Cleancache support in XFS

[James Dingwall]

Konrad Rzeszutek Wilk wrote:
> On Fri, May 24, 2013 at 08:30:16AM +0100, James Dingwall wrote:
>> Konrad Rzeszutek Wilk wrote:
>>> On Thu, May 02, 2013 at 09:24:51AM +0100, James Dingwall wrote:
>>>> Dave Chinner wrote:
>>>>> On Wed, May 01, 2013 at 11:20:44AM -0500, Ben Myers wrote:
>>>>>> Hi James,
>>> Hey folks,
>>> I am walking through my vacation-emails-mbox.
>>>
>>>>>> On Wed, May 01, 2013 at 01:39:09PM +0100, James Dingwall wrote:
>>>>>>> In reference to: http://oss.sgi.com/archives/xfs/2012-05/msg00046.html
>>>>>>>
>>>>>>> $ grep -r cleancache fs/xfs
>>>>>>> on the 3.9 kernel source suggests that no patch was submitted to
>>>>>>> enable cleancache for the XFS filesystem.  Since it was suggested
>>>>>>> that this could be a one liner I've had a go and my first effort is
>>>>>>> inline below.  While this seems to compile OK I have no experience
>>>>>>> in filesystems so I would appreciate it if anyone can point out that
>>>>>>> it is obviously wrong and likely to eat my data before I try booting
>>>>>>> the kernel.
>>>>>>>
>>>>>>> If it seems a reasonable attempt what would be the best way to check
>>>>>>> that it isn't doing nasty things?
>>>>>> Hrm.. Looks like there is a doc in Documentation/vm/cleancache.txt which
>>>>>> includes a list of attributes the filesystem needs to have to work properly
>>>>>> with cleancache.
>>>>> So, those points are:
>>>> I had started to look at these too but I feel very out of my depth!
>>>> I had similar conclusions to what Dave wrote but I don't think my
>>>> thoughts should carry very much (any) weight.  Anyway I gambled and
>>>> booted my xen domU with this patch and so far so good...  xen top
>>>> shows that tmem is now being used where previously it wasn't.  I'll
>>>> try running the xfstests at the weekend after a couple more days up
>>>> time to see what happens.
>>> And how did it go?
>> I am running the patch I created on 3.9.3 on half of my xen guests
>> now and have not noticed any stability or filesystem problems.  xl
>> top with 'T' shows that the guests running with it are using
>> ephemeral pages were those without do not.  I did do some runs with
>> xfstests which had some failures but they were present with and
>> without the patch.  The best I can really offer is that it works for
>> me, ymmv.  The patch is available as commit
>> c725011c4fc5d47e12d131f61bd91a58a40036b5 in
>> https://github.com/JKDingwall/linux.git xfs-enable-cleancache or in
>> the first message of this thread.
> Hey James,
>
> I've run this patch on my local tree and it looks to work right. I am
> saying "looks" as I am hitting some other issue that I believe are
> unralted to the patch - but I need to figure them out before I can
> comfortably say: "Yes, this looks right and works for me as well."
>
> Stay tuned.
Just to add that I have also had no observable problems running this 
patch on 3.10.0 or 3.10.1.

Regards,
James