1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
#[cfg(any(
all(target_os = "linux", target_arch = "x86_64"),
all(target_os = "macos", target_arch = "aarch64")
))]
mod supported {
//! 100/100 reliability gate for multi-vCPU snapshot/restore.
//!
//! THIS IS THE SHIP-GATE for the smpark.ko approach. Bakes a
//! 4-vCPU snapshot with smpark.ko staged + auto-loaded by init-
//! oci, then drives 100 acquire+exec cycles. Expects zero
//! failures. Prior userspace-only attempt (rendezvous + LR drain
//! at HVF level) topped out at 8/10 and was rejected.
//!
//! Why 4 vCPUs: the failure rate scales with secondary count.
//! 1 vCPU: 100/100 trivially. 2 vCPUs: ~95/100 baseline (LR
//! drain is mostly correct). 4 vCPUs: ~80/100 baseline — this
//! is where the failure class becomes visible. 8/16 vCPUs go
//! lower still but few real workloads use that many.
//!
//! Run:
//! cargo run --release --example _park_reliability_gate
//!
//! Optional env vars:
//! PARK_GATE_CYCLES=100 — number of acquire+exec cycles
//! PARK_GATE_VCPUS=4 — vCPUs for the bake
//! PARK_GATE_KO=docs/design/extras/smpark/smpark.ko
//! — path to the kernel module
//!
//! Exit code:
//! 0 — all cycles succeeded (ship gate passed)
//! 1 — at least one cycle failed (ship gate failed)
//! 2 — bake failed (gate inconclusive)
use std::path::PathBuf;
use std::time::{Duration, Instant};
use supermachine::Image;
pub(super) fn main() -> Result<(), Box<dyn std::error::Error>> {
let cycles: usize = std::env::var("PARK_GATE_CYCLES")
.ok()
.and_then(|v| v.parse().ok())
.unwrap_or(100);
let vcpus: u32 = std::env::var("PARK_GATE_VCPUS")
.ok()
.and_then(|v| v.parse().ok())
.unwrap_or(4);
let park_ko = PathBuf::from(
std::env::var("PARK_GATE_KO")
.unwrap_or_else(|_| "docs/design/extras/smpark/smpark.ko".to_string()),
);
if !park_ko.is_file() {
eprintln!(
"[gate] smpark.ko not found at {} — set PARK_GATE_KO or run from repo root",
park_ko.display()
);
std::process::exit(2);
}
eprintln!(
"[gate] baking rust:1-slim with {} vCPUs + smpark.ko staged (pipelined path)...",
vcpus
);
let bake_t0 = Instant::now();
// No-op warmup callback forces the pipelined-bake path,
// which drives `smpark_park` / `smpark_unpark` from the host
// around `SNAPSHOT_ASYNC` and the warm `SNAPSHOT` calls. The
// sequential bake (no warmup) uses the worker's internal
// heartbeat trigger which can't issue host-driven RPCs from
// within take_snapshot — vCPU 0 is the dispatcher and would
// deadlock waiting for the agent it's supposed to be running.
let image = match Image::builder("rust:1-slim")
.with_name(format!("rust_1_slim_park_gate_{}vcpu", vcpus))
.with_memory_mib(512)
.with_vcpus(vcpus)
.with_extra_file(&park_ko, "/supermachine-smpark.ko")
.with_warmup_tag("park_gate")
.with_warmup(|_vm| Ok(()))
.build()
{
Ok(i) => i,
Err(e) => {
eprintln!("[gate] bake failed: {e}");
std::process::exit(2);
}
};
let bake_ms = bake_t0.elapsed().as_millis();
eprintln!("[gate] bake done in {} ms", bake_ms);
// Mode toggle:
// PARK_GATE_FRESH_RESTORE=1 (default) — `restore_on_release(true)`,
// pool restores the worker between cycles. Tests that
// restore-from-snapshot is reliable (the actual ship gate).
// PARK_GATE_FRESH_RESTORE=0 — `restore_on_release(false)`,
// same restored worker reused across cycles. Tests that
// a single restored multi-vCPU guest stays healthy under
// repeated exec — useful for isolating restore-time
// issues vs in-flight-execution issues.
let fresh_restore = std::env::var("PARK_GATE_FRESH_RESTORE").ok().as_deref() != Some("0");
let max_workers: usize = std::env::var("PARK_GATE_MAX")
.ok()
.and_then(|v| v.parse().ok())
.unwrap_or(1);
let pool = image
.pool()
.min(0)
.max(max_workers)
.restore_on_release(fresh_restore)
// 30s per acquire — long enough to spawn+restore on a busy
// host, short enough to surface a hang as FAIL rather than
// wedging the gate. Default would be 60s.
.acquire_timeout(Duration::from_secs(30))
.build()?;
eprintln!("[gate] pool max={max_workers}");
eprintln!(
"[gate] mode: fresh_restore={fresh_restore} (set PARK_GATE_FRESH_RESTORE=0 to reuse)"
);
let mut successes = 0usize;
let mut failures = 0usize;
let mut first_fail: Option<(usize, String)> = None;
let cycles_t0 = Instant::now();
for i in 0..cycles {
let t0 = Instant::now();
let outcome: Result<bool, String> = (|| -> Result<bool, String> {
let vm = pool.acquire().map_err(|e| format!("acquire: {e}"))?;
let out = vm
.exec_builder()
.argv(["sh", "-c", "echo ok && cat /proc/modules | grep -q smpark"])
.timeout(Duration::from_secs(10))
.output()
.map_err(|e| format!("exec: {e}"))?;
let stdout = String::from_utf8_lossy(&out.stdout);
let exit_code = out.status.code().unwrap_or(-1);
if exit_code != 0 || !stdout.contains("ok") {
return Err(format!(
"exit={exit_code} stdout={stdout:?} stderr={:?}",
String::from_utf8_lossy(&out.stderr)
));
}
Ok(true)
})();
let cycle_ms = t0.elapsed().as_millis();
match outcome {
Ok(_) => {
successes += 1;
eprintln!(
"[gate] cycle {:3}/{cycles} OK in {} ms (total OK={successes} FAIL={failures})",
i + 1,
cycle_ms
);
}
Err(e) => {
failures += 1;
eprintln!(
"[gate] cycle {:3}/{cycles} FAIL in {} ms: {e}",
i + 1,
cycle_ms
);
if first_fail.is_none() {
first_fail = Some((i + 1, e));
}
}
}
}
let total_ms = cycles_t0.elapsed().as_millis();
eprintln!();
eprintln!("[gate] === SUMMARY ===");
eprintln!("[gate] cycles : {cycles}");
eprintln!("[gate] successes : {successes}");
eprintln!("[gate] failures : {failures}");
eprintln!("[gate] total ms : {total_ms}");
if let Some((n, e)) = &first_fail {
eprintln!("[gate] first fail : cycle {n}: {e}");
}
if failures == 0 {
eprintln!("[gate] PASS — multi-vCPU snapshot reliability gate met");
Ok(())
} else {
eprintln!("[gate] FAIL — {failures} cycle(s) failed");
std::process::exit(1);
}
}
}
#[cfg(any(
all(target_os = "linux", target_arch = "x86_64"),
all(target_os = "macos", target_arch = "aarch64")
))]
fn main() -> Result<(), Box<dyn std::error::Error>> {
supported::main()
}
#[cfg(not(any(
all(target_os = "linux", target_arch = "x86_64"),
all(target_os = "macos", target_arch = "aarch64")
)))]
fn main() -> Result<(), Box<dyn std::error::Error>> {
eprintln!("this example requires a linux-x86_64 (KVM) or macos-aarch64 (HVF) host; unsupported on this platform");
Ok(())
}