Re: [PATCH net v1] virtio_net: Introduce skb_vnet_common_hdr to avoid typecasting

[Willem de Bruijn]

On Tue, Aug 15, 2023 at 12:29 PM Simon Horman <horms@xxxxxxxxxx> wrote:
>
> On Tue, Aug 15, 2023 at 11:09:02AM -0400, Feng Liu wrote:
> >
> >
> > On 2023-08-15 a.m.6:51, Simon Horman wrote:
> > > External email: Use caution opening links or attachments
> > >
> > >
> > > On Mon, Aug 14, 2023 at 01:18:45PM -0400, Feng Liu wrote:
> > >
> > > + "David S. Miller" <davem@xxxxxxxxxxxxx>
> > >    Eric Dumazet <edumazet@xxxxxxxxxx>
> > >    Jakub Kicinski <kuba@xxxxxxxxxx>
> > >    Paolo Abeni <pabeni@xxxxxxxxxx>
> > >
> > Thanks for adding David S. Miller.
> >
> > > > The virtio_net driver currently deals with different versions and types
> > > > of virtio net headers, such as virtio_net_hdr_mrg_rxbuf,
> > > > virtio_net_hdr_v1_hash, etc. Due to these variations, the code relies
> > > > on multiple type casts to convert memory between different structures,
> > > > potentially leading to bugs when there are changes in these structures.
> > > >
> > > > Introduces the "struct skb_vnet_common_hdr" as a unifying header
> > > > structure using a union. With this approach, various virtio net header
> > > > structures can be converted by accessing different members of this
> > > > structure, thus eliminating the need for type casting and reducing the
> > > > risk of potential bugs.
> > > >
> > > > For example following code:
> > > > static struct sk_buff *page_to_skb(struct virtnet_info *vi,
> > > >                struct receive_queue *rq,
> > > >                struct page *page, unsigned int offset,
> > > >                unsigned int len, unsigned int truesize,
> > > >                unsigned int headroom)
> > > > {
> > > > [...]
> > > >        struct virtio_net_hdr_mrg_rxbuf *hdr;
> > > > [...]
> > > >        hdr_len = vi->hdr_len;
> > > > [...]
> > > > ok:
> > > >        hdr = skb_vnet_hdr(skb);
> > > >        memcpy(hdr, hdr_p, hdr_len);
> > > > [...]
> > > > }
> > > >
> > > > When VIRTIO_NET_F_HASH_REPORT feature is enabled, hdr_len = 20
> > > > But the sizeof(*hdr) is 12,
> > > > memcpy(hdr, hdr_p, hdr_len); will copy 20 bytes to the hdr,
> > > > which make a potential risk of bug. And this risk can be avoided by
> > > > introducing struct virtio_net_hdr_mrg_rxbuf.
> > > >
> > > > Signed-off-by: Feng Liu <feliu@xxxxxxxxxx>
> > > > Reviewed-by: Jiri Pirko <jiri@xxxxxxxxxx>
> > >
> > > I'm unsure if this is 'net' material.
> > >
> >
> > It is about the modification of the virtio_net driver. I think it should be
> > regarded as `net` material.
>
> To clarify: In general new Networking features go via the net-next tree,
> while bug fixes go via the net tree. I was suggesting this
> is more appropriate for net-next, and that should be reflected in the
> subject.
>
>         Subject: [PATCH net-next] ...
>
> Sorry for not being clearer the first time around.

Right, this should go to net-next.

>
> >
> > > > ---
> > > >   drivers/net/virtio_net.c        | 29 ++++++++++++++++-------------
> > > >   include/uapi/linux/virtio_net.h |  7 +++++++
> > > >   2 files changed, 23 insertions(+), 13 deletions(-)
> > > >
> > > > diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
> > > > index 1270c8d23463..6ce0fbcabda9 100644
> > > > --- a/drivers/net/virtio_net.c
> > > > +++ b/drivers/net/virtio_net.c
> > > > @@ -344,9 +344,10 @@ static int rxq2vq(int rxq)
> > > >        return rxq * 2;
> > > >   }
> > > >
> > > > -static inline struct virtio_net_hdr_mrg_rxbuf *skb_vnet_hdr(struct sk_buff *skb)
> > > > +static inline struct virtio_net_common_hdr *
> > > > +skb_vnet_common_hdr(struct sk_buff *skb)
> > > >   {
> > > > -     return (struct virtio_net_hdr_mrg_rxbuf *)skb->cb;
> > > > +     return (struct virtio_net_common_hdr *)skb->cb;
> > > >   }
> > > >
> > > >   /*
> > > > @@ -469,7 +470,7 @@ static struct sk_buff *page_to_skb(struct virtnet_info *vi,
> > > >                                   unsigned int headroom)
> > > >   {
> > > >        struct sk_buff *skb;
> > > > -     struct virtio_net_hdr_mrg_rxbuf *hdr;
> > > > +     struct virtio_net_common_hdr *hdr;
> > > >        unsigned int copy, hdr_len, hdr_padded_len;
> > > >        struct page *page_to_free = NULL;
> > > >        int tailroom, shinfo_size;
> > > > @@ -554,7 +555,7 @@ static struct sk_buff *page_to_skb(struct virtnet_info *vi,
> > > >                give_pages(rq, page);
> > > >
> > > >   ok:
> > > > -     hdr = skb_vnet_hdr(skb);
> > > > +     hdr = skb_vnet_common_hdr(skb);
> > > >        memcpy(hdr, hdr_p, hdr_len);
> > > >        if (page_to_free)
> > > >                put_page(page_to_free);
> > > > @@ -966,7 +967,7 @@ static struct sk_buff *receive_small_build_skb(struct virtnet_info *vi,
> > > >                return NULL;
> > > >
> > > >        buf += header_offset;
> > > > -     memcpy(skb_vnet_hdr(skb), buf, vi->hdr_len);
> > > > +     memcpy(skb_vnet_common_hdr(skb), buf, vi->hdr_len);
> > > >
> > > >        return skb;
> > > >   }
> > > > @@ -1577,7 +1578,8 @@ static void receive_buf(struct virtnet_info *vi, struct receive_queue *rq,
> > > >   {
> > > >        struct net_device *dev = vi->dev;
> > > >        struct sk_buff *skb;
> > > > -     struct virtio_net_hdr_mrg_rxbuf *hdr;
> > > > +     struct virtio_net_common_hdr *common_hdr;
> > > > +     struct virtio_net_hdr_mrg_rxbuf *mrg_hdr;
> > > >
> > > >        if (unlikely(len < vi->hdr_len + ETH_HLEN)) {
> > > >                pr_debug("%s: short packet %i\n", dev->name, len);
> > > > @@ -1597,18 +1599,19 @@ static void receive_buf(struct virtnet_info *vi, struct receive_queue *rq,
> > > >        if (unlikely(!skb))
> > > >                return;
> > > >
> > > > -     hdr = skb_vnet_hdr(skb);
> > > > +     common_hdr = skb_vnet_common_hdr(skb);
> > > >        if (dev->features & NETIF_F_RXHASH && vi->has_rss_hash_report)
> > > > -             virtio_skb_set_hash((const struct virtio_net_hdr_v1_hash *)hdr, skb);
> > > > +             virtio_skb_set_hash(&common_hdr->hash_v1_hdr, skb);
> > > >
> > > > -     if (hdr->hdr.flags & VIRTIO_NET_HDR_F_DATA_VALID)
> > > > +     mrg_hdr = &common_hdr->mrg_hdr;
> > > > +     if (mrg_hdr->hdr.flags & VIRTIO_NET_HDR_F_DATA_VALID)
> > > >                skb->ip_summed = CHECKSUM_UNNECESSARY;
> > > >
> > > > -     if (virtio_net_hdr_to_skb(skb, &hdr->hdr,
> > > > +     if (virtio_net_hdr_to_skb(skb, &mrg_hdr->hdr,
> > > >                                  virtio_is_little_endian(vi->vdev))) {
> > > >                net_warn_ratelimited("%s: bad gso: type: %u, size: %u\n",
> > > > -                                  dev->name, hdr->hdr.gso_type,
> > > > -                                  hdr->hdr.gso_size);
> > > > +                                  dev->name, mrg_hdr->hdr.gso_type,
> > > > +                                  mrg_hdr->hdr.gso_size);
> > > >                goto frame_err;
> > > >        }
> > > >
> > > > @@ -2105,7 +2108,7 @@ static int xmit_skb(struct send_queue *sq, struct sk_buff *skb)
> > > >        if (can_push)
> > > >                hdr = (struct virtio_net_hdr_mrg_rxbuf *)(skb->data - hdr_len);
> > > >        else
> > > > -             hdr = skb_vnet_hdr(skb);
> > > > +             hdr = &skb_vnet_common_hdr(skb)->mrg_hdr;
> > > >
> > > >        if (virtio_net_hdr_from_skb(skb, &hdr->hdr,
> > > >                                    virtio_is_little_endian(vi->vdev), false,
> > > > diff --git a/include/uapi/linux/virtio_net.h b/include/uapi/linux/virtio_net.h
> > > > index 12c1c9699935..db40f93ae8b3 100644
> > > > --- a/include/uapi/linux/virtio_net.h
> > > > +++ b/include/uapi/linux/virtio_net.h
> > > > @@ -201,6 +201,13 @@ struct virtio_net_hdr_mrg_rxbuf {
> > > >        struct virtio_net_hdr hdr;
> > > >        __virtio16 num_buffers; /* Number of merged rx buffers */
> > > >   };
> > > > +
> > > > +struct virtio_net_common_hdr {
> > > > +     union {
> > > > +             struct virtio_net_hdr_mrg_rxbuf mrg_hdr;
> > > > +             struct virtio_net_hdr_v1_hash hash_v1_hdr;
> > > > +     };
> > > > +};
> > >
> > > Does this belong in the UAPI?
> > > I would have assumed it's a Kernel implementation detail.
> > >
> > The existing codes, virtio_net.h is in uapi/linux/, I added the new
> > structure and followed existing code. My modification is related to Kernel
> > implementation detail now.
>
> The header you have modified forms part of the userspace API (UAPI).
> Perhaps there is something about virtio_net that makes this correct, but it
> seems to me that kernel-internal details don't belong there.

FWIW, I ran into similar issues before in a draft that added timestamp
support [1]

If we're going to change this structure, we should do it in a way that
is forward proof to future extensions to the virtio spec and with that
the fields in this struct. Especially in UAPI.

Is virtio_net_hdr_v1_hash the latest virtio-spec compliant header? And
do we expect for v1.3 to just add some fields to this?

The struct comment of virtio_net_hdr_v1 states "This is
bitwise-equivalent to the legacy struct virtio_net_hdr_mrg_rxbuf, only
flattened.". I don't quite understand what the flattening bought, vs
having struct virtio_net_hdr as first member. Another difference may
be the endianness between legacy (0.9) and v1.0+.

Since legacy virtio will no longer be modified, I don't think there is
much value is exposing this new union as UAPI. I do appreciate the
benefit to the implementation.

[1] https://patches.linaro.org/project/netdev/patch/20210208185558.995292-3-willemdebruijn.kernel@xxxxxxxxx/