[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20220624100536.4bbc1156@hermes.local>
Date: Fri, 24 Jun 2022 10:05:36 -0700
From: Stephen Hemminger <stephen@...workplumber.org>
To: James Yonan <james@...nvpn.net>
Cc: netdev@...r.kernel.org, therbert@...gle.com
Subject: Re: [PATCH net-next] rfs: added /proc/sys/net/core/rps_allow_ooo
flag to tweak flow alg
On Fri, 24 Jun 2022 10:54:47 -0600
James Yonan <james@...nvpn.net> wrote:
> @@ -4494,6 +4496,7 @@ static int get_rps_cpu(struct net_device *dev, struct sk_buff *skb,
> * If the desired CPU (where last recvmsg was done) is
> * different from current CPU (one in the rx-queue flow
> * table entry), switch if one of the following holds:
> + * - rps_allow_ooo_sysctl is enabled.
> * - Current CPU is unset (>= nr_cpu_ids).
> * - Current CPU is offline.
> * - The current CPU's queue tail has advanced beyond the
> @@ -4502,7 +4505,7 @@ static int get_rps_cpu(struct net_device *dev, struct sk_buff *skb,
> * have been dequeued, thus preserving in order delivery.
> */
> if (unlikely(tcpu != next_cpu) &&
> - (tcpu >= nr_cpu_ids || !cpu_online(tcpu) ||
> + (rps_allow_ooo_sysctl || tcpu >= nr_cpu_ids || !cpu_online(tcpu) ||
> ((int)(per_cpu(softnet_data, tcpu).input_queue_head -
> rflow->last_qtail)) >= 0)) {
This conditional is getting complex, maybe an inline helper function would
be clearer for developers that decide to add more in future.
Powered by blists - more mailing lists