From: Jouni Malinen (jkmaline_at_cc.hut.fi)
Date: 2002-08-19 17:24:00 UTC
On Sun, Aug 18, 2002 at 08:27:27PM +0200, Ricardo Galli wrote:
> just trying the CVS version with 2.4.18 and 2.4.19. Both -vanilla- kernels
> plus hostap_cs hangs the computer very hard. The screen error is related to
> the "swapper" and "killing an interrupt". OTH, Prism2-2002-05-19 works nicely
> (I know lot of things have changed...).
Do you have an SMP host or uniprocessor setup? It would be helpful, if you could record the oops text and resolve the call trace with ksymoops.
> The only _possible_ pattern I found is that occurs after few minutes of the
> last ethernet package. Before I tried ping floods between de AP and the
> clients and between clients and everything goes smooth, but after few minutes
> the machines dies.
Was the ping directly from a station associated to the AP (i.e., no routing or bridging needed)?
> I am using a Conceptronic card, with WEP enabled and kernel bridging on a
> Digital (Pentium 120) laptop with RedHat 7.2. I've upgraded wireless-tools to
> the rpm provided by RedHat in RawHide (wireless-tools-24-2).
I do not normally use WEP or kernel bridging in my tests. However, now that I enabled both of these, I was able to crash kernel after extended flooding from a station to a host in wired net (i.e., using kernel bridge code). This crash was also in an interrupt handler. It looks like softirq was doing TX and when it freed the packet, something hang.. There might be some sort of race condition in freeing the skb when using bridge code.. There was no Host AP driver functions in the call trace, but this might anyway be because of something the driver did incorrectly. I'll go through the code and check if I can find a reason for this.
-- Jouni Malinen PGP id EFC895FA