[ovs-dev] [gmane.linux.network] [PATCH 3/3] netlink: wake up netlink listeners sooner

[Ben Pfaff]

I wonder whether this would help OVS in some cases.

-------------------- Start of forwarded message --------------------
Path: news.gmane.org!not-for-mail
From: Stephen Hemminger <shemminger at vyatta.com>
Newsgroups: gmane.linux.network
Subject: [PATCH 3/3] netlink: wake up netlink listeners sooner
Date: Wed, 21 Dec 2011 13:49:44 -0800
Organization: Vyatta
Message-ID: <20111221134944.52223cdf at nehalam.linuxnetplumber.net>
Cc: netdev at vger.kernel.org
To: David Miller <davem at davemloft.net>
Archived-At: <http://permalink.gmane.org/gmane.linux.network/215921>

A netlink listening task (such as Quagga/zebra) can easily get overrun
when lots of events happen such as a link state transition with full BGP route
table. This happens because the sender does not yield to the receiver
(unless socket queue is full). The problem is exacerbated because it is
typical for listeners to set large receive buffer to attempt to keep up.

This patch changes it to yield sooner at halfway instead. Still not a cure-all
for listener overrun if listener is slow, but works much reliably.

Signed-off-by: Stephen Hemminger <shemminger at vyatta.com>


--- a/net/netlink/af_netlink.c	2011-12-20 10:25:19.364247598 -0800
+++ b/net/netlink/af_netlink.c	2011-12-20 10:25:46.756570010 -0800
@@ -960,7 +960,7 @@ static int netlink_broadcast_deliver(str
 		skb_set_owner_r(skb, sk);
 		skb_queue_tail(&sk->sk_receive_queue, skb);
 		sk->sk_data_ready(sk, skb->len);
-		return atomic_read(&sk->sk_rmem_alloc) > sk->sk_rcvbuf;
+		return atomic_read(&sk->sk_rmem_alloc) > sk->sk_rcvbuf / 2;
 	}
 	return -1;
 }

-------------------- End of forwarded message --------------------