qiongsiwu1 added a comment. In D139720#3989651 <https://reviews.llvm.org/D139720#3989651>, @jamieschmeiser wrote:
> I'm fine with handling the return for common as a separate change, if > necessary. Sounds good! In this case I will fix it in a different patch once we decide what exactly to do for the `-mcpu=common` case. > Is the error produced now because it passes back the incorrect option rather > than quietly changing it to something appropriate as it did before? Yes precisely. When a `TargetInfo` is created, we have this <https://github.com/llvm/llvm-project/blob/4c3357ad56a0f5f85dfcadc83f719b09954d43fa/clang/lib/Basic/Targets.cpp#L711> code that checks the validity of the CPU string. Currently, we silently change anything weird to something reasonable and use it to create the `TargetInfo`. This patch removes that filtering and pass on the incorrect CPU string so we can catch the error. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D139720/new/ https://reviews.llvm.org/D139720 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
