Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::3:4 as permitted sender) client-ip=2620:137:e000::3:4;
MIME-Version: 1.0
References: <20231008052101.144422-1-akihiko.odaki@daynix.com>
 <20231008052101.144422-6-akihiko.odaki@daynix.com> <CAF=yD-K2MQt4nnfwJrx6h6Nii_rho7j1o6nb_jYaSwcWY45pPw@mail.gmail.com>
 <48e20be1-b658-4117-8856-89ff1df6f48f@daynix.com> <CAF=yD-K4bCBpUVtDR_cv=bagRL+vM4Rusez+uHFTb4_kR8XkpA@mail.gmail.com>
 <6a698c99-6f02-4cfb-a709-ba02296a05f7@daynix.com>
In-Reply-To: <6a698c99-6f02-4cfb-a709-ba02296a05f7@daynix.com>
From:   Willem de Bruijn <willemdebruijn.kernel@gmail.com>
Date:   Mon, 9 Oct 2023 03:07:27 -0700
Message-ID: <CAF=yD-+WFy8us0wUWo-0KpZUKHx2Q82cJ8teO0qRkK-_R1e0cA@mail.gmail.com>
Subject: Re: [RFC PATCH 5/7] tun: Introduce virtio-net hashing feature
To:     Akihiko Odaki <akihiko.odaki@daynix.com>
Cc:     Jason Wang <jasowang@redhat.com>,
        "Michael S. Tsirkin" <mst@redhat.com>, netdev@vger.kernel.org,
        linux-kernel@vger.kernel.org, kvm@vger.kernel.org,
        virtualization@lists.linux-foundation.org,
        linux-kselftest@vger.kernel.org, bpf@vger.kernel.org,
        davem@davemloft.net, kuba@kernel.org, ast@kernel.org,
        daniel@iogearbox.net, andrii@kernel.org, kafai@fb.com,
        songliubraving@fb.com, yhs@fb.com, john.fastabend@gmail.com,
        kpsingh@kernel.org, rdunlap@infradead.org, willemb@google.com,
        gustavoars@kernel.org, herbert@gondor.apana.org.au,
        steffen.klassert@secunet.com, nogikh@google.com,
        pablo@netfilter.org, decui@microsoft.com, cai@lca.pw,
        jakub@cloudflare.com, elver@google.com, pabeni@redhat.com,
        Yuri Benditovich <yuri.benditovich@daynix.com>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Mon, Oct 9, 2023 at 3:05=E2=80=AFAM Akihiko Odaki <akihiko.odaki@daynix.=
com> wrote:
>
>
>
> On 2023/10/09 18:54, Willem de Bruijn wrote:
> > On Mon, Oct 9, 2023 at 3:44=E2=80=AFAM Akihiko Odaki <akihiko.odaki@day=
nix.com> wrote:
> >>
> >> On 2023/10/09 17:13, Willem de Bruijn wrote:
> >>> On Sun, Oct 8, 2023 at 12:22=E2=80=AFAM Akihiko Odaki <akihiko.odaki@=
daynix.com> 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 overc=
ome
> >>>> thse challenges. An alternative solution is to extend the eBPF steer=
ing
> >>>> program so that it will be able to report to the userspace, but it m=
akes
> >>>> 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.
> >>>>
> >>>> The hash value already stored in sk_buff is not used and computed
> >>>> independently since it may have been computed in a way not conforman=
t
> >>>> with the specification.
> >>>>
> >>>> Signed-off-by: Akihiko Odaki <akihiko.odaki@daynix.com>
> >>>
> >>>> @@ -2116,31 +2172,49 @@ static ssize_t tun_put_user(struct tun_struc=
t *tun,
> >>>>           }
> >>>>
> >>>>           if (vnet_hdr_sz) {
> >>>> -               struct virtio_net_hdr gso;
> >>>> +               union {
> >>>> +                       struct virtio_net_hdr hdr;
> >>>> +                       struct virtio_net_hdr_v1_hash v1_hash_hdr;
> >>>> +               } hdr;
> >>>> +               int ret;
> >>>>
> >>>>                   if (iov_iter_count(iter) < vnet_hdr_sz)
> >>>>                           return -EINVAL;
> >>>>
> >>>> -               if (virtio_net_hdr_from_skb(skb, &gso,
> >>>> -                                           tun_is_little_endian(tun=
), true,
> >>>> -                                           vlan_hlen)) {
> >>>> +               if ((READ_ONCE(tun->vnet_hash.flags) & TUN_VNET_HASH=
_REPORT) &&
> >>>> +                   vnet_hdr_sz >=3D sizeof(hdr.v1_hash_hdr) &&
> >>>> +                   skb->tun_vnet_hash) {
> >>>
> >>> Isn't vnet_hdr_sz guaranteed to be >=3D hdr.v1_hash_hdr, by virtue of
> >>> the set hash ioctl failing otherwise?
> >>>
> >>> Such checks should be limited to control path where possible
> >>
> >> There is a potential race since tun->vnet_hash.flags and vnet_hdr_sz a=
re
> >> not read at once.
> >
> > It should not be possible to downgrade the hdr_sz once v1 is selected.
>
> I see nothing that prevents shrinking the header size.
>
> tun->vnet_hash.flags is read after vnet_hdr_sz so the race can happen
> even for the case the header size grows though this can be fixed by
> reordering the two reads.

One option is to fail any control path that tries to re-negotiate
header size once this hash option is enabled?

There is no practical reason to allow feature re-negotiation at any
arbitrary time.