diff --git a/src/cmd/compile/internal/ssa/gen/ARM64.rules b/src/cmd/compile/internal/ssa/gen/ARM64.rules
index 27959d01fc2..80e8c7137b0 100644
--- a/src/cmd/compile/internal/ssa/gen/ARM64.rules
+++ b/src/cmd/compile/internal/ssa/gen/ARM64.rules
@@ -1323,6 +1323,7 @@
 (AND x (MVN y)) -> (BIC x y)
 (XOR x (MVN y)) -> (EON x y)
 (OR  x (MVN y)) -> (ORN x y)
+(MVN (XOR x y)) -> (EON x y)
 (CSEL {cc} x (MOVDconst [0]) flag) -> (CSEL0 {cc} x flag)
 (CSEL {cc} (MOVDconst [0]) y flag) -> (CSEL0 {arm64Negate(cc.(Op))} y flag)
 (SUB x (SUB y z)) -> (SUB (ADD <v.Type> x z) y)
diff --git a/src/cmd/compile/internal/ssa/rewriteARM64.go b/src/cmd/compile/internal/ssa/rewriteARM64.go
index 023d9908c2c..842eddbf4a8 100644
--- a/src/cmd/compile/internal/ssa/rewriteARM64.go
+++ b/src/cmd/compile/internal/ssa/rewriteARM64.go
@@ -14593,6 +14593,18 @@ func rewriteValueARM64_OpARM64MULW(v *Value) bool {
 }
 func rewriteValueARM64_OpARM64MVN(v *Value) bool {
 	v_0 := v.Args[0]
+	// match: (MVN (XOR x y))
+	// result: (EON x y)
+	for {
+		if v_0.Op != OpARM64XOR {
+			break
+		}
+		y := v_0.Args[1]
+		x := v_0.Args[0]
+		v.reset(OpARM64EON)
+		v.AddArg2(x, y)
+		return true
+	}
 	// match: (MVN (MOVDconst [c]))
 	// result: (MOVDconst [^c])
 	for {
diff --git a/test/codegen/bits.go b/test/codegen/bits.go
index 0a5428b55a3..398dd84e9e0 100644
--- a/test/codegen/bits.go
+++ b/test/codegen/bits.go
@@ -310,9 +310,18 @@ func op_bic(x, y uint32) uint32 {
 	return x &^ y
 }
 
-func op_eon(x, y uint32) uint32 {
+func op_eon(x, y, z uint32, a []uint32, n, m uint64) uint64 {
+	// arm64:`EON\t`,-`EOR`,-`MVN`
+	a[0] = x ^ (y ^ 0xffffffff)
+
+	// arm64:`EON\t`,-`EOR`,-`MVN`
+	a[1] = ^(y ^ z)
+
 	// arm64:`EON\t`,-`XOR`
-	return x ^ ^y
+	a[2] = x ^ ^z
+
+	// arm64:`EON\t`,-`EOR`,-`MVN`
+	return n ^ (m ^ 0xffffffffffffffff)
 }
 
 func op_orn(x, y uint32) uint32 {