rsmith added a comment. In D68407#1744486 <https://reviews.llvm.org/D68407#1744486>, @leonardchan wrote:
> Hi. I think this patch is causing some test failures for us: This is breaking us too, in the same way. The "problem" is that this makes the riscv64 target use the compiler-rt crtbegin / crtend files instead of the libgcc ones when configured with `-DCLANG_DEFAULT_RTLIB=compiler-rt`. That's probably the right behavior, but someone should check, and additional tests need to be updated to match. (It's not obvious how to even test this change in the presence of `CLANG_DEFAULT_RTLIB`...) The following tests fail in the `-DCLANG_DEFAULT_RTLIB=compiler-rt` build configuration: - `test/Driver/riscv32-toolchain.c` - `test/Driver/riscv32-toolchain-extra.c` - `test/Driver/riscv64-toolchain.c` - `test/Driver/riscv64-toolchain-extra.c` Reverted in aeaddf9 <https://reviews.llvm.org/rGaeaddf926a74814aa4f33050cfc12b19f419f1b0> to unbreak things for now. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D68407/new/ https://reviews.llvm.org/D68407 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
