[PATCH v6 17/22] usb: dwc2: host: Manage frame nums better in scheduler

Doug Anderson dianders at chromium.org
Wed Feb 3 12:29:55 PST 2016


On Thu, Jan 28, 2016 at 6:20 PM, Douglas Anderson <dianders at chromium.org> wrote:
>  static void dwc2_qh_init(struct dwc2_hsotg *hsotg, struct dwc2_qh *qh,
>                          struct dwc2_hcd_urb *urb)
>  {
> @@ -569,11 +655,6 @@ static void dwc2_qh_init(struct dwc2_hsotg *hsotg, struct dwc2_qh *qh,
>                               qh->ep_type == USB_ENDPOINT_XFER_ISOC,
>                               bytecount));
> -               /* Ensure frame_number corresponds to the reality */
> -               hsotg->frame_number = dwc2_hcd_get_frame_number(hsotg);

In reviewing patches I realized that this is actually a revert of
commit dd81dd7c8178 ("usb: dwc2: host: use correct frame number during
qh init").  IMHO that patch was wrong: hsotg->frame_number is supposed
to be the frame number as of the last start of frame.  If we need to
know a more recent frame number then we should query it ourselves.

Presumably the reason for the original patch was to try to fix some of
the same problems I've addressed in my series, so I'd presume that
this doesn't add any new regressions.  I haven't heard much from
Gregory Herrero about my series, but it would be nice to confirm that
this virtual revert wasn't causing problems.


More information about the linux-rpi-kernel mailing list