[PATCH 3.16 026/202] rbd: don't return 0 on unmap if RBD_DEV_FLAG_REMOVING is set

[Ben Hutchings]

3.16.66-rc1 review patch.  If anyone has any objections, please let me know.

------------------

From: Ilya Dryomov <idryomov@xxxxxxxxx>

commit 85f5a4d666fd9be73856ed16bb36c5af5b406b29 upstream.

There is a window between when RBD_DEV_FLAG_REMOVING is set and when
the device is removed from rbd_dev_list.  During this window, we set
"already" and return 0.

Returning 0 from write(2) can confuse userspace tools because
0 indicates that nothing was written.  In particular, "rbd unmap"
will retry the write multiple times a second:

  10:28:05.463299 write(4, "0", 1)        = 0
  10:28:05.463509 write(4, "0", 1)        = 0
  10:28:05.463720 write(4, "0", 1)        = 0
  10:28:05.463942 write(4, "0", 1)        = 0
  10:28:05.464155 write(4, "0", 1)        = 0

Signed-off-by: Ilya Dryomov <idryomov@xxxxxxxxx>
Tested-by: Dongsheng Yang <dongsheng.yang@xxxxxxxxxxxx>
[bwh: Backported to 3.16: adjust context]
Signed-off-by: Ben Hutchings <ben@xxxxxxxxxxxxxxx>
---
 drivers/block/rbd.c | 9 ++++-----
 1 file changed, 4 insertions(+), 5 deletions(-)

--- a/drivers/block/rbd.c
+++ b/drivers/block/rbd.c
@@ -5422,7 +5422,6 @@ static ssize_t do_rbd_remove(struct bus_
 	struct list_head *tmp;
 	int dev_id;
 	unsigned long ul;
-	bool already = false;
 	int ret;
 
 	ret = kstrtoul(buf, 10, &ul);
@@ -5447,13 +5446,13 @@ static ssize_t do_rbd_remove(struct bus_
 		spin_lock_irq(&rbd_dev->lock);
 		if (rbd_dev->open_count)
 			ret = -EBUSY;
-		else
-			already = test_and_set_bit(RBD_DEV_FLAG_REMOVING,
-							&rbd_dev->flags);
+		else if (test_and_set_bit(RBD_DEV_FLAG_REMOVING,
+					  &rbd_dev->flags))
+			ret = -EINPROGRESS;
 		spin_unlock_irq(&rbd_dev->lock);
 	}
 	spin_unlock(&rbd_dev_list_lock);
-	if (ret < 0 || already)
+	if (ret)
 		return ret;
 
 	rbd_dev_header_unwatch_sync(rbd_dev);