On Mon, Sep 05, 2016 at 08:41:39AM -0400, Tejun Heo wrote: > On Sun, Sep 04, 2016 at 03:29:39AM +0200, Andreas Mohr wrote: > > Reason: any other [early-boot] invoker of cancel_delayed_work_sync() > > would hit the same issue, > > without any fix then available locally each. > > > > This may or may not be intentional. > > Just wanted to point it out. > > idk, invoking a blocking API from early boot is pretty special (as > with everything during early boot), so I think it's fine to handle > them as execeptions.
Yup, this sounds like the rationale that I would have expected. Andreas
