[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20210611163045.GC10632@netronome.com>
Date: Fri, 11 Jun 2021 18:30:46 +0200
From: Simon Horman <simon.horman@...ronome.com>
To: Toke Høiland-Jørgensen <toke@...hat.com>
Cc: bpf@...r.kernel.org, netdev@...r.kernel.org,
Martin KaFai Lau <kafai@...com>,
Hangbin Liu <liuhangbin@...il.com>,
Jesper Dangaard Brouer <brouer@...hat.com>,
Magnus Karlsson <magnus.karlsson@...il.com>,
"Paul E . McKenney" <paulmck@...nel.org>,
oss-drivers@...ronome.com, Jakub Kicinski <kuba@...nel.org>
Subject: Re: [PATCH bpf-next 12/17] nfp: remove rcu_read_lock() around XDP
program invocation
+Jakub
On Wed, Jun 09, 2021 at 12:33:21PM +0200, Toke Høiland-Jørgensen wrote:
> The nfp driver has rcu_read_lock()/rcu_read_unlock() pairs around XDP
> program invocations. However, the actual lifetime of the objects referred
> by the XDP program invocation is longer, all the way through to the call to
> xdp_do_flush(), making the scope of the rcu_read_lock() too small.
>
> While this is not actually an issue for the nfp driver because it doesn't
> support XDP_REDIRECT (and thus doesn't call xdp_do_flush()), the
> rcu_read_lock() is still unneeded. And With the addition of RCU annotations
> to the XDP_REDIRECT map types that take bh execution into account, lockdep
> even understands this to be safe, so there's really no reason to keep it
> around.
>
> Cc: Simon Horman <simon.horman@...ronome.com>
> Cc: oss-drivers@...ronome.com
> Signed-off-by: Toke Høiland-Jørgensen <toke@...hat.com>
Reviewed-by: Simon Horman <simon.horman@...ronome.com>
> ---
> drivers/net/ethernet/netronome/nfp/nfp_net_common.c | 2 --
> 1 file changed, 2 deletions(-)
>
> diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_common.c b/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
> index eeb30680b4dc..5dfa4799c34f 100644
> --- a/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
> +++ b/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
> @@ -1819,7 +1819,6 @@ static int nfp_net_rx(struct nfp_net_rx_ring *rx_ring, int budget)
> struct xdp_buff xdp;
> int idx;
>
> - rcu_read_lock();
> xdp_prog = READ_ONCE(dp->xdp_prog);
> true_bufsz = xdp_prog ? PAGE_SIZE : dp->fl_bufsz;
> xdp_init_buff(&xdp, PAGE_SIZE - NFP_NET_RX_BUF_HEADROOM,
> @@ -2036,7 +2035,6 @@ static int nfp_net_rx(struct nfp_net_rx_ring *rx_ring, int budget)
> if (!nfp_net_xdp_complete(tx_ring))
> pkts_polled = budget;
> }
> - rcu_read_unlock();
>
> return pkts_polled;
> }
> --
> 2.31.1
>
Powered by blists - more mailing lists