Re: 2.6.2 crash after network link failure

[Petr Vandrovec <vandrove@xxxxxxxxxx>]

On Mon, Feb 09, 2004 at 03:37:48PM -0800, Feldman, Scott wrote:
> > I think what might be happening is that somehow the TX queue 
> > is corrupted if
> > e100_config() runs (due to link UP state change) while there 
> > are active normal SKB packets on the TX queue.  Or perhaps 
> > some TX queue handling locking issue.
> > 
> > Scott, any ideas?
> 
> e100 hardware will continue to process the hardware's Tx queue even
> after link is lost, and then cleanup (return skbs) on interrupt.  I
> would expect e100 to be holding no Tx skbs when link returned.
> 
> Petr, -mm kernel has an updated (and much simpler) e100 driver.  Is this
> something you can try?  The switch failure can be simulated by manually
> plugging the cable in/out.

Unfortunately it does not seem easily triggerable. I spent about one
hour plugging/unplugging cable while transmitting UDP packets as fast
as possible, sometime interleaved with 'mii-tool -r', and it refused
to crash...
                                                Petr Vandrovec