[PATCH] D128899: [LoongArch] Add codegen support for fpround, fpextend and converting between signed integer and floating-point

Sun Jul 10 21:25:31 PDT 2022

xen0n accepted this revision.
xen0n added a comment.
This revision is now accepted and ready to land.

I don't know if the LA64-specific DAG nodes could actually be removed for further simplicity, but the rest seems good to me. Someone could enlighten me on this.

================
Comment at: llvm/lib/Target/LoongArch/LoongArchFloat32InstrInfo.td:25
+    : SDNode<"LoongArchISD::MOVGR2FR_W_LA64", SDT_LoongArchMOVGR2FR_W_LA64>;
+def loongarch_movfr2gr_s_anyext_la64
+    : SDNode<"LoongArchISD::MOVFR2GR_S_LA64", SDT_LoongArchMOVFR2GR_S_LA64>;
----------------
What does `anyext` mean? Either this isn't meaningful and should be removed, or an explanatory comment should be around.

================
Comment at: llvm/lib/Target/LoongArch/LoongArchFloat64InstrInfo.td:201
+
+/// Float conversion operations
+
----------------
nit: "FP" or "Floating-point" for consistency.

Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D128899/new/

https://reviews.llvm.org/D128899