1
0
mirror of https://git.FreeBSD.org/ports.git synced 2024-11-15 23:50:44 +00:00
freebsd-ports/sysutils/xperfmon3/files/freebsd_system.c

919 lines
28 KiB
C
Raw Normal View History

/*
* freebsd_system.c f<EFBFBD>r FreeBSD-3.X
* COPYRIGHT 1998 by Lars K<EFBFBD>ller <Lars.Koeller@Uni-Bielefeld.de>
*
* Work has started on 7th Sep 1998 on Northsea island F<EFBFBD>hr.
*
1999-08-31 02:00:46 +00:00
* $FreeBSD$
*/
/*
* Copyright (c) 1998 Lars K<EFBFBD>ller
* 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. The name of the author may not be used to endorse or promote products
* derived from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR 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 AUTHOR 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.
*
*/
/*
*
* Parts of this program are derived from the iostat program:
*
* Copyright (c) 1997, 1998 Kenneth D. Merry.
* 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. The name of the author may not be used to endorse or promote products
* derived from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR 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 AUTHOR 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.
*
*/
/*
* Parts of this program are derived from other original FreeBSD programs:
*
* Copyright (c) 1986, 1991, 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.
* 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
1999-08-31 02:00:46 +00:00
static char rcsid[] = "$FreeBSD$";
#endif
#include "fbsd_vers.h"
#include <X11/IntrinsicP.h>
#include "system.h"
#include "is.h"
#include <sys/file.h>
#include <sys/param.h>
#include <sys/types.h>
#include <sys/socket.h>
#include <sys/sysctl.h>
#include <sys/dkstat.h>
#include <sys/buf.h>
#include <sys/time.h>
#include <vm/vm.h>
#include <net/if.h>
#include <net/if_var.h>
#include <netinet/in.h>
#include <sys/stat.h>
#include <sys/conf.h>
#if __FreeBSD_version < 400000
#include <sys/rlist.h>
#endif
#include <sys/mount.h>
#include <nfs/rpcv2.h>
#include <nfs/nfsproto.h>
#include <nfs/nfs.h>
#include <stdio.h>
#include <stdlib.h>
#include <strings.h>
#include <unistd.h>
#include <ctype.h>
#include <paths.h>
#include <kvm.h>
#include <nlist.h>
#include <limits.h>
#include <errno.h>
#include <err.h>
#include <devstat.h>
struct nlist namelist[] = {
#define X_HZ 0
{ "_hz" },
#define X_STATHZ 1
{ "_stathz" },
#define X_CP_TIME 2
{ "_cp_time" },
#define X_TK_NIN 3
{ "_tk_nin" },
#define X_TK_NOUT 4
{ "_tk_nout" },
#define N_IFNET 5
{ "_ifnet" },
#define X_INTRCNT 6
{ "_intrcnt" },
#define X_EINTRCNT 7
{ "_eintrcnt" },
#define X_BOOTTIME 8
{ "_boottime" },
#if __FreeBSD_version < 400000
#define VM_NSWAP 9
{ "_nswap" },
#define VM_NSWDEV 10
{ "_nswdev" },
#define VM_DMMAX 11
{ "_dmmax" },
#define VM_SWAPLIST 12
{ "_swaplist" },
#define VM_SWDEVT 13
{ "_swdevt" },
#define X_END 13
{ NULL },
#else
#define X_END 8
{ NULL },
#endif /* __FreeBSD_version < 400000 */
};
extern Widget perfmon[NUM_GRAPHS];
extern char *info1;
double current_values[NUM_GRAPHS];
static kvm_t *kd;
static char errbuf[_POSIX2_LINE_MAX];
static int has_nfs;
static int hz, stathz;
static double etime, pct;
static long tmp;
struct vfsconf vfc;
/*
* Variables holding the statistic information
*/
static struct statinfo cur, last;
static struct nfsstats nfsstats;
static struct _nfsStats {
int ServerCalls, ServerCacheHit, ServerCacheMis;
int ClientCalls, ClientCacheHit, ClientCacheMis;
} nfsStats, old_nfsStats;
static struct _packets {
int input, output, collisions;
} packets, old_packets;
static int interrupts, old_interrupts;
/*
* Variables need for devstat to get diskio statistics
*/
static struct statinfo cur, last;
static struct devstat_match *matches = NULL;
static struct device_selection *dev_select = NULL;
static devstat_select_mode select_mode;
static int num_matches = 0, num_devices = 0;
static int num_selected = 0, num_selections = 0;
static int num_devices_specified = 0, maxshowdevs;
static long generation, select_generation;
static char **specified_devices = NULL;
static struct _diskstat {
long double da_trsf, da_mb;
long double cd_trsf, cd_mb;
long double sa_trsf, sa_mb;
} diskstat;
void sys_setup(void);
void update_stats(void);
static void get_namelist(const char *kernel_name, const char *memory_name);
static void kread(int nlx, void *addr, size_t size);
static double get_load(void);
static void get_cpustat(void);
static void get_ttystat(void);
static void get_interrupts(void);
static void init_diskio(void);
static void get_diskio(void);
static double get_swapspace(void);
static void get_netstat(void);
static void get_nfsstat(void);
void
sys_setup(void)
{
get_namelist(getbootfile(), _PATH_MEM);
get_cpustat();
get_load();
get_ttystat();
get_interrupts();
init_diskio();
get_diskio();
/*
* To force first scale of 'free swapspace' to 100%, we need this hack!
* Wee use a value < 0 cause it's not a valid value!
*/
current_values[FREE_MEM] = -0.4 * SCALE_HACK;
get_netstat();
/*
* Check if we have NFS in the kernel
*/
if (getvfsbyname("nfs", &vfc) < 0) {
fprintf(stderr, "xperfmon++: getvfsbyname: NFS not compiled into kernel\n");
has_nfs = FALSE;
} else {
has_nfs = TRUE;
get_nfsstat();
}
}
void
update_stats(void)
{
static int firsttime = 1;
/*
* For any stat we need etime, so get it!
*/
get_cpustat();
current_values[ USER_CPU_PERCENTAGE] = cur.cp_time[CP_USER] * pct * SCALE_HACK;
current_values[ NICE_CPU_PERCENTAGE] = cur.cp_time[CP_NICE] * pct * SCALE_HACK;
current_values[SYSTEM_CPU_PERCENTAGE] = cur.cp_time[CP_SYS] * pct * SCALE_HACK;
current_values[ INTER_CPU_PERCENTAGE] = cur.cp_time[CP_INTR] * pct * SCALE_HACK;
current_values[ IDLE_CPU_PERCENTAGE] = cur.cp_time[CP_IDLE] * pct * SCALE_HACK;
if (perfmon[LOAD])
current_values[LOAD] = get_load() * SCALE_HACK;
if (perfmon[TTY_CHARS_IN] || perfmon[TTY_CHARS_OUT]) {
get_ttystat();
current_values[TTY_IN] = cur.tk_nin/etime * SCALE_HACK;
current_values[TTY_OUT] = cur.tk_nout/etime * SCALE_HACK;
}
if (perfmon[INTERRUPTS]) {
get_interrupts();
current_values[INTERRUPTS] = (interrupts - old_interrupts)/etime * SCALE_HACK;
}
/*
* The first time called, we want to get 100%
* for the full scale of the graph!
*/
if (perfmon[FREE_MEM] && !firsttime)
current_values[FREE_MEM] =
(current_values[FREE_MEM] < 0 ? 100.0 * SCALE_HACK : get_swapspace() * SCALE_HACK);
else
--firsttime;
if (perfmon[DISK_TRANSFERS] || perfmon[DISK_MB] ||
perfmon[TAPE_TRANSFERS] || perfmon[TAPE_MB] ||
perfmon[CDROM_TRANSFERS] || perfmon[CDROM_MB] ) {
get_diskio();
current_values[DISK_TRANSFERS] = diskstat.da_trsf * SCALE_HACK;
current_values[DISK_MB] = diskstat.da_mb * SCALE_HACK;
current_values[TAPE_TRANSFERS] = diskstat.sa_trsf * SCALE_HACK;
current_values[TAPE_MB] = diskstat.sa_mb * SCALE_HACK;
current_values[CDROM_TRANSFERS] = diskstat.cd_trsf * SCALE_HACK;
current_values[CDROM_MB] = diskstat.cd_mb * SCALE_HACK;
}
if (perfmon[INPUT_PACKETS] || perfmon[OUTPUT_PACKETS] ||
perfmon[COLLISION_PACKETS]){
get_netstat();
current_values[ INPUT_PACKETS] =
(packets.input - old_packets.input)/etime * SCALE_HACK;
current_values[ OUTPUT_PACKETS] =
(packets.output - old_packets.output)/etime * SCALE_HACK;
current_values[COLLISION_PACKETS] =
(packets.collisions - old_packets.collisions)/etime * SCALE_HACK;
}
if (has_nfs && (perfmon[NFS_CLIENT_CALLS] ||
perfmon[NFS_CLIENT_HIT] ||
perfmon[NFS_CLIENT_MIS] ||
perfmon[NFS_SERVER_CALLS] ||
perfmon[NFS_SERVER_HIT] ||
perfmon[NFS_SERVER_MIS])) {
get_nfsstat();
current_values[NFS_CLIENT_CALLS] =
(nfsStats.ClientCalls - old_nfsStats.ClientCalls)/etime * SCALE_HACK;
current_values[NFS_SERVER_CALLS] =
(nfsStats.ServerCalls - old_nfsStats.ServerCalls)/etime * SCALE_HACK;
current_values[NFS_CLIENT_HIT] =
(nfsStats.ClientCacheHit - old_nfsStats.ClientCacheHit)/etime * SCALE_HACK;
current_values[NFS_CLIENT_MIS] =
(nfsStats.ClientCacheMis - old_nfsStats.ClientCacheMis)/etime * SCALE_HACK;
current_values[NFS_SERVER_HIT] =
(nfsStats.ServerCacheHit - old_nfsStats.ServerCacheHit)/etime * SCALE_HACK;
current_values[NFS_SERVER_MIS] =
(nfsStats.ServerCacheMis - old_nfsStats.ServerCacheMis)/etime * SCALE_HACK;
}
}
static void
get_namelist(kernel_name, memory_name)
const char *kernel_name, *memory_name;
{
register int i, ret;
time_t now, boottime;
int nintv;
kd = kvm_openfiles(kernel_name, memory_name, NULL, O_RDONLY, errbuf);
if (kd == 0)
errx(1, "xperfmon++: kvm_openfiles: %s", errbuf);
if ((ret = kvm_nlist(kd, namelist)) != 0) {
if (ret > 0) {
(void)fprintf(stderr, "xperfmon++: undefined symbols:");
for (i = 0; i < sizeof(namelist)/sizeof(namelist[0]); i++)
if (namelist[i].n_type == 0 || namelist[i].n_value == 0)
fprintf(stderr, "%s", namelist[i].n_name);
fprintf(stderr, "\n");
} else
errx(1, "xperfmon++: kvm_nlist: %s", kvm_geterr(kd));
exit(1);
}
kread(X_HZ, &hz, sizeof(hz));
kread(X_STATHZ, &stathz, sizeof(stathz));
if (stathz)
hz = stathz;
kread(X_BOOTTIME, &boottime, sizeof(boottime));
time(&now);
nintv = now - boottime;
if (nintv <= 0 || nintv > 60*60*24*365*10)
errx(1, "Time makes no sense ... namelist must be wrong", NULL);
}
static void
kread(nlx, addr, size)
int nlx; void *addr; size_t size;
{
char *sym;
if (kvm_read(kd, namelist[nlx].n_value, addr, size) != size) {
sym = namelist[nlx].n_name;
if (*sym == '_')
++sym;
errx(1, "xperfmon++: %s: %s", sym, kvm_geterr(kd));
}
}
static void
get_cpustat(void)
{
register int i;
double tot;
kread(X_CP_TIME, cur.cp_time, sizeof(cur.cp_time));
etime = 0.0;
for (i = 0; i < CPUSTATES; i++) {
tmp = cur.cp_time[i];
cur.cp_time[i] -= last.cp_time[i];
last.cp_time[i] = tmp;
etime += cur.cp_time[i];
}
tot = etime;
if (etime == 0.0 )
etime = 1.0;
etime /= (float)hz;
/*
* scale to percent
*/
if(tot)
pct = 100.0 / tot;
else
pct = 0;
}
static double
get_load(void)
{
double loadavg[3];
loadavg[0] = 0.0;
if (getloadavg(loadavg, sizeof(loadavg)/sizeof(loadavg[0])) == -1) {
fprintf(stderr, "xperfmon++: getloadavg() returned no values\n");
return;
}
return(loadavg[0]);
}
static void
get_ttystat(void)
{
kread(X_TK_NIN, &cur.tk_nin, sizeof(cur.tk_nin));
kread(X_TK_NOUT, &cur.tk_nout, sizeof(cur.tk_nout));
tmp = cur.tk_nin;
cur.tk_nin -= last.tk_nin;
last.tk_nin = tmp;
tmp = cur.tk_nout;
cur.tk_nout -= last.tk_nout;
last.tk_nout = tmp;
}
static void
get_interrupts(void)
{
register int i;
unsigned long *intrcnt;
int nintr;
nintr = namelist[X_EINTRCNT].n_value - namelist[X_INTRCNT].n_value;
if ((intrcnt = (unsigned long *)malloc((size_t) nintr)) == NULL)
err(1, "xperfmon++: malloc failed in get_interrupts()");
nintr /= sizeof(long);
kread(X_INTRCNT, intrcnt, (size_t)nintr*sizeof(long));
old_interrupts = interrupts;
for (i = 0, interrupts = 0; i < nintr; i++)
interrupts += *(intrcnt + i);
free(intrcnt);
}
/*
* Parts extracted from /usr/src/usr.bin/nfsstat/nfsstat.c
* pay attention Copyrights!
*/
static void
get_nfsstat(void)
{
#if __FreeBSD_version < 400001
int name[3];
size_t size = sizeof(nfsstats);
name[0] = CTL_VFS;
name[1] = vfc.vfc_typenum;
name[2] = NFS_NFSSTATS;
if (sysctl(name, 3, &nfsstats, &size, (void *)0, (size_t)0) < 0) {
#else
size_t size = sizeof(nfsstats);
if (sysctlbyname("vfs.nfs.nfsstats", &nfsstats, &size, (void *)0, (size_t)0) < 0) {
#endif
fprintf(stderr, "xperfmon++: get_nfsstat(): Can?%t get NFS statistics with sysctl()\n");
return;
}
old_nfsStats = nfsStats;
nfsStats.ClientCalls = nfsstats.rpccnt[NFSPROC_GETATTR] +
nfsstats.rpccnt[NFSPROC_SETATTR] +
nfsstats.rpccnt[NFSPROC_LOOKUP] +
nfsstats.rpccnt[NFSPROC_READLINK] +
nfsstats.rpccnt[NFSPROC_READ] +
nfsstats.rpccnt[NFSPROC_WRITE] +
nfsstats.rpccnt[NFSPROC_CREATE] +
nfsstats.rpccnt[NFSPROC_REMOVE] +
nfsstats.rpccnt[NFSPROC_RENAME] +
nfsstats.rpccnt[NFSPROC_LINK] +
nfsstats.rpccnt[NFSPROC_SYMLINK] +
nfsstats.rpccnt[NFSPROC_MKDIR] +
nfsstats.rpccnt[NFSPROC_RMDIR] +
nfsstats.rpccnt[NFSPROC_READDIR] +
nfsstats.rpccnt[NFSPROC_READDIRPLUS]+
nfsstats.rpccnt[NFSPROC_ACCESS] +
nfsstats.rpccnt[NFSPROC_MKNOD] +
nfsstats.rpccnt[NFSPROC_FSSTAT] +
nfsstats.rpccnt[NFSPROC_FSINFO] +
nfsstats.rpccnt[NFSPROC_PATHCONF] +
nfsstats.rpccnt[NFSPROC_COMMIT] +
nfsstats.rpccnt[NQNFSPROC_GETLEASE] +
nfsstats.rpccnt[NQNFSPROC_VACATED] +
nfsstats.rpccnt[NQNFSPROC_EVICTED];
nfsStats.ClientCacheHit = nfsstats.attrcache_hits +
nfsstats.lookupcache_hits +
nfsstats.biocache_reads -
nfsstats.read_bios +
nfsstats.biocache_writes -
nfsstats.write_bios +
nfsstats.biocache_readlinks -
nfsstats.readlink_bios +
nfsstats.biocache_readdirs -
nfsstats.readdir_bios +
nfsstats.direofcache_hits;
nfsStats.ClientCacheMis = nfsstats.attrcache_misses +
nfsstats.lookupcache_misses +
nfsstats.read_bios +
nfsstats.write_bios +
nfsstats.readlink_bios +
nfsstats.readdir_bios +
nfsstats.direofcache_misses;
nfsStats.ServerCalls = nfsstats.srvrpccnt[NFSPROC_GETATTR] +
nfsstats.srvrpccnt[NFSPROC_SETATTR] +
nfsstats.srvrpccnt[NFSPROC_LOOKUP] +
nfsstats.srvrpccnt[NFSPROC_READLINK] +
nfsstats.srvrpccnt[NFSPROC_READ] +
nfsstats.srvrpccnt[NFSPROC_WRITE] +
nfsstats.srvrpccnt[NFSPROC_CREATE] +
nfsstats.srvrpccnt[NFSPROC_REMOVE] +
nfsstats.srvrpccnt[NFSPROC_RENAME] +
nfsstats.srvrpccnt[NFSPROC_LINK] +
nfsstats.srvrpccnt[NFSPROC_SYMLINK] +
nfsstats.srvrpccnt[NFSPROC_MKDIR] +
nfsstats.srvrpccnt[NFSPROC_RMDIR] +
nfsstats.srvrpccnt[NFSPROC_READDIR] +
nfsstats.srvrpccnt[NFSPROC_READDIRPLUS]+
nfsstats.srvrpccnt[NFSPROC_ACCESS] +
nfsstats.srvrpccnt[NFSPROC_MKNOD] +
nfsstats.srvrpccnt[NFSPROC_FSSTAT] +
nfsstats.srvrpccnt[NFSPROC_FSINFO] +
nfsstats.srvrpccnt[NFSPROC_PATHCONF] +
nfsstats.srvrpccnt[NFSPROC_COMMIT] +
nfsstats.srvrpccnt[NQNFSPROC_GETLEASE] +
nfsstats.srvrpccnt[NQNFSPROC_VACATED] +
nfsstats.srvrpccnt[NQNFSPROC_EVICTED];
nfsStats.ServerCacheHit = nfsstats.srvcache_inproghits +
nfsstats.srvcache_idemdonehits +
nfsstats.srvcache_nonidemdonehits;
nfsStats.ServerCacheMis = nfsstats.srvcache_misses;
}
/*
* Stolen from /usr/src/usr.sbin/pstat/pstat.c and
* /usr/src/usr.bin/top/machine.c (pay attention Copyrights!)
*/
static double
get_swapspace(void)
{
#if __FreeBSD_version >= 400000
/* based on swapmode from /usr/src/usr.bin/top/machine.c */
int n;
int percentfree;
struct kvm_swap swapary[1];
n = kvm_getswapinfo(kd, swapary, 1, 0);
if (n < 0)
return(0);
percentfree = ((((double)swapary[0].ksw_total -
(double)swapary[0].ksw_used) * 100.0) /
(double)swapary[0].ksw_total);
return(percentfree);
#else
/* based on /usr/src/usr.sbin/pstat/pstat.c */
char *header;
int hlen, nswap, nswdev, dmmax;
int i, div, avail, nfree, npfree, used;
struct swdevt *sw;
long blocksize, *perdev;
struct rlist head;
struct rlisthdr swaplist;
struct rlist *swapptr;
u_long ptr;
kread(VM_NSWAP, &nswap, sizeof(nswap));
kread(VM_NSWDEV, &nswdev, sizeof(nswdev));
kread(VM_DMMAX, &dmmax, sizeof(dmmax));
kread(VM_SWAPLIST, &swaplist, sizeof(swaplist));
if ((sw = malloc(nswdev * sizeof(*sw))) == NULL ||
(perdev = malloc(nswdev * sizeof(*perdev))) == NULL)
err(1, "xperfmon++ malloc in get_swapspace");
kread(VM_SWDEVT, &ptr, sizeof(ptr));
kvm_read(kd, ptr, sw, nswdev * sizeof(*sw));
/* Count up swap space. */
nfree = 0;
memset(perdev, 0, nswdev * sizeof(*perdev));
swapptr = swaplist.rlh_list;
while (swapptr) {
int top, bottom, next_block;
kvm_read(kd, (u_long)swapptr, &head, sizeof(struct rlist));
top = head.rl_end;
bottom = head.rl_start;
nfree += top - bottom + 1;
/*
* Swap space is split up among the configured disks.
*
* For interleaved swap devices, the first dmmax blocks
* of swap space some from the first disk, the next dmmax
* blocks from the next, and so on up to nswap blocks.
*
* The list of free space joins adjacent free blocks,
* ignoring device boundries. If we want to keep track
* of this information per device, we'll just have to
* extract it ourselves.
*/
while (top / dmmax != bottom / dmmax) {
next_block = ((bottom + dmmax) / dmmax);
perdev[(bottom / dmmax) % nswdev] +=
next_block * dmmax - bottom;
bottom = next_block * dmmax;
}
perdev[(bottom / dmmax) % nswdev] +=
top - bottom + 1;
swapptr = head.rl_next;
}
header = getbsize(&hlen, &blocksize);
div = blocksize / 512;
avail = npfree = 0;
for (i = 0; i < nswdev; i++) {
int xsize, xfree;
/*
* Don't report statistics for partitions which have not
* yet been activated via swapon(8).
*/
if (!(sw[i].sw_flags & SW_FREED))
continue;
/* The first dmmax is never allocated to avoid trashing of
* disklabels
*/
xsize = sw[i].sw_nblks - dmmax;
xfree = perdev[i];
used = xsize - xfree;
npfree++;
avail += xsize;
}
/*
* If only one partition has been set up via swapon(8), we don't
* need to bother with totals.
*/
used = avail - nfree;
free(sw);
free(perdev);
return((100.0*nfree)/avail);
#endif /* __FreeBSD_version >= 400000 */
}
/*
* Stolen from /usr/src/usr.bin/netstat/if.c
* pay attention Copyrights!
*/
static void
get_netstat(void)
{
off_t ifnetaddr;
if ((ifnetaddr = namelist[N_IFNET].n_value) != 0) {
struct ifnet ifnet;
struct ifnethead ifnethead;
u_long ifaddraddr, ifnetfound;
struct ifaddr ifa;
if(kvm_read(kd, ifnetaddr, (char *)&ifnethead, sizeof(ifnethead)) == -1)
return;
ifnetaddr = (u_long)ifnethead.tqh_first;
if(kvm_read(kd, ifnetaddr, (char *)&ifnet, sizeof(ifnet)) == -1)
return;
old_packets = packets;
packets.input = packets.output = packets.collisions = 0;
ifaddraddr = 0;
while (ifnetaddr || ifaddraddr) {
if (ifaddraddr == 0) {
ifnetfound = ifnetaddr;
if(kvm_read(kd, ifnetaddr, (char *)&ifnet, sizeof ifnet) == -1)
return;
ifnetaddr = (u_long)ifnet.if_link.tqe_next;
ifaddraddr = (u_long)ifnet.if_addrhead.tqh_first;
}
if (kvm_read(kd, ifaddraddr, (char *)&ifa, sizeof ifa) == -1) {
ifaddraddr = 0;
continue;
}
ifaddraddr = (u_long)ifa.ifa_link.tqe_next;
packets.input += ifnet.if_ipackets;
packets.output += ifnet.if_opackets;
packets.collisions += ifnet.if_collisions;
}
}
}
/*
* Many thanks to Kenneth Merry <ken@plutotech.com> who helped to code
* the initial disk statistic routines
*/
static void
init_diskio(void)
{
int retval, i;
cur.dinfo = (struct devinfo *) malloc(sizeof(struct devinfo));
last.dinfo = (struct devinfo *) malloc(sizeof(struct devinfo));
bzero(cur.dinfo, sizeof(struct devinfo));
bzero(last.dinfo, sizeof(struct devinfo));
if (checkversion() < 0)
errx(1, "xperfmon++: %s", devstat_errbuf);
/*
* man mu<EFBFBD> <EFBFBD>ber specified devices gehen, um da1,... zu
* selektieren, siehe /usr/src/usr.sbin/iostat/iostat.c
*/
/* printf("%s\n", info1); LK!!! */
if (buildmatch("da", &matches, &num_matches) != 0)
errx(1, "xperfmon++: %s", devstat_errbuf);
if (buildmatch("sa", &matches, &num_matches) != 0)
errx(1, "xperfmon++: %s", devstat_errbuf);
if (buildmatch("cd", &matches, &num_matches) != 0)
errx(1, "xperfmon++: %s", devstat_errbuf);
if (getdevs(&cur) == -1)
errx(1, "xperfmon++: %s", devstat_errbuf);
maxshowdevs = 50;
num_devices = cur.dinfo->numdevs;
generation = cur.dinfo->generation;
retval = selectdevs(&dev_select, &num_selected,
&num_selections, &select_generation,
generation, cur.dinfo->devices,
num_devices, matches, num_matches,
specified_devices, num_devices_specified,
DS_SELECT_ONLY, maxshowdevs, 0);
switch (retval) {
case -1:
errx(1, "xperfmon++: %s", devstat_errbuf);
break;
default:
break;
}
}
static void
get_diskio(void)
{
register int i, dn;
long double busy_seconds;
long double transfers_per_second, mb_per_second;
struct devinfo *tmp_dinfo;
int retval;
tmp_dinfo = last.dinfo;
last.dinfo = cur.dinfo;
cur.dinfo = tmp_dinfo;
last.busy_time = cur.busy_time;
switch (getdevs(&cur)) {
case -1:
errx(1, "xperfmon++: %s", devstat_errbuf);
break;
case 1: {
int retval;
num_devices = cur.dinfo->numdevs;
generation = cur.dinfo->generation;
retval = selectdevs(&dev_select, &num_selected,
&num_selections, &select_generation,
generation, cur.dinfo->devices,
num_devices, matches, num_matches,
specified_devices, num_devices_specified,
DS_SELECT_ONLY, maxshowdevs, 0);
switch (retval) {
case -1:
errx(1, "xperfmon++: %s", devstat_errbuf);
break;
default:
break;
}
break;
}
default:
break;
}
/*
* Calculate elapsed time up front, since it's the same for all
* devices.
*/
diskstat.da_trsf = diskstat.da_mb = 0.0;
diskstat.sa_trsf = diskstat.sa_mb = 0.0;
diskstat.cd_trsf = diskstat.cd_mb = 0.0;
busy_seconds = compute_etime(cur.busy_time, last.busy_time);
for (dn = 0; dn < num_devices; dn++) {
int di;
if (dev_select[dn].selected == 0)
continue;
di = dev_select[dn].position;
if (compute_stats(&cur.dinfo->devices[di],
&last.dinfo->devices[di], busy_seconds,
NULL, NULL,
NULL, NULL,
&transfers_per_second, &mb_per_second,
NULL, NULL) != 0)
errx(1, "xperfmon++: %s", devstat_errbuf);
if (strcmp(cur.dinfo->devices[di].device_name, "da") == 0 ||
strcmp(cur.dinfo->devices[di].device_name, "wd") == 0) {
#ifdef DEBUG
printf( "da%d: %Lf %Lf\n", cur.dinfo->devices[di].unit_number, transfers_per_second, mb_per_second);
#endif
diskstat.da_trsf += transfers_per_second;
diskstat.da_mb += mb_per_second;
}
if (strcmp(cur.dinfo->devices[di].device_name, "sa") == 0) {
#ifdef DEBUG
printf( "sa%d: %Lf %Lf\n", cur.dinfo->devices[di].unit_number, transfers_per_second, mb_per_second);
#endif
diskstat.sa_trsf += transfers_per_second;
diskstat.sa_mb += mb_per_second;
}
if (strcmp(cur.dinfo->devices[di].device_name, "cd") == 0 ||
strcmp(cur.dinfo->devices[di].device_name, "acd") == 0) {
#ifdef DEBUG
printf( "cd%d: %Lf %Lf\n", cur.dinfo->devices[di].unit_number, transfers_per_second, mb_per_second);
#endif
diskstat.cd_trsf += transfers_per_second;
diskstat.cd_mb += mb_per_second;
}
}
}