You are not logged in.
Recently I've been seeing weird behavior where a kworker process seems to be stuck consuming 100% of a single CPU core without any reason. This is fine if it was temporary, but when it happens it just seems to run forever (even had one running all night without noticing) and the weird thing is this:
# cat /proc/<offending_pid>/stack
[<ffffffffffffffff>] 0xffffffffffffffffNothing helps this problem unless I reboot. Is this a kernel bug of some sort? How can I figure out what's causing this?
Offline
You'll get the same output for every kernel thread.
Maybe perf would be able to tell what's going there.
Offline
You'll get the same output for every kernel thread.
Well, not quite:
# cat /proc/<kworker_pid>/stack
[<ffffffff8108e5cf>] worker_thread+0xbf/0x4a0
[<ffffffff81093ff8>] kthread+0xd8/0xf0
[<ffffffff81591e8f>] ret_from_fork+0x3f/0x70
[<ffffffffffffffff>] 0xffffffffffffffffNot to mention other threads that are actually doing something have some more depth to the stack.
Last edited by yuvadm (2016-02-20 09:37:56)
Offline
Sorry, I tried on a different kernel. Anyway, perf record -g.
Offline