[Trusty][Saucy][Raring][Quantal][Precise][PATCH 1/1] (no-up) overlayfs: add OVERLAYFS_SUPER_MAGIC to include/uapi/linux/magic.h

[Joseph Salisbury]

On 11/05/2013 08:42 AM, Andy Whitcroft wrote:
> On Mon, Nov 04, 2013 at 02:44:46PM -0500, Joseph Salisbury wrote:
>> On 11/04/2013 12:54 PM, Andy Whitcroft wrote:
>>> On Mon, Nov 04, 2013 at 12:30:01PM -0500, Joseph Salisbury wrote:
>>>> On 11/04/2013 12:08 PM, Stefan Bader wrote:
>>>>> On 04.11.2013 17:44, Joseph Salisbury wrote:
>>>>>> BugLink: http://bugs.launchpad.net/bugs/1247769
>>>>>>
>>>>>> Signed-off-by: Joseph Salisbury <joseph.salisbury at canonical.com>
>>>>>> ---
>>>>>>  include/uapi/linux/magic.h |    1 +
>>>>>>  1 file changed, 1 insertion(+)
>>>>>>
>>>>>> diff --git a/include/uapi/linux/magic.h b/include/uapi/linux/magic.h
>>>>>> index 2944278..7993c79 100644
>>>>>> --- a/include/uapi/linux/magic.h
>>>>>> +++ b/include/uapi/linux/magic.h
>>>>>> @@ -31,6 +31,7 @@
>>>>>>  #define PSTOREFS_MAGIC		0x6165676C
>>>>>>  #define EFIVARFS_MAGIC		0xde5e81e4
>>>>>>  #define HOSTFS_SUPER_MAGIC	0x00c0ffee
>>>>>> +#define OVERLAYFS_SUPER_MAGIC	0x794c764f
>>>>>>  
>>>>>>  #define MINIX_SUPER_MAGIC	0x137F		/* minix v1 fs, 14 char names */
>>>>>>  #define MINIX_SUPER_MAGIC2	0x138F		/* minix v1 fs, 30 char names */
>>>>>>
>>>>> I wished the bug report would be helpful in a way to say which other piece of sw
>>>>> would like that define in magic.h.
>>>>> It probably works as band-aid but I wonder whether it would be cleaner to remove
>>>>> the define from fs/overlayfs/super.c and include magic.h instead (to have only
>>>>> one define and in the same place)
>>>>>
>>>>> -Stefan
>>>>>
>>>>>
>>>>>
>>>> Thanks for the feedback, Stefan.
>>>>
>>>> I can rework the patch to also remove the define from
>>>> fs/overlayfs/super.c if that is best.  I'll also ask for additional
>>>> info, in the bug report, to identify which other sw would like the
>>>> define in magic.h.
>>> I would like to know why this would be no-up as well.  If overlayfs is
>>> not following a standard (and I am taking your word that this is so, not
>>> confirming it) then it should be going upstream me thinks.
>>>
>>> -apw
>> I added no-op because I didn't see an overlayfs directory in the
>> mainline linux ~/fs tree. 
> Well yes, this isn't Linus' upstream, but there is an upstream, Mikos
> something who looks after it.  We should send it to him still.
>
> -apw

Thanks, Andy.  I'll remove the no-up and send the patch to Mikos. 

James responded in the bug report[0] as to why this is needed.  Due to
another bug, they are changing Upstart tests to detect a /tmp overlalyfs
filesystem and display a warning.

Do you think it's best to remove the define from ~fs/overlayfs/super.c
and include it in magic.h instead?  If so, I'll make that change and
re-send the patch.

[0] https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1247769/comments/5