[llvm] [AMDGPU] Implement IR expansion for frem instruction (PR #130988)

[Björn Pettersson via llvm-commits]

================
@@ -720,6 +719,13 @@ FUNCTION_PASS_WITH_PARAMS(
     },
     parseBoundsCheckingOptions,
     "trap;rt;rt-abort;min-rt;min-rt-abort;merge;guard=N")
+    FUNCTION_PASS_WITH_PARAMS(
+    "expand-fp", "ExpandFpPass",
+    [TM = TM](CodeGenOptLevel OL) {
+      return ExpandFpPass(TM, OL);
+    },
+    parseExpandFpOptions, "opt-level")
----------------
bjope wrote:

I think this could say "opt-level=N" to indicate that the syntax is like that, and not that opt-level should be replaced by a value. Similar to "guard=N" for bounds-checking.

https://github.com/llvm/llvm-project/pull/130988