[mythtv] Patch: Big audio code update (please test)

[Jesper Sörensen]

>>> I'm currently running Alsa driver 1.0.4 (included in Linux 2.6.8) 
>>> and the Debian Alsa libs/utils which are currently at 1.0.6[a]. The 
>>> card is a M-Audio Revolution 7.1 (ice1724). I'm on a DVB sat feed 
>>> here so the audio comes in different formats on different channels 
>>> (and even changes between programmes), and I guess those of you 
>>> using analog TV cards always get the same audio format from the card?
>>
>>
> I am on DVB, and 2.6.7.  I'm sure that 2.6.7 has alsa 1.0.5, so you 
> must be more up to date than that?  Try as David says and get a 
> matching set of alsa libs and drivers.  M-Audio is usually quite well 
> supported?


Yes, I had some initial troubles with the Revo but now it's rocking 
pretty good!

I'm actually not exactly sure which Alsa version is in my kernel, but 
this is what I get:
$ cat /proc/asound/version
Advanced Linux Sound Architecture Driver Version 1.0.4 (Mon May 17 
14:31:44 2004 UTC).
Compiled on Oct  8 2004 for kernel 2.6.8-9-htpc.

>
>>>> The alsa support is likely to be driver problems rather than the 
>>>> myth alsa code to be honest.  But we can try to work around stuff 
>>>> here.  Does mplayer work ok for you via Alsa?
>>>
>>>
>>>
>>> Yes, Alsa works with mplayer and xine. But as I said, I don't think 
>>> my Alsa problems are caused by your patch since it has never been 
>>> stable here. The OSS output works great though (I'm using Alsa's OSS 
>>> emu, with digital out) so I'm not sitting on a sinking ship here... 
>>
>>
>
> Sure.  My only thought here was that if the mplayer code works then we 
> can try to work out what's failing for you.  The mplayer alsa code is 
> quite straightforward and could be easily nicked for myth, or at least 
> compared for differences.  There is nothing too fancy going on here at 
> all, it's all kind of set_rate, set_buffer, off we go kind of stuff.


FYI, I've found the bug causing the crash, but now I've found a new bug 
instead. This one doesn't crash Myth but it kills the audio device (and 
thereby probably triggers the crashing bug). Anyway, I'll hold off on 
the patch and won't investigate this any further until I've seen David's 
patch. No need to do double work and I guess there's a fair chance David 
has already fixed this...

/JS