Re: [PATCH] Fix PR88182

On Mon, 26 Nov 2018, Jakub Jelinek wrote: > On Mon, Nov 26, 2018 at 04:36:26PM +0100, Richard Biener wrote: > > > > With the relatex outer loop reduction support we need to avoid picking > > up a different nested cycles reduction def. That's easy given we > > record the PHI we are looking at - a

Re: [PATCH 10/10] Port testsuite to GCN

On 26/11/2018 21:13, Mike Stump wrote: On Nov 26, 2018, at 12:04 PM, Mike Stump wrote: I'll Ok the signal one, if you prefer it over a dummy signal routine. Though, would be nice for you to add signal if possible/reasonable. Oh, and my long term thinking on signal is that logically, it's f

Re: [PATCH 07/10] Add dg-require-effective-target exceptions

On 26/11/2018 22:08, Mike Stump wrote: On Nov 16, 2018, at 8:28 AM, Andrew Stubbs wrote: [This patch was previously approved by Richard Sandiford (with added documentation I've still not done), but objected to by Mike Stump. I need to figure out who's right.] Since the planned port is done

[testsuite] Require ucn support in gdc.test/compilable/ddoc12.d (PR d/88039)

Some assemblers, including the Solaris one, don't support UTF-8 identifiers, which breaks the gdc.test/compilable/ddoc12.d testcase as reported in the PR. Since we don't want to modify the upstream testcase directly, I've instead modified the gdc.test driver to require ucn support via a dg-require

[c-family] Fix -fdump-ada-spec regressions in C++

This fixes a few regressions introduced by the switch to Ada 2012 for the output of -fdump-ada-spec that are present when the input is in C++. Tested on x86_64-suse-linux, applied on the mainline. 2018-11-27 Eric Botcazou * c-ada-spec.c: Include stringpool.h. (has_static_fie

Re: [PATCH] x86: Add -march=cascadelake

Thanks for the helpful information! But I'm still checking with hardware team about the family/model/stepping numbers for Cascadelake which are not officially disclosed by Intel (to my best knowledge). Wei Martin Liška 于2018年11月26日周一 下午10:00写道： > > On 11/26/18 12:18 PM, Jakub Jelinek wrote: > > O

[PATCH][wwwdocs] Mention Ampere eMAG support in GCC9 changes.html

This patch adds "Ampere eMAG" to the list of new supported AArch64 cores in changes.html for GCC 9. Ok to commit to CVS? Thanks, Christoph --- htdocs/gcc-9/changes.html.orig 2018-11-26 20:40:38.069556986 +0100 +++ htdocs/gcc-9/changes.html 2018-11-26 20:40:44.620096365 +0100 @@ -144,6 +14

Re: [PATCH/coding style] clarify pointers and operators

On 11/26/18 6:59 PM, Martin Sebor wrote: > Martin suggested we update the Coding Conventions to describe > the expected style for function declarations with a pointer > return types, and for overloaded operators.  Below is the patch. > > As an aside, regarding the space convention in casts: a crud

Re: [PATCH/coding style] clarify pointers and operators

On Mon, Nov 26, 2018 at 10:59:47AM -0700, Martin Sebor wrote: > --- htdocs/codingconventions.html 30 Sep 2018 14:38:46 - 1.85 > +++ htdocs/codingconventions.html 26 Nov 2018 17:59:21 - > @@ -803,12 +803,17 @@ >- x > >cast > - (foo) x > - (foo)x > + (type) x > + (t

Re: [PATCH v3] [aarch64] Correct the maximum shift amount for shifted operands.

On 11/26/18 7:50 PM, Christoph Muellner wrote: > The aarch64 ISA specification allows a left shift amount to be applied > after extension in the range of 0 to 4 (encoded in the imm3 field). > > This is true for at least the following instructions: > > * ADD (extend register) > * ADDS (extended reg

Re: [PATCH v3] [aarch64] Correct the maximum shift amount for shifted operands.

Sam, > On 27.11.2018, at 14:06, Sam Tebbs wrote: > > > On 11/26/18 7:50 PM, Christoph Muellner wrote: >> The aarch64 ISA specification allows a left shift amount to be applied >> after extension in the range of 0 to 4 (encoded in the imm3 field). >> >> This is true for at least the following i

Re: [PATCH][RFC] Extend locations where to seach for Fortran pre-include.

On 11/27/18 8:57 AM, Thomas Koenig wrote: > Hi Martin, > >> he header file will be install by glibc (glibc-devel package). > > Why actually glibc-devel?  Needing glibc-devel for fast compilation > of Fortran seems to be counter-intuitive. Maybe glibc would be better. Works for me. Martin > >

Re: [PATCH][RFC] Extend locations where to seach for Fortran pre-include.

On 11/26/18 7:33 PM, Joseph Myers wrote: > On Mon, 26 Nov 2018, Martin Liška wrote: > >> The header file will be install by glibc (glibc-devel package). > > To confirm: you intend to submit a patch to glibc upstream to install this > file (rather than it only being something in distribution pack

Re: [PATCH][RFC] Extend locations where to seach for Fortran pre-include.

On 11/26/18 7:44 PM, Joseph Myers wrote: > On Mon, 26 Nov 2018, Martin Liška wrote: > >>> I don't see how this version ensures that NATIVE_SYSTEM_HEADER_DIR is >>> properly sysrooted. Note there's add_sysrooted_prefix separate from >>> add_prefix (but that's *not* the correct thing to use here

[PATCH][Arm] Fix FPU configurations for Cortex-R7 and Cortex-R8

Hi, This patch fixes the FPU configurations of Cortex-R7 and Cortex-R8, enabling the use of FP16 conversion instructions for both and adding the option to disable double precision instruction support using '+nofp.dp'. Passes the self-check during building for an arm target. Is this OK for trunk

Re: [PATCH v4] PR preprocessor/83173: Enhance -fdump-internal-locations output

On 11/26/18 8:29 PM, David Malcolm wrote: > On Mon, 2018-11-26 at 22:17 +, Mike Gulick wrote: >> On 11/13/18 3:12 PM, David Malcolm wrote: >>> On Tue, 2018-11-13 at 14:54 -0500, Mike Gulick wrote: 2018-11-13 Mike Gulick >>> >>> [...] >>> * gcc/diagnostic-core.h (num_digits): Add

Re: [PATCH][Arm] Fix FPU configurations for Cortex-R7 and Cortex-R8

On 27/11/2018 14:10, Andre Vieira (lists) wrote: > > Hi, > > This patch fixes the FPU configurations of Cortex-R7 and Cortex-R8, > enabling the use of FP16 conversion instructions for both and adding the > option to disable double precision instruction support using '+nofp.dp'. > > Passes the se

Re: [PATCH v4] PR preprocessor/83173: Enhance -fdump-internal-locations output

On Tue, 2018-11-27 at 14:10 +, Mike Gulick wrote: > On 11/26/18 8:29 PM, David Malcolm wrote: > > On Mon, 2018-11-26 at 22:17 +, Mike Gulick wrote: > > > On 11/13/18 3:12 PM, David Malcolm wrote: > > > > On Tue, 2018-11-13 at 14:54 -0500, Mike Gulick wrote: > > > > > 2018-11-13 Mike Gulick

Re: [PATCH v4] PR preprocessor/83173: Enhance -fdump-internal-locations output

On 11/27/18 9:27 AM, David Malcolm wrote: [...] > > I can commit them for you if you like. Please can you repost the > latest version of the patches as one kit, for clarity, so I can commit > them. > Do you want me to repost them as a patch series, or as a single patch? > Thanks > Dave >

Re: [PATCH, OpenACC] Properly handle wait clause with no arguments

On 2018/11/8 3:13 AM, Thomas Schwinge wrote: NACK. Instead let's do the following, similar to C, C++, and also similar to Fortran's OpenACC async clause handling without explicit async-argument: --- gcc/fortran/openmp.c +++ gcc/fortran/openmp.c @@ -1885,7 +1885,19 @@ gfc_match_omp_clauses (gfc_

Re: [PATCH v4] PR preprocessor/83173: Enhance -fdump-internal-locations output

On Tue, 2018-11-27 at 14:37 +, Mike Gulick wrote: > On 11/27/18 9:27 AM, David Malcolm wrote: > > [...] > > > > > I can commit them for you if you like. Please can you repost the > > latest version of the patches as one kit, for clarity, so I can > > commit > > them. > > > > Do you want m

[PATCH v5 1/3] PR preprocessor/83173: Additional check before decrementing highest_location

2018-11-27 Mike Gulick PR preprocessor/83173 * libcpp/files.c (_cpp_stack_include): Check if line_table->highest_location is past current line before decrementing. --- libcpp/files.c | 32 +++- 1 file changed, 23 insertions(+), 9 dele

[PATCH v5 2/3] PR preprocessor/83173: New test

2018-11-27 Mike Gulick PR preprocessor/83173 * gcc.dg/plugin/location-overflow-test-pr83173.c: New test. * gcc.dg/plugin/location-overflow-test-pr83173.h: Header for pr83173.c. * gcc.dg/plugin/location-overflow-test-pr83173-1.h: Header for pr83173

[PATCH v5 3/3] PR preprocessor/83173: Enhance -fdump-internal-locations output

2018-11-27 Mike Gulick PR preprocessor/83173 * gcc/input.c (dump_location_info): Dump reason and included_from fields from line_map_ordinary struct. Fix indentation when location > 5 digits. * libcpp/location-example.txt: Update example -fdump-in

FIO, powerpc-darwin mangling patch [7.x and earlier].

Hi, So it turns out that the Darwin PPC port was broken essentially “forever” (before the tidy-up in 8.2) with respect to mangling the symbols for __ibm128 type (the default long double for the port). In 7.x and earlier (at least back to Apple’s 4.2.1) the use passes right through rs6000_mang

[committed] Trivial fix to microblaze port

The recent changes to the move_by_pieces API were slightly incomplete. They failed to update the mips and microblaze ports (there may be others, that's what has complained so far). This patch fixes the microblaze port. I've verified we can build it through libgcc. The tester will run a more com

Re: [PATCH 4/6] [RS6000] Remove constraints on call rounded_stack_size_rtx arg

On Tue, Nov 13, 2018 at 11:21:30PM +1030, Alan Modra wrote: > Version 2. (Same as before, here for completeness.) > > This call arg is unused on rs6000. > > * config/rs6000/darwin.md (call_indirect_nonlocal_darwin64), > (call_nonlocal_darwin64, call_value_indirect_nonlocal_darwin64),

[committed] Fix mips calls to move_by_pieces

And the same for the mips port. I'll go through all kinds of build tests during the day, including a bootstrap on mipsisa32r2. Committed. Jeff diff --git a/gcc/ChangeLog b/gcc/ChangeLog index 4be85e3c366..a6e8729f370 100644 --- a/gcc/ChangeLog +++ b/gcc/ChangeLog @@ -1,5 +1,8 @@ 2018-11-27 Jef

[committed] lm32 port fix for move_by_pieces

And the lm32 port... It'll go through cross testing at some point today. Committed. Jeff diff --git a/gcc/ChangeLog b/gcc/ChangeLog index a6e8729f370..fe8f6dc6ad7 100644 --- a/gcc/ChangeLog +++ b/gcc/ChangeLog @@ -1,5 +1,8 @@ 2018-11-27 Jeff Law + * config/lm32/lm32.c (lm32_block_mo

[committed] sh and riscv fixes for move_by_pieces change

sh and riscv this time. Sigh. Hopefully that's all the ports... These will also test via bootstraps during the day. Committed. Jeff diff --git a/gcc/ChangeLog b/gcc/ChangeLog index fe8f6dc6ad7..96215ee5cd7 100644 --- a/gcc/ChangeLog +++ b/gcc/ChangeLog @@ -1,5 +1,11 @@ 2018-11-27 Jeff Law

[PATCH] Fix vect/costmodel/ppc/costmodel-vect-33.c testcase (PR middle-end/87157)

Hi! This testcase started FAILing, because function splitting recently started splitting main1 into inline containing the cheap loop and main1.part.0 which contains the verification in abort. I believe the test is meant to test the vectorizer behavior, not how many times that loop has been inline

Go patch committed: Import inlinable functions from package export data

This patch to the Go frontend imports inlinable functions from other packages, when the body is recorded in the package export data. At this point we will inline direct calls to empty functions and methods defined in different packages. Bootstrapped and ran Go testsuite on x86_64-pc-linux-gnu. C

Re: [PATCH v2] MIPS: Add `-mfix-r5900' option for the R5900 short loop erratum

On Fri, 9 Nov 2018, Fredrik Noring wrote: > gcc/ > * config/mips/mips.c (mips_reorg_process_insns) > (mips_option_override): Handle `-mfix-r5900'. > * config/mips/mips.h (ASM_SPEC): Add `mfix-r5900' and > `mno-fix-r5900'. > * config/mips/mips.opt (mfix-r5900): N

[PATCH] Fix slowness in gcov (PR gcov-profile/88225).

Hi. The patch is about addition of 2 maps that significantly speed up gcov for tramp3d from 2.0s to 0.2s. Survives gcov.exp tests, I'm planning to install the patch in couple of days. Martin gcc/ChangeLog: 2018-11-27 Martin Liska PR gcov-profile/88225 * gcov.c(source_info::

Re: [PATCH v4] Repeat jump threading after combine

> Am 26.11.2018 um 16:26 schrieb Ilya Leoshkevich : > >> Am 26.11.2018 um 16:07 schrieb Segher Boessenkool >> : >> >>> # ppc64le-redhat-linux: >>> 511.povray_r -1.29% >>> 482.sphinx3-0.65% >>> 456.hmmer -0.53% >>> 519.lbm_r -0.51% >>> # skip |dt| < 0.5% >>> 549.fo

Re: [PATCH v5 1/3] PR preprocessor/83173: Additional check before decrementing highest_location

On Tue, 2018-11-27 at 09:53 -0500, Mike Gulick wrote: > 2018-11-27 Mike Gulick > > PR preprocessor/83173 > * libcpp/files.c (_cpp_stack_include): Check if > line_table->highest_location is past current line before > decrementing. I've committed these patches to trunk as

Re: [PATCH/coding style] clarify pointers and operators

On 11/27/18 5:57 AM, Jakub Jelinek wrote: On Mon, Nov 26, 2018 at 10:59:47AM -0700, Martin Sebor wrote: --- htdocs/codingconventions.html 30 Sep 2018 14:38:46 - 1.85 +++ htdocs/codingconventions.html 26 Nov 2018 17:59:21 - @@ -803,12 +803,17 @@ - x cast - (fo

[PATCH] Fix PR78444.

Hi, As described in comment #4/5 in the PR, there are [corner] cases where the the compiler incorrectly concludes that only 64b alignment is needed at a call site. This is caught by Darwin’s dynamic linker which enforces the ABI requirement for 128b alignment, causing the exe be aborted. With

Re: [PATCH v5 1/3] PR preprocessor/83173: Additional check before decrementing highest_location

On 11/27/18 11:07 AM, David Malcolm wrote: > On Tue, 2018-11-27 at 09:53 -0500, Mike Gulick wrote: >> 2018-11-27 Mike Gulick >> >> PR preprocessor/83173 >> * libcpp/files.c (_cpp_stack_include): Check if >> line_table->highest_location is past current line before >> decremen

Re: [PATCH][RFC] Extend locations where to seach for Fortran pre-include.

On Mon, 2018-11-26 at 17:35 +0100, Martin Liška wrote: > On 11/26/18 5:19 PM, Matthias Klose wrote: > > On 26.11.18 13:20, Martin Liška wrote: > > > On 11/23/18 7:08 PM, Joseph Myers wrote: > > > > In the multiarch case, do you want > > > > /include/finclude/ or > > > > /include//finclude? (This

[PATCH] gcov: do not ICE on NULL string in JSON export.

Hi. One obvious fix that can be triggered: gcov asdnkjasndjkn asdnkjasndjkn.gcno:cannot open notes file asdnkjasndjkn.gcda:cannot open data file, assuming not executed gcov: internal compiler error: in string, at json.cc:159 0x4038e7 json::string::string(char const*) /home/marxin/Program

Re: [PATCH 5/6] [RS6000] Use standard call patterns for __tls_get_addr calls

Hi! On Tue, Nov 13, 2018 at 11:22:43PM +1030, Alan Modra wrote: > Version 2. > > The current code handling __tls_get_addr calls for powerpc*-linux > generates a call then overwrites the call insn with a special > tls_{gd,ld}_{aix,sysv} pattern. It's done that way to support > !TARGET_TLS_MARKERS

Re: [committed] sh and riscv fixes for move_by_pieces change

On 11/27/18 4:34 PM, Jeff Law wrote: > > sh and riscv this time. Sigh. Hopefully that's all the ports... These > will also test via bootstraps during the day. > > > Committed. > > Jeff > Thank you Jeff for the help. I grepped for a usage in a target code, but apparently wrongly. Sorry. Mar

Re: [PATCH] Fix PR78444.

On Tue, Nov 27, 2018 at 5:17 PM Iain Sandoe wrote: > > Hi, > > As described in comment #4/5 in the PR, there are [corner] cases where the > the compiler incorrectly concludes that only 64b alignment is needed at a > call site. This is caught by Darwin’s dynamic linker which enforces the ABI >

Re: [PATCH] detect missing nuls in address of const char (PR 87756)

On 11/26/18 10:52 AM, Rainer Orth wrote: Hi Martin, I have now committed this patch as r266418. this patch has created a bunch of XPASSes everywhere: +XPASS: gcc.dg/tree-ssa/builtin-fprintf-warn-1.c pr87756 (test for warnings, line 119) +XPASS: gcc.dg/tree-ssa/builtin-fprintf-warn-1.c pr877

Re: [PATCH 6/6] [RS6000] inline plt call sequences

Hi Alan, On Tue, Nov 13, 2018 at 11:23:43PM +1030, Alan Modra wrote: > Finally, the point of the previous patches in this series, support for > inline PLT calls, keyed off -fno-plt. This emits code using new > relocations that tie all insns in the sequence together, so that the > linker can edit

Re: [PING][PATCH] correct handling of EXCESS_PRECISION_EXPR in function calls (PR 88091)

On Mon, 26 Nov 2018, Martin Sebor wrote: > Ping: https://gcc.gnu.org/ml/gcc-patches/2018-11/msg01674.html This patch needs an update to the comment on convert_argument to explain the semantics of the new valtype parameter and how it differs from the previously used TREE_TYPE (val). -- Joseph

Re: [PATCH] PR880088 Enable -Wtrampolines for no-exec-stack targets with -Wall.

Hi Mark, On Mon, Nov 26, 2018 at 10:13:29AM +0100, Mark Wielaard wrote: > With -Wtrampolines a warning is produced whenever gcc generates executable > code on the stack at runtime to support taking a nested function address > that is used to call the nested function indirectly when it needs to acc

Re: [PATCH/coding style] clarify pointers and operators

On Mon, Nov 26, 2018 at 01:41:07PM -0700, Jeff Law wrote: > On 11/26/18 10:59 AM, Martin Sebor wrote: > > As an aside, regarding the space convention in casts: a crude > > grep search yields about 10,000 instances of the "(type)x" kinds > > of casts in GCC sources and 40,000 of the preferred "(type

Re: [PATCH 5/7][MSP430][TESTSUITE] Prune messages about ISO C not supporting __int20 from output of tests

On Nov 26, 2018, at 3:07 PM, Jozef Lawrynowicz wrote: > > On Mon, 26 Nov 2018 12:51:26 -0800 > Mike Stump wrote: > >> On Nov 14, 2018, at 7:56 AM, Jozef Lawrynowicz >> wrote: >>> >>> Patch 5 deals with ISO C errors emitted by tests when the large memory >>> model is >>> used. size_t and ptr

Go patch committed: Change Expression export to use Export_function_body

This patch to the Go frontend changes Expression export to use Export_function_body instead of Export. This is in preparation for writing expressions to inline function bodies. This is a refactoring that doesn't affect compiler output. Bootstrapped and ran Go testsuite on x86_64-pc-linux-gnu. C

Re: FIO, powerpc-darwin mangling patch [7.x and earlier].

On Nov 27, 2018, at 7:05 AM, Iain Sandoe wrote: > > So it turns out that the Darwin PPC port was broken essentially “forever” > (before the tidy-up in 8.2) with respect to mangling the symbols for __ibm128 > type (the default long double for the port). > > In 7.x and earlier (at least back to

Re: [PATCH] Added information about inline assembler in stack calculations (.su files)

Hi! On Mon, Nov 26, 2018 at 02:02:49PM +, Torbjorn SVENSSON wrote: > Attached is a small patch that, in case of inline assembler code, > indicates that the function stack usage is uncertain due to inline > assembler. > > The test suite are using "nop" as an assembler instruction on all > t

Re: [testsuite] Require ucn support in gdc.test/compilable/ddoc12.d (PR d/88039)

On Nov 27, 2018, at 2:18 AM, Rainer Orth wrote: > > Some assemblers, including the Solaris one, don't support UTF-8 > identifiers, which breaks the gdc.test/compilable/ddoc12.d testcase as > reported in the PR. So, another style of fix, would be to change the binding from the language front-end

Re: [PATCH] Fix vect/costmodel/ppc/costmodel-vect-33.c testcase (PR middle-end/87157)

On Tue, Nov 27, 2018 at 04:47:29PM +0100, Jakub Jelinek wrote: > This testcase started FAILing, because function splitting recently started > splitting main1 into inline containing the cheap loop and main1.part.0 which > contains the verification in abort. I believe the test is meant to test > the

Re: [testsuite] Require ucn support in gdc.test/compilable/ddoc12.d (PR d/88039)

Hi Mike, > On Nov 27, 2018, at 2:18 AM, Rainer Orth > wrote: >> >> Some assemblers, including the Solaris one, don't support UTF-8 >> identifiers, which breaks the gdc.test/compilable/ddoc12.d testcase as >> reported in the PR. > > So, another style of fix, would be to change the binding from t

[gcc-10 PATCH, i386]: Use accessible_reg_set to disable unsupported register sets

Hello! This patch is based on the discussion in PR88178. Apparently, MIPS uses accessible_reg_set array to disable register sets, unsupported by currently active ISAs. The patch implements the same approach for x86, which also results in IMO better error messages. The middle-end misses to error ou

[PATCH v4] Add sinh(atanh(x)) and cosh(atanh(x)) optimizations

Only do this optimization if funsafe-math and -fno-math-errno are enabled, as pointed in the previous iteration. Also added one more test case to ensure that fno-math-errno is required for the optimization. Special thanks for Wilco Dijsktra for all his help :-) gcc/ChangeLog 2018-11-27 Giuliano

Re: [PATCH] Added information about inline assembler in stack calculations (.su files)

Hi! Thanks for the feedback. Attached is an updated patch where I switched to the NOP define instead. I'm not sure if stack-usage-naked.c should be moved to gcc.dg-tree, or if it should skip using the nop.h file (it feels wrong to do #include "../../gcc.dg/nop.h" from within gcc.taget-tree). To

Re: [PATCH] PR880088 Enable -Wtrampolines for no-exec-stack targets with -Wall.

Hi, On Tue, 2018-11-27 at 12:37 -0600, Segher Boessenkool wrote: > > Define a new target macro TARGET_HAS_DEFAULT_NOEXEC_STACK for those > > targets > > that have a non-executable default stack based on when they call > > file_end_indicate_exec_stack. > > As Paul says, that name isn't so good. >

Go patch committed: Pass a Location to import_expression

This patch changes the Go frontend to pass a Location to import_expression. This separates the Location that import_expression uses when creating a new Expression from the Location used to report an error. This is a step toward importing expressions for inlined functions. This is a pure refactor

Go patch committed: Change expressions importing to use Import_expression

This patch to the Go frontend changes expression importing to use a new abstract interface class Import_expression, so that we can more easily import expressions from inlinable function bodies. This is a refactoring with no affect on compiler behavior. Bootstrapped and ran Go tests on x86_64-pc-l

Re: [PATCH v3] [aarch64] Correct the maximum shift amount for shifted operands.

> On 27.11.2018, at 14:04, Sam Tebbs wrote: > > > On 11/26/18 7:50 PM, Christoph Muellner wrote: >> The aarch64 ISA specification allows a left shift amount to be applied >> after extension in the range of 0 to 4 (encoded in the imm3 field). >> >> This is true for at least the following instr

Re: C++ PATCH to implement P1094R2, Nested inline namespaces

On 11/21/18 8:46 PM, Marek Polacek wrote: On Tue, Nov 20, 2018 at 04:59:46PM -0500, Jason Merrill wrote: On 11/19/18 5:12 PM, Marek Polacek wrote: On Mon, Nov 19, 2018 at 10:33:17PM +0100, Jakub Jelinek wrote: On Mon, Nov 19, 2018 at 04:21:19PM -0500, Marek Polacek wrote: 2018-11-19 Marek Po

Re: [C++ PATCH] Propagate TYPE_PACKED to template variants (PR c++/88181)

On 11/26/18 3:54 PM, Jakub Jelinek wrote: Hi! On the following patch -fpack-struct forces TYPE_PACKED on all the classes and their variants, but we then create a variant of a class instantiation (const) which doesn't have the TYPE_PACKED set and later finalize the template main variant, but don'

Re: C++ PATCH to implement P1094R2, Nested inline namespaces

On 11/23/18 4:12 PM, Marek Polacek wrote: I wasn't aware you had worked on this. Perhaps we should track the progress of C++20 features in Bugzilla (to keep track of who's working on what). Yes, I think so. I created a couple of PRs, for contracts and cmpxchg with padding bits, but more are

Re: [PATCH][RFC] Extend locations where to seach for Fortran pre-include.

Am 27.11.18 um 17:22 schrieb Steve Ellcey: Why wouldn't clang (flang) want to use the same mechanism as GCC/gfortran? I know there is some interest/work going on here for flang and we would like a consistent way to use pre-includes to define SIMD vector functions in both gfortran and flang. I t

[PATCH] Fix 87380.

Hi This is [intentionally] broken C++ ABI, that was catering for a tool problem that existed in a very old Darwin toolchain. I checked that the issue is not present after Darwin7 (using default Xcode tools). Of course, more modern tools are probably required to build trunk GCC for Darwin7, bu

Go patch committed: Add '$' to names in expression export data

This patch to the Go frontend adds '$' to names in expression export data. For inlined function bodies we're going to need to refer to variables, so change the existing export data to add a '$' to names that look like identifiers: true, false, nil, convert. While we're here drop an unnecessary sp

Re: [C++ PATCH] Fix ICE in grokdeclarator (PR c++/88187)

On 11/26/18 3:49 PM, Jakub Jelinek wrote: Hi! Marek has changed grokdeclarator in r263836, so that in this part of code it is either a funcdecl_p (previously the only allowed one), which implies inner_declarator is non-NULL and therefore unqualified_id too, or newly inner_declarator == NULL. In

Re: Fix hashtable memory leak

I eventually called the new method _M_assign_elements. And yes, tracker_allocator was enough. Committed on trunk for the moment. François On 11/26/18 1:12 PM, Jonathan Wakely wrote: On 24/11/18 22:58 +0100, François Dumont wrote: --- a/libstdc++-v3/testsuite/23_containers/unordered_set/allo

Go patch committed: Finalize types parsed for inline functions

When the Go frontend inlines functions from other packages, we may parse types that we have not seen before inlining. Inlining runs after the finalize_methods pass, so those types will not be finalized, meaning that we don't have an accurate list of which methods they support. Explicitly finalize

Go patch committed: Add types used by inline functions to export data

This Go frontend patch add types used by inline functions to the export data. Bootstrapped and ran Go testsuite on x86_64-pc-linux-gnu. Committed to mainline. Ian Index: gcc/go/gofrontend/MERGE === --- gcc/go/gofrontend/MERGE (r

Go patch committed: Add result parameter names for inlinable functions

This patch to the Go frontend adds result parameter names for inlinable functions. An inlinable function body may need to refer to result parameters, so each result parameter needs a name. We already give them all names in start_function (via create_result_variables). Change the export data so th

Re: [PATCH] target/58397: add host_hooks for NetBSD to make precompiled headers work

Hej! On Sun, 25 Nov 2018, Krister Walfridsson wrote: > On Sun, 25 Nov 2018, Maya Rashish wrote: > > > gcc/config.host | 4 ++ > > gcc/config/host-netbsd.c | 85 > > gcc/config/x-netbsd | 4 ++ > > 3 files changed, 93 insertions(+) > > create mo

Re: Fix hashtable memory leak

On 27/11/18 22:31 +0100, François Dumont wrote: I eventually called the new method _M_assign_elements. Perfect. And yes, tracker_allocator was enough. Great. Committed on trunk for the moment. Great, thanks. Please note that GCC 7.4 RC1 is scheduled for this week: https://gcc.gnu.org/m

Re: [C++ PATCH] Fix wrong-code with generic lambda (PR c++/86943)

On 11/23/18 4:15 PM, Jakub Jelinek wrote: Hi! On the following testcase, the call to operator () is marked as CALL_FROM_THUNK_P and therefore genericization ignores all subtrees thereof. Unfortunately, one of the arguments is a move ctor call which is not itself CALL_FROM_THUNK_P and thus we nee

Re: [C++ PATCH] Don't incorrectly reject {un,}signed char constexpr array initialization in templates (PR c++/87476)

On 11/16/18 4:00 AM, Jakub Jelinek wrote: Hi! The following two testcases, one is a regression from GCC 8 (introduced by the constructor to STRING_CST optimization), the other seems to fail since C++11 support has been introduced (but is accepted by clang++) fail, because during parsing with pro

[build] Fix libgphobos linking on Solaris 11

As mentioned in passing in PR d/87864, libgphobos.so currently fails to link before Solaris 11.4. Until then, you needed to link with -lsocket -lnsl for the networking functions, in S11.4 they were merged into libc. To fix this, I've adapted the check from libgo/configure.ac, for the moment just

Re: [PATCH] Fix vect/costmodel/ppc/costmodel-vect-33.c testcase (PR middle-end/87157)

On 11/27/18 8:47 AM, Jakub Jelinek wrote: > Hi! > > This testcase started FAILing, because function splitting recently started > splitting main1 into inline containing the cheap loop and main1.part.0 which > contains the verification in abort. I believe the test is meant to test > the vectorizer

Re: [C++ PATCH] Fix wrong-code with generic lambda (PR c++/86943)

On Tue, Nov 27, 2018 at 05:20:56PM -0500, Jason Merrill wrote: > On 11/23/18 4:15 PM, Jakub Jelinek wrote: > > Hi! > > > > On the following testcase, the call to operator () is marked as > > CALL_FROM_THUNK_P and therefore genericization ignores all subtrees thereof. > > Unfortunately, one of the

Re: [PATCH] [PR86397] set p_t_decl while canonicalizing eh specs for mangling

On 11/22/18 6:40 PM, Alexandre Oliva wrote: Mangling visits the base template function type, prior to template resolution, and on such types, exception specifications may contain unresolved noexcept expressions. nothrow_spec_p is called on them even when exception specifications are not part of

Re: [PATCH] Fix 87380.

On 11/27/18 2:16 PM, Iain Sandoe wrote: > Hi > > This is [intentionally] broken C++ ABI, that was catering for a tool problem > that existed in a very old Darwin toolchain. > > I checked that the issue is not present after Darwin7 (using default Xcode > tools). Of course, more modern tools are

Re: [PATCH/coding style] clarify pointers and operators

On 11/27/18 11:46 AM, Segher Boessenkool wrote: > On Mon, Nov 26, 2018 at 01:41:07PM -0700, Jeff Law wrote: >> On 11/26/18 10:59 AM, Martin Sebor wrote: >>> As an aside, regarding the space convention in casts: a crude >>> grep search yields about 10,000 instances of the "(type)x" kinds >>> of cast

[PATCH] PR libstdc++/67843 set shared_ptr lock policy at build-time

This resolves a longstanding issue where the lock policy for shared_ptr reference counting depends on compilation options when the header is included, so that different -march options can cause ABI changes. For example, objects compiled with -march=armv7 will use atomics to synchronize reference c

Go patch committed: Record final type for numeric expressions

This patch changes the Go frontend to record the final type for numeric expressions in export data. Inlinable function bodies are generated after the determine_types pass, so we know the type for all constants. Rather than try to determine it again when inlining, record the type in the export dat

[PATCH] Clean up temporary files created by std::filesystem testsuite

* testsuite/27_io/filesystem/operations/canonical.cc: Remove directory created by test. * testsuite/27_io/filesystem/operations/symlink_status.cc: Remove symlink created by test. Tested x86_64-linux, committed to trunk. commit c981ab59c04fe12ffee89bbfe466a40d66e40

Re: [PATCH] [PR85569] skip constexpr target_expr constructor dummy type conversion

On 11/22/18 6:39 PM, Alexandre Oliva wrote: The testcase is the work-around testcase for the PR; even that had started failing. The problem was that, when unqualifying the type of a TARGET_EXPR, we'd create a variant of the type, then request the conversion of the TARGET_EXPR_INITIAL to that var

Go patch committed: Tweaks for importing inline function bodies

This patch to the Go frontend adds some tweaks for importing inline function bodies. We track whether we've seen an error when importing a function; we will use error tracking to avoid knock-on errors. We stop importing identifiers at a ')'. We provide a way to adjust the indentation level while

Re: [PATCH 5/6] [RS6000] Use standard call patterns for __tls_get_addr calls

On Tue, Nov 27, 2018 at 10:29:29AM -0600, Segher Boessenkool wrote: > Hi! Thanks for the review! > > +(define_insn "*tls_gdld_aix" > > + [(match_parallel 3 "" > > A match_parallel without predicate... Does this work?! Yes. In fact, rs6000/predicates.md any_parallel_operand is useless except

Re: [PATCH 6/6] [RS6000] inline plt call sequences

On Tue, Nov 27, 2018 at 11:17:48AM -0600, Segher Boessenkool wrote: > That all sounds great :-) Has this been tested with older binutils, too? Yes, lots of times. Sometimes by accident when I forgot to install a new binutils. :-) > > +(define_insn "*sibcall_indirect_nonlocal_sysv" > > > + (

[PATCH] be more permissive about function alignments (PR 88208)

The tests for the new __builtin_has_attribute function have been failing on a number of targets because of a couple of assumptions that only hold on some. First, they expect that it's safe to apply attribute aligned with a smaller alignment than the target provides when GCC rejects such arguments

Re: [PATCH] [PR86397] set p_t_decl while canonicalizing eh specs for mangling

On Nov 27, 2018, Jason Merrill wrote: > On 11/22/18 6:40 PM, Alexandre Oliva wrote: >> Mangling visits the base template function type, prior to template >> resolution, and on such types, exception specifications may contain >> unresolved noexcept expressions. nothrow_spec_p is called on them >>

Re: [PATCH/coding style] clarify pointers and operators

On 11/26/18 10:59 AM, Martin Sebor wrote: Martin suggested we update the Coding Conventions to describe the expected style for function declarations with a pointer return types, and for overloaded operators.  Below is the patch. As an aside, regarding the space convention in casts: a crude grep

Re: Fix hashtable memory leak

On 11/27/18 11:00 PM, Jonathan Wakely wrote: On 27/11/18 22:31 +0100, François Dumont wrote: I eventually called the new method _M_assign_elements. Perfect. And yes, tracker_allocator was enough. Great. Committed on trunk for the moment. Great, thanks. Please note that GCC 7.4 RC1 is

[PATCH] Use {,v}blendvp{s,d} for [SD]Fmode sse_movcc (PR target/88189)

Hi! This patch implments Marc's idea of using {,v}blenvp{s,d} for scalar [SD]Fmode ix86_expand_sse_movcc for -msse4.1 and above. Without this patch we emit sequences like andpd %xmm2, %xmm0 andnpd %xmm1, %xmm2 orpd%xmm2, %xmm0 or andps %xmm2, %xmm0

[PATCH] Fix UBSAN ICE with C++ attribute types (PR c++/88215)

Hi! On the following testcase we ICE, because op1 doesn't have normal int type, but a distinct attribute type, while their main variants are different, they are still treated as compatible types and the C++ FE doesn't actually try to ensure both arguments have the same type. Bootstrapped/regteste