Running Previous 2.0 w/ NS3.3 - trying to get networking up

I have Previous 2.0 on Windows and a NS3.3 .dd volume I downloaded and after following the instructions for setting up networking that came with the .dd, I see the EN light blink a bit and when I

telnet bbs.fozztexx.com

The command sits there thinking and the EN light blinks, but nothing happens. So it's somewhat setup correctly. I followed these instructions:

QuoteHowto: Setup networking on NeXTstep 3.0 and later:

1. Use Simple Network Starter (you find it in the NextAdmin directory)
  and set up some values:
  a  Check "Provide the services specified below."
  b  Choose a hostname (whatever you want).
  c  Set IP address to "10.0.2.15".
  d  Go to "Network Options..." and set
     Router to "10.0.2.2"
     Broadcast Address to "10.0.2.255"
     Netmask to "255.255.255.0" and
     NIS Domain Name to "None".
  e  Hit configure. You will be prompted to connect the Ethernet
     cable.
  f  If not already done, connect emulated Ethernet now using
     Previous' GUI and hit OK.

2. Using root access rights open and edit these files:
  a  /etc/hostconfig and replace "-ROUTER-" with "10.0.2.2".
  b  /etc/resolv.conf and add "nameserver 10.0.2.3".
     If this file does not exist, you have to first create it.
     Be sure you create a plain text file. If you use Edit.app
     you can do this by selecting Format > Text > Make ASCII.

  Howto: Open the above files using root privileges (NeXTstep 3.3):
  a  Open the /etc directory with this command in Terminal:
     "open /etc"
  b  Select the file/application to open and from Workspace select
     Services > Open Sesame > Open As Root.
  c  For making a new resolv.conf open Edit.app the same way (b)
     and in the save dialog type as name "/etc/resolv.conf"

3. Reboot.

(I used vi from terminal to edit, fwiw.) But -- not quite working. Any ideas? Thanks.


bp[/quote]

start with the basics..

with SLiRP the only thing you can ping is the gateway, see if that works.

ping 10.0.2.2

I think NeXTSTEP should have nslookup, try to resolve a DNS name...

Quote from: "neozeed"start with the basics..

with SLiRP the only thing you can ping is the gateway, see if that works.

ping 10.0.2.2

I think NeXTSTEP should have nslookup, try to resolve a DNS name...

Well, first ping isn't on the system. I'll need to get that solved first. :-) I will look into mounting a shared volume thru Previous, which I'm guessing is possible.


bp

Ping is normally present on a NeXTstep 3.3 installation. But on my system it is not recognized, if I am running it as normal user. I need to run it as super user: #su

If I recall, this is just a path issue - it's not on the search path unless you are a superuser...

Quote from: "neozeed"start with the basics..

with SLiRP the only thing you can ping is the gateway, see if that works.

ping 10.0.2.2

I think NeXTSTEP should have nslookup, try to resolve a DNS name...

As root I can, indeed, ping 10.0.2.2 -- getting successful ping returns. I can't nslookup anything -- just stalls. Any ideas from here? Thanks.


bp

Quote from: "blakespot"

Quote from: "neozeed"start with the basics..

with SLiRP the only thing you can ping is the gateway, see if that works.

ping 10.0.2.2

I think NeXTSTEP should have nslookup, try to resolve a DNS name...

As root I can, indeed, ping 10.0.2.2 -- getting successful ping returns. I can't nslookup anything -- just stalls. Any ideas from here? Thanks.


bp

what is your /etc/resolv.conf

I'd try the following:

10.0.2.3

this should have slirp intercept and ask the OS on your behalf...

8.8.8.8

the google....


1.1.1.1
1.0.0.1  the new cloudflare ones...


also check your route table..

netstat -nr

make sure your gateway is set to 10.0.2.2

Quote from: "neozeed"

what is your /etc/resolv.conf

I'd try the following:

10.0.2.3

this should have slirp intercept and ask the OS on your behalf...

8.8.8.8

the google....


1.1.1.1
1.0.0.1  the new cloudflare ones...


also check your route table..

netstat -nr

make sure your gateway is set to 10.0.2.2

I can ping and get returns on 10.0.2.3 but none of the other IPs you mention -- 100% packet loss there. My /etc/resolv.conf is the one following line:

nameserver 10.0.2.3

And my netstat -nr output is:



Thanks.


bp

Yep you can only ping 10.0.2.2

Try nslookup


nslookup
> server 1.1.1.1
Default server: 1.1.1.1
Address: 1.1.1.1#53
> www.google.com
Server:         1.1.1.1
Address:        1.1.1.1#53

Non-authoritative answer:
Name:   www.google.com
Address: 74.125.201.99
Name:   www.google.com
Address: 74.125.201.103
Name:   www.google.com
Address: 74.125.201.104
Name:   www.google.com
Address: 74.125.201.105
Name:   www.google.com
Address: 74.125.201.106
Name:   www.google.com
Address: 74.125.201.147
>

Try all the servers... See if any of this resolves...

Try telnetting to 71.95.196.34

Quote from: "neozeed"

Try all the servers... See if any of this resolves...

Try telnetting to 71.95.196.34

nslookup just stalls pointing at 1.1.1.1 or 1.0.0.1, etc. Telnet to the .34 IP just stalls.


bp

I'm not sure..  I'm stuck with only my Android phone and an ancient g4 MacBook at the moment.... So I can't really do anything windows like.

Did you boot with the Ethernet connected?

I find it works best with it unconnected at the ROM, and to enable it after the kernel starts to look initialize

Trying it running as Administrator. I don't do much with Windows, myself, but game. I just couldn't find Previous 2.x on macOS. Do you happen to know where it can be had, precompiled? Thanks.


bp

Admin privileges shouldn't matter at all when it comes to slirp...

What version of Windows was it anyway?

Cmake hides everything so I really have no idea if it's doing anything strange while building...

Quote from: "neozeed"Admin privileges shouldn't matter at all when it comes to slirp...

What version of Windows was it anyway?

Cmake hides everything so I really have no idea if it's doing anything strange while building...

A recent version (if not very latest update) of Windows 10 Pro 64-bit.


bp

I've tried only enabling Ethernet in Pervious options at different points in the boot sequence and nothing is working. Have you heard of others running Previous 2, Windows 10 Pro 64 having networking work?

Frustrating.



bp

Wait, I don't need to explicitly run SLiRP under Windows right? It's rolled into Previous, ya? Is that my problem?

Quote from: "blakespot"I just couldn't find Previous 2.x on macOS. Do you happen to know where it can be had, precompiled? Thanks.

Hey Blake - it is available on my Previous site - http://previous.unixdude.net/download.html

Quote from: "eagle"

Quote from: "blakespot"I just couldn't find Previous 2.x on macOS. Do you happen to know where it can be had, precompiled? Thanks.

Hey Blake - it is available on my Previous site - http://previous.unixdude.net/download.html

Thank you! And networking works for me under OS X. Not sure why Windows 10 setup didn't work as far as networking.

(https://flic.kr/p/26taPHU)
Logging into Level 29 BBS on an original NeXT computer (...emulated via Previous 2.0) (https://flic.kr/p/26taPHU) by Blake Patterson (https://www.flickr.com/photos/blakespot/), on Flickr

Cheers.

I generally use NFS or an ISO disk image to transfer files. A while ago, some NFS activity would cause Previous to crash but I have not tried it lately.

Exactly the same situation, but this time with Previous 2.1 + NS3.3 on MacOS (tried on two different machines with different OS). Checked all I could, apparently everything is fine, consistent with what the real NeXT machines are reading (bar for the addresses, of course), but still no way to get networking up and running.

Tried to run the whole thing all over again, but networking still does not fully work. E.g. if I ping www.apple.com, the address is resolved but all packets are lost.

Quote from: "paolo.bertolo"Tried to run the whole thing all over again, but networking still does not fully work. E.g. if I ping www.apple.com, the address is resolved but all packets are lost.

what's your gateway address? do a netstat -rn and look for a 'default' or a '0.0.0.0' (sorry, dont have a NS/OS box handy to tell you exactly what to look for). Or just post the whole output here.

The gateway would need to be local (same network).

The Previous stuff (NeXTComputer - NeXTStep 3.3)

previous: 1# netstat -rn
Routing tables
Destination      Gateway            Flags     Refs  Use     Interface
127.0.0.1        127.0.0.1          UH          2     282     lo0
default            10.0.2.2            UG          0        0      en0
10.2.2             10.0.2.15          U            4     755     en0
previous: 2#


The real thing (NeXTComputer - NeXTStep 3.3)

nextcomputer:6# netstat -rn
Routing tables
Destination        Gateway             Flags     Refs     Use  Interface
127.0.0.1          127.0.0.1            UH          2      300   lo0
default              192.168.88.1       UG          0        4    en0
192.168.88       192.168.88.55      U           4     1207  en0
nextcomputer:7#

My host setup: macOS Catalina, SLiRP networking over the thin interface. Guest is a NeXTStation Color Turbo.

I'm trying to get networking going with Previous and NS3.3, and like the most recent poster in this thread, have got name service working (can lookup hosts), but can't ping or otherwise access them.  Also can't ping machines on my local network. So the guest is somewhat configured, but something is missing.

EDIT: I fired up OmniWeb, and pointed it at a web page served in my local network, and it worked (!). This suggests that Previous hosts some kind of user-level proxy, but I'm a bit puzzled how I'm supposed to contact the guest from the host.

How does SLiRP work from the host-perspective? Is it running a NAT, or some other level? When I run ifconfig on my host machine, I don't see any new interfaces, so I'm not sure how I can / should test the networking setup from the host inward.

Quote from: dylanmcnamee on November 10, 2019, 04:44:04 PMMy host setup: macOS Catalina, SLiRP networking over the thin interface. Guest is a NeXTStation Color Turbo.

I'm trying to get networking going with Previous and NS3.3, and like the most recent poster in this thread, have got name service working (can lookup hosts), but can't ping or otherwise access them.  Also can't ping machines on my local network. So the guest is somewhat configured, but something is missing.

EDIT: I fired up OmniWeb, and pointed it at a web page served in my local network, and it worked (!). This suggests that Previous hosts some kind of user-level proxy, but I'm a bit puzzled how I'm supposed to contact the guest from the host.

How does SLiRP work from the host-perspective? Is it running a NAT, or some other level? When I run ifconfig on my host machine, I don't see any new interfaces, so I'm not sure how I can / should test the networking setup from the host inward.

You have to re-compile Previous after editing one file.  I confirmed you can use this for telnet and SSH and everything I tried.

Quote from: rumbeard on February 21, 2019, 11:36:35 PMSo I have this working according to instructions and I can even telnet back to the host on 42323.  My question is... do I have to rebuild from source to use the other 5 re-direct slots?

Ideally I'd swap out 42323 with 42322 for SSH, port 80, 443, 25, and maybe one other.

I can't seem to figure out even from a glance at the source where the slirp.conf stuff is.

Quote from: neozeed on February 24, 2019, 12:40:53 AMhttps://sourceforge.net/p/previous/code/HEAD/tree/trunk/src/enet_slirp.c

It's in enet_slirp_start

Code Select Expand

void enet_slirp_start(Uint8 *mac) {
    struct in_addr guest_addr;
   
    if (!slirp_inited) {
        Log_Printf(LOG_WARN, "Initializing SLIRP");
        slirp_inited=1;
        slirp_init();
        inet_aton("10.0.2.15", &guest_addr);
        slirp_redir(0, 42323, guest_addr, 23);
    }
    if (slirp_inited && !slirp_started) {
        Log_Printf(LOG_WARN, "Starting SLIRP");
        slirp_started=1;
        slirpq = QueueCreate();
        slirp_mutex=SDL_CreateMutex();
        tick_func_handle=SDL_CreateThread(tick_func,"SLiRPTickThread", (void *)NULL);
    }
}

Specifically this line:

Code Select Expand

slirp_redir(0, 42323, guest_addr, 23);


I can provide a patched binary if you get stuck.  I think it's 80, 443, 22, and I forget what other two.

I made a tweaked version that forwards the telnet, ftp, ssh and http ports, it built just fine, and now I'm happily bi-directionally networking.

I added these lines to my enet_slirp.c:

        slirp_redir(0, 40220, guest_addr, 20); // ftp data
        slirp_redir(0, 40221, guest_addr, 21); // ftp control
        slirp_redir(0, 40222, guest_addr, 22); // ssh
        slirp_redir(0, 40223, guest_addr, 23); // telnet
        slirp_redir(0, 40233, guest_addr, 1023); // nfs ?
        slirp_redir(0, 40280, guest_addr, 80); // http

And on my host machine, I can both telnet and ftp to the guest with these commands:

% ncftp -u <my username> -P 40221 localhost
- and -
% telnet localhost 40223

Quote from: dylanmcnamee on November 17, 2019, 04:23:33 PMI made a tweaked version that forwards the telnet, ftp, ssh and http ports, it built just fine, and now I'm happily bi-directionally networking.

I added these lines to my enet_slirp.c:

        slirp_redir(0, 40220, guest_addr, 20); // ftp data
        slirp_redir(0, 40221, guest_addr, 21); // ftp control
        slirp_redir(0, 40222, guest_addr, 22); // ssh
        slirp_redir(0, 40223, guest_addr, 23); // telnet
        slirp_redir(0, 40233, guest_addr, 1023); // nfs ?
        slirp_redir(0, 40280, guest_addr, 80); // http

And on my host machine, I can both telnet and ftp to the guest with these commands:

% ncftp -u <my username> -P 40221 localhost
- and -
% telnet localhost 40223

Fabulous! Can we get this change into the official builds?

Quote from: stefan on July 06, 2024, 12:31:30 PMFabulous! Can we get this change into the official builds?

This sounds like a good idea. Short question:

There is already this line in the code:
slirp_redir(0, 42323, guest_addr, 23);
Do we still need
slirp_redir(0, 40223, guest_addr, 23); // telnet
and why are those new lines using port 402xx instead of 423xx?

Thank you Andreas for picking this up!

I am not sure why the original commenter changed the port to 402xx but I would imagine that using 423xx for ftp, ssh and http would enable a lot of people's tinkering with services running on NeXTstep

Quote from: stefan on July 07, 2024, 01:24:19 AMThank you Andreas for picking this up!

I am not sure why the original commenter changed the port to 402xx but I would imagine that using 423xx for ftp, ssh and http would enable a lot of people's tinkering with services running on NeXTstep

So adding/keeping these lines would be OK?

slirp_redir(0, 42320, guest_addr, 20); /* ftp data */
slirp_redir(0, 42321, guest_addr, 21); /* ftp control */
slirp_redir(0, 42322, guest_addr, 22); /* ssh */
slirp_redir(0, 42323, guest_addr, 23); /* telnet */
slirp_redir(0, 42380, guest_addr, 80); /* http */


Quote from: andreas_g on July 07, 2024, 01:43:42 AMSo adding/keeping these lines would be OK?

Code Select Expand

slirp_redir(0, 42320, guest_addr, 20); /* ftp data */
slirp_redir(0, 42321, guest_addr, 21); /* ftp control */
slirp_redir(0, 42322, guest_addr, 22); /* ssh */
slirp_redir(0, 42323, guest_addr, 23); /* telnet */
slirp_redir(0, 42380, guest_addr, 80); /* http */


Perfect!

Great! Can you write a short paragraph for the readme to document this feature? I'm afraid I am missing the words to do so.

Quote from: andreas_g on July 07, 2024, 02:10:56 PMGreat! Can you write a short paragraph for the readme to document this feature? I'm afraid I am missing the words to do so.

I've got you covered:
SLiRP Port Forwards
 -------------------

When networking with SLiRP, Previous listens on the following TCP ports, and will forward connections to the guest machine:

Host Port Internal Port Service
42320 20 FTP data
42321 21 FTP control
42322 22 SSH
42323 23 Telnet
42380 80 HTTP

So e.g. to connect to telnet on the emulated machine, you can point your client at localhost port 42323.

Note that the guest OS has to have these services running in order to actually connect to them! NeXTSTEP and OPENSTEP do not include servers for SSH or HTTP.

Oh, we overlapped at posting! I like your table better. Maybe it's worth adding an example though:


diff -u a/networking.howto.txt b/networking.howto.txt
--- a/networking.howto.txt
+++ b/networking.howto.txt
@@ -194,3 +194,32 @@ NOWTO: Setup networking on Mac OS:
      Subnet Mask: 255.255.255.0
      Gateway/Router Address: 10.0.2.2
      Domain Name Server Information: Domain home, IP Address 10.0.2.3
+
+
+NOWTO: Accessing the emulated machine from the host machine through TCP/IP
+
+When using SLiRP, the emulated machine is not directly visible / accessible
+through its own IP address 10.0.2.2. Instead, Previous provides a number of
+port redirects that allow you to access certain ports on the emulated NeXT
+machine:
+
+  FTP: 42320 / 42321
+  SSH: 42322
+  Telnet: 42323
+  HTTP: 42380
+
+Examples:
+
+So, if you wamt to telnet into your emulated NeXT machine, you can do so by
+running the following command on your host machine:
+
+  telnet localhost 42323
+
+In order to connect through SSH (if installed on the NeXT side), do
+
+  ssh me@localhost -p 42322
+
+You can ftp into your emulated NeXT machine with
+
+  ncftp -P 42321 ftp://me@localhost
+


@Rhetorica, @stefan
Thank you very much! The updated readme is online.

To whom it may concern in the future: some FTP clients will have trouble finding the alternate data port. Forcing a passive connection in your connection settings will solve this problem.

In WinSCP, this can be done with the "Force IP address for passive mode connections" checkbox under the "FTP" section.

Just got this port mapping done. It works great to get into the emulator from the host machine.
But anyone knows if there is a way of getting networking to work with pcap? (I tried on OSX Monterey but got some messages related to the something failing in the kernel on boot)

Quote from: daniel.dlds on December 03, 2024, 04:28:52 PMJust got this port mapping done. It works great to get into the emulator from the host machine.
But anyone knows if there is a way of getting networking to work with pcap? (I tried on OSX Monterey but got some messages related to the something failing in the kernel on boot)

Please post the messages here. I'll have a look and see if I can fix it.

Please see the screenshot at :
https://ibb.co/2KkGr47

Quote from: daniel.dlds on December 14, 2024, 03:48:17 AMPlease see the screenshot at :
https://ibb.co/2KkGr47

I am getting the same message. I don't know why we get that message, but there do not seem to be any issues related to this. It is not related to your networking problems.

I think your network configuration might be faulty. For example it seems your netmask is set to 0.0.0.0.