I would post a new patch which includes my RESETTING state earlier
patch + the cancel_work_sync which Sagi suggested after testing.
Sagi: Because my RESETTING patch earlier is subset of your untested
patch with cancel_work_sync, it would be logical to take a signed off
from you as well. May you review/ack/nack the patch? Feel free to
let me know if you want me to change it further or instead you want to
post as author.