kern/73225: [PANIC] Reproducable Page Fault

Paul paul79 at optonline.net
Sun Dec 12 23:10:34 PST 2004


The following reply was made to PR kern/73225; it has been noted by GNATS.

From: Paul <paul79 at optonline.net>
To: Gleb Smirnoff <glebius at cell.sick.ru>
Cc: freebsd-gnats-submit at freebsd.org
Subject: Re: kern/73225: [PANIC] Reproducable Page Fault
Date: Mon, 13 Dec 2004 02:02:07 -0500

 Hi,
 
 I think I may have isolated the problem.  This is a known issue regarding
 SCHED_ULE being broken.  I'm just one of the lucky ones that build everything
 prior to any warning of it being broken.  I believe this was causing the
 problem, because at this time, I am unable to reproduce the problem.  I have
 reverted back to SCHED_4BSD.
 
 For clarity sake, I will go through each of your questions with whatever
 information I have in case this becomes a problem again, once SCHED_ULE is
 fixed.
 
 1) When rebooting the machine, it doesn't matter whether there's traffic
 going through the interfaces while I issue the 'reboot' command, OR if I 
 boot the machine from a cold start.  This happends irregardless.
 
 2) I didn't get to this part to your question, becaues I was unable to
 *currently* reproduce the problem.  Again, I believe this was do to a
 buggy scheduler.
 
 3) I did have a chance to apply your patch, and I have achieve the following
 result with one_pass set to true and *no* queue lines removed from my ruleset:
 
 kernel: ipfw: no next rule
 kernel: ipfw: ouch!, skip past end of rules, denying packet
 
 *Note: this occur over and over and over.
 
 Thanks a Bunch,
 Paul
 
 On Sun, Dec 12, 2004 at 03:44:51PM +0300, Gleb Smirnoff wrote:
 >   Paul,
 > 
 >   I have several questions, to narrow scope of problem search.
 > 
 > 1) Is there active traffic flow when machine reboots?
 > 
 > 2) Is the problem reproducible if you remove all 'queue' rules
 >    from rc.firewall?
 >  If the answer is 'no', can you pls add these rules one by one
 >  and try to reproduce the problem. I mean first try only rule 1000,
 >  then 1000 and 1100, etc..
 > 
 > 3) Can you try to reproduce the problem with this patch. This is
 > not a fix, but it will give some idea on what's going on.
 > 
 > Patch (TABS are lost, apply it by hand):
 > 
 > Index: ip_fw2.c
 > ===================================================================
 > RCS file: /home/ncvs/src/sys/netinet/ip_fw2.c,v
 > retrieving revision 1.85
 > diff -u -r1.85 ip_fw2.c
 > --- ip_fw2.c    10 Dec 2004 02:17:18 -0000      1.85
 > +++ ip_fw2.c    12 Dec 2004 12:38:34 -0000
 > @@ -1909,7 +1909,11 @@
 >  
 >                 f = args->rule->next_rule;
 >                 if (f == NULL)
 > +#if 0
 >                         f = lookup_next_rule(args->rule);
 > +#else
 > +                       print("ipfw: no next rule\n");
 > +#endif
 >         } else {
 >                 /*
 >                  * Find the starting rule. It can be either the first
 >   
 > 
 > -- 
 > Totus tuus, Glebius.
 > GLEBIUS-RIPN GLEB-RIPE
 > 


More information about the freebsd-bugs mailing list