Re: e100 REGRESSION in 2.6.32

[Roger Oksanen]

On Wednesday 16 December 2009 21:50:38 Alan Stern wrote:
> On Wed, 16 Dec 2009, Brandeburg, Jesse wrote:
> >  Alan thanks for the report, I wonder if 
> > you can revert 98468efddb101f8a29af974101c17ba513b07be1 and see if that 
> > fixes it?
> 
> It does indeed.  Thanks for the quick response.
> 
> > Roger on the To: line made that change, which may have broken e100 for 
> > your system.
> 
> It looks simple enough, but obviously something is wrong with it.

I'm looking into this. Quite a baffling problem though. I have no idea what
causes the last 4 bytes  to change slightly (6c 6b 6b a3). Perhaps the patch 
only changed the timings to uncloak a lurking bug.. 
struct sk_buff has users as it's last member, so perhaps there is a preempt
point in dmapool that triggers a race with the rx path.

I'll try to reproduce this..

-- 
Roger Oksanen <roger.oksanen@xxxxxxxxxxxxxx>
http://www.cs.helsinki.fi/u/raoksane
+358 50 355 1990
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/