[PATCH] net: skip RPS if packet is already on target CPU
From: Caleb Sander Mateos <hidden>
Date: 2024-10-29 18:27:50
Also in:
lkml
Subsystem:
networking [general], the rest · Maintainers:
"David S. Miller", Eric Dumazet, Jakub Kicinski, Paolo Abeni, Linus Torvalds
If RPS is enabled, all packets with a CPU flow hint are enqueued to the target CPU's input_pkt_queue and process_backlog() is scheduled on that CPU to dequeue and process the packets. If ARFS has already steered the packets to the correct CPU, this additional queuing is unnecessary and the spinlocks involved incur significant CPU overhead. In netif_receive_skb_internal() and netif_receive_skb_list_internal(), check if the CPU flow hint get_rps_cpu() returns is the current CPU. If so, bypass input_pkt_queue and immediately process the packet(s) on the current CPU. Signed-off-by: Caleb Sander Mateos <redacted> --- net/core/dev.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/net/core/dev.c b/net/core/dev.c
index c682173a7642..714a47897c75 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c@@ -5855,11 +5855,11 @@ static int netif_receive_skb_internal(struct sk_buff *skb) #ifdef CONFIG_RPS if (static_branch_unlikely(&rps_needed)) { struct rps_dev_flow voidflow, *rflow = &voidflow; int cpu = get_rps_cpu(skb->dev, skb, &rflow); - if (cpu >= 0) { + if (cpu >= 0 && cpu != smp_processor_id()) { ret = enqueue_to_backlog(skb, cpu, &rflow->last_qtail); rcu_read_unlock(); return ret; } }
@@ -5884,15 +5884,17 @@ void netif_receive_skb_list_internal(struct list_head *head) list_splice_init(&sublist, head); rcu_read_lock(); #ifdef CONFIG_RPS if (static_branch_unlikely(&rps_needed)) { + int curr_cpu = smp_processor_id(); + list_for_each_entry_safe(skb, next, head, list) { struct rps_dev_flow voidflow, *rflow = &voidflow; int cpu = get_rps_cpu(skb->dev, skb, &rflow); - if (cpu >= 0) { + if (cpu >= 0 && cpu != curr_cpu) { /* Will be handled, remove from list */ skb_list_del_init(skb); enqueue_to_backlog(skb, cpu, &rflow->last_qtail); } }
--
2.45.2