[mythtv] V4L2 bug ??

[Isaac Richards]

On Wednesday 02 February 2005 05:53 pm, Mike Isely wrote:
> This *might* be entirely unrelated, but in early January there was a bug
> in the code which caused the *front end* to crash on channel changing.
> The bug was due to uninitialized data.  One characteristic was that if you
> successfully flip channels once - if the uninitialized fields happened to
> have been "uninitialized the right way" - then it was fine from that point
> forward until you restarted the front end again (and thus hit the
> uninitialized data again).
>
> This is described in bug #181.
>
> The situation discussed in this thread here doesn't very well match the
> problem solved in bug #181.  But that problem was so rampant for me at the
> the time that it basically rendered my mythbox useless; surely others must
> have at least seen this.  So it may still be worth reviewing that bug
> description in light of the discussion above.

Naw, I think it's unrelated - what I was seeing were races in the driver 
itself.  But this _was_ fairly long ago, with early v4l2 stuff patched into 
2.4.x.  I don't have a problem enabling v4l2 for everything that says 
supports it after the 0.17 release, but not before then.

Isaac