feat: add the icmp-echo-reply feature

[stlankes]

smoltcp answers ICMP echo requests by itself once auto-icmp-echo-reply is set, for IPv4 as well as IPv6, so making a Hermit guest respond to ping needs nothing beyond the feature flag.

I tested the configuration on macOS as follows:

$ cargo build --release -Zbuild-std=std,panic_abort --target=aarch64-unknown-hermit -p httpd --features hermit/echo-reply,hermit/pci,hermit/virtio-net,hermit/dhcpv4
$ sudo qemu-system-aarch64 -display none -serial stdio \
  -machine virt,gic-version=3 \
  -cpu max,lpa2=off  \
  -smp 1 -m 1024M \
  -global virtio-mmio.force-legacy=off \
  -kernel hermit-loader-aarch64 \
  -device guest-loader,addr=0x48000000,initrd=target/aarch64-unknown-hermit/release/httpd \
  -netdev vmnet-shared,id=net0,start-address=192.168.68.2,end-address=192.168.68.254,subnet-mask=255.255.255.0 \
  -device virtio-net-pci,netdev=net0,mac=52:54:00:12:34:56,disable-legacy=on,packed=on,mq=on
$ ping -c 5 192.168.68.3
PING 192.168.68.3 (192.168.68.3): 56 data bytes
64 bytes from 192.168.68.3: icmp_seq=0 ttl=64 time=2.790 ms
64 bytes from 192.168.68.3: icmp_seq=1 ttl=64 time=0.485 ms
64 bytes from 192.168.68.3: icmp_seq=2 ttl=64 time=0.724 ms
64 bytes from 192.168.68.3: icmp_seq=3 ttl=64 time=0.621 ms
64 bytes from 192.168.68.3: icmp_seq=4 ttl=64 time=0.855 ms

close #2608

[github-actions]

Benchmark Results

Details

Benchmark	Current: fc53e38	Previous: 2e23902	Performance Ratio
startup_benchmark Build Time	91.37 s	80.34 s	1.14 ❗
startup_benchmark File Size	0.79 MB	0.80 MB	0.99 ❗
Startup Time - 1 core	0.70 s (±0.01 s)	0.75 s (±0.02 s)	0.94 ❗
Startup Time - 2 cores	0.71 s (±0.01 s)	0.74 s (±0.02 s)	0.97 ❗
Startup Time - 4 cores	0.71 s (±0.01 s)	0.74 s (±0.02 s)	0.96 ❗
multithreaded_benchmark Build Time	89.74 s	82.11 s	1.09 ❗
multithreaded_benchmark File Size	0.84 MB	0.86 MB	0.98 ❗
Multithreaded Pi Efficiency - 2 Threads	66.79 % (±4.99 %)	85.89 % (±6.61 %)	0.78 ❗
Multithreaded Pi Efficiency - 4 Threads	41.22 % (±3.29 %)	43.43 % (±2.56 %)	0.95
Multithreaded Pi Efficiency - 8 Threads	20.52 % (±1.75 %)	25.76 % (±1.53 %)	0.80 ❗
micro_benchmarks Build Time	223.08 s	80.40 s	2.77 ❗
micro_benchmarks File Size	0.84 MB	0.86 MB	0.98 ❗
Scheduling time - 1 thread	148.06 ticks (±29.30 ticks)	62.65 ticks (±4.06 ticks)	2.36 ❗
Scheduling time - 2 threads	89.19 ticks (±22.96 ticks)	34.08 ticks (±4.10 ticks)	2.62 ❗
Micro - Time for syscall (getpid)	9.22 ticks (±5.59 ticks)	3.45 ticks (±0.58 ticks)	2.67 ❗
Memcpy speed - (built_in) block size 4096	60702.61 MByte/s (±43101.14 MByte/s)	82448.38 MByte/s (±56997.13 MByte/s)	0.74
Memcpy speed - (built_in) block size 1048576	16441.42 MByte/s (±14149.25 MByte/s)	30585.98 MByte/s (±24707.84 MByte/s)	0.54
Memcpy speed - (built_in) block size 16777216	12999.38 MByte/s (±10991.43 MByte/s)	26340.06 MByte/s (±21720.96 MByte/s)	0.49
Memset speed - (built_in) block size 4096	61284.49 MByte/s (±43494.20 MByte/s)	82292.76 MByte/s (±56891.50 MByte/s)	0.74
Memset speed - (built_in) block size 1048576	16891.17 MByte/s (±14393.80 MByte/s)	31323.85 MByte/s (±25145.86 MByte/s)	0.54
Memset speed - (built_in) block size 16777216	13364.21 MByte/s (±11211.11 MByte/s)	27104.68 MByte/s (±22209.94 MByte/s)	0.49
Memcpy speed - (rust) block size 4096	52876.99 MByte/s (±38563.21 MByte/s)	74097.96 MByte/s (±51811.44 MByte/s)	0.71
Memcpy speed - (rust) block size 1048576	15549.03 MByte/s (±12478.44 MByte/s)	30361.60 MByte/s (±24602.37 MByte/s)	0.51
Memcpy speed - (rust) block size 16777216	12892.17 MByte/s (±10642.24 MByte/s)	27625.34 MByte/s (±22806.88 MByte/s)	0.47
Memset speed - (rust) block size 4096	53536.33 MByte/s (±38994.13 MByte/s)	74373.47 MByte/s (±51976.48 MByte/s)	0.72
Memset speed - (rust) block size 1048576	15638.53 MByte/s (±12522.93 MByte/s)	31110.89 MByte/s (±25033.24 MByte/s)	0.50
Memset speed - (rust) block size 16777216	13134.09 MByte/s (±10754.64 MByte/s)	28386.93 MByte/s (±23265.03 MByte/s)	0.46
alloc_benchmarks Build Time	211.34 s	74.76 s	2.83 ❗
alloc_benchmarks File Size	0.86 MB	0.87 MB	0.99 ❗
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	26624.67 Ticks (±1668.45 Ticks)	5860.58 Ticks (±98.43 Ticks)	4.54 ❗
Allocations - Average Allocation time (no fail)	27505.70 Ticks (±1911.75 Ticks)	6554.81 Ticks (±92.86 Ticks)	4.20 ❗
Allocations - Average Deallocation time	6081.42 Ticks (±1429.98 Ticks)	1805.01 Ticks (±250.35 Ticks)	3.37 ❗
mutex_benchmark Build Time	207.73 s	79.82 s	2.60 ❗
mutex_benchmark File Size	0.84 MB	0.86 MB	0.98 ❗
Mutex Stress Test Average Time per Iteration - 1 Threads	34.22 ns (±7.22 ns)	12.10 ns (±0.41 ns)	2.83 ❗
Mutex Stress Test Average Time per Iteration - 2 Threads	31.16 ns (±9.38 ns)	40.26 ns (±1.68 ns)	0.77

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

[mkroening]

Great! :)

[mkroening]

Actually, since this requires no kernel changes, it might make sense to only add this in the wrapper crate, then.

Users that directly build the kernel can already enable smoltcp/auto-icmp-echo-reply. hermit-os/hermit-rs#1060 adds a forward as smoltcp-auto-icmp-echo-reply similar to what we do for compile-time log levels.

[stlankes]

Actually, since this requires no kernel changes, it might make sense to only add this in the wrapper crate, then.

Users that directly build the kernel can already enable smoltcp/auto-icmp-echo-reply. hermit-os/hermit-rs#1060 adds a forward as smoltcp-auto-icmp-echo-reply similar to what we do for compile-time log levels.

It is for me not clear, why this is the better way. But I am fine with it.