[ovs-dev] [PATCH v2] netdev-afxdp: Best-effort configuration of XDP mode.

[William Tu]

On Tue, Nov 19, 2019 at 09:26:40PM +0100, Ilya Maximets wrote:
> On 19.11.2019 21:09, William Tu wrote:
> > On Tue, Nov 19, 2019 at 08:51:19PM +0100, Ilya Maximets wrote:
> >> On 19.11.2019 20:45, William Tu wrote:
> >>> On Tue, Nov 19, 2019 at 05:52:22PM +0100, Ilya Maximets wrote:
> >>>> On 19.11.2019 17:16, Eelco Chaudron wrote:
> >>>>>
> >>>>>
> >>>>> On 7 Nov 2019, at 12:36, Ilya Maximets wrote:
> >>>>>
> >>>>>> Until now there was only two options for XDP mode in OVS: SKB or DRV.
> >>>>>> i.e. 'generic XDP' or 'native XDP with zero-copy enabled'.
> >>>>>>
> >>>>>> Devices like 'veth' interfaces in Linux supports native XDP, but
> >>>>>> doesn't support zero-copy mode.  This case can not be covered by
> >>>>>> existing API and we have to use slower generic XDP for such devices.
> >>>>>> There are few more issues, e.g. TCP is not supported in generic XDP
> >>>>>> mode for veth interfaces due to kernel limitations, however it is
> >>>>>> supported in native mode.
> >>>>>>
> >>>>>> This change introduces ability to use native XDP without zero-copy
> >>>>>> along with best-effort configuration option that enabled by default.
> >>>>>> In best-effort case OVS will sequentially try different modes starting
> >>>>>> from the fastest one and will choose the first acceptable for current
> >>>>>> interface.  This will guarantee the best possible performance.
> >>>>>>
> >>>>>> If user will want to choose specific mode, it's still possible by
> >>>>>> setting the 'options:xdp-mode'.
> >>>>>>
> >>>>>> This change additionally changes the API by renaming the configuration
> >>>>>> knob from 'xdpmode' to 'xdp-mode' and also renaming the modes
> >>>>>> themselves to be more user-friendly.
> >>>>>>
> >>>>>> The full list of currently supported modes:
> >>>>>>   * native-with-zerocopy - former DRV
> >>>>>>   * native               - new one, DRV without zero-copy
> >>>>>>   * generic              - former SKB
> >>>>>>   * best-effort          - new one, chooses the best available from
> >>>>>>                            3 above modes
> >>>>>>
> >>>>>> Since 'best-effort' is a default mode, users will not need to
> >>>>>> explicitely set 'xdp-mode' in most cases.
> >>>>>>
> >>>>>> TCP related tests enabled back in system afxdp testsuite, because
> >>>>>> 'best-effort' will choose 'native' mode for veth interfaces
> >>>>>> and this mode has no issues with TCP.
> >>>>> Patch in general looks good, two small comments inline.
> >>>>
> >>>> Thanks for review.
> >>>>
> >>>>>
> >>>>> The only thing that bothers me is the worse performance of the TAP interface with the new default config. Can we somehow keep the old behavior for TAP interfaces?
> >>>>
> >>>> Could you check if TCP works over tap interfaces in generic mode?
> >>>> For me the point is that correctness is better than performance.
> >>>> I also hope that native implementation for tap will be improved
> >>>> over time.
> >>>>
> >>>
> >>> I agree that we should first make sure correctness.
> >>> I created a simple TCP test using br0, since br0 is a tap device.
> >>> Unfortunately it does not work...
> >>>
> >>> ---
> >>> ovs-vsctl -- add-br br0 -- set Bridge br0 datapath_type=netdev
> >>>
> >>> ip netns add at_ns0
> >>> ip link add p0 type veth peer name afxdp-p0
> >>> ip link set p0 netns at_ns0
> >>> ip link set dev afxdp-p0 up
> >>>
> >>> ovs-vsctl add-port br0 afxdp-p0
> >>> ovs-vsctl -- set interface afxdp-p0 options:n_rxq=1 type="afxdp" options:xdp-mode=native
> >>>
> >>> ip netns exec at_ns0 sh << NS_EXEC_HEREDOC
> >>> ip addr add "10.1.1.1/24" dev p0
> >>> ip link set dev p0 up
> >>> NS_EXEC_HEREDOC
> >>>
> >>> ovs-vsctl -- set interface br0 options:n_rxq=1 type="afxdp" options:xdp-mode=native
> >>
> >> I'm not sure if this is a valid thing to do.
> >> After changing the netdev type this device is no linger TAP device.
> >> Now it's a permanent tap device that detached from any application
> >> and opened by af_xdp from the other side.
> >>
> >> BTW, changing the type of "internal" port doesn't sound safe.
> > 
> > I see, thanks.
> > 
> > Then if we don't change br0 to afxdp type, the TCP test
> > below still fails, which I think should pass because
> > we are using native mode for veth now.
> > 
> > I think even native mode doesn't work for TCP?
> 
> Disable Tx offloading on both p0 and br0 and it will work.

So is it because when doing tx, af_xdp dev does not support
tcp checksum offload, so kernel is dropping tcp packet?

William