[compiler-rt] Bfi precision (PR #66285)

[Mingming Liu via llvm-commits]

minglotus-6 wrote:

> Seems this got introduced in https://reviews.llvm.org/D34312 with the rough idea that we shouldn't inline into parts of the code that _builtin_expect(...) deems unlikely

@xur-llvm @david-xl This just reminds me of the observed regressions when we did an experiment to override branch probability from `absl_predict_{true,false}` with (manually collected IIRC) sampled PGO profiles.

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