Re: HDMI and Composite capture on Lager, for kernel 4.1, version 3

[Hans Verkuil <hverkuil@xs4all.nl>]

On 06/22/2015 04:41 PM, William Towle wrote:
> 
> Hi Hans,
>> I'm not sure if I've asked this before, but shouldn't soc-camera be extended
>> with support for the DV_TIMINGS ioctls in order to control the adv7604?
>>
>> It's peculiar that that is not included in this patch series...
> 
>    Rob tells me he did some work on this, but we don't have it passing
> muster with v4l2-compliance and since gstreamer tries a number of
> resolutions of its own accord and subsequently produces images and
> video without it we can make do without for our needs. I can include it
> in or alongside the next submission if you'd to see it.

Another reason not to use soc-camera.

It's really not a very suitable framework for this hardware.

> 
>    Regarding your other comments, in particular the specification of the
> following:
>  		.tdms_lock_mask = 0x43,
>  		.cable_det_mask = 0x01,
> ...when testing just the first input, these seem reasonable - in the
> case of tdms_lock_mask the two least significant bits represent
> V_LOCKED_RAW and DE_REGEN_LCK_RAW, and they are set in line with
> the value of TDMSPLL_LCK_A_RAW when queried.
> 
>    While an implementation suitable for testing both of the ADV7612's
> A and B inputs could just add TDMSPLL_LCK_B_RAW to tdms_lock mask,
> the cable detect function isn't so trivial: it would need to query two
> separate (and numerically non-adjacent) registers to do its job.
> 
>    For simplicity we would like to propose that our next iteration
> supports just the first input (which is sufficient for our current
> needs) and has commentary in appropriate places regarding its
> shortcomings, although we're open to alternative suggestions if there
> has already been discussion on the matter here.

I think supporting just the first input and commenting accordingly is the
best approach for now.

Regards,

	Hans
--
To unsubscribe from this list: send the line "unsubscribe linux-media" in