Actually one problem I see with this patch, and it's a recent change which I'm still trying to figure out the reasons for needing. There seems to be a problem running the pvr500/150 over extended periods of time for audio, will just stop working eventually, after a day or so. It seems that if you switch inputs, or reinit the input, things begin working once again. So it seems like each time we start an encoding we need to redo that input register setup, makes sense since each input init the chip autodetects the audio standard and so on. I think this patch, although usually would be a good thing, makes it not happen which my last 0.3.5j actually made it force an input init each capture startup. So something for people to see if they too have this problem, if the last version fixes it, and what should we do for this to work properly, I'm still looking closer at this to see if your patch actually undoes that forced input init, just maybe remove the part which checks if it's already the input, or maybe what I did won't change with this patch and it's fine (I think I void the input to -1, so more I think about it, more it seems like your patch should be just fine). Let me know what you think, if you've had this odd audio drop out problem too, or anyone else seeing it and thoughts on what we can do, or if I've hopefully fixed it already :-).
Thanks, Chris On Fri, May 27, 2005 at 02:29:35PM -0400, Tyler Trafford wrote: > Straight forward patch, does the following: > > * During initialization, instead of iteratively copying the array > values for cx25840_input_layout, just the set a pointer to the proper > array. > > * In DECODER_SET_INPUT, assign state->input after performing operation > successfully. Also tried to make debug messages more like the > saa7115. > > * Move log_status() call to bottom of ioctl command function, since > it's way more useful _after_ we've done something than before. > > * Rearrange DECODER_SET_AUDIO* so that they only return on an error, > just like the others. > > * No reason for functions only called from a single place not to be > 'inline', so I made them so. > > -- > Tyler Trafford -- --- Chris Kennedy / [EMAIL PROTECTED] Engineer KMOS-TV/KTBG-FM Broadcasting Services Department Central Missouri State University ------------------------------------------------------- This SF.Net email is sponsored by Yahoo. Introducing Yahoo! Search Developer Network - Create apps using Yahoo! Search APIs Find out how you can build Yahoo! directly into your own Applications - visit http://developer.yahoo.net/?fr=offad-ysdn-ostg-q22005 _______________________________________________ ivtv-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/ivtv-devel
