On Tue, 17 Jul 2012, Laurent Pinchart wrote:
> Most Logitech UVC webcams (both early models that don't advertise UVC
> compatibility and newer UVC-advertised devices) require the RESET_RESUME
> quirk. Instead of listing each and every model, match the devices based
> on the UVC interface information.
>
> Signed-off-by: Laurent Pinchart <[email protected]>
> ---
> drivers/usb/core/quirks.c | 86
> +++++++++++++--------------------------------
> 1 files changed, 24 insertions(+), 61 deletions(-)
>
> Alan, what do you think about this approach ? I'm not sure whether we need to
> include the early UVC models that advertise a vendor-specific class in the
> list.
The general approach is okay. The details aren't quite right.
Are the 0x08c2 - 0x08c7 products the ones with a vendor-specific class?
> @@ -153,31 +116,31 @@ static const struct usb_device_id usb_quirk_list[] = {
> /* INTEL VALUE SSD */
> { USB_DEVICE(0x8086, 0xf1a5), .driver_info = USB_QUIRK_RESET_RESUME },
>
> + /* Logitech UVC Cameras */
> + { .match_flags = USB_DEVICE_ID_MATCH_VENDOR
> + | USB_DEVICE_ID_MATCH_INT_INFO,
> + .idVendor = 0x046d,
> + .bInterfaceClass = USB_CLASS_VIDEO,
> + .bInterfaceSubClass = 1,
> + .bInterfaceProtocol = 0,
> + .driver_info = USB_QUIRK_RESET_RESUME },
> +
usb_match_device() doesn't pay attention to the MATCH_INT_INFO flag.
Therefore this entry would end up matching every Logitech device.
In fact, at the time usb_detect_quirks() gets called we haven't yet
read in the device's config descriptors, so matching based on interface
info won't work. It looks like we need a second routine,
usb_detect_interface_quirks(), which would be called at the end of
usb_enumerate_device(), along with a second id table.
Alan Stern
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
