[llvm] Bfi precision (PR #66285)

[Matthias Braun via llvm-commits]

MatzeB wrote:

Thanks for the comment.

I still think choosing more aggressive factors in case of small spreads should not hurt anything and we can easily do so. Even if that just makes people writing test cases happier :)

As for overflows I am still trying to get a feeling for how much we need to go and fix those (as in https://github.com/llvm/llvm-project/pull/66280 ) or whether we have to accept them as a factor of (engineering-)life and rather leave enough room in the values...

https://github.com/llvm/llvm-project/pull/66285