[PATCH] USB: g_printer: fix handling zero-length packet

[PATCH] USB: g_printer: fix handling zero-length packet

[sangsu]

g_printer doesn't have to check whether the data size is a multiple of
MaxPacketSize, because device controller driver already make that check.

Signed-off-by: SangSu Park<sangsu@gmail.com>
---
drivers/usb/gadget/printer.c | 3++-
1 files changed, 2 insertion(+), 1 deletions(-)


--- a/drivers/usb/gadget/printer.c	2008-08-12 21:55:39.000000000 -0400
+++ b/drivers/usb/gadget/printer.c	2008-09-19 20:39:44.000000000 -0400
@@ -1264,8 +1264,7 @@
	/* respond with data transfer before status phase? */
	if (value >= 0) {
		req->length = value;
-		req->zero = value < wLength
-				&& (value % gadget->ep0->maxpacket) == 0;
+		req->zero = value < wLength;
		value = usb_ep_queue(gadget->ep0, req, GFP_ATOMIC);
		if (value < 0) {
			DBG(dev, "ep_queue --> %d\n", value);

Re: [PATCH] USB: g_printer: fix handling zero-length packet

[Alan Stern]

On Fri, 19 Sep 2008, sangsu wrote:

> g_printer doesn't have to check whether the data size is a multiple of
> MaxPacketSize, because device controller driver already make that check.
> 
> Signed-off-by: SangSu Park<sangsu@gmail.com>
> ---
> drivers/usb/gadget/printer.c | 3++-
> 1 files changed, 2 insertion(+), 1 deletions(-)
> 
> 
> --- a/drivers/usb/gadget/printer.c	2008-08-12 21:55:39.000000000 -0400
> +++ b/drivers/usb/gadget/printer.c	2008-09-19 20:39:44.000000000 -0400
> @@ -1264,8 +1264,7 @@
> 	/* respond with data transfer before status phase? */
> 	if (value >= 0) {
> 		req->length = value;
> -		req->zero = value < wLength
> -				&& (value % gadget->ep0->maxpacket) == 0;
> +		req->zero = value < wLength;
> 		value = usb_ep_queue(gadget->ep0, req, GFP_ATOMIC);
> 		if (value < 0) {
> 			DBG(dev, "ep_queue --> %d\n", value);

When posting changes like this you should add Greg KH to the CC: list;  
otherwise he won't realize this patch needs to be merged.

You should also CC: the author of the original code -- but in this case
you can't because Craig Nadler didn't include his email address in the
source file.

For what it's worth, this patch looks okay to me.

Alan Stern