On (06/28/17 15:17), Petr Mladek wrote:
> On Thu 2017-06-01 16:21:02, Sergey Senozhatsky wrote:
> > On (05/31/17 16:30), Sergey Senozhatsky wrote:
> > > On (05/29/17 14:12), Jan Kara wrote:
> > > [..]
> > > > Actually I had something very similar in old versions of my patch set.
> > > > And
> > > > it didn't work very well. The problem was that e.g. sometimes scheduler
> > > > decided that printk kthread should run on the same CPU as the process
> > > > currently doing printing and in such case printk kthread never took over
> > > > printing and the machine locked up due to heavy printing.
> > >
> > > hm, interesting.
> >
> > that's a tricky problem to deal with.
> >
> >
> >
> > ... so may be we can have per-CPU printk kthreads then
> >
> > static DEFINE_PER_CPU(struct task_struct *, printk_kthread);
> >
> >
> > SMP hotplug threads, to be precise, the same way as watchdog has it. and
> > then during offloading we can wake_up any printk_kthread that is knowingly
> > not from this-CPU, all of them, let them compete for the console_sem.
> >
> > just a quick idea.
> >
> > thoughts?
>
> I am not sure if this is worth the resources. It think that one
> big win of workqueues was that it reduced the amount of running
> per-CPU kthreads. There are systems with thousands of CPUs.
>
> I am a bit afraid to use workqueues for flushing consoles.
> It would be another dependency and another risk.
>
> Otherwise, per-CPU kthreads/workqueues primary handle per-CPU
> resources. But printk_kthread would handle consoles that
> need to be serialized anyway. It sounds weird to have
> per-CPU task just to increase the chance that it will
> get scheduled.
yeah. was just a quick idea. it has some 'interesting' options, tho.
I'll reply in another thread.
the waste of resources argument is somewhat interesting. I'm not
arguing, and agree that per-CPU kthread for printk seems like a
massive-massive overkill. the point is - I think that 99.999% of
the time printk_safe and printk_nmi buffers are not needed. they
simply waste memory. on a $BIG systems that's, once again, can be
huge. so in terms of resources printk probably must do better,
already.
-ss
