diff --git a/internal/native/neon/f32toa_arm64.s b/internal/native/neon/f32toa_arm64.s
index 7be208dbe..cc6abecd7 100644
--- a/internal/native/neon/f32toa_arm64.s
+++ b/internal/native/neon/f32toa_arm64.s
@@ -7,7 +7,8 @@
 
 TEXT ·__f32toa_entry__(SB), NOSPLIT, $16
 	NO_LOCAL_POINTERS
-	PCALIGN $16
+	WORD $0x910083ff  // add sp, sp, #32
+	JMP _f32toa
 	  // .p2align 2, 0x00
 _f32toa:
 	WORD $0xa9be7bfd  // stp	fp, lr, [sp, #-32]!
@@ -904,7 +905,7 @@ _Digits:
 	WORD $0x37393639  // .ascii 4, '96979899'
 	WORD $0x39393839  // .ascii 4, '9899'
 	WORD $0x00000000  // .p2align 3, 0x00
-_LB_b84e6c90: // _pow10_ceil_sig_f32.g
+_LB_e80ea3d0: // _pow10_ceil_sig_f32.g
 	WORD $0x4b43fcf5; WORD $0x81ceb32c  // .quad -9093133594791772939
 	WORD $0x5e14fc32; WORD $0xa2425ff7  // .quad -6754730975062328270
 	WORD $0x359a3b3f; WORD $0xcad2f7f5  // .quad -3831727700400522433
@@ -996,7 +997,7 @@ _f32toa:
 	MOVD out+0(FP), R0
 	FMOVD val+8(FP), F0
 	WORD $0xf90007fc // str x28, [sp, #8]
-	CALL ·__f32toa_entry__+16(SB)  // _f32toa
+	CALL ·__f32toa_entry__(SB)  // _f32toa
 	WORD $0xf94007fc // ldr x28, [sp, #8]
 	MOVD R0, ret+16(FP)
 	RET
diff --git a/internal/native/neon/f64toa_arm64.s b/internal/native/neon/f64toa_arm64.s
index 5a14e3010..92ce3d329 100644
--- a/internal/native/neon/f64toa_arm64.s
+++ b/internal/native/neon/f64toa_arm64.s
@@ -7,7 +7,8 @@
 
 TEXT ·__f64toa_entry__(SB), NOSPLIT, $16
 	NO_LOCAL_POINTERS
-	PCALIGN $16
+	WORD $0x910083ff  // add sp, sp, #32
+	JMP _f64toa
 	  // .p2align 2, 0x00
 _f64toa:
 	WORD $0xa9be7bfd  // stp	fp, lr, [sp, #-32]!
@@ -1220,7 +1221,7 @@ _Digits:
 	WORD $0x37393639  // .ascii 4, '96979899'
 	WORD $0x39393839  // .ascii 4, '9899'
 	  // .p2align 3, 0x00
-_LB_de8c9345: // _pow10_ceil_sig.g
+_LB_40c6cc40: // _pow10_ceil_sig.g
 	WORD $0xbebcdc4f; WORD $0xff77b1fc  // .quad -38366372719436721
 	WORD $0x13bb0f7b; WORD $0x25e8e89c  // .quad 2731688931043774331
 	WORD $0xf73609b1; WORD $0x9faacf3d  // .quad -6941508010590729807
@@ -2469,7 +2470,7 @@ _f64toa:
 	MOVD out+0(FP), R0
 	FMOVD val+8(FP), F0
 	WORD $0xf90007fc // str x28, [sp, #8]
-	CALL ·__f64toa_entry__+16(SB)  // _f64toa
+	CALL ·__f64toa_entry__(SB)  // _f64toa
 	WORD $0xf94007fc // ldr x28, [sp, #8]
 	MOVD R0, ret+16(FP)
 	RET
diff --git a/internal/native/neon/get_by_path_arm64.s b/internal/native/neon/get_by_path_arm64.s
index 9f5c49c30..1a1769f12 100644
--- a/internal/native/neon/get_by_path_arm64.s
+++ b/internal/native/neon/get_by_path_arm64.s
@@ -7,7 +7,8 @@
 
 TEXT ·__get_by_path_entry__(SB), NOSPLIT, $192
 	NO_LOCAL_POINTERS
-	PCALIGN $16
+	WORD $0x910343ff  // add sp, sp, #208
+	JMP _get_by_path
 	  // .p2align 4, 0x00
 lCPI0_0:
 	WORD $0x08040201
@@ -4566,7 +4567,7 @@ _get_by_path:
 	MOVD path+16(FP), R2
 	MOVD m+24(FP), R3
 	WORD $0xf90007fc // str x28, [sp, #8]
-	CALL ·__get_by_path_entry__+64(SB)  // _get_by_path
+	CALL ·__get_by_path_entry__(SB)  // _get_by_path
 	WORD $0xf94007fc // ldr x28, [sp, #8]
 	MOVD R0, ret+32(FP)
 	RET
diff --git a/internal/native/neon/html_escape_arm64.s b/internal/native/neon/html_escape_arm64.s
index 2227891fc..7f44ff66e 100644
--- a/internal/native/neon/html_escape_arm64.s
+++ b/internal/native/neon/html_escape_arm64.s
@@ -7,7 +7,8 @@
 
 TEXT ·__html_escape_entry__(SB), NOSPLIT, $16
 	NO_LOCAL_POINTERS
-	PCALIGN $16
+	WORD $0x910083ff  // add sp, sp, #32
+	JMP _html_escape
 	  // .p2align 4, 0x00
 lCPI0_0:
 	WORD $0x08040201
@@ -1389,7 +1390,7 @@ _html_escape:
 	MOVD dp+16(FP), R2
 	MOVD dn+24(FP), R3
 	WORD $0xf90007fc // str x28, [sp, #8]
-	CALL ·__html_escape_entry__+48(SB)  // _html_escape
+	CALL ·__html_escape_entry__(SB)  // _html_escape
 	WORD $0xf94007fc // ldr x28, [sp, #8]
 	MOVD R0, ret+32(FP)
 	RET
diff --git a/internal/native/neon/i64toa_arm64.s b/internal/native/neon/i64toa_arm64.s
index 9260bca9d..828e7f73b 100644
--- a/internal/native/neon/i64toa_arm64.s
+++ b/internal/native/neon/i64toa_arm64.s
@@ -7,7 +7,8 @@
 
 TEXT ·__i64toa_entry__(SB), NOSPLIT, $16
 	NO_LOCAL_POINTERS
-	PCALIGN $16
+	WORD $0x910083ff  // add sp, sp, #32
+	JMP _i64toa
 	  // .p2align 3, 0x00
 lCPI0_0:
 	WORD $0x147b20c5
@@ -963,7 +964,7 @@ _i64toa:
 	MOVD out+0(FP), R0
 	MOVD val+8(FP), R1
 	WORD $0xf90007fc // str x28, [sp, #8]
-	CALL ·__i64toa_entry__+64(SB)  // _i64toa
+	CALL ·__i64toa_entry__(SB)  // _i64toa
 	WORD $0xf94007fc // ldr x28, [sp, #8]
 	MOVD R0, ret+16(FP)
 	RET
diff --git a/internal/native/neon/lspace_arm64.s b/internal/native/neon/lspace_arm64.s
index a8d688e98..c62c6f615 100644
--- a/internal/native/neon/lspace_arm64.s
+++ b/internal/native/neon/lspace_arm64.s
@@ -7,7 +7,8 @@
 
 TEXT ·__lspace_entry__(SB), NOSPLIT, $16
 	NO_LOCAL_POINTERS
-	PCALIGN $16
+	WORD $0x910083ff  // add sp, sp, #32
+	JMP _lspace
 	  // .p2align 2, 0x00
 _lspace:
 	WORD $0xa9be7bfd  // stp	fp, lr, [sp, #-32]!
@@ -53,7 +54,7 @@ _lspace:
 	MOVD nb+8(FP), R1
 	MOVD off+16(FP), R2
 	WORD $0xf90007fc // str x28, [sp, #8]
-	CALL ·__lspace_entry__+16(SB)  // _lspace
+	CALL ·__lspace_entry__(SB)  // _lspace
 	WORD $0xf94007fc // ldr x28, [sp, #8]
 	MOVD R0, ret+24(FP)
 	RET
diff --git a/internal/native/neon/quote_arm64.s b/internal/native/neon/quote_arm64.s
index 8f1f5d545..61cd7dd52 100644
--- a/internal/native/neon/quote_arm64.s
+++ b/internal/native/neon/quote_arm64.s
@@ -7,7 +7,8 @@
 
 TEXT ·__quote_entry__(SB), NOSPLIT, $16
 	NO_LOCAL_POINTERS
-	PCALIGN $16
+	WORD $0x910083ff  // add sp, sp, #32
+	JMP _quote
 	  // .p2align 4, 0x00
 lCPI0_0:
 	WORD $0x08040201
@@ -2545,7 +2546,7 @@ _quote:
 	MOVD dn+24(FP), R3
 	MOVD flags+32(FP), R4
 	WORD $0xf90007fc // str x28, [sp, #8]
-	CALL ·__quote_entry__+48(SB)  // _quote
+	CALL ·__quote_entry__(SB)  // _quote
 	WORD $0xf94007fc // ldr x28, [sp, #8]
 	MOVD R0, ret+40(FP)
 	RET
diff --git a/internal/native/neon/skip_array_arm64.s b/internal/native/neon/skip_array_arm64.s
index 9986d49a5..9e258625a 100644
--- a/internal/native/neon/skip_array_arm64.s
+++ b/internal/native/neon/skip_array_arm64.s
@@ -7,7 +7,8 @@
 
 TEXT ·__skip_array_entry__(SB), NOSPLIT, $96
 	NO_LOCAL_POINTERS
-	PCALIGN $16
+	WORD $0x9101c3ff  // add sp, sp, #112
+	JMP _skip_array
 	  // .p2align 4, 0x00
 lCPI0_0:
 	WORD $0x00000001; WORD $0x00000000  // .quad 1
@@ -2369,7 +2370,7 @@ _skip_array:
 	MOVD m+16(FP), R2
 	MOVD flags+24(FP), R3
 	WORD $0xf90007fc // str x28, [sp, #8]
-	CALL ·__skip_array_entry__+64(SB)  // _skip_array
+	CALL ·__skip_array_entry__(SB)  // _skip_array
 	WORD $0xf94007fc // ldr x28, [sp, #8]
 	MOVD R0, ret+32(FP)
 	RET
diff --git a/internal/native/neon/skip_number_arm64.s b/internal/native/neon/skip_number_arm64.s
index 232bb8072..af72ff40c 100644
--- a/internal/native/neon/skip_number_arm64.s
+++ b/internal/native/neon/skip_number_arm64.s
@@ -7,7 +7,8 @@
 
 TEXT ·__skip_number_entry__(SB), NOSPLIT, $32
 	NO_LOCAL_POINTERS
-	PCALIGN $16
+	WORD $0x9100c3ff  // add sp, sp, #48
+	JMP _skip_number
 	  // .p2align 4, 0x00
 lCPI0_0:
 	WORD $0x08040201
@@ -360,7 +361,7 @@ _skip_number:
 	MOVD s+0(FP), R0
 	MOVD p+8(FP), R1
 	WORD $0xf90007fc // str x28, [sp, #8]
-	CALL ·__skip_number_entry__+48(SB)  // _skip_number
+	CALL ·__skip_number_entry__(SB)  // _skip_number
 	WORD $0xf94007fc // ldr x28, [sp, #8]
 	MOVD R0, ret+16(FP)
 	RET
diff --git a/internal/native/neon/skip_object_arm64.s b/internal/native/neon/skip_object_arm64.s
index 7ff58bdd8..eb3cd036d 100644
--- a/internal/native/neon/skip_object_arm64.s
+++ b/internal/native/neon/skip_object_arm64.s
@@ -7,7 +7,8 @@
 
 TEXT ·__skip_object_entry__(SB), NOSPLIT, $96
 	NO_LOCAL_POINTERS
-	PCALIGN $16
+	WORD $0x9101c3ff  // add sp, sp, #112
+	JMP _skip_object
 	  // .p2align 4, 0x00
 lCPI0_0:
 	WORD $0x00000001; WORD $0x00000000  // .quad 1
@@ -2369,7 +2370,7 @@ _skip_object:
 	MOVD m+16(FP), R2
 	MOVD flags+24(FP), R3
 	WORD $0xf90007fc // str x28, [sp, #8]
-	CALL ·__skip_object_entry__+64(SB)  // _skip_object
+	CALL ·__skip_object_entry__(SB)  // _skip_object
 	WORD $0xf94007fc // ldr x28, [sp, #8]
 	MOVD R0, ret+32(FP)
 	RET
diff --git a/internal/native/neon/skip_one_arm64.s b/internal/native/neon/skip_one_arm64.s
index e45105c10..694c38ec4 100644
--- a/internal/native/neon/skip_one_arm64.s
+++ b/internal/native/neon/skip_one_arm64.s
@@ -7,7 +7,8 @@
 
 TEXT ·__skip_one_entry__(SB), NOSPLIT, $96
 	NO_LOCAL_POINTERS
-	PCALIGN $16
+	WORD $0x9101c3ff  // add sp, sp, #112
+	JMP _skip_one
 	  // .p2align 4, 0x00
 lCPI0_0:
 	WORD $0x00000001; WORD $0x00000000  // .quad 1
@@ -2369,7 +2370,7 @@ _skip_one:
 	MOVD m+16(FP), R2
 	MOVD flags+24(FP), R3
 	WORD $0xf90007fc // str x28, [sp, #8]
-	CALL ·__skip_one_entry__+64(SB)  // _skip_one
+	CALL ·__skip_one_entry__(SB)  // _skip_one
 	WORD $0xf94007fc // ldr x28, [sp, #8]
 	MOVD R0, ret+32(FP)
 	RET
diff --git a/internal/native/neon/skip_one_fast_arm64.s b/internal/native/neon/skip_one_fast_arm64.s
index 98dc26c4f..47f883d50 100644
--- a/internal/native/neon/skip_one_fast_arm64.s
+++ b/internal/native/neon/skip_one_fast_arm64.s
@@ -7,7 +7,8 @@
 
 TEXT ·__skip_one_fast_entry__(SB), NOSPLIT, $176
 	NO_LOCAL_POINTERS
-	PCALIGN $16
+	WORD $0x910303ff  // add sp, sp, #192
+	JMP _skip_one_fast
 	  // .p2align 4, 0x00
 lCPI0_0:
 	WORD $0x08040201
@@ -985,7 +986,7 @@ _skip_one_fast:
 	MOVD s+0(FP), R0
 	MOVD p+8(FP), R1
 	WORD $0xf90007fc // str x28, [sp, #8]
-	CALL ·__skip_one_fast_entry__+48(SB)  // _skip_one_fast
+	CALL ·__skip_one_fast_entry__(SB)  // _skip_one_fast
 	WORD $0xf94007fc // ldr x28, [sp, #8]
 	MOVD R0, ret+16(FP)
 	RET
diff --git a/internal/native/neon/u64toa_arm64.s b/internal/native/neon/u64toa_arm64.s
index 7a27ab31a..da0f7018c 100644
--- a/internal/native/neon/u64toa_arm64.s
+++ b/internal/native/neon/u64toa_arm64.s
@@ -7,7 +7,8 @@
 
 TEXT ·__u64toa_entry__(SB), NOSPLIT, $16
 	NO_LOCAL_POINTERS
-	PCALIGN $16
+	WORD $0x910083ff  // add sp, sp, #32
+	JMP _u64toa
 	  // .p2align 3, 0x00
 lCPI0_0:
 	WORD $0x147b20c5
@@ -580,7 +581,7 @@ _u64toa:
 	MOVD out+0(FP), R0
 	MOVD val+8(FP), R1
 	WORD $0xf90007fc // str x28, [sp, #8]
-	CALL ·__u64toa_entry__+64(SB)  // _u64toa
+	CALL ·__u64toa_entry__(SB)  // _u64toa
 	WORD $0xf94007fc // ldr x28, [sp, #8]
 	MOVD R0, ret+16(FP)
 	RET
diff --git a/internal/native/neon/unquote_arm64.s b/internal/native/neon/unquote_arm64.s
index a799bb673..c58cf0052 100644
--- a/internal/native/neon/unquote_arm64.s
+++ b/internal/native/neon/unquote_arm64.s
@@ -7,7 +7,8 @@
 
 TEXT ·__unquote_entry__(SB), NOSPLIT, $96
 	NO_LOCAL_POINTERS
-	PCALIGN $16
+	WORD $0x9101c3ff  // add sp, sp, #112
+	JMP _unquote
 	  // .p2align 4, 0x00
 lCPI0_0:
 	WORD $0x08040201
@@ -600,7 +601,7 @@ _unquote:
 	MOVD ep+24(FP), R3
 	MOVD flags+32(FP), R4
 	WORD $0xf90007fc // str x28, [sp, #8]
-	CALL ·__unquote_entry__+48(SB)  // _unquote
+	CALL ·__unquote_entry__(SB)  // _unquote
 	WORD $0xf94007fc // ldr x28, [sp, #8]
 	MOVD R0, ret+40(FP)
 	RET
diff --git a/internal/native/neon/validate_one_arm64.s b/internal/native/neon/validate_one_arm64.s
index 58c57ea4e..088b02e90 100644
--- a/internal/native/neon/validate_one_arm64.s
+++ b/internal/native/neon/validate_one_arm64.s
@@ -7,7 +7,8 @@
 
 TEXT ·__validate_one_entry__(SB), NOSPLIT, $96
 	NO_LOCAL_POINTERS
-	PCALIGN $16
+	WORD $0x9101c3ff  // add sp, sp, #112
+	JMP _validate_one
 	  // .p2align 4, 0x00
 lCPI0_0:
 	WORD $0x00000001; WORD $0x00000000  // .quad 1
@@ -1845,7 +1846,7 @@ _validate_one:
 	MOVD p+8(FP), R1
 	MOVD m+16(FP), R2
 	WORD $0xf90007fc // str x28, [sp, #8]
-	CALL ·__validate_one_entry__+64(SB)  // _validate_one
+	CALL ·__validate_one_entry__(SB)  // _validate_one
 	WORD $0xf94007fc // ldr x28, [sp, #8]
 	MOVD R0, ret+24(FP)
 	RET
diff --git a/internal/native/neon/validate_utf8_arm64.s b/internal/native/neon/validate_utf8_arm64.s
index ca87d00b3..f87c1eaac 100644
--- a/internal/native/neon/validate_utf8_arm64.s
+++ b/internal/native/neon/validate_utf8_arm64.s
@@ -7,7 +7,8 @@
 
 TEXT ·__validate_utf8_entry__(SB), NOSPLIT, $48
 	NO_LOCAL_POINTERS
-	PCALIGN $16
+	WORD $0x910103ff  // add sp, sp, #64
+	JMP _validate_utf8
 	  // .p2align 2, 0x00
 _validate_utf8:
 	WORD $0xd10103ff  // sub	sp, sp, #64
@@ -209,7 +210,7 @@ _validate_utf8:
 	MOVD p+8(FP), R1
 	MOVD m+16(FP), R2
 	WORD $0xf90007fc // str x28, [sp, #8]
-	CALL ·__validate_utf8_entry__+16(SB)  // _validate_utf8
+	CALL ·__validate_utf8_entry__(SB)  // _validate_utf8
 	WORD $0xf94007fc // ldr x28, [sp, #8]
 	MOVD R0, ret+24(FP)
 	RET
diff --git a/internal/native/neon/validate_utf8_fast_arm64.s b/internal/native/neon/validate_utf8_fast_arm64.s
index 4296b6a4f..95da166eb 100644
--- a/internal/native/neon/validate_utf8_fast_arm64.s
+++ b/internal/native/neon/validate_utf8_fast_arm64.s
@@ -7,7 +7,8 @@
 
 TEXT ·__validate_utf8_fast_entry__(SB), NOSPLIT, $32
 	NO_LOCAL_POINTERS
-	PCALIGN $16
+	WORD $0x9100c3ff  // add sp, sp, #48
+	JMP _validate_utf8_fast
 	  // .p2align 2, 0x00
 _validate_utf8_fast:
 	WORD $0xd100c3ff  // sub	sp, sp, #48
@@ -169,7 +170,7 @@ _entry:
 _validate_utf8_fast:
 	MOVD s+0(FP), R0
 	WORD $0xf90007fc // str x28, [sp, #8]
-	CALL ·__validate_utf8_fast_entry__+16(SB)  // _validate_utf8_fast
+	CALL ·__validate_utf8_fast_entry__(SB)  // _validate_utf8_fast
 	WORD $0xf94007fc // ldr x28, [sp, #8]
 	MOVD R0, ret+8(FP)
 	RET
diff --git a/internal/native/neon/value_arm64.s b/internal/native/neon/value_arm64.s
index a3c79debd..a6cf69959 100644
--- a/internal/native/neon/value_arm64.s
+++ b/internal/native/neon/value_arm64.s
@@ -7,7 +7,8 @@
 
 TEXT ·__value_entry__(SB), NOSPLIT, $96
 	NO_LOCAL_POINTERS
-	PCALIGN $16
+	WORD $0x9101c3ff  // add sp, sp, #112
+	JMP _value
 	  // .p2align 4, 0x00
 lCPI0_0:
 	WORD $0x08040201
@@ -5953,7 +5954,7 @@ _value:
 	MOVD v+24(FP), R3
 	MOVD flags+32(FP), R4
 	WORD $0xf90007fc // str x28, [sp, #8]
-	CALL ·__value_entry__+48(SB)  // _value
+	CALL ·__value_entry__(SB)  // _value
 	WORD $0xf94007fc // ldr x28, [sp, #8]
 	MOVD R0, ret+40(FP)
 	RET
diff --git a/internal/native/neon/vnumber_arm64.s b/internal/native/neon/vnumber_arm64.s
index c9ae8f4fc..a531a49ad 100644
--- a/internal/native/neon/vnumber_arm64.s
+++ b/internal/native/neon/vnumber_arm64.s
@@ -7,7 +7,8 @@
 
 TEXT ·__vnumber_entry__(SB), NOSPLIT, $96
 	NO_LOCAL_POINTERS
-	PCALIGN $16
+	WORD $0x9101c3ff  // add sp, sp, #112
+	JMP _vnumber
 	  // .p2align 2, 0x00
 _vnumber:
 	WORD $0xa9b96ffc  // stp	x28, x27, [sp, #-112]!
@@ -4676,7 +4677,7 @@ _vnumber:
 	MOVD p+8(FP), R1
 	MOVD v+16(FP), R2
 	WORD $0xf90007fc // str x28, [sp, #8]
-	CALL ·__vnumber_entry__+16(SB)  // _vnumber
+	CALL ·__vnumber_entry__(SB)  // _vnumber
 	WORD $0xf94007fc // ldr x28, [sp, #8]
 	RET
 
diff --git a/internal/native/neon/vsigned_arm64.s b/internal/native/neon/vsigned_arm64.s
index debadbd10..8b8acd0c4 100644
--- a/internal/native/neon/vsigned_arm64.s
+++ b/internal/native/neon/vsigned_arm64.s
@@ -7,7 +7,8 @@
 
 TEXT ·__vsigned_entry__(SB), NOSPLIT, $16
 	NO_LOCAL_POINTERS
-	PCALIGN $16
+	WORD $0x910083ff  // add sp, sp, #32
+	JMP _vsigned
 	  // .p2align 2, 0x00
 _vsigned:
 	WORD $0xa9be7bfd  // stp	fp, lr, [sp, #-32]!
@@ -137,7 +138,7 @@ _vsigned:
 	MOVD p+8(FP), R1
 	MOVD v+16(FP), R2
 	WORD $0xf90007fc // str x28, [sp, #8]
-	CALL ·__vsigned_entry__+16(SB)  // _vsigned
+	CALL ·__vsigned_entry__(SB)  // _vsigned
 	WORD $0xf94007fc // ldr x28, [sp, #8]
 	RET
 
diff --git a/internal/native/neon/vstring_arm64.s b/internal/native/neon/vstring_arm64.s
index e47fa4b32..9c259c448 100644
--- a/internal/native/neon/vstring_arm64.s
+++ b/internal/native/neon/vstring_arm64.s
@@ -7,7 +7,8 @@
 
 TEXT ·__vstring_entry__(SB), NOSPLIT, $32
 	NO_LOCAL_POINTERS
-	PCALIGN $16
+	WORD $0x9100c3ff  // add sp, sp, #48
+	JMP _vstring
 	  // .p2align 4, 0x00
 lCPI0_0:
 	WORD $0x08040201
@@ -631,7 +632,7 @@ _vstring:
 	MOVD v+16(FP), R2
 	MOVD flags+24(FP), R3
 	WORD $0xf90007fc // str x28, [sp, #8]
-	CALL ·__vstring_entry__+48(SB)  // _vstring
+	CALL ·__vstring_entry__(SB)  // _vstring
 	WORD $0xf94007fc // ldr x28, [sp, #8]
 	RET
 
diff --git a/internal/native/neon/vunsigned_arm64.s b/internal/native/neon/vunsigned_arm64.s
index 372265a8d..b3c1f3fb3 100644
--- a/internal/native/neon/vunsigned_arm64.s
+++ b/internal/native/neon/vunsigned_arm64.s
@@ -7,7 +7,8 @@
 
 TEXT ·__vunsigned_entry__(SB), NOSPLIT, $16
 	NO_LOCAL_POINTERS
-	PCALIGN $16
+	WORD $0x910083ff  // add sp, sp, #32
+	JMP _vunsigned
 	  // .p2align 2, 0x00
 _vunsigned:
 	WORD $0xa9be7bfd  // stp	fp, lr, [sp, #-32]!
@@ -131,7 +132,7 @@ _vunsigned:
 	MOVD p+8(FP), R1
 	MOVD v+16(FP), R2
 	WORD $0xf90007fc // str x28, [sp, #8]
-	CALL ·__vunsigned_entry__+16(SB)  // _vunsigned
+	CALL ·__vunsigned_entry__(SB)  // _vunsigned
 	WORD $0xf94007fc // ldr x28, [sp, #8]
 	RET
 
diff --git a/tools/asm2arm/arm.py b/tools/asm2arm/arm.py
index 50a5a8faf..e4a5ea8d7 100644
--- a/tools/asm2arm/arm.py
+++ b/tools/asm2arm/arm.py
@@ -2014,11 +2014,12 @@ def _declare(self, protos: PrototypeMap):
         self._declare_functions(protos)
 
     def _declare_body(self, name: str):
-        size = self.code.stacksize(name)-16
-        size = 0 if size < 0 else size
-        self.out.append('TEXT ·_%s_entry__(SB), NOSPLIT, $%d' % (name, size))
+        size = self.code.stacksize(name)
+        gosize = 0 if size < 16 else size-16
+        self.out.append('TEXT ·_%s_entry__(SB), NOSPLIT, $%d' % (name, gosize))
         self.out.append('\tNO_LOCAL_POINTERS')
-        self.out.append('\tPCALIGN $16')
+        self.out.append('\t'+Instruction('add sp, sp, #%d' % size).encoded)
+        self.out.append('\tJMP %s' % name)
         self._LE_4bytes_IntIntr_2_RawIntr()
         self._reloc()
 
@@ -2118,7 +2119,7 @@ def _declare_function(self, name: str, proto: Prototype):
         # self.out.append('\tSUB $16, RSP')
         # self.out.append('\tADD $%d, RSP' % (size + 32))
         self.out.append('\tWORD $0xf90007fc // str x28, [sp, #8]')
-        self.out.append('\tCALL ·_%s_entry__+%d(SB)  // %s' % (subr, addr, subr))
+        self.out.append('\tCALL ·_%s_entry__(SB)  // %s' % (subr, subr))
         self.out.append('\tWORD $0xf94007fc // ldr x28, [sp, #8]')
         # self.out.append('\tSUB $%d, RSP' % (size + 32))
         # self.out.append('\tADD $16, RSP')