Il 8 dicembre 2022 14:21:31 CET, Marek Vasut <marex@xxxxxxx> ha scritto:
>On 12/8/22 11:51, Miquel Raynal wrote:
>> Hi Shawn,
>
>Hi,
>
>> + Thorsten
>>
>> marex@xxxxxxx wrote on Mon, 5 Dec 2022 17:26:53 +0100:
>>
>>> On 12/5/22 16:23, Francesco Dolcini wrote:
>>>> From: Francesco Dolcini <francesco.dolcini@xxxxxxxxxxx>
>>>>
>>>> This reverts commit 753395ea1e45c724150070b5785900b6a44bd5fb.
>>>>
>>>> It introduced a boot regression on colibri-imx7, and potentially any
>>>> other i.MX7 boards with MTD partition list generated into the fdt by
>>>> U-Boot.
>>>>
>>>> While the commit we are reverting here is not obviously wrong, it fixes
>>>> only a dt binding checker warning that is non-functional, while it
>>>> introduces a boot regression and there is no obvious fix ready.
>>>>
>>>> Cc: stable@xxxxxxxxxxxxxxx
>>>> Fixes: 753395ea1e45 ("ARM: dts: imx7: Fix NAND controller size-cells")
>>>> Link: https://lore.kernel.org/all/Y4dgBTGNWpM6SQXI@xxxxxxxxxxxxxxxxxxxxxxxxxxxx/
>>>> Link: https://lore.kernel.org/all/20221205144917.6514168a@xps-13/
>>>> Signed-off-by: Francesco Dolcini <francesco.dolcini@xxxxxxxxxxx>
>> [...]
>>> Reviewed-by: Miquel Raynal <miquel.raynal@xxxxxxxxxxx>
>> [...]
>>> Acked-by: Marek Vasut <marex@xxxxxxx>
>> [...]
>>
>> As discussed in the above links, boot is broken on imx7 Colibri boards,
>> this revert was the most quick and straightforward fix we agreed upon
>> with the hope (~ duty?) it would make it in v6.1. Any chance you could
>> pick this up rapidly and forward it to Linus? Or should we involve
>> him directly (Thorsten?).
>
>It seems neither Francesco nor me agree that this is the right approach and rather the fix should be the two-liner change to the OF partition parser, so maybe this should not be picked ?
I think that the 2 lines change might not be good enough to properly handle the U-Boot generated OF partitions in the general case, even if it fixes my specific issue.
Given that I would do the revert as an immediate first step.
Francesco
