On Tue, Nov 08, 2022, Michael Grzeschik wrote:
> Since the patch b44c0e7fef51 ("usb: dwc3: gadget: conditionally remove
> requests") the driver is now returning different request status values
> in different conditions of disabled endpoint and stopped transfers.
> The patch however has swapped the actual status results they should
> have. We fix this now with this patch.
>
> Fixes: b44c0e7fef51 ("usb: dwc3: gadget: conditionally remove requests")
> Cc: stable@xxxxxxxxxxxxxxx
> Signed-off-by: Michael Grzeschik <m.grzeschik@xxxxxxxxxxxxxx>
> ---
>
> I had this patch prepared already, but forgot about it. Sorry about the
> delay. Thanks for catching up on this, Thinh!
>
> mgr
NP, I wasn't sure you would prepare it any time soon. It breaks some of
our tests, so I went ahead and pushed the fix. Greg already picked it
up.
>
> drivers/usb/dwc3/gadget.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/usb/dwc3/gadget.c b/drivers/usb/dwc3/gadget.c
> index ecddb144871bec..cd5ace438d40ee 100644
> --- a/drivers/usb/dwc3/gadget.c
> +++ b/drivers/usb/dwc3/gadget.c
> @@ -1029,7 +1029,7 @@ static int __dwc3_gadget_ep_disable(struct dwc3_ep *dep)
> dep->endpoint.desc = NULL;
> }
>
> - dwc3_remove_requests(dwc, dep, -ECONNRESET);
> + dwc3_remove_requests(dwc, dep, -ESHUTDOWN);
>
> dep->stream_capable = false;
> dep->type = 0;
> @@ -2375,7 +2375,7 @@ static void dwc3_stop_active_transfers(struct dwc3 *dwc)
> if (!dep)
> continue;
>
> - dwc3_remove_requests(dwc, dep, -ESHUTDOWN);
> + dwc3_remove_requests(dwc, dep, -ECONNRESET);
> }
> }
That doesn't look right. What about soft-disconnect? What conditions
that dwc3_stop_active_transfers() is called that's not supposed to
return -ESHUTDOWN?
Thanks,
Thinh
