[PATCH] arm64: defconfig: Enable missing configs for mt8192-asurada

[AngeloGioacchino Del Regno]

Il 18/01/23 19:15, Bjorn Andersson ha scritto:
> On Tue, Jan 17, 2023 at 03:11:59PM -0500, Nícolas F. R. A. Prado wrote:
>> On Tue, Jan 17, 2023 at 07:58:54AM -0800, Bjorn Andersson wrote:
>>> On Thu, Jan 12, 2023 at 10:12:38AM -0500, Nícolas F. R. A. Prado wrote:
>>>> Enable missing configs in the arm64 defconfig to get all devices probing
>>>> on mt8192-asurada based machines.
>>>>
>>>> The devices enabled are: MediaTek Bluetooth USB controller, MediaTek
>>>> PCIe Gen3 MAC controller, MT7921E wireless adapter, Elan I2C Trackpad,
>>>> MediaTek SPI NOR flash controller, Mediatek SPMI Controller, ChromeOS EC
>>>> regulators, MT6315 PMIC, MediaTek Video Codec, MT8192 sound cards,
>>>> ChromeOS EC rpmsg communication, all MT8192 clocks.
>>>>
>>>> Support for DMA Restricted Pool is also enabled since it is used by the
>>>> WiFi card on this platform.
>>>>
>>>> REGULATOR_CROS_EC is enabled as builtin since it powers the MMC
>>>> controller for the SD card, making it required for booting on some
>>>> setups.
>>>
>>> I presume this implies that there's no ramdisk to carry these additional
>>> modules?
>>
>> There may or may not be, at the discretion of the user, so enable by default.
>>
> 
> As long as the dependency chain is reasonably short this is okay with me.
> 
>>>
>>>>
>>>> By enabling the support for all of this platform's devices on the
>>>> defconfig we make it effortless to test the relevant hardware both by
>>>> developers as well as CI systems like KernelCI.
>>>>
>>>> Signed-off-by: Nícolas F. R. A. Prado <nfraprado at collabora.com>
>>>>
>>>> ---
>>>>
>>>>   arch/arm64/configs/defconfig | 25 +++++++++++++++++++++++++
>>>>   1 file changed, 25 insertions(+)
>>>>
>>>> diff --git a/arch/arm64/configs/defconfig b/arch/arm64/configs/defconfig
>>>> index f3053e7018fe..4e806d8068f6 100644
>>>> --- a/arch/arm64/configs/defconfig
>>>> +++ b/arch/arm64/configs/defconfig
>> [..]
>>>> @@ -1092,6 +1104,18 @@ CONFIG_CLK_IMX8QXP=y
>>>>   CONFIG_CLK_IMX8ULP=y
>>>>   CONFIG_CLK_IMX93=y
>>>>   CONFIG_TI_SCI_CLK=y
>>>> +CONFIG_COMMON_CLK_MT8192_AUDSYS=y
>>>> +CONFIG_COMMON_CLK_MT8192_CAMSYS=y
>>>> +CONFIG_COMMON_CLK_MT8192_IMGSYS=y
>>>> +CONFIG_COMMON_CLK_MT8192_IMP_IIC_WRAP=y
>>>> +CONFIG_COMMON_CLK_MT8192_IPESYS=y
>>>> +CONFIG_COMMON_CLK_MT8192_MDPSYS=y
>>>> +CONFIG_COMMON_CLK_MT8192_MFGCFG=y
>>>> +CONFIG_COMMON_CLK_MT8192_MMSYS=y
>>>> +CONFIG_COMMON_CLK_MT8192_MSDC=y
>>>> +CONFIG_COMMON_CLK_MT8192_SCP_ADSP=y
>>>> +CONFIG_COMMON_CLK_MT8192_VDECSYS=y
>>>> +CONFIG_COMMON_CLK_MT8192_VENCSYS=y
>>>
>>> Are all these clock drivers needed in order to reach a ramdisk, with
>>> working uart, where further kernel modules could be loaded?
>>
>> No, but currently these configs can't be enabled as modules. So my intention is
>> to enable all of them as builtin for now so we get the functionality there, and
>> after the drivers and configs are reworked to allow building as modules (which
>> should happen shortly after [1] is merged), I'll update the defconfig so that
>> the non-essential ones for boot are set to =m.
>>
> 
> Sounds good to me.

Adding context... on MediaTek clock drivers, I'm doing the rework [1].

There's a lot to cleanup and one first rather big series is already there,
well-reviewed and ready to be picked, but that does not change the MT8192,
nor others, non-critical clock drivers to tristate.

I'll do that for most MTK clock drivers in a second (or third) cleanup series,
which is already scheduled on my side, but that will need some time to happen
since, as I said, there's still a bit to cleanup before that and I will make
sure that nobody will forget to send a followup commit that will change the
defconfig to build the non-boot-critical clock drivers as modules as soon as
possible.

Cheers,
Angelo