Brad Campbell
Brad Campbell
> I would be interested to understand how the all-dynamic looks. Having two mechanisms for a low-level, correctness-sensitive part of the execution flow feels like it increases complexity a lot...
> Update: I am still working on fully static assignment of deferred calls, from both capsules and chip peripherals, but there are some tricky subproblems that arise as a result...
If we re-implement IPC to use message passing rather than shared memory, we can revisit the MPU implementation (re: #1532).
I personally have no general issue with making breaking changes to tock-registers. We could make tock-registers 2.0. Having both 1.x and 2.x is also fine with me, although in that...
I added a commit for an effective but maybe not elegant way to remote `unwrap()` from console.
While working on the change to console and thinking about the grant/processbuffer API, as well as trying to reason about what would happen if a process went away in between...
I'm in favor of this change, and am working on converting a capsule to use slice copy functions rather than manual for loops.
I added a few more commits.
I haven't thought a lot about what a wifi hil should look like, or how a wifi hil would fit into a more generic network stack, but this looks like...