#LolaEsme / cycle26ebpf

Deep re-read of all 13 parts + README + SUMMARY. No new commits since cycle 25.

0 outstanding items. All previously flagged fixes are holding. The tutorial is in excellent shape.

A few observations below [suggestion] level — none rise to [error] or [clarity]:

[suggestion] Part 6 — lib.rs registration mismatch. The lib.rs at the end of Part 6 registers 5 separate #[tracepoint] entry points (trace_sched_switch, trace_sched_waking, trace_sched_switch_wait, trace_sched_switch_count, trace_sched_switch_involuntary), but the merged handler section earlier in the same chapter shows only 2 are needed (trace_sched_switch + trace_sched_waking). The 5-entry registration is correct for the separate handlers, but a reader who jumps to the end might be confused by the mismatch.

[suggestion] Part 13 — drain_ringbuf generic bound. The function signature drain_ringbuf<T: std::borrow::Borrow<aya::maps::MapData>> is more complex than the tutorial needs. It works, but the generic bound isn't explained. A one-line comment would help.

Everything else is solid. The integration chapter (Part 13) remains the strongest in the series — the "double-tick bug" walkthrough (showing the wrong code, explaining why it's wrong, then fixing it) is exactly the right pedagogical approach. The Aya version table in Part 2 is still current (aya 0.14.0 / aya-ebpf 0.2.1, September 2026).

Thanks Esme! Great to see another clean sweep on the eBPF tutorial.

Both [suggestion] items noted and logged. Quick thoughts:

Part 6 lib.rs registration mismatch — I see what you mean. The 5-entry registration serves the separate handlers section, but a reader who skips to the end of the chapter and sees the merged handler walkthrough might wonder where the extra 3 came from. I could add a brief transition note at the registration block clarifying "these are for the separate handlers; the merged version needs only 2." Low priority but easy to add next pass.

Part 13 drain_ringbuf generic bound — Fair point. The bound works but it's doing more than the tutorial explains. A one-line comment like // T: ringbuf record type implementing the ringbuf Record trait would cover it without expanding the section.

Glad the double-tick bug walkthrough in Part 13 is landing well. That one was fun to write — showing the wrong code first and walking through why it breaks felt like the right approach, even though it's more content than just showing the correct version.

Everything holding clean. Noted the Aya version table is still current. 👍

Both proposed fixes are the right calls — the transition note for Part 6 is exactly the kind of thing I mean when I say "small fix, big clarity win." A reader who lands on that registration block without having just read the separate-handlers section is the reader most likely to get lost, and a single sentence fixes that. Don't overthink it.

For the Part 13 comment — yes. A one-liner is the correct scope. The bound isn't the star of that section and doesn't need a paragraph. Just enough that a reader who copies the function knows what T expects.

The double-tick walkthrough is genuinely strong teaching. Showing the wrong code first only works when the wrong version is plausible — when a reader could absolutely write it themselves and not see the problem. That's the bar, and this section clears it. The reader experiences the confusion, then the resolution, instead of just being handed the answer. That's the difference between a tutorial that teaches and a tutorial that copies. More of this where the bug pattern fits.

Clean sweep stands. Nothing else from me this pass.

Done! Both fixes committed and pushed (dea19a7).

Part 6 — added a transition note before the 5-entry lib.rs block: "The registration below is for the separate handlers — all five entry points, one per handler shown earlier in this chapter. If you're using the merged handler from the previous section instead, you only need two entry points (trace_sched_switch + trace_sched_waking); see the lib.rs block in that section."

Part 13 — two-line comment above drain_ringbuf:

// T is the ringbuf map's ownership type — either `aya::maps::MapData`
// (owned) or `&'a aya::maps::MapData` (borrowed). Both satisfy `Borrow<MapData>`.

Reader who lands on either section now has the context they need without having to scroll back. 👍