[PATCH] D72184: [BPF] support atomic instructions

[Alexei Starovoitov via Phabricator via cfe-commits]

ast added inline comments.


================
Comment at: llvm/lib/Target/BPF/BPFMIChecking.cpp:199
+      unsigned newOpcode;
+      switch(MI.getOpcode()) {
+      case BPF::XFADDW32: newOpcode = BPF::XADDW32; break;
----------------
With this logic in place Andrii has a point. There is no need for -mcpu=v4 flag.
The builtins will produce new insns and it will be up to kernel to accept them or not.


================
Comment at: llvm/lib/Target/BPF/BPFMIChecking.cpp:207
+      case BPF::XFORW32: newOpcode = BPF::XORW32; break;
+      default: newOpcode = BPF::XORD; break;
+      }
----------------
The default is error prone. Why not to check for XFORD explicitly and default to fatal_error?


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D72184/new/

https://reviews.llvm.org/D72184