spi: Fix crash with double message finalisation on error handling
authorGeert Uytterhoeven <geert+renesas@linux-m68k.org>
Tue, 28 Jan 2014 09:33:03 +0000 (10:33 +0100)
committerMark Brown <broonie@linaro.org>
Tue, 28 Jan 2014 17:48:40 +0000 (17:48 +0000)
This reverts commit e120cc0dcf2880a4c5c0a6cb27b655600a1cfa1d.

It causes a NULL pointer dereference with drivers using the generic
spi_transfer_one_message(), which always calls
spi_finalize_current_message(), which zeroes master->cur_msg.

Drivers implementing transfer_one_message() theirselves must always call
spi_finalize_current_message(), even if the transfer failed:

 * @transfer_one_message: the subsystem calls the driver to transfer a single
 *      message while queuing transfers that arrive in the meantime. When the
 *      driver is finished with this message, it must call
 *      spi_finalize_current_message() so the subsystem can issue the next
 *      transfer

Signed-off-by: Geert Uytterhoeven <geert+renesas@linux-m68k.org>
Signed-off-by: Mark Brown <broonie@linaro.org>
drivers/spi/spi.c

index 63613a96233c3f1eddc77846affea03ffe92a852..26f4afd19fafc77ba44ed9d0eaba679dba2f267c 100644 (file)
@@ -745,9 +745,7 @@ static void spi_pump_messages(struct kthread_work *work)
        ret = master->transfer_one_message(master, master->cur_msg);
        if (ret) {
                dev_err(&master->dev,
-                       "failed to transfer one message from queue: %d\n", ret);
-               master->cur_msg->status = ret;
-               spi_finalize_current_message(master);
+                       "failed to transfer one message from queue\n");
                return;
        }
 }