[Bug target/88892] [8/9 Regression] Double-to-float conversion uses wrong rounding mode when followed by memcpy

[fw at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88892

--- Comment #4 from Florian Weimer <fw at gcc dot gnu.org> ---
Eh, forget what I wrote.  The pattern *is* used.  r253210 looks definitely to
blame.