Re: [PATCH RFC ticketlock] Auto-queued ticketlock

["Paul E. McKenney" <paulmck@linux.vnet.ibm.com>]

On Tue, Jun 11, 2013 at 11:22:45AM -0400, Steven Rostedt wrote:
> On Tue, 2013-06-11 at 03:14 -0700, Paul E. McKenney wrote:
> >  
> > > Off-topic, although I am in this community for several years,
> > > I am not exactly clear with this problem.
> > > 
> > > 1) In general case, which lock is the most competitive in the kernel? what it protects for?
> > > 2) In which special case, which lock is the most competitive in the kernel? what it protects for?
> > > 3) In general case, which list is the most hot list?
> > > 4) In which special case, which list is the most hot list?
> > 
> > Others would know better than I, but mmap_sem has been called out as a
> 
> If the contention is with mmap_sem, then I doubt this is going to help
> much, as that's a sleeping rw semaphore. Now, rw semaphores are
> implemented with raw spinlocks, but I doubt that would be the main point
> of contention, compared to the sleeping part.

If I remember correctly, someone actually hit this earlier this year,
which prompted use of a special-purpose queued lock to guard the
semaphore data.  I don't recall whether it was mmap_sem or not, so
cannot say whether it was a straight mutex or an rw semaphore.

							Thanx, Paul

> -- Steve
> 
> > prime offender for some workloads.  There is of course some debate as
> > to whether the fault lies mmap_sem or with the workloads.  There have
> > been some efforts to solve this one on LKML, plus some in academia have
> > worked on this as well:
> > 
> 
>