Compare commits

..
Author SHA1 Message Date
ChuckBuilds 73fff8d2d5 test(systemd): pin the arena value instead of accepting a range
Review follow-up. The range check accepted 1, 3 and 4, so a change to 4 --
which hands most of the resident saving back -- passed a test whose whole
purpose is to notice that.

Pinned to the value the unit ships, in one named constant. Raising it is still
a legitimate response to a frame-time regression, but it should be a visible
edit here rather than silent drift, and the failure message says so.

Mutation-checked: changing the unit to 4 now fails.
2026-08-20 01:55:14 -04:00
ChuckBuilds 446207ffbc perf(systemd): cap glibc malloc arenas on the display service
Measured on a live rig 2.5 hours after start:

    RSS                          1030 MB
    Private_Dirty                 988 MB
    anonymous mappings > 10 MB       23
    largest        104, 79, 66, 63, 63 MB, on 64 MB-aligned addresses
    threads                           9
    cores                             3   -> glibc ceiling = 8 x 3 = 24 arenas

23 against a ceiling of 24, all 64 MB-aligned: these are glibc's per-thread
malloc arenas, not live objects. The data the process was actually holding
accounts for perhaps 15 MB -- the widest scroll strip observed was 35,746 x 64,
about 7 MB as RGB and the same again for its numpy mirror.

It is bloat rather than a leak: sampled four times over 135 seconds, RSS sat
between 990 and 1030 MB rather than climbing. glibc gives each allocating
thread its own arena, grows them to hold peak demand, and never gives them
back. A process that builds and drops large images across several threads is
exactly the shape that produces this.

The device had 59 MB free at the time, on 1845 MB total.

MALLOC_ARENA_MAX=2 trades a little allocator concurrency for that resident
memory. It is a tuning knob rather than a fix for a defect, so the rationale
and the measurements sit next to it in the unit file, and a test asserts they
stay there -- a bare environment variable invites removal by whoever meets it
next.

Two things this is NOT, both checked rather than assumed:

- Not an OOM problem today. A grep for "oom" in the service journal returned
  24 matches, all of which were the radar logging zoom=9 and zoom=7. The kernel
  OOM killer has not fired: dmesg has zero matches.
- Not currently capped by the unit's MemoryMax=85% either. That directive is in
  this file but absent from the unit actually installed on the rig, which
  reports MemoryMax=infinity, so nothing is enforcing a ceiling there.

The saving is unmeasured on hardware: applying it needs a service restart,
which blanks the panel, so that is the user's call rather than something to do
mid-audit. If p99 frame time regresses -- it sits at 18.4 ms against a 16.7 ms
budget for 60 FPS, so there is not much headroom -- raise the value rather than
remove it.
2026-08-20 00:17:27 -04:00
4 changed files with 109 additions and 184 deletions
+12
View File
@@ -8,6 +8,18 @@ Type=simple
User=root
WorkingDirectory=__PROJECT_ROOT_DIR__
Environment=PYTHONDONTWRITEBYTECODE=1
# glibc gives each allocating thread its own malloc arena, up to 8 x CPU count,
# and an arena that has grown is never handed back to the OS. This process runs
# 9 threads on a 3-core Pi, so the ceiling is 24 arenas -- and a rig measured at
# 1030 MB resident held 23 large anonymous mappings on 64 MB-aligned addresses,
# 920 MB of them, while the live data it was actually holding (widest scroll
# strip seen: 35,746 x 64) accounts for roughly 15 MB. That gap is arena bloat,
# not leaked objects: RSS was flat across repeated sampling, not climbing.
#
# Capping the arenas trades a little allocator concurrency for a large amount of
# resident memory on a device that has neither to spare. 2 is the usual value;
# raise it if frame times regress.
Environment=MALLOC_ARENA_MAX=2
ExecStart=/usr/bin/python3 __PROJECT_ROOT_DIR__/run.py
# Restart=always, not on-failure: run.py exiting 0 (a clean shutdown path taken
# for a reason that no longer applies, e.g. a config reload) would otherwise leave
-143
View File
@@ -1,143 +0,0 @@
"""GET /config/main must not hand out credentials.
The endpoint returned the raw config to anyone who could reach the port, and
this web interface has no authentication of any kind. Measured against a live
rig, an unauthenticated request returned:
github.api_token 40 chars
incoming-packages.ha_token 183 chars
jellyfin-now-playing.api_key 32 chars
ledmatrix-weather.api_key 32 chars
on-air.mqtt_password 8 chars
youtube.api_key 20 chars
youtube-stats.api_key 39 chars
A GitHub token and a Home Assistant long-lived token among them.
The x-secret masking the plugin config endpoints use does not apply here: this
endpoint never consults a schema, and core keys such as github.api_token have
no schema to carry the marker. Several of those fields *are* tagged x-secret in
their plugin's schema and were still returned in full, which is what makes the
schema route the wrong one to rely on for this endpoint.
Matching on field name is blunt. For a whole-config dump it is the right
default: anything named like a credential should not leave the process, and a
new plugin that adds a differently-shaped secret is covered without anyone
remembering to tag it.
"""
import pytest
from web_interface.blueprints.api_v3 import (
_looks_like_a_credential,
_redact_credentials,
)
@pytest.mark.parametrize("name", [
"password", "mqtt_password", "opensky_password", "passwd",
"api_key", "apikey", "API_KEY", "flightaware_api_key",
"token", "ha_token", "api_token", "access_token",
"secret", "client_secret", "spotify_client_secret",
"access_key", "private_key",
])
def test_credential_names_are_recognised(name):
assert _looks_like_a_credential(name)
@pytest.mark.parametrize("name", [
"timezone", "city", "brightness", "enabled", "update_interval",
"favorite_teams", "display_duration", "keyword",
])
def test_ordinary_names_are_left_alone(name):
assert not _looks_like_a_credential(name)
def test_the_measured_leak_is_closed():
"""The exact shape taken off the rig."""
config = {
"github": {"api_token": "ghp_" + "x" * 36},
"incoming-packages": {"ha_token": "y" * 183, "enabled": True},
"jellyfin-now-playing": {"api_key": "z" * 32},
"on-air": {"mqtt_password": "hunter22"},
"youtube": {"api_key": "k" * 20},
"timezone": "America/New_York",
}
out = _redact_credentials(config)
assert out["github"]["api_token"] == ""
assert out["incoming-packages"]["ha_token"] == ""
assert out["jellyfin-now-playing"]["api_key"] == ""
assert out["on-air"]["mqtt_password"] == ""
assert out["youtube"]["api_key"] == ""
# Everything else survives, or the config editor breaks.
assert out["timezone"] == "America/New_York"
assert out["incoming-packages"]["enabled"] is True
def test_nested_and_listed_credentials_are_reached():
config = {"a": {"b": {"c": {"password": "p"}}},
"feeds": [{"name": "x", "api_key": "k"}, {"name": "y"}]}
out = _redact_credentials(config)
assert out["a"]["b"]["c"]["password"] == ""
assert out["feeds"][0]["api_key"] == ""
assert out["feeds"][0]["name"] == "x"
def test_the_original_is_not_mutated():
"""The caller holds the live config; redaction must not edit it in place."""
config = {"github": {"api_token": "keepme"}}
_redact_credentials(config)
assert config["github"]["api_token"] == "keepme"
def test_a_credential_shaped_container_is_still_walked():
"""`secrets: {...}` is a section name, not a value to blank."""
config = {"secrets": {"api_key": "k", "note": "keep"}}
out = _redact_credentials(config)
assert out["secrets"]["api_key"] == ""
assert out["secrets"]["note"] == "keep"
def test_non_dict_input_passes_through():
assert _redact_credentials("plain") == "plain"
assert _redact_credentials(7) == 7
assert _redact_credentials(None) is None
def test_the_endpoint_itself_redacts():
"""Through the view function, not the helper.
The helper tests above all passed with the route still returning
`config` -- reverting the one line that calls the redactor changed
nothing, because nothing exercised the route. A property asserted on a
helper is not a property asserted on the endpoint, and it is the endpoint
that is exposed to the network.
"""
import json as _json
from unittest.mock import MagicMock
import flask
from web_interface.blueprints import api_v3 as mod
raw = {"github": {"api_token": "ghp_secret_value"},
"timezone": "America/New_York"}
manager = MagicMock()
manager.load_config.return_value = raw
previous = getattr(mod.api_v3, "config_manager", None)
mod.api_v3.config_manager = manager
app = flask.Flask(__name__)
try:
with app.test_request_context("/config/main"):
response = mod.get_main_config()
payload = response.get_json() if hasattr(response, "get_json") else _json.loads(response[0].data)
finally:
mod.api_v3.config_manager = previous
data = payload["data"]
assert data["github"]["api_token"] == "", (
"the endpoint returned the token; the redactor is not wired in")
assert data["timezone"] == "America/New_York"
# And the config the manager handed over is untouched.
assert raw["github"]["api_token"] == "ghp_secret_value"
+95
View File
@@ -0,0 +1,95 @@
"""The display unit must cap glibc's malloc arenas.
glibc hands each allocating thread its own malloc arena, up to 8 x CPU count,
and an arena that has grown is never returned to the OS. This process runs
threads for the render loop, the update workers and the background fetchers, so
on a 3-core Pi the ceiling is 24 arenas.
Measured on a live rig, 2.5 hours in:
RSS 1030 MB
Private_Dirty 988 MB
anonymous mappings > 10 MB 23 (ceiling is 8 x 3 = 24)
largest few 104, 79, 66, 63, 63 MB, on 64 MB-aligned addresses
against live data that accounts for perhaps 15 MB -- the widest scroll strip
observed was 35,746 x 64, about 7 MB as RGB and the same again for its numpy
mirror. Repeated sampling showed RSS flat between 990 and 1030 MB rather than
climbing, so this is arena bloat rather than a leak: memory Python has freed
but glibc is holding per-arena.
The device had 59 MB free at the time.
Capping the arena count trades a little allocator concurrency for that resident
memory. The render loop is latency-sensitive, so if p99 frame time regresses the
right response is to raise this rather than remove it.
"""
import re
from pathlib import Path
import pytest
UNIT = (Path(__file__).resolve().parent.parent / "systemd" / "ledmatrix.service")
#: The value the unit is expected to carry. 2 is the usual choice for a
#: threaded Python process; 1-4 all keep some of the saving, but only one of
#: them is what this project ships.
EXPECTED_ARENA_MAX = 2
def _environment(unit_text):
return dict(
line.split("=", 2)[1:3] if line.count("=") >= 2 else (line.split("=", 1)[1], "")
for line in unit_text.splitlines()
if line.startswith("Environment=")
)
def test_the_unit_exists():
assert UNIT.is_file(), f"{UNIT} is missing"
def test_malloc_arena_max_is_capped():
env = _environment(UNIT.read_text(encoding="utf-8"))
assert "MALLOC_ARENA_MAX" in env, (
"the display unit does not cap glibc arenas; on a 3-core Pi the default "
"ceiling is 24 and a measured rig held 23 of them, 920 MB"
)
value = int(env["MALLOC_ARENA_MAX"])
# Pinned, not a range. A range let a change to 4 -- which hands most of the
# saving back -- pass unnoticed, which was the point of the finding that
# prompted this. Raising it is a legitimate response to a frame-time
# regression, but it should be a visible edit here rather than a silent
# drift, so the number lives in one place and changing it shows up in
# review.
assert value == EXPECTED_ARENA_MAX, (
f"MALLOC_ARENA_MAX={value}, expected {EXPECTED_ARENA_MAX}. If this was "
"raised deliberately because frame times regressed, update "
"EXPECTED_ARENA_MAX here and say so in the commit."
)
def test_the_reason_is_recorded_next_to_it():
"""A bare tuning knob invites removal by whoever meets it next."""
text = UNIT.read_text(encoding="utf-8")
index = text.index("Environment=MALLOC_ARENA_MAX")
preamble = text[:index].splitlines()[-12:]
comment = "\n".join(line for line in preamble if line.startswith("#"))
assert "arena" in comment.lower(), "no explanation precedes the setting"
assert re.search(r"\d", comment), (
"the explanation cites no measurement, so a reader cannot tell whether "
"it still applies to their hardware"
)
@pytest.mark.parametrize("unit", ["ledmatrix.service"])
def test_the_unit_still_parses_as_ini(unit):
"""systemd will refuse a malformed unit, and the panel stays dark."""
import configparser
path = UNIT.parent / unit
parser = configparser.ConfigParser(strict=False)
# systemd allows repeated keys; ConfigParser needs them merged, not rejected.
parser.read_string(path.read_text(encoding="utf-8"))
assert parser.has_section("Service")
assert parser.has_option("Service", "ExecStart")
+2 -41
View File
@@ -262,54 +262,15 @@ def _stop_display_service():
result['status'] = status
return result
#: Field names whose value is a credential. Matched by name because this
#: endpoint returns the whole config, core keys included, and core config has
#: no schema to carry x-secret markers.
_CREDENTIAL_NAME_PARTS = ("password", "passwd", "secret", "token", "api_key",
"apikey", "access_key", "private_key", "client_secret")
def _looks_like_a_credential(name: str) -> bool:
lowered = name.lower()
return any(part in lowered for part in _CREDENTIAL_NAME_PARTS)
def _redact_credentials(value):
"""A copy of `value` with credential-named fields blanked.
/config/main returned the raw config to anyone who could reach the port,
and this interface has no authentication. On one rig that meant a 40-char
GitHub token, a 183-char Home Assistant token and five API keys were
readable by anything on the LAN.
The x-secret masking used by the plugin config endpoints does not help
here: this endpoint never consults a schema, and core keys such as
github.api_token have no schema to mark. Matching on the field name is
blunt, but for a whole-config dump the right default is that anything
named like a credential does not leave the process.
Blanked rather than removed, and safe to blank: POST /config/main merges
into the loaded config and only writes the keys it was given, so a client
that round-trips this response cannot erase a secret it never saw.
"""
if isinstance(value, dict):
return {k: ("" if _looks_like_a_credential(k) and not isinstance(v, (dict, list))
else _redact_credentials(v))
for k, v in value.items()}
if isinstance(value, list):
return [_redact_credentials(item) for item in value]
return value
@api_v3.route('/config/main', methods=['GET'])
def get_main_config():
"""Get main configuration, with credentials redacted."""
"""Get main configuration"""
try:
if not api_v3.config_manager:
return jsonify({'status': 'error', 'message': 'Config manager not initialized'}), 500
config = api_v3.config_manager.load_config()
return jsonify({'status': 'success', 'data': _redact_credentials(config)})
return jsonify({'status': 'success', 'data': config})
except Exception as e:
logger.error('Unhandled exception', exc_info=True)
return jsonify({'status': 'error', 'message': 'An error occurred; see logs for details', 'details': describe_exception(e)}), 500