El 13/02/15 a les 0.46, Andrew Thompson ha escrit:
> On 11 February 2015 at 05:22, Roger Pau Monné <roger....@citrix.com> wrote:
> 
>> Hello Andrew,
>>
>> El 10/02/15 a les 10.22, Andrew Thompson ha escrit:
>>> On 10 February 2015 at 16:25, Andrew Thompson <thom...@freebsd.org>
>> wrote:
>>>
>>>> On 9 February 2015 at 23:28, Roger Pau Monné <roger....@citrix.com>
>> wrote:
>>>> A bit of trial and error with dtrace has narrowed this down. I can cause
>>>> the leak by just opening /dev/xen/xenstore
>>
>> Thanks for doing all this legwork! IMHO the best solution is to switch
>> xenstore dev to use cdevpriv in order to store each client data. What we
>> are doing right now (storing client data in dev->si_dvr1) is plain
>> wrong. I've uploaded two patches (one for HEAD and one for stable/10) so
>> that you can try it also, please report back whether this fixes your
>> problem or not:
>>
>> https://people.freebsd.org/~royger/xenstore_fix/
>>
>>
> I have tested this on 10.0 and it does fix the issue. After 36000
> allocations there are just 6 active at 13K
> 
> xenstore     6    13K       -    36699  16,32,64,128,256,512
> 
> 
> I looked in to why this only affected one of my many VMs as they run
> xe-update-guest-attrs and nova-agent which should have triggered the leak.
> They are all stuck on xsdread when reading from the xenstore so the update
> loop has stopped.
> 
>   UID   PID  PPID CPU PRI NI    VSZ   RSS MWCHAN   STAT TT          TIME
> COMMAND
>     0 13250 13249   0  20  0  16552  1968 xsdread  S     -       5:57.60
> /usr/local/bin/xenstore-read domid (xenstore)

Is this with my patch applied? AFAICT xsdread means it's waiting on the
address at dev->si_drv1, which is completely wrong because every time
the device was opened this address changed. With my patch applied it
should no longer be an issue.

Roger.
_______________________________________________
freebsd-xen@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-xen
To unsubscribe, send any mail to "freebsd-xen-unsubscr...@freebsd.org"

Reply via email to