gems-kernel/source/THIRDPARTY/linux-old/fs/proc/array.c
2024-06-03 11:16:15 -05:00

548 lines
12 KiB
C

/*
* linux/fs/proc/array.c
*
* Copyright (C) 1992 by Linus Torvalds
* based on ideas by Darren Senn
*
* stat,statm extensions by Michael K. Johnson, johnsonm@stolaf.edu
*/
#include <linux/types.h>
#include <linux/errno.h>
#include <linux/sched.h>
#include <linux/kernel.h>
#include <linux/kernel_stat.h>
#include <linux/tty.h>
#include <linux/user.h>
#include <linux/a.out.h>
#include <linux/string.h>
#include <linux/mman.h>
#include <asm/segment.h>
#include <asm/io.h>
#define LOAD_INT(x) ((x) >> FSHIFT)
#define LOAD_FRAC(x) LOAD_INT(((x) & (FIXED_1-1)) * 100)
#ifdef CONFIG_DEBUG_MALLOC
int get_malloc(char * buffer);
#endif
static int read_core(struct inode * inode, struct file * file,char * buf, int count)
{
unsigned long p = file->f_pos;
int read;
int count1;
char * pnt;
struct user dump;
memset(&dump, 0, sizeof(struct user));
dump.magic = CMAGIC;
dump.u_dsize = high_memory >> 12;
if (count < 0)
return -EINVAL;
if (p >= high_memory)
return 0;
if (count > high_memory - p)
count = high_memory - p;
read = 0;
if (p < sizeof(struct user) && count > 0) {
count1 = count;
if (p + count1 > sizeof(struct user))
count1 = sizeof(struct user)-p;
pnt = (char *) &dump + p;
memcpy_tofs(buf,(void *) pnt, count1);
buf += count1;
p += count1;
count -= count1;
read += count1;
}
while (p < 2*PAGE_SIZE && count > 0) {
put_fs_byte(0,buf);
buf++;
p++;
count--;
read++;
}
memcpy_tofs(buf,(void *) (p - PAGE_SIZE),count);
read += count;
file->f_pos += read;
return read;
}
static int get_loadavg(char * buffer)
{
int a, b, c;
a = avenrun[0] + (FIXED_1/200);
b = avenrun[1] + (FIXED_1/200);
c = avenrun[2] + (FIXED_1/200);
return sprintf(buffer,"%d.%02d %d.%02d %d.%02d\n",
LOAD_INT(a), LOAD_FRAC(a),
LOAD_INT(b), LOAD_FRAC(b),
LOAD_INT(c), LOAD_FRAC(c));
}
static int get_kstat(char * buffer)
{
return sprintf(buffer, "cpu %u %u %u %lu\n"
"disk %u %u %u %u\n"
"page %u %u\n"
"swap %u %u\n"
"intr %u\n"
"ctxt %u\n"
"btime %lu\n",
kstat.cpu_user,
kstat.cpu_nice,
kstat.cpu_system,
jiffies - (kstat.cpu_user + kstat.cpu_nice + kstat.cpu_system),
kstat.dk_drive[0],
kstat.dk_drive[1],
kstat.dk_drive[2],
kstat.dk_drive[3],
kstat.pgpgin,
kstat.pgpgout,
kstat.pswpin,
kstat.pswpout,
kstat.interrupts,
kstat.context_swtch,
xtime.tv_sec - jiffies / HZ);
}
static int get_uptime(char * buffer)
{
unsigned long uptime;
unsigned long idle;
uptime = jiffies;
idle = task[0]->utime + task[0]->stime;
return sprintf(buffer,"%lu.%02lu %lu.%02lu\n",
uptime / HZ,
uptime % HZ,
idle / HZ,
idle % HZ);
}
static int get_meminfo(char * buffer)
{
struct sysinfo i;
si_meminfo(&i);
si_swapinfo(&i);
return sprintf(buffer, " total: used: free: shared: buffers:\n"
"Mem: %8lu %8lu %8lu %8lu %8lu\n"
"Swap: %8lu %8lu %8lu\n",
i.totalram, i.totalram-i.freeram, i.freeram, i.sharedram, i.bufferram,
i.totalswap, i.totalswap-i.freeswap, i.freeswap);
}
static int get_version(char * buffer)
{
extern char *linux_banner;
strcpy(buffer, linux_banner);
return strlen(buffer);
}
static struct task_struct ** get_task(pid_t pid)
{
struct task_struct ** p;
p = task;
while (++p < task+NR_TASKS) {
if (*p && (*p)->pid == pid)
return p;
}
return NULL;
}
static unsigned long get_phys_addr(struct task_struct ** p, unsigned long ptr)
{
unsigned long page;
if (!p || !*p || ptr >= TASK_SIZE)
return 0;
page = *PAGE_DIR_OFFSET((*p)->tss.cr3,ptr);
if (!(page & 1))
return 0;
page &= PAGE_MASK;
page += PAGE_PTR(ptr);
page = *(unsigned long *) page;
if (!(page & 1))
return 0;
page &= PAGE_MASK;
page += ptr & ~PAGE_MASK;
return page;
}
static int get_array(struct task_struct ** p, unsigned long start, unsigned long end, char * buffer)
{
unsigned long addr;
int size = 0, result = 0;
char c;
if (start >= end)
return result;
for (;;) {
addr = get_phys_addr(p, start);
if (!addr)
return result;
do {
c = *(char *) addr;
if (!c)
result = size;
if (size < PAGE_SIZE)
buffer[size++] = c;
else
return result;
addr++;
start++;
if (start >= end)
return result;
} while (!(addr & ~PAGE_MASK));
}
}
static int get_env(int pid, char * buffer)
{
struct task_struct ** p = get_task(pid);
if (!p || !*p)
return 0;
return get_array(p, (*p)->env_start, (*p)->env_end, buffer);
}
static int get_arg(int pid, char * buffer)
{
struct task_struct ** p = get_task(pid);
if (!p || !*p)
return 0;
return get_array(p, (*p)->arg_start, (*p)->arg_end, buffer);
}
static unsigned long get_wchan(struct task_struct *p)
{
unsigned long ebp, eip;
unsigned long stack_page;
int count = 0;
if (!p || p == current || p->state == TASK_RUNNING)
return 0;
stack_page = p->kernel_stack_page;
if (!stack_page)
return 0;
ebp = p->tss.ebp;
do {
if (ebp < stack_page || ebp >= 4092+stack_page)
return 0;
eip = *(unsigned long *) (ebp+4);
if ((void *)eip != sleep_on &&
(void *)eip != interruptible_sleep_on)
return eip;
ebp = *(unsigned long *) ebp;
} while (count++ < 16);
return 0;
}
#define KSTK_EIP(stack) (((unsigned long *)stack)[1019])
#define KSTK_ESP(stack) (((unsigned long *)stack)[1022])
static int get_stat(int pid, char * buffer)
{
struct task_struct ** p = get_task(pid);
unsigned long sigignore=0, sigcatch=0, bit=1, wchan;
unsigned long vsize, eip, esp;
int i,tty_pgrp;
char state;
if (!p || !*p)
return 0;
if ((*p)->state < 0 || (*p)->state > 5)
state = '.';
else
state = "RSDZTD"[(*p)->state];
eip = esp = 0;
vsize = (*p)->kernel_stack_page;
if (vsize) {
eip = KSTK_EIP(vsize);
esp = KSTK_ESP(vsize);
vsize = (*p)->brk - (*p)->start_code + PAGE_SIZE-1;
if (esp)
vsize += TASK_SIZE - esp;
}
wchan = get_wchan(*p);
for(i=0; i<32; ++i) {
switch((int) (*p)->sigaction[i].sa_handler) {
case 1: sigignore |= bit; break;
case 0: break;
default: sigcatch |= bit;
} bit <<= 1;
}
tty_pgrp = (*p)->tty;
if (tty_pgrp > 0 && tty_table[tty_pgrp])
tty_pgrp = tty_table[tty_pgrp]->pgrp;
else
tty_pgrp = -1;
return sprintf(buffer,"%d (%s) %c %d %d %d %d %d %lu %lu \
%lu %lu %lu %ld %ld %ld %ld %ld %ld %lu %lu %ld %lu %u %u %lu %lu %lu %lu %lu %lu \
%lu %lu %lu %lu\n",
pid,
(*p)->comm,
state,
(*p)->p_pptr->pid,
(*p)->pgrp,
(*p)->session,
(*p)->tty,
tty_pgrp,
(*p)->flags,
(*p)->min_flt,
(*p)->cmin_flt,
(*p)->maj_flt,
(*p)->cmaj_flt,
(*p)->utime,
(*p)->stime,
(*p)->cutime,
(*p)->cstime,
(*p)->counter, /* this is the kernel priority ---
subtract 30 in your user-level program. */
(*p)->priority, /* this is the nice value ---
subtract 15 in your user-level program. */
(*p)->timeout,
(*p)->it_real_value,
(*p)->start_time,
vsize,
(*p)->rss, /* you might want to shift this left 3 */
(*p)->rlim[RLIMIT_RSS].rlim_cur,
(*p)->start_code,
(*p)->end_code,
(*p)->start_stack,
esp,
eip,
(*p)->signal,
(*p)->blocked,
sigignore,
sigcatch,
wchan);
}
static int get_statm(int pid, char * buffer)
{
struct task_struct ** p = get_task(pid);
int i, tpag;
int size=0, resident=0, share=0, trs=0, lrs=0, drs=0, dt=0;
unsigned long ptbl, *buf, *pte, *pagedir, map_nr;
if (!p || !*p)
return 0;
tpag = (*p)->end_code / PAGE_SIZE;
if ((*p)->state != TASK_ZOMBIE) {
pagedir = (unsigned long *) (*p)->tss.cr3;
for (i = 0; i < 0x300; ++i) {
if ((ptbl = pagedir[i]) == 0) {
tpag -= PTRS_PER_PAGE;
continue;
}
buf = (unsigned long *)(ptbl & PAGE_MASK);
for (pte = buf; pte < (buf + PTRS_PER_PAGE); ++pte) {
if (*pte != 0) {
++size;
if (*pte & 1) {
++resident;
if (tpag > 0)
++trs;
else
++drs;
if (i >= 15 && i < 0x2f0) {
++lrs;
if (*pte & 0x40)
++dt;
else
--drs;
}
map_nr = MAP_NR(*pte);
if (map_nr < (high_memory / PAGE_SIZE) && mem_map[map_nr] > 1)
++share;
}
}
--tpag;
}
}
}
return sprintf(buffer,"%d %d %d %d %d %d %d\n",
size, resident, share, trs, lrs, drs, dt);
}
static int get_maps(int pid, char *buf)
{
int sz = 0;
struct task_struct **p = get_task(pid);
struct vm_area_struct *map;
if (!p || !*p)
return 0;
for(map = (*p)->mmap; map != NULL; map = map->vm_next) {
char str[7], *cp = str;
int prot = map->vm_page_prot;
int perms, flags;
int end = sz + 80; /* Length of line */
dev_t dev;
unsigned long ino;
/*
* This tries to get an "rwxsp" string out of silly
* intel page permissions. The vm_area_struct should
* probably have the original mmap args preserved.
*/
flags = perms = 0;
if ((prot & PAGE_READONLY) == PAGE_READONLY)
perms |= PROT_READ | PROT_EXEC;
if (prot & (PAGE_COW|PAGE_RW)) {
perms |= PROT_WRITE | PROT_READ;
flags = prot & PAGE_COW ? MAP_PRIVATE : MAP_SHARED;
}
*cp++ = perms & PROT_READ ? 'r' : '-';
*cp++ = perms & PROT_WRITE ? 'w' : '-';
*cp++ = perms & PROT_EXEC ? 'x' : '-';
*cp++ = flags & MAP_SHARED ? 's' : '-';
*cp++ = flags & MAP_PRIVATE ? 'p' : '-';
*cp++ = 0;
if (end >= PAGE_SIZE) {
sprintf(buf+sz, "...\n");
break;
}
if (map->vm_inode != NULL) {
dev = map->vm_inode->i_dev;
ino = map->vm_inode->i_ino;
} else {
dev = 0;
ino = 0;
}
sz += sprintf(buf+sz, "%08lx-%08lx %s %08lx %02x:%02x %lu\n",
map->vm_start, map->vm_end, str, map->vm_offset,
MAJOR(dev),MINOR(dev), ino);
if (sz > end) {
printk("get_maps: end(%d) < sz(%d)\n", end, sz);
break;
}
}
return sz;
}
asmlinkage int get_module_list( char *);
static int array_read(struct inode * inode, struct file * file,char * buf, int count)
{
char * page;
int length;
int end;
unsigned int type, pid;
if (count < 0)
return -EINVAL;
if (!(page = (char*) __get_free_page(GFP_KERNEL)))
return -ENOMEM;
type = inode->i_ino;
pid = type >> 16;
type &= 0x0000ffff;
switch (type) {
case 2:
length = get_loadavg(page);
break;
case 3:
length = get_uptime(page);
break;
case 4:
length = get_meminfo(page);
break;
case 6:
length = get_version(page);
break;
case 9:
length = get_env(pid, page);
break;
case 10:
length = get_arg(pid, page);
break;
case 11:
length = get_stat(pid, page);
break;
case 12:
length = get_statm(pid, page);
break;
#ifdef CONFIG_DEBUG_MALLOC
case 13:
length = get_malloc(page);
break;
#endif
case 14:
free_page((unsigned long) page);
return read_core(inode, file, buf, count);
case 15:
length = get_maps(pid, page);
break;
case 16:
length = get_module_list(page);
break;
case 17:
length = get_kstat(page);
break;
default:
free_page((unsigned long) page);
return -EBADF;
}
if (file->f_pos >= length) {
free_page((unsigned long) page);
return 0;
}
if (count + file->f_pos > length)
count = length - file->f_pos;
end = count + file->f_pos;
memcpy_tofs(buf, page + file->f_pos, count);
free_page((unsigned long) page);
file->f_pos = end;
return count;
}
static struct file_operations proc_array_operations = {
NULL, /* array_lseek */
array_read,
NULL, /* array_write */
NULL, /* array_readdir */
NULL, /* array_select */
NULL, /* array_ioctl */
NULL, /* mmap */
NULL, /* no special open code */
NULL, /* no special release code */
NULL /* can't fsync */
};
struct inode_operations proc_array_inode_operations = {
&proc_array_operations, /* default base directory file-ops */
NULL, /* create */
NULL, /* lookup */
NULL, /* link */
NULL, /* unlink */
NULL, /* symlink */
NULL, /* mkdir */
NULL, /* rmdir */
NULL, /* mknod */
NULL, /* rename */
NULL, /* readlink */
NULL, /* follow_link */
NULL, /* bmap */
NULL, /* truncate */
NULL /* permission */
};