[llvm] [AMDGPU] Make maximum hard clause size a subtarget feature (PR #81287)

[Jay Foad via llvm-commits]

jayfoad wrote:

> Should we go ahead and land this?

I still think `FeatureMaxHardClauseLength32` and `FeatureMaxHardClauseLength63` need brief comments explaining where 32 and 63 come from. You removed the old comment about 63 in this patch.

https://github.com/llvm/llvm-project/pull/81287