[U-Boot] [PATCH v1 0/9] sunxi: initial upstreamining effort

[Hans de Goede]

Hi,

First of all: Ian: Many many thanks for working on this, you rock!

On 03/14/2014 04:01 PM, Albert ARIBAUD wrote:
> Hi Tom,
> 
> On Fri, 14 Mar 2014 10:19:27 -0400, Tom Rini <trini at ti.com> wrote:
> 
>> OK good, I think Albert and I might end up starting the sun4i/sun5i
>> stuff as those are the boards we've got sitting around.
> 
> As I have a Mele A1000G (A31), that would be sun6i, not sun4i/sun5i

So I'm very interested in getting sunxi support in upstream u-boot too, and
I've a ton of sunxi devices. I'll start working on doing a sun4i patch-set
on top of Ian's sun7i patch-set. Then Tom can start working on sun5i.

Tom, since I've started working on sunxi stuff 2 years ago I've gained a lot
of experience with these little SoCs so let me know if you hit some roadblock
and I can take a look.

Disclaimer I'm more of a kernel / device driver guy, all the low-level
RAM init (and having to think about where in memory to put all the bits)
is mostly new, so have some patience with me please :)

> and yes, I would like to contribute support for it.

If someone can start working on sun6i support that would be awesome.

My sunxi-next u-boot branch:
https://github.com/jwrdegoede/u-boot-sunxi/commits/sunxi-next

Has some minimal patches by Maxime to get u-boot to run as a secondary
boot-loader, using allwinner's boot0 and boot1 as primary loader.

What we really want is to get a working SPL for sun6i, there are 2 ways
to get to a working SPL:

1) Use allwinners special USB boot mode called FEL, this way no MMC / NAND
reading code is necessary.
2) First get MMC to work using uboot as a secondary loader only, and
once we've working MMC code try to get a SPL which can read the
real u-boot directly from MMC.

And then for 2, there is the option to first get MMC for sun6i working in the
upstream Linux kernel, so that we know exactly what changes are needed compared
to sun5i / sun7i MMC (which are identical).

I've done a lot of work on the upstream kernel mmc host driver, and I actually
wanted to start looking into MMC for sun6i in the upstream kernel one of these
days. But now that Ian has published his awesome work to get sunxi support upstream,
my first priority is to get sun4i + sun5i support added on top of Ian's patches.

Regards,

Hans