Remove setproctitle (commaai#32716)

* Custom setproctitle (commaai#32667)

* add custom setproctitle

* add test

* Update poetry.lock

* fix lint

* support only Linux

* test only Linux

* final lint

* Update test_setproctitle.py

* Update setproctitle.py

* convert to threadnames

* delete proctitles

* Check str len and use PR_GET_NAME

* fix poetry.lock

* lint fix

* Update common/threadname.py

---------

Co-authored-by: reddyn12 <[email protected]>
Co-authored-by: Adeeb Shihadeh <[email protected]>

* revert that for now

* use last 15

* fix

* use name

* update those

* and modeld

* rm

---------

Co-authored-by: schlimeszn <[email protected]>
Co-authored-by: reddyn12 <[email protected]>
Co-authored-by: Comma Device <[email protected]>

common/realtime.py

@@ -4,7 +4,7 @@
 import time
 from collections import deque
 
-from setproctitle import getproctitle
+from openpilot.common.threadname import getthreadname
 
 from openpilot.system.hardware import PC
 
@@ -52,7 +52,7 @@ def __init__(self, rate: float, print_delay_threshold: float | None = 0.0) -> No
     self._print_delay_threshold = print_delay_threshold
     self._frame = 0
     self._remaining = 0.0
-    self._process_name = getproctitle()
+    self._thread_name = getthreadname()
     self._dts = deque([self._interval], maxlen=100)
     self._last_monitor_time = time.monotonic()
 
@@ -87,7 +87,7 @@ def monitor_time(self) -> bool:
     remaining = self._next_frame_time - time.monotonic()
     self._next_frame_time += self._interval
     if self._print_delay_threshold is not None and remaining < -self._print_delay_threshold:
-      print(f"{self._process_name} lagging by {-remaining * 1000:.2f} ms")
+      print(f"{self._thread_name} lagging by {-remaining * 1000:.2f} ms")
       lagged = True
     self._frame += 1
     self._remaining = remaining

common/tests/test_threadname.py

@@ -0,0 +1,8 @@
+from openpilot.common.threadname import setthreadname, getthreadname, LINUX
+
+class TestThreadName:
+  def test_set_get_threadname(self):
+    if LINUX:
+      name = 'TESTING'
+      setthreadname(name)
+      assert name == getthreadname()

common/threadname.py

@@ -0,0 +1,19 @@
+import ctypes
+import os
+
+LINUX = os.name == 'posix' and os.uname().sysname == 'Linux'
+
+if LINUX:
+  libc = ctypes.CDLL('libc.so.6')
+
+def setthreadname(name: str) -> None:
+  if LINUX:
+    name = name[-15:] + '\0'
+    libc.prctl(15, str.encode(name), 0, 0, 0)
+
+def getthreadname() -> str:
+  if LINUX:
+    name = ctypes.create_string_buffer(16)
+    libc.prctl(16, name)
+    return name.value.decode('utf-8')
+  return ""

pyproject.toml

@@ -137,7 +137,6 @@ future-fstrings = "*"
 # these should be removed
 psutil = "*"
 timezonefinder = "*"  # just used for nav ETA
-setproctitle = "*"
 pycryptodome = "*"  # used in updated/casync, panda, body, and a test
 
 [tool.poetry.group.dev.dependencies]

scripts/waste.py

@@ -3,7 +3,7 @@
 import time
 import numpy as np
 from multiprocessing import Process
-from setproctitle import setproctitle
+from openpilot.common.threadname import setthreadname
 
 def waste(core):
   os.sched_setaffinity(0, [core,])
@@ -16,7 +16,7 @@ def waste(core):
   j = 0
   while 1:
     if (i % 100) == 0:
-      setproctitle("%3d: %8d" % (core, i))
+      setthreadname("%3d: %8d" % (core, i))
       lt = time.monotonic()
       print("%3d: %8d %f  %.2f" % (core, i, lt-st, j))
       st = lt

selfdrive/modeld/modeld.py

@@ -6,7 +6,7 @@
 import cereal.messaging as messaging
 from cereal import car, log
 from pathlib import Path
-from setproctitle import setproctitle
+from openpilot.common.threadname import setthreadname
 from cereal.messaging import PubMaster, SubMaster
 from msgq.visionipc import VisionIpcClient, VisionStreamType, VisionBuf
 from openpilot.common.swaglog import cloudlog
@@ -24,7 +24,7 @@
 from openpilot.selfdrive.modeld.constants import ModelConstants
 from openpilot.selfdrive.modeld.models.commonmodel_pyx import ModelFrame, CLContext
 
-PROCESS_NAME = "selfdrive.modeld.modeld"
+THREAD_NAME = "selfdrive.modeld.modeld"
 SEND_RAW_PRED = os.getenv('SEND_RAW_PRED')
 
 MODEL_PATHS = {
@@ -114,9 +114,9 @@ def run(self, buf: VisionBuf, wbuf: VisionBuf, transform: np.ndarray, transform_
 def main(demo=False):
   cloudlog.warning("modeld init")
 
-  sentry.set_tag("daemon", PROCESS_NAME)
-  cloudlog.bind(daemon=PROCESS_NAME)
-  setproctitle(PROCESS_NAME)
+  sentry.set_tag("daemon", THREAD_NAME)
+  cloudlog.bind(daemon=THREAD_NAME)
+  setthreadname("modeld")
   config_realtime_process(7, 54)
 
   cloudlog.warning("setting up CL context")
@@ -286,7 +286,7 @@ def main(demo=False):
     args = parser.parse_args()
     main(demo=args.demo)
   except KeyboardInterrupt:
-    cloudlog.warning(f"child {PROCESS_NAME} got SIGINT")
+    cloudlog.warning(f"child {THREAD_NAME} got SIGINT")
   except Exception:
     sentry.capture_exception()
     raise

selfdrive/test/test_onroad.py

@@ -36,26 +36,26 @@
   # Baseline CPU usage by process
   "selfdrive.controls.controlsd": 32.0,
   "selfdrive.car.card": 22.0,
-  "./loggerd": 14.0,
-  "./encoderd": 17.0,
-  "./camerad": 14.5,
-  "./locationd": 11.0,
+  "loggerd": 14.0,
+  "encoderd": 17.0,
+  "camerad": 14.5,
+  "locationd": 11.0,
   "selfdrive.controls.plannerd": 11.0,
-  "./ui": 18.0,
+  "ui": 18.0,
   "selfdrive.locationd.paramsd": 9.0,
-  "./sensord": 7.0,
+  "sensord": 7.0,
   "selfdrive.controls.radard": 7.0,
-  "selfdrive.modeld.modeld": 13.0,
+  "modeld": 13.0,
   "selfdrive.modeld.dmonitoringmodeld": 8.0,
   "system.hardware.hardwared": 3.87,
   "selfdrive.locationd.calibrationd": 2.0,
   "selfdrive.locationd.torqued": 5.0,
   "selfdrive.ui.soundd": 3.5,
   "selfdrive.monitoring.dmonitoringd": 4.0,
-  "./proclogd": 1.54,
+  "proclogd": 1.54,
   "system.logmessaged": 0.2,
   "system.tombstoned": 0,
-  "./logcatd": 0,
+  "logcatd": 0,
   "system.micd": 6.0,
   "system.timed": 0,
   "selfdrive.pandad.pandad": 0,
@@ -67,12 +67,12 @@
 
 PROCS.update({
   "tici": {
-    "./pandad": 4.0,
-    "./ubloxd": 0.02,
+    "pandad": 4.0,
+    "ubloxd": 0.02,
     "system.ubloxd.pigeond": 6.0,
   },
   "tizi": {
-     "./pandad": 19.0,
+     "pandad": 19.0,
     "system.qcomgpsd.qcomgpsd": 1.0,
   }
 }.get(HARDWARE.get_device_type(), {}))
@@ -247,17 +247,17 @@ def test_cpu_usage(self, subtests):
     for pl in self.service_msgs['procLog']:
       for x in pl.procLog.procs:
         if len(x.cmdline) > 0:
-          n = list(x.cmdline)[0]
-          plogs_by_proc[n].append(x)
+          plogs_by_proc[x.name].append(x)
     print(plogs_by_proc.keys())
 
     cpu_ok = True
     dt = (self.service_msgs['procLog'][-1].logMonoTime - self.service_msgs['procLog'][0].logMonoTime) / 1e9
     for proc_name, expected_cpu in PROCS.items():
 
       err = ""
+      exp = "???"
       cpu_usage = 0.
-      x = plogs_by_proc[proc_name]
+      x = plogs_by_proc[proc_name[-15:]]
       if len(x) > 2:
         cpu_time = cputime_total(x[-1]) - cputime_total(x[0])
         cpu_usage = cpu_time / dt * 100.
@@ -309,7 +309,7 @@ def test_memory_usage(self):
     assert max(mems) - min(mems) <= 3.0
 
   def test_gpu_usage(self):
-    assert self.gpu_procs == {"weston", "ui", "camerad", "selfdrive.modeld.modeld"}
+    assert self.gpu_procs == {"weston", "ui", "camerad", "modeld"}
 
   def test_camera_processing_time(self):
     result = "\n"

system/manager/process.py

@@ -8,7 +8,7 @@
 from abc import ABC, abstractmethod
 from multiprocessing import Process
 
-from setproctitle import setproctitle
+from openpilot.common.threadname import setthreadname
 
 from cereal import car, log
 import cereal.messaging as messaging
@@ -27,7 +27,7 @@ def launcher(proc: str, name: str) -> None:
     mod = importlib.import_module(proc)
 
     # rename the process
-    setproctitle(proc)
+    setthreadname(proc)
 
     # create new context since we forked
     messaging.context = messaging.Context()