Date: Wed, 13 Sep 1995 10:52:16 -0700 (MST) From: Terry Lambert <terry@lambert.org> To: roberto@keltia.freenix.fr (Ollivier Robert) Cc: wollman@lcs.mit.edu, terry@lambert.org, current@freefall.freebsd.org Subject: Re: Is nullfs broken in -current? Message-ID: <199509131752.KAA07769@phaeton.artisoft.com> In-Reply-To: <199509122200.AAA19788@keltia.Freenix.FR> from "Ollivier Robert" at Sep 13, 95 00:00:24 am
next in thread | previous in thread | raw e-mail | index | archive | help
> It seems that Garrett A. Wollman said: > > # mount -t null -o union /dsk2/src1 /usr/src > > # mount -t null -o union /dsk3/src2 /usr/src > > # cd /usr/src > > # make world This one is a unionfs, not a nullfs. > This one is a winner too: > > # mount -t null /src/src /usr/src > # cvs update -d -P You are replacing the /usr/src vnode (potentially on cdrom) with the /src/src vnode from a writeable media? This seems more useful and applicable. I'll look at it. I suspect some of the problem is the delay in the vnode alias should not be occuring for the mount point itself. Part of this could be a cache issue. The name cache wants to be moved to the lookup layer instead of being implemented in each subsidiary file system. It doesn't look like the cache is flushed for the top level vnode, so potentially stale data is being returned. Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199509131752.KAA07769>