Hi,
On May 23 2015 00:08, Takashi Iwai wrote:
Here you dropped _ratelimited(). Are you sure that it won't give any problem?
Similar situation as we discussed before: http://mailman.alsa-project.org/pipermail/alsa-devel/2015-May/092192.html
On May 23 2015 03:05, Clemens Ladisch wrote:
An aborted stream ends up in the XRUN state; the application is likely to prepare and start the stream again. So it is likely that there will be message flooding.
Considering about how frequently this message can be generated at receiving such problematic packets, Not so flooding.
In the situation, all of committed drivers (fireworks/bebob/oxfw/dice) stops isochronous contexts and actual streams once, then start new streams and isochronous contexts. This is because our old packet stream falls to error state. Therefore, when devices transfers such problematic packets, different isochronous contexts (old one and new one) generate the error messages. This means that there're a gap between the generated error messages, approximately several hundred seconds because it cost expensive for devices to restart isochronous streams. In my opinion, this is not such flooding.
Regards
Takashi Sakamoto