Re: [PATCH 00/26] clk: amlogic: clock controllers clean-up and factorisation

[Jerome Brunet <jbrunet@baylibre.com>]

On Sun 21 Sep 2025 at 14:21, Martin Blumenstingl <martin.blumenstingl@googlemail.com> wrote:

> Hi Mark,
>
> On Sun, Sep 21, 2025 at 1:25 PM Mark Brown <broonie@kernel.org> wrote:
>>
>> On Wed, Jul 02, 2025 at 05:25:58PM +0200, Jerome Brunet wrote:
>> > The work on this patchset started with the submission of the Amlogic t7
>> > peripheral clock controller [1]. This controller is fairly similar to
>> > existing controllers. Once again, it redefines the peripheral clock macro,
>> > the probe function and composite clock helpers, even if it is almost the
>> > same as in other controllers. This code duplication trend has been going on
>> > for too long and now is the time to properly address the problem.
>>
>> Friday's -next breaks the boot for me on the Libretch Alta.  The system
>> doesn't show any output on the console unfortunately:
>>
>>    https://lava.sirena.org.uk/scheduler/job/1853204#L1606
>>
>> Unfortunately earlycon isn't turned on, I'll try to do so at tomorrow or
>> Tuesday.
>>
>> A bisect points to the patch 4c4e17f27013 ("clk: amlogic: naming
>> consistency alignment") which looks like it's a squash of several of the
>> patches here.  Due to issues with the bootloader it could potentially be
>> an adjacent commit but this looks suspiciously relevant.  Bisect log, my
>> infrastructure does retry the boot a lot on these boards so even if the
>> final error is a bootloader issue there will likely have been boots that
>> start the kernel:
> A few days ago Marek reported problems on an Odroid-N2 and provided a fix: [0]
> It's not queued up for -next yet.
>
> Please give Marek's patch a try
>
>
> [0]
> https://lore.kernel.org/linux-amlogic/175834437006.4354.11933545174004616373@lazor/T/#m29128b0acec2db62d45971d94a1c38fedc8493a0

Hey Mark,

Being on the A311D as well, it is likely to be the same issue.
I suspect you would not see the crash it without earlycon indeed.

The fix is in clk-next already so linux-next should get it soon.
Sorry for the inconvenience and thanks for the report.

-- 
Jerome