The overall design I propose is to keep the hash table global and tag inet_frag_queue with the net. Since the fragments hash is going to be re-sizable, this is OK to keep fragments from different namespace in one hash.
To speedup the evicting process LRU list is made per namespace. As far as the CTL-tuned variables are concerned, the timeout and thresholds are made per namespace, since they have the per namespace sense, but the secret rebuild interval is read-only in sub-namespaces. Since fragment management code is consolidated for ipv4 and ipv6 I make them all in one go. The conntrack_reasm netns-ization is not done - we have to make at least the core netfilter per namespace first, but this reasm code is patched to keep working in the initial namespace. Signed-off-by: Pavel Emelyanov <[EMAIL PROTECTED]> -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
