[COUNTERPATCH] mm: avoid overflowing preempt_count() in mmu_take_all_locks()

[Avi Kivity <avi@redhat.com>]

mmu_take_all_locks() takes a spinlock for each vma, which means we increase
the preempt count by the number of vmas in an address space.  Since the user
controls the number of vmas, they can cause preempt_count to overflow.

Fix by making mmu_take_all_locks() only disable preemption once by making
the spinlocks preempt-neutral.

Signed-off-by: Avi Kivity <avi@redhat.com>
---
 mm/mmap.c |   37 +++++++++++++++++++++++++++++++++++++
 1 files changed, 37 insertions(+), 0 deletions(-)

diff --git a/mm/mmap.c b/mm/mmap.c
index 75557c6..7305510 100644
--- a/mm/mmap.c
+++ b/mm/mmap.c
@@ -2442,6 +2442,12 @@ static void vm_lock_anon_vma(struct mm_struct *mm, struct anon_vma *anon_vma)
 		 */
 		spin_lock_nest_lock(&anon_vma->lock, &mm->mmap_sem);
 		/*
+		 * To avoid O(nr_vmas) preempt_disable()s, we keep each
+		 * vma preemp_count neutral and instead disable preempt
+		 * once per mmu_take_all_lock().
+		 */
+		preempt_enable();
+		/*
 		 * We can safely modify head.next after taking the
 		 * anon_vma->lock. If some other vma in this mm shares
 		 * the same anon_vma we won't take it again.
@@ -2471,6 +2477,12 @@ static void vm_lock_mapping(struct mm_struct *mm, struct address_space *mapping)
 		if (test_and_set_bit(AS_MM_ALL_LOCKS, &mapping->flags))
 			BUG();
 		spin_lock_nest_lock(&mapping->i_mmap_lock, &mm->mmap_sem);
+		/*
+		 * To avoid O(nr_vmas) preempt_disable()s, we keep each
+		 * vma preemp_count neutral and instead disable preempt
+		 * once per mmu_take_all_lock().
+		 */
+		preempt_enable();
 	}
 }
 
@@ -2516,6 +2528,13 @@ int mm_take_all_locks(struct mm_struct *mm)
 
 	mutex_lock(&mm_all_locks_mutex);
 
+	/*
+	 * To avoid O(nr_vmas) preempt_disable()s, we keep each
+	 * vma preemp_count neutral and instead disable preempt
+	 * once per mmu_take_all_lock().
+	 */
+	preempt_disable();
+
 	for (vma = mm->mmap; vma; vma = vma->vm_next) {
 		if (signal_pending(current))
 			goto out_unlock;
@@ -2558,6 +2577,12 @@ static void vm_unlock_anon_vma(struct anon_vma *anon_vma)
 		if (!__test_and_clear_bit(0, (unsigned long *)
 					  &anon_vma->head.next))
 			BUG();
+		/*
+		 * To avoid O(nr_vmas) preempt_disable()s, we keep each
+		 * vma preemp_count neutral and instead disable preempt
+		 * once per mmu_take_all_lock().
+		 */
+		preempt_disable();
 		spin_unlock(&anon_vma->lock);
 	}
 }
@@ -2569,6 +2594,12 @@ static void vm_unlock_mapping(struct address_space *mapping)
 		 * AS_MM_ALL_LOCKS can't change to 0 from under us
 		 * because we hold the mm_all_locks_mutex.
 		 */
+		/*
+		 * To avoid O(nr_vmas) preempt_disable()s, we keep each
+		 * vma preemp_count neutral and instead disable preempt
+		 * once per mmu_take_all_lock().
+		 */
+		preempt_disable();
 		spin_unlock(&mapping->i_mmap_lock);
 		if (!test_and_clear_bit(AS_MM_ALL_LOCKS,
 					&mapping->flags))
@@ -2596,6 +2627,12 @@ void mm_drop_all_locks(struct mm_struct *mm)
 			vm_unlock_mapping(vma->vm_file->f_mapping);
 	}
 
+	/*
+	 * To avoid O(nr_vmas) preempt_disable()s, we keep each
+	 * vma preemp_count neutral and instead disable preempt
+	 * once per mmu_take_all_lock().
+	 */
+	preempt_enable();
 	mutex_unlock(&mm_all_locks_mutex);
 }
 
-- 
1.7.0.2