[Bug target/105613] New: [13 Regression] wrong code with unsigned V1TI comparison

[Bug target/105613] New: [13 Regression] wrong code with unsigned V1TI comparison

[zsojka at seznam dot cz]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105613

            Bug ID: 105613
           Summary: [13 Regression] wrong code with unsigned V1TI
                    comparison
           Product: gcc
           Version: 13.0
            Status: UNCONFIRMED
          Keywords: wrong-code
          Severity: normal
          Priority: P3
         Component: target
          Assignee: unassigned at gcc dot gnu.org
          Reporter: zsojka at seznam dot cz
  Target Milestone: ---
              Host: x86_64-pc-linux-gnu
            Target: x86_64-pc-linux-gnu

Created attachment 52982
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=52982&action=edit
reduced testcase

Output:
$ x86_64-pc-linux-gnu-gcc testcase.c
$ ./a.out 
Aborted

(gdb) p r
$1 = {0}

$ x86_64-pc-linux-gnu-gcc -v
Using built-in specs.
COLLECT_GCC=/repo/gcc-trunk/binary-latest-amd64/bin/x86_64-pc-linux-gnu-gcc
COLLECT_LTO_WRAPPER=/repo/gcc-trunk/binary-trunk-r13-468-20220516083034-g69c4b5c519f-checking-yes-rtl-df-extra-nobootstrap-amd64/bin/../libexec/gcc/x86_64-pc-linux-gnu/13.0.0/lto-wrapper
Target: x86_64-pc-linux-gnu
Configured with: /repo/gcc-trunk//configure --enable-languages=c,c++
--enable-valgrind-annotations --disable-nls --enable-checking=yes,rtl,df,extra
--disable-bootstrap --with-cloog --with-ppl --with-isl
--build=x86_64-pc-linux-gnu --host=x86_64-pc-linux-gnu
--target=x86_64-pc-linux-gnu --with-ld=/usr/bin/x86_64-pc-linux-gnu-ld
--with-as=/usr/bin/x86_64-pc-linux-gnu-as --disable-libstdcxx-pch
--prefix=/repo/gcc-trunk//binary-trunk-r13-468-20220516083034-g69c4b5c519f-checking-yes-rtl-df-extra-nobootstrap-amd64
Thread model: posix
Supported LTO compression algorithms: zlib zstd
gcc version 13.0.0 20220516 (experimental) (GCC)

[Bug target/105613] [13 Regression] wrong code with unsigned V1TI comparison

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105613

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |13.0

[Bug target/105613] [13 Regression] wrong code with unsigned V1TI comparison

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105613

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
     Ever confirmed|0                           |1
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2022-05-16

--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Confirmed.

[Bug target/105613] [13 Regression] wrong code with unsigned V1TI comparison since r13-458-gd75d4293dcc029

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105613

Martin Liška <marxin at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |marxin at gcc dot gnu.org
            Summary|[13 Regression] wrong code  |[13 Regression] wrong code
                   |with unsigned V1TI          |with unsigned V1TI
                   |comparison                  |comparison since
                   |                            |r13-458-gd75d4293dcc029

--- Comment #2 from Martin Liška <marxin at gcc dot gnu.org> ---
Started with r13-458-gd75d4293dcc029.

[Bug target/105613] [13 Regression] wrong code with unsigned V1TI comparison since r13-458-gd75d4293dcc029

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105613

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P1

[Bug target/105613] [13 Regression] wrong code with unsigned V1TI comparison since r13-458-gd75d4293dcc029

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105613

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED
           Assignee|unassigned at gcc dot gnu.org      |jakub at gcc dot gnu.org
                 CC|                            |jakub at gcc dot gnu.org

--- Comment #3 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
We want to use andv4si3 only for EQ, for NE we want to use iorv4si3 instead,
for both expanders.

[Bug target/105613] [13 Regression] wrong code with unsigned V1TI comparison since r13-458-gd75d4293dcc029

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105613

--- Comment #4 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Created attachment 52983
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=52983&action=edit
gcc13-pr105613.patch

Untested fix.

[Bug target/105613] [13 Regression] wrong code with unsigned V1TI comparison since r13-458-gd75d4293dcc029

[roger at nextmovesoftware dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105613

--- Comment #5 from Roger Sayle <roger at nextmovesoftware dot com> ---
Firstly sorry for the inconvenience.  Yep, Jakub's analysis is correct we/I
need iorv4si for NE.  My thinko, X==Y is X.h==Y.h && X.l==Y.l, but X!=Y is
X.h!=Y.h || X.l!=Y.l, sorry.  Many thanks to Jakub for assigning this to
himself so quickly.

[Bug target/105613] [13 Regression] wrong code with unsigned V1TI comparison since r13-458-gd75d4293dcc029

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105613

--- Comment #6 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Jakub Jelinek <jakub@gcc.gnu.org>:

https://gcc.gnu.org/g:254cbf72661b932eb039220fccef9a2546ab8f4e

commit r13-552-g254cbf72661b932eb039220fccef9a2546ab8f4e
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Tue May 17 12:10:30 2022 +0200

    i386: Fix up V2DI and V1TI inequality comparisons [PR105613]

    The recent r13-458 change to introduce vec_cmpeqv1tiv1ti and
    add TARGET_SSE2 support to vec_cmpeqv2div2di works nicely for
    equality comparisons, but as the testcase shows doesn't work
    for inequality comparisons.
    For EQ if we perform comparison with twice as many half-sized elemenets,
    the result should be ~0 when both halves are ~0 only (both halves need
    to be equal for the whole to be equal), otherwise 0, so AND is the
    correct operation for it.
    But for NE, the result should be ~0 when either of the halves is ~0
    (if either half is not equal, the whole is not equal) and so the right
    operation for NE is IOR, not AND.

    2022-05-17  Jakub Jelinek  <jakub@redhat.com>

            PR target/105613
            * config/i386/sse.md (vec_cmpeqv2div2di, vec_cmpeqv1tiv1ti): Use
            andv4si3 only for EQ, for NE use iorv4si3 instead.

            * gcc.c-torture/execute/pr105613.c: New test.

[Bug target/105613] [13 Regression] wrong code with unsigned V1TI comparison since r13-458-gd75d4293dcc029

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105613

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|---                         |FIXED

--- Comment #7 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Fixed.

[Bug target/105613] [13 Regression] wrong code with unsigned V1TI comparison since r13-458-gd75d4293dcc029

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105613

Martin Liška <marxin at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |vsevolod.livinskiy at gmail dot co
                   |                            |m

--- Comment #8 from Martin Liška <marxin at gcc dot gnu.org> ---
*** Bug 105621 has been marked as a duplicate of this bug. ***