diff mbox

BugFix in XHCI controller driver for scatter gather DMA

Message ID 1450699494-4598-1-git-send-email-vikas.bansal@samsung.com
State Changes Requested, archived
Delegated to: David Miller
Headers show

Commit Message

Vikas Bansal Dec. 21, 2015, 12:04 p.m. UTC 
Pre-Condition
At the time of reset resume of a USB device, both self and bus powered devices might go to a low power state or power off state depending on the acceptable suspend time power of the system.    
In case the device experiences a power glitch or completely powers off during suspend-resume, the device will lose its internal state and hence it'll again need a set interface from class driver on reset resume operation.

Issue 
So far our experiments were based on a USB gadget working on cdc_eem protocol. 
We have seen that device is unable to continue the packet transfer on bulk endpoints after the reset resume operation.

Solution
We have added a .reset_resume function for cdc_eem protocol which sends a set interface command on the Control endpoint. And calls the existing usbnet_resume thereafter


Signed-off-by: Vikas Bansal <vikas.bansal@samsung.com>
Signed-off-by: Sumit Batra <sumit.batra@samsung.com>
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Comments

Oliver Neukum Dec. 21, 2015, 12:50 p.m. UTC | #1 
On Mon, 2015-12-21 at 17:34 +0530, Vikas Bansal wrote:
> Pre-Condition
> At the time of reset resume of a USB device, both self and bus powered devices might go to a low power state or power off state depending on the acceptable suspend time power of the system.    
> In case the device experiences a power glitch or completely powers off during suspend-resume, the device will lose its internal state and hence it'll again need a set interface from class driver on reset resume operation.
> 
> Issue 
> So far our experiments were based on a USB gadget working on cdc_eem protocol. 
> We have seen that device is unable to continue the packet transfer on bulk endpoints after the reset resume operation.
> 
> Solution
> We have added a .reset_resume function for cdc_eem protocol which sends a set interface command on the Control endpoint. And calls the existing usbnet_resume thereafte

Hi,

something is wrong with your subject line. Could you fix
that and resend?

	Regards
		Oliver


--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
diff mbox

Patch

diff --git a/drivers/net/usb/cdc_eem.c b/drivers/net/usb/cdc_eem.c
index f7180f8..1f6f7ea 100644
--- a/drivers/net/usb/cdc_eem.c
+++ b/drivers/net/usb/cdc_eem.c
@@ -342,6 +342,19 @@  next:
 	return 1;
 }
 
+static int cdc_eem_resume(struct usb_interface *intf)
+{
+	int ret = 0;
+	struct usbnet *dev = usb_get_intfdata(intf);
+
+	ret = usbnet_get_endpoints(dev, intf);
+	if(ret < 0)
+		goto err;
+	ret = usbnet_resume(intf);
+err:
+	return ret;
+}
+
 static const struct driver_info eem_info = {
 	.description =	"CDC EEM Device",
 	.flags =	FLAG_ETHER | FLAG_POINTTOPOINT,
@@ -371,6 +384,7 @@  static struct usb_driver eem_driver = {
 	.disconnect =	usbnet_disconnect,
 	.suspend =	usbnet_suspend,
 	.resume =	usbnet_resume,
+	.reset_resume =	cdc_eem_resume,
 	.disable_hub_initiated_lpm = 1,
 };