Subject: Re: Fix found for nsphys!
To: Dave Sainty <dave@dtsp.co.nz>
From: John Charles Hayward <John.C.Hayward@wheaton.edu>
List: current-users
Date: 07/28/2000 16:25:43
We also noticed the bug in setting the PCR_FLINK100 bug when looking over
the code.  We tested with the correct setting of PCR_FLINK100 and without
and it did not seem to matter.  The magic bit 10 was the one which
resolved the problem.  This seems to agree with our experience that before
the changing of the magic bit 10 we had problems with all kernels from
1.4.x to 1.5current.  The problem with the PCR_FLINK100 bit had been fixed
in some of the later kernels.

The chip we have is the ns83840 and the documentation we were looking at
was the ns83840A.  Maybe there was a change in the chip from ns83840 to
the ns83840A?

johnh...
> > 
> > --- nsphy.c.orig        Wed Jul 26 13:03:43 2000
> > +++ nsphy.c     Wed Jul 26 14:38:06 2000
> > @@ -218,14 +218,17 @@
> >                  */
> >                 reg &= ~PCR_FLINK100;
> 
> Note that this line is incorrect in your sources and may be
> contributing to your problems in a big way.  "Current" has the
> correct:
> 
> reg |= PCR_FLINK100;
> 
> However, the fact that setting the reserved (and documented read-only)
> bits help is interesting :)
> 
> Cheers,
> 
> Dave
>