On Jan 28, 2008 7:18 AM, Jeff Chua <[EMAIL PROTECTED]> wrote:
I'm sending printing jobs to a network printer (it's actually printing
to the localhost simply creating a file), and running this on
Linux-2.6.24 will cause the printing to slow down to 1 print every 3
seconds after printing 500 times.
I bisected the kernel since the last known good 2.6.23 and zeroed in to
this commit.
commit 17311393f969090ab060540bd9dbe7dc885a76d5
Author: Jozsef Kadlecsik <[EMAIL PROTECTED]>
Date: Thu Oct 11 14:35:52 2007 -0700
[NETFILTER]: nf_conntrack_tcp: fix connection reopening
Reverting this commit solves the problem.
Version 1000 jobs
2.6.23 90 sec
2.6.24 1,492 sec <== with commit
2.6.24(patch) 90 sec <== reverted the commit
Since the code has changed an can't simply revert the commit for the
latest linux git download, I made a patch to revert the above commit. I've
tested and it's working, but don't know if it breaks other things. Please
review.
Thanks,
Jeff.
--- net/netfilter/nf_conntrack_proto_tcp.c.org 2008-01-29 03:34:43 +0800
+++ net/netfilter/nf_conntrack_proto_tcp.c 2008-01-29 03:36:26 +0800
@@ -836,8 +836,8 @@
break;
if ((conntrack->proto.tcp.seen[!dir].flags &
IP_CT_TCP_FLAG_CLOSE_INIT)
- || (conntrack->proto.tcp.last_dir == dir
- && conntrack->proto.tcp.last_index == TCP_RST_SET)) {
+ || after(ntohl(th->seq),
+ conntrack->proto.tcp.seen[dir].td_end)) {
/* Attempt to reopen a closed/aborted connection.
* Delete this connection and look up again. */
write_unlock_bh(&tcp_lock);
@@ -845,6 +845,12 @@
conntrack->timeout.function((unsigned long)
conntrack);
return -NF_REPEAT;
+ } else {
+ write_unlock_bh(&tcp_lock);
+ if (LOG_INVALID(IPPROTO_TCP))
+ nf_log_packet(pf, 0, skb, NULL, NULL,
+ NULL, "nf_ct_tcp: invalid SYN");
+ return -NF_ACCEPT;
}
/* Fall through */
case TCP_CONNTRACK_IGNORE:
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
