On 3/22/24 8:02 AM, Anton Protopopov wrote:
> Extend the bpf_fib_lookup() helper by making it to utilize mark if
> the BPF_FIB_LOOKUP_MARK flag is set. In order to pass the mark the
> four bytes of struct bpf_fib_lookup are used, shared with the
> output-only smac/dmac fields.
>
> Signed-off-by: Anton Protopopov <aspsk@xxxxxxxxxxxxx>
> Acked-by: Daniel Borkmann <daniel@xxxxxxxxxxxxx>
> ---
> include/uapi/linux/bpf.h | 20 ++++++++++++++++++--
> net/core/filter.c | 12 +++++++++---
> tools/include/uapi/linux/bpf.h | 20 ++++++++++++++++++--
> 3 files changed, 45 insertions(+), 7 deletions(-)
>
> diff --git a/include/uapi/linux/bpf.h b/include/uapi/linux/bpf.h
> index 9585f5345353..96d57e483133 100644
> --- a/include/uapi/linux/bpf.h
> +++ b/include/uapi/linux/bpf.h
> @@ -3394,6 +3394,10 @@ union bpf_attr {
> * for the nexthop. If the src addr cannot be derived,
> * **BPF_FIB_LKUP_RET_NO_SRC_ADDR** is returned. In this
> * case, *params*->dmac and *params*->smac are not set either.
> + * **BPF_FIB_LOOKUP_MARK**
> + * Use the mark present in *params*->mark for the fib lookup.
> + * This option should not be used with BPF_FIB_LOOKUP_DIRECT,
> + * as it only has meaning for full lookups.
> *
> * *ctx* is either **struct xdp_md** for XDP programs or
> * **struct sk_buff** tc cls_act programs.
> @@ -7120,6 +7124,7 @@ enum {
> BPF_FIB_LOOKUP_SKIP_NEIGH = (1U << 2),
> BPF_FIB_LOOKUP_TBID = (1U << 3),
> BPF_FIB_LOOKUP_SRC = (1U << 4),
> + BPF_FIB_LOOKUP_MARK = (1U << 5),
> };
>
> enum {
> @@ -7197,8 +7202,19 @@ struct bpf_fib_lookup {
> __u32 tbid;
> };
>
> - __u8 smac[6]; /* ETH_ALEN */
> - __u8 dmac[6]; /* ETH_ALEN */
> + union {
> + /* input */
> + struct {
> + __u32 mark; /* policy routing */
> + /* 2 4-byte holes for input */
> + };
> +
> + /* output: source and dest mac */
> + struct {
> + __u8 smac[6]; /* ETH_ALEN */
> + __u8 dmac[6]; /* ETH_ALEN */
> + };
> + };
> };
>
> struct bpf_redir_neigh {
> diff --git a/net/core/filter.c b/net/core/filter.c
> index 0c66e4a3fc5b..1205dd777dc2 100644
> --- a/net/core/filter.c
> +++ b/net/core/filter.c
> @@ -5884,7 +5884,10 @@ static int bpf_ipv4_fib_lookup(struct net *net, struct bpf_fib_lookup *params,
>
> err = fib_table_lookup(tb, &fl4, &res, FIB_LOOKUP_NOREF);
> } else {
> - fl4.flowi4_mark = 0;
> + if (flags & BPF_FIB_LOOKUP_MARK)
> + fl4.flowi4_mark = params->mark;
> + else
> + fl4.flowi4_mark = 0;
> fl4.flowi4_secid = 0;
> fl4.flowi4_tun_key.tun_id = 0;
> fl4.flowi4_uid = sock_net_uid(net, NULL);
> @@ -6027,7 +6030,10 @@ static int bpf_ipv6_fib_lookup(struct net *net, struct bpf_fib_lookup *params,
> err = ipv6_stub->fib6_table_lookup(net, tb, oif, &fl6, &res,
> strict);
> } else {
> - fl6.flowi6_mark = 0;
> + if (flags & BPF_FIB_LOOKUP_MARK)
> + fl6.flowi6_mark = params->mark;
> + else
> + fl6.flowi6_mark = 0;
> fl6.flowi6_secid = 0;
> fl6.flowi6_tun_key.tun_id = 0;
> fl6.flowi6_uid = sock_net_uid(net, NULL);
> @@ -6105,7 +6111,7 @@ static int bpf_ipv6_fib_lookup(struct net *net, struct bpf_fib_lookup *params,
>
> #define BPF_FIB_LOOKUP_MASK (BPF_FIB_LOOKUP_DIRECT | BPF_FIB_LOOKUP_OUTPUT | \
> BPF_FIB_LOOKUP_SKIP_NEIGH | BPF_FIB_LOOKUP_TBID | \
> - BPF_FIB_LOOKUP_SRC)
> + BPF_FIB_LOOKUP_SRC | BPF_FIB_LOOKUP_MARK)
>
> BPF_CALL_4(bpf_xdp_fib_lookup, struct xdp_buff *, ctx,
> struct bpf_fib_lookup *, params, int, plen, u32, flags)
> diff --git a/tools/include/uapi/linux/bpf.h b/tools/include/uapi/linux/bpf.h
> index bf80b614c4db..4c9b5bfbd9c6 100644
> --- a/tools/include/uapi/linux/bpf.h
> +++ b/tools/include/uapi/linux/bpf.h
> @@ -3393,6 +3393,10 @@ union bpf_attr {
> * for the nexthop. If the src addr cannot be derived,
> * **BPF_FIB_LKUP_RET_NO_SRC_ADDR** is returned. In this
> * case, *params*->dmac and *params*->smac are not set either.
> + * **BPF_FIB_LOOKUP_MARK**
> + * Use the mark present in *params*->mark for the fib lookup.
> + * This option should not be used with BPF_FIB_LOOKUP_DIRECT,
> + * as it only has meaning for full lookups.
> *
> * *ctx* is either **struct xdp_md** for XDP programs or
> * **struct sk_buff** tc cls_act programs.
> @@ -7119,6 +7123,7 @@ enum {
> BPF_FIB_LOOKUP_SKIP_NEIGH = (1U << 2),
> BPF_FIB_LOOKUP_TBID = (1U << 3),
> BPF_FIB_LOOKUP_SRC = (1U << 4),
> + BPF_FIB_LOOKUP_MARK = (1U << 5),
> };
>
> enum {
> @@ -7196,8 +7201,19 @@ struct bpf_fib_lookup {
> __u32 tbid;
> };
>
> - __u8 smac[6]; /* ETH_ALEN */
> - __u8 dmac[6]; /* ETH_ALEN */
> + union {
> + /* input */
> + struct {
> + __u32 mark; /* policy routing */
> + /* 2 4-byte holes for input */
> + };
> +
> + /* output: source and dest mac */
> + struct {
> + __u8 smac[6]; /* ETH_ALEN */
> + __u8 dmac[6]; /* ETH_ALEN */
> + };
> + };
> };
>
> struct bpf_redir_neigh {
It would be good to add
static_assert(sizeof(struct bpf_fib_lookup) == 64, "bpf_fib_lookup size
check");
to ensure this struct never exceeds a cacheline.
The patch itself looks good to me:
Reviewed-by: David Ahern <dsahern@xxxxxxxxxx>
