Merge 1.127 from OpenBSD, that closes a regression from 1.125 (merged

as r242694):
  do better detection of when we have a better version of the tcp sequence
  windows than our peer.

  this resolves the last of the pfsync traffic storm issues ive been able to
  produce, and therefore makes it possible to do usable active-active
  statuful firewalls with pf.
This commit is contained in:
glebius 2012-12-11 08:37:08 +00:00
parent 778f6b756b
commit 6ce9102001

View File

@ -47,7 +47,7 @@
* 1.118, 1.124, 1.148, 1.149, 1.151, 1.171 - fixes to bulk updates
* 1.120, 1.175 - use monotonic time_uptime
* 1.122 - reduce number of updates for non-TCP sessions
* 1.125 - rewrite merge or stale processing
* 1.125, 1.127 - rewrite merge or stale processing
* 1.128 - cleanups
* 1.146 - bzero() mbuf before sparsely filling it with data
* 1.170 - SIOCSIFMTU checks
@ -787,12 +787,15 @@ pfsync_upd_tcp(struct pf_state *st, struct pfsync_state_peer *src,
if ((st->src.state > src->state &&
(st->src.state < PF_TCPS_PROXY_SRC ||
src->state >= PF_TCPS_PROXY_SRC)) ||
SEQ_GT(st->src.seqlo, ntohl(src->seqlo)))
(st->src.state == src->state &&
SEQ_GT(st->src.seqlo, ntohl(src->seqlo))))
sync++;
else
pf_state_peer_ntoh(src, &st->src);
if (st->dst.state > dst->state ||
if ((st->dst.state > dst->state) ||
(st->dst.state >= TCPS_SYN_SENT &&
SEQ_GT(st->dst.seqlo, ntohl(dst->seqlo))))
sync++;