[PATCH v2] f2fs: validate ACL entry sizes in f2fs_acl_from_disk()

[PATCH v2] f2fs: validate ACL entry sizes in f2fs_acl_from_disk()

[Zhang Cen]

f2fs_acl_count() only validates the aggregate ACL xattr length. A
malformed ACL can still place ACL_USER or ACL_GROUP in a slot that only
contains struct f2fs_acl_entry_short bytes, and f2fs_acl_from_disk()
then reads entry->e_id before verifying that a full entry fits.

Require a short entry before reading e_tag and e_perm, and require a
full entry before reading e_id for ACL_USER and ACL_GROUP. Return
-EFSCORRUPTED for malformed on-disk ACL blobs so corrupted ACL xattrs
are reported as filesystem corruption instead of generic invalid
arguments.

Validation reproduced this kernel report:
KASAN slab-out-of-bounds in __f2fs_get_acl+0x6fb/0x7e0
RIP: 0033:0x7f4b835ea7aa
The buggy address belongs to the object at ffff888114589960 which belongs
to the cache kmalloc-8 of size 8
The buggy address is located 0 bytes to the right of allocated 8-byte
region [ffff888114589960, ffff888114589968)
Read of size 4
Call trace:
  dump_stack_lvl+0x66/0xa0 (?:?)
  print_report+0xce/0x630 (?:?)
  __f2fs_get_acl+0x6fb/0x7e0 (fs/f2fs/acl.c:169)
  srso_alias_return_thunk+0x5/0xfbef5 (?:?)
  __virt_addr_valid+0x224/0x430 (?:?)
  kasan_report+0xe0/0x110 (?:?)
  __f2fs_get_acl+0x5/0x7e0 (fs/f2fs/acl.c:169)
  __get_acl+0x281/0x380 (?:?)
  vfs_get_acl+0x10b/0x190 (?:?)
  do_get_acl+0x2a/0x410 (?:?)
  do_get_acl+0x9/0x410 (?:?)
  do_getxattr+0xe8/0x260 (?:?)
  filename_getxattr+0xd1/0x140 (?:?)
  do_getname+0x2d/0x2d0 (?:?)
  path_getxattrat+0x16c/0x200 (?:?)
  lock_release+0xc8/0x290 (?:?)
  cgroup_update_frozen+0x9d/0x320 (?:?)
  lockdep_hardirqs_on_prepare+0xea/0x1a0 (?:?)
  trace_hardirqs_on+0x1a/0x170 (?:?)
  _raw_spin_unlock_irq+0x28/0x50 (?:?)
  do_syscall_64+0x115/0x6a0 (arch/x86/entry/syscall_64.c:87)
  entry_SYSCALL_64_after_hwframe+0x77/0x7f (?:?)

Fixes: af48b85b8cd3 ("f2fs: add xattr and acl functionalities")

Assisted-by: Codex:gpt-5.5
Signed-off-by: Zhang Cen <rollkingzzc@gmail.com>
---
Changes in v2:
- Wrap the new ACL entry-size checks in unlikely().
- Return -EFSCORRUPTED for malformed on-disk ACL blobs.

 fs/f2fs/acl.c | 17 ++++++++++++-----
 1 file changed, 12 insertions(+), 5 deletions(-)

diff --git a/fs/f2fs/acl.c b/fs/f2fs/acl.c
index fa8d81a30fb91..e62ce979be0b7 100644
--- a/fs/f2fs/acl.c
+++ b/fs/f2fs/acl.c
@@ -53,14 +53,14 @@ static struct posix_acl *f2fs_acl_from_disk(const char *value, size_t size)
 	const char *end = value + size;
 
 	if (size < sizeof(struct f2fs_acl_header))
-		return ERR_PTR(-EINVAL);
+		return ERR_PTR(-EFSCORRUPTED);
 
 	if (hdr->a_version != cpu_to_le32(F2FS_ACL_VERSION))
-		return ERR_PTR(-EINVAL);
+		return ERR_PTR(-EFSCORRUPTED);
 
 	count = f2fs_acl_count(size);
 	if (count < 0)
-		return ERR_PTR(-EINVAL);
+		return ERR_PTR(-EFSCORRUPTED);
 	if (count == 0)
 		return NULL;
 
@@ -70,7 +70,8 @@ static struct posix_acl *f2fs_acl_from_disk(const char *value, size_t size)
 
 	for (i = 0; i < count; i++) {
 
-		if ((char *)entry > end)
+		if (unlikely((char *)entry +
+			     sizeof(struct f2fs_acl_entry_short) > end))
 			goto fail;
 
 		acl->a_entries[i].e_tag  = le16_to_cpu(entry->e_tag);
@@ -86,6 +87,9 @@ static struct posix_acl *f2fs_acl_from_disk(const char *value, size_t size)
 			break;
 
 		case ACL_USER:
+			if (unlikely((char *)entry +
+				     sizeof(struct f2fs_acl_entry) > end))
+				goto fail;
 			acl->a_entries[i].e_uid =
 				make_kuid(&init_user_ns,
 						le32_to_cpu(entry->e_id));
@@ -93,6 +97,9 @@ static struct posix_acl *f2fs_acl_from_disk(const char *value, size_t size)
 					sizeof(struct f2fs_acl_entry));
 			break;
 		case ACL_GROUP:
+			if (unlikely((char *)entry +
+				     sizeof(struct f2fs_acl_entry) > end))
+				goto fail;
 			acl->a_entries[i].e_gid =
 				make_kgid(&init_user_ns,
 						le32_to_cpu(entry->e_id));
@@ -108,7 +115,7 @@ static struct posix_acl *f2fs_acl_from_disk(const char *value, size_t size)
 	return acl;
 fail:
 	posix_acl_release(acl);
-	return ERR_PTR(-EINVAL);
+	return ERR_PTR(-EFSCORRUPTED);
 }
 
 static void *f2fs_acl_to_disk(struct f2fs_sb_info *sbi,
-- 
2.43.0

Re: [PATCH v2] f2fs: validate ACL entry sizes in f2fs_acl_from_disk()

[Chao Yu]

On 6/13/26 19:05, Zhang Cen wrote:
> f2fs_acl_count() only validates the aggregate ACL xattr length. A
> malformed ACL can still place ACL_USER or ACL_GROUP in a slot that only
> contains struct f2fs_acl_entry_short bytes, and f2fs_acl_from_disk()
> then reads entry->e_id before verifying that a full entry fits.
> 
> Require a short entry before reading e_tag and e_perm, and require a
> full entry before reading e_id for ACL_USER and ACL_GROUP. Return
> -EFSCORRUPTED for malformed on-disk ACL blobs so corrupted ACL xattrs
> are reported as filesystem corruption instead of generic invalid
> arguments.
> 
> Validation reproduced this kernel report:
> KASAN slab-out-of-bounds in __f2fs_get_acl+0x6fb/0x7e0
> RIP: 0033:0x7f4b835ea7aa
> The buggy address belongs to the object at ffff888114589960 which belongs
> to the cache kmalloc-8 of size 8
> The buggy address is located 0 bytes to the right of allocated 8-byte
> region [ffff888114589960, ffff888114589968)
> Read of size 4
> Call trace:
>   dump_stack_lvl+0x66/0xa0 (?:?)
>   print_report+0xce/0x630 (?:?)
>   __f2fs_get_acl+0x6fb/0x7e0 (fs/f2fs/acl.c:169)
>   srso_alias_return_thunk+0x5/0xfbef5 (?:?)
>   __virt_addr_valid+0x224/0x430 (?:?)
>   kasan_report+0xe0/0x110 (?:?)
>   __f2fs_get_acl+0x5/0x7e0 (fs/f2fs/acl.c:169)
>   __get_acl+0x281/0x380 (?:?)
>   vfs_get_acl+0x10b/0x190 (?:?)
>   do_get_acl+0x2a/0x410 (?:?)
>   do_get_acl+0x9/0x410 (?:?)
>   do_getxattr+0xe8/0x260 (?:?)
>   filename_getxattr+0xd1/0x140 (?:?)
>   do_getname+0x2d/0x2d0 (?:?)
>   path_getxattrat+0x16c/0x200 (?:?)
>   lock_release+0xc8/0x290 (?:?)
>   cgroup_update_frozen+0x9d/0x320 (?:?)
>   lockdep_hardirqs_on_prepare+0xea/0x1a0 (?:?)
>   trace_hardirqs_on+0x1a/0x170 (?:?)
>   _raw_spin_unlock_irq+0x28/0x50 (?:?)
>   do_syscall_64+0x115/0x6a0 (arch/x86/entry/syscall_64.c:87)
>   entry_SYSCALL_64_after_hwframe+0x77/0x7f (?:?)
> 
> Fixes: af48b85b8cd3 ("f2fs: add xattr and acl functionalities")
> 
> Assisted-by: Codex:gpt-5.5
> Signed-off-by: Zhang Cen <rollkingzzc@gmail.com>
> ---
> Changes in v2:
> - Wrap the new ACL entry-size checks in unlikely().
> - Return -EFSCORRUPTED for malformed on-disk ACL blobs.
> 
>  fs/f2fs/acl.c | 17 ++++++++++++-----
>  1 file changed, 12 insertions(+), 5 deletions(-)
> 
> diff --git a/fs/f2fs/acl.c b/fs/f2fs/acl.c
> index fa8d81a30fb91..e62ce979be0b7 100644
> --- a/fs/f2fs/acl.c
> +++ b/fs/f2fs/acl.c
> @@ -53,14 +53,14 @@ static struct posix_acl *f2fs_acl_from_disk(const char *value, size_t size)
>  	const char *end = value + size;
>  
>  	if (size < sizeof(struct f2fs_acl_header))
> -		return ERR_PTR(-EINVAL);
> +		return ERR_PTR(-EFSCORRUPTED);

The change is not related to this patch, right? I suspect you only give the
hint to let LLM to replace EINVAL w/ EFSCORRUPTED in f2fs_acl_from_disk(),
rather than replacing in newly added lines.

>  
>  	if (hdr->a_version != cpu_to_le32(F2FS_ACL_VERSION))
> -		return ERR_PTR(-EINVAL);
> +		return ERR_PTR(-EFSCORRUPTED);

Ditto,

>  
>  	count = f2fs_acl_count(size);
>  	if (count < 0)
> -		return ERR_PTR(-EINVAL);
> +		return ERR_PTR(-EFSCORRUPTED);

Ditto,

>  	if (count == 0)
>  		return NULL;
>  
> @@ -70,7 +70,8 @@ static struct posix_acl *f2fs_acl_from_disk(const char *value, size_t size)
>  
>  	for (i = 0; i < count; i++) {
>  
> -		if ((char *)entry > end)
> +		if (unlikely((char *)entry +
> +			     sizeof(struct f2fs_acl_entry_short) > end))
>  			goto fail;
>  
>  		acl->a_entries[i].e_tag  = le16_to_cpu(entry->e_tag);
> @@ -86,6 +87,9 @@ static struct posix_acl *f2fs_acl_from_disk(const char *value, size_t size)
>  			break;
>  
>  		case ACL_USER:
> +			if (unlikely((char *)entry +
> +				     sizeof(struct f2fs_acl_entry) > end))
> +				goto fail;
>  			acl->a_entries[i].e_uid =
>  				make_kuid(&init_user_ns,
>  						le32_to_cpu(entry->e_id));
> @@ -93,6 +97,9 @@ static struct posix_acl *f2fs_acl_from_disk(const char *value, size_t size)
>  					sizeof(struct f2fs_acl_entry));
>  			break;
>  		case ACL_GROUP:
> +			if (unlikely((char *)entry +
> +				     sizeof(struct f2fs_acl_entry) > end))
> +				goto fail;
>  			acl->a_entries[i].e_gid =
>  				make_kgid(&init_user_ns,
>  						le32_to_cpu(entry->e_id));
> @@ -108,7 +115,7 @@ static struct posix_acl *f2fs_acl_from_disk(const char *value, size_t size)
>  	return acl;
>  fail:
>  	posix_acl_release(acl);
> -	return ERR_PTR(-EINVAL);
> +	return ERR_PTR(-EFSCORRUPTED);

Ditto,

Thanks,

>  }
>  
>  static void *f2fs_acl_to_disk(struct f2fs_sb_info *sbi,

Re: [PATCH v2] f2fs: validate ACL entry sizes in f2fs_acl_from_disk()

[Cen Zhang]

> The change is not related to this patch, right? I suspect you only give the
> hint to let LLM to replace EINVAL w/ EFSCORRUPTED in f2fs_acl_from_disk(),
> rather than replacing in newly added lines.

Hi Chao,

Sorry for the noise.

I did use AI assistance when preparing v2 based on the bug details and
your review comments, but I did not review the resulting semantic scope
carefully enough. I am sorry for this.

I will send a v3 following your suggestion and  I will be more careful with
both the generated patch and my own review before sending future revisions.
Thanks for your guidance and careful review.

Best regards,
Zhang Cen