JIT: Consume FMA intrinsic operands in right order

The operands of the FMA intrinsic are permuted in a non-standard way
during LSRA. Codegen already takes this into account, but the handling
was missing when consuming the operands.

Ideally we would permute these during lowering instead to avoid these
hacks.

Fix dotnet#102773

src/coreclr/jit/hwintrinsiccodegenxarch.cpp

@@ -3053,8 +3053,6 @@ void CodeGen::genFMAIntrinsic(GenTreeHWIntrinsic* node, insOpts instOptions)
 
     regNumber targetReg = node->GetRegNum();
 
-    genConsumeMultiOpOperands(node);
-
     regNumber op1NodeReg = op1->GetRegNum();
     regNumber op2NodeReg = op2->GetRegNum();
     regNumber op3NodeReg = op3->GetRegNum();
@@ -3143,6 +3141,21 @@ void CodeGen::genFMAIntrinsic(GenTreeHWIntrinsic* node, insOpts instOptions)
         }
     }
 
+#ifdef DEBUG
+    // Use nums are assigned in LIR order but this node is special and doesn't
+    // actually use operands. Fix up the use nums here to avoid asserts.
+    unsigned useNum1  = op1->gtUseNum;
+    unsigned useNum2  = op2->gtUseNum;
+    unsigned useNum3  = op3->gtUseNum;
+    emitOp1->gtUseNum = useNum1;
+    emitOp2->gtUseNum = useNum2;
+    emitOp3->gtUseNum = useNum3;
+#endif
+
+    genConsumeRegs(emitOp1);
+    genConsumeRegs(emitOp2);
+    genConsumeRegs(emitOp3);
+
     assert(ins != INS_invalid);
     genHWIntrinsic_R_R_R_RM(ins, attr, targetReg, emitOp1->GetRegNum(), emitOp2->GetRegNum(), emitOp3, instOptions);
     genProduceReg(node);