[PATCH 1/2] process: Avoid bb.utils.timeout

* [PATCH 1/2] process: Avoid bb.utils.timeout
@ 2020-08-25 23:08 Richard Purdie
  2020-08-25 23:08 ` [PATCH 2/2] utils: Drop broken timeout function Richard Purdie
  0 siblings, 1 reply; 2+ messages in thread
From: Richard Purdie @ 2020-08-25 23:08 UTC (permalink / raw)
  To: bitbake-devel

I have a suspicion based on process traces that the flock() call is no
longer interrupted by SIGALRM and hence the timeout doesn't work. We
were relying on EINTR triggering around syscalls but python is likely
protecting us from that in modern versions.

Re-implement this code with a different mechanism which doesn't have
that potential issue.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
---
 lib/bb/server/process.py | 51 +++++++++++++++++++++-------------------
 1 file changed, 27 insertions(+), 24 deletions(-)

diff --git a/lib/bb/server/process.py b/lib/bb/server/process.py
index 6d936ed457..973bc45251 100644
--- a/lib/bb/server/process.py
+++ b/lib/bb/server/process.py
@@ -266,34 +266,37 @@ class ProcessServer():
         lock = None
 
         while not lock:
-            with bb.utils.timeout(3):
-                lock = bb.utils.lockfile(lockfile, shared=False, retry=False, block=True)
-                if lock:
-                    # We hold the lock so we can remove the file (hide stale pid data)
-                    # via unlockfile.
-                    bb.utils.unlockfile(lock)
-                    return
-
-                if not lock:
-                    # Some systems may not have lsof available
-                    procs = None
+            i = 0
+            lock = None
+            while not lock and i < 30:
+                time.sleep(0.1)
+                _, lock = bb.utils.lockfile(lockfile, shared=False, retry=False, block=False)
+                i += 1
+            if lock:
+                # We hold the lock so we can remove the file (hide stale pid data)
+                # via unlockfile.
+                bb.utils.unlockfile(lock)
+                return
+            if not lock:
+                # Some systems may not have lsof available
+                procs = None
+                try:
+                    procs = subprocess.check_output(["lsof", '-w', lockfile], stderr=subprocess.STDOUT)
+                except OSError as e:
+                    if e.errno != errno.ENOENT:
+                        raise
+                if procs is None:
+                    # Fall back to fuser if lsof is unavailable
                     try:
-                        procs = subprocess.check_output(["lsof", '-w', lockfile], stderr=subprocess.STDOUT)
+                        procs = subprocess.check_output(["fuser", '-v', lockfile], stderr=subprocess.STDOUT)
                     except OSError as e:
                         if e.errno != errno.ENOENT:
                             raise
-                    if procs is None:
-                        # Fall back to fuser if lsof is unavailable
-                        try:
-                            procs = subprocess.check_output(["fuser", '-v', lockfile], stderr=subprocess.STDOUT)
-                        except OSError as e:
-                            if e.errno != errno.ENOENT:
-                                raise
-
-                    msg = "Delaying shutdown due to active processes which appear to be holding bitbake.lock"
-                    if procs:
-                        msg += ":\n%s" % str(procs)
-                    print(msg)
+
+                msg = "Delaying shutdown due to active processes which appear to be holding bitbake.lock"
+                if procs:
+                    msg += ":\n%s" % str(procs)
+                print(msg)
 
     def idle_commands(self, delay, fds=None):
         nextsleep = delay
-- 
2.25.1


^ permalink raw reply related	[flat|nested] 2+ messages in thread