lib/bb: Update thread/process locks to use a timeout

The thread/process locks we use translate to futexes in Linux. If a
process dies holding the lock, anything else trying to take the lock
will hang indefinitely. An example would be the OOM killer taking out
a parser process.

To avoid bitbake processes just hanging indefinitely, add a timeout to
our lock calls using a context manager. If we can't obtain the lock
after waiting 5 minutes, hard exit out using os._exit(1). Use _exit()
to avoid locking in any other places trying to write error messages to
event handler queues (which also need locks).

Whilst a bit harsh, this should mean we stop having lots of long running
processes in cases where things are never going to work out and also
avoids hanging builds on the autobuilder.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

1 files changed, 8 insertions, 8 deletions

diff --git a/lib/bb/server/process.py b/lib/bb/server/process.py
index ac7749d36..b5f6faf6f 100644
--- a/lib/bb/server/process.py
+++ b/lib/bb/server/process.py
@@ -113,7 +113,7 @@ class ProcessServer():
     def register_idle_function(self, function, data):
         """Register a function to be called while the server is idle"""
         assert hasattr(function, '__call__')
-        with self._idlefuncsLock:
+        with bb.utils.lock_timeout(self._idlefuncsLock):
             self._idlefuns[function] = data
         serverlog("Registering idle function %s" % str(function))
 
@@ -379,7 +379,7 @@ class ProcessServer():
 
     def idle_thread(self):
         def remove_idle_func(function):
-            with self._idlefuncsLock:
+            with bb.utils.lock_timeout(self._idlefuncsLock):
                 del self._idlefuns[function]
                 self.idle_cond.notify_all()
 
@@ -387,7 +387,7 @@ class ProcessServer():
             nextsleep = 0.1
             fds = []
 
-            with self._idlefuncsLock:
+            with bb.utils.lock_timeout(self._idlefuncsLock):
                 items = list(self._idlefuns.items())
 
             for function, data in items:
@@ -743,7 +743,7 @@ class BBUIEventQueue:
         self.t.start()
 
     def getEvent(self):
-        with self.eventQueueLock:
+        with bb.utils.lock_timeout(self.eventQueueLock):
             if len(self.eventQueue) == 0:
                 return None
 
@@ -758,7 +758,7 @@ class BBUIEventQueue:
         return self.getEvent()
 
     def queue_event(self, event):
-        with self.eventQueueLock:
+        with bb.utils.lock_timeout(self.eventQueueLock):
             self.eventQueue.append(event)
             self.eventQueueNotify.set()
 
@@ -794,7 +794,7 @@ class ConnectionReader(object):
         return self.reader.poll(timeout)
 
     def get(self):
-        with self.rlock:
+        with bb.utils.lock_timeout(self.rlock):
             res = self.reader.recv_bytes()
         return multiprocessing.reduction.ForkingPickler.loads(res)
 
@@ -815,7 +815,7 @@ class ConnectionWriter(object):
 
     def _send(self, obj):
         gc.disable()
-        with self.wlock:
+        with bb.utils.lock_timeout(self.wlock):
             self.writer.send_bytes(obj)
         gc.enable()
 
@@ -828,7 +828,7 @@ class ConnectionWriter(object):
         # pthread_sigmask block/unblock would be nice but doesn't work, https://bugs.python.org/issue47139
         process = multiprocessing.current_process()
         if process and hasattr(process, "queue_signals"):
-            with process.signal_threadlock:
+            with bb.utils.lock_timeout(process.signal_threadlock):
                 process.queue_signals = True
                 self._send(obj)
                 process.queue_signals = False