KarpelesLab / pktkit
53%

DEFAULT BRANCH: master
Repo Added
Token 1PhEffmLefUHSVS0XJvlKqgLP8E9ifvvm regen
Build 27 Last
Files 84
Badge
Embed ▾
README BADGES
x

If you need to use a raster PNG badge, change the '.svg' to '.png' in the link

Markdown

Textile

RDoc

HTML

Rst

LAST BUILD ON BRANCH master
branch: master
CHANGE BRANCH
x
Sync Branches

coverage: 53.337% (-0.09%) from 53.426%
Relevant lines Covered
13966 RELEVANT LINES 7449 COVERED LINES
26.17 HITS PER LINE
Source Files on master

Recent builds

Builds Branch Commit Type Ran Committer Via Coverage
24013095013 master Add multi-key Adapter support and parallel packet processing Adapter now accepts a MultiHandler for serving multiple WireGuard identities on a single UDP port with pktkit networking integration. New methods: AddPeerTo, RemovePeerFrom, ConnectWith... push MagicalTux github
53.34
24007449401 master Fix CI: use bash shell on Windows to avoid PowerShell arg parsing PowerShell splits -coverprofile=coverage.out treating .out as a separate package argument. Using shell: bash fixes the issue. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply... push MagicalTux github
53.43
24007150346 master Add Windows and macOS CI, fix cross-platform build - Add macos-latest and windows-latest to test matrix - Move qemu.Socketpair to unix-only file (socketpair_unix.go) - Add build tags to qemu and qemu_e2e tests that use Socketpair - All packages b... push MagicalTux github
53.36
24006830663 master Add AF_XDP driver package for kernel-bypass L2 frame I/O New afxdp/ package implementing pktkit.L2Device over Linux AF_XDP sockets using only golang.org/x/sys/unix — no external libraries. - UMEM shared memory with split RX/TX frame pools - Lock... push MagicalTux github
53.41
24005892453 master Improve test coverage for nat, ovpn, and vclient packages nat: 14.4% → 36.9% — defrag reassembly, NAT64 IPv6↔IPv4 translation, ALG expectations, port forwards, ConnectL3 namespaces. ovpn: 10.8% → 37.6% — GCM/CBC encrypt/decrypt roundtrip, replay... push MagicalTux github
55.77
24004585469 master Rewrite e2e VPN benchmarks to measure pipelined throughput Previous benchmarks measured per-packet latency (send-wait-receive), reporting 50 MB/s which was dominated by UDP loopback syscall overhead. New benchmarks saturate the pipeline: sender ... push MagicalTux github
44.34
24004525574 master Add end-to-end VPN throughput benchmarks over real UDP loopback Both WireGuard and OpenVPN benchmarks use real UDP sockets on 127.0.0.1 for a fair apples-to-apples comparison of the full encrypt → sendto → recvfrom → decrypt pipeline at 1400B pay... push MagicalTux github
44.29
24004240187 master Fix OpenVPN GCM decrypt benchmark and guard nil replay window The decrypt benchmark was inflated by per-iteration make+copy overhead (~50% of measured time). Fixed to reuse a pre-copied buffer with replay window disabled, revealing the true decry... push MagicalTux github
44.33
24004052060 master Update README benchmarks with current performance numbers Add separate benchmark tables for NAT, WireGuard, and OpenVPN GCM reflecting the recent optimization work (zero-alloc NAT at 36 GB/s, WireGuard EncryptTo at 2.6 GB/s, OpenVPN GCM encrypt a... push MagicalTux github
44.33
24003982542 master Optimize OpenVPN GCM data channel: 18x encrypt, 2x decrypt Rewrite GCM encrypt/decrypt hot paths to eliminate per-packet allocations: Encrypt (sendDataGCM): pre-allocate single output buffer from global pool, build nonce/AD on stack, Seal direct... push MagicalTux github
44.34
See All Builds (27)

Badge your Repo: pktkit

We detected this repo isn’t badged! Grab the embed code to the right, add it to your repo to show off your code coverage, and when the badge is live hit the refresh button to remove this message.
Embed ▾
README BADGES
x

If you need to use a raster PNG badge, change the '.svg' to '.png' in the link

Markdown

Textile

RDoc

HTML

Rst

Refresh