[PATCH 4.4 115/137] drivers/hwspinlock: fix race between radix tree insertion and lookup

From: Greg Kroah-Hartman
Date: Tue Feb 23 2016 - 23:26:10 EST


4.4-stable review patch. If anyone has any objections, please let me know.

------------------

From: Matthew Wilcox <willy@xxxxxxxxxxxxxxx>

commit c6400ba7e13a41539342f1b6e1f9e78419cb0148 upstream.

of_hwspin_lock_get_id() is protected by the RCU lock, which means that
insertions can occur simultaneously with the lookup. If the radix tree
transitions from a height of 0, we can see a slot with the indirect_ptr
bit set, which will cause us to at least read random memory, and could
cause other havoc.

Fix this by using the newly introduced radix_tree_iter_retry().

Signed-off-by: Matthew Wilcox <willy@xxxxxxxxxxxxxxx>
Cc: Hugh Dickins <hughd@xxxxxxxxxx>
Cc: Ohad Ben-Cohen <ohad@xxxxxxxxxx>
Cc: Konstantin Khlebnikov <khlebnikov@xxxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
Signed-off-by: Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>

---
drivers/hwspinlock/hwspinlock_core.c | 4 ++++
1 file changed, 4 insertions(+)

--- a/drivers/hwspinlock/hwspinlock_core.c
+++ b/drivers/hwspinlock/hwspinlock_core.c
@@ -313,6 +313,10 @@ int of_hwspin_lock_get_id(struct device_
hwlock = radix_tree_deref_slot(slot);
if (unlikely(!hwlock))
continue;
+ if (radix_tree_is_indirect_ptr(hwlock)) {
+ slot = radix_tree_iter_retry(&iter);
+ continue;
+ }

if (hwlock->bank->dev->of_node == args.np) {
ret = 0;