diff --git a/sm3/sm3blocks_ppc64x.s b/sm3/sm3blocks_ppc64x.s
index e4b97fa..ba03820 100644
--- a/sm3/sm3blocks_ppc64x.s
+++ b/sm3/sm3blocks_ppc64x.s
@@ -120,7 +120,7 @@ TEXT ·copyResultsBy4(SB),NOSPLIT,$0
 	MOVD	dst+8(FP), R4
 
 #ifdef NEEDS_PERMW	
-	MOVD	$·mask+0x80(SB), R5
+	MOVD	$·mask+0x50(SB), R5
 	LVX	(R5), ESPERMW
 #endif
 
@@ -129,30 +129,30 @@ TEXT ·copyResultsBy4(SB),NOSPLIT,$0
 
 	MOVD	$16, R5
 	LXVD2X 	(R5)(R3), V0
-	PPC64X_STXVD2X(V0, R0, R4)
+	PPC64X_STXVD2X(V0, R5, R4)
 	
 	ADD 	$16, R5
 	LXVD2X 	(R5)(R3), V0
-	PPC64X_STXVD2X(V0, R0, R4)
+	PPC64X_STXVD2X(V0, R5, R4)
 
 	ADD 	$16, R5
 	LXVD2X 	(R5)(R3), V0
-	PPC64X_STXVD2X(V0, R0, R4)
+	PPC64X_STXVD2X(V0, R5, R4)
 
 	ADD 	$16, R5
 	LXVD2X 	(R5)(R3), V0
-	PPC64X_STXVD2X(V0, R0, R4)
+	PPC64X_STXVD2X(V0, R5, R4)
 
 	ADD 	$16, R5
 	LXVD2X 	(R5)(R3), V0
-	PPC64X_STXVD2X(V0, R0, R4)
+	PPC64X_STXVD2X(V0, R5, R4)
 
 	ADD 	$16, R5
 	LXVD2X 	(R5)(R3), V0
-	PPC64X_STXVD2X(V0, R0, R4)
+	PPC64X_STXVD2X(V0, R5, R4)
 
 	ADD 	$16, R5
 	LXVD2X 	(R5)(R3), V0
-	PPC64X_STXVD2X(V0, R0, R4)
+	PPC64X_STXVD2X(V0, R5, R4)
 
 	RET