diff --git a/src/engine/x86-64/X86_64Interpreter.v3 b/src/engine/x86-64/X86_64Interpreter.v3
index 2a1a0c7f..1855cbf3 100644
--- a/src/engine/x86-64/X86_64Interpreter.v3
+++ b/src/engine/x86-64/X86_64Interpreter.v3
@@ -2410,12 +2410,7 @@ class X86_64InterpreterGen(ic: X86_64InterpreterCode, w: DataWriter) {
 		]) {
 			bindHandler(t.0);
 			load_v128_xmm0_tmp0();
-			var width = byte.view(t.2);
-			var mask = (1 << width) - 1;
-			asm.movq_r_r(r_tmp1, r_tmp0);
-			asm.and_r_i(r_tmp1, mask);
-			asm.movq_s_r(r_xmm1, r_tmp1);
-			t.1(r_xmm0, r_xmm1);
+			masm.emit_v128_shift(r_xmm0, r_tmp0, byte.view(t.2), r_tmp1, r_xmm1, t.1);
 			asm.movdqu_m_s(vsph[-2].value, r_xmm0);
 			decrementVsp();
 			endHandler();
diff --git a/src/engine/x86-64/X86_64MacroAssembler.v3 b/src/engine/x86-64/X86_64MacroAssembler.v3
index ea03ec4f..647ab3c7 100644
--- a/src/engine/x86-64/X86_64MacroAssembler.v3
+++ b/src/engine/x86-64/X86_64MacroAssembler.v3
@@ -981,6 +981,14 @@ class X86_64MacroAssembler extends MacroAssembler {
 		asm.psrlw_s_s(dst, tmp3);
 		asm.packuswb_s_s(dst, tmp2);
 	}
+	def emit_v128_shift<T>(dst: X86_64Xmmr, shift: X86_64Gpr, width: byte, gtmp: X86_64Gpr, xtmp: X86_64Xmmr,
+			asm_pshfit_s_s: (X86_64Xmmr, X86_64Xmmr) -> T) {
+		var mask = (1 << width) - 1;
+		asm.movq_r_r(gtmp, shift);
+		asm.and_r_i(gtmp, mask);
+		asm.movq_s_r(xtmp, gtmp);
+		asm_pshfit_s_s(dst, xtmp);
+	}
 	def emit_i64x2_abs(dst: X86_64Xmmr, scratch: X86_64Xmmr) {
 		asm.movshdup_s_s(scratch, dst);
 		asm.psrad_i(scratch, 31);