[Bug bootstrap/124406] [16 Regression] LTO profiledbootstrap fails on s390x-linux due to coverage mismatch

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=124406

--- Comment #2 from Richard Biener <rguenth at gcc dot gnu.org> ---
I'll note this particular build was for r16-7922, but the previous r16-7751 at
least failed the same way.

The arch/tune is the one we use for openSUSE, for SLE we're using the same as
Jakub but there we don't do PGO for reproducible build reasons.

I will see to dig out the VM to reproduce this outside of OBS, but I fear that
will take some time.  I also have no idea where to look for this kind of
failure, hopefully Honza can provide some guidance?