[07/91] block: rescan partitions on invalidated devices on -ENOMEDIA

[Greg KH]

2.6.32-longterm review patch.  If anyone has any objections, please let us know.

------------------
 too

From: Tejun Heo <tj@xxxxxxxxxx>

commit 02e352287a40bd456eb78df705bf888bc3161d3f upstream.

__blkdev_get() doesn't rescan partitions if disk->fops->open() fails,
which leads to ghost partition devices lingering after medimum removal
is known to both the kernel and userland.  The behavior also creates a
subtle inconsistency where O_NONBLOCK open, which doesn't fail even if
there's no medium, clears the ghots partitions, which is exploited to
work around the problem from userland.

Fix it by updating __blkdev_get() to issue partition rescan after
-ENOMEDIA too.

This was reported in the following bz.

 https://bugzilla.kernel.org/show_bug.cgi?id=13029

Stable: 2.6.38

Signed-off-by: Tejun Heo <tj@xxxxxxxxxx>
Reported-by: David Zeuthen <zeuthen@xxxxxxxxx>
Reported-by: Martin Pitt <martin.pitt@xxxxxxxxxx>
Reported-by: Kay Sievers <kay.sievers@xxxxxxxx>
Tested-by: Kay Sievers <kay.sievers@xxxxxxxx>
Cc: Alan Cox <alan@xxxxxxxxxxxxxxxxxxx>
Signed-off-by: Jens Axboe <jaxboe@xxxxxxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxx>

---
 fs/block_dev.c |   27 ++++++++++++++++++---------
 1 file changed, 18 insertions(+), 9 deletions(-)

--- a/fs/block_dev.c
+++ b/fs/block_dev.c
@@ -1203,6 +1203,7 @@ static int __blkdev_get(struct block_dev
 			if (!bdev->bd_part)
 				goto out_clear;
 
+			ret = 0;
 			if (disk->fops->open) {
 				ret = disk->fops->open(bdev, mode);
 				if (ret == -ERESTARTSYS) {
@@ -1218,9 +1219,18 @@ static int __blkdev_get(struct block_dev
 					mutex_unlock(&bdev->bd_mutex);
 					goto restart;
 				}
-				if (ret)
-					goto out_clear;
 			}
+			/*
+			 * If the device is invalidated, rescan partition
+			 * if open succeeded or failed with -ENOMEDIUM.
+			 * The latter is necessary to prevent ghost
+			 * partitions on a removed medium.
+			 */
+			if (bdev->bd_invalidated && (!ret || ret == -ENOMEDIUM))
+				rescan_partitions(disk, bdev);
+			if (ret)
+				goto out_clear;
+
 			if (!bdev->bd_openers) {
 				bd_set_size(bdev,(loff_t)get_capacity(disk)<<9);
 				bdi = blk_get_backing_dev_info(bdev);
@@ -1228,8 +1238,6 @@ static int __blkdev_get(struct block_dev
 					bdi = &default_backing_dev_info;
 				bdev->bd_inode->i_data.backing_dev_info = bdi;
 			}
-			if (bdev->bd_invalidated)
-				rescan_partitions(disk, bdev);
 		} else {
 			struct block_device *whole;
 			whole = bdget_disk(disk, 0);
@@ -1256,13 +1264,14 @@ static int __blkdev_get(struct block_dev
 		put_disk(disk);
 		disk = NULL;
 		if (bdev->bd_contains == bdev) {
-			if (bdev->bd_disk->fops->open) {
+			ret = 0;
+			if (bdev->bd_disk->fops->open)
 				ret = bdev->bd_disk->fops->open(bdev, mode);
-				if (ret)
-					goto out_unlock_bdev;
-			}
-			if (bdev->bd_invalidated)
+			/* the same as first opener case, read comment there */
+			if (bdev->bd_invalidated && (!ret || ret == -ENOMEDIUM))
 				rescan_partitions(bdev->bd_disk, bdev);
+			if (ret)
+				goto out_unlock_bdev;
 		}
 	}
 	bdev->bd_openers++;


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/