Subject: Problems
To: None <port-arm32@NetBSD.ORG>
From: Jonathan Hunt <csuwg@dcs.warwick.ac.uk>
List: port-arm32
Date: 07/14/1998 00:21:07
Hi,

I've just installed RiscBSD 1.3.2 from the NetBSD site and I've got a few
problems.  The installation went well - ide drive on the internal
interface partitioned ok, sets transfered and installed ok, machine
rebooted ok off a native kernel, old home directories from my previous
1.2 RiscBSD installation copied back in, X windows sets installed
(from a 1.3alpha CD-ROM) and X windows working ok.  Then the problems
started.

I have an I3 EtherH network card which seemed to be working ok, i.e.
ping worked, but when I put any heavy load on it, ftping a file on
the local ethernet for example, I get *thousands* of errors similar
to the following:

eh: no carrier
EHDEBUG: Remote 16 wasn't finished : spl level 6
EHDEBUG: SetReg done : status 18 : type 16
EHDEBUG: COM_WRITE : About to SetReg ( EH_DATA_PORT, dummy );
ehstart: failed to load packet
eh: Interrupt not serviced
eh: Receive Error: packet received intact (this should happen)

I know there have been problems with etherh in the past, but I've never
had this much trouble. Is there something wrong here?  I've got
'ehbug' set in the bootloader. Is that right?


The next thing I tried was to run some of my old binaries that I'd
compiled under 1.2.  None of them worked, they all gave
segmentation fault errors.  Even the rc5des client did the same.

All the installed set binaries seemed to work though at first including
all the X stuff, xclock, xlogo, xterm, etc, though something seemed up
with xterm. If I ran it when no DISPLAY variable was set, it crashed
horribly:

Data abort: 'Permission error (page)' status = 00f address = 00000084 PC = 100a3d34
Process = f1580600 pid = 197 comm = xterm
CPSR=60000013 Traceback info (frame=f2b4bfb8)
fp->lr=f0127580 fp->pc=f0130714
fp->lr=f0126014 fp->pc=f0126fc8
Trapframe PC = 100a3d34
Trapframe SPSR = 60000010
Segmentation fault (core dumped)


I've tried compiling some stuff, but didn't get very far.  An emacs compile
died during the ./configure stage with a similar error to the above.


So I'm a bit stuck at the moment.  I've got a RiscBSD setup that looks
fine but which I can't use because I can't compile anything, and none
of my old binaries work.  I've tried doing an install of 1.3alpha from
the CD, but that produced the same problems.  Can anyone help?

My setup is: StrongArm(Early 200MHz) RiscPC, RiscOS 3.7, 32+16+2Mb RAM,
             406MB Harddrive (internal ide)  826 cylinders of which
                       24cyl = ADFS    128cyl = swap    674 cyl = root
             2.5Gig and 4Gig drives on a RapIDE interface (All ADFS)
             I3 EtherH network card


All the best,

-- 
Jonathan Hunt
xlcus@oaktree.co.uk