[PATCH 5.4 104/138] binder: Dont use mmput() from shrinker function.

From: Greg Kroah-Hartman
Date: Mon Jul 27 2020 - 10:18:20 EST


From: Tetsuo Handa <penguin-kernel@xxxxxxxxxxxxxxxxxxx>

commit f867c771f98891841c217fa8459244ed0dd28921 upstream.

syzbot is reporting that mmput() from shrinker function has a risk of
deadlock [1], for delayed_uprobe_add() from update_ref_ctr() calls
kzalloc(GFP_KERNEL) with delayed_uprobe_lock held, and
uprobe_clear_state() from __mmput() also holds delayed_uprobe_lock.

Commit a1b2289cef92ef0e ("android: binder: drop lru lock in isolate
callback") replaced mmput() with mmput_async() in order to avoid sleeping
with spinlock held. But this patch replaces mmput() with mmput_async() in
order not to start __mmput() from shrinker context.

[1] https://syzkaller.appspot.com/bug?id=bc9e7303f537c41b2b0cc2dfcea3fc42964c2d45

Reported-by: syzbot <syzbot+1068f09c44d151250c33@xxxxxxxxxxxxxxxxxxxxxxxxx>
Reported-by: syzbot <syzbot+e5344baa319c9a96edec@xxxxxxxxxxxxxxxxxxxxxxxxx>
Signed-off-by: Tetsuo Handa <penguin-kernel@xxxxxxxxxxxxxxxxxxx>
Reviewed-by: Michal Hocko <mhocko@xxxxxxxx>
Acked-by: Todd Kjos <tkjos@xxxxxxxxxx>
Acked-by: Christian Brauner <christian.brauner@xxxxxxxxxx>
Cc: stable <stable@xxxxxxxxxxxxxxx>
Link: https://lore.kernel.org/r/4ba9adb2-43f5-2de0-22de-f6075c1fab50@xxxxxxxxxxxxxxxxxxx
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>

---
drivers/android/binder_alloc.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

--- a/drivers/android/binder_alloc.c
+++ b/drivers/android/binder_alloc.c
@@ -948,7 +948,7 @@ enum lru_status binder_alloc_free_page(s
trace_binder_unmap_user_end(alloc, index);
}
up_read(&mm->mmap_sem);
- mmput(mm);
+ mmput_async(mm);

trace_binder_unmap_kernel_start(alloc, index);