YDL 3.01 on Wallstreet - clock runs way too fast!

[Alexander Holst]

Hi

I recently onstalled YDL 3.0.1 from scratch on a 266 MHz Wallstreet
PowerBook. Everything went smooth, once the proper kernel arguments for
the display were found.

Now that the machine is up and running, I experience a very strange
behaviour: the system time under Linux is about twice the speed it
should! It takes about 30 minutes in real time, for the internal clock
to go through a full hour.

This is so fast (well, it's internal time ;), that it doesn't even make
sense to use NTP to keep the machine synchronized. How could I get
around this problem?

Her is the first part of the dmesg file:

Memory BAT mapping: BAT2=128Mb, BAT3=64Mb, residual: 128Mb
mem_pieces_remove: [375000,40d8fb) not in any region
Total memory = 192MB; using 512kB for hash table (at c0480000)
Linux version 2.4.22-2g (dburcaw@localhost.localdomain) (gcc version
3.2.2 20030217 (Yellow Dog Linux 3.0 3.2.2-2a)) #1 Tue Dec 2 22:48:31
MST 2003
Found a Gatwick mac-io controller, rev: 1, mapped at 0xfdc80000
Found a Heathrow mac-io controller, rev: 1, mapped at 0xfdc00000
PowerMac motherboard: PowerBook Wallstreet
CPU HID1 : 0xa0000000
Found Grackle (MPC106) PCI host bridge at 0x80000000. Firmware bus
number: 0->0
hose pci, pci_mem_offset: 00000000, start0: 80000000
  io_base_virt: fd3fd000, io_base_phys: fe000000, isa_mem_base: fd000000
PMU driver 2 initialized for PowerBook G3 Series, firmware: 0a
nvram: OF partition at 0x1800
nvram: XP partition at 0x1300
nvram: NR partition at 0x1400
On node 0 totalpages: 49152
zone(0): 49152 pages.
zone(1): 0 pages.
zone(2): 0 pages.
Kernel command line:  root=/dev/hda14 ramdisk_size=8192
video=atyfb:vmode:14,cmode:8
irq: fixed SCC on second controller (79,68,69)
irq: fixed media-bay on second controller (93)
irq: fixed floppy on second controller (83,65)
irq: fixed ide on second controller (78,67)
irq: secondary controller on irq 27
System has 128 possible interrupts
64 interrupts on main controller
GMT Delta read from XPRAM: 120 minutes, DST: on
via_calibrate_decr: ticks per jiffy = 63509 (381058 ticks)
Warning: real time clock seems stuck!
Console: colour dummy device 80x25
Calibrating delay loop... 202.75 BogoMIPS
Memory: 189380k available (2228k kernel code, 1152k data, 176k init, 0k
highmem)

What does "Warning: real time clock seems stuck!" mean? And shouldn't
the BogoMIPS be twice the CPU speed roughly? As for a 266 MHz machine,
it should be around 532, and not 202 as above.

On my 233MHz G3 desktop, the BogoMIPS are almost exactly twice the CPU
speed, as dmesg shows on that machine:

PowerMac motherboard: PowerMac G3 (Gossamer)
[ ... snipped ... ]
via_calibrate_decr: ticks per jiffy = 167079 (1002477 ticks)
Console: colour dummy device 80x25
Calibrating delay loop... 466.94 BogoMIPS


Could that be tha cause for the system time running away? How could it
be fixed?

Any help/ideas appreciated,
Alex



-- 
Alexander Holst
Hergulesse Design
ph: +49 [0]711 887 5195
fx: +49 [0]711 887 5196
<alex.holst [at] hergulesse [dot] de>