Re: [GIT PULL FOR v3.6] Add adv7604/ad9389b drivers

[Hans Verkuil <hverkuil@xs4all.nl>]

On Mon July 23 2012 14:25:38 Laurent Pinchart wrote:
> Hi Hans,
> 
> On Monday 23 July 2012 13:36:15 Hans Verkuil wrote:
> > Hi all!
> > 
> > There haven't been any comments since either RFCv1 or RFCv2.
> > 
> > (http://www.spinics.net/lists/linux-media/msg48529.html and
> > http://www.spinics.net/lists/linux-media/msg50413.html)
> > 
> > So I'm making this pull request now.
> > 
> > The only changes since RFCv2 are some documentation fixes:
> > 
> > - Add a note that the SUBDEV_G/S_EDID ioctls are experimental
> > - Add the proper revision/experimental references.
> > - Update the spec version to 3.6.
> 
> Jumping a bit late on this. Wouldn't it be good to submit the RFCs to the dri-
> devel mailing list before pushing them upstream ? They have been dealing with 
> EDID for ages and might offer good advices.

I've looked at DRM as well, but there is really nothing to share. This API is for
embedded systems: it does *not* parse the EDID, it just transports it from kernel
to userspace and vice versa.

The main difference is that DRM attempts to load all blocks from the EDID, whereas
this API allows you to load blocks on demand (actually, drivers can still decide
to load all blocks initially, but that's an implementation detail). This is more
efficient given the slow i2c bus.

Regards,

	Hans