From mboxrd@z Thu Jan 1 00:00:00 1970
Return-Path: <danny.zhou@intel.com>
Received: from mga09.intel.com (mga09.intel.com [134.134.136.24])
by dpdk.org (Postfix) with ESMTP id B5C945901
for <dev@dpdk.org>; Fri, 12 Sep 2014 22:31:14 +0200 (CEST)
Received: from azsmga001.ch.intel.com ([10.2.17.19])
by orsmga102.jf.intel.com with ESMTP; 12 Sep 2014 13:29:44 -0700
X-ExtLoop1: 1
X-IronPort-AV: E=Sophos;i="5.04,515,1406617200"; d="scan'208";a="477198894"
Received: from fmsmsx103.amr.corp.intel.com ([10.18.124.201])
by azsmga001.ch.intel.com with ESMTP; 12 Sep 2014 13:35:50 -0700
Received: from FMSMSX110.amr.corp.intel.com (10.18.116.10) by
FMSMSX103.amr.corp.intel.com (10.18.124.201) with Microsoft SMTP Server (TLS)
id 14.3.195.1; Fri, 12 Sep 2014 13:35:50 -0700
Received: from shsmsx151.ccr.corp.intel.com (10.239.6.50) by
fmsmsx110.amr.corp.intel.com (10.18.116.10) with Microsoft SMTP Server (TLS)
id 14.3.195.1; Fri, 12 Sep 2014 13:35:49 -0700
Received: from shsmsx104.ccr.corp.intel.com ([169.254.5.230]) by
SHSMSX151.ccr.corp.intel.com ([169.254.3.172]) with mapi id 14.03.0195.001;
Sat, 13 Sep 2014 04:35:48 +0800
From: "Zhou, Danny" <danny.zhou@intel.com>
To: "John W. Linville" <linville@tuxdriver.com>
Thread-Topic: [dpdk-dev] [PATCH v2] librte_pmd_packet: add PMD for
AF_PACKET-based virtual devices
Thread-Index: AQHPn5Gwvcd4A0wcTkeObxakesCY6Jv9ov6AgACKEgD//4OfgIAAmgdg
Date: Fri, 12 Sep 2014 20:35:47 +0000
Message-ID: <DFDF335405C17848924A094BC35766CF0A935E75@SHSMSX104.ccr.corp.intel.com>
References: <1405024369-30058-1-git-send-email-linville@tuxdriver.com>
<1405362290-6753-1-git-send-email-linville@tuxdriver.com>
<20140912180523.GB7145@tuxdriver.com>
<DFDF335405C17848924A094BC35766CF0A935AEE@SHSMSX104.ccr.corp.intel.com>
<20140912185423.GD7145@tuxdriver.com>
In-Reply-To: <20140912185423.GD7145@tuxdriver.com>
Accept-Language: zh-CN, en-US
Content-Language: en-US
X-MS-Has-Attach:
X-MS-TNEF-Correlator:
x-originating-ip: [10.239.127.40]
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
Cc: "dev@dpdk.org" <dev@dpdk.org>
Subject: Re: [dpdk-dev] [PATCH v2] librte_pmd_packet: add PMD for
AF_PACKET-based virtual devices
X-BeenThere: dev@dpdk.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: patches and discussions about DPDK <dev.dpdk.org>
List-Unsubscribe: <http://dpdk.org/ml/options/dev>,
<mailto:dev-request@dpdk.org?subject=unsubscribe>
List-Archive: <http://dpdk.org/ml/archives/dev/>
List-Post: <mailto:dev@dpdk.org>
List-Help: <mailto:dev-request@dpdk.org?subject=help>
List-Subscribe: <http://dpdk.org/ml/listinfo/dev>,
<mailto:dev-request@dpdk.org?subject=subscribe>
X-List-Received-Date: Fri, 12 Sep 2014 20:31:16 -0000
> -----Original Message-----
> From: John W. Linville [mailto:linville@tuxdriver.com]
> Sent: Saturday, September 13, 2014 2:54 AM
> To: Zhou, Danny
> Cc: dev@dpdk.org
> Subject: Re: [dpdk-dev] [PATCH v2] librte_pmd_packet: add PMD for AF_PACK=
ET-based virtual devices
>=20
> On Fri, Sep 12, 2014 at 06:31:08PM +0000, Zhou, Danny wrote:
> > I am concerned about its performance caused by too many
> > memcpy(). Specifically, on Rx side, kernel NIC driver needs to copy
> > packets to skb, then af_packet copies packets to AF_PACKET buffer
> > which are mapped to user space, and then those packets to be copied
> > to DPDK mbuf. In addition, 3 copies needed on Tx side. So to run a
> > simple DPDK L2/L3 forwarding benchmark, each packet needs 6 packet
> > copies which brings significant negative performance impact. We
> > had a bifurcated driver prototype that can do zero-copy and achieve
> > native DPDK performance, but it depends on base driver and AF_PACKET
> > code changes in kernel, John R will be presenting it in coming Linux
> > Plumbers Conference. Once kernel adopts it, the relevant PMD will be
> > submitted to dpdk.org.
>=20
> Admittedly, this is not as good a performer as most of the existing
> PMDs. It serves a different purpose, afterall. FWIW, you did
> previously indicate that it performed better than the pcap-based PMD.
Yes, slightly higher but makes no big difference.
> I look forward to seeing the changes you mention -- they sound very
> exciting. But, they will still require both networking core and
> driver changes in the kernel. And as I understand things today,
> the userland code will still need at least some knowledge of specific
> devices and how they layout their packet descriptors, etc. So while
> those changes sound very promising, they will still have certain
> drawbacks in common with the current situation.
Yes, we would like the DPDK performance optimization techniques such as hug=
e page, efficient rx/tx routines to manipulate device-specific=20
packet descriptors, polling-model can be still used. We have to tradeoff be=
tween performance and commonality. But we believe it will be much easier
to develop DPDK PMD for non-Intel NICs than porting entire kernel drivers t=
o DPDK.
> It seems like the changes you mention will still need some sort of
> AF_PACKET-based PMD driver. Have you implemented that completely
> separate from the code I already posted? Or did you add that work
> on top of mine?
>=20
For userland code, it certainly use some of your code related to raw rocket=
, but highly modified. A layer will be added into eth_dev library to do dev=
ice
probe and support new socket options.
> John
>=20
> > > -----Original Message-----
> > > From: dev [mailto:dev-bounces@dpdk.org] On Behalf Of John W. Linville
> > > Sent: Saturday, September 13, 2014 2:05 AM
> > > To: dev@dpdk.org
> > > Subject: Re: [dpdk-dev] [PATCH v2] librte_pmd_packet: add PMD for AF_=
PACKET-based virtual devices
> > >
> > > Ping? Are there objections to this patch from mid-July?
> > >
> > > John
> > >
> > > On Mon, Jul 14, 2014 at 02:24:50PM -0400, John W. Linville wrote:
> > > > This is a Linux-specific virtual PMD driver backed by an AF_PACKET
> > > > socket. This implementation uses mmap'ed ring buffers to limit cop=
ying
> > > > and user/kernel transitions. The PACKET_FANOUT_HASH behavior of
> > > > AF_PACKET is used for frame reception. In the current implementati=
on,
> > > > Tx and Rx queues are always paired, and therefore are always equal
> > > > in number -- changing this would be a Simple Matter Of Programming.
> > > >
> > > > Interfaces of this type are created with a command line option like
> > > > "--vdev=3Deth_packet0,iface=3D...". There are a number of options =
availabe
> > > > as arguments:
> > > >
> > > > - Interface is chosen by "iface" (required)
> > > > - Number of queue pairs set by "qpairs" (optional, default: 1)
> > > > - AF_PACKET MMAP block size set by "blocksz" (optional, default: 4=
096)
> > > > - AF_PACKET MMAP frame size set by "framesz" (optional, default: 2=
048)
> > > > - AF_PACKET MMAP frame count set by "framecnt" (optional, default:=
512)
> > > >
> > > > Signed-off-by: John W. Linville <linville@tuxdriver.com>
> > > > ---
> > > > This PMD is intended to provide a means for using DPDK on a broad
> > > > range of hardware without hardware-specific PMDs and (hopefully)
> > > > with better performance than what PCAP offers in Linux. This might
> > > > be useful as a development platform for DPDK applications when
> > > > DPDK-supported hardware is expensive or unavailable.
> > > >
> > > > New in v2:
> > > >
> > > > -- fixup some style issues found by check patch
> > > > -- use if_index as part of fanout group ID
> > > > -- set default number of queue pairs to 1
> > > >
> > > > config/common_bsdapp | 5 +
> > > > config/common_linuxapp | 5 +
> > > > lib/Makefile | 1 +
> > > > lib/librte_eal/linuxapp/eal/Makefile | 1 +
> > > > lib/librte_pmd_packet/Makefile | 60 +++
> > > > lib/librte_pmd_packet/rte_eth_packet.c | 826 +++++++++++++++++++++=
++++++++++++
> > > > lib/librte_pmd_packet/rte_eth_packet.h | 55 +++
> > > > mk/rte.app.mk | 4 +
> > > > 8 files changed, 957 insertions(+)
> > > > create mode 100644 lib/librte_pmd_packet/Makefile
> > > > create mode 100644 lib/librte_pmd_packet/rte_eth_packet.c
> > > > create mode 100644 lib/librte_pmd_packet/rte_eth_packet.h
> > > >
> > > > diff --git a/config/common_bsdapp b/config/common_bsdapp
> > > > index 943dce8f1ede..c317f031278e 100644
> > > > --- a/config/common_bsdapp
> > > > +++ b/config/common_bsdapp
> > > > @@ -226,6 +226,11 @@ CONFIG_RTE_LIBRTE_PMD_PCAP=3Dy
> > > > CONFIG_RTE_LIBRTE_PMD_BOND=3Dy
> > > >
> > > > #
> > > > +# Compile software PMD backed by AF_PACKET sockets (Linux only)
> > > > +#
> > > > +CONFIG_RTE_LIBRTE_PMD_PACKET=3Dn
> > > > +
> > > > +#
> > > > # Do prefetch of packet data within PMD driver receive function
> > > > #
> > > > CONFIG_RTE_PMD_PACKET_PREFETCH=3Dy
> > > > diff --git a/config/common_linuxapp b/config/common_linuxapp
> > > > index 7bf5d80d4e26..f9e7bc3015ec 100644
> > > > --- a/config/common_linuxapp
> > > > +++ b/config/common_linuxapp
> > > > @@ -249,6 +249,11 @@ CONFIG_RTE_LIBRTE_PMD_PCAP=3Dn
> > > > CONFIG_RTE_LIBRTE_PMD_BOND=3Dy
> > > >
> > > > #
> > > > +# Compile software PMD backed by AF_PACKET sockets (Linux only)
> > > > +#
> > > > +CONFIG_RTE_LIBRTE_PMD_PACKET=3Dy
> > > > +
> > > > +#
> > > > # Compile Xen PMD
> > > > #
> > > > CONFIG_RTE_LIBRTE_PMD_XENVIRT=3Dn
> > > > diff --git a/lib/Makefile b/lib/Makefile
> > > > index 10c5bb3045bc..930fadf29898 100644
> > > > --- a/lib/Makefile
> > > > +++ b/lib/Makefile
> > > > @@ -47,6 +47,7 @@ DIRS-$(CONFIG_RTE_LIBRTE_I40E_PMD) +=3D librte_pm=
d_i40e
> > > > DIRS-$(CONFIG_RTE_LIBRTE_PMD_BOND) +=3D librte_pmd_bond
> > > > DIRS-$(CONFIG_RTE_LIBRTE_PMD_RING) +=3D librte_pmd_ring
> > > > DIRS-$(CONFIG_RTE_LIBRTE_PMD_PCAP) +=3D librte_pmd_pcap
> > > > +DIRS-$(CONFIG_RTE_LIBRTE_PMD_PACKET) +=3D librte_pmd_packet
> > > > DIRS-$(CONFIG_RTE_LIBRTE_VIRTIO_PMD) +=3D librte_pmd_virtio
> > > > DIRS-$(CONFIG_RTE_LIBRTE_VMXNET3_PMD) +=3D librte_pmd_vmxnet3
> > > > DIRS-$(CONFIG_RTE_LIBRTE_PMD_XENVIRT) +=3D librte_pmd_xenvirt
> > > > diff --git a/lib/librte_eal/linuxapp/eal/Makefile b/lib/librte_eal/=
linuxapp/eal/Makefile
> > > > index 756d6b0c9301..feed24a63272 100644
> > > > --- a/lib/librte_eal/linuxapp/eal/Makefile
> > > > +++ b/lib/librte_eal/linuxapp/eal/Makefile
> > > > @@ -44,6 +44,7 @@ CFLAGS +=3D -I$(RTE_SDK)/lib/librte_ether
> > > > CFLAGS +=3D -I$(RTE_SDK)/lib/librte_ivshmem
> > > > CFLAGS +=3D -I$(RTE_SDK)/lib/librte_pmd_ring
> > > > CFLAGS +=3D -I$(RTE_SDK)/lib/librte_pmd_pcap
> > > > +CFLAGS +=3D -I$(RTE_SDK)/lib/librte_pmd_packet
> > > > CFLAGS +=3D -I$(RTE_SDK)/lib/librte_pmd_xenvirt
> > > > CFLAGS +=3D $(WERROR_FLAGS) -O3
> > > >
> > > > diff --git a/lib/librte_pmd_packet/Makefile b/lib/librte_pmd_packet=
/Makefile
> > > > new file mode 100644
> > > > index 000000000000..e1266fb992cd
> > > > --- /dev/null
> > > > +++ b/lib/librte_pmd_packet/Makefile
> > > > @@ -0,0 +1,60 @@
> > > > +# BSD LICENSE
> > > > +#
> > > > +# Copyright(c) 2014 John W. Linville <linville@redhat.com>
> > > > +# Copyright(c) 2010-2014 Intel Corporation. All rights reserved.
> > > > +# Copyright(c) 2014 6WIND S.A.
> > > > +# All rights reserved.
> > > > +#
> > > > +# Redistribution and use in source and binary forms, with or wit=
hout
> > > > +# modification, are permitted provided that the following condit=
ions
> > > > +# are met:
> > > > +#
> > > > +# * Redistributions of source code must retain the above copyr=
ight
> > > > +# notice, this list of conditions and the following disclaim=
er.
> > > > +# * Redistributions in binary form must reproduce the above co=
pyright
> > > > +# notice, this list of conditions and the following disclaim=
er in
> > > > +# the documentation and/or other materials provided with the
> > > > +# distribution.
> > > > +# * Neither the name of Intel Corporation nor the names of its
> > > > +# contributors may be used to endorse or promote products de=
rived
> > > > +# from this software without specific prior written permissi=
on.
> > > > +#
> > > > +# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIB=
UTORS
> > > > +# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT =
NOT
> > > > +# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITN=
ESS FOR
> > > > +# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COP=
YRIGHT
> > > > +# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCI=
DENTAL,
> > > > +# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT N=
OT
> > > > +# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS =
OF USE,
> > > > +# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND=
ON ANY
> > > > +# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR=
TORT
> > > > +# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF =
THE USE
> > > > +# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH D=
AMAGE.
> > > > +
> > > > +include $(RTE_SDK)/mk/rte.vars.mk
> > > > +
> > > > +#
> > > > +# library name
> > > > +#
> > > > +LIB =3D librte_pmd_packet.a
> > > > +
> > > > +CFLAGS +=3D -O3
> > > > +CFLAGS +=3D $(WERROR_FLAGS)
> > > > +
> > > > +#
> > > > +# all source are stored in SRCS-y
> > > > +#
> > > > +SRCS-$(CONFIG_RTE_LIBRTE_PMD_PACKET) +=3D rte_eth_packet.c
> > > > +
> > > > +#
> > > > +# Export include files
> > > > +#
> > > > +SYMLINK-y-include +=3D rte_eth_packet.h
> > > > +
> > > > +# this lib depends upon:
> > > > +DEPDIRS-$(CONFIG_RTE_LIBRTE_PMD_PACKET) +=3D lib/librte_mbuf
> > > > +DEPDIRS-$(CONFIG_RTE_LIBRTE_PMD_PACKET) +=3D lib/librte_ether
> > > > +DEPDIRS-$(CONFIG_RTE_LIBRTE_PMD_PACKET) +=3D lib/librte_malloc
> > > > +DEPDIRS-$(CONFIG_RTE_LIBRTE_PMD_PACKET) +=3D lib/librte_kvargs
> > > > +
> > > > +include $(RTE_SDK)/mk/rte.lib.mk
> > > > diff --git a/lib/librte_pmd_packet/rte_eth_packet.c b/lib/librte_pm=
d_packet/rte_eth_packet.c
> > > > new file mode 100644
> > > > index 000000000000..9c82d16e730f
> > > > --- /dev/null
> > > > +++ b/lib/librte_pmd_packet/rte_eth_packet.c
> > > > @@ -0,0 +1,826 @@
> > > > +/*-
> > > > + * BSD LICENSE
> > > > + *
> > > > + * Copyright(c) 2014 John W. Linville <linville@tuxdriver.com>
> > > > + *
> > > > + * Originally based upon librte_pmd_pcap code:
> > > > + *
> > > > + * Copyright(c) 2010-2014 Intel Corporation. All rights reserved=
.
> > > > + * Copyright(c) 2014 6WIND S.A.
> > > > + * All rights reserved.
> > > > + *
> > > > + * Redistribution and use in source and binary forms, with or wi=
thout
> > > > + * modification, are permitted provided that the following condi=
tions
> > > > + * are met:
> > > > + *
> > > > + * * Redistributions of source code must retain the above copy=
right
> > > > + * notice, this list of conditions and the following disclai=
mer.
> > > > + * * Redistributions in binary form must reproduce the above c=
opyright
> > > > + * notice, this list of conditions and the following disclai=
mer in
> > > > + * the documentation and/or other materials provided with th=
e
> > > > + * distribution.
> > > > + * * Neither the name of Intel Corporation nor the names of it=
s
> > > > + * contributors may be used to endorse or promote products d=
erived
> > > > + * from this software without specific prior written permiss=
ion.
> > > > + *
> > > > + * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRI=
BUTORS
> > > > + * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT=
NOT
> > > > + * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FIT=
NESS FOR
> > > > + * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE CO=
PYRIGHT
> > > > + * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INC=
IDENTAL,
> > > > + * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT =
NOT
> > > > + * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS=
OF USE,
> > > > + * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AN=
D ON ANY
> > > > + * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, O=
R TORT
> > > > + * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF=
THE USE
> > > > + * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH =
DAMAGE.
> > > > + */
> > > > +
> > > > +#include <rte_mbuf.h>
> > > > +#include <rte_ethdev.h>
> > > > +#include <rte_malloc.h>
> > > > +#include <rte_kvargs.h>
> > > > +#include <rte_dev.h>
> > > > +
> > > > +#include <linux/if_ether.h>
> > > > +#include <linux/if_packet.h>
> > > > +#include <arpa/inet.h>
> > > > +#include <net/if.h>
> > > > +#include <sys/types.h>
> > > > +#include <sys/socket.h>
> > > > +#include <sys/ioctl.h>
> > > > +#include <sys/mman.h>
> > > > +#include <unistd.h>
> > > > +#include <poll.h>
> > > > +
> > > > +#include "rte_eth_packet.h"
> > > > +
> > > > +#define ETH_PACKET_IFACE_ARG "iface"
> > > > +#define ETH_PACKET_NUM_Q_ARG "qpairs"
> > > > +#define ETH_PACKET_BLOCKSIZE_ARG "blocksz"
> > > > +#define ETH_PACKET_FRAMESIZE_ARG "framesz"
> > > > +#define ETH_PACKET_FRAMECOUNT_ARG "framecnt"
> > > > +
> > > > +#define DFLT_BLOCK_SIZE (1 << 12)
> > > > +#define DFLT_FRAME_SIZE (1 << 11)
> > > > +#define DFLT_FRAME_COUNT (1 << 9)
> > > > +
> > > > +struct pkt_rx_queue {
> > > > + int sockfd;
> > > > +
> > > > + struct iovec *rd;
> > > > + uint8_t *map;
> > > > + unsigned int framecount;
> > > > + unsigned int framenum;
> > > > +
> > > > + struct rte_mempool *mb_pool;
> > > > +
> > > > + volatile unsigned long rx_pkts;
> > > > + volatile unsigned long err_pkts;
> > > > +};
> > > > +
> > > > +struct pkt_tx_queue {
> > > > + int sockfd;
> > > > +
> > > > + struct iovec *rd;
> > > > + uint8_t *map;
> > > > + unsigned int framecount;
> > > > + unsigned int framenum;
> > > > +
> > > > + volatile unsigned long tx_pkts;
> > > > + volatile unsigned long err_pkts;
> > > > +};
> > > > +
> > > > +struct pmd_internals {
> > > > + unsigned nb_queues;
> > > > +
> > > > + int if_index;
> > > > + struct ether_addr eth_addr;
> > > > +
> > > > + struct tpacket_req req;
> > > > +
> > > > + struct pkt_rx_queue rx_queue[RTE_PMD_PACKET_MAX_RINGS];
> > > > + struct pkt_tx_queue tx_queue[RTE_PMD_PACKET_MAX_RINGS];
> > > > +};
> > > > +
> > > > +static const char *valid_arguments[] =3D {
> > > > + ETH_PACKET_IFACE_ARG,
> > > > + ETH_PACKET_NUM_Q_ARG,
> > > > + ETH_PACKET_BLOCKSIZE_ARG,
> > > > + ETH_PACKET_FRAMESIZE_ARG,
> > > > + ETH_PACKET_FRAMECOUNT_ARG,
> > > > + NULL
> > > > +};
> > > > +
> > > > +static const char *drivername =3D "AF_PACKET PMD";
> > > > +
> > > > +static struct rte_eth_link pmd_link =3D {
> > > > + .link_speed =3D 10000,
> > > > + .link_duplex =3D ETH_LINK_FULL_DUPLEX,
> > > > + .link_status =3D 0
> > > > +};
> > > > +
> > > > +static uint16_t
> > > > +eth_packet_rx(void *queue, struct rte_mbuf **bufs, uint16_t nb_pkt=
s)
> > > > +{
> > > > + unsigned i;
> > > > + struct tpacket2_hdr *ppd;
> > > > + struct rte_mbuf *mbuf;
> > > > + uint8_t *pbuf;
> > > > + struct pkt_rx_queue *pkt_q =3D queue;
> > > > + uint16_t num_rx =3D 0;
> > > > + unsigned int framecount, framenum;
> > > > +
> > > > + if (unlikely(nb_pkts =3D=3D 0))
> > > > + return 0;
> > > > +
> > > > + /*
> > > > + * Reads the given number of packets from the AF_PACKET socket on=
e by
> > > > + * one and copies the packet data into a newly allocated mbuf.
> > > > + */
> > > > + framecount =3D pkt_q->framecount;
> > > > + framenum =3D pkt_q->framenum;
> > > > + for (i =3D 0; i < nb_pkts; i++) {
> > > > + /* point at the next incoming frame */
> > > > + ppd =3D (struct tpacket2_hdr *) pkt_q->rd[framenum].iov_base;
> > > > + if ((ppd->tp_status & TP_STATUS_USER) =3D=3D 0)
> > > > + break;
> > > > +
> > > > + /* allocate the next mbuf */
> > > > + mbuf =3D rte_pktmbuf_alloc(pkt_q->mb_pool);
> > > > + if (unlikely(mbuf =3D=3D NULL))
> > > > + break;
> > > > +
> > > > + /* packet will fit in the mbuf, go ahead and receive it */
> > > > + mbuf->pkt.pkt_len =3D mbuf->pkt.data_len =3D ppd->tp_snaplen;
> > > > + pbuf =3D (uint8_t *) ppd + ppd->tp_mac;
> > > > + memcpy(mbuf->pkt.data, pbuf, mbuf->pkt.data_len);
> > > > +
> > > > + /* release incoming frame and advance ring buffer */
> > > > + ppd->tp_status =3D TP_STATUS_KERNEL;
> > > > + if (++framenum >=3D framecount)
> > > > + framenum =3D 0;
> > > > +
> > > > + /* account for the receive frame */
> > > > + bufs[i] =3D mbuf;
> > > > + num_rx++;
> > > > + }
> > > > + pkt_q->framenum =3D framenum;
> > > > + pkt_q->rx_pkts +=3D num_rx;
> > > > + return num_rx;
> > > > +}
> > > > +
> > > > +/*
> > > > + * Callback to handle sending packets through a real NIC.
> > > > + */
> > > > +static uint16_t
> > > > +eth_packet_tx(void *queue, struct rte_mbuf **bufs, uint16_t nb_pkt=
s)
> > > > +{
> > > > + struct tpacket2_hdr *ppd;
> > > > + struct rte_mbuf *mbuf;
> > > > + uint8_t *pbuf;
> > > > + unsigned int framecount, framenum;
> > > > + struct pollfd pfd;
> > > > + struct pkt_tx_queue *pkt_q =3D queue;
> > > > + uint16_t num_tx =3D 0;
> > > > + int i;
> > > > +
> > > > + if (unlikely(nb_pkts =3D=3D 0))
> > > > + return 0;
> > > > +
> > > > + memset(&pfd, 0, sizeof(pfd));
> > > > + pfd.fd =3D pkt_q->sockfd;
> > > > + pfd.events =3D POLLOUT;
> > > > + pfd.revents =3D 0;
> > > > +
> > > > + framecount =3D pkt_q->framecount;
> > > > + framenum =3D pkt_q->framenum;
> > > > + ppd =3D (struct tpacket2_hdr *) pkt_q->rd[framenum].iov_base;
> > > > + for (i =3D 0; i < nb_pkts; i++) {
> > > > + /* point at the next incoming frame */
> > > > + if ((ppd->tp_status !=3D TP_STATUS_AVAILABLE) &&
> > > > + (poll(&pfd, 1, -1) < 0))
> > > > + continue;
> > > > +
> > > > + /* copy the tx frame data */
> > > > + mbuf =3D bufs[num_tx];
> > > > + pbuf =3D (uint8_t *) ppd + TPACKET2_HDRLEN -
> > > > + sizeof(struct sockaddr_ll);
> > > > + memcpy(pbuf, mbuf->pkt.data, mbuf->pkt.data_len);
> > > > + ppd->tp_len =3D ppd->tp_snaplen =3D mbuf->pkt.data_len;
> > > > +
> > > > + /* release incoming frame and advance ring buffer */
> > > > + ppd->tp_status =3D TP_STATUS_SEND_REQUEST;
> > > > + if (++framenum >=3D framecount)
> > > > + framenum =3D 0;
> > > > + ppd =3D (struct tpacket2_hdr *) pkt_q->rd[framenum].iov_base;
> > > > +
> > > > + num_tx++;
> > > > + rte_pktmbuf_free(mbuf);
> > > > + }
> > > > +
> > > > + /* kick-off transmits */
> > > > + sendto(pkt_q->sockfd, NULL, 0, MSG_DONTWAIT, NULL, 0);
> > > > +
> > > > + pkt_q->framenum =3D framenum;
> > > > + pkt_q->tx_pkts +=3D num_tx;
> > > > + pkt_q->err_pkts +=3D nb_pkts - num_tx;
> > > > + return num_tx;
> > > > +}
> > > > +
> > > > +static int
> > > > +eth_dev_start(struct rte_eth_dev *dev)
> > > > +{
> > > > + dev->data->dev_link.link_status =3D 1;
> > > > + return 0;
> > > > +}
> > > > +
> > > > +/*
> > > > + * This function gets called when the current port gets stopped.
> > > > + */
> > > > +static void
> > > > +eth_dev_stop(struct rte_eth_dev *dev)
> > > > +{
> > > > + unsigned i;
> > > > + int sockfd;
> > > > + struct pmd_internals *internals =3D dev->data->dev_private;
> > > > +
> > > > + for (i =3D 0; i < internals->nb_queues; i++) {
> > > > + sockfd =3D internals->rx_queue[i].sockfd;
> > > > + if (sockfd !=3D -1)
> > > > + close(sockfd);
> > > > + sockfd =3D internals->tx_queue[i].sockfd;
> > > > + if (sockfd !=3D -1)
> > > > + close(sockfd);
> > > > + }
> > > > +
> > > > + dev->data->dev_link.link_status =3D 0;
> > > > +}
> > > > +
> > > > +static int
> > > > +eth_dev_configure(struct rte_eth_dev *dev __rte_unused)
> > > > +{
> > > > + return 0;
> > > > +}
> > > > +
> > > > +static void
> > > > +eth_dev_info(struct rte_eth_dev *dev, struct rte_eth_dev_info *dev=
_info)
> > > > +{
> > > > + struct pmd_internals *internals =3D dev->data->dev_private;
> > > > +
> > > > + dev_info->driver_name =3D drivername;
> > > > + dev_info->if_index =3D internals->if_index;
> > > > + dev_info->max_mac_addrs =3D 1;
> > > > + dev_info->max_rx_pktlen =3D (uint32_t)ETH_FRAME_LEN;
> > > > + dev_info->max_rx_queues =3D (uint16_t)internals->nb_queues;
> > > > + dev_info->max_tx_queues =3D (uint16_t)internals->nb_queues;
> > > > + dev_info->min_rx_bufsize =3D 0;
> > > > + dev_info->pci_dev =3D NULL;
> > > > +}
> > > > +
> > > > +static void
> > > > +eth_stats_get(struct rte_eth_dev *dev, struct rte_eth_stats *igb_s=
tats)
> > > > +{
> > > > + unsigned i, imax;
> > > > + unsigned long rx_total =3D 0, tx_total =3D 0, tx_err_total =3D 0;
> > > > + const struct pmd_internals *internal =3D dev->data->dev_private;
> > > > +
> > > > + memset(igb_stats, 0, sizeof(*igb_stats));
> > > > +
> > > > + imax =3D (internal->nb_queues < RTE_ETHDEV_QUEUE_STAT_CNTRS ?
> > > > + internal->nb_queues : RTE_ETHDEV_QUEUE_STAT_CNTRS);
> > > > + for (i =3D 0; i < imax; i++) {
> > > > + igb_stats->q_ipackets[i] =3D internal->rx_queue[i].rx_pkts;
> > > > + rx_total +=3D igb_stats->q_ipackets[i];
> > > > + }
> > > > +
> > > > + imax =3D (internal->nb_queues < RTE_ETHDEV_QUEUE_STAT_CNTRS ?
> > > > + internal->nb_queues : RTE_ETHDEV_QUEUE_STAT_CNTRS);
> > > > + for (i =3D 0; i < imax; i++) {
> > > > + igb_stats->q_opackets[i] =3D internal->tx_queue[i].tx_pkts;
> > > > + igb_stats->q_errors[i] =3D internal->tx_queue[i].err_pkts;
> > > > + tx_total +=3D igb_stats->q_opackets[i];
> > > > + tx_err_total +=3D igb_stats->q_errors[i];
> > > > + }
> > > > +
> > > > + igb_stats->ipackets =3D rx_total;
> > > > + igb_stats->opackets =3D tx_total;
> > > > + igb_stats->oerrors =3D tx_err_total;
> > > > +}
> > > > +
> > > > +static void
> > > > +eth_stats_reset(struct rte_eth_dev *dev)
> > > > +{
> > > > + unsigned i;
> > > > + struct pmd_internals *internal =3D dev->data->dev_private;
> > > > +
> > > > + for (i =3D 0; i < internal->nb_queues; i++)
> > > > + internal->rx_queue[i].rx_pkts =3D 0;
> > > > +
> > > > + for (i =3D 0; i < internal->nb_queues; i++) {
> > > > + internal->tx_queue[i].tx_pkts =3D 0;
> > > > + internal->tx_queue[i].err_pkts =3D 0;
> > > > + }
> > > > +}
> > > > +
> > > > +static void
> > > > +eth_dev_close(struct rte_eth_dev *dev __rte_unused)
> > > > +{
> > > > +}
> > > > +
> > > > +static void
> > > > +eth_queue_release(void *q __rte_unused)
> > > > +{
> > > > +}
> > > > +
> > > > +static int
> > > > +eth_link_update(struct rte_eth_dev *dev __rte_unused,
> > > > + int wait_to_complete __rte_unused)
> > > > +{
> > > > + return 0;
> > > > +}
> > > > +
> > > > +static int
> > > > +eth_rx_queue_setup(struct rte_eth_dev *dev,
> > > > + uint16_t rx_queue_id,
> > > > + uint16_t nb_rx_desc __rte_unused,
> > > > + unsigned int socket_id __rte_unused,
> > > > + const struct rte_eth_rxconf *rx_conf __rte_unus=
ed,
> > > > + struct rte_mempool *mb_pool)
> > > > +{
> > > > + struct pmd_internals *internals =3D dev->data->dev_private;
> > > > + struct pkt_rx_queue *pkt_q =3D &internals->rx_queue[rx_queue_id];
> > > > + struct rte_pktmbuf_pool_private *mbp_priv;
> > > > + uint16_t buf_size;
> > > > +
> > > > + pkt_q->mb_pool =3D mb_pool;
> > > > +
> > > > + /* Now get the space available for data in the mbuf */
> > > > + mbp_priv =3D rte_mempool_get_priv(pkt_q->mb_pool);
> > > > + buf_size =3D (uint16_t) (mbp_priv->mbuf_data_room_size -
> > > > + RTE_PKTMBUF_HEADROOM);
> > > > +
> > > > + if (ETH_FRAME_LEN > buf_size) {
> > > > + RTE_LOG(ERR, PMD,
> > > > + "%s: %d bytes will not fit in mbuf (%d bytes)\n",
> > > > + dev->data->name, ETH_FRAME_LEN, buf_size);
> > > > + return -ENOMEM;
> > > > + }
> > > > +
> > > > + dev->data->rx_queues[rx_queue_id] =3D pkt_q;
> > > > +
> > > > + return 0;
> > > > +}
> > > > +
> > > > +static int
> > > > +eth_tx_queue_setup(struct rte_eth_dev *dev,
> > > > + uint16_t tx_queue_id,
> > > > + uint16_t nb_tx_desc __rte_unused,
> > > > + unsigned int socket_id __rte_unused,
> > > > + const struct rte_eth_txconf *tx_conf __rte_unus=
ed)
> > > > +{
> > > > +
> > > > + struct pmd_internals *internals =3D dev->data->dev_private;
> > > > +
> > > > + dev->data->tx_queues[tx_queue_id] =3D &internals->tx_queue[tx_que=
ue_id];
> > > > + return 0;
> > > > +}
> > > > +
> > > > +static struct eth_dev_ops ops =3D {
> > > > + .dev_start =3D eth_dev_start,
> > > > + .dev_stop =3D eth_dev_stop,
> > > > + .dev_close =3D eth_dev_close,
> > > > + .dev_configure =3D eth_dev_configure,
> > > > + .dev_infos_get =3D eth_dev_info,
> > > > + .rx_queue_setup =3D eth_rx_queue_setup,
> > > > + .tx_queue_setup =3D eth_tx_queue_setup,
> > > > + .rx_queue_release =3D eth_queue_release,
> > > > + .tx_queue_release =3D eth_queue_release,
> > > > + .link_update =3D eth_link_update,
> > > > + .stats_get =3D eth_stats_get,
> > > > + .stats_reset =3D eth_stats_reset,
> > > > +};
> > > > +
> > > > +/*
> > > > + * Opens an AF_PACKET socket
> > > > + */
> > > > +static int
> > > > +open_packet_iface(const char *key __rte_unused,
> > > > + const char *value __rte_unused,
> > > > + void *extra_args)
> > > > +{
> > > > + int *sockfd =3D extra_args;
> > > > +
> > > > + /* Open an AF_PACKET socket... */
> > > > + *sockfd =3D socket(AF_PACKET, SOCK_RAW, htons(ETH_P_ALL));
> > > > + if (*sockfd =3D=3D -1) {
> > > > + RTE_LOG(ERR, PMD, "Could not open AF_PACKET socket\n");
> > > > + return -1;
> > > > + }
> > > > +
> > > > + return 0;
> > > > +}
> > > > +
> > > > +static int
> > > > +rte_pmd_init_internals(const char *name,
> > > > + const int sockfd,
> > > > + const unsigned nb_queues,
> > > > + unsigned int blocksize,
> > > > + unsigned int blockcnt,
> > > > + unsigned int framesize,
> > > > + unsigned int framecnt,
> > > > + const unsigned numa_node,
> > > > + struct pmd_internals **internals,
> > > > + struct rte_eth_dev **eth_dev,
> > > > + struct rte_kvargs *kvlist)
> > > > +{
> > > > + struct rte_eth_dev_data *data =3D NULL;
> > > > + struct rte_pci_device *pci_dev =3D NULL;
> > > > + struct rte_kvargs_pair *pair =3D NULL;
> > > > + struct ifreq ifr;
> > > > + size_t ifnamelen;
> > > > + unsigned k_idx;
> > > > + struct sockaddr_ll sockaddr;
> > > > + struct tpacket_req *req;
> > > > + struct pkt_rx_queue *rx_queue;
> > > > + struct pkt_tx_queue *tx_queue;
> > > > + int rc, tpver, discard, bypass;
> > > > + unsigned int i, q, rdsize;
> > > > + int qsockfd, fanout_arg;
> > > > +
> > > > + for (k_idx =3D 0; k_idx < kvlist->count; k_idx++) {
> > > > + pair =3D &kvlist->pairs[k_idx];
> > > > + if (strstr(pair->key, ETH_PACKET_IFACE_ARG) !=3D NULL)
> > > > + break;
> > > > + }
> > > > + if (pair =3D=3D NULL) {
> > > > + RTE_LOG(ERR, PMD,
> > > > + "%s: no interface specified for AF_PACKET ethdev\n",
> > > > + name);
> > > > + goto error;
> > > > + }
> > > > +
> > > > + RTE_LOG(INFO, PMD,
> > > > + "%s: creating AF_PACKET-backed ethdev on numa socket %u\n",
> > > > + name, numa_node);
> > > > +
> > > > + /*
> > > > + * now do all data allocation - for eth_dev structure, dummy pci =
driver
> > > > + * and internal (private) data
> > > > + */
> > > > + data =3D rte_zmalloc_socket(name, sizeof(*data), 0, numa_node);
> > > > + if (data =3D=3D NULL)
> > > > + goto error;
> > > > +
> > > > + pci_dev =3D rte_zmalloc_socket(name, sizeof(*pci_dev), 0, numa_no=
de);
> > > > + if (pci_dev =3D=3D NULL)
> > > > + goto error;
> > > > +
> > > > + *internals =3D rte_zmalloc_socket(name, sizeof(**internals),
> > > > + 0, numa_node);
> > > > + if (*internals =3D=3D NULL)
> > > > + goto error;
> > > > +
> > > > + req =3D &((*internals)->req);
> > > > +
> > > > + req->tp_block_size =3D blocksize;
> > > > + req->tp_block_nr =3D blockcnt;
> > > > + req->tp_frame_size =3D framesize;
> > > > + req->tp_frame_nr =3D framecnt;
> > > > +
> > > > + ifnamelen =3D strlen(pair->value);
> > > > + if (ifnamelen < sizeof(ifr.ifr_name)) {
> > > > + memcpy(ifr.ifr_name, pair->value, ifnamelen);
> > > > + ifr.ifr_name[ifnamelen] =3D '\0';
> > > > + } else {
> > > > + RTE_LOG(ERR, PMD,
> > > > + "%s: I/F name too long (%s)\n",
> > > > + name, pair->value);
> > > > + goto error;
> > > > + }
> > > > + if (ioctl(sockfd, SIOCGIFINDEX, &ifr) =3D=3D -1) {
> > > > + RTE_LOG(ERR, PMD,
> > > > + "%s: ioctl failed (SIOCGIFINDEX)\n",
> > > > + name);
> > > > + goto error;
> > > > + }
> > > > + (*internals)->if_index =3D ifr.ifr_ifindex;
> > > > +
> > > > + if (ioctl(sockfd, SIOCGIFHWADDR, &ifr) =3D=3D -1) {
> > > > + RTE_LOG(ERR, PMD,
> > > > + "%s: ioctl failed (SIOCGIFHWADDR)\n",
> > > > + name);
> > > > + goto error;
> > > > + }
> > > > + memcpy(&(*internals)->eth_addr, ifr.ifr_hwaddr.sa_data, ETH_ALEN)=
;
> > > > +
> > > > + memset(&sockaddr, 0, sizeof(sockaddr));
> > > > + sockaddr.sll_family =3D AF_PACKET;
> > > > + sockaddr.sll_protocol =3D htons(ETH_P_ALL);
> > > > + sockaddr.sll_ifindex =3D (*internals)->if_index;
> > > > +
> > > > + fanout_arg =3D (getpid() ^ (*internals)->if_index) & 0xffff;
> > > > + fanout_arg |=3D (PACKET_FANOUT_HASH | PACKET_FANOUT_FLAG_DEFRAG |
> > > > + PACKET_FANOUT_FLAG_ROLLOVER) << 16;
> > > > +
> > > > + for (q =3D 0; q < nb_queues; q++) {
> > > > + /* Open an AF_PACKET socket for this queue... */
> > > > + qsockfd =3D socket(AF_PACKET, SOCK_RAW, htons(ETH_P_ALL));
> > > > + if (qsockfd =3D=3D -1) {
> > > > + RTE_LOG(ERR, PMD,
> > > > + "%s: could not open AF_PACKET socket\n",
> > > > + name);
> > > > + return -1;
> > > > + }
> > > > +
> > > > + tpver =3D TPACKET_V2;
> > > > + rc =3D setsockopt(qsockfd, SOL_PACKET, PACKET_VERSION,
> > > > + &tpver, sizeof(tpver));
> > > > + if (rc =3D=3D -1) {
> > > > + RTE_LOG(ERR, PMD,
> > > > + "%s: could not set PACKET_VERSION on AF_PACKET "
> > > > + "socket for %s\n", name, pair->value);
> > > > + goto error;
> > > > + }
> > > > +
> > > > + discard =3D 1;
> > > > + rc =3D setsockopt(qsockfd, SOL_PACKET, PACKET_LOSS,
> > > > + &discard, sizeof(discard));
> > > > + if (rc =3D=3D -1) {
> > > > + RTE_LOG(ERR, PMD,
> > > > + "%s: could not set PACKET_LOSS on "
> > > > + "AF_PACKET socket for %s\n", name, pair->value);
> > > > + goto error;
> > > > + }
> > > > +
> > > > + bypass =3D 1;
> > > > + rc =3D setsockopt(qsockfd, SOL_PACKET, PACKET_QDISC_BYPASS,
> > > > + &bypass, sizeof(bypass));
> > > > + if (rc =3D=3D -1) {
> > > > + RTE_LOG(ERR, PMD,
> > > > + "%s: could not set PACKET_QDISC_BYPASS "
> > > > + "on AF_PACKET socket for %s\n", name,
> > > > + pair->value);
> > > > + goto error;
> > > > + }
> > > > +
> > > > + rc =3D setsockopt(qsockfd, SOL_PACKET, PACKET_RX_RING, req, size=
of(*req));
> > > > + if (rc =3D=3D -1) {
> > > > + RTE_LOG(ERR, PMD,
> > > > + "%s: could not set PACKET_RX_RING on AF_PACKET "
> > > > + "socket for %s\n", name, pair->value);
> > > > + goto error;
> > > > + }
> > > > +
> > > > + rc =3D setsockopt(qsockfd, SOL_PACKET, PACKET_TX_RING, req, size=
of(*req));
> > > > + if (rc =3D=3D -1) {
> > > > + RTE_LOG(ERR, PMD,
> > > > + "%s: could not set PACKET_TX_RING on AF_PACKET "
> > > > + "socket for %s\n", name, pair->value);
> > > > + goto error;
> > > > + }
> > > > +
> > > > + rx_queue =3D &((*internals)->rx_queue[q]);
> > > > + rx_queue->framecount =3D req->tp_frame_nr;
> > > > +
> > > > + rx_queue->map =3D mmap(NULL, 2 * req->tp_block_size * req->tp_bl=
ock_nr,
> > > > + PROT_READ | PROT_WRITE, MAP_SHARED | MAP_LOCKED,
> > > > + qsockfd, 0);
> > > > + if (rx_queue->map =3D=3D MAP_FAILED) {
> > > > + RTE_LOG(ERR, PMD,
> > > > + "%s: call to mmap failed on AF_PACKET socket for %s\n",
> > > > + name, pair->value);
> > > > + goto error;
> > > > + }
> > > > +
> > > > + /* rdsize is same for both Tx and Rx */
> > > > + rdsize =3D req->tp_frame_nr * sizeof(*(rx_queue->rd));
> > > > +
> > > > + rx_queue->rd =3D rte_zmalloc_socket(name, rdsize, 0, numa_node);
> > > > + for (i =3D 0; i < req->tp_frame_nr; ++i) {
> > > > + rx_queue->rd[i].iov_base =3D rx_queue->map + (i * framesize);
> > > > + rx_queue->rd[i].iov_len =3D req->tp_frame_size;
> > > > + }
> > > > + rx_queue->sockfd =3D qsockfd;
> > > > +
> > > > + tx_queue =3D &((*internals)->tx_queue[q]);
> > > > + tx_queue->framecount =3D req->tp_frame_nr;
> > > > +
> > > > + tx_queue->map =3D rx_queue->map + req->tp_block_size * req->tp_b=
lock_nr;
> > > > +
> > > > + tx_queue->rd =3D rte_zmalloc_socket(name, rdsize, 0, numa_node);
> > > > + for (i =3D 0; i < req->tp_frame_nr; ++i) {
> > > > + tx_queue->rd[i].iov_base =3D tx_queue->map + (i * framesize);
> > > > + tx_queue->rd[i].iov_len =3D req->tp_frame_size;
> > > > + }
> > > > + tx_queue->sockfd =3D qsockfd;
> > > > +
> > > > + rc =3D bind(qsockfd, (const struct sockaddr*)&sockaddr, sizeof(s=
ockaddr));
> > > > + if (rc =3D=3D -1) {
> > > > + RTE_LOG(ERR, PMD,
> > > > + "%s: could not bind AF_PACKET socket to %s\n",
> > > > + name, pair->value);
> > > > + goto error;
> > > > + }
> > > > +
> > > > + rc =3D setsockopt(qsockfd, SOL_PACKET, PACKET_FANOUT,
> > > > + &fanout_arg, sizeof(fanout_arg));
> > > > + if (rc =3D=3D -1) {
> > > > + RTE_LOG(ERR, PMD,
> > > > + "%s: could not set PACKET_FANOUT on AF_PACKET socket "
> > > > + "for %s\n", name, pair->value);
> > > > + goto error;
> > > > + }
> > > > + }
> > > > +
> > > > + /* reserve an ethdev entry */
> > > > + *eth_dev =3D rte_eth_dev_allocate(name);
> > > > + if (*eth_dev =3D=3D NULL)
> > > > + goto error;
> > > > +
> > > > + /*
> > > > + * now put it all together
> > > > + * - store queue data in internals,
> > > > + * - store numa_node info in pci_driver
> > > > + * - point eth_dev_data to internals and pci_driver
> > > > + * - and point eth_dev structure to new eth_dev_data structure
> > > > + */
> > > > +
> > > > + (*internals)->nb_queues =3D nb_queues;
> > > > +
> > > > + data->dev_private =3D *internals;
> > > > + data->port_id =3D (*eth_dev)->data->port_id;
> > > > + data->nb_rx_queues =3D (uint16_t)nb_queues;
> > > > + data->nb_tx_queues =3D (uint16_t)nb_queues;
> > > > + data->dev_link =3D pmd_link;
> > > > + data->mac_addrs =3D &(*internals)->eth_addr;
> > > > +
> > > > + pci_dev->numa_node =3D numa_node;
> > > > +
> > > > + (*eth_dev)->data =3D data;
> > > > + (*eth_dev)->dev_ops =3D &ops;
> > > > + (*eth_dev)->pci_dev =3D pci_dev;
> > > > +
> > > > + return 0;
> > > > +
> > > > +error:
> > > > + if (data)
> > > > + rte_free(data);
> > > > + if (pci_dev)
> > > > + rte_free(pci_dev);
> > > > + for (q =3D 0; q < nb_queues; q++) {
> > > > + if ((*internals)->rx_queue[q].rd)
> > > > + rte_free((*internals)->rx_queue[q].rd);
> > > > + if ((*internals)->tx_queue[q].rd)
> > > > + rte_free((*internals)->tx_queue[q].rd);
> > > > + }
> > > > + if (*internals)
> > > > + rte_free(*internals);
> > > > + return -1;
> > > > +}
> > > > +
> > > > +static int
> > > > +rte_eth_from_packet(const char *name,
> > > > + int const *sockfd,
> > > > + const unsigned numa_node,
> > > > + struct rte_kvargs *kvlist)
> > > > +{
> > > > + struct pmd_internals *internals =3D NULL;
> > > > + struct rte_eth_dev *eth_dev =3D NULL;
> > > > + struct rte_kvargs_pair *pair =3D NULL;
> > > > + unsigned k_idx;
> > > > + unsigned int blockcount;
> > > > + unsigned int blocksize =3D DFLT_BLOCK_SIZE;
> > > > + unsigned int framesize =3D DFLT_FRAME_SIZE;
> > > > + unsigned int framecount =3D DFLT_FRAME_COUNT;
> > > > + unsigned int qpairs =3D 1;
> > > > +
> > > > + /* do some parameter checking */
> > > > + if (*sockfd < 0)
> > > > + return -1;
> > > > +
> > > > + /*
> > > > + * Walk arguments for configurable settings
> > > > + */
> > > > + for (k_idx =3D 0; k_idx < kvlist->count; k_idx++) {
> > > > + pair =3D &kvlist->pairs[k_idx];
> > > > + if (strstr(pair->key, ETH_PACKET_NUM_Q_ARG) !=3D NULL) {
> > > > + qpairs =3D atoi(pair->value);
> > > > + if (qpairs < 1 ||
> > > > + qpairs > RTE_PMD_PACKET_MAX_RINGS) {
> > > > + RTE_LOG(ERR, PMD,
> > > > + "%s: invalid qpairs value\n",
> > > > + name);
> > > > + return -1;
> > > > + }
> > > > + continue;
> > > > + }
> > > > + if (strstr(pair->key, ETH_PACKET_BLOCKSIZE_ARG) !=3D NULL) {
> > > > + blocksize =3D atoi(pair->value);
> > > > + if (!blocksize) {
> > > > + RTE_LOG(ERR, PMD,
> > > > + "%s: invalid blocksize value\n",
> > > > + name);
> > > > + return -1;
> > > > + }
> > > > + continue;
> > > > + }
> > > > + if (strstr(pair->key, ETH_PACKET_FRAMESIZE_ARG) !=3D NULL) {
> > > > + framesize =3D atoi(pair->value);
> > > > + if (!framesize) {
> > > > + RTE_LOG(ERR, PMD,
> > > > + "%s: invalid framesize value\n",
> > > > + name);
> > > > + return -1;
> > > > + }
> > > > + continue;
> > > > + }
> > > > + if (strstr(pair->key, ETH_PACKET_FRAMECOUNT_ARG) !=3D NULL) {
> > > > + framecount =3D atoi(pair->value);
> > > > + if (!framecount) {
> > > > + RTE_LOG(ERR, PMD,
> > > > + "%s: invalid framecount value\n",
> > > > + name);
> > > > + return -1;
> > > > + }
> > > > + continue;
> > > > + }
> > > > + }
> > > > +
> > > > + if (framesize > blocksize) {
> > > > + RTE_LOG(ERR, PMD,
> > > > + "%s: AF_PACKET MMAP frame size exceeds block size!\n",
> > > > + name);
> > > > + return -1;
> > > > + }
> > > > +
> > > > + blockcount =3D framecount / (blocksize / framesize);
> > > > + if (!blockcount) {
> > > > + RTE_LOG(ERR, PMD,
> > > > + "%s: invalid AF_PACKET MMAP parameters\n", name);
> > > > + return -1;
> > > > + }
> > > > +
> > > > + RTE_LOG(INFO, PMD, "%s: AF_PACKET MMAP parameters:\n", name);
> > > > + RTE_LOG(INFO, PMD, "%s:\tblock size %d\n", name, blocksize);
> > > > + RTE_LOG(INFO, PMD, "%s:\tblock count %d\n", name, blockcount);
> > > > + RTE_LOG(INFO, PMD, "%s:\tframe size %d\n", name, framesize);
> > > > + RTE_LOG(INFO, PMD, "%s:\tframe count %d\n", name, framecount);
> > > > +
> > > > + if (rte_pmd_init_internals(name, *sockfd, qpairs,
> > > > + blocksize, blockcount,
> > > > + framesize, framecount,
> > > > + numa_node, &internals, ð_dev,
> > > > + kvlist) < 0)
> > > > + return -1;
> > > > +
> > > > + eth_dev->rx_pkt_burst =3D eth_packet_rx;
> > > > + eth_dev->tx_pkt_burst =3D eth_packet_tx;
> > > > +
> > > > + return 0;
> > > > +}
> > > > +
> > > > +int
> > > > +rte_pmd_packet_devinit(const char *name, const char *params)
> > > > +{
> > > > + unsigned numa_node;
> > > > + int ret;
> > > > + struct rte_kvargs *kvlist;
> > > > + int sockfd =3D -1;
> > > > +
> > > > + RTE_LOG(INFO, PMD, "Initializing pmd_packet for %s\n", name);
> > > > +
> > > > + numa_node =3D rte_socket_id();
> > > > +
> > > > + kvlist =3D rte_kvargs_parse(params, valid_arguments);
> > > > + if (kvlist =3D=3D NULL)
> > > > + return -1;
> > > > +
> > > > + /*
> > > > + * If iface argument is passed we open the NICs and use them for
> > > > + * reading / writing
> > > > + */
> > > > + if (rte_kvargs_count(kvlist, ETH_PACKET_IFACE_ARG) =3D=3D 1) {
> > > > +
> > > > + ret =3D rte_kvargs_process(kvlist, ETH_PACKET_IFACE_ARG,
> > > > + &open_packet_iface, &sockfd);
> > > > + if (ret < 0)
> > > > + return -1;
> > > > + }
> > > > +
> > > > + ret =3D rte_eth_from_packet(name, &sockfd, numa_node, kvlist);
> > > > + close(sockfd); /* no longer needed */
> > > > +
> > > > + if (ret < 0)
> > > > + return -1;
> > > > +
> > > > + return 0;
> > > > +}
> > > > +
> > > > +static struct rte_driver pmd_packet_drv =3D {
> > > > + .name =3D "eth_packet",
> > > > + .type =3D PMD_VDEV,
> > > > + .init =3D rte_pmd_packet_devinit,
> > > > +};
> > > > +
> > > > +PMD_REGISTER_DRIVER(pmd_packet_drv);
> > > > diff --git a/lib/librte_pmd_packet/rte_eth_packet.h b/lib/librte_pm=
d_packet/rte_eth_packet.h
> > > > new file mode 100644
> > > > index 000000000000..f685611da3e9
> > > > --- /dev/null
> > > > +++ b/lib/librte_pmd_packet/rte_eth_packet.h
> > > > @@ -0,0 +1,55 @@
> > > > +/*-
> > > > + * BSD LICENSE
> > > > + *
> > > > + * Copyright(c) 2010-2014 Intel Corporation. All rights reserved=
.
> > > > + * All rights reserved.
> > > > + *
> > > > + * Redistribution and use in source and binary forms, with or wi=
thout
> > > > + * modification, are permitted provided that the following condi=
tions
> > > > + * are met:
> > > > + *
> > > > + * * Redistributions of source code must retain the above copy=
right
> > > > + * notice, this list of conditions and the following disclai=
mer.
> > > > + * * Redistributions in binary form must reproduce the above c=
opyright
> > > > + * notice, this list of conditions and the following disclai=
mer in
> > > > + * the documentation and/or other materials provided with th=
e
> > > > + * distribution.
> > > > + * * Neither the name of Intel Corporation nor the names of it=
s
> > > > + * contributors may be used to endorse or promote products d=
erived
> > > > + * from this software without specific prior written permiss=
ion.
> > > > + *
> > > > + * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRI=
BUTORS
> > > > + * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT=
NOT
> > > > + * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FIT=
NESS FOR
> > > > + * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE CO=
PYRIGHT
> > > > + * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INC=
IDENTAL,
> > > > + * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT =
NOT
> > > > + * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS=
OF USE,
> > > > + * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AN=
D ON ANY
> > > > + * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, O=
R TORT
> > > > + * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF=
THE USE
> > > > + * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH =
DAMAGE.
> > > > + */
> > > > +
> > > > +#ifndef _RTE_ETH_PACKET_H_
> > > > +#define _RTE_ETH_PACKET_H_
> > > > +
> > > > +#ifdef __cplusplus
> > > > +extern "C" {
> > > > +#endif
> > > > +
> > > > +#define RTE_ETH_PACKET_PARAM_NAME "eth_packet"
> > > > +
> > > > +#define RTE_PMD_PACKET_MAX_RINGS 16
> > > > +
> > > > +/**
> > > > + * For use by the EAL only. Called as part of EAL init to set up a=
ny dummy NICs
> > > > + * configured on command line.
> > > > + */
> > > > +int rte_pmd_packet_devinit(const char *name, const char *params);
> > > > +
> > > > +#ifdef __cplusplus
> > > > +}
> > > > +#endif
> > > > +
> > > > +#endif
> > > > diff --git a/mk/rte.app.mk b/mk/rte.app.mk
> > > > index 34dff2a02a05..a6994c4dbe93 100644
> > > > --- a/mk/rte.app.mk
> > > > +++ b/mk/rte.app.mk
> > > > @@ -210,6 +210,10 @@ ifeq ($(CONFIG_RTE_LIBRTE_PMD_PCAP),y)
> > > > LDLIBS +=3D -lrte_pmd_pcap -lpcap
> > > > endif
> > > >
> > > > +ifeq ($(CONFIG_RTE_LIBRTE_PMD_PACKET),y)
> > > > +LDLIBS +=3D -lrte_pmd_packet
> > > > +endif
> > > > +
> > > > endif # plugins
> > > >
> > > > LDLIBS +=3D $(EXECENV_LDLIBS)
> > > > --
> > > > 1.9.3
> > > >
> > > >
> > >
> > > --
> > > John W. Linville Someday the world will need a hero, and you
> > > linville@tuxdriver.com might be all we have. Be ready.
> >
>=20
> --
> John W. Linville Someday the world will need a hero, and you
> linville@tuxdriver.com might be all we have. Be ready.