[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <CAJ+HfNicARR3ZoH38_ANx9t5cqFn9DEfvegjUSN019un5xcgnA@mail.gmail.com>
Date: Tue, 30 Mar 2021 19:25:29 +0200
From: Björn Töpel <bjorn@...nel.org>
To: Ciara Loftus <ciara.loftus@...el.com>
Cc: Netdev <netdev@...r.kernel.org>, bpf <bpf@...r.kernel.org>,
"Karlsson, Magnus" <magnus.karlsson@...el.com>,
Alexei Starovoitov <alexei.starovoitov@...il.com>
Subject: Re: [PATCH v3 bpf 3/3] libbpf: only create rx and tx XDP rings when necessary
On Tue, 30 Mar 2021 at 14:05, Ciara Loftus <ciara.loftus@...el.com> wrote:
>
> Prior to this commit xsk_socket__create(_shared) always attempted to create
> the rx and tx rings for the socket. However this causes an issue when the
> socket being setup is that which shares the fd with the UMEM. If a
> previous call to this function failed with this socket after the rings were
> set up, a subsequent call would always fail because the rings are not torn
> down after the first call and when we try to set them up again we encounter
> an error because they already exist. Solve this by remembering whether the
> rings were set up by introducing a new flag to struct xsk_umem, and
> checking it before setting up the rings for sockets which share the fd
> with the UMEM.
>
> Fixes: 1cad07884239 ("libbpf: add support for using AF_XDP sockets")
>
> Signed-off-by: Ciara Loftus <ciara.loftus@...el.com>
> ---
> tools/lib/bpf/xsk.c | 16 ++++++++++++++--
> 1 file changed, 14 insertions(+), 2 deletions(-)
>
> diff --git a/tools/lib/bpf/xsk.c b/tools/lib/bpf/xsk.c
> index d4991ddff05a..12110bba4cc0 100644
> --- a/tools/lib/bpf/xsk.c
> +++ b/tools/lib/bpf/xsk.c
> @@ -14,6 +14,7 @@
> #include <unistd.h>
> #include <arpa/inet.h>
> #include <asm/barrier.h>
> +#include <linux/bitops.h>
> #include <linux/compiler.h>
> #include <linux/ethtool.h>
> #include <linux/filter.h>
> @@ -46,6 +47,9 @@
> #define PF_XDP AF_XDP
> #endif
>
> +#define XDP_RX_RING_SETUP_DONE BIT(0)
> +#define XDP_TX_RING_SETUP_DONE BIT(1)
> +
> enum xsk_prog {
> XSK_PROG_FALLBACK,
> XSK_PROG_REDIRECT_FLAGS,
> @@ -59,6 +63,7 @@ struct xsk_umem {
> int fd;
> int refcount;
> struct list_head ctx_list;
> + __u8 ring_setup_status;
Are we envisioning any more flags here? Otherwise, just a simple
bool/__u8 stating ring_setup_complete or ring_is_setup and just use
true/false and a simple check w/o bitwise and.
Björn
Powered by blists - more mailing lists