Pat Pannuto

Continuing to close the loop on this, I just got the ping that the upstream llvm bug is now considered resolved: https://bugs.llvm.org/show_bug.cgi?id=46368 -- though I think we already had evidence...

> I have a pull request prepared, but I don't have access to create a new branch. You don't need to make a branch on the main Tock repo to...

> I wish there were a way to force an interrupt from the USART. With the caveat of never having had occasion to write code that does it, I'm pretty...

What started me down this path was trying to move the `set_client` call for the UART from the top-level board file into the `initialize` method of the console capsule. We...

> There's no way to ensure a CommandReturn contains the expected value. IMO CommandReturn::into_inner should be pub, not pub(crate). I believe we discussed this briefly at TockWorld and agreed it...

If we can, I think I would like to separate this in to two threads of discussion, as I think there are two different cases here. We have some uses...

~FWIW, I think that `volatile` is necessary, but I am not arguing that `volatile` alone is sufficient – we can chat more on the latter tomorrow.~ https://github.com/tock/tock/pull/3095#issuecomment-1208688711

After the conversation that Leon and I had last week, I agree that volatile should not be necessary (or used) for the [un]stacking. I do still think we should split...

re exception handler signature: we can probably pull this discussion to a separate PR, but yeah, I think the 'right' thing to do is to mark them as `aapcs` with...

FWIW, at least once one Tock platform supported GNSS and that was the approach it took: https://github.com/lab11/signpost-software/blob/f13b32e0661849024b4817ceb08f47fe980277ff/signpost/apps/libsignpost-tock/gps.c -- this was the motivation behind the UartAdvanced which listened for a logically...