Port-pc532 archive

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

Re: Netbsd & PC532 ... moving in small steps



Hi again Udo,

<evil top-posting mode>
First, I know that a dumpster is completely wrong for a PC532, which is why
it's not there yet.  The problem is that I'm not quite sure where the right
place is yet, but it is still someplace near my other computers...  My
biggest fear right now is that I used a battery-backed SRAM for the boot ROM,
and that was only supposed to last 10 years.  I don't have an EPROM or a
burner for it, and I might not even have the memory image.  So it might
actually have changed from a computer to a brick in the last few years, and
I am afraid to find out.

As for your progress, that is pretty impressive. As for the odd boot message,
I would expect it is a debug message in one of the drivers that got left in
by mistake, but I don't know. Second, the next thing to happen after the root system filetype is for it to start running init and all the /etc/rc.d scripts,
so you are very close to having it running.

For the "all 8 LUNs report" problem, as I said, back then, some drives did
that, some didn't, and there was a work around built into the kernel.  You
can leave the FPGA that way and add the SCSI driver quirk for your particular
drive, or alter the FPGA to correct that.

Next, while the PC532 board did not have a clock, almost all of them were
built with a small clock from Dallas Semiconductor that fit between the
motherboard and the EPROM chip.  If you sent the right sequence of reads
to it, it would return the time and date instead of EPROM data, but otherwise
would just pass the addresses and data through.  The kernel was written with
that in mind.

Finally, I should have copied this to the port-pc532 mailing list, as I am sure there are more people who might be able to help and would at least be interested
in seeing the progress of your work.  (Simon or Julian, are you there? :)
</evil top-posting mode>

Jon

udo.a.moeller%web.de@localhost wrote:
Hi Jon,

a dumpster is definitly not the right place for a good old piece of computer hardware ... I have some space available for it if you really want to dump it.

There was a bug in the SCSI emulation in my FPGA. Now it continues a little bit 
further :


NS32000 ROM Debugger
Version: Wed Nov 13 21:45:02 1991
Auto-boot enabled.
RAM free above 0x130c

Command (? for help): download 260000
CRC ok, length = 1474560
Command (? for help): run 3BE020

NetBSD/pc532 Boot, Revision 1.1
(jon%bullbox.bullers.net@localhost, Wed Jun 19 21:11:10 PDT 2002)
Boot: [[[sd0a:]netbsd][-abds]] :- md0a:/netbsd.gz
688128+2125824+134232 total=0x2d1c5c

Copyright (c) 1996, 1997, 1998, 1999, 2000, 2001
    The NetBSD Foundation, Inc.  All rights reserved.
Copyright (c) 1982, 1986, 1989, 1991, 1993
    The Regents of the University of California.  All rights reserved.


NetBSD 1.5.3_RC1 (INSTALL) #0: Wed Mar 13 23:28:17 PST 2002
    jon%bullbox.bullers.net@localhost:/usr/src/sys/arch/pc532/compile/INSTALL
total memory = 8192 KB
avail memory = 4568 KB
using 128 buffers containing 512 KB of memory
mainbus0 (root)
clock0 at mainbus0 addr 0xfffffe00, irq 2
ncr0 at mainbus0 addr 0xffd00000, irq 4
scsibus0 at ncr0: 8 targets, 8 luns per target
zero=fffcbfeb, bio=4010, net=34010, tty=34010, clock=34014, imp=34010, 
softclock=fffdffeb, softnet=34000
scsibus0: waiting 2 seconds for devices to settle...
sd0 at scsibus0 target 0 lun 0: <UMoeller, V-SCSI Harddisk, V1.0> SCSI2 
0/direct fixed
sd0: 81920 KB, 5120 cyl, 2 head, 16 sec, 512 bytes/sect x 163840 sectors
sd1 at scsibus0 target 0 lun 1: <UMoeller, V-SCSI Harddisk, V1.0> SCSI2 
0/direct fixed
sd1: could not mode sense (4/5); using fictitious geometry
sd1(ncr0:0:1): could not get size
sd1: 0, 0 cyl, 64 head, 32 sec, 512 bytes/sect x 0 sectors
sd2 at scsibus0 target 0 lun 2: <UMoeller, V-SCSI Harddisk, V1.0> SCSI2 
0/direct fixed
sd2: could not mode sense (4/5); using fictitious geometry
sd2(ncr0:0:2): could not get size
sd2: 0, 0 cyl, 64 head, 32 sec, 512 bytes/sect x 0 sectors
sd3 at scsibus0 target 0 lun 3: <UMoeller, V-SCSI Harddisk, V1.0> SCSI2 
0/direct fixed
sd3: could not mode sense (4/5); using fictitious geometry
sd3(ncr0:0:3): could not get size
sd3: 0, 0 cyl, 64 head, 32 sec, 512 bytes/sect x 0 sectors
sd4 at scsibus0 target 0 lun 4: <UMoeller, V-SCSI Harddisk, V1.0> SCSI2 
0/direct fixed
sd4: could not mode sense (4/5); using fictitious geometry
sd4(ncr0:0:4): could not get size
sd4: 0, 0 cyl, 64 head, 32 sec, 512 bytes/sect x 0 sectors
sd5 at scsibus0 target 0 lun 5: <UMoeller, V-SCSI Harddisk, V1.0> SCSI2 
0/direct fixed
sd5: could not mode sense (4/5); using fictitious geometry
sd5(ncr0:0:5): could not get size
sd5: 0, 0 cyl, 64 head, 32 sec, 512 bytes/sect x 0 sectors
sd6 at scsibus0 target 0 lun 6: <UMoeller, V-SCSI Harddisk, V1.0> SCSI2 
0/direct fixed
sd6: could not mode sense (4/5); using fictitious geometry
sd6(ncr0:0:6): could not get size
sd6: 0, 0 cyl, 64 head, 32 sec, 512 bytes/sect x 0 sectors
sd7 at scsibus0 target 0 lun 7: <UMoeller, V-SCSI Harddisk, V1.0> SCSI2 
0/direct fixed
sd7: could not mode sense (4/5); using fictitious geometry
sd7(ncr0:0:7): could not get size
sd7: 0, 0 cyl, 64 head, 32 sec, 512 bytes/sect x 0 sectors
md0: internal 2048K image area
boot device: <unknown>
root on md0a dumps on md0b
WARNING: bad date in battery clock -- CHECK AND RESET THE DATE!
root file system type: ffs

++++++++++++++++++++++++++++++

The only action after the last message is the timer interrupt every second. Do 
you know what should come next ?
What is meant by the WARNING : bad date in battery clock ? As far as I know 
there is no real-time clock chip in the PC532.
Do you have an explenation for the line with "zero=... bio=... net=... etc." ?

The SCSI device is emulated on the Host. Therefore I can define any device. My 
current setting is that any LUN other than 0
is not accepted. I made a test with the Quantum LPS525S with LUN 0 but it still 
reports the LUNs from 1 to 7 (see above).

Maybe Richard and myself are too optimistic : just emulate the hardware and the 
software will run without troubles. But
first results looked promising and therefore we will not give up.

I attached a photo of my CPU card. Dumping is not an option !
Regards,
Udo

_________________________________________
NEU: Mit WEB.DE DSL über 1000,- ¿ sparen!
http://produkte.web.de/go/02/



------------------------------------------------------------------------



Home | Main Index | Thread Index | Old Index