[llvm] [InstCombine] Add fold combining adjacent extracted bit fields to restore optimization for separate truncations (PR #219598)

[Mauricio Briceno via llvm-commits]

maubri wrote:

Would it be possible to backport this fix to the LLVM 23.x release branch? This fixes the Clang 14 performance regression reported in #215254. The change is limited to InstCombine, has regression tests, and does not affect API/ABI.

I don't appear to have permission to set the LLVM 23.x Release milestone myself.

Many thanks.

https://github.com/llvm/llvm-project/pull/219598