diff --git a/crates/core_arch/src/loongarch64/lasx/generated.rs b/crates/core_arch/src/loongarch64/lasx/generated.rs
index d2e1a87fde..5559c6ad4d 100644
--- a/crates/core_arch/src/loongarch64/lasx/generated.rs
+++ b/crates/core_arch/src/loongarch64/lasx/generated.rs
@@ -5044,7 +5044,7 @@ pub unsafe fn lasx_xvld<const IMM_S12: i32>(mem_addr: *const i8) -> m256i {
 #[unstable(feature = "stdarch_loongarch", issue = "117427")]
 pub unsafe fn lasx_xvst<const IMM_S12: i32>(a: m256i, mem_addr: *mut i8) {
     static_assert_simm_bits!(IMM_S12, 12);
-    transmute(__lasx_xvst(transmute(a), mem_addr, IMM_S12))
+    __lasx_xvst(transmute(a), mem_addr, IMM_S12)
 }
 
 #[inline(always)]
@@ -5054,7 +5054,7 @@ pub unsafe fn lasx_xvst<const IMM_S12: i32>(a: m256i, mem_addr: *mut i8) {
 pub unsafe fn lasx_xvstelm_b<const IMM_S8: i32, const IMM4: u32>(a: m256i, mem_addr: *mut i8) {
     static_assert_simm_bits!(IMM_S8, 8);
     static_assert_uimm_bits!(IMM4, 4);
-    transmute(__lasx_xvstelm_b(transmute(a), mem_addr, IMM_S8, IMM4))
+    __lasx_xvstelm_b(transmute(a), mem_addr, IMM_S8, IMM4)
 }
 
 #[inline(always)]
@@ -5064,7 +5064,7 @@ pub unsafe fn lasx_xvstelm_b<const IMM_S8: i32, const IMM4: u32>(a: m256i, mem_a
 pub unsafe fn lasx_xvstelm_h<const IMM_S8: i32, const IMM3: u32>(a: m256i, mem_addr: *mut i8) {
     static_assert_simm_bits!(IMM_S8, 8);
     static_assert_uimm_bits!(IMM3, 3);
-    transmute(__lasx_xvstelm_h(transmute(a), mem_addr, IMM_S8, IMM3))
+    __lasx_xvstelm_h(transmute(a), mem_addr, IMM_S8, IMM3)
 }
 
 #[inline(always)]
@@ -5074,7 +5074,7 @@ pub unsafe fn lasx_xvstelm_h<const IMM_S8: i32, const IMM3: u32>(a: m256i, mem_a
 pub unsafe fn lasx_xvstelm_w<const IMM_S8: i32, const IMM2: u32>(a: m256i, mem_addr: *mut i8) {
     static_assert_simm_bits!(IMM_S8, 8);
     static_assert_uimm_bits!(IMM2, 2);
-    transmute(__lasx_xvstelm_w(transmute(a), mem_addr, IMM_S8, IMM2))
+    __lasx_xvstelm_w(transmute(a), mem_addr, IMM_S8, IMM2)
 }
 
 #[inline(always)]
@@ -5084,7 +5084,7 @@ pub unsafe fn lasx_xvstelm_w<const IMM_S8: i32, const IMM2: u32>(a: m256i, mem_a
 pub unsafe fn lasx_xvstelm_d<const IMM_S8: i32, const IMM1: u32>(a: m256i, mem_addr: *mut i8) {
     static_assert_simm_bits!(IMM_S8, 8);
     static_assert_uimm_bits!(IMM1, 1);
-    transmute(__lasx_xvstelm_d(transmute(a), mem_addr, IMM_S8, IMM1))
+    __lasx_xvstelm_d(transmute(a), mem_addr, IMM_S8, IMM1)
 }
 
 #[inline(always)]
@@ -5192,7 +5192,7 @@ pub unsafe fn lasx_xvldx(mem_addr: *const i8, b: i64) -> m256i {
 #[target_feature(enable = "lasx")]
 #[unstable(feature = "stdarch_loongarch", issue = "117427")]
 pub unsafe fn lasx_xvstx(a: m256i, mem_addr: *mut i8, b: i64) {
-    transmute(__lasx_xvstx(transmute(a), mem_addr, transmute(b)))
+    __lasx_xvstx(transmute(a), mem_addr, transmute(b))
 }
 
 #[inline(always)]
diff --git a/crates/core_arch/src/loongarch64/lsx/generated.rs b/crates/core_arch/src/loongarch64/lsx/generated.rs
index 679c82079c..faa8859eba 100644
--- a/crates/core_arch/src/loongarch64/lsx/generated.rs
+++ b/crates/core_arch/src/loongarch64/lsx/generated.rs
@@ -5052,7 +5052,7 @@ pub fn lsx_vfrintrm_d(a: m128d) -> m128d {
 pub unsafe fn lsx_vstelm_b<const IMM_S8: i32, const IMM4: u32>(a: m128i, mem_addr: *mut i8) {
     static_assert_simm_bits!(IMM_S8, 8);
     static_assert_uimm_bits!(IMM4, 4);
-    transmute(__lsx_vstelm_b(transmute(a), mem_addr, IMM_S8, IMM4))
+    __lsx_vstelm_b(transmute(a), mem_addr, IMM_S8, IMM4)
 }
 
 #[inline(always)]
@@ -5062,7 +5062,7 @@ pub unsafe fn lsx_vstelm_b<const IMM_S8: i32, const IMM4: u32>(a: m128i, mem_add
 pub unsafe fn lsx_vstelm_h<const IMM_S8: i32, const IMM3: u32>(a: m128i, mem_addr: *mut i8) {
     static_assert_simm_bits!(IMM_S8, 8);
     static_assert_uimm_bits!(IMM3, 3);
-    transmute(__lsx_vstelm_h(transmute(a), mem_addr, IMM_S8, IMM3))
+    __lsx_vstelm_h(transmute(a), mem_addr, IMM_S8, IMM3)
 }
 
 #[inline(always)]
@@ -5072,7 +5072,7 @@ pub unsafe fn lsx_vstelm_h<const IMM_S8: i32, const IMM3: u32>(a: m128i, mem_add
 pub unsafe fn lsx_vstelm_w<const IMM_S8: i32, const IMM2: u32>(a: m128i, mem_addr: *mut i8) {
     static_assert_simm_bits!(IMM_S8, 8);
     static_assert_uimm_bits!(IMM2, 2);
-    transmute(__lsx_vstelm_w(transmute(a), mem_addr, IMM_S8, IMM2))
+    __lsx_vstelm_w(transmute(a), mem_addr, IMM_S8, IMM2)
 }
 
 #[inline(always)]
@@ -5082,7 +5082,7 @@ pub unsafe fn lsx_vstelm_w<const IMM_S8: i32, const IMM2: u32>(a: m128i, mem_add
 pub unsafe fn lsx_vstelm_d<const IMM_S8: i32, const IMM1: u32>(a: m128i, mem_addr: *mut i8) {
     static_assert_simm_bits!(IMM_S8, 8);
     static_assert_uimm_bits!(IMM1, 1);
-    transmute(__lsx_vstelm_d(transmute(a), mem_addr, IMM_S8, IMM1))
+    __lsx_vstelm_d(transmute(a), mem_addr, IMM_S8, IMM1)
 }
 
 #[inline(always)]
@@ -6376,7 +6376,7 @@ pub unsafe fn lsx_vld<const IMM_S12: i32>(mem_addr: *const i8) -> m128i {
 #[unstable(feature = "stdarch_loongarch", issue = "117427")]
 pub unsafe fn lsx_vst<const IMM_S12: i32>(a: m128i, mem_addr: *mut i8) {
     static_assert_simm_bits!(IMM_S12, 12);
-    transmute(__lsx_vst(transmute(a), mem_addr, IMM_S12))
+    __lsx_vst(transmute(a), mem_addr, IMM_S12)
 }
 
 #[inline(always)]
@@ -6455,7 +6455,7 @@ pub unsafe fn lsx_vldx(mem_addr: *const i8, b: i64) -> m128i {
 #[target_feature(enable = "lsx")]
 #[unstable(feature = "stdarch_loongarch", issue = "117427")]
 pub unsafe fn lsx_vstx(a: m128i, mem_addr: *mut i8, b: i64) {
-    transmute(__lsx_vstx(transmute(a), mem_addr, transmute(b)))
+    __lsx_vstx(transmute(a), mem_addr, transmute(b))
 }
 
 #[inline(always)]
diff --git a/crates/stdarch-gen-loongarch/src/main.rs b/crates/stdarch-gen-loongarch/src/main.rs
index fdc9b2b6bc..fe767fc309 100644
--- a/crates/stdarch-gen-loongarch/src/main.rs
+++ b/crates/stdarch-gen-loongarch/src/main.rs
@@ -571,21 +571,21 @@ fn gen_bind_body(
     } else if para_num == 3 && in_t[1] == "CVPOINTER" && in_t[2] == "SI" {
         call_params = match asm_fmts[2].as_str() {
             "si12" => format!(
-                "static_assert_simm_bits!(IMM_S12, 12);\n    {unsafe_start}transmute(__{current_name}(transmute(a), mem_addr, IMM_S12)){unsafe_end}"
+                "static_assert_simm_bits!(IMM_S12, 12);\n    {unsafe_start}__{current_name}(transmute(a), mem_addr, IMM_S12){unsafe_end}"
             ),
             _ => panic!("unsupported assembly format: {}", asm_fmts[2]),
         };
     } else if para_num == 3 && in_t[1] == "CVPOINTER" && in_t[2] == "DI" {
         call_params = match asm_fmts[2].as_str() {
             "rk" => format!(
-                "{unsafe_start}transmute(__{current_name}(transmute(a), mem_addr, transmute(b))){unsafe_end}"
+                "{unsafe_start}__{current_name}(transmute(a), mem_addr, transmute(b)){unsafe_end}"
             ),
             _ => panic!("unsupported assembly format: {}", asm_fmts[2]),
         };
     } else if para_num == 4 {
         call_params = match (asm_fmts[2].as_str(), current_name.chars().last().unwrap()) {
             ("si8", t) => format!(
-                "static_assert_simm_bits!(IMM_S8, 8);\n    static_assert_uimm_bits!(IMM{0}, {0});\n    {unsafe_start}transmute(__{current_name}(transmute(a), mem_addr, IMM_S8, IMM{0})){unsafe_end}",
+                "static_assert_simm_bits!(IMM_S8, 8);\n    static_assert_uimm_bits!(IMM{0}, {0});\n    {unsafe_start}__{current_name}(transmute(a), mem_addr, IMM_S8, IMM{0}){unsafe_end}",
                 type_to_imm(t)
             ),
             (_, _) => panic!(