linux-nfs.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Chuck Lever <chuck.lever@oracle.com>
To: linux-rdma@vger.kernel.org, linux-nfs@vger.kernel.org
Subject: [PATCH v1 02/18] xprtrdma: Bound the inline threshold values
Date: Mon, 11 Apr 2016 16:10:25 -0400	[thread overview]
Message-ID: <20160411201024.20531.77252.stgit@manet.1015granger.net> (raw)
In-Reply-To: <20160411200323.20531.8893.stgit@manet.1015granger.net>

Currently the sysctls that allow setting the inline threshold allow
any value to be set.

Small values only make the transport run slower. The default 1KB
setting is as low as is reasonable. And the logic that decides how
to divide a Send buffer buffer between RPC-over-RDMA header and RPC
message assumes (but does not check) that the lower bound is not
crazy (say, 57 bytes).

Send and receive buffers share a page with some control information.
Values larger than about 3KB can't be supported, currently.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
---
 include/linux/sunrpc/xprtrdma.h |    4 +++-
 net/sunrpc/xprtrdma/transport.c |    6 ++++++
 2 files changed, 9 insertions(+), 1 deletion(-)

diff --git a/include/linux/sunrpc/xprtrdma.h b/include/linux/sunrpc/xprtrdma.h
index 767190b..39267dc 100644
--- a/include/linux/sunrpc/xprtrdma.h
+++ b/include/linux/sunrpc/xprtrdma.h
@@ -52,7 +52,9 @@
 #define RPCRDMA_DEF_SLOT_TABLE	(128U)
 #define RPCRDMA_MAX_SLOT_TABLE	(256U)
 
-#define RPCRDMA_DEF_INLINE  (1024)	/* default inline max */
+#define RPCRDMA_MIN_INLINE  (1024)	/* min inline thresh */
+#define RPCRDMA_DEF_INLINE  (1024)	/* default inline thresh */
+#define RPCRDMA_MAX_INLINE  (3068)	/* max inline thresh */
 
 /* Memory registration strategies, by number.
  * This is part of a kernel / user space API. Do not remove. */
diff --git a/net/sunrpc/xprtrdma/transport.c b/net/sunrpc/xprtrdma/transport.c
index 9954342..16595ff 100644
--- a/net/sunrpc/xprtrdma/transport.c
+++ b/net/sunrpc/xprtrdma/transport.c
@@ -73,6 +73,8 @@ static unsigned int xprt_rdma_memreg_strategy = RPCRDMA_FRMR;
 
 static unsigned int min_slot_table_size = RPCRDMA_MIN_SLOT_TABLE;
 static unsigned int max_slot_table_size = RPCRDMA_MAX_SLOT_TABLE;
+static unsigned int min_inline_size = RPCRDMA_MIN_INLINE;
+static unsigned int max_inline_size = RPCRDMA_MAX_INLINE;
 static unsigned int zero;
 static unsigned int max_padding = PAGE_SIZE;
 static unsigned int min_memreg = RPCRDMA_BOUNCEBUFFERS;
@@ -96,6 +98,8 @@ static struct ctl_table xr_tunables_table[] = {
 		.maxlen		= sizeof(unsigned int),
 		.mode		= 0644,
 		.proc_handler	= proc_dointvec,
+		.extra1		= &min_inline_size,
+		.extra2		= &max_inline_size,
 	},
 	{
 		.procname	= "rdma_max_inline_write",
@@ -103,6 +107,8 @@ static struct ctl_table xr_tunables_table[] = {
 		.maxlen		= sizeof(unsigned int),
 		.mode		= 0644,
 		.proc_handler	= proc_dointvec,
+		.extra1		= &min_inline_size,
+		.extra2		= &max_inline_size,
 	},
 	{
 		.procname	= "rdma_inline_write_padding",


  parent reply	other threads:[~2016-04-11 20:10 UTC|newest]

Thread overview: 29+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-04-11 20:10 [PATCH v1 00/18] NFS/RDMA client patches for 4.7 Chuck Lever
2016-04-11 20:10 ` [PATCH v1 01/18] sunrpc: Advertise maximum backchannel payload size Chuck Lever
2016-04-11 20:10 ` Chuck Lever [this message]
2016-04-12 18:04   ` [PATCH v1 02/18] xprtrdma: Bound the inline threshold values Anna Schumaker
2016-04-12 19:12     ` Chuck Lever
2016-04-11 20:10 ` [PATCH v1 03/18] xprtrdma: Limit number of RDMA segments in RPC-over-RDMA headers Chuck Lever
2016-04-11 20:10 ` [PATCH v1 04/18] xprtrdma: Prevent inline overflow Chuck Lever
2016-04-11 20:10 ` [PATCH v1 05/18] xprtrdma: Avoid using Write list for small NFS READ requests Chuck Lever
2016-04-11 20:35   ` Steve Wise
2016-04-11 20:38     ` Chuck Lever
2016-04-12 14:15       ` Christoph Hellwig
2016-04-12 14:49         ` Chuck Lever
2016-04-12 17:01           ` Christoph Hellwig
2016-04-12 18:08             ` Chuck Lever
2016-04-11 20:10 ` [PATCH v1 06/18] xprtrdma: Update comments in rpcrdma_marshal_req() Chuck Lever
2016-04-11 20:11 ` [PATCH v1 07/18] xprtrdma: Allow Read list and Reply chunk simultaneously Chuck Lever
2016-04-11 20:11 ` [PATCH v1 08/18] xprtrdma: Remove rpcrdma_create_chunks() Chuck Lever
2016-04-11 20:11 ` [PATCH v1 09/18] xprtrdma: Use core ib_drain_qp() API Chuck Lever
2016-04-12  4:49   ` Leon Romanovsky
2016-04-11 20:11 ` [PATCH v1 10/18] xprtrdma: Rename rpcrdma_frwr::sg and sg_nents Chuck Lever
2016-04-11 20:11 ` [PATCH v1 11/18] xprtrdma: Save I/O direction in struct rpcrdma_frwr Chuck Lever
2016-04-11 20:11 ` [PATCH v1 12/18] xprtrdma: Reset MRs in frwr_op_unmap_sync() Chuck Lever
2016-04-11 20:11 ` [PATCH v1 00/18] NFS/RDMA client patches for 4.7 Chuck Lever
2016-04-11 20:11 ` [PATCH v1 13/18] xprtrdma: Refactor the FRWR recovery worker Chuck Lever
2016-04-11 20:12 ` [PATCH v1 14/18] xprtrdma: Move fr_xprt and fr_worker to struct rpcrdma_mw Chuck Lever
2016-04-11 20:12 ` [PATCH v1 15/18] xprtrdma: Refactor __fmr_dma_unmap() Chuck Lever
2016-04-11 20:12 ` [PATCH v1 16/18] xprtrdma: Add ro_unmap_safe memreg method Chuck Lever
2016-04-11 20:12 ` [PATCH v1 17/18] xprtrdma: Remove ro_unmap() from all registration modes Chuck Lever
2016-04-11 20:12 ` [PATCH v1 18/18] xprtrdma: Faster server reboot recovery 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=20160411201024.20531.77252.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).