[RFC PATCH 0/4] Added virtio-net RSS with eBPF support.

Peter Krempa pkrempa at redhat.com
Mon Oct 9 08:22:05 UTC 2023


On Mon, Oct 09, 2023 at 10:15:34 +0200, Peter Krempa wrote:
> On Mon, Oct 09, 2023 at 09:16:10 +0300, Andrew Melnychenko wrote:
> > This series of rfc patches adds support for loading the RSS eBPF program and passing it to the QEMU.
> > Comments and suggestions would be useful.
> > 
> > QEMU with vhost may work with RSS through eBPF. To load eBPF,
> > the capabilities required that Libvirt may provide.
> > eBPF program and maps may be unique for particular QEMU and
> > Libvirt retrieves eBPF through qapi.
> > For now, there is only "RSS" eBPF object in QEMU, in the future,
> > there may be another one(g.e. network filters).
> > That's why in Libvirt added logic to load and store any
> > eBPF object that QEMU provides using qapi schema.
> > 
> > For virtio-net RSS, the document has not changed.
> > ```
> >  <interface type="network">
> >   <model type="virtio"/>
> >   <driver queues="4" rss="on" rss_hash_report="off"/>
> >  <interface type="network">
> > ```
> > 
> > Simplified routine for RSS:
> >  * Libvirt retrieves eBPF "RSS" and load it.
> >  * Libvirt passes file descriptors to virtio-net with property "ebpf_rss_fds" ("rss" property should be "on" too).
> >  * if fds was provided - QEMU using eBPF RSS implementation.
> >  * if fds was not provided - QEMU tries to load eBPF RSS in own context and use it.
> >  * if eBPF RSS was not loaded - QEMU uses "in-qemu" RSS(vhost not supported).
> 
> Hi,
> 

Also what is the status of the qemu patches, specifically the QMP
interface? I see that the patch was in v1 of Jasons 'Net patches' pull
request in September, but was dropped in v2 for some reason.

Specifically our policy is that a feature depending on new qemu
interfaces can be merged to libvirt only after the code was comitted to
qemu.

We do not require that it is released, in fact we follow the development
cycle with the the "qemu capabilities" updates done throughout the the
dev cycle.

Once the qemu code is merged I can do the required update of the
capability test output (tests/qemucapabilitiesdata/...replies) output
file based on your patch so that churn and your manual input is
minimized.


More information about the libvir-list mailing list