[PATCH 02/15] libmultipath: Avoid that thread cancellation causes a hang

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Apparently uevq_stop() hangs every now and then in pthread_mutex_lock()
if the udev listener thread is canceled. Avoid this hang by removing
code from uevq_stop(). This is safe because uevq_stop() is only called
just before uevent_listen() returns and in that case neither clearing
my_uev_trigger nor signaling uev_condp is needed. Additionally, rename
uevq_stop() into uevent_cleanup() to make clear that its purpose is to
clean up resources and not to stop the uevent listener.

Signed-off-by: Bart Van Assche <bart.vanassche@xxxxxxxxxxx>
---
 libmultipath/uevent.c | 10 +++-------
 1 file changed, 3 insertions(+), 7 deletions(-)

diff --git a/libmultipath/uevent.c b/libmultipath/uevent.c
index 6247898..0f32dbc 100644
--- a/libmultipath/uevent.c
+++ b/libmultipath/uevent.c
@@ -124,15 +124,11 @@ service_uevq(struct list_head *tmpq)
 	}
 }
 
-static void uevq_stop(void *arg)
+static void uevent_cleanup(void *arg)
 {
 	struct udev *udev = arg;
 
-	condlog(3, "Stopping uev queue");
-	pthread_mutex_lock(uevq_lockp);
-	my_uev_trigger = NULL;
-	pthread_cond_signal(uev_condp);
-	pthread_mutex_unlock(uevq_lockp);
+	condlog(3, "Releasing uevent_listen() resources");
 	udev_unref(udev);
 }
 
@@ -495,7 +491,7 @@ int uevent_listen(struct udev *udev)
 		return 1;
 	}
 	udev_ref(udev);
-	pthread_cleanup_push(uevq_stop, udev);
+	pthread_cleanup_push(uevent_cleanup, udev);
 
 	monitor = udev_monitor_new_from_netlink(udev, "udev");
 	if (!monitor) {
-- 
2.10.1

--
dm-devel mailing list
dm-devel@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/dm-devel



[Index of Archives]     [DM Crypt]     [Fedora Desktop]     [ATA RAID]     [Fedora Marketing]     [Fedora Packaging]     [Fedora SELinux]     [Yosemite Discussion]     [KDE Users]     [Fedora Docs]

  Powered by Linux