On Sat, 23 Jan 2021 02:35:41 +0100
Daniel Borkmann <daniel@xxxxxxxxxxxxx> wrote:
> > + * The *flags* argument can be a combination of one or more of the
> > + * following values:
> > + *
> > + * **BPF_MTU_CHK_SEGS**
> > + * This flag will only works for *ctx* **struct sk_buff**.
> > + * If packet context contains extra packet segment buffers
> > + * (often knows as GSO skb), then MTU check is harder to
> > + * check at this point, because in transmit path it is
> > + * possible for the skb packet to get re-segmented
> > + * (depending on net device features). This could still be
> > + * a MTU violation, so this flag enables performing MTU
> > + * check against segments, with a different violation
> > + * return code to tell it apart. Check cannot use len_diff.
> > + *
> > + * On return *mtu_len* pointer contains the MTU value of the net
> > + * device. Remember the net device configured MTU is the L3 size,
> > + * which is returned here and XDP and TX length operate at L2.
> > + * Helper take this into account for you, but remember when using
> > + * MTU value in your BPF-code. On input *mtu_len* must be a valid
> > + * pointer and be initialized (to zero), else verifier will reject
> > + * BPF program.
> > + *
> > + * Return
> > + * * 0 on success, and populate MTU value in *mtu_len* pointer.
> > + *
> > + * * < 0 if any input argument is invalid (*mtu_len* not updated)
> > + *
> > + * MTU violations return positive values, but also populate MTU
> > + * value in *mtu_len* pointer, as this can be needed for
> > + * implementing PMTU handing:
> > + *
> > + * * **BPF_MTU_CHK_RET_FRAG_NEEDED**
> > + * * **BPF_MTU_CHK_RET_SEGS_TOOBIG**
> > + *
> > */
> [...]
> > +BPF_CALL_5(bpf_skb_check_mtu, struct sk_buff *, skb,
> > + u32, ifindex, u32 *, mtu_len, s32, len_diff, u64, flags)
> > +{
> > + int ret = BPF_MTU_CHK_RET_FRAG_NEEDED;
> > + struct net_device *dev = skb->dev;
> > + int skb_len, dev_len;
> > + int mtu;
> > +
> > + if (unlikely(flags & ~(BPF_MTU_CHK_SEGS)))
> > + return -EINVAL;
> > +
> > + dev = __dev_via_ifindex(dev, ifindex);
> > + if (unlikely(!dev))
> > + return -ENODEV;
> > +
> > + mtu = READ_ONCE(dev->mtu);
> > +
> > + dev_len = mtu + dev->hard_header_len;
> > + skb_len = skb->len + len_diff; /* minus result pass check */
> > + if (skb_len <= dev_len) {
> > + ret = BPF_MTU_CHK_RET_SUCCESS;
> > + goto out;
> > + }
> > + /* At this point, skb->len exceed MTU, but as it include length of all
> > + * segments, it can still be below MTU. The SKB can possibly get
> > + * re-segmented in transmit path (see validate_xmit_skb). Thus, user
> > + * must choose if segs are to be MTU checked.
> > + */
> > + if (skb_is_gso(skb)) {
> > + ret = BPF_MTU_CHK_RET_SUCCESS;
> > +
> > + if (flags & BPF_MTU_CHK_SEGS &&
> > + !skb_gso_validate_network_len(skb, mtu))
> > + ret = BPF_MTU_CHK_RET_SEGS_TOOBIG;
>
> I think that looks okay overall now. One thing that will easily slip through
> is that in the helper description you mentioned 'Check cannot use len_diff.'
> for BPF_MTU_CHK_SEGS flag. So right now for non-zero len_diff the user
> will still get BPF_MTU_CHK_RET_SUCCESS if the current length check via
> skb_gso_validate_network_len(skb, mtu) passes. If it cannot be checked,
> maybe enforce len_diff == 0 for gso skbs on BPF_MTU_CHK_SEGS?
Ok. Do you want/think this can be enforced by the verifier or are you
simply requesting that the helper will return -EINVAL (or another errno)?
--
Best regards,
Jesper Dangaard Brouer
MSc.CS, Principal Kernel Engineer at Red Hat
LinkedIn: http://www.linkedin.com/in/brouer
