Patchwork [Quantal,SRU,Pull,Request] Fix for LP#1201869

login
register
mail settings
Submitter Chris J Arges
Date Aug. 26, 2013, 2:12 p.m.
Message ID <521B6246.3090902@canonical.com>
Download mbox
Permalink /patch/269896/
State New
Headers show

Pull-request

git://kernel.ubuntu.com/arges/ubuntu-quantal.git lp1201869

Comments

Chris J Arges - Aug. 26, 2013, 2:12 p.m.
Note: this is a resubmission; it has been retested for an extended
period of time. In addition I checked upstream patches around the code
for bugfixes.

SRU Justification:

Impact:
Users of the 3.5/3.8 kernel will have poor network throughput when using
OpenStack Neutron depending on their setup.

Fix:
These upstream patches are necessary to fix the issue:

2681128f0ced8aa4e66f221197e183cc16d244fe
8093315a91340bca52549044975d8c7f673b28a1
d0e2c55e7c940a3ee91e9e23a2683b593690f1e9
2efd32ee1b60b0b31404ca47c1ce70e5a5d24ebc
f45a5c267da35174e22cec955093a7513dc1623d

Testcase:
Setup OpenStack Neutron. Test throughput between internal and external
nodes.

The following explains an example vlan+namespace configuration:
Internal Node: [10.x.x.2]->eth2.123->br123->tap123->qr-123[10.x.x.1]
<--- netns: qrouter-123
netns: qrouter-123 --->
qg-234[10.x.y.1]->tap234->br234->eth2.234->External Node[10.x.y.2]
Where:
1) tap123+qr-123 and tap234+qg-234 are veth pairs
2) qr-123 and qg-234 reside inside the qrouter-123 namespace

--

The following changes since commit 6cf308f67ee50537264b594a1c0c0b8b1fce130f:

  UBUNTU: Ubuntu-3.5.0-40.61 (2013-08-21 13:48:04 -0700)

are available in the git repository at:

  git://kernel.ubuntu.com/arges/ubuntu-quantal.git lp1201869

for you to fetch changes up to 3c57730b63fae6bb82a66d18715f184ef3cf6b5b:

  veth: fix NULL dereference in veth_dellink() (2013-08-21 16:40:14 -0500)

----------------------------------------------------------------
Eric Dumazet (5):
      veth: reduce stat overhead
      veth: extend device features
      veth: avoid a NULL deref in veth_stats_one
      veth: fix a NULL deref in netif_carrier_off
      veth: fix NULL dereference in veth_dellink()

 drivers/net/veth.c        |  177
++++++++++++++++++++++++---------------------
 include/linux/netdevice.h |    1 +
 2 files changed, 94 insertions(+), 84 deletions(-)
Tim Gardner - Aug. 26, 2013, 2:23 p.m.
On 08/26/2013 08:12 AM, Chris J Arges wrote:

> Eric Dumazet (5):
>       veth: reduce stat overhead
>       veth: extend device features
>       veth: avoid a NULL deref in veth_stats_one
>       veth: fix a NULL deref in netif_carrier_off
>       veth: fix NULL dereference in veth_dellink()
> 

The first 3 patches were recently reverted for regression. Do the
additional last 2 patches fix regressions caused by the first 3 ? I
can't tell what the original regression was because there is no
information in the revert commits.

rtg
Chris J Arges - Aug. 26, 2013, 2:35 p.m.
On 08/26/2013 09:23 AM, Tim Gardner wrote:
> On 08/26/2013 08:12 AM, Chris J Arges wrote:
> 
>> Eric Dumazet (5):
>>       veth: reduce stat overhead
>>       veth: extend device features
>>       veth: avoid a NULL deref in veth_stats_one
>>       veth: fix a NULL deref in netif_carrier_off
>>       veth: fix NULL dereference in veth_dellink()
>>
> 
> The first 3 patches were recently reverted for regression. Do the
> additional last 2 patches fix regressions caused by the first 3 ?

Yes. The original regression was a null pointer deference issue. The
last 3 patches (actually) fix these issues introduced by the first 2
patches. Originally I only SRUed the first 3, and missed the other two
corrections. In addition I checked the latest upstream code and did not
find any other fixes to those sections of the code.

> I can't tell what the original regression was because there is no
> information in the revert commits.

The original regression was the null pointer deference issues which were
found during validation of the original bug.

With these 5 patches, and a few weeks worth of testing I'm confident
this will address the original bug and not introduce additional regressions.

> 
> rtg
>
Tim Gardner - Aug. 26, 2013, 2:55 p.m.

Luis Henriques - Aug. 26, 2013, 4:35 p.m.
Chris J Arges <chris.j.arges@canonical.com> writes:

> Note: this is a resubmission; it has been retested for an extended
> period of time. In addition I checked upstream patches around the code
> for bugfixes.

They look fine to me.

Cheers,
Tim Gardner - Aug. 26, 2013, 4:41 p.m.