linux-nfs.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Chuck Lever <chuck.lever@oracle.com>
To: linux-nfs@vger.kernel.org, linux-rdma@vger.kernel.org
Subject: [PATCH v4 21/24] SUNRPC: Move congestion window contants to header file
Date: Wed, 21 May 2014 20:57:15 -0400	[thread overview]
Message-ID: <20140522005715.27190.64456.stgit@manet.1015granger.net> (raw)
In-Reply-To: <20140522004505.27190.58897.stgit@manet.1015granger.net>

I would like to use one of the RPC client's congestion algorithm
constants in transport-specific code.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
---

 include/linux/sunrpc/xprt.h |    6 ++++++
 net/sunrpc/xprt.c           |   28 +++++++++-------------------
 2 files changed, 15 insertions(+), 19 deletions(-)

diff --git a/include/linux/sunrpc/xprt.h b/include/linux/sunrpc/xprt.h
index 3e5efb2..5903d2c 100644
--- a/include/linux/sunrpc/xprt.h
+++ b/include/linux/sunrpc/xprt.h
@@ -24,6 +24,12 @@
 #define RPC_MAX_SLOT_TABLE_LIMIT	(65536U)
 #define RPC_MAX_SLOT_TABLE	RPC_MAX_SLOT_TABLE_LIMIT
 
+#define RPC_CWNDSHIFT		(8U)
+#define RPC_CWNDSCALE		(1U << RPC_CWNDSHIFT)
+#define RPC_INITCWND		RPC_CWNDSCALE
+#define RPC_MAXCWND(xprt)	((xprt)->max_reqs << RPC_CWNDSHIFT)
+#define RPCXPRT_CONGESTED(xprt) ((xprt)->cong >= (xprt)->cwnd)
+
 /*
  * This describes a timeout strategy
  */
diff --git a/net/sunrpc/xprt.c b/net/sunrpc/xprt.c
index d173f79..2d1d5a6 100644
--- a/net/sunrpc/xprt.c
+++ b/net/sunrpc/xprt.c
@@ -71,24 +71,6 @@ static void	 xprt_destroy(struct rpc_xprt *xprt);
 static DEFINE_SPINLOCK(xprt_list_lock);
 static LIST_HEAD(xprt_list);
 
-/*
- * The transport code maintains an estimate on the maximum number of out-
- * standing RPC requests, using a smoothed version of the congestion
- * avoidance implemented in 44BSD. This is basically the Van Jacobson
- * congestion algorithm: If a retransmit occurs, the congestion window is
- * halved; otherwise, it is incremented by 1/cwnd when
- *
- *	-	a reply is received and
- *	-	a full number of requests are outstanding and
- *	-	the congestion window hasn't been updated recently.
- */
-#define RPC_CWNDSHIFT		(8U)
-#define RPC_CWNDSCALE		(1U << RPC_CWNDSHIFT)
-#define RPC_INITCWND		RPC_CWNDSCALE
-#define RPC_MAXCWND(xprt)	((xprt)->max_reqs << RPC_CWNDSHIFT)
-
-#define RPCXPRT_CONGESTED(xprt) ((xprt)->cong >= (xprt)->cwnd)
-
 /**
  * xprt_register_transport - register a transport implementation
  * @transport: transport to register
@@ -446,7 +428,15 @@ EXPORT_SYMBOL_GPL(xprt_release_rqst_cong);
  * @task: recently completed RPC request used to adjust window
  * @result: result code of completed RPC request
  *
- * We use a time-smoothed congestion estimator to avoid heavy oscillation.
+ * The transport code maintains an estimate on the maximum number of out-
+ * standing RPC requests, using a smoothed version of the congestion
+ * avoidance implemented in 44BSD. This is basically the Van Jacobson
+ * congestion algorithm: If a retransmit occurs, the congestion window is
+ * halved; otherwise, it is incremented by 1/cwnd when
+ *
+ *	-	a reply is received and
+ *	-	a full number of requests are outstanding and
+ *	-	the congestion window hasn't been updated recently.
  */
 void xprt_adjust_cwnd(struct rpc_xprt *xprt, struct rpc_task *task, int result)
 {


  parent reply	other threads:[~2014-05-22  0:57 UTC|newest]

Thread overview: 27+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-05-22  0:54 [PATCH v4 00/24] NFS/RDMA client patches for next merge Chuck Lever
2014-05-22  0:54 ` [PATCH v4 01/24] xprtrdma: mind the device's max fast register page list depth Chuck Lever
2014-05-22  0:54 ` [PATCH v4 02/24] nfs-rdma: Fix for FMR leaks Chuck Lever
2014-05-22  0:54 ` [PATCH v4 03/24] xprtrdma: RPC/RDMA must invoke xprt_wake_pending_tasks() in process context Chuck Lever
2014-05-22  0:54 ` [PATCH v4 04/24] xprtrdma: Remove BOUNCEBUFFERS memory registration mode Chuck Lever
2014-05-22  0:54 ` [PATCH v4 05/24] xprtrdma: Remove MEMWINDOWS registration modes Chuck Lever
2014-05-22  0:55 ` [PATCH v4 06/24] xprtrdma: Remove REGISTER memory registration mode Chuck Lever
2014-05-22  0:55 ` [PATCH v4 07/24] xprtrdma: Fall back to MTHCAFMR when FRMR is not supported Chuck Lever
2014-05-22  0:55 ` [PATCH v4 08/24] xprtrdma: mount reports "Invalid mount option" if memreg mode " Chuck Lever
2014-05-22  0:55 ` [PATCH v4 09/24] xprtrdma: Simplify rpcrdma_deregister_external() synopsis Chuck Lever
2014-05-22  0:55 ` [PATCH v4 10/24] xprtrdma: Make rpcrdma_ep_destroy() return void Chuck Lever
2014-05-22  0:55 ` [PATCH v4 11/24] xprtrdma: Split the completion queue Chuck Lever
2014-05-22  0:55 ` [PATCH v4 12/24] xprtrmda: Reduce lock contention in completion handlers Chuck Lever
2014-05-22  0:56 ` [PATCH v4 13/24] xprtrmda: Reduce calls to ib_poll_cq() " Chuck Lever
2014-05-22  0:56 ` [PATCH v4 14/24] xprtrdma: Limit work done by completion handler Chuck Lever
2014-05-22  0:56 ` [PATCH v4 15/24] xprtrdma: Reduce the number of hardway buffer allocations Chuck Lever
2014-05-22  0:56 ` [PATCH v4 16/24] xprtrdma: Ensure ia->ri_id->qp is not NULL when reconnecting Chuck Lever
2014-05-22  0:56 ` [PATCH v4 17/24] xprtrdma: Remove Tavor MTU setting Chuck Lever
2014-05-22  0:56 ` [PATCH v4 18/24] xprtrdma: Allocate missing pagelist Chuck Lever
2014-05-22  0:56 ` [PATCH v4 19/24] xprtrdma: Use macros for reconnection timeout constants Chuck Lever
2014-05-22  0:57 ` [PATCH v4 20/24] xprtrdma: Reset connection timeout after successful reconnect Chuck Lever
2014-05-22  2:07   ` Trond Myklebust
2014-05-22  3:28     ` Chuck Lever
2014-05-22  0:57 ` Chuck Lever [this message]
2014-05-22  0:57 ` [PATCH v4 22/24] xprtrdma: Avoid deadlock when credit window is reset Chuck Lever
2014-05-22  0:57 ` [PATCH v4 23/24] xprtrdma: Remove BUG_ON() call sites Chuck Lever
2014-05-22  0:57 ` [PATCH v4 24/24] xprtrdma: Disconnect on registration failure Chuck Lever

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=20140522005715.27190.64456.stgit@manet.1015granger.net \
    --to=chuck.lever@oracle.com \
    --cc=linux-nfs@vger.kernel.org \
    --cc=linux-rdma@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 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).