2022-11-08 23:25:46

by Nícolas F. R. A. Prado

[permalink] [raw]
Subject: [PATCH 00/13] arm64: defconfig: Enable missing kconfigs for mt8183-kukui-jacuzzi-juniper


mt8183-kukui-jacuzzi-juniper is one of the devices set up to run tests
on KernelCI, but several of its drivers are currently disabled in the
defconfig. This series enables all the missing kconfigs on the defconfig
to get everything probing on that machine so that it can be fully tested
by KernelCI.

Given that all kconfigs added in the series are to enable support for a
MediaTek platform, it seems reasonable for it to be applied through the
MediaTek tree, but the commits themselves are independent (apart from
MTK_CMDQ and MTK_SVS) and could be applied separately.

[1] https://github.com/kernelci/kernelci-core/pull/1352
[2] https://github.com/kernelci/bootrr/pull/22


Nícolas F. R. A. Prado (13):
arm64: defconfig: Enable SPI_MT65XX
arm64: defconfig: Enable ATH10K_SDIO
arm64: defconfig: Enable TOUCHSCREEN_ELAN
arm64: defconfig: Enable sound support for MT8183 based machines
arm64: defconfig: Enable DRM_ANALOGIX_ANX7625
arm64: defconfig: Enable MTK_CMDQ
arm64: defconfig: Enable MTK_SVS
arm64: defconfig: Enable CR50 TPMs
arm64: defconfig: Enable MTK_SCP
arm64: defconfig: Enable VIDEO_MEDIATEK_JPEG
arm64: defconfig: Enable ARM_MEDIATEK_CCI_DEVFREQ
arm64: defconfig: Enable GENERIC_ADC_THERMAL
arm64: defconfig: Enable MEDIATEK_MT6577_AUXADC

arch/arm64/configs/defconfig | 18 ++++++++++++++++++
1 file changed, 18 insertions(+)

--
2.38.1



2022-11-08 23:26:10

by Nícolas F. R. A. Prado

[permalink] [raw]
Subject: [PATCH 04/13] arm64: defconfig: Enable sound support for MT8183 based machines

Enable sound support for the MT8183 platform and for its two sound card
variants.

By enabling this on the defconfig we make it effortless to test the
relevant hardware on CI systems like KernelCI.

Signed-off-by: Nícolas F. R. A. Prado <[email protected]>
---

arch/arm64/configs/defconfig | 3 +++
1 file changed, 3 insertions(+)

diff --git a/arch/arm64/configs/defconfig b/arch/arm64/configs/defconfig
index c6bd2f7cd1b8..3c1351609134 100644
--- a/arch/arm64/configs/defconfig
+++ b/arch/arm64/configs/defconfig
@@ -808,6 +808,9 @@ CONFIG_SND_SOC_IMX_SGTL5000=m
CONFIG_SND_SOC_IMX_SPDIF=m
CONFIG_SND_SOC_FSL_ASOC_CARD=m
CONFIG_SND_SOC_IMX_AUDMIX=m
+CONFIG_SND_SOC_MT8183=m
+CONFIG_SND_SOC_MT8183_MT6358_TS3A227E_MAX98357A=m
+CONFIG_SND_SOC_MT8183_DA7219_MAX98357A=m
CONFIG_SND_MESON_AXG_SOUND_CARD=m
CONFIG_SND_MESON_GX_SOUND_CARD=m
CONFIG_SND_SOC_QCOM=m
--
2.38.1


2022-11-08 23:26:15

by Nícolas F. R. A. Prado

[permalink] [raw]
Subject: [PATCH 05/13] arm64: defconfig: Enable DRM_ANALOGIX_ANX7625

Enable support for the ANX7625 MIPI/DPI to DP bridge. It is used to
drive the display on mt8183-kukui-jacuzzi.

By enabling this on the defconfig we make it effortless to test the
relevant hardware on CI systems like KernelCI.

Signed-off-by: Nícolas F. R. A. Prado <[email protected]>
---

arch/arm64/configs/defconfig | 1 +
1 file changed, 1 insertion(+)

diff --git a/arch/arm64/configs/defconfig b/arch/arm64/configs/defconfig
index 3c1351609134..a0cb700bf0af 100644
--- a/arch/arm64/configs/defconfig
+++ b/arch/arm64/configs/defconfig
@@ -766,6 +766,7 @@ CONFIG_DRM_SII902X=m
CONFIG_DRM_SIMPLE_BRIDGE=m
CONFIG_DRM_THINE_THC63LVD1024=m
CONFIG_DRM_TI_SN65DSI86=m
+CONFIG_DRM_ANALOGIX_ANX7625=m
CONFIG_DRM_I2C_ADV7511=m
CONFIG_DRM_I2C_ADV7511_AUDIO=y
CONFIG_DRM_CDNS_MHDP8546=m
--
2.38.1


2022-11-08 23:26:23

by Nícolas F. R. A. Prado

[permalink] [raw]
Subject: [PATCH 06/13] arm64: defconfig: Enable MTK_CMDQ

Enable support for the MediaTek Command Queue (CMDQ). This implicitly
enables MTK_CMDQ_MBOX, the Command Queue Mailbox, which is used as part
of the display pipeline.

By enabling this on the defconfig we make it effortless to test the
relevant hardware on CI systems like KernelCI.

Signed-off-by: Nícolas F. R. A. Prado <[email protected]>
---

arch/arm64/configs/defconfig | 1 +
1 file changed, 1 insertion(+)

diff --git a/arch/arm64/configs/defconfig b/arch/arm64/configs/defconfig
index a0cb700bf0af..5e752a9afed7 100644
--- a/arch/arm64/configs/defconfig
+++ b/arch/arm64/configs/defconfig
@@ -1137,6 +1137,7 @@ CONFIG_RASPBERRYPI_POWER=y
CONFIG_FSL_DPAA=y
CONFIG_FSL_MC_DPIO=y
CONFIG_FSL_RCPM=y
+CONFIG_MTK_CMDQ=m
CONFIG_MTK_DEVAPC=m
CONFIG_MTK_PMIC_WRAP=y
CONFIG_QCOM_AOSS_QMP=y
--
2.38.1


2022-11-08 23:48:02

by Nícolas F. R. A. Prado

[permalink] [raw]
Subject: [PATCH 10/13] arm64: defconfig: Enable VIDEO_MEDIATEK_JPEG

Enable support for the JPEG hardware codec present on some MediaTek
SoCs.

By enabling this on the defconfig we make it effortless to test the
relevant hardware on CI systems like KernelCI.

Signed-off-by: Nícolas F. R. A. Prado <[email protected]>
---

arch/arm64/configs/defconfig | 1 +
1 file changed, 1 insertion(+)

diff --git a/arch/arm64/configs/defconfig b/arch/arm64/configs/defconfig
index 1070ecc68b08..c0f2068c69b5 100644
--- a/arch/arm64/configs/defconfig
+++ b/arch/arm64/configs/defconfig
@@ -708,6 +708,7 @@ CONFIG_USB_VIDEO_CLASS=m
CONFIG_V4L_PLATFORM_DRIVERS=y
CONFIG_SDR_PLATFORM_DRIVERS=y
CONFIG_V4L_MEM2MEM_DRIVERS=y
+CONFIG_VIDEO_MEDIATEK_JPEG=m
CONFIG_VIDEO_QCOM_CAMSS=m
CONFIG_VIDEO_QCOM_VENUS=m
CONFIG_VIDEO_RCAR_ISP=m
--
2.38.1


2022-11-08 23:49:37

by Nícolas F. R. A. Prado

[permalink] [raw]
Subject: [PATCH 07/13] arm64: defconfig: Enable MTK_SVS

Enable support for the Smart Voltage Scaling (SVS) engine present on
some MediaTek SoCs like mt8183.

By enabling this on the defconfig we make it effortless to test the
relevant hardware on CI systems like KernelCI.

Signed-off-by: Nícolas F. R. A. Prado <[email protected]>
---

arch/arm64/configs/defconfig | 1 +
1 file changed, 1 insertion(+)

diff --git a/arch/arm64/configs/defconfig b/arch/arm64/configs/defconfig
index 5e752a9afed7..9fe43387d83f 100644
--- a/arch/arm64/configs/defconfig
+++ b/arch/arm64/configs/defconfig
@@ -1140,6 +1140,7 @@ CONFIG_FSL_RCPM=y
CONFIG_MTK_CMDQ=m
CONFIG_MTK_DEVAPC=m
CONFIG_MTK_PMIC_WRAP=y
+CONFIG_MTK_SVS=m
CONFIG_QCOM_AOSS_QMP=y
CONFIG_QCOM_COMMAND_DB=y
CONFIG_QCOM_CPR=y
--
2.38.1


2022-11-08 23:51:15

by Nícolas F. R. A. Prado

[permalink] [raw]
Subject: [PATCH 09/13] arm64: defconfig: Enable MTK_SCP

Enable support for the MediaTek System Companion Processor (SCP). It is
a required component in order to have working hardware video decoding on
several MediaTek SoCs.

By enabling this on the defconfig we make it effortless to test the
relevant hardware on CI systems like KernelCI.

Signed-off-by: Nícolas F. R. A. Prado <[email protected]>
---

arch/arm64/configs/defconfig | 1 +
1 file changed, 1 insertion(+)

diff --git a/arch/arm64/configs/defconfig b/arch/arm64/configs/defconfig
index c54a85679d40..1070ecc68b08 100644
--- a/arch/arm64/configs/defconfig
+++ b/arch/arm64/configs/defconfig
@@ -1124,6 +1124,7 @@ CONFIG_ARM_SMMU_V3=y
CONFIG_MTK_IOMMU=y
CONFIG_QCOM_IOMMU=y
CONFIG_REMOTEPROC=y
+CONFIG_MTK_SCP=m
CONFIG_QCOM_Q6V5_ADSP=m
CONFIG_QCOM_Q6V5_MSS=m
CONFIG_QCOM_Q6V5_PAS=m
--
2.38.1


2022-11-08 23:51:19

by Nícolas F. R. A. Prado

[permalink] [raw]
Subject: [PATCH 11/13] arm64: defconfig: Enable ARM_MEDIATEK_CCI_DEVFREQ

Enable support for frequency and voltage scaling of the MediaTek Cache
Coherent Interconnect (CCI), which is present on some MediaTek SoCs like
mt8183 and mt8186.

By enabling this on the defconfig we make it effortless to test the
relevant hardware on CI systems like KernelCI.

Signed-off-by: Nícolas F. R. A. Prado <[email protected]>
---

arch/arm64/configs/defconfig | 1 +
1 file changed, 1 insertion(+)

diff --git a/arch/arm64/configs/defconfig b/arch/arm64/configs/defconfig
index c0f2068c69b5..0a2b76f7a997 100644
--- a/arch/arm64/configs/defconfig
+++ b/arch/arm64/configs/defconfig
@@ -1197,6 +1197,7 @@ CONFIG_ARCH_TEGRA_234_SOC=y
CONFIG_TI_SCI_PM_DOMAINS=y
CONFIG_ARM_IMX_BUS_DEVFREQ=m
CONFIG_ARM_IMX8M_DDRC_DEVFREQ=m
+CONFIG_ARM_MEDIATEK_CCI_DEVFREQ=m
CONFIG_EXTCON_PTN5150=m
CONFIG_EXTCON_USB_GPIO=y
CONFIG_EXTCON_USBC_CROS_EC=y
--
2.38.1


2022-11-08 23:52:15

by Nícolas F. R. A. Prado

[permalink] [raw]
Subject: [PATCH 12/13] arm64: defconfig: Enable GENERIC_ADC_THERMAL

Enable support for generic ADC based thermal sensors. It is used by some
imx6dl based boards as well as mt8183-kukui.

By enabling this on the defconfig we make it effortless to test the
relevant hardware on CI systems like KernelCI.

Signed-off-by: Nícolas F. R. A. Prado <[email protected]>
---

arch/arm64/configs/defconfig | 1 +
1 file changed, 1 insertion(+)

diff --git a/arch/arm64/configs/defconfig b/arch/arm64/configs/defconfig
index 0a2b76f7a997..caa95c908430 100644
--- a/arch/arm64/configs/defconfig
+++ b/arch/arm64/configs/defconfig
@@ -624,6 +624,7 @@ CONFIG_BRCMSTB_THERMAL=m
CONFIG_EXYNOS_THERMAL=y
CONFIG_TEGRA_SOCTHERM=m
CONFIG_TEGRA_BPMP_THERMAL=m
+CONFIG_GENERIC_ADC_THERMAL=m
CONFIG_QCOM_TSENS=y
CONFIG_QCOM_SPMI_ADC_TM5=m
CONFIG_QCOM_SPMI_TEMP_ALARM=m
--
2.38.1


2022-11-09 00:10:06

by Nícolas F. R. A. Prado

[permalink] [raw]
Subject: [PATCH 01/13] arm64: defconfig: Enable SPI_MT65XX

Enable support for the SPI bus on pretty much every MediaTek SoC. Only
MT7621 is handled by a separate driver, all the others use this one.

It is enabled as builtin since on some platforms like mt8195-cherry, the
ChromeOS Embedded Controller is connected through SPI and it is
responsible for the regulators powering the MMC controller used for the
SD card, and thus SPI support is essential for booting.

By enabling this on the defconfig we make it effortless to test the
relevant hardware on CI systems like KernelCI.

Signed-off-by: Nícolas F. R. A. Prado <[email protected]>

---

arch/arm64/configs/defconfig | 1 +
1 file changed, 1 insertion(+)

diff --git a/arch/arm64/configs/defconfig b/arch/arm64/configs/defconfig
index 226bb30ea7e2..171fbecd7372 100644
--- a/arch/arm64/configs/defconfig
+++ b/arch/arm64/configs/defconfig
@@ -505,6 +505,7 @@ CONFIG_SPI_IMX=m
CONFIG_SPI_FSL_DSPI=y
CONFIG_SPI_MESON_SPICC=m
CONFIG_SPI_MESON_SPIFC=m
+CONFIG_SPI_MT65XX=y
CONFIG_SPI_ORION=y
CONFIG_SPI_PL022=y
CONFIG_SPI_ROCKCHIP=y
--
2.38.1


2022-11-09 00:23:30

by Nícolas F. R. A. Prado

[permalink] [raw]
Subject: [PATCH 08/13] arm64: defconfig: Enable CR50 TPMs

Enable support for the CR50 TPMs present on Chromebooks. They can be
connected through SPI or I2C, and there's a different driver for each
case, so enable both.

By enabling this on the defconfig we make it effortless to test the
relevant hardware on CI systems like KernelCI.

Signed-off-by: Nícolas F. R. A. Prado <[email protected]>
---

arch/arm64/configs/defconfig | 4 ++++
1 file changed, 4 insertions(+)

diff --git a/arch/arm64/configs/defconfig b/arch/arm64/configs/defconfig
index 9fe43387d83f..c54a85679d40 100644
--- a/arch/arm64/configs/defconfig
+++ b/arch/arm64/configs/defconfig
@@ -464,6 +464,10 @@ CONFIG_IPMI_HANDLER=m
CONFIG_IPMI_DEVICE_INTERFACE=m
CONFIG_IPMI_SI=m
CONFIG_TCG_TPM=y
+CONFIG_TCG_TIS=m
+CONFIG_TCG_TIS_SPI=m
+CONFIG_TCG_TIS_SPI_CR50=y
+CONFIG_TCG_TIS_I2C_CR50=m
CONFIG_TCG_TIS_I2C_INFINEON=y
CONFIG_I2C_CHARDEV=y
CONFIG_I2C_MUX=y
--
2.38.1


2022-11-09 00:24:45

by Nícolas F. R. A. Prado

[permalink] [raw]
Subject: [PATCH 02/13] arm64: defconfig: Enable ATH10K_SDIO

Enable support for the ATH10K wireless adapter when connected through a
SDIO/MMC bus. It is used by mt8183-kukui based devices.

By enabling this on the defconfig we make it effortless to test the
relevant hardware on CI systems like KernelCI.

Signed-off-by: Nícolas F. R. A. Prado <[email protected]>
---

arch/arm64/configs/defconfig | 1 +
1 file changed, 1 insertion(+)

diff --git a/arch/arm64/configs/defconfig b/arch/arm64/configs/defconfig
index 171fbecd7372..af072b47f762 100644
--- a/arch/arm64/configs/defconfig
+++ b/arch/arm64/configs/defconfig
@@ -392,6 +392,7 @@ CONFIG_USB_NET_PLUSB=m
CONFIG_USB_NET_MCS7830=m
CONFIG_ATH10K=m
CONFIG_ATH10K_PCI=m
+CONFIG_ATH10K_SDIO=m
CONFIG_ATH10K_SNOC=m
CONFIG_WCN36XX=m
CONFIG_ATH11K=m
--
2.38.1


2022-11-09 00:27:33

by Nícolas F. R. A. Prado

[permalink] [raw]
Subject: [PATCH 13/13] arm64: defconfig: Enable MEDIATEK_MT6577_AUXADC

Enable support for the Auxiliary Analog/Digital Converter (AUXADC) found
in some MediaTek SoCs. This allows reading temperature measures from the
SoC.

By enabling this on the defconfig we make it effortless to test the
relevant hardware on CI systems like KernelCI.

Signed-off-by: Nícolas F. R. A. Prado <[email protected]>
---

arch/arm64/configs/defconfig | 1 +
1 file changed, 1 insertion(+)

diff --git a/arch/arm64/configs/defconfig b/arch/arm64/configs/defconfig
index caa95c908430..14ce97d78662 100644
--- a/arch/arm64/configs/defconfig
+++ b/arch/arm64/configs/defconfig
@@ -1206,6 +1206,7 @@ CONFIG_RENESAS_RPCIF=m
CONFIG_IIO=y
CONFIG_EXYNOS_ADC=y
CONFIG_MAX9611=m
+CONFIG_MEDIATEK_MT6577_AUXADC=m
CONFIG_QCOM_SPMI_VADC=m
CONFIG_QCOM_SPMI_ADC5=m
CONFIG_ROCKCHIP_SARADC=m
--
2.38.1


2022-11-09 00:27:37

by Nícolas F. R. A. Prado

[permalink] [raw]
Subject: [PATCH 03/13] arm64: defconfig: Enable TOUCHSCREEN_ELAN

Enable support for Elan eKTH I2C touchscreens. They are used by several
of the arm64 devicetrees.

By enabling this on the defconfig we make it effortless to test the
relevant hardware on CI systems like KernelCI.

Signed-off-by: Nícolas F. R. A. Prado <[email protected]>
---

arch/arm64/configs/defconfig | 1 +
1 file changed, 1 insertion(+)

diff --git a/arch/arm64/configs/defconfig b/arch/arm64/configs/defconfig
index af072b47f762..c6bd2f7cd1b8 100644
--- a/arch/arm64/configs/defconfig
+++ b/arch/arm64/configs/defconfig
@@ -413,6 +413,7 @@ CONFIG_KEYBOARD_CROS_EC=y
CONFIG_INPUT_TOUCHSCREEN=y
CONFIG_TOUCHSCREEN_ATMEL_MXT=m
CONFIG_TOUCHSCREEN_GOODIX=m
+CONFIG_TOUCHSCREEN_ELAN=m
CONFIG_TOUCHSCREEN_EDT_FT5X06=m
CONFIG_INPUT_MISC=y
CONFIG_INPUT_PM8941_PWRKEY=y
--
2.38.1


2022-11-09 08:02:49

by Arnd Bergmann

[permalink] [raw]
Subject: Re: [PATCH 00/13] arm64: defconfig: Enable missing kconfigs for mt8183-kukui-jacuzzi-juniper

On Wed, Nov 9, 2022, at 00:22, Nícolas F. R. A. Prado wrote:
> mt8183-kukui-jacuzzi-juniper is one of the devices set up to run tests
> on KernelCI, but several of its drivers are currently disabled in the
> defconfig. This series enables all the missing kconfigs on the defconfig
> to get everything probing on that machine so that it can be fully tested
> by KernelCI.

The changes all look fine, but I would recommend not separating it
out into 13 patches when you are doing just one thing here.

As a general rule, if you keep saying the same things in each
patch description, it is usually an indication that they should
be combined. Similarly, if you find describing unrelated changes
("also, ..."), that would be an indication that patches should
be split up.

> Given that all kconfigs added in the series are to enable support for a
> MediaTek platform, it seems reasonable for it to be applied through the
> MediaTek tree, but the commits themselves are independent (apart from
> MTK_CMDQ and MTK_SVS) and could be applied separately.

Agreed, merging this through the Mediatek tree is the preferred way.

Arnd

2022-11-09 10:05:33

by Krzysztof Kozlowski

[permalink] [raw]
Subject: Re: [PATCH 00/13] arm64: defconfig: Enable missing kconfigs for mt8183-kukui-jacuzzi-juniper

On 09/11/2022 08:28, Arnd Bergmann wrote:
> On Wed, Nov 9, 2022, at 00:22, Nícolas F. R. A. Prado wrote:
>> mt8183-kukui-jacuzzi-juniper is one of the devices set up to run tests
>> on KernelCI, but several of its drivers are currently disabled in the
>> defconfig. This series enables all the missing kconfigs on the defconfig
>> to get everything probing on that machine so that it can be fully tested
>> by KernelCI.
>
> The changes all look fine, but I would recommend not separating it
> out into 13 patches when you are doing just one thing here.
>
> As a general rule, if you keep saying the same things in each
> patch description, it is usually an indication that they should
> be combined. Similarly, if you find describing unrelated changes
> ("also, ..."), that would be an indication that patches should
> be split up.

I agree. Descriptions you wrote are useful - they explain why you are
doing it - but it got all really too detailed, just for defconfigs. One
commit per one symbol is a bit too much...

Best regards,
Krzysztof


2022-11-09 10:08:13

by Krzysztof Kozlowski

[permalink] [raw]
Subject: Re: [PATCH 02/13] arm64: defconfig: Enable ATH10K_SDIO

On 09/11/2022 00:22, Nícolas F. R. A. Prado wrote:
> Enable support for the ATH10K wireless adapter when connected through a
> SDIO/MMC bus. It is used by mt8183-kukui based devices.
>
> By enabling this on the defconfig we make it effortless to test the
> relevant hardware on CI systems like KernelCI.

You duplicate this reason in every commit. I propose to skip it.
Defconfig is not for KernelCI strictly. Defconfig is for developers, but
KernelCI is a system which can easily customize before deployments.
While it is an argument, I am not sure if it is worth keeping in these
13 commits...

Best regards,
Krzysztof


2022-11-09 15:02:49

by Nícolas F. R. A. Prado

[permalink] [raw]
Subject: Re: [PATCH 00/13] arm64: defconfig: Enable missing kconfigs for mt8183-kukui-jacuzzi-juniper

On Wed, Nov 09, 2022 at 10:08:27AM +0100, Krzysztof Kozlowski wrote:
> On 09/11/2022 08:28, Arnd Bergmann wrote:
> > On Wed, Nov 9, 2022, at 00:22, N?colas F. R. A. Prado wrote:
> >> mt8183-kukui-jacuzzi-juniper is one of the devices set up to run tests
> >> on KernelCI, but several of its drivers are currently disabled in the
> >> defconfig. This series enables all the missing kconfigs on the defconfig
> >> to get everything probing on that machine so that it can be fully tested
> >> by KernelCI.
> >
> > The changes all look fine, but I would recommend not separating it
> > out into 13 patches when you are doing just one thing here.
> >
> > As a general rule, if you keep saying the same things in each
> > patch description, it is usually an indication that they should
> > be combined. Similarly, if you find describing unrelated changes
> > ("also, ..."), that would be an indication that patches should
> > be split up.
>
> I agree. Descriptions you wrote are useful - they explain why you are
> doing it - but it got all really too detailed, just for defconfigs. One
> commit per one symbol is a bit too much...

Okay, thank you both for the feedback. Given that this is a contentious file,
splitting independent symbols throughout the commits seemed to be more friendly
to possible merge conflicts. But you're right that there's a single overarching
goal for all the changes, so I'll join them all into a single commit as
suggested.

Thanks,
N?colas