NetBSD/sys/miscfs/procfs/procfs_linux.c

607 lines
15 KiB
C
Raw Normal View History

/* $NetBSD: procfs_linux.c,v 1.58 2009/10/19 01:25:29 dholland Exp $ */
/*
* Copyright (c) 2001 Wasabi Systems, Inc.
* All rights reserved.
*
* Written by Frank van der Linden for Wasabi Systems, Inc.
*
* 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 for the NetBSD Project by
* Wasabi Systems, Inc.
* 4. The name of Wasabi Systems, Inc. may not be used to endorse
* or promote products derived from this software without specific prior
* written permission.
*
* THIS SOFTWARE IS PROVIDED BY WASABI SYSTEMS, INC. ``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 WASABI SYSTEMS, INC
* 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.
*/
2001-11-10 16:33:40 +03:00
#include <sys/cdefs.h>
__KERNEL_RCSID(0, "$NetBSD: procfs_linux.c,v 1.58 2009/10/19 01:25:29 dholland Exp $");
2001-11-10 16:33:40 +03:00
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/time.h>
#include <sys/kernel.h>
#include <sys/proc.h>
#include <sys/vnode.h>
#include <sys/exec.h>
#include <sys/resource.h>
#include <sys/resourcevar.h>
#include <sys/signal.h>
#include <sys/signalvar.h>
#include <sys/tty.h>
#include <sys/malloc.h>
#include <sys/mount.h>
2006-09-20 12:09:05 +04:00
#include <sys/conf.h>
#include <miscfs/procfs/procfs.h>
#include <compat/linux/common/linux_exec.h>
#include <uvm/uvm_extern.h>
#include <uvm/uvm.h>
2006-09-20 12:09:05 +04:00
extern struct devsw_conv *devsw_conv;
extern int max_devsw_convs;
#define PGTOB(p) ((unsigned long)(p) << PAGE_SHIFT)
#define PGTOKB(p) ((unsigned long)(p) << (PAGE_SHIFT - 10))
#define LBFSZ (8 * 1024)
static void
get_proc_size_info(struct lwp *l, unsigned long *stext, unsigned long *etext, unsigned long *sstack)
{
struct proc *p = l->l_proc;
struct vmspace *vm;
struct vm_map *map;
struct vm_map_entry *entry;
*stext = 0;
*etext = 0;
*sstack = 0;
proc_vmspace_getref(p, &vm);
map = &vm->vm_map;
vm_map_lock_read(map);
for (entry = map->header.next; entry != &map->header;
entry = entry->next) {
if (UVM_ET_ISSUBMAP(entry))
continue;
/* assume text is the first entry */
if (*stext == *etext) {
*stext = entry->start;
*etext = entry->end;
break;
}
}
#ifdef LINUX_USRSTACK32
if (strcmp(p->p_emul->e_name, "linux32") == 0 &&
LINUX_USRSTACK32 < USRSTACK32)
*sstack = (unsigned long)LINUX_USRSTACK32;
else
#endif
#ifdef LINUX_USRSTACK
if (strcmp(p->p_emul->e_name, "linux") == 0 &&
LINUX_USRSTACK < USRSTACK)
*sstack = (unsigned long)LINUX_USRSTACK;
else
#endif
#ifdef USRSTACK32
if (strstr(p->p_emul->e_name, "32") != NULL)
*sstack = (unsigned long)USRSTACK32;
else
#endif
*sstack = (unsigned long)USRSTACK;
/*
* jdk 1.6 compares low <= addr && addr < high
* if we put addr == high, then the test fails
* so eat one page.
*/
*sstack -= PAGE_SIZE;
vm_map_unlock_read(map);
uvmspace_free(vm);
}
/*
* Linux compatible /proc/meminfo. Only active when the -o linux
* mountflag is used.
*/
int
procfs_domeminfo(struct lwp *curl, struct proc *p,
struct pfsnode *pfs, struct uio *uio)
{
char *bf;
2004-04-22 04:31:00 +04:00
int len;
int error = 0;
bf = malloc(LBFSZ, M_TEMP, M_WAITOK);
len = snprintf(bf, LBFSZ,
" total: used: free: shared: buffers: cached:\n"
"Mem: %8lu %8lu %8lu %8lu %8lu %8lu\n"
"Swap: %8lu %8lu %8lu\n"
"MemTotal: %8lu kB\n"
"MemFree: %8lu kB\n"
"MemShared: %8lu kB\n"
"Buffers: %8lu kB\n"
"Cached: %8lu kB\n"
2005-02-27 01:58:54 +03:00
"SwapTotal: %8lu kB\n"
"SwapFree: %8lu kB\n",
PGTOB(uvmexp.npages),
PGTOB(uvmexp.npages - uvmexp.free),
PGTOB(uvmexp.free),
0L,
PGTOB(uvmexp.filepages),
PGTOB(uvmexp.anonpages + uvmexp.filepages + uvmexp.execpages),
PGTOB(uvmexp.swpages),
PGTOB(uvmexp.swpginuse),
PGTOB(uvmexp.swpages - uvmexp.swpginuse),
PGTOKB(uvmexp.npages),
PGTOKB(uvmexp.free),
0L,
PGTOKB(uvmexp.filepages),
PGTOKB(uvmexp.anonpages + uvmexp.filepages + uvmexp.execpages),
PGTOKB(uvmexp.swpages),
PGTOKB(uvmexp.swpages - uvmexp.swpginuse));
if (len == 0)
goto out;
error = uiomove_frombuf(bf, len, uio);
out:
free(bf, M_TEMP);
return error;
}
2006-09-20 12:09:05 +04:00
/*
* Linux compatible /proc/devices. Only active when the -o linux
* mountflag is used.
*/
int
procfs_dodevices(struct lwp *curl, struct proc *p,
struct pfsnode *pfs, struct uio *uio)
2006-09-20 12:09:05 +04:00
{
char *bf;
2006-09-20 12:09:05 +04:00
int offset = 0;
int i, error = ENAMETOOLONG;
/* XXX elad - may need filtering. */
bf = malloc(LBFSZ, M_TEMP, M_WAITOK);
2006-09-20 12:09:05 +04:00
offset += snprintf(&bf[offset], LBFSZ - offset, "Character devices:\n");
if (offset >= LBFSZ)
goto out;
2006-09-20 12:09:05 +04:00
mutex_enter(&device_lock);
2006-09-20 12:09:05 +04:00
for (i = 0; i < max_devsw_convs; i++) {
if ((devsw_conv[i].d_name == NULL) ||
(devsw_conv[i].d_cmajor == -1))
continue;
offset += snprintf(&bf[offset], LBFSZ - offset,
2006-09-20 12:09:05 +04:00
"%3d %s\n", devsw_conv[i].d_cmajor, devsw_conv[i].d_name);
if (offset >= LBFSZ) {
mutex_exit(&device_lock);
goto out;
}
2006-09-20 12:09:05 +04:00
}
offset += snprintf(&bf[offset], LBFSZ - offset, "\nBlock devices:\n");
if (offset >= LBFSZ) {
mutex_exit(&device_lock);
goto out;
}
2006-09-20 12:09:05 +04:00
for (i = 0; i < max_devsw_convs; i++) {
if ((devsw_conv[i].d_name == NULL) ||
(devsw_conv[i].d_bmajor == -1))
continue;
offset += snprintf(&bf[offset], LBFSZ - offset,
2006-09-20 12:09:05 +04:00
"%3d %s\n", devsw_conv[i].d_bmajor, devsw_conv[i].d_name);
if (offset >= LBFSZ) {
mutex_exit(&device_lock);
goto out;
}
2006-09-20 12:09:05 +04:00
}
mutex_exit(&device_lock);
2006-09-20 12:09:05 +04:00
error = uiomove_frombuf(bf, offset, uio);
out:
free(bf, M_TEMP);
return error;
2006-09-20 12:09:05 +04:00
}
/*
* Linux compatible /proc/stat. Only active when the -o linux
* mountflag is used.
*/
int
procfs_docpustat(struct lwp *curl, struct proc *p,
struct pfsnode *pfs, struct uio *uio)
{
char *bf;
int error;
int len;
#if defined(MULTIPROCESSOR)
struct cpu_info *ci;
CPU_INFO_ITERATOR cii;
#endif
int i;
error = ENAMETOOLONG;
bf = malloc(LBFSZ, M_TEMP, M_WAITOK);
len = snprintf(bf, LBFSZ,
"cpu %" PRIu64 " %" PRIu64 " %" PRIu64 " %" PRIu64 "\n",
curcpu()->ci_schedstate.spc_cp_time[CP_USER],
curcpu()->ci_schedstate.spc_cp_time[CP_NICE],
curcpu()->ci_schedstate.spc_cp_time[CP_SYS] /*+ [CP_INTR]*/,
curcpu()->ci_schedstate.spc_cp_time[CP_IDLE]);
if (len == 0)
goto out;
#if defined(MULTIPROCESSOR)
#define ALLCPUS CPU_INFO_FOREACH(cii, ci)
#define CPUNAME ci
#else
#define ALLCPUS ; i < 1 ;
#define CPUNAME curcpu()
#endif
i = 0;
for (ALLCPUS) {
len += snprintf(&bf[len], LBFSZ - len,
"cpu%d %" PRIu64 " %" PRIu64 " %" PRIu64 " %" PRIu64
"\n", i,
CPUNAME->ci_schedstate.spc_cp_time[CP_USER],
CPUNAME->ci_schedstate.spc_cp_time[CP_NICE],
CPUNAME->ci_schedstate.spc_cp_time[CP_SYS],
CPUNAME->ci_schedstate.spc_cp_time[CP_IDLE]);
if (len >= LBFSZ)
goto out;
i += 1;
}
len += snprintf(&bf[len], LBFSZ - len,
"disk 0 0 0 0\n"
"page %u %u\n"
"swap %u %u\n"
"intr %u\n"
"ctxt %u\n"
"btime %lld\n",
uvmexp.pageins, uvmexp.pdpageouts,
uvmexp.pgswapin, uvmexp.pgswapout,
uvmexp.intrs,
uvmexp.swtch,
(long long)boottime.tv_sec);
if (len >= LBFSZ)
goto out;
error = uiomove_frombuf(bf, len, uio);
out:
free(bf, M_TEMP);
return error;
}
/*
* Linux compatible /proc/loadavg. Only active when the -o linux
* mountflag is used.
*/
int
procfs_doloadavg(struct lwp *curl, struct proc *p,
struct pfsnode *pfs, struct uio *uio)
{
char *bf;
int error;
int len;
error = ENAMETOOLONG;
bf = malloc(LBFSZ, M_TEMP, M_WAITOK);
averunnable.fscale = FSCALE;
len = snprintf(bf, LBFSZ,
"%d.%02d %d.%02d %d.%02d %d/%d %d\n",
(int)(averunnable.ldavg[0] / averunnable.fscale),
(int)(averunnable.ldavg[0] * 100 / averunnable.fscale % 100),
(int)(averunnable.ldavg[1] / averunnable.fscale),
(int)(averunnable.ldavg[1] * 100 / averunnable.fscale % 100),
(int)(averunnable.ldavg[2] / averunnable.fscale),
(int)(averunnable.ldavg[2] * 100 / averunnable.fscale % 100),
1, /* number of ONPROC processes */
nprocs,
30000); /* last pid */
if (len == 0)
goto out;
error = uiomove_frombuf(bf, len, uio);
out:
free(bf, M_TEMP);
return error;
}
/*
* Linux compatible /proc/<pid>/statm. Only active when the -o linux
* mountflag is used.
*/
int
procfs_do_pid_statm(struct lwp *curl, struct lwp *l,
struct pfsnode *pfs, struct uio *uio)
{
struct vmspace *vm;
struct proc *p = l->l_proc;
struct rusage *ru = &p->p_stats->p_ru;
char *bf;
int error;
int len;
error = ENAMETOOLONG;
bf = malloc(LBFSZ, M_TEMP, M_WAITOK);
/* XXX - we use values from vmspace, since dsl says that ru figures
are always 0 except for zombies. See kvm_proc.c::kvm_getproc2() */
if ((error = proc_vmspace_getref(p, &vm)) != 0) {
goto out;
}
len = snprintf(bf, LBFSZ,
"%lu %lu %lu %lu %lu %lu %lu\n",
(unsigned long)(vm->vm_tsize + vm->vm_dsize + vm->vm_ssize), /* size */
(unsigned long)(vm->vm_rssize), /* resident */
(unsigned long)(ru->ru_ixrss), /* shared */
(unsigned long)(vm->vm_tsize), /* text size in pages */
(unsigned long)(vm->vm_dsize), /* data size in pages */
(unsigned long)(vm->vm_ssize), /* stack size in pages */
(unsigned long) 0);
uvmspace_free(vm);
if (len == 0)
goto out;
error = uiomove_frombuf(bf, len, uio);
out:
free(bf, M_TEMP);
return error;
}
#define USEC_2_TICKS(x) ((x) / 10000)
/*
* Linux compatible /proc/<pid>/stat. Only active when the -o linux
* mountflag is used.
*/
int
procfs_do_pid_stat(struct lwp *curl, struct lwp *l,
struct pfsnode *pfs, struct uio *uio)
{
char *bf;
struct proc *p = l->l_proc;
2004-04-22 04:31:00 +04:00
int len;
struct tty *tty = p->p_session->s_ttyp;
struct rusage *ru = &p->p_stats->p_ru;
struct rusage *cru = &p->p_stats->p_cru;
unsigned long stext = 0, etext = 0, sstack = 0;
2007-02-10 00:55:00 +03:00
struct timeval rt;
struct vmspace *vm;
int error = 0;
bf = malloc(LBFSZ, M_TEMP, M_WAITOK);
if ((error = proc_vmspace_getref(p, &vm)) != 0) {
goto out;
}
get_proc_size_info(l, &stext, &etext, &sstack);
2007-02-10 00:55:00 +03:00
mutex_enter(proc_lock);
mutex_enter(p->p_lock);
2007-02-10 00:55:00 +03:00
calcru(p, NULL, NULL, NULL, &rt);
len = snprintf(bf, LBFSZ,
2009-01-11 05:45:45 +03:00
"%d (%s) %c %d %d %d %lld %d "
"%u "
"%lu %lu %lu %lu %lu %lu %lu %lu "
"%d %d %d "
2009-01-11 05:45:45 +03:00
"%lld %lld %lu %lu %" PRIu64 " "
"%lu %lu %lu "
"%u %u "
"%u %u %u %u "
"%lu %lu %lu %d %d\n",
p->p_pid,
p->p_comm,
"0IR3SZD"[(p->p_stat > 6) ? 0 : (int)p->p_stat],
(p->p_pptr != NULL) ? p->p_pptr->p_pid : 0,
p->p_pgid,
p->p_session->s_sid,
2009-01-11 06:16:33 +03:00
(unsigned long long)(tty ? tty->t_dev : 0),
2003-10-30 17:51:01 +03:00
(tty && tty->t_pgrp) ? tty->t_pgrp->pg_id : 0,
p->p_flag,
ru->ru_minflt,
cru->ru_minflt,
ru->ru_majflt,
cru->ru_majflt,
2009-01-11 05:45:45 +03:00
(long)USEC_2_TICKS(ru->ru_utime.tv_usec),
(long)USEC_2_TICKS(ru->ru_stime.tv_usec),
(long)USEC_2_TICKS(cru->ru_utime.tv_usec),
(long)USEC_2_TICKS(cru->ru_stime.tv_usec),
l->l_priority, /* XXX: priority */
p->p_nice - 20,
0,
2009-01-11 05:45:45 +03:00
(long long)rt.tv_sec,
(long long)p->p_stats->p_start.tv_sec,
(unsigned long)(vm->vm_tsize + vm->vm_dsize + vm->vm_ssize), /* size */
(unsigned long)(vm->vm_rssize), /* resident */
p->p_rlimit[RLIMIT_RSS].rlim_cur,
stext, /* start code */
etext, /* end code */
sstack, /* mm start stack */
0, /* XXX: pc */
0, /* XXX: sp */
2007-02-10 00:55:00 +03:00
p->p_sigpend.sp_set.__bits[0], /* XXX: pending */
0, /* XXX: held */
p->p_sigctx.ps_sigignore.__bits[0], /* ignored */
p->p_sigctx.ps_sigcatch.__bits[0], /* caught */
(unsigned long)(intptr_t)l->l_wchan,
ru->ru_nvcsw,
ru->ru_nivcsw,
p->p_exitsig,
0); /* XXX: processor */
mutex_exit(p->p_lock);
mutex_exit(proc_lock);
2007-02-10 00:55:00 +03:00
uvmspace_free(vm);
if (len == 0)
goto out;
error = uiomove_frombuf(bf, len, uio);
out:
free(bf, M_TEMP);
return error;
}
int
procfs_docpuinfo(struct lwp *curl, struct proc *p,
struct pfsnode *pfs, struct uio *uio)
{
int len = LBFSZ;
char *bf = malloc(len, M_TEMP, M_WAITOK);
int error;
if (procfs_getcpuinfstr(bf, &len) < 0) {
error = ENOSPC;
goto done;
}
if (len == 0) {
error = 0;
goto done;
}
error = uiomove_frombuf(bf, len, uio);
done:
free(bf, M_TEMP);
return error;
}
int
procfs_douptime(struct lwp *curl, struct proc *p,
struct pfsnode *pfs, struct uio *uio)
{
char *bf;
2004-04-22 04:31:00 +04:00
int len;
struct timeval runtime;
u_int64_t idle;
int error = 0;
bf = malloc(LBFSZ, M_TEMP, M_WAITOK);
microuptime(&runtime);
idle = curcpu()->ci_schedstate.spc_cp_time[CP_IDLE];
len = snprintf(bf, LBFSZ,
2009-01-11 05:45:45 +03:00
"%lld.%02lu %" PRIu64 ".%02" PRIu64 "\n",
(long long)runtime.tv_sec, (long)runtime.tv_usec / 10000,
2004-04-22 04:31:00 +04:00
idle / hz, (((idle % hz) * 100) / hz) % 100);
if (len == 0)
goto out;
error = uiomove_frombuf(bf, len, uio);
out:
free(bf, M_TEMP);
return error;
}
int
procfs_domounts(struct lwp *curl, struct proc *p,
struct pfsnode *pfs, struct uio *uio)
{
char *bf, *mtab = NULL;
const char *fsname;
size_t len, mtabsz = 0;
struct mount *mp, *nmp;
struct statvfs *sfs;
int error = 0;
bf = malloc(LBFSZ, M_TEMP, M_WAITOK);
mutex_enter(&mountlist_lock);
for (mp = CIRCLEQ_FIRST(&mountlist); mp != (void *)&mountlist;
mp = nmp) {
if (vfs_busy(mp, &nmp)) {
continue;
}
sfs = &mp->mnt_stat;
/* Linux uses different names for some filesystems */
fsname = sfs->f_fstypename;
if (strcmp(fsname, "procfs") == 0)
fsname = "proc";
else if (strcmp(fsname, "ext2fs") == 0)
fsname = "ext2";
2005-02-27 01:58:54 +03:00
len = snprintf(bf, LBFSZ, "%s %s %s %s%s%s%s%s%s 0 0\n",
sfs->f_mntfromname,
sfs->f_mntonname,
fsname,
(mp->mnt_flag & MNT_RDONLY) ? "ro" : "rw",
(mp->mnt_flag & MNT_NOSUID) ? ",nosuid" : "",
(mp->mnt_flag & MNT_NOEXEC) ? ",noexec" : "",
(mp->mnt_flag & MNT_NODEV) ? ",nodev" : "",
(mp->mnt_flag & MNT_SYNCHRONOUS) ? ",sync" : "",
(mp->mnt_flag & MNT_NOATIME) ? ",noatime" : ""
);
mtab = realloc(mtab, mtabsz + len, M_TEMP, M_WAITOK);
memcpy(mtab + mtabsz, bf, len);
mtabsz += len;
vfs_unbusy(mp, false, &nmp);
}
mutex_exit(&mountlist_lock);
free(bf, M_TEMP);
if (mtabsz > 0) {
error = uiomove_frombuf(mtab, mtabsz, uio);
free(mtab, M_TEMP);
}
return error;
}