[RFC][PATCH] server/process: Avoid hanging if a parser process is terminated

[Peter Kjellerstedt <pkj@axis.com>]

If a parser process is terminated while holding a write lock, then it
will lead to a deadlock (see
https://docs.python.org/3/library/multiprocessing.html#multiprocessing.Process.terminate).

Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
---

After the discussion on IRC about the hanging parsing processes, I
just had to understand the code and what happens. This is a solution to
the problem, though I am not sure it is THE solution. It may also be
that the rlock in ConnectionReader should be handled the same way.
Anyway, even if this is not accepted as a solution to the problem, I now
at least know a lot more about how the parsing processes in bitbake
work... :)

 bitbake/lib/bb/server/process.py | 9 +++++++++
 1 file changed, 9 insertions(+)

diff --git a/bitbake/lib/bb/server/process.py b/bitbake/lib/bb/server/process.py
index 1636616660..544b00f2cd 100644
--- a/bitbake/lib/bb/server/process.py
+++ b/bitbake/lib/bb/server/process.py
@@ -20,6 +20,7 @@ import os
 import sys
 import time
 import select
+import signal
 import socket
 import subprocess
 import errno
@@ -728,6 +729,10 @@ class ConnectionReader(object):
     def close(self):
         return self.reader.close()
 
+terminated = False
+def catch_sigterm(signum, frame):
+    global terminated
+    terminated = True
 
 class ConnectionWriter(object):
 
@@ -739,8 +744,12 @@ class ConnectionWriter(object):
 
     def send(self, obj):
         obj = multiprocessing.reduction.ForkingPickler.dumps(obj)
+        oldsig = signal.signal(signal.SIGTERM, catch_sigterm)
         with self.wlock:
             self.writer.send_bytes(obj)
+        signal.signal(signal.SIGTERM, oldsig)
+        if terminated:
+            os.kill(os.getpid(), signal.SIGTERM)
 
     def fileno(self):
         return self.writer.fileno()