Re: WAPBL fix for deallocation exhaustion + slow file removal

[David Holland <dholland-tech%netbsd.org@localhost>]

On Sat, Oct 01, 2016 at 05:00:10PM +0000, Taylor R Campbell wrote:
 > It's also suboptimal that we sleep while holding rwlocks for vnode
 > locks, since rw_enter is uninterruptable, so if a wait inside the
 > kernel hangs with a vnode lock held, anyone else trying to examine
 > that vnode will hang uninterruptably too.  But it will take some
 > effort to undo that state of affairs by teaching all callers of
 > vn_lock to accept failure.

...even if someone can persuade the developer consensus to accept that
change, which has failed in the past.

(more on the rest when I have some time, I hope)

-- 
David A. Holland
dholland%netbsd.org@localhost