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

[uwu at icenowy dot me via Gcc-bugs]

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

--- Comment #8 from Icenowy Zheng <uwu at icenowy dot me> ---
Or should I create a pr126049.c test instead of changing existing ones?