Received-SPF: pass (google.com: domain of linux-nfs+bounces-2244-linux.lists.archive=gmail.com@vger.kernel.org designates 139.178.88.99 as permitted sender) client-ip=139.178.88.99;
From: trondmy@gmail.com
To: Steve Dickson <SteveD@redhat.com>
Cc: linux-nfs@vger.kernel.org
Subject: [PATCH v2] nfsd: allow more than 64 backlogged connections
Date: Fri,  8 Mar 2024 13:02:23 -0500
Message-ID: <20240308180223.2965601-1-trond.myklebust@hammerspace.com>
Precedence: bulk
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit

From: Trond Myklebust <trond.myklebust@hammerspace.com>

When creating a listener socket to be handed to /proc/fs/nfsd/portlist,
we currently limit the number of backlogged connections to 64. Since
that value was chosen in 2006, the scale at which data centres operate
has changed significantly. Given a modern server with many thousands of
clients, a limit of 64 connections can create bottlenecks, particularly
at at boot time.
Let's use the POSIX-sanctioned maximum value of SOMAXCONN.

Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
---
v2: Use SOMAXCONN instead of a value of -1.

 utils/nfsd/nfssvc.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/utils/nfsd/nfssvc.c b/utils/nfsd/nfssvc.c
index 46452d972407..9650cecee986 100644
--- a/utils/nfsd/nfssvc.c
+++ b/utils/nfsd/nfssvc.c
@@ -205,7 +205,8 @@ nfssvc_setfds(const struct addrinfo *hints, const char *node, const char *port)
 			rc = errno;
 			goto error;
 		}
-		if (addr->ai_protocol == IPPROTO_TCP && listen(sockfd, 64)) {
+		if (addr->ai_protocol == IPPROTO_TCP &&
+		    listen(sockfd, SOMAXCONN)) {
 			xlog(L_ERROR, "unable to create listening socket: "
 				"errno %d (%m)", errno);
 			rc = errno;
-- 
2.44.0