[PATCH 4/8] powerpc/mm: Fix possible race on task->mm

[Anton Vorontsov <anton.vorontsov@linaro.org>]

Checking for task->mm is dangerous as ->mm might disappear (exit_mm()
assigns NULL under task_lock(), so tasklist lock is not enough).

We can't use get_task_mm()/mmput() pair as mmput() might sleep,
so let's take the task lock while we care about its mm.

Signed-off-by: Anton Vorontsov <anton.vorontsov@linaro.org>
---
 arch/powerpc/mm/mmu_context_nohash.c |    3 +++
 1 files changed, 3 insertions(+), 0 deletions(-)

diff --git a/arch/powerpc/mm/mmu_context_nohash.c b/arch/powerpc/mm/mmu_context_nohash.c
index 5b63bd3..13ec484 100644
--- a/arch/powerpc/mm/mmu_context_nohash.c
+++ b/arch/powerpc/mm/mmu_context_nohash.c
@@ -43,6 +43,7 @@
 #include <linux/kernel.h>
 #include <linux/mm.h>
 #include <linux/init.h>
+#include <linux/sched.h>
 #include <linux/spinlock.h>
 #include <linux/bootmem.h>
 #include <linux/notifier.h>
@@ -360,8 +361,10 @@ static int __cpuinit mmu_context_cpu_notify(struct notifier_block *self,
 		/* We also clear the cpu_vm_mask bits of CPUs going away */
 		read_lock(&tasklist_lock);
 		for_each_process(p) {
+			task_lock(p);
 			if (p->mm)
 				cpumask_clear_cpu(cpu, mm_cpumask(p->mm));
+			task_unlock(p);
 		}
 		read_unlock(&tasklist_lock);
 	break;
-- 
1.7.7.6