Hmm... If there is a bug in kerneld, so that it dies when it loads a module,
there might be a schedule() at the wrong moment...
If someone with the problem has the opportunity to check if kerneld has
died immediately after the "Aiee" message, that might give me some input
for a bug-hunt in modprobe and/or kerneld.
Try the included patch.
Bjorn <bj0rn@blox.se>
--- linux/ipc/msg.c.org Wed Feb 14 13:11:33 1996
+++ linux/ipc/msg.c Wed Feb 14 13:13:13 1996
@@ -412,7 +412,8 @@
wake_up (&msq->rwait);
if (msq->wwait)
wake_up (&msq->wwait);
- schedule();
+ if (!intr_count) /* In case kerneld died "with his boots on" */
+ schedule();
}
for (msgp = msq->msg_first; msgp; msgp = msgh ) {
msgh = msgp->msg_next;