2020-12-18 20:53:32

by Randy Dunlap

[permalink] [raw]
Subject: [PATCH v2] JFS: more checks for invalid superblock

syzbot is feeding invalid superblock data to JFS for mount testing.
JFS does not check several of the fields -- just assumes that they
are good since the JFS_MAGIC and version fields are good.

In this case (syzbot reproducer), we have s_l2bsize == 0xda0c,
pad == 0xf045, and s_state == 0x50, all of which are invalid IMO.
Having s_l2bsize == 0xda0c causes this UBSAN warning:
UBSAN: shift-out-of-bounds in fs/jfs/jfs_mount.c:373:25
shift exponent -9716 is negative

s_l2bsize can be tested for correctness. pad can be tested for non-0
and punted. s_state can be tested for its valid values and punted.

Do those 3 tests and if any of them fails, report the superblock as
invalid/corrupt and let fsck handle it.

With this patch, chkSuper() says this when JFS_DEBUG is enabled:
jfs_mount: Mount Failure: superblock is corrupt!
Mount JFS Failure: -22
jfs_mount failed w/return code = -22

The obvious problem with this method is that next week there could
be another syzbot test that uses different fields for invalid values,
this making this like a game of whack-a-mole.

syzkaller link: https://syzkaller.appspot.com/bug?extid=36315852ece4132ec193

Reported-by: [email protected]
Reported-by: kernel test robot <[email protected]> # v2
Signed-off-by: Randy Dunlap <[email protected]>
Cc: Dave Kleikamp <[email protected]>
Cc: [email protected]
---
v2: fix sparse __le32 warning (lkp robot)

fs/jfs/jfs_filsys.h | 1 +
fs/jfs/jfs_mount.c | 10 ++++++++++
2 files changed, 11 insertions(+)

--- lnx-510.orig/fs/jfs/jfs_mount.c
+++ lnx-510/fs/jfs/jfs_mount.c
@@ -37,6 +37,7 @@
#include <linux/fs.h>
#include <linux/buffer_head.h>
#include <linux/blkdev.h>
+#include <linux/log2.h>

#include "jfs_incore.h"
#include "jfs_filsys.h"
@@ -366,6 +367,15 @@ static int chkSuper(struct super_block *
sbi->bsize = bsize;
sbi->l2bsize = le16_to_cpu(j_sb->s_l2bsize);

+ /* check some fields for possible corruption */
+ if (sbi->l2bsize != ilog2((u32)bsize) ||
+ j_sb->pad != 0 ||
+ le32_to_cpu(j_sb->s_state) > FM_STATE_MAX) {
+ rc = -EINVAL;
+ jfs_err("jfs_mount: Mount Failure: superblock is corrupt!");
+ goto out;
+ }
+
/*
* For now, ignore s_pbsize, l2bfactor. All I/O going through buffer
* cache.
--- lnx-510.orig/fs/jfs/jfs_filsys.h
+++ lnx-510/fs/jfs/jfs_filsys.h
@@ -268,5 +268,6 @@
* fsck() must be run to repair
*/
#define FM_EXTENDFS 0x00000008 /* file system extendfs() in progress */
+#define FM_STATE_MAX 0x0000000f /* max value of s_state */

#endif /* _H_JFS_FILSYS */


2020-12-18 20:55:40

by Dave Kleikamp

[permalink] [raw]
Subject: Re: [PATCH v2] JFS: more checks for invalid superblock

Thanks! This looks good and reasonable. I'll try to get it pushed out to
-next in the next few days.

Shaggy

On 12/18/20 2:17 PM, Randy Dunlap wrote:
> syzbot is feeding invalid superblock data to JFS for mount testing.
> JFS does not check several of the fields -- just assumes that they
> are good since the JFS_MAGIC and version fields are good.
>
> In this case (syzbot reproducer), we have s_l2bsize == 0xda0c,
> pad == 0xf045, and s_state == 0x50, all of which are invalid IMO.
> Having s_l2bsize == 0xda0c causes this UBSAN warning:
> UBSAN: shift-out-of-bounds in fs/jfs/jfs_mount.c:373:25
> shift exponent -9716 is negative
>
> s_l2bsize can be tested for correctness. pad can be tested for non-0
> and punted. s_state can be tested for its valid values and punted.
>
> Do those 3 tests and if any of them fails, report the superblock as
> invalid/corrupt and let fsck handle it.
>
> With this patch, chkSuper() says this when JFS_DEBUG is enabled:
> jfs_mount: Mount Failure: superblock is corrupt!
> Mount JFS Failure: -22
> jfs_mount failed w/return code = -22
>
> The obvious problem with this method is that next week there could
> be another syzbot test that uses different fields for invalid values,
> this making this like a game of whack-a-mole.
>
> syzkaller link: https://syzkaller.appspot.com/bug?extid=36315852ece4132ec193
>
> Reported-by: [email protected]
> Reported-by: kernel test robot <[email protected]> # v2
> Signed-off-by: Randy Dunlap <[email protected]>
> Cc: Dave Kleikamp <[email protected]>
> Cc: [email protected]
> ---
> v2: fix sparse __le32 warning (lkp robot)
>
> fs/jfs/jfs_filsys.h | 1 +
> fs/jfs/jfs_mount.c | 10 ++++++++++
> 2 files changed, 11 insertions(+)
>
> --- lnx-510.orig/fs/jfs/jfs_mount.c
> +++ lnx-510/fs/jfs/jfs_mount.c
> @@ -37,6 +37,7 @@
> #include <linux/fs.h>
> #include <linux/buffer_head.h>
> #include <linux/blkdev.h>
> +#include <linux/log2.h>
>
> #include "jfs_incore.h"
> #include "jfs_filsys.h"
> @@ -366,6 +367,15 @@ static int chkSuper(struct super_block *
> sbi->bsize = bsize;
> sbi->l2bsize = le16_to_cpu(j_sb->s_l2bsize);
>
> + /* check some fields for possible corruption */
> + if (sbi->l2bsize != ilog2((u32)bsize) ||
> + j_sb->pad != 0 ||
> + le32_to_cpu(j_sb->s_state) > FM_STATE_MAX) {
> + rc = -EINVAL;
> + jfs_err("jfs_mount: Mount Failure: superblock is corrupt!");
> + goto out;
> + }
> +
> /*
> * For now, ignore s_pbsize, l2bfactor. All I/O going through buffer
> * cache.
> --- lnx-510.orig/fs/jfs/jfs_filsys.h
> +++ lnx-510/fs/jfs/jfs_filsys.h
> @@ -268,5 +268,6 @@
> * fsck() must be run to repair
> */
> #define FM_EXTENDFS 0x00000008 /* file system extendfs() in progress */
> +#define FM_STATE_MAX 0x0000000f /* max value of s_state */
>
> #endif /* _H_JFS_FILSYS */
>