[llvm] [IVDescriptors] Don't require nsz/nnan for (min|max)num. (PR #137003)

[Matt Arsenault via llvm-commits]

arsenm wrote:

> > One thing to note is that GCC uses fmaxnm on AArch64 for fmax, which appears incorrect as per the discussion above: https://godbolt.org/z/W3z1G1z3c
> 
> No, not incorrect. Or depends on your reading of the spec and at what point in time. glibc had the snan-as-qnan behavior until https://sourceware.org/bugzilla/show_bug.cgi?id=20947. The C99 wording was less clear than the C23 wording, which made it more clear that fmin/fmax should follow the IEEE-754 2008 behavior that llvm.maxnum is now supposed to have

Or maybe it was just clarified that signaling nans will be quieted, I'm not sure without rereading all that 

https://github.com/llvm/llvm-project/pull/137003