On 7/6/21 8:40 AM, Tetsuo Handa wrote:
> Commit 6cc8e7430801fa23 ("loop: scale loop device by introducing per
> device lock") re-opened a race window for NULL pointer dereference at
> loop_validate_file() where commit 310ca162d779efee ("block/loop: Use
> global lock for ioctl() operation.") has closed.
>
> Although we need to guarantee that other loop devices will not change
> during traversal, we can't take remote "struct loop_device"->lo_mutex
> inside loop_validate_file() in order to avoid AB-BA deadlock. Therefore,
> introduce a global lock dedicated for loop_validate_file() which is
> conditionally taken before local "struct loop_device"->lo_mutex is taken.
I'll queue this up for next weeks merging. Christoph, are you happy with
it at this point? Can't say it's a thing of beauty, but the problem does
seem real.
--
Jens Axboe
