Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 29 May 2009 11:59:01 +0100
From:      Bruce Simpson <bms@incunabulum.net>
To:        deeptech71@gmail.com
Cc:        freebsd-current@freebsd.org
Subject:   Re: panic: igmp_v3_dispatch_general_query: called when version 2
Message-ID:  <4A1FBFF5.6090103@incunabulum.net>
In-Reply-To: <4A1EC8FB.6090206@gmail.com>
References:  <4A1DD57A.7010704@gmail.com> <4A1E9831.4010606@incunabulum.net> <4A1EC8FB.6090206@gmail.com>

next in thread | previous in thread | raw e-mail | index | archive | help
deeptech71@gmail.com wrote:
> Bruce Simpson wrote:
>   
>> Are you running a multicast router of any kind on your network?
>>     
>
> I have no idea what IGMP/multicast is, and have no router whatsoever.
>   

No matter. I think I know where the problem is -- as I fixed a similiar
issue in MLDv2 this week before the final round of fixes were committed
for that feature. The RFCs for SSM are somewhat ambiguously worded about
the treatment of queries from legacy routers. The backtrace you posted
points towards this.

I may not get free time to fix this right away, are you able to test a
patch when I can make one available?

It would be helpful if you could confirm, with a tcpdump capture forked
from single user mode, if your FreeBSD host is receiving an IGMPv2 query
when this problem is triggered.

I had an automated test suite crafted around QEMU to catch problems like
this, however, perhaps due to the TTY refactoring, I have been unable to
get QEMU virtualized FreeBSD 8.x images to respond on the emulated
serial port, which blew away my test suite functionality. FreeBSD 7.x
does not have this issue.

> Well... I'm using FreeBSD-current as a desktop system for gaming and
> C/C++ development. The system is receiving a DHCP address (not a
> reserved one like 10.x.x.x) in a university's local network with lots of
> computers mostly tossing around Windows NetBIOS packets. If that matters.
>   

It sounds like your university LAN may well have an IGMPv2 capable
multicast router present on the network. If you could double confirm
this, via packet capture, that would be great.

f there is anything unusual about the format of the IGMPv2 queries from
your university's vendor equipment, that would be good to know too.

> Sounds like that is all useless, so I'd send the vmcore. Who wants it?
>   

Unfortunately, just posting backtraces or vmcores does not always
constitute a useful bug report in of itself :-) Although I appreciate
the effort you have gone to in making such available...

thanks
BMS



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4A1FBFF5.6090103>