Skip to contentSkip to conversation

arm: imx6q_logic: Enable I2C PMIC

PATCH New Adam Ford 2 participants 4 messages Aug 18, 2020 – Aug 19, 2020

Subsystems (2)
lore
Share & tools
lore

Adam Ford

author
#0 lore raw

PATCH 1/1 arm: imx6q_logic: Enable I2C PMIC

The PMIC which powers the whole board is connected to the I2C1 bus. This patch enables I2C. With this enabled, the PMIC needs to told to keep the rail enabled, or it gets shutdown before the system boots.

arch/arm/dts/imx6q-logicpd-u-boot.dtsi +4 -0
index ee44ed91fe..94b53d569d 100644
@@ -28,3 +28,7 @@
 &pinctrl_usdhc2 {
 	u-boot,dm-spl;
 };
+
+&sw2_reg {
+	regulator-always-on;
+};
configs/imx6q_logic_defconfig +2 -0
index 35dfd8a303..d02c81449c 100644
@@ -9,6 +9,7 @@ CONFIG_ENV_SIZE=0x100000
 CONFIG_ENV_OFFSET=0x400000
 CONFIG_MX6_OCRAM_256KB=y
 CONFIG_TARGET_MX6LOGICPD=y
+CONFIG_SYS_I2C_MXC_I2C1=y
 CONFIG_SPL_TEXT_BASE=0x00908000
 CONFIG_SPL_MMC_SUPPORT=y
 CONFIG_SPL_SERIAL_SUPPORT=y
@@ -70,6 +71,7 @@ CONFIG_SPL_DM=y
 CONFIG_SPL_DM_SEQ_ALIAS=y
 CONFIG_SPL_OF_TRANSLATE=y
 CONFIG_PCF8575_GPIO=y
+CONFIG_SYS_I2C_MXC=y
 CONFIG_LED=y
 CONFIG_LED_GPIO=y
 CONFIG_FSL_USDHC=y
Signature · 1 line
2.25.1
reviewer
#1 lore raw

PATCH 1/1 arm: imx6q_logic: Enable I2C PMIC

Hi Adam,

Quoted 4 lines · Adam Ford

The PMIC which powers the whole board is connected to the I2C1 bus. This patch enables I2C. With this enabled, the PMIC needs to told to keep the rail enabled, or it gets shutdown before the

Looks like a typo.

Quoted 9 lines

system boots.

arch/arm/dts/imx6q-logicpd-u-boot.dtsi +0 -0
index ee44ed91fe..94b53d569d 100644

arch/arm/dts/imx6q-logicpd-u-boot.dtsi line 34

Quoted 7 lines
arch/arm/dts/imx6q-logicpd-u-boot.dtsi +4 -0
@@ -28,3 +28,7 @@
 &pinctrl_usdhc2 {
        u-boot,dm-spl;
 };
+
+&sw2_reg {
+       regulator-always-on;
+};

It seems that this could be fixed in the main dts instead of the U-Boot one since this dts description is not U-Boot specific.

author
#2 lore raw

PATCH 1/1 arm: imx6q_logic: Enable I2C PMIC

Quoted 10 lines · Fabio Estevam

Hi Adam,

Quoted 4 lines · Adam Ford

The PMIC which powers the whole board is connected to the I2C1 bus. This patch enables I2C. With this enabled, the PMIC needs to told to keep the rail enabled, or it gets shutdown before the

Looks like a typo.

Oops.

Quoted 5 lines
Quoted 5 lines

system boots.

Quoted 15 lines

arch/arm/dts/imx6q-logicpd-u-boot.dtsi line 34

Quoted 12 lines
arch/arm/dts/imx6q-logicpd-u-boot.dtsi +4 -0
index ee44ed91fe..94b53d569d 100644
@@ -28,3 +28,7 @@
 &pinctrl_usdhc2 {
        u-boot,dm-spl;
 };
+
+&sw2_reg {
+       regulator-always-on;
+};

It seems that this could be fixed in the main dts instead of the U-Boot one since this dts description is not U-Boot specific.

The issue I am seeing is U-Boot specific, because Linux doesn't power this rail down. I'd rather not keep this regulator-always-on in Linux, and I want to keep the devices trees between U-Boot and Linux the same. In the device tree, I have associated this rail to the SDHC interface , but for some reason, U-Boot still powers it down, so when it tries to go and read from the SD card, the system hangs.

If you're ok with this rationale, I can update the commit message to add some of this background info to V2 when I fix the typo above.

adam

reviewer
#3 lore raw

PATCH 1/1 arm: imx6q_logic: Enable I2C PMIC

Hi Adam,

Quoted 3 lines · Adam Ford

The issue I am seeing is U-Boot specific, because Linux doesn't power this rail down. I'd rather not keep this regulator-always-on in Linux, and I want to keep the devices trees between U-Boot and Linux

Yes, I also prefer to always keep them in sync.

Quoted 3 lines

the same. In the device tree, I have associated this rail to the SDHC interface , but for some reason, U-Boot still powers it down, so when it tries to go and read from the SD card, the system hangs.

Interesting. Then this needs to be investigated, otherwise we will just paper over the real problem without understanding its root cause.

Why is U-Boot powering the regulator down?

The device tree representation of the regulators is agnostic to U-Boot or kernel, as it just represents the hardware, so I prefer that the correct regulator description is passed in the main device tree instead of hacking the specific U-Boot device tree.