At Thu, 03 Sep 2009 13:50:26 +0200, Clemens Ladisch wrote:
(please don't top-post)
Giovanni Maruzzelli wrote:
In centos is 1000HZ, in the kernels I compiled is 100HZ
I would have expected these to be the other way around, since higher timer frequency should result in more accurate simulated sound card interrupts.
Actually, it's maybe because of a bad code in snd-dummy driver. It issues the timer update at each tick. Thus with more HZ, it can loose more ticks and be less accurate.
I worked on it ago, and changed the code to update in a better way. But forgot to apply to the upstream.
Now I digged down, and resurrected it with the highres timer support as a quick hack.
The patches are found in sound-unstable GIT tree topic/dummy-hrtimer branch: git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-unstable-2.6.git topic/dummy-hrtimer
and in alsa-driver-unstable snapshot as well: ftp://ftp.kernel.org/pub/linux/kernel/people/tiwai/snapshot/alsa-driver-unstable-snapshot.tar.gz
Giovanni, give it a try. If it works, I'm going to add it to 2.6.32.
thanks,
Takashi