https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109093
Uroš Bizjak <ubizjak at gmail dot com> changed:
What |Removed |Added
----------------------------------------------------------------------------
Summary|[15/16 regression] csmith: |[15 regression] csmith: a
|a February runtime bug ? |February runtime bug ?
--- Comment #48 from Uroš Bizjak <ubizjak at gmail dot com> ---
(In reply to David Binderman from comment #47)
> > What's the status here? I have difficulties to reproduce anywhere. Is this
> > fixed for sure on trunk at least?
>
> Looks fixed to me on trunk:
This is fixed on trunk. The compiler now tracks all registers that depend on
stack/frame registers and marks current function for realignment if one of the
dependent registers is used in the insn that requires increased alignment.
