[Bug target/126049] RISC-V auto-vectorization misoptimizes fmax()/fmin() regarding NaN

[uwu at icenowy dot me]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=126049

--- Comment #7 from Icenowy Zheng <uwu at icenowy dot me> ---
Oh I found that reduc_run-10 test exists, I will add check for all NaN case
there.