linux-nfs.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Chuck Lever <chuck.lever@oracle.com>
To: anna.schumaker@netapp.com
Cc: linux-rdma@vger.kernel.org, linux-nfs@vger.kernel.org
Subject: [PATCH v5 08/30] xprtrdma: Reduce max_frwr_depth
Date: Wed, 19 Dec 2018 10:58:51 -0500	[thread overview]
Message-ID: <20181219155851.11602.24862.stgit@manet.1015granger.net> (raw)
In-Reply-To: <20181219155152.11602.18605.stgit@manet.1015granger.net>

Some devices advertise a large max_fast_reg_page_list_len
capability, but perform optimally when MRs are significantly smaller
than that depth -- probably when the MR itself is no larger than a
page.

By default, the RDMA R/W core API uses max_sge_rd as the maximum
page depth for MRs. For some devices, the value of max_sge_rd is
1, which is also not optimal. Thus, when max_sge_rd is larger than
1, use that value. Otherwise use the value of the
max_fast_reg_page_list_len attribute.

I've tested this with CX-3 Pro, FastLinq, and CX-5 devices. It
reproducibly improves the throughput of large I/Os by several
percent.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
---
 net/sunrpc/xprtrdma/frwr_ops.c |   15 +++++++++++----
 1 file changed, 11 insertions(+), 4 deletions(-)

diff --git a/net/sunrpc/xprtrdma/frwr_ops.c b/net/sunrpc/xprtrdma/frwr_ops.c
index f587e44..16976b0 100644
--- a/net/sunrpc/xprtrdma/frwr_ops.c
+++ b/net/sunrpc/xprtrdma/frwr_ops.c
@@ -193,10 +193,17 @@
 	if (attrs->device_cap_flags & IB_DEVICE_SG_GAPS_REG)
 		ia->ri_mrtype = IB_MR_TYPE_SG_GAPS;
 
-	ia->ri_max_frwr_depth =
-			min_t(unsigned int, RPCRDMA_MAX_DATA_SEGS,
-			      attrs->max_fast_reg_page_list_len);
-	dprintk("RPC:       %s: device's max FR page list len = %u\n",
+	/* Quirk: Some devices advertise a large max_fast_reg_page_list_len
+	 * capability, but perform optimally when the MRs are not larger
+	 * than a page.
+	 */
+	if (attrs->max_sge_rd > 1)
+		ia->ri_max_frwr_depth = attrs->max_sge_rd;
+	else
+		ia->ri_max_frwr_depth = attrs->max_fast_reg_page_list_len;
+	if (ia->ri_max_frwr_depth > RPCRDMA_MAX_DATA_SEGS)
+		ia->ri_max_frwr_depth = RPCRDMA_MAX_DATA_SEGS;
+	dprintk("RPC:       %s: max FR page list depth = %u\n",
 		__func__, ia->ri_max_frwr_depth);
 
 	/* Add room for frwr register and invalidate WRs.


  parent reply	other threads:[~2018-12-19 15:58 UTC|newest]

Thread overview: 31+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-12-19 15:58 [PATCH v5 00/30] NFS/RDMA client for next (4.21) Chuck Lever
2018-12-19 15:58 ` [PATCH v5 01/30] xprtrdma: Yet another double DMA-unmap Chuck Lever
2018-12-19 15:58 ` [PATCH v5 02/30] xprtrdma: Ensure MRs are DMA-unmapped when posting LOCAL_INV fails Chuck Lever
2018-12-19 15:58 ` [PATCH v5 03/30] xprtrdma: Refactor Receive accounting Chuck Lever
2018-12-19 15:58 ` [PATCH v5 04/30] xprtrdma: Replace rpcrdma_receive_wq with a per-xprt workqueue Chuck Lever
2018-12-19 15:58 ` [PATCH v5 05/30] xprtrdma: No qp_event disconnect Chuck Lever
2018-12-19 15:58 ` [PATCH v5 06/30] xprtrdma: Don't wake pending tasks until disconnect is done Chuck Lever
2018-12-19 15:58 ` [PATCH v5 07/30] xprtrdma: Fix ri_max_segs and the result of ro_maxpages Chuck Lever
2018-12-19 15:58 ` Chuck Lever [this message]
2018-12-19 15:58 ` [PATCH v5 09/30] xprtrdma: Remove support for FMR memory registration Chuck Lever
2018-12-19 15:59 ` [PATCH v5 10/30] xprtrdma: Remove rpcrdma_memreg_ops Chuck Lever
2018-12-19 15:59 ` [PATCH v5 11/30] xprtrdma: Plant XID in on-the-wire RDMA offset (FRWR) Chuck Lever
2018-12-19 15:59 ` [PATCH v5 12/30] NFS: Make "port=" mount option optional for RDMA mounts Chuck Lever
2018-12-19 15:59 ` [PATCH v5 13/30] xprtrdma: Recognize XDRBUF_SPARSE_PAGES Chuck Lever
2018-12-19 15:59 ` [PATCH v5 14/30] xprtrdma: Remove request_module from backchannel Chuck Lever
2018-12-19 15:59 ` [PATCH v5 15/30] xprtrdma: Expose transport header errors Chuck Lever
2018-12-19 15:59 ` [PATCH v5 16/30] xprtrdma: Simplify locking that protects the rl_allreqs list Chuck Lever
2018-12-19 15:59 ` [PATCH v5 17/30] xprtrdma: Cull dprintk() call sites Chuck Lever
2018-12-19 15:59 ` [PATCH v5 18/30] xprtrdma: Remove unused fields from rpcrdma_ia Chuck Lever
2018-12-19 15:59 ` [PATCH v5 19/30] xprtrdma: Clean up of xprtrdma chunk trace points Chuck Lever
2018-12-19 15:59 ` [PATCH v5 20/30] xprtrdma: Relocate the xprtrdma_mr_map " Chuck Lever
2018-12-19 16:00 ` [PATCH v5 21/30] xprtrdma: Add trace points for calls to transport switch methods Chuck Lever
2018-12-19 16:00 ` [PATCH v5 22/30] xprtrdma: Trace mapping, alloc, and dereg failures Chuck Lever
2018-12-19 16:00 ` [PATCH v5 23/30] NFS: Fix NFSv4 symbolic trace point output Chuck Lever
2018-12-19 16:00 ` [PATCH v5 24/30] SUNRPC: Simplify defining common RPC trace events Chuck Lever
2018-12-19 16:00 ` [PATCH v5 25/30] SUNRPC: Fix some kernel doc complaints Chuck Lever
2018-12-19 16:00 ` [PATCH v5 26/30] xprtrdma: Update comments in frwr_op_send Chuck Lever
2018-12-19 16:00 ` [PATCH v5 27/30] xprtrdma: Replace outdated comment for rpcrdma_ep_post Chuck Lever
2018-12-19 16:00 ` [PATCH v5 28/30] xprtrdma: Add documenting comment for rpcrdma_buffer_destroy Chuck Lever
2018-12-19 16:00 ` [PATCH v5 29/30] xprtrdma: Clarify comments in rpcrdma_ia_remove Chuck Lever
2018-12-19 16:00 ` [PATCH v5 30/30] xprtrdma: Don't leak freed MRs 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=20181219155851.11602.24862.stgit@manet.1015granger.net \
    --to=chuck.lever@oracle.com \
    --cc=anna.schumaker@netapp.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).