Re: [PATCH v4] softlockup: remove hung_task_check_count

[Mandeep Singh Baines <msb@google.com>]

Mandeep Singh Baines (msb@google.com) wrote:
> Peter Zijlstra (peterz@infradead.org) wrote:
> > 
> > Why break out at all? Are you that worried about khungtaskd introducing
> > latencies?
> 
> Yes, I was worried about disabling preemption for an unbounded amount of
> time.
> 
> > Is using preemptible RCU an option for you?
> > 
> 
> I had not even considered that. To be honest, I had not even heard of it
> till now. So I spent another morning at LWN grokking preemptible RCU;)
> 
> I think it can work. I'm a little worried about the OOM risk. It could take
> a really long time to iterate over the task list. A lot of pending kfree()s
> could build up in that time.
> 

I misunderstood preemptible RCU. I assumed it was a new API but its not. So
I don't think preemptible RCU is an option since it would force a dependency
on CONFIG_PREEMPT_RCU.

I'm going to break up this patch in two. One patch for converting to rcu.
A second patch which will support checking all tasks. To support checking
all tasks I reverted back to a design similar to Frédéric original proposal.

I'll send the patches out right after this email.

[PATCH 1/2] softlockup: convert read_lock in hung_task to rcu_read_lock
[PATCH 2/2] softlockup: check all tasks in hung_task

Regards,
Mandeep