From 27af4245b6ce99e08c6a7c38825383bf51119cc9 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Mon, 1 Dec 2008 14:18:13 -0800 Subject: posix-timers: use "struct pid*" instead of "struct task_struct*" Impact: restructure, clean up code k_itimer holds the ref to the ->it_process until sys_timer_delete(). This allows to pin up to RLIMIT_SIGPENDING dead task_struct's. Change the code to use "struct pid *" instead. The patch doesn't kill ->it_process, it places ->it_pid into the union. ->it_process is still used by do_cpu_nanosleep() as before. It would be trivial to change the nanosleep code as well, but since it uses it_process in a special way I think it is better to keep this field for grep. The patch bloats the kernel by 104 bytes and it also adds the new pointer, ->it_signal, to k_itimer. It is used by lock_timer() to verify that the found timer was not created by another process. It is not clear why do we use the global database (and thus the global idr_lock) for posix timers. We still need the signal_struct->posix_timers which contains all useable timers, perhaps it is better to use some form of per-process array instead. Signed-off-by: Oleg Nesterov Signed-off-by: Andrew Morton Signed-off-by: Thomas Gleixner --- kernel/posix-timers.c | 43 +++++++++++++++++++++++-------------------- 1 file changed, 23 insertions(+), 20 deletions(-) (limited to 'kernel/posix-timers.c') diff --git a/kernel/posix-timers.c b/kernel/posix-timers.c index 5e79c662294..42a39afd694 100644 --- a/kernel/posix-timers.c +++ b/kernel/posix-timers.c @@ -116,7 +116,7 @@ static DEFINE_SPINLOCK(idr_lock); * must supply functions here, even if the function just returns * ENOSYS. The standard POSIX timer management code assumes the * following: 1.) The k_itimer struct (sched.h) is used for the - * timer. 2.) The list, it_lock, it_clock, it_id and it_process + * timer. 2.) The list, it_lock, it_clock, it_id and it_pid * fields are not modified by timer code. * * At this time all functions EXCEPT clock_nanosleep can be @@ -313,7 +313,8 @@ void do_schedule_next_timer(struct siginfo *info) int posix_timer_event(struct k_itimer *timr, int si_private) { - int shared, ret; + struct task_struct *task; + int shared, ret = -1; /* * FIXME: if ->sigq is queued we can race with * dequeue_signal()->do_schedule_next_timer(). @@ -327,8 +328,13 @@ int posix_timer_event(struct k_itimer *timr, int si_private) */ timr->sigq->info.si_sys_private = si_private; - shared = !(timr->it_sigev_notify & SIGEV_THREAD_ID); - ret = send_sigqueue(timr->sigq, timr->it_process, shared); + rcu_read_lock(); + task = pid_task(timr->it_pid, PIDTYPE_PID); + if (task) { + shared = !(timr->it_sigev_notify & SIGEV_THREAD_ID); + ret = send_sigqueue(timr->sigq, task, shared); + } + rcu_read_unlock(); /* If we failed to send the signal the timer stops. */ return ret > 0; } @@ -405,7 +411,7 @@ static enum hrtimer_restart posix_timer_fn(struct hrtimer *timer) return ret; } -static struct task_struct * good_sigevent(sigevent_t * event) +static struct pid *good_sigevent(sigevent_t * event) { struct task_struct *rtn = current->group_leader; @@ -419,7 +425,7 @@ static struct task_struct * good_sigevent(sigevent_t * event) ((event->sigev_signo <= 0) || (event->sigev_signo > SIGRTMAX))) return NULL; - return rtn; + return task_pid(rtn); } void register_posix_clock(const clockid_t clock_id, struct k_clock *new_clock) @@ -471,7 +477,7 @@ sys_timer_create(const clockid_t which_clock, { struct k_itimer *new_timer; int error, new_timer_id; - struct task_struct *process; + struct pid *it_pid; sigevent_t event; int it_id_set = IT_ID_NOT_SET; @@ -525,11 +531,9 @@ sys_timer_create(const clockid_t which_clock, goto out; } rcu_read_lock(); - process = good_sigevent(&event); - if (process) - get_task_struct(process); + it_pid = get_pid(good_sigevent(&event)); rcu_read_unlock(); - if (!process) { + if (!it_pid) { error = -EINVAL; goto out; } @@ -537,8 +541,7 @@ sys_timer_create(const clockid_t which_clock, event.sigev_notify = SIGEV_SIGNAL; event.sigev_signo = SIGALRM; event.sigev_value.sival_int = new_timer->it_id; - process = current->group_leader; - get_task_struct(process); + it_pid = get_pid(task_tgid(current)); } new_timer->it_sigev_notify = event.sigev_notify; @@ -548,7 +551,8 @@ sys_timer_create(const clockid_t which_clock, new_timer->sigq->info.si_code = SI_TIMER; spin_lock_irq(¤t->sighand->siglock); - new_timer->it_process = process; + new_timer->it_pid = it_pid; + new_timer->it_signal = current->signal; list_add(&new_timer->list, ¤t->signal->posix_timers); spin_unlock_irq(¤t->sighand->siglock); @@ -583,8 +587,7 @@ static struct k_itimer *lock_timer(timer_t timer_id, unsigned long *flags) timr = idr_find(&posix_timers_id, (int)timer_id); if (timr) { spin_lock(&timr->it_lock); - if (timr->it_process && - same_thread_group(timr->it_process, current)) { + if (timr->it_pid && timr->it_signal == current->signal) { spin_unlock(&idr_lock); return timr; } @@ -831,8 +834,8 @@ retry_delete: * This keeps any tasks waiting on the spin lock from thinking * they got something (see the lock code above). */ - put_task_struct(timer->it_process); - timer->it_process = NULL; + put_pid(timer->it_pid); + timer->it_pid = NULL; unlock_timer(timer, flags); release_posix_timer(timer, IT_ID_SET); @@ -858,8 +861,8 @@ retry_delete: * This keeps any tasks waiting on the spin lock from thinking * they got something (see the lock code above). */ - put_task_struct(timer->it_process); - timer->it_process = NULL; + put_pid(timer->it_pid); + timer->it_pid = NULL; unlock_timer(timer, flags); release_posix_timer(timer, IT_ID_SET); -- cgit v1.2.3 From 899921025b406a71a8aeb2d7855658ea0cf0ed23 Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Mon, 1 Dec 2008 14:18:15 -0800 Subject: posix-timers: check ->it_signal instead of ->it_pid to validate the timer Impact: clean up, speed up ->it_pid (was ->it_process) has also a special meaning: if it is NULL, the timer is under deletion or it wasn't initialized yet. We can check ->it_signal != NULL instead, this way we can - simplify sys_timer_create() a bit - remove yet another check from lock_timer() - move put_pid(->it_pid) into release_posix_timer() which runs outside of ->it_lock Signed-off-by: Oleg Nesterov Signed-off-by: Andrew Morton Signed-off-by: Thomas Gleixner --- kernel/posix-timers.c | 17 +++++++---------- 1 file changed, 7 insertions(+), 10 deletions(-) (limited to 'kernel/posix-timers.c') diff --git a/kernel/posix-timers.c b/kernel/posix-timers.c index 42a39afd694..aa922bbb640 100644 --- a/kernel/posix-timers.c +++ b/kernel/posix-timers.c @@ -464,6 +464,7 @@ static void release_posix_timer(struct k_itimer *tmr, int it_id_set) idr_remove(&posix_timers_id, tmr->it_id); spin_unlock_irqrestore(&idr_lock, flags); } + put_pid(tmr->it_pid); sigqueue_free(tmr->sigq); kmem_cache_free(posix_timers_cache, tmr); } @@ -477,7 +478,6 @@ sys_timer_create(const clockid_t which_clock, { struct k_itimer *new_timer; int error, new_timer_id; - struct pid *it_pid; sigevent_t event; int it_id_set = IT_ID_NOT_SET; @@ -531,9 +531,9 @@ sys_timer_create(const clockid_t which_clock, goto out; } rcu_read_lock(); - it_pid = get_pid(good_sigevent(&event)); + new_timer->it_pid = get_pid(good_sigevent(&event)); rcu_read_unlock(); - if (!it_pid) { + if (!new_timer->it_pid) { error = -EINVAL; goto out; } @@ -541,7 +541,7 @@ sys_timer_create(const clockid_t which_clock, event.sigev_notify = SIGEV_SIGNAL; event.sigev_signo = SIGALRM; event.sigev_value.sival_int = new_timer->it_id; - it_pid = get_pid(task_tgid(current)); + new_timer->it_pid = get_pid(task_tgid(current)); } new_timer->it_sigev_notify = event.sigev_notify; @@ -551,7 +551,6 @@ sys_timer_create(const clockid_t which_clock, new_timer->sigq->info.si_code = SI_TIMER; spin_lock_irq(¤t->sighand->siglock); - new_timer->it_pid = it_pid; new_timer->it_signal = current->signal; list_add(&new_timer->list, ¤t->signal->posix_timers); spin_unlock_irq(¤t->sighand->siglock); @@ -587,7 +586,7 @@ static struct k_itimer *lock_timer(timer_t timer_id, unsigned long *flags) timr = idr_find(&posix_timers_id, (int)timer_id); if (timr) { spin_lock(&timr->it_lock); - if (timr->it_pid && timr->it_signal == current->signal) { + if (timr->it_signal == current->signal) { spin_unlock(&idr_lock); return timr; } @@ -834,8 +833,7 @@ retry_delete: * This keeps any tasks waiting on the spin lock from thinking * they got something (see the lock code above). */ - put_pid(timer->it_pid); - timer->it_pid = NULL; + timer->it_signal = NULL; unlock_timer(timer, flags); release_posix_timer(timer, IT_ID_SET); @@ -861,8 +859,7 @@ retry_delete: * This keeps any tasks waiting on the spin lock from thinking * they got something (see the lock code above). */ - put_pid(timer->it_pid); - timer->it_pid = NULL; + timer->it_signal = NULL; unlock_timer(timer, flags); release_posix_timer(timer, IT_ID_SET); -- cgit v1.2.3