tra added a comment. Would that be OK to use target_sdk_version to pass *CUDA* SDK version to the CC1 compilations? I have upcoming changes that need to know the version to generate correct glue IR for CUDA. The driver currently figures out detected CUDA version in lib/Driver/ToolChains/Cuda.cpp and I could use -target-sdk-version to pass it on to CC1 instances.
On one hand CUDA is a target SDK and the option appears to be accessible exactly where I need it. On the other hand, my use case is not *exactly* the case `-target-sdk-version` was intended for (i.e. it's not darwin and it has nothing to do with module metadata, though it *may* be eventually useful there even for CUDA). If using the option for CUDA sounds like too much of a hack, I guess I can add a separate `-cuda-sdk-version=`, though it would effectively replicate some of this patch. Opinions? Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D55673/new/ https://reviews.llvm.org/D55673 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
