[alsa-devel] Timer instability

[Lennart Poettering]

On Mon, 23.02.09 20:20, Lennart Poettering (mznyfn at 0pointer.de) wrote:

> On Mon, 23.02.09 03:56, Lennart Poettering (mznyfn at 0pointer.de) wrote:
> 
> > Hmm, for now my little testing experiment I've now found es1969,
> > ens1371, intel8x0, snd-emu10k1 and some intel-hda to have unreliable
> > _avail() or _delay().
> 
> And here's the output for HDA STAC92xx:
> 
> http://pastebin.mandriva.com/7149
> 
> The first dump is the interesting one: when audio starts to play the
> buffer abruptly runs empty in just 20us after the device started. (The
> rightmost column is the state btw, 3 == playing).
> 
> A quick overview of cards where _delay() and/or _avail() are unreliable:
> 
>  es1969
>  ens1371
>  intel8x0
>  emu10k1
>  intel-hda on STAC92xx
>  intel-hda on AD1989B

And here's another one, for intel-hda on ALC883:

Sometimes snd_pcm_avail() returns values like 4611686018427387098.

Smells like another overflow to me. 

Lennart

-- 
Lennart Poettering                        Red Hat, Inc.
lennart [at] poettering [dot] net         ICQ# 11060553
http://0pointer.net/lennart/           GnuPG 0x1A015CC4