On Mon, Sep 21, 2020 at 11:03:42PM +0900, Yasushi Asano wrote:
> From: Yasushi Asano <yasano@xxxxxxxxxxxxxx>
>
> Dear Alan
>
> Thank you very much for providing the patch.
> I really appreciate your kindness.
>
> When I added a pseudo error code and checked it,
> the for statement of the "operation" in the new scheme
> runs unconditionally three times. Therefore It doesn't
> seem to meet the requirements (30seconds).
>
> After applying your patch, I added a patch that can
> change the loop number of "operation" as shown below,
> and it worked fine in the pseudo error environment.
> Is this modification acceptable?
Ah, I missed that change. Yes, this is the right approach.
> If it is good, I'll add this patch and test it.
> Since the PET tool is only in the customer,
> I will ask the customer to test it. I will report the
> result when I receive the results.
>
> Signed-off-by: Yasushi Asano <yasano@xxxxxxxxxxxxxx>
> ---
> drivers/usb/core/hub.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/usb/core/hub.c b/drivers/usb/core/hub.c
> index 61effd5..2f07f7c 100644
> --- a/drivers/usb/core/hub.c
> +++ b/drivers/usb/core/hub.c
> @@ -2709,12 +2709,14 @@ static unsigned hub_is_wusb(struct usb_hub *hub)
> #define PORT_RESET_TRIES 2
> #define SET_ADDRESS_TRIES 1
> #define GET_DESCRIPTOR_TRIES 1
> +#define GET_DESCRIPTOR_OPERATIONS 1
> #define PORT_INIT_TRIES 5
>
> #else
> #define PORT_RESET_TRIES 5
> #define SET_ADDRESS_TRIES 2
> #define GET_DESCRIPTOR_TRIES 2
> +#define GET_DESCRIPTOR_OPERATIONS 3
> #define PORT_INIT_TRIES 4
> #endif /* CONFIG_USB_FEW_INIT_RETRIES */
>
> @@ -4699,7 +4701,7 @@ hub_port_init(struct usb_hub *hub, struct usb_device *udev, int port1,
> * 255 is for WUSB devices, we actually need to use
> * 512 (WUSB1.0[4.8.1]).
> */
> - for (operations = 0; operations < 3; ++operations) {
> + for (operations = 0; operations < GET_DESCRIPTOR_OPERATIONS; ++operations) {
> buf->bMaxPacketSize0 = 0;
> r = usb_control_msg(udev, usb_rcvaddr0pipe(),
> USB_REQ_GET_DESCRIPTOR, USB_DIR_IN,
Okay. I will merge your change into my patch.
Alan Stern
