On Fri, Sep 27, 2024 at 10:11 AM Akihiko Odaki [off-list ref] wrote:

On 2024/09/25 12:30, Jason Wang wrote:

quoted

On Tue, Sep 24, 2024 at 5:01 PM Akihiko Odaki [off-list ref] wrote:

quoted

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 is based on context
rewrites, which is in feature freeze. We can adopt kfuncs, but they will
not be UAPIs. We opt to ioctl to align with other relevant UAPIs (KVM
and vhost_net).

I wonder if we could clone the skb and reuse some to store the hash,
then the steering eBPF program can access these fields without
introducing full RSS in the kernel?

I don't get how cloning the skb can solve the issue.

We can certainly implement Toeplitz function in the kernel or even with
tc-bpf to store a hash value that can be used for eBPF steering program
and virtio hash reporting. However we don't have a means of storing a
hash type, which is specific to virtio hash reporting and lacks a
corresponding skb field.

I may miss something but looking at sk_filter_is_valid_access(). It
looks to me we can make use of skb->cb[0..4]?

Thanks

Regards,
Akihiko Odaki