[PATCH v2] fs/ntfs3: Fix infinite loop in hdr_find_split due to zero-sized entry

Jiasheng Jiang posted 1 patch 2 weeks, 5 days ago
fs/ntfs3/index.c | 3 +++
1 file changed, 3 insertions(+)
[PATCH v2] fs/ntfs3: Fix infinite loop in hdr_find_split due to zero-sized entry
Posted by Jiasheng Jiang 2 weeks, 5 days ago
The function hdr_find_split iterates over index entries to calculate a
split point. The loop increments the offset 'o' by 'esize', which is
derived directly from the on-disk 'e->size' field.

If a corrupted or malicious filesystem image contains an index entry
with a size of 0, the variable 'o' will fail to advance. This results
in an infinite loop if the condition 'o < used_2' remains true, causing
a kernel hang (Denial of Service).

This patch adds a sanity check to ensure 'esize' is at least the size
of the NTFS_DE structure, consistent with validation logic in sibling
functions like hdr_find_e.

Fixes: 82cae269cfa9 ("fs/ntfs3: Add initialization of super block")
Signed-off-by: Jiasheng Jiang <jiashengjiangcool@gmail.com>
---
Changelog:

v1 -> v2:

1. Add a Fixes tag.
---
 fs/ntfs3/index.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/fs/ntfs3/index.c b/fs/ntfs3/index.c
index 7157cfd70fdc..da6927e6d360 100644
--- a/fs/ntfs3/index.c
+++ b/fs/ntfs3/index.c
@@ -577,6 +577,9 @@ static const struct NTFS_DE *hdr_find_split(const struct INDEX_HDR *hdr)
 			return p;
 
 		esize = le16_to_cpu(e->size);
+
+		if (esize < sizeof(struct NTFS_DE))
+			return NULL;
 	}
 
 	return e;
-- 
2.25.1
Re: [PATCH v2] fs/ntfs3: Fix infinite loop in hdr_find_split due to zero-sized entry
Posted by Markus Elfring 2 weeks, 5 days ago
…

> This patch adds a sanity check …

Thanks for another contribution.

* Does anything hinder to follow a corresponding wording requirement?
  https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/process/submitting-patches.rst?h=v6.19-rc5#n94

* Can it be helpful to append parentheses to function names?


Regards,
Markus
Re: [PATCH v2] fs/ntfs3: Fix infinite loop in hdr_find_split due to zero-sized entry
Posted by Greg KH 2 weeks, 5 days ago
On Sun, Jan 18, 2026 at 07:01:45PM +0000, Jiasheng Jiang wrote:
> The function hdr_find_split iterates over index entries to calculate a
> split point. The loop increments the offset 'o' by 'esize', which is
> derived directly from the on-disk 'e->size' field.
> 
> If a corrupted or malicious filesystem image contains an index entry
> with a size of 0, the variable 'o' will fail to advance. This results
> in an infinite loop if the condition 'o < used_2' remains true, causing
> a kernel hang (Denial of Service).
> 
> This patch adds a sanity check to ensure 'esize' is at least the size
> of the NTFS_DE structure, consistent with validation logic in sibling
> functions like hdr_find_e.
> 
> Fixes: 82cae269cfa9 ("fs/ntfs3: Add initialization of super block")
> Signed-off-by: Jiasheng Jiang <jiashengjiangcool@gmail.com>
> ---
> Changelog:
> 
> v1 -> v2:
> 
> 1. Add a Fixes tag.
> ---
>  fs/ntfs3/index.c | 3 +++
>  1 file changed, 3 insertions(+)
> 
> diff --git a/fs/ntfs3/index.c b/fs/ntfs3/index.c
> index 7157cfd70fdc..da6927e6d360 100644
> --- a/fs/ntfs3/index.c
> +++ b/fs/ntfs3/index.c
> @@ -577,6 +577,9 @@ static const struct NTFS_DE *hdr_find_split(const struct INDEX_HDR *hdr)
>  			return p;
>  
>  		esize = le16_to_cpu(e->size);
> +
> +		if (esize < sizeof(struct NTFS_DE))
> +			return NULL;
>  	}
>  
>  	return e;
> -- 
> 2.25.1
> 
> 

<formletter>

This is not the correct way to submit patches for inclusion in the
stable kernel tree.  Please read:
    https://www.kernel.org/doc/html/latest/process/stable-kernel-rules.html
for how to do this properly.

</formletter>