hverkuil: I have another annoying question for you
when setting multiple controls with VIDIOC_S_EXT_CTRLS, the control framework validates the value of each of those controls before proceeding
modifying some controls may affect the min/max values for other controls
for instance, the EXPOSURE control of a sensor is bound by height + VBLANK, so modifying VBLANK changes the maximum value for EXPOSURE
if both VBLANK and EXPOSURE are set, in that order, EXPOSURE will be validated against the current maximum, based on the current VBLANK value, not the new one
is that something that could be fixed in the control framework ?
jernej would you please have a look at my cedrus patch if you have the time? I had your feedback in mind from last time but still no dice, meaning that only the first slice will decode right unfortunately :/
sometimes I am able to decode a few slices in a row, but the absolutely majority will fail with CEDRUS_IRQ_ERROR 
btw I checked the offsets and them seem ok, i.e. the same as I get on the userspace side 
hverkuil: any chance that you could have some time to take a look at the latest vb2 non-consistent series from Sergey?
it's been looking fine for me from the previous revision, but it seems like you had some comments
hverkuil: I've asked this last night, but you got disconnected with a ping timeout right after so I'm not sure you have received the messages
when setting multiple controls with VIDIOC_S_EXT_CTRLS, the control framework validates the value of each of those controls before proceeding
modifying some controls may affect the min/max values for other controls 
for instance, the EXPOSURE control of a sensor is bound by height + VBLANK, so modifying VBLANK changes the maximum value for EXPOSURE 
if both VBLANK and EXPOSURE are set, in that order, EXPOSURE will be validated against the current maximum, based on the current VBLANK value, not the new one
is that something that could be fixed in the control framework ?
tfiga: the last series I saw was v2, and I commented on that. I do see a reply from Sergey on 25/06 that I didn't answer (it's been quite busy lately). I'll try to reply early next week.