The Linux Kernel Mailing List
 help / color / mirror / Atom feed
* [PATCH v2 0/2] ocfs2: add checks in ocfs2_xattr_find_entry() to avoid potential out-of-bound access.
@ 2024-05-17  9:41 Ferry Meng
  2024-05-17  9:41 ` [PATCH v2 1/2] ocfs2: add bounds checking to ocfs2_xattr_find_entry() Ferry Meng
  2024-05-17  9:41 ` [PATCH v2 2/2] ocfs2: strict bound check before memcmp in ocfs2_xattr_find_entry() Ferry Meng
  0 siblings, 2 replies; 6+ messages in thread
From: Ferry Meng @ 2024-05-17  9:41 UTC (permalink / raw)
  To: Mark Fasheh, Joel Becker, Joseph Qi, ocfs2-devel; +Cc: linux-kernel, Ferry Meng

Hi, all:

This patch series attempts to address a scenario where accessing user-defined
xattrs in a carefully crafted image can lead to out-of-bound access.(To speak
truthfully, I do not think this vehavior would occur under proper usage.)

In my testing environment, I constructed an OCFS2 image, created a file with
several user-defined xattrs(long name attributes, this will cause a "Non-INLINE"
xattr, which requires additional space for storage), and then forcibly modified
the xe_name_offset using a binary editing tool (e.g "hexedit"). Upon remounting
the image and running 'getfattr -d /path/to/file', this patchset was able to 
detect "partial" malicious modification.

In v2, I make these changes:
- (1/2) use xs->end directly, no need to parse a parameter.
- (2/2) define a local var 'name_offset'.

Ferry Meng (2):
  ocfs2: add bounds checking to ocfs2_xattr_find_entry()
  ocfs2: strict bound check before memcmp in ocfs2_xattr_find_entry()

 fs/ocfs2/xattr.c | 27 ++++++++++++++++++---------
 1 file changed, 18 insertions(+), 9 deletions(-)

-- 
2.32.0.3.g01195cf9f


^ permalink raw reply	[flat|nested] 6+ messages in thread

* [PATCH v2 1/2] ocfs2: add bounds checking to ocfs2_xattr_find_entry()
  2024-05-17  9:41 [PATCH v2 0/2] ocfs2: add checks in ocfs2_xattr_find_entry() to avoid potential out-of-bound access Ferry Meng
@ 2024-05-17  9:41 ` Ferry Meng
  2024-05-17 10:52   ` Joseph Qi
  2024-05-17  9:41 ` [PATCH v2 2/2] ocfs2: strict bound check before memcmp in ocfs2_xattr_find_entry() Ferry Meng
  1 sibling, 1 reply; 6+ messages in thread
From: Ferry Meng @ 2024-05-17  9:41 UTC (permalink / raw)
  To: Mark Fasheh, Joel Becker, Joseph Qi, ocfs2-devel; +Cc: linux-kernel, Ferry Meng

Add a paranoia check to make sure it doesn't stray beyond valid memory
region containing ocfs2 xattr entries when scanning for a match.
It will prevent out-of-bound access in case of crafted images.

Signed-off-by: Ferry Meng <mengferry@linux.alibaba.com>
---
 fs/ocfs2/xattr.c | 12 ++++++++----
 1 file changed, 8 insertions(+), 4 deletions(-)

diff --git a/fs/ocfs2/xattr.c b/fs/ocfs2/xattr.c
index 3b81213ed7b8..8aea94c90739 100644
--- a/fs/ocfs2/xattr.c
+++ b/fs/ocfs2/xattr.c
@@ -1062,7 +1062,7 @@ ssize_t ocfs2_listxattr(struct dentry *dentry,
 	return i_ret + b_ret;
 }
 
-static int ocfs2_xattr_find_entry(int name_index,
+static int ocfs2_xattr_find_entry(struct inode *inode, int name_index,
 				  const char *name,
 				  struct ocfs2_xattr_search *xs)
 {
@@ -1076,6 +1076,10 @@ static int ocfs2_xattr_find_entry(int name_index,
 	name_len = strlen(name);
 	entry = xs->here;
 	for (i = 0; i < le16_to_cpu(xs->header->xh_count); i++) {
+		if ((void *)entry >= xs->end) {
+			ocfs2_error(inode->i_sb, "corrupted xattr entries");
+			return -EFSCORRUPTED;
+		}
 		cmp = name_index - ocfs2_xattr_get_type(entry);
 		if (!cmp)
 			cmp = name_len - entry->xe_name_len;
@@ -1166,7 +1170,7 @@ static int ocfs2_xattr_ibody_get(struct inode *inode,
 	xs->base = (void *)xs->header;
 	xs->here = xs->header->xh_entries;
 
-	ret = ocfs2_xattr_find_entry(name_index, name, xs);
+	ret = ocfs2_xattr_find_entry(inode, name_index, name, xs);
 	if (ret)
 		return ret;
 	size = le64_to_cpu(xs->here->xe_value_size);
@@ -2698,7 +2702,7 @@ static int ocfs2_xattr_ibody_find(struct inode *inode,
 
 	/* Find the named attribute. */
 	if (oi->ip_dyn_features & OCFS2_INLINE_XATTR_FL) {
-		ret = ocfs2_xattr_find_entry(name_index, name, xs);
+		ret = ocfs2_xattr_find_entry(inode, name_index, name, xs);
 		if (ret && ret != -ENODATA)
 			return ret;
 		xs->not_found = ret;
@@ -2833,7 +2837,7 @@ static int ocfs2_xattr_block_find(struct inode *inode,
 		xs->end = (void *)(blk_bh->b_data) + blk_bh->b_size;
 		xs->here = xs->header->xh_entries;
 
-		ret = ocfs2_xattr_find_entry(name_index, name, xs);
+		ret = ocfs2_xattr_find_entry(inode, name_index, name, xs);
 	} else
 		ret = ocfs2_xattr_index_block_find(inode, blk_bh,
 						   name_index,
-- 
2.32.0.3.g01195cf9f


^ permalink raw reply related	[flat|nested] 6+ messages in thread

* [PATCH v2 2/2] ocfs2: strict bound check before memcmp in ocfs2_xattr_find_entry()
  2024-05-17  9:41 [PATCH v2 0/2] ocfs2: add checks in ocfs2_xattr_find_entry() to avoid potential out-of-bound access Ferry Meng
  2024-05-17  9:41 ` [PATCH v2 1/2] ocfs2: add bounds checking to ocfs2_xattr_find_entry() Ferry Meng
@ 2024-05-17  9:41 ` Ferry Meng
  2024-05-17 10:53   ` Joseph Qi
  1 sibling, 1 reply; 6+ messages in thread
From: Ferry Meng @ 2024-05-17  9:41 UTC (permalink / raw)
  To: Mark Fasheh, Joel Becker, Joseph Qi, ocfs2-devel; +Cc: linux-kernel, Ferry Meng

xattr in ocfs2 maybe 'non-indexed', which saved with additional space
requested. It's better to check if the memory is out of bound before
memcmp, although this possibility mainly comes from crafted poisonous
images.

Signed-off-by: Ferry Meng <mengferry@linux.alibaba.com>
---
 fs/ocfs2/xattr.c | 15 ++++++++++-----
 1 file changed, 10 insertions(+), 5 deletions(-)

diff --git a/fs/ocfs2/xattr.c b/fs/ocfs2/xattr.c
index 8aea94c90739..35c0cc2a51af 100644
--- a/fs/ocfs2/xattr.c
+++ b/fs/ocfs2/xattr.c
@@ -1068,7 +1068,7 @@ static int ocfs2_xattr_find_entry(struct inode *inode, int name_index,
 {
 	struct ocfs2_xattr_entry *entry;
 	size_t name_len;
-	int i, cmp = 1;
+	int i, name_offset, cmp = 1;
 
 	if (name == NULL)
 		return -EINVAL;
@@ -1083,10 +1083,15 @@ static int ocfs2_xattr_find_entry(struct inode *inode, int name_index,
 		cmp = name_index - ocfs2_xattr_get_type(entry);
 		if (!cmp)
 			cmp = name_len - entry->xe_name_len;
-		if (!cmp)
-			cmp = memcmp(name, (xs->base +
-				     le16_to_cpu(entry->xe_name_offset)),
-				     name_len);
+		if (!cmp) {
+			name_offset = le16_to_cpu(entry->xe_name_offset);
+			if ((xs->base + name_offset + name_len) > xs->end) {
+				ocfs2_error(inode->i_sb,
+					    "corrupted xattr entries");
+				return -EFSCORRUPTED;
+			}
+			cmp = memcmp(name, (xs->base + name_offset), name_len);
+		}
 		if (cmp == 0)
 			break;
 		entry += 1;
-- 
2.32.0.3.g01195cf9f


^ permalink raw reply related	[flat|nested] 6+ messages in thread

* Re: [PATCH v2 1/2] ocfs2: add bounds checking to ocfs2_xattr_find_entry()
  2024-05-17  9:41 ` [PATCH v2 1/2] ocfs2: add bounds checking to ocfs2_xattr_find_entry() Ferry Meng
@ 2024-05-17 10:52   ` Joseph Qi
  0 siblings, 0 replies; 6+ messages in thread
From: Joseph Qi @ 2024-05-17 10:52 UTC (permalink / raw)
  To: Ferry Meng, lei lu, akpm
  Cc: linux-kernel, Mark Fasheh, Joel Becker, ocfs2-devel



On 5/17/24 5:41 PM, Ferry Meng wrote:
> Add a paranoia check to make sure it doesn't stray beyond valid memory
> region containing ocfs2 xattr entries when scanning for a match.
> It will prevent out-of-bound access in case of crafted images.
> 
> Signed-off-by: Ferry Meng <mengferry@linux.alibaba.com>

Reported-by: lei lu <llfamsec@gmail.com>
Reviewed-by: Joseph Qi <joseph.qi@linux.alibaba.com>

> ---
>  fs/ocfs2/xattr.c | 12 ++++++++----
>  1 file changed, 8 insertions(+), 4 deletions(-)
> 
> diff --git a/fs/ocfs2/xattr.c b/fs/ocfs2/xattr.c
> index 3b81213ed7b8..8aea94c90739 100644
> --- a/fs/ocfs2/xattr.c
> +++ b/fs/ocfs2/xattr.c
> @@ -1062,7 +1062,7 @@ ssize_t ocfs2_listxattr(struct dentry *dentry,
>  	return i_ret + b_ret;
>  }
>  
> -static int ocfs2_xattr_find_entry(int name_index,
> +static int ocfs2_xattr_find_entry(struct inode *inode, int name_index,
>  				  const char *name,
>  				  struct ocfs2_xattr_search *xs)
>  {
> @@ -1076,6 +1076,10 @@ static int ocfs2_xattr_find_entry(int name_index,
>  	name_len = strlen(name);
>  	entry = xs->here;
>  	for (i = 0; i < le16_to_cpu(xs->header->xh_count); i++) {
> +		if ((void *)entry >= xs->end) {
> +			ocfs2_error(inode->i_sb, "corrupted xattr entries");
> +			return -EFSCORRUPTED;
> +		}
>  		cmp = name_index - ocfs2_xattr_get_type(entry);
>  		if (!cmp)
>  			cmp = name_len - entry->xe_name_len;
> @@ -1166,7 +1170,7 @@ static int ocfs2_xattr_ibody_get(struct inode *inode,
>  	xs->base = (void *)xs->header;
>  	xs->here = xs->header->xh_entries;
>  
> -	ret = ocfs2_xattr_find_entry(name_index, name, xs);
> +	ret = ocfs2_xattr_find_entry(inode, name_index, name, xs);
>  	if (ret)
>  		return ret;
>  	size = le64_to_cpu(xs->here->xe_value_size);
> @@ -2698,7 +2702,7 @@ static int ocfs2_xattr_ibody_find(struct inode *inode,
>  
>  	/* Find the named attribute. */
>  	if (oi->ip_dyn_features & OCFS2_INLINE_XATTR_FL) {
> -		ret = ocfs2_xattr_find_entry(name_index, name, xs);
> +		ret = ocfs2_xattr_find_entry(inode, name_index, name, xs);
>  		if (ret && ret != -ENODATA)
>  			return ret;
>  		xs->not_found = ret;
> @@ -2833,7 +2837,7 @@ static int ocfs2_xattr_block_find(struct inode *inode,
>  		xs->end = (void *)(blk_bh->b_data) + blk_bh->b_size;
>  		xs->here = xs->header->xh_entries;
>  
> -		ret = ocfs2_xattr_find_entry(name_index, name, xs);
> +		ret = ocfs2_xattr_find_entry(inode, name_index, name, xs);
>  	} else
>  		ret = ocfs2_xattr_index_block_find(inode, blk_bh,
>  						   name_index,

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [PATCH v2 2/2] ocfs2: strict bound check before memcmp in ocfs2_xattr_find_entry()
  2024-05-17  9:41 ` [PATCH v2 2/2] ocfs2: strict bound check before memcmp in ocfs2_xattr_find_entry() Ferry Meng
@ 2024-05-17 10:53   ` Joseph Qi
  2024-05-28 11:30     ` lei lu
  0 siblings, 1 reply; 6+ messages in thread
From: Joseph Qi @ 2024-05-17 10:53 UTC (permalink / raw)
  To: Ferry Meng, lei lu, akpm
  Cc: linux-kernel, Mark Fasheh, Joel Becker, ocfs2-devel



On 5/17/24 5:41 PM, Ferry Meng wrote:
> xattr in ocfs2 maybe 'non-indexed', which saved with additional space
> requested. It's better to check if the memory is out of bound before
> memcmp, although this possibility mainly comes from crafted poisonous
> images.
> 
> Signed-off-by: Ferry Meng <mengferry@linux.alibaba.com>

Reported-by: lei lu <llfamsec@gmail.com>
Reviewed-by: Joseph Qi <joseph.qi@linux.alibaba.com>

> ---
>  fs/ocfs2/xattr.c | 15 ++++++++++-----
>  1 file changed, 10 insertions(+), 5 deletions(-)
> 
> diff --git a/fs/ocfs2/xattr.c b/fs/ocfs2/xattr.c
> index 8aea94c90739..35c0cc2a51af 100644
> --- a/fs/ocfs2/xattr.c
> +++ b/fs/ocfs2/xattr.c
> @@ -1068,7 +1068,7 @@ static int ocfs2_xattr_find_entry(struct inode *inode, int name_index,
>  {
>  	struct ocfs2_xattr_entry *entry;
>  	size_t name_len;
> -	int i, cmp = 1;
> +	int i, name_offset, cmp = 1;
>  
>  	if (name == NULL)
>  		return -EINVAL;
> @@ -1083,10 +1083,15 @@ static int ocfs2_xattr_find_entry(struct inode *inode, int name_index,
>  		cmp = name_index - ocfs2_xattr_get_type(entry);
>  		if (!cmp)
>  			cmp = name_len - entry->xe_name_len;
> -		if (!cmp)
> -			cmp = memcmp(name, (xs->base +
> -				     le16_to_cpu(entry->xe_name_offset)),
> -				     name_len);
> +		if (!cmp) {
> +			name_offset = le16_to_cpu(entry->xe_name_offset);
> +			if ((xs->base + name_offset + name_len) > xs->end) {
> +				ocfs2_error(inode->i_sb,
> +					    "corrupted xattr entries");
> +				return -EFSCORRUPTED;
> +			}
> +			cmp = memcmp(name, (xs->base + name_offset), name_len);
> +		}
>  		if (cmp == 0)
>  			break;
>  		entry += 1;

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [PATCH v2 2/2] ocfs2: strict bound check before memcmp in ocfs2_xattr_find_entry()
  2024-05-17 10:53   ` Joseph Qi
@ 2024-05-28 11:30     ` lei lu
  0 siblings, 0 replies; 6+ messages in thread
From: lei lu @ 2024-05-28 11:30 UTC (permalink / raw)
  To: joseph.qi
  Cc: akpm, jlbec, linux-kernel, llfamsec, mark, mengferry, ocfs2-devel

Hi,

Maybe we should add the same check for ocfs2_xattr_list_entries.

^ permalink raw reply	[flat|nested] 6+ messages in thread

end of thread, other threads:[~2024-05-28 11:30 UTC | newest]

Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2024-05-17  9:41 [PATCH v2 0/2] ocfs2: add checks in ocfs2_xattr_find_entry() to avoid potential out-of-bound access Ferry Meng
2024-05-17  9:41 ` [PATCH v2 1/2] ocfs2: add bounds checking to ocfs2_xattr_find_entry() Ferry Meng
2024-05-17 10:52   ` Joseph Qi
2024-05-17  9:41 ` [PATCH v2 2/2] ocfs2: strict bound check before memcmp in ocfs2_xattr_find_entry() Ferry Meng
2024-05-17 10:53   ` Joseph Qi
2024-05-28 11:30     ` lei lu

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