hvc_console: Fix race between hvc_close and hvc_remove
authorAnton Blanchard <anton@samba.org>
Tue, 6 Apr 2010 11:42:38 +0000 (21:42 +1000)
committerRusty Russell <rusty@rustcorp.com.au>
Thu, 8 Apr 2010 00:16:20 +0000 (09:46 +0930)
commit320718ee074acce5ffced6506cb51af1388942aa
tree0c57b9337452263e8f032cd0f37069f70921c490
parentb7a413015d2986edf020fba765c906cc9cbcbfc9
hvc_console: Fix race between hvc_close and hvc_remove

I don't claim to understand the tty layer, but it seems like hvc_open and
hvc_close should be balanced in their kref reference counting.

Right now we get a kref every call to hvc_open:

        if (hp->count++ > 0) {
                tty_kref_get(tty); <----- here
                spin_unlock_irqrestore(&hp->lock, flags);
                hvc_kick();
                return 0;
        } /* else count == 0 */

        tty->driver_data = hp;

        hp->tty = tty_kref_get(tty); <------ or here if hp->count was 0

But hvc_close has:

        tty_kref_get(tty);

        if (--hp->count == 0) {
...
                /* Put the ref obtained in hvc_open() */
                tty_kref_put(tty);
...
        }

        tty_kref_put(tty);

Since the outside kref get/put balance we only do a single kref_put when
count reaches 0.

The patch below changes things to call tty_kref_put once for every
hvc_close call, and with that my machine boots fine.

Signed-off-by: Anton Blanchard <anton@samba.org>
Acked-by: Amit Shah <amit.shah@redhat.com>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
drivers/char/hvc_console.c