Hi Olof,
On mar., déc. 22 2015, Olof Johansson <olof@xxxxxxxxx> wrote:
> Hi,
>
> On Mon, Dec 21, 2015 at 11:33:44AM +0100, Geert Uytterhoeven wrote:
>> Hi,
>>
>> As of commit e488ca9f8d4f62c2 ("doc: dt: mtd: partitions: add compatible
>> property to "partitions" node"), which is in v4.4-rc6, the "partitions"
>> subnode of an SPI FLASH device node must have a compatible property. The
>> partitions are no longer detected if it is not present.
>>
>> However, several DTSes in -next have already been converted to the
>> "partitions" subnode without "compatible" property, introduced by
>> commits 5cfdedb7b9a0fe38 ("mtd: ofpart: move ofpart partitions to a
>> dedicated dt node") and fe2585e9c29a650a ("doc: dt: mtd: support
>> partitions in a special 'partitions' subnode"). Hence all of these are
>> now broken in -next, and will be broken in upstream during the merge
>> window.
>
> So, if I understand this correctly, the partitions format was added for v4.4,
> then this non-backwards compatible change was added in -rc6. But, there were
> also DT files that had the new-for-v4.4 partitions nodes in them that then
> stopped working in -rc6?
At least for the mvebu dts, the change was added for v4.5 not
v4.4. Currently it is only in the next branches (and also in your
arm-soc tree in next/dt).
>
> That sounds like a regression, so this should be picked up as fixes
> for v4.4.
It won't be a regression but a breakage of the bissectability between
the patch adding the partition subnode and the one using the compatible
"fixed-partitions".
If you really want avoiding this you could squash the patches, however
from my point of view it's not worth doing it.
>
> Please confirm that I've understood the setup correctly, and I'll apply the
> series directly to fixes.
I think it won't apply on v4.4 because the partition change was not made
at this moment.
Gregory
--
Gregory Clement, Free Electrons
Kernel, drivers, real-time and embedded Linux
development, consulting, training and support.
http://free-electrons.com
