Re: [Patch, testsuite, mips] Fix two failing MIPS tests.

Steve Ellcey sell...@mips.com writes: Two MIPS specific tests started failing a few days ago, this patch fixes them. The tests are trying to check that a performance optimization is done and that one constant is derived from another and not simply loaded as a completely new constant. Post

Re: [wide-int] Add more assertions

Kenneth Zadeck zad...@naturalbridge.com writes: These are fine. Thanks. I'll hold off applying it until the java fix has been reviewed (which could be after the merge -- no need to hold it up for this IMO). Richard On 05/02/2014 03:20 PM, Richard Sandiford wrote: This patch adds some

Re: Contributing new gcc targets: i386-*-dragonfly and x86-64-*-dragonfly

On 5/2/2014 22:15, Joseph S. Myers wrote: On Fri, 2 May 2014, John Marino wrote: 1) I don't know which type definitions are missing (iow, the important ones from sys/type.h that are required to build gcc) The default presumption should be: * stddef.h from GCC provides what it needs to

Ping: RFA: speeding up dg-extract-results.sh

Ping for this old patch. I didn't ping it before because I can imagine it wasn't a good idea at that stage in 4.8. But I've been using it locally since and it really does make a big difference when testing lots of multilibs. Thanks, Richard Richard Sandiford rdsandif...@googlemail.com writes:

[SH, committed] Fix PR 61026 sh-*-* Fails to Compile on FreeBSD

On Thu, 2014-05-01 at 23:39 +0200, Oleg Endo wrote: On May 1, 2014, at 11:17 PM, Joel Sherrill joel.sherr...@oarcorp.com wrote: On 5/1/2014 3:29 PM, Oleg Endo wrote: On 01 May 2014, at 22:08, Joel Sherrill joel.sherr...@oarcorp.com wrote: Hi gcc-4.8.2 targeting sh-*-* fails

[Fortran, patch, committed] Fix PR61025

I have committed the following patch in order to fix PR61025 (preapproved at http://gcc.gnu.org/ml/fortran/2014-05/msg1.html). Dominique 2014-05-03 Dominique d'Humieres domi...@lps.ens.fr PR fortran/61025 * gfortran.dg/coarray_lib_this_image_1.f90: Adjust the dg-final

Re: [C PATCH] proposal to add new warning -Wsizeof-array-argument

On Fri, May 2, 2014 at 5:50 AM, Joseph S. Myers jos...@codesourcery.com wrote: The following apply to all versions of this patch: * New options need documenting in invoke.texi. Added. * New options need nonempty help text in c.opt. (It's unfortunate that the -Wsizeof-pointer-memaccess

Re: [C++ Patch/RFC] PR 58582

OK. Jason

[patch, avr] handle JUMP_TABLE_DATA

Committed patch. 2014-05-03 Denis Chertykov cherty...@gmail.com * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA. Index: gcc/config/avr/avr.c === --- gcc/config/avr/avr.c (revision 210034) +++

Re: [wide-int] Handle zero-precision INTEGER_CSTs again

The doc at wide-int.h:136 needs work.The doc currently says that the precision and length are always greater than 0. This is now not correct. It also says that the bits above the precision are defined to be the sign extension if the precision does not cover that block. I do not know

[Patch, Fortran] Add support for TS18508's CO_SUM/MAX/MIN (part 1/2)

Dear all, this patch adds support for the most important of TS18508's collective intrinsics, namely co_sum (for int, real, cmplx) and co_min/co_max (int, real, char). The feature is described at ftp://ftp.nag.co.uk/sc22wg5/N2001-N2050/N2007.pdf ; N2007 has an optional result= argument,

[SH, committed] Fix formatting in sh-mem.cc

Hello, Attached patch tested with 'make all-gcc', committed as r210037. Cheers, Oleg gcc/ChangeLog: * config/sh/sh-mem.cc: Use tabs instead of spaces. (prob_unlikely, prob_likely): Make variables const. Index: gcc/config/sh/sh-mem.cc

[Fortran-CAF] Merged trunk into the branch

I have now merged the trunk into the branch; committed as Rev. 210036. Due to the recent changes on the trunk, this reduces some of the differences and it also brings some coarray-related bug fixes to the branch. Tobias

Re: [wide-int] Handle zero-precision INTEGER_CSTs again

Kenneth Zadeck zad...@naturalbridge.com writes: The doc at wide-int.h:136 needs work.The doc currently says that the precision and length are always greater than 0. This is now not correct. It also says that the bits above the precision are defined to be the sign extension if the

Re: [PATCH, RS6000] Fix __builtin_{pack,unpack}_longdouble and __builtin_{pack,unpack}_dec128 builtins

On Fri, 2014-05-02 at 18:34 -0500, Peter Bergner wrote: This passed bootstrap and regtesting on powerpc64-linux with no regressions and the pack0[23].c tests now pass. Ok for trunk? Is this also ok for the 4.8/4.9 branches once my testing on those are complete? FYI, this passed bootstrap

[committed] Add a mips_isa_rev variable

In view of possible upcoming changes for new ISA revisions, I committed this patch to redo the checks for MIPS32rN and MIPS64rN. It's really just a drop in the ocean of what needs to be done to clean up the ISA selection code, but at least it's a start... Tested on mipsisa64-sde-elf. Richard

Re: [RFC][PATCH][MIPS] Patch to enable LRA for MIPS backend

Vladimir Makarov vmaka...@redhat.com writes: Not sure how the constraint would/should exclude $sp-based address in LRA. In this particular case, a spilled pseudo is changed to memory giving the following RTL form: (insn 30 29 31 4 (set (reg:SI 4 $4) (and:SI (mem/c:SI (plus:SI

Re: [wide-int] Handle zero-precision INTEGER_CSTs again

Then with a fixed comment, this patch is fine. kenny On 05/03/2014 02:59 PM, Richard Sandiford wrote: Kenneth Zadeck zad...@naturalbridge.com writes: The doc at wide-int.h:136 needs work.The doc currently says that the precision and length are always greater than 0. This is now not

[PATCH] Clean up and extend VRP edge-assertion code

Hi, This patch refactors the VRP edge-assertion code to make it unconditionally traverse SSA-name definitions in order to find suitable edge assertions to insert. Currently SSA-name definitions get traversed only when the orginal conditional is a bitwise AND or OR operation, which is a strange

Re: RFA: speeding up dg-extract-results.sh

On Feb 13, 2014, at 1:18 AM, Richard Sandiford rdsandif...@googlemail.com wrote: This patch tries to reduce that by providing an alternative single-script version. Python isn't yet required and I'm pretty sure this script needs 2.6 or later. I'm also worried that the seek/tell stuff might

[SH, committed] add missing function* argument in sh_optimize_sett_clrt::execute

Hi, I've committed the attached patch as r210040, which reinstates the RTL pass. Trevor, could you please double check that you haven't missed any other cases in your pass refactoring patches? Cheers, Oleg gcc/ChangeLog: * config/sh/sh_optimize_sett_clrt.cc

[PATCH, libgfortran] PR 61035 Stack overflow in GETCWD

Hello, the attached patch avoids a stack overflow crash due to not trying to create a null-terminated duplicate of the argument char array on the stack. Also, for the common case it avoids an extra allocation and an extra memcpy. Regtested on x86_64-unknown-linux-gnu, Ok for trunk? 2014-05-03

[PATCH, libgfortran, committed] PR 56919 Backport from 4.9 to 4.7/4.8

Hi, I committed the patch below as obvious to the 4.7 (r210042) and 4.8 (r210041) branches. 2014-05-03 Janne Blomqvist j...@gcc.gnu.org Backport from 4.9 PR libfortran/56919 * intrinsics/system_clock.c (GF_CLOCK_MONOTONIC): Check _POSIX_MONOTONIC_CLOCK as well. Index:

Re: [C++ Patch] PR 60999

On 04/30/2014 05:15 AM, Paolo Carlini wrote: decl = (CLASSTYPE_TEMPLATE_INFO (DECL_CONTEXT (decl)) +uses_template_parms (DECL_CONTEXT (decl)) Do you want CLASSTYPE_IS_TEMPLATE here? Jason

[Fortran, Patch] Some prep patches for coarray communication

This patch is a teaser: It adds some preparation for coarray communication to the trunk, but it doesn't yet added the library calls. (In particular, note the 0 in the conditions, which add the intrinsics.) The idea is that coindexed variables are wrapped in the intrinsic function

Re: [C++ Patch] PR 60999

Hi, On 05/03/2014 11:14 PM, Jason Merrill wrote: On 04/30/2014 05:15 AM, Paolo Carlini wrote: decl = (CLASSTYPE_TEMPLATE_INFO (DECL_CONTEXT (decl)) + uses_template_parms (DECL_CONTEXT (decl)) Do you want CLASSTYPE_IS_TEMPLATE here? Uhm, you mean something simple like the

[PATCH] VRP: simplify assert location check

We can check if any assertions were found by simply inspecting the need_assert_for bitmap. This eliminates the need to pass all these bools around everywhere. 2014-05-03 Patrick Palka patr...@parcs.ath.cx * tree-vrp.c (register_edge_assert_for_1): Change return type to void.