Allow the compiler to micro-optimize byte swapping functions by

evaluating them at compile time rather than at run time.  As for x86
and amd64, this requires GCC and it's enabled only if __OPTIMIZE__ is
defined (ie, if at least -O is used).

Reviewed by:	jake

1 files changed, 24 insertions, 15 deletions

diff --git a/sys/sparc64/include/endian.h b/sys/sparc64/include/endian.h
index f10032c..a2ff62b 100644
--- a/sys/sparc64/include/endian.h
+++ b/sys/sparc64/include/endian.h
@@ -67,17 +67,30 @@
 #define	BYTE_ORDER	_BYTE_ORDER
 #endif
 
-#ifdef __GNUC__
+#if defined(__GNUC__) && defined(__OPTIMIZE__)
+#define	__is_constant(x)	__builtin_constant_p(x)
+#else
+#define	__is_constant(x)	0
+#endif
+
+#define	__bswap16_const(x)	((x >> 8) | ((x << 8) & 0xff00))
+#define	__bswap32_const(x)	((x >> 24) | ((x >> 8) & 0xff00) |	\
+	((x << 8) & 0xff0000) | ((x << 24) & 0xff000000))
+#define	__bswap64_const(x)	((x >> 56) | ((x >> 40) & 0xff00) |	\
+	((x >> 24) & 0xff0000) | ((x >> 8) & 0xff000000) |		\
+	((x << 8) & ((__uint64_t)0xff << 32)) |				\
+	((x << 24) & ((__uint64_t)0xff << 40)) |			\
+	((x << 40) & ((__uint64_t)0xff << 48)) | ((x << 56)))
 
 static __inline __uint16_t
-__bswap16(__uint16_t _x)
+__bswap16_var(__uint16_t _x)
 {
 
 	return ((_x >> 8) | ((_x << 8) & 0xff00));
 }
 
 static __inline __uint32_t
-__bswap32(__uint32_t _x)
+__bswap32_var(__uint32_t _x)
 {
 
 	return ((_x >> 24) | ((_x >> 8) & 0xff00) | ((_x << 8) & 0xff0000) |
@@ -85,7 +98,7 @@ __bswap32(__uint32_t _x)
 }
 
 static __inline __uint64_t
-__bswap64(__uint64_t _x)
+__bswap64_var(__uint64_t _x)
 {
 
 	return ((_x >> 56) | ((_x >> 40) & 0xff00) | ((_x >> 24) & 0xff0000) |
@@ -94,20 +107,16 @@ __bswap64(__uint64_t _x)
 	    ((_x << 40) & ((__uint64_t)0xff << 48)) | ((_x << 56)));
 }
 
+#define	__bswap16(x)	(__is_constant(x) ? __bswap16_const(x) : \
+	__bswap16_var(x))
+#define	__bswap32(x)	(__is_constant(x) ? __bswap32_const(x) : \
+	__bswap32_var(x))
+#define	__bswap64(x)	(__is_constant(x) ? __bswap64_const(x) : \
+	__bswap64_var(x))
+
 #define	__htonl(x)	((__uint32_t)(x))
 #define	__htons(x)	((__uint16_t)(x))
 #define	__ntohl(x)	((__uint32_t)(x))
 #define	__ntohs(x)	((__uint16_t)(x))
 
-#else /* !__GNUC__ */
-
-/*
- * No optimizations are available for this compiler.  Fall back to
- * non-optimized functions by defining the constant usually used to prevent
- * redefinition.
- */
-#define	_BYTEORDER_FUNC_DEFINED
-
-#endif /* __GNUC__ */
-
 #endif /* !_MACHINE_ENDIAN_H_ */