Dear nvdia teams:
I want to use kdump to grab kernel panic stack information. Please help confirm how to use the kdump tool.
Default configuration:
kernel version: 5.10.104-tegra
tegra_defconfig: CONFIG_CRASH_DUMP=y
As you can see, kdump is enabled by default. But there is no information in the / var/crash directory when it crashes.
I haven’t used those tools in years, and not with a 5.x kernel. However, it is very likely you need to know about the Magic SysRq feature. The following was related to using sysrq, although in a different context:
https://forums.developer.nvidia.com/t/enter-oem-interface-after-reboot/235298/34
This is related, and might be the addition to the above which you need:
https://forums.developer.nvidia.com/t/save-kernel-crash-and-panic-dmesg/235540
Note that not all sysrq features are implemented on every o/s, and some may not even be enabled without a setting in “/etc”, but here is a more general list of sysrq information:
https://en.wikipedia.org/wiki/Magic_SysRq_key
Note that either key bindings or an echo of the right value to a file in “/proc” will trigger the bindings which are active. Various kernel information dumps will be available depending on architecture and o/s.
Hi,
We have checked and confirmed this is not supported on Jetpack 5.1. Will evaluate to support it in future release.
This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.