[linux-dvb] Re: m920x crash

Aapo Tahkola aet at rasterburn.org
Sat Sep 30 20:53:16 CEST 2006


On Thu, 28 Sep 2006 23:31:17 +0200 (CEST)
Patrick Boettcher <patrick.boettcher at desy.de> wrote:

> On Thu, 28 Sep 2006, Michael Krufky wrote:
> 
> > Martin Schwier wrote:
> > > Hi!
> > 
> > Martin, I hope you dont mind that I added cc to the linux-dvb list.
> > 
> > I pushed up a few patches about an hour ago, tip changeset is:
> > 
> > c0f7293d1ac6 - m920x: move qt1010_tuner_attach function into
> > qt1010.h
> > 
> > if you do, 'hg heads', is this the revision that you're using?  If
> > not, please update and try again.
> > 
> > > 
> > > I'm not a developer but I just tried out your driver
> > > (http://linuxtv.org/hg/~mkrufky/m920x) with my MSI Megasky 580 on
> > > linux 2.6.18 and i got this:
> > 
> > Aapo, I just pushed up some more cleanup patches... Is the problem
> > described below my fault?  Can you test the current tree on your
> > device and reproduce this error?
> > 
> > Patrick has given his Ack on the new dvb-usb module... I was just
> > about to email you asking to test the tree before I request a merge
> > to Mauro, but then I got this email from Martin.  Since I don't
> > have this device, myself, I dont have any way to test it.
> > 
> > Can you look into this, Aapo?
> 
> I looked at it:
> 
> line 165 in m920x.c is accessing the adapter[0]. Because
> frontend_attach is called (which again is using i2c to identify
> zarlink) during/before the initialization of the adapter.

Right, accessing adapter[0] does not work. That's not why it's oopsing
though.

> The if in 165 is very ugly. (What is the meaning of the value
> assigned to w_len? I2C clock/speed? I2C port?).

You cannot change or leave them out as far as I can remember. I would
assume it's some sort of read wait time as it doesn't appear with
writes.

-- 
Aapo Tahkola



More information about the linux-dvb mailing list