[RFC] soc-camera: remove device and bus abstractions

[RFC] soc-camera: remove device and bus abstractions

[Guennadi Liakhovetski]

I'm currently working on removing "struct device" from "struct 
soc_camera_device" and the "struct bus_type soc_camera_bus_type" bus. 
Reasoning - well, I've just got enough of them;-) Everything, what they 
are currently providing, can also be implemented directly. They served 
their purpose during the active soc-camera API development as a nice 
abstraction to build around, as debugging means, etc. But not, that also 
subdevices have got their own "struct video_device" and with "struct 
device" nodes embedded, I thi, we in v4l2 have (more than) enough devices 
to satisfy all sysfs fans;-)

So, this is just to let everyone know about my plans, don't know whether I 
manage it for 3.1, or will have to postpone to 3.2. In any case if anyone 
has any principle objections against this plan, please shout now!

Thanks
Guennadi
---
Guennadi Liakhovetski, Ph.D.
Freelance Open-Source Software Developer
http://www.open-technology.de/

Re: [RFC] soc-camera: remove device and bus abstractions

[Mauro Carvalho Chehab]

Em 15-07-2011 12:49, Guennadi Liakhovetski escreveu:
> I'm currently working on removing "struct device" from "struct 
> soc_camera_device" and the "struct bus_type soc_camera_bus_type" bus. 
> Reasoning - well, I've just got enough of them;-) Everything, what they 
> are currently providing, can also be implemented directly. They served 
> their purpose during the active soc-camera API development as a nice 
> abstraction to build around, as debugging means, etc. But not, that also 
> subdevices have got their own "struct video_device" and with "struct 
> device" nodes embedded, I thi, we in v4l2 have (more than) enough devices 
> to satisfy all sysfs fans;-)

Very welcome plan.
> 
> So, this is just to let everyone know about my plans, don't know whether I 
> manage it for 3.1, or will have to postpone to 3.2. In any case if anyone 
> has any principle objections against this plan, please shout now!

It is probably too late for 3.1, as the merge window will likely start next
week. So, all patches for 3.1 are likely already submitted.

Cheers,
Mauro.