Re: PR65742: OpenACC acc_on_device fixes

On July 14, 2015 10:15:30 PM GMT+02:00, Thomas Schwinge wrote: >Hi! > >Per your request in >, »can you >please >fix the gcc 5 branch«, I'm planning to apply the following to >gcc-5-branch tomorrow (but wanted to give you a chance to veto, giv

Re: [PR 66854 powerpc regression] Fix RTEMS powerpc build issue

On Tue, Jul 14, 2015 at 8:09 PM, Michael Meissner wrote: > My IEEE 128-bit floating point infrastructure patches resulted in a > segmentation violation on RTEMS when building libgcc. I had put a test using > a > type tree variable before the test whether it was NULL or not. > > This patch bootst

Re: [PATCH] rs6000: Revamp rotate-and-mask and insert

On Sun, Jul 12, 2015 at 04:18:31PM -0400, David Edelsohn wrote: > On Sun, Jul 12, 2015 at 1:08 PM, Segher Boessenkool > wrote: > > This rewrites all the rotate-and-mask and insert patterns. > > This is great! I'm glad that you completed this feature. Compared to mainline the results do look goo

[PR 66854 powerpc regression] Fix RTEMS powerpc build issue

My IEEE 128-bit floating point infrastructure patches resulted in a segmentation violation on RTEMS when building libgcc. I had put a test using a type tree variable before the test whether it was NULL or not. This patch bootstraps on 64-bit Linux, and it also fixes the problem in question when b

Re: [gomp] Move openacc vector& worker single handling to RTL

On 07/14/15 04:25, Thomas Schwinge wrote: addr = gen_rtx_MEM (mode, addr); addr = gen_rtx_UNSPEC (mode, gen_rtvec (1, addr), UNSPEC_SHARED_DATA); - if (pm & PM_read) + if (pm == PM_read) res = gen_rtx_SET (addr, reg); - if (pm & PM_write) + else

[gomp] constify device data & fix cleanup

This patch turned out a little larger than expected as I ran into an API limitation between libgomp and the plugins. The patch changes GOMP_offload_{,un}register to take a pointer to constant target data. I've fixed up the two mkoffloads to constify their target data. The other thing this do

[v3 PATCH] Implement N4280, Non-member size() and more (Revision 2)

Tested on Linux-PPC64. 2015-07-15 Ville Voutilainen Implement N4280, Non-member size() and more (Revision 2) * include/bits/range_access.h: Change class to typename in every template. * include/bits/range_access.h(size, empty, data): New. * testsuite/24_iterators/container_a

[nios2] [7/7] Add new intrinsics

This patch adds a set of new built-in functions to the Nios II back end. Some of these are specific to R2 (wrpie, eni, and the MPX load/store exclusive instructions) and others (rdprs, flushd, flushda) correspond to instructions also present in R1 but not previously exposed as intrinsics. Commit

[nios2] [6/7] Update function prologues/epilogues for R2 CDX

This patch re-works the function prologue and epilogue generation code to emit CDX push.n/stwm and pop.n/ldwm instructions, respectively. A CDX function prologue always uses push.n to push the callee-saved registers -- the stack frame will be padded if necessary to ensure a contiguous register ra

Re: [PATCH][C++] Fix PR65091

Hi, On 07/14/2015 09:10 PM, Jason Merrill wrote: On 07/12/2015 01:53 PM, Paolo Carlini wrote: On 07/11/2015 09:46 PM, Paolo Carlini wrote: I'm going to ping this one too: a tad less trivial than the other one - a little explanation here or in a comment would definitely help - but certainly it

[nios2] [5/7] Support R2 CDX load/store multiple instructions

This installment of the Nios II R2 patch series adds support for the new CDX load/store multiple instructions (ldwm, stwm, pop.n, push.n). The implementation approach we used here is similar to that in the ARM backend, with all the insn patterns and peephole optimizers generated by a Standard ML

[nios2] [4/7] Support new R2 instructions

This patch adds GCC support for the bulk of the new R2 instructions -- everything except the new CDX load/store multiple instructions (part 5) and the instructions that are only emitted by new builtins (part 7). CDX adds a group of 16-bit instructions similar to existing 32-bit instructions, but

[nios2] [3/7] Correct nested function trampolines for R2 encodings

This patch adds R2 encodings for the instructions used in nested function trampolines in libgcc. Committed as r225794. -Sandra 2015-07-14 Sandra Loosemore Cesar Philippidis Chung-Lin Tang libgcc/ * config/nios2/tramp.c (MOVHI, ORI, JMP): Conditionalize for __nios2_arch__ le

[nios2] [2/7] Adjust for reduced offsets in R2 load/store IO insns

Nios II has a group of load/store IO instructions that bypass the normal memory cache; they're intended to be used for accessing memory-mapped IO peripherals. In the R2 re-encoding of the Nios II instruction set, the constant offset field for these instructions has been reduced from 16 bits to 12

Re: patch to fix PR66626

On Tue, Jul 14, 2015 at 1:58 PM, Vladimir Makarov wrote: > The following patch fixes > > https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66626 > > The patch was bootstrapped and tested on x86/x86-64. > > Committed as rev. 225789. > > 2015-07-14 Vladimir Makarov > > PR rtl-optimization/6

[nios2] [1/7] Add -march=, -mbmx, -mcdx flags

This patch adds command-line options to select the new R2 variants of the Nios II architecture. Aside from checking for conflicting options, the only thing that this patch does with the new flags is pass -march= through to the assembler. The -mbmx and -mcdx options will enable code generation fo

[nios2] [0/7] Support for Nios II R2

I will shortly begin committing a patch series to add GCC support for Nios II R2, a revision of the original Nios II instruction set. I previously wrote up some notes on the technical changes from R1 to R2 when I posted the corresponding binutils patches, here: https://sourceware.org/ml/binutils

RE: [PATCH] [aarch64] Implemented reciprocal square root (rsqrt) estimation in -ffast-math

I ran a simple test on A57 rev. 0, looping a million times around sqrt{,f} and the respective series iterations with the values in the sequence 1..100 and got these results: sqrt(x):36593844/s 1/sqrt(x): 18283875/s 3 Steps:47922557/s 3 Steps:49005194/s

RE: [PATCH] [aarch64] Implemented reciprocal square root (rsqrt) estimation in -ffast-math

For both FRECPE and FRSQRTE the ARMv8 ISA guide states in their pseudo-code that: "Result is double-precision and a multiple of 1/256 in the range 1 to 511/256." This suggests that the estimate is merely 8 bits long. IIRC, x86 returns 12 bits for its equivalent insns, requiring then a single s

Re: [C/C++ PATCH] Implement -Wtautological-compare (PR c++/66555, c/54979)

On Tue, Jul 14, 2015 at 10:30:06PM +0100, Richard Sandiford wrote: > Marek Polacek writes: > > + /* Don't warn for e.g. > > + HOST_WIDE_INT n; > > + ... > > + if (n == (long) n) ... > > + */ > > + if ((CONVERT_EXPR_P (lhs) || TREE_CODE (lhs) == NON_LVALUE_EXPR) > > + ^ (CONVER

Re: [C/C++ PATCH] Implement -Wtautological-compare (PR c++/66555, c/54979)

Marek Polacek writes: > + /* Don't warn for e.g. > + HOST_WIDE_INT n; > + ... > + if (n == (long) n) ... > + */ > + if ((CONVERT_EXPR_P (lhs) || TREE_CODE (lhs) == NON_LVALUE_EXPR) > + ^ (CONVERT_EXPR_P (rhs) || TREE_CODE (rhs) == NON_LVALUE_EXPR)) > +return; I might be m

RE: [PATCH] MIPS: Correctly update the isa and arch_test_option_p variables after the arch dependency handling code in mips.exp

> -Original Message- > From: Andrew Bennett [mailto:andrew.benn...@imgtec.com] > Sent: Tuesday, July 14, 2015 9:24 AM > To: Richard Sandiford; Matthew Fortune > Cc: gcc-patches@gcc.gnu.org; Moore, Catherine > Subject: RE: [PATCH] MIPS: Correctly update the isa and arch_test_option_p > var

RE: [PATCH, MIPS] Support new interrupt handler options

> -Original Message- > From: Robert Suchanek [mailto:robert.sucha...@imgtec.com] > Sent: Tuesday, July 14, 2015 11:14 AM > To: Moore, Catherine; Matthew Fortune; gcc-patches@gcc.gnu.org > Subject: RE: [PATCH, MIPS] Support new interrupt handler options > > Hi Catherine, > > > I'm gettin

Re: [gomp4, fortran] Patch to fix continuation checks of OpenACC and OpenMP directives

On 07/14/2015 02:20 PM, Ilmir Usmanov wrote: > Ping Sorry, I thought I had already approved this. It's fine for gomp-4_0-branch. Cesar > On 07.07.2015 14:27, Ilmir Usmanov wrote: >> Ping >> >> 30.06.2015, 03:43, "Ilmir Usmanov" : >>> Hi Cesar! >>> >>> Thanks for your review! >>> >>> 08.06.2015,

Re: [gomp4, fortran] Patch to fix continuation checks of OpenACC and OpenMP directives

Ping On 07.07.2015 14:27, Ilmir Usmanov wrote: Ping 30.06.2015, 03:43, "Ilmir Usmanov" : Hi Cesar! Thanks for your review! 08.06.2015, 17:59, "Cesar Philippidis" : On 06/07/2015 02:05 PM, Ilmir Usmanov wrote: Fixed fortran mail-list address. Sorry for inconvenience. 08.06.2015, 00

patch to fix PR66626

The following patch fixes https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66626 The patch was bootstrapped and tested on x86/x86-64. Committed as rev. 225789. 2015-07-14 Vladimir Makarov PR rtl-optimization/66626 * lra-constraints.c (lra_constraints): Prevent equivalence

Re: [PATCH] PR fortran/66864 -- Use the right precision in FLOOR()

On Tue, Jul 14, 2015 at 10:20 PM, Steve Kargl wrote: > The attached patch fixes an issue in the constant > folding of FLOOR(). In the old code, the precision > of an temporary MPFR variable was set according to > the requested kind of the INTEGER result, ie., the > wrong precision. The patch set

[PR 66861 5/6 regression] Mingw null pointer crash

Hi, the attached patch fixes a null pointer crash on the mingw target (or to be precise, targets where HAVE_WORKING_STAT is not defined). Committed as obvious to trunk as r225788. Richi, is this Ok for 5.2? 2015-07-14 Janne Blomqvist PR libfortran/66861 * io/unix.c (compare_file_fi

[patch] Adjust *-streamer.h

THIs patch just does a minor cleanup.. gimple-pretty-print.h doesn't need to include pretty-print.h since tree-pretty-print.h already does. I also cleaned up the 4 streamer files, such that each one includes just the previous one, and then make each source file include just he one it needs.

PR65742: OpenACC acc_on_device fixes

Hi! Per your request in , »can you please fix the gcc 5 branch«, I'm planning to apply the following to gcc-5-branch tomorrow (but wanted to give you a chance to veto, given that your backport request pre-dates the branch freeze by a week): c

Re: [nvptx offloading] Only 64-bit configurations are currently supported

On July 14, 2015 10:05:52 PM GMT+02:00, Thomas Schwinge wrote: >Hi! > >OK for gcc-5-branch? OK Richard >On Wed, 8 Jul 2015 17:03:02 +0200, I wrote: >> On Wed, 18 Feb 2015 09:50:15 +0100, I wrote: >> > So far, we have concentrated only on the 64-bit x86_64 >configuration; >> > 32-bit has severa

Re: [PATCH 6/n] OpenMP 4.0 offloading infrastructure: option handling

On July 14, 2015 10:03:32 PM GMT+02:00, Thomas Schwinge wrote: >Hi! > >OK for gcc-5-branch? OK Richard >On Wed, 29 Apr 2015 18:26:06 +0200, I wrote: >> Committed in r222583: >> >> commit df615909263269988fd9611f8d007902580829d9 >> Author: tschwinge >> Date: Wed Apr 29 16:23:26 2015 + >

Re: [nvptx offloading] Only 64-bit configurations are currently supported

Hi! OK for gcc-5-branch? On Wed, 8 Jul 2015 17:03:02 +0200, I wrote: > On Wed, 18 Feb 2015 09:50:15 +0100, I wrote: > > So far, we have concentrated only on the 64-bit x86_64 configuration; > > 32-bit has several known issues to be resolved. > > filed. > > I have co

Re: [PATCH 6/n] OpenMP 4.0 offloading infrastructure: option handling

Hi! OK for gcc-5-branch? On Wed, 29 Apr 2015 18:26:06 +0200, I wrote: > Committed in r222583: > > commit df615909263269988fd9611f8d007902580829d9 > Author: tschwinge > Date: Wed Apr 29 16:23:26 2015 + > > [PR libgomp/65099] nvptx mkoffload: pass "-m32" or "-m64" to the compiler >

[patch, nios2] stack limit checking improvements

In preparation for some other changes to the Nios II prologue code that are in my queue, I've checked in the attached patch to the stack limit checking logic. With this patch, the check generation code now takes an offset so that the check can be emitted before the SP adjustment; this allows c

[PATCH] PR fortran/66864 -- Use the right precision in FLOOR()

The attached patch fixes an issue in the constant folding of FLOOR(). In the old code, the precision of an temporary MPFR variable was set according to the requested kind of the INTEGER result, ie., the wrong precision. The patch sets the precision to the precision of the arg of FLOOR(). Regres

Re: [PATCH][C++] Fix PR65091

On 07/12/2015 01:53 PM, Paolo Carlini wrote: On 07/11/2015 09:46 PM, Paolo Carlini wrote: I'm going to ping this one too: a tad less trivial than the other one - a little explanation here or in a comment would definitely help - but certainly it looks much simpler than my own tries a while ago...

[gomp4] New test loop independent clause

Hi, The attached adds testing for the independent clause with the loop directive in Fortran. Committed to gomp-4_0-branch. Jim diff --git a/libgomp/testsuite/libgomp.oacc-fortran/kernels-independent.f90 b/libgomp/testsuite/libgomp.oacc-fortran/kernels-independent.f90 new file mode 100644 index

Re: [PATCH][C++] PR 65071

On 07/11/2015 03:40 PM, Paolo Carlini wrote: I'm going to ping this on behalf of Andrea: the patch seems almost obvious to me and so small to not immediately require a copyright assignment (not sure whether Andrea already has it on file?!?). I also double checked that it still applies cleanly and

Re: Tests for libgomp based on OpenMP Examples 4.0.2.

On Tue, Jul 14, 2015 at 08:40:50PM +0300, Maxim Blumental wrote: > The patch replaces all FP comparisons with inequalities and epsilons > in those tests for libgomp. Ok, thanks. Jakub

Re: [PING] Re: [PATCH] c/66516 - missing diagnostic on taking the address of a builtin function

On 07/14/2015 09:01 AM, Jason Merrill wrote: I wonder if it would make sense to handle this when we actually try to emit a reference to one of these functions in the back end, rather than in many places through the front end. If it's going to stay in the front end, the C and C++ front-ends shoul

Re: [PATCH g++ driver] Push -static-libstdc++ back onto the command line to allow spec substitutions to use it.

On 14 Jul 2015, at 18:24, Jason Merrill wrote: > On 06/18/2015 04:12 AM, Iain Sandoe wrote: >> The patch below pushes -static-libstdc++ onto the output command line (for >> targets without -Bstatic/dynamic) so that such specs have an opportunity to >> fire. > > Won't that produce an unrecogni

Re: Tests for libgomp based on OpenMP Examples 4.0.2.

The patch replaces all FP comparisons with inequalities and epsilons in those tests for libgomp. 2015-07-14 20:37 GMT+03:00 Maxim Blumental : > Sorry, a little misunderstanding between me and Ilya ) I'll send the > new patch now. > > 2015-07-14 20:29 GMT+03:00 Jakub Jelinek : >> On Tue, Jul 14, 2

Re: [PATCH g++ driver] Push -static-libstdc++ back onto the command line to allow spec substitutions to use it.

On 06/18/2015 04:12 AM, Iain Sandoe wrote: The patch below pushes -static-libstdc++ onto the output command line (for targets without -Bstatic/dynamic) so that such specs have an opportunity to fire. Won't that produce an unrecognized flag error from the linker? Jason

Re: [PATCH] Fix PR c++/66850 (Adding a forward declaration causes ICE on valid code)

OK. Jason

Re: [PATCH, ARM] stop changing signedness in PROMOTE_MODE

On July 14, 2015 6:49:18 PM GMT+02:00, Jim Wilson wrote: >On Tue, Jul 14, 2015 at 9:13 AM, Richard Earnshaw > wrote: >> We went through this a couple of weeks back. The backend >documentation >> for PROMOTE_MODE says: > >I disagree that this is a fully resolved issue. I see clear problems >with

Re: Committed: partial sync with src repo: bfd depends on zlib

On Tue, Jul 14, 2015 at 6:46 AM, Hans-Peter Nilsson wrote: > On Tue, 14 Jul 2015, H.J. Lu wrote: >> The shared files between gcc and binutils are way out of sync. > > Yeah I noticed, but didn't go for a full sync as that'd require > testing other builds too. Thankfully your imported src commit >

Re: [PATCH, ARM] stop changing signedness in PROMOTE_MODE

On Tue, Jul 14, 2015 at 9:13 AM, Richard Earnshaw wrote: > We went through this a couple of weeks back. The backend documentation > for PROMOTE_MODE says: I disagree that this is a fully resolved issue. I see clear problems with how the ARM port uses PROMOTE_MODE. > " For most machines, the ma

Re: [PATCH, ARM] stop changing signedness in PROMOTE_MODE

On July 14, 2015 6:13:13 PM GMT+02:00, Richard Earnshaw wrote: >On 13/07/15 16:29, Michael Matz wrote: >> Hi, >> >> On Mon, 13 Jul 2015, Richard Biener wrote: >> >>> On Fri, Jul 10, 2015 at 5:46 PM, Jim Wilson >wrote: On Tue, Jul 7, 2015 at 2:35 PM, Richard Biener wrote: > On Ju

Re: [C/C++ PATCH] PR c++/66572. Fix Wlogical-op false positive

Sorry it's taken so long to respond. On Wed, Jun 24, 2015 at 05:43:05PM +0300, Mikhail Maltsev wrote: > > That looks wrong, because with TREE_CONSTANT we'd warn in C but not in C++ > > for the following: > > > > const int a = 4; > > void > > f (void) > > { > > const int b = 4; > > static cons

Re: [PATCH, ARM] stop changing signedness in PROMOTE_MODE

On 13/07/15 16:29, Michael Matz wrote: > Hi, > > On Mon, 13 Jul 2015, Richard Biener wrote: > >> On Fri, Jul 10, 2015 at 5:46 PM, Jim Wilson wrote: >>> On Tue, Jul 7, 2015 at 2:35 PM, Richard Biener >>> wrote: On July 7, 2015 6:29:21 PM GMT+02:00, Jim Wilson wrote: > signed sub-

[committed, PATCH] Update copyright year in include

--- include/ansidecl.h | 4 +--- include/demangle.h | 5 ++--- include/dwarf2.def | 4 +--- include/dwarf2.h| 4 +--- include/dyn-string.h| 3 +-- include/fibheap.h | 3 +-- include/filenames.h | 2 +- inc

RE: [PATCH, MIPS] Fix restoration of hi/lo in MIPS64R2 interrupt handlers

Robert Suchanek writes: > > Hi Robert, > > The patch is OK, but will you please name the test something other than the > > date? > > OK. I'll change it to interrupt_handler-5.c, add a comment and commit after > approval for the new interrupt handler options. I believe this change is independent

[gomp4.1] Parsing of {use,is}_device_ptr clauses

Hi! I've committed this patch to parse/gimplify/omp lower {use,is}_device_ptr clauses which I understood are meant to act like firstprivate clause with special side-effects, but they are ignored for now during OpenMP expansion. I'm committing this now, since it still shouldn't break the tests. I

Re: [PATCH, PR testsuite/66734] Support MPX tests in lto.exp

On 07/14/2015 09:49 AM, Ilya Enkovich wrote: Hi, This patch initializes MPX paths in lto.exp to give check_effective_target_mpx a chance. Is it OK for trunk? Thanks, Ilya -- gcc/testsuite/ 2015-07-14 Ilya Enkovich PR testsuite/66734 * gcc.dg/lto/lto.exp: Initialize MPX.

[PATCH, PR testsuite/66734] Support MPX tests in lto.exp

Hi, This patch initializes MPX paths in lto.exp to give check_effective_target_mpx a chance. Is it OK for trunk? Thanks, Ilya -- gcc/testsuite/ 2015-07-14 Ilya Enkovich PR testsuite/66734 * gcc.dg/lto/lto.exp: Initialize MPX. diff --git a/gcc/testsuite/gcc.dg/lto/lto.exp

Re: [C++ Patch] Prefer error + inform to two errors in check_template_shadow

Hi, On 07/14/2015 05:07 PM, Jason Merrill wrote: On 07/13/2015 09:41 AM, Paolo Carlini wrote: +++ testsuite/g++.dg/template/crash81.C (working copy) @@ -3,6 +3,6 @@ struct A { template struct X; // { dg-error "'T' has not been declared" "T" } - // { dg-error "declaration of 'template

Re: [PATCH] Disable all target libraries if not building gcc

On Tue, Mar 17, 2015 at 5:08 PM, H.J. Lu wrote: > On Mon, Mar 16, 2015 at 8:01 PM, H.J. Lu wrote: >> On Mon, Mar 16, 2015 at 7:26 PM, H.J. Lu wrote: >>> On Mon, Mar 16, 2015 at 4:43 PM, H.J. Lu wrote: On Mon, Mar 16, 2015 at 4:41 PM, Joseph Myers wrote: > On Mon, 16 Mar 2015, H.

[committed, PATCH] Sync toplevel configure with binutils-gdb

Sync with binutils-gdb: 2015-05-13 John David Anglin * configure.ac: Disable configuration of GDB for HPUX targets. * configure: Regenerate. 2015-04-01 H.J. Lu * configure.ac: Add --with-system-zlib. * configure: Regenerated.

Re: [PATCH][RFC] Consolidate -O3 torture options

On 07/14/2015 05:58 AM, Richard Biener wrote: The following patch tries to consolidate the -O3 torture testing options in the attempt to reduce testing time while not losing coverage. It drops testing of -funroll-all-loops (which nobody should use) and retains only one non-default -O3 set of op

RE: [PATCH, MIPS] Fix restoration of hi/lo in MIPS64R2 interrupt handlers

Hi Catherine, > Hi Robert, > The patch is OK, but will you please name the test something other than the > date? OK. I'll change it to interrupt_handler-5.c, add a comment and commit after approval for the new interrupt handler options. Regards, Robert diff --git a/gcc/testsuite/gcc.target/mips

[C/C++ PATCH] Implement -Wtautological-compare (PR c++/66555, c/54979)

Code such as "if (i == i)" is hardly ever desirable, so we should be able to warn about this to prevent dumb mistakes. This ought to be easy in principle: when we see a comparison, just check for operand_equal_p of operands, and if they're the same, warn. In reality, it of course ain't that simpl

RE: [PATCH, MIPS] Support new interrupt handler options

Hi Catherine, > I'm getting build errors with the current TOT and your patch. > > The first errors that I encounter are: > gcc/config/mips/mips.c:1355:1: warning: 'mips_int_mask > mips_interrupt_mask(tree)' defined but not used [-Wunused-function] > gcc/config/mips/mips.c:1392:1: warning: 'mips_s

Re: [C++ Patch] Prefer error + inform to two errors in check_template_shadow

On 07/13/2015 09:41 AM, Paolo Carlini wrote: +++ testsuite/g++.dg/template/crash81.C (working copy) @@ -3,6 +3,6 @@ struct A { template struct X; // { dg-error "'T' has not been declared" "T" } - // { dg-error "declaration of 'template struct A::X'" "A::X" { target *-*-* } 5 } - // { d

Re: [PING] Re: [PATCH] c/66516 - missing diagnostic on taking the address of a builtin function

I wonder if it would make sense to handle this when we actually try to emit a reference to one of these functions in the back end, rather than in many places through the front end. If it's going to stay in the front end, the C and C++ front-ends should share an implementation of this function,

Re: [PATCH] PR ld/18355: --enable-shared doesn't work

On Fri, May 1, 2015 at 8:32 AM, H.J. Lu wrote: > On Wed, Apr 29, 2015 at 9:12 AM, H.J. Lu wrote: >> When bfd is configured as a shared library, we need to configure zlib >> with --enable-host-shared since zlib is used by bfd. Any comments, >> feedbacks, objections? >> >> >> H.J. >> -- >>

[PATCH] gcc: fix building w/isl-0.15

--- gcc/config.in | 6 ++ gcc/configure | 31 +++ gcc/configure.ac| 14 ++ gcc/graphite-dependences.c | 14 +++--- gcc/graphite-optimize-isl.c | 8 ++-- gcc/graphite-poly.h | 5 + 6 file

Re: [RFC, PR target/65105] Use vector instructions for scalar 64bit computations on 32bit target

Hi, Any comments on this? Thanks, Ilya 2015-06-19 16:21 GMT+03:00 Ilya Enkovich : > Hi, > > This patch tries to improve 64bit integer computations on 32bit target. This > is achieved by an additional i386 target pass which searches for all > conversion candidates and tries to transform them i

Re: [PATCH] [gomp] Recycle non-nested team if possible

Hi! On Tue, Jul 14, 2015 at 01:47:41PM +0200, Sebastian Huber wrote: > With --disable-linux-futex (without asserts) I got several failures, but > none of them is related to my patch, e.g. they are of the following type > > /tmp/ccw4RofR.o: In function `main': > data-already-3.f:(.text+0x56): unde

Re: [PATCH] Sync and use zlib.m4 with binutils-gdb

On Tue, Apr 14, 2015 at 12:11 PM, H.J. Lu wrote: > This patch syncs zlib.m4 with binutils-gdb and uses AM_ZLIB from zlib.m4 > in gcc/configure.ac. OK for trunk? > > Thanks. > > H.J. > -- > config/ > > * zlib.m4: Sync with binutils-gdb. > > gcc/ > > * Makefile.in (top_srcdir): New.

Re: [gomp4.1] depend(sink) and depend(source) parsing for C

On Tue, Jul 14, 2015 at 07:06:52AM -0700, Aldy Hernandez wrote: > On 07/13/2015 06:56 AM, Jakub Jelinek wrote: > >On Sat, Jul 11, 2015 at 11:35:36AM -0700, Aldy Hernandez wrote: > > Everything addressed except this, which I'll address as a follow-up: > > >If you want to spend time on something st

Re: [gomp4.1] depend(sink) and depend(source) parsing for C

On 07/13/2015 06:56 AM, Jakub Jelinek wrote: On Sat, Jul 11, 2015 at 11:35:36AM -0700, Aldy Hernandez wrote: Everything addressed except this, which I'll address as a follow-up: If you want to spend time on something still in the FE, it would be nice to resolve the C++ iteration var issue (i.

Re: Committed: partial sync with src repo: bfd depends on zlib

On Tue, 14 Jul 2015, H.J. Lu wrote: > The shared files between gcc and binutils are way out of sync. Yeah I noticed, but didn't go for a full sync as that'd require testing other builds too. Thankfully your imported src commit was local. > I have synced binutils with gcc. I started submitting G

Re: Tests for libgomp based on OpenMP Examples 4.0.2.

Hi! Note Richard reported that simd-8.c and simd-8.f90 tests fail on i686, can you please adjust them so that they don't perform direct comparison of floating point values, but allow some small epsilon (0.00390625 or something similar)? The last iteration multiplies 998 by 1.8, adds that to 1498.

Re: [patch] install missing rs6000 header, needed to build plugins

On Tue, Jul 14, 2015 at 5:57 AM, Matthias Klose wrote: > This installs the rs6000-cpus.def file for powerpc targets (included by > default64.h), which is needed to build plugins on powerpc64* targets. > > Ok for the trunk, and the 5 branch? Tested with a build and installation. Okay. Thanks, Dav

RE: [PATCH] MIPS: Correctly update the isa and arch_test_option_p variables after the arch dependency handling code in mips.exp

> Yeah, I agree that this doesn't really fit the model that well, > but like you say, we're stretching the logic a bit :-). When I wrote it, > the architectures formed a nice tree in which moving to leaf nodes only > added features. So in the pre-r6 days: > > # Handle dependencies between th

Re: [PATCH][AArch64] Handle -|x| case using a single csneg

On 14/07/15 13:55, Segher Boessenkool wrote: On Tue, Jul 14, 2015 at 09:18:06AM +0100, Kyrill Tkachov wrote: Before combine, you have two insns, a negation and an abs, so that is not so very strange :-) Oh, hrm, my aarch64 cross was three months old, and this now changed. Or I messed up. Sorr

Re: [PATCH, RFC] combine: Don't create insv insns unless HAVE_insv

On Mon, Jul 13, 2015 at 10:50:28PM -0600, Jeff Law wrote: > On 07/12/2015 07:56 AM, Segher Boessenkool wrote: > >Currently combine tries to make assignments to bitfields (of a register) > >whenever it can. If the target has no insv pattern, the result will not > >ever match (if the MD is sane at a

[PATCH] Add 'switch' statement to match.pd language

The following as promised adds a 'switch' statement. This way (if (A) B (if (B) C (if (C) D E))) can now be written as (switch (if (A) B) (if (B) C) (if (C) D) E) the ifs immediately nested in the switch cannot have else clauses and I reject switches that

Re: [PATCH][AArch64] Handle -|x| case using a single csneg

On Tue, Jul 14, 2015 at 09:18:06AM +0100, Kyrill Tkachov wrote: > >Before combine, you have two insns, a negation and an abs, so that is > >not so very strange :-) Oh, hrm, my aarch64 cross was three months old, and this now changed. Or I messed up. Sorry for the noise. It does look like the if_

Re: [gomp4.1] Handle linear clause modifiers in declare simd

On Tue, Jul 14, 2015 at 02:41:04PM +0300, Ilya Verbin wrote: > This caused: > > gcc/tree-vect-stmts.c: In function ‘bool vectorizable_simd_clone_call(gimple, > gimple_stmt_iterator*, gimple_statement_base**, slp_tree)’: > gcc/tree-vect-stmts.c:2810:13: error: enumeration value > ‘SIMD_CLONE_ARG_

Re: Committed: partial sync with src repo: bfd depends on zlib

On Tue, Jul 14, 2015 at 5:22 AM, Hans-Peter Nilsson wrote: > I tried the good old combined-tree approach (which some call > ubertree), and was bitten by bfd not depending on zlib *in the > gcc version of toplevel Makefile.in*. So, I did a partial sync > to include the particular commit from the b

Re: [PATCH][AArch64] Handle -|x| case using a single csneg

On Tue, Jul 14, 2015 at 07:04:13PM +0800, pins...@gmail.com wrote: > Combine knows how to use define_split without being an insn. Combine uses define_split in very different circumstances than it uses define_insn. In this case, define_split will only do anything if the "nabs" is combined from th

Re: [Patch wwwdocs] gcc-6/changes.html : Document AMD monitorx and mwaitx

On Tue, 14 Jul 2015, Kumar, Venkataramanan wrote: > Please let me know if it is good to commit. This looks good to me. Gerald

Committed: partial sync with src repo: bfd depends on zlib

I tried the good old combined-tree approach (which some call ubertree), and was bitten by bfd not depending on zlib *in the gcc version of toplevel Makefile.in*. So, I did a partial sync to include the particular commit from the binutils-gdb git and committed as obvious after a auccessful build of

[PATCH][RFC] Consolidate -O3 torture options

The following patch tries to consolidate the -O3 torture testing options in the attempt to reduce testing time while not losing coverage. It drops testing of -funroll-all-loops (which nobody should use) and retains only one non-default -O3 set of options - namely -O3 plus those flags that would b

Re: [PATCH] [gomp] Recycle non-nested team if possible

On 14/07/15 12:44, Jakub Jelinek wrote: On Tue, Jul 14, 2015 at 11:28:18AM +0200, Sebastian Huber wrote: >If I destroy the barrier of the last team, and then initialize it later in >gomp_new_team() > >+static inline struct gomp_team * >+get_last_team (unsigned nthreads) >+{ >+ struct gomp_thr

Re: [gomp4.1] Handle linear clause modifiers in declare simd

On Wed, Jul 01, 2015 at 12:55:38 +0200, Jakub Jelinek wrote: > * cgraph.h (enum cgraph_simd_clone_arg_type): Add > SIMD_CLONE_ARG_TYPE_LINEAR_REF_CONSTANT_STEP, > SIMD_CLONE_ARG_TYPE_LINEAR_UVAL_CONSTANT_STEP, > and SIMD_CLONE_ARG_TYPE_LINEAR_VAL_CONSTANT_STEP. > (stru

Re: [PATCH][4/n] Remove GENERIC stmt combining from SCCVN

On Tue, 14 Jul 2015, Richard Biener wrote: > On Mon, 13 Jul 2015, Jeff Law wrote: > > > > 2015-07-13 Richard Biener > > > > > > * tree-ssa-dom.c (record_temporary_equivalences): Merge > > > wideing type conversion case from > > > record_equivalences_from_incoming_edge > > > and use reco

[PATCH] Fix PR66863

The following fixes PR66863. Bootstrapped and tested on x86_64-unknown-linux-gnu. Richard. 2015-07-14 Richard Biener PR tree-optimization/66863 * tree-vrp.c (register_edge_assert_for_2): Properly restrict what we record for conversion use stmt lhs inequalities.

Re: [Bug fortran/52846] [F2008] Support submodules - part 2/3

PS I forgot to add that I have added a submodule file cleanup procedure to fortran-modules.exp and cleanup lines to each of the submodule testcases, such as: ! { dg-final { cleanup-submodules "color_points_a" } } Paul On 14 July 2015 at 13:10, Paul Richard Thomas wrote: > Dear All, > > Reinhold

Re: [PATCH][AArch64] Handle -|x| case using a single csneg

On 14/07/15 11:40, Andrew Pinski wrote: On Tue, Jul 14, 2015 at 3:13 AM, Kyrill Tkachov wrote: On 14/07/15 11:06, Andrew Pinski wrote: On Tue, Jul 14, 2015 at 1:18 AM, Kyrill Tkachov wrote: Hi Segher, On 14/07/15 01:38, Segher Boessenkool wrote: On Mon, Jul 13, 2015 at 10:48:19AM +0100, K

[Bug fortran/52846] [F2008] Support submodules - part 2/3

Dear All, Reinhold Bader has pointed out the naming the submodule files after the submodule name and using .mod as the extension can potentially lead to clashes. Therefore, I have written a patch to change gfortran to follow the naming convention of another leading brand: submodule filename = mod

Re: [PATCH] Unswitching outer loops.

On Fri, Jul 10, 2015 at 12:02 PM, Yuri Rumyantsev wrote: > Hi All, > > Here is presented simple transformation which tries to hoist out of > outer-loop a check on zero trip count for inner-loop. This is very > restricted transformation since it accepts outer-loops with very > simple cfg, as for ex

Re: [PATCH, gomp4] Propagate independent clause for OpenACC kernels pass

Hi! On Tue, 14 Jul 2015 11:36:24 +0200, I wrote: > On Tue, 14 Jul 2015 13:46:04 +0800, Chung-Lin Tang > wrote: > > this patch provides a 'bool independent' field in struct loop, which > > will be switched on by an "independent" clause in a #pragma acc loop > > directive. > > Thanks! > > > Th

Re: [PATCH][AArch64] Handle -|x| case using a single csneg

> On Jul 13, 2015, at 5:48 PM, Kyrill Tkachov wrote: > > Hi all, > > For the testcase in the patch we were generating an extra neg instruction: >cmp w0, wzr >csneg w0, w0, w0, ge >neg w0, w0 >ret > > instead of the optimal: >cmp w0, wzr

Re: [PATCH] Yet another simple fix to enhance outer-loop vectorization.

On Wed, Jun 17, 2015 at 6:35 PM, Yuri Rumyantsev wrote: > Richard, > > I attached updated patch. > You asked me to explain why I did changes for renaming. > If we do not change PHI arguments for inner loop header we can get the > following IR: > source outer loop: > : outer-loop header > # i_4

Re: [PATCH] Yet another simple fix to enhance outer-loop vectorization.

On Mon, Jun 29, 2015 at 6:15 PM, Yuri Rumyantsev wrote: > Hi All, > > Here is updated patch containing missed change in > slpeel_tree_peel_loop_to_edge which prevents renaming of exit PHI uses > in inner loop. Ok. Thanks, Richard. > ChangeLog: > 2015-06-29 Yuri Rumyantsev > > * tree-vect-loo

Re: [PATCH 4/5] Downgrade value_expr_for_decl to non-cache

On Tue, Jul 14, 2015 at 12:45 PM, Richard Biener wrote: > On Sun, Jul 12, 2015 at 5:53 PM, Tom de Vries wrote: >> On 12/07/15 17:45, Tom de Vries wrote: >>> >>> Hi, >>> >>> this patch series implements the forbidding of multi-step garbage >>> collection liveness dependencies between caches. >>> >

Re: conditional lim

On Mon, Jun 29, 2015 at 4:21 PM, Evgeniya Maenkova wrote: > On Mon, Jun 29, 2015 at 5:10 PM, Richard Biener > wrote: >> On Tue, Jun 9, 2015 at 10:11 PM, Evgeniya Maenkova >> wrote: >>> On Tue, Jun 9, 2015 at 3:46 PM, Richard Biener >>> wrote: On Fri, May 29, 2015 at 3:14 PM, Evgeniya Maenk

  1   2   >