linux-nfs.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: NeilBrown <neilb@suse.com>
To: Olga Kornievskaia <aglo@umich.edu>,
	Chuck Lever <chuck.lever@oracle.com>,
	Schumaker Anna <Anna.Schumaker@netapp.com>,
	Trond Myklebust <trondmy@hammerspace.com>
Cc: linux-nfs@vger.kernel.org
Subject: [PATCH 2/9] SUNRPC: Allow creation of RPC clients with multiple connections
Date: Thu, 30 May 2019 10:41:28 +1000	[thread overview]
Message-ID: <155917688859.3988.12349634364426174850.stgit@noble.brown> (raw)
In-Reply-To: <155917564898.3988.6096672032831115016.stgit@noble.brown>

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

Add an argument to struct rpc_create_args that allows the specification
of how many transport connections you want to set up to the server.

Multiple transports allow hardware parallelism on the network path to
be fully exploited.  For example if there are multiple network
connections that are bonded together into a single virtual interface,
the bonding mechanism usually distributes flows, rather than packets,
across the multiple connections.  A single TCP flow will then only use
a single connection, while multiple TCP flows can use more of the
connections.
Similarly, when there are multiple DMA engines, multiple offload
engines, or multiple lanes for a network connection, using multiple
flows can allow more of the bandwidth to be utilized.

As an example, Olga Kornievskaia tested NFS read traffic to a
NetApp A700 using a 25GigE Ethernet port.
With a single TCP connection, throughput is limited to about 800
MB/sec
With 4 connections, 2400MB/sec is possible.  With 8, 3000 MB/sec
(24Gb/sec) can be achieved.

Tested-by: Olga Kornievskaia <aglo@umich.edu>
Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
Signed-off-by: NeilBrown <neilb@suse.com>
---
 include/linux/sunrpc/clnt.h |    1 +
 net/sunrpc/clnt.c           |   17 ++++++++++++++++-
 net/sunrpc/xprtmultipath.c  |    3 +--
 3 files changed, 18 insertions(+), 3 deletions(-)

diff --git a/include/linux/sunrpc/clnt.h b/include/linux/sunrpc/clnt.h
index 6e8073140a5d..4619098affa3 100644
--- a/include/linux/sunrpc/clnt.h
+++ b/include/linux/sunrpc/clnt.h
@@ -124,6 +124,7 @@ struct rpc_create_args {
 	u32			prognumber;	/* overrides program->number */
 	u32			version;
 	rpc_authflavor_t	authflavor;
+	u32			nconnect;
 	unsigned long		flags;
 	char			*client_name;
 	struct svc_xprt		*bc_xprt;	/* NFSv4.1 backchannel */
diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c
index 371080ad698a..3619dd5e9e0e 100644
--- a/net/sunrpc/clnt.c
+++ b/net/sunrpc/clnt.c
@@ -528,6 +528,8 @@ struct rpc_clnt *rpc_create(struct rpc_create_args *args)
 		.bc_xprt = args->bc_xprt,
 	};
 	char servername[48];
+	struct rpc_clnt *clnt;
+	int i;
 
 	if (args->bc_xprt) {
 		WARN_ON_ONCE(!(args->protocol & XPRT_TRANSPORT_BC));
@@ -590,7 +592,15 @@ struct rpc_clnt *rpc_create(struct rpc_create_args *args)
 	if (args->flags & RPC_CLNT_CREATE_NONPRIVPORT)
 		xprt->resvport = 0;
 
-	return rpc_create_xprt(args, xprt);
+	clnt = rpc_create_xprt(args, xprt);
+	if (IS_ERR(clnt) || args->nconnect <= 1)
+		return clnt;
+
+	for (i = 0; i < args->nconnect - 1; i++) {
+		if (rpc_clnt_add_xprt(clnt, &xprtargs, NULL, NULL) < 0)
+			break;
+	}
+	return clnt;
 }
 EXPORT_SYMBOL_GPL(rpc_create);
 
@@ -2748,6 +2758,10 @@ int rpc_clnt_test_and_add_xprt(struct rpc_clnt *clnt,
 		return -ENOMEM;
 	data->xps = xprt_switch_get(xps);
 	data->xprt = xprt_get(xprt);
+	if (rpc_xprt_switch_has_addr(data->xps, (struct sockaddr *)&xprt->addr)) {
+		rpc_cb_add_xprt_release(data);
+		goto success;
+	}
 
 	task = rpc_call_null_helper(clnt, xprt, NULL,
 			RPC_TASK_SOFT|RPC_TASK_SOFTCONN|RPC_TASK_ASYNC|RPC_TASK_NULLCREDS,
@@ -2755,6 +2769,7 @@ int rpc_clnt_test_and_add_xprt(struct rpc_clnt *clnt,
 	if (IS_ERR(task))
 		return PTR_ERR(task);
 	rpc_put_task(task);
+success:
 	return 1;
 }
 EXPORT_SYMBOL_GPL(rpc_clnt_test_and_add_xprt);
diff --git a/net/sunrpc/xprtmultipath.c b/net/sunrpc/xprtmultipath.c
index 394e427533be..9d66ce53355d 100644
--- a/net/sunrpc/xprtmultipath.c
+++ b/net/sunrpc/xprtmultipath.c
@@ -52,8 +52,7 @@ void rpc_xprt_switch_add_xprt(struct rpc_xprt_switch *xps,
 	if (xprt == NULL)
 		return;
 	spin_lock(&xps->xps_lock);
-	if ((xps->xps_net == xprt->xprt_net || xps->xps_net == NULL) &&
-	    !rpc_xprt_switch_has_addr(xps, (struct sockaddr *)&xprt->addr))
+	if (xps->xps_net == xprt->xprt_net || xps->xps_net == NULL)
 		xprt_switch_add_xprt_locked(xps, xprt);
 	spin_unlock(&xps->xps_lock);
 }



  reply	other threads:[~2019-05-30  0:43 UTC|newest]

Thread overview: 66+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-05-30  0:41 [PATCH 0/9] Multiple network connections for a single NFS mount NeilBrown
2019-05-30  0:41 ` NeilBrown [this message]
2019-05-30  0:41 ` [PATCH 9/9] NFS: Allow multiple connections to a NFSv2 or NFSv3 server NeilBrown
2019-05-30  0:41 ` [PATCH 4/9] SUNRPC: enhance rpc_clnt_show_stats() to report on all xprts NeilBrown
2019-05-30  0:41 ` [PATCH 5/9] SUNRPC: add links for all client xprts to debugfs NeilBrown
2019-05-30  0:41 ` [PATCH 3/9] NFS: send state management on a single connection NeilBrown
2019-07-23 18:11   ` Schumaker, Anna
2019-07-23 22:54     ` NeilBrown
2019-07-31  2:05     ` [PATCH] NFS: add flags arg to nfs4_call_sync_sequence() NeilBrown
2019-05-30  0:41 ` [PATCH 8/9] pNFS: Allow multiple connections to the DS NeilBrown
2019-05-30  0:41 ` [PATCH 1/9] SUNRPC: Add basic load balancing to the transport switch NeilBrown
2019-05-30  0:41 ` [PATCH 7/9] NFSv4: Allow multiple connections to NFSv4.x servers NeilBrown
2019-05-30  0:41 ` [PATCH 6/9] NFS: Add a mount option to specify number of TCP connections to use NeilBrown
2019-05-30 17:05 ` [PATCH 0/9] Multiple network connections for a single NFS mount Tom Talpey
2019-05-30 17:20   ` Olga Kornievskaia
2019-05-30 17:41     ` Tom Talpey
2019-05-30 18:41       ` Olga Kornievskaia
2019-05-31  1:45         ` Tom Talpey
2019-05-30 22:38       ` NeilBrown
2019-05-31  1:48         ` Tom Talpey
2019-05-31  2:31           ` NeilBrown
2019-05-31 12:39             ` Tom Talpey
2019-05-30 23:53     ` Rick Macklem
2019-05-31  0:15       ` J. Bruce Fields
2019-05-31  1:01       ` NeilBrown
2019-05-31  2:20         ` Rick Macklem
2019-05-31 12:36           ` Tom Talpey
2019-05-31 13:33             ` Trond Myklebust
2019-05-30 17:56 ` Chuck Lever
2019-05-30 18:59   ` Olga Kornievskaia
2019-05-30 22:56   ` NeilBrown
2019-05-31 13:46     ` Chuck Lever
2019-05-31 15:38       ` J. Bruce Fields
2019-06-11  1:09       ` NeilBrown
2019-06-11 14:51         ` Chuck Lever
2019-06-11 15:05           ` Tom Talpey
2019-06-11 15:20           ` Trond Myklebust
2019-06-11 15:35             ` Chuck Lever
2019-06-11 16:41               ` Trond Myklebust
2019-06-11 17:32                 ` Chuck Lever
2019-06-11 17:44                   ` Trond Myklebust
2019-06-12 12:34                     ` Steve Dickson
2019-06-12 12:47                       ` Trond Myklebust
2019-06-12 13:10                         ` Trond Myklebust
2019-06-11 15:34           ` Olga Kornievskaia
2019-06-11 17:46             ` Chuck Lever
2019-06-11 19:13               ` Olga Kornievskaia
2019-06-11 20:02                 ` Tom Talpey
2019-06-11 20:09                   ` Chuck Lever
2019-06-11 21:10                     ` Olga Kornievskaia
2019-06-11 21:35                       ` Tom Talpey
2019-06-11 22:55                         ` NeilBrown
2019-06-12 12:55                           ` Tom Talpey
2019-06-11 23:02                       ` NeilBrown
2019-06-11 23:21                   ` NeilBrown
2019-06-12 12:52                     ` Tom Talpey
2019-06-11 23:42               ` NeilBrown
2019-06-12 12:39                 ` Steve Dickson
2019-06-12 17:36                 ` Chuck Lever
2019-06-12 23:03                   ` NeilBrown
2019-06-13 16:13                     ` Chuck Lever
2019-06-12  1:49           ` NeilBrown
2019-06-12 18:32             ` Chuck Lever
2019-06-12 23:37               ` NeilBrown
2019-06-13 16:27                 ` Chuck Lever
2019-05-31  0:24 ` J. Bruce Fields

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=155917688859.3988.12349634364426174850.stgit@noble.brown \
    --to=neilb@suse.com \
    --cc=Anna.Schumaker@netapp.com \
    --cc=aglo@umich.edu \
    --cc=chuck.lever@oracle.com \
    --cc=linux-nfs@vger.kernel.org \
    --cc=trondmy@hammerspace.com \
    /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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).