Re: [PATCH v2] arm64: dts: allwinner: a64: Enable eMMC on A64-OLinuXino

[Sunil Mohan Adapa <sunil@medhas.org>]

[-- Attachment #1.1.1: Type: text/plain, Size: 4460 bytes --]



On 07/08/19 10:12 pm, Chen-Yu Tsai wrote:
> On Wed, Aug 7, 2019 at 10:45 PM Maxime Ripard <maxime.ripard@bootlin.com> wrote:
>>
>> Hi,
>>
>> On Wed, Aug 07, 2019 at 08:09:19PM +0800, Chen-Yu Tsai wrote:
>>> On Wed, Aug 7, 2019 at 8:01 PM Maxime Ripard <maxime.ripard@bootlin.com> wrote:
>>>>
>>>> On Tue, Aug 06, 2019 at 02:25:17PM +0800, Chen-Yu Tsai wrote:
>>>>> On Mon, Aug 5, 2019 at 8:58 PM Martin Ayotte <martinayotte@gmail.com> wrote:
>>>>>>
>>>>>> Fine for me too.
>>>>>>
>>>>>> Thanks .
>>>>>>
>>>>>> -----Message d'origine-----
>>>>>> De : Sunil Mohan Adapa [mailto:sunil@medhas.org]
>>>>>> Envoyé : Monday, August 05, 2019 1:25 AM
>>>>>> Ą : Chen-Yu Tsai
>>>>>> Cc : Maxime Ripard; Martin Ayotte; linux-arm-kernel
>>>>>> Objet : Re: [PATCH v2] arm64: dts: allwinner: a64: Enable eMMC on
>>>>>> A64-OLinuXino
>>>>>>
>>>>>> On 04/08/19 8:33 pm, Chen-Yu Tsai wrote:
>>>>>>> On Fri, Aug 2, 2019 at 2:47 AM Sunil Mohan Adapa <sunil@medhas.org> wrote:
>>>>>>>>
>>>>>>>> On 01/08/19 6:49 am, Martin Ayotte wrote:
>>>>>>>>> If my SOB could help here, I don't mind since I've done the commit
>>>>>>>>> more than a year ago for Armbian ...
>>>>>>>>>
>>>>>>>>> Signed-off-by: Martin Ayotte <martinayotte@gmai.com>
>>>>>>>>> Tested-by: Martin Ayotte <martinayotte@gmai.com>
>>>>>>>> gmai.com is likely a typo.
>>>>>>>>
>>>>>>>>> On Wed, Jul 31, 2019 at 10:42 PM Chen-Yu Tsai <wens@csie.org
>>>>>>>>>
>>>>>>>>>> Thanks. The patch looks good overall. The authorship is a little
>>>>>>>>>> confusing though. If it was initially done by Martin (CC-ed), then
>>>>>>>>>> he should be the author, and we should get his Signed-off-by if
>>>>>>>>>> possible.
>>>>>>>>
>>>>>>>> Martin is indeed the original author of the patch. Thank you for
>>>>>> reviewing.
>>>>>>>
>>>>>>> I'd like to apply this patch with Martin as the author, if that's OK with
>>>>>> you
>>>>>>> both?
>>>>>>
>>>>>> That is completely okay with me.
>>>>>
>>>>> Applied for 5.4.
>>>>>
>>>>> I reordered the tags so they make more sense:
>>>>>
>>>>> https://git.kernel.org/pub/scm/linux/kernel/git/sunxi/linux.git/commit/?h=sunxi/dt-for-5.4&id=0834887732df5af41b59b2e4d530fc1f5478965f
>>>>
>>>> Sorry for being late on this, but it looks like the eMMC, NAND and SPI
>>>> pins are conflicting on the A64-Olinuxino design.
>>>>
>>>> There's no configuration with a NAND, so we don't really need to worry
>>>> about that, but if we merge this in the main DT, we'll prevent anyone
>>>> from using that DT on an olinuxino with a SPI flash.
>>>>
>>>> I think we should just create emmc and SPI-flash variants of that DT.
>>>
>>> Actually they aren't. Olimex specifically uses eMMC modules that don't
>>> use the data strobe line, so SPI can be used together.
>>
>> Ah, right.
>>
>> Still, this creates a precedent that I'm not really comfortable
>> with. Three actually.
>>
>> Merging this in the main DT means three things:
>>   - We're not consistent anymore, including within the olinuxino
>>     boards only. A20 Olinuxino is pretty much in the same situation,
>>     yet we dealt with it differently.
>>   - This means that this will create a spurious device and report
>>     errors in the kernel message and whenever someone will try to
>>     access the device on boards that don't have it wired. This
>>     shouldn't happen and we really shouldn't expose devices that just
>>     aren't there, just like you don't have all the devices that are
>>     not connected on your USB connector.
>>   - Finally, this means that in order to keep it somewhat consistent,
>>     we would have to merge the SPI flash in the main DT too. This will
>>     prevent people without a SPI flash to use the SPI signals on the
>>     UEXT connector for something else, which again goes against the
>>     policy we've had for basically any other board.
> 
> OK. Shall we back it out and figure out something else?

I can try to help with alternate implementation with some guidance. I
can also test any patches on an A64-OLinuXino model with eMMC. So, do we
create a new -emmc.dts like in case of A20 OLinuXino?

BTW, a basic question: how does u-boot know which variant of dtb (with
-emmc or without) it has load before booting Linux? Does this need to be
hardcoded now into the boot script? Currently, in Debian the DTB name is
chosen by u-boot and not present in the boot script.

Thanks,

-- 
Sunil


[-- Attachment #1.2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

[-- Attachment #2: Type: text/plain, Size: 176 bytes --]

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel