freebsd-nq/usr.bin/netstat/main.c

704 lines
16 KiB
C
Raw Normal View History

1994-05-27 12:33:43 +00:00
/*
* Copyright (c) 1983, 1988, 1993
* 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.
* 3. All advertising materials mentioning features or use of this software
* must display the following acknowledgement:
* This product includes software developed by the University of
* California, Berkeley and its contributors.
* 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.
*/
#ifndef lint
char const copyright[] =
1994-05-27 12:33:43 +00:00
"@(#) Copyright (c) 1983, 1988, 1993\n\
Regents of the University of California. All rights reserved.\n";
#endif /* not lint */
#ifndef lint
#if 0
1994-05-27 12:33:43 +00:00
static char sccsid[] = "@(#)main.c 8.4 (Berkeley) 3/1/94";
#endif
static const char rcsid[] =
1999-08-28 01:08:13 +00:00
"$FreeBSD$";
1994-05-27 12:33:43 +00:00
#endif /* not lint */
#include <sys/param.h>
#include <sys/file.h>
#include <sys/protosw.h>
#include <sys/socket.h>
#include <netinet/in.h>
#include <netgraph/ng_socket.h>
1994-05-27 12:33:43 +00:00
#include <ctype.h>
#include <err.h>
1994-05-27 12:33:43 +00:00
#include <errno.h>
#include <kvm.h>
#include <limits.h>
#include <netdb.h>
#include <nlist.h>
#include <paths.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>
#include "netstat.h"
1998-08-08 08:13:04 +00:00
static struct nlist nl[] = {
#define N_IFNET 0
1994-05-27 12:33:43 +00:00
{ "_ifnet" },
#define N_RTSTAT 1
1994-05-27 12:33:43 +00:00
{ "_rtstat" },
#define N_RTREE 2
1994-05-27 12:33:43 +00:00
{ "_rt_tables"},
#define N_MRTSTAT 3
1994-05-27 12:33:43 +00:00
{ "_mrtstat" },
#define N_MFCTABLE 4
{ "_mfctable" },
#define N_VIFTABLE 5
1994-05-27 12:33:43 +00:00
{ "_viftable" },
#define N_IPX 6
{ "_ipxpcb"},
#define N_IPXSTAT 7
{ "_ipxstat"},
#define N_SPXSTAT 8
{ "_spx_istat"},
#define N_DDPSTAT 9
{ "_ddpstat"},
#define N_DDPCB 10
{ "_ddpcb"},
#define N_NGSOCKS 11
{ "_ngsocklist"},
#define N_IP6STAT 12
{ "_ip6stat" },
#define N_ICMP6STAT 13
{ "_icmp6stat" },
#define N_IPSECSTAT 14
{ "_ipsecstat" },
#define N_IPSEC6STAT 15
{ "_ipsec6stat" },
#define N_PIM6STAT 16
{ "_pim6stat" },
#define N_MRT6STAT 17
{ "_mrt6stat" },
#define N_MF6CTABLE 18
{ "_mf6ctable" },
#define N_MIF6TABLE 19
{ "_mif6table" },
#define N_PFKEYSTAT 20
2000-07-05 02:02:54 +00:00
{ "_pfkeystat" },
#define N_MBSTAT 21
{ "_mbstat" },
#define N_MBTYPES 22
{ "_mbtypes" },
#define N_NMBCLUSTERS 23
{ "_nmbclusters" },
#define N_NMBUFS 24
{ "_nmbufs" },
#define N_MBHI 25
{ "_mbuf_hiwm" },
#define N_CLHI 26
{ "_clust_hiwm" },
#define N_NCPUS 27
Introduce numerous SMP friendly changes to the mbuf allocator. Namely, introduce a modified allocation mechanism for mbufs and mbuf clusters; one which can scale under SMP and which offers the possibility of resource reclamation to be implemented in the future. Notable advantages: o Reduce contention for SMP by offering per-CPU pools and locks. o Better use of data cache due to per-CPU pools. o Much less code cache pollution due to excessively large allocation macros. o Framework for `grouping' objects from same page together so as to be able to possibly free wired-down pages back to the system if they are no longer needed by the network stacks. Additional things changed with this addition: - Moved some mbuf specific declarations and initializations from sys/conf/param.c into mbuf-specific code where they belong. - m_getclr() has been renamed to m_get_clrd() because the old name is really confusing. m_getclr() HAS been preserved though and is defined to the new name. No tree sweep has been done "to change the interface," as the old name will continue to be supported and is not depracated. The change was merely done because m_getclr() sounds too much like "m_get a cluster." - TEMPORARILY disabled mbtypes statistics displaying in netstat(1) and systat(1) (see TODO below). - Fixed systat(1) to display number of "free mbufs" based on new per-CPU stat structures. - Fixed netstat(1) to display new per-CPU stats based on sysctl-exported per-CPU stat structures. All infos are fetched via sysctl. TODO (in order of priority): - Re-enable mbtypes statistics in both netstat(1) and systat(1) after introducing an SMP friendly way to collect the mbtypes stats under the already introduced per-CPU locks (i.e. hopefully don't use atomic() - it seems too costly for a mere stat update, especially when other locks are already present). - Optionally have systat(1) display not only "total free mbufs" but also "total free mbufs per CPU pool." - Fix minor length-fetching issues in netstat(1) related to recently re-enabled option to read mbuf stats from a core file. - Move reference counters at least for mbuf clusters into an unused portion of the cluster itself, to save space and need to allocate a counter. - Look into introducing resource freeing possibly from a kproc. Reviewed by (in parts): jlemon, jake, silby, terry Tested by: jlemon (Intel & Alpha), mjacob (Intel & Alpha) Preliminary performance measurements: jlemon (and me, obviously) URL: http://people.freebsd.org/~bmilekic/mb_alloc/
2001-06-22 06:35:32 +00:00
{ "_smp_cpus" },
#define N_PAGESZ 28
Introduce numerous SMP friendly changes to the mbuf allocator. Namely, introduce a modified allocation mechanism for mbufs and mbuf clusters; one which can scale under SMP and which offers the possibility of resource reclamation to be implemented in the future. Notable advantages: o Reduce contention for SMP by offering per-CPU pools and locks. o Better use of data cache due to per-CPU pools. o Much less code cache pollution due to excessively large allocation macros. o Framework for `grouping' objects from same page together so as to be able to possibly free wired-down pages back to the system if they are no longer needed by the network stacks. Additional things changed with this addition: - Moved some mbuf specific declarations and initializations from sys/conf/param.c into mbuf-specific code where they belong. - m_getclr() has been renamed to m_get_clrd() because the old name is really confusing. m_getclr() HAS been preserved though and is defined to the new name. No tree sweep has been done "to change the interface," as the old name will continue to be supported and is not depracated. The change was merely done because m_getclr() sounds too much like "m_get a cluster." - TEMPORARILY disabled mbtypes statistics displaying in netstat(1) and systat(1) (see TODO below). - Fixed systat(1) to display number of "free mbufs" based on new per-CPU stat structures. - Fixed netstat(1) to display new per-CPU stats based on sysctl-exported per-CPU stat structures. All infos are fetched via sysctl. TODO (in order of priority): - Re-enable mbtypes statistics in both netstat(1) and systat(1) after introducing an SMP friendly way to collect the mbtypes stats under the already introduced per-CPU locks (i.e. hopefully don't use atomic() - it seems too costly for a mere stat update, especially when other locks are already present). - Optionally have systat(1) display not only "total free mbufs" but also "total free mbufs per CPU pool." - Fix minor length-fetching issues in netstat(1) related to recently re-enabled option to read mbuf stats from a core file. - Move reference counters at least for mbuf clusters into an unused portion of the cluster itself, to save space and need to allocate a counter. - Look into introducing resource freeing possibly from a kproc. Reviewed by (in parts): jlemon, jake, silby, terry Tested by: jlemon (Intel & Alpha), mjacob (Intel & Alpha) Preliminary performance measurements: jlemon (and me, obviously) URL: http://people.freebsd.org/~bmilekic/mb_alloc/
2001-06-22 06:35:32 +00:00
{ "_pagesize" },
#define N_MBPSTAT 29
Introduce numerous SMP friendly changes to the mbuf allocator. Namely, introduce a modified allocation mechanism for mbufs and mbuf clusters; one which can scale under SMP and which offers the possibility of resource reclamation to be implemented in the future. Notable advantages: o Reduce contention for SMP by offering per-CPU pools and locks. o Better use of data cache due to per-CPU pools. o Much less code cache pollution due to excessively large allocation macros. o Framework for `grouping' objects from same page together so as to be able to possibly free wired-down pages back to the system if they are no longer needed by the network stacks. Additional things changed with this addition: - Moved some mbuf specific declarations and initializations from sys/conf/param.c into mbuf-specific code where they belong. - m_getclr() has been renamed to m_get_clrd() because the old name is really confusing. m_getclr() HAS been preserved though and is defined to the new name. No tree sweep has been done "to change the interface," as the old name will continue to be supported and is not depracated. The change was merely done because m_getclr() sounds too much like "m_get a cluster." - TEMPORARILY disabled mbtypes statistics displaying in netstat(1) and systat(1) (see TODO below). - Fixed systat(1) to display number of "free mbufs" based on new per-CPU stat structures. - Fixed netstat(1) to display new per-CPU stats based on sysctl-exported per-CPU stat structures. All infos are fetched via sysctl. TODO (in order of priority): - Re-enable mbtypes statistics in both netstat(1) and systat(1) after introducing an SMP friendly way to collect the mbtypes stats under the already introduced per-CPU locks (i.e. hopefully don't use atomic() - it seems too costly for a mere stat update, especially when other locks are already present). - Optionally have systat(1) display not only "total free mbufs" but also "total free mbufs per CPU pool." - Fix minor length-fetching issues in netstat(1) related to recently re-enabled option to read mbuf stats from a core file. - Move reference counters at least for mbuf clusters into an unused portion of the cluster itself, to save space and need to allocate a counter. - Look into introducing resource freeing possibly from a kproc. Reviewed by (in parts): jlemon, jake, silby, terry Tested by: jlemon (Intel & Alpha), mjacob (Intel & Alpha) Preliminary performance measurements: jlemon (and me, obviously) URL: http://people.freebsd.org/~bmilekic/mb_alloc/
2001-06-22 06:35:32 +00:00
{ "_mb_statpcpu" },
#define N_RTTRASH 30
{ "_rttrash" },
#define N_MBLO 31
{ "_mbuf_lowm" },
#define N_CLLO 32
{ "_clust_lowm" },
{ "" },
1994-05-27 12:33:43 +00:00
};
struct protox {
u_char pr_index; /* index into nlist of cb head */
u_char pr_sindex; /* index into nlist of stat block */
u_char pr_wanted; /* 1 if wanted, 0 otherwise */
void (*pr_cblocks)(u_long, const char *, int);
2001-06-15 23:35:13 +00:00
/* control blocks printing routine */
void (*pr_stats)(u_long, const char *, int);
2001-06-15 23:35:13 +00:00
/* statistics printing routine */
void (*pr_istats)(char *); /* per/if statistics printing routine */
2002-04-28 12:12:49 +00:00
const char *pr_name; /* well-known name */
u_long pr_usesysctl; /* true if we use sysctl, not kvm */
1994-05-27 12:33:43 +00:00
} protox[] = {
{ -1, -1, 1, protopr,
tcp_stats, NULL, "tcp", IPPROTO_TCP },
{ -1, -1, 1, protopr,
udp_stats, NULL, "udp", IPPROTO_UDP },
{ -1, -1, 1, protopr,
NULL, NULL, "divert",IPPROTO_DIVERT },
{ -1, -1, 1, protopr,
ip_stats, NULL, "ip", IPPROTO_RAW },
{ -1, -1, 1, protopr,
icmp_stats, NULL, "icmp", IPPROTO_ICMP },
{ -1, -1, 1, protopr,
igmp_stats, NULL, "igmp", IPPROTO_IGMP },
#ifdef IPSEC
{ -1, N_IPSECSTAT, 1, 0,
ipsec_stats, NULL, "ipsec", 0},
#endif
{ -1, -1, 1, 0,
bdg_stats, NULL, "bdg", 1 /* bridging... */ },
{ -1, -1, 1, protopr,
pim_stats, NULL, "pim", IPPROTO_PIM },
1994-05-27 12:33:43 +00:00
{ -1, -1, 0, 0,
0, NULL, 0, 0 }
1994-05-27 12:33:43 +00:00
};
#ifdef INET6
struct protox ip6protox[] = {
{ -1, -1, 1, protopr,
tcp_stats, NULL, "tcp", IPPROTO_TCP },
{ -1, -1, 1, protopr,
udp_stats, NULL, "udp", IPPROTO_UDP },
{ -1, N_IP6STAT, 1, protopr,
ip6_stats, ip6_ifstats, "ip6", IPPROTO_RAW },
{ -1, N_ICMP6STAT, 1, protopr,
icmp6_stats, icmp6_ifstats, "icmp6",IPPROTO_ICMPV6 },
#ifdef IPSEC
{ -1, N_IPSEC6STAT, 1, 0,
ipsec_stats, NULL, "ipsec6",0 },
#endif
#ifdef notyet
{ -1, N_PIM6STAT, 1, 0,
pim6_stats, NULL, "pim6", 0 },
#endif
{ -1, -1, 1, 0,
rip6_stats, NULL, "rip6", 0 },
{ -1, -1, 1, 0,
bdg_stats, NULL, "bdg", 1 /* bridging... */ },
{ -1, -1, 0, 0,
0, NULL, 0, 0 }
};
#endif /*INET6*/
2000-07-05 02:02:54 +00:00
#ifdef IPSEC
struct protox pfkeyprotox[] = {
{ -1, N_PFKEYSTAT, 1, 0,
pfkey_stats, NULL, "pfkey", 0 },
{ -1, -1, 0, 0,
0, NULL, 0, 0 }
};
#endif
struct protox atalkprotox[] = {
{ N_DDPCB, N_DDPSTAT, 1, atalkprotopr,
ddp_stats, NULL, "ddp", 0 },
{ -1, -1, 0, 0,
0, NULL, 0, 0 }
};
struct protox netgraphprotox[] = {
{ N_NGSOCKS, -1, 1, netgraphprotopr,
NULL, NULL, "ctrl", 0 },
{ N_NGSOCKS, -1, 1, netgraphprotopr,
NULL, NULL, "data", 0 },
{ -1, NULL, 0, 0,
0, NULL, 0, 0 }
};
struct protox ipxprotox[] = {
{ N_IPX, N_IPXSTAT, 1, ipxprotopr,
ipx_stats, NULL, "ipx", 0 },
{ N_IPX, N_SPXSTAT, 1, ipxprotopr,
spx_stats, NULL, "spx", 0 },
{ -1, -1, 0, 0,
0, NULL, 0, 0 }
};
struct protox *protoprotox[] = {
protox,
#ifdef INET6
ip6protox,
2000-07-05 02:02:54 +00:00
#endif
#ifdef IPSEC
pfkeyprotox,
#endif
ipxprotox, atalkprotox, NULL };
1994-05-27 12:33:43 +00:00
2002-04-28 12:12:49 +00:00
static void printproto (struct protox *, const char *);
2001-06-15 23:35:13 +00:00
static void usage (void);
static struct protox *name2protox (char *);
static struct protox *knownname (char *);
1994-05-27 12:33:43 +00:00
1998-08-08 08:13:04 +00:00
static kvm_t *kvmd;
static char *nlistf = NULL, *memf = NULL;
int Aflag; /* show addresses of protocol control block */
int aflag; /* show all sockets (including servers) */
int bflag; /* show i/f total bytes in/out */
int cflag; /* show mbuf cache information */
int dflag; /* show i/f dropped packets */
int gflag; /* show group (multicast) routing or stats */
int iflag; /* show interfaces */
int Lflag; /* show size of listen queues */
int mflag; /* show memory stats */
int numeric_addr; /* show addresses numerically */
int numeric_port; /* show ports numerically */
static int pflag; /* show given protocol */
int rflag; /* show routing tables (or routing stats) */
int sflag; /* show protocol statistics */
int tflag; /* show i/f watchdog timers */
int Wflag; /* wide display */
int zflag; /* zero stats */
int interval; /* repeat interval for i/f stats */
char *interface; /* desired i/f for stats, or NULL for all i/fs */
int unit; /* unit number for above */
int af; /* address family */
1994-05-27 12:33:43 +00:00
int
main(int argc, char *argv[])
1994-05-27 12:33:43 +00:00
{
struct protox *tp = NULL; /* for printing cblocks & stats */
1994-05-27 12:33:43 +00:00
int ch;
af = AF_UNSPEC;
while ((ch = getopt(argc, argv, "Aabcdf:gI:iLlM:mN:np:rSstuWw:z")) != -1)
1994-05-27 12:33:43 +00:00
switch(ch) {
case 'A':
Aflag = 1;
break;
case 'a':
aflag = 1;
break;
case 'b':
bflag = 1;
break;
case 'c':
cflag = 1;
break;
1994-05-27 12:33:43 +00:00
case 'd':
dflag = 1;
break;
case 'f':
1996-02-13 17:30:37 +00:00
if (strcmp(optarg, "ipx") == 0)
af = AF_IPX;
1994-05-27 12:33:43 +00:00
else if (strcmp(optarg, "inet") == 0)
af = AF_INET;
#ifdef INET6
else if (strcmp(optarg, "inet6") == 0)
af = AF_INET6;
2000-07-05 02:02:54 +00:00
#endif /*INET6*/
#ifdef INET6
else if (strcmp(optarg, "pfkey") == 0)
af = PF_KEY;
#endif /*INET6*/
1994-05-27 12:33:43 +00:00
else if (strcmp(optarg, "unix") == 0)
af = AF_UNIX;
else if (strcmp(optarg, "atalk") == 0)
af = AF_APPLETALK;
else if (strcmp(optarg, "ng") == 0
|| strcmp(optarg, "netgraph") == 0)
af = AF_NETGRAPH;
else if (strcmp(optarg, "link") == 0)
af = AF_LINK;
1994-05-27 12:33:43 +00:00
else {
errx(1, "%s: unknown address family", optarg);
1994-05-27 12:33:43 +00:00
}
break;
case 'g':
gflag = 1;
break;
case 'I': {
char *cp;
iflag = 1;
for (cp = interface = optarg; isalpha(*cp); cp++)
continue;
unit = atoi(cp);
break;
}
case 'i':
iflag = 1;
break;
case 'L':
Lflag = 1;
break;
1994-05-27 12:33:43 +00:00
case 'M':
memf = optarg;
break;
case 'm':
mflag = 1;
break;
case 'N':
nlistf = optarg;
break;
case 'n':
numeric_addr = numeric_port = 1;
1994-05-27 12:33:43 +00:00
break;
case 'p':
if ((tp = name2protox(optarg)) == NULL) {
errx(1,
"%s: unknown or uninstrumented protocol",
optarg);
1994-05-27 12:33:43 +00:00
}
pflag = 1;
break;
case 'r':
rflag = 1;
break;
case 's':
++sflag;
break;
case 'S':
numeric_addr = 1;
break;
1994-05-27 12:33:43 +00:00
case 't':
tflag = 1;
break;
case 'u':
af = AF_UNIX;
break;
case 'W':
case 'l':
Wflag = 1;
break;
1994-05-27 12:33:43 +00:00
case 'w':
interval = atoi(optarg);
iflag = 1;
break;
case 'z':
zflag = 1;
break;
1994-05-27 12:33:43 +00:00
case '?':
default:
usage();
}
argv += optind;
argc -= optind;
#define BACKWARD_COMPATIBILITY
#ifdef BACKWARD_COMPATIBILITY
if (*argv) {
if (isdigit(**argv)) {
interval = atoi(*argv);
if (interval <= 0)
usage();
++argv;
iflag = 1;
}
if (*argv) {
nlistf = *argv;
if (*++argv)
memf = *argv;
}
}
#endif
/*
* Discard setgid privileges if not the running kernel so that bad
* guys can't print interesting stuff from kernel memory.
*/
if (nlistf != NULL || memf != NULL)
setgid(getgid());
if (cflag && !mflag) {
(void)fprintf(stderr, "-c only valid with -m\n");
usage();
}
1994-05-27 12:33:43 +00:00
if (mflag) {
if (memf != NULL) {
if (kread(0, 0, 0) == 0)
mbpr(nl[N_MBSTAT].n_value,
nl[N_MBTYPES].n_value,
nl[N_NMBCLUSTERS].n_value,
Introduce numerous SMP friendly changes to the mbuf allocator. Namely, introduce a modified allocation mechanism for mbufs and mbuf clusters; one which can scale under SMP and which offers the possibility of resource reclamation to be implemented in the future. Notable advantages: o Reduce contention for SMP by offering per-CPU pools and locks. o Better use of data cache due to per-CPU pools. o Much less code cache pollution due to excessively large allocation macros. o Framework for `grouping' objects from same page together so as to be able to possibly free wired-down pages back to the system if they are no longer needed by the network stacks. Additional things changed with this addition: - Moved some mbuf specific declarations and initializations from sys/conf/param.c into mbuf-specific code where they belong. - m_getclr() has been renamed to m_get_clrd() because the old name is really confusing. m_getclr() HAS been preserved though and is defined to the new name. No tree sweep has been done "to change the interface," as the old name will continue to be supported and is not depracated. The change was merely done because m_getclr() sounds too much like "m_get a cluster." - TEMPORARILY disabled mbtypes statistics displaying in netstat(1) and systat(1) (see TODO below). - Fixed systat(1) to display number of "free mbufs" based on new per-CPU stat structures. - Fixed netstat(1) to display new per-CPU stats based on sysctl-exported per-CPU stat structures. All infos are fetched via sysctl. TODO (in order of priority): - Re-enable mbtypes statistics in both netstat(1) and systat(1) after introducing an SMP friendly way to collect the mbtypes stats under the already introduced per-CPU locks (i.e. hopefully don't use atomic() - it seems too costly for a mere stat update, especially when other locks are already present). - Optionally have systat(1) display not only "total free mbufs" but also "total free mbufs per CPU pool." - Fix minor length-fetching issues in netstat(1) related to recently re-enabled option to read mbuf stats from a core file. - Move reference counters at least for mbuf clusters into an unused portion of the cluster itself, to save space and need to allocate a counter. - Look into introducing resource freeing possibly from a kproc. Reviewed by (in parts): jlemon, jake, silby, terry Tested by: jlemon (Intel & Alpha), mjacob (Intel & Alpha) Preliminary performance measurements: jlemon (and me, obviously) URL: http://people.freebsd.org/~bmilekic/mb_alloc/
2001-06-22 06:35:32 +00:00
nl[N_NMBUFS].n_value,
nl[N_MBHI].n_value,
nl[N_CLHI].n_value,
nl[N_MBLO].n_value,
nl[N_CLLO].n_value,
Introduce numerous SMP friendly changes to the mbuf allocator. Namely, introduce a modified allocation mechanism for mbufs and mbuf clusters; one which can scale under SMP and which offers the possibility of resource reclamation to be implemented in the future. Notable advantages: o Reduce contention for SMP by offering per-CPU pools and locks. o Better use of data cache due to per-CPU pools. o Much less code cache pollution due to excessively large allocation macros. o Framework for `grouping' objects from same page together so as to be able to possibly free wired-down pages back to the system if they are no longer needed by the network stacks. Additional things changed with this addition: - Moved some mbuf specific declarations and initializations from sys/conf/param.c into mbuf-specific code where they belong. - m_getclr() has been renamed to m_get_clrd() because the old name is really confusing. m_getclr() HAS been preserved though and is defined to the new name. No tree sweep has been done "to change the interface," as the old name will continue to be supported and is not depracated. The change was merely done because m_getclr() sounds too much like "m_get a cluster." - TEMPORARILY disabled mbtypes statistics displaying in netstat(1) and systat(1) (see TODO below). - Fixed systat(1) to display number of "free mbufs" based on new per-CPU stat structures. - Fixed netstat(1) to display new per-CPU stats based on sysctl-exported per-CPU stat structures. All infos are fetched via sysctl. TODO (in order of priority): - Re-enable mbtypes statistics in both netstat(1) and systat(1) after introducing an SMP friendly way to collect the mbtypes stats under the already introduced per-CPU locks (i.e. hopefully don't use atomic() - it seems too costly for a mere stat update, especially when other locks are already present). - Optionally have systat(1) display not only "total free mbufs" but also "total free mbufs per CPU pool." - Fix minor length-fetching issues in netstat(1) related to recently re-enabled option to read mbuf stats from a core file. - Move reference counters at least for mbuf clusters into an unused portion of the cluster itself, to save space and need to allocate a counter. - Look into introducing resource freeing possibly from a kproc. Reviewed by (in parts): jlemon, jake, silby, terry Tested by: jlemon (Intel & Alpha), mjacob (Intel & Alpha) Preliminary performance measurements: jlemon (and me, obviously) URL: http://people.freebsd.org/~bmilekic/mb_alloc/
2001-06-22 06:35:32 +00:00
nl[N_NCPUS].n_value,
nl[N_PAGESZ].n_value,
nl[N_MBPSTAT].n_value);
} else
mbpr(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0);
1994-05-27 12:33:43 +00:00
exit(0);
}
#if 0
1994-05-27 12:33:43 +00:00
/*
* Keep file descriptors open to avoid overhead
* of open/close on each call to get* routines.
*/
sethostent(1);
setnetent(1);
#else
/*
* This does not make sense any more with DNS being default over
* the files. Doing a setXXXXent(1) causes a tcp connection to be
* used for the queries, which is slower.
*/
#endif
if (iflag && !sflag) {
1998-08-08 08:13:04 +00:00
kread(0, 0, 0);
intpr(interval, nl[N_IFNET].n_value, NULL);
1994-05-27 12:33:43 +00:00
exit(0);
}
if (rflag) {
1998-08-08 08:13:04 +00:00
kread(0, 0, 0);
1994-05-27 12:33:43 +00:00
if (sflag)
rt_stats(nl[N_RTSTAT].n_value, nl[N_RTTRASH].n_value);
1994-05-27 12:33:43 +00:00
else
routepr(nl[N_RTREE].n_value);
exit(0);
}
if (gflag) {
1998-08-08 08:13:04 +00:00
kread(0, 0, 0);
if (sflag) {
if (af == AF_INET || af == AF_UNSPEC)
mrt_stats(nl[N_MRTSTAT].n_value);
#ifdef INET6
if (af == AF_INET6 || af == AF_UNSPEC)
mrt6_stats(nl[N_MRT6STAT].n_value);
#endif
} else {
if (af == AF_INET || af == AF_UNSPEC)
mroutepr(nl[N_MFCTABLE].n_value,
nl[N_VIFTABLE].n_value);
#ifdef INET6
if (af == AF_INET6 || af == AF_UNSPEC)
mroute6pr(nl[N_MF6CTABLE].n_value,
nl[N_MIF6TABLE].n_value);
#endif
1994-05-27 12:33:43 +00:00
}
exit(0);
1994-05-27 12:33:43 +00:00
}
kread(0, 0, 0);
if (tp) {
printproto(tp, tp->pr_name);
exit(0);
}
if (af == AF_INET || af == AF_UNSPEC)
for (tp = protox; tp->pr_name; tp++)
printproto(tp, tp->pr_name);
#ifdef INET6
if (af == AF_INET6 || af == AF_UNSPEC)
for (tp = ip6protox; tp->pr_name; tp++)
printproto(tp, tp->pr_name);
#endif /*INET6*/
2000-07-05 02:02:54 +00:00
#ifdef IPSEC
if (af == PF_KEY || af == AF_UNSPEC)
for (tp = pfkeyprotox; tp->pr_name; tp++)
printproto(tp, tp->pr_name);
#endif /*IPSEC*/
if (af == AF_IPX || af == AF_UNSPEC) {
kread(0, 0, 0);
for (tp = ipxprotox; tp->pr_name; tp++)
printproto(tp, tp->pr_name);
}
if (af == AF_APPLETALK || af == AF_UNSPEC)
for (tp = atalkprotox; tp->pr_name; tp++)
printproto(tp, tp->pr_name);
if (af == AF_NETGRAPH || af == AF_UNSPEC)
for (tp = netgraphprotox; tp->pr_name; tp++)
printproto(tp, tp->pr_name);
if ((af == AF_UNIX || af == AF_UNSPEC) && !Lflag && !sflag)
unixpr();
1994-05-27 12:33:43 +00:00
exit(0);
}
/*
* Print out protocol statistics or control blocks (per sflag).
* If the interface was not specifically requested, and the symbol
* is not in the namelist, ignore this one.
*/
static void
printproto(tp, name)
struct protox *tp;
2002-04-28 12:12:49 +00:00
const char *name;
1994-05-27 12:33:43 +00:00
{
void (*pr)(u_long, const char *, int);
1994-05-27 12:33:43 +00:00
u_long off;
if (sflag) {
if (iflag) {
if (tp->pr_istats)
intpr(interval, nl[N_IFNET].n_value,
tp->pr_istats);
else if (pflag)
printf("%s: no per-interface stats routine\n",
tp->pr_name);
return;
}
else {
pr = tp->pr_stats;
if (!pr) {
if (pflag)
printf("%s: no stats routine\n",
tp->pr_name);
return;
}
off = tp->pr_usesysctl ? tp->pr_usesysctl
: nl[tp->pr_sindex].n_value;
}
1994-05-27 12:33:43 +00:00
} else {
pr = tp->pr_cblocks;
if (!pr) {
if (pflag)
printf("%s: no PCB routine\n", tp->pr_name);
return;
}
off = tp->pr_usesysctl ? tp->pr_usesysctl
: nl[tp->pr_index].n_value;
1994-05-27 12:33:43 +00:00
}
if (pr != NULL && (off || af != AF_UNSPEC))
(*pr)(off, name, af);
1994-05-27 12:33:43 +00:00
}
/*
* Read kernel memory, return 0 on success.
*/
int
2001-06-15 23:35:13 +00:00
kread(u_long addr, char *buf, int size)
1994-05-27 12:33:43 +00:00
{
if (kvmd == 0) {
/*
* XXX.
*/
kvmd = kvm_openfiles(nlistf, memf, NULL, O_RDONLY, buf);
setgid(getgid());
if (kvmd != NULL) {
if (kvm_nlist(kvmd, nl) < 0) {
if(nlistf)
errx(1, "%s: kvm_nlist: %s", nlistf,
kvm_geterr(kvmd));
else
errx(1, "kvm_nlist: %s", kvm_geterr(kvmd));
}
if (nl[0].n_type == 0) {
if(nlistf)
errx(1, "%s: no namelist", nlistf);
else
errx(1, "no namelist");
}
} else {
warnx("kvm not available");
return(-1);
}
}
1998-08-08 08:13:04 +00:00
if (!buf)
return (0);
1994-05-27 12:33:43 +00:00
if (kvm_read(kvmd, addr, buf, size) != size) {
warnx("%s", kvm_geterr(kvmd));
1994-05-27 12:33:43 +00:00
return (-1);
}
return (0);
}
const char *
2001-06-15 23:35:13 +00:00
plural(int n)
1994-05-27 12:33:43 +00:00
{
return (n != 1 ? "s" : "");
}
const char *
2001-06-15 23:35:13 +00:00
plurales(int n)
1994-05-27 12:33:43 +00:00
{
return (n != 1 ? "es" : "");
}
const char *
pluralies(int n)
{
return (n != 1 ? "ies" : "y");
}
1994-05-27 12:33:43 +00:00
/*
* Find the protox for the given "well-known" name.
*/
static struct protox *
2001-06-15 23:35:13 +00:00
knownname(char *name)
1994-05-27 12:33:43 +00:00
{
struct protox **tpp, *tp;
for (tpp = protoprotox; *tpp; tpp++)
for (tp = *tpp; tp->pr_name; tp++)
if (strcmp(tp->pr_name, name) == 0)
return (tp);
return (NULL);
}
/*
* Find the protox corresponding to name.
*/
static struct protox *
2001-06-15 23:35:13 +00:00
name2protox(char *name)
1994-05-27 12:33:43 +00:00
{
struct protox *tp;
char **alias; /* alias from p->aliases */
struct protoent *p;
/*
* Try to find the name in the list of "well-known" names. If that
* fails, check if name is an alias for an Internet protocol.
*/
if ((tp = knownname(name)) != NULL)
1994-05-27 12:33:43 +00:00
return (tp);
setprotoent(1); /* make protocol lookup cheaper */
while ((p = getprotoent()) != NULL) {
1994-05-27 12:33:43 +00:00
/* assert: name not same as p->name */
for (alias = p->p_aliases; *alias; alias++)
if (strcmp(name, *alias) == 0) {
endprotoent();
return (knownname(p->p_name));
}
}
endprotoent();
return (NULL);
}
static void
2001-06-15 23:35:13 +00:00
usage(void)
1994-05-27 12:33:43 +00:00
{
2001-09-11 15:21:36 +00:00
(void)fprintf(stderr, "%s\n%s\n%s\n%s\n%s\n%s\n%s\n%s\n%s\n%s\n",
"usage: netstat [-AaLnSW] [-f protocol_family | -p protocol]\n"
" [-M core] [-N system]",
" netstat -i | -I interface [-abdnt] [-f address_family]\n"
" [-M core] [-N system]",
2001-09-11 15:21:36 +00:00
" netstat -w wait [-I interface] [-d] [-M core] [-N system]",
" netstat -s [-s] [-z] [-f protocol_family | -p protocol] [-M core]",
" netstat -i | -I interface -s [-f protocol_family | -p protocol]\n"
" [-M core] [-N system]",
" netstat -m [-c] [-M core] [-N system]",
" netstat -r [-AenW] [-f address_family] [-M core] [-N system]",
2001-09-11 15:21:36 +00:00
" netstat -rs [-s] [-M core] [-N system]",
" netstat -g [-W] [-f address_family] [-M core] [-N system]",
" netstat -gs [-s] [-f address_family] [-M core] [-N system]");
1994-05-27 12:33:43 +00:00
exit(1);
}