Re: [PATCH][AArch64] Remove '*' from movsi/di/ti patterns

On Tue, Sep 12, 2017 at 9:10 AM, James Greenhalgh wrote: > On Wed, Jul 26, 2017 at 02:46:14PM +0100, Wilco Dijkstra wrote: >> Remove the remaining uses of '*' from the movsi/di/ti patterns. >> Using '*' in alternatives is typically incorrect at it tells the register >>

Re: [PATCH version 2, rs6000] Add builtins to convert from float/double to int/long using current rounding mode

On Tue, Sep 12, 2017 at 05:41:34PM -0500, Segher Boessenkool wrote: > This needs "TARGET_HARD_FLOAT && TARGET_DOUBLE_FLOAT" I think? Which > is the same as "TARGET_DF_FPR". "lrintdi2" also has "TARGET_FPRND" > but that is a mistake I think. Cc:ing Mike for that. TARGET_FPRND is for ISA 2.02

Re: [PATCH] Factor out division by squares and remove division around comparisons (0/2)

On 09/06/2017 03:54 AM, Jackson Woodruff wrote: > Hi all, > > This patch is split from part (1/2). It includes the patterns that have > been moved out of fold-const.c > > > It also removes an (almost entirely) redundant pattern: > > (A / C1) +- (A / C2) -> A * (1 / C1 +- 1 / C2) > > which

Re: [PATCH version 2, rs6000] Add builtins to convert from float/double to int/long using current rounding mode

On Tue, Sep 12, 2017 at 02:07:43PM -0700, Carl Love wrote: > On Tue, 2017-09-12 at 11:49 -0500, Segher Boessenkool wrote: > > On Fri, Sep 08, 2017 at 11:41:13AM -0700, Carl Love wrote: > > > The following patch adds support for a couple of requested builtins that > > > convert from float/double to

Re: Turn SECONDARY_MEMORY_NEEDED into a hook

On 09/12/2017 12:50 PM, Richard Sandiford wrote: > Since the patch is going through all the definitions anyway, it seemed > like a good opportunity to put the mode argument first, to match the > order for register_move_cost. > > Tested on aarch64-linux-gnu, x86_64-linux-gnu and

Re: Turn SECONDARY_MEMORY_NEEDED_MODE into a target hook

On 09/12/2017 12:47 PM, Richard Sandiford wrote: > It feels like SECONDARY_MEMORY_NEEDED and SECONDARY_MEMORY_NEEDED_MODE > should be a single hook that returns the memory mode if memory is needed > and an empty opt_mode otherwise. The snag is this code in reload.c: > > /* If we need a

Re: [PATCH, rs6000] [v2] Folding of vector loads in GIMPLE

> On Sep 12, 2017, at 4:08 PM, Will Schmidt wrote: > > Hi, > > [PATCH, rs6000] [v2] Folding of vector loads in GIMPLE > > Folding of vector loads in GIMPLE. > > Add code to handle gimple folding for the vec_ld builtins. > Remove the now obsoleted folding code for

Re: [PATCH, rs6000] [v2] Folding of vector loads in GIMPLE

Hi, [PATCH, rs6000] [v2] Folding of vector loads in GIMPLE Folding of vector loads in GIMPLE. Add code to handle gimple folding for the vec_ld builtins. Remove the now obsoleted folding code for vec_ld from rs6000-c.c. Surrounding comments have been adjusted slightly so they continue to

Re: [PATCH version 2, rs6000] Add builtins to convert from float/double to int/long using current rounding mode

On Tue, 2017-09-12 at 11:49 -0500, Segher Boessenkool wrote: > Hi Carl, > > On Fri, Sep 08, 2017 at 11:41:13AM -0700, Carl Love wrote: > > The following patch adds support for a couple of requested builtins that > > convert from float/double to int / long using the current rounding > > mode. I

Re: std::forward_list optim for always equal allocator

On 12/09/2017 00:10, Jonathan Wakely wrote: On 11/09/17 22:39 +0200, Daniel Krügler wrote: 2017-09-11 22:36 GMT+02:00 François Dumont : [..] So my remark was rather for the: _Fwd_list_iterator() noexcept : _M_node() { } that could simply be

[SPARC] Small source location information fix

output_return and output_sibcall manually emit the instruction present in the delay slot (if any) but fail to emit its source location information. Tested on SPARC64/Linux, applied on the mainline. 2017-09-12 Eric Botcazou * config/sparc/sparc.c

Re: [PATCH] Add -static-pie to GCC driver to create static PIE

On Tue, Sep 12, 2017 at 11:48 AM, Aaron Sawdey wrote: > On Tue, 2017-09-12 at 16:20 +, Joseph Myers wrote: >> On Mon, 28 Aug 2017, H.J. Lu wrote: >> >> > Here is the updated patch. OK for trunk? >> >> OK. > > This seems to be causing an issue for me on powerpc:

Turn SECONDARY_MEMORY_NEEDED into a hook

Since the patch is going through all the definitions anyway, it seemed like a good opportunity to put the mode argument first, to match the order for register_move_cost. Tested on aarch64-linux-gnu, x86_64-linux-gnu and powerpc64le-linux-gnu. Also tested by comparing the testsuite assembly output

Re: [PATCH] Add -static-pie to GCC driver to create static PIE

On Tue, 2017-09-12 at 16:20 +, Joseph Myers wrote: > On Mon, 28 Aug 2017, H.J. Lu wrote: > > > Here is the updated patch.   OK for trunk? > > OK. This seems to be causing an issue for me on powerpc: /home/sawdey/src/gcc/trunk/build/./prev-gcc/xg++

Turn SECONDARY_MEMORY_NEEDED_MODE into a target hook

It feels like SECONDARY_MEMORY_NEEDED and SECONDARY_MEMORY_NEEDED_MODE should be a single hook that returns the memory mode if memory is needed and an empty opt_mode otherwise. The snag is this code in reload.c: /* If we need a memory location to copy between the two reload regs,

Re: [AArch64] Merge stores of D register values of different modes

Thanks for doing this, looks good to me FWIW. I was just wondering: Jackson Woodruff writes: > @@ -14712,6 +14712,11 @@ aarch64_operands_ok_for_ldpstp (rtx *operands, bool > load, >if (!rtx_equal_p (base_1, base_2)) > return false; > > + /* Check that

Re: [PATCH, rs6000] Fix PR target/80210: ICE in extract_insn

On 9/9/17 3:44 AM, Andreas Schwab wrote: > On Sep 08 2017, Peter Bergner wrote: > >> The following patch fixes the problem I saw on Linux and bootstraps and >> regtests >> with no regressions on LE and BE (running testsuite in both 32-bit and 64-bit >> modes). I was

Re: [Patch, fortran] PR82173 (PDT) - [meta-bug] Parameterized derived type errors

Hi Janus et al., I had to do this quickly because of time pressure and I thought it to be most efficient while the main patch was fresh in my mind. Committed as revision 252039 with attributions suitably modified. Thanks Paul On 11 September 2017 at 20:50, Janus Weil

[AArch64, patch] Refactor of aarch64-ldpstp.md

Hi all, This patch removes a lot of duplicated code in aarch64-ldpstp.md. The patterns that did not previously generate a base register now do not check for aarch64_mem_pair_operand in the pattern. This has been extracted to a check in aarch64_operands_ok_for_ldpstp. All patterns in the file

Re: [PATCH] Fix fortran vector tests on powerpc and aarch64, PR tree-optimization/80925

Hi Steve, This looks fine from the fortran side - especially since it involves removing checks :-) Thanks Paul On 12 September 2017 at 18:34, Steve Ellcey wrote: > Ping, also adding fort...@gcc.gnu.org which I seem to left out when > sending this to

Re: [PATCH, rs6000] Folding of vector loads in GIMPLE

On Tue, 2017-09-12 at 10:22 -0500, Bill Schmidt wrote: > > On Sep 12, 2017, at 9:41 AM, Will Schmidt wrote: > > > > Hi > > > > [PATCH, rs6000] Folding of vector loads in GIMPLE > > > > Folding of vector loads in GIMPLE. > > > > - Add code to handle gimple folding

Re: [PATCH] Fix fortran vector tests on powerpc and aarch64, PR tree-optimization/80925

Ping,  also adding fort...@gcc.gnu.org which I seem to left out when sending this to gcc-patches@gcc.gnu.org. Steve On Fri, 2017-08-25 at 09:46 -0700, Steve Ellcey wrote: > My earlier patch to update tests and resolve PR tree- > optimization/80925 > did not include FORTRAN, just C and C++.  

Re: [Patch, fortran] PR82173 (PDT) - [meta-bug] Parameterized derived type errors

Dear Dominique, That error is perfectly correct. Change the order of the assignment and the declaration for 'b' and you will see that all is well. The matching of type parameter specification list follows the same rules as those of actual arguments, except that deferred and assumed expressions

Re: [RFC, vectorizer] Allow single element vector types for vector reduction operations

I've been able to build spec again with this patch. Thanks Richard! Tamar. From: Richard Biener Sent: Tuesday, September 12, 2017 5:36:10 PM To: Andreas Schwab Cc: Tamar Christina; Vidya Praveen; Andrew Pinski; Jon Beniston;

Re: [PATCH v2 11/13] D: GCC builtins and runtime support.

On 09/11/2017 12:04 PM, Iain Buclaw wrote: > On 11 September 2017 at 18:40, Jeff Law wrote: >> On 06/24/2017 11:53 AM, Iain Buclaw wrote: >>> On 28 May 2017 at 23:17, Iain Buclaw wrote: This patch adds GCC builtins and runtime support for GDC

Re: [PATCH 3/13] D: The front-end (GDC) changelogs.

On 09/11/2017 11:39 AM, Iain Buclaw wrote: > On 11 September 2017 at 17:43, Jeff Law wrote: >> On 05/28/2017 03:11 PM, Iain Buclaw wrote: >>> This patch just includes all changelogs for the D front-end (GDC), >>> going back to the dawn of time itself. >>> >>> Change logs for the

Re: [PATCH 1/3] improve detection of attribute conflicts (PR 81544)

On 08/17/2017 10:03 AM, Martin Sebor wrote: > > First, there is the risk that someone will write code based > on the pure declaration even if there's no intervening call > to the function between it and the const one. Code tends to > be sloppy, and it's also not uncommon to declare the same >

Re: [PATCH] Fix PR 81096 (ttest failures)

On Tue, 2017-09-12 at 09:39 -0700, Ian Lance Taylor wrote: > On Tue, Sep 12, 2017 at 3:59 AM, Wilco Dijkstra om> wrote: > > > > Steve Ellcey wrote: > > > > > > This patch fixes the ttest failures on aarch64 by adding > > > AM_CFLAGS to > > > the test options, like btest

Re: [PATCH version 2, rs6000] Add builtins to convert from float/double to int/long using current rounding mode

Hi Carl, On Fri, Sep 08, 2017 at 11:41:13AM -0700, Carl Love wrote: > The following patch adds support for a couple of requested builtins that > convert from float/double to int / long using the current rounding > mode. I initially posted an early version of this patch which generated >

Re: [C PATCH] field_decl_cmp

On Tue, 12 Sep 2017, Nathan Sidwell wrote: > On 09/12/2017 12:06 PM, Joseph Myers wrote: > > On Tue, 12 Sep 2017, Nathan Sidwell wrote: > > > > This patch removes that checking, and also asserts that when we see > > > identically named decls, exactly one is a TYPE_DECL. > > > > When do you get

Re: [PATCH] Fix PR 81096 (ttest failures)

On Tue, 2017-09-12 at 09:39 -0700, Ian Lance Taylor wrote: > On Tue, Sep 12, 2017 at 3:59 AM, Wilco Dijkstra om> wrote: > > > > Steve Ellcey wrote: > > > > > > This patch fixes the ttest failures on aarch64 by adding > > > AM_CFLAGS to > > > the test options, like btest

Re: [C PATCH] field_decl_cmp

On 09/12/2017 12:06 PM, Joseph Myers wrote: On Tue, 12 Sep 2017, Nathan Sidwell wrote: This patch removes that checking, and also asserts that when we see identically named decls, exactly one is a TYPE_DECL. When do you get TYPE_DECLs here, for C? I wouldn't expect them to be possible.

Re: [PATCH] Fix PR 81096 (ttest failures)

On Tue, Sep 12, 2017 at 3:59 AM, Wilco Dijkstra wrote: > Steve Ellcey wrote: >> This patch fixes the ttest failures on aarch64 by adding AM_CFLAGS to >> the test options, like btest already does and as Wilco says works for >> him in Comment #4 of the bug report. > > Thanks

Re: [RFC, vectorizer] Allow single element vector types for vector reduction operations

On September 12, 2017 6:25:32 PM GMT+02:00, Andreas Schwab wrote: >On Sep 12 2017, Richard Biener wrote: > >> Index: gcc/tree-vect-generic.c >> === >> --- gcc/tree-vect-generic.c (revision

Re: Add support to trace comparison instructions and switch statements

On Tue, Sep 12, 2017 at 7:32 AM, Dmitry Vyukov wrote: > On Thu, Sep 7, 2017 at 9:02 AM, 吴潍浠(此彼) wrote: >> Hi >> The trace-div and trace-gep options seems be used to evaluate corpus >> to trigger specific kind of bugs. And they don't have strong effect to

Re: [AArch64, PATCH] Improve Neon store of zero

On Wed, Sep 06, 2017 at 10:02:52AM +0100, Jackson Woodruff wrote: > Hi all, > > I've attached a new patch that addresses some of the issues raised with > my original patch. > > On 08/23/2017 03:35 PM, Wilco Dijkstra wrote: > > Richard Sandiford wrote: > >> > >> Sorry for only noticing now, but

Re: [RFC, vectorizer] Allow single element vector types for vector reduction operations

On Sep 12 2017, Richard Biener wrote: > Index: gcc/tree-vect-generic.c > === > --- gcc/tree-vect-generic.c (revision 251997) > +++ gcc/tree-vect-generic.c (working copy) > @@ -1638,7 +1638,11 @@

Re: [PATCH 1/2] Don't warn function alignment if warn_if_not_aligned_p is true

On Mon, 21 Aug 2017, H.J. Lu wrote: > When warn_if_not_aligned_p is true, a warning will be issued on function > declaration later. There is no need to warn function alignment when > warn_if_not_aligned_p is true. > > OK for trunk? > > H.J. > -- > * c-attribs.c

Re: [PATCH] Add -static-pie to GCC driver to create static PIE

On Mon, 28 Aug 2017, H.J. Lu wrote: > Here is the updated patch. OK for trunk? OK. -- Joseph S. Myers jos...@codesourcery.com

Re: [PATCH] detect incompatible aliases (PR c/81854)

On Thu, 17 Aug 2017, Martin Sebor wrote: > + || (prototype_p (t1) > + && prototype_p (t2) > + && !types_compatible_p (t1, t2 Why the restriction to prototyped types? I'd expect a warning for an alias between unprototyped functions of types

Re: [PATCH][AArch64] Remove '*' from movsi/di/ti patterns

On Wed, Jul 26, 2017 at 02:46:14PM +0100, Wilco Dijkstra wrote: > Remove the remaining uses of '*' from the movsi/di/ti patterns. > Using '*' in alternatives is typically incorrect at it tells the register > allocator to ignore those alternatives. So remove these from all the > integer move

Re: [PATCH version 2, rs6000] Add builtins to convert from float/double to int/long using current rounding mode

On 09/08/2017 01:41 PM, Carl Love wrote: > +(define_insn "fctid" > + [(set (match_operand:DI 0 "gpc_reg_operand" "=d") > +(unspec:DI [(match_operand:DF 1 "gpc_reg_operand" "d")] > + UNSPEC_FCTID_INST))] > + "" > + "fctid %0,%1") > + > +(define_insn "fctiw" > + [(set

Re: [C PATCH] field_decl_cmp

On Tue, 12 Sep 2017, Nathan Sidwell wrote: > Joseph, > in moving field_decl_cmp to the C FE, I noticed it checks for NULL DECL_NAMES. > Those don't occur. To be clear: they don't occur in the case where field_decl_cmp is used; they can occur in other cases. > This patch removes that checking,

RE: 0001-Part-1.-Add-generic-part-for-Intel-CET-enabling

> -Original Message- > From: Jeff Law [mailto:l...@redhat.com] > Sent: Friday, August 25, 2017 10:50 PM > To: Tsimbalist, Igor V ; 'gcc- > patc...@gcc.gnu.org' > Subject: Re: 0001-Part-1.-Add-generic-part-for-Intel-CET-enabling > >

Re: [PATCH 1/3] improve detection of attribute conflicts (PR 81544)

On 08/09/2017 01:55 PM, Joseph Myers wrote: > On Wed, 9 Aug 2017, Martin Sebor wrote: > >> the same function with the other of this pair attributes. I'd >> also be okay with not diagnosing this combination if I could >> convince myself that it's safe (or can be made safe) and treated >>

Re: [PATCH 1/3] improve detection of attribute conflicts (PR 81544)

On 08/10/2017 03:55 PM, Joseph Myers wrote: > On Wed, 9 Aug 2017, Martin Sebor wrote: > >> The problem isn't that the declarations aren't merged at the call >> site but rather that the middle-end gives const precedence over >> pure so when both attributes are provided the former wins. > > But

RE: 0001-Part-1.-Add-generic-part-for-Intel-CET-enabling

> -Original Message- > From: Jeff Law [mailto:l...@redhat.com] > Sent: Friday, August 25, 2017 10:32 PM > To: Richard Biener ; Tsimbalist, Igor V > > Cc: gcc-patches@gcc.gnu.org > Subject: Re:

RE: 0001-Part-1.-Add-generic-part-for-Intel-CET-enabling

> -Original Message- > From: Tsimbalist, Igor V > Sent: Friday, August 18, 2017 4:43 PM > To: 'Richard Biener' > Cc: gcc-patches@gcc.gnu.org; Tsimbalist, Igor V > > Subject: RE: 0001-Part-1.-Add-generic-part-for-Intel-CET-enabling

Re: [PATCH]: PR target/80204 (Darwin macosx-version-min problem)

On 09/04/2017 01:48 PM, Jakub Jelinek wrote: > On Mon, Sep 04, 2017 at 08:47:07PM +0100, Simon Wright wrote: >> On 1 Sep 2017, at 23:05, Simon Wright wrote: >>> >>> 2017-09-01 Simon Wright >>> >>> PR target/80204 >>> * config/darwin-driver.c

[C PATCH] field_decl_cmp

Joseph, in moving field_decl_cmp to the C FE, I noticed it checks for NULL DECL_NAMES. Those don't occur. This patch removes that checking, and also asserts that when we see identically named decls, exactly one is a TYPE_DECL. ok? nathan -- Nathan Sidwell 2017-09-12 Nathan Sidwell

Re: [PATCH, rs6000] Folding of vector loads in GIMPLE

> On Sep 12, 2017, at 9:41 AM, Will Schmidt wrote: > > Hi > > [PATCH, rs6000] Folding of vector loads in GIMPLE > > Folding of vector loads in GIMPLE. > > - Add code to handle gimple folding for the vec_ld builtins. > - Remove the now obsoleted folding code for

Re: [PATCH] Fix emission of exception dispatch (PR middle-end/82154).

On 09/12/2017 01:43 AM, Martin Liška wrote: > Hello. > > In transition to simple_case_node, I forgot to initialize m_high to m_low if > a case > does not have CASE_HIGH. > > Patch can bootstrap on ppc64le-redhat-linux and survives regression tests. > > Ready to be installed? > Martin > >

Re: [PATCH] Fix powerpc ICE with __builtin_vec_ld on an array (PR target/82112, take 2)

> On Sep 12, 2017, at 10:00 AM, Segher Boessenkool > wrote: > > Hi Jakub, > > On Tue, Sep 12, 2017 at 04:25:48PM +0200, Jakub Jelinek wrote: >> On Thu, Sep 07, 2017 at 10:40:30AM +0200, Jakub Jelinek wrote: >>> The C and C++ FE handle resolve_overloaded_builtin

Re: [PATCH, rs6000] testcase coverage for vector load builtins

On Tue, Sep 12, 2017 at 09:36:41AM -0500, Will Schmidt wrote: > [PATCH, rs6000] testcase coverage for vector load builtins > Add testcase coverage for the vec_ld intrinsic builtins. > > Tested across power platforms (p6 and newer). OK for trunk? Looks good, please apply. Thanks! Segher [

Re: [PATCH] Fix powerpc ICE with __builtin_vec_ld on an array (PR target/82112, take 2)

On Tue, Sep 12, 2017 at 10:00:44AM -0500, Segher Boessenkool wrote: > > --- gcc/testsuite/g++.dg/ext/altivec-18.C.jj2017-09-08 > > 09:15:20.593774717 +0200 > > +++ gcc/testsuite/g++.dg/ext/altivec-18.C 2017-09-08 09:15:20.593774717 > > +0200 > > @@ -0,0 +1,14 @@ > > +// PR target/82112

Re: [PATCH] Fix powerpc ICE with __builtin_vec_ld on an array (PR target/82112, take 2)

Hi Jakub, On Tue, Sep 12, 2017 at 04:25:48PM +0200, Jakub Jelinek wrote: > On Thu, Sep 07, 2017 at 10:40:30AM +0200, Jakub Jelinek wrote: > > The C and C++ FE handle resolve_overloaded_builtin differently, the C FE > > calls it when e.g. array-to-pointer and function-to-pointer conversions > >

[Patch AArch64 2/2] Fix memory sizes to load/store patterns

On Mon, Jun 12, 2017 at 02:54:00PM +0100, James Greenhalgh wrote: > > Hi, > > There seems to be a partial misconception in the AArch64 backend that > load1/load2 referred to the number of registers to load, rather than the > number of words to load. This patch fixes that using the new "number of

Re: [Mechanical Patch ARM/AArch64 1/2] Rename load/store scheduling types to encode data size

On Mon, Jun 12, 2017 at 02:53:59PM +0100, James Greenhalgh wrote: > > Hi, > > In the AArch64 backend and scheduling models there is some confusion as to > what the load1/load2 etc. scheduling types refer to. This leads to us using > load1/load2 in two contexts - for a variety of 32-bit, 64-bit

Re: C++ PATCH to reduced_constant_expression for partially-initialized objects

On 12 September 2017 at 16:21, Andrew Pinski wrote: > On Tue, Sep 12, 2017 at 7:17 AM, Christophe Lyon > wrote: >> Hi Jason >> >> On 10 September 2017 at 11:09, Jason Merrill wrote: >>> A few months back I queued this patch to

Re: Add support to trace comparison instructions and switch statements

Some stats from kernel build for number of trace_cmp callbacks: gcc non-const: 38051 const: 272726 total: 310777 clang: non-const: 45944 const: 266299 total: 312243 The total is quite close. Gcc seems to emit more const callbacks, which is good. On Tue, Sep 12, 2017 at 4:32 PM, Dmitry Vyukov

Re: [PATCH] Fix PR82129

On Tue, 12 Sep 2017, Jakub Jelinek wrote: > On Tue, Sep 12, 2017 at 04:09:41PM +0200, Richard Biener wrote: > > > > The following "fixes" PR82129 (hides the issue). It folds > > writes from uninitialized data to CLOBBERs: > > > > - *h5_26(D) = tv_24(D); > > + *h5_26(D) ={v} {CLOBBER}; > > >

Re: [PATCH, rs6000] add vectorization to vec_mule and vec_mulo builtins

Hi Carl, Just some nits: On Mon, Sep 11, 2017 at 09:18:12AM -0700, Carl Love wrote: > +(define_expand "vec_widen_smult_odd_v4si" > + [(use (match_operand:V2DI 0 "register_operand" "")) > + (use (match_operand:V4SI 1 "register_operand" "")) > + (use (match_operand:V4SI 2 "register_operand"

[PATCH, rs6000] Folding of vector loads in GIMPLE

Hi [PATCH, rs6000] Folding of vector loads in GIMPLE Folding of vector loads in GIMPLE. - Add code to handle gimple folding for the vec_ld builtins. - Remove the now obsoleted folding code for vec_ld from rs6000-c.c. Surrounding comments have been adjusted slightly so they continue to read

[PATCH, rs6000] testcase coverage for vector load builtins

Hi, [PATCH, rs6000] testcase coverage for vector load builtins Add testcase coverage for the vec_ld intrinsic builtins. Tested across power platforms (p6 and newer). OK for trunk? Thanks, -Will [gcc/testsuite] 2017-09-12 Will Schmidt *

Re: [PATCH] Fix GIMPLE FE test (PR testsuite/82114)

On 09/12/2017 02:29 PM, Aldy Hernandez wrote: > I'm going to abuse either my *gimpl* maintainer hat or my obvious hat > and approve this. > > However, since we already have another test that looks the same > (gimplefe-error-2.c) but without the 0 case, it would probably be best > to add a comment

Re: Add support to trace comparison instructions and switch statements

On Thu, Sep 7, 2017 at 9:02 AM, 吴潍浠(此彼) wrote: > Hi > The trace-div and trace-gep options seems be used to evaluate corpus > to trigger specific kind of bugs. And they don't have strong effect to > coverage. > > The trace-pc-guard is useful, but it may be much more complex

[PATCH] move sorted field handling to C FE

Now that I've excised sorted fields from the C++ FE (as a distinct concept), we can move some helper functions from c-family/ to c/ this patch does exactly that. Applied to trunk. nathan -- Nathan Sidwell 2017-09-12 Nathan Sidwell c-family/ * c-common.c (field_decl_cmp,

Re: [RFC][PATCH] Do refactoring of attribute functions and move them to attribs.[hc].

On 09/12/2017 01:39 PM, Jakub Jelinek wrote: > On Tue, Sep 12, 2017 at 01:31:47PM +0200, Martin Liška wrote: >> >From a40c06fc06afcb7bb886d7a3106e6da631a48430 Mon Sep 17 00:00:00 2001 >> From: marxin >> Date: Tue, 12 Sep 2017 13:30:39 +0200 >> Subject: [PATCH] Reduce

[PATCH] Fix powerpc ICE with __builtin_vec_ld on an array (PR target/82112, take 2)

Hi! On Thu, Sep 07, 2017 at 10:40:30AM +0200, Jakub Jelinek wrote: > The C and C++ FE handle resolve_overloaded_builtin differently, the C FE > calls it when e.g. array-to-pointer and function-to-pointer conversions > are already done on the arguments, while C++ FE does that only much later. >

Re: C++ PATCH to reduced_constant_expression for partially-initialized objects

On Tue, Sep 12, 2017 at 7:17 AM, Christophe Lyon wrote: > Hi Jason > > On 10 September 2017 at 11:09, Jason Merrill wrote: >> A few months back I queued this patch to bring back for GCC 8. >> Unfortunately I don't remember the context that it came up

Re: C++ PATCH to reduced_constant_expression for partially-initialized objects

Hi Jason On 10 September 2017 at 11:09, Jason Merrill wrote: > A few months back I queued this patch to bring back for GCC 8. > Unfortunately I don't remember the context that it came up in, but it > affects for instance cases of self-assignment, which can't have a > constant

Re: [PATCH] Fix PR82129

On Tue, Sep 12, 2017 at 04:09:41PM +0200, Richard Biener wrote: > > The following "fixes" PR82129 (hides the issue). It folds > writes from uninitialized data to CLOBBERs: > > - *h5_26(D) = tv_24(D); > + *h5_26(D) ={v} {CLOBBER}; > > Bootstrap & regtest running on x86_64-unknown-linux-gnu.

[PATCH] Fix PR82129

The following "fixes" PR82129 (hides the issue). It folds writes from uninitialized data to CLOBBERs: - *h5_26(D) = tv_24(D); + *h5_26(D) ={v} {CLOBBER}; Bootstrap & regtest running on x86_64-unknown-linux-gnu. Richard. 2017-09-12 Richard Biener PR

Re: [PATCH] PR libstdc++/79433 no #error for including headers with wrong -std

On 07/09/17 15:18 +0100, Jonathan Wakely wrote: As discussed in PR 79433, the recommended way to test for new features such as std::optional has problems. The current version of SD-6 at https://isocpp.org/std/standing-documents/sd-6-sg10-feature-test-recommendations says to simply check

Re: [PATCH][PR sanitizer/77631] Support separate debug info in libbacktrace

On Mon, Sep 11, 2017 at 3:06 AM, Denis Khalikov wrote: > Thanks for answer. > I understood all points which you mentioned, but can't > find last one >> It seems to work >> out the file name a second time, even though the file name must >> already be known. > > Can

[PATCH] Fix PR82128

This is the one with the folding fix... Bootstrapped on x86_64-unknown-linux-gnu, testing in progress. Richard. 2017-09-12 Richard Biener PR middle-end/82128 * gimple-fold.c (gimple_fold_call): Update iterator properly. Index: gcc/gimple-fold.c

[PATCH] Fix PR82157

The following fixes a latent issue in fold_stmt uncovered by PRE re-org. We muck with the gsi in inapprorpate ways. Bootstrapped on x86_64-unknown-linux-gnu, testing in progress. Richard. 2017-09-12 Richard Biener PR tree-optimization/82157 *

[PATCH] Fix PR82144

The following avoids adding DW_AT_alignment twice by not doing it for incomplete types. Bootstrapped on x86_64-unknown-linux-gnu, testing in progress. Alex, is that ok or do we want DW_AT_alignment for incomplete types as well? Thanks, Richard. 2017-09-12 Richard Biener

Re: [PATCH] PR libstdc++/70483 make std::string_view fully constexpr

On 11/09/17 22:48 +0100, Jonathan Wakely wrote: This adds 'constexpr' everywhere it's missing from std::basic_string_view. And this does it for std::experimental::basic_string_view, fixing the bug as reported. Tested powerpc64le-linux, committed to trunk. commit

[PATCH] Fix typo in recent match.pd pattern

Bootstrapped and tested on x86_64-unknown-linux-gnu, applied. Richard. 2017-09-12 Richard Biener PR middle-end/82149 * match.pd ((FTYPE) N CMP CST): Fix typo. Index: gcc/match.pd === ---

[gotools] Fix some gotools testing problems

A couple of gotools test FAIL on Solaris, and there are several issues that make investigation particularly tedious. * The one invocation of gotest doesn't support passing additional flags (--keep in particular). Added below. * The order of some of the summaries isn't stable (cmd/go and

[C++ PATCH] Kill SORTED_FIELDS

This patch kills the SORTED_FIELDS vector from the C++ FE. After a type is completed, we now hold everthing on METHOD_VEC, and thus only need 1 binary search, rather than 2 (or worse, under certain conditions). As I said at the Cauldron, my earlier attempt to approach this from a different

[libgo] Support 32-bit SPARC relocs

A couple of gotools tests were FAILing on 32-bit Solaris/SPARC like this: FAIL: TestCgoHandlesWlORIGIN go_test.go:267: running testgo [build origin] go_test.go:286: standard error: go_test.go:287: # origin cannot load DWARF output from

RE: [PATCH] Add 'short_call' attribute for MIPS targets

Simon Atanasyan writes: > On Mon, Sep 11, 2017 at 03:26:52PM +, Matthew Fortune wrote: > > Simon Atanasyan writes: > > > Here is the updated patch with chnaged e-mail address and fixed > > > indentation issues: > > > -8< > > >

Re: [Patch, fortran] PR82173 (PDT) - [meta-bug] Parameterized derived type errors

Dear Paul, I have been playing with the patch and found that the following variant of the test for pr82168 module mod implicit none integer, parameter :: dp = kind (0.0d0) type, public :: v(z, k) integer, len :: z integer, kind :: k = kind(0.0) real(kind = k) :: e(z) end type v end

Re: [PATCH] Add 'short_call' attribute for MIPS targets

On Mon, Sep 11, 2017 at 03:26:52PM +, Matthew Fortune wrote: > Simon Atanasyan writes: > > Here is the updated patch with chnaged e-mail address and fixed > > indentation issues: > > -8< > > Currently GCC supports 'long_call', 'far', and 'near'

Re: [PATCH] Fix GIMPLE FE test (PR testsuite/82114)

I'm going to abuse either my *gimpl* maintainer hat or my obvious hat and approve this. However, since we already have another test that looks the same (gimplefe-error-2.c) but without the 0 case, it would probably be best to add a comment in gimplefe-14.c as to why there is a 0 case. In the

[C++ PATCH] small METHOD_VEC cleanup

It turned out I wasn't quite at the point of committing my sorted_fields killing patch. This one just cleans up some functions affected by that, and I've applied it separately to reduce the churn. nathan -- Nathan Sidwell 2017-09-12 Nathan Sidwell * name-lookup.c

Re: [RFC][PATCH] Do refactoring of attribute functions and move them to attribs.[hc].

On Tue, Sep 12, 2017 at 01:31:47PM +0200, Martin Liška wrote: > >From a40c06fc06afcb7bb886d7a3106e6da631a48430 Mon Sep 17 00:00:00 2001 > From: marxin > Date: Tue, 12 Sep 2017 13:30:39 +0200 > Subject: [PATCH] Reduce lookup_attribute memory footprint. > > gcc/ChangeLog: > >

[libgfortran] Replace implicit conversions between enums in io/transfer.c by explicit casts.

Hi, I am working on patch for PR78736 (https://gcc.gnu.org/ml/gcc-patches/2017-09/msg00011.html), which adds a new warning -Wenum-conversion to C front-end to warn for implicit conversion between different enums. The warning in that patch triggered on io/transfer.c for following implicit

Re: [RFC][PATCH] Do refactoring of attribute functions and move them to attribs.[hc].

On 09/12/2017 09:54 AM, Jakub Jelinek wrote: > On Thu, Jul 13, 2017 at 03:51:31PM +0200, Martin Liška wrote: >> It's request for comment where I mechanically moved attribute-related >> function to attribs.[hc]. >> >> Patch can bootstrap on ppc64le-redhat-linux and survives regression tests. >> >>

[PATCH] [ARC] Check the assembler for gdwarf2 support.

From: claziss This small patch enables the gcc driver to pass dwarf related options to the assembler. Ok to apply? Claudiu gcc/ 2017-06-21 Claudiu Zissulescu * configure.ac: Add arc and check if assembler supports gdwarf2. *

RE: [PATCH] Fix PR 81096 (ttest failures)

Steve Ellcey wrote: > This patch fixes the ttest failures on aarch64 by adding AM_CFLAGS to > the test options, like btest already does and as Wilco says works for > him in Comment #4 of the bug report. Thanks for picking this up, this looks OK. > Tested by me on aarch64. Ok to checkin? This

RE: [RFC, vectorizer] Allow single element vector types for vector reduction operations

On Tue, 12 Sep 2017, Richard Biener wrote: > On Tue, 12 Sep 2017, Tamar Christina wrote: > > > Hi Jon, Richard, > > > > > > > > > > > > Testing the following instead: > > > > > > > > Any news on this? > > > > > > Didn't work out as expected. I think the logic in tree-vect-generic is > > >

Re: [patch, fortran, RFC] warn about out-of-bounds errors in DO loops

Thanks for doing this, looks really useful. Thomas Koenig writes: > Well, here's a version which actually throws a hard error in > obvious cases; the other cases are reserved for -Wextra. Is it OK to throw a hard error for this? Maybe the rules are different from C and

[Ada] Enforce check on wrong usage names in aspect definitions

The frontend does not report an error on aspect clases defined in the public part of a nested package that reference names defined in the private part. After this patch the error is reported on the following sources: pragma Ada_2012; with Ada.Text_IO; use Ada.Text_IO; procedure Small is

[Ada] Spurious warning in Containers.Bounded_Doubly_Linked_Lists

This patch fixes a bug in which an instantiation of Containers.Bounded_Doubly_Linked_Lists can give warnings if the actual type passed to Element_Type is not fully default-initialized. The following test should compile quietly. with Ada.Containers.Bounded_Doubly_Linked_Lists; procedure Test is

[Ada] Ignore both Global and Refined_Global in CodePeer mode

Only the first pragma was ignored in CodePeer mode, to allow analysis of legacy code, which caused errors when the second was also present. Now both are ignored in CodePeer mode. Tested on x86_64-pc-linux-gnu, committed on trunk 2017-09-12 Yannick Moy * gnat1drv.adb

[Ada] Dynamically tagged expr of expr function is illegal

It is illegal to return a dynamically tagged expression from an expression function that returns a specific type. See RM-3.9.2(9). This patch fixes a bug that caused the compiler to fail to detect this illegality. Tested on x86_64-pc-linux-gnu, committed on trunk 2017-09-12 Bob Duff

RE: [RFC, vectorizer] Allow single element vector types for vector reduction operations

On Tue, 12 Sep 2017, Tamar Christina wrote: > Hi Jon, Richard, > > > > > > > > > Testing the following instead: > > > > > > Any news on this? > > > > Didn't work out as expected. I think the logic in tree-vect-generic is the > > one > > to be fixed but I have to carve out some time to look

  1   2   >