We’ve received some complaints from our customers of in-field devices which sometimes seem to hang and need a full power-cycle to start up again. We’ve determined that the problem itself is that the TX2 hangs in Cboot, after a software reboot (linux kernel watchdog reboot).
[0001.866] W> Return default slot: _a
[0001.869] I> A/B: bin_type (24) slot 0
[0001.877] I> Boot image size read from image header: f2d10
[0001.882] I> Boot image load address: 0x0x80400000
[0001.887] I> Loading partition kernel at 0x80400000 from device(0x1)
[0002.636] I> Kernel hdr @0x80400000
[0002.639] I> Kernel dtb @0x80000000
[0002.643] I> T18x: Authenticate boot.img (bin_type 24), max size 0x4000000
[0001.869] W> Return default slot: _a
[0001.872] I> A/B: 0
Tegra186 (P2771-0000-500) #
What is the cause of this? And more importantly, how can we fix this/avoid this? Is there a watchdog in Cboot that we can enable? Because it starts correctly again after such an error when a power cycle is done.
Custom carrier board is used
Custom device tree is used
Cboot is unmodified
Jetpack 4.2.3 is used
Hopefully someone can help us with this, as our end-devices are not easily reached and thus a power-cycle can’t be easily done!