diff --git a/kernel/loongarch64/cscal_lasx.S b/kernel/loongarch64/cscal_lasx.S
index e32071def7..d44e978a64 100644
--- a/kernel/loongarch64/cscal_lasx.S
+++ b/kernel/loongarch64/cscal_lasx.S
@@ -94,7 +94,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
     CMPEQ  $fcc1, ALPHAI, a1
     bge $r0, I, .L19
 ///////   INCX == 1 && N >= 4   ////////
-    bnez  DUMMY2, .L17 // if DUMMPY2 == 1, called from c/zscal.
+    bnez  DUMMY2, .L17 // if DUMMY2 == 1, called from c/zscal.
 
     bceqz $fcc0, .L17
 
@@ -146,6 +146,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
     addi.d  I, I, -1
     blt $r0, I, .L17
     b .L19
+
     .align 3
 
 ///////  INCX == 1 && N < 8   ///////
@@ -156,7 +157,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
     andi   I,   N,   7
 #endif
     beqz   I,   .L999
-    bnez  DUMMY2, .L998 // if DUMMPY2 == 1, called from c/zscal.
+    bnez  DUMMY2, .L998 // if DUMMY2 == 1, called from c/zscal.
 
     bceqz $fcc0, .L998
 
@@ -171,7 +172,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
     CMPEQ  $fcc1, ALPHAI, a1
     move XX, X
     bge $r0, I, .L29
-    bnez  DUMMY2, .L25 // if DUMMPY2 == 1, called from c/zscal.
+    bnez  DUMMY2, .L25 // if DUMMY2 == 1, called from c/zscal.
     bceqz $fcc0, .L25
 
     bceqz $fcc1, .L25
@@ -341,7 +342,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
     andi   I,   N,   7
 #endif
     beqz   I,   .L999
-    bnez  DUMMY2, .L998 // if DUMMPY2 == 1, called from c/zscal.
+    bnez  DUMMY2, .L998 // if DUMMY2 == 1, called from c/zscal.
 
     bceqz $fcc0, .L998
 
diff --git a/kernel/loongarch64/cscal_lsx.S b/kernel/loongarch64/cscal_lsx.S
index 241d3d16e4..c235a206a6 100644
--- a/kernel/loongarch64/cscal_lsx.S
+++ b/kernel/loongarch64/cscal_lsx.S
@@ -33,6 +33,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 #define ALPHAI $f1
 #define X      $r7
 #define INCX   $r8
+#define DUMMY2 $r9
 
 #define I      $r12
 #define TEMP   $r13
@@ -65,6 +66,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 
     bge $r0, N, .L999
     bge $r0, INCX, .L999
+    ld.d DUMMY2, $sp, 0
     li.d TEMP, 1
     movgr2fr.d a1, $r0
     FFINT    a1, a1
@@ -84,24 +86,20 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
     srai.d I, N, 2
     bne INCX, TEMP, .L22
 
+///////    INCX == 1   ////////
 .L11:
-    bge $r0, I, .L997
     CMPEQ  $fcc0, ALPHAR, a1
     CMPEQ  $fcc1, ALPHAI, a1
-    bceqz $fcc0, .L13
-    b .L14
-    .align 3
+    bge $r0, I, .L19
 
-.L13:
-    bceqz $fcc1, .L114 //alpha_r != 0.0 && alpha_i != 0.0
-    b .L113 //alpha_r != 0.0 && alpha_i == 0.0
+///////   INCX == 1 && N >= 4   ////////
+    bnez  DUMMY2, .L17 // if DUMMPY2 == 1, called from c/zscal.
 
-.L14:
-    bceqz $fcc1, .L114  //alpha_r == 0.0 && alpha_i != 0.0
-    b .L111 //alpha_r == 0.0 && alpha_i == 0.0
-    .align 3
+    bceqz $fcc0, .L17
 
-.L111:  //alpha_r == 0.0 && alpha_i == 0.0
+    bceqz $fcc1, .L17
+
+.L15:  //alpha_r == 0.0 && alpha_i == 0.0
     vst VXZ, X, 0 * SIZE
 #ifdef DOUBLE
     vst VXZ, X, 2 * SIZE
@@ -112,50 +110,11 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 #endif
     addi.d X, X, 8 * SIZE
     addi.d  I, I, -1
-    blt $r0, I, .L111
-    b .L997
-    .align 3
-
-.L113: //alpha_r != 0.0 && alpha_i == 0.0
-    vld VX0, X, 0 * SIZE
-#ifdef DOUBLE
-    vld VX1, X, 2 * SIZE
-    vpickev.d x1, VX1, VX0
-    vpickod.d x2, VX1, VX0
-    vfmul.d x3, VXAR, x1
-    vfmul.d x4, VXAR, x2
-    vilvl.d VX2, x4 ,x3
-    vilvh.d VX3, x4, x3
-    vst VX2, X, 0 * SIZE
-    vst VX3, X, 2 * SIZE
-    vld VX0, X, 4 * SIZE
-    vld VX1, X, 6 * SIZE
-    vpickev.d x1, VX1, VX0
-    vpickod.d x2, VX1, VX0
-    vfmul.d x3, VXAR, x1
-    vfmul.d x4, VXAR, x2
-    vilvl.d VX2, x4 ,x3
-    vilvh.d VX3, x4, x3
-    vst VX2, X, 4 * SIZE
-    vst VX3, X, 6 * SIZE
-#else
-    vld VX1, X, 4 * SIZE
-    vpickev.w x1, VX1, VX0
-    vpickod.w x2, VX1, VX0
-    vfmul.s x3, VXAR, x1
-    vfmul.s x4, VXAR, x2
-    vilvl.w VX2, x4 ,x3
-    vilvh.w VX3, x4, x3
-    vst VX2, X, 0 * SIZE
-    vst VX3, X, 4 * SIZE
-#endif
-    addi.d X, X, 8 * SIZE
-    addi.d  I, I, -1
-    blt $r0, I, .L113
-    b .L997
+    blt $r0, I, .L15
+    b .L19
     .align 3
 
-.L114:  //alpha_r != 0.0 && alpha_i != 0.0
+.L17:
     vld VX0, X, 0 * SIZE
 #ifdef DOUBLE
     vld VX1, X, 2 * SIZE
@@ -196,29 +155,35 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 #endif
     addi.d X, X, 8 * SIZE
     addi.d  I, I, -1
-    blt $r0, I, .L114
-    b .L997
+    blt $r0, I, .L17
+    b .L19
     .align 3
 
+///////  INCX == 1 && N < 8   ///////
+.L19:
+    andi   I,   N,   3
+    beqz   I,   .L999
+    bnez  DUMMY2, .L998 // if DUMMPY2 == 1, called from c/zscal.
+
+    bceqz $fcc0, .L998
+
+    bceqz $fcc1, .L998
+
+    b .L995            // alpha_r == 0.0 && alpha_i == 0.0
+
+///////    INCX != 1   ////////
 .L22:
-    bge $r0, I, .L997
-    move XX, X
     CMPEQ  $fcc0, ALPHAR, a1
     CMPEQ  $fcc1, ALPHAI, a1
-    bceqz $fcc0, .L23
-    b .L24
-    .align 3
+    move XX, X
+    bge $r0, I, .L29
+    bnez  DUMMY2, .L25 // if DUMMPY2 == 1, called from c/zscal.
 
-.L23:
-    bceqz $fcc1, .L224 //alpha_r != 0.0 && alpha_i != 0.0
-    b .L223 //alpha_r != 0.0 && alpha_i == 0.0
+    bceqz $fcc0, .L25
 
-.L24:
-    bceqz $fcc1, .L224  //alpha_r == 0.0 && alpha_i != 0.0
-    b .L221 //alpha_r == 0.0 && alpha_i == 0.0
-    .align 3
+    bceqz $fcc1, .L25
 
-.L221:  //alpha_r == 0.0 && alpha_i == 0.0
+.L27:  //alpha_r == 0.0 && alpha_i == 0.0
 #ifdef DOUBLE
     vstelm.d VXZ, X, 0, 0
     vstelm.d VXZ, X, 1 * SIZE, 0
@@ -246,92 +211,11 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 #endif
     add.d X, X, INCX
     addi.d  I, I, -1
-    blt $r0, I, .L221
-    b .L997
+    blt $r0, I, .L27
+    b .L29
     .align 3
 
-.L223: //alpha_r != 0.0 && alpha_i == 0.0
-#ifdef DOUBLE
-    ld.d t1, X, 0 * SIZE
-    ld.d t2, X, 1 * SIZE
-    add.d X, X, INCX
-    ld.d t3, X, 0 * SIZE
-    ld.d t4, X, 1 * SIZE
-    add.d X, X, INCX
-    vinsgr2vr.d x1, t1, 0
-    vinsgr2vr.d x2, t2, 0
-    vinsgr2vr.d x1, t3, 1
-    vinsgr2vr.d x2, t4, 1
-    vfmul.d x3, VXAR, x1
-    vfmul.d x4, VXAR, x2
-    vstelm.d x3, XX, 0 * SIZE, 0
-    vstelm.d x4, XX, 1 * SIZE, 0
-    add.d XX, XX, INCX
-    vstelm.d x3, XX, 0 * SIZE, 1
-    vstelm.d x4, XX, 1 * SIZE, 1
-    add.d XX, XX, INCX
-
-    ld.d t1, X, 0 * SIZE
-    ld.d t2, X, 1 * SIZE
-    add.d X, X, INCX
-    ld.d t3, X, 0 * SIZE
-    ld.d t4, X, 1 * SIZE
-    vinsgr2vr.d x1, t1, 0
-    vinsgr2vr.d x2, t2, 0
-    vinsgr2vr.d x1, t3, 1
-    vinsgr2vr.d x2, t4, 1
-    add.d X, X, INCX
-    vfmul.d x3, VXAR, x1
-    vfmul.d x4, VXAR, x2
-    addi.d  I, I, -1
-    vstelm.d x3, XX, 0 * SIZE, 0
-    vstelm.d x4, XX, 1 * SIZE, 0
-    add.d XX, XX, INCX
-    vstelm.d x3, XX, 0 * SIZE, 1
-    vstelm.d x4, XX, 1 * SIZE, 1
-#else
-    ld.w t1, X, 0 * SIZE
-    ld.w t2, X, 1 * SIZE
-    add.d X, X, INCX
-    ld.w t3, X, 0 * SIZE
-    ld.w t4, X, 1 * SIZE
-    add.d X, X, INCX
-    vinsgr2vr.w x1, t1, 0
-    vinsgr2vr.w x2, t2, 0
-    vinsgr2vr.w x1, t3, 1
-    vinsgr2vr.w x2, t4, 1
-    ld.w t1, X, 0 * SIZE
-    ld.w t2, X, 1 * SIZE
-    add.d X, X, INCX
-    ld.w t3, X, 0 * SIZE
-    ld.w t4, X, 1 * SIZE
-    vinsgr2vr.w x1, t1, 2
-    vinsgr2vr.w x2, t2, 2
-    vinsgr2vr.w x1, t3, 3
-    vinsgr2vr.w x2, t4, 3
-    add.d X, X, INCX
-
-    vfmul.s x3, VXAR, x1
-    vfmul.s x4, VXAR, x2
-    addi.d  I, I, -1
-    vstelm.w x3, XX, 0 * SIZE, 0
-    vstelm.w x4, XX, 1 * SIZE, 0
-    add.d XX, XX, INCX
-    vstelm.w x3, XX, 0 * SIZE, 1
-    vstelm.w x4, XX, 1 * SIZE, 1
-    add.d XX, XX, INCX
-    vstelm.w x3, XX, 0 * SIZE, 2
-    vstelm.w x4, XX, 1 * SIZE, 2
-    add.d XX, XX, INCX
-    vstelm.w x3, XX, 0 * SIZE, 3
-    vstelm.w x4, XX, 1 * SIZE, 3
-#endif
-    add.d XX, XX, INCX
-    blt $r0, I, .L223
-    b .L997
-    .align 3
-
-.L224:  //alpha_r != 0.0 && alpha_i != 0.0
+.L25:
 #ifdef DOUBLE
     ld.d t1, X, 0 * SIZE
     ld.d t2, X, 1 * SIZE
@@ -414,15 +298,29 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
     vstelm.w x4, XX, 1 * SIZE, 3
 #endif
     add.d XX, XX, INCX
-    blt $r0, I, .L224
-    b .L997
+    blt $r0, I, .L25
+    b .L29
     .align 3
 
-.L997:
-    andi I, N, 3
-    bge $r0, I, .L999
-    .align 3
+///////  INCX != 1 && N < 8   ///////
+.L29:
+    andi   I,   N,   3
+    beqz   I,   .L999
+    bnez  DUMMY2, .L998 // if DUMMPY2 == 1, called from c/zscal.
+
+    bceqz $fcc0, .L998
+
+    bceqz $fcc1, .L998
 
+    b .L995            // alpha_r == 0.0 && alpha_i == 0.0
+
+.L995: // alpha_r == 0.0 && alpha_i == 0.0
+    ST   a1, X, 0 * SIZE
+    ST   a1, X, 1 * SIZE
+    addi.d I, I, -1
+    add.d  X, X, INCX
+    blt $r0, I, .L995
+    b .L999
 .L998:
     LD   a1, X, 0 * SIZE
     LD   a2, X, 1 * SIZE
@@ -435,7 +333,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
     ST   s2, X, 1 * SIZE
     add.d X, X, INCX
     blt $r0, I, .L998
-    .align 3
+    b .L999
 
 .L999:
     move $r4, $r12
diff --git a/kernel/loongarch64/zscal.S b/kernel/loongarch64/zscal.S
index a12e527a56..f6213b1591 100644
--- a/kernel/loongarch64/zscal.S
+++ b/kernel/loongarch64/zscal.S
@@ -53,6 +53,8 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
    PROLOGUE
 
    li.d  TEMP, 2 * SIZE
+   ld.d XX, $sp, 0 // Load dummy2
+   slli.d XX, XX, ZBASE_SHIFT
    MTC  a1, $r0
    slli.d INCX, INCX, ZBASE_SHIFT
    bge $r0,    N, .L999
@@ -60,6 +62,7 @@ USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
    CMPEQ   $fcc1, ALPHA_I, a1
    bceqz   $fcc0, .L50
    bceqz   $fcc1, .L50
+   beq     XX,    TEMP,  .L50  // if dummp2 == 1, do not directly copy 0
    srai.d I, N, 2
    bne INCX, TEMP, .L20
    bge $r0,    I, .L15