>Yes! It's true -- RISC OS is dead in the Croft household! I haven't used
>RISC OS for the past fortnight, only RiscBSD, and I'm considering changing
>things so I have to hold down ALT to enter RISC OS!

Try the latest version of the bootloader in the 1.2-release dir.

You can use the bootloader config window to set a CMOS bit so the fast boot
code defaults to RiscBSD and ALT will override and give you RiscOS.

>I've been having great fun installing sets, compiling software, etc.
>(and doing all my *real* work in Unix too!), and now it's time to ask
>all the questions that have been bugging me...

Fire away ;-)

>Setup: ARM710, 3.2Gig IDE Quantum Fireball (600ish Mb for RiscBSD)
>Cumana SCSI II with proTeus, 32+8+2Mb RAM.
>Kernel bsd-4634-64MB. 1.2-beta sets (including cc2721)
oh, whats the 3.2 Quantum like ? I need a new drive ASAP and was thinking about
either the 3.2GB Quantum (10.5ms) or the 3GB IBM (9.5ms)

>1. Occasionally my current console goes berserk and starts inserting
>loads of ^@ characters... and keeps on adding them until I hit another key.
>Is this a known problem?

>I don't seem to be able to tie down exactly what is causing it - it happens
>both whilst I'm typing a command, and also when I'm at a prompt. It has
>even happened whilst in a MORE session! (At least, I assume it happened.
>I got the repeated bleeping of bad keypress even though I wasn't typing).

Yep this is actually the fault of the kbd device rather than the console.
It is a bug related to the timing problems when communicating with the keyboard
and sending resend commands.
It is on my fix it list.

>Speaking of which, is there anyway of turning down the bleep volume? :-)

Simplest way is the mkbeep binary in the local set

mkbeep -v 50

for example. There are other options (must write the man page) that allow the
frequency to be changed. What it does is recalculates a 4KB sine wave for
the beep generator.

>2. Compiling Pine. I would assume this is a RiscBSD-specific problem
>as I'm sure others would have noticed it if it wasn't...
>
>I am trying to compile Pine 3.95 for NEB (NetBSD). It compiles
>imapd, mtest, pico and pilot fine, then when it gets to compiling
>pine...
>
>134 root@hawk# make -f makefile.neb
>cc -O  -g -DDEBUG  -DNETBSD  -DSYSTYPE=\"NEB\" -DMOUSE  -c mailview.c
>mailview.c:In function `format_header_text':
>mailview.c:1387:internal error-unrecognised insn
>(insn 1078 1077 569 (set (reg:SI 3 r3)
>        (smin:SI (reg:SI 3 r3)
>            (const_int 1024))) -1 (nil)
>    (nil))
>cc: Internal compiler error: program cc1 got fatal signal 6
>*** Error code 1
Ok yagb (Yet Another Gcc Bug)

This is GCC at fault not pine.

Can you try preprocessing it

cc -O -g -DDEBUG -DNETBSD -DSYSTYPE=\"NEB\" -DMOUSE -c mailview.c -E
>mailview.i

and send me mailview.i
That way I can add it to the bug list and can test it against gcc patches.
NOTE: GCC bugs should be passed to me atm as NetBSD/arm32 is not yet supported
by GNU GCC or the GCC ARM maintainer.

Try also compiling without -O as these bugs to frequently tend to be due to the
optimiser.

>3. Will this new StrongARM kernel work on an ARM710? And does it fix
>the pmap_enter problem?

Nope. The SA kernels have the pmap_enter debug printing removed. I thin kI may
have put a new bsd kernel online with that debug message removed as well.

CUrrently the architecture difss between SA and ARM710 mean that a common
kernel is not possible yet but is a compile time option.

>5. I often get "Possible process deadlock due to shortage of L1 page tables".
>And, indeed, the process I initiated locks. My typical setup is 5
>console logins, 2 of them su'd, 1 run, 2 telnets, X, fvwm, xfm, xconsole, and
a
>Nedit sessions. This doesn't seem unreasonable for a 42MB machine, but if
>I go beyond this, I get the aforementioned message. Is this a fixed limit
>or can I increase something?

Increase MAX PROCESSES in the bootloader config window.

This value dictates the amount of memory reserved for L1 pages tables.
Fixing this limit is also on my todo list.

>6. I configured a swap partition when I installed, but there is no
>swap partition in /etc/fstab. Do I need to add it, or has it decided
>I have enough memory to not require a swap partition?

swapon will pick up the swap partition that was specified in the bootloader
or configured into a custom kernel.

>On a related note, what are all the wd0a-wd0e letters? wd0a is my /
>partition, wd0e is /etc. What are b, c, and d?

Partition c is the whole disc, partitions b, d, e, f, g, h are the other
partitions you could have. 8 partitions are supported and can be created in the
disklabel. By default the installer will use partition a for / and partition e
for /usr and partition b for swap.
As a said part is always refers to the whole disc i.e. overlaps parts a-h
partition d is normally used to describe the primary adfs partition on the
disc.
CUrrently this is not used but when adfs is available for RiscBSD ...
f, g, h can be assigned to other partitions you may have on the disc.

>7. Correction to the (original) installation guide: the command to mount a
>cdrom is mount -r -t cd9660 /dev/cd0a /cdrom

That looks right, I'll check what the guide says (probably a typo)

Hope this helps.

Cheers,
				Mark

-- 
Mark Brinicombe				amb@physig.ph.kcl.ac.uk
Research Associate			http://www.ph.kcl.ac.uk/~amb/
Department of Physics			tel: 0171 873 2894
King's College London			fax: 0171 873 2716