percpu: fix synchronization between synchronous map extension and chunk destruction

[ Upstream commit 6710e594 ]

For non-atomic allocations, pcpu_alloc() can try to extend the area
map synchronously after dropping pcpu_lock; however, the extension
wasn't synchronized against chunk destruction and the chunk might get
freed while extension is in progress.

This patch fixes the bug by putting most of non-atomic allocations
under pcpu_alloc_mutex to synchronize against pcpu_balance_work which
is responsible for async chunk management including destruction.
Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-and-tested-by: Alexei Starovoitov <alexei.starovoitov@gmail.com>
Reported-by: Vlastimil Babka <vbabka@suse.cz>
Reported-by: Sasha Levin <sasha.levin@oracle.com>
Cc: stable@vger.kernel.org # v3.18+
Fixes: 1a4d7607 ("percpu: implement asynchronous chunk population")
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>

mm/percpu.c

@@ -160,7 +160,7 @@ static struct pcpu_chunk *pcpu_reserved_chunk;
 static int pcpu_reserved_chunk_limit;
 
 static DEFINE_SPINLOCK(pcpu_lock);	/* all internal data structures */
-static DEFINE_MUTEX(pcpu_alloc_mutex);	/* chunk create/destroy, [de]pop */
+static DEFINE_MUTEX(pcpu_alloc_mutex);	/* chunk create/destroy, [de]pop, map ext */
 
 static struct list_head *pcpu_slot __read_mostly; /* chunk list slots */
 
@@ -446,6 +446,8 @@ static int pcpu_extend_area_map(struct pcpu_chunk *chunk, int new_alloc)
 	size_t old_size = 0, new_size = new_alloc * sizeof(new[0]);
 	unsigned long flags;
 
+	lockdep_assert_held(&pcpu_alloc_mutex);
+
 	new = pcpu_mem_zalloc(new_size);
 	if (!new)
 		return -ENOMEM;
@@ -892,6 +894,9 @@ static void __percpu *pcpu_alloc(size_t size, size_t align, bool reserved,
 		return NULL;
 	}
 
+	if (!is_atomic)
+		mutex_lock(&pcpu_alloc_mutex);
+
 	spin_lock_irqsave(&pcpu_lock, flags);
 
 	/* serve reserved allocations from the reserved chunk if available */
@@ -964,12 +969,9 @@ static void __percpu *pcpu_alloc(size_t size, size_t align, bool reserved,
 	if (is_atomic)
 		goto fail;
 
-	mutex_lock(&pcpu_alloc_mutex);
-
 	if (list_empty(&pcpu_slot[pcpu_nr_slots - 1])) {
 		chunk = pcpu_create_chunk();
 		if (!chunk) {
-			mutex_unlock(&pcpu_alloc_mutex);
 			err = "failed to allocate new chunk";
 			goto fail;
 		}
@@ -980,7 +982,6 @@ static void __percpu *pcpu_alloc(size_t size, size_t align, bool reserved,
 		spin_lock_irqsave(&pcpu_lock, flags);
 	}
 
-	mutex_unlock(&pcpu_alloc_mutex);
 	goto restart;
 
 area_found:
@@ -990,8 +991,6 @@ static void __percpu *pcpu_alloc(size_t size, size_t align, bool reserved,
 	if (!is_atomic) {
 		int page_start, page_end, rs, re;
 
-		mutex_lock(&pcpu_alloc_mutex);
-
 		page_start = PFN_DOWN(off);
 		page_end = PFN_UP(off + size);
 
@@ -1002,7 +1001,6 @@ static void __percpu *pcpu_alloc(size_t size, size_t align, bool reserved,
 
 			spin_lock_irqsave(&pcpu_lock, flags);
 			if (ret) {
-				mutex_unlock(&pcpu_alloc_mutex);
 				pcpu_free_area(chunk, off, &occ_pages);
 				err = "failed to populate";
 				goto fail_unlock;
@@ -1042,6 +1040,8 @@ static void __percpu *pcpu_alloc(size_t size, size_t align, bool reserved,
 		/* see the flag handling in pcpu_blance_workfn() */
 		pcpu_atomic_alloc_failed = true;
 		pcpu_schedule_balance_work();
+	} else {
+		mutex_unlock(&pcpu_alloc_mutex);
 	}
 	return NULL;
 }