Re: [2/5] C-SKY port v3: Backend implementation

On 08/05/2018 10:34 PM, Sandra Loosemore wrote: > > csky-gcc-2.log > > > 2018-08-05 Jojo > Huibin Wang > Sandra Loosemore > Chung-Lin Tang > > C-SKY port: Backend implementation > > gcc/ > * config/csky/*: New. > *

Re: [PATCH] DWARF: add DW_AT_count to zero-length arrays

On Thu, Aug 16, 2018 at 9:29 PM Omar Sandoval wrote: > > Hi, > > This fixes the issue that it is impossible to distinguish a zero-length array > type from a flexible array type given the DWARF produced by GCC (which I > reported here [1]). We do so by adding a DW_AT_count attribute with a value

Re: PING [PATCH] warn for strlen of arrays with missing nul (PR 86552, 86711, 86714) )

On 08/01/2018 08:44 PM, Martin Sebor wrote: > Since the foundation of the patch is detecting and avoiding > the overly aggressive folding of unterminated char arrays, > besides issuing a warning for such arguments to strlen, > the patch also fixes pr86711 - wrong folding of memchr, and > pr86714 -

Re: [PATCH] Make strlen range computations more conservative

On 08/09/2018 12:27 AM, Richard Biener wrote: > On August 9, 2018 7:26:19 AM GMT+02:00, Jeff Law wrote: >> On 07/24/2018 05:18 PM, Bernd Edlinger wrote: >>> On 07/24/18 23:46, Jeff Law wrote: On 07/24/2018 01:59 AM, Bernd Edlinger wrote: > Hi! > > This patch makes strlen range

Re: [PATCH, Darwin] Move fixed _Unwind_find_Enclosing_func to a crt.

On 08/15/2018 12:50 PM, Iain Sandoe wrote: > Hi, > > This is a pre-cursor to other tidy-ups aimed at being able to dispense with > the “ext” library that we introduced (a lng time ago) to deal with the > differences between Darwin’s installed libgcc_s and the current compiler. > > Since

Re: [PATCH] Check the STRING_CSTs in varasm.c

On 08/05/2018 04:28 AM, Bernd Edlinger wrote: > Hi, > > I would like to do a minor tweak to the patch. > While staring at the other patch I realized that I should > better pass size and not thissize to the check > function, instead of making use of how thissize is > computed using MIN above.

[PATCH] DWARF: add DW_AT_count to zero-length arrays

Hi, This fixes the issue that it is impossible to distinguish a zero-length array type from a flexible array type given the DWARF produced by GCC (which I reported here [1]). We do so by adding a DW_AT_count attribute with a value of zero only for zero-length arrays (this is what clang does in

Re: [PATCH] v2: Formatted printing for dump_* in the middle-end

On 08/02/2018 11:54 AM, David Malcolm wrote: > On Tue, 2018-07-31 at 19:56 +, Joseph Myers wrote: >> On Tue, 31 Jul 2018, David Malcolm wrote: >> >>> I didn't exhaustively check every callsite to the changed calls; >>> I'm >>> assuming that -Wformat during bootstrap has effectively checked >>>

Re: [PATCH] assume sprintf formatting of wide characters may fail (PR 86853)

On 08/04/2018 12:46 PM, Martin Sebor wrote: > The sprintf handling of wide characters neglects to consider > that calling the function may fail due to a conversion error > (when the wide character is invalid or not representable in > the current locale).  The handling also misinterprets > the 

C++ PATCH for c++/67012, c++/86942, detect invalid cases with function return type deduction

As I promised in , this patch fixes a couple of invalid cases we weren't detecting. It's got testcases from two PRs and another case I found out; they're intertwined so I think it makes sense to fix them in one go. Bootstrapped/regtested

Re: [PATCH] Add a character size parameter to c_strlen/get_range_strlen

On 08/16/2018 05:01 PM, Joseph Myers wrote: > On Thu, 16 Aug 2018, Jeff Law wrote: > >> restores previous behavior. The sprintf bits want to count element >> sized chunks, which for wchars is 4 bytes (that count will then be > >>/* Compute the range the argument's length can be in. */ >> -

Re: [PATCH, RFC, rs6000, v2] folding of vec_splat

On Thu, 2018-08-16 at 15:51 -0500, Segher Boessenkool wrote: > Hi Will, > > On Thu, Aug 16, 2018 at 10:50:45AM -0500, Will Schmidt wrote: > > 2018-08-16 Will Schmidt > > > > * config/rs6000/rs6000.c (rs6000_gimple_fold_builtin): Add support for > > early gimple folding of

Re: [PATCH] Add a character size parameter to c_strlen/get_range_strlen

On Thu, 16 Aug 2018, Jeff Law wrote: > restores previous behavior. The sprintf bits want to count element > sized chunks, which for wchars is 4 bytes (that count will then be >/* Compute the range the argument's length can be in. */ > - fmtresult slen = get_string_length (arg); > + int

Re: [PATCH] assume sprintf formatting of wide characters may fail (PR 86853)

On 08/09/2018 01:29 PM, Bernd Edlinger wrote: >>> + bool one_2_one_ascii >>> + = (target_to_host_charmap[0] == 1 && target_to_host ('a') == >>> 97); >> Hmm. Is this really sufficient?I have nowhere near enough knowledge >> of the potential target character sets to know if

Re: [PATCH] Add a character size parameter to c_strlen/get_range_strlen

On 08/14/2018 04:25 AM, Bernd Edlinger wrote: > 2018-08-14 Bernd Edlinger > > * builtins.c (c_strlen): Add new parameter eltsize. > * builtins.h (c_strlen): Adjust prototype. > * expr.c (string_constant): Add new parameter mem_size. > * expr.h (string_constant): Adjust

[committed] diagnostics: fix bad interaction between line spans and line numbers

Without this patch, the "line span" markers and the line numbering interacted badly, leading to stray copies of the line-span markers appearing as prefixes on the first source line in a span: missing-header-fixit-3.c: In function 'test': missing-header-fixit-3.c:9:3: warning: implicit declaration

[committed] diagnostics: tweak to line-insertion fix-it hints with line-numbering

This commit slightly tweaks line-insertion fix-it hints, so that with line-numbering, rather than e.g.: 99 | x = a; |+ break; 110 | case 'b': | ^~~~ we fill the margin with "+": 99 | x = a; +++ |+ break; 110 | case 'b': | ^~~~ to

Re: [ARM/FDPIC v2 00/21] FDPIC ABI for ARM

Ping? Le mer. 1 août 2018 à 10:03, Christophe Lyon a écrit : > Ping? > > > On 13/07/2018 18:10, christophe.l...@st.com wrote: > > From: Christophe Lyon > > > > Hello, > > > > This patch series implements the GCC contribution of the FDPIC ABI for > > ARM targets. > > > > This ABI enables to run

Re: [PATCH] Add a character size parameter to c_strlen/get_range_strlen

On 08/16/18 23:27, Jeff Law wrote: > On 08/16/2018 12:47 PM, Bernd Edlinger wrote: > > Parameterizing the function to return either the number of > bytes or the number of elements makes sense as an enhancement. > It makes less sense (and could be the source of bugs) to let >

Re: [PATCH] Add a character size parameter to c_strlen/get_range_strlen

On 08/16/2018 12:47 PM, Bernd Edlinger wrote: Parameterizing the function to return either the number of bytes or the number of elements makes sense as an enhancement. It makes less sense (and could be the source of bugs) to let callers pass in anything else.  A boolean flag

Re: [PATCH, RFC, rs6000, v2] folding of vec_splat

Hi Will, On Thu, Aug 16, 2018 at 10:50:45AM -0500, Will Schmidt wrote: > 2018-08-16 Will Schmidt > > * config/rs6000/rs6000.c (rs6000_gimple_fold_builtin): Add support for > early gimple folding of vec_splat(). Continuation lines should be indented to the *, not to the text

Re: [PATCH, Darwin, drivers] Split DWARF is not supported on Darwin.

On Aug 16, 2018, at 6:55 AM, Iain Sandoe wrote: > > The Darwin toolchains have a separate debug linker (dsymutil) so that the > link-time penalty for debug data is not usually seen. At present, it's not > clear how we would support split DWARF on Darwin (or if it would bring > any additional

[PATCH] RFC: reduce likelihood of fully-dynamic-string throwing on move

With --enable-fully-dynamic-string the COW basic_string move constructor is noexcept(false), because it has to allocate a new empty rep for the moved-from string. If we did this, it would only throw when the string we're moving from is "leaked" (that is, there are potentially pointers,

[PATCH] Macro definition parameter parsing

This next patch in the macro cleanup changes the internal interface to the parameter parsing logic. Rather than pass a macro pointer in, we pass explicit variable pointers in. This'll allow later creation of the macro object itself. While there, I cleaned up the parsing logic for more

[PATCH, testsuite]: Loosen scan-assembler strings in gcc.target/i386/avx{,2}-cvt-2.c

Hello! These instructions can take memory operands and current scan-assembler strings were too tight to accept them. 2018-08-16 Uros Bizjak PR testsuite/86745 * gcc.target/i386/avx-cvt-2.c: Loosen scan-assembler strings. * gcc.target/i386/avx2-cvt-2.c: Ditto. Tested on

Re: [PATCH] Add a character size parameter to c_strlen/get_range_strlen

On 08/16/18 19:35, Jeff Law wrote: > On 08/15/2018 12:51 PM, Bernd Edlinger wrote: > [ snip -- comment attribution is likely lost... ] > > > 2018-08-14  Bernd Edlinger  > > * builtins.c (c_strlen): Add new parameter eltsize. > * builtins.h (c_strlen): Adjust

Re: [PATCH] [C] Warn when calculating abs(unsigned_value)

On 08/16/2018 10:49 AM, Joseph Myers wrote: On Wed, 15 Aug 2018, Martin Sebor wrote: Detecting some of these bugs without too much noise would require moving the warning out of the front-end and to some later point after VRP has run. But you need the information about whether the conversion

Re: [PATCH] Add a character size parameter to c_strlen/get_range_strlen

On 08/15/2018 12:51 PM, Bernd Edlinger wrote: [ snip -- comment attribution is likely lost... ] 2018-08-14  Bernd Edlinger  * builtins.c (c_strlen): Add new parameter eltsize. * builtins.h (c_strlen): Adjust prototype. * expr.c (string_constant): Add new

[PATCH] Fix warning with -Wsign-compare -Wsystem-headers in __sph_legendre

Ed, I'm checking this in as it looks correct anyway, quite apart from the fact it fixes a warning, but could you please double check it? * include/tr1/legendre_function.tcc (__sph_legendre): Avoid warning about signed/unsigned comparison. Tested x86_64-linux, committed to trunk.

[PATCH 4/4] rs6000: Delete old add+cmp patterns

There are some patterns that recognise the parallel of an add and a compare, and split it back to the same two insns. This apparently helped RIOS machines before RTL scheduling existed? Either way, it isn't helpful anymore, and even hurts a tiny bit. So, delete it. 2018-08-16 Segher

[PATCH 3/4] rs6000: Remove "length 4" from other insns

There were many insns that set "length 4" explicitly while that does not make anything clearer to the reader. So, simplify the code. 2018-08-16 Segher Boessenkool * config/rs6000/altivec.md: Don't set length attribute to the default value. * config/rs6000/darwin.md:

[PATCH 2/4] rs6000: Remove "length 4" from branch insns

Now that it is the default for branch insns like for all other insns, we don't need to set it explicitly so often anymore. 2018-08-16 Segher Boessenkool * config/rs6000/altivec.md: Don't set length attribute to the default value, for branch instructions. *

[PATCH 1/4] rs6000: Change the length attribute default

This moves what is currently the default of the length attribute to the only branch instruction patterns where it applies, namely, the B-form instructions. It was used for the "jump" instruction as well before, but that is an I-form instruction and always has length 4. 2018-08-16 Segher

[PATCH 0/4] rs6000: length attribute

This series makes some changes to the length attribute. Currently, the length attribute for branch instructions defaults to what is needed for B-form instructions, that is, branches with an immediate 16-bit displacement field. There is only one such machine insn, and only three such RTL

[committed] -Wmisleading-indentation: fix ICE in get_visual_column (PR c++/70693)

PR c++/70693 reports a crash within -Wmisleading-indentation in get_visual_column, reading past the end of a source line. The issue occurs due to a stray carriage return aka '\r' aka ^M, occurring towards the end of line 35 of attachment 38289 - but not at the end itself. This carriage return

[PATCH] Avoid deprecation warning with -Wsystem-headers

C++17 says to use std::uncaught_exceptions() here instead of std::uncaught_exception() but since we only care whether the result is non-zero (and we aren't planning to remove the deprecated version) we can just keep using std::uncaught_exception() and suppress the warning. *

Re: [PATCH] Merge Ignore and Deprecated in .opt files.

On Thu, 16 Aug 2018, Martin Liška wrote: > On 08/15/2018 06:38 PM, Joseph Myers wrote: > > On Wed, 15 Aug 2018, Martin Liška wrote: > > > >> Ok, so you have very similar opinion as Jakub. Thus I'm sending new > >> version that preserves status quo, it only does: > > > > This is removing

Re: [PATCH] [C] Warn when calculating abs(unsigned_value)

On Wed, 15 Aug 2018, Martin Sebor wrote: > Detecting some of these bugs without too much noise would require > moving the warning out of the front-end and to some later point > after VRP has run. But you need the information about whether the conversion was explicit or implicit. If someone

Re: [RFC][PATCH][mid-end] Optimize immediate choice in comparisons.

On 08/16/2018 10:46 AM, Vlad Lazar wrote: >> Thanks.  I fixed up the ChangeLog entry and installed this on the trunk. >> >> Richard S -- thanks for working with Vlad to push this forward. >> >> jeff >> > Thanks for committing.  Sorry about the ChangeLog. No worries. Just trivial stuff that we

Re: [RFC][PATCH][mid-end] Optimize immediate choice in comparisons.

On 16/08/18 17:35, Jeff Law wrote: On 08/14/2018 11:01 AM, Vlad Lazar wrote: On 13/08/18 15:00, Richard Sandiford wrote: Vlad Lazar writes: diff --git a/gcc/expmed.h b/gcc/expmed.h index 2890d9c9bbd034f01030dd551d544bf73e73b784..86a32a643fdd0fc9f396bd2c7904244bd484df16 100644 ---

Re: [RFC][PATCH][mid-end] Optimize immediate choice in comparisons.

On 08/14/2018 11:01 AM, Vlad Lazar wrote: > On 13/08/18 15:00, Richard Sandiford wrote: >> Vlad Lazar writes: >>> diff --git a/gcc/expmed.h b/gcc/expmed.h >>> index >>> 2890d9c9bbd034f01030dd551d544bf73e73b784..86a32a643fdd0fc9f396bd2c7904244bd484df16 >>> 100644 >>> --- a/gcc/expmed.h >>> +++

[PATCH] Fix bootstrap with --enable-fully-dynamic-string

PR libstdc++/86447 * src/c++11/cow-stdexcept.cc [_GLIBCXX_FULLY_DYNAMIC_STRING] (logic_error::logic_error(logic_error&&)) (logic_error::operator=(logic_error&&)) (runtime_error::runtime_error(runtime_error&&))

[PATCH, RFC, rs6000, v2] folding of vec_splat

Hi Enable GIMPLE folding of the vec_splat() intrinsic. (v2). This uses the tree_vec_extract() function out of tree-vect-generic.c to retrieve the splat value, which is a BIT_FIELD_REF. That function is made non-static as part of this change. Testcases are already in-tree. V2

Re: [PATCH, OpenACC] Add support for gang local storage allocation in shared memory

On Wed, 15 Aug 2018 21:56:54 +0200 Bernhard Reutner-Fischer wrote: > On 15 August 2018 18:46:37 CEST, Julian Brown > wrote: > >On Mon, 13 Aug 2018 12:06:21 -0700 > >Cesar Philippidis wrote: > > atttribute has more t than strictly necessary. > Don't like signed integer levels where they

Re: [PATCH] convert braced initializers to strings (PR 71625)

On 08/16/2018 09:23 AM, Martin Sebor wrote: > On 08/15/2018 03:34 PM, Jeff Law wrote: >> On 08/15/2018 03:02 PM, Martin Sebor wrote: >>> On 08/15/2018 06:07 AM, Joseph Myers wrote: On Tue, 14 Aug 2018, Martin Sebor wrote: >> This is with Bison 3.0.4, should the version used to

Re: [PATCH] convert braced initializers to strings (PR 71625)

On 08/15/2018 03:34 PM, Jeff Law wrote: On 08/15/2018 03:02 PM, Martin Sebor wrote: On 08/15/2018 06:07 AM, Joseph Myers wrote: On Tue, 14 Aug 2018, Martin Sebor wrote: This is with Bison 3.0.4, should the version used to produce intl/plural.c prove relevant. Can you send me the

Re: [PATCH, testsuite] Make pr80263 work for Darwin by suppressing pubnames/types.

On 08/16/2018 08:35 AM, Iain Sandoe wrote: > > Darwin emits pubnames/types by default which masks the intended check. > > OK for trunk? > Iain > > gcc/testsuite > > * gcc.dg/debug/dwarf2/pr80263.c: Suppress pubtypes output for Darwin. OK jeff

Re: [PATCH][GCC][DOCS] Remove rtl.texi references to old RTX code class names

On 08/16/2018 08:44 AM, Sandra Loosemore wrote: > On 08/16/2018 02:42 AM, Matthew Malcomson wrote: >> >> On 15/08/18 20:26, Sandra Loosemore wrote: >>> >>> Use @item for the first item in a group, @itemx for all the others. >> >> Thanks for the spot, updated patch attached. > > This version looks

Re: [Patch wwwdocs] Document arm-8-branch

On Wed, 15 Aug 2018, Ramana Radhakrishnan wrote: > As $subject. > > Ok ? Yes, and you do not need anyone's approval. :-) Though since you asked: + This branch provides bug fixes and minor enhancements for GCC when + used targeting the AArch64 and the Arm architecture. Most patches Could

Re: [PATCH][GCC][DOCS] Remove rtl.texi references to old RTX code class names

On 08/16/2018 02:42 AM, Matthew Malcomson wrote: On 15/08/18 20:26, Sandra Loosemore wrote: Use @item for the first item in a group, @itemx for all the others. Thanks for the spot, updated patch attached. This version looks fine with me. Nobody's made any objection about technical

[PATCH, testsuite] Make pr80263 work for Darwin by suppressing pubnames/types.

Darwin emits pubnames/types by default which masks the intended check. OK for trunk? Iain gcc/testsuite * gcc.dg/debug/dwarf2/pr80263.c: Suppress pubtypes output for Darwin. --- gcc/testsuite/gcc.dg/debug/dwarf2/pr80263.c | 2 ++ 1 file changed, 2 insertions(+) diff --git

[PATCH, Darwin, drivers] Split DWARF is not supported on Darwin.

The Darwin toolchains have a separate debug linker (dsymutil) so that the link-time penalty for debug data is not usually seen. At present, it's not clear how we would support split DWARF on Darwin (or if it would bring any additional benefit over dsymutil). This patch produces diagnostic

Re: [2/2] Add AddressSanitizer annotations to std::string.

^^ gentle ping. On 07/16/2018 07:16 PM, Mikhail Kashkarov wrote: > Rebased and update patch (typos, add missing annotations), > add ASan teststo verify string annotation. > > > On 06/28/2018 11:09 AM, Mikhail Kashkarov wrote: >> ^ gentle ping. >> >> >> On 06/08/2018 05:54 PM, Mikhail Kashkarov

Re: [RFC][PATCH][mid-end] Optimize immediate choice in comparisons.

Vlad Lazar writes: > On 13/08/18 15:00, Richard Sandiford wrote: >> Vlad Lazar writes: >>> diff --git a/gcc/expmed.h b/gcc/expmed.h >>> index >>> 2890d9c9bbd034f01030dd551d544bf73e73b784..86a32a643fdd0fc9f396bd2c7904244bd484df16 >>> 100644 >>> --- a/gcc/expmed.h >>> +++ b/gcc/expmed.h >>> @@

Re: [GCC][PATCH][Aarch64] Make common aarch64 options target-dependent

On 08/16/2018 10:46 AM, Richard Earnshaw (lists) wrote: I've fixed this up and applied. R. PS, if you can use hard tabs for indenting your ChangeLog text it will make my life a lot easier, as I don't have to fix it up manually. Thanks for the tips and for committing. Sam

Re: GCC 7 backport

Hi. I'm going to install one more patch. Martin >From 1a1934d31636a8bd72a6d93143dbe216ff6e6ad4 Mon Sep 17 00:00:00 2001 From: marxin Date: Thu, 2 Aug 2018 09:32:58 + Subject: Backport r263246 libsanitizer/ChangeLog: 2018-08-02 Martin Liska PR sanitizer/86022 *

Re: GCC 8 backports

Hi. Another 2 patches that I tested. Martin >From 733fad0936b46794f9e1ba6742471bd9f351a4c9 Mon Sep 17 00:00:00 2001 From: marxin Date: Thu, 2 Aug 2018 10:17:34 + Subject: Backport r263248 gcc/ChangeLog: 2018-08-02 Martin Liska PR gcov-profile/86817 * gcov.c

Re: [GCC][PATCH][Aarch64] Make common aarch64 options target-dependent

On 16/08/18 09:45, Sam Tebbs wrote: > Hi all, > > This patch replaces the "Common" attribute in some aarch64 options with > "Target" > to make them target-dependent. This doesn't affect their behaviour or > how they > are documented with "--help", but does mean that they follow the option >

Re: [PATCH v2 2/4] libgcc: add crt{begin,end} for powerpc-wrs-vxworks target

On 2018-06-28 10:43, Rasmus Villemoes wrote: > In order to allow ZCX on VxWorks, we need the frame_dummy function to do > the register_frame_info(). So make sure crtbegin.o and crtend.o are > available for use with a custom spec file. Hi Olivier Can I also have your explicit ok for patch 2/4

Re: [PATCH] Merge Ignore and Deprecated in .opt files.

On 08/15/2018 06:38 PM, Joseph Myers wrote: > On Wed, 15 Aug 2018, Martin Liška wrote: > >> Ok, so you have very similar opinion as Jakub. Thus I'm sending new >> version that preserves status quo, it only does: > > This is removing RejectNegative from some Deprecated options. Won't that >

[GCC][PATCH][Aarch64] Make common aarch64 options target-dependent

Hi all, This patch replaces the "Common" attribute in some aarch64 options with "Target" to make them target-dependent. This doesn't affect their behaviour or how they are documented with "--help", but does mean that they follow the option specification that target-specific options are given the

Re: [PATCH][GCC][DOCS] Remove rtl.texi references to old RTX code class names

On 15/08/18 20:26, Sandra Loosemore wrote: Use @item for the first item in a group, @itemx for all the others. -Sandra Thanks for the spot, updated patch attached. diff --git a/gcc/doc/rtl.texi b/gcc/doc/rtl.texi index

Re: [PATCH] convert braced initializers to strings (PR 71625)

Jeff Law wrote: > I wonder if the change to how we set up the initializers is ultimately > changing the section those go into and ultimately causing an overflow of > the .sdata section. Yes, that is definitely the case. Due to the -fmerge-all-constants option used named arrays with brace