netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Jesper Dangaard Brouer <brouer@redhat.com>
To: Jesper Dangaard Brouer <brouer@redhat.com>, netdev@vger.kernel.org
Cc: Alexander Duyck <alexander.h.duyck@intel.com>,
	Jeff Kirsher <jeffrey.t.kirsher@intel.com>,
	Daniel Borkmann <dborkman@redhat.com>,
	Florian Westphal <fw@strlen.de>,
	"David S. Miller" <davem@davemloft.net>,
	Stephen Hemminger <shemminger@vyatta.com>,
	"Paul E. McKenney" <paulmck@linux.vnet.ibm.com>,
	Robert Olsson <robert@herjulf.se>,
	Ben Greear <greearb@candelatech.com>,
	John Fastabend <john.r.fastabend@intel.com>,
	danieltt@kth.se, zhouzhouyi@gmail.com
Subject: [net-next PATCH 2/5] ixgbe: increase default TX ring buffer to 1024
Date: Wed, 14 May 2014 16:17:48 +0200	[thread overview]
Message-ID: <20140514141748.20309.83121.stgit@dragon> (raw)
In-Reply-To: <20140514141545.20309.28343.stgit@dragon>

Using pktgen I'm seeing the ixgbe driver "push-back", due TX ring
running full.  Thus, the TX ring is artificially limiting pktgen.

Diagnose via "ethtool -S", look for "tx_restart_queue" or "tx_busy"
counters.

Increasing the TX ring buffer should be done carefully, as it comes at
a higher memory cost, which can also negatively influence performance.
E.g. ring buffer array of struct ixgbe_tx_buffer (current size 48bytes)
increase from 512*48=24576bytes to 1024*48=49152bytes which is larger
than the L1 data cache (32KB on my E5-2630), thus increasing the L1->L2
cache-references.

Adjusting the TX ring buffer (TXSZ) measured over 10 sec with ifpps
 (single CPU performance, ixgbe 10Gbit/s, E5-2630)
 * cmd: ethtool -G eth8 tx $TXSZ
 * 3,930,065 pps -- TXSZ= 512
 * 5,312,249 pps -- TXSZ= 768
 * 5,362,722 pps -- TXSZ=1024
 * 5,361,390 pps -- TXSZ=1536
 * 5,362,439 pps -- TXSZ=2048
 * 5,359,744 pps -- TXSZ=4096

Choosing size 1024 because for the next optimizations 768 is not
enough.

Notice after commit 6f25cd47d (pktgen: fix xmit test for BQL enabled
devices) pktgen uses netif_xmit_frozen_or_drv_stopped() and ignores
the BQL "stack" pause (QUEUE_STATE_STACK_XOFF) flag.  This allow us to put
more pressure on the TX ring buffers.

It is the ixgbe_maybe_stop_tx() call that stops the transmits, and
pktgen respecting this in the call to netif_xmit_frozen_or_drv_stopped(txq).

Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
---

 drivers/net/ethernet/intel/ixgbe/ixgbe.h |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe.h b/drivers/net/ethernet/intel/ixgbe/ixgbe.h
index c688c8a..bf078fe 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe.h
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe.h
@@ -63,7 +63,7 @@
 #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
 
 /* TX/RX descriptor defines */
-#define IXGBE_DEFAULT_TXD		    512
+#define IXGBE_DEFAULT_TXD		   1024
 #define IXGBE_DEFAULT_TX_WORK		    256
 #define IXGBE_MAX_TXD			   4096
 #define IXGBE_MIN_TXD			     64

  parent reply	other threads:[~2014-05-14 14:18 UTC|newest]

Thread overview: 22+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-05-14 14:17 [net-next PATCH 0/5] Optimizing "pktgen" for single CPU performance Jesper Dangaard Brouer
2014-05-14 14:17 ` [net-next PATCH 1/5] ixgbe: trivial fixes while reading code Jesper Dangaard Brouer
2014-05-14 14:17 ` Jesper Dangaard Brouer [this message]
2014-05-14 14:28   ` [net-next PATCH 2/5] ixgbe: increase default TX ring buffer to 1024 David Laight
2014-05-14 19:25     ` Jesper Dangaard Brouer
2014-05-14 16:28   ` Alexander Duyck
2014-05-14 17:49     ` David Miller
2014-05-14 19:09       ` Jesper Dangaard Brouer
2014-05-14 19:54         ` David Miller
2014-05-15  9:16         ` David Laight
2014-05-29 15:29         ` Jesper Dangaard Brouer
2014-05-14 14:17 ` [net-next PATCH 3/5] pktgen: avoid atomic_inc per packet in xmit loop Jesper Dangaard Brouer
2014-05-14 14:35   ` Eric Dumazet
2014-05-14 15:13     ` Jesper Dangaard Brouer
2014-05-14 15:35       ` Eric Dumazet
2014-05-14 14:17 ` [net-next PATCH 4/5] pktgen: avoid expensive set_current_state() call in loop Jesper Dangaard Brouer
2014-05-14 14:18 ` [net-next PATCH 5/5] pktgen: RCU'ify "if_list" to remove lock in next_to_run() Jesper Dangaard Brouer
2014-06-26 11:16 ` [net-next PATCH V2 0/3] Optimizing pktgen for single CPU performance Jesper Dangaard Brouer
2014-06-26 11:16   ` [net-next PATCH V2 1/3] pktgen: document tuning for max NIC performance Jesper Dangaard Brouer
2014-06-26 11:16   ` [net-next PATCH V2 2/3] pktgen: avoid expensive set_current_state() call in loop Jesper Dangaard Brouer
2014-06-26 11:16   ` [net-next PATCH V2 3/3] pktgen: RCU-ify "if_list" to remove lock in next_to_run() Jesper Dangaard Brouer
2014-07-01 22:51   ` [net-next PATCH V2 0/3] Optimizing pktgen for single CPU performance David Miller

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=20140514141748.20309.83121.stgit@dragon \
    --to=brouer@redhat.com \
    --cc=alexander.h.duyck@intel.com \
    --cc=danieltt@kth.se \
    --cc=davem@davemloft.net \
    --cc=dborkman@redhat.com \
    --cc=fw@strlen.de \
    --cc=greearb@candelatech.com \
    --cc=jeffrey.t.kirsher@intel.com \
    --cc=john.r.fastabend@intel.com \
    --cc=netdev@vger.kernel.org \
    --cc=paulmck@linux.vnet.ibm.com \
    --cc=robert@herjulf.se \
    --cc=shemminger@vyatta.com \
    --cc=zhouzhouyi@gmail.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).