optimization possibilities...

Before I convert this code into AT&T syntax for GAS to assemble, and to
MASM 
I was wondering if there are any possible optimizations I can perform on
the 
following code, I will show the C header first:


cpusync_i686.h
________________________________________________________________
#if ! defined(CPUSYNC_i686_H)
# define CPUSYNC_i686_H
# undef EXTERN
# if defined(__cplusplus)
#   define EXTERN extern "C"
# else
#   define EXTERN extern
# endif
/*************************************************************/

typedef int atomicword_i686;
typedef atomicword_i686 spinlock_i686;
typedef struct nbnode_i686_s nbnode_i686;
typedef struct nblifo_i686_s nblifo_i686;

struct nbnode_i686_s {
  nbnode_i686* next;
};

struct nblifo_i686_s {
  nbnode_i686* head;
  atomicword_i686 aba;
};

#define SPINLOCK_I686_STATICINIT() { 0 }
#define NBLIFO_I686_STATICINIT() { 0 }

typedef char cpusync_i686_static_assert[
  ((sizeof(atomicword_i686) == 4) &&
   (sizeof(spinlock_i686) == 4) &&
   (sizeof(nbnode_i686) == 4) &&
   (sizeof(nblifo_i686) == 8)
  ) ? 1 : -1
];

EXTERN void
spinlock_i686_acquire(
 atomicword_i686* const
);

EXTERN void
spinlock_i686_release(
 atomicword_i686* const
);

EXTERN void
nblifo_i686_push(
 nblifo_i686* const,
 nbnode_i686* const
);

EXTERN nbnode_i686*
nblifo_i686_flush(
 nblifo_i686* const
);

EXTERN nbnode_i686*
nblifo_i686_pop(
 nblifo_i686* const
);

/*************************************************************/
#endif

________________________________________________________________




Here is the temporary implementation which compiles with VC++:

cpusync_i686.c
________________________________________________________________
__declspec(naked) void
spinlock_i686_acquire(
 atomicword_i686* const _this
) {
  _asm {
    MOV ECX, [ESP + 4]
    MOV EAX, 1

spinlock_i686_acquire_retry:
    XCHG [ECX], EAX
    TEST EAX, EAX
    JNZ spinlock_i686_acquire_failed
    RET

spinlock_i686_acquire_failed:
    PAUSE
    JMP spinlock_i686_acquire_retry
  }
}


__declspec(naked) void
spinlock_i686_release(
 atomicword_i686* const _this
) {
  _asm {
    MOV ECX, [ESP + 4]
    MOV EAX, 0
    MOV [ECX], EAX
    RET
  }
}


__declspec(naked) void
nblifo_i686_push(
 nblifo_i686* const _this,
 nbnode_i686* const node
) {
  _asm {
    MOV EDX, [ESP + 4]
    MOV EAX, [EDX]
    MOV ECX, [ESP + 8]

nblifo_i686_push_retry:
    MOV [ECX], EAX
    LOCK CMPXCHG [EDX], ECX
    JNE nblifo_i686_push_retry
    RET
  }
}


__declspec(naked) nbnode_i686*
nblifo_i686_pop(
 nblifo_i686* const _this
) {
  _asm {
    PUSH ESI
    PUSH EBX
    MOV ESI, [ESP + 12]
    MOV EDX, [ESI + 4]
    MOV EAX, [ESI]

nblifo_i686_pop_retry:
    TEST EAX, EAX
    JZ nblifo_i686_pop_fail
    MOV EBX, [EAX]
    LEA ECX, [EDX + 1]
    LOCK CMPXCHG8B QWORD PTR [ESI]
    JNE nblifo_i686_pop_retry

nblifo_i686_pop_fail:
    POP EBX
    POP ESI
    RET
  }
}


__declspec(naked) nbnode_i686*
nblifo_i686_flush(
 nblifo_i686* const _this
) {
  _asm {
    MOV ECX, [ESP + 4]
    MOV EAX, 0
    XCHG [ECX], EAX
    RET
  }
}
________________________________________________________________




AFAICT, everything seems to look okay. However, I am not an expert x86 
programmer! Any comments/suggestions would be highly appreciated.


Thanks.


-- 
Chris M. Thomasson
http://appcore.home.comcast.net