Merge bk://ldm.bkbits.net/linux-2.5

into home.transmeta.com:/home/torvalds/v2.5/linux

6 files changed, 109 insertions, 31 deletions

diff --git a/kernel/exit.c b/kernel/exit.c
index 7189e9bce6d4..6ed07def4c62 100644
--- a/kernel/exit.c
+++ b/kernel/exit.c
@@ -32,6 +32,7 @@ int getrusage(struct task_struct *, int, struct rusage *);
 static struct dentry * __unhash_process(struct task_struct *p)
 {
 	struct dentry *proc_dentry;
+
 	nr_threads--;
 	detach_pid(p, PIDTYPE_PID);
 	detach_pid(p, PIDTYPE_TGID);
@@ -57,31 +58,31 @@ static struct dentry * __unhash_process(struct task_struct *p)
 void release_task(struct task_struct * p)
 {
 	struct dentry *proc_dentry;
+	task_t *leader;
 
-	if (p->state != TASK_ZOMBIE)
+	if (p->state < TASK_ZOMBIE)
 		BUG();
 	if (p != current)
 		wait_task_inactive(p);
 	atomic_dec(&p->user->processes);
 	security_ops->task_free_security(p);
 	free_uid(p->user);
-	if (unlikely(p->ptrace)) {
-		write_lock_irq(&tasklist_lock);
+	write_lock_irq(&tasklist_lock);
+	if (unlikely(p->ptrace))
 		__ptrace_unlink(p);
-		write_unlock_irq(&tasklist_lock);
-	}
 	BUG_ON(!list_empty(&p->ptrace_list) || !list_empty(&p->ptrace_children));
-	write_lock_irq(&tasklist_lock);
 	__exit_sighand(p);
 	proc_dentry = __unhash_process(p);
 
 	/*
 	 * If we are the last non-leader member of the thread
 	 * group, and the leader is zombie, then notify the
-	 * group leader's parent process.
+	 * group leader's parent process. (if it wants notification.)
 	 */
-	if (p->group_leader != p && thread_group_empty(p))
-		do_notify_parent(p->group_leader, p->group_leader->exit_signal);
+	leader = p->group_leader;
+	if (leader != p && thread_group_empty(leader) &&
+		    leader->state == TASK_ZOMBIE && leader->exit_signal != -1)
+		do_notify_parent(leader, leader->exit_signal);
 
 	p->parent->cutime += p->utime + p->cutime;
 	p->parent->cstime += p->stime + p->cstime;
@@ -159,7 +160,7 @@ static int __will_become_orphaned_pgrp(int pgrp, task_t *ignored_task)
 
 	for_each_task_pid(pgrp, PIDTYPE_PGID, p, l, pid) {
 		if (p == ignored_task
-				|| p->state == TASK_ZOMBIE 
+				|| p->state >= TASK_ZOMBIE 
 				|| p->real_parent->pid == 1)
 			continue;
 		if (p->real_parent->pgrp != pgrp
@@ -435,8 +436,11 @@ void exit_mm(struct task_struct *tsk)
 
 static inline void choose_new_parent(task_t *p, task_t *reaper, task_t *child_reaper)
 {
-	/* Make sure we're not reparenting to ourselves.  */
-	if (p == reaper)
+	/*
+	 * Make sure we're not reparenting to ourselves and that
+	 * the parent is not a zombie.
+	 */
+	if (p == reaper || reaper->state >= TASK_ZOMBIE)
 		p->real_parent = child_reaper;
 	else
 		p->real_parent = reaper;
@@ -774,9 +778,10 @@ static int eligible_child(pid_t pid, int options, task_t *p)
 
 asmlinkage long sys_wait4(pid_t pid,unsigned int * stat_addr, int options, struct rusage * ru)
 {
-	int flag, retval;
 	DECLARE_WAITQUEUE(wait, current);
 	struct task_struct *tsk;
+	unsigned long state;
+	int flag, retval;
 
 	if (options & ~(WNOHANG|WUNTRACED|__WNOTHREAD|__WCLONE|__WALL))
 		return -EINVAL;
@@ -827,7 +832,15 @@ repeat:
 				 */
 				if (ret == 2)
 					continue;
+				/*
+				 * Try to move the task's state to DEAD
+				 * only one thread is allowed to do this:
+				 */
+				state = xchg(&p->state, TASK_DEAD);
+				if (state != TASK_ZOMBIE)
+					continue;
 				read_unlock(&tasklist_lock);
+
 				retval = ru ? getrusage(p, RUSAGE_BOTH, ru) : 0;
 				if (!retval && stat_addr) {
 					if (p->sig->group_exit)
@@ -835,13 +848,16 @@ repeat:
 					else
 						retval = put_user(p->exit_code, stat_addr);
 				}
-				if (retval)
-					goto end_wait4; 
+				if (retval) {
+					p->state = TASK_ZOMBIE;
+					goto end_wait4;
+				}
 				retval = p->pid;
 				if (p->real_parent != p->parent) {
 					write_lock_irq(&tasklist_lock);
 					__ptrace_unlink(p);
 					do_notify_parent(p, SIGCHLD);
+					p->state = TASK_ZOMBIE;
 					write_unlock_irq(&tasklist_lock);
 				} else
 					release_task(p);
diff --git a/kernel/fork.c b/kernel/fork.c
index 062a4d1f9c3e..5880309f3fee 100644
--- a/kernel/fork.c
+++ b/kernel/fork.c
@@ -103,6 +103,52 @@ void remove_wait_queue(wait_queue_head_t *q, wait_queue_t * wait)
 	spin_unlock_irqrestore(&q->lock, flags);
 }
 
+void prepare_to_wait(wait_queue_head_t *q, wait_queue_t *wait, int state)
+{
+	unsigned long flags;
+
+	__set_current_state(state);
+	wait->flags &= ~WQ_FLAG_EXCLUSIVE;
+	spin_lock_irqsave(&q->lock, flags);
+	if (list_empty(&wait->task_list))
+		__add_wait_queue(q, wait);
+	spin_unlock_irqrestore(&q->lock, flags);
+}
+
+void
+prepare_to_wait_exclusive(wait_queue_head_t *q, wait_queue_t *wait, int state)
+{
+	unsigned long flags;
+
+	__set_current_state(state);
+	wait->flags |= WQ_FLAG_EXCLUSIVE;
+	spin_lock_irqsave(&q->lock, flags);
+	if (list_empty(&wait->task_list))
+		__add_wait_queue_tail(q, wait);
+	spin_unlock_irqrestore(&q->lock, flags);
+}
+
+void finish_wait(wait_queue_head_t *q, wait_queue_t *wait)
+{
+	unsigned long flags;
+
+	__set_current_state(TASK_RUNNING);
+	if (!list_empty(&wait->task_list)) {
+		spin_lock_irqsave(&q->lock, flags);
+		list_del_init(&wait->task_list);
+		spin_unlock_irqrestore(&q->lock, flags);
+	}
+}
+
+int autoremove_wake_function(wait_queue_t *wait, unsigned mode, int sync)
+{
+	int ret = default_wake_function(wait, mode, sync);
+
+	if (ret)
+		list_del_init(&wait->task_list);
+	return ret;
+}
+
 void __init fork_init(unsigned long mempages)
 {
 	/* create a slab on which task_structs can be allocated */
diff --git a/kernel/ksyms.c b/kernel/ksyms.c
index 557ae8f7ded2..0409fc676f29 100644
--- a/kernel/ksyms.c
+++ b/kernel/ksyms.c
@@ -400,6 +400,10 @@ EXPORT_SYMBOL(irq_stat);
 EXPORT_SYMBOL(add_wait_queue);
 EXPORT_SYMBOL(add_wait_queue_exclusive);
 EXPORT_SYMBOL(remove_wait_queue);
+EXPORT_SYMBOL(prepare_to_wait);
+EXPORT_SYMBOL(prepare_to_wait_exclusive);
+EXPORT_SYMBOL(finish_wait);
+EXPORT_SYMBOL(autoremove_wake_function);
 
 /* completion handling */
 EXPORT_SYMBOL(wait_for_completion);
@@ -493,7 +497,9 @@ EXPORT_SYMBOL(jiffies_64);
 EXPORT_SYMBOL(xtime);
 EXPORT_SYMBOL(do_gettimeofday);
 EXPORT_SYMBOL(do_settimeofday);
-
+#ifdef CONFIG_DEBUG_KERNEL
+EXPORT_SYMBOL(__might_sleep);
+#endif
 #if !defined(__ia64__)
 EXPORT_SYMBOL(loops_per_jiffy);
 #endif
diff --git a/kernel/pid.c b/kernel/pid.c
index b4da62f0aef2..0005a8cc36cb 100644
--- a/kernel/pid.c
+++ b/kernel/pid.c
@@ -53,6 +53,8 @@ static pidmap_t pidmap_array[PIDMAP_ENTRIES] =
 
 static pidmap_t *map_limit = pidmap_array + PIDMAP_ENTRIES;
 
+static spinlock_t pidmap_lock __cacheline_aligned_in_smp = SPIN_LOCK_UNLOCKED;
+
 inline void free_pidmap(int pid)
 {
 	pidmap_t *map = pidmap_array + pid / BITS_PER_PAGE;
@@ -77,8 +79,13 @@ static inline pidmap_t *next_free_map(pidmap_t *map, int *max_steps)
 			 * Free the page if someone raced with us
 			 * installing it:
 			 */
-			if (cmpxchg(&map->page, NULL, (void *) page))
+			spin_lock(&pidmap_lock);
+			if (map->page)
 				free_page(page);
+			else
+				map->page = (void *)page;
+			spin_unlock(&pidmap_lock);
+
 			if (!map->page)
 				break;
 		}
diff --git a/kernel/sched.c b/kernel/sched.c
index 304f90fd4bdf..9965e5f7549e 100644
--- a/kernel/sched.c
+++ b/kernel/sched.c
@@ -2150,3 +2150,20 @@ void __init sched_init(void)
 	enter_lazy_tlb(&init_mm, current, smp_processor_id());
 }
 
+#ifdef CONFIG_DEBUG_KERNEL
+void __might_sleep(char *file, int line)
+{
+#if defined(in_atomic)
+	static unsigned long prev_jiffy;	/* ratelimiting */
+
+	if (in_atomic()) {
+		if (time_before(jiffies, prev_jiffy + HZ))
+			return;
+		prev_jiffy = jiffies;
+		printk("Sleeping function called from illegal"
+				" context at %s:%d\n", file, line);
+		dump_stack();
+	}
+#endif
+}
+#endif
diff --git a/kernel/timer.c b/kernel/timer.c
index 3b4be840f931..55c14c11c901 100644
--- a/kernel/timer.c
+++ b/kernel/timer.c
@@ -888,20 +888,6 @@ asmlinkage long sys_nanosleep(struct timespec *rqtp, struct timespec *rmtp)
 	if (t.tv_nsec >= 1000000000L || t.tv_nsec < 0 || t.tv_sec < 0)
 		return -EINVAL;
 
-
-	if (t.tv_sec == 0 && t.tv_nsec <= 2000000L &&
-	    current->policy != SCHED_NORMAL)
-	{
-		/*
-		 * Short delay requests up to 2 ms will be handled with
-		 * high precision by a busy wait for all real-time processes.
-		 *
-		 * Its important on SMP not to do this holding locks.
-		 */
-		udelay((t.tv_nsec + 999) / 1000);
-		return 0;
-	}
-
 	expire = timespec_to_jiffies(&t) + (t.tv_sec || t.tv_nsec);
 
 	current->state = TASK_INTERRUPTIBLE;