Re: Emu10k1x driver

[Takashi Iwai <tiwai@suse.de>]

At Sat, 15 May 2004 17:30:52 -0400,
Francisco Moraes wrote:
> 
> Takashi Iwai wrote:
> 
> >>Is your patch for kernel 2.6?
> >>    
> >>
> >
> >to the latest ALSA cvs.
> >you'd better to get the CVS tree (either via cvs or snapshot).
> >it already incldues emu10k1x.c.
> >  
> >
> I got it from there but your changes to add the module params don't 
> compile on my system (Red Hat Fedora Core 1 - 2.4.22-2188 kernel).

did you rebuild via cvscompile script?


> There is also a problem with mono streams, so the hw params need to 
> indicate that only stereo is supported.

that's fine.  many hardwares support only stereo indeed.

>  Also, is there any advantage in 
> enabling both interrupts in the buffer (loop and half loop)?

yes, definitely.  (IIRC, it IS enabled?)
otherwise the app won't work proprely in most cases.
remember that the sound system is driven by interrupts.
an irq per buffer means that we have only one chance to update the
data in the accurate timinig.

>  Also, how 
> important is the definition of period bytes? It seems it should be the 
> same as the buffer size.

no, it should be a half of the buffer size.

>  Maybe you can just explain it to me how it 
> relates to the total buffer size being used.

the buffer size is the total size of the memory space you callocate.
when the amount of processed PCM data becomes to the period size, an
irq is issued.  in the case of emu10k1x, hence, period size is ALWAYS
buffer size / 2 (as long as half-interrupt bit is set).


Takashi


-------------------------------------------------------
This SF.Net email is sponsored by: SourceForge.net Broadband
Sign-up now for SourceForge Broadband and get the fastest
6.0/768 connection for only $19.95/mo for the first 3 months!
http://ads.osdn.com/?ad_id=2562&alloc_id=6184&op=click