o Don't require long long support in bswap64() functions.

o In i386's <machine/endian.h>, macros have some advantages over
  inlines, so change some inlines to macros.
o In i386's <machine/endian.h>, ungarbage collect word_swap_int()
  (previously __uint16_swap_uint32), it has some uses on i386's with
  PDP endianness.

Submitted by:	bde

o Move a comment up in <machine/endian.h> that was accidentially moved
  down a few revisions ago.
o Reenable userland's use of optimized inline-asm versions of
  byteorder(3) functions.
o Fix ordering of prototypes vs. redefinition of byteorder(3)
  functions, so that the non-GCC (libc asm) case has proper
  prototypes.
o Add proper prototypes for byteorder(3) functions in <sys/param.h>.
o Prevent redundant duplicate prototypes by making use of the
  _BYTEORDER_PROTOTYPED define.
o Move the bswap16(), bswap32(), bswap64() C functions into MD space
  for platforms in which asm versions don't exist.  This significantly
  reduces the complexity of some things at the cost of duplicate code.

Reviewed by:	bde

7 files changed, 163 insertions, 84 deletions

diff --git a/sys/alpha/include/endian.h b/sys/alpha/include/endian.h
index 28f7dd0..6781ceda 100644
--- a/sys/alpha/include/endian.h
+++ b/sys/alpha/include/endian.h
@@ -46,11 +46,11 @@
 #define _QUAD_HIGHWORD 1
 #define _QUAD_LOWWORD 0
 
+#ifndef _POSIX_SOURCE
 /*
  * Definitions for byte order, according to byte significance from low
  * address to high.
  */
-#ifndef _POSIX_SOURCE
 #define	LITTLE_ENDIAN	1234	/* LSB first: i386, vax */
 #define	BIG_ENDIAN	4321	/* MSB first: 68000, ibm, net */
 #define	PDP_ENDIAN	3412	/* LSB first in word, MSW first in long */
@@ -58,10 +58,18 @@
 #define	BYTE_ORDER	LITTLE_ENDIAN
 #endif /* !_POSIX_SOURCE */
 
-#ifdef _KERNEL
 #ifdef __GNUC__
 
-#define _BSWAP32_DEFINED
+static __inline __uint64_t
+__bswap64(__uint64_t _x)
+{
+
+	return ((_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 __uint32_t
 __bswap32(__uint32_t __x)
 {
@@ -81,7 +89,6 @@ __bswap32(__uint32_t __x)
 	return (__r);
 }
 
-#define _BSWAP16_DEFINED
 static __inline __uint16_t 
 __bswap16(__uint16_t __x)
 {
@@ -95,8 +102,11 @@ __bswap16(__uint16_t __x)
 	return (__r);
 }
 
-#endif /* _KERNEL */
-
 #endif /* __GNUC__ */
 
+#define	__htonl(x)	__bswap32(x)
+#define	__htons(x)	__bswap16(x)
+#define	__ntohl(x)	__bswap32(x)
+#define	__ntohs(x)	__bswap16(x)
+
 #endif /* !_MACHINE_ENDIAN_H_ */
diff --git a/sys/i386/include/endian.h b/sys/i386/include/endian.h
index 4076422..115511f 100644
--- a/sys/i386/include/endian.h
+++ b/sys/i386/include/endian.h
@@ -46,11 +46,11 @@
 #define	_QUAD_HIGHWORD 1
 #define	_QUAD_LOWWORD 0
 
+#ifndef _POSIX_SOURCE
 /*
  * Definitions for byte order, according to byte significance from low
  * address to high.
  */
-#ifndef _POSIX_SOURCE
 #define	LITTLE_ENDIAN	1234	/* LSB first: i386, vax */
 #define	BIG_ENDIAN	4321	/* MSB first: 68000, ibm, net */
 #define	PDP_ENDIAN	3412	/* LSB first in word, MSW first in long */
@@ -58,35 +58,61 @@
 #define	BYTE_ORDER	LITTLE_ENDIAN
 #endif /* ! _POSIX_SOURCE */
 
-#ifdef _KERNEL
 #ifdef __GNUC__
 
-#define	_BSWAP32_DEFINED
-static __inline __uint32_t
-__bswap32(__uint32_t __x)
-{
+#define __word_swap_int(x) \
+__extension__ ({ register __uint32_t __X = (x); \
+   __asm ("rorl $16, %0" : "+r" (__X)); \
+   __X; })
+
 #if defined(_KERNEL) && (defined(I486_CPU) || defined(I586_CPU) || defined(I686_CPU)) && !defined(I386_CPU)
-	__asm ("bswap %0" : "+r" (__x));
+
+#define __byte_swap_int(x) \
+__extension__ ({ register __uint32_t __X = (x); \
+   __asm ("bswap %0" : "+r" (__X)); \
+   __X; })
 #else
-	__asm ("xchgb %h0, %b0\n\t"
-	       "rorl $16, %0\n\t"
-	       "xchgb %h0, %b0"
-	       : "+q" (__x));
+
+#define __byte_swap_int(x) \
+__extension__ ({ register __uint32_t __X = (x); \
+   __asm ("xchgb %h0, %b0\n\trorl $16, %0\n\txchgb %h0, %b0" \
+       : "+q" (__X)); \
+   __X; })
 #endif
-	return __x;
+
+#define __byte_swap_word(x) \
+__extension__ ({ register __uint16_t __X = (x); \
+   __asm ("xchgb %h0, %b0" : "+q" (__X)); \
+   __X; })
+
+static __inline __uint64_t
+__bswap64(__uint64_t _x)
+{
+	return ((_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)));
 }
 
-#define	_BSWAP16_DEFINED
-static __inline __uint16_t
-__bswap16(__uint16_t __x)
+static __inline __uint32_t
+__bswap32(__uint32_t _x)
 {
-	__asm ("xchgb %h0, %b0" : "+q" (__x));
 
-	return __x;
+	return (__byte_swap_int(_x));
 }
 
-#endif /* _KERNEL */
+static __inline __uint16_t
+__bswap16(__uint16_t _x)
+{
+
+	return (__byte_swap_word(_x));
+}
 
 #endif /* __GNUC__ */
 
+#define	__htonl(x)	__bswap32(x)
+#define	__htons(x)	__bswap16(x)
+#define	__ntohl(x)	__bswap32(x)
+#define	__ntohs(x)	__bswap16(x)
+
 #endif /* !_MACHINE_ENDIAN_H_ */
diff --git a/sys/ia64/include/endian.h b/sys/ia64/include/endian.h
index fe927a2..826fa54 100644
--- a/sys/ia64/include/endian.h
+++ b/sys/ia64/include/endian.h
@@ -47,11 +47,11 @@
 #define _QUAD_HIGHWORD 1
 #define _QUAD_LOWWORD 0
 
+#ifndef _POSIX_SOURCE
 /*
  * Definitions for byte order, according to byte significance from low
  * address to high.
  */
-#ifndef _POSIX_SOURCE
 #define	LITTLE_ENDIAN	1234	/* LSB first: i386, vax */
 #define	BIG_ENDIAN	4321	/* MSB first: 68000, ibm, net */
 #define	PDP_ENDIAN	3412	/* LSB first in word, MSW first in long */
@@ -59,10 +59,8 @@
 #define	BYTE_ORDER	LITTLE_ENDIAN
 #endif /* !_POSIX_SOURCE */
 
-#ifdef _KERNEL
 #ifdef __GNUC__
 
-#define	_BSWAP64_DEFINED
 static __inline __uint64_t
 __bswap64(__uint64_t __x)
 {
@@ -72,7 +70,6 @@ __bswap64(__uint64_t __x)
 	return __r;
 }
 
-#define	_BSWAP32_DEFINED
 static __inline __uint32_t
 __bswap32(__uint32_t __x)
 {
@@ -80,7 +77,6 @@ __bswap32(__uint32_t __x)
 	return (__bswap64(__x) >> 32);
 }
 
-#define	_BSWAP16_DEFINED
 static __inline __uint16_t
 __bswap16(__uint16_t __x)
 {
@@ -88,14 +84,11 @@ __bswap16(__uint16_t __x)
 	return (__bswap64(__x) >> 48);
 }
 
-#else /* !__GNUC__ */
-/* XXX: use the libkern versions for now; these might go away soon. */
-#define	_BSWAP16_DEFINED
-__uint16_t __bswap16(__uint16_t);
-#define	_BSWAP32_DEFINED
-__uint32_t __bswap32(__uint32_t);
 #endif /* __GNUC__ */
 
-#endif /* _KERNEL */
+#define	__htonl(x)	__bswap32(x)
+#define	__htons(x)	__bswap16(x)
+#define	__ntohl(x)	__bswap32(x)
+#define	__ntohs(x)	__bswap16(x)
 
 #endif /* !_MACHINE_ENDIAN_H_ */
diff --git a/sys/netinet/in.h b/sys/netinet/in.h
index cfd55ca..819a827 100644
--- a/sys/netinet/in.h
+++ b/sys/netinet/in.h
@@ -498,6 +498,16 @@ char	*inet_ntoa_r __P((struct in_addr ina, char *buf)); /* in libkern */
 
 #else /* !_KERNEL */
 
+#ifndef _BYTEORDER_PROTOTYPED
+#define	_BYTEORDER_PROTOTYPED
+__BEGIN_DECLS
+__uint32_t	htonl __P((__uint32_t));
+__uint16_t	htons __P((__uint16_t));
+__uint32_t	ntohl __P((__uint32_t));
+__uint16_t	ntohs __P((__uint16_t));
+__END_DECLS
+#endif
+
 #ifndef _BYTEORDER_FUNC_DEFINED
 #define	_BYTEORDER_FUNC_DEFINED
 #define	htonl(x)	__htonl(x)
@@ -506,13 +516,6 @@ char	*inet_ntoa_r __P((struct in_addr ina, char *buf)); /* in libkern */
 #define	ntohs(x)	__ntohs(x)
 #endif
 
-__BEGIN_DECLS
-__uint32_t	htonl __P((__uint32_t));
-__uint16_t	htons __P((__uint16_t));
-__uint32_t	ntohl __P((__uint32_t));
-__uint16_t	ntohs __P((__uint16_t));
-__END_DECLS
-
 #endif /* _KERNEL */
 
 #endif
diff --git a/sys/powerpc/include/endian.h b/sys/powerpc/include/endian.h
index 4acd586..4bccc4e 100644
--- a/sys/powerpc/include/endian.h
+++ b/sys/powerpc/include/endian.h
@@ -58,8 +58,38 @@
 #define	BYTE_ORDER	BIG_ENDIAN
 #endif /* !_POSIX_SOURCE */
 
-#ifndef _KERNEL
-#include <sys/cdefs.h>
-#endif /* _KERNEL */
+#ifdef __GNUC__
+
+static __inline __uint16_t
+__bswap16(__uint16_t _x)
+{
+
+	return ((_x >> 8) | ((_x << 8) & 0xff00));
+}
+
+static __inline __uint32_t
+__bswap32(__uint32_t _x)
+{
+
+	return ((_x >> 24) | ((_x >> 8) & 0xff00) | ((_x << 8) & 0xff0000) |
+	    ((_x << 24) & 0xff000000));
+}
+
+static __inline __uint64_t
+__bswap64(__uint64_t _x)
+{
+
+	return ((_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)));
+}
+
+#endif /* __GNUC__ */
+
+#define	__htonl(x)	((__uint32_t)(x))
+#define	__htons(x)	((__uint16_t)(x))
+#define	__ntohl(x)	((__uint32_t)(x))
+#define	__ntohs(x)	((__uint16_t)(x))
 
 #endif /* !_MACHINE_ENDIAN_H_ */
diff --git a/sys/sparc64/include/endian.h b/sys/sparc64/include/endian.h
index fc51c7e..b3f27d8 100644
--- a/sys/sparc64/include/endian.h
+++ b/sys/sparc64/include/endian.h
@@ -45,11 +45,11 @@
 #define _QUAD_HIGHWORD 0
 #define _QUAD_LOWWORD 1
 
+#ifndef _POSIX_SOURCE
 /*
  * Definitions for byte order, according to byte significance from low
  * address to high.
  */
-#ifndef _POSIX_SOURCE
 #define	LITTLE_ENDIAN	1234	/* LSB first: i386, vax */
 #define	BIG_ENDIAN	4321	/* MSB first: 68000, ibm, net */
 #define	PDP_ENDIAN	3412	/* LSB first in word, MSW first in long */
@@ -57,4 +57,38 @@
 #define	BYTE_ORDER	BIG_ENDIAN
 #endif /* !_POSIX_SOURCE */
 
+#ifdef __GNUC__
+
+static __inline __uint16_t
+__bswap16(__uint16_t _x)
+{
+
+	return ((_x >> 8) | ((_x << 8) & 0xff00));
+}
+
+static __inline __uint32_t
+__bswap32(__uint32_t _x)
+{
+
+	return ((_x >> 24) | ((_x >> 8) & 0xff00) | ((_x << 8) & 0xff0000) |
+	    ((_x << 24) & 0xff000000));
+}
+
+static __inline __uint64_t
+__bswap64(__uint64_t _x)
+{
+
+	return ((_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)));
+}
+
+#endif /* __GNUC__ */
+
+#define	__htonl(x)	((__uint32_t)(x))
+#define	__htons(x)	((__uint16_t)(x))
+#define	__ntohl(x)	((__uint32_t)(x))
+#define	__ntohs(x)	((__uint16_t)(x))
+
 #endif /* !_MACHINE_ENDIAN_H_ */
diff --git a/sys/sys/param.h b/sys/sys/param.h
index d954ed1..6799984 100644
--- a/sys/sys/param.h
+++ b/sys/sys/param.h
@@ -188,43 +188,31 @@
 #define	MAX(a,b) (((a)>(b))?(a):(b))
 #endif
 
-#ifdef _KERNEL
 /*
- * Extended byte order support functions, for kernel use only currently.
- * First, generic implementation of the byte swapping functions for those
- * architectures that do not have optimized variants of each.
+ * Basic byte order function prototypes for non-inline functions.
+ *
+ * XXX temporarily exposed to userland for bogus software.
  */
-#ifndef _BSWAP16_DEFINED
-#define	_BSWAP16_DEFINED
-static __inline __uint16_t
-__bswap16(__uint16_t x)
-{
-	return ((x >> 8) | ((x << 8) & 0xff00U));
-}
+#ifndef _BYTEORDER_PROTOTYPED
+#define	_BYTEORDER_PROTOTYPED
+__BEGIN_DECLS
+__uint32_t	 htonl __P((__uint32_t));
+__uint16_t	 htons __P((__uint16_t));
+__uint32_t	 ntohl __P((__uint32_t));
+__uint16_t	 ntohs __P((__uint16_t));
+__END_DECLS
 #endif
 
-#ifndef _BSWAP32_DEFINED
-#define	_BSWAP32_DEFINED
-static __inline __uint32_t
-__bswap32(__uint32_t x)
-{
-	return ((x >> 24) | ((x >> 8) & 0xff00U) | ((x << 8) & 0xff0000U) |
-	    ((x << 24) & 0xff000000U));
-}
-#endif
-
-#ifndef _BSWAP64_DEFINED
-#define	_BSWAP64_DEFINED
-static __inline __uint64_t
-__bswap64(__uint64_t x)
-{
-	return ((x >> 56) | ((x >> 40) & 0xff00UL) | ((x >> 24) & 0xff0000UL) |
-	    ((x >> 8) & 0xff000000UL) | ((x << 8) & 0xff00000000UL) |
-	    ((x << 24) & 0xff0000000000UL) | ((x << 40) & 0xff000000000000UL) |
-	    ((x << 56)));
-}
-#endif
+/* XXX temporarily exposed to userland for bogus software. */
+#ifndef _BYTEORDER_FUNC_DEFINED
+#define	_BYTEORDER_FUNC_DEFINED
+#define	htonl(x)	__htonl(x)
+#define	htons(x)	__htons(x)
+#define	ntohl(x)	__ntohl(x)
+#define	ntohs(x)	__ntohs(x)
+#endif /* !_BYTEORDER_FUNC_DEFINED */
 
+#ifdef _KERNEL
 #define	bswap16(x)	__bswap16(x)
 #define	bswap32(x)	__bswap32(x)
 #define	bswap64(x)	__bswap64(x)
@@ -259,11 +247,6 @@ __bswap64(__uint64_t x)
 #define	le64toh(x)	bswap64((x))
 #endif /* BYTE_ORDER */
 
-#define	htonl(x)	htobe32((x))
-#define	htons(x)	htobe16((x))
-#define	ntohl(x)	be32toh((x))
-#define	ntohs(x)	be16toh((x))
-
 #endif /* _KERNEL */
 
 /*