Change the second argument of vflush() to an integer that specifies

the number of references on the filesystem root vnode to be both
expected and released. Many filesystems hold an extra reference on
the filesystem root vnode, which must be accounted for when
determining if the filesystem is busy and then released if it isn't
busy. The old `skipvp' approach required individual filesystem
xxx_unmount functions to re-implement much of vflush()'s logic to
deal with the root vnode.

All 9 filesystems that hold an extra reference on the root vnode
got the logic wrong in the case of forced unmounts, so `umount -f'
would always fail if there were any extra root vnode references.
Fix this issue centrally in vflush(), now that we can.

This commit also fixes a vnode reference leak in devfs, which could
result in idle devfs filesystems that refuse to unmount.

Reviewed by:	phk, bp

1 files changed, 2 insertions, 12 deletions

diff --git a/sys/fs/portalfs/portal_vfsops.c b/sys/fs/portalfs/portal_vfsops.c
index 5146456..423cd38 100644
--- a/sys/fs/portalfs/portal_vfsops.c
+++ b/sys/fs/portalfs/portal_vfsops.c
@@ -155,7 +155,6 @@ portal_unmount(mp, mntflags, p)
 	int mntflags;
 	struct proc *p;
 {
-	struct vnode *rootvp = VFSTOPORTAL(mp)->pm_root;
 	int error, flags = 0;
 
 
@@ -172,21 +171,12 @@ portal_unmount(mp, mntflags, p)
 	if (mntinvalbuf(mp, 1))
 		return (EBUSY);
 #endif
-	if (rootvp->v_usecount > 1)
-		return (EBUSY);
-	error = vflush(mp, rootvp, flags);
+	/* There is 1 extra root vnode reference (pm_root). */
+	error = vflush(mp, 1, flags);
 	if (error)
 		return (error);
 
 	/*
-	 * Release reference on underlying root vnode
-	 */
-	vrele(rootvp);
-	/*
-	 * And blow it away for future re-use
-	 */
-	vgone(rootvp);
-	/*
 	 * Shutdown the socket.  This will cause the select in the
 	 * daemon to wake up, and then the accept will get ECONNABORTED
 	 * which it interprets as a request to go and bury itself.