[PATCH] knfsd: fix hash function for IP addresses on 64bit little-endian machines.

The hash.h hash_long function, when used on a 64 bit machine, ignores many
of the middle-order bits.  (The prime chosen it too bit-sparse).

IP addresses for clients of an NFS server are very likely to differ only in
the low-order bits.  As addresses are stored in network-byte-order, these
bits become middle-order bits in a little-endian 64bit 'long', and so do
not contribute to the hash.  Thus you can have the situation where all
clients appear on one hash chain.

So, until hash_long is fixed (or maybe forever), us a hash function that
works well on IP addresses - xor the bytes together.

Thanks to "Iozone" <capps@iozone.org> for identifying this problem.

Cc: "Iozone" <capps@iozone.org>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

diff --git a/net/sunrpc/svcauth_unix.c b/net/sunrpc/svcauth_unix.c
index cac2e77..3e6c694 100644 (file)
--- a/net/sunrpc/svcauth_unix.c
+++ b/net/sunrpc/svcauth_unix.c
@@ -101,10 +101,22 @@ static void ip_map_put(struct cache_head *item, struct cache_detail *cd)
        }
 }
 
+#if IP_HASHBITS == 8
+/* hash_long on a 64 bit machine is currently REALLY BAD for
+ * IP addresses in reverse-endian (i.e. on a little-endian machine).
+ * So use a trivial but reliable hash instead
+ */
+static inline int hash_ip(unsigned long ip)
+{
+       int hash = ip ^ (ip>>16);
+       return (hash ^ (hash>>8)) & 0xff;
+}
+#endif
+
 static inline int ip_map_hash(struct ip_map *item)
 {
        return hash_str(item->m_class, IP_HASHBITS) ^ 
-               hash_long((unsigned long)item->m_addr.s_addr, IP_HASHBITS);
+               hash_ip((unsigned long)item->m_addr.s_addr);
 }
 static inline int ip_map_match(struct ip_map *item, struct ip_map *tmp)
 {