Re: [PATCH] usb: gadget: dummy: don't disconnect on gadget removal

[Alan Stern <stern@xxxxxxxxxxxxxxxxxxx>]

On Thu, 19 Apr 2012, Alexander Shishkin wrote:

> I'm using linux-next, and as of today, the call order is still like I
> listed before. The change that introduced this order is 8ae8090c (usb:
> gadget: udc-core: fix asymmetric calls in remove_driver), which is also
> present in "usb-linus" branch of linux-usb tree (that's how, I assume
> it's got to linux-next) and Felipe's "fixes" and "master".
> 
> So no, I didn't make this up. ;)
> 
> The change seems logical though: udc_stop() may have disabled the usb
> controller, and calling pullup() after that doesn't look very useful.

Have you tried out this patch?  After I upgraded to 3.4-rc4 I 
encountered the same problem as you did, and the patch fixed it.

The two disconnect calls may be redundant; if the UDC driver implements
a pullup method then setting the pullup to 0 ought to cause the gadget
driver's disconnect method to be called.

Alan Stern



Index: usb-3.4/drivers/usb/gadget/udc-core.c
===================================================================
--- usb-3.4.orig/drivers/usb/gadget/udc-core.c
+++ usb-3.4/drivers/usb/gadget/udc-core.c
@@ -263,8 +263,8 @@ static void usb_gadget_remove_driver(str
 
 	if (udc_is_newstyle(udc)) {
 		udc->driver->disconnect(udc->gadget);
-		udc->driver->unbind(udc->gadget);
 		usb_gadget_disconnect(udc->gadget);
+		udc->driver->unbind(udc->gadget);
 		usb_gadget_udc_stop(udc->gadget, udc->driver);
 	} else {
 		usb_gadget_stop(udc->gadget, udc->driver);

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