From 0c1eecfb345401629aa57c9d3b077273e56c45a7 Mon Sep 17 00:00:00 2001 From: "Rafael J. Wysocki" Date: Thu, 19 Jul 2007 01:47:33 -0700 Subject: Freezer: avoid freezing kernel threads prematurely Kernel threads should not have TIF_FREEZE set when user space processes are being frozen, since otherwise some of them might be frozen prematurely. To prevent this from happening we can (1) make exit_mm() unset TIF_FREEZE unconditionally just after clearing tsk->mm and (2) make try_to_freeze_tasks() check if p->mm is different from zero and PF_BORROWED_MM is unset in p->flags when user space processes are to be frozen. Namely, when user space processes are being frozen, we only should set TIF_FREEZE for tasks that have p->mm different from NULL and don't have PF_BORROWED_MM set in p->flags. For this reason task_lock() must be used to prevent try_to_freeze_tasks() from racing with use_mm()/unuse_mm(), in which p->mm and p->flags.PF_BORROWED_MM are changed under task_lock(p). Also, we need to prevent the following scenario from happening: * daemonize() is called by a task spawned from a user space code path * freezer checks if the task has p->mm set and the result is positive * task enters exit_mm() and clears its TIF_FREEZE * freezer sets TIF_FREEZE for the task * task calls try_to_freeze() and goes to the refrigerator, which is wrong at that point This requires us to acquire task_lock(p) before p->flags.PF_BORROWED_MM and p->mm are examined and release it after TIF_FREEZE is set for p (or it turns out that TIF_FREEZE should not be set). Signed-off-by: Rafael J. Wysocki Cc: Gautham R Shenoy Cc: Pavel Machek Cc: Nigel Cunningham Cc: Oleg Nesterov Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/exit.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'kernel/exit.c') diff --git a/kernel/exit.c b/kernel/exit.c index e8af8d0c248..464c2b172f0 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -45,6 +45,7 @@ #include #include #include +#include #include #include @@ -594,6 +595,8 @@ static void exit_mm(struct task_struct * tsk) tsk->mm = NULL; up_read(&mm->mmap_sem); enter_lazy_tlb(mm, current); + /* We don't want this task to be frozen prematurely */ + clear_freeze_flag(tsk); task_unlock(tsk); mmput(mm); } -- cgit v1.2.3-70-g09d2 From 247284481ca40288bd120cf0707681c3bdbee78f Mon Sep 17 00:00:00 2001 From: Oleg Nesterov Date: Sat, 4 Aug 2007 01:04:41 +0400 Subject: Kill some obsolete sub-thread-ptrace stuff There is a couple of subtle checks which were needed to handle ptracing from the same thread group. This was deprecated a long ago, imho this code just complicates the understanding. And, the "->parent->signal->flags & SIGNAL_GROUP_EXIT" check in exit_notify() is not right. SIGNAL_GROUP_EXIT can mean exec(), not exit_group(). This means ptracer can lose a ptraced zombie on exec(). Minor problem, but still the bug. Signed-off-by: Oleg Nesterov Acked-by: Roland McGrath Signed-off-by: Linus Torvalds --- kernel/exit.c | 8 ++------ kernel/signal.c | 4 ---- 2 files changed, 2 insertions(+), 10 deletions(-) (limited to 'kernel/exit.c') diff --git a/kernel/exit.c b/kernel/exit.c index 464c2b172f0..9578c1ae19c 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -813,7 +813,7 @@ static void exit_notify(struct task_struct *tsk) __kill_pgrp_info(SIGCONT, SEND_SIG_PRIV, pgrp); } - /* Let father know we died + /* Let father know we died * * Thread signals are configurable, but you aren't going to use * that to send signals to arbitary processes. @@ -826,9 +826,7 @@ static void exit_notify(struct task_struct *tsk) * If our self_exec id doesn't match our parent_exec_id then * we have changed execution domain as these two values started * the same after a fork. - * */ - if (tsk->exit_signal != SIGCHLD && tsk->exit_signal != -1 && ( tsk->parent_exec_id != t->self_exec_id || tsk->self_exec_id != tsk->parent_exec_id) @@ -848,9 +846,7 @@ static void exit_notify(struct task_struct *tsk) } state = EXIT_ZOMBIE; - if (tsk->exit_signal == -1 && - (likely(tsk->ptrace == 0) || - unlikely(tsk->parent->signal->flags & SIGNAL_GROUP_EXIT))) + if (tsk->exit_signal == -1 && likely(!tsk->ptrace)) state = EXIT_DEAD; tsk->exit_state = state; diff --git a/kernel/signal.c b/kernel/signal.c index ef8156a6aad..b27c01a6644 100644 --- a/kernel/signal.c +++ b/kernel/signal.c @@ -1561,10 +1561,6 @@ static inline int may_ptrace_stop(void) (current->ptrace & PT_ATTACHED))) return 0; - if (unlikely(current->signal == current->parent->signal) && - unlikely(current->signal->flags & SIGNAL_GROUP_EXIT)) - return 0; - /* * Are we in the middle of do_coredump? * If so and our tracer is also part of the coredump stopping -- cgit v1.2.3-70-g09d2 From f2ab6d8889422c1f5354f014e8bef337b1d1bade Mon Sep 17 00:00:00 2001 From: Jonathan Lim Date: Thu, 30 Aug 2007 23:56:23 -0700 Subject: Assign task_struct.exit_code before taskstats_exit() taskstats.ac_exitcode is assigned to task_struct.exit_code in bacct_add_tsk() through the following kernel function calls: do_exit() taskstats_exit() fill_pid() bacct_add_tsk() The problem is that in do_exit(), task_struct.exit_code is set to 'code' only after taskstats_exit() has been called. So we need to move the assignment before taskstats_exit(). Signed-off-by: Jonathan Lim Cc: Balbir Singh Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/exit.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel/exit.c') diff --git a/kernel/exit.c b/kernel/exit.c index 9578c1ae19c..06b24b3aa37 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -975,6 +975,7 @@ fastcall NORET_TYPE void do_exit(long code) if (unlikely(tsk->audit_context)) audit_free(tsk); + tsk->exit_code = code; taskstats_exit(tsk, group_dead); exit_mm(tsk); @@ -996,7 +997,6 @@ fastcall NORET_TYPE void do_exit(long code) if (tsk->binfmt) module_put(tsk->binfmt->module); - tsk->exit_code = code; proc_exit_connector(tsk); exit_task_namespaces(tsk); exit_notify(tsk); -- cgit v1.2.3-70-g09d2 From b8fceee17a310f189188599a8fa5e9beaff57eb0 Mon Sep 17 00:00:00 2001 From: Davide Libenzi Date: Thu, 20 Sep 2007 12:40:16 -0700 Subject: signalfd simplification This simplifies signalfd code, by avoiding it to remain attached to the sighand during its lifetime. In this way, the signalfd remain attached to the sighand only during poll(2) (and select and epoll) and read(2). This also allows to remove all the custom "tsk == current" checks in kernel/signal.c, since dequeue_signal() will only be called by "current". I think this is also what Ben was suggesting time ago. The external effect of this, is that a thread can extract only its own private signals and the group ones. I think this is an acceptable behaviour, in that those are the signals the thread would be able to fetch w/out signalfd. Signed-off-by: Davide Libenzi Signed-off-by: Linus Torvalds --- fs/exec.c | 3 - fs/signalfd.c | 190 +++++++--------------------------------------- include/linux/init_task.h | 2 +- include/linux/sched.h | 2 +- include/linux/signalfd.h | 40 +--------- kernel/exit.c | 9 --- kernel/fork.c | 2 +- kernel/signal.c | 8 +- 8 files changed, 39 insertions(+), 217 deletions(-) (limited to 'kernel/exit.c') diff --git a/fs/exec.c b/fs/exec.c index c21a8cc0627..073b0b8c6d0 100644 --- a/fs/exec.c +++ b/fs/exec.c @@ -50,7 +50,6 @@ #include #include #include -#include #include #include @@ -784,7 +783,6 @@ static int de_thread(struct task_struct *tsk) * and we can just re-use it all. */ if (atomic_read(&oldsighand->count) <= 1) { - signalfd_detach(tsk); exit_itimers(sig); return 0; } @@ -923,7 +921,6 @@ static int de_thread(struct task_struct *tsk) sig->flags = 0; no_thread_group: - signalfd_detach(tsk); exit_itimers(sig); if (leader) release_task(leader); diff --git a/fs/signalfd.c b/fs/signalfd.c index a8e293d3003..aefb0be0794 100644 --- a/fs/signalfd.c +++ b/fs/signalfd.c @@ -11,8 +11,10 @@ * Now using anonymous inode source. * Thanks to Oleg Nesterov for useful code review and suggestions. * More comments and suggestions from Arnd Bergmann. - * Sat May 19, 2007: Davi E. M. Arnaut + * Sat May 19, 2007: Davi E. M. Arnaut * Retrieve multiple signals with one read() call + * Sun Jul 15, 2007: Davide Libenzi + * Attach to the sighand only during read() and poll(). */ #include @@ -27,102 +29,12 @@ #include struct signalfd_ctx { - struct list_head lnk; - wait_queue_head_t wqh; sigset_t sigmask; - struct task_struct *tsk; }; -struct signalfd_lockctx { - struct task_struct *tsk; - unsigned long flags; -}; - -/* - * Tries to acquire the sighand lock. We do not increment the sighand - * use count, and we do not even pin the task struct, so we need to - * do it inside an RCU read lock, and we must be prepared for the - * ctx->tsk going to NULL (in signalfd_deliver()), and for the sighand - * being detached. We return 0 if the sighand has been detached, or - * 1 if we were able to pin the sighand lock. - */ -static int signalfd_lock(struct signalfd_ctx *ctx, struct signalfd_lockctx *lk) -{ - struct sighand_struct *sighand = NULL; - - rcu_read_lock(); - lk->tsk = rcu_dereference(ctx->tsk); - if (likely(lk->tsk != NULL)) - sighand = lock_task_sighand(lk->tsk, &lk->flags); - rcu_read_unlock(); - - if (!sighand) - return 0; - - if (!ctx->tsk) { - unlock_task_sighand(lk->tsk, &lk->flags); - return 0; - } - - if (lk->tsk->tgid == current->tgid) - lk->tsk = current; - - return 1; -} - -static void signalfd_unlock(struct signalfd_lockctx *lk) -{ - unlock_task_sighand(lk->tsk, &lk->flags); -} - -/* - * This must be called with the sighand lock held. - */ -void signalfd_deliver(struct task_struct *tsk, int sig) -{ - struct sighand_struct *sighand = tsk->sighand; - struct signalfd_ctx *ctx, *tmp; - - BUG_ON(!sig); - list_for_each_entry_safe(ctx, tmp, &sighand->signalfd_list, lnk) { - /* - * We use a negative signal value as a way to broadcast that the - * sighand has been orphaned, so that we can notify all the - * listeners about this. Remember the ctx->sigmask is inverted, - * so if the user is interested in a signal, that corresponding - * bit will be zero. - */ - if (sig < 0) { - if (ctx->tsk == tsk) { - ctx->tsk = NULL; - list_del_init(&ctx->lnk); - wake_up(&ctx->wqh); - } - } else { - if (!sigismember(&ctx->sigmask, sig)) - wake_up(&ctx->wqh); - } - } -} - -static void signalfd_cleanup(struct signalfd_ctx *ctx) -{ - struct signalfd_lockctx lk; - - /* - * This is tricky. If the sighand is gone, we do not need to remove - * context from the list, the list itself won't be there anymore. - */ - if (signalfd_lock(ctx, &lk)) { - list_del(&ctx->lnk); - signalfd_unlock(&lk); - } - kfree(ctx); -} - static int signalfd_release(struct inode *inode, struct file *file) { - signalfd_cleanup(file->private_data); + kfree(file->private_data); return 0; } @@ -130,23 +42,15 @@ static unsigned int signalfd_poll(struct file *file, poll_table *wait) { struct signalfd_ctx *ctx = file->private_data; unsigned int events = 0; - struct signalfd_lockctx lk; - poll_wait(file, &ctx->wqh, wait); + poll_wait(file, ¤t->sighand->signalfd_wqh, wait); - /* - * Let the caller get a POLLIN in this case, ala socket recv() when - * the peer disconnects. - */ - if (signalfd_lock(ctx, &lk)) { - if ((lk.tsk == current && - next_signal(&lk.tsk->pending, &ctx->sigmask) > 0) || - next_signal(&lk.tsk->signal->shared_pending, - &ctx->sigmask) > 0) - events |= POLLIN; - signalfd_unlock(&lk); - } else + spin_lock_irq(¤t->sighand->siglock); + if (next_signal(¤t->pending, &ctx->sigmask) || + next_signal(¤t->signal->shared_pending, + &ctx->sigmask)) events |= POLLIN; + spin_unlock_irq(¤t->sighand->siglock); return events; } @@ -219,59 +123,46 @@ static ssize_t signalfd_dequeue(struct signalfd_ctx *ctx, siginfo_t *info, int nonblock) { ssize_t ret; - struct signalfd_lockctx lk; DECLARE_WAITQUEUE(wait, current); - if (!signalfd_lock(ctx, &lk)) - return 0; - - ret = dequeue_signal(lk.tsk, &ctx->sigmask, info); + spin_lock_irq(¤t->sighand->siglock); + ret = dequeue_signal(current, &ctx->sigmask, info); switch (ret) { case 0: if (!nonblock) break; ret = -EAGAIN; default: - signalfd_unlock(&lk); + spin_unlock_irq(¤t->sighand->siglock); return ret; } - add_wait_queue(&ctx->wqh, &wait); + add_wait_queue(¤t->sighand->signalfd_wqh, &wait); for (;;) { set_current_state(TASK_INTERRUPTIBLE); - ret = dequeue_signal(lk.tsk, &ctx->sigmask, info); - signalfd_unlock(&lk); + ret = dequeue_signal(current, &ctx->sigmask, info); if (ret != 0) break; if (signal_pending(current)) { ret = -ERESTARTSYS; break; } + spin_unlock_irq(¤t->sighand->siglock); schedule(); - ret = signalfd_lock(ctx, &lk); - if (unlikely(!ret)) { - /* - * Let the caller read zero byte, ala socket - * recv() when the peer disconnect. This test - * must be done before doing a dequeue_signal(), - * because if the sighand has been orphaned, - * the dequeue_signal() call is going to crash - * because ->sighand will be long gone. - */ - break; - } + spin_lock_irq(¤t->sighand->siglock); } + spin_unlock_irq(¤t->sighand->siglock); - remove_wait_queue(&ctx->wqh, &wait); + remove_wait_queue(¤t->sighand->signalfd_wqh, &wait); __set_current_state(TASK_RUNNING); return ret; } /* - * Returns either the size of a "struct signalfd_siginfo", or zero if the - * sighand we are attached to, has been orphaned. The "count" parameter - * must be at least the size of a "struct signalfd_siginfo". + * Returns a multiple of the size of a "struct signalfd_siginfo", or a negative + * error code. The "count" parameter must be at least the size of a + * "struct signalfd_siginfo". */ static ssize_t signalfd_read(struct file *file, char __user *buf, size_t count, loff_t *ppos) @@ -287,7 +178,6 @@ static ssize_t signalfd_read(struct file *file, char __user *buf, size_t count, return -EINVAL; siginfo = (struct signalfd_siginfo __user *) buf; - do { ret = signalfd_dequeue(ctx, &info, nonblock); if (unlikely(ret <= 0)) @@ -300,7 +190,7 @@ static ssize_t signalfd_read(struct file *file, char __user *buf, size_t count, nonblock = 1; } while (--count); - return total ? total : ret; + return total ? total: ret; } static const struct file_operations signalfd_fops = { @@ -309,20 +199,13 @@ static const struct file_operations signalfd_fops = { .read = signalfd_read, }; -/* - * Create a file descriptor that is associated with our signal - * state. We can pass it around to others if we want to, but - * it will always be _our_ signal state. - */ asmlinkage long sys_signalfd(int ufd, sigset_t __user *user_mask, size_t sizemask) { int error; sigset_t sigmask; struct signalfd_ctx *ctx; - struct sighand_struct *sighand; struct file *file; struct inode *inode; - struct signalfd_lockctx lk; if (sizemask != sizeof(sigset_t) || copy_from_user(&sigmask, user_mask, sizeof(sigmask))) @@ -335,17 +218,7 @@ asmlinkage long sys_signalfd(int ufd, sigset_t __user *user_mask, size_t sizemas if (!ctx) return -ENOMEM; - init_waitqueue_head(&ctx->wqh); ctx->sigmask = sigmask; - ctx->tsk = current->group_leader; - - sighand = current->sighand; - /* - * Add this fd to the list of signal listeners. - */ - spin_lock_irq(&sighand->siglock); - list_add_tail(&ctx->lnk, &sighand->signalfd_list); - spin_unlock_irq(&sighand->siglock); /* * When we call this, the initialization must be complete, since @@ -364,23 +237,18 @@ asmlinkage long sys_signalfd(int ufd, sigset_t __user *user_mask, size_t sizemas fput(file); return -EINVAL; } - /* - * We need to be prepared of the fact that the sighand this fd - * is attached to, has been detched. In that case signalfd_lock() - * will return 0, and we'll just skip setting the new mask. - */ - if (signalfd_lock(ctx, &lk)) { - ctx->sigmask = sigmask; - signalfd_unlock(&lk); - } - wake_up(&ctx->wqh); + spin_lock_irq(¤t->sighand->siglock); + ctx->sigmask = sigmask; + spin_unlock_irq(¤t->sighand->siglock); + + wake_up(¤t->sighand->signalfd_wqh); fput(file); } return ufd; err_fdalloc: - signalfd_cleanup(ctx); + kfree(ctx); return error; } diff --git a/include/linux/init_task.h b/include/linux/init_task.h index cab741c2d60..f8abfa349ef 100644 --- a/include/linux/init_task.h +++ b/include/linux/init_task.h @@ -86,7 +86,7 @@ extern struct nsproxy init_nsproxy; .count = ATOMIC_INIT(1), \ .action = { { { .sa_handler = NULL, } }, }, \ .siglock = __SPIN_LOCK_UNLOCKED(sighand.siglock), \ - .signalfd_list = LIST_HEAD_INIT(sighand.signalfd_list), \ + .signalfd_wqh = __WAIT_QUEUE_HEAD_INITIALIZER(sighand.signalfd_wqh), \ } extern struct group_info init_groups; diff --git a/include/linux/sched.h b/include/linux/sched.h index 3de79016f2a..a01ac6dd5f5 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -438,7 +438,7 @@ struct sighand_struct { atomic_t count; struct k_sigaction action[_NSIG]; spinlock_t siglock; - struct list_head signalfd_list; + wait_queue_head_t signalfd_wqh; }; struct pacct_struct { diff --git a/include/linux/signalfd.h b/include/linux/signalfd.h index 51042949569..4c9ff0910ae 100644 --- a/include/linux/signalfd.h +++ b/include/linux/signalfd.h @@ -45,49 +45,17 @@ struct signalfd_siginfo { #ifdef CONFIG_SIGNALFD /* - * Deliver the signal to listening signalfd. This must be called - * with the sighand lock held. Same are the following that end up - * calling signalfd_deliver(). - */ -void signalfd_deliver(struct task_struct *tsk, int sig); - -/* - * No need to fall inside signalfd_deliver() if no signal listeners - * are available. + * Deliver the signal to listening signalfd. */ static inline void signalfd_notify(struct task_struct *tsk, int sig) { - if (unlikely(!list_empty(&tsk->sighand->signalfd_list))) - signalfd_deliver(tsk, sig); -} - -/* - * The signal -1 is used to notify the signalfd that the sighand - * is on its way to be detached. - */ -static inline void signalfd_detach_locked(struct task_struct *tsk) -{ - if (unlikely(!list_empty(&tsk->sighand->signalfd_list))) - signalfd_deliver(tsk, -1); -} - -static inline void signalfd_detach(struct task_struct *tsk) -{ - struct sighand_struct *sighand = tsk->sighand; - - if (unlikely(!list_empty(&sighand->signalfd_list))) { - spin_lock_irq(&sighand->siglock); - signalfd_deliver(tsk, -1); - spin_unlock_irq(&sighand->siglock); - } + if (unlikely(waitqueue_active(&tsk->sighand->signalfd_wqh))) + wake_up(&tsk->sighand->signalfd_wqh); } #else /* CONFIG_SIGNALFD */ -#define signalfd_deliver(t, s) do { } while (0) -#define signalfd_notify(t, s) do { } while (0) -#define signalfd_detach_locked(t) do { } while (0) -#define signalfd_detach(t) do { } while (0) +static inline void signalfd_notify(struct task_struct *tsk, int sig) { } #endif /* CONFIG_SIGNALFD */ diff --git a/kernel/exit.c b/kernel/exit.c index 06b24b3aa37..993369ee94d 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -24,7 +24,6 @@ #include #include #include -#include #include #include #include @@ -86,14 +85,6 @@ static void __exit_signal(struct task_struct *tsk) sighand = rcu_dereference(tsk->sighand); spin_lock(&sighand->siglock); - /* - * Notify that this sighand has been detached. This must - * be called with the tsk->sighand lock held. Also, this - * access tsk->sighand internally, so it must be called - * before tsk->sighand is reset. - */ - signalfd_detach_locked(tsk); - posix_cpu_timers_exit(tsk); if (atomic_dec_and_test(&sig->count)) posix_cpu_timers_exit_group(tsk); diff --git a/kernel/fork.c b/kernel/fork.c index 7332e236d36..33f12f48684 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -1438,7 +1438,7 @@ static void sighand_ctor(void *data, struct kmem_cache *cachep, struct sighand_struct *sighand = data; spin_lock_init(&sighand->siglock); - INIT_LIST_HEAD(&sighand->signalfd_list); + init_waitqueue_head(&sighand->signalfd_wqh); } void __init proc_caches_init(void) diff --git a/kernel/signal.c b/kernel/signal.c index 3169bed0b4d..9fb91a32edd 100644 --- a/kernel/signal.c +++ b/kernel/signal.c @@ -378,8 +378,7 @@ int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info) /* We only dequeue private signals from ourselves, we don't let * signalfd steal them */ - if (likely(tsk == current)) - signr = __dequeue_signal(&tsk->pending, mask, info); + signr = __dequeue_signal(&tsk->pending, mask, info); if (!signr) { signr = __dequeue_signal(&tsk->signal->shared_pending, mask, info); @@ -407,8 +406,7 @@ int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info) } } } - if (likely(tsk == current)) - recalc_sigpending(); + recalc_sigpending(); if (signr && unlikely(sig_kernel_stop(signr))) { /* * Set a marker that we have dequeued a stop signal. Our @@ -425,7 +423,7 @@ int dequeue_signal(struct task_struct *tsk, sigset_t *mask, siginfo_t *info) if (!(tsk->signal->flags & SIGNAL_GROUP_EXIT)) tsk->signal->flags |= SIGNAL_STOP_DEQUEUED; } - if (signr && likely(tsk == current) && + if (signr && ((info->si_code & __SI_MASK) == __SI_TIMER) && info->si_sys_private){ /* -- cgit v1.2.3-70-g09d2