Hello Nvidia Team,
I have a question regarding the expected behavior of the SOC_ERROR (K37) pin on the Jetson AGX Orin.
.
My expectation was that this pin would remain high by default and pull low (active low) when a critical system error occurs, such as a kernel panic. To test this, I triggered a kernel crash using:
sudo echo c > /proc/sysrq-trigger
However, the SOC_ERROR pin did not active to low after running this command. Instead, the system hung and rebooted. Additionally, I observed that the pin was pulled low starting from the bootloader stage, which seems unexpected.
.
Here is my pinmux configuration for pin K37 (soc_error_pu0):
soc_error_pu0 {
nvidia,pins = "soc_error_pu0";
nvidia,function = "soc";
nvidia,pull = <TEGRA_PIN_PULL_NONE>;
nvidia,tristate = <TEGRA_PIN_DISABLE>;
nvidia,enable-input = <TEGRA_PIN_DISABLE>;
nvidia,io-high-voltage = <TEGRA_PIN_DISABLE>;
nvidia,lpdr = <TEGRA_PIN_DISABLE>;
};
Is SOC_ERROR expected to trigger on kernel panics (sysrq-trigger), or what else can I do to test this SOC_EEROR?
My Jetpack is 6.2. kernel 5.15.148