1.7 KiB
CVE-2023-52609
Description
In the Linux kernel, the following vulnerability has been resolved:binder: fix race between mmput() and do_exit()Task A calls binder_update_page_range() to allocate and insert pages ona remote address space from Task B. For this, Task A pins the remote mmvia mmget_not_zero() first. This can race with Task B do_exit() and thefinal 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 Aas TWA_RESUME. So in theory, Task A returns to userspace and the cleanupwork gets executed. However, Task A instead sleep, waiting for a replyfrom Task B that never comes (it's dead).This means the binder_deferred_release() is blocked until an unrelatedbinder event forces Task A to go back to userspace. All the associateddeath notifications will also be delayed until then.In order to fix this use mmput_async() that will schedule the work inthe corresponding mm->async_put_work WQ instead of Task A.
POC
Reference
No PoCs from references.