Hi, On Tue, Jun 12, 2012 at 03:29:53PM +0530, Pratyush Anand wrote: > Hi Felipe, > > I see one more and hopefully last issue with multiple ways of calling > stop active transfer which might generate annoying warnings. > > It is called from ep_dequeue and ep_disable both. indeed. Great catch!!! > Do you see an optimal way to resolve this? I guess we could just assume gadget driver is a nice citizen and it doesn't try to disable an EP before killing the requests ? A quick look at the gadget drivers shows that they dequeue requests much before they disable endpoints. Maybe that's all we need to do then ? Could you cook a patch and test it out ? -- balbi
Attachment:
signature.asc
Description: Digital signature
