ipmi: handle run_to_completion properly in deliver_recv_msg()

If run_to_completion flag is set, it means that we are running in a
single-threaded mode, and thus no locks are held.

This fixes a deadlock when IPMI notifier is being called during panic.

Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Acked-by: Corey Minyard <minyard@acm.org>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

1 files changed, 8 insertions, 3 deletions

diff --git a/drivers/char/ipmi/ipmi_si_intf.c b/drivers/char/ipmi/ipmi_si_intf.c
index f052c481327..35603dd4e6c 100644
--- a/drivers/char/ipmi/ipmi_si_intf.c
+++ b/drivers/char/ipmi/ipmi_si_intf.c
@@ -323,9 +323,14 @@ static void deliver_recv_msg(struct smi_info *smi_info,
 {
         /* Deliver the message to the upper layer with the lock
            released. */
-        spin_unlock(&(smi_info->si_lock));
-        ipmi_smi_msg_received(smi_info->intf, msg);
-        spin_lock(&(smi_info->si_lock));
+
+        if (smi_info->run_to_completion) {
+                ipmi_smi_msg_received(smi_info->intf, msg);
+        } else {
+                spin_unlock(&(smi_info->si_lock));
+                ipmi_smi_msg_received(smi_info->intf, msg);
+                spin_lock(&(smi_info->si_lock));
+        }
 }
 
 static void return_hosed_msg(struct smi_info *smi_info, int cCode)