mikef@hpsad.HP.COM (Mike Ferrara) (09/01/88)
Does anybody out there have minix running on an HP Vectra? We've got a vectra ES/12 and a plain old vectra, (AT compatibles) neither of which will boot the minix 1.2 for AT/512K distribution from Prentice-Hall. During the boot, I get as far as the first menu, so I can fsck disks, and start loading the root file system, but upon the first disk access, I get the error: unexpected trap vector >=16 This may be due to accidentally including a non-MINIX library routine that is trying to make a system call. pc=0x154 text+data+bss=0x900 Any clues?? If somebody's fixed this problem, I'd sure like to get a copy of his bits. Mike Ferrara HP Signal Analysis Div Rohnert Park, CA mikef@hpsadle.hp.com (707) 794-4479
dsb@cidam.rmit.oz (David Burren [Athos]) (09/05/88)
In article <590001@hpsad.HP.COM>, mikef@hpsad.HP.COM (Mike Ferrara) writes: > Does anybody out there have minix running on an HP Vectra? > We've got a vectra ES/12 and a plain old vectra, (AT compatibles) neither > of which will boot the minix 1.2 for AT/512K distribution from Prentice-Hall. > During the boot, I get as far as the first menu, so I can fsck disks, and > start loading the root file system, but upon the first disk access, I get > the error: > > unexpected trap vector >=16 This may be due to accidentally including > a non-MINIX library routine that is trying to make a system call. > pc=0x154 text+data+bss=0x900 > I had similar problems when getting Minix first running on my AT-clown. It can run at four speeds, 6 or 12 MHz, with 0 or 1 wait-states. At any speed above 6 MHz 1 wait, it would work perfectly until just after the "RAM disk loaded" message, when it would get an unexpected trap. If I booted at that speed and then switched to any higher speed, it would get the unexpected trap message at the next command. I found that although the machine appeared to continue working, attempts to mount filesystems caused it to hang. As for the addresses it reported, they were different at each speed. (I can't at the moment remember exactly what they were). I've "fixed" this temporarily by removing the controller for my tape drive (a Sysgen QIC-File 60). At the moment it's running at top speed without a hitch (although it doesn't want to know about my hard-disk, which is another problem altogether). I'm trying to get a tech manual for the tape drive to investigate changing interrupt lines or somesuch. I know some other people here at RMIT who were getting unexpected traps until they changed the interrupt used by their mouse interface. Does the Vectra use any unusual interrupts? ----------------------------------------------------------------------- David S. Burren (Athos) ACSnet: dsb@cidam.oz UUCP: ...!uunet!munnari!cidam.oz!dsb ARPA: dsb%cidam.oz@uunet.uu.net
kcollinsthom@lion.waterloo.edu (Kevyn Collins-Thompson) (09/21/88)
> Does anybody out there have minix running on an HP Vectra?
Well, I also have attempted running MINIX 1.2 on a Vectra, but soon came to the
conclusion that it's not possible without a lot of work and more detailed
info about exactly how HP's Human Interface Loop works.
The Vectra uses its HP-HIL hardware to grab data from the keyboard, and then
attempts, through a chain of interrupts, to make it look like the keycode was
procured in standard PC-compatible fashion. Only problem is, HP-HIL uses
a couple extra non-standard interrupt vectors in this chain. (MINIX overwrites
these with a pointer to a trap routine, hence the TRAP message as soon as you
press the '=' key in the first menu). Special routines in the Enhanced
Vectra BIOS are also used, I believe, at points in the chain.
Modifying one line in the MINIX kernal so that these vectors are left alone
fixed the problem temporarily, but I was only able to use MINIX for a couple
of minutes before the machine inevitably hung. This isn't surprising,
considering the HP-HIL driver was still using the BIOS.
One true working solution to this HP-HIL mess would be a bigger mess:
to rewrite the HP-HIL driver as part of MINIX and get around the BIOS.
That would not only be a pain, but you would need the manuals containing
the proprietary info on HP-HIL.
Sorry about lack of details; I did this way back last March while working at HP.
There might be other problems in addition to those caused by HP-HIL, but this
is definately a major one.
------
Kevyn Collins-Thompson
...uunet!watmath!lion!kcollinsthomdrl@vuse.vanderbilt.edu (David R. Linn) (09/23/88)
Contact war@vuse.vanderbilt.edu for info on how to run MINIX on an old Vectra.