On Mon, Jun 18, 2012 at 04:31:36PM +0530, Pratyush Anand wrote: > On 6/12/2012 4:17 PM, Felipe Balbi wrote: > >correct. Gadget driver should dequeue each request, but the fact is that > >stop transfer command should be sent only once, no matter how many TRBs > >we queued together, so that'll be a bit tricky, I guess. > > We still need to have solution for it. I do not see any straight > forward solution for avoiding calling of stop transfer if more than > one request is in req_queued list. Do you see something? > > I think WARN_ON(!dep->res_trans_idx); in dwc3_stop_active_transfer is > not needed now. What do you say? probably not needed anymore. Indeed. We know very well how those things are wired up ;-) -- balbi
Attachment:
signature.asc
Description: Digital signature
