Am Mittwoch, den 11.03.2020, 06:25 +0000 schrieb Jonas Karlsson:
> Hi Fabio,
>
> > Hi Jonas,
> >
> > On Tue, Mar 10, 2020 at 1:07 PM Jonas Karlsson <jonas.karlsson@xxxxxxxx>
> > wrote:
> >
> > > I have also _reverted_ this patch after recommendation from NXP to avoid
> >
> > RCU stall
> > > crashes:
> > >
> > > commit 077506972ba23772b752e08b1ab7052cf5f04511
> > > Author: Paul E. McKenney <paulmck@xxxxxxxxxxxxxxxxxx>
> > > Date: Mon Jul 9 13:47:30 2018 -0700
> > >
> > > rcu: Make need_resched() respond to urgent RCU-QS needs
> >
> > Could you please test without this revert?
> >
> > Thanks
>
> I removed the revert and it still works fine.
Hi,
it is good that we have something that works.
It would be even better if we understood exactly how
it works. In fact that these patches work and are needed
may very well indicate that error handling on at least
some XHCs does not work as expected.
So a question and a request, if I may.
Did you run the test with autosuspend disabled? If so could
you retest with it enabled?
Secondly could you run tests with
commit 7c8f7af078a4eda73f347667d12584736e613062
Author: Oliver Neukum <oneukum@xxxxxxxx>
Date: Thu Mar 5 11:16:02 2020 +0100
cdc-acm: close race betrween suspend() and acm_softint
not applied (respectively reverted) with and without autosuspend?
Regards
Oliver
PS: When I submit upstream, may I add your 'Tested-by'?
