zhouguangyuan0718 wrote: > I'm not really opposed in general, I guess? It's just a matter of whether > it's actually practical for the specific registers in question. > > ESI in particular has a problem this patch doesn't address: we use it as the > base register, to refer to the stack in certain functions. See > X86RegisterInfo::X86RegisterInfo .
Thanks for review, Now I removed reserving ESI support, only EDI are supported to reserve. And when reserving EDI, rep stos/movs are disabled. Also, I fixed the logic in `validateGlobalRegisterVariable` and added a testcase for it. https://github.com/llvm/llvm-project/pull/186123 _______________________________________________ cfe-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
