RE: [PATCH, ping 1] Move insns without introducing new temporaries in loop2_invariant

> From: Jeff Law [mailto:l...@redhat.com] > Sent: Tuesday, May 12, 2015 4:17 AM > > On 05/06/2015 03:47 AM, Thomas Preud'homme wrote: > > Ping? > Something to consider as future work -- I'm pretty sure PRE sets up the > same kind of problematical pattern with a new pseudo (reaching reg) > holding

Re: [PATCH, AArch64] Add Cortex-A53 erratum 843419 configure-time option

Hi, On 5 May 2015 at 10:11, Marcus Shawcroft wrote: > On 4 May 2015 at 09:58, Yvan Roux wrote: > >> Yes this is a better formulation. >> >>> +corresponding flag to the linker. It can be explicitly disabled >>> during >>> +compilation by passing the @option{-mno-fix-cortex-a53-835769} option. >>

Re: [PATCH][AArch64] PR target/65491: Classify V1TF vectors as AAPCS64 short vectors rather than composite types

Ping^2. Thanks, Kyrill On 06/05/15 10:57, Kyrill Tkachov wrote: Ping. https://gcc.gnu.org/ml/gcc-patches/2015-04/msg01004.html Thanks, Kyrill On 20/04/15 11:16, Kyrill Tkachov wrote: Hi all, The ICE in the PR happens when we pass a 1x(128-bit float) vector as an argument. The aarch64 backen

Re: [PATCH][ARM/AArch64] Properly cost rev16 operand

Ping. https://gcc.gnu.org/ml/gcc-patches/2015-05/msg00013.html Thanks, Kyrill On 01/05/15 09:24, Kyrill Tkachov wrote: Hi all, It occurs to me that in the IOR-of-shifts form of the rev16 operation we should be costing the operand properly. For that we'd want to reuse the aarch_rev16_p function

Re: [PATCH][ARM] Fix PR 65955: Do not take REGNO on non-REG operand in movcond_addsi

On 05/05/15 09:22, Kyrill Tkachov wrote: Hi all, As the PR says, the movcond_addsi pattern takes the REGNO of an operand that may be a CONST_INT. The fix for that is rather simple (perhaps even obvious?). Unfortunately the testcase is in ada, and I'm not sure how to integrate this into the

Re: [PATCH][ARM] Handle UNSPEC_VOLATILE in rtx costs and don't recurse inside the unspec

Ping^2. Thanks, Kyrill On 30/04/15 13:01, Kyrill Tkachov wrote: Ping. https://gcc.gnu.org/ml/gcc-patches/2015-04/msg01047.html Thanks, Kyrill On 20/04/15 17:28, Kyrill Tkachov wrote: Hi all, A pet project of mine is to get to the point where backend rtx costs functions won't have to handle

Re: [PATCH][ARM] Fix PR 65955: Do not take REGNO on non-REG operand in movcond_addsi

Ping. https://gcc.gnu.org/ml/gcc-patches/2015-05/msg00284.html Thanks, Kyrill On 05/05/15 11:50, Kyrill Tkachov wrote: On 05/05/15 09:22, Kyrill Tkachov wrote: Hi all, As the PR says, the movcond_addsi pattern takes the REGNO of an operand that may be a CONST_INT. The fix for that is rather

Re: [PATCH][AArch64] Increase static buffer size in aarch64_rewrite_selected_cpu

Ping on the backport. Is this ok for the GCC 5 branch? Thanks, Kyrill On 29/04/15 14:59, Kyrill Tkachov wrote: On 20/04/15 21:30, James Greenhalgh wrote: On Mon, Apr 20, 2015 at 05:24:39PM +0100, Kyrill Tkachov wrote: Hi all, When trying to compile a testcase with -mcpu=cortex-a57+crypto+noc

Re: [PATCH][ARM][stage-1] Initialise cost to COSTS_N_INSNS (1) and increment in arm rtx costs

Ping^2. Thanks, Kyrill On 30/04/15 13:00, Kyrill Tkachov wrote: Ping. https://gcc.gnu.org/ml/gcc-patches/2015-04/msg01130.html Thanks, Kyrill On 21/04/15 10:11, Kyrill Tkachov wrote: Hi all, This is the first of a series to clean up and simplify the arm rtx costs function. This patch initi

Re: [PATCH][AArch64] Increase static buffer size in aarch64_rewrite_selected_cpu

On Tue, May 12, 2015 at 10:08:55AM +0100, Kyrill Tkachov wrote: > Ping on the backport. > Is this ok for the GCC 5 branch? > As below, yes this is OK for the GCC 5 branch. > On 29/04/15 14:59, Kyrill Tkachov wrote: > > On 20/04/15 21:30, James Greenhalgh wrote: > >> On Mon, Apr 20, 2015 at 05:24

[Ada] Generic_Dispatching_Constructor and multiple interfaces.

This patch fixes a regression in the handling of the generic_dispatching_ constructor in the presence of several levels of interfaces. Previous to this patch, a dispatching call might call the wrong primitive of an object whose type overrides a primitive inherited from an interface that has several

RE: [PATCH, ping 1] Move insns without introducing new temporaries in loop2_invariant

> From: gcc-patches-ow...@gcc.gnu.org [mailto:gcc-patches- > ow...@gcc.gnu.org] On Behalf Of Thomas Preud'homme > > > From: Jeff Law [mailto:l...@redhat.com] > > Sent: Tuesday, May 12, 2015 4:17 AM > > > > >> > > >> + > > >> + /* Check that all uses reached by the def in insn would still be > > r

Re: Improve LTO type checking during symtab merging

On Mon, May 11, 2015 at 6:42 AM, Jan Hubicka wrote: > Hi, > this is the updated version of patch with testsuite compensation and > some changes: > - As discussed with Richard on IRC, I added -Wlto-type-mismatch to control > the warnings (enabled by default) and split ODR warnings from warnin

Re: [Patch, Fortran] Simplify lbound

Le 12/05/2015 08:43, Thomas Koenig a écrit : > Hi Mikael, > > >> To be honest, both patches look fragile to me. Yours because it leaves >> gfc_current_ns to its value, leaving the door open to other problems. >> Mine, well, because it's playing with a global variable, with the >> possible side-ef

Re: [PATCH][PR66010] Don't take address of ap unless necessary

On 12-05-15 09:45, Richard Biener wrote: On Mon, 11 May 2015, Tom de Vries wrote: On 11-05-15 09:47, Richard Biener wrote: Bootstrapped and reg-tested on x86_64, with and without -m32. OK for trunk? [ FWIW, I suspect this patch will make life easier for the reimplementation of the pass_st

Re: [PATCH][PR66010] Don't take address of ap unless necessary

On Tue, 12 May 2015, Tom de Vries wrote: > On 12-05-15 09:45, Richard Biener wrote: > > On Mon, 11 May 2015, Tom de Vries wrote: > > > > > On 11-05-15 09:47, Richard Biener wrote: > > > > > Bootstrapped and reg-tested on x86_64, with and without -m32. > > > > > > > > > > > > OK for trunk? > > >

[PATCH, i386, PR target/66048] Fix mode switching ICE for functions returning bounds

Hi, This patch moves up use insns for returned bounds and also marks bnd1 register as holding return value. This fixes problem in create_pre_exit for functions with no returned bounds initialization. Bootstrapped and regtested for x86_64-unknown-linux-gnu. Is it OK for trunk/gcc-5? Thanks,

Re: [PATCH][PR66010] Don't take address of ap unless necessary

Hi, On Fri, 8 May 2015, Tom de Vries wrote: > III. > > Using the patch, before inlining we can see the address operator has been > removed in va_arg: > ... > f2_1 (struct * apD.1832) > { > intD.6 _4; > > # .MEM_3 = VDEF <.MEM_1(D)> > # USE = anything > # CLB = anything > > _4 = VA_A

[PATCH, CHKP] Hardcode wrapper names

Hi, This patch hardcodes wrapper names because builtin assembler name may vary and cannot be used to build a wrapper name. Also don't use alias chain for wrappers. Bootstrapped and regtested for x86_64-unknown-linux-gnu. Going to apply it to trunk and gcc-5 if no objections appear. Thanks,

Re: [PATCH][PR66010] Don't take address of ap unless necessary

On 12-05-15 12:12, Michael Matz wrote: Hi, On Fri, 8 May 2015, Tom de Vries wrote: III. Using the patch, before inlining we can see the address operator has been removed in va_arg: ... f2_1 (struct * apD.1832) { intD.6 _4; # .MEM_3 = VDEF <.MEM_1(D)> # USE = anything # CLB = any

Re: [PATCH, i386, PR target/66048] Fix mode switching ICE for functions returning bounds

On Tue, May 12, 2015 at 12:07 PM, Ilya Enkovich wrote: > This patch moves up use insns for returned bounds and also marks bnd1 > register as holding return value. This fixes problem in create_pre_exit for > functions with no returned bounds initialization. Bootstrapped and regtested > for x8

Re: [PATCH 2/6] combine: If recog fails, try again with zero_ext{ract,end} simplified

On Tue, May 12, 2015 at 09:20:18AM +0100, Kyrill Tkachov wrote: > Does this patch means we can remove any patterns in > the backend that look like: > > - [(set (match_operand:SI 0 "gpc_reg_operand" "=r") > - (zero_extract:SI (match_operand:SI 1 "gpc_reg_operand" "r") > -

[PINGv6][PATCH] ASan on unaligned accesses

On 04/07/2015 03:22 PM, Jakub Jelinek wrote: How are the automatic misaligned variables different from say heap allocated ones, or global vars etc.? No difference you are right Jakub. Shadow memory initialization for heap values and globals of course also should be changed but it is a task for

Re: [PINGv6][PATCH] ASan on unaligned accesses

On 05/12/2015 02:16 PM, Marat Zakirov wrote: On 04/07/2015 03:22 PM, Jakub Jelinek wrote: How are the automatic misaligned variables different from say heap allocated ones, or global vars etc.? No difference you are right Jakub. Shadow memory initialization for heap values and globals of course

Re: [PATCH][PR66010] Don't take address of ap unless necessary

On 12-05-15 12:04, Tom de Vries wrote: Committed with comments below added. Hmm, this causes an ice for s390 in gcc.dg/tree-ssa/stdarg-2.c: ... gimplification failed: &ap type 0x7f132f46b888 __va_list_tag> sizes-gimplified BLK size unit size

[PATCH] Support vectorizing SLP with mixed plus/minus

The following makes it possible to vectorize complex multiplication without unrolling by allowing mixed operations in a SLP node (plus/minus for now). We vectorize that by computing both plus and minus and then merge the result. Something needs to be done in the backends to fully leverage this a

Re: [tz] [C/C++ PATCH] Implement -Wshift-negative-value (PR c/65179)

On May 11, 3:18pm, sell...@imgtec.com (Steve Ellcey) wrote: -- Subject: Re: [tz] [C/C++ PATCH] Implement -Wshift-negative-value (PR c/651 | FYI: I put this into my glibc sources (the private.h, zdump.c, and zic.c | parts, glibc doesn't have localtime.c in the timezone directory and the | one in t

[Ada] Add constants to g-sercom for slow bit rates

This change introduces new literals B75, B110, B150, B300, and B600 for enumerated type GNAT.Serial_Communications.Data_Rate, denoting the corresponding bit rates. This allows communication with legacy devices that require them. Tested on x86_64-pc-linux-gnu, committed on trunk 2015-05-12 Thomas

Re: Fwd: [PING 2][PATCH] libgcc: Add CFI directives to the soft floating point support code for ARM

On Tue, May 12, 2015 at 5:49 AM, Ramana Radhakrishnan wrote: > That's what I mean when I say email clients "munged it" : email clients and > / or some popular email servers appear to end up munging white spaces and > patches don't apply cleanly. > > So, no it doesn't work - once you've sent it thr

[PATCH] Fix PR66101

This fixes PR66101. Bootstrapped and tested on x86_64-unknown-linux-gnu, applied. Richard. 2015-05-12 Richard Biener PR tree-optimization/66101 * tree-ssa-dce.c (remove_dead_stmt): Properly mark loops for fixup if we turn a loop exit edge to a fallthru edge.

My patch for GCC 5 directory names

I promised to send out my pat^Whack. Before building I introduce gcc/FULL-VER as copy of gcc/BASE-VER and adjust gcc/BASE-VER to just the major number. Then I only need the following small patch (where I don't speak enough tcl for fixing libjava.exp "properly"). Without the FULL-VER trick the p

[PATCH] Fix PR66110

The following patch fixes PR66110 - there is no reason why two alias-sets conflict just because one or the other has a child with alias-set zero. We need to paper over strict-aliasing bugs in dfp.c (re-using struct real_value unsigned long sig[] as decimal128) by adding -Wno-strict-aliasing, othe

Re: False ODR violation positives on anonymous namespace types

On 05/11/2015 04:39 PM, Jan Hubicka wrote: What happens in LTO is that lto-symtab decide to merge the two declarations of foo. At this time it has chance to output the warning. For that it needs to be able to work out that these declarations are having different types, but because LTO merge can

Re: My patch for GCC 5 directory names

On Tue, May 12, 2015 at 6:42 AM, Richard Biener wrote: > > I promised to send out my pat^Whack. Before building I introduce > gcc/FULL-VER as copy of gcc/BASE-VER and adjust gcc/BASE-VER to > just the major number. Then I only need the following small > patch (where I don't speak enough tcl for

Re: My patch for GCC 5 directory names

On Tue, 12 May 2015, H.J. Lu wrote: > On Tue, May 12, 2015 at 6:42 AM, Richard Biener wrote: > > > > I promised to send out my pat^Whack. Before building I introduce > > gcc/FULL-VER as copy of gcc/BASE-VER and adjust gcc/BASE-VER to > > just the major number. Then I only need the following sma

Re: My patch for GCC 5 directory names

On Tue, May 12, 2015 at 6:54 AM, Richard Biener wrote: > On Tue, 12 May 2015, H.J. Lu wrote: > >> On Tue, May 12, 2015 at 6:42 AM, Richard Biener wrote: >> > >> > I promised to send out my pat^Whack. Before building I introduce >> > gcc/FULL-VER as copy of gcc/BASE-VER and adjust gcc/BASE-VER to

Re: [PATCH 0/6] Getting rid of some zero_ext* patterns

On Sun, May 10, 2015 at 09:13:49AM -0700, Segher Boessenkool wrote: > This series teaches combine to behave a bit better, and then takes > advantage of that in the rs6000 backend. > > I'll test the combine patches on x86_64-linux before committing, > and will wait a bit for comments anyway. Done

Re: My patch for GCC 5 directory names

On Tue, 12 May 2015, H.J. Lu wrote: > On Tue, May 12, 2015 at 6:54 AM, Richard Biener wrote: > > On Tue, 12 May 2015, H.J. Lu wrote: > > > >> On Tue, May 12, 2015 at 6:42 AM, Richard Biener wrote: > >> > > >> > I promised to send out my pat^Whack. Before building I introduce > >> > gcc/FULL-VER

Re: False ODR violation positives on anonymous namespace types

> On 05/11/2015 04:39 PM, Jan Hubicka wrote: > >What happens in LTO is that lto-symtab decide to merge the two declarations > >of > >foo. At this time it has chance to output the warning. For that it needs to > >be able to work out that these declarations are having different types, but > >becau

Re: [PATCH, PR target/65103, 3/3] Change rtx cost for i386 address constants

Ping 2015-03-10 19:14 GMT+03:00 Ilya Enkovich : > On 10 Mar 18:11, Ilya Enkovich wrote: >> Hi, >> >> This patch changes rtx cost for address constants to 0 similar to what we >> have in address cost hook beacuse we expect them to be propagated into >> address expression anyway. Bootstrapped and

[PATCH, libgcc]: Commited: Add space between string literal and macro name.

... to avoid "C++11 requires a space between string literal and macro" warning. Plus fix lines over 80 chars. 2015-05-12 Uros Bizjak * libgcov-util.c: Add space between string literal and macro name. Tested on x86_64-linux-gnu and committed to mainline SVN. Uros. Index: ChangeLog ==

Re: My patch for GCC 5 directory names

On Tue, May 12, 2015 at 7:01 AM, Richard Biener wrote: > On Tue, 12 May 2015, H.J. Lu wrote: > >> On Tue, May 12, 2015 at 6:54 AM, Richard Biener wrote: >> > On Tue, 12 May 2015, H.J. Lu wrote: >> > >> >> On Tue, May 12, 2015 at 6:42 AM, Richard Biener wrote: >> >> > >> >> > I promised to send o

Re: False ODR violation positives on anonymous namespace types

> /* Record, union and enumeration type have linkage that allows use >to check type_in_anonymous_namespace_p. Compound types >can be always compared structurally. >To save some work we compare builtin types properties of its main >variant. A special case are i

Re: [PATCH] Fix PR ipa/65557

On 04/30/2015 11:11 PM, Uros Bizjak wrote: > Hello! > >> 2015-03-29 Martin Liska >> >>PR ipa/65557 >>* ipa-icf.c (sem_function::equals_wpa): Check if IPA CP >>has already filled up function summary. >>(sem_item_optimizer::update_hash_by_addr_refs): Likewise. >> >> gcc/testsuite/

Re: [PATCH] Fix PR ipa/65557

On Tue, May 12, 2015 at 4:54 PM, Martin Liška wrote: > On 04/30/2015 11:11 PM, Uros Bizjak wrote: >> Hello! >> >>> 2015-03-29 Martin Liska >>> >>>PR ipa/65557 >>>* ipa-icf.c (sem_function::equals_wpa): Check if IPA CP >>>has already filled up function summary. >>>(sem_item_optim

Re: [PATCH, 1/5] check_GNU_style.sh: Don't use filename prefix for one patch

On 05/12/2015 01:27 AM, Tom de Vries wrote: Hi, this patch makes sure that we don't print the patch filename in front of the error messages, if we're checking just one file. Making the lines shorter makes things easier to read. And since we're only processing one file, there's no information los

[Ada] Incomplete types and null procedures.

An untagged incomeplete type is legal in the profile of a null procedure. This patch remotes a spurious error on such, because the legality check for the use of an incomplete type was also applied to the subprogram body generated during expansion of a null procedure declaration. The following must

[Ada] Legality checks on a formal derived type derived from previous formal.

This patch adds a missing legality check on the instantiation of a formal derived type whose parent type is a previous formal of the same generic unit, that is not a derived type. Compiling generic_testm.adb must yield: generic_testm.adb:7:26: expect type derived from "Integer" in instantiation

Re: [PATCH, 2/5] check_GNU_style.sh: Check file presence

On 05/12/2015 01:28 AM, Tom de Vries wrote: Hi, this patch checks that the files specified on the check_GNU_style.sh command line are present. OK for trunk? Thanks, - Tom 0002-check_GNU_style.sh-Check-file-presence.patch [PATCH 2/5] check_GNU_style.sh: Check file presence 2015-05-11 Tom

[Ada] Implement new pragma/aspect Volatile_Full_Access

A new pragma (and equivalent aspect) Volatile_Full_Access is implemented. This is similar to Volatile except that there is a guarantee that every read and write access to an object with this aspect will always use a single instruction which reads or writes all the bits of the object. Note that this

Re: [PATCH, 3/5] check_GNU_style.sh: Read either from stdin, or from files

On 05/12/2015 01:29 AM, Tom de Vries wrote: Hi, this patch makes sure we either handle: ... $ cat temp.patch | ./check_GNU_style.sh - ... or ... $ ./check_GNU_style.sh temp.patch ... but not: ... $ cat temp.patch | ./check_GNU_style.sh - temp2.patch ... OK for trunk? Thanks, - Tom 0003-check

Re: [PATCH, 4/5] check_GNU_style.sh: Put stdin into temp file

On 05/12/2015 01:29 AM, Tom de Vries wrote: Hi, this patch makes sure we dump stdin into a file before we handle it in the rest of the script. This makes the stdin handling more like the handling of normal files. OK for trunk? Thanks, - Tom 0004-check_GNU_style.sh-Put-stdin-into-temp-file.pa

[PATCH, alpha]: Switch alpha to TARGET_SUPPORTS_WIDE_INT

Hello! The patch also further removes dead HWI < 64 code in a couple of places. 2015-05-12 Uros Bizjak * config/alpha/alpha.h (TARGET_SUPPORTS_WIDE_INT): New define. * config/alpha/alpha.c (alpha_rtx_costs): Handle CONST_WIDE_INT. (alpha_extract_integer): Ditto. (alpha_legitim

Re: [PATCH, 5/5] check_GNU_style.sh: Fix tab size in 80 characters check

On 05/12/2015 01:29 AM, Tom de Vries wrote: Hi, this patch fixes a problem in the 80 characters length check. Currently tab width is not properly calculated. The patch uses expand to interpret tabs properly. OK for trunk? Thanks, - Tom 0005-check_GNU_style.sh-Fix-tab-size-in-80-characters-

[PATCH][PING^2] Skip preprocessor directives in mklog

On 04/30/2015 12:03 PM, Yury Gribov wrote: On 04/21/2015 02:26 PM, Yury Gribov wrote: Hi all, Contrib/mklog is currently faked by preprocessor directives inside functions to produce invalid ChangeLog. The attached patch fixes this. Tested with my local mklog testsuite and http://paste.debian.

Re: [PATCH][PING^2] Skip preprocessor directives in mklog

The patch looks fine to me. I'm not really involved in GCC development anymore. I would suggest that this script should be maintained by whoever's been hacking on it the most. It's a simple script, so it shouldn't be hard to find a new maintainer for it. Diegop. On Tue, May 12, 2015 at 11:19 AM

Re: [patch 1/28] top-level: Use automake-1.11.6

On Tue, 12 May 2015, Michael Haubenwallner wrote: > On 05/11/2015 07:55 PM, Joseph Myers wrote: > > On Sat, 9 May 2015, Bernd Edlinger wrote: > > > >> But maybe you would like it better if we update, for instance, to: > >> automake-1.14 _and_ autoconf-2.69 ? > > > > Updating to current automak

Re: My patch for GCC 5 directory names

Hi, On Tue, 12 May 2015, H.J. Lu wrote: > >> So we have > >> > >> experimental > >> release > >> post-release > >> > >> Why not just rename prerelease to post-release? That is a one-line > >> change. > > > > Why print anything at all? 5.1.1 is after 5.1.0 in obvious ways. > > > > How can you te

Re: [PATCH, PR target/65103, 3/3] Change rtx cost for i386 address constants

Hello! >> 2015-03-10 Ilya Enkovich >> >> PR target/65103 >> * config/i386/i386.c (ix86_rtx_costs): We want to propagate >> link time constants into adress expressions and therefore set >> their cost to 0. >> >> gcc/testsuite/ >> >> 2015-03-10 Ilya Enkovich >>

Re: [PATCH][PING^2] Skip preprocessor directives in mklog

On 05/12/2015 06:23 PM, Diego Novillo wrote: The patch looks fine to me. I'm not really involved in GCC development anymore. I would suggest that this script should be maintained by whoever's been hacking on it the most. It's a simple script, so it shouldn't be hard to find a new maintainer for

Re: [PATCH, alpha]: Switch alpha to TARGET_SUPPORTS_WIDE_INT

On 05/12/2015 08:06 AM, Uros Bizjak wrote: > 2015-05-12 Uros Bizjak > > * config/alpha/alpha.h (TARGET_SUPPORTS_WIDE_INT): New define. > * config/alpha/alpha.c (alpha_rtx_costs): Handle CONST_WIDE_INT. > (alpha_extract_integer): Ditto. > (alpha_legitimate_constant_p): Ditto. >

Re: [PATCH][PING^2] Skip preprocessor directives in mklog

On 12-05-15 17:33, Yury Gribov wrote: On 05/12/2015 06:23 PM, Diego Novillo wrote: The patch looks fine to me. I'm not really involved in GCC development anymore. I would suggest that this script should be maintained by whoever's been hacking on it the most. It's a simple script, so it shouldn'

Re: My patch for GCC 5 directory names

On Tue, May 12, 2015 at 8:28 AM, Michael Matz wrote: > Hi, > > On Tue, 12 May 2015, H.J. Lu wrote: > >> >> So we have >> >> >> >> experimental >> >> release >> >> post-release >> >> >> >> Why not just rename prerelease to post-release? That is a one-line >> >> change. >> > >> > Why print anything

Re: [PATCH][PING^2] Skip preprocessor directives in mklog

On Tue, May 12, 2015 at 06:33:28PM +0300, Yury Gribov wrote: > On 05/12/2015 06:23 PM, Diego Novillo wrote: > >The patch looks fine to me. > > > >I'm not really involved in GCC development anymore. I would suggest > >that this script should be maintained by whoever's been hacking on it > >the most.

Re: False ODR violation positives on anonymous namespace types

Looks good. Jason

[patch, nios2] add IJMP_REGS reg class for indirect jumps

While testing something else, I noticed a couple compilation errors like: /tmp/ccCxOOoN.s: Assembler messages: /tmp/ccCxOOoN.s:164: Error: r31 cannot be used with jmp; use ret instead The assembler is correct in rejecting this code, since this is a documented restriction of the JMP instruction.

Re: My patch for GCC 5 directory names

On May 12, 2015 5:58:07 PM GMT+02:00, "H.J. Lu" wrote: >On Tue, May 12, 2015 at 8:28 AM, Michael Matz wrote: >> Hi, >> >> On Tue, 12 May 2015, H.J. Lu wrote: >> >>> >> So we have >>> >> >>> >> experimental >>> >> release >>> >> post-release >>> >> >>> >> Why not just rename prerelease to post-rel

Re: My patch for GCC 5 directory names

On Tue, May 12, 2015 at 9:09 AM, Richard Biener wrote: > On May 12, 2015 5:58:07 PM GMT+02:00, "H.J. Lu" wrote: >>On Tue, May 12, 2015 at 8:28 AM, Michael Matz wrote: >>> Hi, >>> >>> On Tue, 12 May 2015, H.J. Lu wrote: >>> >> So we have >> >> experimental >> release >> p

Re: [PATCH][PING^2] Skip preprocessor directives in mklog

On Tue, May 12, 2015 at 11:50 AM, Tom de Vries wrote: > I'm not a good choice to be the maintainer of a perl script. I'm all kinds of sorry about the original choice of scripting language. I'd just spend a couple of hours re-writing it in python.

Re: [PATCH][PING^2] Skip preprocessor directives in mklog

On 05/12/2015 06:57 PM, Trevor Saunders wrote: On Tue, May 12, 2015 at 06:33:28PM +0300, Yury Gribov wrote: On 05/12/2015 06:23 PM, Diego Novillo wrote: The patch looks fine to me. I'm not really involved in GCC development anymore. I would suggest that this script should be maintained by whoe

Re: False ODR violation positives on anonymous namespace types

> Looks good. Thank you! Testing looks good, too, no false positives on Firefox. I suppose the next challenge is to make type_with_linkage_p to return false on types that was not built by C++ FE as discussed in https://gcc.gnu.org/ml/gcc-patches/2015-05/msg01023.html What would be the preferred so

Re: [PATCH 4/n] OpenMP 4.0 offloading infrastructure: lto-wrapper

Hi Bernd! One question to you, as the one who has originally written this code (): On Thu, 2 Oct 2014 19:14:57 +0400, Ilya Verbin wrote: > With this patch lto-wrapper performs invocation of mkoffload tool for each > offload target. [...]

Re: My patch for GCC 5 directory names

On May 12, 2015 6:11:45 PM GMT+02:00, "H.J. Lu" wrote: >On Tue, May 12, 2015 at 9:09 AM, Richard Biener > wrote: >> On May 12, 2015 5:58:07 PM GMT+02:00, "H.J. Lu" >wrote: >>>On Tue, May 12, 2015 at 8:28 AM, Michael Matz wrote: Hi, On Tue, 12 May 2015, H.J. Lu wrote: > >>

Re: My patch for GCC 5 directory names

On Tue, May 12, 2015 at 9:32 AM, Richard Biener wrote: > On May 12, 2015 6:11:45 PM GMT+02:00, "H.J. Lu" wrote: >>On Tue, May 12, 2015 at 9:09 AM, Richard Biener >> wrote: >>> On May 12, 2015 5:58:07 PM GMT+02:00, "H.J. Lu" >>wrote: On Tue, May 12, 2015 at 8:28 AM, Michael Matz wrote: >

[patch, nios2] use "trap" instruction instead of "break"

In connection with this discussion https://gcc.gnu.org/ml/gcc/2015-02/msg00163.html about bad behavior of -fisolate-erroneous-paths-dereference on nios2 bare-metal targets, I realized that we also had a serious problem on nios2-linux-gnu targets as well. The "trap" and "ctrapsi4" insns were

Re: [patch 1/28] top-level: Use automake-1.11.6

On 05/12/2015 10:45 AM, Michael Haubenwallner wrote: > On 05/11/2015 07:55 PM, Joseph Myers wrote: >> On Sat, 9 May 2015, Bernd Edlinger wrote: >> >>> But maybe you would like it better if we update, for instance, to: >>> automake-1.14 _and_ autoconf-2.69 ? >> >> Updating to current automake and

[PATCH, ARM] Make tune params tables more self-documenting

This patch modifies the way that the tune_params data structures are constructed so that: 1) The initializers are (mostly) self documenting. 2) The compiler does a better job of detecting merge failures. Mostly this resolves around using C++ enums, such that we get type clashes if elements do not

[patch] Clean up detection of SJLJ exceptions in target libraries

Hi, 6 target libraries in the tree detect whether they are being compiled by a compiler configured for setjmp/longjmp exceptions: libada, libgcc, libgo, libjava, libobjc and libstdc++. They can be divided into 3 categories: 1) libada only checks the preprocessor macro __USING_SJLJ_EXCEPTIONS__

RE: [RFC]: Remove Mem/address type assumption in combiner

On Tue, 2015-05-12 at 05:21 +, Kumar, Venkataramanan wrote: > Hi Steve, > > Yes this is expected. As Segher pointed out, we need to change .md patterns > in > target to be based on shifts instead of mults. > > Regards, > Venkat. Here is my patch to change this. I tested it with the mips-

Re: [PATCH] Support vectorizing SLP with mixed plus/minus

On Tue, May 12, 2015 at 4:48 AM, Richard Biener wrote: > > The following makes it possible to vectorize complex multiplication > without unrolling by allowing mixed operations in a SLP node > (plus/minus for now). We vectorize that by computing both plus > and minus and then merge the result. > >

Re: [PATCH 4/n] OpenMP 4.0 offloading infrastructure: lto-wrapper

On 05/12/2015 06:27 PM, Thomas Schwinge wrote: Patch variant 1: @@ -4266,7 +4266,7 @@ process_command (unsigned int decoded_op } gcc_assert (!IS_ABSOLUTE_PATH (tooldir_base_prefix)); - tooldir_prefix2 = concat (tooldir_base_prefix, spec_host_machine, + tooldir_prefix2 = concat (too

Re: [patch 1/28] top-level: Use automake-1.11.6

On Tue, 12 May 2015, Michael Haubenwallner wrote: > It turns out that besides the top-level ones there are more > copied-from-automake > files in various library dirs - where some of them seem quite outdated > actually: > > libffi/mdate-sh > libjava/classpath/INSTALL > libjava/libltdl/con

[PATCH, committed] Unbreak the JIT after conversion of md_asm_adjust to use vec<>

The patch: https://gcc.gnu.org/ml/gcc-patches/2015-05/msg00600.html broke the build of the jit: In file included from ../../src/gcc/jit/jit-builtins.c:23:0: ../../src/gcc/target.h:188:60: error: expected ‘;’ at end of member declaration #define DEFHOOK(NAME, DOC, TYPE, PARAMS, INIT) TYPE (* NA

Re: False ODR violation positives on anonymous namespace types

Hi, here is patch I LTO-bootstrapped and regtested on x86_64-linux and I am going to commit shortly. Thank you for useful discussion - I did not really realize before that types may or may not have linkage. This caused surprises indeed ;) Honza * ipa-devirt.c (type_with_linkage_p): New f

[PATCH] Fix get_min_precision in __builtin_*_overflow expansion (PR target/66112)

Hi! This patch fixes a thinko in get_min_precision. If we want UNSIGNED number and arg_min is signed, but non-negative, calling wi::min_precision with SIGNED resulted in one bigger precision than necessary. If get_min_precision is originally called with SIGNED and sign is then changed into UNSIG

[PATCH] Add {u,}mulvhi4 patterns on i?86 (PR target/66112)

Hi! This patch improves expansion of __builtin_mul_overflow for HImode, both signed and unsigned, on x86_64/i686. Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk? 2015-05-12 Jakub Jelinek PR target/66112 * config/i386/i386.md (mulv4, umulv4, *umulv4):

Re: [PATCH] Add {u,}mulvhi4 patterns on i?86 (PR target/66112)

On Tue, May 12, 2015 at 7:43 PM, Jakub Jelinek wrote: > Hi! > > This patch improves expansion of __builtin_mul_overflow for HImode, both > signed and unsigned, on x86_64/i686. > > Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk? > > 2015-05-12 Jakub Jelinek > > PR ta

Re: [PATCH, ping 1] Move insns without introducing new temporaries in loop2_invariant

On 05/12/2015 02:55 AM, Thomas Preud'homme wrote: From: Jeff Law [mailto:l...@redhat.com] Sent: Tuesday, May 12, 2015 4:17 AM On 05/06/2015 03:47 AM, Thomas Preud'homme wrote: Ping? Something to consider as future work -- I'm pretty sure PRE sets up the same kind of problematical pattern with

Re: [PATCH][PING^2] Skip preprocessor directives in mklog

On 05/12/2015 10:14 AM, Yury Gribov wrote: On 05/12/2015 06:57 PM, Trevor Saunders wrote: On Tue, May 12, 2015 at 06:33:28PM +0300, Yury Gribov wrote: On 05/12/2015 06:23 PM, Diego Novillo wrote: The patch looks fine to me. I'm not really involved in GCC development anymore. I would suggest t

Re: [PATCH] Add {u,}mulvhi4 patterns on i?86 (PR target/66112)

On Tue, May 12, 2015 at 8:43 PM, Jakub Jelinek wrote: > Hi! > > This patch improves expansion of __builtin_mul_overflow for HImode, both > signed and unsigned, on x86_64/i686. > > Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk? > > 2015-05-12 Jakub Jelinek > > PR ta

Re: [PATCH, i386, PR target/66048] Fix mode switching ICE for functions returning bounds

On 05/12/2015 04:07 AM, Ilya Enkovich wrote: Hi, This patch moves up use insns for returned bounds and also marks bnd1 register as holding return value. This fixes problem in create_pre_exit for functions with no returned bounds initialization. Bootstrapped and regtested for x86_64-unknown-

Re: [PATCH, ping 1] Move insns without introducing new temporaries in loop2_invariant

On 05/12/2015 03:32 AM, Thomas Preud'homme wrote: From: gcc-patches-ow...@gcc.gnu.org [mailto:gcc-patches- ow...@gcc.gnu.org] On Behalf Of Thomas Preud'homme From: Jeff Law [mailto:l...@redhat.com] Sent: Tuesday, May 12, 2015 4:17 AM + + /* Check that all uses reached by the def in insn wou

[patch, fortran] Fix PR 66111

Hello world, this patch fixes a regression from the inline matmul patch by not inlining a case that is not yet handled, namely vector subscripts. Regression-tested. OK for trunk? Regards Thomas 2015-05-12 Thomas Koenig PR fortran/66111 * frontend-passes.c (has_dime

Re: [PATCH, RFC]: Next stage1, refactoring: propagating rtx subclasses

On 05/11/2015 02:41 PM, Mikhail Maltsev wrote: On 09.05.2015 0:54, Jeff Law wrote: Both patches are approved. Please install onto the trunk. jeff Sorry for delay. When I started to work on this task, I wrote that I'll test the patches on couple of other platforms (not just x86). Probably I

Re: [patch] Clean up detection of SJLJ exceptions in target libraries

On 05/12/2015 10:42 AM, Eric Botcazou wrote: Hi, 6 target libraries in the tree detect whether they are being compiled by a compiler configured for setjmp/longjmp exceptions: libada, libgcc, libgo, libjava, libobjc and libstdc++. They can be divided into 3 categories: 1) libada only checks th

Re: [PATCH][PING^2] Skip preprocessor directives in mklog

On 05/12/2015 09:19 AM, Yury Gribov wrote: On 04/30/2015 12:03 PM, Yury Gribov wrote: On 04/21/2015 02:26 PM, Yury Gribov wrote: Hi all, Contrib/mklog is currently faked by preprocessor directives inside functions to produce invalid ChangeLog. The attached patch fixes this. Tested with my lo

Re: [PATCH] Fix get_min_precision in __builtin_*_overflow expansion (PR target/66112)

On May 12, 2015 8:40:32 PM GMT+02:00, Jakub Jelinek wrote: >Hi! > >This patch fixes a thinko in get_min_precision. If we want UNSIGNED >number and arg_min is signed, but non-negative, calling >wi::min_precision >with SIGNED resulted in one bigger precision than necessary. If >get_min_precision i

[Patch ARM-AArch64/testsuite 00/13] Neon intrinsics executable tests

This patch series is a follow-up to the tests I already contributed, converted from my original testsuite. This series consists in 13 new files, which can be committed independently. Another series (hopefully final) will follow. Tested with qemu on arm*linux, aarch64-linux. I couldn't test on a

[Patch ARM-AArch64/testsuite Neon intrinsics 03/13] Add vqrdmulh tests.

diff --git a/gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/vqrdmulh.c b/gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/vqrdmulh.c new file mode 100644 index 000..8875c53 --- /dev/null +++ b/gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/vqrdmulh.c @@ -0,0 +1,161 @@ +#include

[Patch ARM-AArch64/testsuite Neon intrinsics 01/13] Add vqmovn tests.

diff --git a/gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/vqmovn.c b/gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/vqmovn.c new file mode 100644 index 000..b4d3198 --- /dev/null +++ b/gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/vqmovn.c @@ -0,0 +1,134 @@ +#include +#incl

  1   2   >