@mark1russell7/harness 0.11.9 → 0.11.10

package/README.md

@@ -226,7 +226,7 @@ Each run starts from kernels that keep nothing from the run before it (`lab.forg
 `measure(cell, run)` (`observe.ts`) reads a run from the records of the kernel:
 
 - **The socket.** The socket of the run is the socket of its observer. For the client, it is the socket that the client connected from, in its namespace. In a topology, the address of a hop serves each run. Thus the socket of the target or of a proxy is the socket that became ESTABLISHED in the run (`socketOf`).
-- **The time.** It times the detection from the reference event of the prediction to the error report of the kernel. eBPF stamps the two events. For keepalive, the reference is the last received segment. For the give-up, it is the first retransmission of the RTO timer, not the tail loss probe one RTO before it. For a reset, it is the send of the client, which no kernel record marks. The client stamps the send before its write.
+- **The time.** It times the detection from the reference event of the prediction to the error report of the kernel. eBPF stamps the two events. For keepalive, the reference is the last received segment. For the give-up, it is the first retransmission of the RTO timer after the fault holds, not the tail loss probe one RTO before it. For a reset, it is the send of the client, which no kernel record marks. The client stamps the send before its write.
 - **The other views.** It also gives the same interval from the `TCP_INFO` samples of the client, and the delay before the client saw the error. It also gives the steps of the kernel: each keepalive probe, or each expiry of the RTO timer.
 - **Void runs.** A run is void when the timer that drives the measurement was two jiffies or more behind its wheel slot in the timed interval. The wheel does a slot in the tick when it is due. Thus such a firing measured the host, which stopped a vCPU, not the kernel. A keepalive timer that found its socket held also makes the run void. A run whose fault came too late is void too.
   - The rule is in time when the run knows its HZ (`Run.hz`, from the manifest). Then the delay must reach two jiffies and δ_abs, the larger of the two (`delayVoids`). Each band includes δ_abs, thus a smaller delay does not change a result. At HZ 250 two jiffies are 8 ms, more than a δ_abs of 4 ms, and the rule stays the same. At HZ 1000 two jiffies are 2 ms. A run without its HZ keeps the rule in jiffies.

package/build/index.js

@@ -1677,7 +1677,7 @@ const measureSocket = (cell, run) => {
 		if (from === "fault") return run.fault?.t0;
 		if (from === "client-start") return (events.find((event) => event.ev === "connecting") ?? events[0])?.t;
 		if (from === "ping-sent") return firstUnanswered(pings, socket, ended)?.at;
-		return (from === "last-received" ? socket.filter((each) => each.ev === "rx" && each.mono <= before).at(-1) : from === "first-retransmission" ? retransmitted.find((each) => each.why === "rto") : from === "first-sent" ? socket.find((each) => each.ev === "tx" && each.mono > before && (each.seg?.len.value ?? 0) > 0) : kernel.find((each) => each.netns === peerNetns && each.cookie === peerCookie && each.ev === "state" && oldState(each) === 1))?.mono;
+		return (from === "last-received" ? socket.filter((each) => each.ev === "rx" && each.mono <= before).at(-1) : from === "first-retransmission" ? retransmitted.find((each) => each.why === "rto" && each.mono > before) : from === "first-sent" ? socket.find((each) => each.ev === "tx" && each.mono > before && (each.seg?.len.value ?? 0) > 0) : kernel.find((each) => each.netns === peerNetns && each.cookie === peerCookie && each.ev === "state" && oldState(each) === 1))?.mono;
 	};
 	const lastRx = kernel.filter((each) => each.netns === run.netns && each.src === local && each.ev === "rx" && each.mono <= before).at(-1)?.mono ?? ready.t;
 	const chain = run.topology === void 0 ? {} : {