[1/2] cdc-acm: do not reset notification buffer index upon urb unlinking

[1/2] cdc-acm: do not reset notification buffer index upon urb unlinking

[Oliver Neukum]

On Sa, 2018-09-22 at 22:11 +0200, Tobias Herzog wrote:
> Resetting the write index of the notification buffer on urb unlink (e.g.
> closing a cdc-acm device from userspace) may lead to wrong interpretation
> of further received notifications, in case the index is not 0 when urb
> unlink happens (i.e. when parts of a notification already have been
> transferred). On the device side there is no "reset" of the notification
> transimission and thus we would get out of sync with the device.
> 
> Signed-off-by: Tobias Herzog <t-herzog@gmx.de>
Acked-by: Oliver Neukum <oneukum@suse.com>

[1/2] cdc-acm: do not reset notification buffer index upon urb unlinking

[Oliver Neukum]

On Sa, 2018-09-22 at 22:11 +0200, Tobias Herzog wrote:
> Resetting the write index of the notification buffer on urb unlink (e.g.
> closing a cdc-acm device from userspace) may lead to wrong interpretation
> of further received notifications, in case the index is not 0 when urb
> unlink happens (i.e. when parts of a notification already have been
> transferred). On the device side there is no "reset" of the notification
> transimission and thus we would get out of sync with the device.

Hi,

you are right, when you are describing the problem. The fix seems to me
to fix a part, albeit a large part of the issue. As far as I can tell,
the transfer could be partially succesful, meaning that the driver
loses a part in the middle of the notification.

	Regards
		Oliver

[1/2] cdc-acm: do not reset notification buffer index upon urb unlinking

[Tobias Herzog]

Resetting the write index of the notification buffer on urb unlink (e.g.
closing a cdc-acm device from userspace) may lead to wrong interpretation
of further received notifications, in case the index is not 0 when urb
unlink happens (i.e. when parts of a notification already have been
transferred). On the device side there is no "reset" of the notification
transimission and thus we would get out of sync with the device.

Signed-off-by: Tobias Herzog <t-herzog@gmx.de>
---
 drivers/usb/class/cdc-acm.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/drivers/usb/class/cdc-acm.c b/drivers/usb/class/cdc-acm.c
index 27346d6..50339ca 100644
--- a/drivers/usb/class/cdc-acm.c
+++ b/drivers/usb/class/cdc-acm.c
@@ -355,7 +355,6 @@ static void acm_ctrl_irq(struct urb *urb)
 	case -ENOENT:
 	case -ESHUTDOWN:
 		/* this urb is terminated, clean up */
-		acm->nb_index = 0;
 		dev_dbg(&acm->control->dev,
 			"%s - urb shutting down with status: %d\n",
 			__func__, status);