From: Changli Gao <xiaosuo@gmail.com>
To: "David S. Miller" <davem@davemloft.net>
Cc: Eric Dumazet <eric.dumazet@gmail.com>,
netdev@vger.kernel.org, Changli Gao <xiaosuo@gmail.com>
Subject: [PATCH v3] net: batch skb dequeueing from softnet input_pkt_queue
Date: Wed, 14 Apr 2010 17:52:18 +0800 [thread overview]
Message-ID: <1271238738-8386-1-git-send-email-xiaosuo@gmail.com> (raw)
batch skb dequeueing from softnet input_pkt_queue
batch skb dequeueing from softnet input_pkt_queue to reduce potential lock
contention and irq disabling/enabling.
Signed-off-by: Changli Gao <xiaosuo@gmail.com>
----
include/linux/netdevice.h | 1
net/core/dev.c | 56 ++++++++++++++++++++++++++++++++--------------
2 files changed, 40 insertions(+), 17 deletions(-)
diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h
index d1a21b5..898bc62 100644
--- a/include/linux/netdevice.h
+++ b/include/linux/netdevice.h
@@ -1335,6 +1335,7 @@ struct softnet_data {
struct call_single_data csd ____cacheline_aligned_in_smp;
#endif
struct sk_buff_head input_pkt_queue;
+ struct sk_buff_head processing_queue;
struct napi_struct backlog;
};
diff --git a/net/core/dev.c b/net/core/dev.c
index a10a216..c635a71 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -131,6 +131,7 @@
#include <linux/random.h>
#include <trace/events/napi.h>
#include <linux/pci.h>
+#include <linux/stop_machine.h>
#include "net-sysfs.h"
@@ -2332,6 +2333,7 @@ static int enqueue_to_backlog(struct sk_buff *skb, int cpu)
{
struct softnet_data *queue;
unsigned long flags;
+ u32 qlen;
queue = &per_cpu(softnet_data, cpu);
@@ -2339,8 +2341,9 @@ static int enqueue_to_backlog(struct sk_buff *skb, int cpu)
__get_cpu_var(netdev_rx_stat).total++;
rps_lock(queue);
- if (queue->input_pkt_queue.qlen <= netdev_max_backlog) {
- if (queue->input_pkt_queue.qlen) {
+ qlen = queue->input_pkt_queue.qlen + queue->processing_queue.qlen;
+ if (qlen <= netdev_max_backlog) {
+ if (qlen) {
enqueue:
__skb_queue_tail(&queue->input_pkt_queue, skb);
rps_unlock(queue);
@@ -2791,19 +2794,31 @@ int netif_receive_skb(struct sk_buff *skb)
EXPORT_SYMBOL(netif_receive_skb);
/* Network device is going away, flush any packets still pending */
-static void flush_backlog(void *arg)
+static void __flush_backlog(struct sk_buff_head *head, struct net_device *dev)
{
- struct net_device *dev = arg;
- struct softnet_data *queue = &__get_cpu_var(softnet_data);
struct sk_buff *skb, *tmp;
- rps_lock(queue);
- skb_queue_walk_safe(&queue->input_pkt_queue, skb, tmp)
+ skb_queue_walk_safe(head, skb, tmp) {
if (skb->dev == dev) {
- __skb_unlink(skb, &queue->input_pkt_queue);
+ __skb_unlink(skb, head);
kfree_skb(skb);
}
- rps_unlock(queue);
+ }
+}
+
+static int flush_backlog(void *arg)
+{
+ struct net_device *dev = arg;
+ struct softnet_data *queue;
+ int cpu;
+
+ for_each_online_cpu(cpu) {
+ queue = &per_cpu(softnet_data, cpu);
+ __flush_backlog(&queue->input_pkt_queue, dev);
+ __flush_backlog(&queue->processing_queue, dev);
+ }
+
+ return 0;
}
static int napi_gro_complete(struct sk_buff *skb)
@@ -3118,20 +3133,23 @@ static int process_backlog(struct napi_struct *napi, int quota)
local_irq_disable();
rps_lock(queue);
- skb = __skb_dequeue(&queue->input_pkt_queue);
- if (!skb) {
+ skb_queue_splice_tail_init(&queue->input_pkt_queue,
+ &queue->processing_queue);
+ if (skb_queue_empty(&queue->processing_queue)) {
__napi_complete(napi);
rps_unlock(queue);
local_irq_enable();
- break;
+ return work;
}
rps_unlock(queue);
local_irq_enable();
- __netif_receive_skb(skb);
- } while (++work < quota && jiffies == start_time);
-
- return work;
+ while ((skb = __skb_dequeue(&queue->processing_queue))) {
+ __netif_receive_skb(skb);
+ if (++work >= quota || jiffies != start_time)
+ return work;
+ }
+ } while (1);
}
/**
@@ -5027,7 +5045,7 @@ void netdev_run_todo(void)
dev->reg_state = NETREG_UNREGISTERED;
- on_each_cpu(flush_backlog, dev, 1);
+ stop_machine(flush_backlog, dev, NULL);
netdev_wait_allrefs(dev);
@@ -5487,6 +5505,9 @@ static int dev_cpu_callback(struct notifier_block *nfb,
raise_softirq_irqoff(NET_TX_SOFTIRQ);
local_irq_enable();
+ while ((skb = __skb_dequeue(&oldsd->processing_queue)))
+ netif_rx(skb);
+
/* Process offline CPU's input_pkt_queue */
while ((skb = __skb_dequeue(&oldsd->input_pkt_queue)))
netif_rx(skb);
@@ -5709,6 +5730,7 @@ static int __init net_dev_init(void)
queue = &per_cpu(softnet_data, i);
skb_queue_head_init(&queue->input_pkt_queue);
+ skb_queue_head_init(&queue->processing_queue);
queue->completion_queue = NULL;
INIT_LIST_HEAD(&queue->poll_list);
next reply other threads:[~2010-04-14 9:52 UTC|newest]
Thread overview: 13+ messages / expand[flat|nested] mbox.gz Atom feed top
2010-04-14 9:52 Changli Gao [this message]
2010-04-14 11:58 ` [PATCH v3] net: batch skb dequeueing from softnet input_pkt_queue jamal
2010-04-14 12:13 ` Changli Gao
2010-04-14 12:28 ` jamal
2010-04-14 15:20 ` Eric Dumazet
2010-04-14 23:13 ` Changli Gao
2010-04-21 23:05 ` Eric Dumazet
2010-04-21 23:23 ` Tom Herbert
2010-04-22 1:35 ` Changli Gao
2010-04-22 6:33 ` Changli Gao
2010-04-22 7:13 ` Eric Dumazet
2010-04-22 7:17 ` David Miller
2010-04-22 12:13 ` jamal
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1271238738-8386-1-git-send-email-xiaosuo@gmail.com \
--to=xiaosuo@gmail.com \
--cc=davem@davemloft.net \
--cc=eric.dumazet@gmail.com \
--cc=netdev@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.