[PATCH 1/2] rbd: fix and simplify rbd_ioctl_set_ro()

* [PATCH 1/2] rbd: fix and simplify rbd_ioctl_set_ro()
@ 2017-11-06 15:49 Ilya Dryomov
  2017-11-06 15:49 ` [PATCH 2/2] rbd: get rid of rbd_mapping::read_only Ilya Dryomov
  0 siblings, 1 reply; 5+ messages in thread
From: Ilya Dryomov @ 2017-11-06 15:49 UTC (permalink / raw)
  To: ceph-devel

->open_count/-EBUSY check is bogus and wrong: when an open device is
set read-only, blkdev_write_iter() refuses further writes with -EPERM.
This is standard behaviour and all other block devices allow this.

set_disk_ro() call is also problematic: we affect the entire device
when called on a single partition.

All rbd_ioctl_set_ro() needs to do is refuse ro -> rw transition for
mapped snapshots.  Everything else can be handled by generic code.

Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
---
 drivers/block/rbd.c | 34 ++++++----------------------------
 1 file changed, 6 insertions(+), 28 deletions(-)

diff --git a/drivers/block/rbd.c b/drivers/block/rbd.c
index 6347ac194e90..832b235b2dc3 100644
--- a/drivers/block/rbd.c
+++ b/drivers/block/rbd.c
@@ -640,46 +640,24 @@ static void rbd_release(struct gendisk *disk, fmode_t mode)
 
 static int rbd_ioctl_set_ro(struct rbd_device *rbd_dev, unsigned long arg)
 {
-	int ret = 0;
-	int val;
-	bool ro;
-	bool ro_changed = false;
+	int ro;
 
-	/* get_user() may sleep, so call it before taking rbd_dev->lock */
-	if (get_user(val, (int __user *)(arg)))
+	if (get_user(ro, (int __user *)arg))
 		return -EFAULT;
 
-	ro = val ? true : false;
-	/* Snapshot doesn't allow to write*/
+	/* Snapshots can't be marked read-write */
 	if (rbd_dev->spec->snap_id != CEPH_NOSNAP && !ro)
 		return -EROFS;
 
-	spin_lock_irq(&rbd_dev->lock);
-	/* prevent others open this device */
-	if (rbd_dev->open_count > 1) {
-		ret = -EBUSY;
-		goto out;
-	}
-
-	if (rbd_dev->mapping.read_only != ro) {
-		rbd_dev->mapping.read_only = ro;
-		ro_changed = true;
-	}
-
-out:
-	spin_unlock_irq(&rbd_dev->lock);
-	/* set_disk_ro() may sleep, so call it after releasing rbd_dev->lock */
-	if (ret == 0 && ro_changed)
-		set_disk_ro(rbd_dev->disk, ro ? 1 : 0);
-
-	return ret;
+	/* Let blkdev_roset() handle it */
+	return -ENOTTY;
 }
 
 static int rbd_ioctl(struct block_device *bdev, fmode_t mode,
 			unsigned int cmd, unsigned long arg)
 {
 	struct rbd_device *rbd_dev = bdev->bd_disk->private_data;
-	int ret = 0;
+	int ret;
 
 	switch (cmd) {
 	case BLKROSET:
-- 
2.4.3


^ permalink raw reply related	[flat|nested] 5+ messages in thread