From what I read on the Apple site that d3cbgf linked in his post and from what is written in the panic log file, I don‘t think this issue is a kernel panic (KP) of macOS. In case of a kernel panic we would have seen a typical function call backtrace of the kernel in the panic log, but we don‘t see this here.
It looks more that the secure enclave processor or the application processor (they both work close together on the SoC) have a kind of „panic“. As Apple it self points out in the linked site these are two processors that are completely independent from the CPU processors (ARM) where macOS and the kernel is running on. On the SoC runs its own operating system called sepOS and I think this OS have a „panic“ for whatever reason.
They point out a few „reset“ events when SoC needs to restart (and maybe macOS too, since secure enclave seems very important for its security), but it could be not all "reset" reasons are mentioned on the linked site. These are the mentioned „reset“ reasons as far as I understand them:
- a Signature failure during booting of sepOS
- Power and clock monitors of SoC detect an illegal operating point of voltage or frequency
- Memory Protection Engine in SoC detect an memory authentication error
I found the mentioned power and clock monitors interesting that supervise the voltage and frequency of the SoC to prevent external attacks of the SoC. I wonder if some external USB devices, like the very often by users here reported connected external USB devices, could have a negative impact to the voltage of the power rail and if the power monitor misinterprets this as an external attack and trigger this reset? But without any schematics and an expert expert analysing them, we will not know this.
Could be also a bug in sepOS that leads to this also mentioned memory authentication error or there is a bug in the way macOS communicate with the SoC over the I2C bus that leads to this memory authentication error? In the end I have no idea what the reason(or reasons) is that is causing this SoC reset.
Only the vendor can analyse this issue deeper and find the exact reason and hopefully will fix it.
