From owner-cvs-all Wed Nov 14 1:26:26 2001 Delivered-To: cvs-all@freebsd.org Received: from apollo.backplane.com (apollo.backplane.com [216.240.41.2]) by hub.freebsd.org (Postfix) with ESMTP id C662237B41C; Wed, 14 Nov 2001 01:26:21 -0800 (PST) Received: (from dillon@localhost) by apollo.backplane.com (8.11.6/8.9.1) id fAE9QKd79635; Wed, 14 Nov 2001 01:26:20 -0800 (PST) (envelope-from dillon) Date: Wed, 14 Nov 2001 01:26:20 -0800 (PST) From: Matthew Dillon Message-Id: <200111140926.fAE9QKd79635@apollo.backplane.com> To: Alfred Perlstein Cc: John Baldwin , cvs-all@FreeBSD.ORG, cvs-committers@FreeBSD.ORG Subject: Re: cvs commit: src/sys/kern kern_descrip.c kern_event.c sys_gen References: <200111140630.fAE6UaS32367@freefall.freebsd.org> <20011114021538.V13393@elvis.mu.org> Sender: owner-cvs-all@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG :... :> > uipc_syscalls.c :> > sys/sys file.h :> > Log: :> > remove holdfp() :> :> Erm, did you coordinate with Alfred on this? There has already been a good :> deal of file descriptor work done. : :I keep stalling on it, I've given him my broken patches and I'll be :keeping an eye on him anyhow. :) : :-Alfred Well, the patches are a bit out of date but they represent a pretty good roadmap for about half of what we need to accomplish in that area. There's a huge amount of work we have to do beyond the patches or anything I've comitted so far. -- I have to again note that the best way to deal with the huge job of unwinding Giant is with an incremental approach. There is no point throwing locks around struct file or filedesc operations without first cleaning up the APIs used to access those structures. There is also no point spending days putting together huge patches that wind up languishing for months and never getting comitted, or wind up becoming so large that they aren't safe to commit, or they become stale. An incremental approach is far more effective. For example, all I am doing right now is cleaning up APIs, one at a time, and comitting the changes in small chunks to give people time to test them to make sure there's no cockpit trouble. But if I were to attempt to commit a large patch set that tries to do everything at once... the API, the mutexing, and Giant unwinding... well good luck finding the bugs! -Matt Matthew Dillon To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe cvs-all" in the body of the message