OpenCloudOS-Kernel/kernel/torture.c

952 lines
25 KiB
C
Raw Normal View History

// SPDX-License-Identifier: GPL-2.0+
/*
* Common functions for in-kernel torture tests.
*
* Copyright (C) IBM Corporation, 2014
*
* Author: Paul E. McKenney <paulmck@linux.ibm.com>
* Based on kernel/rcu/torture.c.
*/
#define pr_fmt(fmt) fmt
#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/init.h>
#include <linux/module.h>
#include <linux/kthread.h>
#include <linux/err.h>
#include <linux/spinlock.h>
#include <linux/smp.h>
#include <linux/interrupt.h>
#include <linux/sched.h>
#include <linux/sched/clock.h>
#include <linux/atomic.h>
#include <linux/bitops.h>
#include <linux/completion.h>
#include <linux/moduleparam.h>
#include <linux/percpu.h>
#include <linux/notifier.h>
#include <linux/reboot.h>
#include <linux/freezer.h>
#include <linux/cpu.h>
#include <linux/delay.h>
#include <linux/stat.h>
#include <linux/slab.h>
#include <linux/trace_clock.h>
#include <linux/ktime.h>
#include <asm/byteorder.h>
#include <linux/torture.h>
#include <linux/sched/rt.h>
#include "rcu/rcu.h"
MODULE_LICENSE("GPL");
MODULE_AUTHOR("Paul E. McKenney <paulmck@linux.ibm.com>");
static bool disable_onoff_at_boot;
module_param(disable_onoff_at_boot, bool, 0444);
static bool ftrace_dump_at_shutdown;
module_param(ftrace_dump_at_shutdown, bool, 0444);
static int verbose_sleep_frequency;
module_param(verbose_sleep_frequency, int, 0444);
static int verbose_sleep_duration = 1;
module_param(verbose_sleep_duration, int, 0444);
static int random_shuffle;
module_param(random_shuffle, int, 0444);
static char *torture_type;
static int verbose;
/* Mediate rmmod and system shutdown. Concurrent rmmod & shutdown illegal! */
#define FULLSTOP_DONTSTOP 0 /* Normal operation. */
#define FULLSTOP_SHUTDOWN 1 /* System shutdown with torture running. */
#define FULLSTOP_RMMOD 2 /* Normal rmmod of torture. */
static int fullstop = FULLSTOP_RMMOD;
static DEFINE_MUTEX(fullstop_mutex);
static atomic_t verbose_sleep_counter;
/*
* Sleep if needed from VERBOSE_TOROUT*().
*/
void verbose_torout_sleep(void)
{
if (verbose_sleep_frequency > 0 &&
verbose_sleep_duration > 0 &&
!(atomic_inc_return(&verbose_sleep_counter) % verbose_sleep_frequency))
schedule_timeout_uninterruptible(verbose_sleep_duration);
}
EXPORT_SYMBOL_GPL(verbose_torout_sleep);
/*
* Schedule a high-resolution-timer sleep in nanoseconds, with a 32-bit
* nanosecond random fuzz. This function and its friends desynchronize
* testing from the timer wheel.
*/
int torture_hrtimeout_ns(ktime_t baset_ns, u32 fuzzt_ns, const enum hrtimer_mode mode,
struct torture_random_state *trsp)
{
ktime_t hto = baset_ns;
if (trsp)
hto += torture_random(trsp) % fuzzt_ns;
set_current_state(TASK_IDLE);
return schedule_hrtimeout(&hto, mode);
}
EXPORT_SYMBOL_GPL(torture_hrtimeout_ns);
/*
* Schedule a high-resolution-timer sleep in microseconds, with a 32-bit
* nanosecond (not microsecond!) random fuzz.
*/
int torture_hrtimeout_us(u32 baset_us, u32 fuzzt_ns, struct torture_random_state *trsp)
{
ktime_t baset_ns = baset_us * NSEC_PER_USEC;
return torture_hrtimeout_ns(baset_ns, fuzzt_ns, HRTIMER_MODE_REL, trsp);
}
EXPORT_SYMBOL_GPL(torture_hrtimeout_us);
/*
* Schedule a high-resolution-timer sleep in milliseconds, with a 32-bit
* microsecond (not millisecond!) random fuzz.
*/
int torture_hrtimeout_ms(u32 baset_ms, u32 fuzzt_us, struct torture_random_state *trsp)
{
ktime_t baset_ns = baset_ms * NSEC_PER_MSEC;
u32 fuzzt_ns;
if ((u32)~0U / NSEC_PER_USEC < fuzzt_us)
fuzzt_ns = (u32)~0U;
else
fuzzt_ns = fuzzt_us * NSEC_PER_USEC;
return torture_hrtimeout_ns(baset_ns, fuzzt_ns, HRTIMER_MODE_REL, trsp);
}
EXPORT_SYMBOL_GPL(torture_hrtimeout_ms);
/*
* Schedule a high-resolution-timer sleep in jiffies, with an
* implied one-jiffy random fuzz. This is intended to replace calls to
* schedule_timeout_interruptible() and friends.
*/
int torture_hrtimeout_jiffies(u32 baset_j, struct torture_random_state *trsp)
{
ktime_t baset_ns = jiffies_to_nsecs(baset_j);
return torture_hrtimeout_ns(baset_ns, jiffies_to_nsecs(1), HRTIMER_MODE_REL, trsp);
}
EXPORT_SYMBOL_GPL(torture_hrtimeout_jiffies);
/*
* Schedule a high-resolution-timer sleep in milliseconds, with a 32-bit
* millisecond (not second!) random fuzz.
*/
int torture_hrtimeout_s(u32 baset_s, u32 fuzzt_ms, struct torture_random_state *trsp)
{
ktime_t baset_ns = baset_s * NSEC_PER_SEC;
u32 fuzzt_ns;
if ((u32)~0U / NSEC_PER_MSEC < fuzzt_ms)
fuzzt_ns = (u32)~0U;
else
fuzzt_ns = fuzzt_ms * NSEC_PER_MSEC;
return torture_hrtimeout_ns(baset_ns, fuzzt_ns, HRTIMER_MODE_REL, trsp);
}
EXPORT_SYMBOL_GPL(torture_hrtimeout_s);
#ifdef CONFIG_HOTPLUG_CPU
/*
* Variables for online-offline handling. Only present if CPU hotplug
* is enabled, otherwise does nothing.
*/
static struct task_struct *onoff_task;
static long onoff_holdoff;
static long onoff_interval;
static torture_ofl_func *onoff_f;
static long n_offline_attempts;
static long n_offline_successes;
static unsigned long sum_offline;
static int min_offline = -1;
static int max_offline;
static long n_online_attempts;
static long n_online_successes;
static unsigned long sum_online;
static int min_online = -1;
static int max_online;
static int torture_online_cpus = NR_CPUS;
/*
* Some torture testing leverages confusion as to the number of online
* CPUs. This function returns the torture-testing view of this number,
* which allows torture tests to load-balance appropriately.
*/
int torture_num_online_cpus(void)
{
return READ_ONCE(torture_online_cpus);
}
EXPORT_SYMBOL_GPL(torture_num_online_cpus);
/*
* Attempt to take a CPU offline. Return false if the CPU is already
* offline or if it is not subject to CPU-hotplug operations. The
* caller can detect other failures by looking at the statistics.
*/
bool torture_offline(int cpu, long *n_offl_attempts, long *n_offl_successes,
unsigned long *sum_offl, int *min_offl, int *max_offl)
{
unsigned long delta;
int ret;
char *s;
unsigned long starttime;
if (!cpu_online(cpu) || !cpu_is_hotpluggable(cpu))
return false;
if (num_online_cpus() <= 1)
return false; /* Can't offline the last CPU. */
if (verbose > 1)
pr_alert("%s" TORTURE_FLAG
"torture_onoff task: offlining %d\n",
torture_type, cpu);
starttime = jiffies;
(*n_offl_attempts)++;
ret = remove_cpu(cpu);
if (ret) {
s = "";
if (!rcu_inkernel_boot_has_ended() && ret == -EBUSY) {
// PCI probe frequently disables hotplug during boot.
(*n_offl_attempts)--;
s = " (-EBUSY forgiven during boot)";
}
if (verbose)
pr_alert("%s" TORTURE_FLAG
"torture_onoff task: offline %d failed%s: errno %d\n",
torture_type, cpu, s, ret);
} else {
if (verbose > 1)
pr_alert("%s" TORTURE_FLAG
"torture_onoff task: offlined %d\n",
torture_type, cpu);
if (onoff_f)
onoff_f();
(*n_offl_successes)++;
delta = jiffies - starttime;
*sum_offl += delta;
if (*min_offl < 0) {
*min_offl = delta;
*max_offl = delta;
}
if (*min_offl > delta)
*min_offl = delta;
if (*max_offl < delta)
*max_offl = delta;
WRITE_ONCE(torture_online_cpus, torture_online_cpus - 1);
WARN_ON_ONCE(torture_online_cpus <= 0);
}
return true;
}
EXPORT_SYMBOL_GPL(torture_offline);
/*
* Attempt to bring a CPU online. Return false if the CPU is already
* online or if it is not subject to CPU-hotplug operations. The
* caller can detect other failures by looking at the statistics.
*/
bool torture_online(int cpu, long *n_onl_attempts, long *n_onl_successes,
unsigned long *sum_onl, int *min_onl, int *max_onl)
{
unsigned long delta;
int ret;
char *s;
unsigned long starttime;
if (cpu_online(cpu) || !cpu_is_hotpluggable(cpu))
return false;
if (verbose > 1)
pr_alert("%s" TORTURE_FLAG
"torture_onoff task: onlining %d\n",
torture_type, cpu);
starttime = jiffies;
(*n_onl_attempts)++;
ret = add_cpu(cpu);
if (ret) {
s = "";
if (!rcu_inkernel_boot_has_ended() && ret == -EBUSY) {
// PCI probe frequently disables hotplug during boot.
(*n_onl_attempts)--;
s = " (-EBUSY forgiven during boot)";
}
if (verbose)
pr_alert("%s" TORTURE_FLAG
"torture_onoff task: online %d failed%s: errno %d\n",
torture_type, cpu, s, ret);
} else {
if (verbose > 1)
pr_alert("%s" TORTURE_FLAG
"torture_onoff task: onlined %d\n",
torture_type, cpu);
(*n_onl_successes)++;
delta = jiffies - starttime;
*sum_onl += delta;
if (*min_onl < 0) {
*min_onl = delta;
*max_onl = delta;
}
if (*min_onl > delta)
*min_onl = delta;
if (*max_onl < delta)
*max_onl = delta;
WRITE_ONCE(torture_online_cpus, torture_online_cpus + 1);
}
return true;
}
EXPORT_SYMBOL_GPL(torture_online);
/*
* Get everything online at the beginning and ends of tests.
*/
static void torture_online_all(char *phase)
{
int cpu;
int ret;
for_each_possible_cpu(cpu) {
if (cpu_online(cpu))
continue;
ret = add_cpu(cpu);
if (ret && verbose) {
pr_alert("%s" TORTURE_FLAG
"%s: %s online %d: errno %d\n",
__func__, phase, torture_type, cpu, ret);
}
}
}
/*
* Execute random CPU-hotplug operations at the interval specified
* by the onoff_interval.
*/
static int
torture_onoff(void *arg)
{
int cpu;
int maxcpu = -1;
DEFINE_TORTURE_RANDOM(rand);
VERBOSE_TOROUT_STRING("torture_onoff task started");
for_each_online_cpu(cpu)
maxcpu = cpu;
WARN_ON(maxcpu < 0);
torture_online_all("Initial");
if (maxcpu == 0) {
VERBOSE_TOROUT_STRING("Only one CPU, so CPU-hotplug testing is disabled");
goto stop;
}
if (onoff_holdoff > 0) {
VERBOSE_TOROUT_STRING("torture_onoff begin holdoff");
torture_hrtimeout_jiffies(onoff_holdoff, &rand);
VERBOSE_TOROUT_STRING("torture_onoff end holdoff");
}
while (!torture_must_stop()) {
if (disable_onoff_at_boot && !rcu_inkernel_boot_has_ended()) {
torture_hrtimeout_jiffies(HZ / 10, &rand);
continue;
}
cpu = torture_random(&rand) % (maxcpu + 1);
if (!torture_offline(cpu,
&n_offline_attempts, &n_offline_successes,
&sum_offline, &min_offline, &max_offline))
torture_online(cpu,
&n_online_attempts, &n_online_successes,
&sum_online, &min_online, &max_online);
torture_hrtimeout_jiffies(onoff_interval, &rand);
}
stop:
torture_kthread_stopping("torture_onoff");
torture_online_all("Final");
return 0;
}
#endif /* #ifdef CONFIG_HOTPLUG_CPU */
/*
* Initiate online-offline handling.
*/
int torture_onoff_init(long ooholdoff, long oointerval, torture_ofl_func *f)
{
#ifdef CONFIG_HOTPLUG_CPU
onoff_holdoff = ooholdoff;
onoff_interval = oointerval;
onoff_f = f;
if (onoff_interval <= 0)
return 0;
return torture_create_kthread(torture_onoff, NULL, onoff_task);
#else /* #ifdef CONFIG_HOTPLUG_CPU */
return 0;
#endif /* #else #ifdef CONFIG_HOTPLUG_CPU */
}
EXPORT_SYMBOL_GPL(torture_onoff_init);
/*
* Clean up after online/offline testing.
*/
static void torture_onoff_cleanup(void)
{
#ifdef CONFIG_HOTPLUG_CPU
if (onoff_task == NULL)
return;
VERBOSE_TOROUT_STRING("Stopping torture_onoff task");
kthread_stop(onoff_task);
onoff_task = NULL;
#endif /* #ifdef CONFIG_HOTPLUG_CPU */
}
/*
* Print online/offline testing statistics.
*/
void torture_onoff_stats(void)
{
#ifdef CONFIG_HOTPLUG_CPU
pr_cont("onoff: %ld/%ld:%ld/%ld %d,%d:%d,%d %lu:%lu (HZ=%d) ",
n_online_successes, n_online_attempts,
n_offline_successes, n_offline_attempts,
min_online, max_online,
min_offline, max_offline,
sum_online, sum_offline, HZ);
#endif /* #ifdef CONFIG_HOTPLUG_CPU */
}
EXPORT_SYMBOL_GPL(torture_onoff_stats);
/*
* Were all the online/offline operations successful?
*/
bool torture_onoff_failures(void)
{
#ifdef CONFIG_HOTPLUG_CPU
return n_online_successes != n_online_attempts ||
n_offline_successes != n_offline_attempts;
#else /* #ifdef CONFIG_HOTPLUG_CPU */
return false;
#endif /* #else #ifdef CONFIG_HOTPLUG_CPU */
}
EXPORT_SYMBOL_GPL(torture_onoff_failures);
#define TORTURE_RANDOM_MULT 39916801 /* prime */
#define TORTURE_RANDOM_ADD 479001701 /* prime */
#define TORTURE_RANDOM_REFRESH 10000
/*
* Crude but fast random-number generator. Uses a linear congruential
* generator, with occasional help from cpu_clock().
*/
unsigned long
torture_random(struct torture_random_state *trsp)
{
if (--trsp->trs_count < 0) {
trsp->trs_state += (unsigned long)local_clock() + raw_smp_processor_id();
trsp->trs_count = TORTURE_RANDOM_REFRESH;
}
trsp->trs_state = trsp->trs_state * TORTURE_RANDOM_MULT +
TORTURE_RANDOM_ADD;
return swahw32(trsp->trs_state);
}
EXPORT_SYMBOL_GPL(torture_random);
/*
* Variables for shuffling. The idea is to ensure that each CPU stays
* idle for an extended period to test interactions with dyntick idle,
* as well as interactions with any per-CPU variables.
*/
struct shuffle_task {
struct list_head st_l;
struct task_struct *st_t;
};
static long shuffle_interval; /* In jiffies. */
static struct task_struct *shuffler_task;
static cpumask_var_t shuffle_tmp_mask;
static int shuffle_idle_cpu; /* Force all torture tasks off this CPU */
static struct list_head shuffle_task_list = LIST_HEAD_INIT(shuffle_task_list);
static DEFINE_MUTEX(shuffle_task_mutex);
/*
* Register a task to be shuffled. If there is no memory, just splat
* and don't bother registering.
*/
void torture_shuffle_task_register(struct task_struct *tp)
{
struct shuffle_task *stp;
if (WARN_ON_ONCE(tp == NULL))
return;
stp = kmalloc(sizeof(*stp), GFP_KERNEL);
if (WARN_ON_ONCE(stp == NULL))
return;
stp->st_t = tp;
mutex_lock(&shuffle_task_mutex);
list_add(&stp->st_l, &shuffle_task_list);
mutex_unlock(&shuffle_task_mutex);
}
EXPORT_SYMBOL_GPL(torture_shuffle_task_register);
/*
* Unregister all tasks, for example, at the end of the torture run.
*/
static void torture_shuffle_task_unregister_all(void)
{
struct shuffle_task *stp;
struct shuffle_task *p;
mutex_lock(&shuffle_task_mutex);
list_for_each_entry_safe(stp, p, &shuffle_task_list, st_l) {
list_del(&stp->st_l);
kfree(stp);
}
mutex_unlock(&shuffle_task_mutex);
}
/* Shuffle tasks such that we allow shuffle_idle_cpu to become idle.
* A special case is when shuffle_idle_cpu = -1, in which case we allow
* the tasks to run on all CPUs.
*/
static void torture_shuffle_tasks(void)
{
DEFINE_TORTURE_RANDOM(rand);
struct shuffle_task *stp;
cpumask_setall(shuffle_tmp_mask);
cpus_read_lock();
/* No point in shuffling if there is only one online CPU (ex: UP) */
if (num_online_cpus() == 1) {
cpus_read_unlock();
return;
}
/* Advance to the next CPU. Upon overflow, don't idle any CPUs. */
shuffle_idle_cpu = cpumask_next(shuffle_idle_cpu, shuffle_tmp_mask);
if (shuffle_idle_cpu >= nr_cpu_ids)
shuffle_idle_cpu = -1;
else
cpumask_clear_cpu(shuffle_idle_cpu, shuffle_tmp_mask);
mutex_lock(&shuffle_task_mutex);
list_for_each_entry(stp, &shuffle_task_list, st_l) {
if (!random_shuffle || torture_random(&rand) & 0x1)
set_cpus_allowed_ptr(stp->st_t, shuffle_tmp_mask);
}
mutex_unlock(&shuffle_task_mutex);
cpus_read_unlock();
}
/* Shuffle tasks across CPUs, with the intent of allowing each CPU in the
* system to become idle at a time and cut off its timer ticks. This is meant
* to test the support for such tickless idle CPU in RCU.
*/
static int torture_shuffle(void *arg)
{
DEFINE_TORTURE_RANDOM(rand);
VERBOSE_TOROUT_STRING("torture_shuffle task started");
do {
torture_hrtimeout_jiffies(shuffle_interval, &rand);
torture_shuffle_tasks();
torture_shutdown_absorb("torture_shuffle");
} while (!torture_must_stop());
torture_kthread_stopping("torture_shuffle");
return 0;
}
/*
* Start the shuffler, with shuffint in jiffies.
*/
int torture_shuffle_init(long shuffint)
{
shuffle_interval = shuffint;
shuffle_idle_cpu = -1;
if (!alloc_cpumask_var(&shuffle_tmp_mask, GFP_KERNEL)) {
TOROUT_ERRSTRING("Failed to alloc mask");
return -ENOMEM;
}
/* Create the shuffler thread */
return torture_create_kthread(torture_shuffle, NULL, shuffler_task);
}
EXPORT_SYMBOL_GPL(torture_shuffle_init);
/*
* Stop the shuffling.
*/
static void torture_shuffle_cleanup(void)
{
torture_shuffle_task_unregister_all();
if (shuffler_task) {
VERBOSE_TOROUT_STRING("Stopping torture_shuffle task");
kthread_stop(shuffler_task);
free_cpumask_var(shuffle_tmp_mask);
}
shuffler_task = NULL;
}
/*
* Variables for auto-shutdown. This allows "lights out" torture runs
* to be fully scripted.
*/
static struct task_struct *shutdown_task;
static ktime_t shutdown_time; /* time to system shutdown. */
static void (*torture_shutdown_hook)(void);
/*
* Absorb kthreads into a kernel function that won't return, so that
* they won't ever access module text or data again.
*/
void torture_shutdown_absorb(const char *title)
{
while (READ_ONCE(fullstop) == FULLSTOP_SHUTDOWN) {
pr_notice("torture thread %s parking due to system shutdown\n",
title);
schedule_timeout_uninterruptible(MAX_SCHEDULE_TIMEOUT);
}
}
EXPORT_SYMBOL_GPL(torture_shutdown_absorb);
/*
* Cause the torture test to shutdown the system after the test has
* run for the time specified by the shutdown_secs parameter.
*/
static int torture_shutdown(void *arg)
{
ktime_t ktime_snap;
VERBOSE_TOROUT_STRING("torture_shutdown task started");
ktime_snap = ktime_get();
while (ktime_before(ktime_snap, shutdown_time) &&
!torture_must_stop()) {
if (verbose)
pr_alert("%s" TORTURE_FLAG
"torture_shutdown task: %llu ms remaining\n",
torture_type,
ktime_ms_delta(shutdown_time, ktime_snap));
set_current_state(TASK_INTERRUPTIBLE);
schedule_hrtimeout(&shutdown_time, HRTIMER_MODE_ABS);
ktime_snap = ktime_get();
}
if (torture_must_stop()) {
torture_kthread_stopping("torture_shutdown");
return 0;
}
/* OK, shut down the system. */
VERBOSE_TOROUT_STRING("torture_shutdown task shutting down system");
shutdown_task = NULL; /* Avoid self-kill deadlock. */
if (torture_shutdown_hook)
torture_shutdown_hook();
else
VERBOSE_TOROUT_STRING("No torture_shutdown_hook(), skipping.");
if (ftrace_dump_at_shutdown)
rcu_ftrace_dump(DUMP_ALL);
kernel_power_off(); /* Shut down the system. */
return 0;
}
/*
* Start up the shutdown task.
*/
int torture_shutdown_init(int ssecs, void (*cleanup)(void))
{
torture_shutdown_hook = cleanup;
if (ssecs > 0) {
shutdown_time = ktime_add(ktime_get(), ktime_set(ssecs, 0));
return torture_create_kthread(torture_shutdown, NULL,
shutdown_task);
}
return 0;
}
EXPORT_SYMBOL_GPL(torture_shutdown_init);
/*
* Detect and respond to a system shutdown.
*/
static int torture_shutdown_notify(struct notifier_block *unused1,
unsigned long unused2, void *unused3)
{
mutex_lock(&fullstop_mutex);
if (READ_ONCE(fullstop) == FULLSTOP_DONTSTOP) {
VERBOSE_TOROUT_STRING("Unscheduled system shutdown detected");
WRITE_ONCE(fullstop, FULLSTOP_SHUTDOWN);
} else {
pr_warn("Concurrent rmmod and shutdown illegal!\n");
}
mutex_unlock(&fullstop_mutex);
return NOTIFY_DONE;
}
static struct notifier_block torture_shutdown_nb = {
.notifier_call = torture_shutdown_notify,
};
/*
* Shut down the shutdown task. Say what??? Heh! This can happen if
* the torture module gets an rmmod before the shutdown time arrives. ;-)
*/
static void torture_shutdown_cleanup(void)
{
unregister_reboot_notifier(&torture_shutdown_nb);
if (shutdown_task != NULL) {
VERBOSE_TOROUT_STRING("Stopping torture_shutdown task");
kthread_stop(shutdown_task);
}
shutdown_task = NULL;
}
/*
* Variables for stuttering, which means to periodically pause and
* restart testing in order to catch bugs that appear when load is
* suddenly applied to or removed from the system.
*/
static struct task_struct *stutter_task;
rcutorture: Fix stuttering races and other issues [ Upstream commit cca42bd8eb1b54a4c9bbf48c79d120e66619a3e4 ] The stuttering code isn't functioning as expected. Ideally, it should pause the torture threads for a designated period before resuming. Yet, it fails to halt the test for the correct duration. Additionally, a race condition exists, potentially causing the stuttering code to pause for an extended period if the 'spt' variable is non-zero due to the stutter orchestration thread's inadequate CPU time. Moreover, over-stuttering can hinder RCU's progress on TREE07 kernels. This happens as the stuttering code may run within a softirq due to RCU callbacks. Consequently, ksoftirqd keeps a CPU busy for several seconds, thus obstructing RCU's progress. This situation triggers a warning message in the logs: [ 2169.481783] rcu_torture_writer: rtort_pipe_count: 9 This warning suggests that an RCU torture object, although invisible to RCU readers, couldn't make it past the pipe array and be freed -- a strong indication that there weren't enough grace periods during the stutter interval. To address these issues, this patch sets the "stutter end" time to an absolute point in the future set by the main stutter thread. This is then used for waiting in stutter_wait(). While the stutter thread still defines this absolute time, the waiters' waiting logic doesn't rely on the stutter thread receiving sufficient CPU time to halt the stuttering as the halting is now self-controlled. Cc: stable@vger.kernel.org Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org> Signed-off-by: Paul E. McKenney <paulmck@kernel.org> Signed-off-by: Frederic Weisbecker <frederic@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
2023-07-29 22:27:31 +08:00
static ktime_t stutter_till_abs_time;
static int stutter;
static int stutter_gap;
/*
* Block until the stutter interval ends. This must be called periodically
* by all running kthreads that need to be subject to stuttering.
*/
bool stutter_wait(const char *title)
{
bool ret = false;
rcutorture: Fix stuttering races and other issues [ Upstream commit cca42bd8eb1b54a4c9bbf48c79d120e66619a3e4 ] The stuttering code isn't functioning as expected. Ideally, it should pause the torture threads for a designated period before resuming. Yet, it fails to halt the test for the correct duration. Additionally, a race condition exists, potentially causing the stuttering code to pause for an extended period if the 'spt' variable is non-zero due to the stutter orchestration thread's inadequate CPU time. Moreover, over-stuttering can hinder RCU's progress on TREE07 kernels. This happens as the stuttering code may run within a softirq due to RCU callbacks. Consequently, ksoftirqd keeps a CPU busy for several seconds, thus obstructing RCU's progress. This situation triggers a warning message in the logs: [ 2169.481783] rcu_torture_writer: rtort_pipe_count: 9 This warning suggests that an RCU torture object, although invisible to RCU readers, couldn't make it past the pipe array and be freed -- a strong indication that there weren't enough grace periods during the stutter interval. To address these issues, this patch sets the "stutter end" time to an absolute point in the future set by the main stutter thread. This is then used for waiting in stutter_wait(). While the stutter thread still defines this absolute time, the waiters' waiting logic doesn't rely on the stutter thread receiving sufficient CPU time to halt the stuttering as the halting is now self-controlled. Cc: stable@vger.kernel.org Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org> Signed-off-by: Paul E. McKenney <paulmck@kernel.org> Signed-off-by: Frederic Weisbecker <frederic@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
2023-07-29 22:27:31 +08:00
ktime_t till_ns;
rcu: Rename cond_resched_rcu_qs() to cond_resched_tasks_rcu_qs() Commit e31d28b6ab8f ("trace: Eliminate cond_resched_rcu_qs() in favor of cond_resched()") substituted cond_resched() for the earlier call to cond_resched_rcu_qs(). However, the new-age cond_resched() does not do anything to help RCU-tasks grace periods because (1) RCU-tasks is only enabled when CONFIG_PREEMPT=y and (2) cond_resched() is a complete no-op when preemption is enabled. This situation results in hangs when running the trace benchmarks. A number of potential fixes were discussed on LKML (https://lkml.kernel.org/r/20180224151240.0d63a059@vmware.local.home), including making cond_resched() not be a no-op; making cond_resched() not be a no-op, but only when running tracing benchmarks; reverting the aforementioned commit (which works because cond_resched_rcu_qs() does provide an RCU-tasks quiescent state; and adding a call to the scheduler/RCU rcu_note_voluntary_context_switch() function. All were deemed unsatisfactory, either due to added cond_resched() overhead or due to magic functions inviting cargo culting. This commit renames cond_resched_rcu_qs() to cond_resched_tasks_rcu_qs(), which provides a clear hint as to what this function is doing and why and where it should be used, and then replaces the call to cond_resched() with cond_resched_tasks_rcu_qs() in the trace benchmark's benchmark_event_kthread() function. Reported-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com> Tested-by: Nicholas Piggin <npiggin@gmail.com>
2018-03-03 08:35:27 +08:00
cond_resched_tasks_rcu_qs();
rcutorture: Fix stuttering races and other issues [ Upstream commit cca42bd8eb1b54a4c9bbf48c79d120e66619a3e4 ] The stuttering code isn't functioning as expected. Ideally, it should pause the torture threads for a designated period before resuming. Yet, it fails to halt the test for the correct duration. Additionally, a race condition exists, potentially causing the stuttering code to pause for an extended period if the 'spt' variable is non-zero due to the stutter orchestration thread's inadequate CPU time. Moreover, over-stuttering can hinder RCU's progress on TREE07 kernels. This happens as the stuttering code may run within a softirq due to RCU callbacks. Consequently, ksoftirqd keeps a CPU busy for several seconds, thus obstructing RCU's progress. This situation triggers a warning message in the logs: [ 2169.481783] rcu_torture_writer: rtort_pipe_count: 9 This warning suggests that an RCU torture object, although invisible to RCU readers, couldn't make it past the pipe array and be freed -- a strong indication that there weren't enough grace periods during the stutter interval. To address these issues, this patch sets the "stutter end" time to an absolute point in the future set by the main stutter thread. This is then used for waiting in stutter_wait(). While the stutter thread still defines this absolute time, the waiters' waiting logic doesn't rely on the stutter thread receiving sufficient CPU time to halt the stuttering as the halting is now self-controlled. Cc: stable@vger.kernel.org Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org> Signed-off-by: Paul E. McKenney <paulmck@kernel.org> Signed-off-by: Frederic Weisbecker <frederic@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
2023-07-29 22:27:31 +08:00
till_ns = READ_ONCE(stutter_till_abs_time);
if (till_ns && ktime_before(ktime_get(), till_ns)) {
torture_hrtimeout_ns(till_ns, 0, HRTIMER_MODE_ABS, NULL);
ret = true;
}
rcutorture: Fix stuttering races and other issues [ Upstream commit cca42bd8eb1b54a4c9bbf48c79d120e66619a3e4 ] The stuttering code isn't functioning as expected. Ideally, it should pause the torture threads for a designated period before resuming. Yet, it fails to halt the test for the correct duration. Additionally, a race condition exists, potentially causing the stuttering code to pause for an extended period if the 'spt' variable is non-zero due to the stutter orchestration thread's inadequate CPU time. Moreover, over-stuttering can hinder RCU's progress on TREE07 kernels. This happens as the stuttering code may run within a softirq due to RCU callbacks. Consequently, ksoftirqd keeps a CPU busy for several seconds, thus obstructing RCU's progress. This situation triggers a warning message in the logs: [ 2169.481783] rcu_torture_writer: rtort_pipe_count: 9 This warning suggests that an RCU torture object, although invisible to RCU readers, couldn't make it past the pipe array and be freed -- a strong indication that there weren't enough grace periods during the stutter interval. To address these issues, this patch sets the "stutter end" time to an absolute point in the future set by the main stutter thread. This is then used for waiting in stutter_wait(). While the stutter thread still defines this absolute time, the waiters' waiting logic doesn't rely on the stutter thread receiving sufficient CPU time to halt the stuttering as the halting is now self-controlled. Cc: stable@vger.kernel.org Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org> Signed-off-by: Paul E. McKenney <paulmck@kernel.org> Signed-off-by: Frederic Weisbecker <frederic@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
2023-07-29 22:27:31 +08:00
torture_shutdown_absorb(title);
return ret;
}
EXPORT_SYMBOL_GPL(stutter_wait);
/*
* Cause the torture test to "stutter", starting and stopping all
* threads periodically.
*/
static int torture_stutter(void *arg)
{
rcutorture: Fix stuttering races and other issues [ Upstream commit cca42bd8eb1b54a4c9bbf48c79d120e66619a3e4 ] The stuttering code isn't functioning as expected. Ideally, it should pause the torture threads for a designated period before resuming. Yet, it fails to halt the test for the correct duration. Additionally, a race condition exists, potentially causing the stuttering code to pause for an extended period if the 'spt' variable is non-zero due to the stutter orchestration thread's inadequate CPU time. Moreover, over-stuttering can hinder RCU's progress on TREE07 kernels. This happens as the stuttering code may run within a softirq due to RCU callbacks. Consequently, ksoftirqd keeps a CPU busy for several seconds, thus obstructing RCU's progress. This situation triggers a warning message in the logs: [ 2169.481783] rcu_torture_writer: rtort_pipe_count: 9 This warning suggests that an RCU torture object, although invisible to RCU readers, couldn't make it past the pipe array and be freed -- a strong indication that there weren't enough grace periods during the stutter interval. To address these issues, this patch sets the "stutter end" time to an absolute point in the future set by the main stutter thread. This is then used for waiting in stutter_wait(). While the stutter thread still defines this absolute time, the waiters' waiting logic doesn't rely on the stutter thread receiving sufficient CPU time to halt the stuttering as the halting is now self-controlled. Cc: stable@vger.kernel.org Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org> Signed-off-by: Paul E. McKenney <paulmck@kernel.org> Signed-off-by: Frederic Weisbecker <frederic@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
2023-07-29 22:27:31 +08:00
ktime_t till_ns;
VERBOSE_TOROUT_STRING("torture_stutter task started");
do {
if (!torture_must_stop() && stutter > 1) {
rcutorture: Fix stuttering races and other issues [ Upstream commit cca42bd8eb1b54a4c9bbf48c79d120e66619a3e4 ] The stuttering code isn't functioning as expected. Ideally, it should pause the torture threads for a designated period before resuming. Yet, it fails to halt the test for the correct duration. Additionally, a race condition exists, potentially causing the stuttering code to pause for an extended period if the 'spt' variable is non-zero due to the stutter orchestration thread's inadequate CPU time. Moreover, over-stuttering can hinder RCU's progress on TREE07 kernels. This happens as the stuttering code may run within a softirq due to RCU callbacks. Consequently, ksoftirqd keeps a CPU busy for several seconds, thus obstructing RCU's progress. This situation triggers a warning message in the logs: [ 2169.481783] rcu_torture_writer: rtort_pipe_count: 9 This warning suggests that an RCU torture object, although invisible to RCU readers, couldn't make it past the pipe array and be freed -- a strong indication that there weren't enough grace periods during the stutter interval. To address these issues, this patch sets the "stutter end" time to an absolute point in the future set by the main stutter thread. This is then used for waiting in stutter_wait(). While the stutter thread still defines this absolute time, the waiters' waiting logic doesn't rely on the stutter thread receiving sufficient CPU time to halt the stuttering as the halting is now self-controlled. Cc: stable@vger.kernel.org Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org> Signed-off-by: Paul E. McKenney <paulmck@kernel.org> Signed-off-by: Frederic Weisbecker <frederic@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
2023-07-29 22:27:31 +08:00
till_ns = ktime_add_ns(ktime_get(),
jiffies_to_nsecs(stutter));
WRITE_ONCE(stutter_till_abs_time, till_ns);
torture_hrtimeout_jiffies(stutter - 1, NULL);
}
if (!torture_must_stop())
torture_hrtimeout_jiffies(stutter_gap, NULL);
torture_shutdown_absorb("torture_stutter");
} while (!torture_must_stop());
torture_kthread_stopping("torture_stutter");
return 0;
}
/*
* Initialize and kick off the torture_stutter kthread.
*/
int torture_stutter_init(const int s, const int sgap)
{
stutter = s;
stutter_gap = sgap;
return torture_create_kthread(torture_stutter, NULL, stutter_task);
}
EXPORT_SYMBOL_GPL(torture_stutter_init);
/*
* Cleanup after the torture_stutter kthread.
*/
static void torture_stutter_cleanup(void)
{
if (!stutter_task)
return;
VERBOSE_TOROUT_STRING("Stopping torture_stutter task");
kthread_stop(stutter_task);
stutter_task = NULL;
}
/*
* Initialize torture module. Please note that this is -not- invoked via
* the usual module_init() mechanism, but rather by an explicit call from
* the client torture module. This call must be paired with a later
* torture_init_end().
*
* The runnable parameter points to a flag that controls whether or not
* the test is currently runnable. If there is no such flag, pass in NULL.
*/
bool torture_init_begin(char *ttype, int v)
{
mutex_lock(&fullstop_mutex);
if (torture_type != NULL) {
pr_alert("%s: Refusing %s init: %s running.\n",
__func__, ttype, torture_type);
pr_alert("%s: One torture test at a time!\n", __func__);
mutex_unlock(&fullstop_mutex);
return false;
}
torture_type = ttype;
verbose = v;
fullstop = FULLSTOP_DONTSTOP;
return true;
}
EXPORT_SYMBOL_GPL(torture_init_begin);
/*
* Tell the torture module that initialization is complete.
*/
torture: Remove __init from torture_init_begin/end Loading rcutorture as a module (as opposed to building it directly into the kernel) results in the following splat: [Wed Apr 16 15:29:33 2014] BUG: unable to handle kernel paging request at ffffffffa0003000 [Wed Apr 16 15:29:33 2014] IP: [<ffffffffa0003000>] 0xffffffffa0003000 [Wed Apr 16 15:29:33 2014] PGD 1c0f067 PUD 1c10063 PMD 378a6067 PTE 0 [Wed Apr 16 15:29:33 2014] Oops: 0010 [#1] SMP [Wed Apr 16 15:29:33 2014] Modules linked in: rcutorture(+) torture [Wed Apr 16 15:29:33 2014] CPU: 0 PID: 4257 Comm: modprobe Not tainted 3.15.0-rc1 #10 [Wed Apr 16 15:29:33 2014] Hardware name: innotek GmbH VirtualBox, BIOS VirtualBox 12/01/2006 [Wed Apr 16 15:29:33 2014] task: ffff8800db1e88d0 ti: ffff8800db25c000 task.ti: ffff8800db25c000 [Wed Apr 16 15:29:33 2014] RIP: 0010:[<ffffffffa0003000>] [<ffffffffa0003000>] 0xffffffffa0003000 [Wed Apr 16 15:29:33 2014] RSP: 0018:ffff8800db25dca0 EFLAGS: 00010282 [Wed Apr 16 15:29:33 2014] RAX: 0000000000000000 RBX: 0000000000000000 RCX: 0000000000000000 [Wed Apr 16 15:29:33 2014] RDX: ffffffffa00090a8 RSI: 0000000000000001 RDI: ffffffffa0008337 [Wed Apr 16 15:29:33 2014] RBP: ffff8800db25dd50 R08: 0000000000000000 R09: 0000000000000000 [Wed Apr 16 15:29:33 2014] R10: ffffea000357b680 R11: ffffffff8113257a R12: ffffffffa000d000 [Wed Apr 16 15:29:33 2014] R13: ffffffffa00094c0 R14: ffffffffa0009510 R15: 0000000000000001 [Wed Apr 16 15:29:33 2014] FS: 00007fee30ce5700(0000) GS:ffff88021fc00000(0000) knlGS:0000000000000000 [Wed Apr 16 15:29:33 2014] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b [Wed Apr 16 15:29:33 2014] CR2: ffffffffa0003000 CR3: 00000000d5eb1000 CR4: 00000000000006f0 [Wed Apr 16 15:29:33 2014] Stack: [Wed Apr 16 15:29:33 2014] ffffffffa000d02c 0000000000000000 ffff88021700d400 0000000000000000 [Wed Apr 16 15:29:33 2014] ffff8800db25dd40 ffffffff81647951 ffff8802162bd000 ffff88021541846c [Wed Apr 16 15:29:33 2014] 0000000000000000 ffffffff817dbe2d ffffffff817dbe2d 0000000000000001 [Wed Apr 16 15:29:33 2014] Call Trace: [Wed Apr 16 15:29:33 2014] [<ffffffffa000d02c>] ? rcu_torture_init+0x2c/0x8b4 [rcutorture] [Wed Apr 16 15:29:33 2014] [<ffffffff81647951>] ? netlink_broadcast_filtered+0x121/0x3a0 [Wed Apr 16 15:29:33 2014] [<ffffffff817dbe2d>] ? mutex_lock+0xd/0x2a [Wed Apr 16 15:29:33 2014] [<ffffffff817dbe2d>] ? mutex_lock+0xd/0x2a [Wed Apr 16 15:29:33 2014] [<ffffffff810e7022>] ? trace_module_notify+0x62/0x1d0 [Wed Apr 16 15:29:33 2014] [<ffffffffa000d000>] ? 0xffffffffa000cfff [Wed Apr 16 15:29:33 2014] [<ffffffff8100034a>] do_one_initcall+0xfa/0x140 [Wed Apr 16 15:29:33 2014] [<ffffffff8106b4ce>] ? __blocking_notifier_call_chain+0x5e/0x80 [Wed Apr 16 15:29:33 2014] [<ffffffff810b3481>] load_module+0x1931/0x21b0 [Wed Apr 16 15:29:33 2014] [<ffffffff810b0330>] ? show_initstate+0x50/0x50 [Wed Apr 16 15:29:33 2014] [<ffffffff810b3d9e>] SyS_init_module+0x9e/0xc0 [Wed Apr 16 15:29:33 2014] [<ffffffff817e4c22>] system_call_fastpath+0x16/0x1b [Wed Apr 16 15:29:33 2014] Code: Bad RIP value. [Wed Apr 16 15:29:33 2014] RIP [<ffffffffa0003000>] 0xffffffffa0003000 [Wed Apr 16 15:29:33 2014] RSP <ffff8800db25dca0> [Wed Apr 16 15:29:33 2014] CR2: ffffffffa0003000 [Wed Apr 16 15:29:33 2014] ---[ end trace 3e88c173037af84b ]--- This splat is due to the fact that torture_init_begin() and torture_init_end() are both marked with __init, despite their use at runtime. This commit therefore removes __init from both functions. Signed-off-by: Pranith Kumar <bobby.prani@gmail.com> Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com> Reviewed-by: Josh Triplett <josh@joshtriplett.org>
2014-04-17 04:42:09 +08:00
void torture_init_end(void)
{
mutex_unlock(&fullstop_mutex);
register_reboot_notifier(&torture_shutdown_nb);
}
EXPORT_SYMBOL_GPL(torture_init_end);
/*
* Clean up torture module. Please note that this is -not- invoked via
* the usual module_exit() mechanism, but rather by an explicit call from
* the client torture module. Returns true if a race with system shutdown
* is detected, otherwise, all kthreads started by functions in this file
* will be shut down.
*
* This must be called before the caller starts shutting down its own
* kthreads.
*
* Both torture_cleanup_begin() and torture_cleanup_end() must be paired,
* in order to correctly perform the cleanup. They are separated because
* threads can still need to reference the torture_type type, thus nullify
* only after completing all other relevant calls.
*/
bool torture_cleanup_begin(void)
{
mutex_lock(&fullstop_mutex);
if (READ_ONCE(fullstop) == FULLSTOP_SHUTDOWN) {
pr_warn("Concurrent rmmod and shutdown illegal!\n");
mutex_unlock(&fullstop_mutex);
schedule_timeout_uninterruptible(10);
return true;
}
WRITE_ONCE(fullstop, FULLSTOP_RMMOD);
mutex_unlock(&fullstop_mutex);
torture_shutdown_cleanup();
torture_shuffle_cleanup();
torture_stutter_cleanup();
torture_onoff_cleanup();
return false;
}
EXPORT_SYMBOL_GPL(torture_cleanup_begin);
void torture_cleanup_end(void)
{
mutex_lock(&fullstop_mutex);
torture_type = NULL;
mutex_unlock(&fullstop_mutex);
}
EXPORT_SYMBOL_GPL(torture_cleanup_end);
/*
* Is it time for the current torture test to stop?
*/
bool torture_must_stop(void)
{
return torture_must_stop_irq() || kthread_should_stop();
}
EXPORT_SYMBOL_GPL(torture_must_stop);
/*
* Is it time for the current torture test to stop? This is the irq-safe
* version, hence no check for kthread_should_stop().
*/
bool torture_must_stop_irq(void)
{
return READ_ONCE(fullstop) != FULLSTOP_DONTSTOP;
}
EXPORT_SYMBOL_GPL(torture_must_stop_irq);
/*
* Each kthread must wait for kthread_should_stop() before returning from
* its top-level function, otherwise segfaults ensue. This function
* prints a "stopping" message and waits for kthread_should_stop(), and
* should be called from all torture kthreads immediately prior to
* returning.
*/
void torture_kthread_stopping(char *title)
{
char buf[128];
snprintf(buf, sizeof(buf), "%s is stopping", title);
VERBOSE_TOROUT_STRING(buf);
while (!kthread_should_stop()) {
torture_shutdown_absorb(title);
torture: Fix hang during kthread shutdown phase During rcutorture shutdown, the rcu_torture_cleanup() function calls torture_cleanup_begin(), which sets the fullstop global variable to FULLSTOP_RMMOD. This causes the rcutorture threads for readers and fakewriters to exit all of their "while" loops and start shutting down. They then call torture_kthread_stopping(), which in turn waits for kthread_stop() to be called. However, rcu_torture_cleanup() has not yet called kthread_stop() on those threads, and before it gets a chance to do so, multiple instances of torture_kthread_stopping() invoke schedule_timeout_interruptible(1) in a tight loop. Tracing confirms that TIMER_SOFTIRQ can then continuously execute timer callbacks. If that TIMER_SOFTIRQ preempts the task executing rcu_torture_cleanup(), that task might never invoke kthread_stop(). This commit improves this situation by increasing the timeout passed to schedule_timeout_interruptible() from one jiffy to 1/20th of a second. This change prevents TIMER_SOFTIRQ from monopolizing its CPU, thus allowing rcu_torture_cleanup() to carry out the needed kthread_stop() invocations. Testing has shown 100 runs of TREE07 passing reliably, as oppose to the tens-of-percent failure rates seen beforehand. Cc: Paul McKenney <paulmck@kernel.org> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Zhouyi Zhou <zhouzhouyi@gmail.com> Cc: <stable@vger.kernel.org> # 6.0.x Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org> Tested-by: Zhouyi Zhou <zhouzhouyi@gmail.com> Reviewed-by: Davidlohr Bueso <dave@stgolabs.net> Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
2023-01-01 14:15:55 +08:00
schedule_timeout_uninterruptible(HZ / 20);
}
}
EXPORT_SYMBOL_GPL(torture_kthread_stopping);
/*
* Create a generic torture kthread that is immediately runnable. If you
* need the kthread to be stopped so that you can do something to it before
* it starts, you will need to open-code your own.
*/
int _torture_create_kthread(int (*fn)(void *arg), void *arg, char *s, char *m,
char *f, struct task_struct **tp, void (*cbf)(struct task_struct *tp))
{
int ret = 0;
VERBOSE_TOROUT_STRING(m);
*tp = kthread_create(fn, arg, "%s", s);
if (IS_ERR(*tp)) {
ret = PTR_ERR(*tp);
TOROUT_ERRSTRING(f);
*tp = NULL;
return ret;
}
if (cbf)
cbf(*tp);
wake_up_process(*tp); // Process is sleeping, so ordering provided.
torture_shuffle_task_register(*tp);
return ret;
}
EXPORT_SYMBOL_GPL(_torture_create_kthread);
/*
* Stop a generic kthread, emitting a message.
*/
void _torture_stop_kthread(char *m, struct task_struct **tp)
{
if (*tp == NULL)
return;
VERBOSE_TOROUT_STRING(m);
kthread_stop(*tp);
*tp = NULL;
}
EXPORT_SYMBOL_GPL(_torture_stop_kthread);