[polly] [InstCombine] Canonicalise SextADD + GEP (PR #69581)

[via llvm-commits]

LiqinWeng wrote:

> This is also a regression on the simple instruction count heuristic, as we go from add+sext+gep to 2_sext+2_gep. Possibly what you want to do is limit this to the case where the RHS of the add is a constant, so the sext folds away?
yes, I have fixed it, thanks



https://github.com/llvm/llvm-project/pull/69581