Re: [Alsa-devel] Re: [2.6 patch] schedule obsolete OSS drivers for removal

[James Courtier-Dutton <James@superbug.demon.co.uk>]

Thorsten Knabe wrote:

> On Mon, 1 Aug 2005, Andrew Haninger wrote:
>
> > Thorsten: Please remember to include the list(s) when emailing those
> > links/numbers. I'd like to be able to watch it, too, and add any
> > information that I can, rather than entering a duplicate bug.
>
>
> Hello.
>
> I have taken a closer look at the ALSA AD1816 sound driver during the 
> last weekend. Here are my findings:
>
> On vanilla Linux 2.6.12.3 and 2.6.13-rc4 modprobe hangs in D-state 
> when loading the snd-ad1816a module. No messages have been logged to 
> the syslog and the system is otherwise stable. Of course the sound 
> card is unusable.
> On Linux 2.6.8 (as shipped with current Debian Sarge), vanilla Linux 
> 2.6.10 and Linux 2.6.11.12 the module loads fine.
>
> I have done some tests with xmms(Debian), kphone(VoIP-Phone/Debian) 
> and iaxcomm(VoIP-Phone/self-made). Audio playback with xmms is always 
> fine using either ALSA or OSS emulation. Using OSS emulation with one 
> of the VoIP phones, playback and recording stop a few seconds after 
> the call is started. Using the ALSA interface with kphone works, but 
> there is a continuous clicking approximately 3 times per second. Also 
> audio latency is poor compared to the OSS driver. iaxcomm does not 
> support the ALSA audio interface, thus no problems here. :-)
> The native OSS driver is fine on all kernels with all tested 
> applications.
>
> Also the ALSA driver does not have an equivalent for the 
> "ad1816_clockfreq" option of the OSS driver. The AD1816 chip requires 
> a 33MHz reference clock, however some cards use a different (mostly 
> 32.125MHz) clock, thus the audio sample rate has to be corrected 
> before it is written to the hardware registers for proper playback and 
> recording speed.
>
> I have not filed any bug reports to the ALSA bug tracking system so 
> far, but will do so tomorrow and add the corresponding bug numbers to 
> this thread.
>
> Thorsten
It sounds to me that the best way to fix this is either:
a) Detect sound card subversion number and select different clock based 
on that.
b) Some how auto detect the clock, much like the intel8x0 driver does.
c) Provide a manual option like the OSS driver. (We should probably have 
this as well as (a) for the cases where (a) does not know about 
particular soundcard X yet.

James

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/