On 13.12.2006 Stephen Hemminger wrote: > On Tue, 12 Dec 2006 Moire <sec at shee.org> wrote: > >> ... >> Scenario: >> >> Every now and then i get this urgent warning: >> >> 1 Time(s): Dead loop on virtual device br0, fix it urgently! >> >> Stephen Hemminger wrote (16.08.2006 20:09) this: >>> It looks like a bridge ended up including itself some how. >> >> Its for me unclear when this happens and what trigger this. >> >> I would appreciate any info to some resources. For example >> a technical document covering this issue or similar. >> >> My setup is a kernel 2.4.31 with bridge-utils 1.0.4. The box >> have three network interfaces eth[0-2]. Bridged interfaces >> are eth[1-2] (router,dmz), eth0 is my intranet. >> I use DLINK network cards with sk98lin driver. > > Could it be the fact that sk98lin driver has some code in it that > sends spanning tree packets itself! This crap is because the driver > has some vendor specific code to do failover on the dual port versions > of the driver. Since SysKonnect was big in the FDDI world, I suspect > they wanted the same management features in their gigabit cards. > > You have some choices: > 1) use a different card > 2) comment out the code in the sk98lin driver that generates and > receives > the management frames > 3) use 2.6 and the skge driver that works on the same hardware. > 4) backport the skge driver to 2.4. Hello, thanks for your reply. I noticed that you provide a backport of the skge driver. I will test it and do report if it is successful. I hope that it is possible to build it into the kernel because i don't use a modularized kernel. Thanks C. Moire
