Re: ATI TV Wonder regression since at least 3.19.6

[Hans Verkuil <hverkuil@xs4all.nl>]

On 05/15/2015 05:12 PM, Patrice Levesque wrote:
> 
> Hi Hans,
> 
> 
>>> Function isn't used; when compiling I get:
>> That makes no sense. This function is most definitely used.
> 
> Idiot guy here did not follow simple instructions and didn't patch the
> right kernel source.  He just did, and function is used.
> 
> 
>> Did you start a capturing video first before running dmesg? I want to
>> see if capturing video will generate messages in dmesg.
> 
> Sending you again my (truncated) dmesg, but here's the annotated salient bit:
> 
> Starting video capture first time:
> [Fri May 15 11:01:43 2015] restart_video_queue
> [Fri May 15 11:01:44 2015] restart_video_queue
> [Fri May 15 11:01:55 2015] restart_video_queue
> [Fri May 15 11:01:56 2015] restart_video_queue
> [Fri May 15 11:01:56 2015] restart_video_queue
> [Fri May 15 11:02:00 2015] restart_video_queue
> [Fri May 15 11:02:05 2015] restart_video_queue
> [Fri May 15 11:02:06 2015] restart_video_queue
> [Fri May 15 11:02:06 2015] restart_video_queue
> [Fri May 15 11:02:06 2015] restart_video_queue
> [Fri May 15 11:02:07 2015] restart_video_queue
> [Fri May 15 11:02:07 2015] restart_video_queue
> [Fri May 15 11:02:07 2015] restart_video_queue
> [Fri May 15 11:02:09 2015] restart_video_queue
> Stopping video capture:
> [Fri May 15 11:03:26 2015] restart_video_queue
> Re-Starting video capture:
> [Fri May 15 11:03:40 2015] restart_video_queue
> Stopping video capture:
> [Fri May 15 11:04:18 2015] restart_video_queue
> 
> Changing channels didn't provoke restart_video_queue events.

FYI: I've bought this card on ebay and I am waiting for it to arrive.
Hopefully I can reproduce it and then I'll fix it. It's a fair bit of
work, unfortunately.

The fact that this function is called means that the DMA stalls every
so often, and I'd like to know why that happens.

Regards,

	Hans