[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <willemdebruijn.kernel.32fdbcb96baeb@gmail.com>
Date: Sun, 31 Aug 2025 21:21:46 -0400
From: Willem de Bruijn <willemdebruijn.kernel@...il.com>
To: Xin Zhao <jackzxcui1989@....com>,
willemdebruijn.kernel@...il.com,
edumazet@...gle.com,
ferenc@...es.dev
Cc: davem@...emloft.net,
kuba@...nel.org,
pabeni@...hat.com,
horms@...nel.org,
netdev@...r.kernel.org,
linux-kernel@...r.kernel.org,
Xin Zhao <jackzxcui1989@....com>
Subject: Re: [PATCH net-next v10 2/2] net: af_packet: Use hrtimer to do the
retire operation
Xin Zhao wrote:
> In a system with high real-time requirements, the timeout mechanism of
> ordinary timers with jiffies granularity is insufficient to meet the
> demands for real-time performance. Meanwhile, the optimization of CPU
> usage with af_packet is quite significant. Use hrtimer instead of timer
> to help compensate for the shortcomings in real-time performance.
> In HZ=100 or HZ=250 system, the update of TP_STATUS_USER is not real-time
> enough, with fluctuations reaching over 8ms (on a system with HZ=250).
> This is unacceptable in some high real-time systems that require timely
> processing of network packets. By replacing it with hrtimer, if a timeout
> of 2ms is set, the update of TP_STATUS_USER can be stabilized to within
> 3 ms.
>
> Signed-off-by: Xin Zhao <jackzxcui1989@....com>
Tiny style point that is probably not worth respinning for.
Otherwise
Reviewed-by: Willem de Bruijn <willemb@...gle.com>
> ---
> Changes in v8:
> - Simplify the logic related to setting timeout.
>
> Changes in v7:
> - Only update the hrtimer expire time within the hrtimer callback.
>
> Changes in v1:
> - Do not add another config for the current changes.
>
> ---
> net/packet/af_packet.c | 79 +++++++++---------------------------------
> net/packet/diag.c | 2 +-
> net/packet/internal.h | 6 ++--
> 3 files changed, 20 insertions(+), 67 deletions(-)
>
> diff --git a/net/packet/af_packet.c b/net/packet/af_packet.c
> index d4eb4a4fe..3e3bb4216 100644
> --- a/net/packet/af_packet.c
> +++ b/net/packet/af_packet.c
> @@ -203,8 +203,7 @@ static void prb_retire_current_block(struct tpacket_kbdq_core *,
> static int prb_queue_frozen(struct tpacket_kbdq_core *);
> static void prb_open_block(struct tpacket_kbdq_core *,
> struct tpacket_block_desc *);
> -static void prb_retire_rx_blk_timer_expired(struct timer_list *);
> -static void _prb_refresh_rx_retire_blk_timer(struct tpacket_kbdq_core *);
> +static enum hrtimer_restart prb_retire_rx_blk_timer_expired(struct hrtimer *);
> static void prb_fill_rxhash(struct tpacket_kbdq_core *, struct tpacket3_hdr *);
> static void prb_clear_rxhash(struct tpacket_kbdq_core *,
> struct tpacket3_hdr *);
> @@ -579,33 +578,13 @@ static __be16 vlan_get_protocol_dgram(const struct sk_buff *skb)
> return proto;
> }
>
> -static void prb_del_retire_blk_timer(struct tpacket_kbdq_core *pkc)
> -{
> - timer_delete_sync(&pkc->retire_blk_timer);
> -}
> -
> static void prb_shutdown_retire_blk_timer(struct packet_sock *po,
> struct sk_buff_head *rb_queue)
> {
> struct tpacket_kbdq_core *pkc;
>
> pkc = GET_PBDQC_FROM_RB(&po->rx_ring);
> -
> - spin_lock_bh(&rb_queue->lock);
> - pkc->delete_blk_timer = 1;
> - spin_unlock_bh(&rb_queue->lock);
> -
> - prb_del_retire_blk_timer(pkc);
> -}
> -
> -static void prb_setup_retire_blk_timer(struct packet_sock *po)
> -{
> - struct tpacket_kbdq_core *pkc;
> -
> - pkc = GET_PBDQC_FROM_RB(&po->rx_ring);
> - timer_setup(&pkc->retire_blk_timer, prb_retire_rx_blk_timer_expired,
> - 0);
> - pkc->retire_blk_timer.expires = jiffies;
> + hrtimer_cancel(&pkc->retire_blk_timer);
> }
>
> static int prb_calc_retire_blk_tmo(struct packet_sock *po,
> @@ -671,29 +650,22 @@ static void init_prb_bdqc(struct packet_sock *po,
> p1->version = po->tp_version;
> po->stats.stats3.tp_freeze_q_cnt = 0;
> if (req_u->req3.tp_retire_blk_tov)
> - p1->retire_blk_tov = req_u->req3.tp_retire_blk_tov;
> + p1->interval_ktime = ms_to_ktime(req_u->req3.tp_retire_blk_tov);
> else
> - p1->retire_blk_tov = prb_calc_retire_blk_tmo(po,
> - req_u->req3.tp_block_size);
> - p1->tov_in_jiffies = msecs_to_jiffies(p1->retire_blk_tov);
> + p1->interval_ktime = ms_to_ktime(prb_calc_retire_blk_tmo(po,
> + req_u->req3.tp_block_size));
req_u is not aligned with the line above.
Powered by blists - more mailing lists