Merge branch 'for-5.5' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup

Pull cgroup updates from Tejun Heo:
 "There are several notable changes here:

   - Single thread migrating itself has been optimized so that it
     doesn't need threadgroup rwsem anymore.

   - Freezer optimization to avoid unnecessary frozen state changes.

   - cgroup ID unification so that cgroup fs ino is the only unique ID
     used for the cgroup and can be used to directly look up live
     cgroups through filehandle interface on 64bit ino archs. On 32bit
     archs, cgroup fs ino is still the only ID in use but it is only
     unique when combined with gen.

   - selftest and other changes"

* 'for-5.5' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup: (24 commits)
  writeback: fix -Wformat compilation warnings
  docs: cgroup: mm: Fix spelling of "list"
  cgroup: fix incorrect WARN_ON_ONCE() in cgroup_setup_root()
  cgroup: use cgrp->kn->id as the cgroup ID
  kernfs: use 64bit inos if ino_t is 64bit
  kernfs: implement custom exportfs ops and fid type
  kernfs: combine ino/id lookup functions into kernfs_find_and_get_node_by_id()
  kernfs: convert kernfs_node->id from union kernfs_node_id to u64
  kernfs: kernfs_find_and_get_node_by_ino() should only look up activated nodes
  kernfs: use dumber locking for kernfs_find_and_get_node_by_ino()
  netprio: use css ID instead of cgroup ID
  writeback: use ino_t for inodes in tracepoints
  kernfs: fix ino wrap-around detection
  kselftests: cgroup: Avoid the reuse of fd after it is deallocated
  cgroup: freezer: don't change task and cgroups status unnecessarily
  cgroup: use cgroup->last_bstat instead of cgroup->bstat_pending for consistency
  cgroup: remove cgroup_enable_task_cg_lists() optimization
  cgroup: pids: use atomic64_t for pids->limit
  selftests: cgroup: Run test_core under interfering stress
  selftests: cgroup: Add task migration tests
  ...
This commit is contained in:
Linus Torvalds
2019-11-25 19:23:46 -08:00
32 changed files with 746 additions and 534 deletions

View File

@@ -1,8 +1,10 @@
# SPDX-License-Identifier: GPL-2.0
CFLAGS += -Wall
CFLAGS += -Wall -pthread
all:
TEST_FILES := with_stress.sh
TEST_PROGS := test_stress.sh
TEST_GEN_PROGS = test_memcontrol
TEST_GEN_PROGS += test_core
TEST_GEN_PROGS += test_freezer

View File

@@ -158,6 +158,22 @@ long cg_read_key_long(const char *cgroup, const char *control, const char *key)
return atol(ptr + strlen(key));
}
long cg_read_lc(const char *cgroup, const char *control)
{
char buf[PAGE_SIZE];
const char delim[] = "\n";
char *line;
long cnt = 0;
if (cg_read(cgroup, control, buf, sizeof(buf)))
return -1;
for (line = strtok(buf, delim); line; line = strtok(NULL, delim))
cnt++;
return cnt;
}
int cg_write(const char *cgroup, const char *control, char *buf)
{
char path[PATH_MAX];
@@ -282,10 +298,12 @@ int cg_enter(const char *cgroup, int pid)
int cg_enter_current(const char *cgroup)
{
char pidbuf[64];
return cg_write(cgroup, "cgroup.procs", "0");
}
snprintf(pidbuf, sizeof(pidbuf), "%d", getpid());
return cg_write(cgroup, "cgroup.procs", pidbuf);
int cg_enter_current_thread(const char *cgroup)
{
return cg_write(cgroup, "cgroup.threads", "0");
}
int cg_run(const char *cgroup,
@@ -410,11 +428,25 @@ int set_oom_adj_score(int pid, int score)
return 0;
}
char proc_read_text(int pid, const char *item, char *buf, size_t size)
ssize_t proc_read_text(int pid, bool thread, const char *item, char *buf, size_t size)
{
char path[PATH_MAX];
snprintf(path, sizeof(path), "/proc/%d/%s", pid, item);
if (!pid)
snprintf(path, sizeof(path), "/proc/%s/%s",
thread ? "thread-self" : "self", item);
else
snprintf(path, sizeof(path), "/proc/%d/%s", pid, item);
return read_text(path, buf, size);
}
int proc_read_strstr(int pid, bool thread, const char *item, const char *needle)
{
char buf[PAGE_SIZE];
if (proc_read_text(pid, thread, item, buf, sizeof(buf)) < 0)
return -1;
return strstr(buf, needle) ? 0 : -1;
}

View File

@@ -1,4 +1,5 @@
/* SPDX-License-Identifier: GPL-2.0 */
#include <stdbool.h>
#include <stdlib.h>
#define PAGE_SIZE 4096
@@ -29,12 +30,14 @@ extern int cg_read_strstr(const char *cgroup, const char *control,
const char *needle);
extern long cg_read_long(const char *cgroup, const char *control);
long cg_read_key_long(const char *cgroup, const char *control, const char *key);
extern long cg_read_lc(const char *cgroup, const char *control);
extern int cg_write(const char *cgroup, const char *control, char *buf);
extern int cg_run(const char *cgroup,
int (*fn)(const char *cgroup, void *arg),
void *arg);
extern int cg_enter(const char *cgroup, int pid);
extern int cg_enter_current(const char *cgroup);
extern int cg_enter_current_thread(const char *cgroup);
extern int cg_run_nowait(const char *cgroup,
int (*fn)(const char *cgroup, void *arg),
void *arg);
@@ -45,4 +48,5 @@ extern int is_swap_enabled(void);
extern int set_oom_adj_score(int pid, int score);
extern int cg_wait_for_proc_count(const char *cgroup, int count);
extern int cg_killall(const char *cgroup);
extern char proc_read_text(int pid, const char *item, char *buf, size_t size);
extern ssize_t proc_read_text(int pid, bool thread, const char *item, char *buf, size_t size);
extern int proc_read_strstr(int pid, bool thread, const char *item, const char *needle);

View File

@@ -5,6 +5,9 @@
#include <unistd.h>
#include <stdio.h>
#include <errno.h>
#include <signal.h>
#include <string.h>
#include <pthread.h>
#include "../kselftest.h"
#include "cgroup_util.h"
@@ -354,6 +357,147 @@ cleanup:
return ret;
}
static void *dummy_thread_fn(void *arg)
{
return (void *)(size_t)pause();
}
/*
* Test threadgroup migration.
* All threads of a process are migrated together.
*/
static int test_cgcore_proc_migration(const char *root)
{
int ret = KSFT_FAIL;
int t, c_threads, n_threads = 13;
char *src = NULL, *dst = NULL;
pthread_t threads[n_threads];
src = cg_name(root, "cg_src");
dst = cg_name(root, "cg_dst");
if (!src || !dst)
goto cleanup;
if (cg_create(src))
goto cleanup;
if (cg_create(dst))
goto cleanup;
if (cg_enter_current(src))
goto cleanup;
for (c_threads = 0; c_threads < n_threads; ++c_threads) {
if (pthread_create(&threads[c_threads], NULL, dummy_thread_fn, NULL))
goto cleanup;
}
cg_enter_current(dst);
if (cg_read_lc(dst, "cgroup.threads") != n_threads + 1)
goto cleanup;
ret = KSFT_PASS;
cleanup:
for (t = 0; t < c_threads; ++t) {
pthread_cancel(threads[t]);
}
for (t = 0; t < c_threads; ++t) {
pthread_join(threads[t], NULL);
}
cg_enter_current(root);
if (dst)
cg_destroy(dst);
if (src)
cg_destroy(src);
free(dst);
free(src);
return ret;
}
static void *migrating_thread_fn(void *arg)
{
int g, i, n_iterations = 1000;
char **grps = arg;
char lines[3][PATH_MAX];
for (g = 1; g < 3; ++g)
snprintf(lines[g], sizeof(lines[g]), "0::%s", grps[g] + strlen(grps[0]));
for (i = 0; i < n_iterations; ++i) {
cg_enter_current_thread(grps[(i % 2) + 1]);
if (proc_read_strstr(0, 1, "cgroup", lines[(i % 2) + 1]))
return (void *)-1;
}
return NULL;
}
/*
* Test single thread migration.
* Threaded cgroups allow successful migration of a thread.
*/
static int test_cgcore_thread_migration(const char *root)
{
int ret = KSFT_FAIL;
char *dom = NULL;
char line[PATH_MAX];
char *grps[3] = { (char *)root, NULL, NULL };
pthread_t thr;
void *retval;
dom = cg_name(root, "cg_dom");
grps[1] = cg_name(root, "cg_dom/cg_src");
grps[2] = cg_name(root, "cg_dom/cg_dst");
if (!grps[1] || !grps[2] || !dom)
goto cleanup;
if (cg_create(dom))
goto cleanup;
if (cg_create(grps[1]))
goto cleanup;
if (cg_create(grps[2]))
goto cleanup;
if (cg_write(grps[1], "cgroup.type", "threaded"))
goto cleanup;
if (cg_write(grps[2], "cgroup.type", "threaded"))
goto cleanup;
if (cg_enter_current(grps[1]))
goto cleanup;
if (pthread_create(&thr, NULL, migrating_thread_fn, grps))
goto cleanup;
if (pthread_join(thr, &retval))
goto cleanup;
if (retval)
goto cleanup;
snprintf(line, sizeof(line), "0::%s", grps[1] + strlen(grps[0]));
if (proc_read_strstr(0, 1, "cgroup", line))
goto cleanup;
ret = KSFT_PASS;
cleanup:
cg_enter_current(root);
if (grps[2])
cg_destroy(grps[2]);
if (grps[1])
cg_destroy(grps[1]);
if (dom)
cg_destroy(dom);
free(grps[2]);
free(grps[1]);
free(dom);
return ret;
}
#define T(x) { x, #x }
struct corecg_test {
int (*fn)(const char *root);
@@ -366,6 +510,8 @@ struct corecg_test {
T(test_cgcore_parent_becomes_threaded),
T(test_cgcore_invalid_domain),
T(test_cgcore_populated),
T(test_cgcore_proc_migration),
T(test_cgcore_thread_migration),
};
#undef T

View File

@@ -72,6 +72,7 @@ static int cg_prepare_for_wait(const char *cgroup)
if (ret == -1) {
debug("Error: inotify_add_watch() failed\n");
close(fd);
fd = -1;
}
return fd;
@@ -701,7 +702,7 @@ static int proc_check_stopped(int pid)
char buf[PAGE_SIZE];
int len;
len = proc_read_text(pid, "stat", buf, sizeof(buf));
len = proc_read_text(pid, 0, "stat", buf, sizeof(buf));
if (len == -1) {
debug("Can't get %d stat\n", pid);
return -1;

View File

@@ -0,0 +1,4 @@
#!/bin/bash
# SPDX-License-Identifier: GPL-2.0
./with_stress.sh -s subsys -s fork ./test_core

View File

@@ -0,0 +1,101 @@
#!/bin/bash
# SPDX-License-Identifier: GPL-2.0
# Kselftest framework requirement - SKIP code is 4.
ksft_skip=4
stress_fork()
{
while true ; do
/usr/bin/true
sleep 0.01
done
}
stress_subsys()
{
local verb=+
while true ; do
echo $verb$subsys_ctrl >$sysfs/cgroup.subtree_control
[ $verb = "+" ] && verb=- || verb=+
# incommensurable period with other stresses
sleep 0.011
done
}
init_and_check()
{
sysfs=`mount -t cgroup2 | head -1 | awk '{ print $3 }'`
if [ ! -d "$sysfs" ]; then
echo "Skipping: cgroup2 is not mounted" >&2
exit $ksft_skip
fi
if ! echo +$subsys_ctrl >$sysfs/cgroup.subtree_control ; then
echo "Skipping: cannot enable $subsys_ctrl in $sysfs" >&2
exit $ksft_skip
fi
if ! echo -$subsys_ctrl >$sysfs/cgroup.subtree_control ; then
echo "Skipping: cannot disable $subsys_ctrl in $sysfs" >&2
exit $ksft_skip
fi
}
declare -a stresses
declare -a stress_pids
duration=5
rc=0
subsys_ctrl=cpuset
sysfs=
while getopts c:d:hs: opt; do
case $opt in
c)
subsys_ctrl=$OPTARG
;;
d)
duration=$OPTARG
;;
h)
echo "Usage $0 [ -s stress ] ... [ -d duration ] [-c controller] cmd args .."
echo -e "\t default duration $duration seconds"
echo -e "\t default controller $subsys_ctrl"
exit
;;
s)
func=stress_$OPTARG
if [ "x$(type -t $func)" != "xfunction" ] ; then
echo "Unknown stress $OPTARG"
exit 1
fi
stresses+=($func)
;;
esac
done
shift $((OPTIND - 1))
init_and_check
for s in ${stresses[*]} ; do
$s &
stress_pids+=($!)
done
time=0
start=$(date +%s)
while [ $time -lt $duration ] ; do
$*
rc=$?
[ $rc -eq 0 ] || break
time=$(($(date +%s) - $start))
done
for pid in ${stress_pids[*]} ; do
kill -SIGTERM $pid
wait $pid
done
exit $rc