dm: avoid _hash_lock deadlock
authorMikulas Patocka <mpatocka@redhat.com>
Thu, 10 Dec 2009 23:51:52 +0000 (23:51 +0000)
committerAlasdair G Kergon <agk@redhat.com>
Thu, 10 Dec 2009 23:51:52 +0000 (23:51 +0000)
Fix a reported deadlock if there are still unprocessed multipath events
on a device that is being removed.

_hash_lock is held during dev_remove while trying to send the
outstanding events.  Sending the events requests the _hash_lock
again in dm_copy_name_and_uuid.

This patch introduces a separate lock around regions that modify the
link to the hash table (dm_set_mdptr) or the name or uuid so that
dm_copy_name_and_uuid no longer needs _hash_lock.

Additionally, dm_copy_name_and_uuid can only be called if md exists
so we can drop the dm_get() and dm_put() which can lead to a BUG()
while md is being freed.

The deadlock:
 #0 [ffff8106298dfb48] schedule at ffffffff80063035
 #1 [ffff8106298dfc20] __down_read at ffffffff8006475d
 #2 [ffff8106298dfc60] dm_copy_name_and_uuid at ffffffff8824f740
 #3 [ffff8106298dfc90] dm_send_uevents at ffffffff88252685
 #4 [ffff8106298dfcd0] event_callback at ffffffff8824c678
 #5 [ffff8106298dfd00] dm_table_event at ffffffff8824dd01
 #6 [ffff8106298dfd10] __hash_remove at ffffffff882507ad
 #7 [ffff8106298dfd30] dev_remove at ffffffff88250865
 #8 [ffff8106298dfd60] ctl_ioctl at ffffffff88250d80
 #9 [ffff8106298dfee0] do_ioctl at ffffffff800418c4
#10 [ffff8106298dff00] vfs_ioctl at ffffffff8002fab9
#11 [ffff8106298dff40] sys_ioctl at ffffffff8004bdaf
#12 [ffff8106298dff80] tracesys at ffffffff8005d28d (via system_call)

Cc: stable@kernel.org
Reported-by: guy keren <choo@actcom.co.il>
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
drivers/md/dm-ioctl.c
drivers/md/dm-uevent.c

index a679429315828eef8e4cbe896b8cdfba00f5eda2..d19854c98184de70a6ba7bb5432543eecd212a04 100644 (file)
@@ -56,6 +56,11 @@ static void dm_hash_remove_all(int keep_open_devices);
  */
 static DECLARE_RWSEM(_hash_lock);
 
+/*
+ * Protects use of mdptr to obtain hash cell name and uuid from mapped device.
+ */
+static DEFINE_MUTEX(dm_hash_cells_mutex);
+
 static void init_buckets(struct list_head *buckets)
 {
        unsigned int i;
@@ -206,7 +211,9 @@ static int dm_hash_insert(const char *name, const char *uuid, struct mapped_devi
                list_add(&cell->uuid_list, _uuid_buckets + hash_str(uuid));
        }
        dm_get(md);
+       mutex_lock(&dm_hash_cells_mutex);
        dm_set_mdptr(md, cell);
+       mutex_unlock(&dm_hash_cells_mutex);
        up_write(&_hash_lock);
 
        return 0;
@@ -224,7 +231,9 @@ static void __hash_remove(struct hash_cell *hc)
        /* remove from the dev hash */
        list_del(&hc->uuid_list);
        list_del(&hc->name_list);
+       mutex_lock(&dm_hash_cells_mutex);
        dm_set_mdptr(hc->md, NULL);
+       mutex_unlock(&dm_hash_cells_mutex);
 
        table = dm_get_table(hc->md);
        if (table) {
@@ -321,7 +330,9 @@ static int dm_hash_rename(uint32_t cookie, const char *old, const char *new)
         */
        list_del(&hc->name_list);
        old_name = hc->name;
+       mutex_lock(&dm_hash_cells_mutex);
        hc->name = new_name;
+       mutex_unlock(&dm_hash_cells_mutex);
        list_add(&hc->name_list, _name_buckets + hash_str(new_name));
 
        /*
@@ -1582,8 +1593,7 @@ int dm_copy_name_and_uuid(struct mapped_device *md, char *name, char *uuid)
        if (!md)
                return -ENXIO;
 
-       dm_get(md);
-       down_read(&_hash_lock);
+       mutex_lock(&dm_hash_cells_mutex);
        hc = dm_get_mdptr(md);
        if (!hc || hc->md != md) {
                r = -ENXIO;
@@ -1596,8 +1606,7 @@ int dm_copy_name_and_uuid(struct mapped_device *md, char *name, char *uuid)
                strcpy(uuid, hc->uuid ? : "");
 
 out:
-       up_read(&_hash_lock);
-       dm_put(md);
+       mutex_unlock(&dm_hash_cells_mutex);
 
        return r;
 }
index 6f65883aef124828c20a6ba1b29a26c541a941c8..c7c555a8c7b2ef619ab4671f023a6eb1b8953f33 100644 (file)
@@ -139,14 +139,13 @@ void dm_send_uevents(struct list_head *events, struct kobject *kobj)
                list_del_init(&event->elist);
 
                /*
-                * Need to call dm_copy_name_and_uuid from here for now.
-                * Context of previous var adds and locking used for
-                * hash_cell not compatable.
+                * When a device is being removed this copy fails and we
+                * discard these unsent events.
                 */
                if (dm_copy_name_and_uuid(event->md, event->name,
                                          event->uuid)) {
-                       DMERR("%s: dm_copy_name_and_uuid() failed",
-                             __func__);
+                       DMINFO("%s: skipping sending uevent for lost device",
+                              __func__);
                        goto uevent_free;
                }