[U-Boot] [PATCH] sunxi: Machine id hack to prevent loading buggy sunxi-3.4 kernels

[Hans de Goede]

Hi,

On 21-02-15 11:26, Ian Campbell wrote:
> On Sat, 2015-02-21 at 10:41 +0100, Hans de Goede wrote:
>> By the right thing I mean having things just work for the end user,
>> whatever I'm doing having things just work is always my end goal.
>> This is also why I'm mostly focussing upstream, which has resulted in
>> Fedora 21 and the upcoming Debian release supporting allwinner devices
>> ootb (albeit only headless) and the upcoming F-22 release will have
>> much improved allwinner device support, supporting more devices, and
>> supporting video output, etc.
>
> Speaking of which, if someone were to identify a suitable set of
> simple-fb backports for 3.16 and they are reasonably self contained I'd
> happily backport them to the kenrel which is going to be in the next
> Debian release too...

You will need all patches by me and by "Luc Verhaegen" from here:
https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/log/drivers/video/fbdev/simplefb.c

As well as these dts patches:

https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit?id=5790d4ee1e425fefef0e012d1415c25baeb9dae8
https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit?id=d501841fc42a97a8b69449135eadf6e9c4589a87
https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit?id=e53a8b2201384ee1a2cfd1609c9ffdf951b041a1
https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit?id=8efc5c2be5b29f7ceec83bb471da81690273e141
https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit?id=a9f8cda32a3acc3150755fe973f7fa1867a7a5d6
https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit?id=8cedd6628472aa2d8f2f54dfcc36633e46a59db8
https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit?id=fd18c7eac03540654abdbfedd52e6dfb3718b489

Those combined with a new enough u-boot should get you video output on most devices
supported in upstream u-boot.

>> Ian, can you live with overloading the high machine-id bits to
>> deliberately break old kernels when not compiled with old kernel
>> compatibility?
>
> It kind of sucks, but so do all the options it seems, I guess we can
> live with it...

Ok, I'll go and merge the v2 patch then. Siarhei, I'm working on a
linux-sunxi-3.4 fix for the sun5i PLL6 issue, please do not add a
patch to the linux-sunxi-3.4 kernels claiming "COMPAT_REV 1" compatibility
until that patch is done and merged.

Thanks & Regards,

Hans