internal/task: use a lock-free queue

[niaow]

trafficstars

This changes task.Stack and task.Queue to use atomic compare-and-swap operations. With this change, Push can also be executed safely from multiple threads/cores. Additionally, the queue is actually LIFO now (not sure what I was thinking before).

Not entirely sure about this yet.

[niaow]

Oops, i meant FIFO not LIFO. Also xtensa seems to still need work?

[niaow]

Size impact statistics:

• Arduino Uno Blinky (tasks): adds 172 bytes
• Arduino Nano 33 Blinky: adds 84 bytes
• Metro M4 Airlift Blinky: adds 40 bytes
• ESP32-Mini32 Blinky: adds 40 bytes
• Linux Channel Test: adds 76 bytes

[niaow]

Out of curiosity: do you have plans to use the changes in this PR?

I am not intending to do anything specific right now, but this might be useful once we want to implement multicore support for something.

Any new features this will add?

Right now:

1. The queue. . . is actually a queue instead of a weird mostly-stack
2. On devices with support for atomics, this does not need any critical sections

The Push operation can also be safely called concurrently. . . but only if the tasks are pinned safely (which is currently an issue, but this is more of a future thing).

[niaow]

This has been rebased now that #2307 is merged

[niaow]

I still need to test this fully on AVR though

[niaow]

Now that #2458 and #2459 are in, I consider this ready.

[deadprogram]

Seems like this is perhaps not working correctly on the HiFive1b? I tried several times, and also ran some other test code successfully.

[jclab-joseph]

@aykevl If this is correct, would you be willing to merge it? I'm going to try implementing threads again.

[aykevl]

I think we can close this now? I don't see what features it adds or what bugs it fixes, while it does increase binary size.