Automatic merge of /spare/repo/netdev-2.6 branch master

8 files changed, 84 insertions, 56 deletions

diff --git a/kernel/cpuset.c b/kernel/cpuset.c
index 961d740..00e8f25 100644
--- a/kernel/cpuset.c
+++ b/kernel/cpuset.c
@@ -166,9 +166,8 @@ static struct super_block *cpuset_sb = NULL;
  * The hooks from fork and exit, cpuset_fork() and cpuset_exit(), don't
  * (usually) grab cpuset_sem.  These are the two most performance
  * critical pieces of code here.  The exception occurs on exit(),
- * if the last task using a cpuset exits, and the cpuset was marked
- * notify_on_release.  In that case, the cpuset_sem is taken, the
- * path to the released cpuset calculated, and a usermode call made
+ * when a task in a notify_on_release cpuset exits.  Then cpuset_sem
+ * is taken, and if the cpuset count is zero, a usermode call made
  * to /sbin/cpuset_release_agent with the name of the cpuset (path
  * relative to the root of cpuset file system) as the argument.
  *
@@ -1404,6 +1403,18 @@ void cpuset_fork(struct task_struct *tsk)
  *
  * Description: Detach cpuset from @tsk and release it.
  *
+ * Note that cpusets marked notify_on_release force every task
+ * in them to take the global cpuset_sem semaphore when exiting.
+ * This could impact scaling on very large systems.  Be reluctant
+ * to use notify_on_release cpusets where very high task exit
+ * scaling is required on large systems.
+ *
+ * Don't even think about derefencing 'cs' after the cpuset use
+ * count goes to zero, except inside a critical section guarded
+ * by the cpuset_sem semaphore.  If you don't hold cpuset_sem,
+ * then a zero cpuset use count is a license to any other task to
+ * nuke the cpuset immediately.
+ *
  **/
 
 void cpuset_exit(struct task_struct *tsk)
@@ -1415,10 +1426,13 @@ void cpuset_exit(struct task_struct *tsk)
 	tsk->cpuset = NULL;
 	task_unlock(tsk);
 
-	if (atomic_dec_and_test(&cs->count)) {
+	if (notify_on_release(cs)) {
 		down(&cpuset_sem);
-		check_for_release(cs);
+		if (atomic_dec_and_test(&cs->count))
+			check_for_release(cs);
 		up(&cpuset_sem);
+	} else {
+		atomic_dec(&cs->count);
 	}
 }
 
diff --git a/kernel/irq/handle.c b/kernel/irq/handle.c
index 2fb0e46..06b5a63 100644
--- a/kernel/irq/handle.c
+++ b/kernel/irq/handle.c
@@ -30,6 +30,7 @@
  */
 irq_desc_t irq_desc[NR_IRQS] __cacheline_aligned = {
 	[0 ... NR_IRQS-1] = {
+		.status = IRQ_DISABLED,
 		.handler = &no_irq_type,
 		.lock = SPIN_LOCK_UNLOCKED
 	}
diff --git a/kernel/power/main.c b/kernel/power/main.c
index 7960ddf..4cdebc9 100644
--- a/kernel/power/main.c
+++ b/kernel/power/main.c
@@ -156,14 +156,14 @@ static int enter_state(suspend_state_t state)
 		goto Unlock;
 	}
 
-	pr_debug("PM: Preparing system for suspend\n");
+	pr_debug("PM: Preparing system for %s sleep\n", pm_states[state]);
 	if ((error = suspend_prepare(state)))
 		goto Unlock;
 
-	pr_debug("PM: Entering state.\n");
+	pr_debug("PM: Entering %s sleep\n", pm_states[state]);
 	error = suspend_enter(state);
 
-	pr_debug("PM: Finishing up.\n");
+	pr_debug("PM: Finishing wakeup.\n");
 	suspend_finish(state);
  Unlock:
 	up(&pm_sem);
diff --git a/kernel/printk.c b/kernel/printk.c
index 290a07c..01b58d7 100644
--- a/kernel/printk.c
+++ b/kernel/printk.c
@@ -160,42 +160,6 @@ static int __init console_setup(char *str)
 
 __setup("console=", console_setup);
 
-/**
- * add_preferred_console - add a device to the list of preferred consoles.
- *
- * The last preferred console added will be used for kernel messages
- * and stdin/out/err for init.  Normally this is used by console_setup
- * above to handle user-supplied console arguments; however it can also
- * be used by arch-specific code either to override the user or more
- * commonly to provide a default console (ie from PROM variables) when
- * the user has not supplied one.
- */
-int __init add_preferred_console(char *name, int idx, char *options)
-{
-	struct console_cmdline *c;
-	int i;
-
-	/*
-	 *	See if this tty is not yet registered, and
-	 *	if we have a slot free.
-	 */
-	for(i = 0; i < MAX_CMDLINECONSOLES && console_cmdline[i].name[0]; i++)
-		if (strcmp(console_cmdline[i].name, name) == 0 &&
-			  console_cmdline[i].index == idx) {
-				selected_console = i;
-				return 0;
-		}
-	if (i == MAX_CMDLINECONSOLES)
-		return -E2BIG;
-	selected_console = i;
-	c = &console_cmdline[i];
-	memcpy(c->name, name, sizeof(c->name));
-	c->name[sizeof(c->name) - 1] = 0;
-	c->options = options;
-	c->index = idx;
-	return 0;
-}
-
 static int __init log_buf_len_setup(char *str)
 {
 	unsigned long size = memparse(str, &str);
@@ -671,6 +635,42 @@ static void call_console_drivers(unsigned long start, unsigned long end) {}
 #endif
 
 /**
+ * add_preferred_console - add a device to the list of preferred consoles.
+ *
+ * The last preferred console added will be used for kernel messages
+ * and stdin/out/err for init.  Normally this is used by console_setup
+ * above to handle user-supplied console arguments; however it can also
+ * be used by arch-specific code either to override the user or more
+ * commonly to provide a default console (ie from PROM variables) when
+ * the user has not supplied one.
+ */
+int __init add_preferred_console(char *name, int idx, char *options)
+{
+	struct console_cmdline *c;
+	int i;
+
+	/*
+	 *	See if this tty is not yet registered, and
+	 *	if we have a slot free.
+	 */
+	for(i = 0; i < MAX_CMDLINECONSOLES && console_cmdline[i].name[0]; i++)
+		if (strcmp(console_cmdline[i].name, name) == 0 &&
+			  console_cmdline[i].index == idx) {
+				selected_console = i;
+				return 0;
+		}
+	if (i == MAX_CMDLINECONSOLES)
+		return -E2BIG;
+	selected_console = i;
+	c = &console_cmdline[i];
+	memcpy(c->name, name, sizeof(c->name));
+	c->name[sizeof(c->name) - 1] = 0;
+	c->options = options;
+	c->index = idx;
+	return 0;
+}
+
+/**
  * acquire_console_sem - lock the console system for exclusive use.
  *
  * Acquires a semaphore which guarantees that the caller has
diff --git a/kernel/profile.c b/kernel/profile.c
index 0221a50..ad8cbb7 100644
--- a/kernel/profile.c
+++ b/kernel/profile.c
@@ -49,15 +49,19 @@ static DECLARE_MUTEX(profile_flip_mutex);
 
 static int __init profile_setup(char * str)
 {
+	static char __initdata schedstr[] = "schedule";
 	int par;
 
-	if (!strncmp(str, "schedule", 8)) {
+	if (!strncmp(str, schedstr, strlen(schedstr))) {
 		prof_on = SCHED_PROFILING;
-		printk(KERN_INFO "kernel schedule profiling enabled\n");
-		if (str[7] == ',')
-			str += 8;
-	}
-	if (get_option(&str,&par)) {
+		if (str[strlen(schedstr)] == ',')
+			str += strlen(schedstr) + 1;
+		if (get_option(&str, &par))
+			prof_shift = par;
+		printk(KERN_INFO
+			"kernel schedule profiling enabled (shift: %ld)\n",
+			prof_shift);
+	} else if (get_option(&str, &par)) {
 		prof_shift = par;
 		prof_on = CPU_PROFILING;
 		printk(KERN_INFO "kernel profiling enabled (shift: %ld)\n",
diff --git a/kernel/sched.c b/kernel/sched.c
index 0dc3158..66b2ed7 100644
--- a/kernel/sched.c
+++ b/kernel/sched.c
@@ -4243,7 +4243,7 @@ static void move_task_off_dead_cpu(int dead_cpu, struct task_struct *tsk)
 
 	/* No more Mr. Nice Guy. */
 	if (dest_cpu == NR_CPUS) {
-		tsk->cpus_allowed = cpuset_cpus_allowed(tsk);
+		cpus_setall(tsk->cpus_allowed);
 		dest_cpu = any_online_cpu(tsk->cpus_allowed);
 
 		/*
diff --git a/kernel/signal.c b/kernel/signal.c
index 8f3debc..b3c24c7 100644
--- a/kernel/signal.c
+++ b/kernel/signal.c
@@ -522,7 +522,16 @@ static int __dequeue_signal(struct sigpending *pending, sigset_t *mask,
 {
 	int sig = 0;
 
-	sig = next_signal(pending, mask);
+	/* SIGKILL must have priority, otherwise it is quite easy
+	 * to create an unkillable process, sending sig < SIGKILL
+	 * to self */
+	if (unlikely(sigismember(&pending->signal, SIGKILL))) {
+		if (!sigismember(mask, SIGKILL))
+			sig = SIGKILL;
+	}
+
+	if (likely(!sig))
+		sig = next_signal(pending, mask);
 	if (sig) {
 		if (current->notifier) {
 			if (sigismember(current->notifier_mask, sig)) {
diff --git a/kernel/spinlock.c b/kernel/spinlock.c
index e15ed17..0c3f9d8 100644
--- a/kernel/spinlock.c
+++ b/kernel/spinlock.c
@@ -294,7 +294,7 @@ EXPORT_SYMBOL(_spin_unlock_irq);
 void __lockfunc _spin_unlock_bh(spinlock_t *lock)
 {
 	_raw_spin_unlock(lock);
-	preempt_enable();
+	preempt_enable_no_resched();
 	local_bh_enable();
 }
 EXPORT_SYMBOL(_spin_unlock_bh);
@@ -318,7 +318,7 @@ EXPORT_SYMBOL(_read_unlock_irq);
 void __lockfunc _read_unlock_bh(rwlock_t *lock)
 {
 	_raw_read_unlock(lock);
-	preempt_enable();
+	preempt_enable_no_resched();
 	local_bh_enable();
 }
 EXPORT_SYMBOL(_read_unlock_bh);
@@ -342,7 +342,7 @@ EXPORT_SYMBOL(_write_unlock_irq);
 void __lockfunc _write_unlock_bh(rwlock_t *lock)
 {
 	_raw_write_unlock(lock);
-	preempt_enable();
+	preempt_enable_no_resched();
 	local_bh_enable();
 }
 EXPORT_SYMBOL(_write_unlock_bh);
@@ -354,7 +354,7 @@ int __lockfunc _spin_trylock_bh(spinlock_t *lock)
 	if (_raw_spin_trylock(lock))
 		return 1;
 
-	preempt_enable();
+	preempt_enable_no_resched();
 	local_bh_enable();
 	return 0;
 }