Re: [ath5k-devel] ath5k_tasklet_rx BUG_ON(bf->skb == NULL)

[Hugh Dickins <hugh@veritas.com>]

On Fri, 9 Jan 2009, Bob Copeland wrote:
> On Thu, Jan 8, 2009 at 1:41 PM, Bob Copeland <bcopeland@gmail.com> wrote:
> > On Thu, Jan 8, 2009 at 12:55 PM, Hugh Dickins <hugh@veritas.com> wrote:
> >> Let me see if I can reproduce the ath5k BUG with a straightforward
> >> memhog, repeatedly dirtying more anon memory than RAM can provide.
> >> Is there something suitable I could run to exercise that wireless
> >> path concurrently?  It was just idling when I hit the BUGs before.
> >
> > Receiving any packets will do it, in your case it was probably
> > beacons from nearby APs if you weren't using the wifi.  So iperf
> > or similar will work.
> 
> Actually it's probably enough to just intentionally fail one
> of the allocations in the driver.  I'll do that and let you know
> if I can easily break it.

Good, that should be a lot quicker.

My overnight try with simple memhog, and iperf -s receiving from
another machine doing iperf -c at intervals, didn't hit it (but my
more involved test wouldn't necessarily have hit it in that time either).

Currently trying just a 2.6.28 defconfig kernel build in tmpfs
(other filesystems should get their dirty pages pdflushed sooner)
along with the iperf, but haven't got the sizes suitably tuned yet.

Hugh