[Bug libffi/109447] New: test case libffi.closures/cls_align_longdouble_split.c fails

[Bug libffi/109447] New: test case libffi.closures/cls_align_longdouble_split.c fails

[seurer at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109447

            Bug ID: 109447
           Summary: test case libffi.closures/cls_align_longdouble_split.c
                    fails
           Product: gcc
           Version: 12.2.1
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: libffi
          Assignee: unassigned at gcc dot gnu.org
          Reporter: seurer at gcc dot gnu.org
  Target Milestone: ---

When running make check on libffi there are two failures.  Note that this only
occurs on systems where the compiler was built with
--with-long-double-format=ieee.  I see these on both gcc 12 and trunk.  I don't
know yet if it is a regression but I am checking.

On powerpc the libffi library is built and tested if go is included in the
languages.

FAIL: libffi.closures/cls_align_longdouble_split.c -W -Wall -Wno-psabi -O0
output pattern test
FAIL: libffi.closures/cls_align_longdouble_split.c -W -Wall -Wno-psabi -O2
output pattern test

                === libffi Summary ===

# of expected passes            1460
# of unexpected failures        2
# of unsupported tests          28


spawn -ignore SIGHUP /home/seurer/gcc/git/build/gcc-test/./gcc/xgcc
-B/home/seurer/gcc/git/build/gcc-test/./gcc/
-B/home/seurer/gcc/git/install/gcc-test/powerpc64le-unknown-linux-gnu/bin/
-B/home/seurer/gcc/git/install/gcc-test/powerpc64le-unknown-linux-gnu/lib/
-isystem
/home/seurer/gcc/git/install/gcc-test/powerpc64le-unknown-linux-gnu/include
-isystem
/home/seurer/gcc/git/install/gcc-test/powerpc64le-unknown-linux-gnu/sys-include
/home/seurer/gcc/git/gcc-test/libffi/testsuite/libffi.closures/cls_align_longdouble_split.c
-W -Wall -Wno-psabi -O0 -mlong-double-128
-I/home/seurer/gcc/git/build/gcc-test/powerpc64le-unknown-linux-gnu/./libffi/include
-I/home/seurer/gcc/git/gcc-test/libffi/testsuite/../include
-I/home/seurer/gcc/git/build/gcc-test/powerpc64le-unknown-linux-gnu/./libffi/include/..
-L/home/seurer/gcc/git/build/gcc-test/powerpc64le-unknown-linux-gnu/./libffi/.libs
-L/home/seurer/gcc/git/build/gcc-test/powerpc64le-unknown-linux-gnu/./libstdc++-v3/src/.libs
-lffi -lm -o ./cls_align_longdouble_split.exe^M
PASS: libffi.closures/cls_align_longdouble_split.c -W -Wall -Wno-psabi -O0
(test for excess errors)
Setting LD_LIBRARY_PATH to
.:/home/seurer/gcc/git/build/gcc-test/gcc:/home/seurer/gcc/git/build/gcc-test/powerpc64le-unknown-linux-gnu/./libffi/.libs:/home/seurer/gcc/git/build/gcc-test/powerpc64le-unknown-linux-gnu/./libstdc++-v3/src/.libs:.:/home/seurer/gcc/git/build/gcc-test/gcc:/home/seurer/gcc/git/build/gcc-test/powerpc64le-unknown-linux-gnu/./libffi/.libs:/home/seurer/gcc/git/build/gcc-test/powerpc64le-unknown-linux-gnu/./libstdc++-v3/src/.libs:
Execution timeout is: 300
spawn [open ...]^M
1 1 1 1 1 1 1 8 8 8 8 8 8 9: 9 9 9 9 9 9 10
res: 9 9 9 9 9 9 10
1 2 3 4 5 6 7 8 9 10 11 12 13 14: 9 11 13 15 17 19 21
res: 9 11 13 15 17 19 21
PASS: libffi.closures/cls_align_longdouble_split.c -W -Wall -Wno-psabi -O0
execution test
FAIL: libffi.closures/cls_align_longdouble_split.c -W -Wall -Wno-psabi -O0
output pattern test
Output was:
1 1 1 1 1 1 1 8 8 8 8 8 8 9: 9 9 9 9 9 9 10
res: 9 9 9 9 9 9 10
1 2 3 4 5 6 7 8 9 10 11 12 13 14: 9 11 13 15 17 19 21
res: 9 11 13 15 17 19 21

Should match:
1 2 3 4 5 6 7 8 9 10 11 12 13 14: 9 11 13 15 17 19 21^M?
res: 9 11 13 15 17 19 21^M?
1 2 3 4 5 6 7 8 9 10 11 12 13 14: 9 11 13 15 17 19 21^M?
res: 9 11 13 15 17 19 21

spawn -ignore SIGHUP /home/seurer/gcc/git/build/gcc-test/./gcc/xgcc
-B/home/seurer/gcc/git/build/gcc-test/./gcc/
-B/home/seurer/gcc/git/install/gcc-test/powerpc64le-unknown-linux-gnu/bin/
-B/home/seurer/gcc/git/install/gcc-test/powerpc64le-unknown-linux-gnu/lib/
-isystem
/home/seurer/gcc/git/install/gcc-test/powerpc64le-unknown-linux-gnu/include
-isystem
/home/seurer/gcc/git/install/gcc-test/powerpc64le-unknown-linux-gnu/sys-include
/home/seurer/gcc/git/gcc-test/libffi/testsuite/libffi.closures/cls_align_longdouble_split.c
-W -Wall -Wno-psabi -O2 -mlong-double-128
-I/home/seurer/gcc/git/build/gcc-test/powerpc64le-unknown-linux-gnu/./libffi/include
-I/home/seurer/gcc/git/gcc-test/libffi/testsuite/../include
-I/home/seurer/gcc/git/build/gcc-test/powerpc64le-unknown-linux-gnu/./libffi/include/..
-L/home/seurer/gcc/git/build/gcc-test/powerpc64le-unknown-linux-gnu/./libffi/.libs
-L/home/seurer/gcc/git/build/gcc-test/powerpc64le-unknown-linux-gnu/./libstdc++-v3/src/.libs
-lffi -lm -o ./cls_align_longdouble_split.exe^M
PASS: libffi.closures/cls_align_longdouble_split.c -W -Wall -Wno-psabi -O2
(test for excess errors)
Setting LD_LIBRARY_PATH to
.:/home/seurer/gcc/git/build/gcc-test/gcc:/home/seurer/gcc/git/build/gcc-test/powerpc64le-unknown-linux-gnu/./libffi/.libs:/home/seurer/gcc/git/build/gcc-test/powerpc64le-unknown-linux-gnu/./libstdc++-v3/src/.libs:.:/home/seurer/gcc/git/build/gcc-test/gcc:/home/seurer/gcc/git/build/gcc-test/powerpc64le-unknown-linux-gnu/./libffi/.libs:/home/seurer/gcc/git/build/gcc-test/powerpc64le-unknown-linux-gnu/./libstdc++-v3/src/.libs:
Execution timeout is: 300
spawn [open ...]^M
1 1 1 1 1 1 1 8 8 8 8 8 8 9: 9 9 9 9 9 9 10
res: 9 9 9 9 9 9 10
1 2 3 4 5 6 7 8 9 10 11 12 13 14: 9 11 13 15 17 19 21
res: 9 11 13 15 17 19 21
PASS: libffi.closures/cls_align_longdouble_split.c -W -Wall -Wno-psabi -O2
execution test
FAIL: libffi.closures/cls_align_longdouble_split.c -W -Wall -Wno-psabi -O2
output pattern test
Output was:
1 1 1 1 1 1 1 8 8 8 8 8 8 9: 9 9 9 9 9 9 10
res: 9 9 9 9 9 9 10
1 2 3 4 5 6 7 8 9 10 11 12 13 14: 9 11 13 15 17 19 21
res: 9 11 13 15 17 19 21

Should match:
1 2 3 4 5 6 7 8 9 10 11 12 13 14: 9 11 13 15 17 19 21^M?
res: 9 11 13 15 17 19 21^M?
1 2 3 4 5 6 7 8 9 10 11 12 13 14: 9 11 13 15 17 19 21^M?
res: 9 11 13 15 17 19 21

[Bug libffi/109447] test case libffi.closures/cls_align_longdouble_split.c fails

[schwab@linux-m68k.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109447

--- Comment #1 from Andreas Schwab <schwab@linux-m68k.org> ---
Does that also happen with the latest sources at
<http://github.com/libffi/libffi>?

[Bug libffi/109447] test case libffi.closures/cls_align_longdouble_split.c fails

[bergner at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109447

--- Comment #2 from Peter Bergner <bergner at gcc dot gnu.org> ---
(In reply to Andreas Schwab from comment #1)
> Does that also happen with the latest sources at
> <http://github.com/libffi/libffi>?

I downloaded current sources on the same system (a long double == IEEE128
system) Bill ran on and I do not see that error in the testsuite:

PASS: libffi.closures/cls_align_longdouble_split.c -W -Wall -Wno-psabi -O0
(test for excess errors)
PASS: libffi.closures/cls_align_longdouble_split.c -W -Wall -Wno-psabi -O0
execution test
PASS: libffi.closures/cls_align_longdouble_split.c -W -Wall -Wno-psabi -O0
output pattern test

PASS: libffi.closures/cls_align_longdouble_split.c -W -Wall -Wno-psabi -O2
(test for excess errors)
PASS: libffi.closures/cls_align_longdouble_split.c -W -Wall -Wno-psabi -O2
execution test
PASS: libffi.closures/cls_align_longdouble_split.c -W -Wall -Wno-psabi -O2
output pattern test

...so it looks like GCC's libffi is just missing a fix already in upstream
sources.

[Bug libffi/109447] test case libffi.closures/cls_align_longdouble_split.c fails

[bergner at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109447

--- Comment #3 from Peter Bergner <bergner at gcc dot gnu.org> ---
Hmmm, libffi commit 464b4b66e3cf3b5489e730c1466ee1bf825560e0 seems to be
missing from GCC's sources.  I'll try applying that and see if it fixes the
problem.

commit 464b4b66e3cf3b5489e730c1466ee1bf825560e0
Author:     Dan Horák <dan@danny.cz>
AuthorDate: Mon Jan 31 14:08:09 2022 +0100
Commit:     GitHub <noreply@github.com>
CommitDate: Mon Jan 31 08:08:09 2022 -0500

    powerpc64: fix handling of homogeneous float128 structs (#689)

    If there is a homogeneous struct with float128 members, they should be
    copied to vector register save area. The current code incorrectly copies
    only the value of the first member, not increasing the pointer with each
    iteration. Fix this.

Given the output Bill is seeing and the description of the git log entry, it
seems like this may be the culprit.

[Bug libffi/109447] test case libffi.closures/cls_align_longdouble_split.c fails

[bergner at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109447

Peter Bergner <bergner at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Last reconfirmed|                            |2023-4-8
                 CC|                            |segher at gcc dot gnu.org

--- Comment #4 from Peter Bergner <bergner at gcc dot gnu.org> ---
(In reply to Peter Bergner from comment #3)
> commit 464b4b66e3cf3b5489e730c1466ee1bf825560e0
> Author:     Dan Horák <dan@danny.cz>
> AuthorDate: Mon Jan 31 14:08:09 2022 +0100
> Commit:     GitHub <noreply@github.com>
> CommitDate: Mon Jan 31 08:08:09 2022 -0500
> 
>     powerpc64: fix handling of homogeneous float128 structs (#689)
>     
>     If there is a homogeneous struct with float128 members, they should be
>     copied to vector register save area. The current code incorrectly copies
>     only the value of the first member, not increasing the pointer with each
>     iteration. Fix this.

Applying this commit does fix the test case FAILs.

[Bug libffi/109447] test case libffi.closures/cls_align_longdouble_split.c fails

[bergner at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109447

Peter Bergner <bergner at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
     Ever confirmed|0                           |1
             Status|UNCONFIRMED                 |NEW

[Bug libffi/109447] test case libffi.closures/cls_align_longdouble_split.c fails

[segher at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109447

Segher Boessenkool <segher at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |green at gcc dot gnu.org

--- Comment #5 from Segher Boessenkool <segher at gcc dot gnu.org> ---
That
(<https://github.com/libffi/libffi/commit/464b4b66e3cf3b5489e730c1466ee1bf825560e0>)
looks like an obvious fix.  Can we get it applied to GCC please?

(cc:ing the libffi author)

[Bug libffi/109447] test case libffi.closures/cls_align_longdouble_split.c fails

[bergner at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109447

Peter Bergner <bergner at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
      Known to fail|                            |12.0, 13.0

--- Comment #6 from Peter Bergner <bergner at gcc dot gnu.org> ---
(In reply to Segher Boessenkool from comment #5)
> That
> (<https://github.com/libffi/libffi/commit/
> 464b4b66e3cf3b5489e730c1466ee1bf825560e0>) looks like an obvious fix.  Can
> we get it applied to GCC please?
> 
> (cc:ing the libffi author)

I just checked and GCC12 needs the fix too.  I don't know the process for how
we apply upstream libffi changes to GCC sources.

[Bug libffi/109447] test case libffi.closures/cls_align_longdouble_split.c fails

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109447

--- Comment #7 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Peter Bergner <bergner@gcc.gnu.org>:

https://gcc.gnu.org/g:043550bceea7881163bba5d8a0486bb100a05809

commit r14-542-g043550bceea7881163bba5d8a0486bb100a05809
Author: Dan Horák <dan@danny.cz>
Date:   Wed May 3 14:29:09 2023 -0500

    libffi: fix handling of homogeneous float128 structs (#689)

    If there is a homogeneous struct with float128 members, they should be
    copied to vector register save area. The current code incorrectly copies
    only the value of the first member, not increasing the pointer with each
    iteration. Fix this.

    Merged from upstream libffi commit:
464b4b66e3cf3b5489e730c1466ee1bf825560e0

    2023-05-03  Dan Horák <dan@danny.cz>

    libffi/
            PR libffi/109447
            * src/powerpc/ffi_linux64.c (ffi_prep_args64): Update arg.f128
pointer.

[Bug libffi/109447] test case libffi.closures/cls_align_longdouble_split.c fails

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109447

--- Comment #8 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-13 branch has been updated by Peter Bergner
<bergner@gcc.gnu.org>:

https://gcc.gnu.org/g:36629645d5ba6a6612c87af1fe7249ea1fb8cdc5

commit r13-7302-g36629645d5ba6a6612c87af1fe7249ea1fb8cdc5
Author: Dan Horák <dan@danny.cz>
Date:   Wed May 3 14:29:09 2023 -0500

    libffi: fix handling of homogeneous float128 structs (#689)

    If there is a homogeneous struct with float128 members, they should be
    copied to vector register save area. The current code incorrectly copies
    only the value of the first member, not increasing the pointer with each
    iteration. Fix this.

    Merged from upstream libffi commit:
464b4b66e3cf3b5489e730c1466ee1bf825560e0

    2023-05-03  Dan Horák <dan@danny.cz>

    libffi/
            PR libffi/109447
            * src/powerpc/ffi_linux64.c (ffi_prep_args64): Update arg.f128
pointer.

    (cherry picked from commit 043550bceea7881163bba5d8a0486bb100a05809)

[Bug libffi/109447] test case libffi.closures/cls_align_longdouble_split.c fails

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109447

--- Comment #9 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-12 branch has been updated by Peter Bergner
<bergner@gcc.gnu.org>:

https://gcc.gnu.org/g:0069ce6631880776a44fb8408c6428c0812285cf

commit r12-9526-g0069ce6631880776a44fb8408c6428c0812285cf
Author: Dan Horák <dan@danny.cz>
Date:   Wed May 3 14:29:09 2023 -0500

    libffi: fix handling of homogeneous float128 structs (#689)

    If there is a homogeneous struct with float128 members, they should be
    copied to vector register save area. The current code incorrectly copies
    only the value of the first member, not increasing the pointer with each
    iteration. Fix this.

    Merged from upstream libffi commit:
464b4b66e3cf3b5489e730c1466ee1bf825560e0

    2023-05-03  Dan Horák <dan@danny.cz>

    libffi/
            PR libffi/109447
            * src/powerpc/ffi_linux64.c (ffi_prep_args64): Update arg.f128
pointer.

    (cherry picked from commit 043550bceea7881163bba5d8a0486bb100a05809)