Re: [PATCH] Add support for sq905 based cameras to gspca

[Jean-Francois Moine <moinejf@free.fr>]

On Tue, 3 Feb 2009 12:21:55 -0600 (CST)
kilgota@banach.math.auburn.edu wrote:

> I should add to the above that now I have tested, and indeed this
> change does not solve the problem of kernel oops after disconnect
> while streaming. It does make one change. The xterm does not go wild
> with error messages. But it is still not possible to close the svv
> window. Moreover, ps ax reveals that [svv] is running as an
> unkillable process, with [sq905/0] and [sq905/1], equally unkillable,
> in supporting roles. And dmesg reveals an oops. The problem is after
> all notorious by now, so I do not see much need for yet another log
> of debug output unless specifically asked for such.

Why is there 2 sq905 processes?

What is the oops? (Your last trace was good, because it gave the last
gspca/sq905 messages and the full oops)

> Perhaps we also need to do what Adam suggested yesterday, and add a
> call to destroy_urbs() in gspca_disconnect()?

Surely not! The destroy_urbs() must be called at the right time, i.e.
on close().

-- 
Ken ar c'hentan	|	      ** Breizh ha Linux atav! **
Jef		|		http://moinejf.free.fr/