Re: [Query] Preemption (hogging) of the work handler

[Viresh Kumar]

On 15-07-16, 22:11, Sergey Senozhatsky wrote:
> Hello,
> 
> On (07/14/16 16:52), Viresh Kumar wrote:
> > On 12-07-16, 23:03, Sergey Senozhatsky wrote:
> > > so, I'm looking at this thing now:
> > > 
> > > : [   12.874909] sched: RT throttling activated for rt_rq ffffffc0ac13fcd0 (cpu 0)
> > > : [   12.874909] potential CPU hogs:
> > > : [   12.874909]  printk (292)
> > > 
> [..]
> > - But before that it dumps the kernel-messages from last boot and the prints
> >   looked like this:
> 
> kmsg_dump()?

I am happy that this code is available in open :)

https://chromium.googlesource.com/aosp/platform/system/core/+/master/healthd/healthd_mode_charger.cpp#194

> a wild guess... any chance that you kmsg dumper iterates
> log_buf records under logbuf_lock long enough to cause throttling on
> other CPU, because printk_kthread is simply has to spin on logbuf_lock?

Hmm, I am not sure of it yet.

-- 
viresh