[pups] 2.11 on an 11/44

[pups] 2.11 on an 11/44

[Bill Gunshannon]

I am trying to put 2.11 on one of my 11/44's.  I picked up the RL02/RK
images from the archive.  I put the RL02 images on real RL02's (Yes,
some of us still have and use them!! :-)  My intent was to boot this
and then use it to build a system on a bigger disk and then go on from
there.

Here's my configuration:

  11/44 CPU
  CIS
  EIS
  FP11
  4M memory (actually 3840KB)
  MMU

  3 RL02 disks
  A CDU/720-TM SCSI Controller with 4 MAXTOR 340M disks and a QIC tape

Now the problem.
The system boots fine.  And it will mount /dev/rl1a.  But it won't mount
/dev/rl2a.  I get "/dev/rl2a on /vol2: No such device or address".
I get the same error when I try to do a disklabel.  What's more, I also
get this same error when I try to access any of the RA devices.  I am
using the GENERIC Kernel which I assume has all the devices in it.

Anybody have any suggestions??

Of course, if I find that the SCSI Controller isn't going to work
I have another controller and a FUJI Eagle I could use too. But I
am certain I will need to get all three RL's working in order to
have enough of a system to do this.

Thanks in advance,

bill

-- 
Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolves
bill at cs.scranton.edu     |  and a sheep voting on what's for dinner.
University of Scranton   |
Scranton, Pennsylvania   |         #include <std.disclaimer.h>

[pups] 2.11 on an 11/44

[David Evans]

On Thu, Feb 27, 2003 at 10:08:26AM -0500, Bill Gunshannon wrote:
> 
>   A CDU/720-TM SCSI Controller with 4 MAXTOR 340M disks and a QIC tape
> 

  You could also stick a bigger disk image (I think I used one for an RA80)
on a SCSI disk and go that way.

-- 
David Evans          (NeXTMail/MIME OK)             dfevans at bbcr.uwaterloo.ca
Ph.D. Candidate, Computer/Synth Junkie     http://bbcr.uwaterloo.ca/~dfevans/
University of Waterloo         "Default is the value selected by the composer
Ontario, Canada           overridden by your command." - Roland TR-707 Manual

[pups] 2.11 on an 11/44

[Johnny Billquist]

On Thu, 27 Feb 2003, Bill Gunshannon wrote:

[...]

> Now the problem.
> The system boots fine.  And it will mount /dev/rl1a.  But it won't mount
> /dev/rl2a.  I get "/dev/rl2a on /vol2: No such device or address".
> I get the same error when I try to do a disklabel.  What's more, I also
> get this same error when I try to access any of the RA devices.  I am
> using the GENERIC Kernel which I assume has all the devices in it.
> 
> Anybody have any suggestions??

What do you see at boot time? Does it show all three RL02s and the
mscp-controller along with disks?

Second, do you have the entries in /dev?

	Johnny

Johnny Billquist                  || "I'm on a bus
                                  ||  on a psychedelic trip
email: bqt at update.uu.se           ||  Reading murder books
pdp is alive!                     ||  tryin' to stay hip" - B. Idol

[pups] 2.11 on an 11/44

[Bill Gunshannon]

On Thu, 27 Feb 2003, Johnny Billquist wrote:

> On Thu, 27 Feb 2003, Bill Gunshannon wrote:
>
> [...]
>
> > Now the problem.
> > The system boots fine.  And it will mount /dev/rl1a.  But it won't mount
> > /dev/rl2a.  I get "/dev/rl2a on /vol2: No such device or address".
> > I get the same error when I try to do a disklabel.  What's more, I also
> > get this same error when I try to access any of the RA devices.  I am
> > using the GENERIC Kernel which I assume has all the devices in it.
> >
> > Anybody have any suggestions??
>
> What do you see at boot time? Does it show all three RL02s and the
> mscp-controller along with disks?

Actually, I am surprised by the lack of information provided.  I guess
the newer BSD's have spoiled me.  :-)

------------------
>>>boot dl0

(Program)
44Boot from rl(0,0,0) at 0174400
:
: rl(0,0,0)unix
Boot: bootdev=03400 bootcsr=0174400
2.11 BSD UNIX #5: Thu Jan 11 21:32:08 PST 1996
    sms1 at sms.sms.iipo.gtegsc.com:/usr/src/sys/GENERIC


phys mem  = 3932160
avail mem = 3719616
user mem  = 307200

February 29 23:09:36 init: configure system

rk ? csr 177400 vector 220 skipped:  No autoconfig routines.
rl 0 csr 174400 vector 160 attached
erase, kill ^U, intr ^C
#

------------------------

And that's it!!
I have commented out the RK entries in /etc/fstab as I certainly don't
have any real RK disks.  :-)

Next, I usually do:
---------------
# mount -a
/dev/rl2a on /vol2: No such device or address
#
-------------------

And so it goes.

I should probably mention again, these are the 2.11_on_rl02 images from
the archive.  It consists of 4 RL02's and 6 RK's and it boots and runs
fine under SIMH.

>
> Second, do you have the entries in /dev?
>

Maybe I'll try running MAKEDEV just to be sure there isn't something
corrupted, but being as the boot disk and vol1 both transfered and
copyed to real packs OK I have no reason to suspect there is anything
wrong with the data on the pack.  Hmmmmm... Maybe I'm just paranoid,
but I don;t think I will make this assumption.  I think I will make
a new vol2 pack anyway.

But that still leaves why I see no RA's either.

Any more suggestions?

Is there anyone running a webserver on a PDP-11??  Or will I be the
first when I get this going??  I was going to do it with Ultrix-11
but I think it may be easier to get up and definitely more stable on
2.11BSD.  Ultrix-11 still needs more work getting it back up to real
production quality. (But I plan on doing it!!)

All the best.

bill

-- 
Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolves
bill at cs.scranton.edu     |  and a sheep voting on what's for dinner.
University of Scranton   |
Scranton, Pennsylvania   |         #include <std.disclaimer.h>

[pups] 2.11 on an 11/44

[Johnny Billquist]

On Thu, 27 Feb 2003, Bill Gunshannon wrote:

> On Thu, 27 Feb 2003, Johnny Billquist wrote:
> 
> > What do you see at boot time? Does it show all three RL02s and the
> > mscp-controller along with disks?
> 
> Actually, I am surprised by the lack of information provided.  I guess
> the newer BSD's have spoiled me.  :-)

[stuff deleted...]

Hmmm. It's been a while since I ran 2.11, but I'm pretty sure you should
see something about the controller at boot. That it attaches, it's csr and
what vector it will use. The same way you observe the RL11 controller.

Seems like you don't have the controller available from the OS.
Have you used the controller anywhere else, so that you know it's
functioning, and that the CSR is set right?

About the RL drives, I guess we'll just have to assume that you don't have
more than two units generated in the system, unless you have some very
funny hardware problem.


Johnny Billquist                  || "I'm on a bus
                                  ||  on a psychedelic trip
email: bqt at update.uu.se           ||  Reading murder books
pdp is alive!                     ||  tryin' to stay hip" - B. Idol

[pups] 2.11 on an 11/44

[Bill Gunshannon]

On Thu, 27 Feb 2003, Johnny Billquist wrote:

> On Thu, 27 Feb 2003, Bill Gunshannon wrote:
>
> > On Thu, 27 Feb 2003, Johnny Billquist wrote:
> >
> > > What do you see at boot time? Does it show all three RL02s and the
> > > mscp-controller along with disks?
> >
> > Actually, I am surprised by the lack of information provided.  I guess
> > the newer BSD's have spoiled me.  :-)
>
> [stuff deleted...]
>
> Hmmm. It's been a while since I ran 2.11, but I'm pretty sure you should
> see something about the controller at boot. That it attaches, it's csr and
> what vector it will use. The same way you observe the RL11 controller.

Hmmmm.  Curious.  I looked at the conf for GENERIC and it had all the
disks.  Maybe I need to build a kernel under SIMH and then move it over
to the real box.  Should be a trip, I've never built a kernel under 2.x.
Wonder how close it is to todays BSD's where I do it al the time.

>
> Seems like you don't have the controller available from the OS.
> Have you used the controller anywhere else, so that you know it's
> functioning, and that the CSR is set right?

Yup.  Had RSTS on it when I got it and I tried RT-11 from an
RL and was able to init Packs, copy files and even boot as a DU.
HA LI from the RSTS Startup reports iit's offspring as RA8x disks.


>
> About the RL drives, I guess we'll just have to assume that you don't have
> more than two units generated in the system,

The same kernel mounts all 4 RL02's under SIMH.  :-(

>                                             unless you have some very
> funny hardware problem.

Maybe I need to crawl in back and check out everything.  Could the
terminator getting knocked loose cause this??  It's interesting that
it isn't a particular drive that doesn't whow up, it is the highest
numbered unit.  Unless I am looking at this wrong.  Do the numbers
assigned by the Unit Plug coincide with the dev number or does it
just assign devices sequentially for each drive it sees?  In other
words, if I had a unit with a 0 plug and a 3 plug would that show
up as /dev/rl0 and /dev/rl3 or as /dev/rl0 and /dev/rl1?
Hope that's understandable....


bill

-- 
Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolves
bill at cs.scranton.edu     |  and a sheep voting on what's for dinner.
University of Scranton   |
Scranton, Pennsylvania   |         #include <std.disclaimer.h>

[pups] 2.11 on an 11/44

[Bill Gunshannon]

Well, the fun continues. I guess it's not going to be as simple as
it is with newer BSD's.

When I try to build a custom kernel I get this:

---------------

# make|more
make -f Make.sys I=/usr/include H=../h M=../machine AS="/bin/as -V" CPP="/lib/cpp -P -DKERNEL -DUOFS -I. -I../h"  CFLAGS="-O -DKERNEL -DUOFS -I. -I../h" ED="/bin/ed"
cc -O -DKERNEL -DUOFS -I. -I../h -S ../sys/init_main.c
/bin/ed - < SPLFIX init_main.s
/bin/as -V - -o init_main.o init_main.s
rm -f init_main.s
cc -O -DKERNEL -DUOFS -I. -I../h -S ../sys/init_sysent.c
/bin/ed - < SPLFIX init_sysent.s
/bin/as -V - -o init_sysent.o init_sysent.s
rm -f init_sysent.s
cc -O -DKERNEL -DUOFS -I. -I../h -S ../sys/kern_acct.c
/bin/ed - < SPLFIX kern_acct.s
/bin/as -V - -o kern_acct.o kern_acct.s
rm -f kern_acct.s
cc -O -DKERNEL -DUOFS -I. -I../h -S ../sys/kern_clock.c
/bin/ed - < SPLFIX kern_clock.s
?
?
?
?
?
----------------------

And the "?" go on forever. Can't even break out of it.  Have to
kill the simulation and start all over.  Anybody run into this??
Seems to be in the clock code.  Is there something I might have
missed in the CONFIG file that could cause this??  All I basicly
did was comment out any hardware I was pretty sure wasn't going
to show up on my machine.

bill

-- 
Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolves
bill at cs.scranton.edu     |  and a sheep voting on what's for dinner.
University of Scranton   |
Scranton, Pennsylvania   |         #include <std.disclaimer.h>

[pups] 2.11 on an 11/44

[Robin Birch]

I haven't got a live 2.11 system anywhere near but the following occurs 
to me.

1) I think that the GENERIC kernel only has 2 RL02s configured and will 
expect these on device selects 0 and 1.  The idea is to get the generic 
going and then rebuild for your specific config.

2) Are all of your RL02s on one controller?.  If not then two things 
occur. The first is that I think the GENERIC kernel is only configured 
for one controller.  Secondly, if the second drive is on the second 
controller then the physical select will probably be 8 and the device 
will probably be rl8a.

3) If the SCSI controller emulates an MSCP device then I think that 2 
MSCP discs are configured for the GENERIC kernel.  Again, if you need 
more then you will need to rebuild the kernel.

Although the generic kernel is configured for a lot of devices they will 
not necessarily be attached if the appropriate file in /etc doesn't have 
the rems taken out.  Can't remember the file name but that will need the 
appropriate controllers to have the rems taken out.

I assume that you can mount rl0 and rl1 from your note.  Load the kernel 
sources onto these two discs, You should be able to get everything 
except the usr src onto the two rl02s.  Then create a kernel that has 
only the devices that you need and then you should see everything that 
you have.

The autoconfig system is not as informative as newer OS's but it will 
tell you if something is detected that the kernel is configured for.  If 
it exists but the kernel knows nothing about it then autoconfig will not 
display it.

If I am attempting to teach my grandmother to suck eggs then forgive me.

Cheers

Robin


In message <20030227093041.Y43050-100000 at server1.cs.uofs.edu>, Bill 
Gunshannon <bill at cs.scranton.edu> writes
>
>I am trying to put 2.11 on one of my 11/44's.  I picked up the RL02/RK
>images from the archive.  I put the RL02 images on real RL02's (Yes,
>some of us still have and use them!! :-)  My intent was to boot this
>and then use it to build a system on a bigger disk and then go on from
>there.
>
>Here's my configuration:
>
>  11/44 CPU
>  CIS
>  EIS
>  FP11
>  4M memory (actually 3840KB)
>  MMU
>
>  3 RL02 disks
>  A CDU/720-TM SCSI Controller with 4 MAXTOR 340M disks and a QIC tape
>
>Now the problem.
>The system boots fine.  And it will mount /dev/rl1a.  But it won't mount
>/dev/rl2a.  I get "/dev/rl2a on /vol2: No such device or address".
>I get the same error when I try to do a disklabel.  What's more, I also
>get this same error when I try to access any of the RA devices.  I am
>using the GENERIC Kernel which I assume has all the devices in it.
>
>Anybody have any suggestions??
>
>Of course, if I find that the SCSI Controller isn't going to work
>I have another controller and a FUJI Eagle I could use too. But I
>am certain I will need to get all three RL's working in order to
>have enough of a system to do this.
>
>Thanks in advance,
>
>bill
>

-- 
Robin Birch

[pups] 2.11 on an 11/44

[Fred N. van Kempen]

Bill,

kernel probably only supports one unit, _OR_ the filesystem doesnt
have all device nodes (/dev/rlXXX).

--f

> -----Original Message-----
> From: Bill Gunshannon [mailto:bill at cs.scranton.edu]
> Sent: Thursday, February 27, 2003 4:08 PM
> To: pups at minnie.tuhs.org
> Subject: [pups] 2.11 on an 11/44
> 
> 
> 
> I am trying to put 2.11 on one of my 11/44's.  I picked up the RL02/RK
> images from the archive.  I put the RL02 images on real RL02's (Yes,
> some of us still have and use them!! :-)  My intent was to boot this
> and then use it to build a system on a bigger disk and then go on from
> there.
> 
> Here's my configuration:
> 
>   11/44 CPU
>   CIS
>   EIS
>   FP11
>   4M memory (actually 3840KB)
>   MMU
> 
>   3 RL02 disks
>   A CDU/720-TM SCSI Controller with 4 MAXTOR 340M disks and a QIC tape
> 
> Now the problem.
> The system boots fine.  And it will mount /dev/rl1a.  But it 
> won't mount
> /dev/rl2a.  I get "/dev/rl2a on /vol2: No such device or address".
> I get the same error when I try to do a disklabel.  What's 
> more, I also
> get this same error when I try to access any of the RA devices.  I am
> using the GENERIC Kernel which I assume has all the devices in it.
> 
> Anybody have any suggestions??
> 
> Of course, if I find that the SCSI Controller isn't going to work
> I have another controller and a FUJI Eagle I could use too. But I
> am certain I will need to get all three RL's working in order to
> have enough of a system to do this.
> 
> Thanks in advance,
> 
> bill
> 
> -- 
> Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  
> Three wolves
> bill at cs.scranton.edu     |  and a sheep voting on what's for dinner.
> University of Scranton   |
> Scranton, Pennsylvania   |         #include <std.disclaimer.h>
> 
> _______________________________________________
> PUPS mailing list
> PUPS at minnie.tuhs.org
> http://minnie.tuhs.org/mailman/listinfo/pups
> 

[pups] 2.11 on an 11/44

[Bill Gunshannon]

On Thu, 27 Feb 2003, Fred N. van Kempen wrote:

> Bill,
>
> kernel probably only supports one unit, _OR_ the filesystem doesnt
> have all device nodes (/dev/rlXXX).

Same set of images (and kernel) works fine under SIMH.  Mounts 4 RL's
without a problem.  It is not till I put it on the real PDP that I have
the problem.

I should also mention it is not hardware as I can shift the Unit Plugs
around and the behavior is the same.  RL0 and RL1 mount, nothing else
is visible.

bill

-- 
Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolves
bill at cs.scranton.edu     |  and a sheep voting on what's for dinner.
University of Scranton   |
Scranton, Pennsylvania   |         #include <std.disclaimer.h>

[pups] 2.11 on an 11/44

[Steven M. Schultz]

Hi -

> From: Bill Gunshannon <bill at cs.scranton.edu>
> 
> I am trying to put 2.11 on one of my 11/44's.  I picked up the RL02/RK
> images from the archive.  I put the RL02 images on real RL02's (Yes,
> some of us still have and use them!! :-)  My intent was to boot this

	RL drives were reliable (compared to RA81 and RK06/7 drives) if not 
	exactly spacious ;)

>   3 RL02 disks
>   A CDU/720-TM SCSI Controller with 4 MAXTOR 340M disks and a QIC tape
> 
> Now the problem.
> The system boots fine.  And it will mount /dev/rl1a.  But it won't mount
> /dev/rl2a.  I get "/dev/rl2a on /vol2: No such device or address".

	What rev of the kernel do you have?   First line in /VERSION should
	have the magic number.

> I get the same error when I try to do a disklabel.  What's more, I also
> get this same error when I try to access any of the RA devices.  I am
> using the GENERIC Kernel which I assume has all the devices in it.
> 
> Anybody have any suggestions??
	
	What I think is happening is that the system is only configured for
	2 RL drives - the change from 2 to 4 happened quite late (patch #439
	I think).

	The RA problem sounds like the MSCP driver either isn't in the
	kernel or wasn't probed/attached at boot time.     You can see if
	MSCP support's present with something like 'nm -g /unix | grep _raintr"
	and seeing if you "040364 T _raintr" for the interrupt handler.   If
	you do get that then check /etc/dtab for a line like:

ra      ? 172150 0      5       raintr          # uda50, rqdx1/2/3

	The '0' for the vector says for the probe/attach logic to assign
	a vector and tell the controller what value was used.  You could put
	anything in there (154 or 150 I think is the assigned value for the
	first MSCP controller).

	If none of that works then I'm stumped as to why the RA drives can't
	be accessed.

	Good Luck.

	Steven Schultz

[pups] 2.11 on an 11/44

[Norman Wilson]

It seems to me it's time to go back to basics:

Tell us what happens if you do
	ls -l /dev/rl2a
and, if that file exists (as seems likely),
	od /dev/rl2a

Diagnosing a problem of this sort solely by the mutterings
of the mount command is a bit like trying to decide what
is causing your back pain and therefore how to treat it
by the tone of your voice when you say `ouch.'  (Hmm,
perhaps I have just invented a lucrative new paramedical
discipline, on a par with chiropractic and cold-laser
therapy and homeopathy and software consulting.)

Norman Wilson
Toronto ON

[pups] 2.11 on an 11/44

[Steven M. Schultz]

Hi -

> From: Bill Gunshannon <bill at cs.scranton.edu>
> Well, the fun continues. I guess it's not going to be as simple as
> 
> When I try to build a custom kernel I get this:
> 
> cc -O -DKERNEL -DUOFS -I. -I../h -S ../sys/kern_clock.c
> /bin/ed - < SPLFIX kern_clock.s
> ?
> ?
> ?
> ?
> ?
> ----------------------

> And the "?" go on forever. Can't even break out of it.  Have to
> kill the simulation and start all over.  Anybody run into this??

	I have an extremely vague memory I might have seen it eons upon 
	eons ago but I might be imaginging it.

> Seems to be in the clock code.  Is there something I might have

	Hmmm, '?' is "ed"s error indication.   I wonder if the '?' is
	coming out of ed and not SIMH?

> missed in the CONFIG file that could cause this??  All I basicly

	Not at this stage of the build - a config file error wouldn't allow
	the first few compiles to succeed and then start causing errors
	on kern_clock.c

	One thing I did notice though was the use of 'ed' - that tells me
	the patchlevel of the system is very low (i.e. old).   'ed' was
	replaced with 'sed' which sped things up a _lot_ - and that took
	place back around #325 in 1996.

	Have you tried P11 instead of SIMH?  When I'm too lazy to fire up
	the 11/73 I use P11 to do the testing/patching and so on.   It's a
	bit a pain to configure (really arcane configfile syntax) but it works
	very well - never had a problem with it.

	Good Luck.

	Steven Schultz