Re: [PATCH v3 0/3] arm64: meson-sm1: add support for the SM1 based VIM3L

From: Kevin Hilman
Date: Thu Aug 29 2019 - 13:09:50 EST


Neil Armstrong <narmstrong@xxxxxxxxxxxx> writes:

> On 28/08/2019 19:55, Kevin Hilman wrote:
>> Neil Armstrong <narmstrong@xxxxxxxxxxxx> writes:
>>
>>> This patchset adds support for the Amlogic SM1 based Khadas VIM3L variant.
>>>
>>> The S903D3 package variant of SM1 is pin-to-pin compatible with the
>>> S922X and A311d, so only internal DT changes are needed :
>>> - DVFS support is different
>>> - Audio support not yet available for SM1
>>>
>>> This patchset moved all the non-g12b nodes to meson-khadas-vim3.dtsi
>>> and add the sm1 specific nodes into meson-sm1-khadas-vim3l.dts.
>>
>> Reviewed-by: Kevin Hilman <khilman@xxxxxxxxxxxx>
>> Tested-by: Kevin Hilman <khilman@xxxxxxxxxxxx>
>>
>> Basic boot test + suspend/resume test OK on my vim3L (thanks to Khadas
>> for the board!)
>>
>>> Display has a color conversion bug on SM1 by using a more recent vendor
>>> bootloader on the SM1 based VIM3, this is out of scope of this patchset
>>> and will be fixed in the drm/meson driver.
>>>
>>> Dependencies:
>>> - patch 1,2: None
>>> - patch 3: Depends on the "arm64: meson-sm1: add support for DVFS" patchset at [1]
>>
>> I tested in my integ branch where this series is applied, but I'm not
>> seeing any OPPs created (/sys/devices/system/cpu/cpufreq/)
>
> These patches were sent from your integ branch, on top of :
> commit 395df5af4c782ad19fb34b9a2009ca240eeb9749 (khilman-amlogic/v5.4/integ)
> Merge: 2fcc5666dd45 9557737987bb
> Author: Kevin Hilman <khilman@xxxxxxxxxxxx>
> Date: Tue Aug 27 15:39:46 2019 -0700
>
> Merge branch 'v5.4/testing' into tmp/aml-rebuild
>
> Rebuilt and retested, and I get the OPPs just fine :
> # cat /sys/bus/cpu/devices/cpu0/cpufreq/scaling_available_frequencies
> 100000 250000 500000 666666 1000000 1200000 1404000 1500000 1608000 1704000 1800000 1908000

Thanks for confirming.

Indeed, there was an issue with my most recent `integ` branch (it was
missing the driver side of some SM1 clocks.) Fixing that issue, and
retesting this series it all works well.

Queuing for v5.4,

Thanks,

Kevin