[PATCH] [RFC PATCH] DRM: add DRM Driver for Samsung SoC EXYNOS4210.

[Sascha Hauer]

On Mon, Aug 01, 2011 at 12:52:37PM +0900, daeinki wrote:
> Hi, Sascha Hauer.
> thank you for your comments and below is my answer.
> 
> Sascha Hauer wrote:
> >Hi,
> >
> >On Fri, Jul 29, 2011 at 04:24:35PM +0900, Inki Dae wrote:
> >>This patch is a DRM Driver(only including FIMD Driver yet)
> >>for Samsung SoC Exynos4210. and as RFC, I am sending only DRM driver part.
> >>
> >>this patch is based on git repository below:
> >>git://git.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6.git,
> >>branch: drm-next
> >>commit-id: 5a96a899bbdee86024ab9ea6d02b9e242faacbed
> >>
> >>We tried to re-use lowlevel codes of the FIMD driver(s3c-fb.c
> >>based on Linux framebuffer) but couldn't so because lowlevel codes
> >>of s3c-fb.c are included internally and so this driver shares only
> >>platform device.
> >>
> >>Sub drivers such as fimd or hdmi have indenpendent platform device and
> >>Platform driver and when driver's probe is called, the driver object
> >>including callbacks(for hardware control) would be registered to
> >>Samsung drm driver. and then when samsung drm driver is probed,
> >>each probe callback of the driver object registered is called so that
> >>additional callbacks for drm framework would be set at this time.
> >>
> >>We used GEM framework for buffer management and this driver supports
> >>only physically continuous memory yet(non-iommu). and for buffer allocation,
> >>we used DMA APIs(dma_alloc_writecombine) but we will change it to CMA instead
> >>of DMA APIs later.
> >>
> >>Refer to this link for CMA(Continuous Memory Allocator):
> >>http://lkml.org/lkml/2011/7/20/45
> >>
> >>Future works:
> >>- HDMI support.
> >>- drm plane feature support.
> >>  refer to this link for drm plane feature:
> >>  http://www.spinics.net/lists/dri-devel/msg11778.html
> >>- change the allocator to CMA.
> >>- iommu support.(for non-continuous physical memory usage)
> >>- fimd driver update.
> >>- add exception codes and code clean.
> >>
> >>to support all features above, we need long time and hard work.
> >>so we wish that only some features(fimd and non-iommu) are applied to
> >>mainline first.
> >>
> >>We would be pleased you to give us your comments.
> >
> >So far I only had a quick look over the driver. You might know that
> >I wrote a Freescale i.MX drm driver which I posted to the list some time
> >ago.
> >
> >My driver lacks GEM support which your driver has, so I specifically
> >looked at this part. It seems we could reuse the GEM code on i.MX and
> >probably on most other ARMs aswell. Can you split out this code
> >and remove the samsung_ namespace?
> >
> By any chance, you mean drm_gem_free_mmap_offset()?
> this patch had already been posted by Rob Clark, Omap TI as RFC.
> 
> you can refer to this link for it.
> http://www.spinics.net/lists/dri-devel/msg13018.html
> 
> if so, I am aware of it but his patch isn't applied to drm-next yet
> and so my drm driver doesn't include his patch. of course I will
> reuse it and remove the samsung_ namespace as you pointed out if the
> patch is applied to drm-next.

Not only. I meant the full content of samsaung_drm_gem.c,
samsung_drm_buf.c and maybe even samsung_drm_fb.c.

Except for the samsung namespace these functions could look the same
on i.MX and other dumb framebuffer drm drivers.

Sascha

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |