carol saah

Hi,

I replaced an Intel P4 D845PESV motherboard
with an Intel P4 D865PERL motherboard -- as a
result of lightning that ended up going through
a working APC Smart-UPS 700 and knocking out the
motherboard/memory.

I had SCO OSR504 installed in that machine on an
Adaptec 29160 with 2 Seagate LVDs, which were
NOT affected by the hit.

/dev/boot is on hd0a and
/dev/root is on /dev/hd24. (Another HD off the 50-pin
port on the 29160 had been installed BEFORE the 2nd
LVD drive - so that's why OSR504 root is on /dev/hd24.)

I am using the ad160 driver that is for
OSR505 since there was none for OSR504 and it worked on OSR504. The system
was bootable both
from floppies and the hard disk with the Intel P4
D845PESV motherboard before the hit.

Solaris 9 is installed on the same hd as OSR504 root
and boots just fine with the new motherboard.
Microsoft Small Business Server is also on the 1st hd and
it also boots just fine.

Now, I cannot boot OSR504 from hd(40)unix nor from
fd(60)unix. I have tried 2 SCO OSR504 boot/root
floppies and here are the results:

When booting with fd(60)unix, boot consistently STOPS
at the 1st dot after the line:

Loading kernel fd(60)unix.Z .text
.. [stops here]

If I boot from the floppy but use the following bootstring
hd(40)unix swap=hd(161) dump=hd(161) root=hd(162),
then boot stops with 1 complete line of dots and about
half-way through the next line of dots after the following
line:

Loading kernel hd(40)unix.Z .text

If I boot from the hard disk, boot consistently stops after
4 complete lines of dots plus half of the 5th line.

If I try to BOOT ONLY from an SCO OSR505 boot
floppy withOUT the ad160 driver and with an OSR505
kernel from a Pentium 100 PC, the boot completes to
the root filesystem diskette request. I do NOT go any
further since the controller driver is NOT ad160 on the
OSR505 system and it isn't the same OS version.

If I disable cache in the bootstring, boot still stops
after the "Loading... .text" msg.

When I remove the Adaptec 29160 and the 2 LVD
drives and hook to a new Dell P4 Dimension 8250,
I have no problem booting/rooting from the hard drives.

The D865PERL motherboard is running with:

DDR333 -> single channel dynamic paging mode

Memory is 512 Mb (DDR400) single channel (setup)

Two serial-ATA ports are also available on this
motherboard, but not being used.

If I try to change the ATA/IDE config to legacy mode --
up to 2 IDE devices max, booting from the hd stops
sooner after the "Loading... .text" line.

Under chipset configuration, the following is set on the
Intel motherboard:

ISA Enable Bit - Enable
PCI Latency Timer 32
Extended config:
SDRAM Frequency Auto
CPC Override Auto
SDRAM Timing Control AUTO
SDRAM RAS Act. to Pre. 7
SDRAM CAS# LATENCY 2.5
SDRAM RAS# to CAS# delay 3
SDRAM RAS# Precharge 3

On the OSR504 system, I have many patches and
rs504c: oss470a,oss485a,oss471f and
oss469d,oss496a,oss621b,oss644b,oss642a,
oss480a,oss601a,oss472a,oss475a,oss459b,oss473a,
oss481a,oss605a,oss644b

If there is anything I can do to change the kernel (while
the disks are hooked to the Dell machine) or a bootstring
option to boot the SCO OSR504 on the new motherboard, please let me know.

I would prefer to work on the P4 rather than the Pentium
100.

I understand that I need to upgrade to OSR507 and
maybe that would solve my problem, but I need an
interim solution if possible.

Thanks in advance.

Carol Saah

bob bailin

Are you using one of the newer P4 processors (3.04GHz) that supports
hyperthreading?
If so, make sure hyperthreading is disabled in the BIOS.

carol saah

My processor is 2.53 GHz and the motherboard
manual indicates that that processor does not fall
into the processor category that features Hyper-
Threading.

I do not see a setting to disable hyper-threading in setup,
though. Maybe that is automatic depending on the
processor.

Thank you for your response.

Carol Saah

carol saah

Bela,

I just tried to boot my OSR504 and OSR505 boot diskettes in an
D875PBZLK machine. In both cases, the diskette was accessed
briefly and boot control passed to the next device in the boot
sequence.

I also want to mention that the Dell Dimension 8250 which I could
boot OSR504 both from OSR504 boot/root floppies and from the HD
off the Adaptec 29160 had problems with "doscp" and "dosdir".
The floppy drive is a 1.44 Mb drive. When I tried to use doscp, the result
was:

# doscp -r filename a:filename
# killed

# dosdir a:
# killed

# dosdir /dev/fd0135ds18
# killed

I am not planning on loading OSR504 in the Dell so I did not report
this earlier.

The Dell Dimension 8250 has:
bus speed 533 MHz
processor P4 2.40 GHz
and Dual channel memory: 2 RIMM_1 of 256 Mb each (setup)
Processor ID: F27 (I don't know what this means.)

Carol Saah

bela lubkin

Again, it looks like this is due to the BIOS enforcing the floppy boot
signature of 0x55 0xAA at address 0x1FE of a diskette. No other modern
BIOS that I am aware of does this.

I will send you private email with a possible fix for this.


Those messages make it look like your `doscp` and `dosdir` binaries are
corrupt. Has nothing to do with the floppy drive. If you run:

# dosdir
# dosdir c:

you'll see the same message. This is so abnormal as to make me think
there is something much more fundamentally wrong with the install.


Those are the three digits "family, model, stepping" returned from one
of the subfunctions of the `CPUID' instruction. Family 'F' is the
signature of the Pentium 4 family of processors. F27 is the right
signature for a Pentium 4 "Northwood" 478-pin 0.18 micron 533MHz FSB
CPU. (I _think_ those are the right code name and line size, but those
are just from memory...)

Basically, that's all fine _except_ that you're using OSR504 on a
Pentium 4, which is unsupported and expected to have trouble. The
various boot symptoms you're having do not appear to be CPU-related. I
am much less certain about the `doscp` problem. I think it very well
_could_ be related to running OSR504 on a CPU that shipped 5 years after
the OS.