From owner-freebsd-hackers Sat Mar 6 11:23:54 1999 Delivered-To: freebsd-hackers@freebsd.org Received: from smtp03.primenet.com (smtp03.primenet.com [206.165.6.133]) by hub.freebsd.org (Postfix) with ESMTP id 4FBA5152BA for ; Sat, 6 Mar 1999 11:23:27 -0800 (PST) (envelope-from tlambert@usr06.primenet.com) Received: (from daemon@localhost) by smtp03.primenet.com (8.8.8/8.8.8) id MAA00217; Sat, 6 Mar 1999 12:23:09 -0700 (MST) Received: from usr06.primenet.com(206.165.6.206) via SMTP by smtp03.primenet.com, id smtpd000199; Sat Mar 6 12:23:09 1999 Received: (from tlambert@localhost) by usr06.primenet.com (8.8.5/8.8.5) id MAA09558; Sat, 6 Mar 1999 12:23:08 -0700 (MST) From: Terry Lambert Message-Id: <199903061923.MAA09558@usr06.primenet.com> Subject: Re: lockf and kernel threads To: dyson@iquest.net Date: Sat, 6 Mar 1999 19:23:07 +0000 (GMT) Cc: dillon@apollo.backplane.com, hasty@rah.star-gate.com, tlambert@primenet.com, dick@tar.com, jplevyak@inktomi.com, hackers@FreeBSD.ORG In-Reply-To: <199903051902.OAA00724@y.dyson.net> from "John S. Dyson" at Mar 5, 99 02:02:36 pm X-Mailer: ELM [version 2.4 PL25] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-hackers@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG > > :Thats probably true however for delivery of an AST I don't thing that we > > :need priviliged instructions --- I could be wrong. > > > > Little things like, ohhhhh disabling interrupts. Accessing the MMU > > registers, flushing the TLB, etc..... believe me, ring 1 and ring 2 > > is utterly useless for anything FreeBSD wants to run in supervisor mode. > > AST's could be simulated, but it is just more hair added to an already > sufficient set of hair. Let me state, categorically, and for the records, that I think that system call completion functions are an architectural dead-end. The point of an async call gate is to allow queueing of operations which are later specifically waited on and/or for which a wait-any is executed to reap their status. They differ primarily from signals in that completion is an event that modifies data rather than triggering procedural code, and an explicit (potentially non-blocking, but *not* asynchronous) test must be done to determine completion. Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message