mirror of
https://github.com/MCKero6423/uv-k5-v3-emulator.git
synced 2026-10-02 11:07:31 +00:00
mckero
65e50c0065
Two faults that combined to take the web UI down twice in one session, each time
surfacing to the user as a 502 through the reverse proxy.
`pkill -f 'M uv-k5-v3'` in run.sh and trace_run.sh matched far more than intended.
-f tests the whole command line, so it also matched the shell running the pkill
(the pattern sits in its own argv), any script mentioning the machine type, and the
QEMU child of a running webui.py. Cleanup now lives in tools/lib_kill_emulator.sh:
pgrep -x on the binary name, confirm uv-k5-v3 in /proc/PID/cmdline, and optionally
scope to one QMP socket so a caller only stops the instance it owns.
The supervisor then could not recover from it. power_on() began with
`if self._client is not None: return False`, but a client object is not proof of a
live guest -- after an external kill the stale client made power_on refuse forever,
so the Power button was dead until the whole service was restarted. It now checks
whether the process actually exited and relaunches, logging why.
Tests:
- tools/test_kill_emulator.sh checks a plain process, a process whose command line
merely mentions uv-k5-v3, and the calling script all survive; that a real emulator
on a named socket is stopped; that one on another socket is not; and that an
unscoped call still clears everything. Verified it leaves a live webui.py alone.
- Two supervisor unit tests cover relaunch-after-external-kill and the case that
must still refuse, so this cannot regress into starting two emulators at once.
Verified end to end against the running web UI: kill the emulator from outside,
status reports unreachable, and pressing Power brings it back to
{"status":"running"} where before it stayed dead.
While writing the first version of the test I modelled the failure as a client
raising BrokenPipeError, which is not what is_running() looks at -- it checks
poll(). The mock was wrong, not the code; the test now has the process report an
exit status, which is what really happens.
112 lines
3.6 KiB
Bash
Executable File
112 lines
3.6 KiB
Bash
Executable File
#!/usr/bin/env bash
|
|
# The cleanup in run.sh must not kill anything other than an emulator.
|
|
#
|
|
# Regression test for a real incident: `pkill -f 'M uv-k5-v3'` matched a running
|
|
# webui.py's QEMU child and the shell issuing the pkill, taking the web UI offline
|
|
# and returning 502 through the reverse proxy. Twice.
|
|
set -u
|
|
|
|
HERE="$(cd "$(dirname "$0")" && pwd)"
|
|
# shellcheck source=tools/lib_kill_emulator.sh
|
|
. "$HERE/lib_kill_emulator.sh"
|
|
|
|
fails=0
|
|
note() { printf ' %s\n' "$1"; }
|
|
fail() { printf 'FAIL %s\n' "$1"; fails=$((fails + 1)); }
|
|
|
|
# A decoy whose command line mentions the machine type but is not the emulator.
|
|
# The old pkill pattern would have killed this.
|
|
sleep 120 &
|
|
decoy=$!
|
|
# Rename via a wrapper so the cmdline contains the trap string.
|
|
bash -c 'exec -a "sleep -M uv-k5-v3 decoy" sleep 120' &
|
|
decoy2=$!
|
|
sleep 1
|
|
|
|
note "decoy pids: $decoy $decoy2"
|
|
|
|
kill_emulators
|
|
|
|
if ! kill -0 "$decoy" 2>/dev/null; then
|
|
fail "killed a plain sleep"
|
|
else
|
|
note "PASS plain process untouched"
|
|
fi
|
|
|
|
if ! kill -0 "$decoy2" 2>/dev/null; then
|
|
fail "killed a process merely mentioning uv-k5-v3 in its command line"
|
|
else
|
|
note "PASS non-emulator process mentioning uv-k5-v3 untouched"
|
|
fi
|
|
|
|
# The script running this must survive too; the old pattern matched its own argv.
|
|
note "PASS this script survived (pid $$)"
|
|
|
|
kill "$decoy" "$decoy2" 2>/dev/null || true
|
|
wait 2>/dev/null || true
|
|
|
|
# And a real emulator must actually be terminated.
|
|
QEMU="$HOME/qemu-build/qemu-7.2+dfsg/build/qemu-system-arm"
|
|
ELF="$HOME/uvk5-port/uvk5-sat/build/CW/nr7y.cw.elf"
|
|
IMG=$(mktemp /tmp/kill-test-XXXX.img)
|
|
gzip -dc "$HERE/../assets/pristine/flash-pristine.img.gz" > "$IMG"
|
|
|
|
IMG2=$(mktemp /tmp/kill-test2-XXXX.img)
|
|
gzip -dc "$HERE/../assets/pristine/flash-pristine.img.gz" > "$IMG2"
|
|
|
|
if [ -x "$QEMU" ] && [ -f "$ELF" ]; then
|
|
# Two instances on different sockets, standing in for "the test's own" and
|
|
# "the one a running web UI owns".
|
|
"$QEMU" -M "uv-k5-v3,flash-image=$IMG" -nographic -monitor none \
|
|
-qmp "unix:/tmp/kill-test-a.sock,server=on,wait=off" -kernel "$ELF" \
|
|
>/dev/null 2>&1 &
|
|
mine=$!
|
|
"$QEMU" -M "uv-k5-v3,flash-image=$IMG2" -nographic -monitor none \
|
|
-qmp "unix:/tmp/kill-test-b.sock,server=on,wait=off" -kernel "$ELF" \
|
|
>/dev/null 2>&1 &
|
|
theirs=$!
|
|
sleep 3
|
|
|
|
if ! kill -0 "$mine" 2>/dev/null || ! kill -0 "$theirs" 2>/dev/null; then
|
|
fail "test emulators did not start"
|
|
else
|
|
kill_emulators /tmp/kill-test-a.sock
|
|
sleep 2
|
|
|
|
if kill -0 "$mine" 2>/dev/null; then
|
|
fail "the targeted emulator was left running"
|
|
else
|
|
note "PASS the emulator on the named socket is terminated"
|
|
fi
|
|
|
|
# This is the property that protects a live session: someone else's
|
|
# instance must survive a scoped cleanup.
|
|
if kill -0 "$theirs" 2>/dev/null; then
|
|
note "PASS an emulator on another socket survives"
|
|
else
|
|
fail "killed an emulator belonging to a different socket"
|
|
fi
|
|
|
|
# Unscoped still clears everything.
|
|
kill_emulators
|
|
sleep 2
|
|
if kill -0 "$theirs" 2>/dev/null; then
|
|
fail "unscoped cleanup left an emulator running"
|
|
kill "$theirs" 2>/dev/null || true
|
|
else
|
|
note "PASS unscoped cleanup terminates all emulators"
|
|
fi
|
|
fi
|
|
else
|
|
note "SKIP emulator binary or firmware missing"
|
|
fi
|
|
|
|
rm -f "$IMG" "$IMG2" /tmp/kill-test-a.sock /tmp/kill-test-b.sock
|
|
wait 2>/dev/null || true
|
|
|
|
if [ "$fails" -gt 0 ]; then
|
|
echo "$fails check(s) failed"
|
|
exit 1
|
|
fi
|
|
echo "cleanup only ever kills emulators"
|