On 05/16/2018 12:30 AM, David Sterba wrote:
On Thu, Apr 12, 2018 at 10:29:35AM +0800, Anand Jain wrote:
close_fs_devices() closes devices of a given fsid, and it is limited
to all the devices of a fsid, so we don't have to hold the global
uuid_mutex, instead we need the device_list_mutex as the device state is
being changed.

Signed-off-by: Anand Jain <anand.j...@oracle.com>
---
  fs/btrfs/volumes.c | 9 ++++-----
  1 file changed, 4 insertions(+), 5 deletions(-)

diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
index dfebf8f29916..4c29214e0c18 100644
--- a/fs/btrfs/volumes.c
+++ b/fs/btrfs/volumes.c
@@ -995,7 +995,6 @@ static void btrfs_prepare_close_one_device(struct 
btrfs_device *device)
                                        device->uuid);
        BUG_ON(IS_ERR(new_device)); /* -ENOMEM */
- /* Safe because we are under uuid_mutex */
        if (device->name) {
                name = rcu_string_strdup(device->name->str, GFP_NOFS);
                BUG_ON(!name); /* -ENOMEM */
@@ -1013,10 +1012,12 @@ static int close_fs_devices(struct btrfs_fs_devices 
*fs_devices)
INIT_LIST_HEAD(&pending_put); - if (--fs_devices->opened > 0)
+       mutex_lock(&fs_devices->device_list_mutex);
+       if (--fs_devices->opened > 0) {
+               mutex_unlock(&fs_devices->device_list_mutex);
                return 0;
+       }
- mutex_lock(&fs_devices->device_list_mutex);
        list_for_each_entry_safe(device, tmp, &fs_devices->devices, dev_list) {
                btrfs_prepare_close_one_device(device);
                list_add(&device->dev_list, &pending_put);
@@ -1050,13 +1051,11 @@ int btrfs_close_devices(struct btrfs_fs_devices 
*fs_devices)
        struct btrfs_fs_devices *seed_devices = NULL;
        int ret;
- mutex_lock(&uuid_mutex);
        ret = close_fs_devices(fs_devices);
        if (!fs_devices->opened) {
                seed_devices = fs_devices->seed;
                fs_devices->seed = NULL;

This still touches ->seed, and also reads ->opened, some locking is
required here or it has to be clearly explained why it's not.

 The ->seed here is %fs_devices local so we don't need the
 global uuid_mutex. We cloned it in btrfs_prepare_sprout().

 Will add comments and send v2.

Thanks, Anand

        }
-       mutex_unlock(&uuid_mutex);
while (seed_devices) {
                fs_devices = seed_devices;
--
2.7.0

--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Reply via email to