On 17/02/26(Tue) 11:28, Theo de Raadt wrote:
> > Per that post, the primary problem concerns page allocation and the
> > way mutexes are implemented
> 
> While looking at things in the pagedaemon, I've become concerned
> about how uvm_lock_pageq() is used

Managed page have an associated lock which is asserted to be locked in
these operation.
 
> Look at how uvm_pageactivate() and uvm_pagedeactivate() must be called
> unlocked

The lock is grabbed inside these functions and is not recursive, that's
why they must be called unlocked.

Most of the loop in the page daemon doesn't need the lock.