lebedev.ri added a comment. So, we can't really teach SCEV about this: D88788 <https://reviews.llvm.org/D88788> (not without the https://bugs.llvm.org/show_bug.cgi?id=47592 at lease) And we can't recover the situation post-inlining in instcombine: D88842 <https://reviews.llvm.org/D88842>.
It really does look like this fold is actively breaking otherwise-good IR, in a way that is not recoverable. And that means, this fold isn't helpful in exposing the passes that are otherwise unaware of these patterns it produces. I'm proceeding with this patch. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D88789/new/ https://reviews.llvm.org/D88789 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
