Date: Wed, 20 Mar 1996 08:29:25 -0500 (EST) From: "matthew c. mead" <mmead@Glock.COM> To: current@freebsd.org Subject: Re: cvs-all-digest V1 #300 Message-ID: <199603201329.IAA00770@neon.Glock.COM> In-Reply-To: <199603191927.LAA23880@freefall.freebsd.org> from "owner-cvs-all-digest@freefall.freebsd.org" at Mar 19, 96 11:27:18 am
next in thread | previous in thread | raw e-mail | index | archive | help
owner-cvs-all-digest@freefall.freebsd.org writes: > From: John Dyson <dyson> > Date: Mon, 18 Mar 1996 21:13:24 -0800 (PST) > Subject: cvs commit: src/sys/miscfs/specfs spec_vnops.c src/sys/vm vnode_pager.c > > dyson 96/03/18 21:13:24 > > Modified: sys/miscfs/specfs spec_vnops.c > sys/vm vnode_pager.c > Log: > Fix the problem that unmounting filesystems that are backed by a VMIO > device have reference count problems. We mark the underlying object > ono-persistent, and account for the reference count that the VM system > maintainsfor the special device close. This should fix the removable > device problem. > > Revision Changes Path > 1.29 +3 -2 src/sys/miscfs/specfs/spec_vnops.c > 1.59 +5 -2 src/sys/vm/vnode_pager.c > > ------------------------------ I've thoroughly tested this code change, and from what I can tell, it fixes all the problems I've found using a JAZ drive, save one. The one that it doesn't fix is if you mount a UFS filesystem on the JAZ drive and allow it to sleep, then before waking it up via a standard filesystem call you fsck the raw device (ok, ok, so I did this by accident the first time... :-), then try to use the mounted ufs, you cannot unmount it or use it until you reboot. The solution to that problem is to not try to fsck it while it's mounted. Other than that, the unsleep code and the unmount/unlocking for ejection code works great. -matt -- Matthew C. Mead mmead@Glock.COM http://www.Glock.COM/~mmead/
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199603201329.IAA00770>