CVE-2023-54107

In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> blk-cgroup: dropping parent refcount after pd_free_fn() is done<br /> <br /> Some cgroup policies will access parent pd through child pd even<br /> after pd_offline_fn() is done. If pd_free_fn() for parent is called<br /> before child, then UAF can be triggered. Hence it&amp;#39;s better to guarantee<br /> the order of pd_free_fn().<br /> <br /> Currently refcount of parent blkg is dropped in __blkg_release(), which<br /> is before pd_free_fn() is called in blkg_free_work_fn() while<br /> blkg_free_work_fn() is called asynchronously.<br /> <br /> This patch make sure pd_free_fn() called from removing cgroup is ordered<br /> by delaying dropping parent refcount after calling pd_free_fn() for<br /> child.<br /> <br /> BTW, pd_free_fn() will also be called from blkcg_deactivate_policy()<br /> from deleting device, and following patches will guarantee the order.