[Bug ipa/111036] [13 Regression] Code generation error in handling __builtin_constant_p

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111036

--- Comment #12 from GCC Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-13 branch has been updated by Richard Biener
<rguenth@gcc.gnu.org>:

https://gcc.gnu.org/g:453b952a787d28e9a3081f0b0786cd3ed7ffa7b9

commit r13-10145-g453b952a787d28e9a3081f0b0786cd3ed7ffa7b9
Author: Richard Biener <rguenther@suse.de>
Date:   Wed Jan 28 10:55:56 2026 +0100

    ipa/111036 - strip nop conversions around __builtin_constant_p arguments

    The PR is about inconsistent behavior wrt inline predicate analysis
    and later folding of __builtin_constant_p which ultimatively results
    from fold_builtin_constant_p stripping nops off its argument but
    this not being done on GIMPLE.  The following adds a match.pd pattern
    for this.

            PR ipa/111036
            * match.pd (__builtin_constant_p ((T)x)): Strip nop-conversions
            from __builtin_constant_p arguments.

            * gcc.dg/torture/pr111036.c: New testcase.

    (cherry picked from commit 8c0458505a9b1b276fbad81d20849b88eda12d24)