docs: document environment variables

[mkroening]

Depends on a decision on #2415.

[github-actions]

Benchmark Results

Details

Benchmark	Current: 2bde53c	Previous: 07d3301	Performance Ratio
startup_benchmark Build Time	89.59 s	89.18 s	1.00 ❗
startup_benchmark File Size	0.78 MB	0.78 MB	1.00 ❗
Startup Time - 1 core	0.81 s (±0.03 s)	0.80 s (±0.04 s)	1.02
Startup Time - 2 cores	0.81 s (±0.03 s)	0.81 s (±0.03 s)	1.00
Startup Time - 4 cores	0.82 s (±0.03 s)	0.83 s (±0.03 s)	0.98
multithreaded_benchmark Build Time	90.49 s	92.43 s	0.98 ❗
multithreaded_benchmark File Size	0.89 MB	0.89 MB	1.00 ❗
Multithreaded Pi Efficiency - 2 Threads	89.92 % (±6.97 %)	90.75 % (±6.76 %)	0.99
Multithreaded Pi Efficiency - 4 Threads	45.02 % (±2.69 %)	44.75 % (±3.33 %)	1.01
Multithreaded Pi Efficiency - 8 Threads	26.03 % (±1.43 %)	25.74 % (±1.81 %)	1.01
micro_benchmarks Build Time	98.59 s	99.63 s	0.99 ❗
micro_benchmarks File Size	0.89 MB	0.90 MB	0.99 ❗
Scheduling time - 1 thread	71.52 ticks (±4.23 ticks)	71.75 ticks (±4.55 ticks)	1.00
Scheduling time - 2 threads	38.66 ticks (±4.00 ticks)	39.08 ticks (±4.44 ticks)	0.99
Micro - Time for syscall (getpid)	3.05 ticks (±0.26 ticks)	3.13 ticks (±0.27 ticks)	0.97
Memcpy speed - (built_in) block size 4096	75574.66 MByte/s (±52315.99 MByte/s)	75616.90 MByte/s (±52577.63 MByte/s)	1.00
Memcpy speed - (built_in) block size 1048576	30259.88 MByte/s (±24993.79 MByte/s)	29990.09 MByte/s (±24687.99 MByte/s)	1.01
Memcpy speed - (built_in) block size 16777216	22880.37 MByte/s (±19056.83 MByte/s)	24067.71 MByte/s (±19970.69 MByte/s)	0.95
Memset speed - (built_in) block size 4096	75904.96 MByte/s (±52526.33 MByte/s)	75500.18 MByte/s (±52500.81 MByte/s)	1.01
Memset speed - (built_in) block size 1048576	31024.08 MByte/s (±25423.94 MByte/s)	30765.36 MByte/s (±25121.72 MByte/s)	1.01
Memset speed - (built_in) block size 16777216	23561.65 MByte/s (±19510.11 MByte/s)	24807.43 MByte/s (±20464.41 MByte/s)	0.95
Memcpy speed - (rust) block size 4096	67423.74 MByte/s (±47102.30 MByte/s)	69070.89 MByte/s (±48290.91 MByte/s)	0.98
Memcpy speed - (rust) block size 1048576	30212.31 MByte/s (±25044.88 MByte/s)	29768.99 MByte/s (±24513.26 MByte/s)	1.01
Memcpy speed - (rust) block size 16777216	22370.41 MByte/s (±18578.12 MByte/s)	24637.88 MByte/s (±20514.70 MByte/s)	0.91
Memset speed - (rust) block size 4096	67693.05 MByte/s (±47273.95 MByte/s)	69563.89 MByte/s (±48684.23 MByte/s)	0.97
Memset speed - (rust) block size 1048576	30967.65 MByte/s (±25445.30 MByte/s)	30549.28 MByte/s (±24960.18 MByte/s)	1.01
Memset speed - (rust) block size 16777216	22987.39 MByte/s (±18977.50 MByte/s)	25362.78 MByte/s (±20979.00 MByte/s)	0.91
alloc_benchmarks Build Time	96.39 s	91.86 s	1.05 ❗
alloc_benchmarks File Size	0.86 MB	0.85 MB	1.01 ❗
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	11977.75 Ticks (±171.15 Ticks)	8634.49 Ticks (±216.63 Ticks)	1.39 ❗
Allocations - Average Allocation time (no fail)	11977.75 Ticks (±171.15 Ticks)	8634.49 Ticks (±216.63 Ticks)	1.39 ❗
Allocations - Average Deallocation time	986.67 Ticks (±213.05 Ticks)	1135.91 Ticks (±535.66 Ticks)	0.87
mutex_benchmark Build Time	93.14 s	92.02 s	1.01 ❗
mutex_benchmark File Size	0.89 MB	0.89 MB	1.00 ❗
Mutex Stress Test Average Time per Iteration - 1 Threads	13.38 ns (±0.77 ns)	13.24 ns (±0.84 ns)	1.01
Mutex Stress Test Average Time per Iteration - 2 Threads	14.00 ns (±0.77 ns)	18.04 ns (±3.62 ns)	0.78 ❗

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

[Gelbpunkt]

We should document that this definition of MTU includes the 14-byte ethernet header, contrary to e.g. Linux where the configured MTU is the IP MTU and excludes the ethernet header

See also the comment here:

kernel/src/drivers/net/mod.rs

Line 50 in a9e28f9

/// This is 1500 IP MTU and a 14-byte ethernet header.

[Gelbpunkt]

Alternatively it would probably be more intuitive to match Linux and make the user configure the IP MTU and default to 1500.

[cagatay-y]

Should this PR also change the README? HERMIT_LOG_LEVEL_FILTER is documented there too. Is it redundant or should they both stay? If both stay, should this one refer to the README for more information?

[cagatay-y]

Is there a preferred alternative for the deprecated variable? Might be useful to mention if there is one.

[cagatay-y]

We should probably clarify that the environment variable is only relevant for virtio-net devices and with a specific feature negotiated.

[cagatay-y]

Since the DHCPv4 feature is enabled by default, it may be helpful to mention that the IP address set may be overwritten if addresses are successfully negotiated.