On Sat, Jul 09, 2016 at 10:24:59PM +0200, Mark Kettenis wrote:

> > Date: Sat, 9 Jul 2016 22:07:16 +0200
> > From: Marcus Glocker <[email protected]>
> > 
> > On every boot on my G5 the 'smu-pulsar' node hangs for around 66 seconds
> > in autoconfs config_search() it seems before the boot process continues:
> > 
> >     "smu-pulsar" at iic2 addr 0x6a not configured
> > 
> > I don't know why.
> > 
> > Anyone else seeing this with a 'smu-pulsar' node in his FDT?
> > Anything speaking against it to skip it from autoconf for now, similar
> > as we already do for 'deq' and 'tas3004'?
> 
> That makes no sense.  The delay is in no way related to this
> "smu-pulsar" device.  As the "not configured" message indicates, we
> don't have a driver for this device.  All your diff does is eliminate
> that message.  The delay will still be there.

With this diff the delay is gone.

But as said, I don't understand yet where it exactly happens.
 
> In all likelyhood, the delay will be caused by the device that comes next.
>
> 
> > Index: maci2c.c
> > ===================================================================
> > RCS file: /cvs/src/sys/arch/macppc/dev/maci2c.c,v
> > retrieving revision 1.11
> > diff -u -p -u -p -r1.11 maci2c.c
> > --- maci2c.c        23 May 2016 15:23:20 -0000      1.11
> > +++ maci2c.c        9 Jul 2016 19:55:04 -0000
> > @@ -59,6 +59,9 @@ maciic_scan(struct device *self, struct 
> >                the i2s port.  For now hide them.  */
> >             if (strcmp(name, "deq") == 0 || strcmp(name, "tas3004") == 0)
> >                     continue;
> > +           /* Skip the smu-pulsar node which delays the boot process. */
> > +           if (strcmp(name, "smu-pulsar") == 0)
> > +                   continue;
> >             if (ia.ia_name)
> >                     config_found(self, &ia, iic_print);
> >     }
> > 
> > 

Reply via email to