[Bug other/105183] New: New test case gcc.dg/ipa/remref-7.c fails with -m32

[Bug other/105183] New: New test case gcc.dg/ipa/remref-7.c fails with -m32

[seurer at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105183

            Bug ID: 105183
           Summary: New test case gcc.dg/ipa/remref-7.c fails with -m32
           Product: gcc
           Version: 12.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: other
          Assignee: unassigned at gcc dot gnu.org
          Reporter: seurer at gcc dot gnu.org
  Target Milestone: ---

g:f6d65e803623c7ba6c8eb92ce5975fc1b90cd91e, r12-7936-gf6d65e803623c7

It works fine for 64 bits.

make  -k check-gcc RUNTESTFLAGS="--target_board=unix'{-m32}'
ipa.exp=gcc.dg/ipa/remref-7.c"
FAIL: gcc.dg/ipa/remref-7.c scan-ipa-dump inline "Removed a reference"
# of expected passes            2
# of unexpected failures        1

f6d65e803623c7ba6c8eb92ce5975fc1b90cd91e is the first bad commit
commit f6d65e803623c7ba6c8eb92ce5975fc1b90cd91e
Author: Martin Jambor <mjambor@suse.cz>
Date:   Thu Mar 31 16:50:24 2022 +0200

    ipa: Create LOAD references when necessary during inlining (PR 103171)

[Bug testsuite/105183] New test case gcc.dg/ipa/remref-7.c fails with -m32

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105183

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Target|powerpc64-linux-gnu         |powerpc64-linux-gnu,
                   |                            |i?86-*-*

--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> ---
Also on i?86-*-* with -mno-sse

[Bug testsuite/105183] New test case gcc.dg/ipa/remref-7.c fails with -m32

[hp at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105183

Hans-Peter Nilsson <hp at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Target|powerpc64-linux-gnu,        |powerpc64-linux-gnu,
                   |i?86-*-*                    |i?86-*-*, m68k-*, cris-*
                 CC|                            |hp at gcc dot gnu.org

--- Comment #2 from Hans-Peter Nilsson <hp at gcc dot gnu.org> ---
...and cris-elf and m68k according to
https://gcc.gnu.org/pipermail/gcc-testresults/2022-April/758712.html so I guess
all or most ILP32.

[Bug testsuite/105183] New test case gcc.dg/ipa/remref-7.c fails with -m32

[jamborm at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105183

Martin Jambor <jamborm at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
     Ever confirmed|0                           |1
   Last reconfirmed|                            |2022-04-08
             Status|UNCONFIRMED                 |ASSIGNED
           Assignee|unassigned at gcc dot gnu.org      |jamborm at gcc dot gnu.org

--- Comment #3 from Martin Jambor <jamborm at gcc dot gnu.org> ---
I'm having a look.

[Bug testsuite/105183] New test case gcc.dg/ipa/remref-7.c fails with -m32

[jamborm at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105183

--- Comment #4 from Martin Jambor <jamborm at gcc dot gnu.org> ---
I expected that the only call of bar in the testcase to be always
inlined everywhere but apparently it is not at least on i?86-*-* with
-mno-sse (and I expect the problem to be the same on the other
reported platforms).

Using --param max-inline-insns-auto=100 "fixes" it (again at least on
i686 without SSE but I expect elsewhere too).

[Bug testsuite/105183] New test case gcc.dg/ipa/remref-7.c fails with -m32

[jamborm at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105183

--- Comment #5 from Martin Jambor <jamborm at gcc dot gnu.org> ---
I proposed to increase the parameter specification in the test in:
https://gcc.gnu.org/pipermail/gcc-patches/2022-April/592994.html

[Bug testsuite/105183] New test case gcc.dg/ipa/remref-7.c fails with -m32

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105183

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

https://gcc.gnu.org/g:dd61ee6fde5fcf1d503b679bb9b5e0a0ba7a515d

commit r12-8139-gdd61ee6fde5fcf1d503b679bb9b5e0a0ba7a515d
Author: Martin Jambor <mjambor@suse.cz>
Date:   Wed Apr 13 15:08:59 2022 +0200

    testsuite: Increase auto-inlining param in gcc.dg/ipa/remref-7.c (PR
105183)

    A scan dump of testsuite gcc.dg/ipa/remref-7.c fails on a number of
    platforms.  I investigated only i?86-*-* with -mno-sse but assume the
    issue is the same on all of the affected platform.

    Because function bar is not inlined there even though it is only
    called once, the process that is being tested is simply not triggered.
    This can be "fixed" by increasing parameter max-inline-insns-auto to
    something high, I randomly picked 100.

    I have only manually tested the change but hopefully that is enough.

    gcc/testsuite/ChangeLog:

    2022-04-08  Martin Jambor  <mjambor@suse.cz>

            PR testsuite/105183
            * gcc.dg/ipa/remref-7.c: Add --param max-inline-insns-auto=100 to
options.

[Bug testsuite/105183] New test case gcc.dg/ipa/remref-7.c fails with -m32

[jamborm at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105183

Martin Jambor <jamborm at gcc dot gnu.org> changed:

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

--- Comment #7 from Martin Jambor <jamborm at gcc dot gnu.org> ---
Please re-open if the fix turns out insufficient for any of the affected
platforms.  Thanks for reporting.