Two T14 judges judge what their boots stage: hda_client_stall the second stream's resume, lan_dhcp_lease this boot's own lease

issues/audio/hda-client-stall-prints-a-stream-error-and-passes.md

@@ -0,0 +1,23 @@
+---
+status: open
+kind: tooling
+opened: 2026-10-03
+---
+
+# `hda_client_stall` prints a stream error and passes
+
+`tests/toyos-rust-tests/src/bin/hda_client_stall.rs` hands cpal an error
+callback that prints `audio error: …` and does nothing else. The fork reports
+two things through it: a stream whose signal pipe went away, and a dropped
+stream soundd did not let go of within its fade and a ring of periods. Either
+leaves the job exiting 0, and the `hda_client_stall` T14 row reads the job's
+exit code. `tone.rs`'s `play_tone` asserts on the same callback.
+
+The `testcases` readback of the T14 run of #638's head carries no
+`audio error` line. Whether the T14 ever reports one under this job is not
+measured.
+
+## Exit condition
+
+The job fails when its stream reports an error, and `hda_client_stall` passes
+on a T14 run of that head.

issues/audio/the-tone-client-waits-a-flat-200-ms-for-its-tail.md

@@ -0,0 +1,22 @@
+---
+status: open
+kind: tooling
+opened: 2026-10-03
+---
+
+# The tone client waits a flat 200 ms for its tail, and for its end with no bound
+
+`tests/toyos-rust-tests/src/tone.rs`'s `play_tone` is what `test_rs_audio_tone`
+and `test_rs_soundd_log_stall` play, so the `hda_tone` and `soundd_log_stall`
+T14 rows ride it. It sleeps 50 ms at a time until its callback has written the
+tone's last sample, with no bound of its own, and then sleeps a flat 200 ms for
+the tail to drain through soundd and the device before it drops the stream.
+
+Root `CLAUDE.md` allows neither: a wait is on the event, bounded by a timeout
+that fails loudly.
+
+## Exit condition
+
+`play_tone` holds no sleep: what it waits on is an event, bounded by a timeout
+that panics by name, and `hda_tone` and `soundd_log_stall` pass on a T14 run
+of that head.

issues/build/lan-hold-holds-two-boots-open-for-a-flat-twenty-seconds.md

@@ -0,0 +1,23 @@
+---
+status: open
+kind: tooling
+opened: 2026-10-03
+---
+
+# `lan_hold` holds two boots open for a flat twenty seconds
+
+`tests/toyos-rust-tests/src/bin/lan_hold.rs` sleeps `toyos_tco::LEASE_BOUND_MS`
+and exits. Two T14 boots run it as their one job: `lanleasecase`, so the boot
+does not end before netd's `--exit-with-lease` has, and `testcases-deaf`, so
+it does not end before the `dump-deaf-cpu` actuator has armed and dumped. Each
+is a fixed delay standing in for an event the job does not wait on, which root
+`CLAUDE.md` forbids in a test.
+
+On the T14 run of #638's head netd exited at 19.212 s of the `lanleasecase`
+boot and init stopped the machine at 21.202 s.
+
+## Exit condition
+
+Neither boot's one job holds a sleep: each waits on the event it is held open
+for, bounded by a timeout that panics by name, and `lan_lease_report` and
+`dump_nmi_probe` pass on a T14 run of that head.