Re: [PATCH 0/2] Final cleanup in move to ISL

On 08/27/2015 12:14 AM, Sebastian Pop wrote: Hi, Richi suggested at the Cauldron that it would be good to have graphite more automatic and with fewer flags. The first patch removes the -funroll-and-jam pass that does not seem very stable or useful for now. The second patch removes the other -f

RE: [Patch,tree-optimization]: Add new path Splitting pass on tree ssa representation

-Original Message- From: Jeff Law [mailto:l...@redhat.com] Sent: Thursday, August 20, 2015 9:19 PM To: Ajit Kumar Agarwal; Richard Biener Cc: GCC Patches; Vinod Kathail; Shail Aditya Gupta; Vidhumouli Hunsigida; Nagaraju Mekala Subject: Re: [Patch,tree-optimization]: Add new path Splitt

[PATCH] fix --with-cpu for sh targets

A missing * in the pattern for sh targets prevents the --with-cpu configure option from being accepted for certain targets (e.g. ones with explicit endianness, like sh2eb). The latest config.sub should also be pulled from upstream since it has a fix for related issues. Rich --- gcc-5.2.0.orig/gcc

Re: [PATCH] Fix and simplify (Re: Fix libbacktrace -fPIC breakage from "Use libbacktrace in libgfortran")

"Ulrich Weigand" writes: > I've verified that this works on x86_64: the resulting > libgfortran.so uses the -fPIC version of the libbacktrace > object, while libgfortran.a uses the non-PIC versions. > > On SPU, libtool will now automatically only generate the > non-PIC versions since the target d

[c++-delayed-folding] fold_simple

Why does fold_simple fold so many patterns? I thought we wanted something that would just fold conversions and negations of constant values. Jason

Re: C++ delayed folding branch review

On 08/24/2015 03:15 AM, Kai Tietz wrote: 2015-08-03 17:39 GMT+02:00 Jason Merrill : On 08/03/2015 05:42 AM, Kai Tietz wrote: 2015-08-03 5:49 GMT+02:00 Jason Merrill : On 07/31/2015 05:54 PM, Kai Tietz wrote: The "STRIP_NOPS-requirement in 'reduced_constant_expression_p'" I could remove, but

[Patch, libstdc++/67362] Fix non-special character for POSIX basic syntax in regex

Bootstrapped and tested on x86_64-pc-linux-gnu. Thanks! -- Regards, Tim Shen commit e134e1a835ad15900686351cade36774593b91ea Author: Tim Shen Date: Wed Aug 26 17:51:29 2015 -0700 PR libstdc++/67362 * include/bits/regex_scanner.tcc (_Scanner<>::_M_scan_normal): Always

Re: [Patch, libstdc++/67362] Fix non-special character for POSIX basic syntax in regex

On Wed, Aug 26, 2015 at 6:41 PM, Tim Shen wrote: > Bootstrapped and tested on x86_64-pc-linux-gnu. Also plan to backport to 4.9 and 5. -- Regards, Tim Shen

[gomp4, wip] remove references to ganglocal shared memory inside gcc

This patch strips out all of the references to ganglocal memory in gcc. Unfortunately, the runtime api still takes a shared memory parameter, so I haven't made any changes there yet. Perhaps we could still keep the shared memory argument to GOACC_parallel, but remove all of the support for gangloca

[gomp4] initialize worker reduction locks

This patch teaches omplow how to emit function calls to IFN_GOACC_LOCK_INIT so that the worker mutex has a proper initial value. On nvptx targets, shared memory isn't initialized (and that's where the lock is located for OpenACC workers), so this makes it explicit. Nathan added the internal functio

Re: [gomp4] lowering OpenACC reductions

On 08/21/2015 02:00 PM, Cesar Philippidis wrote: > This patch teaches omplower how to utilize the new OpenACC reduction > framework described in Nathan's document, which was posted here > . Here is the > infrastructure patch >

Go patch committed: Don't crash on invalid builtin calls

This patch by Chris Manghane fixes the Go compiler to not crash when it sees invalid builtin calls. This fixes https://golang.org/issue/11544 . Bootstrapped and ran Go testsuite on x86_64-unknown-linux-gnu. Committed to mainline. Ian Index: gcc/go/gofrontend/MERGE ==

[PATCH 2/2] remove -floop-* flags

--- gcc/Makefile.in|2 - gcc/common.opt | 16 +- gcc/doc/invoke.texi| 108 +- gcc/graphite-blocking.c| 270 - gcc/graphite-interchange.c

[PATCH 1/2] remove -floop-unroll-and-jam

--- gcc/common.opt | 4 +- gcc/doc/invoke.texi | 8 +- gcc/graphite-isl-ast-to-gimple.c | 102 +- gcc/graphite-optimize-isl.c | 179 --- gcc/graphite-poly.c | 3 +- gcc/graphite-poly.h

[PATCH 0/2] Final cleanup in move to ISL

Hi, Richi suggested at the Cauldron that it would be good to have graphite more automatic and with fewer flags. The first patch removes the -funroll-and-jam pass that does not seem very stable or useful for now. The second patch removes the other -floop-* flags that were part of the old graphite

Re: [PATCH] [AVX512F] Add scatter support for vectorizer

On Wed, Aug 26, 2015 at 7:39 PM, Petr Murzin wrote: > On Wed, Aug 26, 2015 at 10:41 AM, Richard Biener > wrote: >> @@ -3763,32 +3776,46 @@ again: >>if (vf > *min_vf) >> *min_vf = vf; >> >> - if (gather) >> + if (gatherscatter != SG_NONE) >> { >> tree of

[v3 patch] try_emplace and insert_or_assign for Debug Mode.

These new members need to be defined in Debug Mode, because the iterators passed in as hints and returned as results need to be safe iterators. No new tests, because we already have tests for these members, and they're failing in debug mode. Tested powerpc64le-linux, committed to trunk. commit

Re: [libvtv] Fix formatting errors

On 08/26/2015 01:50 PM, Caroline Tice wrote: As far as I know vtv is working just fine...is there something I don't know about? I'm not aware of anything that isn't working, but I'm also not aware of vtv in widespread use, typical performance hit experienced, etc. jeff

[patch] libstdc++/66902 Make _S_debug_messages static.

This patch removes a public symbol from the .so, which is generally a bad thing, but there should be no users of this anywhere (it's never declared in any public header). For targets using symbol versioning this isn't exported at all, as it isn't in the linker script, so this really just makes ot

[Patch, fortran] F2008 - implement pointer function assignment

Dear All, The attached patch more or less implements the assignment of expressions to the result of a pointer function. To wit: my_ptr_fcn (arg1, arg2...) = expr arg1 would usually be the target, pointed to by the function. The patch parses these statements and resolves them into: temp_ptr => m

Fwd: [libvtv] Fix formatting errors

-- Forwarded message -- From: Caroline Tice Date: Wed, Aug 26, 2015 at 12:50 PM Subject: Re: [libvtv] Fix formatting errors To: Jeff Law Cc: Rainer Orth , GCC Patches As far as I know vtv is working just fine...is there something I don't know about? -- Caroline cmt...@google.c

Re: [libvtv] Fix formatting errors

On 08/26/2015 07:30 AM, Rainer Orth wrote: While looking at libvtv for the Solaris port, I noticed all sorts of GNU Coding Standard violations: * ChangeLog entries attributed to the committer instead of the author and with misformatted PR references, entries only giving a vague rational in

Go patch committed: don't crash on invalid numeric type

This patch by Chris Manghane fixes a compiler crash on an invalid program when the compiler tries to set a numeric constant to an invalid type. This fixes https://golang.org/issue/11537. Bootstrapped and ran Go testsuite on x86_64-unknown-linux-gnu. Committed to mainline. Ian Index: gcc/go/gofron

Re: [PATCH 2/5] completely_scalarize arrays as well as records

On August 26, 2015 6:08:55 PM GMT+02:00, Alan Lawrence wrote: >Richard Biener wrote: > One extra question is does the way we limit total scalarization >work >>> well for arrays? I suppose we have either sth like the maximum size of >an aggregate we scalarize or the maximum number

Re: [PATCH] [AVX512F] Add scatter support for vectorizer

On Wed, Aug 26, 2015 at 10:41 AM, Richard Biener wrote: > @@ -3763,32 +3776,46 @@ again: >if (vf > *min_vf) > *min_vf = vf; > > - if (gather) > + if (gatherscatter != SG_NONE) > { > tree off; > + if (vect_check_gather_scatter (stmt, loop_vinfo, N

Re: [PATCH] 2015-07-31 Benedikt Huber Philipp Tomsich

ping [PATCH v4][aarch64] Implemented reciprocal square root (rsqrt) estimation in -ffast-math https://gcc.gnu.org/ml/gcc-patches/2015-07/msg02698.html > On 31 Jul 2015, at 19:05, Benedikt Huber > wrote: > > * config/aarch64/aarch64-builtins.c: Builtins for rsqrt and > rsqrtf. >

Re: [Scalar masks 2/x] Use bool masks in if-conversion

On 08/26/2015 05:13 AM, Ilya Enkovich wrote: 2015-08-26 0:42 GMT+03:00 Jeff Law : On 08/21/2015 04:49 AM, Ilya Enkovich wrote: I want a work with bitmasks to be expressed in a natural way using regular integer operations. Currently all masks manipulations are emulated via vector statements (m

Re: [PATCH 2/5] completely_scalarize arrays as well as records

Richard Biener wrote: One extra question is does the way we limit total scalarization work well for arrays? I suppose we have either sth like the maximum size of an aggregate we scalarize or the maximum number of component accesses we create? Only the former and that would be kept intact.

Re: [Scalar masks 2/x] Use bool masks in if-conversion

2015-08-26 17:56 GMT+03:00 Richard Biener : > On Wed, Aug 26, 2015 at 4:38 PM, Ilya Enkovich wrote: >> 2015-08-26 16:02 GMT+03:00 Richard Biener : >>> On Fri, Aug 21, 2015 at 2:17 PM, Ilya Enkovich >>> wrote: 2015-08-21 14:00 GMT+03:00 Richard Biener : > > Hmm, I don't see how vecto

Re: [RFC 4/5] Handle constant-pool entries

Jeff Law wrote: The question I have is why this differs from the effects of patch #5. That would seem to indicate that there's things we're not getting into the candidate tables with this approach?!? I'll answer this first, as I think (Richard and) Martin have identified enough other issues

[PATCH] Fix and simplify (Re: Fix libbacktrace -fPIC breakage from "Use libbacktrace in libgfortran")

Hans-Peter Nilsson wrote: > I don't feel very confused, but I understand you've investigated > things down to a point where we can conclude that libtool can't > do what SPU needs without also at least fiddling with > compilation options. Well, looks like I was confused after all. I missed one ex

Re: [Scalar masks 2/x] Use bool masks in if-conversion

On Wed, Aug 26, 2015 at 04:56:23PM +0200, Richard Biener wrote: > >> How do you declare those? > > > > Something like this: > > > > #pragma omp declare simd inbranch > > int foo(int*); > > The 'inbranch' is the thing that matters? And all of foo is then > implicitely predicated? If it is #pragma

Re: [Scalar masks 2/x] Use bool masks in if-conversion

On Wed, Aug 26, 2015 at 4:38 PM, Ilya Enkovich wrote: > 2015-08-26 16:02 GMT+03:00 Richard Biener : >> On Fri, Aug 21, 2015 at 2:17 PM, Ilya Enkovich >> wrote: >>> 2015-08-21 14:00 GMT+03:00 Richard Biener : Hmm, I don't see how vector masks are more difficult to operate with. >>> >>>

[PATCH] s390: Add emit_barrier() after trap.

^_^ ^_^ -- Dominik Vogt IBM Germany gcc/ChangeLog * config/s390/s390.c (s390_emit_prologue): Add emit_barrier() after trap to fix ICE. gcc/testsuite/ChangeLog * gcc.target/s390/20150826-1.c: New test. >From ec6b88cd51234d138bd559271def086156fcae07 Mon Sep 17 00:00:00 2001 F

[patch] libstdc++/64351 Ensure std::generate_canonical doesn't return 1.

Ed posted this patch to https://gcc.gnu.org/PR64351 in January, I've tested it and am committing it to trunk with a test. commit 45f154a5f9172a17f6226b99b41cb9c0bd8d15ec Author: Jonathan Wakely Date: Wed Aug 26 12:53:08 2015 +0100 Ensure std::generate_canonical doesn't return 1.

Re: [PATCH][4/N] Introduce new inline functions for GET_MODE_UNIT_SIZE and GET_MODE_UNIT_PRECISION

On 26 Aug 2015, at 23:27, Oleg Endo wrote: > > On 19 Aug 2015, at 22:35, Jeff Law wrote: > >> On 08/19/2015 06:29 AM, David Sherwood wrote: I asked Richard S. to give this a once-over which he did. However, he technically can't approve due to the way his maintainership position was

[v3 patch] Only set std::enable_shared_from_this member once.

This adds a check to weak_ptr::_M_assign() so that calling __enable_shared_from_this_helper twice with the same pointer won't change which shared_ptr object the weak_ptr shares ownership with. On the lib reflector Peter Dimov convinced me that the boost::enable_shared_from_this behaviour is prefe

Re: [Scalar masks 2/x] Use bool masks in if-conversion

2015-08-26 16:02 GMT+03:00 Richard Biener : > On Fri, Aug 21, 2015 at 2:17 PM, Ilya Enkovich wrote: >> 2015-08-21 14:00 GMT+03:00 Richard Biener : >>> >>> Hmm, I don't see how vector masks are more difficult to operate with. >> >> There are just no instructions for that but you have to pretend you

Re: [libvtv] Update copyrights

On Wed, 26 Aug 2015, Rainer Orth wrote: > While working on the Solaris libvtv port, I noticed that many of the > libvtv copyright years hadn't been updated, were misformtted, or both. > It turns out that libvtv isn't listed in contrib/update-copyright.py at > all. This patch fixes this and includ

Re: [Scalar masks 2/x] Use bool masks in if-conversion

On Wed, Aug 26, 2015 at 3:35 PM, Jakub Jelinek wrote: > On Wed, Aug 26, 2015 at 03:21:52PM +0200, Richard Biener wrote: >> On Wed, Aug 26, 2015 at 3:16 PM, Jakub Jelinek wrote: >> > On Wed, Aug 26, 2015 at 03:02:02PM +0200, Richard Biener wrote: >> >> > AVX-512 is such target. Current representat

[PATCH][AArch64 array_mode 8/8] Add d-registers to TARGET_ARRAY_MODE_SUPPORTED_P

This adds an AARCH64_VALID_SIMD_DREG_MODE exactly paralleling the existing ...QREG... macro, and as a driveby fixes mode->(MODE) in the latter. The new test now compiles (at -O3) to: test_1: add v1.2s, v1.2s, v5.2s add v2.2s, v2.2s, v6.2s add v3.2s, v3.2s, v7.2

Re: [PATCH][4/N] Introduce new inline functions for GET_MODE_UNIT_SIZE and GET_MODE_UNIT_PRECISION

On 19 Aug 2015, at 22:35, Jeff Law wrote: > On 08/19/2015 06:29 AM, David Sherwood wrote: >>> I asked Richard S. to give this a once-over which he did. However, he >>> technically can't approve due to the way his maintainership position was >>> worded. >>> >>> The one request would be a functi

Re: [gomp4] teach the tracer pass to ignore more blocks for OpenACC

On 08/26/15 09:57, Cesar Philippidis wrote: I hit a problem in on one of my reduction test cases where the GOACC_JOIN was getting cloned. Nvptx requires FORK and JOIN to be single-entry, single-exit regions, or some form of thread divergence may occur. When that happens, we cannot use the shfl in

Re: [RFC 4/5] Handle constant-pool entries

Hi, On Tue, Aug 25, 2015 at 12:06:16PM +0100, Alan Lawrence wrote: > This makes SRA replace loads of records/arrays from constant pool entries, > with elementwise assignments of the constant values, hence, overcoming the > fundamental problem in PR/63679. > > As a first pass, the approach I took

[AArch64/testsuite] Add more TLS local executable testcases

This patch cover tlsle tiny model tests, tls size truncation for tiny & small model included also. All testcases pass native test. OK for trunk? 2015-08-26 Jiong Wang gcc/testsuite/ * gcc.target/aarch64/tlsle12_tiny_1.c: New testcase for tiny model. * gcc.target/aarch64/tlsle24_tiny_1.c

Re: [PATCH] PR66870 PowerPC64 Enable gold linker with split stack

I am working on a new patch to address some of the previous concerns and plan to post it soon after some final testing. On 08/25/2015 05:51 PM, Ian Lance Taylor wrote: On Tue, Aug 18, 2015 at 1:36 PM, Lynn A. Boger wrote: libgo/ PR target/66870 configure.ac: When gccgo for b

[gomp4] teach the tracer pass to ignore more blocks for OpenACC

I hit a problem in on one of my reduction test cases where the GOACC_JOIN was getting cloned. Nvptx requires FORK and JOIN to be single-entry, single-exit regions, or some form of thread divergence may occur. When that happens, we cannot use the shfl instruction for reductions or broadcasting (if t

Re: [PATCH], PowerPC IEEE 128-bit patch #6

On Fri, Aug 14, 2015 at 11:47 AM, Michael Meissner wrote: > This is patch #6: > > 2015-08-13 Michael Meissner > > * config/rs6000/rs6000-protos.h (rs6000_expand_float128_convert): > Add declaration. > > * config/rs6000/rs6000.c (rs6000_emit_le_vsx_store): Fix a >

[PATCH][AArch64 array_mode 7/8] Combine the expanders using VSTRUCT:nregs

The previous patches leave ld[234]_lane, st[234]_lane, and ld[234]r expanders all nearly identical, so we can easily parameterize across the number of lanes and combine them. For the ld_lane pattern, I switched from the VCONQ attribute to just using the MODE attribute, this is identical for all

[PATCH][AArch64 array_mode 5/8] Remove V_FOUR_ELEM, again using BLKmode + set_mem_size.

This removes V_FOUR_ELEM in the same way that patch 3 removed V_THREE_ELEM, again using BLKmode + set_mem_size. (This makes the four-lane expanders very similar to the three-lane expanders, and they will be combined in patch 7.) bootstrapped and check-gcc on aarch64-none-linux-gnu gcc/ChangeLog:

[PATCH][AArch64 array_mode 3/8] Stop using EImode in aarch64-simd.md and iterators.md

The V_THREE_ELEM attribute used BLKmode for most sizes, but occasionally EImode. This patch changes to BLKmode in all cases, explicitly setting memory size (thus, preserving size for the cases that were EImode, and setting size for the first time for cases that were already BLKmode). The patterns

[PATCH][AArch64 array_mode 6/8] Remove V_TWO_ELEM, again using BLKmode + set_mem_size.

Same logic as previous; this makes the 2-, 3-, and 4-lane expanders all follow the same pattern. bootstrapped and check-gcc on aarch64-none-linux-gnu. gcc/ChangeLog: * config/aarch64/aarch64-simd.md (aarch64_simd_ld2r, aarch64_vec_load_lanesoi_lane, aarch64_vec_store_lan

[PATCH][AArch64 array_mode 2/8] Remove VSTRUCT_DREG, use BLKmode for d-reg aarch64_st/ld expands

aarch64_st and aarch64_ld expanders back onto 12 insns aarch64_{ld,st}{2,3,4}_dreg (for VD and DX modes), using the VSTRUCT_DREG iterator over TI/EI/OI modes to represent the block of memory transferred. Instead, use BLKmode for all memory transfers, explicitly setting mem_size. Bootstrapped and c

[PATCH][AArch64 array_mode 4/8] Remove EImode

This removes EImode from the (AArch64) compiler, and all mention of or support for it. bootstrapped and check-gcc on aarch64-none-linux-gnu gcc/ChangeLog: * config/aarch64/aarch64.c (aarch64_simd_attr_length_rglist): Update comment. * config/aarch64/aarch64-builtins.c (e

[PATCH][AArch64 array_mode 1/8] Rename vec_store_lanes_lane to aarch64_vec_store_lanes_lane

vec_store_lanes{oi,ci,xi}_lane are not standard pattern names, so using them in aarch64-simd.md is misleading. This adds an aarch64_ prefix to those pattern names, paralleling aarch64_vec_load_lanes_lane. bootstrapped and check-gcc on aarch64-none-linux-gnu gcc/ChangeLog: * config/aarc

[PATCH][AArch64 0/8] Add D-registers to TARGET_ARRAY_MODE_SUPPORTED_P

The end goal of this series of patches is to enable 64bit vector modes for TARGET_ARRAY_MODE_SUPPORTED_P, achieved in the last patch. At present, doing so causes ICEs with illegal subregs (e.g. returning the middle bits from a large int mode covering 3 vectors); the patchset avoids these by first r

Re: [Scalar masks 2/x] Use bool masks in if-conversion

On Wed, Aug 26, 2015 at 03:21:52PM +0200, Richard Biener wrote: > On Wed, Aug 26, 2015 at 3:16 PM, Jakub Jelinek wrote: > > On Wed, Aug 26, 2015 at 03:02:02PM +0200, Richard Biener wrote: > >> > AVX-512 is such target. Current representation forces multiple scalar > >> > mask -> vector mask and ba

[libvtv] Fix formatting errors

While looking at libvtv for the Solaris port, I noticed all sorts of GNU Coding Standard violations: * ChangeLog entries attributed to the committer instead of the author and with misformatted PR references, entries only giving a vague rational instead of what changed * overlong lines * tons

Re: [gomp4.1] comment some stuff

On Tue, Aug 25, 2015 at 10:35:37AM -0700, Aldy Hernandez wrote: > diff --git a/libgomp/env.c b/libgomp/env.c > index 65a6851..0569521 100644 > --- a/libgomp/env.c > +++ b/libgomp/env.c > @@ -69,7 +69,7 @@ struct gomp_task_icv gomp_global_icv = { > > unsigned long gomp_max_active_levels_var = INT

Re: [Scalar masks 2/x] Use bool masks in if-conversion

On Wed, Aug 26, 2015 at 3:16 PM, Jakub Jelinek wrote: > On Wed, Aug 26, 2015 at 03:02:02PM +0200, Richard Biener wrote: >> > AVX-512 is such target. Current representation forces multiple scalar >> > mask -> vector mask and back transformations which are artificially >> > introduced by current boo

Re: [Scalar masks 2/x] Use bool masks in if-conversion

On Wed, Aug 26, 2015 at 03:02:02PM +0200, Richard Biener wrote: > > AVX-512 is such target. Current representation forces multiple scalar > > mask -> vector mask and back transformations which are artificially > > introduced by current bool patterns and are hard to optimize out. > > I dislike the

Re: [PATCH], PowerPC IEEE 128-bit patch #5

On Tue, Aug 25, 2015 at 7:20 PM, Michael Meissner wrote: > Here is the revised patch. Is it ok to install? > > 2015-08-25 Michael Meissner > > * config/rs6000/predicates.md (int_reg_operand_not_pseudo): New > predicate for only GPR hard registers. > > * config/rs6000/rs

Re: [Scalar masks 2/x] Use bool masks in if-conversion

On Wed, Aug 26, 2015 at 1:13 PM, Ilya Enkovich wrote: > 2015-08-26 0:42 GMT+03:00 Jeff Law : >> On 08/21/2015 04:49 AM, Ilya Enkovich wrote: >>> >>> >>> I want a work with bitmasks to be expressed in a natural way using >>> regular integer operations. Currently all masks manipulations are >>> emul

Re: [Scalar masks 2/x] Use bool masks in if-conversion

On Fri, Aug 21, 2015 at 2:17 PM, Ilya Enkovich wrote: > 2015-08-21 14:00 GMT+03:00 Richard Biener : >> On Fri, Aug 21, 2015 at 12:49 PM, Ilya Enkovich >> wrote: >>> 2015-08-21 11:15 GMT+03:00 Richard Biener : On Thu, Aug 20, 2015 at 8:46 PM, Jeff Law wrote: > On 08/17/2015 10:25 AM, Il

Re: [PATCH][AARCH64]Fix for branch offsets over 1 MiB

On 25 August 2015 at 14:12, Andre Vieira wrote: > gcc/ChangeLog: > 2015-08-07 Ramana Radhakrishnan > Andre Vieira > > * config/aarch64/aarch64.md (*condjump): Handle functions > 1 Mib. > (*cb1): Likewise. > (*tb1): Likewise. > (*cb1): Likewise. >

[libgfortran,committed] Fix SHAPE intrinsic with KIND values 1 and 2

Attached patch fixes the SHAPE intrinsic with option argument KIND values of 1 and 2. While we already accept and emit code for SHAPE with KIND values, the runtime versions with integer kinds 1 and 2 are missing (while values of 4, 8 and 10 are present). The patch adds the necessary generated f

Re: [AArch64][TLSLE][1/3] Add the option "-mtls-size" for AArch64

On 25 August 2015 at 15:15, Jiong Wang wrote: > 2015-08-25 Jiong Wang > > gcc/ > * config/aarch64/aarch64.opt (mtls-size): New entry. > * config/aarch64/aarch64.c (initialize_aarch64_tls_size): New function. > (aarch64_override_options_internal): Call initialize_aarch64_tls_size. > * d

Re: Fix libbacktrace -fPIC breakage from "Use libbacktrace in libgfortran"

> From: Ulrich Weigand > Date: Wed, 26 Aug 2015 13:45:35 +0200 > Hans-Peter Nilsson wrote: > > > From: Ulrich Weigand > > > Date: Tue, 25 Aug 2015 19:45:06 +0200 > > > > > However, neither works for the SPU, because in both cases libtool > > > will only do the test whether the target supports t

[gomp4] loop partition optimization

I've committed this patch, which implements a simple partioned execution optimization. A loop over both worker and vector dimensions is emits separate FORK and JOIN markers for the two dimensions -- there may be reduction pieces between them, as Cesar will shortly be committing. However, if

[libvtv] Update copyrights

While working on the Solaris libvtv port, I noticed that many of the libvtv copyright years hadn't been updated, were misformtted, or both. It turns out that libvtv isn't listed in contrib/update-copyright.py at all. This patch fixes this and includes the result of running update-copyright.py --th

Re: Fix libbacktrace -fPIC breakage from "Use libbacktrace in libgfortran"

Hans-Peter Nilsson wrote: > > From: Ulrich Weigand > > Date: Tue, 25 Aug 2015 19:45:06 +0200 > > > However, neither works for the SPU, because in both cases libtool > > will only do the test whether the target supports the -fPIC option. > > It will not test whether the target supports dynamic lib

Re: [PATCH] Remove reference to undefined documentation node.

On Wed, Aug 26, 2015 at 11:05:09AM +0100, Dominik Vogt wrote: > This patch removes a menu entry that points to an undefined node > in the documentation. The faulty entry has been introduced with > git commit id 3aabc45f2, subversion id 138bc75d-0d04-0410-96. It > looks like the entry is a remnant

[PATCH][3/n] dwarf2out refactoring for early (LTO) debug

The following fixes a GC issue I run into when doing prune_unused_types_prune early. The issue is that the DIE struct has a chain_circular marked field (die_sib) which cannot tolerate spurious extra entries from old removed entries into the circular chain. Otherwise we fail to properly mark par

[build] Use __cxa_atexit on Solaris 12+

Solaris 12 introduced __cxa_atexit in libc. The following patch makes use of it, and also removes the strange failures seen with gld reported in PR c++/51923. Bootstrapped without regressions on i386-pc-solaris2.1[12] and sparc-sun-solaris2.1[12], will installl on mainline. Will backport to the

[boehm-gc] Avoid unstructured procfs on Solaris

boehm-gc doesn't currently build on Solaris 12 since that release finally removed the old unstructured /proc, thus the PIOCOPENPD ioctl. This is already mentioned in the Solaris 11 EOF list: http://www.oracle.com/technetwork/systems/end-of-notices/eonsolaris11-392732.html Since the repla

[libgo] Use stat_atim.go on Solaris 12+

Solaris 12 changes the stat_[amc]tim members of struct stat from timestruc_t to timespec_t for XPG7 compatiblity, thus breaking the libgo build. The following patch checks for this change and uses the common stat_atim.go if appropriate. Btw., I noticed that go/os/stat_atim.go and stat_dragonfly.g

Re: [Scalar masks 2/x] Use bool masks in if-conversion

2015-08-26 0:42 GMT+03:00 Jeff Law : > On 08/21/2015 04:49 AM, Ilya Enkovich wrote: >> >> >> I want a work with bitmasks to be expressed in a natural way using >> regular integer operations. Currently all masks manipulations are >> emulated via vector statements (mostly using a bunch of vec_cond).

Re: [Scalar masks 2/x] Use bool masks in if-conversion

2015-08-26 0:26 GMT+03:00 Jeff Law : > On 08/21/2015 06:17 AM, Ilya Enkovich wrote: >>> >>> >>> Hmm, I don't see how vector masks are more difficult to operate with. >> >> >> There are just no instructions for that but you have to pretend you >> have to get code vectorized. >> >>> Also accordi

Re: [PATCH][1/n] dwarf2out refactoring for early (LTO) debug

On Wed, 19 Aug 2015, Richard Biener wrote: > On Tue, 18 Aug 2015, Aldy Hernandez wrote: > > > On 08/18/2015 07:20 AM, Richard Biener wrote: > > > > > > This starts a series of patches (still in development) to refactor > > > dwarf2out.c to better cope with early debug (and LTO debug). > > > > A

[PATCH] Remove reference to undefined documentation node.

This patch removes a menu entry that points to an undefined node in the documentation. The faulty entry has been introduced with git commit id 3aabc45f2, subversion id 138bc75d-0d04-0410-96. It looks like the entry is a remnant of an earlier version of the documentation introduced with that chang

Re: [PATCH 2/5] completely_scalarize arrays as well as records

On August 26, 2015 11:30:26 AM GMT+02:00, Martin Jambor wrote: >Hi, > >On Wed, Aug 26, 2015 at 09:07:33AM +0200, Richard Biener wrote: >> On Tue, Aug 25, 2015 at 11:44 PM, Jeff Law wrote: >> > On 08/25/2015 03:42 PM, Martin Jambor wrote: >> >> >> >> Hi, >> >> >> >> On Tue, Aug 25, 2015 at 12:06:1

Re: [PATCH 2/5] completely_scalarize arrays as well as records

Hi, On Wed, Aug 26, 2015 at 09:07:33AM +0200, Richard Biener wrote: > On Tue, Aug 25, 2015 at 11:44 PM, Jeff Law wrote: > > On 08/25/2015 03:42 PM, Martin Jambor wrote: > >> > >> Hi, > >> > >> On Tue, Aug 25, 2015 at 12:06:14PM +0100, Alan Lawrence wrote: > >>> > >>> This changes the completely_s

RE: [PATCH] MIPS: If a test in the MIPS testsuite requires standard library support check the sysroot supports the required test options.

Moore, Catherine writes: > > The recent changes to the MIPS GCC Linux sysroot > > (https://gcc.gnu.org/ml/gcc-patches/2015-06/msg01014.html) have meant > > that the include directory is now not global and is provided only for > > each multi-lib configuration. This means that for any test in the >

Re: [PATCH][ARM]Tighten the conditions for arm_movw, arm_movt

> > I have tested that, arm-none-linux-gnueabi bootstraps Okay on trunk code. JFTR, this is ok to backport to gcc-5 in case there are no regressions. regards Ramana > >> >> Thanks, >> Kyrill >> >> >

Re: [PING^2][PATCH, PR46193] Handle mix/max pointer reductions in parloops

On Mon, Aug 24, 2015 at 5:10 PM, Tom de Vries wrote: > On 22-07-15 20:15, Tom de Vries wrote: >> >> On 13/07/15 13:02, Tom de Vries wrote: >>> >>> Hi, >>> >>> this patch fixes PR46193. >>> >>> It handles min and max reductions of pointer type in parloops. >>> >>> Bootstrapped and reg-tested on x86

Re: [PATCH] [AVX512F] Add scatter support for vectorizer

On Fri, Aug 21, 2015 at 2:18 PM, Petr Murzin wrote: > Hello, > Please have a look at updated patch. > > On Tue, Aug 4, 2015 at 3:15 PM, Richard Biener wrote: >> On Fri, 31 Jul 2015, Petr Murzin wrote: >> @@ -5586,8 +5770,6 @@ vectorizable_store (gimple stmt, >> gimple_stmt_iterator *gsi, gimple *

Re: [PATCH 3/5] Build ARRAY_REFs when the base is of ARRAY_TYPE.

On Wed, Aug 26, 2015 at 3:29 PM, Richard Biener wrote: > On Wed, 26 Aug 2015, Bin.Cheng wrote: > >> On Wed, Aug 26, 2015 at 3:50 AM, Jeff Law wrote: >> > On 08/25/2015 05:06 AM, Alan Lawrence wrote: >> >> >> >> When SRA completely scalarizes an array, this patch changes the >> >> generated access

Re: [PATCH 3/5] Build ARRAY_REFs when the base is of ARRAY_TYPE.

On Wed, 26 Aug 2015, Bin.Cheng wrote: > On Wed, Aug 26, 2015 at 3:50 AM, Jeff Law wrote: > > On 08/25/2015 05:06 AM, Alan Lawrence wrote: > >> > >> When SRA completely scalarizes an array, this patch changes the > >> generated accesses from e.g. > >> > >> MEM[(int[8] *)&a + 4B] = 1; > >> > >> to

Re: [RFC 5/5] Always completely replace constant pool entries

On Tue, Aug 25, 2015 at 9:54 PM, Jeff Law wrote: > On 08/25/2015 05:06 AM, Alan Lawrence wrote: >> >> I used this as a means of better-testing the previous changes, as it >> exercises >> the constant replacement code a whole lot more. Indeed, quite a few tests >> are >> now optimized away to nothi

Re: [RFC 4/5] Handle constant-pool entries

On Tue, Aug 25, 2015 at 10:13 PM, Jeff Law wrote: > On 08/25/2015 05:06 AM, Alan Lawrence wrote: >> >> This makes SRA replace loads of records/arrays from constant pool entries, >> with elementwise assignments of the constant values, hence, overcoming the >> fundamental problem in PR/63679. >> >>

Re: [PATCH 3/5] Build ARRAY_REFs when the base is of ARRAY_TYPE.

On Tue, Aug 25, 2015 at 9:50 PM, Jeff Law wrote: > On 08/25/2015 05:06 AM, Alan Lawrence wrote: >> >> When SRA completely scalarizes an array, this patch changes the >> generated accesses from e.g. >> >> MEM[(int[8] *)&a + 4B] = 1; >> >> to >> >> a[1] = 1; >> >> This overcomes a limitation in dom2

Re: [PATCH 2/5] completely_scalarize arrays as well as records

On Tue, Aug 25, 2015 at 11:44 PM, Jeff Law wrote: > On 08/25/2015 03:42 PM, Martin Jambor wrote: >> >> Hi, >> >> On Tue, Aug 25, 2015 at 12:06:14PM +0100, Alan Lawrence wrote: >>> >>> This changes the completely_scalarize_record path to also work on arrays >>> (thus >>> allowing records containing