| From: | Max Laier <max@xxxxxxxxxxxxxx> |
| Date: | Tue, 4 Jul 2006 18:28:24 +0200 |
On Sunday 02 July 2006 17:17, Simon 'corecode' Schubert wrote: > about every month i'm getting a panic in pf_purge_expired_states on some > RB_* function. > > I talked to dhartmei and he said that there are no problems known to > this code, except for reentrancy issues. > > So I wonder: how can the RB-tree be destroyed? Probably some race > between the timer and the network. But can that be? The timer runs in > a critical section, and shouldn't the network too? > > several crash dumps available on request. Can you show me a trace instead? > and clues? We had rb-tree inconsistencies in FreeBSD as well. I think the basic behind the fix was pf.c, rev. 1.35 but it took me some intermediate steps to get it final. -- /"\ Best regards, | mlaier@xxxxxxxxxxx \ / Max Laier | ICQ #67774661 X http://pf4freebsd.love2party.net/ | mlaier@EFnet / \ ASCII Ribbon Campaign | Against HTML Mail and News
Attachment:
pgp00000.pgp
Description: PGP signature