Date: Thu, 23 Apr 1998 14:57:26 +0900 (JST) From: Michael Hancock <michaelh@cet.co.jp> To: Terry Lambert <tlambert@primenet.com> Cc: "Jordan K. Hubbard" <jkh@time.cdrom.com>, ken@plutotech.com, syssgm@dtir.qld.gov.au, freebsd-current@FreeBSD.ORG Subject: Re: NFS corruption Message-ID: <Pine.SV4.3.95.980423132225.17092A-100000@parkplace.cet.co.jp> In-Reply-To: <199804221912.MAA29389@usr08.primenet.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, 22 Apr 1998, Terry Lambert wrote: > There probably needs to be some consensus brought to Michael Hancock's > patches, as well. They seem good, and they seem in line with simplifying > what an NFS maintainer (or any other person doing FS developement) needs > to have to be able to hold in their head at one time to get useful work > done. I feel confident that the patches DTRT. Kirk wanted to see them fixed and after going through the code I can see why. They weren't done earlier because the scope of the changes seemed very large, but it wasn't all that bad. There were over 700 vrele and vput calls, but it appears that only about 10 to 15 percent of them were bogusly placed. I'm pretty sure we have concensus, but I would like to have someone test NFS and ext2fs since I've only done compile testing on those file systems. Note the patches for these two file systems were very straight forward compared to the ones for unionfs so I'm confident that they are fine. If anyone has the resources to test then please download the patches at... http://www.freebsd.org/~mch/vop1a.diff Then do cd /sys patch < /tmp/vop1a.diff rebuild your kernel To reverse cd /sys patch -R < /tmp/vop1a.diff If I can't find any testers for nfs/ext2 then I suggest we just commit the patch sometime next week. There is a national holiday in Japan, but I'm not going anywhere so I'll be around in the unlikely case there's a big problem. After this I will have a little time to help out on softupdates and in a month or so I'll have an NFS environment up and running so I can help out there as well. I'm also going to look more into VM/VFS interaction for all file systems and some issues with nullfs. I decided to put the VFS_VRELE stuff on hold after mulling through all the discussions with you, John, Kirk, Poul and recently John Heidemann. Basically, instead of abstracting up into vn_rele I'll probably just keep vrele and make a VFS_VALLOC/VFS_VFREE. We should backout my previous set of patches. I'll probably revisit this after softupdates and NFS have stabilized. Regards, Mike To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.SV4.3.95.980423132225.17092A-100000>