Changes to Functions to include inline and moved benches to criterion to use stable toolchain

[infosechoudini]

pent_macro_support/src/packet.rs

• added inlining to all functions available to increase speed at cost of size

pnet_macro/src/decorator.rs

• added inlining to all functions to increase speed at cost of size

pnet_packet/

• moved benches from ethernet.rs to new folder ~/benches/packet_benchmarks.rs

Benchmarks Before Changes

test bench_ipv4_parsing                  ... bench:           8 ns/iter (+/- 0)
test bench_packet_get_source             ... bench:          16 ns/iter (+/- 0)
test bench_packet_immutable_to_immutable ... bench:          10 ns/iter (+/- 1)
test bench_packet_mutable_to_immutable   ... bench:          10 ns/iter (+/- 0)
test bench_packet_new_constructor        ... bench:           8 ns/iter (+/- 0)
test bench_packet_set_source_black_box   ... bench:          18 ns/iter (+/- 1)
Benchmarks After Changes

running 6 tests
test bench_ipv4_parsing                  ... bench:           1 ns/iter (+/- 0)
test bench_packet_get_source             ... bench:           7 ns/iter (+/- 0)
test bench_packet_immutable_to_immutable ... bench:           2 ns/iter (+/- 0)
test bench_packet_mutable_to_immutable   ... bench:           2 ns/iter (+/- 0)
test bench_packet_new_constructor        ... bench:           1 ns/iter (+/- 0)
test bench_packet_set_source_black_box   ... bench:           9 ns/iter (+/- 0)

[infosechoudini]

Criterion Output

Gnuplot not found, using plotters backend
EthernetPacket New Packet                                                                             
                        time:   [1.9560 ns 1.9666 ns 1.9787 ns]
                        change: [-0.5771% +0.3666% +1.2536%] (p = 0.43 > 0.05)
                        No change in performance detected.
Found 7 outliers among 100 measurements (7.00%)
  5 (5.00%) high mild
  2 (2.00%) high severe

EthernetPacket Get Source                                                                             
                        time:   [4.5342 ns 4.5548 ns 4.5782 ns]
                        change: [+0.1741% +1.5730% +2.5907%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 13 outliers among 100 measurements (13.00%)
  8 (8.00%) high mild
  5 (5.00%) high severe

EthernetPacket Set Source                                                                             
                        time:   [637.33 ps 638.82 ps 640.36 ps]
                        change: [-3.0706% -2.4471% -1.8802%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 4 outliers among 100 measurements (4.00%)
  4 (4.00%) high mild

Mutable to Immutable    time:   [2.5357 ns 2.5486 ns 2.5620 ns]                                  
                        change: [-1.5676% -0.7663% -0.0239%] (p = 0.06 > 0.05)
                        No change in performance detected.
Found 7 outliers among 100 measurements (7.00%)
  6 (6.00%) high mild
  1 (1.00%) high severe

Immutable to Immutable  time:   [2.5807 ns 2.5930 ns 2.6076 ns]                                    
                        change: [-1.4937% -0.6850% +0.0496%] (p = 0.09 > 0.05)
                        No change in performance detected.
Found 9 outliers among 100 measurements (9.00%)
  5 (5.00%) high mild
  4 (4.00%) high severe

IPV4 Parsing            time:   [1.9412 ns 1.9472 ns 1.9535 ns]                          
                        change: [+0.0893% +0.6524% +1.2177%] (p = 0.03 < 0.05)
                        Change within noise threshold.
Found 8 outliers among 100 measurements (8.00%)
  4 (4.00%) high mild
  4 (4.00%) high severe

[infosechoudini]

@mrmonday et al., can i get a workflow approval?

[stappersg]

@mrmonday et al., can i get a workflow approval?

The request is be seen by a human.
My advice is to ask again in about week.
Consider meanwhile the idea of something like easter vacaction or any other possible reason for a delay in response.

My real message is

@infosechoudini aim for the long run, avoid to get worn down on the short run

Other thing

Day changed to 20 apr 2022
00:22 -!- infosechoudini [~infosecho@h597.436.156.298.static.ip.redacted.net] has joined #libpnet
00:24 < infosechoudini> hey, i'm looking to help contribute to libpnet, was wondering what's the process
02:21 -!- infosechoudini [~infosecho@h597.436.156.298.static.ip.redacted.net] has quit [Quit: Connection closed]

Feel welcome to hang-out at libera/#libpnet IRC channel.

[infosechoudini]

@stappersg thanks, i apologize for the ping. I forgot it was easter week!!!

[mrmonday]

I'm surprised the annotations make so much difference - in the past we've removed lots of them because the compiler is "sufficiently smart" that it made no difference.

It would be good to get the benchmarks in CI, no need to do that here though.

Thanks!