[GloriousAlpaca]

Fixes #2748
Before ConsoleStdin::poll() did not register a waker so if no input has been received before calling select(), the thread is stuck, even when receiving input afterwards.

I have edited the poll function to be closer to ConsoleStdin::read(), so it now correctly registers the waker and checks for input when a UART Interrupt is received.

This has fixed the reproducer that did not function before:

#include <stdio.h>
#include <sys/select.h>

int main()
{
    fd_set readfds;

    FD_ZERO(&readfds);
    FD_SET(0, &readfds);
    printf("Select Test: Before select() call\n");
    fflush(stdout);
    int result = select(1, &readfds, (fd_set*)NULL, NULL, NULL);
    printf("Select Test: After select() call, result = %d\n", result);
    return 0;
}

[github-actions]

Benchmark Results

Details

Benchmark	Current: b2097d0	Previous: 2e23902	Performance Ratio
startup_benchmark Build Time	82.90 s	80.34 s	1.03 ❗
startup_benchmark File Size	0.84 MB	0.80 MB	1.05 ❗
Startup Time - 1 core	0.70 s (±0.02 s)	0.75 s (±0.02 s)	0.94 ❗
Startup Time - 2 cores	0.73 s (±0.02 s)	0.74 s (±0.02 s)	0.99
Startup Time - 4 cores	0.74 s (±0.01 s)	0.74 s (±0.02 s)	0.99
multithreaded_benchmark Build Time	84.46 s	82.11 s	1.03 ❗
multithreaded_benchmark File Size	0.94 MB	0.86 MB	1.10 ❗
Multithreaded Pi Efficiency - 2 Threads	92.23 % (±8.64 %)	85.89 % (±6.61 %)	1.07
Multithreaded Pi Efficiency - 4 Threads	44.56 % (±3.71 %)	43.43 % (±2.56 %)	1.03
Multithreaded Pi Efficiency - 8 Threads	26.17 % (±1.64 %)	25.76 % (±1.53 %)	1.02
micro_benchmarks Build Time	84.19 s	80.40 s	1.05 ❗
micro_benchmarks File Size	0.94 MB	0.86 MB	1.10 ❗
Scheduling time - 1 thread	64.60 ticks (±3.39 ticks)	62.65 ticks (±4.06 ticks)	1.03
Scheduling time - 2 threads	36.20 ticks (±3.55 ticks)	34.08 ticks (±4.10 ticks)	1.06
Micro - Time for syscall (getpid)	3.72 ticks (±0.60 ticks)	3.45 ticks (±0.58 ticks)	1.08
Memcpy speed - (built_in) block size 4096	81604.86 MByte/s (±56575.84 MByte/s)	82448.38 MByte/s (±56997.13 MByte/s)	0.99
Memcpy speed - (built_in) block size 1048576	30140.45 MByte/s (±24405.16 MByte/s)	30585.98 MByte/s (±24707.84 MByte/s)	0.99
Memcpy speed - (built_in) block size 16777216	28803.89 MByte/s (±23700.74 MByte/s)	26340.06 MByte/s (±21720.96 MByte/s)	1.09
Memset speed - (built_in) block size 4096	81915.23 MByte/s (±56797.42 MByte/s)	82292.76 MByte/s (±56891.50 MByte/s)	1.00
Memset speed - (built_in) block size 1048576	30858.39 MByte/s (±24822.25 MByte/s)	31323.85 MByte/s (±25145.86 MByte/s)	0.99
Memset speed - (built_in) block size 16777216	29563.62 MByte/s (±24141.28 MByte/s)	27104.68 MByte/s (±22209.94 MByte/s)	1.09
Memcpy speed - (rust) block size 4096	74069.60 MByte/s (±51678.12 MByte/s)	74097.96 MByte/s (±51811.44 MByte/s)	1.00
Memcpy speed - (rust) block size 1048576	30052.47 MByte/s (±24377.01 MByte/s)	30361.60 MByte/s (±24602.37 MByte/s)	0.99
Memcpy speed - (rust) block size 16777216	28913.43 MByte/s (±23816.93 MByte/s)	27625.34 MByte/s (±22806.88 MByte/s)	1.05
Memset speed - (rust) block size 4096	74532.83 MByte/s (±52018.63 MByte/s)	74373.47 MByte/s (±51976.48 MByte/s)	1.00
Memset speed - (rust) block size 1048576	30786.20 MByte/s (±24799.64 MByte/s)	31110.89 MByte/s (±25033.24 MByte/s)	0.99
Memset speed - (rust) block size 16777216	29674.71 MByte/s (±24253.88 MByte/s)	28386.93 MByte/s (±23265.03 MByte/s)	1.05
alloc_benchmarks Build Time	82.02 s	74.76 s	1.10 ❗
alloc_benchmarks File Size	0.92 MB	0.87 MB	1.05 ❗
Allocations - Allocation success	91.37 %	91.31 %	1.00 ❗
Allocations - Deallocation success	100.00 %	100.00 %	1
Allocations - Pre-fail Allocations	61.61 %	61.44 %	1.00 ❗
Allocations - Average Allocation time	6314.74 Ticks (±82.44 Ticks)	5860.58 Ticks (±98.43 Ticks)	1.08 ❗
Allocations - Average Allocation time (no fail)	6880.72 Ticks (±108.38 Ticks)	6554.81 Ticks (±92.86 Ticks)	1.05 ❗
Allocations - Average Deallocation time	1268.95 Ticks (±253.67 Ticks)	1805.01 Ticks (±250.35 Ticks)	0.70 ❗
mutex_benchmark Build Time	81.79 s	79.82 s	1.02 ❗
mutex_benchmark File Size	0.94 MB	0.86 MB	1.10 ❗
Mutex Stress Test Average Time per Iteration - 1 Threads	12.22 ns (±0.46 ns)	12.10 ns (±0.41 ns)	1.01
Mutex Stress Test Average Time per Iteration - 2 Threads	69.24 ns (±3.80 ns)	40.26 ns (±1.68 ns)	1.72 ❗

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

[mkroening]

We only need to register the waker when retuning Pending.

[GloriousAlpaca]

I think there's a possibility of receiving an interrupt between checking for readiness and registering a waker. I'm not sure how high the chance for this to occur is during normal execution, or if there's something in place preventing this?

[mkroening]

You are right. We could either check again before returning Pending to optimize the fast path or restore your original proposal. Both are fine with me. Performance should not really be an issue here.

[GloriousAlpaca]

I think I'd prefer the original version, as it looks a bit cleaner.
I believe ConsoleStdin::read() has the same possibility, should I just do a separate commit in this pr?

[mkroening]

Sure, go ahead! :)

[mkroening]

Looks good to me, thanks! :)