[RFC,0/7] tun: Introduce virtio-net hashing feature

Message

Akihiko Odaki Oct. 8, 2023, 5:20 a.m. UTC

virtio-net have two usage of hashes: one is RSS and another is hash
reporting. Conventionally the hash calculation was done by the VMM.
However, computing the hash after the queue was chosen defeats the
purpose of RSS.

Another approach is to use eBPF steering program. This approach has
another downside: it cannot report the calculated hash due to the
restrictive nature of eBPF.

Introduce the code to compute hashes to the kernel in order to overcome
thse challenges.

An alternative solution is to extend the eBPF steering program so that it
will be able to report to the userspace, but it makes little sense to
allow to implement different hashing algorithms with eBPF since the hash
value reported by virtio-net is strictly defined by the specification.
An implementation of this alternative solution was submitted as RFC
patches in the past:
https://lore.kernel.org/lkml/[e-mail]/

QEMU patched to use this new feature is available at:
https://github.com/daynix/qemu/tree/akihikodaki/rss

The QEMU patches will soon be submitted to the upstream as RFC too.

Akihiko Odaki (7):
  net: skbuff: Add tun_vnet_hash flag
  net/core: Ensure qdisc_skb_cb will not be overwritten
  net: sched: Add members to qdisc_skb_cb
  virtio_net: Add functions for hashing
  tun: Introduce virtio-net hashing feature
  selftest: tun: Add tests for virtio-net hashing
  vhost_net: Support VIRTIO_NET_F_HASH_REPORT

 drivers/net/tun.c                    | 187 ++++++++-
 drivers/vhost/net.c                  |  16 +-
 include/linux/skbuff.h               |   2 +
 include/linux/virtio_net.h           | 157 ++++++++
 include/net/sch_generic.h            |  12 +-
 include/uapi/linux/if_tun.h          |  16 +
 net/core/dev.c                       |   1 +
 tools/testing/selftests/net/Makefile |   2 +-
 tools/testing/selftests/net/tun.c    | 578 ++++++++++++++++++++++++++-
 9 files changed, 933 insertions(+), 38 deletions(-)

Comments

Willem de Bruijn Oct. 8, 2023, 6:36 p.m. UTC | #1

On Sun, Oct 8, 2023 at 7:21 AM Akihiko Odaki <[e-mail]> wrote:
>
> virtio-net have two usage of hashes: one is RSS and another is hash
> reporting. Conventionally the hash calculation was done by the VMM.
> However, computing the hash after the queue was chosen defeats the
> purpose of RSS.
>
> Another approach is to use eBPF steering program. This approach has
> another downside: it cannot report the calculated hash due to the
> restrictive nature of eBPF.
>
> Introduce the code to compute hashes to the kernel in order to overcome
> thse challenges.
>
> An alternative solution is to extend the eBPF steering program so that it
> will be able to report to the userspace, but it makes little sense to
> allow to implement different hashing algorithms with eBPF since the hash
> value reported by virtio-net is strictly defined by the specification.

But using the existing BPF steering may have the benefit of requiring
a lot less new code.

There is ample precedence for BPF programs that work this way. The
flow dissector comes to mind.