Hi,

On 7/28/20 12:25 AM, Laurent Pinchart wrote:

Hi Hans,

Thank you for the patch.

On Fri, Jul 24, 2020 at 01:48:23PM +0200, Hans de Goede wrote:

...

uvc_ctrl_add_info() calls uvc_ctrl_get_flags() which will override the fixed-up flags set by uvc_ctrl_fixup_xu_info().

This commit fixes this by adding a is_xu argument to uvc_ctrl_add_info() and skipping the uvc_ctrl_get_flags() call for xu ctrls.

Note that the xu path has already called uvc_ctrl_get_flags() from uvc_ctrl_fill_xu_info() before calling uvc_ctrl_add_info().

This fixes the xu motor controls not working properly on a Logitech 046d:08cc, and presumably also on the other Logitech models which have a quirk for this in the uvc_ctrl_fixup_xu_info() function.

Cc: stable@vger.kernel.org Signed-off-by: Hans de Goede hdegoede@redhat.com

---

drivers/media/usb/uvc/uvc_ctrl.c | 11 ++++++----- 1 file changed, 6 insertions(+), 5 deletions(-)

diff --git a/drivers/media/usb/uvc/uvc_ctrl.c b/drivers/media/usb/uvc/uvc_ctrl.c index e399b9fad757..4bdea5814d6a 100644 --- a/drivers/media/usb/uvc/uvc_ctrl.c +++ b/drivers/media/usb/uvc/uvc_ctrl.c @@ -1815,7 +1815,7 @@ static int uvc_ctrl_fill_xu_info(struct uvc_device *dev, } static int uvc_ctrl_add_info(struct uvc_device *dev, struct uvc_control *ctrl,

• const struct uvc_control_info *info);

• const struct uvc_control_info *info, bool is_xu);

static int uvc_ctrl_init_xu_ctrl(struct uvc_device *dev, struct uvc_control *ctrl) @@ -1830,7 +1830,7 @@ static int uvc_ctrl_init_xu_ctrl(struct uvc_device *dev, if (ret < 0) return ret;

• ret = uvc_ctrl_add_info(dev, ctrl, &info);

• ret = uvc_ctrl_add_info(dev, ctrl, &info, true); if (ret < 0) uvc_trace(UVC_TRACE_CONTROL, "Failed to initialize control " "%pUl/%u on device %s entity %u\n", info.entity,

@@ -2009,7 +2009,7 @@ int uvc_ctrl_restore_values(struct uvc_device *dev)

• Add control information to a given control.

*/ static int uvc_ctrl_add_info(struct uvc_device *dev, struct uvc_control *ctrl,

• const struct uvc_control_info *info)

• const struct uvc_control_info *info, bool is_xu) { int ret = 0;

@@ -2029,7 +2029,8 @@ static int uvc_ctrl_add_info(struct uvc_device *dev, struct uvc_control *ctrl, * default flag values from the uvc_ctrl array when the device doesn't * properly implement GET_INFO on standard controls. */

• uvc_ctrl_get_flags(dev, ctrl, &ctrl->info);

• if (!is_xu)

• uvc_ctrl_get_flags(dev, ctrl, &ctrl->info);
  

Would it make sense to instead move this line (and the above comment) to uvc_ctrl_init_ctrl(), right after the uvc_ctrl_add_info() call ?

I was thinking about the same lines, since that seems like the obvious / logical fix. I was thinking about doing it before the uvc_ctrl_add_info() call, but that will not work, because info is const before the call.

Doing it after the add_info() call, as you suggest, we can pass &ctrl->info to the get_flags call. So yes that will work and is the logical thing to do.

I'll spin a v2 with this change.

If you would prefer keeping it here, I think is_xu should be renamed to update_flags (with the true/false values switched) to make it clearer. It would then also add a comment to uvc_ctrl_init_xu_ctrl() right before the call to uvc_ctrl_add_info() to state that we don't update flags to avoid overwriting the value set by uvc_ctrl_fixup_xu_info() in uvc_ctrl_fill_xu_info().

What's your preference ?

See above.

Regards,

Hans

...

ctrl->initialized = 1; @@ -2252,7 +2253,7 @@ static void uvc_ctrl_init_ctrl(struct uvc_device *dev, struct uvc_control *ctrl) for (; info < iend; ++info) { if (uvc_entity_match_guid(ctrl->entity, info->entity) && ctrl->index == info->index) {

• 	uvc_ctrl_add_info(dev, ctrl, info);
  

• 	uvc_ctrl_add_info(dev, ctrl, info, false);
  break;
  

  } }