Xavier reboot issue

Autonomous Machines Jetson & Embedded Systems Jetson AGX Xavier
1

Dear Sir,

We have a problem while running Xavier: Probabilistic restart.
Look to the kernel log, we find pmc reset information, but we are not sure about it:

Before reboot:

tegra-pmc c360000.pmc: scratch reg offset dts data not present
tegra-pmc: get_secure_pmc_setting: done secure_pmc=0
tegra-pmc: ### PMC reset source: TEGRA_SOFTWARE_RESET
tegra-pmc: ### PMC reset level: TEGRA_RESET_LEVEL_L1
tegra-pmc: ### PMC reset status reg: 0x2d
tegra-pmc: PMC Prod config success

After reboot:

tegra-pmc c360000.pmc: scratch reg offset dts data not present
tegra-pmc: get_secure_pmc_setting: done secure_pmc=0
tegra-pmc: ### PMC reset source: TEGRA_BCCPLEX_WATCHDOG
tegra-pmc: ### PMC reset level: TEGRA_RESET_LEVEL_L1
tegra-pmc: ### PMC reset status reg: 0x9
tegra-pmc: PMC Prod config success

kern.log (197 KB)
syslog.log (502 KB)

2

Hi dennis_Leetop,

Issue was happened on NVIDIA devkit? What’s the L4T SW version? What kinds of test are running to repro this issue?

Thanks

3

Dear kayccc,

Not NVIDIA Devkit, this issue was happened on our own carrier board, we are trying to do more tests to locate the issue root cause. If there is any other information, we will push here.

Thanks.

4

For what it’s worth, I’ve found the Jetson to be super finicky about changes in voltage, both on the input to the devkit board (which contains switching regulators,) and coming out of the regulators (which presumably you built your own if you have your own carrier board.)

Try hooking up a digital storage scope with a trigger set just below the steady-state 5V level (or the other levels for the other rails,) and see if there’s a voltage glitch right when it restarts. You may need to use a faster switching controller and/or more, bigger, lower-ESR capacitors.