Re: Changes to reduce pressure on uvm_pageqlock

[Mindaugas Rasiukevicius <rmind%netbsd.org@localhost>]

Chuck Silvers <chuq%chuq.com@localhost> wrote:
> In your new diff, the pageq lock is still used by the pagedaemon
> (for uvmpd_trylockowner() as you noted in the comment).
> It would be good not to have a global lock involved in changing page
> identity, and with the pageq's now protected by a separate lock, we
> should be able to eliminate the need for the pageq lock when freeing a
> page with an RCUy thing to prevent freeing uvm_object and uvm_anon
> structures while the pagedaemon might still be accessing them via the
> vm_page fields.  We should be able to improve the locking for
> pg->loan_count with a similar RCUy thing... my recollection is that the
> only reason loan_count uses the pageq lock currently is to stabilize the
> page identity while locking the owning object.  It might make sense to
> have pg->wire_count protected by the pdpol lock too (though I have some
> thoughts about changing the nature of wire_count that would probaby make
> that not make sense, I haven't had time to think that through yet).

Just a small note: since vm_page::wire_count and vm_page::loan_count are
kind of mutually exclusive, I think suggestion by yamt@ (long time ago) to
merge them (positive value meaning wired count and negative -- loan) is a
useful one.  It would also free up those 16-bits, since Andrew seems to be
in need of some.

-- 
Mindaugas