Char: moxa, fix locking in moxa_write

moxa_write can be called from atomic context with irqs disabled (from
ppp_async_push). Don't enable interrupts by spin_unlock_bh as this
might cause deadlocks in the ppp layer.

Instead, use irqsave/irqrestore spin_lock functions.

Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

1 files changed, 3 insertions, 2 deletions

diff --git a/drivers/tty/moxa.c b/drivers/tty/moxa.c
index a290e9ebafe0..6255561dd2a3 100644
--- a/drivers/tty/moxa.c
+++ b/drivers/tty/moxa.c
@@ -1202,14 +1202,15 @@ static int moxa_write(struct tty_struct *tty,
                       const unsigned char *buf, int count)
 {
         struct moxa_port *ch = tty->driver_data;
+        unsigned long flags;
         int len;
 
         if (ch == NULL)
                 return 0;
 
-        spin_lock_bh(&moxa_lock);
+        spin_lock_irqsave(&moxa_lock, flags);
         len = MoxaPortWriteData(tty, buf, count);
-        spin_unlock_bh(&moxa_lock);
+        spin_unlock_irqrestore(&moxa_lock, flags);
 
         set_bit(LOWWAIT, &ch->statusflags);
         return len;