void added a comment. In D69868#1848327 <https://reviews.llvm.org/D69868#1848327>, @MaskRay wrote:
> @rnk Thoughts on passing live-in formation (a bit similar to WinEH catchpad) > to `callbr` at the SelectionDAG stage? To help with the review, here are some of the assumptions I'm making about the `INLINEASM_BR` and the copy block this patch creates: 1. The two blocks (CallBrBB and CopyBB) are "tightly coupled". I.e., you can't split the edge between them (not that you should want to). 2. The CopyBB block will *always* be the fall-through block. This implies that `INLINEASM_BR` is the only terminator in CallBrBB. 3. If one block moves, then both must move. I would be happy to add code in the machine instruction verifier to ensure that these assumptions are met if you think it's necessary. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D69868/new/ https://reviews.llvm.org/D69868 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
