Hi Sakari, On 22-09-19, Sakari Ailus wrote: > On Mon, Sep 19, 2022 at 08:37:23PM +0300, Laurent Pinchart wrote: > > > > > + ctrl = v4l2_ctrl_new_int_menu(&tc358746->ctrl_hdl, NULL, > > > > > + V4L2_CID_LINK_FREQ, 0, 0, > > > > > > > > Shouldn't the max argument be set to the number of items minus 1 ? > > > > > > Right now I would keep it that way since the driver only supports one > > > link-frequencies setting. So the ctrl don't let the userspace assume > > > that there are more than one link-frequency. > > > > Good point. Can you add a short comment above the call to explain this ? > > Wouldn't it be just easier to do what Laurent suggested originally? The end > result is the same, isn't it, and no comment needed? The end result depends on the device-tree given "link-frequencies" property. The driver currently takes only one frequency but the system-integrator of course can specify many more. In such case the 1st is used. If I go with Laurent's comment, all frequencies would be shown to the user-space but IMHO this shouldn't be the case since the driver supports only 1 frequency. > > > > > > > + link_frequencies); > > > > > + if (ctrl) > > > > > + ctrl->flags |= V4L2_CTRL_FLAG_READ_ONLY; > > Now that this is a bridge, this value presumably doesn't need to change. The value can change, e.g. if you have a few freq. to allow a wide range of frame sizes. This can be the case to conform the emv tests. So there can be a freq. for smaller frame sizes and one for larger frame sizes. > There will just be more blanking if the source sub-device pixel rate is > slower, right? Please see above. Regards, Marco
