Subject: Re: poll(2) oddity
To: None <tech-userlevel@netbsd.org>
From: David Laight <david@l8s.co.uk>
List: tech-userlevel
Date: 07/08/2002 13:45:05
While on the subject of poll(), I've just found another bug in it!

Consider:

process a polls on fd
process b polls on fd
process a times out
process c polls on fd
fd is woken up
now both processes should wake up but...
process c continues
process b stays sleeping

bug is that selrecord() mustn't clear SI_COLL.

	David

-- 
David Laight: david@l8s.co.uk