From: Jesper Dangaard Brouer <brouer@redhat.com>
To: "Toke Høiland-Jørgensen" <toke@redhat.com>
Cc: David Miller <davem@davemloft.net>,
netdev@vger.kernel.org, Daniel Borkmann <daniel@iogearbox.net>,
Alexei Starovoitov <ast@kernel.org>,
brouer@redhat.com, Jonathan Lemon <jonathan.lemon@gmail.com>
Subject: Re: [PATCH net-next 1/2] bpf_xdp_redirect_map: Add flag to return XDP_PASS on map lookup failure
Date: Wed, 5 Jun 2019 12:39:41 +0200 [thread overview]
Message-ID: <20190605123941.5b1d36ab@carbon> (raw)
In-Reply-To: <155966185069.9084.1926498690478259792.stgit@alrua-x1>
On Tue, 04 Jun 2019 17:24:10 +0200
Toke Høiland-Jørgensen <toke@redhat.com> wrote:
> The bpf_redirect_map() helper used by XDP programs doesn't return any
> indication of whether it can successfully redirect to the map index it was
> given. Instead, BPF programs have to track this themselves, leading to
> programs using duplicate maps to track which entries are populated in the
> devmap.
>
> This adds a flag to the XDP version of the bpf_redirect_map() helper, which
> makes the helper do a lookup in the map when called, and return XDP_PASS if
> there is no value at the provided index. This enables two use cases:
To Jonathan Lemon, notice this approach of adding a flag to the helper
call, it actually also works for your use-case of XSK AF_XDP maps.
> - A BPF program can check the return code from the helper call and react if
> it is XDP_PASS (by, for instance, redirecting out a different interface).
>
> - Programs that just return the value of the bpf_redirect() call will
> automatically fall back to the regular networking stack, simplifying
> programs that (for instance) build a router with the fib_lookup() helper.
>
> Signed-off-by: Toke Høiland-Jørgensen <toke@redhat.com>
> ---
> include/uapi/linux/bpf.h | 8 ++++++++
> net/core/filter.c | 10 +++++++++-
> 2 files changed, 17 insertions(+), 1 deletion(-)
>
> diff --git a/include/uapi/linux/bpf.h b/include/uapi/linux/bpf.h
> index 7c6aef253173..4c41482b7604 100644
> --- a/include/uapi/linux/bpf.h
> +++ b/include/uapi/linux/bpf.h
> @@ -3098,6 +3098,14 @@ enum xdp_action {
> XDP_REDIRECT,
> };
>
> +/* Flags for bpf_xdp_redirect_map helper */
> +
> +/* If set, the help will check if the entry exists in the map and return
> + * XDP_PASS if it doesn't.
> + */
> +#define XDP_REDIRECT_PASS_ON_INVALID BIT(0)
> +#define XDP_REDIRECT_ALL_FLAGS XDP_REDIRECT_PASS_ON_INVALID
> +
> /* user accessible metadata for XDP packet hook
> * new fields must be added to the end of this structure
> */
> diff --git a/net/core/filter.c b/net/core/filter.c
> index 55bfc941d17a..dfab8478f66c 100644
> --- a/net/core/filter.c
> +++ b/net/core/filter.c
> @@ -3755,9 +3755,17 @@ BPF_CALL_3(bpf_xdp_redirect_map, struct bpf_map *, map, u32, ifindex,
> {
> struct bpf_redirect_info *ri = this_cpu_ptr(&bpf_redirect_info);
>
> - if (unlikely(flags))
> + if (unlikely(flags & ~XDP_REDIRECT_ALL_FLAGS))
> return XDP_ABORTED;
>
> + if (flags & XDP_REDIRECT_PASS_ON_INVALID) {
> + struct net_device *fwd;
It is slightly misguiding that '*fwd' is a 'struct net_device', as the
__xdp_map_lookup_elem() call works for all the supported redirect-map
types.
People should realize that this patch is a general approach for all the
redirect-map types.
> +
> + fwd = __xdp_map_lookup_elem(map, ifindex);
> + if (unlikely(!fwd))
> + return XDP_PASS;
> + }
> +
> ri->ifindex = ifindex;
> ri->flags = flags;
> WRITE_ONCE(ri->map, map);
--
Best regards,
Jesper Dangaard Brouer
MSc.CS, Principal Kernel Engineer at Red Hat
LinkedIn: http://www.linkedin.com/in/brouer
next prev parent reply other threads:[~2019-06-05 10:40 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-06-04 15:24 [PATCH net-next 0/2] xdp: Allow lookup into devmaps before redirect Toke Høiland-Jørgensen
2019-06-04 15:24 ` [PATCH net-next 1/2] bpf_xdp_redirect_map: Add flag to return XDP_PASS on map lookup failure Toke Høiland-Jørgensen
2019-06-05 10:39 ` Jesper Dangaard Brouer [this message]
2019-06-05 15:09 ` Jonathan Lemon
2019-06-06 10:01 ` Toke Høiland-Jørgensen
2019-06-04 15:24 ` [PATCH net-next 2/2] devmap: Allow map lookups from eBPF Toke Høiland-Jørgensen
2019-06-04 16:35 ` Jesper Dangaard Brouer
2019-06-04 18:42 ` Toke Høiland-Jørgensen
2019-06-04 19:41 ` Jonathan Lemon
2019-06-04 20:00 ` Toke Høiland-Jørgensen
2019-06-04 20:22 ` Toke Høiland-Jørgensen
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20190605123941.5b1d36ab@carbon \
--to=brouer@redhat.com \
--cc=ast@kernel.org \
--cc=daniel@iogearbox.net \
--cc=davem@davemloft.net \
--cc=jonathan.lemon@gmail.com \
--cc=netdev@vger.kernel.org \
--cc=toke@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).