craig.topper added inline comments.
================ Comment at: llvm/test/CodeGen/RISCV/rvv/fixed-vectors-bitcast.ll:158 ; CHECK-NEXT: vmv.x.s a0, v8 +; CHECK-NEXT: lui a1, 1048560 +; CHECK-NEXT: or a0, a0, a1 ---------------- frasercrmck wrote: > What's going on here, do you know? I believe this is coming from the nan-boxing for f16 in RISCVTargetLowering::splitValueIntoRegisterParts. The addition of +f must have changed PartVT from i32/i64 to f32. Even though we're using i32 for the return due to ABI. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D112408/new/ https://reviews.llvm.org/D112408 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits