* 2.6.0-test1: autofs4 doesn't expire
@ 2003-07-20 10:32 Dick Streefland
2003-07-20 13:00 ` Dick Streefland
0 siblings, 1 reply; 3+ messages in thread
From: Dick Streefland @ 2003-07-20 10:32 UTC (permalink / raw)
To: linux-kernel
In 2.6.0-test1, the autofs4 automounter doesn't expire mounts anymore,
both NFS and CDROM mounts. It did work in my previous kernel, 2.5.72.
When I try to unmount manually, I get "device is busy", although
"fuser -m" doesn't report anything. I've also searched /proc/*/fd/,
but there are no open files below the mount points.
Is anybody else seeing this? Any ideas what could be the cause?
--
Dick Streefland //// De Bilt
dick.streefland@xs4all.nl (@ @) The Netherlands
------------------------------oOO--(_)--OOo------------------
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: 2.6.0-test1: autofs4 doesn't expire
2003-07-20 10:32 2.6.0-test1: autofs4 doesn't expire Dick Streefland
@ 2003-07-20 13:00 ` Dick Streefland
2003-07-20 16:19 ` Jeremy Fitzhardinge
0 siblings, 1 reply; 3+ messages in thread
From: Dick Streefland @ 2003-07-20 13:00 UTC (permalink / raw)
To: linux-kernel
spam@streefland.xs4all.nl (Dick Streefland) wrote:
| In 2.6.0-test1, the autofs4 automounter doesn't expire mounts anymore,
| both NFS and CDROM mounts. It did work in my previous kernel, 2.5.72.
| When I try to unmount manually, I get "device is busy", although
| "fuser -m" doesn't report anything. I've also searched /proc/*/fd/,
| but there are no open files below the mount points.
|
| Is anybody else seeing this? Any ideas what could be the cause?
Update: 2.5.75 is OK, so the problem was introduced in 2.6.0-test1.
I have not yet verified it, but I suspect that the addition of the
mntget() call is the cause:
==> fs/namespace.c <==
+/*
+ * Now, lookup_mnt increments the ref count before returning
+ * the vfsmount struct.
+ */
struct vfsmount *lookup_mnt(struct vfsmount *mnt, struct dentry *dentry)
{
struct list_head * head = mount_hashtable + hash(mnt, dentry);
struct list_head * tmp = head;
- struct vfsmount *p;
+ struct vfsmount *p, *found = NULL;
+ spin_lock(&vfsmount_lock);
for (;;) {
tmp = tmp->next;
p = NULL;
if (tmp == head)
break;
p = list_entry(tmp, struct vfsmount, mnt_hash);
- if (p->mnt_parent == mnt && p->mnt_mountpoint == dentry)
+ if (p->mnt_parent == mnt && p->mnt_mountpoint == dentry) {
+ found = mntget(p);
break;
+ }
}
- return p;
+ spin_unlock(&vfsmount_lock);
+ return found;
}
--
Dick Streefland //// De Bilt
dick.streefland@xs4all.nl (@ @) The Netherlands
------------------------------oOO--(_)--OOo------------------
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: 2.6.0-test1: autofs4 doesn't expire
2003-07-20 13:00 ` Dick Streefland
@ 2003-07-20 16:19 ` Jeremy Fitzhardinge
0 siblings, 0 replies; 3+ messages in thread
From: Jeremy Fitzhardinge @ 2003-07-20 16:19 UTC (permalink / raw)
To: Dick Streefland; +Cc: linux-kernel
On Sun, 2003-07-20 at 06:00, Dick Streefland wrote:
> spam@streefland.xs4all.nl (Dick Streefland) wrote:
> | In 2.6.0-test1, the autofs4 automounter doesn't expire mounts anymore,
> | both NFS and CDROM mounts. It did work in my previous kernel, 2.5.72.
> | When I try to unmount manually, I get "device is busy", although
> | "fuser -m" doesn't report anything. I've also searched /proc/*/fd/,
> | but there are no open files below the mount points.
> |
> | Is anybody else seeing this? Any ideas what could be the cause?
>
> Update: 2.5.75 is OK, so the problem was introduced in 2.6.0-test1.
> I have not yet verified it, but I suspect that the addition of the
> mntget() call is the cause:
Hm, that does look like it would upset autofs4. I'll take a look.
J
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2003-07-20 16:04 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2003-07-20 10:32 2.6.0-test1: autofs4 doesn't expire Dick Streefland
2003-07-20 13:00 ` Dick Streefland
2003-07-20 16:19 ` Jeremy Fitzhardinge
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).