This ensures the grant region is created for the app whenever it uses
the GPIO driver. This in turn ensures that upcalls work, even if the app
uses Yield-WaitFor.

The function comment did not get updated for Yield-WaitFor.

Specify that this queues even if the upcall is null. Also our internal
note about using a processid can just be a // comment.

Add a brief note about how the kernel avoids an issue where an upcall is
lost between an app asking for an event and calling YWF.

Perhaps this is an ok place to document that any capsule that
implements SyscallDriver and uses upcalls MUST ensure the grant is
allocated to ensure that YWF will work. Capsules can't schedule upcalls
if the grant region is not allocated.

Trying to enable/disable interrupts will only work if the grant was
allocated to ensure upcalls are possible.

Also update the gpio syscall doc.

Core working group call notes: April 08, 2026

Network working group call notes: March 30, 2026

…ation

doc: kernel: sysdriver: add note about grants

kernel: upcall: update schedule doc

doc: trd104: add note about YWF race condition avoidance

The `PanicWriter` trait allows creation of a synchronous,
reinitialized Uart-like peripheral for use in panic handlers.

This supports passing in a Writer

tested with the panic command in the process console

kernel: move write to binary to capsules

Core working group call notes: April 29, 2026

kernel: migrate to Rust 2024 edition

This nightly predates the CLI redesign, hopefully this fixes CI.

flux-ci-runner: pin fixpoint's version

Bumps [urllib3](https://github.com/urllib3/urllib3) from 2.6.3 to 2.7.0.
- [Release notes](https://github.com/urllib3/urllib3/releases)
- [Changelog](https://github.com/urllib3/urllib3/blob/main/CHANGES.rst)
- [Commits](urllib3/urllib3@2.6.3...2.7.0)

---
updated-dependencies:
- dependency-name: urllib3
  dependency-version: 2.7.0
  dependency-type: direct:production
...

Signed-off-by: dependabot[bot] <support@github.com>

Checks the m0 crate

These crates define their start trap functions twice:

1. Once `#[cfg]`'d to only show up when compiled for their devices, implemented as a naked function that jumps to `rv32i::_start_trap`.
2. Once `#[cfg]`'d to show up in all other cases, so e.g. `cargo check` passes without `--target`.

`rv32i::_start_trap` (in both compilation modes) and the group 2 definitions have return type `()`, while the group 1 definitions have return type `!`. This makes the function cast from PR tock#4801 fail.

This changes all of the functions to have return type `()`. Ultimately, I don't believe the return type matters here -- for the naked implementatinos the return type doesn't really matter, and for the non-naked implementation both are fine. But they should be consistent.

This is crashing the llvm-objdump included with newer nightly Rust toolchains. It also doesn't seem to be needed, at least with LLVM's objdump?

…t/urllib3-2.7.0

build(deps): bump urllib3 from 2.6.3 to 2.7.0 in /tools/ci/tockbot

Update Rust to 2026-04

rp2040, rp2350: Remove `resolve_dependencies()`

lpc55s6x: use static refs, not optionalcells