https://gcc.gnu.org/bugzilla/show_bug.cgi?id=92462
Aleksei Voitylov <aleksei.voity...@bell-sw.com> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|RESOLVED |UNCONFIRMED
Resolution|FIXED |---
--- Comment #13 from Aleksei Voitylov <aleksei.voity...@bell-sw.com> ---
I'm reopening it since the new testcase that doesn't have anything to do with
the actual atomic operations still fails in the same manner.
I changed the name "Atomic" so that to make make it less distractive but even
the source itself is quite self-evident now.
It could be the case that the optimizer wrongly considers uint8_t type (which
is unsigned char) as not aliasing uint32_t in this particular scenario.
