On 29 November 2008 11:10:38 GWater wrote: > Vasily Khoruzhick schrieb: > > On 28 November 2008 22:12:07 GWater wrote: > >> Oh, > >> I never got that patch I think. (Otherwise I would have pushed it ;) .) > > > > Patch attached :) > > > >> Anyway - it is correct that we currently don't have any more functions > >> to cover with V4L2 CIDs but this is easy to change. I still can only > >> think of black-level calibration but if we go over the datasheets we may > >> find some other expert stuff. > > > > Sounds OK, btw it seems that datasheet for my ov9650 lies too often, even > > for gain registers :( so I'm afraid I will not be very usefull here. > > > > First thing we should think about is how to add V4L2-controls specific > > for sensors, but not globally. > > > >> I just keep talking about these things because I really like madwifi and > >> these guys also implemented everything they could find and a bit more. > >> > >> Makes it easier for the users to decide they want and does not bind them > >> to the default settings some developer hardcoded into the driver. > >> > >> GWater > >> > >> BTW The AUTO-exposure CID I put in there only works for >=2.6.25 . I'll > >> try to fix this. > >> GWater > > > > Vasily > > I tested your patch and while it basically change exposure it also made > some problems. Mostly because the registers of the sensorsare not the same. > (Do you have all the omnivision datasheets - if not I can mailthem to > you, including one that has always worked for SOI968.) > > So my question is - did you check whether your function ov_set_gain > actually works for all these sensors? If not I would remove the "links". > > GWater > > I attached a changed version of your patch that works for me. > > GWater
It should work at least for ov96xx and ov76xx sensors, I have no datasheet for SOI968, so I've assumed it's the same
signature.asc
Description: This is a digitally signed message part.
