On Mon, Jul 11, 2011 at 5:31 PM, Felipe Balbi <balbi@xxxxxx> wrote:
> Hi,
>
> On Mon, Jul 11, 2011 at 05:18:18PM +0800, Bob Liu wrote:
>> g_zero test during loopback mode will fail like this:
>> zero gadget: can't loop ep6out to ep5in: -22
>>
>> The reason is this check "if (request->ep != musb_ep)" in musb_gadget,
>> this patch rm it.
>>
>> Signed-off-by: Bob Liu <lliubbo@xxxxxxxxx>
>> ---
>> drivers/usb/musb/musb_gadget.c | 3 ---
>> 1 files changed, 0 insertions(+), 3 deletions(-)
>>
>> diff --git a/drivers/usb/musb/musb_gadget.c b/drivers/usb/musb/musb_gadget.c
>> index b67a062..7318a20 100644
>> --- a/drivers/usb/musb/musb_gadget.c
>> +++ b/drivers/usb/musb/musb_gadget.c
>> @@ -1289,9 +1289,6 @@ static int musb_gadget_queue(struct usb_ep *ep, struct usb_request *req,
>> request = to_musb_request(req);
>> request->musb = musb;
>>
>> - if (request->ep != musb_ep)
>> - return -EINVAL;
>
> I believe the problem is somewhere else. If request->ep != musb_ep we
> would be queueing a request to the wrong endpoint. Is g_zero simply
Oh, I found after this patch
./testusb -t5 -c 2 -s 256 -D /proc/bus/usb/002/019
will hang although
./testusb -t5 -c 1 -s 256 -D /proc/bus/usb/002/019
can pass.
> re-using a request it just received from ep6out and queueing it straight
> away to ep5in ?? I mean.. I'm not sure we should allow:
I think so, see loopback_complete in f_loopback.c
200 if (ep == loop->out_ep) {
201 /* loop this OUT packet back IN to the host */
202 req->zero = (req->actual < req->length);
203 req->length = req->actual;
204 status = usb_ep_queue(loop->in_ep, req, GFP_ATOMIC);
205 if (status == 0)
206 return;
207
208 /* "should never get here" */
209 ERROR(cdev, "can't loop %s to %s: %d\n",
210 ep->name, loop->in_ep->name,
211 status);
212 }
BTW: Would you please do g_zero testing on your platform to check
whether this error
"zero gadget: can't loop xxx to xxx: -22" also happen.
Thanks.
>
> usb_ep_alloc_request(ep6out, GFP_KERNEL);
> usb_ep_queue(ep5in, GFP_KERNEL);
>
> The request, clearly, wasn't queued for that endpoint.
>
> --
> balbi
>
--
Regards,
--Bob
--
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
