New dtschema v2025.6 enforces different naming on I2C nodes thus new
dtbs_check warnings appeared for I2C GPIO nodes:
exynos4212-tab3-lte8.dtb: i2c-gpio-3 (i2c-gpio):
$nodename:0: 'i2c-gpio-3' does not match '^i2c(@.+|-[a-z0-9]+)?$'
exynos4212-tab3-lte8.dtb: i2c-gpio-3 (i2c-gpio):
Unevaluated properties are not allowed ('#address-cells', '#size-cells', 'magnetometer@2e' were unexpected)
Rename the nodes to a generic i2c-[0-9]+ style with numbers continuing
the SoC I2C controller indexing (Exynos3250: 8 controllers, Exynos4: 9
controllers) for simplicity and obviousness, even if the SoC I2C
controller is not enabled on given board. The names anyway would not
conflict with SoC ones because of unit addresses.
Verified with comparing two fdt (after fdtdump).
Reported-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Closes: https://lore.kernel.org/all/aCtD7BH5N_uPGkq7@shikoro/
Link: https://lore.kernel.org/r/20250612094807.62532-3-krzysztof.kozlowski@linaro.org
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
This was initially copied from the Midas DTSI, but there is no
proof that the same interrupt is also used on the Tab 3. The pin
listed as the interrupt here is GPIO_HDMI_CEC on the Midas,
but for the Tab 3 it is the headset button GPIO - GPIO_EAR_SEND_END.
Drop the interrupt, since there is no proof that it is used.
Signed-off-by: Artur Weber <aweber.kernel@gmail.com>
Link: https://lore.kernel.org/r/20240816-midas-audio-tab3-v2-5-48ee7f2293b3@gmail.com
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
In the schematics, the MCLK2 pin is shown as connected to CODEC_CLK32K,
which is derived from the same 32KHZ_PMIC clock as Bluetooth/WiFi and
GPS clocks. 32KHZ_PMIC is connected to the BTCLK pin, represented in
mainline as S2MPS11_CLK_BT.
Add the MCLK2 clock to the WM1811 codec clock property to properly
describe the hardware.
Signed-off-by: Artur Weber <aweber.kernel@gmail.com>
Link: https://lore.kernel.org/r/20240816-midas-audio-tab3-v2-4-48ee7f2293b3@gmail.com
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Set up headset mic bias regulator and add the necessary properties to
the samsung,midas-audio node to allow for headset jack detection.
Signed-off-by: Artur Weber <aweber.kernel@gmail.com>
Link: https://lore.kernel.org/r/20240816-midas-audio-tab3-v2-3-48ee7f2293b3@gmail.com
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
The stock bootloader on the Samsung Galaxy Tab 3 8.0 provides an
incorrect available memory range over ATAG_MEM. Limit the usable
memory in the DTS to prevent it from doing so, without having to
disable ATAG support.
Signed-off-by: Artur Weber <aweber.kernel@gmail.com>
Reviewed-by: Henrik Grimler <henrik@grimler.se>
Link: https://lore.kernel.org/r/20240217-tab3-limit-usable-memory-range-v1-1-49cc9c86a5cc@gmail.com
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
After more investigation, I've found that it's not the panel driver
config that needs to be modified to invert the data polarity, but
the FIMD config.
Add the missing invert-vclk option that is required to get the display
to work correctly.
Fixes: ee37a457af ("ARM: dts: exynos: Add Samsung Galaxy Tab 3 8.0 boards")
Signed-off-by: Artur Weber <aweber.kernel@gmail.com>
Link: https://lore.kernel.org/r/20240105-tab3-display-fixes-v2-1-904d1207bf6f@gmail.com
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
