On 6.03.2023 15:52, Guilherme G. Piccoli wrote:
> On 06/03/2023 07:14, Konrad Dybcio wrote:
>> [...]
>>> Curiosity questions, since I'm far from expert in the device-tree world:
>>> what happens in case the device-tree doesn't export/contain this
>>> property (exactly the case we have right now)? Does the device work
>>> fine? Also, having it "wrong" (based on the other OnePlus 3) is worse
>>> than not having it? In other words, what's the default value picked if
>>> none is provided in the DT?
>> Basically:
>>
>> if (msm_id) {
>> if is_close_enough(msm_id, internal_msm_id)
>> boot()
>> else
>> die()
>> } else
>> die()
>>
>> Konrad
>>>
>
> Thanks, interesting...
>
> So, if I understand correctly, currently we end-up in the 2nd else
> block, and we straight die(), right?
>
> With this patch, we have some chance to boot, by falling in the if()
> block or...if we're not lucky, we also die() heh
>
> With that said, any con in accepting this as-is if we don't find anybody
> to test? Notice it at least has the advantage of fixing the device-tree
> blob creation...
You're using some magic tools that shouldn't be necessary.. Some device
owner should try booting with simply appending the device tree to the
compressed kernel image.
Konrad
>
> Ah, of course by "as-is" I mean I'd send a V2 fixing what you mentioned
> before ("...use the preprocessor constant...etc").
> Thanks,
>
>
> Guilherme
