Re: [patch, fortran] Fix PR 92004, restore Lapack compilation

Hi Tobias, function ("o" missing); I think it is not clause 14 but paragraph 14. Fixed. (That one was easy :-) +   warning for this can be suppressed later.  */ + +bool +maybe_dummy_array_arg (gfc_expr *e) +{ +  gfc_symbol *s; + +  if (e->rank > 0) +    return false; + +  if (e->ts.type ==

[ARM] Enable DF only when TARGET_VFP_DOUBLE

As reported in Linaro bug report (https://bugs.linaro.org/show_bug.cgi?id=4636 ; there is no reproducible testcase provided), for some applications, we see (insn 126 125 127 9 (set (reg:DF 189) (fma:DF (reg:DF 126 [ _74 ]) (reg:DF 190) (reg:DF 191))) "ops.c":30 -1

Re: [PATCH], V4, patch #9 [part of patch #4.2], Add prefixed address offset checks

On Wed, Oct 09, 2019 at 04:56:48PM -0500, Segher Boessenkool wrote: > Hi! > > On Fri, Oct 04, 2019 at 08:29:11AM -0400, Michael Meissner wrote: > > @@ -8651,6 +8675,11 @@ rs6000_legitimate_address_p (machine_mod > >&& mode_supports_pre_incdec_p (mode) > >&&

Re: [PATCH], V4, patch #9 [part of patch #4.2], Add prefixed address offset checks

On Wed, Oct 09, 2019 at 04:56:48PM -0500, Segher Boessenkool wrote: > Hi! > > On Fri, Oct 04, 2019 at 08:29:11AM -0400, Michael Meissner wrote: > > @@ -8651,6 +8675,11 @@ rs6000_legitimate_address_p (machine_mod > >&& mode_supports_pre_incdec_p (mode) > >&&

Re: C++ PATCH to merge concepts-cxx2a branch

Le mer. 9 oct. 2019 à 19:20, Jason Merrill a écrit : > > Andrew has done a lot of work to update the GCC implementation of C++ > concepts to conform to the upcoming C++20 standard, which clarifies, > removes, and changes various aspects of the earlier concepts TS. It can > still use some

Re: [PATCH], V4, patch #9 [part of patch #4.2], Add prefixed address offset checks

Hi! On Fri, Oct 04, 2019 at 08:29:11AM -0400, Michael Meissner wrote: > @@ -8651,6 +8675,11 @@ rs6000_legitimate_address_p (machine_mod >&& mode_supports_pre_incdec_p (mode) >&& legitimate_indirect_address_p (XEXP (x, 0), reg_ok_strict)) > return 1; > + > + /* Handle

[PATCH] Remove redundant code from reduction vectorization

Some cleanup. Bootstrapped & tested on x86_64-unknown-linux-gnu, applied to trunk. Richard. 2019-10-08 Richard Biener * tree-vectorizer.h (_stmt_vec_info::reduc_vectype_in): New. (_stmt_vec_info::force_single_cycle): Likewise. (STMT_VINFO_FORCE_SINGLE_CYCLE): New.

Re: [patch][Fortran] Actually permit OpenMP's 'target simd'

On 8 October 2019 14:12:47 CEST, Jakub Jelinek wrote: >On Tue, Oct 08, 2019 at 02:04:17PM +0200, Tobias Burnus wrote: >> Seemingly, 'target simd' was forgotten – which yielded the error: >> "Unexpected !$OMP TARGET SIMD statement" >> >> OK for the trunk? >Ok, with moving the test to

[PATCH, rs6000] Lower vec_promote_demote vectorization cost for P8/P9

Hi, This patch is to lower vec_promote_demote vectorization cost in rs6000_builtin_vectorization_cost. It's similar to what we committed for vec_perm, the current cost for vec_promote_demote is also overpriced for Power8 and Power9 since Power8 and Power9 has supported more units for

Re: [PATCH 1/2][vect]PR 88915: Vectorize epilogues when versioning loops

On Tue, 8 Oct 2019, Andre Vieira (lists) wrote: > Hi Richard, > > As I mentioned in the IRC channel, I managed to get "most" of the regression > testsuite working for x86_64 (avx512) and aarch64. > > On x86_64 I get a failure that I can't explain, was hoping you might be able > to have a look

Re: [PATCH, rs6000] Lower vec_promote_demote vectorization cost for P8/P9

Hi Kewen, On Wed, Oct 09, 2019 at 02:43:02PM +0800, Kewen.Lin wrote: > This patch is to lower vec_promote_demote vectorization cost in > rs6000_builtin_vectorization_cost. It's similar to what we committed > for vec_perm, the current cost for vec_promote_demote is also > overpriced for Power8

[committed] Fix libgomp use_device_* handling (PR libgomp/92028)

Hi! The following is what I've intended to do but actually ended up committing a weird if (!not_found_cnt) { } else ... The code used to be at the indentation level of if, so my suspicion is that I've marked the block in an editor and tried

Re: Type representation in CTF and DWARF

On Wed, Oct 09, 2019 at 09:41:09AM +0200, Richard Biener wrote: > There's a mechanism to get type (and decl - I suppose CTF also > contains debug info > for function declarations not only its type?) info as part of early > debug generation. > The attached "hack" simply mangles dwarf2out to output

[Patch] gcc-10/changes.html – mention OpenMP's use_device_addr support

This mentioned the recently added OpenMP 5 feature "use_device_addr" which complements the (for Fortran) underspecified "use_device_ptr". The rest of the message, I have taken and modified from gcc-7/changes.html. – [On the C/C++, ME and libgomp side, OpenMP 4.5 is supported and 5.0

Re: [PATCH 2/2][MSP430] Optimize zero_extend insns and PSImode pointer manipulation

Hi Jozef, On Tue, Oct 08, 2019 at 11:39:57AM +0100, Jozef Lawrynowicz wrote: > +;; > +;; ZERO EXTEND INSTRUCTIONS > +;; Byte-writes to registers clear bits 19:8 > +;; * Byte-writes to memory do not affect bits 15:8 > +;; Word-writes to registers clear bits 19:16 > +;;

Re: Type representation in CTF and DWARF

On Tue, Oct 08, 2019 at 10:26:13PM -0700, Indu Bhagat wrote: > The justification for CTF is and will remain - a compact, faster debug > format > for type information and support some online debugging use-cases (like > backtraces) in future. Approximate backtraces, sure. (It cannot know if

Re: [Patch, fortran] PR91926 - assumed rank optional

Hi, On Sat, 5 Oct 2019 at 20:31, Paul Richard Thomas < paul.richard.tho...@gmail.com> wrote: > I must apologise not posting this before committing. I left for a > vacation this morning and I thought that this problem and the one > posted by Gilles were best fixed before departing. The patch

Re: [patch][Fortran] Actually permit OpenMP's 'target simd'

Hi! Tobias, if that's not inconvenient for you for any reson (which?), would you please adopt the habit of using the same text for a commit's first line in the log message (r276698: "Fortran - fix OpenMP 'target simd'), and the email's "Subject" line ("[Fortran] Actually permit OpenMP's 'target

Re: [patch][Fortran] Actually permit OpenMP's 'target simd'

On Wed, Oct 09, 2019 at 11:46:37AM +0200, Thomas Schwinge wrote: > On 2019-10-09T08:04:29+0200, Bernhard Reutner-Fischer > wrote: > > On 8 October 2019 14:12:47 CEST, Jakub Jelinek wrote: > >>On Tue, Oct 08, 2019 at 02:04:17PM +0200, Tobias Burnus wrote: > >>> Seemingly, 'target simd' was

Re: [Patch, fortran] PR91926 - assumed rank optional

Hi Christophe, Thanks for flagging this up - I am back at base on Saturday and will take it up then. Regards Paul On Wed, 9 Oct 2019 at 11:13, Christophe Lyon wrote: > > Hi, > > > On Sat, 5 Oct 2019 at 20:31, Paul Richard Thomas > wrote: >> >> I must apologise not posting this before

Re: [PATCH] Come up with ipa passes introduction in gccint documentation

> On Tue, Oct 8, 2019 at 10:06 PM Sandra Loosemore > wrote: > > > > On 10/8/19 2:52 AM, luoxhu wrote: > > > Hi, > > > > > > This is the formal documentation patch for IPA passes. Thanks. > > > > > > > > > None of the IPA passes are documented in passes.texi. This patch adds > > > a section IPA

Re: Type representation in CTF and DWARF

On Wed, Oct 9, 2019 at 7:26 AM Indu Bhagat wrote: > > > > On 10/08/2019 08:37 AM, Pedro Alves wrote: > > On 10/4/19 8:23 PM, Indu Bhagat wrote: > >> Hello, > >> > >> At GNU Tools Cauldron this year, some folks were curious to know more on > >> how > >> the "type representation" in CTF compares

Re: GCC wwwdocs move to git done

Hi Joseph! On 2019-10-09T00:27:38+, Joseph Myers wrote: > I've done the move of GCC wwwdocs to git Thanks! And now, ;-) as promised: # apt-get --purge remove cvs Reading package lists... Done Building dependency tree Reading state information... Done The

Re: [libgomp][testsuite] PR testsuite/91884 Add -lquadmath if available

On Tue, Oct 08, 2019 at 08:33:03PM +0200, Tobias Burnus wrote: > libgomp/ > * testsuite/libgomp.fortran/fortran.exp: Add -lquadmath if available. > * testsuite/libgomp.oacc-fortran/fortran.exp: Ditto. > > diff --git a/libgomp/testsuite/libgomp.fortran/fortran.exp >

Re: [PR 70929] IPA call type compatibility fix/workaround

On Tue, 8 Oct 2019, Martin Jambor wrote: > Hi, > > I've been looking at PR 70929 and at the newly reported duplicate PR > 91988 and would like to propose the following patch to address them. > Basically, the idea is that if the types or arguments in TYPE_ARG_TYPES > (as opposed to

Re: [libgomp][testsuite] PR testsuite/91884 Add -lquadmath if available

I have also an alternative version: * For in-build-dir test runs, do as with previous patch, i.e. add "-lquadmath" to the "xgcc" call if the directory is available. * For installed-compiler test runs, the new patch uses GFORTRAN_UNDER_TEST / GXX_UNDER_TEST (It would also use them with

Re: [libgomp][testsuite] PR testsuite/91884 Add -lquadmath if available

Thanks for the review. I have now committed this version as r276754 (with splitting the long line). If someone prefers the other variant or bits of it, I will do it as follow up. Cheers, Tobias On 10/9/19 9:43 AM, Jakub Jelinek wrote: On Tue, Oct 08, 2019 at 08:33:03PM +0200, Tobias

Re: GCC wwwdocs move to git done

On Wed, 9 Oct 2019 at 01:28, Joseph Myers wrote: > > I've done the move of GCC wwwdocs to git (using the previously posted and > discussed scripts), including setting up the post-receive hook to do the > same things previously covered by the old CVS hooks, and minimal updates > to the web pages

Re: [Patch 1/2][ipa] Add target hook to sanitize cloned declaration's attributes

Hi Martin, Thanks for the suggestion. I have moved the target hook to the following functions: 'create_clone' 'create_version_clone' 'create_version_clone_with_body' I looked for all 'cgraph_node::create_*' functions and added the target hook to them. The exception is 'create_virtual_clone'

Re: [PATCH] Come up with ipa passes introduction in gccint documentation

On Tue, Oct 8, 2019 at 10:06 PM Sandra Loosemore wrote: > > On 10/8/19 2:52 AM, luoxhu wrote: > > Hi, > > > > This is the formal documentation patch for IPA passes. Thanks. > > > > > > None of the IPA passes are documented in passes.texi. This patch adds > > a section IPA passes just before

Re: [Patch] gcc-10/changes.html – mention OpenMP's use_device_addr support

On Wed, 9 Oct 2019, Tobias Burnus wrote: > This mentioned the recently added OpenMP 5 feature "use_device_addr" > which complements the (for Fortran) underspecified "use_device_ptr". The > rest of the message, I have taken and modified from gcc-7/changes.html. Please use instead of and break

Re: C++ PATCH for C++20 P0388R4 (conversions to arrays of unknown bounds) and CWG 1307 (c++/91364, c++/69531)

On 07/10/19 14:56 -0400, Jason Merrill wrote: On 10/7/19 1:42 PM, Marek Polacek wrote: @@ -7401,8 +7432,20 @@ convert_like_real (conversion *convs, tree expr, tree fn, int argnum, error_at (loc, "cannot bind non-const lvalue reference of " "type %qH to an

Re: [Patch 1/2][ipa] Add target hook to sanitize cloned declaration's attributes

Hi, On Wed, Oct 09 2019, Andre Vieira (lists) wrote: > Hi Martin, > > Thanks for the suggestion. I have moved the target hook to the > following functions: > 'create_clone' > 'create_version_clone' I am not sure that calling the hook in create_version_clone is what you want. It should perhaps

Re: Make C2X imply -fno-fp-int-builtin-inexact

Hi Joseph, > gcc/testsuite: > 2019-10-08 Joseph Myers > > * gcc.dg/torture/builtin-fp-int-inexact-c2x.c: New test. [...] > Index: gcc/testsuite/gcc.dg/torture/builtin-fp-int-inexact-c2x.c > === > ---

[PATCH] Relax nested cycle vectorization further

This simplifies and refactors vect_is_simple_reduction to make sure to not reject nested cycle vectorization just beacuse there are calls in the innermost loop. This lets us vectorize the new testcase using outer loop vectorization. Bootstrapped on x86_64-unknown-linux-gnu, testing in

Re: GCC wwwdocs move to git done

On Wed, Oct 09, 2019 at 01:25:30PM +0100, Iain Sandoe wrote: >Jonathan Wakely wrote: > >>On Wed, 9 Oct 2019 at 01:28, Joseph Myers wrote: >>>I've done the move of GCC wwwdocs to git (using the previously posted and >>>discussed scripts), including setting up the post-receive hook to do the

PR92036 "OpenACC 'firstprivate' clause: initial value"

Hi! In r276757 "[PR92036] Add 'libgomp.oacc-c-c++-common/data-firstprivate-1.c'", I committed the attached to document the status quo, and then what I think it instead should be per PR92036 "OpenACC 'firstprivate' clause: initial value". Grüße Thomas From

Re: [PATCH 2/2][MSP430] Optimize zero_extend insns and PSImode pointer manipulation

Hi Segher! On Wed, 9 Oct 2019 03:47:32 -0500 Segher Boessenkool wrote: > Hi Jozef, > > On Tue, Oct 08, 2019 at 11:39:57AM +0100, Jozef Lawrynowicz wrote: > > +;; > > +;; ZERO EXTEND INSTRUCTIONS > > +;; Byte-writes to registers clear bits 19:8 > > +;; * Byte-writes

Re: GCC wwwdocs move to git done

On Wed, 2019-10-09 at 09:37 -0400, Christopher Faylor wrote: > On Wed, Oct 09, 2019 at 01:25:30PM +0100, Iain Sandoe wrote: > > Jonathan Wakely wrote: > > > > > On Wed, 9 Oct 2019 at 01:28, Joseph Myers wrote: > > > > I've done the move of GCC wwwdocs to git (using the previously > > > > posted

Re: GCC wwwdocs move to git done

On Wed, 2019-10-09 at 10:44 +0100, Jonathan Wakely wrote: > On Wed, 9 Oct 2019 at 01:28, Joseph Myers wrote: > > > > I've done the move of GCC wwwdocs to git (using the previously > > posted and > > discussed scripts), including setting up the post-receive hook to > > do the > > same things

Re: [PATCH, nvptx] Expand OpenACC child function arguments to use CUDA params space

On 01-10-2019 14:45, Chung-Lin Tang wrote: > Index: gcc/config/nvptx/nvptx.c > === > --- gcc/config/nvptx/nvptx.c (revision 276406) > +++ gcc/config/nvptx/nvptx.c (working copy) > @@ -68,6 +68,10 @@ > #include "attribs.h" >

Re: [libgomp][testsuite] PR testsuite/91884 Add -lquadmath if available

Hi Tobias, > I have also an alternative version: > > * For in-build-dir test runs, do as with previous patch, i.e. add > "-lquadmath" to the "xgcc" call if the directory is available. > > * For installed-compiler test runs, the new patch uses GFORTRAN_UNDER_TEST > / GXX_UNDER_TEST > (It would

Re: GCC wwwdocs move to git done

Christopher Faylor wrote: On Wed, Oct 09, 2019 at 01:25:30PM +0100, Iain Sandoe wrote: Jonathan Wakely wrote: On Wed, 9 Oct 2019 at 01:28, Joseph Myers wrote: I've done the move of GCC wwwdocs to git (using the previously posted and discussed scripts), including setting up the

Re: GCC wwwdocs move to git done

Jonathan Wakely wrote: On Wed, 9 Oct 2019 at 01:28, Joseph Myers wrote: I've done the move of GCC wwwdocs to git (using the previously posted and discussed scripts), including setting up the post-receive hook to do the same things previously covered by the old CVS hooks, and minimal updates

[PATCH] PR libstdc++/78552 only construct std::locale for C locale once

PR libstdc++/78552 * src/c++98/locale_init.cc (locale::classic()): Do not construct a new locale object for every call. (locale::_S_initialize_once()): Construct C locale here. Tested x86_64-linux, committed to trunk. This should be safe to backport too, but I'll

Re: [PATCH, nvptx] Expand OpenACC child function arguments to use CUDA params space

-BEGIN PGP SIGNED MESSAGE- Hash: SHA256 On 08-10-2019 16:05, Thomas Schwinge wrote: > Hi Chung-Lin! > > While we're all waiting for Tom to comment on this ;-) Ack, thanks for the ping ... > -- here's another item I realized: > > On 2019-09-10T19:41:59+0800, Chung-Lin Tang > wrote: >>

[PATCH][LTO] Do not merge anonymous NAMESPACE_DECLs

The following does $subject which likely fixes some debuginfo issues with anonymous namespaces (but likely artificial and not too noticable?! we'd get bogus contexts for DIEs) Bootstrapped and tested on x86_64-unknown-linux-gnu. I'm doin a LTO bootstrap right now (but IIRC it failed a few

Re: GCC wwwdocs move to git done

On Wed, Oct 09, 2019 at 09:37:54AM -0400, Christopher Faylor wrote: >The binary is 20 years old and, somehow, the source code used to build >it seems to have disappeared. Sorry. Not 20 years: -rwxr-xr-x. 1 root root 574221 Mar 22 2013 /usr/local/bin/mhc cgf

Re: [C++ Patch] Fix cp_build_binary_op locations

On 10/9/19 10:23 AM, Paolo Carlini wrote: Hi, this largely mechanical patch fixes those errors in cp_build_binary_op which weren't exploiting yet the op_location_t argument, thus aligning the locations to those of the C front-end. Plus an additional use for DECL_SOURCE_LOCATION in decl.c.

[C++ Patch] Fix cp_build_binary_op locations

Hi, this largely mechanical patch fixes those errors in cp_build_binary_op which weren't exploiting yet the op_location_t argument, thus aligning the locations to those of the C front-end. Plus an additional use for DECL_SOURCE_LOCATION in decl.c. Tested x86_64-linux. Thanks, Paolo.

Re: [Patch 1/2][ipa] Add target hook to sanitize cloned declaration's attributes

Hi Martin, I see thank you. For my particular use case I only need to strip attributes for local symbols. However, if I were to claim it is only used for non-externally visible clones than I would also need to make sure they are non-externally visible now and put something in place that

[PATCH][gcov-profile/91971]Profile directory concatenated with object file path

Hi, This is a patch for fix PR 91971: Profile directory concatenated with object file path https://gcc.gnu.org/bugzilla/show_bug.cgi?id=91971 Gcc/ChangeLog: +2019-10-09 qing zhao + + * coverage.c (coverage_init): Mangle the full path of filename when + filename is a absolute

[PATCH] PR libstdc++/91057 set locale::id::_M_index atomically

If two threads see _M_index==0 concurrently they will both try to set it, potentially storing the facet at two different indices in the array. Either set the _M_index data member using an atomic compare-exchange operation or while holding a mutex. Also move the LONG_DOUBLE_COMPAT code into a

Re: GCC wwwdocs move to git done

On Wed, 2019-10-09 at 15:06 +, Joseph Myers wrote: > On Tue, 8 Oct 2019, Frank Ch. Eigler wrote: > > > Hi - > > > > Thanks - good job with moving this to git! +1 > > > Note 1: someone with the right access needs to create the > > > symlink > > > /sourceware/git/gcc-wwwdocs.git -> > > >

Re: GCC wwwdocs move to git done

On Tue, 8 Oct 2019, Frank Ch. Eigler wrote: > Hi - > > Thanks - good job with moving this to git! > > > Note 1: someone with the right access needs to create the symlink > > /sourceware/git/gcc-wwwdocs.git -> > > /sourceware/projects/gcc-home/wwwdocs.git (and anything else needed for > >

Re: [Patch 1/2][ipa] Add target hook to sanitize cloned declaration's attributes

Adding Jakub to CC' On 09/10/2019 16:29, Andre Vieira (lists) wrote: Hi Martin, I see thank you. For my particular use case I only need to strip attributes for local symbols.  However, if I were to claim it is only used for non-externally visible clones than I would also need to make sure

Re: C++ PATCH for c++/92032 - DR 1601: Promotion of enum with fixed underlying type

OK. On 10/8/19 6:14 PM, Marek Polacek wrote: I've been messing with compare_ics recently and noticed that we don't implement CWG 1601, which should be fairly easy. The motivating example is enum E : char { e }; void f(char); void f(int); void g() { f(e); } where the call

[PR target/85401][v2] Add test-cases

On Fri, Oct 04, 2019 at 02:28:55PM -0600, Jeff Law wrote: > On 10/4/19 1:43 PM, co...@sdf.org wrote: > > On Tue, Oct 01, 2019 at 01:26:16PM -0600, Jeff Law wrote: > >> On 9/30/19 2:45 PM, co...@sdf.org wrote: > >>> On Mon, Sep 30, 2019 at 11:46:24AM -0400, Vladimir Makarov wrote: > Yes, the

PowerPC future machine patches, version 5

I'm going to re-issue the outstanding 'future' patches as separate threads under this message. Compared to the original V4 patches, I have tried to break the larger patches (particularly #4) into smaller chunks. One of the V4 patches that I submitted was wrong, and this series of patches

[PATCH] V5, #1 of 15: Support 34-bit offsets for prefixed instructions

This patch adds support in the various functions that check memory offsets for the 34-bit offset with prefixed instructions on the 'future' machine. Along with the other patches, I have done bootstraps on a little endian power8 system, and there were no regressions in the test suite. Can I check

[PATCH] V5, #2 of 15: Fix prefixed instruction length for some 128-bit types

This patch fixes the prefixed and non-prefixed instruction sizes for the 128-bit types that aren't loaded into 128-bit vectors (TDmode, TFmode, IFmode, PTImode). Along with the other patches, I have done bootstraps on a little endian power8 system, and there were no regressions in the test suite.

[PATCH] V5, #3 of 15: Deal with prefixed instructions in rs6000_insn_cost

This patch is a simplification of the previous patch to adjust the rtx insn cost for prefixed instructions. Rather than making it a separate function with only one caller, I folded the code into rs6000_insn_cost. You had asked for this to be a separate patch, so it is in this patch. The basic

[PATCH] V5, #4 of 15: Update predicates

This patch updates the predicates. It allows lwa_operand (used by extendsidi2) to generate offsets that can be used with PLWA but not LWA (due to LWA using the DS instruction format). It also adds two new predicates that disallow prefixed memory that will used in patches #5 and #7. Along with

[PATCH] Modify simplify_truncation to handle extended CONST_INT.

This addresses PR 91860 which has four testcases triggering internal errors. The problem here is that in combine when handling debug insns, we are trying to substitute (sign_extend:DI (const_int 8160 [0x1fe0])) as the value for (reg:DI 78 [ _9 ]) in the debug insn (debug_insn 29 28 30 2

Re: Add std::copy_n overload for istreambuf_iterator

On Fri, 4 Oct 2019 at 07:01, François Dumont wrote: > On 9/27/19 1:00 PM, Jonathan Wakely wrote: > > On 19/07/19 23:37 +0200, François Dumont wrote: > >> It sounds reasonable to overload std::copy_n for istreambuf_iterator. > > I wonder whether it's worth doing: > > > > #if __cplusplus >=

[PATCH] V5, #5 of 15: Support -fstack-protect and large stack frames

This patch allows -fstack-protect to work with large stack frames. It does so by adding a new constraint (em) that matches normal memory that does not use a prefixed address (including PC-relative addresses). The stack protect test and set functions now can take indexed address forms, and there

[PATCH] PR fortran/92019 -- BOZ cannot be an array index

The attach patch fixes an ICE caused by the use of a bOZ as an array index. Tested on x86_64-*-freebsd. OK to commit? 2019-10-09 Steven G. Kargl PR fortran/92019 * array.c (match_subscript): BOZ cannot be an array subscript. 2019-10-09 Steven G. Kargl PR

[PATCH] V5, #6 of 15: Make vector load/store instruction length correct with prefixed addresses

This patch updates the instruction length for 128-bit move insns for types that go into a single vector register. It is a lot simpler than the previous patch, which was written before I added the prefixed_length and non_prefixed_length attributes. Along with the other patches, I have done

[OG9][committed] Fix OpenMP's use_device_ptr with Fortran array descriptors

The committed/attached patch fixes OpenMP 4.5's use_device_ptr(var) for variables with array descriptor. This is the OG9 version of the GCC-10 trunk patch: [1] https://gcc.gnu.org/ml/fortran/2019-09/msg00088.html [still awaiting review] [On trunk, it makes sense to have first: [2]

Re: C++ PATCH for C++20 P0388R4 (conversions to arrays of unknown bounds) and CWG 1307 (c++/91364, c++/69531)

On Tue, Oct 08, 2019 at 01:50:17PM -0400, Jason Merrill wrote: > > > That sounds like a bug in grok_reference_init; it isn't properly > > > implementing > > > > > > "Otherwise, if the initializer list has a single element of type E and > > > either T is not a reference type or its > > >

{PATCH v3, rs6000] Replace X-form addressing with D-form addressing in new pass for Power9

This patch is a refinement of a patch first submitted to this list on Nov. 10, 2018, with revisions submitted this list on Dec. 13, 2018 and Sep. 3, 2019. This new pass scans existing rtl expressions and replaces them with rtl expressions that favor selection of the D-form instructions in

Re: Add std::copy_n overload for istreambuf_iterator

On 10/9/19 10:18 PM, Christophe Lyon wrote: On Fri, 4 Oct 2019 at 07:01, François Dumont > wrote: On 9/27/19 1:00 PM, Jonathan Wakely wrote: > On 19/07/19 23:37 +0200, François Dumont wrote: >> It sounds reasonable to overload std::copy_n for

Re: C++ PATCH for C++20 P0388R4 (conversions to arrays of unknown bounds) and CWG 1307 (c++/91364, c++/69531)

On Wed, Oct 09, 2019 at 11:01:39AM +0100, Jonathan Wakely wrote: > On 07/10/19 14:56 -0400, Jason Merrill wrote: > > On 10/7/19 1:42 PM, Marek Polacek wrote: > > > @@ -7401,8 +7432,20 @@ convert_like_real (conversion *convs, tree expr, > > > tree fn, int argnum, > > > error_at (loc,

[PATCH] V5, #7 of 15: Restrict vector extract to not use prefixed memory

This patch fixes an issue Aaron Sawdey with vector extracts. The vector address adjust functions have a single base register temporary. However for PC-relative addresses where the vector element is not constant, it logically will need two temporary registers, one to load up the PC-relative

[Darwin, committed] Improve Objective-C NeXT ABI version check.

While doing some wider testing on a patch to remove a build warning, I noticed we were missing a check for the case that user's ABI version was set > 2 and the codegen was for 64 bit. Fixed thus, tested on x86_64-dawin16, Applied to mainline, thanks Iain gcc/ChangeLog: 2019-10-09 Iain Sandoe

[Darwin, machopic 3/n, committed] Set a SYMBOL flag for indirections.

We are able to treat these specially where needed in legitimate address tests (specifically, they are guaranteed to be pointer-aligned). tested on x86_64-darwin16, powerpc-darwin9. applied to mainline, thanks Iain gcc/ChangeLog: 2019-10-09 Iain Sandoe * config/darwin.c

[PATCH] V5, #8 of 15: Use PADDI/PLI to load up 34-bit DImode constants

This patch uses PADDI (PLI) to load up 34-bit DImode constants. In the V3 time frame, it was part of a larger patch, but you asked that this be a separate patch. Along with the other patches, I have done bootstraps on a little endian power8 system, and there were no regressions in the test

[PATCH] V5, #9 of 15: Add PADDI (PLI) support to load up 32-bit SImode constants

This patch is a companion patch to patch #8, and it allows using the PADDI (PLI) instruction to load up 32-bit SImode constants rather than generating separate ADDI + ADDIS instructions. Along with the other patches, I have done bootstraps on a little endian power8 system, and there were no

C++ PATCH to fix typo in test name

Applying as obvious: A +initlist-array1.C > moved from initlist-arrray1.C D initlist-arrray1.C > moved to initlist-array1.C -- Marek Polacek • Red Hat, Inc. • 300 A St, Boston, MA

Re: C++ PATCH for C++20 P0388R4 (conversions to arrays of unknown bounds) and CWG 1307 (c++/91364, c++/69531)

On 10/9/19 4:26 PM, Marek Polacek wrote: On Tue, Oct 08, 2019 at 01:50:17PM -0400, Jason Merrill wrote: That sounds like a bug in grok_reference_init; it isn't properly implementing "Otherwise, if the initializer list has a single element of type E and either T is not a reference type or its

[PATCH] V5, #10 of 15: Use PADDI to add large 34-bit constants

This patch enables generating PADDI to add 34-bit constants. This is the last of the patches in V5 to add the necessary support for the 'future' machine. Along with the other patches, I have done bootstraps on a little endian power8 system, and there were no regressions in the test suite. Can I

[PATCH] V5, #11 of 15: Make -mpcrel default on Linux 64-bit

This patch assumes patches #1 through #10 have been applied. Once all of those patches have been applied, this patch will change the default for Linux 64-bit systems to enable -mpcrel and -mprefixed-addr being default when -mcpu=future is used. The other OS/bit-size targets (currently) do not

Re: Make C2X imply -fno-fp-int-builtin-inexact

On Wed, 9 Oct 2019, Rainer Orth wrote: > the new test FAILs on Solaris/x86, both 32 and 64-bit: > > FAIL: gcc.dg/torture/builtin-fp-int-inexact-c2x.c -O0 execution test > FAIL: gcc.dg/torture/builtin-fp-int-inexact-c2x.c -O1 execution test > FAIL: