Generate compiler compatible assembly symbols

Typically when compiling a function `foo`,  GCC will add a prefix `_` to the symbol, i.e., generate `_foo`. But on some platforms, the compiler will not add prefix. option `ENABLE_ASM_UNDERSCORE_PREFIX` change the default name of global symbols in assembly code.

src/gf128_aarch64.S

@@ -8,6 +8,9 @@
  */
 
 
+#include <gmssl/asm.h>
+
+
 /* GF(2^128) defined by f(x) = x^128 + x^7 + x^2 + x + 1
 
 	f0 = x^128 = x^7 + x^2 + x + 1
@@ -29,9 +32,11 @@
 	= c + (e0 + w0) * x^64 + (d0 + w1) * f0
 */
 .text
-.globl	_gf128_mul
+
+.globl	func(gf128_mul)
 .align	4
-_gf128_mul:
+
+func(gf128_mul):
 	// load (a0, a1)
 	ld1	{v1.2d},[x1]
 	// load (b0, b1)

src/sm4_aarch64.S

@@ -7,6 +7,7 @@
  *  http://www.apache.org/licenses/LICENSE-2.0
  */
 
+#include <gmssl/asm.h>
 
 .align	7
 
@@ -62,9 +63,10 @@ Llshift:
 .byte	4,5,6,7,  8,9,10,11,  12,13,14,15,  0,1,2,3
 
 
-.globl	_sm4_set_encrypt_key
+.globl	func(sm4_set_encrypt_key)
 .align	4
-_sm4_set_encrypt_key:
+
+func(sm4_set_encrypt_key):
 
 	// load const v16..v31 = SBox
 	adr	x3, LSBOX
@@ -140,9 +142,10 @@ _sm4_set_encrypt_key:
 	ret
 
 
-.globl	_sm4_set_decrypt_key
+.globl	func(sm4_set_decrypt_key)
 .align	4
-_sm4_set_decrypt_key:
+
+func(sm4_set_decrypt_key):
 
 	// load const v16..v31 = SBox
 	adr	x3,LSBOX
@@ -221,10 +224,10 @@ _sm4_set_decrypt_key:
 	ret
 
 
-.globl	_sm4_encrypt
+.globl	func(sm4_encrypt)
 
 .align	5
-_sm4_encrypt:
+func(sm4_encrypt):
 
 	// load sbox
 	adr	x3, LSBOX