SLIP client in ancient RS6000-320H

[Carlos Murillo]

Hello, everyone;

Please bear with me, as this is a long post, if you have any expertise with SLIP and old AIX.

About a year ago I posted about this headless RS6000 320H with AIX 3.1-005 in it; the thread was

https://groups.google.com/forum/#!topic/comp.unix.aix/kvSWJI-ykmI

At the time, I solved the issue with the missing mount binary, added a second hard drive, and built a toolchain up to gcc-g++-g77 2.95.3 (3.x proved too difficult), Python 2.6.1, perl 5.6.2 among others. Then work took all of my free time until now, when I am trying to connect the machine (which has no network adapter) via SLIP to a Raspberry Pi in order to network it, since I had been doing all file transfers to/from it using xmodem or kermit on the serial console.

I have not had luck in setting the SLIP connection. AIX 3.1 does not have PPP functionality; only uncompressed SLIP is supported. It comes with an slattach command. I did the following on the AIX side:

1) I made available tty6 using smit/Devices/tty/Add a TTY, changed its characteristics to rts/cts and other things using the comp.unix.aix FAQ, and configured it to make it available. XON/XOFF was disabled.

2) Created the sl0 interface using smit/Communications Applications and Services/TCPIP/Further Configuration/Network Interfaces/Network Interface Selection/Add a Network Interface/Add a Serial Line INTERNET Network Interface, and selected tty6 for it. The local address was set to 192.168.0.201 and the remote to 192.168.0.200 .

3) Changed the tty6 inittab entry to
tty6:2:off:/etc/getty /dev/tty6

4) Added the following entry to /usr/lib/uucp/Devices
Direct tty6 - 38400 direct

On the Raspberry Pi side, I installed net-tools (which contains the Linux version of slattach).

I connected the two systems with a full null-modem cable (DTR on one side goes to both DSR and CD on the other side), and verified that the systems could talk with each other at 38400 baud using "screen" on the Raspberry Pi side and cu on the AIX side. The AIX system asserts DTR when using the serial line; the Raspberry Pi sees that on DSR and CD, and viceversa.

Following the information about Linux's slattach and AIX's slattach, the commands that I've used to try to connect the systems are as follows:

Raspberry Pi:
sudo slattach -v -d -p slip -s 38400 /dev/tty_USB3 &
sudo ifconfig sl0 192.168.0.200 pointopoint 192.168.0.201 netmask 255.255.255.0 up

AIX 3.1: (as root)
ifconfig sl0 192.168.0.201 192.168.0.200 up
slattach tty6 38400

If I do that, then on the Raspberry Pi side, I can ping 192.168.0.200, but not 192.168.0.201,
and on the AIX side I get
ping 192.168.0.201
PING 192.168.0.201: 56 data bytes
0821-069 ping: sendto: The network is not currently available
ping: wrote 192.168.0.201 64 chars, ret=1

Same goes when pinging 192.168.0.200 .

In the AIX side I get
ifconfig sl0
sl0: flags=31<UP,POINTOPOINT,NOTRAILERS>
inet 192.168.0.201 --> 192.168.0.200 netmask 0xffffff00

In the Raspberry Pi side I get
ifconfig sl0
sl0: flags=4305<IP,POINTOPOINT,RUNNING,NOARP,MULTICAST> mtu 296
inet 192.168.0.200 netmask 255.255.255.0 destination 192.168.0.201
slip txqueuelen 10 (Serial Line IP)
RX packets 0 bytes (0.0 B)
RX errors 0 dropped 0 overruns 0 frame 0
TX packets 8 bytes 672 (672.0 B)
TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0

I have an RS232 line level monitor (basically LEDs) connected and I can see that slattach asserts DTR on the AIX side (the Raspberry's side is always asserted), but I see no flickering whatsoever on either RxD or TxD, so it seems that neither side is doing anything at all to initiate the SLIP protocol (no probing packets, although the ifconfig stats in the Raspberry would seem to indicate that some packets were sent from the Raspberry Pi to the AIX system; I just did not see them). I have tried several variations in the slattach command on the AIX side, such as

slattach tty6
slattach tty6 38400 '"" "CONNECT"' 4
slattach tty6 38400 '"" "\p\b"' 4 (one of these two is supposed to assert a BREAK
slattach tty6 38400 '"" "\p\k"' 4 in the RS232 line for the Raspberry to see)

Supposedly, in a no-authentication, no-modem dedicated direct line, what I've done should be enough to start the protocol. But neither side initiates it (or the AIX side doesn't reply) , and I don't see any other command line options in either Linux's slattach or AIX's that would help.

Any ideas about how to proceed?

Thanks in advance,

Carlos.

[Grant Taylor]

On 12/11/19 10:08 AM, Carlos Murillo wrote:
> Hello, everyone;

Hi,

> About a year ago I posted about this headless RS6000 320H with AIX
> 3.1-005 in it; the thread was

I've not yet read that thread. I'll have to pull it up and read when
time permits.

> I had been doing all file transfers to/from it using xmodem or kermit
> on the serial console.

I'm curious, does AIX 3.1-005 support UUCP? That might be another
option that supports more than discrete manual file transfer.

> Any ideas about how to proceed?

I've done very little with SLIP. But I /thought/ that SLIP was
initiated from the communications channel that would be carrying the
SLIP data.

> Thanks in advance,

Good luck.



--
Grant. . . .
unix || die

[Carlos Murillo]

On Wednesday, December 11, 2019 at 12:27:15 PM UTC-5, Grant Taylor wrote:
> I'm curious, does AIX 3.1-005 support UUCP? That might be another
> option that supports more than discrete manual file transfer.

AIX does have UUCP, but I'd really like to network the thing so I can telnet into it and use it from anywhere in the house.

Carlos.

[Grant Taylor]

On 12/11/19 4:56 PM, Carlos Murillo wrote:
> AIX does have UUCP,

Thank you for confirming.

> but I'd really like to network the thing so I can telnet into it and
> use it from anywhere in the house.

That seems reasonable enough.

I now wonder if telneting to something like a serial console server
(read: Raspberry Pi w/ an RS-232 serial port) would count as a stop gap
measure. Especially if you could telnet / ssh to the console server and
have it spawn the serial session for you.

I know it's not as nice as IP connectivity to AIX itself. But it is
accessible from anywhere in the house. ;-)

[Carlos Murillo]

On Wednesday, December 11, 2019 at 7:19:58 PM UTC-5, Grant Taylor wrote:
> > but I'd really like to network the thing so I can telnet into it and
> > use it from anywhere in the house.
>
> That seems reasonable enough.
>
> I now wonder if telneting to something like a serial console server
> (read: Raspberry Pi w/ an RS-232 serial port) would count as a stop gap
> measure. Especially if you could telnet / ssh to the console server and
> have it spawn the serial session for you.
>
> I know it's not as nice as IP connectivity to AIX itself. But it is
> accessible from anywhere in the house. ;-)
>
> --
> Grant. . . .
> unix || die

As a matter of fact, I am building a Raspberry Pi-powered box that will

1) provide serial console access to several old machines using ser2net; so yes, there will be a console server;
2) switch the mains on and off to those machines
3) provide SLIP connectivity to the 320H
4) host several shared file systems (for some old 68K macs and for several unix boxes); for this, I need it to boot from an USB hard drive (even it it has an SD card with the /boot partition in it but that is only read one at boot time)

But I still want the 320H networked.

Carlos.

[Carlos Murillo]

On Wednesday, December 11, 2019 at 10:55:37 PM UTC-5, Carlos Murillo wrote:
> But I still want the 320H networked.

By the way, I have confirmed that the TX packet count in the Raspberry Pi interface was the result of pinging the address local to the Raspberry Pi from itself; so indeed, there are no packets from any machine present in the RS232 medium.

carlos.

[Grant Taylor]

On 12/11/19 8:55 PM, Carlos Murillo wrote:
> As a matter of fact, I am building a Raspberry Pi-powered box that will
>
> 1) provide serial console access to several old machines using ser2net;
> so yes, there will be a console server;
> 2) switch the mains on and off to those machines
> 3) provide SLIP connectivity to the 320H
> 4) host several shared file systems (for some old 68K macs and for
> several unix boxes); for this, I need it to boot from an USB hard
> drive (even it it has an SD card with the /boot partition in it but
> that is only read one at boot time)

I question why you need to boot from the USB hard drive vs just use it
as an additional drive. But if that's what you want to do, then okay.

> But I still want the 320H networked.

Fair enough.

You have options and you are knowingly choosing to make something work.
Seems reasonable to me.

[Grant Taylor]

On 12/11/19 10:08 AM, Carlos Murillo wrote:

> 1) I made available tty6

Naive question: Why tty6?

Am I completely misinterpreting something that should be obvious to
someone that has spent more time working on AIX than I have?

I find the number of the tty, 6, to be unexpected.

How many physical serial ports does the machine have?

Are you using tty6 for your xmodem / kermit connections over the null modem?

I don't mean any disrespect. I've made those types of mistakes before
and kick myself for them. So, I'm just asking.

[Carlos Murillo]

On Thursday, December 12, 2019 at 12:26:44 AM UTC-5, Grant Taylor wrote:
> On 12/11/19 8:55 PM, Carlos Murillo wrote:
> > 4) host several shared file systems (for some old 68K macs and for
> > several unix boxes); for this, I need it to boot from an USB hard
> > drive (even it it has an SD card with the /boot partition in it but
> > that is only read one at boot time)
>
> I question why you need to boot from the USB hard drive vs just use it
> as an additional drive. But if that's what you want to do, then okay.

> --
> Grant. . . .
> unix || die

I've had several SD cards die in Raspberry Pi's that were on continuously; it seems that some SD cards just don't handle continuous reading/writing very well. Real HDs fare much better in this regard.

carlos.

[Carlos Murillo]

On Thursday, December 12, 2019 at 12:29:50 AM UTC-5, Grant Taylor wrote:
> Naive question: Why tty6?

> --
> Grant. . . .
> unix || die

Oh, there's nothing special about tty6. This machine originally came with an 8 port RS232 async adapter in addition to the two built-in serial ports. Apparently, the first built-in port was configured as tty0 and used originally for a printer, while five of the 8 ports in the async adapter were configured as tty1-tty5. In AIX, you configure each entry in /dev/ttyX using a hardware configuration utility and assign each /dev/ttyX to a physical port. Now, I could not use the ports in the 8-port async adapter because I am missing the special cable that connects to it (DA-78 to eight DB-25, I have the pinout, I'll build it later), so for SLIP I had to configure the second built-in port as an additional tty, and, not wanting to mess with the existing configuration for tty1-tty5, the next device to assign was tty6. The console is currently on the first built-in port at /dev/tty0 .

I used AIX only from userland back in the 90's in an SP2 cluster. It is only in the last year that I have learned a little about AIX administration, so there are still many things that are new to me. If I knew more, I would have probably solved the SLIP issue by now.

carlos.

[Grant Taylor]

On 12/12/19 7:42 AM, Carlos Murillo wrote:
> I've had several SD cards die in Raspberry Pi's that were on
> continuously; it seems that some SD cards just don't handle continuous
> reading/writing very well. Real HDs fare much better in this regard.

Fair.

I was assuming that any heavy I/O would be to the external USB drive.

Thus just the base Raspbian (et al.) would not kill good quality SD
cards. But, I can't say as I'm surprised if that's not the case.

[Grant Taylor]

On 12/12/19 8:01 AM, Carlos Murillo wrote:
> Oh, there's nothing special about tty6. This machine originally
> came with an 8 port RS232 async adapter in addition to the two
> built-in serial ports. Apparently, the first built-in port was
> configured as tty0 and used originally for a printer, while five
> of the 8 ports in the async adapter were configured as tty1-tty5.
> In AIX, you configure each entry in /dev/ttyX using a hardware
> configuration utility and assign each /dev/ttyX to a physical port.
> Now, I could not use the ports in the 8-port async adapter because
> I am missing the special cable that connects to it (DA-78 to eight
> DB-25, I have the pinout, I'll build it later), so for SLIP I had
> to configure the second built-in port as an additional tty, and,
> not wanting to mess with the existing configuration for tty1-tty5,
> the next device to assign was tty6. The console is currently on the
> first built-in port at /dev/tty0 .

Ah. That makes much more sense. Thank you for explaining.

> I used AIX only from userland back in the 90's in an SP2 cluster.
> It is only in the last year that I have learned a little about AIX
> administration, so there are still many things that are new to me.
> If I knew more, I would have probably solved the SLIP issue by now.

#learningOpportunity

[Kevin Bowling]

This is a non-answer to your question since I don't know how to
configure AIX SLIP, but why not pick up a Type 2-1.

http://ps-2.kev009.com/rsinfo/adap21.htm

Regards,
Kevin

[cmhe...@gmail.com]

On Sunday, December 22, 2019 at 5:18:06 PM UTC-8, Kevin Bowling wrote:
> This is a non-answer to your question since I don't know how to
> configure AIX SLIP, but why not pick up a Type 2-1.

I’d suggest that as well, I think it even supports netboot for even the oldest RS/6000 systems.