freelance-project-34-market.../deps/cpulimit-fork/src/process_group.c
2022-12-31 19:45:03 +03:00

252 lines
7.2 KiB
C

/**
*
* cpulimit - a CPU limiter for Linux
*
* Copyright (C) 2005-2012, by: Angelo Marletta <angelo dot marletta at gmail dot com>
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version 2
* of the License, or (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
*/
#include <string.h>
#include <stdlib.h>
#include <limits.h>
#include <stdio.h>
#include <sys/time.h>
#include <signal.h>
#include <assert.h>
#include "process_iterator.h"
#include "process_group.h"
#include "list.h"
// look for a process by pid
// search_pid : pid of the wanted process
// return: pid of the found process, if successful
// negative pid, if the process does not exist or if the signal fails
int find_process_by_pid(pid_t pid)
{
return (kill(pid,0)==0) ? pid : -pid;
}
// look for a process with a given name
// process: the name of the wanted process. it can be an absolute path name to the executable file
// or just the file name
// return: pid of the found process, if it is found
// 0, if it's not found
// negative pid, if it is found but it's not possible to control it
int find_process_by_name(const char *process_name)
{
//pid of the target process
pid_t pid = -1;
//process iterator
struct process_iterator it;
struct process proc;
struct process_filter filter;
filter.pid = 0;
filter.include_children = 0;
init_process_iterator(&it, &filter);
while (get_next_process(&it, &proc) != -1)
{
//process found
if (strncmp(basename(proc.command), process_name, strlen(process_name))==0 && kill(pid,SIGCONT)==0) {
//process is ok!
pid = proc.pid;
break;
}
}
if (close_process_iterator(&it) != 0) exit(1);
if (pid >= 0) {
//ok, the process was found
return pid;
}
else {
//process not found
return 0;
}
}
int init_process_group(struct process_group *pgroup, int target_pid, int include_children)
{
//hashtable initialization
memset(&pgroup->proctable, 0, sizeof(pgroup->proctable));
pgroup->target_pid = target_pid;
pgroup->include_children = include_children;
pgroup->proclist = (struct list*)malloc(sizeof(struct list));
init_list(pgroup->proclist, 4);
memset(&pgroup->last_update, 0, sizeof(pgroup->last_update));
update_process_group(pgroup);
return 0;
}
int close_process_group(struct process_group *pgroup)
{
int i;
int size = sizeof(pgroup->proctable) / sizeof(struct process*);
for (i=0; i<size; i++) {
if (pgroup->proctable[i] != NULL) {
//free() history for each process
destroy_list(pgroup->proctable[i]);
free(pgroup->proctable[i]);
pgroup->proctable[i] = NULL;
}
}
clear_list(pgroup->proclist);
free(pgroup->proclist);
pgroup->proclist = NULL;
return 0;
}
void remove_terminated_processes(struct process_group *pgroup)
{
//TODO
}
//return t1-t2 in microseconds (no overflow checks, so better watch out!)
static inline unsigned long timediff(const struct timeval *t1,const struct timeval *t2)
{
return (t1->tv_sec - t2->tv_sec) * 1000000 + (t1->tv_usec - t2->tv_usec);
}
//parameter in range 0-1
#define ALFA 0.08
#define MIN_DT 20
void update_process_group_fast(struct process_group *pgroup, int cycle, int verbose)
{
struct list_node *node;
struct process tmp_process;
struct timeval now;
long dt;
double sample;
gettimeofday(&now, NULL);
//time elapsed from previous sample (in ms)
dt = timediff(&now, &pgroup->last_update) / 1000;
//estimate how much the controlled processes are using the cpu in the working interval
for (node = pgroup->proclist->first; node != NULL; node = node->next) {
struct process *proc = (struct process*)(node->data);
if (read_process_info(proc->pid, &tmp_process) != 0)
{
continue;
}
sample = 1.0 * (tmp_process.cputime - proc->cputime) / dt;
if (proc->cpu_usage == -1) {
//initialization
proc->cpu_usage = sample;
}
else {
//usage adjustment
proc->cpu_usage = (1.0-ALFA) * proc->cpu_usage + ALFA * sample;
}
proc->cputime = tmp_process.cputime;
if (verbose && cycle % 10 == 0)
{
printf(
"%d %5.2lf, ",
proc->pid,
proc->cpu_usage * 100
);
}
proc->cputime = tmp_process.cputime;
}
pgroup->last_update = now;
}
void update_process_group(struct process_group *pgroup)
{
struct process_iterator it;
struct process tmp_process;
struct process_filter filter;
struct timeval now;
gettimeofday(&now, NULL);
//time elapsed from previous sample (in ms)
long dt = timediff(&now, &pgroup->last_update) / 1000;
filter.pid = pgroup->target_pid;
filter.include_children = pgroup->include_children;
init_process_iterator(&it, &filter);
clear_list(pgroup->proclist);
init_list(pgroup->proclist, 4);
while (get_next_process(&it, &tmp_process) != -1)
{
// struct timeval t;
// gettimeofday(&t, NULL);
// printf("T=%ld.%ld PID=%d PPID=%d START=%d CPUTIME=%d\n", t.tv_sec, t.tv_usec, tmp_process.pid, tmp_process.ppid, tmp_process.starttime, tmp_process.cputime);
int hashkey = pid_hashfn(tmp_process.pid);
if (pgroup->proctable[hashkey] == NULL)
{
//empty bucket
pgroup->proctable[hashkey] = malloc(sizeof(struct list));
struct process *new_process = malloc(sizeof(struct process));
tmp_process.cpu_usage = -1;
memcpy(new_process, &tmp_process, sizeof(struct process));
init_list(pgroup->proctable[hashkey], 4);
add_elem(pgroup->proctable[hashkey], new_process);
add_elem(pgroup->proclist, new_process);
}
else
{
//existing bucket
struct process *p = (struct process*)locate_elem(pgroup->proctable[hashkey], &tmp_process);
if (p == NULL)
{
//process is new. add it
struct process *new_process = malloc(sizeof(struct process));
tmp_process.cpu_usage = -1;
memcpy(new_process, &tmp_process, sizeof(struct process));
add_elem(pgroup->proctable[hashkey], new_process);
add_elem(pgroup->proclist, new_process);
}
else
{
assert(tmp_process.pid == p->pid);
assert(tmp_process.starttime == p->starttime);
add_elem(pgroup->proclist, p);
if (dt < MIN_DT) continue;
//process exists. update CPU usage
double sample = 1.0 * (tmp_process.cputime - p->cputime) / dt;
if (p->cpu_usage == -1) {
//initialization
p->cpu_usage = sample;
}
else {
//usage adjustment
p->cpu_usage = (1.0-ALFA) * p->cpu_usage + ALFA * sample;
}
p->cputime = tmp_process.cputime;
}
}
}
close_process_iterator(&it);
if (dt < MIN_DT) return;
pgroup->last_update = now;
}
int remove_process(struct process_group *pgroup, int pid)
{
int hashkey = pid_hashfn(pid);
if (pgroup->proctable[hashkey] == NULL) return 1; //nothing to delete
struct list_node *node = (struct list_node*)locate_node(pgroup->proctable[hashkey], &pid);
if (node == NULL) return 2;
delete_node(pgroup->proctable[hashkey], node);
return 0;
}