[PATCH] vfs_glusterfs: Make sure posix acls are formatted "before setting"/"after getting"

[Christopher R. Hertel]

On 08/12/2013 01:49 PM, Jeremy Allison wrote:
> On Mon, Aug 12, 2013 at 12:56:19PM -0500, Christopher R. Hertel wrote:
>> On 08/12/2013 10:58 AM, Jeremy Allison wrote:
>>> On Sun, Aug 11, 2013 at 03:12:34PM -0400, Raghavendra Talur wrote:
>>>> Simo, Andrew, Jeremy and Chris,
>>>>
>>>> Thank you all for the feedback. I apologize for not replying
>>>> earlier, I was on a vacation.
>>>>
>>>> 1. I agree with the formatting changes.
>>>> 2. When I first read the plugin code I had the same questions as
>>>>     Jeremy has, but, when I looked at the libacl code:
>>>>
>>>>     a. http://acl.sourcearchive.com/documentation/2.2.49-1/____acl__from__xattr_8c-source.html
>>>>     b. http://acl.sourcearchive.com/documentation/2.2.49-1/acl__ea_8h-source.html
>>>>
>>>>     I was convinced that this was a acceptable way of handling acls. As the plugin interacts
>>>>     only with Samba on one side and libgfapi of Gluster on the other, both residing on the
>>>>     same machine, I could not think of any case where plugin would need to handle anything
>>>>     extra than what libacl does.
>>>>
>>>> Not to undermine any of the reviews, I am asking this just to make it clear to any new
>>>> person who would ask the same question of "why din't we do what libacl does?"
>>>
>>> This is libacl code, not Samba code. libacl is a Linux
>>> specific library, and can do whatever it likes in terms
>>> of portability.
>>>
>>> I'd prefer higher standard for Samba code please.
>>
>> Jeremy,
>>
>> The question was aimed at developing a better understanding of the
>> risks/shortcomings of the code as it was submitted originally, to
>> avoid similar problems in the future.  Can you elaborate on your
>> explanation?
>
> I can't be any clearer really. To pass C structures over
> a network they have to be marshalled (unless you're depending
> on #pragma's to freeze the layout). It's just basic computer
> science.

There's more that could be said, but I'll take that up with Talur offline.

>> Also, Avati has submitted a fourth version of his rewrite of Talur's
>> original patch.  Jose is testing it, and your review would be
>> appreciated.
>
> Yes, will do asap !

Thanks.  As mentioned, Jose is going through a couple of tests we've put 
together and I'll have a look at it all once you're done.

Chris -)-----

-- 
"Implementing CIFS - the Common Internet FileSystem" ISBN: 013047116X
Samba Team -- http://www.samba.org/     -)-----   Christopher R. Hertel
jCIFS Team -- http://jcifs.samba.org/   -)-----   ubiqx development, uninq.
ubiqx Team -- http://www.ubiqx.org/     -)-----   crh at ubiqx.mn.org
OnLineBook -- http://ubiqx.org/cifs/    -)-----   crh at ubiqx.org