cgroup: fix an off-by-one bug which may trigger BUG_ON()

commit 3ac1707 upstream.

The 3rd parameter of flex_array_prealloc() is the number of elements,
not the index of the last element.

The effect of the bug is, when opening cgroup.procs, a flex array will
be allocated and all elements of the array is allocated with
GFP_KERNEL flag, but the last one is GFP_ATOMIC, and if we fail to
allocate memory for it, it'll trigger a BUG_ON().

Signed-off-by: Li Zefan <lizefan@huawei.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

kernel/cgroup.c

@@ -2026,7 +2026,7 @@ int cgroup_attach_proc(struct cgroup *cgrp, struct task_struct *leader)
 	if (!group)
 		return -ENOMEM;
 	/* pre-allocate to guarantee space while iterating in rcu read-side. */
-	retval = flex_array_prealloc(group, 0, group_size - 1, GFP_KERNEL);
+	retval = flex_array_prealloc(group, 0, group_size, GFP_KERNEL);
 	if (retval)
 		goto out_free_group_list;