fix(mmio)

[edef1c]

1. In check_linux_args, we return VAs to the free list while a VolatileRef to them is active. This breaks multi-device configurations. A similar bug exists in guess_device, but the current version of this patch doesn't fix it, since my reference target supplies mmio args.
2. Per virtio spec 4.2.3.1, after writing 0 to the status register the driver must wait for the read-back to return 0 before considering the reset complete. We should spin until it does.

[github-actions]

Benchmark Results

Details

Benchmark	Current: 253b948	Previous: 4912505	Performance Ratio
startup_benchmark Build Time	102.52 s	99.11 s	1.03 ❗
startup_benchmark File Size	0.86 MB	0.86 MB	1.00 ❗
Startup Time - 1 core	0.96 s (±0.04 s)	0.96 s (±0.03 s)	1.00
Startup Time - 2 cores	0.96 s (±0.03 s)	0.96 s (±0.03 s)	0.99
Startup Time - 4 cores	0.96 s (±0.03 s)	0.96 s (±0.02 s)	1.00
multithreaded_benchmark Build Time	100.83 s	102.01 s	0.99 ❗
multithreaded_benchmark File Size	0.95 MB	0.95 MB	1.00 ❗
Multithreaded Pi Efficiency - 2 Threads	88.53 % (±7.35 %)	85.65 % (±8.90 %)	1.03
Multithreaded Pi Efficiency - 4 Threads	43.74 % (±3.81 %)	43.12 % (±3.13 %)	1.01
Multithreaded Pi Efficiency - 8 Threads	25.20 % (±2.21 %)	25.03 % (±1.95 %)	1.01
micro_benchmarks Build Time	104.99 s	114.00 s	0.92 ❗
micro_benchmarks File Size	0.96 MB	0.96 MB	1.00 ❗
Scheduling time - 1 thread	69.68 ticks (±3.25 ticks)	70.19 ticks (±3.58 ticks)	0.99
Scheduling time - 2 threads	38.35 ticks (±4.15 ticks)	38.99 ticks (±4.43 ticks)	0.98
Micro - Time for syscall (getpid)	2.89 ticks (±0.26 ticks)	2.96 ticks (±0.24 ticks)	0.98
Memcpy speed - (built_in) block size 4096	62722.07 MByte/s (±45057.38 MByte/s)	64026.03 MByte/s (±45553.92 MByte/s)	0.98
Memcpy speed - (built_in) block size 1048576	29877.34 MByte/s (±24635.09 MByte/s)	29743.12 MByte/s (±24826.10 MByte/s)	1.00
Memcpy speed - (built_in) block size 16777216	28459.33 MByte/s (±23707.11 MByte/s)	24007.89 MByte/s (±20230.77 MByte/s)	1.19
Memset speed - (built_in) block size 4096	63280.76 MByte/s (±45440.59 MByte/s)	64637.10 MByte/s (±45951.78 MByte/s)	0.98
Memset speed - (built_in) block size 1048576	30666.10 MByte/s (±25102.97 MByte/s)	30516.60 MByte/s (±25260.76 MByte/s)	1.00
Memset speed - (built_in) block size 16777216	29240.52 MByte/s (±24158.52 MByte/s)	24779.16 MByte/s (±20739.68 MByte/s)	1.18
Memcpy speed - (rust) block size 4096	61864.35 MByte/s (±45240.77 MByte/s)	60077.78 MByte/s (±44331.51 MByte/s)	1.03
Memcpy speed - (rust) block size 1048576	29667.37 MByte/s (±24499.80 MByte/s)	29870.69 MByte/s (±24878.06 MByte/s)	0.99
Memcpy speed - (rust) block size 16777216	28477.88 MByte/s (±23719.86 MByte/s)	24212.89 MByte/s (±20297.89 MByte/s)	1.18
Memset speed - (rust) block size 4096	62696.50 MByte/s (±45746.48 MByte/s)	60925.02 MByte/s (±44834.41 MByte/s)	1.03
Memset speed - (rust) block size 1048576	30438.95 MByte/s (±24950.58 MByte/s)	30651.78 MByte/s (±25309.43 MByte/s)	0.99
Memset speed - (rust) block size 16777216	29255.32 MByte/s (±24167.74 MByte/s)	24986.00 MByte/s (±20809.17 MByte/s)	1.17
alloc_benchmarks Build Time	103.84 s	105.28 s	0.99 ❗
alloc_benchmarks File Size	0.93 MB	0.93 MB	1.00 ❗
Allocations - Allocation success	100.00 %	100.00 %	1
Allocations - Deallocation success	100.00 %	100.00 %	1
Allocations - Pre-fail Allocations	100.00 %	100.00 %	1
Allocations - Average Allocation time	12088.91 Ticks (±192.04 Ticks)	9727.01 Ticks (±267.20 Ticks)	1.24 ❗
Allocations - Average Allocation time (no fail)	12088.91 Ticks (±192.04 Ticks)	9727.01 Ticks (±267.20 Ticks)	1.24 ❗
Allocations - Average Deallocation time	1087.80 Ticks (±546.23 Ticks)	2219.93 Ticks (±959.95 Ticks)	0.49 ❗
mutex_benchmark Build Time	103.31 s	104.33 s	0.99 ❗
mutex_benchmark File Size	0.96 MB	0.96 MB	1.00 ❗
Mutex Stress Test Average Time per Iteration - 1 Threads	12.82 ns (±0.59 ns)	12.84 ns (±0.76 ns)	1.00
Mutex Stress Test Average Time per Iteration - 2 Threads	15.66 ns (±0.74 ns)	15.44 ns (±0.78 ns)	1.01

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