scheduler: add optional preemptive scheduling

[stlankes]

The scheduler is tickless (a TSC-deadline one-shot timer), so a CPU-bound task that never blocks or yields keeps its core indefinitely. Add an optional preemptive feature that round-robins between ready tasks via a time slice.

A new Source::Preemption timer slot is armed for now + 10ms whenever the scheduler switches to a task while others are ready, and whenever custom_wakeup makes a task ready (so a freshly woken task preempts a non-yielding runner soon). When the slice expires, the timer handler's existing reschedule() switches to the next ready task. With no contention nothing is armed, so the system stays tickless.

[github-actions]

Benchmark Results

Details

Benchmark	Current: 11bf411	Previous: 2e23902	Performance Ratio
startup_benchmark Build Time	93.54 s	80.34 s	1.16 ❗
startup_benchmark File Size	0.79 MB	0.80 MB	1.00 ❗
Startup Time - 1 core	0.76 s (±0.03 s)	0.75 s (±0.02 s)	1.02
Startup Time - 2 cores	0.76 s (±0.02 s)	0.74 s (±0.02 s)	1.04
Startup Time - 4 cores	0.78 s (±0.02 s)	0.74 s (±0.02 s)	1.05 ❗
multithreaded_benchmark Build Time	95.96 s	82.11 s	1.17 ❗
multithreaded_benchmark File Size	0.85 MB	0.86 MB	0.99 ❗
Multithreaded Pi Efficiency - 2 Threads	90.77 % (±8.39 %)	85.89 % (±6.61 %)	1.06
Multithreaded Pi Efficiency - 4 Threads	44.87 % (±3.35 %)	43.43 % (±2.56 %)	1.03
Multithreaded Pi Efficiency - 8 Threads	26.00 % (±1.51 %)	25.76 % (±1.53 %)	1.01
micro_benchmarks Build Time	93.83 s	80.40 s	1.17 ❗
micro_benchmarks File Size	0.86 MB	0.86 MB	0.99 ❗
Scheduling time - 1 thread	67.55 ticks (±2.79 ticks)	62.65 ticks (±4.06 ticks)	1.08
Scheduling time - 2 threads	36.28 ticks (±3.61 ticks)	34.08 ticks (±4.10 ticks)	1.06
Micro - Time for syscall (getpid)	3.85 ticks (±0.60 ticks)	3.45 ticks (±0.58 ticks)	1.12
Memcpy speed - (built_in) block size 4096	74947.50 MByte/s (±51969.17 MByte/s)	82448.38 MByte/s (±56997.13 MByte/s)	0.91
Memcpy speed - (built_in) block size 1048576	29505.92 MByte/s (±23983.51 MByte/s)	30585.98 MByte/s (±24707.84 MByte/s)	0.96
Memcpy speed - (built_in) block size 16777216	24039.37 MByte/s (±19938.63 MByte/s)	26340.06 MByte/s (±21720.96 MByte/s)	0.91
Memset speed - (built_in) block size 4096	74711.39 MByte/s (±51798.70 MByte/s)	82292.76 MByte/s (±56891.50 MByte/s)	0.91
Memset speed - (built_in) block size 1048576	30249.36 MByte/s (±24420.09 MByte/s)	31323.85 MByte/s (±25145.86 MByte/s)	0.97
Memset speed - (built_in) block size 16777216	24751.82 MByte/s (±20406.28 MByte/s)	27104.68 MByte/s (±22209.94 MByte/s)	0.91
Memcpy speed - (rust) block size 4096	67439.24 MByte/s (±47243.29 MByte/s)	74097.96 MByte/s (±51811.44 MByte/s)	0.91
Memcpy speed - (rust) block size 1048576	29348.87 MByte/s (±23966.27 MByte/s)	30361.60 MByte/s (±24602.37 MByte/s)	0.97
Memcpy speed - (rust) block size 16777216	24254.13 MByte/s (±20065.72 MByte/s)	27625.34 MByte/s (±22806.88 MByte/s)	0.88
Memset speed - (rust) block size 4096	67951.00 MByte/s (±47588.57 MByte/s)	74373.47 MByte/s (±51976.48 MByte/s)	0.91
Memset speed - (rust) block size 1048576	30104.63 MByte/s (±24450.61 MByte/s)	31110.89 MByte/s (±25033.24 MByte/s)	0.97
Memset speed - (rust) block size 16777216	25011.11 MByte/s (±20581.05 MByte/s)	28386.93 MByte/s (±23265.03 MByte/s)	0.88
alloc_benchmarks Build Time	92.14 s	74.76 s	1.23 ❗
alloc_benchmarks File Size	0.87 MB	0.87 MB	1.00 ❗
Allocations - Allocation success	91.38 %	91.31 %	1.00 ❗
Allocations - Deallocation success	100.00 %	100.00 %	1
Allocations - Pre-fail Allocations	61.60 %	61.44 %	1.00 ❗
Allocations - Average Allocation time	11391.38 Ticks (±411.00 Ticks)	5860.58 Ticks (±98.43 Ticks)	1.94 ❗
Allocations - Average Allocation time (no fail)	11885.29 Ticks (±468.35 Ticks)	6554.81 Ticks (±92.86 Ticks)	1.81 ❗
Allocations - Average Deallocation time	2792.86 Ticks (±910.99 Ticks)	1805.01 Ticks (±250.35 Ticks)	1.55 ❗
mutex_benchmark Build Time	85.72 s	79.82 s	1.07 ❗
mutex_benchmark File Size	0.86 MB	0.86 MB	0.99 ❗
Mutex Stress Test Average Time per Iteration - 1 Threads	13.14 ns (±0.49 ns)	12.10 ns (±0.41 ns)	1.09 ❗
Mutex Stress Test Average Time per Iteration - 2 Threads	89.46 ns (±2.73 ns)	40.26 ns (±1.68 ns)	2.22 ❗

This comment was automatically generated by workflow using github-action-benchmark.

[Gelbpunkt]

Suggested change

	let deadline = processor::get_timer_ticks() + PREEMPTION_SLICE_US;
	timer_interrupts::create_timer_abs(timer_interrupts::Source::Preemption, deadline);
	timer_interrupts::create_timer(timer_interrupts::Source::Preemption, PREEMPTION_SLICE_US);

same for all the other calls. The function create_timer needs its #[cfg(feature = "net")] replaced with #[cfg(any(feature = "net", feature = "preemptive"))]

[mkroening]

Please address @Gelbpunkt's comment. Then, this is fine with me. 👍

[stlankes]

Why is feature net required?