All of lore.kernel.org
 help / color / mirror / Atom feed
* [PATCH] btrfs: Fix UAF
@ 2018-01-26 13:20 Nikolay Borisov
  2018-01-29  2:38 ` Anand Jain
  2018-01-30 14:07 ` [RESEND PATCH] btrfs: Fix UAF when cleaning up fs_devs with a single stale device Nikolay Borisov
  0 siblings, 2 replies; 6+ messages in thread
From: Nikolay Borisov @ 2018-01-26 13:20 UTC (permalink / raw)
  To: linux-btrfs; +Cc: dsterba, anand.jain, Nikolay Borisov

Commit 4fde46f0cc71 ("Btrfs: free the stale device") introduced
btrfs_free_stale_device which iterates the device lists for all
registered btrfs filesystems and deletes those devices which aren't
mounted. In a btrfs_devices structure has only 1 device attached to it
and it is unused then btrfs_free_stale_devices will proceed to also
free the btrfs_fs_devices struct itself. Currently this leads to a UAF
since list_for_each_entry will try to perform a check on the already-
freed memory to see if it has to terminated the loop.

The fix is to use 'break' when we know we are freeing the current
fs_devs.

Fixes: 4fde46f0cc71 ("Btrfs: free the stale device")
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
---
 fs/btrfs/volumes.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
index f7147740b68e..c3ab55336ee0 100644
--- a/fs/btrfs/volumes.c
+++ b/fs/btrfs/volumes.c
@@ -645,6 +645,7 @@ static void btrfs_free_stale_devices(const char *path,
 				btrfs_sysfs_remove_fsid(fs_devs);
 				list_del(&fs_devs->list);
 				free_fs_devices(fs_devs);
+				break;
 			} else {
 				fs_devs->num_devices--;
 				list_del(&dev->dev_list);
-- 
2.7.4


^ permalink raw reply related	[flat|nested] 6+ messages in thread
* Re: [PATCH] btrfs: Fix UAF
  2018-01-26 13:20 [PATCH] btrfs: Fix UAF Nikolay Borisov
@ 2018-01-29  2:38 ` Anand Jain
  2018-01-29  7:01   ` Nikolay Borisov
  2018-01-30 14:07 ` [RESEND PATCH] btrfs: Fix UAF when cleaning up fs_devs with a single stale device Nikolay Borisov
  1 sibling, 1 reply; 6+ messages in thread
From: Anand Jain @ 2018-01-29  2:38 UTC (permalink / raw)
  To: Nikolay Borisov, linux-btrfs; +Cc: dsterba



On 01/26/2018 09:20 PM, Nikolay Borisov wrote:
> Commit 4fde46f0cc71 ("Btrfs: free the stale device") introduced
> btrfs_free_stale_device which iterates the device lists for all
> registered btrfs filesystems and deletes those devices which aren't
> mounted. In a btrfs_devices structure has only 1 device attached to it
> and it is unused then btrfs_free_stale_devices will proceed to also
> free the btrfs_fs_devices struct itself. Currently this leads to a UAF
> since list_for_each_entry will try to perform a check on the already-
> freed memory to see if it has to terminated the loop.
> 
> The fix is to use 'break' when we know we are freeing the current
> fs_devs.

  No break is needed as we need to iterate all stale devices and delete
  the found stale entry, so commit [1] used list_for_each_entry_safe()
  and removed the break,

  [1]
   commit 38cf665d338fca33af4b16f9ec7cad6637fc0fec
   Author: Anand Jain <Anand.Jain@oracle.com>
     btrfs: make btrfs_free_stale_device() to iterate all stales


  I am guessing UAF might be in[2], instead ?

  [2]
     free_fs_devices(fs_devs)
::
         while (!list_empty(&fs_devices->devices)) {
                 device = list_entry(fs_devices->devices.next,
                                     struct btrfs_device, dev_list);

Thanks, Anand

> Fixes: 4fde46f0cc71 ("Btrfs: free the stale device")
> Signed-off-by: Nikolay Borisov <nborisov@suse.com>
 >
> ---
>   fs/btrfs/volumes.c | 1 +
>   1 file changed, 1 insertion(+)
> 
> diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
> index f7147740b68e..c3ab55336ee0 100644
> --- a/fs/btrfs/volumes.c
> +++ b/fs/btrfs/volumes.c
> @@ -645,6 +645,7 @@ static void btrfs_free_stale_devices(const char *path,
>   				btrfs_sysfs_remove_fsid(fs_devs);
>   				list_del(&fs_devs->list);
>   				free_fs_devices(fs_devs);
> +				break;
>   			} else {
>   				fs_devs->num_devices--;
>   				list_del(&dev->dev_list);
> 

^ permalink raw reply	[flat|nested] 6+ messages in thread
* Re: [PATCH] btrfs: Fix UAF
  2018-01-29  2:38 ` Anand Jain
@ 2018-01-29  7:01   ` Nikolay Borisov
  2018-01-29 12:55     ` Anand Jain
  0 siblings, 1 reply; 6+ messages in thread
From: Nikolay Borisov @ 2018-01-29  7:01 UTC (permalink / raw)
  To: Anand Jain, linux-btrfs; +Cc: dsterba



On 29.01.2018 04:38, Anand Jain wrote:
> 
> 
> On 01/26/2018 09:20 PM, Nikolay Borisov wrote:
>> Commit 4fde46f0cc71 ("Btrfs: free the stale device") introduced
>> btrfs_free_stale_device which iterates the device lists for all
>> registered btrfs filesystems and deletes those devices which aren't
>> mounted. In a btrfs_devices structure has only 1 device attached to it
>> and it is unused then btrfs_free_stale_devices will proceed to also
>> free the btrfs_fs_devices struct itself. Currently this leads to a UAF
>> since list_for_each_entry will try to perform a check on the already-
>> freed memory to see if it has to terminated the loop.
>>
>> The fix is to use 'break' when we know we are freeing the current
>> fs_devs.
> 
>  No break is needed as we need to iterate all stale devices and delete
>  the found stale entry, so commit [1] used list_for_each_entry_safe()
>  and removed the break,

We only do the break if we know we have a single device in the current
fs_devs struct. And executing free_fs_devices would have already freed
the device + fs_devs.

> 
>  [1]
>   commit 38cf665d338fca33af4b16f9ec7cad6637fc0fec
>   Author: Anand Jain <Anand.Jain@oracle.com>
>     btrfs: make btrfs_free_stale_device() to iterate all stales
> 
> 
>  I am guessing UAF might be in[2], instead ?
> 
>  [2]
>     free_fs_devices(fs_devs)
> ::
>         while (!list_empty(&fs_devices->devices)) {
>                 device = list_entry(fs_devices->devices.next,
>                                     struct btrfs_device, dev_list);

It's not that, I thought so at first. But here using list_empty you are
awlays accessing the head the of the list, which is guaranteed to be
valid since you do the freeing of fs_devices outside of the while loop.


> 
> Thanks, Anand
> 
>> Fixes: 4fde46f0cc71 ("Btrfs: free the stale device")
>> Signed-off-by: Nikolay Borisov <nborisov@suse.com>
>>
>> ---
>>   fs/btrfs/volumes.c | 1 +
>>   1 file changed, 1 insertion(+)
>>
>> diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
>> index f7147740b68e..c3ab55336ee0 100644
>> --- a/fs/btrfs/volumes.c
>> +++ b/fs/btrfs/volumes.c
>> @@ -645,6 +645,7 @@ static void btrfs_free_stale_devices(const char
>> *path,
>>                   btrfs_sysfs_remove_fsid(fs_devs);
>>                   list_del(&fs_devs->list);
>>                   free_fs_devices(fs_devs);
>> +                break;
>>               } else {
>>                   fs_devs->num_devices--;
>>                   list_del(&dev->dev_list);
>>
> 

^ permalink raw reply	[flat|nested] 6+ messages in thread
* Re: [PATCH] btrfs: Fix UAF
  2018-01-29  7:01   ` Nikolay Borisov
@ 2018-01-29 12:55     ` Anand Jain
  0 siblings, 0 replies; 6+ messages in thread
From: Anand Jain @ 2018-01-29 12:55 UTC (permalink / raw)
  To: Nikolay Borisov, linux-btrfs; +Cc: dsterba



On 01/29/2018 03:01 PM, Nikolay Borisov wrote:
> 
> 
> On 29.01.2018 04:38, Anand Jain wrote:
>>
>>
>> On 01/26/2018 09:20 PM, Nikolay Borisov wrote:
>>> Commit 4fde46f0cc71 ("Btrfs: free the stale device") introduced
>>> btrfs_free_stale_device which iterates the device lists for all
>>> registered btrfs filesystems and deletes those devices which aren't
>>> mounted. In a btrfs_devices structure has only 1 device attached to it
>>> and it is unused then btrfs_free_stale_devices will proceed to also
>>> free the btrfs_fs_devices struct itself. Currently this leads to a UAF
>>> since list_for_each_entry will try to perform a check on the already-
>>> freed memory to see if it has to terminated the loop.
>>>
>>> The fix is to use 'break' when we know we are freeing the current
>>> fs_devs.
>>
>>   No break is needed as we need to iterate all stale devices and delete
>>   the found stale entry, so commit [1] used list_for_each_entry_safe()
>>   and removed the break,
> 
> We only do the break if we know we have a single device in the current
> fs_devs struct. And executing free_fs_devices would have already freed
> the device + fs_devs.
> 
>>
>>   [1]
>>    commit 38cf665d338fca33af4b16f9ec7cad6637fc0fec
>>    Author: Anand Jain <Anand.Jain@oracle.com>
>>      btrfs: make btrfs_free_stale_device() to iterate all stales
>>
>>
>>   I am guessing UAF might be in[2], instead ?
>>
>>   [2]
>>      free_fs_devices(fs_devs)
>> ::
>>          while (!list_empty(&fs_devices->devices)) {
>>                  device = list_entry(fs_devices->devices.next,
>>                                      struct btrfs_device, dev_list);
> 
> It's not that, I thought so at first. But here using list_empty you are
> awlays accessing the head the of the list, which is guaranteed to be
> valid since you do the freeing of fs_devices outside of the while loop.

break when num_devices == 1 is fine;

   Reviewed-by: Anand Jain <anand.jain@oracle.com>

Thanks, Anand

>>
>> Thanks, Anand
>>
>>> Fixes: 4fde46f0cc71 ("Btrfs: free the stale device")
>>> Signed-off-by: Nikolay Borisov <nborisov@suse.com>
>>>
>>> ---
>>>    fs/btrfs/volumes.c | 1 +
>>>    1 file changed, 1 insertion(+)
>>>
>>> diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
>>> index f7147740b68e..c3ab55336ee0 100644
>>> --- a/fs/btrfs/volumes.c
>>> +++ b/fs/btrfs/volumes.c
>>> @@ -645,6 +645,7 @@ static void btrfs_free_stale_devices(const char
>>> *path,
>>>                    btrfs_sysfs_remove_fsid(fs_devs);
>>>                    list_del(&fs_devs->list);
>>>                    free_fs_devices(fs_devs);
>>> +                break;
>>>                } else {
>>>                    fs_devs->num_devices--;
>>>                    list_del(&dev->dev_list);
>>>
>>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> 

^ permalink raw reply	[flat|nested] 6+ messages in thread
* [RESEND PATCH] btrfs: Fix UAF when cleaning up fs_devs with a single stale device
  2018-01-26 13:20 [PATCH] btrfs: Fix UAF Nikolay Borisov
  2018-01-29  2:38 ` Anand Jain
@ 2018-01-30 14:07 ` Nikolay Borisov
  2018-02-02 16:03   ` David Sterba
  1 sibling, 1 reply; 6+ messages in thread
From: Nikolay Borisov @ 2018-01-30 14:07 UTC (permalink / raw)
  To: dsterba; +Cc: linux-btrfs, Nikolay Borisov

Commit 4fde46f0cc71 ("Btrfs: free the stale device") introduced
btrfs_free_stale_device which iterates the device lists for all
registered btrfs filesystems and deletes those devices which aren't
mounted. In a btrfs_devices structure has only 1 device attached to it
and it is unused then btrfs_free_stale_devices will proceed to also
free the btrfs_fs_devices struct itself. Currently this leads to a UAF
since list_for_each_entry will try to perform a check on the already-
freed memory to see if it has to terminated the loop.

The fix is to use 'break' when we know we are freeing the current
fs_devs.

Fixes: 4fde46f0cc71 ("Btrfs: free the stale device")
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
---

Changed the subject to make it more descriptive 

 fs/btrfs/volumes.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
index f7147740b68e..c3ab55336ee0 100644
--- a/fs/btrfs/volumes.c
+++ b/fs/btrfs/volumes.c
@@ -645,6 +645,7 @@ static void btrfs_free_stale_devices(const char *path,
 				btrfs_sysfs_remove_fsid(fs_devs);
 				list_del(&fs_devs->list);
 				free_fs_devices(fs_devs);
+				break;
 			} else {
 				fs_devs->num_devices--;
 				list_del(&dev->dev_list);
-- 
2.7.4


^ permalink raw reply related	[flat|nested] 6+ messages in thread
* Re: [RESEND PATCH] btrfs: Fix UAF when cleaning up fs_devs with a single stale device
  2018-01-30 14:07 ` [RESEND PATCH] btrfs: Fix UAF when cleaning up fs_devs with a single stale device Nikolay Borisov
@ 2018-02-02 16:03   ` David Sterba
  0 siblings, 0 replies; 6+ messages in thread
From: David Sterba @ 2018-02-02 16:03 UTC (permalink / raw)
  To: Nikolay Borisov; +Cc: dsterba, linux-btrfs

On Tue, Jan 30, 2018 at 04:07:37PM +0200, Nikolay Borisov wrote:
> Commit 4fde46f0cc71 ("Btrfs: free the stale device") introduced
> btrfs_free_stale_device which iterates the device lists for all
> registered btrfs filesystems and deletes those devices which aren't
> mounted. In a btrfs_devices structure has only 1 device attached to it
> and it is unused then btrfs_free_stale_devices will proceed to also
> free the btrfs_fs_devices struct itself. Currently this leads to a UAF
> since list_for_each_entry will try to perform a check on the already-
> freed memory to see if it has to terminated the loop.
> 
> The fix is to use 'break' when we know we are freeing the current
> fs_devs.
> 
> Fixes: 4fde46f0cc71 ("Btrfs: free the stale device")
> Signed-off-by: Nikolay Borisov <nborisov@suse.com>

Added to next, thanks.

^ permalink raw reply	[flat|nested] 6+ messages in thread
end of thread, other threads:[~2018-02-02 16:05 UTC | newest]

Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2018-01-26 13:20 [PATCH] btrfs: Fix UAF Nikolay Borisov
2018-01-29  2:38 ` Anand Jain
2018-01-29  7:01   ` Nikolay Borisov
2018-01-29 12:55     ` Anand Jain
2018-01-30 14:07 ` [RESEND PATCH] btrfs: Fix UAF when cleaning up fs_devs with a single stale device Nikolay Borisov
2018-02-02 16:03   ` David Sterba

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.