Re: cgroup trace events acquire sleeping locks

[Peter Zijlstra]

On Mon, Jul 09, 2018 at 04:30:10PM -0400, Steven Rostedt wrote:
> On Mon, 9 Jul 2018 22:22:15 +0200
> Sebastian Andrzej Siewior <bigeasy@xxxxxxxxxxxxx> wrote:
> 
> > On 2018-07-09 15:01:54 [-0400], Steven Rostedt wrote:
> > > > which is the trace_cgroup_rmdir() trace event in cgroup_rmdir(). The
> > > > trace event invokes cgroup_path() which acquires a spin_lock_t and this
> > > > is invoked within a preempt_disable()ed section.   
> > > 
> > > Correct. And I wish no trace event took spin locks.  
> > 
> > is there an easy way to detect this? I mean instead hitting the trace
> > event with debug enabled and doing a review of each of them.
> 
> Hmm, good question. I could possibly make all the tracepoint code into
> its own section. And then look to see if any spin locks exist in them.
> That wouldn't be too trivial to implement though.

pick a bit from the preempt_count (say right above NMI_MASK) and set it
inside a trace-event and add in_trace().

Then make lockdep explode when in_trace().

Or something along those lines.