Received-SPF: pass (google.com: domain of linux-kernel+bounces-134845-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45d1:ec00::1 as permitted sender) client-ip=2604:1380:45d1:ec00::1;
Precedence: bulk
MIME-Version: 1.0
References: <20240202121151.65710-1-liangchen.linux@gmail.com>
 <c8d59e75-d0bb-4a03-9ef4-d6de65fa9356@kernel.org> <CAKhg4tJFpG5nUNdeEbXFLonKkFUP0QCh8A9CpwU5OvtnBuz4Sw@mail.gmail.com>
 <5297dad6499f6d00f7229e8cf2c08e0eacb67e0c.camel@redhat.com>
 <CAKhg4tLbF8SfYD4dU9U9Nhii4FY2dftjPKYz-Emrn-CRwo10mg@mail.gmail.com>
 <73c242b43513bde04eebb4eb581deb189443c26b.camel@redhat.com>
 <CAKhg4tJPjcShkw4-FHFkKOcgzHK27A5pMu9FP7OWj4qJUX1ApA@mail.gmail.com>
 <1b2d471a5d06ecadcb75e3d9155b6d566afb2767.camel@redhat.com>
 <1708652254.1517398-1-xuanzhuo@linux.alibaba.com> <CACGkMEuUeQTJYpZDx8ggqwBWULQS1Fjd_DgPvVMLq-_cjYfm7g@mail.gmail.com>
 <65dcf7a775437_20e0a2087f@john.notmuch> <CAKhg4t+dzRPjyRXAifS_TCGPv3SfMMm1CF3pCs18OR+o9v+S_Q@mail.gmail.com>
 <CAKhg4tLO7vG5jEYZ3xnzm=xKDHO0SNgDw=JT-j7gb5bjiQOqsw@mail.gmail.com>
In-Reply-To: <CAKhg4tLO7vG5jEYZ3xnzm=xKDHO0SNgDw=JT-j7gb5bjiQOqsw@mail.gmail.com>
From: Jason Wang <jasowang@redhat.com>
Date: Mon, 8 Apr 2024 14:41:21 +0800
Message-ID: <CACGkMEsfW-j=W8fdTofUsF2a9rRujCs8TH6wkFvZkpLc68ZEEg@mail.gmail.com>
Subject: Re: [PATCH net-next v5] virtio_net: Support RX hash XDP hint
To: Liang Chen <liangchen.linux@gmail.com>
Cc: John Fastabend <john.fastabend@gmail.com>, Xuan Zhuo <xuanzhuo@linux.alibaba.com>, 
	Paolo Abeni <pabeni@redhat.com>, Jesper Dangaard Brouer <hawk@kernel.org>, mst@redhat.com, 
	hengqi@linux.alibaba.com, davem@davemloft.net, edumazet@google.com, 
	kuba@kernel.org, netdev@vger.kernel.org, virtualization@lists.linux.dev, 
	linux-kernel@vger.kernel.org, bpf@vger.kernel.org, daniel@iogearbox.net, 
	ast@kernel.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

On Mon, Apr 1, 2024 at 11:38=E2=80=AFAM Liang Chen <liangchen.linux@gmail.c=
om> wrote:
>
> On Thu, Feb 29, 2024 at 4:37=E2=80=AFPM Liang Chen <liangchen.linux@gmail=
com> wrote:
> >
> > On Tue, Feb 27, 2024 at 4:42=E2=80=AFAM John Fastabend <john.fastabend@=
gmail.com> wrote:
> > >
> > > Jason Wang wrote:
> > > > On Fri, Feb 23, 2024 at 9:42=E2=80=AFAM Xuan Zhuo <xuanzhuo@linux.a=
libaba.com> wrote:
> > > > >
> > > > > On Fri, 09 Feb 2024 13:57:25 +0100, Paolo Abeni <pabeni@redhat.co=
m> wrote:
> > > > > > On Fri, 2024-02-09 at 18:39 +0800, Liang Chen wrote:
> > > > > > > On Wed, Feb 7, 2024 at 10:27=E2=80=AFPM Paolo Abeni <pabeni@r=
edhat.com> wrote:
> > > > > > > >
> > > > > > > > On Wed, 2024-02-07 at 10:54 +0800, Liang Chen wrote:
> > > > > > > > > On Tue, Feb 6, 2024 at 6:44=E2=80=AFPM Paolo Abeni <paben=
i@redhat.com> wrote:
> > > > > > > > > >
> > > > > > > > > > On Sat, 2024-02-03 at 10:56 +0800, Liang Chen wrote:
> > > > > > > > > > > On Sat, Feb 3, 2024 at 12:20=E2=80=AFAM Jesper Dangaa=
rd Brouer <hawk@kernel.org> wrote:
> > > > > > > > > > > > On 02/02/2024 13.11, Liang Chen wrote:
> > > > > > > > > > [...]
> > > > > > > > > > > > > @@ -1033,6 +1039,16 @@ static void put_xdp_frags(=
struct xdp_buff *xdp)
> > > > > > > > > > > > >       }
> > > > > > > > > > > > >   }
> > > > > > > > > > > > >
> > > > > > > > > > > > > +static void virtnet_xdp_save_rx_hash(struct virt=
net_xdp_buff *virtnet_xdp,
> > > > > > > > > > > > > +                                  struct net_dev=
ice *dev,
> > > > > > > > > > > > > +                                  struct virtio_=
net_hdr_v1_hash *hdr_hash)
> > > > > > > > > > > > > +{
> > > > > > > > > > > > > +     if (dev->features & NETIF_F_RXHASH) {
> > > > > > > > > > > > > +             virtnet_xdp->hash_value =3D hdr_has=
h->hash_value;
> > > > > > > > > > > > > +             virtnet_xdp->hash_report =3D hdr_ha=
sh->hash_report;
> > > > > > > > > > > > > +     }
> > > > > > > > > > > > > +}
> > > > > > > > > > > > > +
> > > > > > > > > > > >
> > > > > > > > > > > > Would it be possible to store a pointer to hdr_hash=
 in virtnet_xdp_buff,
> > > > > > > > > > > > with the purpose of delaying extracting this, until=
 and only if XDP
> > > > > > > > > > > > bpf_prog calls the kfunc?
> > > > > > > > > > > >
> > > > > > > > > > >
> > > > > > > > > > > That seems to be the way v1 works,
> > > > > > > > > > > https://lore.kernel.org/all/20240122102256.261374-1-l=
iangchen.linux@gmail.com/
> > > > > > > > > > > . But it was pointed out that the inline header may b=
e overwritten by
> > > > > > > > > > > the xdp prog, so the hash is copied out to maintain i=
ts integrity.
> > > > > > > > > >
> > > > > > > > > > Why? isn't XDP supposed to get write access only to the=
 pkt
> > > > > > > > > > contents/buffer?
> > > > > > > > > >
> > > > > > > > >
> > > > > > > > > Normally, an XDP program accesses only the packet data. H=
owever,
> > > > > > > > > there's also an XDP RX Metadata area, referenced by the d=
ata_meta
> > > > > > > > > pointer. This pointer can be adjusted with bpf_xdp_adjust=
_meta to
> > > > > > > > > point somewhere ahead of the data buffer, thereby grantin=
g the XDP
> > > > > > > > > program access to the virtio header located immediately b=
efore the
> > > > > > > >
> > > > > > > > AFAICS bpf_xdp_adjust_meta() does not allow moving the meta=
_data before
> > > > > > > > xdp->data_hard_start:
> > > > > > > >
> > > > > > > > https://elixir.bootlin.com/linux/latest/source/net/core/fil=
ter.c#L4210
> > > > > > > >
> > > > > > > > and virtio net set such field after the virtio_net_hdr:
> > > > > > > >
> > > > > > > > https://elixir.bootlin.com/linux/latest/source/drivers/net/=
virtio_net.c#L1218
> > > > > > > > https://elixir.bootlin.com/linux/latest/source/drivers/net/=
virtio_net.c#L1420
> > > > > > > >
> > > > > > > > I don't see how the virtio hdr could be touched? Possibly e=
ven more
> > > > > > > > important: if such thing is possible, I think is should be =
somewhat
> > > > > > > > denied (for the same reason an H/W nic should prevent XDP f=
rom
> > > > > > > > modifying its own buffer descriptor).
> > > > > > >
> > > > > > > Thank you for highlighting this concern. The header layout di=
ffers
> > > > > > > slightly between small and mergeable mode. Taking 'mergeable =
mode' as
> > > > > > > an example, after calling xdp_prepare_buff the layout of xdp_=
buff
> > > > > > > would be as depicted in the diagram below,
> > > > > > >
> > > > > > >                       buf
> > > > > > >                        |
> > > > > > >                        v
> > > > > > >         +--------------+--------------+-------------+
> > > > > > >         | xdp headroom | virtio header| packet      |
> > > > > > >         | (256 bytes)  | (20 bytes)   | content     |
> > > > > > >         +--------------+--------------+-------------+
> > > > > > >         ^                             ^
> > > > > > >         |                             |
> > > > > > >  data_hard_start                    data
> > > > > > >                                   data_meta
> > > > > > >
> > > > > > > If 'bpf_xdp_adjust_meta' repositions the 'data_meta' pointer =
a little
> > > > > > > towards 'data_hard_start', it would point to the inline heade=
r, thus
> > > > > > > potentially allowing the XDP program to access the inline hea=
der.
> > >
> > > Fairly late to the thread sorry. Given above layout does it make sens=
e to
> > > just delay extraction to the kfunc as suggested above? Sure the XDP p=
rogram
> > > could smash the entry in virtio header, but this is already the case =
for
> > > anything else there. A program writing over the virtio header is like=
ly
> > > buggy anyways. Worse that might happen is bad rss values and mappings=
?
> >
> > Thank you for raising the concern. I am not quite sure if the XDP
> > program is considered buggy, as it is agnostic to the layout of the
> > inline header.
> > Let's say an XDP program calls bpf_xdp_adjust_meta to adjust data_meta
> > to point to the inline header and overwrites it without even knowing
> > of its existence. Later, when the XDP program invokes the kfunc to
> > retrieve the hash, incorrect data would be returned. In this case, the
> > XDP program seems to be doing everything legally but ends up with the
> > wrong hash data.
> >
> > Thanks,
> > Liang
> >
>
> I haven=E2=80=99t received any feedback yet, so I=E2=80=99m under the imp=
ression that
> the XDP program is still considered buggy in the scenario mentioned
> above, and the overall behavior is as designed from XDP perspective.
> Looking up the intel igc driver, it also does not bother with this
> particular aspect.

So let's post a new version with all the detailed explanations as above and=
 see?

>
> Given this context, we don't need to be concerned about the hash value
> being overwritten. So if there aren't any objections, I plan to remove
> the preservation of the hash value in the next iteration.
>
> Thanks,
> Liang

Thanks

>
> > >
> > > I like seeing more use cases for the hints though.
> > >
> > > Thanks!
> > > John
>