30 Nov
2013
30 Nov
'13
9:23 a.m.
Hi,
I've been making progress with my pdp11/53 project. Once I figured out
how to get into Emulex F.R.D. mode from ODT, I was able to configure,
format, and test my QD01 connected MFM hard drive.
I'd like to try installing 2.11BSD. Although I've got a TK50 drive
and controller, I don't have any tapes, or anything else here that
can write to them.
I have a Quantum DLT8000 SCSI drive, and an Emulex UC07 scsi controller.
I was able to set the UC07 as the first TMSCP controller, and F.R.D on
the UC07 reports the DLT8000 drive, and tests writing and reading to
tapes loaded in the drive with no errors.
I followed the 2.11BSD setup.ps and HOWTO file instructions to create
a tape with dd, using a Sun Solaris SPARC server connected to the DLT8000.
It seems silly trying to install an OS on a 70MB MFM drive from an
80GB DLT tape, but that's what I've got to work with so far, and my
UC07 can't do TMSCP and MSCP at once, so no SCSI disks if I'm installing
from SCSI tape.
Back on the 11/53, if I try to "BOOT MU0", the DLT tape spins for a few
seconds encouragingly, but then I get:
KDJ11-D/S E.03
Media not bootable
Should I be able to boot the 2.11BSD boot block on an 11/53 from a
TMSCP connected SCSI tape drive? The BSD setup.ps file includes bootstrap
monitor examples for TM and TS tape drives. Do I need to use one of
these instead of the normal 11/53 bootstrap ROM? Or, maybe there is
something about block sizes and writing the tapes from Sun Solaris SPARC
dd that means they aren't compatible?
Alternate paths I could take might be installing 2.11BSD using VTserver,
or maybe using PUTR or an emulator to set up OS directly onto a SCSI hard
drive or zip drive for use on the 11/53, but I liked the idea of doing it
with tape if I can figure out how. I have successfully booted the 11/53
from an RT11 RX33 floppy I created with PUTR, for what
it's worth.
Mark
--
Mark G. Thomas (Mark at Misty.com)
30 Nov
30 Nov
8:39 a.m.
Is the UC07 supported by 2.11BSD? I know that when I considered
installing it on my PDP-11/73 that it had issues with my Viking QDT
SCSI Adapter.
As far as tape issues, I use a DEC TLZ06 4mm DAT on my PDP-11/73, as
well as a TK50. I was able to install RSTS/E from 4mm without any
problem, but couldn't install layered products from 4mm, for those I
had to use TK50's. I don't know if there could be any strange issues
like that or not. A DLT8000 drive is quite a bit different from a
TK50.
I'll check with a couple people I know, and see if they have any
idea. I am wondering if you can even boot from a DLT8000 drive.
BTW, is the drive HVD, or something else? I believe that all of the
DLT8000 drives we had were HVD.
Zane
At 12:23 PM -0500 11/30/13, Mark G Thomas wrote:
Hi,
I've been making progress with my pdp11/53 project. Once I figured out
how to get into Emulex F.R.D. mode from ODT, I was able to configure,
format, and test my QD01 connected MFM hard drive.
I'd like to try installing 2.11BSD. Although I've got a TK50 drive
and controller, I don't have any tapes, or anything else here that
can write to them.
I have a Quantum DLT8000 SCSI drive, and an Emulex UC07 scsi controller.
I was able to set the UC07 as the first TMSCP controller, and F.R.D on
the UC07 reports the DLT8000 drive, and tests writing and reading to
tapes loaded in the drive with no errors.
I followed the 2.11BSD setup.ps and HOWTO file instructions to create
a tape with dd, using a Sun Solaris SPARC server connected to the DLT8000.
It seems silly trying to install an OS on a 70MB MFM drive from an
80GB DLT tape, but that's what I've got to work with so far, and my
UC07 can't do TMSCP and MSCP at once, so no SCSI disks if I'm installing
from SCSI tape.
Back on the 11/53, if I try to "BOOT MU0", the DLT tape spins for a few
seconds encouragingly, but then I get:
KDJ11-D/S E.03
Media not bootable
Should I be able to boot the 2.11BSD boot block on an 11/53 from a
TMSCP connected SCSI tape drive? The BSD setup.ps file includes bootstrap
monitor examples for TM and TS tape drives. Do I need to use one of
these instead of the normal 11/53 bootstrap ROM? Or, maybe there is
something about block sizes and writing the tapes from Sun Solaris SPARC
dd that means they aren't compatible?
Alternate paths I could take might be installing 2.11BSD using VTserver,
or maybe using PUTR or an emulator to set up OS directly onto a SCSI hard
drive or zip drive for use on the 11/53, but I liked the idea of doing it
with tape if I can figure out how. I have successfully booted the 11/53
from an RT11 RX33 floppy I created with PUTR, for what it's worth.
Mark
--
Mark G. Thomas (Mark at Misty.com)
--
| Zane H. Healy | UNIX Systems Administrator |
| healyzh at aracnet.com | OpenVMS Enthusiast |
| | Photographer |
+----------------------------------+----------------------------+
| My flickr Photostream |
| http://www.flickr.com/photos/33848088 at N03/ |
| My Photography Website |
| http://www.zanesphotography.com |
10:10 a.m.
Hi,
On Sat, Nov 30, 2013 at 09:39:57AM -0700, Zane H. Healy wrote:
Is the UC07 supported by 2.11BSD? I know that when I
considered
installing it on my PDP-11/73 that it had issues with my Viking QDT
SCSI Adapter.
http://minnie.tuhs.org/PUPS/Setup/2.11bsd_setup.html
"MSCP disks and TMSCP tapes include SCSI drives attached to the RQZX1
controller on the PDP-11/93. MSCP disks and TMSCP tapes also include
SCSI drives attached to the Emulex UC07 or UC08 Q-BUS controllers on
Q-bus systems as well as the UC17 and UC18 controllers on UNIBUS systems."
But, I'm starting to think maybe the boostrap is more restrictive.
As far as tape issues, I use a DEC TLZ06 4mm DAT on my
PDP-11/73, as
well as a TK50. I was able to install RSTS/E from 4mm without any
problem, but couldn't install layered products from 4mm, for those I
had to use TK50's. I don't know if there could be any strange
issues like that or not. A DLT8000 drive is quite a bit different
from a TK50.
I'll check with a couple people I know, and see if they have any
idea. I am wondering if you can even boot from a DLT8000 drive.
BTW, is the drive HVD, or something else? I believe that all of the
DLT8000 drives we had were HVD.
My DLT8000 is LVD, and will run SE.
When connected to the pdp11/53, the "SE" LED lights on the terminator
plugged into the drive, The UC07 is also correctly identifying:
"Drive #1 is Online , QUANTUM , DLT8000"
The UC07 is successfully able to perform it's F.R.D. tape read and write tests.
When connected to the Sun, the drive operates in LVD mode, and tapes can be
written and read fine. I don't have a HVD controller in the Sun.
Mark
--
Mark G. Thomas (Mark at Misty.com)
4:12 p.m.
Mark G Thomas wrote:
Hi,
On Sat, Nov 30, 2013 at 09:39:57AM -0700, Zane H. Healy wrote:
Oh yes, the Quarter Inch SCSI Drives above 525 MB are working too...
Regards,
Holm
--
Technik Service u. Handel Tiffe, www.tsht.de, Holm Tiffe,
Freiberger Stra?e 42, 09600 Obersch?na, USt-Id: DE253710583
www.tsht.de, info at tsht.de, Fax +49 3731 74200, Mobil: 0172 8790 741
Is the UC07 supported by 2.11BSD? I know that
when I considered
installing it on my PDP-11/73 that it had issues with my Viking QDT
SCSI Adapter.
http://minnie.tuhs.org/PUPS/Setup/2.11bsd_setup.html
"MSCP disks and TMSCP tapes include SCSI drives attached to the RQZX1
controller on the PDP-11/93. MSCP disks and TMSCP tapes also include
SCSI drives attached to the Emulex UC07 or UC08 Q-BUS controllers on
Q-bus systems as well as the UC17 and UC18 controllers on UNIBUS systems."
But, I'm starting to think maybe the boostrap is more restrictive.
As far as tape issues, I use a DEC TLZ06 4mm DAT
on my PDP-11/73, as
well as a TK50. I was able to install RSTS/E from 4mm without any
problem, but couldn't install layered products from 4mm, for those I
had to use TK50's. I don't know if there could be any strange
issues like that or not. A DLT8000 drive is quite a bit different
from a TK50.
I'll check with a couple people I know, and see if they have any
idea. I am wondering if you can even boot from a DLT8000 drive.
BTW, is the drive HVD, or something else? I believe that all of the
DLT8000 drives we had were HVD.
My DLT8000 is LVD, and will run SE.
When connected to the pdp11/53, the "SE" LED lights on the terminator
plugged into the drive, The UC07 is also correctly identifying:
"Drive #1 is Online , QUANTUM , DLT8000"
The UC07 is successfully able to perform it's F.R.D. tape read and write tests.
When connected to the Sun, the drive operates in LVD mode, and tapes can be
written and read fine. I don't have a HVD controller in the Sun.
Mark
--
Mark G. Thomas (Mark at Misty.com) 
2 p.m.
If you can get a SCSI card for a PC, or a unix system with SCSI disk interface,
you are golden.
Here is what I did to get 2.11BSD over to my 11/44 with an Emulex UC18 SCSI
controller:
(1) Use SIMH to configure/build a working 2.11BSD system from scratch tapes, if
you desire.
(2) Or you can get pre-built 2.11BSD image from here:
http://www.ak6dn.dyndns.org/PDP-11/2.11BSD/
or other locations on the net as well. The image I built is current thru
the latest patches.
(3) Mount the SCSI disk on which you will be installing the 2.11BSD image on
your PC/linux box. Use 'dd' to blast the whole image (from (2)) to the disk.
(4) Dismount the SCSI disk from your PC/unix box, carry it over to your PDP-11,
install it, and boot it up.
Any way you can do a byte-for-byte copy of the SIMH .dsk file to your target
SCSI drive should probably work.
Don
On 11/30/2013 8:39 AM, Zane H. Healy wrote:
Is the UC07 supported by 2.11BSD? I know that when I
considered installing it
on my PDP-11/73 that it had issues with my Viking QDT SCSI Adapter.
As far as tape issues, I use a DEC TLZ06 4mm DAT on my PDP-11/73, as well as a
TK50. I was able to install RSTS/E from 4mm without any problem, but couldn't
install layered products from 4mm, for those I had to use TK50's. I don't
know if there could be any strange issues like that or not. A DLT8000 drive
is quite a bit different from a TK50.
I'll check with a couple people I know, and see if they have any idea. I am
wondering if you can even boot from a DLT8000 drive.
BTW, is the drive HVD, or something else? I believe that all of the DLT8000
drives we had were HVD.
Zane
At 12:23 PM -0500 11/30/13, Mark G Thomas wrote:
Hi,
I've been making progress with my pdp11/53 project. Once I figured out
how to get into Emulex F.R.D. mode from ODT, I was able to configure,
format, and test my QD01 connected MFM hard drive.
I'd like to try installing 2.11BSD. Although I've got a TK50 drive
and controller, I don't have any tapes, or anything else here that
can write to them.
I have a Quantum DLT8000 SCSI drive, and an Emulex UC07 scsi controller.
I was able to set the UC07 as the first TMSCP controller, and F.R.D on
the UC07 reports the DLT8000 drive, and tests writing and reading to
tapes loaded in the drive with no errors.
I followed the 2.11BSD setup.ps and HOWTO file instructions to create
a tape with dd, using a Sun Solaris SPARC server connected to the DLT8000.
It seems silly trying to install an OS on a 70MB MFM drive from an
80GB DLT tape, but that's what I've got to work with so far, and my
UC07 can't do TMSCP and MSCP at once, so no SCSI disks if I'm installing
from SCSI tape.
Back on the 11/53, if I try to "BOOT MU0", the DLT tape spins for a few
seconds encouragingly, but then I get:
KDJ11-D/S E.03
Media not bootable
Should I be able to boot the 2.11BSD boot block on an 11/53 from a
TMSCP connected SCSI tape drive? The BSD setup.ps file includes bootstrap
monitor examples for TM and TS tape drives. Do I need to use one of
these instead of the normal 11/53 bootstrap ROM? Or, maybe there is
something about block sizes and writing the tapes from Sun Solaris SPARC
dd that means they aren't compatible?
Alternate paths I could take might be installing 2.11BSD using VTserver,
or maybe using PUTR or an emulator to set up OS directly onto a SCSI hard
drive or zip drive for use on the 11/53, but I liked the idea of doing it
with tape if I can figure out how. I have successfully booted the 11/53
from an RT11 RX33 floppy I created with PUTR, for what it's worth.
Mark
--
Mark G. Thomas (Mark at Misty.com)
2:24 p.m.
On Sat, Nov 30, 2013 at 02:00:42PM -0800, Don North wrote:
(3) Mount the SCSI disk on which you will be installing
the 2.11BSD
image on your PC/linux box. Use 'dd' to blast the whole image (from
(2)) to the disk.
(4) Dismount the SCSI disk from your PC/unix box, carry it over to
your PDP-11, install it, and boot it up.
FWIW, this technique will also work with Ersatz-11 with the raw SCSI drive
mounted directly as a PDP-11 device (e.g. "MOUNT DU0: SCSI0:", on a Linux
system with /dev/sg*, or a DOS or Windows system with an ASPI driver), so you
can install directly on the SCSI disk with no copying. Raw SCSI disks aren't
subject to the 256 MB file limit so it's OK if it's bigger.
John Wilson
D Bit
3:11 p.m.
On Sat, Nov 30, 2013 at 5:00 PM, Don North <ak6dn at mindspring.com> wrote:
If you can get a SCSI card for a PC, or a unix system
with SCSI disk
interface, you are golden.
(2) Or you can get pre-built 2.11BSD image from here:
http://www.ak6dn.dyndns.org/PDP-11/2.11BSD/
or other locations on the net as well. The image I built is current
thru the latest patches.
If you use this image and you have a CQD-220M in your pdp-11, update boot
in block 0. The source for rauboot was patched, but it doesn't look like
the boot block on the disk image was updated.
4:10 p.m.
Don North wrote:
If you can get a SCSI card for a PC, or a unix system
with SCSI disk
interface, you are golden.
Here is what I did to get 2.11BSD over to my 11/44 with an Emulex UC18 SCSI
controller:
(1) Use SIMH to configure/build a working 2.11BSD system from scratch
tapes, if you desire.
(2) Or you can get pre-built 2.11BSD image from here:
http://www.ak6dn.dyndns.org/PDP-11/2.11BSD/
or other locations on the net as well. The image I built is current
thru the latest patches.
(3) Mount the SCSI disk on which you will be installing the 2.11BSD image
on your PC/linux box. Use 'dd' to blast the whole image (from (2)) to the
disk.
(4) Dismount the SCSI disk from your PC/unix box, carry it over to your
PDP-11, install it, and boot it up.
Any way you can do a byte-for-byte copy of the SIMH .dsk file to your
target SCSI drive should probably work.
Don
May be this works wit an CQD220 Controller, but I don't think that this
will work with an Emulex UC07.
The Emulex can partition the disk in some slices, the information where the
partitions are, is somwhere on the disk.
I wrote out Tapes with SIMH to a Unix file and converted those files with
the containing Block length information to a real tape (on some qarter Inch
Tapes with an Tandberg drive [525 MB and up, the smaller Tapes don't
support variable block lenghts]). Done this on FreeBSD.
This way I've created install tapes for 2.11 BSD, RT11, RSX11 and
XXDP.
Regards,
Holm
--
Technik Service u. Handel Tiffe, www.tsht.de, Holm Tiffe,
Freiberger Stra?e 42, 09600 Obersch?na, USt-Id: DE253710583
www.tsht.de, info at tsht.de, Fax +49 3731 74200, Mobil: 0172 8790 741
9:51 a.m.
I have neither an Emulex UC07 nor a DLT8000 so I cannot comment on
that exact configuration.
I have used Exabyte EXB-8200 8mm tape drives with CMD CQD-200 and
CQD-220 controllers to install 2.11BSD on an M7554-SD KDJ11-SD
(DECServer 550 CPU but with EPROMs replace with 11/53 version).
If you are manually creating the tape with DD make sure you pad the
tape blocks to the correct size.
If you grab the 2.11BSD distribution files from here to build your
tape the mtboot file size is 490 bytes:
http://www.tuhs.org/Archive/PDP-11/Distributions/ucb/2.11BSD/
That needs to get padded to a 512 byte block before writing it to tape.
If you look at what happens using maketape.c and maketape.data the
files are padded to multiples of 512-byte blocks before being written
to tape.
-Glen
3:01 p.m.
Hi,
Am 30.11.13 18:23, schrieb Mark G Thomas:
>I followed the 2.11BSD setup.ps and HOWTO file instructions to create
>a tape with dd, using a Sun Solaris SPARC server connected to the DLT8000.
On Sat, Nov 30, 2013 at 08:04:12PM +0100, Jochen Kunz wrote:
You need to use maketape.c to get the right block
structure.
Jochen
I used maketape.c instead of the dd instructions, and that worked perfectly.
Thanks everyone for the help!
Mark
--
Mark G. Thomas (Mark at Misty.com)
20 Dec
20 Dec
5:22 a.m.
New subject: usb to unibus - was installing 2.11BSD on a pdp11/53
On a semi related note. Is there a way to write RL02 packs
from a mac/pc/whatever? This would require some kind of
interface adapter. Has anyone ever done this? Has anyone
built a USB/unibus adapter? I realize that this might be
the extreme path to writing RL02 packs from a laptop, but
it would also allow other unibus devices to be attached.
--tom
6:58 a.m.
New subject: usb to unibus - was installing 2.11BSD on a pdp11/53
There are pci to unibus bridges. I think they sre made for
use with some emulator. Perhaps it was qbus.. i'm not sure
now.
/p
On Fri, Dec 20, 2013 at 07:22:24AM -0600, Tom Uban wrote:
On a semi related note. Is there a way to write RL02
packs
from a mac/pc/whatever? This would require some kind of
interface adapter. Has anyone ever done this? Has anyone
built a USB/unibus adapter? I realize that this might be
the extreme path to writing RL02 packs from a laptop, but
it would also allow other unibus devices to be attached.
--tom
8:51 a.m.
New subject: usb to unibus - was installing 2.11BSD on a pdp11/53
On Fri, Dec 20, 2013 at 03:58:48PM +0100, Pontus Pihlgren wrote:
There are pci to unibus bridges. I think they sre made
for
use with some emulator. Perhaps it was qbus.. i'm not sure
now.
Both exist, but unfortunately they're quite expensive ($3500 a pop).
I've got some other USB projects in the works and have given quite a bit
of thought to Unibus -- a general-purpose, emulator-compatible NPR-capable
interface would be impossible because of the slow 1-msec response time
(NPR masters time out if the PDP-11 memory doesn't respond within ~10 usec),
but some sort of oddball doodad that has its own on-board memory and isn't
compatible with anything (and so would need its own drivers for every possible
disk/tape) could be done.
An EPP parallel port might work a lot better as the host (they're still
available as PCIe cards which may be better anyway since ISA-like timing
would be too slow), although the interrupt response still might not be
fast enough to handle NPRs.
A direct USB-to-RL01/02 interface is on my to-do list but there's a bunch
of stuff higher up on the list. Eventually I'd like to build USB interfaces
for all the basic DEC disks, to make them look like USB flash disks, so that
they'll work with any OS or emulator or utility or just the Weenix "dd"
command. That way you can easily deal with foreign disks whether or not you
like the software "nest" that's been built for that interface (I'm
looking at
you, Catweasel!). Bunch of other stuff to finish debugging first though.
John Wilson
D Bit
9:05 a.m.
New subject: usb to unibus - was installing 2.11BSD on a pdp11/53
On Fri, Dec 20, 2013 at 11:51 AM, John Wilson <wilson at dbit.com> wrote:
A direct USB-to-RL01/02 interface is on my to-do list
but there's a bunch
of stuff higher up on the list.
That sounds like a neat thing to work on. Way back in the day, c. 1986
or so, I had an Amiga, and an RL01 on an RL8A on my PDP-8/a and
worked with RL02s on RL11s at work. We made 68000-based Unibus
and Qbus interfaces, so I naturally wondered what it would take to
adapt the Zorro Bus (or the 86-pin SOTS bus on the A1000, really)
to the Unibus so I could use an RL11 on it. We had all the necessary
components for a simple Amiga-Unibus adapter, but the Amiga doesn't
have a DMA controller, so NPRs added a dimension of complexity I
wasn't prepared to deal with.
A year later, I solved the problem by buying a "Wedge" from
Canada - it was an 8-bit ISA bridge that included AmigaDOS
drivers for the WX-1 MFM interface. 20MB for $500 total, and
a lot less work than trying to make a Unibus card work.
Still... I occasionally wish to be able to easily read and write
RL02 packs with fewer active parts between me and it (right
now, my "simple" solution is a modern machine, a serial port,
a PDP-11, and vtserver).
If you ever do work on a USB RL02 interface, I'll be happy
to help test it.
-ethan
23 Dec
23 Dec
8:22 a.m.
New subject: usb to unibus - was installing 2.11BSD on a pdp11/53
On Fri, Dec 20, 2013 at 12:05:34PM -0500, Ethan Dicks wrote:
[...] I naturally wondered what it would take to adapt
the Zorro Bus (or the
86-pin SOTS bus on the A1000, really) to the Unibus so I could use an RL11 on
it. We had all the necessary components for a simple Amiga-Unibus adapter,
but the Amiga doesn't have a DMA controller, [...]
That's because it didn't need one. Zorro I was little more than the CPU lines
being exposed on an edge connector. So you could tri-state the CPU and then
access memory directly without needing to jump through the hoops of a DMA
controller.
Your hardware effectively *is* the CPU for the duration, and there was at least
one CPU accelerator that exploited this ability, and didn't need one to void
the warranty by cracking open the case and pulling out the 68000 like some
accelerators.
(It would perhaps be entertaining to build one such accelerator that contains
a modern CPU running a 68k emulator.)
10:05 a.m.
New subject: usb to unibus - was installing 2.11BSD on a pdp11/53
On Mon, Dec 23, 2013 at 11:22 AM, Peter Corlett <abuse at cabal.org.uk> wrote:
On Fri, Dec 20, 2013 at 12:05:34PM -0500, Ethan Dicks
wrote:
I disagree. I used to make 68000-based serial cards in the 80s and 90s
(COMBOARD). The board had a host interface that implemented Unibus,
Qbus or VAXBI (3 different cards, naturally, but largely the same internally).
This board did not have an on-board DMA engine, so a software loop read
its RAM and transferred the words to the right addresses on its own bus
to initiate DMA cycles in the host. The tightest loop we could manage
(a couple of 68000 instructions) got us about 200KB/sec transfer rates,
far below what the host bus could manage (usually over 1MB/sec).
It was common in the 68000 days to put the DMA engines on the
peripherals, largely because they were expensive and not every
peripheral needed one (serial ports, for example). This is the opposite
of what IBM did with the PC where *every* PC had a multi-channel
DMA engine that any peripheral could use (floppy, hard disk, sound
card...)
[...] I naturally wondered what it would take to
adapt the Zorro Bus (or the
86-pin SOTS bus on the A1000, really) to the Unibus so I could use an RL11 on
it. We had all the necessary components for a simple Amiga-Unibus adapter,
but the Amiga doesn't have a DMA controller, [...]
That's because it didn't need one. Zorro I was little more than the CPU lines
being exposed on an edge connector. So you could tri-state the CPU and then
access memory directly without needing to jump through the hoops of a DMA
controller.
Your hardware effectively *is* the CPU for the duration, and there was at least
one CPU accelerator that exploited this ability, and didn't need one to void
the warranty by cracking open the case and pulling out the 68000 like some
accelerators.
While I don't disagree with the model you've described when building hardware
from scratch, that doesn't help attaching an
existing Unibus peripheral (RL11
disk controller) to the Zorro bus. *Some* bit of
hardware has to implement the
Unibus NPR mechanism or the RL11 can't do its job, and in the 68000 world,
we call that a DMA controller, whether it's a Motorola chip (MC68440 or
MC68450) or a wad of TTL. *That* is what always slowed my intentions
down to a crawl.
My first several SCSI interfaces for the Amiga were PIO and could only
transfer a couple hundred Kbytes/sec. I knew how they worked from my
time writing embedded 68000 code with transfer loops and was not
surprised at their speed.
-ethan
11:04 a.m.
New subject: usb to unibus - was installing 2.11BSD on a pdp11/53
On Mon, Dec 23, 2013 at 01:05:44PM -0500, Ethan Dicks wrote:
[...]
My first several SCSI interfaces for the Amiga were
PIO and could only
transfer a couple hundred Kbytes/sec. I knew how they worked from my time
writing embedded 68000 code with transfer loops and was not surprised at
their speed.
If you had contrived the hardware design to have suitable incomplete decoding,
you could have exploited the movem hack that copies 56 bytes in 72 cycles,
which would give you north of 5MB/s on a 7.09MHz CPU.
11:30 a.m.
New subject: usb to unibus - was installing 2.11BSD on a pdp11/53
On Mon, Dec 23, 2013 at 2:04 PM, Peter Corlett <abuse at cabal.org.uk> wrote:
On Mon, Dec 23, 2013 at 01:05:44PM -0500, Ethan Dicks
wrote:
[...]
I think you'd need a lot of external hardware to do that when your choices
for SCSI chips were the 5380 or 33C93 - 8-bit chips for an 8-bit SCSI bus
on a 16-bit 68000 bus that did byte transfers in a funny way (I've hung
$20,000 bus analyzers off of the 68K looking for a hardware bug the designer
claimed was a bug in my firmware - he swapped UDS and LDS and
because of how the 68K is built internally, somethings worked and somethings
failed... so I'm pretty sure I'm on firm ground saying it's not practical
to try to make MOVEM work with an 8-bit peripheral - anything is
possible, but elaborate schemes are often more expensive than just
doing it the right way in the first place).
-ethan
My first several SCSI interfaces for the Amiga
were PIO and could only
transfer a couple hundred Kbytes/sec. I knew how they worked from my time
writing embedded 68000 code with transfer loops and was not surprised at
their speed.
If you had contrived the hardware design to have suitable incomplete decoding,
you could have exploited the movem hack that copies 56 bytes in 72 cycles,
which would give you north of 5MB/s on a 7.09MHz CPU.
2:20 p.m.
New subject: usb to unibus - was installing 2.11BSD on a pdp11/53
On Mon, Dec 23, 2013 at 4:15 PM, Al Kossow <aek at bitsavers.org> wrote:
On 12/23/13 11:30 AM, Ethan Dicks wrote:
Yeah. Byte Writes still worked (because of how the internal bus
latches worked), but Byte reads failed (for the same cause - you got a
ghost of whatever the previous bus transaction was - that was how I
proved it wasn't software, the hardware trace didn't match what the
code was clearly doing).
-ethan
he swapped UDS and LDS and
because of how the 68K is built internally, somethings worked and
somethings failed.
'byte smearing'
9:30 p.m.
New subject: looking for a diode boot m792 board
If anyone has one, please contact me offlist...
--tnx
--tom
1:13 p.m.
New subject: usb to unibus - was installing 2.11BSD on a pdp11/53
On 12/23/13 10:05 AM, Ethan Dicks wrote:
This is the opposite
of what IBM did with the PC where *every* PC had a multi-channel
DMA engine that any peripheral could use (floppy, hard disk, sound
card...)
Which has come to bite us in the ass 30 years later when modern systems
can't support legacy peripherals that used those
DMA channels (re. the software compatible floppy interface thread).
The performance of those MB DMA channels were pretty poor, too, along
with the fact that they couldn't be upgraded due to backwards compatibility
and the stupidity of the ISA signaling polarity of DMAs and IRQs
oh.. I forgot about DMA/IRQ conficts.
They did us no favors by having them.
20 Dec
20 Dec
10:38 a.m.
New subject: usb to unibus - was installing 2.11BSD on a pdp11/53
On 12/20/13 8:51 AM, John Wilson wrote:
A direct USB-to-RL01/02 interface is on my to-do list
but there's a bunch
of stuff higher up on the list. Eventually I'd like to build USB interfaces
for all the basic DEC disks, to make them look like USB flash disks, so that
they'll work with any OS or emulator or utility or just the Weenix "dd"
command.
I mentioned this to John and Brad off-list. Brad's RL02 Unibus interface got
me off my butt on a project I've been thinking about for a long time to interface
Omnibus, Unibus and Qbus to peripherals using 10/100 Ethernet as the transport.
For example, for a DECtape you have an Omnibus card that looks like a TC08 to the
computer, and talks at the block level to a TU56 through Ethernet. The same for
a TC11, or a QBus TC11 (which never existed in real life). The topology is however
you plug the Ethernet together, so you could have a simulated peripheral running
off of a Linux box talking to a real computer or a simulator talking to a real
peripheral, or a PDP11 and PDP8 sharing a TU56 (obviously not at the same time).
The two storage devices I was going to start on were the TU56 (the drive-side adapter
would
have integral G888 replacements) and RK05.
The Omnibus card is going to have a few extra things on it (32k of memory and an RF08
emulator)
10:54 a.m.
New subject: usb to unibus - was installing 2.11BSD on a pdp11/53
On Fri, Dec 20, 2013 at 10:38:54AM -0800, Al Kossow wrote:
The Omnibus card is going to have a few extra things on
it (32k of memory and
an RF08 emulator)
NOW you're talking my language! I've got my own back-burnered posi/negibus
project waiting for attention but my favorite part (and the only thing actually
prototyped so far) is the teensy little daughtercard (PCB was $4) with two
FRAM chips on it and a selector, for emulating an RF08. A PDP-8 without TSS/8
is like a day without sunshine...
John Wilson
D Bit
11:12 a.m.
New subject: usb to unibus - was installing 2.11BSD on a pdp11/53
On 12/20/13 10:54 AM, John Wilson wrote:
A PDP-8 without TSS/8
is like a day without sunshine...
One of the things that got me started thinking about this again was
Dale Luck just got a really nice 8/e that he wants to run TSS/8 on.
I had to break the news to him that RF08s are impossible to find.
Did you see that someone else has turned up with a TSS/8 listing?
http://www.vintage-computer.com/vcforum/showthread.php?40507-Can-I-run-an-R…
11:22 a.m.
New subject: usb to unibus - was installing 2.11BSD on a pdp11/53
On Fri, Dec 20, 2013 at 2:12 PM, Al Kossow <aek at bitsavers.org> wrote:
On 12/20/13 10:54 AM, John Wilson wrote:
Didn't TSS/8 run on DF32s? (maybe it took 2 or 4 of them...)
A PDP-8 without TSS/8
is like a day without sunshine..
One of the things that got me started thinking about this again was
Dale Luck just got a really nice 8/e that he wants to run TSS/8 on.
I had to break the news to him that RF08s are impossible to find. Did you see that someone else has turned up with a
TSS/8 listing?
http://www.vintage-computer.com/vcforum/showthread.php?40507-Can-I-run-an-R…
Nice!
I have a lot of time spent on OS/8, and a bit of time spent on RTS-8,
but I have yet to try to get TSS/8 running, even under emulation.
-ethan
11:24 a.m.
New subject: usb to unibus - was installing 2.11BSD on a pdp11/53
On Fri, Dec 20, 2013 at 11:12:36AM -0800, Al Kossow wrote:
One of the things that got me started thinking about
this again was
Dale Luck just got a really nice 8/e that he wants to run TSS/8 on.
I had to break the news to him that RF08s are impossible to find.
IIRC, DECUS used to have someone's patches for somehow running TSS/8 on
an RK8E/RK05 (only). I'm suspicious, since the DF32/RF08's word-oriented
behavior is exposed in the RFILE/WFILE calls so they'd need to find space
for sector buffers, but, whatever. Anyway it's no RL8A but if true it solves
the same basic problem. (TSS/8 officially supported the RK8E/RK05 as a
data-only device but not as a system device.)
Did you see that someone else has turned up with a
TSS/8 listing?
No! I'm surprised they bothered with 8.22 since I've had 8.24 online for
years and years (although I think mine does have a typo or two) and I don't
think 8.22 has any advantages over the later version. But hey, whatever!
It makes sense that people want to duplicate the system they used in the
old days.
John Wilson
D Bit
6:56 p.m.
New subject: usb to unibus - was installing 2.11BSD on a pdp11/53
At 02:24 PM 12/20/2013, John Wilson wrote:
On Fri, Dec 20, 2013 at 11:12:36AM -0800, Al Kossow
wrote:
Yes, there was a version of TSS/8 that ran on RK05 media. The way I
implemented this was to use OS/8 on the "RKA0" half of the disk,
compiling the TSS/8 bits, and installed TSS/8 on the "RKB0" side.
As you mention, TSS/8 is highly dependent upon the RF08's (or DF32's)
ability to read/write single words, so those I/O operations have to be
buffered. Given an 8/E or later base, it's not hard to find stuff that
can be deleted to make room for a buffer to manage that. This
definitely had lower performance than a RF08, but this was a hack that
never saw real operation.
Bob Shelley wrote a UUO for TSS/8 that allowed an application to
operate what would be called pseudo terminals today: send input to a
detached terminal's input buffer and read whatever was in the
terminal's output buffer. With that, I could fire up a system with 24
virtual terminals and run memory tests in all 24. That was a very
effective stress test.
All of this was during my field service support days at DEC. One
example was trying to support a school system that had been promised
that they could have 24 terminals running at once on their TSS/8
system. Their 8/E system with expansion box to handle 24 KL8Es, bus
converter, 32 Mb of memory, etc. was just too fragile (too many bus
loads) and failed all the time. Eventually, DEC replaced the expansion
box with an 8A box and KL8A multiport serial controllers. I implemented
support for the KL8/A in TSS/8.
Sadly, none of that code ever saw the light of day outside of DEC.
-Rick
One of the things that got me started thinking
about this again was
Dale Luck just got a really nice 8/e that he wants to run TSS/8 on.
I had to break the news to him that RF08s are impossible to find.
IIRC, DECUS used to have someone's patches for somehow running TSS/8 on
an RK8E/RK05 (only). I'm suspicious, since the DF32/RF08's word-oriented
behavior is exposed in the RFILE/WFILE calls so they'd need to find space
for sector buffers, but, whatever. Anyway it's no RL8A but if true it
solves
the same basic problem. (TSS/8 officially supported the RK8E/RK05 as a
data-only device but not as a system device.)
9:59 p.m.
New subject: usb to unibus - was installing 2.11BSD on a pdp11/53
On Fri, Dec 20, 2013 at 09:56:02PM -0500, Rick Murphy wrote:
Yes, there was a version of TSS/8 that ran on RK05
media. The way I
implemented this was to use OS/8 on the "RKA0" half of the disk,
compiling the TSS/8 bits, and installed TSS/8 on the "RKB0" side.
Oh wicked cool, so it's true!! Way to go! Plus you were building more
or less natively? (The listings I had for TSS/8.24 built for Melrose H.S.
were from PAL-10.)
Bob Shelley wrote a UUO for TSS/8 that allowed an
application to
operate what would be called pseudo terminals today: send input to a
detached terminal's input buffer and read whatever was in the
terminal's output buffer.
Wow ... so Telnet would be possible! OK OK or a batch processor. Nice.
Their 8/E system with expansion box to handle 24 KL8Es,
bus
converter, 32 Mb of memory, etc. was just too fragile (too many bus
loads) and failed all the time.
My 8/E came with something like 19 KL8Es and an LC8E and the one time I
could afford to have F-S come look at it, the first thing they said was
to dump the second box and move the core into the main box, or I'd go
crazy with crashes. Guess there was a lot of that going around! It's
not like I really ran more than a handful of TTYs anyway.
Sadly, none of that code ever saw the light of day
outside of DEC.
By any chance did you save it? The only reason I knew about the RK8E stuff
was because it was once in the DECUS catalog so that much must have been
allowed out the door at least ... semi-legitimately, even.
I never get tired of how amazingly well TSS/8 squeezed every drop out
of a PDP-8. And it put on quite a lights show doing it.
John Wilson
D Bit
22 Dec
22 Dec
3:55 a.m.
New subject: usb to unibus - was installing 2.11BSD on a pdp11/53
At 12:59 AM 12/21/2013, John Wilson wrote:
Batch, yes. Telnet? No way, in tiny little PDP-8 that needs to support
24 users.
Unfortunately, no. If I had only held on to a copy of the pack with
that stuff.
At the time (ca. 30 years ago), I just wanted to get rid of the
machine, rack, and everything else. Sigh.
On Fri, Dec 20, 2013 at 09:56:02PM -0500, Rick Murphy
wrote:
Yup. There's nothing special in the TSS/8 code that required PAL-10.
I think that was used more for the nicely formatting listings. :)
Yes, there was a version of TSS/8 that ran on RK05
media. The way I
implemented this was to use OS/8 on the "RKA0" half of the disk,
compiling the TSS/8 bits, and installed TSS/8 on the "RKB0" side.
Oh wicked cool, so it's true!! Way to go! Plus you were building more
or less natively? (The listings I had for TSS/8.24 built for Melrose H.S.
were from PAL-10.) Bob Shelley
wrote a UUO for TSS/8 that allowed an application to
operate what would be called pseudo terminals today: send input to a
detached terminal's input buffer and read whatever was in the
terminal's output buffer.
Wow ... so Telnet would be possible! OK OK or a batch processor. Nice. Their 8/E
system with expansion box to handle 24 KL8Es, bus
converter, 32 Mb of memory, etc. was just too fragile (too many bus
loads) and failed all the time.
My 8/E came with something like 19 KL8Es and an LC8E and the one time I
could afford to have F-S come look at it, the first thing they said was
to dump the second box and move the core into the main box, or I'd go
crazy with crashes. Guess there was a lot of that going around! It's
not like I really ran more than a handful of TTYs anyway.
Sadly, none of that code ever saw the light of day
outside of DEC.
By any chance did you save it? The only reason I knew about the RK8E
stuff
was because it was once in the DECUS catalog so that much must have been
allowed out the door at least ... semi-legitimately, even. I never get tired of how amazingly well TSS/8 squeezed
every drop out
of a PDP-8. And it put on quite a lights show doing it.
Yup. It was amazing to see in action at some of the schools with a
dozen or more people working away.
-Rick
3780
days inactive
3804
days old
31 comments
15 participants
participants (15)
-
abuse＠cabal.org.uk -
aek＠bitsavers.org -
ak6dn＠mindspring.com -
chd＠chdickman.com -
ethan.dicks＠gmail.com -
glen.slick＠gmail.com -
healyzh＠aracnet.com -
holm＠freibergnet.de -
jkunz＠unixag-kl.fh-kl.de -
Mark＠Misty.com -
pontus＠Update.UU.SE -
rick＠rickmurphy.net -
uban＠ubanproductions.com -
useddec＠gmail.com -
wilson＠dbit.com