On Thu, 8 Jun 2017, Benjamin Herrenschmidt wrote:
> This looks wrong:
>
> #define EndpointRequest \
> ((USB_DIR_IN|USB_TYPE_STANDARD|USB_RECIP_INTERFACE)<<8)
> #define EndpointOutRequest \
> ((USB_DIR_OUT|USB_TYPE_STANDARD|USB_RECIP_INTERFACE)<<8)
>
> Shouldn't it be USB_RECIP_ENDPOINT ?
Yes, indeed it should. Looks like copy-and-paste errors.
> If I'm right, we had issues with some of the set/clear feature
> "len" on root hub, not that anybody noticed ... I'm using these
> definitions in the aspeed vhub though, and this breaks it, so
> in case I'm right:
>
> Subject: [PATCH] usb/hcd: Fix definition of Endpoint{out}Request
>
> The definition for these is incorrect. Thankfully the only user
> today is the root hub emulation code and only in corner cases
> that seem to have never hit anybody so far. My upcoming driver
> for the Aspeed virtual hub will rely on them being correct
> however.
>
> Signed-off-by: Benjamin Herrenschmidt <benh@xxxxxxxxxxxxxxxxxxx>
> --
>
> diff --git a/include/linux/usb/hcd.h b/include/linux/usb/hcd.h
> index 40edf6a..8a1552e 100644
> --- a/include/linux/usb/hcd.h
> +++ b/include/linux/usb/hcd.h
> @@ -561,9 +561,9 @@ extern void usb_ep0_reinit(struct usb_device *);
> ((USB_DIR_IN|USB_TYPE_STANDARD|USB_RECIP_INTERFACE)<<8)
>
> #define EndpointRequest \
> - ((USB_DIR_IN|USB_TYPE_STANDARD|USB_RECIP_INTERFACE)<<8)
> + ((USB_DIR_IN|USB_TYPE_STANDARD|USB_RECIP_ENDPOINT)<<8)
> #define EndpointOutRequest \
> - ((USB_DIR_OUT|USB_TYPE_STANDARD|USB_RECIP_INTERFACE)<<8)
> + ((USB_DIR_OUT|USB_TYPE_STANDARD|USB_RECIP_ENDPOINT)<<8)
>
> /* class requests from the USB 2.0 hub spec, table 11-15 */
> #define HUB_CLASS_REQ(dir, type, request) ((((dir) | (type)) << 8) |
> (request))
Acked-by: Alan Stern <stern@xxxxxxxxxxxxxxxxxxx>
--
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
