Re: [Xen-devel] [linux-3.10 test] 21490: regressions - FAIL

[Wei Liu <wei.liu2@xxxxxxxxxx>]

On Wed, Nov 06, 2013 at 10:48:37AM +0000, Ian Campbell wrote:
> On Wed, 2013-11-06 at 09:03 +0000, xen.org wrote:
> > flight 21490 linux-3.10 real [real]
> > http://www.chiark.greenend.org.uk/~xensrcts/logs/21490/
> > 
> > Regressions :-(
> > 
> > Tests which did not succeed and are blocking,
> > including tests which could not be run:
> >  test-amd64-i386-pv           18 leak-check/check          fail REGR. vs. 
> > 21026
> >  test-amd64-i386-qemut-rhel6hvm-intel 11 leak-check/check  fail REGR. vs. 
> > 21026
> >  test-amd64-amd64-pv          18 leak-check/check          fail REGR. vs. 
> > 21026
> >  test-amd64-i386-qemuu-rhel6hvm-intel 11 leak-check/check  fail REGR. vs. 
> > 21026
> >  test-amd64-amd64-xl          18 leak-check/check          fail REGR. vs. 
> > 21026
> >  test-amd64-i386-pair         20 leak-check/check/src_host fail REGR. vs. 
> > 21026
> >  test-amd64-i386-pair         21 leak-check/check/dst_host fail REGR. vs. 
> > 21026
> >  test-amd64-amd64-pair        21 leak-check/check/dst_host fail REGR. vs. 
> > 21026
> >  test-amd64-amd64-pair        20 leak-check/check/src_host fail REGR. vs. 
> > 21026
> >  test-amd64-i386-rhel6hvm-intel 11 leak-check/check        fail REGR. vs. 
> > 21026
> 
> The bisector says, in <E1Ve0Tg-0008Kj-Il@xxxxxxxxxxxxxxxxxxxxxxx>:
> 

David posted a patch to fix this, but that patch is not in that tree
yet.

<1381150519-14557-1-git-send-email-david.vrabel@xxxxxxxxxx>

From: David Vrabel <david.vrabel@xxxxxxxxxx>

If a guest is destroyed without transitioning its frontend to CLOSED,
the domain becomes a zombie as netback was not grant unmapping the
shared rings.

When removing a VIF, transition the backend to CLOSED so the VIF is
disconnected if necessary (which will unmap the shared rings etc).

This fixes a regression introduced by
279f438e36c0a70b23b86d2090aeec50155034a9 (xen-netback: Don't destroy
the netdev until the vif is shut down).

Signed-off-by: David Vrabel <david.vrabel@xxxxxxxxxx>
Cc: Ian Campbell <ian.campbell@xxxxxxxxxx>
Cc: Wei Liu <wei.liu2@xxxxxxxxxx>
Cc: Paul Durrant <Paul.Durrant@xxxxxxxxxx>

Wei.

> *** Found and reproduced problem changeset ***
> 
>   Bug is in tree:  linux 
> git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable.git
>   Bug introduced:  f495ddc46f97dd0054ff2d5d3c7493d59f7511fb
>   Bug not present: b9396c4c9e7f499b1dd8080e901c88705f2efa99
> 
> 
>   commit f495ddc46f97dd0054ff2d5d3c7493d59f7511fb
>   Author: Paul Durrant <paul.durrant@xxxxxxxxxx>
>   Date:   Tue Oct 8 14:22:56 2013 +0100
>   
>       xen-netback: Don't destroy the netdev until the vif is shut down
> 
> http://www.chiark.greenend.org.uk/~xensrcts/logs/21490/test-amd64-amd64-xl/18.ts-leak-check.log
>  shows:
> 
>         2013-11-06 04:23:59 Z executing ssh ... root@xxxxxxxxxxxx xl
>         list
>         2013-11-06 04:23:59 Z LEAKED [domain 1] domain: (null)                
>                        1     0     1     --pssd       4.6
>         2013-11-06 04:23:59 Z LEAKED [domain 2] domain: (null)                
>                        2     0     1     --pssd       0.9
>         2013-11-06 04:23:59 Z LEAKED [domain 3] domain: (null)                
>                        3     0     1     --pssd       0.5
>         2013-11-06 04:23:59 Z LEAKED [domain 4] domain: (null)                
>                        4     0     1     --pssd       0.9
>         2013-11-06 04:23:59 Z LEAKED [domain 5] domain: (null)                
>                        5     0     1     --pssd       1.1
>         2013-11-06 04:23:59 Z LEAKED [domain 6] domain: (null)                
>                        6     0     1     --pssd       0.4
>         2013-11-06 04:23:59 Z LEAKED [domain 7] domain: (null)                
>                        7     0     1     --pssd       0.4
>         2013-11-06 04:23:59 Z LEAKED [domain 8] domain: (null)                
>                        8     0     1     --pssd       0.4
>         2013-11-06 04:23:59 Z LEAKED [domain 9] domain: (null)                
>                        9     0     1     --pssd       0.4
>         2013-11-06 04:23:59 Z LEAKED [domain 10] domain: (null)               
>                        10     0     1     --pssd       0.4
>         2013-11-06 04:23:59 Z LEAKED [domain 11] domain: (null)               
>                        11     0     1     --pssd       0.4
>         2013-11-06 04:23:59 Z LEAKED [domain 12] domain: (null)               
>                        12     0     1     --pssd       0.4
>         2013-11-06 04:23:59 Z LEAKED [domain 13] domain: (null)               
>                        13     0     1     --pssd       0.4
>         2013-11-06 04:23:59 Z LEAKED [domain 14] domain: (null)               
>                        14     0     1     --pssd       0.4
>         2013-11-06 04:23:59 Z LEAKED [domain 16] domain: (null)               
>                        16     0     1     --p--d       4.7
>         
> I imagine this is because the vif isn't being properly torn down on
> shutdown... ifconfig and brctl etc logs don't show the devices living on
> in dom0, I expect it is failing to release the grant maps.
> 
> http://www.chiark.greenend.org.uk/~xensrcts/logs/21490/test-amd64-amd64-xl/serial-fire-frog.log
>  shows each domain still owns two pages. Which I would bet are the vif 
> rings... Also:
> 
> Nov  6 04:24:19.449065 (XEN) [ref] localdom mfn      pin          localdom 
> gmfn     flags
> Nov  6 04:24:19.449101 (XEN) grant-table for remote domain:    1 (v2)
> Nov  6 04:24:19.461055 (XEN) [ 10]        0 0x415f36 0x00000001          0 
> 0x415f36 0x18
> Nov  6 04:24:19.461091 (XEN) [ 11]        0 0x415f35 0x00000001          0 
> 0x415f35 0x18
> 
> Ian.

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel