On Monday 24 May 2010 11:44:16 Hans Verkuil wrote: > Hi Laurent, > > Thanks for your review! As always, it was very useful. I've incorporated most points in my ctrlfw3 branch: http://git.linuxtv.org/hverkuil/v4l-dvb.git?a=shortlog;h=refs/heads/ctrlfw3 Main changes: - Replaced 'bridge driver' by 'V4L2 driver'. - Added is_volatile and is_uninitialized flags (to control whether g_volatile_ctrl or init should be called). - Added is_volatile, is_uninitialized and is_private flags to v4l2_ctrl_config. - If the name field in struct v4l2_ctrl_config is NULL, then assume it is a standard control and fill in the defaults accordingly. These two changes together make it possible to make an array of struct v4l2_ctrl_config to create all controls. Perhaps v4l2_ctrl_new_custom should be renamed to v4l2_ctrl_new? - v4l2_ctrl_new_std_menu now has a 'max' argument. - v4l2_ctrl_new_std can no longer be used to create a standard menu control. This should prevent confusion regarding step and skip_mask. - v4l2_ctrl_active and v4l2_ctrl_grab set the flag atomically, so can be called from anywhere. I did not yet change anything regarding the init return type. I'm really not sure what userspace is supposed to do here. If you fail initializing a control, does that mean that -EIO should be returned? The only way this can fail is if there is a hardware problem, right? Can you give some background info on how this is currently handled in uvc? Regards, Hans -- Hans Verkuil - video4linux developer - sponsored by TANDBERG, part of Cisco -- To unsubscribe from this list: send the line "unsubscribe linux-media" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
