[PATCH] if_sk(4) rx/tx "hangs"

[David Kleiner]

Bjoern,

Thank you for the update.  I re-sync'ed to the latest -HEAD, re-added
sk to the kernel statically and it's been running for 5 hours fine,
although it "feels" a bit sluggish over the network.  Will see how it
goes.

Cheers,

David


On Tue, 16 Nov 2004 18:55:35 +0000 (UTC), Bjoern A. Zeeb
<bzeeb-lists at lists.zabbadoz.net> wrote:
> On Mon, 15 Nov 2004, David Kleiner wrote:
> 
> Hi,
> 
> > Here is what I get when trying to load up the if_sk.ko module:
> >
> ...
> > Nov 15 15:31:58 macropus kernel: sk0: no memory for jumbo buffers!
> > Nov 15 15:31:58 macropus kernel: sk0: jumbo buffer allocation failed
> > Nov 15 15:31:58 macropus kernel: device_attach: sk0 attach returned 12
> 
> I had seen this too some weeks ago when loading and unloading the
> if_sk.ko repeatedly but I haven't been able to reproduce afterwards.
> 
> > This is an Asus "A8V Deluxe (v. 2.0) K8T800 No-WiFi/A" and 6.0-CURRENT
> > built on Sat Nov 13 12:08:59 PST 2004.
> >
> > I also had it crashed while running a 700MB bittorrent transfer (stock
> > py-bittorrent) but did not get a crash dump.  The stack trace pointed
> > at problem with networking and calling "panic" at the db> prompt
> > wedged it solid.
> 
> you may perhaps want to update to lastest version which adds some
> locking in the jumbo buffer memory functions. This should hopefully
> help you.
> 
> --
> 
> 
> Bjoern A. Zeeb                          bzeeb at Zabbadoz dot NeT
>