Hello everybody,

To continue my previous message above (I'm on a hp340c+ with 8 Mb RAM
running NetBSD 1.3.2 from a MO. All my devices are connected to the
HP-IB bus):


1. I can see that the netbsd file is about 1.2 Mb; why then, when I
boot, the computer says:

real mem = 8,300,000
avail mem = 4,700,000

Where do the extra 2 Mb go? Is there a way to change this and make more
memory available to my applications? And if it is possible, is it worth
doing it or not?


2. I have a HP7320 hard disk at code 0 and a Bering 5600 MO at code 1;
none is "supported" in this release, according to the list of supported
hardware in the installation file. However, I managed to install NetBSD
1.3.2 on the MO and boot from it. But it proved impossible to do that
with the hd. Concerning my hd, I had Pascal 3.22 previously installed on
it. I now use the Pascal utilities from another 50 MB hd at code 7 to
format the hd I try to install NetBSD to.

How come I managed to install and boot NetBSD to/from the MO? Hopefully,
I have the Bering 5600 MO manual and it contains a list of
specifications as follows:

Formatted capacity	295	322	MB
Sector size		512	1024	bytes
Track size		31	17	sectors
Logical cylinder size	8	8	tracks
Logical cylinders	2326	2312	cylinders
Physical heads		1	1
Rotation speed			2400
...

One can immediately notice that the track size with 512 bytes per sector
is not 34--i.e., the double of the value available with 1024 bytes per
sector--but 31, that is, slightly lower, and that there are more
cylinders with 512 bytes/sector than with 1024 bytes/sector!

I just disklabelled the MO as indicated in the list above and
installation worked. But I also have to mention an inconsistency between
these data and the MKHFS printout I got upon formatting the MO with the
Pascal mediainit utility:

Device				HP5600
Formatted capacity		322 MB
Sectors (1 kb/track)		17
Tracks per cylinder		4
Block size in bytes		8129
Fragment size in bytes		1024
Rotations per second		40
data bytes per inode		2048
Cylinders			4624
Total sectors			314432

So it says 4 tracks/cylinder instead of 8 and 4624 cylinders instead of
2312. Whom to believe? I wouldn't worry about it if I didn't have
problems with NetBSD installation to the hd--and I don't have any manual
for the hd, only the printout from MKHFS.

Pascal installation utility put an HFS filesystem on the hd and the
MKHFS program printed the following params:

Device				HP7320
Device size in 1Kb blocks	199640		(204.4 MB)
Sectors (1Kb) per track		23
Tracks per cylinder		70
Block size in bytes		8192
Fragment size in bytes		1024
Rotations per second		587
data bytes per inode		2048

I first attempted an installation with 46 sectors per track--it seemed
logical to double the number of sectors per track upon reducing by half
the number of bytes per sector. The Miniroot Installer said it
"successfully" copied the miniroot file system to rd0b but when I tried
to boot it said (as described in my previous message above) "rd0b:netbsd
-s \ boot: No such file or directory".

Then I tried to install from the MO, i.e., to boot from rd1b. Everything
went fine until the installation program tried to create and mount the
file system on rd0a. Here it said it found "hard" and "read/write"
errors at various sectors, if I understood well.

Given all these facts, and especially the discrepancy I found between
the data in the Bering 5600 manual and the actual printout I got upon
formatting it, I wonder if the numbers of sectors per track and tracks
per cylinder I used with disklabel for my hard disk (because they were
printed as such by MKHFS after formatting of the hd) are reliable and
what to try best if they are not. I cannot imagine that installation of
NetBSD to the hd that had HP Pascal previously installed on it is
impossible. Any suggestion is welcome.


3. I can't figure out how to mount my floppy disk. I looked in the
NetBSD FAQ, and in the *BSD FAQ, and I also tried to find the BSD manual
(which I couldn't find; these file(s) seem to have vanished from
ftp://(ftp.)ocf.berkeley.edu/pub/Library/Computer/doc4.3) and I have
been unable to find out how. Maybe the answer to this question is so
plainly obvious and I'm unable to get it. But this doesn't change
anything for me. So I need help. I have Linux on an i686 and in this
case I mount /dev/floppy but the funny thing is that it seems to be a
"part" of the device containing the zip drive; the zip drive is just
/dev/hdc4. I imagine the system somehow figured out this dual feature of
the device and knows where and how to find the floppy.

Now I come to my floppy drive on the hp340 machine. Actually, it's in
the same module as the HP-IB hard disk (are all hp hard disks/floppy
drives like this?). In HP Pascal 3.22, the hd/floppy module is at code
7; the hard disk is logical number 11 and the floppy is logical number 3
(this number 3 is assigned automatically). But what do I do from NetBSD
when I have a hard disk at code 0 if I want to mount and access it's
floppy drive? I looked in /dev/*, there is no such device as /dev/floppy
or /dev/fd*.


4. A similar question about the printer. My computer came with a HP-IB
PaintJet which prints well from Pascal 3.22. But when I'm in NetBSD, if
I leave the printer on-line, its keeps on printing strange symbols
despite the fact that I didn't request anything to be printed. Actually,
I don't know how to print. I looked at the /etc/printcap file and
everything was commented so I uncommented the lp lines but device
/dev/lp doesn't exist in /dev/*, anyway. Do I have to create it--and
how? And why does the printer spontaneously print strange symbols?


Thank you very much for any help you can provide on any of the items
above.


--laurentiu
 
Laurentiu COCEA, M.D., M.Sc.
INSERM U.373, Faculte de Medecine Necker         cocea@necker.fr
156 rue de Vaugirard, 8e etage,               ph: (33)14061-5384
75730 Paris Cedex 15, France                  fax:(33)14061-5590