Nathaniel Wesley Filardo

Whoop, I see I missed being tagged here. Sorry! The choice was not completely arbitrary, though I suspect we could live with either option. Summarizing [the PR discussion](https://github.com/CHERIoT-Platform/cheriot-sail/pull/83), we think...

Based on our (admittedly, not, y'know, decades of) experience with CHERIoT software, we don't see a lot of code that is polymorphic in whether a pointer is sealed or not....

Two thoughts: 1. We haven't, yet, needed generic helpers for domain crossings like that. We have had some utility functions for checking pointers against permission sets (the RTOS's `check_pointer` and...

Moreover, I think this approach does not scale. CHERIoT is surely not the only thing that wants to be "standard CHERI with some extra flare" (i.e., an extension to the...

> If we can't fully define the behaviour of an instruction, which we can't if the encoding format or permissions change, then it needs to be a new base. That...

FWIW, the comparison to CHERIoT and research-CHERI only for the two-level version. In particular, AFAIK, there has been no work demonstrating the practical sensibility/utility/viability of multi-level schemes. From hallway chatter...

FWIW, Zstid is significantly less "research". At the very least, it doesn't have configuration knobs like `Zcherilevels` and we know, concretely, how we might use it if we had it...

I am not sure why you are bringing those points up on _this_ issue. In any case, my current proposal, in #716, leaves cbld in the base. It doesn't hurt...

(Hilariously, running `node` under `qemu-x86_64`'s user emulation allows the build step to work)

FWIW, libusb1 as of 1.0.16 sports a [`libusb_get_port_numbers()`](https://libusb.sourceforge.io/api-1.0/group__libusb__dev.html#gaa4b7b2b50a9ce2aa396b0af2b979544d) API that returns the entire device path, rather than just the terminal port number. That would be a stable way to refer...