On 11/02/2019 12.39, Peter Suetterlin wrote:
Carlos E. R. wrote:
So, what do you think?
Two things came to my mind looking at those:
- I first suspected thunderbird, as it was at higher load than firefox, but then realized that of course a process waiting for swap IO won't use too much CPU...
Right. And that top runs once per minute - possibly the thing tells the average of the minute, or perhaps just that instant.
- I don't think alsa is the cause. The messages you get are rather a consequence, as audio is extremely timing critical, and is also affected by the general slowness.
It is certainly probable.
I did see similar in my cases of hitting the swap deathmill: Sound started stuttering, until it would eventually stop completely. But it came back once swapping stopped.
Maybe not too enlightening, but all I can come up with.
Thanks. I think I will report in bugzilla what I know, and try to update to the tumbleweed kernel. Unless someone knows some command to try when it crashes to obtain more info, and I write it in a postit... Maybe log iotop output as well. -- Cheers / Saludos, Carlos E. R. (from openSUSE 15.0 (Legolas))