qeth: ungrouping a device must not be interruptible
authorUrsula Braun <braunu@de.ibm.com>
Wed, 29 Aug 2007 09:26:52 +0000 (11:26 +0200)
committerJeff Garzik <jeff@garzik.org>
Fri, 31 Aug 2007 10:52:57 +0000 (06:52 -0400)
Problem:
A recovery thread must not be active when device is removed.
In qeth_remove_device() an interruptible wait operation is used
to wait until a qeth recovery thread is finished. If a user really
interrupts the ungroup operation of a qeth device while a recovery
is running, cio and qeth are out of sync (device already removed
from cio, but kept in qeth). A following module unload of qeth
results in a kernel OOPS here.

Solution:
Do not allow interruption of ungroup operation to guarantee
finishing of a potentially running qeth recovery thread.

Signed-off-by: Ursula Braun <braunu@de.ibm.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
drivers/s390/net/qeth_main.c

index 57f69434fbf9f2319b5b9a6ac33692c80474d546..ba10d42a1a56cd5d655a6de68abf91bbfae0f5d2 100644 (file)
@@ -561,7 +561,7 @@ qeth_set_offline(struct ccwgroup_device *cgdev)
 }
 
 static int
-qeth_wait_for_threads(struct qeth_card *card, unsigned long threads);
+qeth_threads_running(struct qeth_card *card, unsigned long threads);
 
 
 static void
@@ -576,8 +576,7 @@ qeth_remove_device(struct ccwgroup_device *cgdev)
        if (!card)
                return;
 
-       if (qeth_wait_for_threads(card, 0xffffffff))
-               return;
+       wait_event(card->wait_q, qeth_threads_running(card, 0xffffffff) == 0);
 
        if (cgdev->state == CCWGROUP_ONLINE){
                card->use_hard_stop = 1;