Linux Kernel Programming: interruptible_sleep_on vs wait_event_interruptible - linux

I am writing a linux device driver (on linux kernel version 5.10.11-v7l+) and I am using O'Reilly's Linux Device Drivers, second edition. In the section on blocking I/O, it explains the difference between interruptible_sleep_on and wait_event_interruptible and how the former is interruptible by a signal and the latter is interruptible by evaluating a condition. For my driver, I have an ISR that signals to a thread that the device is ready to be read from and it seems to me that interruptible_sleep_on would be preferable as it just sends a signal to the wait queue and doesn't need a condition variable to be set by the ISR and then cleared by the thread (which, at least it seems to me, would require another access control method for that variable).
But, interruptible_sleep_on is no longer in the kernel (at least in my version) and wait_event_interruptible is the only option. My question is: is there a way to duplicate the behavior of interruptible_sleep_on with wait_event_interruptible and essentially eschew using a condition variable and just rely on the signal being pushed to the wait queue?
Relevant code:
static int data_ready = 0;
static irqreturn_t irq_handler(int irq, void* dev_id)
{
static unsigned long flags = 0;
local_irq_restore(flags);
data_ready = 1;
wake_up_interruptible(&dataint);
return IRQ_HANDLED;
}
static int data_thread_fn(void* param)
{
allow_signal(SIGKILL);
while(!kthread_should_stop())
{
wait_event_interruptible(dataint, data_ready == 1);
data_lock(); //lock to share with main thread
data_buffer(); //read data from device
data_unlock();
wake_up_interruptible(&databuffer); //tell poll that data is ready
data_ready = 0;
if (signal_pending(data_thread))
break;
}
do_exit(0);
return 0;
}

Related

Does poll/epoll handling is in interrupt context?

This is probably trivial question for some people, but somehow I'm not sure about it.
When waiting with poll for event from kernel, is it that the handling of new event is done in interrupt context ?
If not, does it mean we can sleep/wait (using other commands in handler) in the handler ?
int main (void)
{
struct pollfd fds[2];
int ret;
fds[0].fd = FILENO;
fds[0].events = POLLIN;
fds[1].fd = FILENO;
fds[1].events = POLLOUT;
ret = poll(fds, 2, TIMEOUT * 1000);
if (ret == -1) {
perror ("poll");
return 1;
}
if (!ret) {
return 0;
}
if (fds[0].revents & POLLIN)
{
/********** HANDLING EVENTS HERE ***************/
printf ("FILENO is POLLIN\n");
}
if (fds[1].revents & POLLOUT)
{
/********** HANDLING EVENTS HERE ***************/
printf ("FILENO is POLLOUT\n");
}
return 0;
}
Thank you,
Ran
No (in general).
When you call poll(), the processor context switches to a kernel context, and other processes (and kernel threads) run. Your process will be context switched back in at some point after at least one of your FDs is ready. In general (consider for instance a pipe), interrupt context is not required for this, though note some I/O requires interrupt context to happen (not directly connected to poll()).

How to join a thread in Linux kernel?

The main question is: How we can wait for a thread in Linux kernel to complete? I have seen a few post concerned about proper way of handling threads in Linux kernel but i'm not sure how we can wait for a single thread in the main thread to be completed (suppose we need the thread[3] be done then proceed):
#include <linux/kernel.h>
#include <linux/string.h>
#include <linux/errno.h>
#include <linux/sched.h>
#include <linux/kthread.h>
#include <linux/slab.h>
void *func(void *arg) {
// doing something
return NULL;
}
int init_module(void) {
struct task_struct* thread[5];
int i;
for(i=0; i<5; i++) {
thread[i] = kthread_run(func, (void*) arg, "Creating thread");
}
return 0;
}
void cleanup_module(void) {
printk("cleaning up!\n");
}
AFAIK there is no equivalent of pthread_join() in kernel. Also, I feel like your pattern (of starting bunch of threads and waiting only for one of them) is not really common in kernel. That being said, there kernel does have few synchronization mechanism that may be used to accomplish your goal.
Note that those mechanisms will not guarantee that the thread finished, they will only let main thread know that they finished doing the work they were supposed to do. It may still take some time to really stop this tread and free all resources.
Semaphores
You can create a locked semaphore, then call down in your main thread. This will put it to sleep. Then you will up this semaphore inside of your thread just before exiting. Something like:
struct semaphore sem;
int func(void *arg) {
struct semaphore *sem = (struct semaphore*)arg; // you could use global instead
// do something
up(sem);
return 0;
}
int init_module(void) {
// some initialization
init_MUTEX_LOCKED(&sem);
kthread_run(&func, (void*) &sem, "Creating thread");
down(&sem); // this will block until thread runs up()
}
This should work but is not the most optimal solution. I mention this as it's a known pattern that is also used in userspace. Semaphores in kernel are designed for cases where it's mostly available and this case has high contention. So a similar mechanism optimized for this case was created.
Completions
You can declare completions using:
struct completion comp;
init_completion(&comp);
or:
DECLARE_COMPLETION(comp);
Then you can use wait_for_completion(&comp); instead of down() to wait in main thread and complete(&comp); instead of up() in your thread.
Here's the full example:
DECLARE_COMPLETION(comp);
struct my_data {
int id;
struct completion *comp;
};
int func(void *arg) {
struct my_data *data = (struct my_data*)arg;
// doing something
if (data->id == 3)
complete(data->comp);
return 0;
}
int init_module(void) {
struct my_data *data[] = kmalloc(sizeof(struct my_data)*N, GFP_KERNEL);
// some initialization
for (int i=0; i<N; i++) {
data[i]->comp = &comp;
data[i]->id = i;
kthread_run(func, (void*) data[i], "my_thread%d", i);
}
wait_for_completion(&comp); // this will block until some thread runs complete()
}
Multiple threads
I don't really see why you would start 5 identical threads and only want to wait for 3rd one but of course you could send different data to each thread, with a field describing it's id, and then call up or complete only if this id equals 3. That's shown in the completion example. There are other ways to do this, this is just one of them.
Word of caution
Go read some more about those mechanisms before using any of them. There are some important details I did not write about here. Also those examples are simplified and not tested, they are here just to show the overall idea.
kthread_stop() is a kernel's way for wait thread to end.
Aside from waiting, kthread_stop() also sets should_stop flag for waited thread and wake up it, if needed. It is usefull for threads which repeat some actions infinitely.
As for single-shot tasks, it is usually simpler to use works for them, instead of kthreads.
EDIT:
Note: kthread_stop() can be called only when kthread(task_struct) structure is not freed.
Either thread function should return only after it found kthread_should_stop() return true, or get_task_struct() should be called before start thread (and put_task_struct() should be called after kthread_stop()).

File segment/section/record locks in Linux threads

I have a multi-threaded process where a file is shared (read and written) by multiple threads. Is there any way a thread can lock one file segment so that other threads cannot access it?
I have tried fcntl(fd, F_SETLKW, &flock), but this lock only works for processes, not threads (a lock is shared between all threads in an process).
Yes - but not with the same mechanism. You'll have to use something like pthread mutexes, and keep track of the bookkeeping yourself.
Possible outline for how to make this work
Wait on and claim a process-level mutex over a bookkeeping structure
make sure no other threads within your process are trying to use that segment
mark yourself as using the file segment
Release the process-level mutex
Grab fnctl lock for process (if necessary)
Do your writing
Release fnctl lock to allow other processes to use the segment (if necessary)
Wait again on process-levelbookkeeping structure mutex (may not be necessary if you can mark it unused atomically)
mark segment as unused within your process.
Release process-level mutex
No. The region-locking feature you're asking about has surprising semantics and it is not being further developed because it is controlled by POSIX. (In fact, it is Kirk McKusick's preferred example of what's wrong with POSIX.) If there is a non-POSIX byte-range lock facility in Linux, I can't find it.
There is discussion of the problems of POSIX byte-range locking in a multithreaded world here: http://www.samba.org/samba/news/articles/low_point/tale_two_stds_os2.html.
However, if you're concerned only with threads within one process, you can build your own region-locking using semaphores. For example:
#include <stdbool.h>
#include <pthread.h>
#include <sys/types.h>
// A record indicating an active lock.
struct threadlock {
int fd; // or -1 for unused entries.
off_t start;
off_t length;
};
// A table of all active locks (and the unused entries).
static struct threadlock all_locks[100];
// Mutex housekeeping.
static pthread_mutex_t mutex;
static pthread_cond_t some_lock_released;
static pthread_once_t once_control = PTHREAD_ONCE_INIT;
static void threadlock_init(void) {
for (int i = 0; i < sizeof(all_locks)/sizeof(all_locks[0]); ++i)
all_locks[i].fd = -1;
pthread_mutex_init(&mutex, (pthread_mutexattr_t *)0);
pthread_cond_init(&some_lock_released, (pthread_condattr_t *)0);
}
// True iff the given region overlaps one that is already locked.
static bool region_overlaps_lock(int fd, off_t start, off_t length) {
for (int i = 0; i < sizeof(all_locks)/sizeof(all_locks[0]); ++i) {
const struct threadlock *t = &all_locks[i];
if (t->fd == fd &&
t->start < start + length &&
start < t->start + t->length)
return true;
}
return false;
}
// Returns a pointer to an unused entry, or NULL if there isn't one.
static struct threadlock *find_unused_entry(void) {
for (int i = 0; i < sizeof(all_locks)/sizeof(all_locks[0]); ++i) {
if (-1 == all_locks[i].fd)
return &all_locks[i];
}
return 0;
}
// True iff the lock table is full.
static inline bool too_many_locks(void) {
return 0 == find_unused_entry();
}
// Wait until no thread has a lock for the given region
// [start, start+end) of the given file descriptor, and then lock
// the region. Keep the return value for threadunlock.
// Warning: if you open two file descriptors on the same file
// (including hard links to the same file), this function will fail
// to notice that they're the same file, and it will happily hand out
// two locks for the same region.
struct threadlock *threadlock(int fd, off_t start, off_t length) {
pthread_once(&once_control, &threadlock_init);
pthread_mutex_lock(&mutex);
while (region_overlaps_lock(fd, start, length) || too_many_locks())
pthread_cond_wait(&some_lock_released, &mutex);
struct threadlock *newlock = find_unused_entry();
newlock->fd = fd;
newlock->start = start;
newlock->length = length;
pthread_mutex_unlock(&mutex);
return newlock;
}
// Unlocks a region locked by threadlock.
void threadunlock(struct threadlock *what_threadlock_returned) {
pthread_mutex_lock(&mutex);
what_threadlock_returned->fd = -1;
pthread_cond_broadcast(&some_lock_released);
pthread_mutex_unlock(&mutex);
}
Caution: the code compiles but I haven't tested it even a little.
If you don't need file locks between different processes, avoid the file locks (which are one of the worst designed parts of the POSIX API) and just use mutexes or other shared memory concurrency primitives.
There are 2 ways you can do it:
Use Mutex to get a record's lock in a thread within the same process. Once the lock is acquired, any other thread in the process, mapping the file that tries to acquire the lock is blocked until the lock is released.(Preferable and only most straightforward solution available in Linux).
Semaphores and mutexes on a shared memory or a memory mapped file.

In pthread, how to reliably pass signal to another thread?

I'm trying to write a simple thread pool program in pthread. However, it seems that pthread_cond_signal doesn't block, which creates a problem. For example, let's say I have a "producer-consumer" program:
pthread_cond_t my_cond = PTHREAD_COND_INITIALIZER;
pthread_mutex_t my_cond_m = PTHREAD_MUTEX_INITIALIZER;
void * liberator(void * arg)
{
// XXX make sure he is ready to be freed
sleep(1);
pthread_mutex_lock(&my_cond_m);
pthread_cond_signal(&my_cond);
pthread_mutex_unlock(&my_cond_m);
return NULL;
}
int main()
{
pthread_t t1;
pthread_create(&t1, NULL, liberator, NULL);
// XXX Don't take too long to get ready. Otherwise I'll miss
// the wake up call forever
//sleep(3);
pthread_mutex_lock(&my_cond_m);
pthread_cond_wait(&my_cond, &my_cond_m);
pthread_mutex_unlock(&my_cond_m);
pthread_join(t1, NULL);
return 0;
}
As described in the two XXX marks, if I take away the sleep calls, then main() may stall because it has missed the wake up call from liberator(). Of course, sleep isn't a very robust way to ensure that either.
In real life situation, this would be a worker thread telling the manager thread that it is ready for work, or the manager thread announcing that new work is available.
How would you do this reliably in pthread?
Elaboration
#Borealid's answer kind of works, but his explanation of the problem could be better. I suggest anyone looking at this question to read the discussion in the comments to understand what's going on.
In particular, I myself would amend his answer and code example like this, to make this clearer. (Since Borealid's original answer, while compiled and worked, confused me a lot)
// In main
pthread_mutex_lock(&my_cond_m);
// If the flag is not set, it means liberator has not
// been run yet. I'll wait for him through pthread's signaling
// mechanism
// If it _is_ set, it means liberator has been run. I'll simply
// skip waiting since I've already synchronized. I don't need to
// use pthread's signaling mechanism
if(!flag) pthread_cond_wait(&my_cond, &my_cond_m);
pthread_mutex_unlock(&my_cond_m);
// In liberator thread
pthread_mutex_lock(&my_cond_m);
// Signal anyone who's sleeping. If no one is sleeping yet,
// they should check this flag which indicates I have already
// sent the signal. This is needed because pthread's signals
// is not like a message queue -- a sent signal is lost if
// nobody's waiting for a condition when it's sent.
// You can think of this flag as a "persistent" signal
flag = 1;
pthread_cond_signal(&my_cond);
pthread_mutex_unlock(&my_cond_m);
Use a synchronization variable.
In main:
pthread_mutex_lock(&my_cond_m);
while (!flag) {
pthread_cond_wait(&my_cond, &my_cond_m);
}
pthread_mutex_unlock(&my_cond_m);
In the thread:
pthread_mutex_lock(&my_cond_m);
flag = 1;
pthread_cond_broadcast(&my_cond);
pthread_mutex_unlock(&my_cond_m);
For a producer-consumer problem, this would be the consumer sleeping when the buffer is empty, and the producer sleeping when it is full. Remember to acquire the lock before accessing the global variable.
I found out the solution here. For me, the tricky bit to understand the problem is that:
Producers and consumers must be able to communicate both ways. Either way is not enough.
This two-way communication can be packed into one pthread condition.
To illustrate, the blog post mentioned above demonstrated that this is actually meaningful and desirable behavior:
pthread_mutex_lock(&cond_mutex);
pthread_cond_broadcast(&cond):
pthread_cond_wait(&cond, &cond_mutex);
pthread_mutex_unlock(&cond_mutex);
The idea is that if both the producers and consumers employ this logic, it will be safe for either of them to be sleeping first, since the each will be able to wake the other role up. Put it in another way, in a typical producer-consumer sceanrio -- if a consumer needs to sleep, it's because a producer needs to wake up, and vice versa. Packing this logic in a single pthread condition makes sense.
Of course, the above code has the unintended behavior that a worker thread will also wake up another sleeping worker thread when it actually just wants to wake the producer. This can be solved by a simple variable check as #Borealid suggested:
while(!work_available) pthread_cond_wait(&cond, &cond_mutex);
Upon a worker broadcast, all worker threads will be awaken, but one-by-one (because of the implicit mutex locking in pthread_cond_wait). Since one of the worker threads will consume the work (setting work_available back to false), when other worker threads awake and actually get to work, the work will be unavailable so the worker will sleep again.
Here's some commented code I tested, for anyone interested:
// gcc -Wall -pthread threads.c -lpthread
#include <stdio.h>
#include <pthread.h>
#include <stdlib.h>
#include <assert.h>
pthread_cond_t my_cond = PTHREAD_COND_INITIALIZER;
pthread_mutex_t my_cond_m = PTHREAD_MUTEX_INITIALIZER;
int * next_work = NULL;
int all_work_done = 0;
void * worker(void * arg)
{
int * my_work = NULL;
while(!all_work_done)
{
pthread_mutex_lock(&my_cond_m);
if(next_work == NULL)
{
// Signal producer to give work
pthread_cond_broadcast(&my_cond);
// Wait for work to arrive
// It is wrapped in a while loop because the condition
// might be triggered by another worker thread intended
// to wake up the producer
while(!next_work && !all_work_done)
pthread_cond_wait(&my_cond, &my_cond_m);
}
// Work has arrived, cache it locally so producer can
// put in next work ASAP
my_work = next_work;
next_work = NULL;
pthread_mutex_unlock(&my_cond_m);
if(my_work)
{
printf("Worker %d consuming work: %d\n", (int)(pthread_self() % 100), *my_work);
free(my_work);
}
}
return NULL;
}
int * create_work()
{
int * ret = (int *)malloc(sizeof(int));
assert(ret);
*ret = rand() % 100;
return ret;
}
void * producer(void * arg)
{
int i;
for(i = 0; i < 10; i++)
{
pthread_mutex_lock(&my_cond_m);
while(next_work != NULL)
{
// There's still work, signal a worker to pick it up
pthread_cond_broadcast(&my_cond);
// Wait for work to be picked up
pthread_cond_wait(&my_cond, &my_cond_m);
}
// No work is available now, let's put work on the queue
next_work = create_work();
printf("Producer: Created work %d\n", *next_work);
pthread_mutex_unlock(&my_cond_m);
}
// Some workers might still be waiting, release them
pthread_cond_broadcast(&my_cond);
all_work_done = 1;
return NULL;
}
int main()
{
pthread_t t1, t2, t3, t4;
pthread_create(&t1, NULL, worker, NULL);
pthread_create(&t2, NULL, worker, NULL);
pthread_create(&t3, NULL, worker, NULL);
pthread_create(&t4, NULL, worker, NULL);
producer(NULL);
pthread_join(t1, NULL);
pthread_join(t2, NULL);
pthread_join(t3, NULL);
pthread_join(t4, NULL);
return 0;
}

CPU Registers and Cache Coherence

What's the relation between CPU registers and CPU cache when it comes to cache coherence protocols such as MESI? If a certain value is stored in the CPU's cache, and is also stored in a register, then what will happen if the cache line will be marked as "dirty"? To my understanding there is no guarantee that the register will update it's value even though the cache was updated (due to MESI).
Hench this code:
static void Main()
{
bool complete = false;
var t = new Thread (() =>
{
bool toggle = false;
while (!complete) toggle = !toggle;
});
t.Start();
Thread.Sleep (1000);
complete = true;
t.Join(); // Blocks indefinitely
}
(let's assume the compiler didn't optimized the load for 'complete' outside the loop)
to my understanding, the update to "complete" isn't visible to the second thread since it's value is held inside a register (CPU 2's cache was update however).
Does placing a memory barrier forces to "flush" all of the registers? What's the relation of registers to the cache? and what about registers and memory barriers?
There is no relationship. Use the "volatile" keyword.
Modern C++ | C++11
You should use std::atomic
#include <thread>
#include <atomic>
#include <chrono>
std::atomic_bool g_exit{ false }, g_exited{ false };
using namespace std::chrono_literals;
void fn()
{
while (!g_exit)
{
// do something (lets say it took 5s)
std::this_thread::sleep_for(5s);
}
g_exited = true;
}
int main()
{
std::thread wt(fn);
wt.detach();
// do something (lets say it took 2s)
std::this_thread::sleep_for(2s);
// Exit
g_exit = true;
for (int i = 0; i < 5; i++) {
std::this_thread::sleep_for(1s);
if (g_exited) {
break;
}
}
}
MESI protocol used in x86 platform guarantees cache coherence, i.e. changes in one CPU cache are automatically propagated to other CPU caches.
Therefore volatile keyword on x86 and x64 is useful only to prevent reordering.

Resources