[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAJ8uoz2oNHtyBaHtKJv-37oKRZykAtC3Q_ok8CmvO7RmdJoTNg@mail.gmail.com>
Date: Fri, 8 Nov 2019 19:09:27 +0100
From: Magnus Karlsson <magnus.karlsson@...il.com>
To: William Tu <u9012063@...il.com>
Cc: Magnus Karlsson <magnus.karlsson@...el.com>,
Björn Töpel <bjorn.topel@...el.com>,
Alexei Starovoitov <ast@...nel.org>,
Daniel Borkmann <daniel@...earbox.net>,
Network Development <netdev@...r.kernel.org>,
Jonathan Lemon <jonathan.lemon@...il.com>,
bpf <bpf@...r.kernel.org>
Subject: Re: [PATCH bpf-next 0/5] Extend libbpf to support shared umems and
Rx|Tx-only sockets
On Fri, Nov 8, 2019 at 3:58 PM William Tu <u9012063@...il.com> wrote:
>
> On Thu, Nov 07, 2019 at 06:47:35PM +0100, Magnus Karlsson wrote:
> > This patch set extends libbpf and the xdpsock sample program to
> > demonstrate the shared umem mode (XDP_SHARED_UMEM) as well as Rx-only
> > and Tx-only sockets. This in order for users to have an example to use
> > as a blue print and also so that these modes will be exercised more
> > frequently.
> >
> > Note that the user needs to supply an XDP program with the
> > XDP_SHARED_UMEM mode that distributes the packets over the sockets
> > according to some policy. There is an example supplied with the
> > xdpsock program, but there is no default one in libbpf similarly to
> > when XDP_SHARED_UMEM is not used. The reason for this is that I felt
> > that supplying one that would work for all users in this mode is
> > futile. There are just tons of ways to distribute packets, so whatever
> > I come up with and build into libbpf would be wrong in most cases.
> >
> Hi Magnus,
>
> Thanks for the patch.
> I look at the sample code and it's sharing a umem among multiple queues in
> the same netdev. Is it possible to shared one umem across multiple netdevs?
It should be possible to register the same umem area multiple times
(wasting memory in the current implementation though). I have not
tried this though, so I might be surprised. You really have to make
sure that you only give a buffer (through the Tx or fill rings) to a
single device. If you do not, your packets will be garbled. But this
needs some testing first and some extension to libbpf to make it
simple. I can try it out, but this will be another patch set.
/Magnus
> For example in OVS, one might create multiple tap/veth devices (using skb-mode
> or native-mode). And I want to save memory by having just one shared umem for
> these devices.
>
> Thanks
> --William
>
> > This patch has been applied against commit 30ee348c1267 ("Merge branch 'bpf-libbpf-fixes'")
> >
> > Structure of the patch set:
> >
> > Patch 1: Adds shared umem support to libbpf
> > Patch 2: Shared umem support and example XPD program added to xdpsock sample
> > Patch 3: Adds Rx-only and Tx-only support to libbpf
> > Patch 4: Uses Rx-only sockets for rxdrop and Tx-only sockets for txpush in
> > the xdpsock sample
> > Patch 5: Add documentation entries for these two features
> >
> > Thanks: Magnus
> >
> > Magnus Karlsson (5):
> > libbpf: support XDP_SHARED_UMEM with external XDP program
> > samples/bpf: add XDP_SHARED_UMEM support to xdpsock
> > libbpf: allow for creating Rx or Tx only AF_XDP sockets
> > samples/bpf: use Rx-only and Tx-only sockets in xdpsock
> > xsk: extend documentation for Rx|Tx-only sockets and shared umems
> >
> > Documentation/networking/af_xdp.rst | 28 +++++--
> > samples/bpf/Makefile | 1 +
> > samples/bpf/xdpsock.h | 11 +++
> > samples/bpf/xdpsock_kern.c | 24 ++++++
> > samples/bpf/xdpsock_user.c | 158 ++++++++++++++++++++++++++----------
> > tools/lib/bpf/xsk.c | 32 +++++---
> > 6 files changed, 195 insertions(+), 59 deletions(-)
> > create mode 100644 samples/bpf/xdpsock.h
> > create mode 100644 samples/bpf/xdpsock_kern.c
> >
> > --
> > 2.7.4
Powered by blists - more mailing lists