Upstream patched binary package from cilium

[lmb]

Currently, marshaling a 24 byte via binary allocates quite a lot:

name                      time/op
Marshalling/reflection-8  1.27µs ± 8%
Marshalling/custom-8       730ns ± 6%
Marshalling/unsafe-8       662ns ± 5%

name                      alloc/op
Marshalling/reflection-8    152B ± 0%
Marshalling/custom-8       32.0B ± 0%
Marshalling/unsafe-8       0.00B

name                      allocs/op
Marshalling/reflection-8    8.00 ± 0%
Marshalling/custom-8        1.00 ± 0%
Marshalling/unsafe-8        0.00

Cilium proper has a patched binary package which reduces allocations: https://github.com/cilium/cilium/tree/master/pkg/bpf/binary

[lmb]

I managed to improve the performance of binary.Read and Write for structs. Improvement of go tip over 1.13:

name                    old time/op    new time/op    delta
Marshalling/reflection    1.24µs ± 1%    1.00µs ± 1%  -19.38%  (p=0.008 n=5+5)

name                    old alloc/op   new alloc/op   delta
Marshalling/reflection      152B ± 0%      112B ± 0%  -26.32%  (p=0.008 n=5+5)

name                    old allocs/op  new allocs/op  delta
Marshalling/reflection      8.00 ± 0%      3.00 ± 0%  -62.50%  (p=0.008 n=5+5)

See golang/go#34471

[tklauser]

Is there anything else to do here? It seems @lmb's performance improvement made it upstream (golang/go@c9d89f6) and will be part of Go 1.14.

[lmb]

I'm happy to close this as a wontfix. If cilium needs more perf you can implement BinaryMarshaler, or use unsafe.Pointer for every last drop.