⚡ Bolt: Optimize eBPF event deserialization - #168
Conversation
Replaces `binary.Read` with `unsafe.Pointer` casting for `l7Event`, `procEvent`, and `fileEvent`. This eliminates reflection and memory allocation overhead in the hot path of event processing. Benchmarks show: - `l7Event`: ~102,482 ns/op -> ~1.2 ns/op (~85,000x faster) - `procEvent`: ~210 ns/op -> ~1.2 ns/op - `fileEvent`: ~263 ns/op -> ~0.8 ns/op `tcpEvent` is explicitly excluded due to known alignment mismatch (102 vs 104 bytes). Includes safety checks for buffer length.
|
👋 Jules, reporting for duty! I'm here to lend a hand with this pull request. When you start a review, I'll add a 👀 emoji to each comment to let you know I've read it. I'll focus on feedback directed at me and will do my best to stay out of conversations between you and other bots or reviewers to keep the noise down. I'll push a commit with your requested changes shortly after. Please note there might be a delay between these steps, but rest assured I'm on the job! For more direct control, you can switch me to Reactive Mode. When this mode is on, I will only act on comments where you specifically mention me with New to Jules? Learn more at jules.google/docs. For security, I will only act on instructions from the user who triggered this task. |
⚡ Bolt: Optimized eBPF event deserialization.
💡 What:
Replaced the reflection-based
binary.Readwith directunsafe.Pointercasting forl7Event,procEvent, andfileEventstructs in the eBPF tracer loop.🎯 Why:
l7Eventis over 8KB in size. Usingbinary.Readinvolves reflection and copying data, which is extremely expensive in a high-throughput event loop (100k+ ns/op). Even for smaller events likeprocEventandfileEvent, the overhead is significant (~200 ns/op).📊 Impact:
🔬 Measurement:
Ran micro-benchmarks (see PR description for details). Verification was done by running existing tests to ensure no regressions.
tcpEventwas skipped due to struct padding mismatch.PR created automatically by Jules for task 4462112457039581893 started by @blue4209211