Fall back to naive built-in byte-swap routines

EL5 lacks htobe64 and friends.  Using its byteswap.h requires an endianness
check in our headers, which is fragile and compiler-dependent.  It's a fair
amount of code [1], and is only useful on EL5.

So instead, let's include a fallback implementation of these functions, and use
it whenever we can't find the (now-)standard Linux or OS X routines.  The
fallback is endianness-independent and should work on any platform.

[1] https://github.com/rurban/mosh/commit/da1a5abd1e09f0c1e9295357bb480a643d9dc8e3

configure.ac

@@ -272,7 +272,7 @@ AC_CHECK_DECL([htobe64],
   [AC_CHECK_DECL([OSSwapHostToBigInt64],
      [AC_DEFINE([HAVE_OSX_SWAP], [1],
         [Define if OSSwapHostToBigInt64 and friends exist.])],
-     [AC_MSG_ERROR([Unable to find byte swapping functions])],
+     [AC_MSG_WARN([Unable to find byte swapping functions; using built-in routines.])],
      [[#include <libkern/OSByteOrder.h>]])],
   [[#if defined(HAVE_ENDIAN_H)
 #include <endian.h>

src/crypto/byteorder.h

@@ -27,12 +27,84 @@
 # elif defined(HAVE_SYS_ENDIAN_H)
 #  include <sys/endian.h>
 # endif
+
 #elif HAVE_OSX_SWAP
 # include <libkern/OSByteOrder.h>
 # define htobe64 OSSwapHostToBigInt64
 # define be64toh OSSwapBigToHostInt64
 # define htobe16 OSSwapHostToBigInt16
 # define be16toh OSSwapBigToHostInt16
+
+#else
+
+/* Use our fallback implementation, which is correct for any endianness. */
+
+#include <stdint.h>
+
+/* Make sure they aren't macros */
+#undef htobe64
+#undef be64toh
+#undef htobe16
+#undef be16toh
+
+/* Use unions rather than casts, to comply with strict aliasing rules. */
+
+inline uint64_t htobe64( uint64_t x ) {
+  uint8_t xs[ 8 ] = {
+    ( x >> 56 ) & 0xFF,
+    ( x >> 48 ) & 0xFF,
+    ( x >> 40 ) & 0xFF,
+    ( x >> 32 ) & 0xFF,
+    ( x >> 24 ) & 0xFF,
+    ( x >> 16 ) & 0xFF,
+    ( x >>  8 ) & 0xFF,
+      x         & 0xFF };
+  union {
+    const uint8_t  *p8;
+    const uint64_t *p64;
+  } u;
+  u.p8 = xs;
+  return *u.p64;
+}
+
+inline uint64_t be64toh( uint64_t x ) {
+  union {
+    const uint8_t  *p8;
+    const uint64_t *p64;
+  } u;
+  u.p64 = &x;
+  return ( uint64_t( u.p8[ 0 ] ) << 56 )
+       | ( uint64_t( u.p8[ 1 ] ) << 48 )
+       | ( uint64_t( u.p8[ 2 ] ) << 40 )
+       | ( uint64_t( u.p8[ 3 ] ) << 32 )
+       | ( uint64_t( u.p8[ 4 ] ) << 24 )
+       | ( uint64_t( u.p8[ 5 ] ) << 16 )
+       | ( uint64_t( u.p8[ 6 ] ) <<  8 )
+       | ( uint64_t( u.p8[ 7 ] ) );
+}
+
+inline uint16_t htobe16( uint16_t x ) {
+  uint8_t xs[ 2 ] = {
+    ( x >> 8 ) & 0xFF,
+      x        & 0xFF };
+  union {
+    const uint8_t  *p8;
+    const uint16_t *p16;
+  } u;
+  u.p8 = xs;
+  return *u.p16;
+}
+
+inline uint16_t be16toh( uint16_t x ) {
+  union {
+    const uint8_t  *p8;
+    const uint16_t *p16;
+  } u;
+  u.p16 = &x;
+  return ( uint16_t( u.p8[ 0 ] ) << 8 )
+       | ( uint16_t( u.p8[ 1 ] ) );
+}
+
 #endif
 
 #endif