[dpdk-dev] [PATCH v7 1/3] lib: add Generic Receive Offload API framework
Jiayu Hu
jiayu.hu at intel.com
Wed Jun 28 04:17:10 CEST 2017
Hi Konstantin,
On Wed, Jun 28, 2017 at 07:42:01AM +0800, Ananyev, Konstantin wrote:
> Hi Jiayu,
>
> > -----Original Message-----
> > From: Hu, Jiayu
> > Sent: Monday, June 26, 2017 7:44 AM
> > To: dev at dpdk.org
> > Cc: Tan, Jianfeng <jianfeng.tan at intel.com>; Ananyev, Konstantin <konstantin.ananyev at intel.com>; stephen at networkplumber.org;
> > yliu at fridaylinux.org; Wu, Jingjing <jingjing.wu at intel.com>; Yao, Lei A <lei.a.yao at intel.com>; Wiles, Keith <keith.wiles at intel.com>; Bie,
> > Tiwei <tiwei.bie at intel.com>; Hu, Jiayu <jiayu.hu at intel.com>
> > Subject: [PATCH v7 1/3] lib: add Generic Receive Offload API framework
> >
> > Generic Receive Offload (GRO) is a widely used SW-based offloading
> > technique to reduce per-packet processing overhead. It gains
> > performance by reassembling small packets into large ones. This
> > patchset is to support GRO in DPDK. To support GRO, this patch
> > implements a GRO API framework.
> >
> > To enable more flexibility to applications, DPDK GRO is implemented as
> > a user library. Applications explicitly use the GRO library to merge
> > small packets into large ones. DPDK GRO provides two reassembly modes.
> > One is called lightweigth mode, the other is called heavyweight mode.
> > If applications want to merge packets in a simple way and the number
> > of packets is relatively small, they can use the lightweigth mode.
> > If applications need more fine-grained controls, they can choose the
> > heavyweigth mode.
> >
> > rte_gro_reassemble_burst is the main reassembly API which is used in
> > lightweigth mode and processes N packets at a time. For applications,
> > performing GRO in lightweigth mode is simple. They just need to invoke
> > rte_gro_reassemble_burst. Applications can get GROed packets as soon as
> > rte_gro_reassemble_burst returns.
> >
> > rte_gro_reassemble is the main reassembly API which is used in
> > heavyweight mode and processes one packet at a time. For applications,
> > performing GRO in heavyweigth mode is relatively complicated. Before
> > performing GRO, applications need to create a GRO table by
> > rte_gro_tbl_create. Then they can use rte_gro_reassemble to process
> > packets one by one. The processed packets are in the GRO table. If
> > applications want to get them, applications need to manually flush
> > them by flush APIs.
> >
> > Signed-off-by: Jiayu Hu <jiayu.hu at intel.com>
> > ---
> > config/common_base | 5 +
> > lib/Makefile | 2 +
> > lib/librte_gro/Makefile | 50 +++++++++
> > lib/librte_gro/rte_gro.c | 125 ++++++++++++++++++++++
> > lib/librte_gro/rte_gro.h | 205 +++++++++++++++++++++++++++++++++++++
> > lib/librte_gro/rte_gro_version.map | 12 +++
> > mk/rte.app.mk | 1 +
> > 7 files changed, 400 insertions(+)
> > create mode 100644 lib/librte_gro/Makefile
> > create mode 100644 lib/librte_gro/rte_gro.c
> > create mode 100644 lib/librte_gro/rte_gro.h
> > create mode 100644 lib/librte_gro/rte_gro_version.map
> >
> > diff --git a/config/common_base b/config/common_base
> > index f6aafd1..167f5ef 100644
> > --- a/config/common_base
> > +++ b/config/common_base
> > @@ -712,6 +712,11 @@ CONFIG_RTE_LIBRTE_VHOST_DEBUG=n
> > CONFIG_RTE_LIBRTE_PMD_VHOST=n
> >
> > #
> > +# Compile GRO library
> > +#
> > +CONFIG_RTE_LIBRTE_GRO=y
> > +
> > +#
> > #Compile Xen domain0 support
> > #
> > CONFIG_RTE_LIBRTE_XEN_DOM0=n
> > diff --git a/lib/Makefile b/lib/Makefile
> > index 07e1fd0..ac1c2f6 100644
> > --- a/lib/Makefile
> > +++ b/lib/Makefile
> > @@ -106,6 +106,8 @@ DIRS-$(CONFIG_RTE_LIBRTE_REORDER) += librte_reorder
> > DEPDIRS-librte_reorder := librte_eal librte_mempool librte_mbuf
> > DIRS-$(CONFIG_RTE_LIBRTE_PDUMP) += librte_pdump
> > DEPDIRS-librte_pdump := librte_eal librte_mempool librte_mbuf librte_ether
> > +DIRS-$(CONFIG_RTE_LIBRTE_GRO) += librte_gro
> > +DEPDIRS-librte_gro := librte_eal librte_mbuf librte_ether librte_net
> >
> > ifeq ($(CONFIG_RTE_EXEC_ENV_LINUXAPP),y)
> > DIRS-$(CONFIG_RTE_LIBRTE_KNI) += librte_kni
> > diff --git a/lib/librte_gro/Makefile b/lib/librte_gro/Makefile
> > new file mode 100644
> > index 0000000..7e0f128
> > --- /dev/null
> > +++ b/lib/librte_gro/Makefile
> > @@ -0,0 +1,50 @@
> > +# BSD LICENSE
> > +#
> > +# Copyright(c) 2017 Intel Corporation. All rights reserved.
> > +# All rights reserved.
> > +#
> > +# Redistribution and use in source and binary forms, with or without
> > +# modification, are permitted provided that the following conditions
> > +# are met:
> > +#
> > +# * Redistributions of source code must retain the above copyright
> > +# notice, this list of conditions and the following disclaimer.
> > +# * Redistributions in binary form must reproduce the above copyright
> > +# notice, this list of conditions and the following disclaimer 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 derived
> > +# from this software without specific prior written permission.
> > +#
> > +# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
> > +# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
> > +# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
> > +# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
> > +# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
> > +# 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 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 DAMAGE.
> > +
> > +include $(RTE_SDK)/mk/rte.vars.mk
> > +
> > +# library name
> > +LIB = librte_gro.a
> > +
> > +CFLAGS += -O3
> > +CFLAGS += $(WERROR_FLAGS) -I$(SRCDIR)
> > +
> > +EXPORT_MAP := rte_gro_version.map
> > +
> > +LIBABIVER := 1
> > +
> > +# source files
> > +SRCS-$(CONFIG_RTE_LIBRTE_GRO) += rte_gro.c
> > +
> > +# install this header file
> > +SYMLINK-$(CONFIG_RTE_LIBRTE_GRO)-include += rte_gro.h
> > +
> > +include $(RTE_SDK)/mk/rte.lib.mk
> > diff --git a/lib/librte_gro/rte_gro.c b/lib/librte_gro/rte_gro.c
> > new file mode 100644
> > index 0000000..33275e8
> > --- /dev/null
> > +++ b/lib/librte_gro/rte_gro.c
> > @@ -0,0 +1,125 @@
> > +/*-
> > + * BSD LICENSE
> > + *
> > + * Copyright(c) 2017 Intel Corporation. All rights reserved.
> > + *
> > + * Redistribution and use in source and binary forms, with or without
> > + * modification, are permitted provided that the following conditions
> > + * are met:
> > + *
> > + * * Redistributions of source code must retain the above copyright
> > + * notice, this list of conditions and the following disclaimer.
> > + * * Redistributions in binary form must reproduce the above copyright
> > + * notice, this list of conditions and the following disclaimer 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 derived
> > + * from this software without specific prior written permission.
> > + *
> > + * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
> > + * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
> > + * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
> > + * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
> > + * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
> > + * 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 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 DAMAGE.
> > + */
> > +
> > +#include <rte_malloc.h>
> > +#include <rte_mbuf.h>
> > +
> > +#include "rte_gro.h"
> > +
> > +static gro_tbl_create_fn tbl_create_functions[GRO_TYPE_MAX_NUM];
> > +static gro_tbl_destroy_fn tbl_destroy_functions[GRO_TYPE_MAX_NUM];
> > +
> > +struct rte_gro_tbl *rte_gro_tbl_create(uint16_t socket_id,
> > + uint16_t max_flow_num,
> > + uint16_t max_item_per_flow,
> > + uint32_t max_packet_size,
> > + uint64_t max_timeout_cycles,
> > + uint64_t desired_gro_types)
> > +{
> > + gro_tbl_create_fn create_tbl_fn;
> > + struct rte_gro_tbl *gro_tbl;
> > + uint64_t gro_type_flag = 0;
> > + uint8_t i;
> > +
> > + gro_tbl = rte_zmalloc_socket(__func__,
> > + sizeof(struct rte_gro_tbl),
> > + RTE_CACHE_LINE_SIZE,
> > + socket_id);
> > + gro_tbl->max_packet_size = max_packet_size;
> > + gro_tbl->max_timeout_cycles = max_timeout_cycles;
> > + gro_tbl->desired_gro_types = desired_gro_types;
> > +
> > + for (i = 0; i < GRO_TYPE_MAX_NUM; i++) {
> > + gro_type_flag = 1 << i;
> > + if (desired_gro_types & gro_type_flag) {
> > + create_tbl_fn = tbl_create_functions[i];
> > + if (create_tbl_fn)
> > + create_tbl_fn(socket_id,
> > + max_flow_num,
> > + max_item_per_flow);
>
> As I understand, create_tbl_fn(0 can fail.
> You should handle such situation correctly.
Thanks, I will add failure check.
>
>
> > + else
> > + gro_tbl->tbls[i] = NULL;
> > + }
> > + }
> > + return gro_tbl;
> > +}
> > +
> > +void rte_gro_tbl_destroy(struct rte_gro_tbl *gro_tbl)
> > +{
> > + gro_tbl_destroy_fn destroy_tbl_fn;
> > + uint64_t gro_type_flag;
> > + uint8_t i;
> > +
> > + if (gro_tbl == NULL)
> > + return;
> > + for (i = 0; i < GRO_TYPE_MAX_NUM; i++) {
> > + gro_type_flag = 1 << i;
> > + if (gro_tbl->desired_gro_types & gro_type_flag) {
> > + destroy_tbl_fn = tbl_destroy_functions[i];
> > + if (destroy_tbl_fn)
> > + destroy_tbl_fn(gro_tbl->tbls[i]);
> > + gro_tbl->tbls[i] = NULL;
> > + }
> > + }
> > + rte_free(gro_tbl);
> > +}
> > +
> > +uint16_t
> > +rte_gro_reassemble_burst(struct rte_mbuf **pkts __rte_unused,
> > + const uint16_t nb_pkts,
> > + const struct rte_gro_param param __rte_unused)
> > +{
> > + return nb_pkts;
> > +}
> > +
> > +int rte_gro_reassemble(struct rte_mbuf *pkt __rte_unused,
> > + struct rte_gro_tbl *gro_tbl __rte_unused)
> > +{
> > + return -1;
> > +}
> > +
> > +uint16_t rte_gro_flush(struct rte_gro_tbl *gro_tbl __rte_unused,
> > + uint64_t desired_gro_types __rte_unused,
> > + struct rte_mbuf **out __rte_unused,
> > + const uint16_t max_nb_out __rte_unused)
> > +{
> > + return 0;
> > +}
> > +
> > +uint16_t
> > +rte_gro_timeout_flush(struct rte_gro_tbl *gro_tbl __rte_unused,
> > + uint64_t desired_gro_types __rte_unused,
> > + struct rte_mbuf **out __rte_unused,
> > + const uint16_t max_nb_out __rte_unused)
> > +{
> > + return 0;
> > +}
> > diff --git a/lib/librte_gro/rte_gro.h b/lib/librte_gro/rte_gro.h
> > new file mode 100644
> > index 0000000..f9d36e8
> > --- /dev/null
> > +++ b/lib/librte_gro/rte_gro.h
> > @@ -0,0 +1,205 @@
> > +/*-
> > + * BSD LICENSE
> > + *
> > + * Copyright(c) 2017 Intel Corporation. All rights reserved.
> > + *
> > + * Redistribution and use in source and binary forms, with or without
> > + * modification, are permitted provided that the following conditions
> > + * are met:
> > + *
> > + * * Redistributions of source code must retain the above copyright
> > + * notice, this list of conditions and the following disclaimer.
> > + * * Redistributions in binary form must reproduce the above copyright
> > + * notice, this list of conditions and the following disclaimer 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 derived
> > + * from this software without specific prior written permission.
> > + *
> > + * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
> > + * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
> > + * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
> > + * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
> > + * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
> > + * 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 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 DAMAGE.
> > + */
> > +
> > +#ifndef _RTE_GRO_H_
> > +#define _RTE_GRO_H_
> > +
> > +#ifdef __cplusplus
> > +extern "C" {
> > +#endif
> > +
> > +/**
> > + * the max packets number that rte_gro_reassemble_burst can
> > + * process in each invocation.
> > + */
> > +#define GRO_MAX_BURST_ITEM_NUM 1024UL
> > +
> > +/* max number of supported GRO types */
> > +#define GRO_TYPE_MAX_NUM 64
> > +#define GRO_TYPE_SUPPORT_NUM 0 /**< current supported GRO num */
>
> Herre and everywhere public macros should start with RTE_ prefix to follow
> DPDK coding style.
Thanks, I will change the names.
>
> > +
> > +/**
> > + * GRO table, which is used to merge packets. It keeps many reassembly
> > + * tables of desired GRO types. Applications need to create GRO tables
> > + * before using rte_gro_reassemble to perform GRO.
> > + */
> > +struct rte_gro_tbl {
> > + uint64_t desired_gro_types; /**< GRO types to perform */
> > + /* max TTL measured in nanosecond */
> > + uint64_t max_timeout_cycles;
> > + /* max length of merged packet measured in byte */
> > + uint32_t max_packet_size;
> > + /* reassebly tables of desired GRO types */
> > + void *tbls[GRO_TYPE_MAX_NUM];
> > +};
>
> Not sure why do you need to define that structure here.
> As I understand it is internal to the library.
> Just declaration should be enough.
This structure defines a GRO table, which is used by rte_gro_reassemble
to merge packets. Applications need to create this table before calling
rte_gro_reassemble. So I define it in rte_gro.h.
>
> > +
> > +struct rte_gro_param {
> > + uint64_t desired_gro_types; /**< desired GRO types */
> > + uint32_t max_packet_size; /**< max length of merged packets */
> > + uint16_t max_flow_num; /**< max flow number */
> > + uint16_t max_item_per_flow; /**< max packet number per flow */
> > +};
> > +
> > +typedef void *(*gro_tbl_create_fn)(uint16_t socket_id,
> > + uint16_t max_flow_num,
> > + uint16_t max_item_per_flow);
> > +typedef void (*gro_tbl_destroy_fn)(void *tbl);
>
> Same here - user probably shouldn't see these typedefs,
> so better to hide them inside.
Thanks, I will hide them inside.
>
> > +
> > +/**
> > + * This function create a GRO table, which is used to merge packets.
> > + *
> > + * @param socket_id
> > + * socket index where the Ethernet port connects to.
> > + * @param max_flow_num
> > + * max number of flows in the GRO table.
> > + * @param max_item_per_flow
> > + * max packet number per flow. We use the value of (max_flow_num *
> > + * max_item_per_fow) to calculate table size.
> > + * @param max_packet_size
> > + * max length of merged packets. Measured in byte.
> > + * @param max_timeout_cycles
> > + * max TTL for a packet in the GRO table. It's measured in nanosecond.
> > + * @param desired_gro_types
> > + * GRO types to perform.
> > + * @return
> > + * if create successfully, return a pointer which points to the GRO
> > + * table. Otherwise, return NULL.
> > + */
> > +struct rte_gro_tbl *rte_gro_tbl_create(uint16_t socket_id,
> > + uint16_t max_flow_num,
> > + uint16_t max_item_per_flow,
> > + uint32_t max_packet_size,
> > + uint64_t max_timeout_cycles,
> > + uint64_t desired_gro_types);
>
> Hm, couldn't we have here struct rte_gro_tbl_param * instead of dozen arguments?
Thanks, I will change it.
>
> > +/**
> > + * This function destroys a GRO table.
> > + */
> > +void rte_gro_tbl_destroy(struct rte_gro_tbl *gro_tbl);
> > +
> > +/**
> > + * This is one of the main reassembly APIs, which merges numbers of
> > + * packets at a time. It assumes that all inputted packets are with
> > + * correct checksums. That is, applications should guarantee all
> > + * inputted packets are correct. Besides, it doesn't re-calculate
> > + * checksums for merged packets. If inputted packets are IP fragmented,
> > + * this function assumes them are complete (i.e. with L4 header). After
> > + * finishing processing, it returns all GROed packets to applications
> > + * immediately.
> > + *
> > + * @param pkts
> > + * a pointer array which points to the packets to reassemble. Besides,
> > + * it keeps addresses of GROed packets.
> > + * @param nb_pkts
> > + * the number of packets to reassemble.
> > + * @param param
> > + * applications use it to tell rte_gro_reassemble_burst what rules
> > + * are demanded.
> > + * @return
> > + * the number of packets after GROed.
> > + */
> > +uint16_t rte_gro_reassemble_burst(struct rte_mbuf **pkts,
> > + const uint16_t nb_pkts,
>
> Here and everywhere - no much point in defining intger input parameter
> (or any other that is passed by value) as const.
Thanks. I will modify it.
>
> > + const struct rte_gro_param param);
>
> You probably meant 'const struct rte_gro_param *param' here?
'const struct rte_gro_param *param' is better. I will modify it.
Thanks.
>
> > +
> > +/**
> > + * Reassembly function, which tries to merge the inputted packet with
> > + * one packet in a given GRO table. This function assumes the inputted
> > + * packet is with correct checksums. And it won't update checksums if
> > + * two packets are merged. Besides, if the inputted packet is IP
> > + * fragmented, this function assumes it's a complete packet (i.e. with
> > + * L4 header).
> > + *
> > + * If the inputted packet doesn't have data or it's with unsupported GRO
> > + * type, function returns immediately. Otherwise, the inputted packet is
> > + * either merged or inserted into the table. If applications want get
> > + * packets in the table, they need to call flush APIs.
> > + *
> > + * @param pkt
> > + * packet to reassemble.
> > + * @param gro_tbl
> > + * a pointer points to a GRO table.
> > + * @return
> > + * if merge the packet successfully, return a positive value. If fail
> > + * to merge, return zero. If the packet doesn't have data, or its GRO
> > + * type is unsupported, return a negative value.
> > + */
> > +int rte_gro_reassemble(struct rte_mbuf *pkt,
> > + struct rte_gro_tbl *gro_tbl);
>
>
> Ok, and why tbl one can't do bursts?
In current design, if applications want to do bursts, they don't need to
create gro_tbl. rte_gro_reassemble_burst will create a temporary table
in stack. So when do bursts (we call it lightweight mode), the operations
of applications is very simple: calling rte_gro_reassemble_burst. And
after rte_gro_reassemble_burst returns, applications can get all merged
packets. rte_gro_reassemble is another mode API, called heavyweight mode.
The gro_tbl is just used in rte_gro_reassemble. rte_gro_reassemble just
processes one packet at a time.
So you mean: we should enable rte_gro_reassemble to merge N inputted
packets with the packets in a given gro_tbl?
>
>
> > +
> > +/**
> > + * This function flushed packets from reassembly tables of desired GRO
> > + * types. It won't re-calculate checksums for merged packets in the
> > + * tables. That is, the returned packets may be with wrong checksums.
> > + *
> > + * @param gro_tbl
> > + * a pointer points to a GRO table object.
> > + * @param desired_gro_types
> > + * GRO types whose packets will be flushed.
> > + * @param out
> > + * a pointer array that is used to keep flushed packets.
> > + * @param nb_out
> > + * the size of out.
> > + * @return
> > + * the number of flushed packets. If no packets are flushed, return 0.
> > + */
> > +uint16_t rte_gro_flush(struct rte_gro_tbl *gro_tbl,
> > + uint64_t desired_gro_types,
> > + struct rte_mbuf **out,
> > + const uint16_t max_nb_out);
> > +
> > +/**
> > + * This function flushes the timeout packets from reassembly tables of
> > + * desired GRO types. It won't re-calculate checksums for merged packets
> > + * in the tables. That is, the returned packets may be with wrong
> > + * checksums.
> > + *
> > + * @param gro_tbl
> > + * a pointer points to a GRO table object.
> > + * @param desired_gro_types
> > + * rte_gro_timeout_flush only processes packets which belong to the
> > + * GRO types specified by desired_gro_types.
> > + * @param out
> > + * a pointer array that is used to keep flushed packets.
> > + * @param nb_out
> > + * the size of out.
> > + * @return
> > + * the number of flushed packets. If no packets are flushed, return 0.
> > + */
> > +uint16_t rte_gro_timeout_flush(struct rte_gro_tbl *gro_tbl,
> > + uint64_t desired_gro_types,
> > + struct rte_mbuf **out,
> > + const uint16_t max_nb_out);
>
> No point to have 2 flush() functions.
> I suggest to merge them together.
rte_gro_flush flush all packets from table, but rte_gro_timeout_flush only
flush timeout packets. They have different operations. But if we merge them
together, we need to flush all or only timeout ones?
>
> > +#ifdef __cplusplus
> > +}
> > +#endif
> > +
> > +#endif
> > diff --git a/lib/librte_gro/rte_gro_version.map b/lib/librte_gro/rte_gro_version.map
> > new file mode 100644
> > index 0000000..827596b
> > --- /dev/null
> > +++ b/lib/librte_gro/rte_gro_version.map
> > @@ -0,0 +1,12 @@
> > +DPDK_17.08 {
> > + global:
> > +
> > + rte_gro_tbl_create;
> > + rte_gro_tbl_destroy;
> > + rte_gro_reassemble_burst;
> > + rte_gro_reassemble;
> > + rte_gro_flush;
> > + rte_gro_timeout_flush;
> > +
> > + local: *;
> > +};
> > diff --git a/mk/rte.app.mk b/mk/rte.app.mk
> > index bcaf1b3..fc3776d 100644
> > --- a/mk/rte.app.mk
> > +++ b/mk/rte.app.mk
> > @@ -98,6 +98,7 @@ _LDLIBS-$(CONFIG_RTE_LIBRTE_RING) += -lrte_ring
> > _LDLIBS-$(CONFIG_RTE_LIBRTE_EAL) += -lrte_eal
> > _LDLIBS-$(CONFIG_RTE_LIBRTE_CMDLINE) += -lrte_cmdline
> > _LDLIBS-$(CONFIG_RTE_LIBRTE_REORDER) += -lrte_reorder
> > +_LDLIBS-$(CONFIG_RTE_LIBRTE_GRO) += -lrte_gro
> >
> > ifeq ($(CONFIG_RTE_EXEC_ENV_LINUXAPP),y)
> > _LDLIBS-$(CONFIG_RTE_LIBRTE_KNI) += -lrte_kni
> > --
> > 2.7.4
More information about the dev
mailing list