Re: [PATCH net-next 3/9] selftests/net: Add env for container based tests

[David Wei <dw@davidwei.uk>]

On 2026-02-24 10:10, Bobby Eshleman wrote:
> On Mon, Feb 23, 2026 at 04:24:03PM -0800, Wei Wang wrote:
>> From: David Wei <dw@davidwei.uk>
>>
>> Add an env NetDrvContEnv for container based selftests. This automates
>> the setup of a netns, netkit pair with one inside the netns, and a BPF
>> program that forwards skbs from the NETIF host inside the container.
>>
>> Currently only netkit is used, but other virtual netdevs e.g. veth can
>> be used too.
>>
>> Expect netkit container datapath selftests to have a publicly routable
>> IP prefix to assign to netkit in a container, such that packets will
>> land on eth0. The BPF skb forward program will then forward such packets
>> from the host netns to the container netns.
>>
>> Signed-off-by: David Wei <dw@davidwei.uk>
>> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
>> ---
>>   .../testing/selftests/drivers/net/README.rst  |  19 ++
>>   .../drivers/net/hw/lib/py/__init__.py         |   7 +-
>>   .../selftests/drivers/net/lib/py/__init__.py  |   7 +-
>>   .../selftests/drivers/net/lib/py/env.py       | 163 ++++++++++++++++++
>>   4 files changed, 190 insertions(+), 6 deletions(-)
>>
>> diff --git a/tools/testing/selftests/drivers/net/README.rst b/tools/testing/selftests/drivers/net/README.rst
>> index eb838ae94844..39370a83f238 100644
>> --- a/tools/testing/selftests/drivers/net/README.rst
>> +++ b/tools/testing/selftests/drivers/net/README.rst
>> @@ -62,6 +62,25 @@ LOCAL_V4, LOCAL_V6, REMOTE_V4, REMOTE_V6
>>   
>>   Local and remote endpoint IP addresses.
>>   
>> +LOCAL_PREFIX_V4, LOCAL_PREFIX_V6
>> +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> 
> Still not sure we want LOCAL_PREFIX_V4 here, since it is not supported?

Thanks, I'll remove LOCAL_PREFIX_V4.

> 
>> +
>> +Local IP prefix/subnet which can be used to allocate extra IP addresses (for
>> +network name spaces behind macvlan, veth, netkit devices). DUT must be
>> +reachable using these addresses from the endpoint.
>> +
>> +  +-------------+        +----------------------------+
>> +  | INIT NS     |        | TEST NS                    |
>> +  | +---------+ |        | +------------------------+ |
>> +  | | NETIF   | |  bpf   | | Netkit                 | |
>> +  | |         |-|--------|>| nk_guest               | |
>> +  | +---------+ |        | | {LOCAL_PREFIX_V6}::2:2 | |
>> +  | +---------+ |        | +------------------------+ |
>> +  | | Netkit  | |        +----------------------------+
>> +  | | nk_host | |
>> +  | +---------+ |
>> +  +-------------+
>> +
> 
> One thing I noticed when I tested this was the flow broke down when
> NETIF's address happened to match LOCAL_PREFIX_V6. I didn't fully reason
> through why this was the case, but maybe it's worth mentioning in the
> docs here?

If NETIF's address matches LOCAL_PREFIX_V6, the installed bpf prog is
going to forward _all_ packets to Netkit in the netns. That's likely why
it doesn't work.

I'll mention that NETIF must not match LOCAL_PREFIX_V6 in the kdocs and
the diagram, then assert that in NetDrvContEnv as well. Thanks.

> 
>>   REMOTE_TYPE
>>   ~~~~~~~~~~~
>>   
[...]
>> diff --git a/tools/testing/selftests/drivers/net/lib/py/env.py b/tools/testing/selftests/drivers/net/lib/py/env.py
>> index 41cc248ac848..857ae0f37516 100644
>> --- a/tools/testing/selftests/drivers/net/lib/py/env.py
>> +++ b/tools/testing/selftests/drivers/net/lib/py/env.py
[...]
>> @@ -289,3 +292,163 @@ class NetDrvEpEnv(NetDrvEnvBase):
>>                   data.get('stats-block-usecs', 0) / 1000 / 1000
>>   
>>           time.sleep(self._stats_settle_time)
>> +
>> +
>> +class NetDrvContEnv(NetDrvEpEnv):
>> +    """
>> +    Class for an environment with a netkit pair setup for forwarding traffic
>> +    between the physical interface and a network namespace.
>> +    +-------------+        +----------------------------+
>> +    | INIT NS     |        | TEST NS                    |
>> +    | +---------+ |        | +------------------------+ |
>> +    | | NETIF   | |  bpf   | | Netkit                 | |
>> +    | |         |-|--------|>| nk_guest               | |
>> +    | +---------+ |        | | {LOCAL_PREFIX_V6}::2:2 | |
>> +    | +---------+ |        | +------------------------+ |
>> +    | | Netkit  | |        +----------------------------+
>> +    | | nk_host | |
>> +    | +---------+ |
>> +    +-------------+
>> +    """
>> +
>> +    def __init__(self, src_path, rxqueues=1, **kwargs):
>> +        super().__init__(src_path, **kwargs)
>> +
>> +        self.netns = None
>> +        self._nk_host_ifname = None
>> +        self._nk_guest_ifname = None
>> +        self._tc_clsact_added = False
>> +        self._tc_attached = False
>> +        self._bpf_prog_pref = None
>> +        self._bpf_prog_id = None
>> +        self._init_ns_attached = False
> 
> One problem with these attrs is that if super().__init__() fails
> (NetDrvEpEnv.__init__()) then these will raise attribute errors in
> __del__().
> 
> This can happen if NetDrvEpEnv.remote_remote_ifc() (for example) throws
> an exception because it couldn't find the address the user configured.
> 
> One alternative to getattr() is moving these assignments before
> super().__init__()?

Sounds good, I'll move it above super().__init__().

> 
> Best,
> Bobby
>