc port currently does not build on the mainline because > of: > > In file included from gcc/unwind-dw2-fde.c:35:0: > gcc/unwind-pe.h: In functi

[PATCH] Fix PR48437

This fixes PR48437, we shouldn't unify function-local externs as that confuses the DECL_CHAIN lists. It seems they can be identified by looking at TREE_STATIC and decl_function_context, if not we'll find out (maybe the hard way, with wrong-code because of unmerged decls). Eventually we should si

[PATCH] Fix PR48467

This avoids unlinking the assembler output file at the end of WPA stage (we don't produce anything useful there) which eventually clobbers an assembler output from a previous stage. Instead simply don't open the file. That requires not performing debug init which we shouldn't need either here (m

Go patch committed: Use backend interface for if statements

This patch to the Go frontend uses the backend interface for if statements. I also renamed some of the temporary conversion functions to use shorter names. Bootstrapped and ran Go testsuite on x86_64-unknown-linux-gnu. Committed to mainline. Ian 2011-04-06 Ian Lance Taylor * go-gcc

Go patch committed: Use backend interface for expression statements

This patch to the Go frontend uses the backend interface for expression statements. Bootstrapped and ran Go testsuite on x86_64-unknown-linux-gnu. Committed to mainline. Ian 2011-04-06 Ian Lance Taylor * go-gcc.cc (Gcc_backend::expression_statement): New function. Index: gcc/go/g

Re: [cxx-mem-model] bitfield tests

Hi, On Mon, 4 Apr 2011, Aldy Hernandez wrote: > > > > (5) Do we agree that all such cpus use a byte-granular modification mask? > > > Now, as of (0) I might agree to disregard the original Alpha, but as the > > embedded world moves to SMP I'm not sure we can disregard > > non-cache coherent NUM

[testsuite] Account for absolute path to ranlib on Darwin (PR testsuite/48480)

As reported in the PR, we can get ranlib messages with an absolute path to ranlib when running the boehm-gc testsuite on x86_64-apple-darwin10. This differs from what Iain had reported for powerpc-apple-darwin9. The following patch accounts for that, tested by myself in tclsh and by Dominique on t

[PATCH, ARM] Switch to EABI version 5 for RTEMS

Hello, there were several requests for ARM Cortex-M support on RTEMS recently. The first step towards this is a suitable ARM tool chain. I want to use this event to clean up the multilibs and switch to the EABI version 5. The benefit of EABI version 5 is that this brings RTEMS more in line with

Re: PowerPC64 non-delegitimized unspecs

On Wed, Apr 06, 2011 at 02:45:49PM +0200, Jakub Jelinek wrote: > On Wed, Apr 06, 2011 at 10:06:11PM +0930, Alan Modra wrote: > > I've noticed a number of tests fail on mainline with complaints about > > non-delegitimized UNSPECs. The reason why rs6000_delegitimze_address > > isn't working is that

Re: [Patch][AVR]: Initial builtins support

Patch for user documentation. 2011-04-06 Georg-Johann Lay * doc/extend.texi (Target Builtins): Add documentation of AVR built-in functions. (AVR Built-in Functions): New node Index: doc/extend.texi ===

Re: [patch libjava]: Do some fixes for windows native targets

2011/4/6 Kai Tietz : > 2011/4/5 Tom Tromey : >>> "Kai" == Kai Tietz writes: >> >> Kai> this patch fixes some issues in libjava for windows targets.  It >> Kai> addresses for now only 32-bit target, but for JNICALL definition I >> Kai> introduced already logic for 64-bit windows target too as p

[Patch][AVR]: Initial builtins support

This patch adds builtin support for some RTL builtins to avr backend. One builtin implements loop for delay of specific number of ticks (under the assumption IRQs are off), others simply map to machine instructions like SEI, CLI, NOP, SLEEP, WDR, FMUL, FMAULS, FMULSU. In addition, builtin macros a

Re: [patch, ARM] Make 128 bits the default vector size for NEON

> Date: Thu, 31 Mar 2011 13:39:05 +0200 > From: Ira Rosen > This patch changes NEON's default vector size from 64 to 128 bits. I'm wondering, are there NEON-specific measurements to support this change? A colleague of mine implemented support for 64- and 128-bit NEON for RAPP

Re: [Patch,AVR]: Fix PR45263

Georg-Johann Lay schrieb: > Weddington, Eric schrieb: >>> -Original Message- >>> From: Georg-Johann Lay [mailto:a...@gjlay.de] >>> Sent: Friday, April 01, 2011 4:01 PM >>> To: Georg-Johann Lay >>> Cc: gcc-patches@gcc.gnu.org; Denis Chertykov; Anatoly Sokolov; Weddington, >>> Eric >>> Subjec

Re: [Ping #2][Patch, testsuite]: Don't xfail sibcalls on AVR

Weddington, Eric schrieb: > >> -Original Message- >> From: Georg-Johann Lay [mailto:a...@gjlay.de] >> To: Weddington, Eric >>> Please commit, Johann. Thanks. >>> >> I don't have write privilege, and I don't think I want write access at >> the moment. > > Oh! I didn't know. I thought you

Re: PowerPC64 non-delegitimized unspecs

On Wed, Apr 06, 2011 at 10:06:11PM +0930, Alan Modra wrote: > I've noticed a number of tests fail on mainline with complaints about > non-delegitimized UNSPECs. The reason why rs6000_delegitimze_address > isn't working is that our insn matching doesn't allow a stack slot in > place of the REG. I

Re: [patch libjava]: Do some fixes for windows native targets

2011/4/5 Tom Tromey : >> "Kai" == Kai Tietz writes: > > Kai> this patch fixes some issues in libjava for windows targets.  It > Kai> addresses for now only 32-bit target, but for JNICALL definition I > Kai> introduced already logic for 64-bit windows target too as preparation. > > Most of this

PowerPC64 non-delegitimized unspecs

I've noticed a number of tests fail on mainline with complaints about non-delegitimized UNSPECs. The reason why rs6000_delegitimze_address isn't working is that our insn matching doesn't allow a stack slot in place of the REG. I could have expanded the condition to include stack slots, but it see

Re: [PATCH, PR43920, 2/9] ARM specific part - test case

On 04/06/2011 12:48 PM, Ramana Radhakrishnan wrote: > >> Is this updated version ( >> http://gcc.gnu.org/ml/gcc-patches/2011-04/msg00058.html ) ok for trunk? > > I think you accidentally applied the patch twice and committed it. That is exactly what happened, according to my history: ... 1877

RE: [Ping #2][Patch, testsuite]: Don't xfail sibcalls on AVR

> -Original Message- > From: Georg-Johann Lay [mailto:a...@gjlay.de] > Sent: Wednesday, April 06, 2011 4:55 AM > To: Weddington, Eric > Cc: gcc-patches@gcc.gnu.org; Anatoly Sokolov; Denis Chertykov > Subject: Re: [Ping #2][Patch, testsuite]: Don't xfail sibcalls on AVR > > >>> Target avr

Fix powerpc rs6000_stack_info ICE

The ENABLE_CHECKING tests I added at the end of rs6000_stack_info are now asserting on gcc.c-torture/execute/20041011-1.c and other testcases. The cause is the late running DSE pass removing a set of LR, which means lr_save_p changes. Why that happens ultimately goes back to a register allocation

Re: Remove old toplevel subdirectory support

On 04/06/2011 02:07 PM, Joseph S. Myers wrote: This patch removes toplevel support for various subdirectories not present in the gcc or src repositories, as agreed in the thread including. Some other directories may in fact also be no longer usefu

Remove old toplevel subdirectory support

This patch removes toplevel support for various subdirectories not present in the gcc or src repositories, as agreed in the thread including . Some other directories may in fact also be no longer useful to support building at toplevel, or may be bui

[4.5] Backport various selective scheduler patches to 4.5 branch

Hello, As we discussed in PR 43603, we need to do another round of sel-sched patches' backporting to 4.5 -- there were a few reports about the problems already fixed on trunk. I have successfully bootstrapped and tested on x86-64 and ia64 the unified patch (attached) that backports fixes for

Re: [Ping #2][Patch, testsuite]: Don't xfail sibcalls on AVR

Weddington, Eric schrieb: > >> -Original Message- >> From: Georg-Johann Lay [mailto:a...@gjlay.de] >> Sent: Tuesday, April 05, 2011 3:06 AM >> To: gcc-patches@gcc.gnu.org >> Cc: Anatoly Sokolov; Denis Chertykov; Weddington, Eric >> Subject: Re: [Ping #2][Patch, testsuite]: Don't xfail sibc

Re: [PATCH, PR43920, 2/9] ARM specific part - test case

Is this updated version ( http://gcc.gnu.org/ml/gcc-patches/2011-04/msg00058.html ) ok for trunk? I think you accidentally applied the patch twice and committed it. Yeah unfortunately there is no equivalent of git clean -f -d in svn unless you write something up with svn status and some shel

[PATCH] Merge early inliner cleanup from pretty-ipa

This merges the last patch from pretty-ipa, the early inliner cleanup. Bootstrapped and tested on x86_64-unknown-linux-gnu, applied to trunk. Richard. 2011-02-22 Richard Guenther * ipa-inline.c (enum inlining_mode): Remove. (cgraph_flatten): Use some other token. (c

Re: [PATCH][ARM] Tweak arm_class_likely_spilled_p, MODE_BASE_REG_CLASS for Thumb-2

On 30/03/11 16:13, Richard Earnshaw wrote: 2011-02-14 Andrew Stubbs Julian Brown Mark Shinwell gcc/ * config/arm/arm.h (arm_class_likely_spilled_p): Check against LO_REGS only for Thumb-1. (MODE_BASE_REG_CLASS): Restrict base registe

[Patch ARM] Committed fix for PR43920 testcase

Hi, I've committed the following patch under the obvious rule for the testcase for PR43920. There seemed to have been an accidental duplication of the function in the patch committed. Tested in a cross compiler to arm-linux-gnueabi with just that test to armv7-a and made sure that it passes.

RFA: MN10300: Handle MDR register in movqi/movhi pattern.

Hi Jeff, Hi Alex, Hi Richard, The MN10300 gcc port currently does not build on the mainline because of: In file included from gcc/unwind-dw2-fde.c:35:0: gcc/unwind-pe.h: In function 'read_encoded_value_with_base': gcc/unwind-pe.h:271:1: error: insn does not satisfy its con

Re: Separate C-family hooks better from general hooks

On Wed, Apr 6, 2011 at 03:27, Joseph S. Myers wrote: > 2011-04-05  Joseph Myers   > >        * doc/tm.texi.in: Document C target hooks as separate from general >        target hooks. >        * doc/tm.texi: Regenerate. >        * genhooks.c (struct hook_desc): Add docname field. >        (HOOK_VE

Re: [PATCH] fix PR bootstrap/48471

On Wed, Apr 6, 2011 at 9:44 AM, Tristan Gingold wrote: > > On Apr 6, 2011, at 3:06 AM, Nathan Froyd wrote: > >> My changes that moved bits of final.c to dbxout.c caused compilation >> failures on targets that *didn't* enable DBX_DEBUGGING_INFO, since some >> now-static symbols could be unused.  To

Re: [patch] Preserve return statements at -O0

On Wed, Apr 6, 2011 at 8:57 AM, Eric Botcazou wrote: > Hi, > > this patchlet ensures that return statements are preserved at -O0 so that you > can put a breakpoint on them in the debugger. > > Bootstrapped/regtested on x86_64-suse-linux, OK for the mainline? Ok. Thanks, Richard. > > 2011-04-06

Re: [PATCH] Fix -m32 -g -O0 debug info (PR debug/48466)

On Wed, Apr 6, 2011 at 12:43 AM, Jakub Jelinek wrote: > Hi! > > On the pr48466.c testcase below on i?86 -m32 -g -O0 we generate wrong debug > info for some of the variables in main.  This was caused by PR36977 fix, > crtl->stack_realign_tried is true, but fde->drap_reg is INVALID_REGNUM, > but fra

[PATCH][2/2] Simplify call stmt handling in stmt costs

With the call stmt cost merge from pretty-IPA we can now simplify their handling in estimate_num_insns. Instead of fancy ways to look at types simply account for the actual arguments and returns. Bootstrapped and tested on x86_64-unknown-linux-gnu, applied to trunk. Richard. 2011-04-06 Richar

Re: [testsuite] Fix gcc.dg/torture/pr47917.c on IRIX 6.5

Eric Botcazou writes: >> The new gcc.dg/torture/pr47917.c test currently fails the execution test >> on IRIX 6.5. To get a C99-conformant snprintf, one needs to include >> with _XOPEN_SOURCE defined as 500. The following patch does >> this. > > Likewise on Solaris 8 (It's deprecated, I know, b

[PATCH][1/2] Merge call stmt cost handling from pretty-IPA

This merges the call stmt cost handling from pretty IPA together with dependent cleanups. Bootstrapped and tested on x86_64-unknown-linux-gnu, applied to trunk. Richard. 2011-04-05 Richard Guenther PR tree-optimization/47663 * cgraph.h (struct cgraph_edge): Add call_stmt_si

Define MAX_OFILE_ALIGNMENT for Tru64 UNIX

The recent failure of FAIL: gcc.dg/torture/pr47975.c -O0 (test for excess errors) /vol/gcc/src/hg/trunk/local/gcc/testsuite/gcc.dg/torture/pr47975.c:4:40: warning: alignment of 'x' is greater than maximum object file alignment. Using 16 [enabled by default] prompted me to finally look at t

Re: typo in gcc manual

On 6 April 2011 04:15, ali hagigat wrote: > 3.15 Specifying subprocesses and the switches to pass to them > *[spec_name]: > If this results in an empty string then the spec will be deleted. (Or, > if the spec did not exist, then nothing will happened.) >

Re: [PATCH] fix PR bootstrap/48471

On Apr 6, 2011, at 3:06 AM, Nathan Froyd wrote: > My changes that moved bits of final.c to dbxout.c caused compilation > failures on targets that *didn't* enable DBX_DEBUGGING_INFO, since some > now-static symbols could be unused. To fix this, the patch below moves > all the new static variables

[patch] Preserve return statements at -O0

Hi, this patchlet ensures that return statements are preserved at -O0 so that you can put a breakpoint on them in the debugger. Bootstrapped/regtested on x86_64-suse-linux, OK for the mainline? 2011-04-06 Eric Botcazou * gimple-low.c (lower_gimple_return): When not optimizing, forc