Re: [OE-core] [PATCH 0/4] kernel-yocto: -stable and new reference kernel

["Mathieu Dubois-Briand" <mathieu.dubois-briand@bootlin.com>]

On Tue Aug 26, 2025 at 4:03 AM CEST, Bruce Ashfield wrote:
> On Fri, Aug 22, 2025 at 12:52 PM Mathieu Dubois-Briand <
> mathieu.dubois-briand@bootlin.com> wrote:
>
>> On Thu Aug 21, 2025 at 4:03 PM CEST, Bruce Ashfield via
>> lists.openembedded.org wrote:
>> > From: Bruce Ashfield <bruce.ashfield@gmail.com>
>> >
>> > Richard,
>> >
>> > Here's the next set of updates to the linux-yocto kernel.
>> >
>> > The plan is to introduce 6.16 as the latest reference kernel, and keep
>> > 6.12 around (since it is a LTS).
>> >
>> > This isn't ready to be the default for builds, but it has passed my
>> > core-iamge-kernel-dev testing, so I wanted to make the recipes available
>> > as visibility to where we are headed for the release (plus, I'm out of
>> > the office on holidays until the first week of September, so making it
>> > the default would be a bad idea from me being able to support issues).
>> >
>>
>> Hi Bruce,
>>
>> Thanks for the series.
>>
>> I understand this isn't fully ready, but I did launch a full build
>> anyway: most builders succeed, we just have one warning on config
>> fragments:
>>
>> WARNING: linux-yocto-6.16.2+git-r0 do_kernel_configcheck: [kernel config]:
>> specified values did not make it into the kernel's final configuration:
>> ...
>> [INFO]: config 'CONFIG_CRYPTO_LIB_CHACHA' was set, but it wasn't
>> assignable, check (parent) dependencies
>> ...
>> [INFO]: config 'CONFIG_CRYPTO_CHACHA20_NEON' was set, but it wasn't
>> assignable, check (parent) dependencies
>> ...
>> [INFO]: the following symbols were not found in the active configuration:
>>      - CONFIG_CRYPTO_SHA2_ARM_CE
>>      - CONFIG_CRYPTO_CRC32_ARM_CE
>>
>> https://autobuilder.yoctoproject.org/valkyrie/#/builders/40/builds/2248
>
>
> I always find it odd that I can't click on any of the AB tabs (in the link
> you sent) and see what the MACHINE value is for that build.
>

Hum, that's raising an interesting point, I had the same issue at the
very beginning. I'm now familiar with each steps and know where I have
to look at, but maybe we can do something better about it.

I will add this idea somewhere on my TODO list.

> I looked at the writeconfig step of the build and it looks to be qemuarm.
> Which is odd, since I'm not seeing the same warning for that here.
>
Yes, I confirm, we have MACHINE = "qemuarm".

Best regards,
Mathieu

-- 
Mathieu Dubois-Briand, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com