refactor(fs/mem): move RomFileInner's RwLock into it

[fogti]

No description provided.

[github-actions]

Benchmark Results

Benchmark	Current: 905b093	Previous: 1caea89	Performance Ratio
startup_benchmark Build Time	113.75 s	113.59 s	1.00
startup_benchmark File Size	0.87 MB	0.88 MB	1.00
Startup Time - 1 core	0.98 s (±0.03 s)	1.00 s (±0.02 s)	0.98
Startup Time - 2 cores	0.97 s (±0.03 s)	1.00 s (±0.03 s)	0.97
Startup Time - 4 cores	1.00 s (±0.03 s)	1.00 s (±0.03 s)	1.00
multithreaded_benchmark Build Time	116.15 s	112.31 s	1.03
multithreaded_benchmark File Size	0.97 MB	0.97 MB	1.00
Multithreaded Pi Efficiency - 2 Threads	87.66 % (±9.85 %)	91.87 % (±5.48 %)	0.95
Multithreaded Pi Efficiency - 4 Threads	42.95 % (±3.62 %)	45.62 % (±1.52 %)	0.94
Multithreaded Pi Efficiency - 8 Threads	24.62 % (±1.70 %)	26.92 % (±1.71 %)	0.91
micro_benchmarks Build Time	302.72 s	300.90 s	1.01
micro_benchmarks File Size	0.98 MB	0.98 MB	1.00
Scheduling time - 1 thread	178.07 ticks (±22.40 ticks)	188.89 ticks (±35.95 ticks)	0.94
Scheduling time - 2 threads	106.89 ticks (±20.93 ticks)	109.29 ticks (±17.74 ticks)	0.98
Micro - Time for syscall (getpid)	12.25 ticks (±7.22 ticks)	11.44 ticks (±6.11 ticks)	1.07
Memcpy speed - (built_in) block size 4096	56693.40 MByte/s (±42008.30 MByte/s)	58218.35 MByte/s (±41749.13 MByte/s)	0.97
Memcpy speed - (built_in) block size 1048576	13548.59 MByte/s (±11179.24 MByte/s)	13322.31 MByte/s (±11078.57 MByte/s)	1.02
Memcpy speed - (built_in) block size 16777216	8693.69 MByte/s (±7053.20 MByte/s)	9632.94 MByte/s (±7763.00 MByte/s)	0.90
Memset speed - (built_in) block size 4096	56846.04 MByte/s (±42112.77 MByte/s)	58627.12 MByte/s (±42001.83 MByte/s)	0.97
Memset speed - (built_in) block size 1048576	13883.76 MByte/s (±11369.29 MByte/s)	13664.11 MByte/s (±11297.52 MByte/s)	1.02
Memset speed - (built_in) block size 16777216	8870.74 MByte/s (±7151.43 MByte/s)	9828.35 MByte/s (±7867.40 MByte/s)	0.90
Memcpy speed - (rust) block size 4096	49003.02 MByte/s (±37540.62 MByte/s)	51942.01 MByte/s (±38359.80 MByte/s)	0.94
Memcpy speed - (rust) block size 1048576	13090.14 MByte/s (±10692.77 MByte/s)	12839.06 MByte/s (±10483.71 MByte/s)	1.02
Memcpy speed - (rust) block size 16777216	8937.15 MByte/s (±7275.69 MByte/s)	9531.71 MByte/s (±7678.70 MByte/s)	0.94
Memset speed - (rust) block size 4096	49957.20 MByte/s (±38225.46 MByte/s)	53011.14 MByte/s (±39115.85 MByte/s)	0.94
Memset speed - (rust) block size 1048576	13450.00 MByte/s (±10910.67 MByte/s)	13244.14 MByte/s (±10750.32 MByte/s)	1.02
Memset speed - (rust) block size 16777216	9152.74 MByte/s (±7404.77 MByte/s)	9722.67 MByte/s (±7778.13 MByte/s)	0.94
alloc_benchmarks Build Time	317.84 s	302.31 s	1.05
alloc_benchmarks File Size	0.94 MB	0.95 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	17995.50 Ticks (±879.50 Ticks)	26215.84 Ticks (±1230.11 Ticks)	0.69
Allocations - Average Allocation time (no fail)	17995.50 Ticks (±879.50 Ticks)	26215.84 Ticks (±1230.11 Ticks)	0.69
Allocations - Average Deallocation time	2805.12 Ticks (±823.86 Ticks)	3319.78 Ticks (±1756.11 Ticks)	0.84
mutex_benchmark Build Time	296.39 s	301.33 s	0.98
mutex_benchmark File Size	0.97 MB	0.98 MB	1.00
Mutex Stress Test Average Time per Iteration - 1 Threads	38.16 ns (±3.32 ns)	38.48 ns (±4.89 ns)	0.99
Mutex Stress Test Average Time per Iteration - 2 Threads	30.92 ns (±2.74 ns)	31.68 ns (±3.36 ns)	0.98

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

[mkroening]

Sorry for the wait! :)

I will move the lock to the outside soonish (hopefully), but for now this is fine. Thanks! 👍

[fogti]

Given that there can be multiple references to the same ram/romfile, one can't easily get rid of the inner locks (unless one wants to take a lock on all in-memory files). But one could get rid of the position Mutex<usize>.

[mkroening]

Given that there can be multiple references to the same ram/romfile, one can't easily get rid of the inner locks (unless one wants to take a lock on all in-memory files). But one could get rid of the position Mutex<usize>.

True. Even better, then.