2011/10/8 Richard Braun <rbr...@sceen.net>: > On Sat, Oct 08, 2011 at 02:06:04AM +0200, Sergio López wrote: >> Reducing lock contention would surely help. > > vm_page_queue_lock is a simple lock if I'm not mistaken, which means > it's a nop in our case. We shouldn't change anything we can't measure > and test easily unless it's really simple enough.
You're right, it's a simple_lock.
