[v2,01/11] examples/l3fwd: add framework for event device
Checks
Commit Message
From: Sunil Kumar Kori <skori@marvell.com>
Add framework to enable event device as a producer of packets.
To switch between event mode and poll mode the following options
have been added:
`--mode="eventdev"` or `--mode="poll"`
Also, allow the user to select the schedule type to be either
RTE_SCHED_TYPE_ORDERED, RTE_SCHED_TYPE_ATOMIC or RTE_SCHED_TYPE_PARALLEL
through:
`--eventq-sched="ordered"` or `--eventq-sched="atomic"` or
`--eventq-sched="parallel"`
Poll mode is still the default operation mode.
Signed-off-by: Sunil Kumar Kori <skori@marvell.com>
---
examples/l3fwd/Makefile | 2 +-
examples/l3fwd/l3fwd.h | 6 +++
examples/l3fwd/l3fwd_event.c | 75 ++++++++++++++++++++++++++++++++++++
examples/l3fwd/l3fwd_event.h | 54 ++++++++++++++++++++++++++
examples/l3fwd/main.c | 41 +++++++++++++++++---
examples/l3fwd/meson.build | 4 +-
6 files changed, 174 insertions(+), 8 deletions(-)
create mode 100644 examples/l3fwd/l3fwd_event.c
create mode 100644 examples/l3fwd/l3fwd_event.h
Comments
> Add framework to enable event device as a producer of packets.
> To switch between event mode and poll mode the following options
> have been added:
> `--mode="eventdev"` or `--mode="poll"`
> Also, allow the user to select the schedule type to be either
> RTE_SCHED_TYPE_ORDERED, RTE_SCHED_TYPE_ATOMIC or RTE_SCHED_TYPE_PARALLEL
> through:
> `--eventq-sched="ordered"` or `--eventq-sched="atomic"` or
> `--eventq-sched="parallel"`
>
> Poll mode is still the default operation mode.
>
> Signed-off-by: Sunil Kumar Kori <skori@marvell.com>
> ---
> examples/l3fwd/Makefile | 2 +-
> examples/l3fwd/l3fwd.h | 6 +++
> examples/l3fwd/l3fwd_event.c | 75 ++++++++++++++++++++++++++++++++++++
> examples/l3fwd/l3fwd_event.h | 54 ++++++++++++++++++++++++++
> examples/l3fwd/main.c | 41 +++++++++++++++++---
> examples/l3fwd/meson.build | 4 +-
> 6 files changed, 174 insertions(+), 8 deletions(-)
> create mode 100644 examples/l3fwd/l3fwd_event.c
> create mode 100644 examples/l3fwd/l3fwd_event.h
>
> diff --git a/examples/l3fwd/Makefile b/examples/l3fwd/Makefile
> index b2dbf2607..c892b867b 100644
> --- a/examples/l3fwd/Makefile
> +++ b/examples/l3fwd/Makefile
> @@ -5,7 +5,7 @@
> APP = l3fwd
>
> # all source are stored in SRCS-y
> -SRCS-y := main.c l3fwd_lpm.c l3fwd_em.c
> +SRCS-y := main.c l3fwd_lpm.c l3fwd_em.c l3fwd_event.c
>
> # Build using pkg-config variables if possible
> ifeq ($(shell pkg-config --exists libdpdk && echo 0),0)
> diff --git a/examples/l3fwd/l3fwd.h b/examples/l3fwd/l3fwd.h
> index 293fb1fa2..cd17a41b3 100644
> --- a/examples/l3fwd/l3fwd.h
> +++ b/examples/l3fwd/l3fwd.h
> @@ -5,6 +5,9 @@
> #ifndef __L3_FWD_H__
> #define __L3_FWD_H__
>
> +#include <stdbool.h>
> +
Why do we need it here?
> +#include <rte_ethdev.h>
> #include <rte_vect.h>
>
> #define DO_RFC_1812_CHECKS
> @@ -169,6 +172,9 @@ is_valid_ipv4_pkt(struct rte_ipv4_hdr *pkt, uint32_t link_len)
> }
> #endif /* DO_RFC_1812_CHECKS */
>
> +void
> +print_usage(const char *prgname);
> +
> /* Function pointers for LPM or EM functionality. */
> void
> setup_lpm(const int socketid);
> diff --git a/examples/l3fwd/l3fwd_event.c b/examples/l3fwd/l3fwd_event.c
> new file mode 100644
> index 000000000..3892720be
> --- /dev/null
> +++ b/examples/l3fwd/l3fwd_event.c
> @@ -0,0 +1,75 @@
> +/* SPDX-License-Identifier: BSD-3-Clause
> + * Copyright(C) 2019 Marvell International Ltd.
> + */
> +
> +#include <stdbool.h>
> +#include <getopt.h>
> +
> +#include "l3fwd.h"
> +#include "l3fwd_event.h"
> +
> +static void
> +parse_mode(const char *optarg)
> +{
> + struct l3fwd_event_resources *evt_rsrc = l3fwd_get_eventdev_rsrc();
> +
> + if (!strncmp(optarg, "poll", 4))
That looks a bit clumsy and error-prone.
Just strcmp(optarg, "poll") seems enough here.
Same for other similar places.
> + evt_rsrc->enabled = false;
> + else if (!strncmp(optarg, "eventdev", 8))
> + evt_rsrc->enabled = true;
> +}
> +
> +static void
> +parse_eventq_sync(const char *optarg)
> +{
> + struct l3fwd_event_resources *evt_rsrc = l3fwd_get_eventdev_rsrc();
> +
> + if (!strncmp(optarg, "ordered", 7))
> + evt_rsrc->sched_type = RTE_SCHED_TYPE_ORDERED;
> + if (!strncmp(optarg, "atomic", 6))
> + evt_rsrc->sched_type = RTE_SCHED_TYPE_ATOMIC;
> + if (!strncmp(optarg, "parallel", 8))
> + evt_rsrc->sched_type = RTE_SCHED_TYPE_PARALLEL;
> +}
> +
> +static void
> +l3fwd_parse_eventdev_args(char **argv, int argc)
> +{
> + const struct option eventdev_lgopts[] = {
> + {CMD_LINE_OPT_MODE, 1, 0, CMD_LINE_OPT_MODE_NUM},
> + {CMD_LINE_OPT_EVENTQ_SYNC, 1, 0, CMD_LINE_OPT_EVENTQ_SYNC_NUM},
> + {NULL, 0, 0, 0}
> + };
> + char *prgname = argv[0];
> + char **argvopt = argv;
> + int32_t option_index;
> + int32_t opt;
> +
> + while ((opt = getopt_long(argc, argvopt, "", eventdev_lgopts,
> + &option_index)) != EOF) {
> + switch (opt) {
> + case CMD_LINE_OPT_MODE_NUM:
> + parse_mode(optarg);
> + break;
> +
> + case CMD_LINE_OPT_EVENTQ_SYNC_NUM:
> + parse_eventq_sync(optarg);
> + break;
> +
> + default:
> + print_usage(prgname);
> + exit(1);
> + }
> + }
> +}
> +
> +void
> +l3fwd_event_resource_setup(void)
> +{
> + struct l3fwd_event_resources *evt_rsrc = l3fwd_get_eventdev_rsrc();
> +
> + /* Parse eventdev command line options */
> + l3fwd_parse_eventdev_args(evt_rsrc->args, evt_rsrc->nb_args);
> + if (!evt_rsrc->enabled)
> + return;
> +}
> diff --git a/examples/l3fwd/l3fwd_event.h b/examples/l3fwd/l3fwd_event.h
> new file mode 100644
> index 000000000..c95296c38
> --- /dev/null
> +++ b/examples/l3fwd/l3fwd_event.h
> @@ -0,0 +1,54 @@
> +/* SPDX-License-Identifier: BSD-3-Clause
> + * Copyright(C) 2019 Marvell International Ltd.
> + */
> +
> +#ifndef __L3FWD_EVENTDEV_H__
> +#define __L3FWD_EVENTDEV_H__
> +
> +#include <rte_common.h>
> +#include <rte_eventdev.h>
> +#include <rte_spinlock.h>
> +
> +#include "l3fwd.h"
> +
> +#define CMD_LINE_OPT_MODE "mode"
> +#define CMD_LINE_OPT_EVENTQ_SYNC "eventq-sched"
> +
> +enum {
> + CMD_LINE_OPT_MODE_NUM = 265,
> + CMD_LINE_OPT_EVENTQ_SYNC_NUM,
> +};
> +
> +struct l3fwd_event_resources {
> + uint8_t sched_type;
> + uint8_t enabled;
> + uint8_t nb_args;
> + char **args;
> +};
> +
> +static inline struct l3fwd_event_resources *
> +l3fwd_get_eventdev_rsrc(void)
> +{
> + static const char name[RTE_MEMZONE_NAMESIZE] = "l3fwd_event_rsrc";
> + const struct rte_memzone *mz;
> +
> + mz = rte_memzone_lookup(name);
> +
> + if (mz != NULL)
> + return mz->addr;
> +
> + mz = rte_memzone_reserve(name, sizeof(struct l3fwd_event_resources),
> + 0, 0);
> + if (mz != NULL) {
> + memset(mz->addr, 0, sizeof(struct l3fwd_event_resources));
> + return mz->addr;
> + }
> +
> + rte_exit(EXIT_FAILURE, "Unable to allocate memory for eventdev cfg\n");
> +
> + return NULL;
> +}
Does this function really need to be inline?
It wouldn't be fast anyway.
Another question - do you really need memzone here?
Wouldn't just rte_malloc() be enough?
> +
> +void l3fwd_event_resource_setup(void);
> +
> +#endif /* __L3FWD_EVENTDEV_H__ */
> diff --git a/examples/l3fwd/main.c b/examples/l3fwd/main.c
> index 4dea12a65..19ca4483c 100644
> --- a/examples/l3fwd/main.c
> +++ b/examples/l3fwd/main.c
> @@ -13,12 +13,12 @@
> #include <errno.h>
> #include <getopt.h>
> #include <signal.h>
> -#include <stdbool.h>
>
> #include <rte_common.h>
> #include <rte_vect.h>
> #include <rte_byteorder.h>
> #include <rte_log.h>
> +#include <rte_malloc.h>
> #include <rte_memory.h>
> #include <rte_memcpy.h>
> #include <rte_eal.h>
> @@ -33,7 +33,6 @@
> #include <rte_random.h>
> #include <rte_debug.h>
> #include <rte_ether.h>
> -#include <rte_ethdev.h>
> #include <rte_mempool.h>
> #include <rte_mbuf.h>
> #include <rte_ip.h>
> @@ -46,6 +45,7 @@
> #include <cmdline_parse_etheraddr.h>
>
> #include "l3fwd.h"
> +#include "l3fwd_event.h"
>
> /*
> * Configurable number of RX/TX ring descriptors
> @@ -274,7 +274,7 @@ init_lcore_rx_queues(void)
> }
>
> /* display usage */
> -static void
> +void
> print_usage(const char *prgname)
> {
> fprintf(stderr, "%s [EAL options] --"
> @@ -289,7 +289,9 @@ print_usage(const char *prgname)
> " [--hash-entry-num]"
> " [--ipv6]"
> " [--parse-ptype]"
> - " [--per-port-pool]\n\n"
> + " [--per-port-pool]"
> + " [--mode]"
> + " [--eventq-sched]\n\n"
>
> " -p PORTMASK: Hexadecimal bitmask of ports to configure\n"
> " -P : Enable promiscuous mode\n"
> @@ -304,7 +306,13 @@ print_usage(const char *prgname)
> " --hash-entry-num: Specify the hash entry number in hexadecimal to be setup\n"
> " --ipv6: Set if running ipv6 packets\n"
> " --parse-ptype: Set to use software to analyze packet type\n"
> - " --per-port-pool: Use separate buffer pool per port\n\n",
> + " --per-port-pool: Use separate buffer pool per port\n"
> + " --mode: Packet transfer mode for I/O, poll or eventdev\n"
> + " Default mode = poll\n"
> + " --eventq-sched: Event queue synchronization method "
> + " ordered, atomic or parallel.\n\t\t"
> + " Default: atomic\n\t\t"
> + " Valid only if --mode=eventdev\n\n",
> prgname);
> }
>
> @@ -504,11 +512,19 @@ static const struct option lgopts[] = {
> static int
> parse_args(int argc, char **argv)
> {
> + struct l3fwd_event_resources *evt_rsrc;
> int opt, ret;
> char **argvopt;
> int option_index;
> char *prgname = argv[0];
>
> + evt_rsrc = l3fwd_get_eventdev_rsrc();
> + evt_rsrc->args = rte_zmalloc("l3fwd_event_args", sizeof(char *), 0);
> + if (evt_rsrc->args == NULL)
> + rte_exit(EXIT_FAILURE,
> + "Unable to allocate memory for eventdev arg");
> + evt_rsrc->args[0] = argv[0];
> + evt_rsrc->nb_args++;
> argvopt = argv;
>
> /* Error or normal output strings. */
> @@ -538,6 +554,15 @@ parse_args(int argc, char **argv)
> l3fwd_lpm_on = 1;
> break;
>
> + case '?':
> + /* Eventdev options are encountered skip for
> + * now and processed later.
> + */
> + evt_rsrc->args[evt_rsrc->nb_args] =
> + argv[optind - 1];
> + evt_rsrc->nb_args++;
> + break;
> +
All this construction with first allocating space for eventdev specific params
copying them and parsing in a special function - looks like an overkill to me.
Why not just to call parse_mode()/parse_eventq.. functions straight from here?
> /* long options */
> case CMD_LINE_OPT_CONFIG_NUM:
> ret = parse_config(optarg);
> @@ -811,6 +836,7 @@ prepare_ptype_parser(uint16_t portid, uint16_t queueid)
> int
> main(int argc, char **argv)
> {
> + struct l3fwd_event_resources *evt_rsrc;
> struct lcore_conf *qconf;
> struct rte_eth_dev_info dev_info;
> struct rte_eth_txconf *txconf;
> @@ -839,11 +865,16 @@ main(int argc, char **argv)
> *(uint64_t *)(val_eth + portid) = dest_eth_addr[portid];
> }
>
> + evt_rsrc = l3fwd_get_eventdev_rsrc();
> + RTE_SET_USED(evt_rsrc);
> /* parse application arguments (after the EAL ones) */
> ret = parse_args(argc, argv);
> if (ret < 0)
> rte_exit(EXIT_FAILURE, "Invalid L3FWD parameters\n");
>
> + /* Configure eventdev parameters if user has requested */
> + l3fwd_event_resource_setup();
> +
> if (check_lcore_params() < 0)
> rte_exit(EXIT_FAILURE, "check_lcore_params failed\n");
>
> diff --git a/examples/l3fwd/meson.build b/examples/l3fwd/meson.build
> index 6dd4b9022..864327c7b 100644
> --- a/examples/l3fwd/meson.build
> +++ b/examples/l3fwd/meson.build
> @@ -6,7 +6,7 @@
> # To build this example as a standalone application with an already-installed
> # DPDK instance, use 'make'
>
> -deps += ['hash', 'lpm']
> +deps += ['hash', 'lpm', 'eventdev']
> sources = files(
> - 'l3fwd_em.c', 'l3fwd_lpm.c', 'main.c'
> + 'l3fwd_em.c', 'l3fwd_lpm.c', 'l3fwd_event.c', 'main.c'
> )
> --
> 2.17.1
>-----Original Message-----
>From: dev <dev-bounces@dpdk.org> On Behalf Of Ananyev,
>Konstantin
>Sent: Friday, January 3, 2020 6:19 PM
>To: Pavan Nikhilesh Bhagavatula <pbhagavatula@marvell.com>; Jerin
>Jacob Kollanukkaran <jerinj@marvell.com>; Kovacevic, Marko
><marko.kovacevic@intel.com>; Ori Kam <orika@mellanox.com>;
>Richardson, Bruce <bruce.richardson@intel.com>; Nicolau, Radu
><radu.nicolau@intel.com>; Akhil Goyal <akhil.goyal@nxp.com>;
>Kantecki, Tomasz <tomasz.kantecki@intel.com>; Sunil Kumar Kori
><skori@marvell.com>
>Cc: dev@dpdk.org
>Subject: Re: [dpdk-dev] [PATCH v2 01/11] examples/l3fwd: add
>framework for event device
>
>
>> Add framework to enable event device as a producer of packets.
>> To switch between event mode and poll mode the following options
>> have been added:
>> `--mode="eventdev"` or `--mode="poll"`
>> Also, allow the user to select the schedule type to be either
>> RTE_SCHED_TYPE_ORDERED, RTE_SCHED_TYPE_ATOMIC or
>RTE_SCHED_TYPE_PARALLEL
>> through:
>> `--eventq-sched="ordered"` or `--eventq-sched="atomic"` or
>> `--eventq-sched="parallel"`
>>
>> Poll mode is still the default operation mode.
>>
>> Signed-off-by: Sunil Kumar Kori <skori@marvell.com>
>> ---
>> examples/l3fwd/Makefile | 2 +-
>> examples/l3fwd/l3fwd.h | 6 +++
>> examples/l3fwd/l3fwd_event.c | 75
>++++++++++++++++++++++++++++++++++++
>> examples/l3fwd/l3fwd_event.h | 54
>++++++++++++++++++++++++++
>> examples/l3fwd/main.c | 41 +++++++++++++++++---
>> examples/l3fwd/meson.build | 4 +-
>> 6 files changed, 174 insertions(+), 8 deletions(-)
>> create mode 100644 examples/l3fwd/l3fwd_event.c
>> create mode 100644 examples/l3fwd/l3fwd_event.h
>>
>> diff --git a/examples/l3fwd/Makefile b/examples/l3fwd/Makefile
>> index b2dbf2607..c892b867b 100644
>> --- a/examples/l3fwd/Makefile
>> +++ b/examples/l3fwd/Makefile
>> @@ -5,7 +5,7 @@
>> APP = l3fwd
>>
>> # all source are stored in SRCS-y
>> -SRCS-y := main.c l3fwd_lpm.c l3fwd_em.c
>> +SRCS-y := main.c l3fwd_lpm.c l3fwd_em.c l3fwd_event.c
>>
>> # Build using pkg-config variables if possible
>> ifeq ($(shell pkg-config --exists libdpdk && echo 0),0)
>> diff --git a/examples/l3fwd/l3fwd.h b/examples/l3fwd/l3fwd.h
>> index 293fb1fa2..cd17a41b3 100644
>> --- a/examples/l3fwd/l3fwd.h
>> +++ b/examples/l3fwd/l3fwd.h
>> @@ -5,6 +5,9 @@
>> #ifndef __L3_FWD_H__
>> #define __L3_FWD_H__
>>
>> +#include <stdbool.h>
>> +
>
>Why do we need it here?
Looks like an artifact I will remove it in next version.
>
>> +#include <rte_ethdev.h>
>> #include <rte_vect.h>
>>
>> #define DO_RFC_1812_CHECKS
>> @@ -169,6 +172,9 @@ is_valid_ipv4_pkt(struct rte_ipv4_hdr *pkt,
>uint32_t link_len)
>> }
>> #endif /* DO_RFC_1812_CHECKS */
>>
>> +void
>> +print_usage(const char *prgname);
>> +
>> /* Function pointers for LPM or EM functionality. */
>> void
>> setup_lpm(const int socketid);
>> diff --git a/examples/l3fwd/l3fwd_event.c
>b/examples/l3fwd/l3fwd_event.c
>> new file mode 100644
>> index 000000000..3892720be
>> --- /dev/null
>> +++ b/examples/l3fwd/l3fwd_event.c
>> @@ -0,0 +1,75 @@
>> +/* SPDX-License-Identifier: BSD-3-Clause
>> + * Copyright(C) 2019 Marvell International Ltd.
>> + */
>> +
>> +#include <stdbool.h>
>> +#include <getopt.h>
>> +
>> +#include "l3fwd.h"
>> +#include "l3fwd_event.h"
>> +
>> +static void
>> +parse_mode(const char *optarg)
>> +{
>> + struct l3fwd_event_resources *evt_rsrc =
>l3fwd_get_eventdev_rsrc();
>> +
>> + if (!strncmp(optarg, "poll", 4))
>
>That looks a bit clumsy and error-prone.
>Just strcmp(optarg, "poll") seems enough here.
>Same for other similar places.
Will simplify in next version.
>
>> + evt_rsrc->enabled = false;
>> + else if (!strncmp(optarg, "eventdev", 8))
>> + evt_rsrc->enabled = true;
>> +}
>> +
>> +static void
>> +parse_eventq_sync(const char *optarg)
>> +{
>> + struct l3fwd_event_resources *evt_rsrc =
>l3fwd_get_eventdev_rsrc();
>> +
>> + if (!strncmp(optarg, "ordered", 7))
>> + evt_rsrc->sched_type = RTE_SCHED_TYPE_ORDERED;
>> + if (!strncmp(optarg, "atomic", 6))
>> + evt_rsrc->sched_type = RTE_SCHED_TYPE_ATOMIC;
>> + if (!strncmp(optarg, "parallel", 8))
>> + evt_rsrc->sched_type = RTE_SCHED_TYPE_PARALLEL;
>> +}
>> +
>> +static void
>> +l3fwd_parse_eventdev_args(char **argv, int argc)
>> +{
>> + const struct option eventdev_lgopts[] = {
>> + {CMD_LINE_OPT_MODE, 1, 0,
>CMD_LINE_OPT_MODE_NUM},
>> + {CMD_LINE_OPT_EVENTQ_SYNC, 1, 0,
>CMD_LINE_OPT_EVENTQ_SYNC_NUM},
>> + {NULL, 0, 0, 0}
>> + };
>> + char *prgname = argv[0];
>> + char **argvopt = argv;
>> + int32_t option_index;
>> + int32_t opt;
>> +
>> + while ((opt = getopt_long(argc, argvopt, "", eventdev_lgopts,
>> + &option_index)) != EOF) {
>> + switch (opt) {
>> + case CMD_LINE_OPT_MODE_NUM:
>> + parse_mode(optarg);
>> + break;
>> +
>> + case CMD_LINE_OPT_EVENTQ_SYNC_NUM:
>> + parse_eventq_sync(optarg);
>> + break;
>> +
>> + default:
>> + print_usage(prgname);
>> + exit(1);
>> + }
>> + }
>> +}
>> +
>> +void
>> +l3fwd_event_resource_setup(void)
>> +{
>> + struct l3fwd_event_resources *evt_rsrc =
>l3fwd_get_eventdev_rsrc();
>> +
>> + /* Parse eventdev command line options */
>> + l3fwd_parse_eventdev_args(evt_rsrc->args, evt_rsrc-
>>nb_args);
>> + if (!evt_rsrc->enabled)
>> + return;
>> +}
>> diff --git a/examples/l3fwd/l3fwd_event.h
>b/examples/l3fwd/l3fwd_event.h
>> new file mode 100644
>> index 000000000..c95296c38
>> --- /dev/null
>> +++ b/examples/l3fwd/l3fwd_event.h
>> @@ -0,0 +1,54 @@
>> +/* SPDX-License-Identifier: BSD-3-Clause
>> + * Copyright(C) 2019 Marvell International Ltd.
>> + */
>> +
>> +#ifndef __L3FWD_EVENTDEV_H__
>> +#define __L3FWD_EVENTDEV_H__
>> +
>> +#include <rte_common.h>
>> +#include <rte_eventdev.h>
>> +#include <rte_spinlock.h>
>> +
>> +#include "l3fwd.h"
>> +
>> +#define CMD_LINE_OPT_MODE "mode"
>> +#define CMD_LINE_OPT_EVENTQ_SYNC "eventq-sched"
>> +
>> +enum {
>> + CMD_LINE_OPT_MODE_NUM = 265,
>> + CMD_LINE_OPT_EVENTQ_SYNC_NUM,
>> +};
>> +
>> +struct l3fwd_event_resources {
>> + uint8_t sched_type;
>> + uint8_t enabled;
>> + uint8_t nb_args;
>> + char **args;
>> +};
>> +
>> +static inline struct l3fwd_event_resources *
>> +l3fwd_get_eventdev_rsrc(void)
>> +{
>> + static const char name[RTE_MEMZONE_NAMESIZE] =
>"l3fwd_event_rsrc";
>> + const struct rte_memzone *mz;
>> +
>> + mz = rte_memzone_lookup(name);
>> +
>> + if (mz != NULL)
>> + return mz->addr;
>> +
>> + mz = rte_memzone_reserve(name, sizeof(struct
>l3fwd_event_resources),
>> + 0, 0);
>> + if (mz != NULL) {
>> + memset(mz->addr, 0, sizeof(struct
>l3fwd_event_resources));
>> + return mz->addr;
>> + }
>> +
>> + rte_exit(EXIT_FAILURE, "Unable to allocate memory for
>eventdev cfg\n");
>> +
>> + return NULL;
>> +}
>
>Does this function really need to be inline?
>It wouldn't be fast anyway.
>Another question - do you really need memzone here?
>Wouldn't just rte_malloc() be enough?
Will remove inline in next version.
rte_malloc would call for a global variable which I'm
trying to avoid. I don't think there is any harm in using
named memzone.
>
>> +
>> +void l3fwd_event_resource_setup(void);
>> +
>> +#endif /* __L3FWD_EVENTDEV_H__ */
>> diff --git a/examples/l3fwd/main.c b/examples/l3fwd/main.c
>> index 4dea12a65..19ca4483c 100644
>> --- a/examples/l3fwd/main.c
>> +++ b/examples/l3fwd/main.c
>> @@ -13,12 +13,12 @@
>> #include <errno.h>
>> #include <getopt.h>
>> #include <signal.h>
>> -#include <stdbool.h>
>>
>> #include <rte_common.h>
>> #include <rte_vect.h>
>> #include <rte_byteorder.h>
>> #include <rte_log.h>
>> +#include <rte_malloc.h>
>> #include <rte_memory.h>
>> #include <rte_memcpy.h>
>> #include <rte_eal.h>
>> @@ -33,7 +33,6 @@
>> #include <rte_random.h>
>> #include <rte_debug.h>
>> #include <rte_ether.h>
>> -#include <rte_ethdev.h>
>> #include <rte_mempool.h>
>> #include <rte_mbuf.h>
>> #include <rte_ip.h>
>> @@ -46,6 +45,7 @@
>> #include <cmdline_parse_etheraddr.h>
>>
>> #include "l3fwd.h"
>> +#include "l3fwd_event.h"
>>
>> /*
>> * Configurable number of RX/TX ring descriptors
>> @@ -274,7 +274,7 @@ init_lcore_rx_queues(void)
>> }
>>
>> /* display usage */
>> -static void
>> +void
>> print_usage(const char *prgname)
>> {
>> fprintf(stderr, "%s [EAL options] --"
>> @@ -289,7 +289,9 @@ print_usage(const char *prgname)
>> " [--hash-entry-num]"
>> " [--ipv6]"
>> " [--parse-ptype]"
>> - " [--per-port-pool]\n\n"
>> + " [--per-port-pool]"
>> + " [--mode]"
>> + " [--eventq-sched]\n\n"
>>
>> " -p PORTMASK: Hexadecimal bitmask of ports to
>configure\n"
>> " -P : Enable promiscuous mode\n"
>> @@ -304,7 +306,13 @@ print_usage(const char *prgname)
>> " --hash-entry-num: Specify the hash entry number in
>hexadecimal to be setup\n"
>> " --ipv6: Set if running ipv6 packets\n"
>> " --parse-ptype: Set to use software to analyze packet
>type\n"
>> - " --per-port-pool: Use separate buffer pool per
>port\n\n",
>> + " --per-port-pool: Use separate buffer pool per port\n"
>> + " --mode: Packet transfer mode for I/O, poll or
>eventdev\n"
>> + " Default mode = poll\n"
>> + " --eventq-sched: Event queue synchronization
>method "
>> + " ordered, atomic or parallel.\n\t\t"
>> + " Default: atomic\n\t\t"
>> + " Valid only if --mode=eventdev\n\n",
>> prgname);
>> }
>>
>> @@ -504,11 +512,19 @@ static const struct option lgopts[] = {
>> static int
>> parse_args(int argc, char **argv)
>> {
>> + struct l3fwd_event_resources *evt_rsrc;
>> int opt, ret;
>> char **argvopt;
>> int option_index;
>> char *prgname = argv[0];
>>
>> + evt_rsrc = l3fwd_get_eventdev_rsrc();
>> + evt_rsrc->args = rte_zmalloc("l3fwd_event_args", sizeof(char
>*), 0);
>> + if (evt_rsrc->args == NULL)
>> + rte_exit(EXIT_FAILURE,
>> + "Unable to allocate memory for
>eventdev arg");
>> + evt_rsrc->args[0] = argv[0];
>> + evt_rsrc->nb_args++;
>> argvopt = argv;
>>
>> /* Error or normal output strings. */
>> @@ -538,6 +554,15 @@ parse_args(int argc, char **argv)
>> l3fwd_lpm_on = 1;
>> break;
>>
>> + case '?':
>> + /* Eventdev options are encountered skip for
>> + * now and processed later.
>> + */
>> + evt_rsrc->args[evt_rsrc->nb_args] =
>> + argv[optind - 1];
>> + evt_rsrc->nb_args++;
>> + break;
>> +
>
>All this construction with first allocating space for eventdev specific
>params
>copying them and parsing in a special function - looks like an overkill to
>me.
>Why not just to call parse_mode()/parse_eventq.. functions straight
>from here?
We were trying to make minimal changes to l3fwd but I guess we could
have common cmdline parse functions.
>
>
>> /* long options */
>> case CMD_LINE_OPT_CONFIG_NUM:
>> ret = parse_config(optarg);
>> @@ -811,6 +836,7 @@ prepare_ptype_parser(uint16_t portid,
>uint16_t queueid)
>> int
>> main(int argc, char **argv)
>> {
>> + struct l3fwd_event_resources *evt_rsrc;
>> struct lcore_conf *qconf;
>> struct rte_eth_dev_info dev_info;
>> struct rte_eth_txconf *txconf;
>> @@ -839,11 +865,16 @@ main(int argc, char **argv)
>> *(uint64_t *)(val_eth + portid) =
>dest_eth_addr[portid];
>> }
>>
>> + evt_rsrc = l3fwd_get_eventdev_rsrc();
>> + RTE_SET_USED(evt_rsrc);
>> /* parse application arguments (after the EAL ones) */
>> ret = parse_args(argc, argv);
>> if (ret < 0)
>> rte_exit(EXIT_FAILURE, "Invalid L3FWD parameters\n");
>>
>> + /* Configure eventdev parameters if user has requested */
>> + l3fwd_event_resource_setup();
>> +
>> if (check_lcore_params() < 0)
>> rte_exit(EXIT_FAILURE, "check_lcore_params
>failed\n");
>>
>> diff --git a/examples/l3fwd/meson.build
>b/examples/l3fwd/meson.build
>> index 6dd4b9022..864327c7b 100644
>> --- a/examples/l3fwd/meson.build
>> +++ b/examples/l3fwd/meson.build
>> @@ -6,7 +6,7 @@
>> # To build this example as a standalone application with an already-
>installed
>> # DPDK instance, use 'make'
>>
>> -deps += ['hash', 'lpm']
>> +deps += ['hash', 'lpm', 'eventdev']
>> sources = files(
>> - 'l3fwd_em.c', 'l3fwd_lpm.c', 'main.c'
>> + 'l3fwd_em.c', 'l3fwd_lpm.c', 'l3fwd_event.c', 'main.c'
>> )
>> --
>> 2.17.1
> >> +struct l3fwd_event_resources {
> >> + uint8_t sched_type;
> >> + uint8_t enabled;
> >> + uint8_t nb_args;
> >> + char **args;
> >> +};
> >> +
> >> +static inline struct l3fwd_event_resources *
> >> +l3fwd_get_eventdev_rsrc(void)
> >> +{
> >> + static const char name[RTE_MEMZONE_NAMESIZE] =
> >"l3fwd_event_rsrc";
> >> + const struct rte_memzone *mz;
> >> +
> >> + mz = rte_memzone_lookup(name);
> >> +
> >> + if (mz != NULL)
> >> + return mz->addr;
> >> +
> >> + mz = rte_memzone_reserve(name, sizeof(struct
> >l3fwd_event_resources),
> >> + 0, 0);
> >> + if (mz != NULL) {
> >> + memset(mz->addr, 0, sizeof(struct
> >l3fwd_event_resources));
> >> + return mz->addr;
> >> + }
> >> +
> >> + rte_exit(EXIT_FAILURE, "Unable to allocate memory for
> >eventdev cfg\n");
> >> +
> >> + return NULL;
> >> +}
> >
> >Does this function really need to be inline?
> >It wouldn't be fast anyway.
> >Another question - do you really need memzone here?
> >Wouldn't just rte_malloc() be enough?
>
> Will remove inline in next version.
> rte_malloc would call for a global variable which I'm
> trying to avoid.
If you plan to move that function into .c file,
you don't need a global var. it could be static local one.
> I don't think there is any harm in using
> named memzone.
I don't see any harm, though malloc+var will be faster I think.
Though up to you - no strong opinion here.
>> >> +struct l3fwd_event_resources {
>> >> + uint8_t sched_type;
>> >> + uint8_t enabled;
>> >> + uint8_t nb_args;
>> >> + char **args;
>> >> +};
>> >> +
>> >> +static inline struct l3fwd_event_resources *
>> >> +l3fwd_get_eventdev_rsrc(void)
>> >> +{
>> >> + static const char name[RTE_MEMZONE_NAMESIZE] =
>> >"l3fwd_event_rsrc";
>> >> + const struct rte_memzone *mz;
>> >> +
>> >> + mz = rte_memzone_lookup(name);
>> >> +
>> >> + if (mz != NULL)
>> >> + return mz->addr;
>> >> +
>> >> + mz = rte_memzone_reserve(name, sizeof(struct
>> >l3fwd_event_resources),
>> >> + 0, 0);
>> >> + if (mz != NULL) {
>> >> + memset(mz->addr, 0, sizeof(struct
>> >l3fwd_event_resources));
>> >> + return mz->addr;
>> >> + }
>> >> +
>> >> + rte_exit(EXIT_FAILURE, "Unable to allocate memory for
>> >eventdev cfg\n");
>> >> +
>> >> + return NULL;
>> >> +}
>> >
>> >Does this function really need to be inline?
>> >It wouldn't be fast anyway.
>> >Another question - do you really need memzone here?
>> >Wouldn't just rte_malloc() be enough?
>>
>> Will remove inline in next version.
>> rte_malloc would call for a global variable which I'm
>> trying to avoid.
>
>If you plan to move that function into .c file,
>you don't need a global var. it could be static local one.
>
>> I don't think there is any harm in using
>> named memzone.
>
>I don't see any harm, though malloc+var will be faster I think.
>Though up to you - no strong opinion here.
Maybe we can cut down some init time. I will move it to .c next version.
Thanks,
Pavan
@@ -5,7 +5,7 @@
APP = l3fwd
# all source are stored in SRCS-y
-SRCS-y := main.c l3fwd_lpm.c l3fwd_em.c
+SRCS-y := main.c l3fwd_lpm.c l3fwd_em.c l3fwd_event.c
# Build using pkg-config variables if possible
ifeq ($(shell pkg-config --exists libdpdk && echo 0),0)
@@ -5,6 +5,9 @@
#ifndef __L3_FWD_H__
#define __L3_FWD_H__
+#include <stdbool.h>
+
+#include <rte_ethdev.h>
#include <rte_vect.h>
#define DO_RFC_1812_CHECKS
@@ -169,6 +172,9 @@ is_valid_ipv4_pkt(struct rte_ipv4_hdr *pkt, uint32_t link_len)
}
#endif /* DO_RFC_1812_CHECKS */
+void
+print_usage(const char *prgname);
+
/* Function pointers for LPM or EM functionality. */
void
setup_lpm(const int socketid);
new file mode 100644
@@ -0,0 +1,75 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(C) 2019 Marvell International Ltd.
+ */
+
+#include <stdbool.h>
+#include <getopt.h>
+
+#include "l3fwd.h"
+#include "l3fwd_event.h"
+
+static void
+parse_mode(const char *optarg)
+{
+ struct l3fwd_event_resources *evt_rsrc = l3fwd_get_eventdev_rsrc();
+
+ if (!strncmp(optarg, "poll", 4))
+ evt_rsrc->enabled = false;
+ else if (!strncmp(optarg, "eventdev", 8))
+ evt_rsrc->enabled = true;
+}
+
+static void
+parse_eventq_sync(const char *optarg)
+{
+ struct l3fwd_event_resources *evt_rsrc = l3fwd_get_eventdev_rsrc();
+
+ if (!strncmp(optarg, "ordered", 7))
+ evt_rsrc->sched_type = RTE_SCHED_TYPE_ORDERED;
+ if (!strncmp(optarg, "atomic", 6))
+ evt_rsrc->sched_type = RTE_SCHED_TYPE_ATOMIC;
+ if (!strncmp(optarg, "parallel", 8))
+ evt_rsrc->sched_type = RTE_SCHED_TYPE_PARALLEL;
+}
+
+static void
+l3fwd_parse_eventdev_args(char **argv, int argc)
+{
+ const struct option eventdev_lgopts[] = {
+ {CMD_LINE_OPT_MODE, 1, 0, CMD_LINE_OPT_MODE_NUM},
+ {CMD_LINE_OPT_EVENTQ_SYNC, 1, 0, CMD_LINE_OPT_EVENTQ_SYNC_NUM},
+ {NULL, 0, 0, 0}
+ };
+ char *prgname = argv[0];
+ char **argvopt = argv;
+ int32_t option_index;
+ int32_t opt;
+
+ while ((opt = getopt_long(argc, argvopt, "", eventdev_lgopts,
+ &option_index)) != EOF) {
+ switch (opt) {
+ case CMD_LINE_OPT_MODE_NUM:
+ parse_mode(optarg);
+ break;
+
+ case CMD_LINE_OPT_EVENTQ_SYNC_NUM:
+ parse_eventq_sync(optarg);
+ break;
+
+ default:
+ print_usage(prgname);
+ exit(1);
+ }
+ }
+}
+
+void
+l3fwd_event_resource_setup(void)
+{
+ struct l3fwd_event_resources *evt_rsrc = l3fwd_get_eventdev_rsrc();
+
+ /* Parse eventdev command line options */
+ l3fwd_parse_eventdev_args(evt_rsrc->args, evt_rsrc->nb_args);
+ if (!evt_rsrc->enabled)
+ return;
+}
new file mode 100644
@@ -0,0 +1,54 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(C) 2019 Marvell International Ltd.
+ */
+
+#ifndef __L3FWD_EVENTDEV_H__
+#define __L3FWD_EVENTDEV_H__
+
+#include <rte_common.h>
+#include <rte_eventdev.h>
+#include <rte_spinlock.h>
+
+#include "l3fwd.h"
+
+#define CMD_LINE_OPT_MODE "mode"
+#define CMD_LINE_OPT_EVENTQ_SYNC "eventq-sched"
+
+enum {
+ CMD_LINE_OPT_MODE_NUM = 265,
+ CMD_LINE_OPT_EVENTQ_SYNC_NUM,
+};
+
+struct l3fwd_event_resources {
+ uint8_t sched_type;
+ uint8_t enabled;
+ uint8_t nb_args;
+ char **args;
+};
+
+static inline struct l3fwd_event_resources *
+l3fwd_get_eventdev_rsrc(void)
+{
+ static const char name[RTE_MEMZONE_NAMESIZE] = "l3fwd_event_rsrc";
+ const struct rte_memzone *mz;
+
+ mz = rte_memzone_lookup(name);
+
+ if (mz != NULL)
+ return mz->addr;
+
+ mz = rte_memzone_reserve(name, sizeof(struct l3fwd_event_resources),
+ 0, 0);
+ if (mz != NULL) {
+ memset(mz->addr, 0, sizeof(struct l3fwd_event_resources));
+ return mz->addr;
+ }
+
+ rte_exit(EXIT_FAILURE, "Unable to allocate memory for eventdev cfg\n");
+
+ return NULL;
+}
+
+void l3fwd_event_resource_setup(void);
+
+#endif /* __L3FWD_EVENTDEV_H__ */
@@ -13,12 +13,12 @@
#include <errno.h>
#include <getopt.h>
#include <signal.h>
-#include <stdbool.h>
#include <rte_common.h>
#include <rte_vect.h>
#include <rte_byteorder.h>
#include <rte_log.h>
+#include <rte_malloc.h>
#include <rte_memory.h>
#include <rte_memcpy.h>
#include <rte_eal.h>
@@ -33,7 +33,6 @@
#include <rte_random.h>
#include <rte_debug.h>
#include <rte_ether.h>
-#include <rte_ethdev.h>
#include <rte_mempool.h>
#include <rte_mbuf.h>
#include <rte_ip.h>
@@ -46,6 +45,7 @@
#include <cmdline_parse_etheraddr.h>
#include "l3fwd.h"
+#include "l3fwd_event.h"
/*
* Configurable number of RX/TX ring descriptors
@@ -274,7 +274,7 @@ init_lcore_rx_queues(void)
}
/* display usage */
-static void
+void
print_usage(const char *prgname)
{
fprintf(stderr, "%s [EAL options] --"
@@ -289,7 +289,9 @@ print_usage(const char *prgname)
" [--hash-entry-num]"
" [--ipv6]"
" [--parse-ptype]"
- " [--per-port-pool]\n\n"
+ " [--per-port-pool]"
+ " [--mode]"
+ " [--eventq-sched]\n\n"
" -p PORTMASK: Hexadecimal bitmask of ports to configure\n"
" -P : Enable promiscuous mode\n"
@@ -304,7 +306,13 @@ print_usage(const char *prgname)
" --hash-entry-num: Specify the hash entry number in hexadecimal to be setup\n"
" --ipv6: Set if running ipv6 packets\n"
" --parse-ptype: Set to use software to analyze packet type\n"
- " --per-port-pool: Use separate buffer pool per port\n\n",
+ " --per-port-pool: Use separate buffer pool per port\n"
+ " --mode: Packet transfer mode for I/O, poll or eventdev\n"
+ " Default mode = poll\n"
+ " --eventq-sched: Event queue synchronization method "
+ " ordered, atomic or parallel.\n\t\t"
+ " Default: atomic\n\t\t"
+ " Valid only if --mode=eventdev\n\n",
prgname);
}
@@ -504,11 +512,19 @@ static const struct option lgopts[] = {
static int
parse_args(int argc, char **argv)
{
+ struct l3fwd_event_resources *evt_rsrc;
int opt, ret;
char **argvopt;
int option_index;
char *prgname = argv[0];
+ evt_rsrc = l3fwd_get_eventdev_rsrc();
+ evt_rsrc->args = rte_zmalloc("l3fwd_event_args", sizeof(char *), 0);
+ if (evt_rsrc->args == NULL)
+ rte_exit(EXIT_FAILURE,
+ "Unable to allocate memory for eventdev arg");
+ evt_rsrc->args[0] = argv[0];
+ evt_rsrc->nb_args++;
argvopt = argv;
/* Error or normal output strings. */
@@ -538,6 +554,15 @@ parse_args(int argc, char **argv)
l3fwd_lpm_on = 1;
break;
+ case '?':
+ /* Eventdev options are encountered skip for
+ * now and processed later.
+ */
+ evt_rsrc->args[evt_rsrc->nb_args] =
+ argv[optind - 1];
+ evt_rsrc->nb_args++;
+ break;
+
/* long options */
case CMD_LINE_OPT_CONFIG_NUM:
ret = parse_config(optarg);
@@ -811,6 +836,7 @@ prepare_ptype_parser(uint16_t portid, uint16_t queueid)
int
main(int argc, char **argv)
{
+ struct l3fwd_event_resources *evt_rsrc;
struct lcore_conf *qconf;
struct rte_eth_dev_info dev_info;
struct rte_eth_txconf *txconf;
@@ -839,11 +865,16 @@ main(int argc, char **argv)
*(uint64_t *)(val_eth + portid) = dest_eth_addr[portid];
}
+ evt_rsrc = l3fwd_get_eventdev_rsrc();
+ RTE_SET_USED(evt_rsrc);
/* parse application arguments (after the EAL ones) */
ret = parse_args(argc, argv);
if (ret < 0)
rte_exit(EXIT_FAILURE, "Invalid L3FWD parameters\n");
+ /* Configure eventdev parameters if user has requested */
+ l3fwd_event_resource_setup();
+
if (check_lcore_params() < 0)
rte_exit(EXIT_FAILURE, "check_lcore_params failed\n");
@@ -6,7 +6,7 @@
# To build this example as a standalone application with an already-installed
# DPDK instance, use 'make'
-deps += ['hash', 'lpm']
+deps += ['hash', 'lpm', 'eventdev']
sources = files(
- 'l3fwd_em.c', 'l3fwd_lpm.c', 'main.c'
+ 'l3fwd_em.c', 'l3fwd_lpm.c', 'l3fwd_event.c', 'main.c'
)