[1/5] ext4: preserve i_mode if __ext4_set_acl() fails

Message ID 52152ec4b83fe0c406475bfee39efafaaa2a88fb.1499805486.git.ernesto.mnd.fernandez@gmail.com
State Awaiting Upstream
Headers show

Commit Message

Ernesto A. Fernández July 12, 2017, 9:53 a.m.
When changing a file's acl mask, __ext4_set_acl() will first set the group
bits of i_mode to the value of the mask, and only then set the actual
extended attribute representing the new acl.

If the second part fails (due to lack of space, for example) and the file
had no acl attribute to begin with, the system will from now on assume
that the mask permission bits are actual group permission bits, potentially
granting access to the wrong users.

Prevent this by only changing the inode mode after the acl has been set.

Signed-off-by: Ernesto A. Fernández <ernesto.mnd.fernandez@gmail.com>
---
 fs/ext4/acl.c | 15 +++++++++++----
 1 file changed, 11 insertions(+), 4 deletions(-)

Comments

Jan Kara July 17, 2017, 3:33 p.m. | #1
On Wed 12-07-17 06:53:51, Ernesto A. Fernández wrote:
> When changing a file's acl mask, __ext4_set_acl() will first set the group
> bits of i_mode to the value of the mask, and only then set the actual
> extended attribute representing the new acl.
> 
> If the second part fails (due to lack of space, for example) and the file
> had no acl attribute to begin with, the system will from now on assume
> that the mask permission bits are actual group permission bits, potentially
> granting access to the wrong users.
> 
> Prevent this by only changing the inode mode after the acl has been set.
> 
> Signed-off-by: Ernesto A. Fernández <ernesto.mnd.fernandez@gmail.com>

This clashes with my fix of another bug in that area [1] but the fixup
should be straightforward. The patch looks good to me so feel free to add

Reviewed-by: Jan Kara <jack@suse.cz>

[1] https://www.spinics.net/lists/stable/msg178111.html

								Honza

> ---
>  fs/ext4/acl.c | 15 +++++++++++----
>  1 file changed, 11 insertions(+), 4 deletions(-)
> 
> diff --git a/fs/ext4/acl.c b/fs/ext4/acl.c
> index 09441ae..2985cd0 100644
> --- a/fs/ext4/acl.c
> +++ b/fs/ext4/acl.c
> @@ -189,16 +189,17 @@ __ext4_set_acl(handle_t *handle, struct inode *inode, int type,
>  	void *value = NULL;
>  	size_t size = 0;
>  	int error;
> +	int update_mode = 0;
> +	umode_t mode = inode->i_mode;
>  
>  	switch (type) {
>  	case ACL_TYPE_ACCESS:
>  		name_index = EXT4_XATTR_INDEX_POSIX_ACL_ACCESS;
>  		if (acl) {
> -			error = posix_acl_update_mode(inode, &inode->i_mode, &acl);
> +			error = posix_acl_update_mode(inode, &mode, &acl);
>  			if (error)
>  				return error;
> -			inode->i_ctime = current_time(inode);
> -			ext4_mark_inode_dirty(handle, inode);
> +			update_mode = 1;
>  		}
>  		break;
>  
> @@ -221,8 +222,14 @@ __ext4_set_acl(handle_t *handle, struct inode *inode, int type,
>  				      value, size, xattr_flags);
>  
>  	kfree(value);
> -	if (!error)
> +	if (!error) {
>  		set_cached_acl(inode, type, acl);
> +		if (update_mode) {
> +			inode->i_mode = mode;
> +			inode->i_ctime = current_time(inode);
> +			ext4_mark_inode_dirty(handle, inode);
> +		}
> +	}
>  
>  	return error;
>  }
> -- 
> 2.1.4
> 
>
Theodore Ts'o July 31, 2017, 2:44 a.m. | #2
On Wed, Jul 12, 2017 at 06:53:51AM -0300, Ernesto A. Fernández wrote:
> When changing a file's acl mask, __ext4_set_acl() will first set the group
> bits of i_mode to the value of the mask, and only then set the actual
> extended attribute representing the new acl.
> 
> If the second part fails (due to lack of space, for example) and the file
> had no acl attribute to begin with, the system will from now on assume
> that the mask permission bits are actual group permission bits, potentially
> granting access to the wrong users.
> 
> Prevent this by only changing the inode mode after the acl has been set.
> 
> Signed-off-by: Ernesto A. Fernández <ernesto.mnd.fernandez@gmail.com>

Thanks, applied.

						- Ted

Patch

diff --git a/fs/ext4/acl.c b/fs/ext4/acl.c
index 09441ae..2985cd0 100644
--- a/fs/ext4/acl.c
+++ b/fs/ext4/acl.c
@@ -189,16 +189,17 @@  __ext4_set_acl(handle_t *handle, struct inode *inode, int type,
 	void *value = NULL;
 	size_t size = 0;
 	int error;
+	int update_mode = 0;
+	umode_t mode = inode->i_mode;
 
 	switch (type) {
 	case ACL_TYPE_ACCESS:
 		name_index = EXT4_XATTR_INDEX_POSIX_ACL_ACCESS;
 		if (acl) {
-			error = posix_acl_update_mode(inode, &inode->i_mode, &acl);
+			error = posix_acl_update_mode(inode, &mode, &acl);
 			if (error)
 				return error;
-			inode->i_ctime = current_time(inode);
-			ext4_mark_inode_dirty(handle, inode);
+			update_mode = 1;
 		}
 		break;
 
@@ -221,8 +222,14 @@  __ext4_set_acl(handle_t *handle, struct inode *inode, int type,
 				      value, size, xattr_flags);
 
 	kfree(value);
-	if (!error)
+	if (!error) {
 		set_cached_acl(inode, type, acl);
+		if (update_mode) {
+			inode->i_mode = mode;
+			inode->i_ctime = current_time(inode);
+			ext4_mark_inode_dirty(handle, inode);
+		}
+	}
 
 	return error;
 }