Antwort: Re: [ANN] U-Boot v2020.04-rc4 released

[Wolfgang Wallner]

Hi Andy,

> On Wed, Apr 1, 2020 at 2:54 PM Andy Shevchenko
> <andy.shevchenko at gmail.com> wrote:
> >
> > On Tue, Mar 31, 2020 at 2:31 AM Tom Rini <trini at konsulko.com> wrote:
> > >
> > > Hey all,
> > >
> > > So, I've gotten out of the habit of tagging rcs on cycle.  This month I
> > > have been taking in changes sparingly and I'm not concerned that things
> > > have gone in that wouldn't have, if I had tagged -rc4/rc5 on schedule.
> > >
> > > Things have been going to the -next branch as expected, and I think
> > > that will work out when things open up for v2020.07
> > >
> > > So here now is -rc4 and I plan to make the final release on April 6th,
> > > as scheduled.  Thanks all and stay safe out there!
> >
> > It becomes a tradition. Intel Edison is broken again (now serial driver).
> > This is show stopper.
> >
> > I'll become later with the culprit commit out of these three:
> >    This reverts commit 1c16606aac8f03be80d9c38ada00fd03123e0dbe.
> >    This reverts commit 62cbde4c4e462e5147d07cf1424492bb39e7bf94.
> >    This reverts commit 720f9e1fdb0c92d3fd16e1bfc25bcbd35612675c.
> 
> And the winner is...
> 
> 720f9e1fdb0c92d3fd16e1bfc25bcbd35612675c.
> 
> Wolfgang, please, fix ASAP, or revert ASAP!

What exactly does "broken" mean in this case?
What happens on Intel Edison?

I don't think just reversing commit 720f9e1fdb0c [1] is a good idea, as this
commit is itself a bug fix for commit 82de42fa1468 [2]. Commit 82de42fa1468
did break the serial driver on Intel Apollo Lake, see the description of
commit 720f9e1fdb0c for details on what happened on that platform.

regards, Wolfgang

[1] commit 720f9e1fdb0c ("serial: ns16550: Move PCI access from
    ofdata_to_platdata() to probe()")
[2] commit 82de42fa1468 ("dm: core: Allocate parent data separate from
    probing parent")