nx_fabric_cc_gate.nx
buildroot/runtime/nx_fabric_cc_gate.nx
about
nx_fabric_cc_gate.nx -- LIVE GATE for R2 (congestion control + telemetry over RTT).
Proves: a DCTCP-style ECN-telemetry controller holds the bottleneck queue near a low
setpoint K at FULL utilisation over an RTT link, where an open-loop window bufferbloats
-- with negative controls so the result cannot be rigged.
criteria:
1 R2 utilisation >= 90% (keeps the RTT link full)
2 R2 q_p99 <= 3*K (queue BOUNDED near the ECN setpoint)
3 baseline util >= 90% AND baseline q_p99 >= 3*R2 q_p99 (open-loop BLOATS at equal util)
4 R2 telemetry error == 0 (reported queue tracks the true queue)
negative controls:
NC1 ignore ECN marks -> queue bloats (>= 3*R2) (the REACTION holds the queue)
NC2 never mark (K huge) -> queue bloats (>= 3*R2) (the TELEMETRY SIGNAL is essential)
NC3 telemetry recount matches on every run (no fabricated telemetry)
expect_exit: 0 license_tier: ORIGINAL
dependencies 1 imports · 0 importers
imports: nx_fabric_cc.nx
imported by: nobody (leaf or entry point)
call flow from main pre-order; caps 40 nodes / depth 6 declared; ↻ = already shown
structs
| none |
consts
| none |
functions
| 19 | func gp(s: *u8) -> i64 { var n: i64 = 0; while s[n] != (0 as u8) { n = n + 1 } sys_write(1, s, n); return 0 } |
| 20 | func gn(v: i64) -> i64 |
| 30 | func fdn(fd: i64, v: i64) -> i64 called by 1: main |
| 39 | func chk(name: *u8, ok: i64) -> i64 |
| 43 | func show(tag: *u8, m: *CcMetrics) -> i64 |
| 49 | func main() -> i64 |