jrtc27 added inline comments.
================ Comment at: test/Driver/riscv64-toolchain.c:71 +// RUN: %clang %s -### -no-canonical-prefixes -fuse-ld=ld \ +// RUN: -target riscv64-linux-unknown-elf \ +// RUN: --gcc-toolchain=%S/Inputs/multilib_riscv_linux_sdk \ ---------------- This (and below) are the only instances of `CPU-linux-unknown-elf` in the whole of clang (and there are only two instances of `CPU-linux-elf`, both in `test/Modules/pch_container.m`), apart from the riscv32 ones they were copied from. Perhaps they should be `riscv64-linux-unknown-gnu` instead? Notably, when parsed as `CPU-company-kernel-OS` this in fact sets *company* to linux and *kernel* to unknown, whereas if you really mean the full 4-tuple should the linux and unknown not be interchanged? ================ Comment at: test/Driver/riscv64-toolchain.c:87 +// RUN: %clang %s -### -no-canonical-prefixes -fuse-ld=ld \ +// RUN: -target riscv64-linux-unknown-elf -march=rv64imafd -mabi=lp64d \ +// RUN: --gcc-toolchain=%S/Inputs/multilib_riscv_linux_sdk \ ---------------- `riscv64-linux-unknown-gnu` as before. Repository: rC Clang https://reviews.llvm.org/D53392 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits