NVIDIA Tegra
NVIDIA DRIVE OS 5.1 Linux

Developer Guide
5.1.0.2 Release


 
Probing RDC Pins
 
Probing RDC Pins in Quickboot
Probing RDC Pins in Kernel
RDC pins are located at the corners of your VCM and are connected in a daisy chain. They are used to determine whether the VCM is still attached to the base board. The RDC pins should be probed during boot up and periodically thereafter for quick and effective debug for any lifted balls or bump cracks. The corner balls that are most likely to develop problems are used as early indicators of any problems.
For hardware VCM30T124 details, see NVIDIA VCM Tegra K1 32 Bit Automotive Products Data Sheet.
Probing RDC Pins in Quickboot
Applies to: The information in this topic applies only S02P version of MCM.
During boot, Quickboot checks the RDC_IN ball signal:
Signal 0 indicates that connections are good.
Signal 1 indicates there is loose connection.
You can also define your own alert mechanism.
To detect the chain-break during boot
1. Build Quickboot with Debugging mode enabled, which provides debug logging and a debug console with utilities.
For more information, see Quickboot: Customizing.
2. Ensure the terminal is setup before proceeding with flashing.
3. Flash the debug Quickboot images with -C option.
4. Check the terminal where you can see the status of the RDC_IN pin:
If the VCM is not connected properly the following displays:
<cpu>: RDC_IN value: 1 MCM/VCM Revision Version: 1
<cpu>: FAILURE : VCM3.0/MCM is not attached to the board
If the VCM is connected properly the terminal displays:
<cpu>: RDC_IN value: 0
MCM/VCM Revision Version: 1
To define a custom alert mechanism to indicate a loose connection
1. Locate and edit the following file:
bootloader/qb/quickboot-partner/arch/<chip>/core/cop_init.c
 
2. Add the error handling code, such as:
if (RegVal) {
// MCM is not attached to the board.
// Implement your system's alert
// notification mechanism here.
}
Probing RDC Pins in Kernel
You should periodically monitor the RDC chain to ensure that the attachment to the PCB is still intact.
Note:
NVIDIA does not provide an OS runtime kernel implementation that responds to a break. You are responsible for implementing such a failure handler.
To detect the chain-break at runtime
In kernel, probe GPIO_PV7(ball CRT_VSYNC).