[RFC] btrfs: fix null pointer deference at btrfs_sysfs_add_one+0x105
diff mbox

Message ID 1389777748-4341-1-git-send-email-Anand.Jain@oracle.com
State Accepted
Headers show

Commit Message

Anand Jain Jan. 15, 2014, 9:22 a.m. UTC
bdev is null when disk has disappeared and mounted with
the degrade option

stack trace
---------
btrfs_sysfs_add_one+0x105/0x1c0 [btrfs]
open_ctree+0x15f3/0x1fe0 [btrfs]
btrfs_mount+0x5db/0x790 [btrfs]
? alloc_pages_current+0xa4/0x160
mount_fs+0x34/0x1b0
vfs_kern_mount+0x62/0xf0
do_mount+0x22e/0xa80
? __get_free_pages+0x9/0x40
? copy_mount_options+0x31/0x170
SyS_mount+0x7e/0xc0
system_call_fastpath+0x16/0x1b
---------

reproducer:
-------
mkfs.btrfs -draid1 -mraid1 /dev/sdc /dev/sdd
(detach a disk)
devmgt detach /dev/sdc [1]
mount -o degrade /dev/sdd /btrfs
-------

[1] github.com/anajain/devmgt.git

Signed-off-by: Anand Jain <Anand.Jain@oracle.com>
---
 fs/btrfs/sysfs.c | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)

Comments

Hidetoshi Seto Feb. 14, 2014, 1:33 a.m. UTC | #1
I still see this trouble on v3.14-rc2.
I confirmed that we cannot do mount with -o degraded without
this patch. Could you pick this up, Chris?

Thanks,
H.Seto

Feel free to add:
Tested-by: Hidetoshi Seto <seto.hidetoshi@jp.fujitsu.com>

(2014/01/15 18:22), Anand Jain wrote:
> bdev is null when disk has disappeared and mounted with
> the degrade option
> 
> stack trace
> ---------
> btrfs_sysfs_add_one+0x105/0x1c0 [btrfs]
> open_ctree+0x15f3/0x1fe0 [btrfs]
> btrfs_mount+0x5db/0x790 [btrfs]
> ? alloc_pages_current+0xa4/0x160
> mount_fs+0x34/0x1b0
> vfs_kern_mount+0x62/0xf0
> do_mount+0x22e/0xa80
> ? __get_free_pages+0x9/0x40
> ? copy_mount_options+0x31/0x170
> SyS_mount+0x7e/0xc0
> system_call_fastpath+0x16/0x1b
> ---------
> 
> reproducer:
> -------
> mkfs.btrfs -draid1 -mraid1 /dev/sdc /dev/sdd
> (detach a disk)
> devmgt detach /dev/sdc [1]
> mount -o degrade /dev/sdd /btrfs
> -------
> 
> [1] github.com/anajain/devmgt.git
> 
> Signed-off-by: Anand Jain <Anand.Jain@oracle.com>
> ---
>  fs/btrfs/sysfs.c | 10 ++++++++--
>  1 file changed, 8 insertions(+), 2 deletions(-)
> 
> diff --git a/fs/btrfs/sysfs.c b/fs/btrfs/sysfs.c
> index ba94b27..f20bc1f 100644
> --- a/fs/btrfs/sysfs.c
> +++ b/fs/btrfs/sysfs.c
> @@ -578,8 +578,14 @@ static int add_device_membership(struct btrfs_fs_info *fs_info)
>  		return -ENOMEM;
>  
>  	list_for_each_entry(dev, &fs_devices->devices, dev_list) {
> -		struct hd_struct *disk = dev->bdev->bd_part;
> -		struct kobject *disk_kobj = &part_to_dev(disk)->kobj;
> +		struct hd_struct *disk;
> +		struct kobject *disk_kobj;
> +
> +		if (!dev->bdev)
> +			continue;
> +
> +		disk = dev->bdev->bd_part;
> +		disk_kobj = &part_to_dev(disk)->kobj;
>  
>  		error = sysfs_create_link(fs_info->device_dir_kobj,
>  					  disk_kobj, disk_kobj->name);
> 

--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Chris Mason Feb. 14, 2014, 1:39 a.m. UTC | #2
On 02/13/2014 08:33 PM, Hidetoshi Seto wrote:
> I still see this trouble on v3.14-rc2.
> I confirmed that we cannot do mount with -o degraded without
> this patch. Could you pick this up, Chris?
> 

Thanks for catching this, it'll be in the next pull.

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

Patch
diff mbox

diff --git a/fs/btrfs/sysfs.c b/fs/btrfs/sysfs.c
index ba94b27..f20bc1f 100644
--- a/fs/btrfs/sysfs.c
+++ b/fs/btrfs/sysfs.c
@@ -578,8 +578,14 @@  static int add_device_membership(struct btrfs_fs_info *fs_info)
 		return -ENOMEM;
 
 	list_for_each_entry(dev, &fs_devices->devices, dev_list) {
-		struct hd_struct *disk = dev->bdev->bd_part;
-		struct kobject *disk_kobj = &part_to_dev(disk)->kobj;
+		struct hd_struct *disk;
+		struct kobject *disk_kobj;
+
+		if (!dev->bdev)
+			continue;
+
+		disk = dev->bdev->bd_part;
+		disk_kobj = &part_to_dev(disk)->kobj;
 
 		error = sysfs_create_link(fs_info->device_dir_kobj,
 					  disk_kobj, disk_kobj->name);