| Message ID | 20091027102251.244ee681@nehalam |
|---|---|
| State | Changes Requested, archived |
| Delegated to: | David Miller |
| Headers | show |
On Tuesday 27 October 2009 19:22:51 you wrote: > The full_name_hash does not produce a value that is evenly distributed > over the lower 8 bits. This causes name hash to be unbalanced with large > number of names. A simple fix is to just fold in the higher bits > with XOR. > > This is independent of possible improvements to full_name_hash() > in future. > I can confirm that the distribution looks good now for our most common cases. Thanks, tavi -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
--- a/net/core/dev.c 2009-10-27 09:21:46.127252547 -0700 +++ b/net/core/dev.c 2009-10-27 09:25:14.593313378 -0700 @@ -199,7 +199,11 @@ EXPORT_SYMBOL(dev_base_lock); static inline struct hlist_head *dev_name_hash(struct net *net, const char *name) { unsigned hash = full_name_hash(name, strnlen(name, IFNAMSIZ)); - return &net->dev_name_head[hash & ((1 << NETDEV_HASHBITS) - 1)]; + + hash ^= (hash >> NETDEV_HASHBITS); + hash &= NETDEV_HASHENTRIES - 1; + + return &net->dev_name_head[hash]; } static inline struct hlist_head *dev_index_hash(struct net *net, int ifindex)
The full_name_hash does not produce a value that is evenly distributed over the lower 8 bits. This causes name hash to be unbalanced with large number of names. A simple fix is to just fold in the higher bits with XOR. This is independent of possible improvements to full_name_hash() in future. Signed-off-by: Stephen Hemminger <shemminger@vyatta.com> -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html