bitbake: server/process: Add bitbake.sock race handling

We've seen cases where the bitbake.sock file appears to disappear but the
server continues to hold bitbake.lock. The most likely explaination is
that some previous build directory was moved out the way, a server there
kept running, eventually exited and removed the sock file from the wrong
directory.

To guard against this, save the inode information for the sock file and check
it before deleting the file. The new code isn't entirely race free but should
guard against what is a rare but annoying potential issue.

(Bitbake rev: 52b6b099a47555811b8d0b311f62af712dd6eb8e)

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit b02ebbffdae27e564450446bf84c4e98d094ee4a)
Signed-off-by: Steve Sakoman <steve@sakoman.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This commit is contained in:
Richard Purdie
2023-01-14 04:12:02 -10:00
parent bfa114bfa8
commit da318dd088

View File

@@ -28,6 +28,7 @@ import datetime
import pickle
import traceback
import gc
import stat
import bb.server.xmlrpcserver
from bb import daemonize
from multiprocessing import queues
@@ -64,6 +65,9 @@ class ProcessServer():
self.bitbake_lock_name = lockname
self.sock = sock
self.sockname = sockname
# It is possible the directory may be renamed. Cache the inode of the socket file
# so we can tell if things changed.
self.sockinode = os.stat(self.sockname)[stat.ST_INO]
self.server_timeout = server_timeout
self.timeout = self.server_timeout
@@ -246,8 +250,14 @@ class ProcessServer():
serverlog("Exiting")
# Remove the socket file so we don't get any more connections to avoid races
# The build directory could have been renamed so if the file isn't the one we created
# we shouldn't delete it.
try:
os.unlink(self.sockname)
sockinode = os.stat(self.sockname)[stat.ST_INO]
if sockinode == self.sockinode:
os.unlink(self.sockname)
else:
serverlog("bitbake.sock inode mismatch (%s vs %s), not deleting." % (sockinode, self.sockinode))
except Exception as err:
serverlog("Removing socket file '%s' failed (%s)" % (self.sockname, err))
self.sock.close()