[Bug target/96402] [10/11 Regression] Wrong code with -moutline-atomics

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96402

--- Comment #2 from Martin Liška <marxin at gcc dot gnu.org> ---
(In reply to ktkachov from comment #1)
> Thanks, can you share the hardware you ran it on? Depending on whether LSE
> is present different code paths are taken at runtime...

Well, I tested that on an experimental hardware we were given by AMD.

cpuinfo says something like:
processor       : 6
BogoMIPS        : 500.00
Features        : fp asimd evtstrm aes pmull sha1 sha2 crc32 cpuid
CPU implementer : 0x41
CPU architecture: 8
CPU variant     : 0x1
CPU part        : 0xd07
CPU revision    : 2