_cpu_down() changes the current task's affinity and then recovers it at
the end. The problems are well known: we can't restore old_allowed if it
was bound to the now-dead-cpu, and we can race with the userspace which
can change cpu-affinity during unplug.

_cpu_down() should not play with current->cpus_allowed at all. Instead,
take_cpu_down() can migrate the caller of _cpu_down() after __cpu_disable()
removes the dying cpu from cpu_online_mask.

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
---

 include/linux/sched.h |    1 +
 kernel/sched.c        |    2 +-
 kernel/cpu.c          |   18 ++++++------------
 3 files changed, 8 insertions(+), 13 deletions(-)

--- 34-rc1/include/linux/sched.h~4_CPU_DOWN_AFFINITY	2010-03-15 09:37:46.000000000 +0100
+++ 34-rc1/include/linux/sched.h	2010-03-15 09:41:51.000000000 +0100
@@ -1843,6 +1843,7 @@ extern void sched_clock_idle_sleep_event
 extern void sched_clock_idle_wakeup_event(u64 delta_ns);
 
 #ifdef CONFIG_HOTPLUG_CPU
+extern void move_task_off_dead_cpu(int dead_cpu, struct task_struct *p);
 extern void idle_task_exit(void);
 #else
 static inline void idle_task_exit(void) {}
--- 34-rc1/kernel/sched.c~4_CPU_DOWN_AFFINITY	2010-03-15 09:41:28.000000000 +0100
+++ 34-rc1/kernel/sched.c	2010-03-15 09:41:51.000000000 +0100
@@ -5503,7 +5503,7 @@ static int migration_thread(void *data)
 /*
  * Figure out where task on dead CPU should go, use force if necessary.
  */
-static void move_task_off_dead_cpu(int dead_cpu, struct task_struct *p)
+void move_task_off_dead_cpu(int dead_cpu, struct task_struct *p)
 {
 	struct rq *rq = cpu_rq(dead_cpu);
 	int needs_cpu, dest_cpu;
--- 34-rc1/kernel/cpu.c~4_CPU_DOWN_AFFINITY	2010-03-15 09:37:46.000000000 +0100
+++ 34-rc1/kernel/cpu.c	2010-03-15 09:41:51.000000000 +0100
@@ -163,6 +163,7 @@ static inline void check_for_tasks(int c
 }
 
 struct take_cpu_down_param {
+	struct task_struct *caller;
 	unsigned long mod;
 	void *hcpu;
 };
@@ -171,6 +172,7 @@ struct take_cpu_down_param {
 static int ...
[ message continues ]

--

[ message continues ]

Commit-ID:  6a1bdc1b577ebcb65f6603c57f8347309bc4ab13
Gitweb:     http://git.kernel.org/tip/6a1bdc1b577ebcb65f6603c57f8347309bc4ab13
Author:     Oleg Nesterov <oleg@redhat.com>
AuthorDate: Mon, 15 Mar 2010 10:10:23 +0100
Committer:  Ingo Molnar <mingo@elte.hu>
CommitDate: Fri, 2 Apr 2010 20:12:03 +0200

sched: _cpu_down(): Don't play with current->cpus_allowed

_cpu_down() changes the current task's affinity and then recovers it at
the end. The problems are well known: we can't restore old_allowed if it
was bound to the now-dead-cpu, and we can race with the userspace which
can change cpu-affinity during unplug.

_cpu_down() should not play with current->cpus_allowed at all. Instead,
take_cpu_down() can migrate the caller of _cpu_down() after __cpu_disable()
removes the dying cpu from cpu_online_mask.

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Acked-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <20100315091023.GA9148@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
---
 include/linux/sched.h |    1 +
 kernel/cpu.c          |   18 ++++++------------
 kernel/sched.c        |    2 +-
 3 files changed, 8 insertions(+), 13 deletions(-)

diff --git a/include/linux/sched.h b/include/linux/sched.h
index 43c9451..8bea407 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -1843,6 +1843,7 @@ extern void sched_clock_idle_sleep_event(void);
 extern void sched_clock_idle_wakeup_event(u64 delta_ns);
 
 #ifdef CONFIG_HOTPLUG_CPU
+extern void move_task_off_dead_cpu(int dead_cpu, struct task_struct *p);
 extern void idle_task_exit(void);
 #else
 static inline void idle_task_exit(void) {}
diff --git a/kernel/cpu.c b/kernel/cpu.c
index f8cced2..8d340fa 100644
--- a/kernel/cpu.c
+++ b/kernel/cpu.c
@@ -163,6 +163,7 @@ static inline void check_for_tasks(int cpu)
 }
 
 struct take_cpu_down_param {
+	struct task_struct *caller;
 	unsigned long mod;
 	void *hcpu;
 };
@@ -171,6 +172,7 @@ struct ...
[ message continues ]