Port-macppc archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Clock disciplining



I have a macppc machine whose clock is substantially out.  The macppc
FAQ pointed me to a port-macppc message from 2001 from Adam Lebsack
about how to fix this by bashing the timebase-frequency property on the
CPU.  (OF identifies itself as 1.0.5 when bannering after reset.)

This works for me; I've found a value that brings the clock well into
the range NTP can correct for.  But I'd rather not have to issue all
these commands before booting, especially since OF can't accept more
than a few bytes of input at full console baudrate.

I could, of course, hack on the kernel to make it ignore the
timebase-frequency property, using a hardwired value instead.  But I
consider that a last resort.

Based on my experience with Suns, and the presence of nvramrc and
use-nvramrc? properties in the OF, I tried these settings (excerpted
from printenv output):

use-nvramrc?        true                false

nvramrc             " /PowerPC,60?" find-device 0b1c37a encode-int " 
timebase-frequency" property 

These settings survive a reset-all, but they don't work: the
timebase-frequency property is still 00B34A70.  But, if I type the
nvramrc command manually at the ok prompt, it takes effect, as
expected.

Anyone have any idea what I'm doing wrong?  (And, yes, there is a
question mark in the CPU device name - it surprised me too, but it
works with dev and it works with find-device.)

/~\ The ASCII                             Mouse
\ / Ribbon Campaign
 X  Against HTML                mouse%rodents-montreal.org@localhost
/ \ Email!           7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B


Home | Main Index | Thread Index | Old Index