[Bug testsuite/110490] New: tree-ssa/clz-* and tree-ssa/ctz-* fail on s390x with arch14

[Bug testsuite/110490] New: tree-ssa/clz-* and tree-ssa/ctz-* fail on s390x with arch14

[mpolacek at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110490

            Bug ID: 110490
           Summary: tree-ssa/clz-* and tree-ssa/ctz-* fail on s390x with
                    arch14
           Product: gcc
           Version: 14.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: testsuite
          Assignee: unassigned at gcc dot gnu.org
          Reporter: mpolacek at gcc dot gnu.org
  Target Milestone: ---

(Not sure what component is most appropriate for this.)

The following FAILs can be seen in
<https://gcc.gnu.org/pipermail/gcc-testresults/2023-June/786039.html>
(s390x-ibm-linux-gnu arch14)
but not in
<https://gcc.gnu.org/pipermail/gcc-testresults/2023-June/786004.html>
(s390x-ibm-linux-gnu default)

It would be great to fix them.

FAIL: gcc.dg/tree-ssa/clz-char.c scan-tree-dump-times optimized
"__builtin_clz|\\\\.CLZ" 1
FAIL: gcc.dg/tree-ssa/clz-complement-char.c scan-tree-dump-times optimized
"__builtin_clz|\\\\.CLZ" 1
FAIL: gcc.dg/tree-ssa/clz-complement-int.c scan-tree-dump-times optimized
"__builtin_clz|\\\\.CLZ" 1
FAIL: gcc.dg/tree-ssa/clz-complement-long.c scan-tree-dump-times optimized
"__builtin_clz|\\\\.CLZ" 1
FAIL: gcc.dg/tree-ssa/clz-int.c scan-tree-dump-times optimized
"__builtin_clz|\\\\.CLZ" 1
FAIL: gcc.dg/tree-ssa/clz-long.c scan-tree-dump-times optimized
"__builtin_clz|\\\\.CLZ" 1
FAIL: gcc.dg/tree-ssa/ctz-char.c scan-tree-dump-times optimized
"__builtin_ctz|\\\\.CTZ" 1
FAIL: gcc.dg/tree-ssa/ctz-complement-char.c scan-tree-dump-times optimized
"__builtin_ctz|\\\\.CTZ" 1
FAIL: gcc.dg/tree-ssa/ctz-complement-int.c scan-tree-dump-times optimized
"__builtin_ctz|\\\\.CTZ" 1
FAIL: gcc.dg/tree-ssa/ctz-complement-long-long.c scan-tree-dump-times optimized
"__builtin_ctz|\\\\.CTZ" 1
FAIL: gcc.dg/tree-ssa/ctz-complement-long.c scan-tree-dump-times optimized
"__builtin_ctz|\\\\.CTZ" 1
FAIL: gcc.dg/tree-ssa/ctz-int.c scan-tree-dump-times optimized
"__builtin_ctz|\\\\.CTZ" 1
FAIL: gcc.dg/tree-ssa/ctz-long-long.c scan-tree-dump-times optimized
"__builtin_ctz|\\\\.CTZ" 1
FAIL: gcc.dg/tree-ssa/ctz-long.c scan-tree-dump-times optimized
"__builtin_ctz|\\\\.CTZ" 1

[Bug testsuite/110490] tree-ssa/clz-* and tree-ssa/ctz-* fail on s390x with arch14

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110490

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

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

--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
I suspect this is a testsuite issue:
>/* { dg-require-effective-target clzl } */

basically the check_effective_target_clz* (ctz*) test is passing for some
reason when it should not be ...

[Bug tree-optimization/110490] tree-ssa/clz-* and tree-ssa/ctz-* fail on s390x with arch14

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110490

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|testsuite                   |tree-optimization
     Ever confirmed|0                           |1
   Last reconfirmed|                            |2023-06-29
           Keywords|                            |missed-optimization
             Status|UNCONFIRMED                 |NEW

--- Comment #2 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
So the problem is a missing optimization.
In this case, arch14 (which enables TARGET_EXTIMM) only defines a clzdi2 and
not a clzsi2 so the code in build_cltz_expr does not realize that it could
still support clzsi2 ...

[Bug tree-optimization/110490] tree-ssa/clz-* and tree-ssa/ctz-* fail on s390x with arch14

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110490

--- Comment #3 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
I was wrong about where the problem was but not wrong about the missed
optimization.

Anyways the issue is in expression_expensive_p where the issue of clzdi2 and
not having clzsi2 ...

expression_expensive_p should check if you can do a widening of the mode ...

[Bug tree-optimization/110490] tree-ssa/clz-* and tree-ssa/ctz-* fail on s390x with arch14 (expression_expensive_p not handling if there is wider clz/ctz/popcount optab)

[stefansf at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110490

Stefan Schulze Frielinghaus <stefansf at gcc dot gnu.org> changed:

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

--- Comment #4 from Stefan Schulze Frielinghaus <stefansf at gcc dot gnu.org> ---
https://gcc.gnu.org/pipermail/gcc-patches/2024-April/650001.html

[Bug tree-optimization/110490] tree-ssa/clz-* and tree-ssa/ctz-* fail on s390x with arch14 (expression_expensive_p not handling if there is wider clz/ctz/popcount optab)

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110490

--- Comment #5 from GCC Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Stefan Schulze Frielinghaus
<stefansf@gcc.gnu.org>:

https://gcc.gnu.org/g:e1f56c67a82172730c377a96a46e8d75445e6a48

commit r15-274-ge1f56c67a82172730c377a96a46e8d75445e6a48
Author: Stefan Schulze Frielinghaus <stefansf@linux.ibm.com>
Date:   Tue May 7 14:12:55 2024 +0200

    tree-optimization/110490 - bitcount for narrow modes

    Bitcount operations popcount, clz, and ctz are emulated for narrow modes
    in case an operation is only supported for wider modes.  Beside that ctz
    may be emulated via clz in expand_ctz.  Reflect this in
    expression_expensive_p.

    I considered the emulation of ctz via clz as not expensive since this
    basically reduces to ctz (x) = c - (clz (x & ~x)) where c is the mode
    precision minus 1 which should be faster than a loop.

    gcc/ChangeLog:

            PR tree-optimization/110490
            * tree-scalar-evolution.cc (expression_expensive_p): Also
            consider mode widening for popcount, clz, and ctz.

[Bug tree-optimization/110490] tree-ssa/clz-* and tree-ssa/ctz-* fail on s390x with arch14 (expression_expensive_p not handling if there is wider clz/ctz/popcount optab)

[stefansf at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110490

Stefan Schulze Frielinghaus <stefansf at gcc dot gnu.org> changed:

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

--- Comment #6 from Stefan Schulze Frielinghaus <stefansf at gcc dot gnu.org> ---
Fixed.