Re: ALSA snd_xferi_t buffer arrangment

[Takashi Iwai <tiwai@suse.de>]

At Wed, 11 Oct 2006 16:00:54 +0100,
James Courtier-Dutton wrote:
> 
> Assaf Hoffman wrote:
> > Hi,
> > Thanks again for clear answers.
> > In case the HW strip the IEC958 frames, where the ALSA stack expect to
> > find the various frame data fields such as perr, valid bit, audio
> > sample, user data and channel status? Is there a data structure the HW
> > should follow when DMAing the striped data to the DRAM buffer?
> > What do you think of such acceleration? Does this saves a significant
> > CPU time processing the ICE958 frames?
> > Many thanks.
> > P.S.
> > I was looking into the ASLA stack and I could find
> > SNDRV_PCM_FMTBIT_IEC958_SUBFRAME_XX only in snd_pcm_hardware structure
> > initialization. I could not see where it is used to tell the ALSA to
> > process the data as IEC958 frames.
> > 
> 
> A majority of sound cards strip the frame, and only put the audio
> samples in the DMA buffer. The sound card then stores the valid bits,
> non-audio bit, user data etc. in a separate register on the sound card.
> 
> The processing of the IEC958 frames is done in alsa-lib.
> ALSA has two parts, and kernel part, and a user space lib.
> The kernel part does as little as possible, with the user space part
> doing all the format conversion etc.
> 
> I have not looked closely, but I don't know if the alsa-lib decodes the
> IEC958 frames or not yet.

It can.  iec958 plugin does encode/decode IEC958 frames.


Takashi

-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642