On Tue, 27 Apr 2004, Max Laier wrote: > Date: Tue, 27 Apr 2004 22:33:49 +0000 > From: Max Laier <max_at_love2party.net> > To: freebsd-current_at_freebsd.org > Cc: Michael Reifenberger <mike_at_Reifenberger.com> > Subject: Re: panics using pf > > On Tuesday 27 April 2004 01:04, Michael Reifenberger wrote: > > Hi, > > I get the attached panics after some time ( max. 1 hour )when using pf(4). > > I do not get them, when using ipf(4). > > > > Any clues? > > > #6 0xc0641d92 in nd6_slowtimo (ignored_arg=0x0) > > at ../../../netinet6/nd6.c:1800 > > nd6if = (struct nd_ifinfo *) 0xc43221c0 > > ifp = (struct ifnet *) 0xc4529c00 > > Okay ... this indicads that ifp->if_afdata[AF_INET6] is uninitialized, which > should not be the case when the slowtimo runs. As you told me off-list > (before I realized that you sent it to -current as well) this happens for the > pflog0 interface. Still that should not happen and I wonder if it's got to do > with the lately reported problems with a cleanup commit to nd6.c. > > So the question is, are you seeing this with nd6.c, rev. 1.42? Does upping to > 1.43 help? Does rolling back to April, 24th help? > > For the record, on my system (week-old -currnet) pflog0->if_afdata[AF_INET6] > if initialized okay. > I've seen panics with pf(4) for some weeks (see -current's archive for one report from me). But at that time I couldn't get a gdb backtrace because of a bug in the binutils... The last reports where with -current sources of that time. Should be the latest rev of nd6.c BTW: Disabling 'options INET6' fixes the panics as well as switching back to ipf(4), so I'm using pf(4) now since this morning... Bye/2 --- Michael Reifenberger, Business Development Manager SAP-Basis, Plaut Consulting Comp: Michael.Reifenberger_at_plaut.de | Priv: Michael_at_Reifenberger.com http://www.plaut.de | http://www.Reifenberger.comReceived on Tue Apr 27 2004 - 15:57:04 UTC
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:37:52 UTC