android: binder: fix lockdep check on clearing vma

When munmapping a vma, the mmap_lock can be degraded to a write before
calling close() on the file handle.  The binder close() function calls
binder_alloc_set_vma() to clear the vma address, which now has a lock dep
check for writing on the mmap_lock.  Change the lockdep check to ensure
the reading lock is held while clearing and keep the write check while
writing.

Link: https://lkml.kernel.org/r/20220627151857.2316964-1-Liam.Howlett@oracle.com
Fixes: 472a68df605b ("android: binder: stop saving a pointer to the VMA")
Signed-off-by: Liam R. Howlett <Liam.Howlett@oracle.com>
Reported-by: syzbot+da54fa8d793ca89c741f@syzkaller.appspotmail.com
Acked-by: Todd Kjos <tkjos@google.com>
Cc: "Arve Hjønnevåg" <arve@android.com>
Cc: Christian Brauner (Microsoft) <brauner@kernel.org>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Hridya Valsaraju <hridya@google.com>
Cc: Joel Fernandes <joel@joelfernandes.org>
Cc: Martijn Coenen <maco@android.com>
Cc: Suren Baghdasaryan <surenb@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

diff --git a/drivers/android/binder_alloc.c b/drivers/android/binder_alloc.c
index f555eebceef6bf4a026732baab10e1fcb636e2d1..1014beb128025706ac6d98d7e09a6030acb5c23f 100644 (file)
--- a/drivers/android/binder_alloc.c
+++ b/drivers/android/binder_alloc.c
@@ -315,12 +315,19 @@ static inline void binder_alloc_set_vma(struct binder_alloc *alloc,
 {
        unsigned long vm_start = 0;
 
+       /*
+        * Allow clearing the vma with holding just the read lock to allow
+        * munmapping downgrade of the write lock before freeing and closing the
+        * file using binder_alloc_vma_close().
+        */
        if (vma) {
                vm_start = vma->vm_start;
                alloc->vma_vm_mm = vma->vm_mm;
+               mmap_assert_write_locked(alloc->vma_vm_mm);
+       } else {
+               mmap_assert_locked(alloc->vma_vm_mm);
        }
 
-       mmap_assert_write_locked(alloc->vma_vm_mm);
        alloc->vma_addr = vm_start;
 }