Task A calls binder_update_page_range() to allocate and insert pages on
a remote address space from Task B. For this, Task A pins the remote mm
via mmget_not_zero() first. This can race with Task B do_exit() and the
final mmput() refcount decrement will come from Task A.

  Task A            | Task B
  ------------------+------------------
  mmget_not_zero()  |
                    |  do_exit()
                    |    exit_mm()
                    |      mmput()
  mmput()           |
    exit_mmap()     |
      remove_vma()  |
        fput()      |

In this case, the work of ____fput() from Task B is queued up in Task A
as TWA_RESUME. So in theory, Task A returns to userspace and the cleanup
work gets executed. However, Task A instead sleep, waiting for a reply
from Task B that never comes (it's dead).

This means the binder_deferred_release() is blocked until an unrelated
binder event forces Task A to go back to userspace. All the associated
death notifications will also be delayed until then.

In order to fix this use mmput_async() that will schedule the work in
the corresponding mm->async_put_work WQ instead of Task A.

Fixes: 457b9a6f09f0 ("Staging: android: add binder driver")
Signed-off-by: Carlos Llamas <cmllamas@google.com>
---
 drivers/android/binder_alloc.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/android/binder_alloc.c b/drivers/android/binder_alloc.c
index c4d60d81221b..046c3ae9eb8f 100644
--- a/drivers/android/binder_alloc.c
+++ b/drivers/android/binder_alloc.c
@@ -271,7 +271,7 @@ static int binder_update_page_range(struct binder_alloc *alloc, int allocate,
 	}
 	if (mm) {
 		mmap_write_unlock(mm);
-		mmput(mm);
+		mmput_async(mm);
 	}
 	return 0;
 
@@ -304,7 +304,7 @@ static int binder_update_page_range(struct binder_alloc *alloc, int allocate,
 err_no_vma:
 	if (mm) {
 		mmap_write_unlock(mm);
-		mmput(mm);
+		mmput_async(mm);
 	}
 	return vma ? -ENOMEM : -ESRCH;
 }
-- 
2.42.0.869.gea05f2083d-goog

Carlos Llamas <cmllamas@google.com> writes:
> Task A calls binder_update_page_range() to allocate and insert pages on
> a remote address space from Task B. For this, Task A pins the remote mm
> via mmget_not_zero() first. This can race with Task B do_exit() and the
> final mmput() refcount decrement will come from Task A.
>
>   Task A            | Task B
>   ------------------+------------------
>   mmget_not_zero()  |
>                     |  do_exit()
>                     |    exit_mm()
>                     |      mmput()
>   mmput()           |
>     exit_mmap()     |
>       remove_vma()  |
>         fput()      |
>
> In this case, the work of ____fput() from Task B is queued up in Task A
> as TWA_RESUME. So in theory, Task A returns to userspace and the cleanup
> work gets executed. However, Task A instead sleep, waiting for a reply
> from Task B that never comes (it's dead).
>
> This means the binder_deferred_release() is blocked until an unrelated
> binder event forces Task A to go back to userspace. All the associated
> death notifications will also be delayed until then.
>
> In order to fix this use mmput_async() that will schedule the work in
> the corresponding mm->async_put_work WQ instead of Task A.
>
> Fixes: 457b9a6f09f0 ("Staging: android: add binder driver")
> Signed-off-by: Carlos Llamas <cmllamas@google.com>

Nice catch!

Reviewed-by: Alice Ryhl <aliceryhl@google.com>

The code really could use a comment, though.

Alice