Help running 2.11 on Supnik 2.3d - TUHS

12 Apr 2000

I've tried running the image in /Boot_Images/2.11_on_rl02 on
Supnik's emulator version 2.3d, but the root filsystem seems
corrupted, with many binaries being unexecutable, bad block
error messages, etc.

Does anyone know how to successfully boot this image?

If this is the wrong place for this kind of question, or if
there's a FAQ on this, then please point me in the right
direction.

Received: (from major@localhost)
	by minnie.cs.adfa.edu.au (8.9.3/8.9.3) id IAA92766
	for pups-liszt; Thu, 13 Apr 2000 08:33:33 +1000 (EST)
...
 From Warren Toomey &lt;wkt(a)cs.adfa.edu.au&gt; Thu Apr
13 08:33:16 2000 Received: from henry.cs.adfa.edu.au (henry.cs.adfa.edu.au
[131.236.21.158])
	by minnie.cs.adfa.edu.au (8.9.3/8.9.3) with ESMTP id IAA92762
	for &lt;pups(a)minnie.cs.adfa.edu.au&gt;; Thu, 13 Apr 2000 08:33:27 +1000 (EST)
Received: (from wkt@localhost)
	by henry.cs.adfa.edu.au (8.9.2/8.9.3) id IAA06177;
	Thu, 13 Apr 2000 08:33:16 +1000 (EST)
	(envelope-from wkt)
From: Warren Toomey &lt;wkt(a)cs.adfa.edu.au&gt;
Message-Id: &lt;200004122233.IAA06177(a)henry.cs.adfa.edu.au&gt;
Subject: Re: Help running 2.11 on Supnik 2.3d
In-Reply-To: &lt;85ln2jsewz.fsf(a)junk.nocrew.org&gt; from lars brinkhoff at "Apr 12,
2000  2: 5:16 pm"
To: lars(a)nocrew.org (lars brinkhoff)
Date: Thu, 13 Apr 2000 08:33:16 +1000 (EST)
Cc: pups(a)minnie.cs.adfa.edu.au (Unix Heritage Society)
Reply-To: wkt(a)cs.adfa.edu.au
X-Mailer: ELM [version 2.4ME+ PL43 (25)]
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Sender: owner-pups(a)minnie.cs.adfa.edu.au
Precedence: bulk

In article by lars brinkhoff:
...
  I've tried running the image in
/Boot_Images/2.11_on_rl02 [from the PUPS
 Archive] on Supnik's emulator version 2.3d, but the root filsystem seems
 corrupted, with many binaries being unexecutable, bad block
 error messages, etc.

 Does anyone know how to successfully boot this image? 
Hmm, if I get a chance I'll try it here. Has anybody used this image
successfully?

	Warren

Received: (from major@localhost)
	by minnie.cs.adfa.edu.au (8.9.3/8.9.3) id NAA94344
	for pups-liszt; Thu, 13 Apr 2000 13:23:56 +1000 (EST)
...
 From "Steven M. Schultz"
&lt;sms(a)moe.2bsd.com&gt; Thu Apr 13 13:22:56 2000 
Received: from moe.2bsd.com
(0(a)MOE.2BSD.COM [206.139.202.200])
	by minnie.cs.adfa.edu.au (8.9.3/8.9.3) with ESMTP id NAA94340
	for &lt;pups(a)minnie.cs.adfa.edu.au&gt;; Thu, 13 Apr 2000 13:23:38 +1000 (EST)
Received: (from sms@localhost)
	by moe.2bsd.com (8.9.3/8.9.3) id UAA16968;
	Wed, 12 Apr 2000 20:22:56 -0700 (PDT)
Date: Wed, 12 Apr 2000 20:22:56 -0700 (PDT)
From: "Steven M. Schultz" &lt;sms(a)moe.2bsd.com&gt;
Message-Id: &lt;200004130322.UAA16968(a)moe.2bsd.com&gt;
To: lars(a)nocrew.org, pups(a)minnie.cs.adfa.edu.au
Subject: Re: Help running 2.11 on Supnik 2.3d
Sender: owner-pups(a)minnie.cs.adfa.edu.au
Precedence: bulk

Hi --

...
  From: lars brinkhoff &lt;lars(a)nocrew.org&gt;
 I've tried running the image in /Boot_Images/2.11_on_rl02 on
 Supnik's emulator version 2.3d, but the root filsystem seems
 corrupted, with many binaries being unexecutable, bad block error messages, 
 etc.

 Does anyone know how to successfully boot this image? 
	Yes.  The problem is not with the images (although the whole "on rl02"
	is a pain - you're far better off using the "xp" or large disk support
	that Bob added).

...
  If this is the wrong place for this kind of question,
or if
 there's a FAQ on this, then please point me in the right direction. 
	This is a very good place to ask this type of question.

	And now the moment you've been waiting for:  "the Answer" ;)

	2.11BSD is *very* upset at having to run in 256kb of memory.  since
	the kernel plus buffer cache (and other data structures) can easily
	exceed 200kb there is not enough memory left over to run a program.
	Programs such as 'fsck' are fairly large split I/D programs and won't
	fit in the remaining ~56kb or so.

	The quick fix to the problem is adding the line:

		set cpu 2048K

	to the config file before running the simulator.  That will give the
	simulated PDP-11 2Mb of memory which is a real nice size.  Oh, if
	memory is a concern on the system then "set cpu 1024K" will work
	well.  Since there's no networking involved 1Mb will be quite adequate.
	If you were using "P11" (the Begemot emulator) and had the full IP/TCP
	stack, etc then the kernel+networking+buffers can reach close to 400Kb
	and you might want to use 2Mb for the memory size.

	You might also look into the latest version (2.5) of the Begemot
	emulator.  The two key advantages of P11 are:  1) a emulated DEQNA so
	you can place the PDP11 on a network, 2) It keeps _good_ time
	(version 2.4 and earlier had severe timeskew when running compute
	bound programs, 2.5 is awesomely better and within range of "ntp"
	to keep the clock correct).  P11 also supports (as does Bob Supnik's
	simulator) large disks such as the RP06 which is much nicer than
	4 RL02s and a batch of RK05s.

	Hmmm, I'm not sure which rev level of 2.11BSD is in the "on_rl02"
	images - I hope it has the "bounce buffer support" to handle the 18bit
	RK controller on a 22bit bus...   If the RK images show corrupt or the
	kernel crashes then I would suspect the kernel is a bit too old.

	Unpack the "211bsd_on_rl02" images from the .gz images, edit the
	"script" file to increase the system memory and you should be all
	set to go.

	Steven

Received: (from major@localhost)
	by minnie.cs.adfa.edu.au (8.9.3/8.9.3) id OAA94566
	for pups-liszt; Thu, 13 Apr 2000 14:56:35 +1000 (EST)
...
 From Greg Lehey &lt;grog(a)lemis.com&gt; Thu Apr 13
14:56:23 2000 
Received: from freebie.lemis.com (root(a)freebie.lemis.com
[192.109.197.137])
	by minnie.cs.adfa.edu.au (8.9.3/8.9.3) with ESMTP id OAA94562
	for &lt;pups(a)minnie.cs.adfa.oz.au&gt;; Thu, 13 Apr 2000 14:56:26 +1000 (EST)
Received: (from grog@localhost)
	by freebie.lemis.com (8.9.3/8.9.0) id OAA45453
	for pups(a)minnie.cs.adfa.oz.au; Thu, 13 Apr 2000 14:26:23 +0930 (CST)
Date: Thu, 13 Apr 2000 14:26:23 +0930
From: Greg Lehey &lt;grog(a)lemis.com&gt;
To: UNIX Heritage Society &lt;pups(a)minnie.cs.adfa.oz.au&gt;
Subject: Early file system layouts (was: Splitting / and /usr)
Message-ID: &lt;20000413142623.A45386(a)freebie.lemis.com&gt;
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-Mailer: Mutt 1.0pre2i
WWW-Home-Page: http://www.lemis.com/~grog
X-PGP-Fingerprint: 6B 7B C3 8C 61 CD 54 AF  13 24 52 F8 6D A4 95 EF
Organization: LEMIS, PO Box 460, Echunga SA 5153, Australia
Phone: +61-8-8388-8286
Fax: +61-8-8388-8725
Mobile: +61-41-739-7062
Sender: owner-pups(a)minnie.cs.adfa.edu.au
Precedence: bulk

Saw this on a NetBSD list.

Greg

----- Forwarded message from "Alistair G. Crooks" &lt;agc(a)ftp.netbsd.org&gt;
-----

...
  Date: Wed, 12 Apr 2000 01:31:41 -0700 (PDT)
 To: kre(a)munnari.oz.au
 Cc: current-users(a)netbsd.org
 Precedence: list
 Delivered-To: current-users(a)netbsd.org

 Robert,

 [Off topic warning]

   Did you
not know that /usr was split off only because the disks were too
 small to keep everything on one way back in the early days 
 That's how I heard it too - but this split must have occurred way back
 very early in the days before anyone outside Bell Labs had ever heard of
 unix (as I remember it, even the CACM paper had /usr in it). 
 I believe that the topic of splitting / and /usr was discussed
 at the Glasgow University meeting of the UKUUG, which was around
 1978, if my memory serves me well. As the first copy of V5 and V6
 came out of the labs in the 1975/1976 timeframe, I suspect it came
 later. I don't have my copy of the CACM paper to hand, so I can't
 check dates.

 Whilst I attented the University there at that time, I didn't attend
 the conference - more fool me.

 I suspect that Alistair Kilgour or Zdravko Podolski could provide
 more information, or any of the Bell Labs alumni who were there.

 agc 
----- End forwarded message -----
----- Forwarded message from Robert Elz &lt;kre(a)munnari.OZ.AU&gt; -----

...
  To: "Alistair G. Crooks"
&lt;agc(a)ftp.netbsd.org&gt;
 Cc: current-users(a)netbsd.org
 Date: Thu, 13 Apr 2000 10:36:15 +1000
 Precedence: list
 Delivered-To: current-users(a)netbsd.org

     Date:        Wed, 12 Apr 2000 01:31:41 -0700 (PDT)
     From:        "Alistair G. Crooks" &lt;agc(a)ftp.netbsd.org&gt;
     Message-ID:  &lt;200004120831.BAA06860(a)nbftp.isc.org&gt;

  [Off topic warning] 
 Ditto - but recording history sometimes has its uses...

  I believe that the topic of splitting / and /usr
was discussed
 at the Glasgow University meeting of the UKUUG, which was around
 1978, 
 It was definitely done before that.

 Unfortunately, I can't find a 5th edition manual (or even a reprinted
 facsimilie thereof at the minute), but the 6th edition manual for sh(1)
 says ...

 	If the first argument [ on a command line ] is the name of an
 	executable file, it is invoked; otherwise the string `/bin' is
 	prepended to the argument.  (In this way most standard commands,
 	which reside in `/bin', are found.)  If no such command is found,
 	the string `/usr' is further prepended (to give `/usr/bin/command')
 	and another attempt is made to execute the resulting file.  (Certain
 	lesser-used commands live in `/usr/bin'.)

 The sixth edition manual is dated May 75, but the date on the sh man page
 is 5/15/74 (which I interpret as the 15th of May, 1974).

 For those who are new to unix (within the last 20 years) note that there
 was no notion of a user settable path...

  I don't have my copy of the CACM paper to
hand, so I can't
 check dates. 
 I have checked now, and it says nothing either way, so that is no help.
 Kernighan's "Unix for Beginners" (of a generally similar vintage) gives
 a diagrammatic view of the filesystem tree, in which all that exists in
 /usr are user directories, though that is not really conclusive.

  I suspect that Alistair Kilgour or Zdravko
Podolski could provide
 more information, or any of the Bell Labs alumni who were there. 
 I will see if Dennis will tell me...

 kre

----- End forwarded message -----
----- Forwarded message from Robert Elz &lt;kre(a)munnari.OZ.AU&gt; -----

...
  To: "Alistair G. Crooks"
&lt;agc(a)ftp.netbsd.org&gt;, current-users(a)netbsd.org
 Date: Thu, 13 Apr 2000 12:45:05 +1000
 Precedence: list
 Delivered-To: current-users(a)netbsd.org

     Date:        Thu, 13 Apr 2000 10:36:15 +1000
     From:        Robert Elz &lt;kre(a)munnari.OZ.AU&gt;
     Message-ID:  &lt;353.955586175(a)munnari.OZ.AU&gt;

  I will see if Dennis will tell me... 
 He did, ...

   But early; definitely by the time of the "nsys" system, the
   first C version, which was 1973.

 And ...

   The point of /usr/bin was really to find a place to put
   those binaries.  (The .5MB disk was pretty cramped even
   with with two of them, as we later had).

 which (if we ever needed it) is confirmation or the original reason...

 kre

----- End forwarded message -----

--
Finger grog(a)lemis.com for PGP public key
See complete headers for address and phone numbers

Received: (from major@localhost)
	by minnie.cs.adfa.edu.au (8.9.3/8.9.3) id PAA94737
	for pups-liszt; Thu, 13 Apr 2000 15:32:38 +1000 (EST)
...
 From Warren Toomey &lt;wkt(a)cs.adfa.edu.au&gt; Thu Apr
13 15:32:02 2000 Received: from henry.cs.adfa.edu.au (henry.cs.adfa.edu.au
[131.236.21.158])
	by minnie.cs.adfa.edu.au (8.9.3/8.9.3) with ESMTP id PAA94732
	for &lt;pups(a)minnie.cs.adfa.edu.au&gt;; Thu, 13 Apr 2000 15:32:26 +1000 (EST)
Received: (from wkt@localhost)
	by henry.cs.adfa.edu.au (8.9.2/8.9.3) id PAA08492;
	Thu, 13 Apr 2000 15:32:02 +1000 (EST)
	(envelope-from wkt)
From: Warren Toomey &lt;wkt(a)cs.adfa.edu.au&gt;
Message-Id: &lt;200004130532.PAA08492(a)henry.cs.adfa.edu.au&gt;
Subject: Re: Early file system layouts (/ and /usr split)
In-Reply-To: &lt;20000413142623.A45386(a)freebie.lemis.com&gt; from Greg Lehey at "Apr
13, 2000  2:26:23 pm"
To: grog(a)lemis.com (Greg Lehey)
Date: Thu, 13 Apr 2000 15:32:02 +1000 (EST)
Cc: current-users(a)netbsd.org
Reply-To: wkt(a)cs.adfa.edu.au
X-Mailer: ELM [version 2.4ME+ PL43 (25)]
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Sender: owner-pups(a)minnie.cs.adfa.edu.au
Precedence: bulk

[ This came to me from a NetBSD mailing list, via Greg Lehey ]

Someone said.....
...
  >>> Did you not know that /usr was split off
only because the disks were too
 >>> small to keep everything on one way back in the early days 
Someone else said....
...
  > I believe that the topic of splitting / and /usr
was discussed
 > at the Glasgow University meeting of the UKUUG, which was around
 > 1978, if my memory serves me well. As the first copy of V5 and V6
 > came out of the labs in the 1975/1976 timeframe, I suspect it came
 > later. I don't have my copy of the CACM paper to hand, so I can't
 > check dates. 
The answer is: UNIX had / and /usr split by at least the time of the
July 1974 CACM paper ``The UNIX Time-sharing system''.

Here is the evidence:

At http://minnie.cs.adfa.edu.au/PUPS/Images/ken-and-den.txt and
   http://minnie.cs.adfa.edu.au/PUPS/Images/ken-and-den.jpg

you will find a picture of Ken and Dennis at the PDP-11/20 around 1972.
The commentary in the text file from John Holden tell us that the disk
drives are RF-11 and RK03 drives.

RF-11 drives were fixed head drives with 512K of storage with fast access.
RK03s and RK05s could store 2M, but were not as fast as RF-11s.

The source code to (nearly) 3rd Edition UNIX, dated August 31, 1973, only
has drivers for two disks, RF-11s and RK05s. This source code is in the
PUPS Archive, http://minnie.cs.adfa.edu.au/PUPS. You need a Unix src license.

Now, the July 1974 CACM paper says this:

	In our installation, for example, the root directory resides
	on the fixed-head disk, and the large disk drive, which contains
	user's files, is mounted by the system initialization program; ....

To me, this strongly indicates that / and /usr were split by at least
July 1974, if not the earlier date of August 1973.

Cheers all,
	Warren

Received: (from major@localhost)
	by minnie.cs.adfa.edu.au (8.9.3/8.9.3) id PAA94865
	for pups-liszt; Thu, 13 Apr 2000 15:50:31 +1000 (EST)
...
 From Warren Toomey &lt;wkt(a)cs.adfa.edu.au&gt; Thu Apr
13 15:50:17 2000 Received: from henry.cs.adfa.edu.au (henry.cs.adfa.edu.au
[131.236.21.158])
	by minnie.cs.adfa.edu.au (8.9.3/8.9.3) with ESMTP id PAA94861
	for &lt;pups(a)minnie.cs.adfa.edu.au&gt;; Thu, 13 Apr 2000 15:50:24 +1000 (EST)
Received: (from wkt@localhost)
	by henry.cs.adfa.edu.au (8.9.2/8.9.3) id PAA08605;
	Thu, 13 Apr 2000 15:50:17 +1000 (EST)
	(envelope-from wkt)
From: Warren Toomey &lt;wkt(a)cs.adfa.edu.au&gt;
Message-Id: &lt;200004130550.PAA08605(a)henry.cs.adfa.edu.au&gt;
Subject: Re: Early file system layouts (/ and /usr split)
In-Reply-To: &lt;200004130532.PAA08492(a)henry.cs.adfa.edu.au&gt; from Warren Toomey at
"Apr 13, 2000  3:32: 2 pm"
To: wkt(a)cs.adfa.edu.au
Date: Thu, 13 Apr 2000 15:50:17 +1000 (EST)
Cc: grog(a)lemis.com, current-users(a)netbsd.org
Reply-To: wkt(a)cs.adfa.edu.au
X-Mailer: ELM [version 2.4ME+ PL43 (25)]
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Sender: owner-pups(a)minnie.cs.adfa.edu.au
Precedence: bulk

In article by Warren Toomey:
...
  The answer is: UNIX had / and /usr split by at least
the time of the
 July 1974 CACM paper ``The UNIX Time-sharing system''. 
I just found some more evidence. The 2nd Edition UNIX manual is dated
June 1972, but the actual man pages have their date of last modification.

The manual for init(7), dated 15th June 1972, says:

[ If console switches are set to 173030, a shell is attached to the
  console immediately, i.e single-user mode ]

	Otherwise, init does some housekeeping: the mode of each DECtape
	file is changed to [read-write] (in case the system crashed during
	a tap command); directory /usr is mounted on the RK0 disk; directory
	/sys is mounted on the RK1 disk.

Cheers,
	Warren

Received: (from major@localhost)
	by minnie.cs.adfa.edu.au (8.9.3/8.9.3) id PAA94952
	for pups-liszt; Thu, 13 Apr 2000 15:59:25 +1000 (EST)
...
 From Warren Toomey &lt;wkt(a)cs.adfa.edu.au&gt; Thu Apr
13 15:59:06 2000 Received: from henry.cs.adfa.edu.au (henry.cs.adfa.edu.au
[131.236.21.158])
	by minnie.cs.adfa.edu.au (8.9.3/8.9.3) with ESMTP id PAA94946
	for &lt;pups(a)minnie.cs.adfa.edu.au&gt;; Thu, 13 Apr 2000 15:59:16 +1000 (EST)
Received: (from wkt@localhost)
	by henry.cs.adfa.edu.au (8.9.2/8.9.3) id PAA08671;
	Thu, 13 Apr 2000 15:59:06 +1000 (EST)
	(envelope-from wkt)
From: Warren Toomey &lt;wkt(a)cs.adfa.edu.au&gt;
Message-Id: &lt;200004130559.PAA08671(a)henry.cs.adfa.edu.au&gt;
Subject: Re: Early file system layouts (/ and /usr split)
In-Reply-To: &lt;200004130550.PAA08605(a)henry.cs.adfa.edu.au&gt; from Warren Toomey at
"Apr 13, 2000  3:50:17 pm"
To: wkt(a)cs.adfa.edu.au
Date: Thu, 13 Apr 2000 15:59:06 +1000 (EST)
Cc: wkt(a)cs.adfa.edu.au, grog(a)lemis.com, current-users(a)netbsd.org
Reply-To: wkt(a)cs.adfa.edu.au
X-Mailer: ELM [version 2.4ME+ PL43 (25)]
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Sender: owner-pups(a)minnie.cs.adfa.edu.au
Precedence: bulk

In article by Warren Toomey:
...
  In article by Warren Toomey:
  The answer is: UNIX had / and /usr split by at
least the time of the
 July 1974 CACM paper ``The UNIX Time-sharing system''.  
 I just found some more evidence. 
And more, from the 1st Edition init(7) man page dated 3rd November, 1971.

	Directory usr is assigned via sys mount as resident on the RK disk.

and sys mount means the mount(2) system call.

Cheers,
	Warren