================
@@ -3199,4 +3199,22 @@ def CIR_ExpectOp : CIR_Op<"expect", [
}];
}
+//===----------------------------------------------------------------------===//
+// Floating Point Ops
+//===----------------------------------------------------------------------===//
+
+class CIR_UnaryFPToFPBuiltinOp<string mnemonic, string llvmOpName>
+ : CIR_Op<mnemonic, [Pure, SameOperandsAndResultType]>
+{
+ let arguments = (ins CIR_AnyFloatOrVecOfFloatType:$src);
+ let results = (outs CIR_AnyFloatOrVecOfFloatType:$result);
+ let summary = "libc builtin equivalent ignoring "
+ "floating point exceptions and errno";
----------------
andykaylor wrote:
It doesn't really make sense to say that this is libc equivalent, when one of
its primary characteristics is a difference from the libc function. Do we need
a summary here at all? I'd prefer to see individual summaries on the operations
that derive from this, even knowing that they will be very similar.
https://github.com/llvm/llvm-project/pull/151750
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
