Merge with master.kernel.org:/pub/scm/linux/kernel/git/torvalds/linux-2.6.git

6 files changed, 40 insertions, 13 deletions

diff --git a/kernel/cpuset.c b/kernel/cpuset.c
index 961d74044de..00e8f257551 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 06b5a632399..436c7d93c00 100644
--- a/kernel/irq/handle.c
+++ b/kernel/irq/handle.c
@@ -119,8 +119,6 @@ fastcall unsigned int __do_IRQ(unsigned int irq, struct pt_regs *regs)
 		 */
 		desc->handler->ack(irq);
 		action_ret = handle_IRQ_event(irq, regs, desc->action);
-		if (!noirqdebug)
-			note_interrupt(irq, desc, action_ret);
 		desc->handler->end(irq);
 		return 1;
 	}
diff --git a/kernel/module.c b/kernel/module.c
index 5734ab09d3f..83b3d376708 100644
--- a/kernel/module.c
+++ b/kernel/module.c
@@ -1758,6 +1758,7 @@ sys_init_module(void __user *umod,
 		const char __user *uargs)
 {
 	struct module *mod;
+	mm_segment_t old_fs = get_fs();
 	int ret = 0;
 
 	/* Must have permission */
@@ -1775,6 +1776,9 @@ sys_init_module(void __user *umod,
 		return PTR_ERR(mod);
 	}
 
+	/* flush the icache in correct context */
+	set_fs(KERNEL_DS);
+
 	/* Flush the instruction cache, since we've played with text */
 	if (mod->module_init)
 		flush_icache_range((unsigned long)mod->module_init,
@@ -1783,6 +1787,8 @@ sys_init_module(void __user *umod,
 	flush_icache_range((unsigned long)mod->module_core,
 			   (unsigned long)mod->module_core + mod->core_size);
 
+	set_fs(old_fs);
+
 	/* Now sew it into the lists.  They won't access us, since
            strong_try_module_get() will fail. */
 	stop_machine_run(__link_module, mod, NR_CPUS);
diff --git a/kernel/sched.c b/kernel/sched.c
index 0dc3158667a..66b2ed78482 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 293e189d8bc..c89821b69ae 100644
--- a/kernel/signal.c
+++ b/kernel/signal.c
@@ -523,7 +523,16 @@ 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 e15ed17863f..0c3f9d8bbe1 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;
 }