Dario Nieuwenhuis

while we're doing breaking changes, perhaps remove the [prelude](https://github.com/rust-embedded/cortex-m/blob/master/src/prelude.rs), since embedded-hal 1.0 has [removed it](https://github.com/rust-embedded/embedded-hal/pull/282)?

I just had to fight CMRT to workaround the mandatory trampoline because I needed access to more things (like r4-r11), so +1 to make it optional. Alternative idea: why not...

this seems to be a fallback to get the stlink "unstuck", see https://github.com/probe-rs/probe-rs/blob/ebe2f08ca245c532ff23e6bb02655ae63ed89b91/probe-rs/src/probe/stlink/mod.rs#L522-L523 in normal operation stlink shouldn't need a reset, and it works fine for me on Windows. I'd...

> When connecting to a onboard ST-LINK V2-1 nucleo board which has not closed proberby by the previous program. @vDorst could you provide exact steps to reproduce this? Maybe there's...

I think there's two somewhat orthogonal issues here: A. Packets are not dropped from the socket queue on neighbor discovery timeout. B. NeighborState assumes "1 socket = 1 peer". Fixing...

Bumping packets to the back of the queue would mean packets destined for the same peer could get reordered, though. I was thinking something more along the lines of: iterate...

Sorry for the :snail: response! - I think the ID should be probably a `u32`, so that it doesn't change between 32bit and 64bit platforms. - Ideally, the feature shouldn't...

> A u32 is probably big enough? At any rate it'd be relatively easy to change in the future. u32 sounds good to me. > how many tests should I...

I think the "assign IDs incrementally vs let user/device assign them concern" concern I initially raised is still unsolved. You mention: > In stm32-eth we only guarantee that the timestamp...

I'm not sure? the current code will ignore expired routes, but it won't actually remove them from the route table.