Add tool: sysctl.

This commit contains an ipc library implemented by dpdk rte_ring and
sysctl tool ported from FreeBSD.

With this commit we can get and set FreeBSD kernel state in runtime.
dev
logwang 2017-05-23 23:13:49 +08:00
parent 7d25ffc210
commit 7abd0fb2a9
14 changed files with 2360 additions and 21 deletions

View File

@ -33,7 +33,7 @@ CORE_LIBS+=" -g -Wl,--no-as-needed -fvisibility=default -pthread -lm -lrt"
CORE_LIBS+=" -Wl,--whole-archive -lrte_pmd_vmxnet3_uio -lrte_pmd_i40e -lrte_pmd_ixgbe -lrte_pmd_e1000 -lrte_pmd_ring"
CORE_LIBS+=" -Wl,--whole-archive -lrte_hash -lrte_kvargs -Wl,-lrte_mbuf -lethdev -lrte_eal -Wl,-lrte_mempool"
CORE_LIBS+=" -lrte_ring -lrte_cmdline -lrte_cfgfile -lrte_kni -lrte_timer -Wl,-lrte_pmd_virtio"
CORE_LIBS+=" -Wl,--no-whole-archive -lrt -lm -ldl -lm -lcrypto"
CORE_LIBS+=" -Wl,--no-whole-archive -lrt -lm -ldl -lcrypto"
cat << END > $NGX_MAKEFILE

View File

@ -1,9 +1,19 @@
TOPDIR=..
ifeq ($(FF_PATH),)
FF_PATH=${TOPDIR}
endif
ifeq ($(FF_DPDK),)
FF_DPDK=${TOPDIR}/dpdk/x86_64-native-linuxapp-gcc
endif
LIBS+= -L${FF_PATH}/lib -L${FF_DPDK}/lib -Wl,--whole-archive,-lfstack,--no-whole-archive
LIBS+= -g -Wl,--no-as-needed -fvisibility=default -pthread -lm -lrt
LIBS+= -Wl,--whole-archive -lrte_pmd_vmxnet3_uio -lrte_pmd_i40e -lrte_pmd_ixgbe -lrte_pmd_e1000 -lrte_pmd_ring
LIBS+= -Wl,--whole-archive -lrte_hash -lrte_kvargs -Wl,-lrte_mbuf -lethdev -lrte_eal -Wl,-lrte_mempool
LIBS+= -lrte_ring -lrte_cmdline -lrte_cfgfile -lrte_kni -lrte_timer -Wl,-lrte_pmd_virtio
LIBS+= -Wl,--no-whole-archive -lrt -lm -ldl -lm -lcrypto
LIBS+= -Wl,--no-whole-archive -lrt -lm -ldl -lcrypto
TARGET="helloworld"
all:

View File

@ -55,11 +55,15 @@
#include "ff_config.h"
#include "ff_veth.h"
#include "ff_host_interface.h"
#include "ff_msg.h"
#include "ff_api.h"
#define MEMPOOL_CACHE_SIZE 256
#define ARP_RING_SIZE 2048
#define MSG_RING_SIZE 32
/*
* Configurable number of RX/TX ring descriptors
*/
@ -153,6 +157,16 @@ static struct rte_mempool *pktmbuf_pool[NB_SOCKETS];
static struct rte_ring **arp_ring[RTE_MAX_LCORE];
struct ff_msg_ring {
char ring_name[2][RTE_RING_NAMESIZE];
/* ring[0] for lcore recv msg, other send */
/* ring[1] for lcore send msg, other read */
struct rte_ring *ring[2];
} __rte_cache_aligned;
static struct ff_msg_ring msg_ring[RTE_MAX_LCORE];
static struct rte_mempool *message_pool;
struct ff_dpdk_if_context {
void *sc;
void *ifp;
@ -441,6 +455,25 @@ init_mem_pool(void)
return 0;
}
static struct rte_ring *
create_ring(const char *name, unsigned count, int socket_id, unsigned flags)
{
struct rte_ring *ring;
if (name == NULL)
return NULL;
/* If already create, just attached it */
if (likely((ring = rte_ring_lookup(name)) != NULL))
return ring;
if (rte_eal_process_type() == RTE_PROC_PRIMARY) {
return rte_ring_create(name, count, socket_id, flags);
} else {
return rte_ring_lookup(name);
}
}
static int
init_arp_ring(void)
{
@ -472,21 +505,14 @@ init_arp_ring(void)
uint8_t port_id = ff_global_cfg.dpdk.port_cfgs[j].port_id;
for(i = 0; i < nb_procs; ++i) {
snprintf(name_buf, RTE_RING_NAMESIZE, "ring_%d_%d", i, port_id);
if (rte_eal_process_type() == RTE_PROC_PRIMARY) {
arp_ring[i][port_id] = rte_ring_create(name_buf,
ARP_RING_SIZE, socketid,
RING_F_SC_DEQ);
if (rte_ring_lookup(name_buf) != arp_ring[i][port_id])
rte_panic("lookup arp ring:%s failed!\n", name_buf);
} else {
arp_ring[i][port_id] = rte_ring_lookup(name_buf);
}
snprintf(name_buf, RTE_RING_NAMESIZE, "arp_ring_%d_%d", i, port_id);
arp_ring[i][port_id] = create_ring(name_buf, ARP_RING_SIZE,
socketid, RING_F_SC_DEQ);
if (arp_ring[i][port_id] == NULL)
rte_panic("create arp ring::%s failed!\n", name_buf);
rte_panic("create ring:%s failed!\n", name_buf);
printf("create arp ring:%s success, %u ring entries are now free!\n",
printf("create ring:%s success, %u ring entries are now free!\n",
name_buf, rte_ring_free_count(arp_ring[i][port_id]));
}
}
@ -494,6 +520,58 @@ init_arp_ring(void)
return 0;
}
static void
ff_msg_init(struct rte_mempool *mp,
__attribute__((unused)) void *opaque_arg,
void *obj, __attribute__((unused)) unsigned i)
{
struct ff_msg *msg = (struct ff_msg *)obj;
msg->buf_addr = (char *)msg + sizeof(struct ff_msg);
msg->buf_len = mp->elt_size - sizeof(struct ff_msg);
}
static int
init_msg_ring(void)
{
uint16_t i;
uint16_t nb_procs = ff_global_cfg.dpdk.nb_procs;
unsigned socketid = lcore_conf.socket_id;
/* Create message buffer pool */
if (rte_eal_process_type() == RTE_PROC_PRIMARY) {
message_pool = rte_mempool_create(FF_MSG_POOL,
MSG_RING_SIZE * 2 * nb_procs,
MAX_MSG_BUF_SIZE, MSG_RING_SIZE / 2, 0,
NULL, NULL, ff_msg_init, NULL,
socketid, 0);
} else {
message_pool = rte_mempool_lookup(FF_MSG_POOL);
}
if (message_pool == NULL) {
rte_panic("Create msg mempool failed\n");
}
for(i = 0; i < nb_procs; ++i) {
snprintf(msg_ring[i].ring_name[0], RTE_RING_NAMESIZE,
"%s%u", FF_MSG_RING_IN, i);
snprintf(msg_ring[i].ring_name[1], RTE_RING_NAMESIZE,
"%s%u", FF_MSG_RING_OUT, i);
msg_ring[i].ring[0] = create_ring(msg_ring[i].ring_name[0],
MSG_RING_SIZE, socketid, RING_F_SP_ENQ | RING_F_SC_DEQ);
if (msg_ring[i].ring[0] == NULL)
rte_panic("create ring::%s failed!\n", msg_ring[i].ring_name[0]);
msg_ring[i].ring[1] = create_ring(msg_ring[i].ring_name[1],
MSG_RING_SIZE, socketid, RING_F_SP_ENQ | RING_F_SC_DEQ);
if (msg_ring[i].ring[1] == NULL)
rte_panic("create ring::%s failed!\n", msg_ring[i].ring_name[0]);
}
return 0;
}
static int
init_kni(void)
{
@ -730,6 +808,8 @@ ff_dpdk_init(int argc, char **argv)
init_arp_ring();
init_msg_ring();
enable_kni = ff_global_cfg.kni.enable;
if (enable_kni) {
init_kni();
@ -872,12 +952,61 @@ process_arp_ring(uint8_t port_id, uint16_t queue_id,
struct rte_mbuf **pkts_burst, const struct ff_dpdk_if_context *ctx)
{
/* read packet from ring buf and to process */
uint16_t nb_tx;
nb_tx = rte_ring_dequeue_burst(arp_ring[queue_id][port_id],
uint16_t nb_rb;
nb_rb = rte_ring_dequeue_burst(arp_ring[queue_id][port_id],
(void **)pkts_burst, MAX_PKT_BURST);
if(nb_tx > 0) {
process_packets(port_id, queue_id, pkts_burst, nb_tx, ctx, 1);
if(nb_rb > 0) {
process_packets(port_id, queue_id, pkts_burst, nb_rb, ctx, 1);
}
return 0;
}
static inline void
handle_sysctl_msg(struct ff_msg *msg, uint16_t proc_id)
{
int ret = ff_sysctl(msg->sysctl.name, msg->sysctl.namelen,
msg->sysctl.old, msg->sysctl.oldlenp, msg->sysctl.new,
msg->sysctl.newlen);
if (ret < 0) {
msg->result = errno;
} else {
msg->result = 0;
}
rte_ring_enqueue(msg_ring[proc_id].ring[1], msg);
}
static inline void
handle_default_msg(struct ff_msg *msg, uint16_t proc_id)
{
msg->result = EINVAL;
rte_ring_enqueue(msg_ring[proc_id].ring[1], msg);
}
static inline void
handle_msg(struct ff_msg *msg, uint16_t proc_id)
{
switch (msg->msg_type) {
case FF_SYSCTL:
handle_sysctl_msg(msg, proc_id);
break;
default:
handle_default_msg(msg, proc_id);
break;
}
}
static inline int
process_msg_ring(uint16_t proc_id)
{
void *msg;
int ret = rte_ring_dequeue(msg_ring[proc_id].ring[0], &msg);
if (unlikely(ret == 0)) {
handle_msg((struct ff_msg *)msg, proc_id);
}
return 0;
@ -1104,6 +1233,8 @@ main_loop(void *arg)
}
}
process_msg_ring(qconf->proc_id);
if (likely(lr->loop != NULL)) {
lr->loop(lr->arg);
}

View File

@ -383,6 +383,9 @@ ff_kni_alloc(uint8_t port_id, unsigned socket_id,
if (rte_eal_process_type() == RTE_PROC_PRIMARY) {
kni_rp[port_id] = rte_ring_create(ring_name, KNI_QUEUE_SIZE,
socket_id, RING_F_SC_DEQ);
if (rte_ring_lookup(ring_name) != kni_rp[port_id])
rte_panic("lookup kni ring failed!\n");
} else {
kni_rp[port_id] = rte_ring_lookup(ring_name);
}
@ -390,9 +393,6 @@ ff_kni_alloc(uint8_t port_id, unsigned socket_id,
if (kni_rp[port_id] == NULL)
rte_panic("create kni ring failed!\n");
if (rte_ring_lookup(ring_name) != kni_rp[port_id])
rte_panic("lookup kni ring failed!\n");
printf("create kni ring success, %u ring entries are now free!\n",
rte_ring_free_count(kni_rp[port_id]));
}

68
lib/ff_msg.h Normal file
View File

@ -0,0 +1,68 @@
/*
* Copyright (C) 2017 THL A29 Limited, a Tencent company.
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions are met:
*
* 1. Redistributions of source code must retain the above copyright notice, this
* list of conditions and the following disclaimer.
* 2. 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.
*
* 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 _FF_MSG_H_
#define _FF_MSG_H_
#include <rte_memory.h>
#define FF_MSG_RING_IN "ff_msg_ring_in_"
#define FF_MSG_RING_OUT "ff_msg_ring_out_"
#define FF_MSG_POOL "ff_msg_pool"
/* MSG TYPE: sysctl, sysctlbyname, etc.. */
enum FF_MSG_TYPE {
FF_UNKNOWN = 0,
FF_SYSCTL,
};
struct ff_sysctl_args {
int *name;
unsigned namelen;
void *old;
size_t *oldlenp;
void *new;
size_t newlen;
};
#define MAX_MSG_BUF_SIZE 10240
/* structure of ipc msg */
struct ff_msg {
enum FF_MSG_TYPE msg_type;
/* Result of msg processing */
int result;
/* Length of segment buffer. */
uint16_t buf_len;
/* Address of segment buffer. */
char *buf_addr;
union {
struct ff_sysctl_args sysctl;
};
} __attribute__((packed)) __rte_cache_aligned;
#endif

7
tools/Makefile Normal file
View File

@ -0,0 +1,7 @@
SUBDIRS=ipc sysctl
all:
for d in $(SUBDIRS); do ( cd $$d; $(MAKE) all ) ; done
clean:
for d in $(SUBDIRS); do ( cd $$d; $(MAKE) clean ) ; done

120
tools/README.md Normal file
View File

@ -0,0 +1,120 @@
# Introduction
Directory `ipc` implements an ipc library using dpdk `rte_ring`, can be used to communicate with F-Stack processes.
All other directories are useful tools ported from FreeBSD.
# ipc
This is a simple implemention using dpdk `rte_ring`.
`ff_ipc_msg_alloc` get msg structure from rte_mempool.
`ff_ipc_msg_free` put msg to rte_mempool.
`ff_ipc_send` enqueue msg to rte_ring.
`ff_ipc_recv` dequeue msg from rte_ring.
Since F-Stack is multi-process architecture and every process has an independent stack, so we must communicate with every F-Stack process.
# sysctl
Usage:
`sysctl -p <f-stack proc_id> [-bdehiNnoqTtWx] [ -B <bufsize> ] [-f filename] name[=value] ...`
`sysctl -p <f-stack proc_id> [-bdehNnoqTtWx] [ -B <bufsize> ] -a`
-p Which F-Stack process to communicate with, default 0.
Except this option, it is same with the original FreeBSD sysctl, see [Manual page](https://www.freebsd.org/cgi/man.cgi?sysctl).
# how to implement a custom tool for communicating with F-Stack process
Add a new FF_MSG_TYPE in ff_msg.h:
```
enum FF_MSG_TYPE {
FF_UNKNOWN = 0,
FF_SYSCTL,
FF_HELLOWORLD,
};
```
Define a structure used to communicate:
```
struct ff_helloworld_args {
void *request;
size_t req_len;
void *reply;
size_t rep_len;
};
```
Note that, when using struct ff_helloworld_args, pointers in this structure must point to the addresses range from ff_msg.buf_addr and ff_msg.buf_addr+ff_msg.buf_len, ff_msg.buf_len is (10240 - sizeof(struct ff_msg)).
And add it to ff_msg:
```
struct ff_msg {
...
union {
struct ff_sysctl_args sysctl;
struct ff_helloworld_args helloworld;
};
};
```
Modify ff_dpdk_if.c, add a handle function:
```
static inline void
handle_helloworld_msg(struct ff_msg *msg, uint16_t proc_id)
{
printf("helloworld msg recved.\n");
msg->result = 0;
rte_ring_enqueue(msg_ring[proc_id].ring[1], msg);
}
static inline void
handle_msg(struct ff_msg *msg, uint16_t proc_id)
{
switch (msg->msg_type) {
case FF_SYSCTL:
handle_sysctl_msg(msg, proc_id);
break;
case FF_HELLOWORLD:
handle_helloworld_msg(msg, proc_id);
default:
handle_default_msg(msg, proc_id);
break;
}
}
```
Create helloworld.c:
```
int main()
{
struct ff_msg *msg = ff_ipc_msg_alloc();
char *buf = msg->buf_addr;
msg->helloworld.request = buf;
memcpy(msg->helloworld.request, "hello", 5);
msg->helloworld.req_len = 5;
buf += 5;
msg->helloworld.reply = buf;
msg->helloworld.rep_len = 10;
ff_ipc_send(msg, 0);
struct ff_msg *retmsg;
ff_ipc_recv(retmsg, 0);
assert(remsg==msg);
ff_ipc_msg_free(msg);
}
```
The Makefile may like this:
```
TOPDIR?=${CURDIR}/../..
PROG=helloworld
include ${TOPDIR}/tools/prog.mk
```

32
tools/ipc/Makefile Normal file
View File

@ -0,0 +1,32 @@
TOPDIR?=${CURDIR}/../..
ifeq ($(FF_DPDK),)
FF_DPDK=${TOPDIR}/dpdk/x86_64-native-linuxapp-gcc
endif
TARGET=libfstack_ipc.a
DPDK_CFLAGS= -g -Wall -Werror -include ${FF_DPDK}/include/rte_config.h
DPDK_CFLAGS+= -march=native -DRTE_MACHINE_CPUFLAG_SSE -DRTE_MACHINE_CPUFLAG_SSE2 -DRTE_MACHINE_CPUFLAG_SSE3
DPDK_CFLAGS+= -DRTE_MACHINE_CPUFLAG_SSSE3 -DRTE_MACHINE_CPUFLAG_SSE4_1 -DRTE_MACHINE_CPUFLAG_SSE4_2
DPDK_CFLAGS+= -DRTE_COMPILE_TIME_CPUFLAGS=RTE_CPUFLAG_SSE,RTE_CPUFLAG_SSE2,RTE_CPUFLAG_SSE3,RTE_CPUFLAG_SSSE3,RTE_CPUFLAG_SSE4_1,RTE_CPUFLAG_SSE4_2
DPDK_CFLAGS+= -I${FF_DPDK}/include
CFLAGS+= ${DPDK_CFLAGS}
CFLAGS+= -I${TOPDIR}/lib
SRCS=ff_ipc.c
OBJS=$(patsubst %.c,%.o,${SRCS})
all: ${TARGET}
${TARGET}: ${OBJS}
ar -cqs $@ ${OBJS}
${OBJS}: %.o: %.c
${CC} -c $< ${CFLAGS} -o $@
.PHONY: clean
clean:
rm -f *.o ${TARGET}

161
tools/ipc/ff_ipc.c Normal file
View File

@ -0,0 +1,161 @@
/*
* Copyright (C) 2017 THL A29 Limited, a Tencent company.
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions are met:
*
* 1. Redistributions of source code must retain the above copyright notice, this
* list of conditions and the following disclaimer.
* 2. 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.
*
* 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_common.h>
#include <rte_memory.h>
#include <rte_config.h>
#include <rte_eal.h>
#include <rte_ring.h>
#include <rte_mempool.h>
#include <rte_malloc.h>
#include <unistd.h>
#include "ff_ipc.h"
static int inited;
static struct rte_mempool *message_pool;
static int
ff_ipc_init(void)
{
if (inited) {
return 0;
}
char *dpdk_argv[] = {
"-c1", "-n4",
"--proc-type=secondary",
"--log-level=0",
};
int ret = rte_eal_init(sizeof(dpdk_argv)/sizeof(dpdk_argv[0]), dpdk_argv);
if (ret < 0) {
rte_exit(EXIT_FAILURE, "Error with EAL initialization\n");
}
message_pool = rte_mempool_lookup(FF_MSG_POOL);
if (message_pool == NULL) {
rte_exit(EXIT_FAILURE, "lookup message pool:%s failed!\n", FF_MSG_POOL);
}
inited = 1;
return 0;
}
struct ff_msg *
ff_ipc_msg_alloc(void)
{
if (inited == 0) {
int ret = ff_ipc_init();
if (ret < 0) {
return NULL;
}
}
void *msg;
if (rte_mempool_get(message_pool, &msg) < 0) {
printf("get buffer from message pool failed.\n");
return NULL;
}
return (struct ff_msg *)msg;
}
int
ff_ipc_msg_free(struct ff_msg *msg)
{
if (inited == 0) {
printf("ff ipc not inited\n");
return -1;
}
rte_mempool_put(message_pool, msg);
return 0;
}
int
ff_ipc_send(const struct ff_msg *msg, uint16_t proc_id)
{
int ret;
if (inited == 0) {
printf("ff ipc not inited\n");
return -1;
}
char name[RTE_RING_NAMESIZE];
snprintf(name, RTE_RING_NAMESIZE, "%s%u",
FF_MSG_RING_IN, proc_id);
struct rte_ring *ring = rte_ring_lookup(name);
if (ring == NULL) {
printf("lookup message ring:%s failed!\n", name);
return -1;
}
ret = rte_ring_enqueue(ring, (void *)msg);
if (ret < 0) {
printf("ff_ipc_send failed\n");
return ret;
}
return 0;
}
int
ff_ipc_recv(struct ff_msg **msg, uint16_t proc_id)
{
int ret, i;
if (inited == 0) {
printf("ff ipc not inited\n");
return -1;
}
char name[RTE_RING_NAMESIZE];
snprintf(name, RTE_RING_NAMESIZE, "%s%u",
FF_MSG_RING_OUT, proc_id);
struct rte_ring *ring = rte_ring_lookup(name);
if (ring == NULL) {
printf("lookup message ring:%s failed!\n", name);
return -1;
}
void *obj;
#define MAX_ATTEMPTS_NUM 1000
for (i = 0; i < MAX_ATTEMPTS_NUM; i++) {
ret = rte_ring_dequeue(ring, &obj);
if (ret == 0) {
*msg = (struct ff_msg *)obj;
break;
}
usleep(1000);
}
return ret;
}

38
tools/ipc/ff_ipc.h Normal file
View File

@ -0,0 +1,38 @@
/*
* Copyright (C) 2017 THL A29 Limited, a Tencent company.
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions are met:
*
* 1. Redistributions of source code must retain the above copyright notice, this
* list of conditions and the following disclaimer.
* 2. 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.
*
* 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 _FF_IPC_H_
#define _FF_IPC_H_
#include "ff_msg.h"
struct ff_msg *ff_ipc_msg_alloc(void);
int ff_ipc_msg_free(struct ff_msg *msg);
int ff_ipc_send(const struct ff_msg *msg, uint16_t proc_id);
int ff_ipc_recv(struct ff_msg **msg, uint16_t proc_id);
#endif

71
tools/prog.mk Normal file
View File

@ -0,0 +1,71 @@
#
# Derived from FreeBSD src/share/mk/bsd.prog.mk
#
ifdef DEBUG_FLAGS
CFLAGS+=${DEBUG_FLAGS}
CXXFLAGS+=${DEBUG_FLAGS}
endif
ifdef NO_SHARED
ifneq (${NO_SHARED},no)
ifneq (${NO_SHARED},NO)
LDFLAGS+= -static
endif
endif
endif
ifdef PROG_CXX
PROG= ${PROG_CXX}
endif
ifndef PROG
$(error PROG or PROG_CXX must be defined.)
endif
ifndef TOPDIR
$(error TOPDIR must be defined.)
endif
ifndef SRCS
ifdef PROG_CXX
SRCS= ${PROG}.cc
else
SRCS= ${PROG}.c
endif
endif
OBJS+= $(patsubst %.cc,%.o,$(patsubst %.c,%.o,${SRCS}))
ifeq ($(FF_DPDK),)
FF_DPDK=${TOPDIR}/dpdk/x86_64-native-linuxapp-gcc
endif
FF_PROG_CFLAGS:= -g -Wall -Werror -DFSTACK
FF_PROG_CFLAGS+= -I${TOPDIR}/lib -I${TOPDIR}/tools/ipc
FF_PROG_CFLAGS+= -include ${FF_DPDK}/include/rte_config.h
FF_PROG_CFLAGS+= -I${FF_DPDK}/include
FF_PROG_LIBS:= -L${TOPDIR}/tools/ipc -lfstack_ipc
FF_PROG_LIBS+= -L${FF_DPDK}/lib
FF_PROG_LIBS+= -g -Wl,--no-as-needed -fvisibility=default -pthread -lm -lrt
FF_PROG_LIBS+= -Wl,--whole-archive -lrte_eal -Wl,-lrte_mempool -lrte_ring
FF_PROG_LIBS+= -Wl,--no-whole-archive -lrt -lm -ldl -lcrypto
CFLAGS+= ${FF_PROG_CFLAGS}
CXXFLAGS+= ${FF_PROG_CFLAGS}
LIBS+= ${FF_PROG_LIBS}
${PROG}: ${OBJS}
ifdef PROG_CXX
${CXX} ${CXXFLAGS} ${LDFLAGS} -o $@ ${OBJS} ${LIBS}
else
${CC} ${CFLAGS} ${LDFLAGS} -o $@ ${OBJS} ${LIBS}
endif
clean:
@rm -f ${PROG} ${OBJS}
all: ${PROG}

5
tools/sysctl/Makefile Normal file
View File

@ -0,0 +1,5 @@
TOPDIR?=${CURDIR}/../..
PROG=sysctl
include ${TOPDIR}/tools/prog.mk

326
tools/sysctl/sysctl.8 Normal file
View File

@ -0,0 +1,326 @@
.\" Copyright (c) 1993
.\" The Regents of the University of California. All rights reserved.
.\"
.\" Redistribution and use in source and binary forms, with or without
.\" modification, are permitted provided that the following conditions
.\" are met:
.\" 1. Redistributions of source code must retain the above copyright
.\" notice, this list of conditions and the following disclaimer.
.\" 2. 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.
.\" 4. Neither the name of the University 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 REGENTS 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 REGENTS 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.
.\"
.\" From: @(#)sysctl.8 8.1 (Berkeley) 6/6/93
.\" $FreeBSD$
.\"
.Dd December 10, 2015
.Dt SYSCTL 8
.Os
.Sh NAME
.Nm sysctl
.Nd get or set kernel state
.Sh SYNOPSIS
.Nm
.Op Fl bdehiNnoRTtqx
.Op Fl B Ar bufsize
.Op Fl f Ar filename
.Ar name Ns Op = Ns Ar value
.Ar ...
.Nm
.Op Fl bdehNnoRTtqx
.Op Fl B Ar bufsize
.Fl a
.Sh DESCRIPTION
The
.Nm
utility retrieves kernel state and allows processes with appropriate
privilege to set kernel state.
The state to be retrieved or set is described using a
.Dq Management Information Base
.Pq Dq MIB
style name, described as a dotted set of
components.
.Pp
The following options are available:
.Bl -tag -width indent
.It Fl A
Equivalent to
.Fl o a
(for compatibility).
.It Fl a
List all the currently available non-opaque values.
This option is ignored if one or more variable names are specified on
the command line.
.It Fl b
Force the value of the variable(s) to be output in raw, binary format.
No names are printed and no terminating newlines are output.
This is mostly useful with a single variable.
.It Fl B Ar bufsize
Set the buffer size to read from the
.Nm
to
.Ar bufsize .
This is necessary for a
.Nm
that has variable length, and the probe value of 0 is a valid length, such as
.Va kern.arandom .
.It Fl d
Print the description of the variable instead of its value.
.It Fl e
Separate the name and the value of the variable(s) with
.Ql = .
This is useful for producing output which can be fed back to the
.Nm
utility.
This option is ignored if either
.Fl N
or
.Fl n
is specified, or a variable is being set.
.It Fl f Ar filename
Specify a file which contains a pair of name and value in each line.
.Nm
reads and processes the specified file first and then processes the name
and value pairs in the command line argument.
.It Fl h
Format output for human, rather than machine, readability.
.It Fl i
Ignore unknown OIDs.
The purpose is to make use of
.Nm
for collecting data from a variety of machines (not all of which
are necessarily running exactly the same software) easier.
.It Fl N
Show only variable names, not their values.
This is particularly useful with shells that offer programmable
completion.
To enable completion of variable names in
.Xr zsh 1 Pq Pa ports/shells/zsh ,
use the following code:
.Bd -literal -offset indent
listsysctls () { set -A reply $(sysctl -AN ${1%.*}) }
compctl -K listsysctls sysctl
.Ed
.Pp
To enable completion of variable names in
.Xr tcsh 1 ,
use:
.Pp
.Dl "complete sysctl 'n/*/`sysctl -Na`/'"
.It Fl n
Show only variable values, not their names.
This option is useful for setting shell variables.
For instance, to save the pagesize in variable
.Va psize ,
use:
.Pp
.Dl "set psize=`sysctl -n hw.pagesize`"
.It Fl o
Show opaque variables (which are normally suppressed).
The format and length are printed, as well as a hex dump of the first
sixteen bytes of the value.
.It Fl q
Suppress some warnings generated by
.Nm
to standard error.
.It Fl T
Display only variables that are settable via loader (CTLFLAG_TUN).
.It Fl t
Print the type of the variable.
.It Fl W
Display only writable variables that are not statistical.
Useful for determining the set of runtime tunable sysctls.
.It Fl X
Equivalent to
.Fl x a
(for compatibility).
.It Fl x
As
.Fl o ,
but prints a hex dump of the entire value instead of just the first
few bytes.
.El
.Pp
The information available from
.Nm
consists of integers, strings, and opaque types.
The
.Nm
utility
only knows about a couple of opaque types, and will resort to hexdumps
for the rest.
The opaque information is much more useful if retrieved by special
purpose programs such as
.Xr ps 1 ,
.Xr systat 1 ,
and
.Xr netstat 1 .
.Pp
Some of the variables which cannot be modified during normal system
operation can be initialized via
.Xr loader 8
tunables.
This can for example be done by setting them in
.Xr loader.conf 5 .
Please refer to
.Xr loader.conf 5
for more information on which tunables are available and how to set them.
.Pp
The string and integer information is summarized below.
For a detailed description of these variable see
.Xr sysctl 3 .
.Pp
The changeable column indicates whether a process with appropriate
privilege can change the value.
String and integer values can be set using
.Nm .
.Bl -column security.bsd.unprivileged_read_msgbuf integerxxx
.It Sy "Name Type Changeable"
.It "kern.ostype string no"
.It "kern.osrelease string no"
.It "kern.osrevision integer no"
.It "kern.version string no"
.It "kern.maxvnodes integer yes"
.It "kern.maxproc integer no"
.It "kern.maxprocperuid integer yes"
.It "kern.maxfiles integer yes"
.It "kern.maxfilesperproc integer yes"
.It "kern.argmax integer no"
.It "kern.securelevel integer raise only"
.It "kern.hostname string yes"
.It "kern.hostid integer yes"
.It "kern.clockrate struct no"
.It "kern.posix1version integer no"
.It "kern.ngroups integer no"
.It "kern.job_control integer no"
.It "kern.saved_ids integer no"
.It "kern.boottime struct no"
.It "kern.domainname string yes"
.It "kern.filedelay integer yes"
.It "kern.dirdelay integer yes"
.It "kern.metadelay integer yes"
.It "kern.osreldate string no"
.It "kern.bootfile string yes"
.It "kern.corefile string yes"
.It "kern.logsigexit integer yes"
.It "security.bsd.suser_enabled integer yes"
.It "security.bsd.see_other_uids integer yes"
.It "security.bsd.unprivileged_proc_debug integer yes"
.It "security.bsd.unprivileged_read_msgbuf integer yes"
.It "vm.loadavg struct no"
.It "hw.machine string no"
.It "hw.model string no"
.It "hw.ncpu integer no"
.It "hw.byteorder integer no"
.It "hw.physmem integer no"
.It "hw.usermem integer no"
.It "hw.pagesize integer no"
.It "hw.floatingpoint integer no"
.It "hw.machine_arch string no"
.It "hw.realmem integer no"
.It "machdep.adjkerntz integer yes"
.It "machdep.disable_rtc_set integer yes"
.It "machdep.guessed_bootdev string no"
.It "user.cs_path string no"
.It "user.bc_base_max integer no"
.It "user.bc_dim_max integer no"
.It "user.bc_scale_max integer no"
.It "user.bc_string_max integer no"
.It "user.coll_weights_max integer no"
.It "user.expr_nest_max integer no"
.It "user.line_max integer no"
.It "user.re_dup_max integer no"
.It "user.posix2_version integer no"
.It "user.posix2_c_bind integer no"
.It "user.posix2_c_dev integer no"
.It "user.posix2_char_term integer no"
.It "user.posix2_fort_dev integer no"
.It "user.posix2_fort_run integer no"
.It "user.posix2_localedef integer no"
.It "user.posix2_sw_dev integer no"
.It "user.posix2_upe integer no"
.It "user.stream_max integer no"
.It "user.tzname_max integer no"
.El
.Sh FILES
.Bl -tag -width ".In netinet/icmp_var.h" -compact
.It In sys/sysctl.h
definitions for top level identifiers, second level kernel and hardware
identifiers, and user level identifiers
.It In sys/socket.h
definitions for second level network identifiers
.It In sys/gmon.h
definitions for third level profiling identifiers
.It In vm/vm_param.h
definitions for second level virtual memory identifiers
.It In netinet/in.h
definitions for third level Internet identifiers and
fourth level IP identifiers
.It In netinet/icmp_var.h
definitions for fourth level ICMP identifiers
.It In netinet/udp_var.h
definitions for fourth level UDP identifiers
.El
.Sh EXAMPLES
For example, to retrieve the maximum number of processes allowed
in the system, one would use the following request:
.Pp
.Dl "sysctl kern.maxproc"
.Pp
To set the maximum number of processes allowed
per uid to 1000, one would use the following request:
.Pp
.Dl "sysctl kern.maxprocperuid=1000"
.Pp
Information about the system clock rate may be obtained with:
.Pp
.Dl "sysctl kern.clockrate"
.Pp
Information about the load average history may be obtained with:
.Pp
.Dl "sysctl vm.loadavg"
.Pp
More variables than these exist, and the best and likely only place
to search for their deeper meaning is undoubtedly the source where
they are defined.
.Sh COMPATIBILITY
The
.Fl w
option has been deprecated and is silently ignored.
.Sh SEE ALSO
.Xr sysctl 3 ,
.Xr loader.conf 5 ,
.Xr sysctl.conf 5 ,
.Xr loader 8
.Sh HISTORY
A
.Nm
utility first appeared in
.Bx 4.4 .
.Pp
In
.Fx 2.2 ,
.Nm
was significantly remodeled.
.Sh BUGS
The
.Nm
utility presently exploits an undocumented interface to the kernel
sysctl facility to traverse the sysctl tree and to retrieve format
and name information.
This correct interface is being thought about for the time being.

1370
tools/sysctl/sysctl.c Normal file

File diff suppressed because it is too large Load Diff