patches for DPDK stable branches
 help / color / mirror / Atom feed
From: Yuanhan Liu <yuanhan.liu@linux.intel.com>
To: Yuanhan Liu <yuanhan.liu@linux.intel.com>
Cc: Olivier Matz <olivier.matz@6wind.com>,
	Maxime Coquelin <maxime.coquelin@redhat.com>,
	"Michael S. Tsirkin" <mst@redhat.com>,
	dpdk stable <stable@dpdk.org>
Subject: [dpdk-stable] patch 'net/virtio: fix performance regression due to TSO' has been queued to stable release 16.11.1
Date: Wed, 15 Feb 2017 14:26:17 +0800	[thread overview]
Message-ID: <1487140012-13314-5-git-send-email-yuanhan.liu@linux.intel.com> (raw)
In-Reply-To: <1487140012-13314-1-git-send-email-yuanhan.liu@linux.intel.com>

Hi,

FYI, your patch has been queued to stable release 16.11.1

Note it hasn't been pushed to http://dpdk.org/browse/dpdk-stable
yet. It will be pushed if I get no objections before 02/18/17.
So please shout if anyone has objections.

Thanks.

	--yliu

---
>From 43ce94d1a20a60e7a0406bc6586379b23bf5bd18 Mon Sep 17 00:00:00 2001
From: Yuanhan Liu <yuanhan.liu@linux.intel.com>
Date: Wed, 11 Jan 2017 12:27:11 +0800
Subject: [PATCH] net/virtio: fix performance regression due to TSO

[ upstream commit c9ea670c1dc7e3f111d8139f915082b60c9c1ffe ]

TSO is now enabled, but it's not actually being used by default in a
simple L2 forward mode. In such case, we have to zero the virtio net
headers, to inform the vhost backend that no offload is being used:

    hdr->csum_start = 0;
    hdr->csum_offset = 0;
    hdr->flags = 0;

    hdr->gso_type = 0;
    hdr->gso_size = 0;
    hdr->hdr_len = 0;

Such writes could be very costly; it introduces severe cache issues:
The above operations introduce cache write for each packet, which
stalls the read operation from the vhost backend.

The fact that virtio net header is initiated to zero in PMD driver
init stage means that these costly writes are unnecessary and could
be avoided:

    if (hdr->csum_start != 0)
        hdr->csum_start = 0;

And that's what the macro ASSIGN_UNLESS_EQUAL does. With this, the
performance drop introduced by TSO enabling is recovered: it could
be up to 20% in micro benchmarking.

Fixes: 58169a9c8153 ("net/virtio: support Tx checksum offload")
Fixes: 696573046e9e ("net/virtio: support TSO")

Cc: Olivier Matz <olivier.matz@6wind.com>
Cc: Maxime Coquelin <maxime.coquelin@redhat.com>
Cc: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Yuanhan Liu <yuanhan.liu@linux.intel.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Reviewed-by: Olivier Matz <olivier.matz@6wind.com>
---
 drivers/net/virtio/virtio_rxtx.c | 18 ++++++++++++------
 1 file changed, 12 insertions(+), 6 deletions(-)

diff --git a/drivers/net/virtio/virtio_rxtx.c b/drivers/net/virtio/virtio_rxtx.c
index 22d97a4..edbd3cd 100644
--- a/drivers/net/virtio/virtio_rxtx.c
+++ b/drivers/net/virtio/virtio_rxtx.c
@@ -258,6 +258,12 @@ tx_offload_enabled(struct virtio_hw *hw)
 		vtpci_with_feature(hw, VIRTIO_NET_F_HOST_TSO6);
 }
 
+/* avoid write operation when necessary, to lessen cache issues */
+#define ASSIGN_UNLESS_EQUAL(var, val) do {	\
+	if ((var) != (val))			\
+		(var) = (val);			\
+} while (0)
+
 static inline void
 virtqueue_enqueue_xmit(struct virtnet_tx *txvq, struct rte_mbuf *cookie,
 		       uint16_t needed, int use_indirect, int can_push)
@@ -337,9 +343,9 @@ virtqueue_enqueue_xmit(struct virtnet_tx *txvq, struct rte_mbuf *cookie,
 			break;
 
 		default:
-			hdr->csum_start = 0;
-			hdr->csum_offset = 0;
-			hdr->flags = 0;
+			ASSIGN_UNLESS_EQUAL(hdr->csum_start, 0);
+			ASSIGN_UNLESS_EQUAL(hdr->csum_offset, 0);
+			ASSIGN_UNLESS_EQUAL(hdr->flags, 0);
 			break;
 		}
 
@@ -355,9 +361,9 @@ virtqueue_enqueue_xmit(struct virtnet_tx *txvq, struct rte_mbuf *cookie,
 				cookie->l3_len +
 				cookie->l4_len;
 		} else {
-			hdr->gso_type = 0;
-			hdr->gso_size = 0;
-			hdr->hdr_len = 0;
+			ASSIGN_UNLESS_EQUAL(hdr->gso_type, 0);
+			ASSIGN_UNLESS_EQUAL(hdr->gso_size, 0);
+			ASSIGN_UNLESS_EQUAL(hdr->hdr_len, 0);
 		}
 	}
 
-- 
1.9.0

  parent reply	other threads:[~2017-02-15  6:24 UTC|newest]

Thread overview: 45+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-02-15  6:26 [dpdk-stable] patch 'devargs: reset driver name pointer on parsing failure' " Yuanhan Liu
2017-02-15  6:26 ` [dpdk-stable] patch 'vhost: fix dead loop in enqueue path' " Yuanhan Liu
2017-02-15  6:26 ` [dpdk-stable] patch 'vhost: fix long stall of negotiation' " Yuanhan Liu
2017-02-15  6:26 ` [dpdk-stable] patch 'vhost: do not GSO when no header is present' " Yuanhan Liu
2017-02-15  6:26 ` Yuanhan Liu [this message]
2017-02-15  6:26 ` [dpdk-stable] patch 'net/virtio: optimize header reset on any layout' " Yuanhan Liu
2017-02-15  6:26 ` [dpdk-stable] patch 'net/i40e: fix Rx checksum flag' " Yuanhan Liu
2017-02-15  6:26 ` [dpdk-stable] patch 'net/mlx5: fix memory leak when parsing device params' " Yuanhan Liu
2017-02-15  6:26 ` [dpdk-stable] patch 'net/qede/base: fix FreeBSD build' " Yuanhan Liu
2017-02-15  6:26 ` [dpdk-stable] patch 'net/i40e: fix deletion of all macvlan filters' " Yuanhan Liu
2017-02-15  6:26 ` [dpdk-stable] patch 'net/bnx2x: fix Rx mode configuration' " Yuanhan Liu
2017-02-15  6:26 ` [dpdk-stable] patch 'net/cxgbe/base: initialize variable before reading EEPROM' " Yuanhan Liu
2017-02-15  6:26 ` [dpdk-stable] patch 'net/i40e: fix checksum flag in x86 vector Rx' " Yuanhan Liu
2017-02-15  6:26 ` [dpdk-stable] patch 'net/i40e: fix crash in close' " Yuanhan Liu
2017-02-15  6:26 ` [dpdk-stable] patch 'vfio: fix file descriptor leak in multi-process' " Yuanhan Liu
2017-02-15  6:26 ` [dpdk-stable] patch 'sched: fix crash when freeing port' " Yuanhan Liu
2017-02-15  6:26 ` [dpdk-stable] patch 'net/enic: fix memory leak with oversized Tx packets' " Yuanhan Liu
2017-02-15  6:26 ` [dpdk-stable] patch 'net/ena: fix setting host attributes' " Yuanhan Liu
2017-02-15  6:26 ` [dpdk-stable] patch 'net/i40e: fix ethertype filter on X722' " Yuanhan Liu
2017-02-15  6:26 ` [dpdk-stable] patch 'net/i40evf: fix reporting of imissed packets' " Yuanhan Liu
2017-02-15  6:26 ` [dpdk-stable] patch 'net/i40e: fix link update delay' " Yuanhan Liu
2017-02-15  6:26 ` [dpdk-stable] patch 'net/enic: fix hardcoding of some flow director masks' " Yuanhan Liu
2017-02-15  6:26 ` [dpdk-stable] patch 'net/ixgbe: fix received packets number for ARM' " Yuanhan Liu
2017-02-15  6:26 ` [dpdk-stable] patch 'net/ixgbe: fix received packets number for ARM NEON' " Yuanhan Liu
2017-02-15  6:26 ` [dpdk-stable] patch 'net/i40e: fix TC bandwidth definition' " Yuanhan Liu
2017-02-15  6:26 ` [dpdk-stable] patch 'drivers/crypto: fix different auth/cipher keys' " Yuanhan Liu
2017-02-15  6:26 ` [dpdk-stable] patch 'examples/l2fwd-crypto: fix overflow' " Yuanhan Liu
2017-02-15  6:26 ` [dpdk-stable] patch 'examples/l2fwd-crypto: fix padding' " Yuanhan Liu
2017-02-16  8:01   ` Yuanhan Liu
2017-02-16  9:10     ` De Lara Guarch, Pablo
2017-02-17  7:44       ` Yuanhan Liu
2017-02-15  6:26 ` [dpdk-stable] patch 'vhost: allow many vhost-user ports' " Yuanhan Liu
2017-02-15  6:26 ` [dpdk-stable] patch 'mempool: fix stack handler dequeue' " Yuanhan Liu
2017-02-15  6:26 ` [dpdk-stable] patch 'ethdev: fix port data mismatched in multiple process model' " Yuanhan Liu
2017-02-15  6:26 ` [dpdk-stable] patch 'net/virtio: fix wrong Rx/Tx method for secondary process' " Yuanhan Liu
2017-02-15  6:26 ` [dpdk-stable] patch 'net/virtio: store PCI operators pointer locally' " Yuanhan Liu
2017-02-15  6:26 ` [dpdk-stable] patch 'net/virtio: store IO port info " Yuanhan Liu
2017-02-15  6:26 ` [dpdk-stable] patch 'net/virtio: fix multiple process support' " Yuanhan Liu
2017-02-15  6:26 ` [dpdk-stable] patch 'net/virtio: fix build without virtio-user' " Yuanhan Liu
2017-02-15  6:26 ` [dpdk-stable] patch 'net/virtio: fix crash when number of virtio devices > 1' " Yuanhan Liu
2017-02-15  6:26 ` [dpdk-stable] patch 'usertools: fix active interface detection when binding' " Yuanhan Liu
2017-02-15  6:26 ` [dpdk-stable] patch 'net/vhost: fix unix socket not removed as closing' " Yuanhan Liu
2017-02-15  7:22   ` Tan, Jianfeng
2017-02-15  7:28     ` Yuanhan Liu
2017-02-15  6:26 ` [dpdk-stable] patch 'net/ixgbevf: fix max packet length' " Yuanhan Liu

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=1487140012-13314-5-git-send-email-yuanhan.liu@linux.intel.com \
    --to=yuanhan.liu@linux.intel.com \
    --cc=maxime.coquelin@redhat.com \
    --cc=mst@redhat.com \
    --cc=olivier.matz@6wind.com \
    --cc=stable@dpdk.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).