mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
* [PATCH v3] ocfs2: validate inline xattr size and entry count in ocfs2_xattr_ibody_list
@ 2025-11-17 11:42 Deepanshu Kartikey
  2025-11-19 15:36 ` Heming Zhao
  0 siblings, 1 reply; 3+ messages in thread
From: Deepanshu Kartikey @ 2025-11-17 11:42 UTC (permalink / raw)
  To: mark, jlbec, joseph.qi, ocfs2-devel, linux-kernel
  Cc: Deepanshu Kartikey, syzbot+ab0ad25088673470d2d9, Heming Zhao

Add comprehensive validation of inline xattr metadata in
ocfs2_xattr_ibody_list() to prevent out-of-bounds access and
use-after-free bugs when processing corrupted inline xattrs.

The patch adds two critical validations:

1. Validates i_xattr_inline_size before use:
   - Ensures it does not exceed block size
   - Ensures it is at least large enough for xattr header
   - Prevents pointer arithmetic with corrupted size values that could
     point outside the inode block

2. Validates xattr entry count (xh_count):
   - Calculates maximum entries that can fit in the inline space
   - Rejects counts that exceed this limit
   - Prevents out-of-bounds array access in subsequent code

Without these checks, a corrupted filesystem with invalid inline xattr
metadata can cause the code to access memory beyond the allocated space.
For example:
- A corrupted i_xattr_inline_size of 0 would cause header pointer
  calculation to point past the end of the block
- A corrupted xh_count of 20041 with inline_size of 256 would cause
  array access beyond the 7 entries that actually fit, leading to
  use-after-free when accessing freed memory pages

The validation uses the correct inline_size (from di->i_xattr_inline_size)
rather than block size, ensuring accurate bounds checking for inline
xattrs specifically.

Reported-by: syzbot+ab0ad25088673470d2d9@syzkaller.appspotmail.com
Closes: https://syzkaller.appspot.com/bug?extid=ab0ad25088673470d2d9
Tested-by: syzbot+ab0ad25088673470d2d9@syzkaller.appspotmail.com
Suggested-by: Heming Zhao <heming.zhao@suse.com>
Link: https://lore.kernel.org/all/20251111073831.2027072-1-kartikey406@gmail.com/ [v1]
Link: https://lore.kernel.org/all/20251117063217.5690-1-kartikey406@gmail.com/T/ [v2]
Signed-off-by: Deepanshu Kartikey <kartikey406@gmail.com>
---
Changes in v3:
 - Moved validation from ocfs2_xattr_list_entries() to
   ocfs2_xattr_ibody_list() to use correct inline size calculation
   (suggested by Heming Zhao)
 - Added validation of i_xattr_inline_size before use
 - Added validation of xattr entry count against inline space
 - Changed return value to -EFSCORRUPTED for consistency
---
 fs/ocfs2/xattr.c | 30 ++++++++++++++++++++++++++++--
 1 file changed, 28 insertions(+), 2 deletions(-)

diff --git a/fs/ocfs2/xattr.c b/fs/ocfs2/xattr.c
index d70a20d29e3e..98fd4f3f2d2d 100644
--- a/fs/ocfs2/xattr.c
+++ b/fs/ocfs2/xattr.c
@@ -971,13 +971,39 @@ static int ocfs2_xattr_ibody_list(struct inode *inode,
 	struct ocfs2_xattr_header *header = NULL;
 	struct ocfs2_inode_info *oi = OCFS2_I(inode);
 	int ret = 0;
+	u16 xattr_count;
+	size_t max_entries;
+	u16 inline_size;
 
 	if (!(oi->ip_dyn_features & OCFS2_INLINE_XATTR_FL))
 		return ret;
 
+	inline_size = le16_to_cpu(di->i_xattr_inline_size);
+
+	/* Validate inline size is reasonable */
+	if (inline_size > inode->i_sb->s_blocksize ||
+	    inline_size < sizeof(struct ocfs2_xattr_header)) {
+		ocfs2_error(inode->i_sb,
+			    "Invalid xattr inline size %u in inode %llu\n",
+			    inline_size,
+			    (unsigned long long)OCFS2_I(inode)->ip_blkno);
+		return -EFSCORRUPTED;
+	}
+
 	header = (struct ocfs2_xattr_header *)
-		 ((void *)di + inode->i_sb->s_blocksize -
-		 le16_to_cpu(di->i_xattr_inline_size));
+		 ((void *)di + inode->i_sb->s_blocksize - inline_size);
+
+	xattr_count = le16_to_cpu(header->xh_count);
+	max_entries = (inline_size - sizeof(struct ocfs2_xattr_header)) /
+		       sizeof(struct ocfs2_xattr_entry);
+
+	if (xattr_count > max_entries) {
+		ocfs2_error(inode->i_sb,
+			    "xattr entry count %u exceeds maximum %zu in inode %llu\n",
+			    xattr_count, max_entries,
+			    (unsigned long long)OCFS2_I(inode)->ip_blkno);
+		return -EFSCORRUPTED;
+	}
 
 	ret = ocfs2_xattr_list_entries(inode, header, buffer, buffer_size);
 
-- 
2.43.0


^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [PATCH v3] ocfs2: validate inline xattr size and entry count in ocfs2_xattr_ibody_list
  2025-11-17 11:42 [PATCH v3] ocfs2: validate inline xattr size and entry count in ocfs2_xattr_ibody_list Deepanshu Kartikey
@ 2025-11-19 15:36 ` Heming Zhao
  2025-11-20  1:36   ` Deepanshu Kartikey
  0 siblings, 1 reply; 3+ messages in thread
From: Heming Zhao @ 2025-11-19 15:36 UTC (permalink / raw)
  To: Deepanshu Kartikey
  Cc: mark, jlbec, joseph.qi, ocfs2-devel, linux-kernel,
	syzbot+ab0ad25088673470d2d9

On Mon, Nov 17, 2025 at 05:12:24PM +0530, Deepanshu Kartikey wrote:
> Add comprehensive validation of inline xattr metadata in
> ocfs2_xattr_ibody_list() to prevent out-of-bounds access and
> use-after-free bugs when processing corrupted inline xattrs.
> 
> The patch adds two critical validations:
> 
> 1. Validates i_xattr_inline_size before use:
>    - Ensures it does not exceed block size
>    - Ensures it is at least large enough for xattr header
>    - Prevents pointer arithmetic with corrupted size values that could
>      point outside the inode block
> 
> 2. Validates xattr entry count (xh_count):
>    - Calculates maximum entries that can fit in the inline space
>    - Rejects counts that exceed this limit
>    - Prevents out-of-bounds array access in subsequent code
> 
> Without these checks, a corrupted filesystem with invalid inline xattr
> metadata can cause the code to access memory beyond the allocated space.
> For example:
> - A corrupted i_xattr_inline_size of 0 would cause header pointer
>   calculation to point past the end of the block
> - A corrupted xh_count of 20041 with inline_size of 256 would cause
>   array access beyond the 7 entries that actually fit, leading to
>   use-after-free when accessing freed memory pages

The following patch code is LGTM, but I can't understand the meaning of above
xh_cont:20041. Please show more details for the value 20041.

In my view,
the sizeof(struct ocfs2_xattr_header) is 16 (by: 2+2+2+2+(4+2+2)+0)
the sizeof(struct ocfs2_xattr_entry) is 16 (by: 4+2+1+1+8).
when the inline_size is 256.
the max number of entries for inline mode: (256-16) / 16 = 15.
therefore, the xh_count of 22 (by: 15+7) corresponds to a value that is 7
entries beyond the boundary.

Thanks,
Heming
> 
> The validation uses the correct inline_size (from di->i_xattr_inline_size)
> rather than block size, ensuring accurate bounds checking for inline
> xattrs specifically.
> 
> Reported-by: syzbot+ab0ad25088673470d2d9@syzkaller.appspotmail.com
> Closes: https://syzkaller.appspot.com/bug?extid=ab0ad25088673470d2d9
> Tested-by: syzbot+ab0ad25088673470d2d9@syzkaller.appspotmail.com
> Suggested-by: Heming Zhao <heming.zhao@suse.com>
> Link: https://lore.kernel.org/all/20251111073831.2027072-1-kartikey406@gmail.com/ [v1]
> Link: https://lore.kernel.org/all/20251117063217.5690-1-kartikey406@gmail.com/T/ [v2]
> Signed-off-by: Deepanshu Kartikey <kartikey406@gmail.com>
> ---
> Changes in v3:
>  - Moved validation from ocfs2_xattr_list_entries() to
>    ocfs2_xattr_ibody_list() to use correct inline size calculation
>    (suggested by Heming Zhao)
>  - Added validation of i_xattr_inline_size before use
>  - Added validation of xattr entry count against inline space
>  - Changed return value to -EFSCORRUPTED for consistency
> ---
>  fs/ocfs2/xattr.c | 30 ++++++++++++++++++++++++++++--
>  1 file changed, 28 insertions(+), 2 deletions(-)
> 
> diff --git a/fs/ocfs2/xattr.c b/fs/ocfs2/xattr.c
> index d70a20d29e3e..98fd4f3f2d2d 100644
> --- a/fs/ocfs2/xattr.c
> +++ b/fs/ocfs2/xattr.c
> @@ -971,13 +971,39 @@ static int ocfs2_xattr_ibody_list(struct inode *inode,
>  	struct ocfs2_xattr_header *header = NULL;
>  	struct ocfs2_inode_info *oi = OCFS2_I(inode);
>  	int ret = 0;
> +	u16 xattr_count;
> +	size_t max_entries;
> +	u16 inline_size;
>  
>  	if (!(oi->ip_dyn_features & OCFS2_INLINE_XATTR_FL))
>  		return ret;
>  
> +	inline_size = le16_to_cpu(di->i_xattr_inline_size);
> +
> +	/* Validate inline size is reasonable */
> +	if (inline_size > inode->i_sb->s_blocksize ||
> +	    inline_size < sizeof(struct ocfs2_xattr_header)) {
> +		ocfs2_error(inode->i_sb,
> +			    "Invalid xattr inline size %u in inode %llu\n",
> +			    inline_size,
> +			    (unsigned long long)OCFS2_I(inode)->ip_blkno);
> +		return -EFSCORRUPTED;
> +	}
> +
>  	header = (struct ocfs2_xattr_header *)
> -		 ((void *)di + inode->i_sb->s_blocksize -
> -		 le16_to_cpu(di->i_xattr_inline_size));
> +		 ((void *)di + inode->i_sb->s_blocksize - inline_size);
> +
> +	xattr_count = le16_to_cpu(header->xh_count);
> +	max_entries = (inline_size - sizeof(struct ocfs2_xattr_header)) /
> +		       sizeof(struct ocfs2_xattr_entry);
> +
> +	if (xattr_count > max_entries) {
> +		ocfs2_error(inode->i_sb,
> +			    "xattr entry count %u exceeds maximum %zu in inode %llu\n",
> +			    xattr_count, max_entries,
> +			    (unsigned long long)OCFS2_I(inode)->ip_blkno);
> +		return -EFSCORRUPTED;
> +	}
>  
>  	ret = ocfs2_xattr_list_entries(inode, header, buffer, buffer_size);
>  
> -- 
> 2.43.0
> 

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [PATCH v3] ocfs2: validate inline xattr size and entry count in ocfs2_xattr_ibody_list
  2025-11-19 15:36 ` Heming Zhao
@ 2025-11-20  1:36   ` Deepanshu Kartikey
  0 siblings, 0 replies; 3+ messages in thread
From: Deepanshu Kartikey @ 2025-11-20  1:36 UTC (permalink / raw)
  To: Heming Zhao
  Cc: mark, jlbec, joseph.qi, ocfs2-devel, linux-kernel,
	syzbot+ab0ad25088673470d2d9

On Wed, Nov 19, 2025 at 9:06 PM Heming Zhao <heming.zhao@suse.com> wrote:

> The following patch code is LGTM, but I can't understand the meaning of above
> xh_cont:20041. Please show more details for the value 20041.

The value 20041 is the actual corrupted xh_count from the syzbot
reproducer's filesystem image.

> In my view,
> the sizeof(struct ocfs2_xattr_header) is 16 (by: 2+2+2+2+(4+2+2)+0)
> the sizeof(struct ocfs2_xattr_entry) is 16 (by: 4+2+1+1+8).
> when the inline_size is 256.
> the max number of entries for inline mode: (256-16) / 16 = 15.
> therefore, the xh_count of 22 (by: 15+7) corresponds to a value that is 7
> entries beyond the boundary.

You're absolutely right! I made an error in my commit message - the
max is 15 entries, not 7.
I'll send v4 with the corrected example using xh_count of 22 as you
suggested, which is clearer than the raw syzbot value of 20041.

Thanks for catching this!

Deepanshu

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2025-11-20  1:36 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2025-11-17 11:42 [PATCH v3] ocfs2: validate inline xattr size and entry count in ocfs2_xattr_ibody_list Deepanshu Kartikey
2025-11-19 15:36 ` Heming Zhao
2025-11-20  1:36   ` Deepanshu Kartikey

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®