From owner-freebsd-hackers Sat Feb 20 2:17: 1 1999 Delivered-To: freebsd-hackers@freebsd.org Received: from apollo.backplane.com (apollo.backplane.com [209.157.86.2]) by hub.freebsd.org (Postfix) with ESMTP id 01F4611740 for ; Sat, 20 Feb 1999 02:15:23 -0800 (PST) (envelope-from dillon@apollo.backplane.com) Received: (from dillon@localhost) by apollo.backplane.com (8.9.3/8.9.1) id CAA07195; Sat, 20 Feb 1999 02:15:22 -0800 (PST) (envelope-from dillon) Date: Sat, 20 Feb 1999 02:15:22 -0800 (PST) From: Matthew Dillon Message-Id: <199902201015.CAA07195@apollo.backplane.com> To: Terry Lambert Cc: julian@whistle.com (Julian Elischer), andre.albsmeier@mchp.siemens.de, sanewo@ba2.so-net.ne.jp, freebsd-hackers@FreeBSD.ORG Subject: Re: softupdate panic, anyone seen this? References: <199902192352.QAA11969@usr02.primenet.com> Sender: owner-freebsd-hackers@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG :> WHY would you async mount an MFS? :> it's an MFS! : :So you don't have to stall the caller copying the page-off-vnode :memory to the swappable-page memory. : : Terry Lambert : terry@lambert.org So you don't have to do a context switch for each individual strategy call. -- Progress is being made on a number of fronts. I have about 9 commits ready to go as soon as they are approved which fix a bunch of things, including the VN device. And I have two patches that 'appear' to fix softupdates that have been sent to the softupdates crew for fodder. With them, I no longer get lockups or panics on a softupdates-enabled UFS filesystem sitting on a NFS-file-backed VN partition. ( But I've no idea if I'm solving the problems the right way. Probably not :-) ). I can now reliably make buildworld on the above said VN configuration ( as /usr/obj ). It's sweet, too... softupdates does an incredible job balancing the network load. It doles the file writes out to the network so smoothly I at first didn't realize that it was working. MFS, in contrast, eats system memory until the pageout daemon is forced to start paging it to swap in weird-looking bursts. The network I/O doesn't get in the way of the build at all - the system is 0% idle through most of it. I think the ultimate 'mfs' is going to be a direct-swap-backed VN partition running a softupdates-enabled UFS. The beast doesn't exist yet, but it wouldn't be too hard - we'd just use a vm_object instead of a vnode and vm_pager_get/put_pages(). -Matt Matthew Dillon To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message