On Thu, 1 Feb 2018, Peter Chen wrote:
> At former code, the SETUP stage does not enable interrupt
> for qtd completion, it relies on IAA watchdog to complete
> interrupt, then the transcation would be considered timeout
> if the flag need_io_watchdog is cleared by platform code.
>
> In this commit, we always add enable interrupt for qtd completion,
> then the qtd completion can be notified by hardware interrupt.
>
> Signed-off-by: Peter Chen <peter.c...@nxp.com>
> ---
> drivers/usb/host/ehci-q.c | 6 +++---
> 1 file changed, 3 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/usb/host/ehci-q.c b/drivers/usb/host/ehci-q.c
> index 8815832..4cc2609 100644
> --- a/drivers/usb/host/ehci-q.c
> +++ b/drivers/usb/host/ehci-q.c
> @@ -1192,6 +1192,8 @@ static int submit_single_step_set_feature(
> qtd_fill(ehci, qtd, urb->setup_dma,
> sizeof(struct usb_ctrlrequest),
> token | (2 /* "setup" */ << 8), 8);
> + /* Always enable interrupt on qtd completion */
> + qtd->hw_token |= cpu_to_hc32(ehci, QTD_IOC);
This would be cleaner if you didn't make it a separate statement.
Before the call to qtd_fill(), you can do:
token |= QTD_IOC; /* Interrupt after SETUP completion */
>
> submit_async(ehci, urb, &qtd_list, GFP_ATOMIC);
> return 0; /*Return now; we shall come back after 15 seconds*/
> @@ -1231,9 +1233,7 @@ static int submit_single_step_set_feature(
> /* dont fill any data in such packets */
> qtd_fill(ehci, qtd, 0, 0, token, 0);
>
> - /* by default, enable interrupt on urb completion */
> - if (likely(!(urb->transfer_flags & URB_NO_INTERRUPT)))
> - qtd->hw_token |= cpu_to_hc32(ehci, QTD_IOC);
> + qtd->hw_token |= cpu_to_hc32(ehci, QTD_IOC);
This could be improved in the same way.
Alan Stern
>
> submit_async(ehci, urb, &qtd_list, GFP_KERNEL);
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
