Network Trouble

I have experienced the following network problem under QNX4. Any
solutions are greatly appreciated.

I attempted to connect to an AIX based server from a panel pc with
touchscreen using QNX 4. I could not successfully ping the server.
Parameters are as follows:

Server IP: 152.116.28.152
Gateway IP: 152.116.28.129
QNX IP: 152.116.28.157

I could not ping the gateway either. I added the following to
/etc/netstart.
(after the ifconfig statements).

route add default 152.116.28.129

I still could not ping the server or the gateway.

What is interesting is that if I configured a notebook PC running
windows 98 to the same IP address as the QNX PC and replace the QNX PC
with it, I can successfully ping the server and the gateway (I can ping
the server regardless of whether I configure the gateway or not in
windows). Furthermore, if I change the IP of the windows notebook PC to
the server’s IP address and use a cross-wired network cable, I can
successfully ping the QNX PC and vice versa.

Any suggestions as to what the problem might be?

Thank You

Dan Szymanski

Previously, Daniel A. Szymanski wrote in qdn.public.qnx4.devtools:

I have experienced the following network problem under QNX4. Any
solutions are greatly appreciated.

I attempted to connect to an AIX based server from a panel pc with
touchscreen using QNX 4. I could not successfully ping the server.
Parameters are as follows:

Server IP: 152.116.28.152
Gateway IP: 152.116.28.129
QNX IP: 152.116.28.157

Ok, what happens if you run:

ifconfig en1

I could not ping the gateway either. I added the following to
/etc/netstart.
(after the ifconfig statements).

route add default 152.116.28.129

You should be able to ping your Server and Gateway without this
statement. This statement should allow you to ping the rest
of the world.

I still could not ping the server or the gateway.

What is interesting is that if I configured a notebook PC running
windows 98 to the same IP address as the QNX PC and replace the QNX PC
with it, I can successfully ping the server and the gateway (I can ping
the server regardless of whether I configure the gateway or not in
windows). Furthermore, if I change the IP of the windows notebook PC to
the server’s IP address and use a cross-wired network cable, I can
successfully ping the QNX PC and vice versa.

Oooooooh! So tell us more about what everthing is plugged into.
Are all three devices plugged into a hub or a switch? Were you
using the cross-wired network cable direct from the QNX PC NIC to
the server’s?

Some hubs have a special port that is cross wired either in a fixed
manner, or can be switched. Try switching how the devices are plugged
in.

Lastly, can you ping 152.116.28.157 ?


Mitchell Schoenbrun --------- maschoen@pobox.com

Mitchell,

I was only trying to run the cross wired cable from the QNX NIC to the windows
PC at this time. I do not have any information at this time about how the
server’s network is configured at this time.


Mitchell Schoenbrun wrote:

Previously, Daniel A. Szymanski wrote in qdn.public.qnx4.devtools:
I have experienced the following network problem under QNX4. Any
solutions are greatly appreciated.

I attempted to connect to an AIX based server from a panel pc with
touchscreen using QNX 4. I could not successfully ping the server.
Parameters are as follows:

Server IP: 152.116.28.152
Gateway IP: 152.116.28.129
QNX IP: 152.116.28.157

Ok, what happens if you run:

ifconfig en1

I could not ping the gateway either. I added the following to
/etc/netstart.
(after the ifconfig statements).

route add default 152.116.28.129

You should be able to ping your Server and Gateway without this
statement. This statement should allow you to ping the rest
of the world.

I still could not ping the server or the gateway.

What is interesting is that if I configured a notebook PC running
windows 98 to the same IP address as the QNX PC and replace the QNX PC
with it, I can successfully ping the server and the gateway (I can ping
the server regardless of whether I configure the gateway or not in
windows). Furthermore, if I change the IP of the windows notebook PC to
the server’s IP address and use a cross-wired network cable, I can
successfully ping the QNX PC and vice versa.

Oooooooh! So tell us more about what everthing is plugged into.
Are all three devices plugged into a hub or a switch? Were you
using the cross-wired network cable direct from the QNX PC NIC to
the server’s?

Some hubs have a special port that is cross wired either in a fixed
manner, or can be switched. Try switching how the devices are plugged
in.

If this was the case, then how come I can ping the server and the gateway from
the actual hardware (not a simple crossed cable) in windows, but not from the
QNX NIC. If the hardware works with windows, it should work with QNX, right?


Lastly, can you ping 152.116.28.157 ?

Yes, I can in essence “ping myself” in both cases (windows and QNX).


Mitchell Schoenbrun --------- > maschoen@pobox.com

Is you MAC Address in /etc/config/netmap specified correctly for the card
that is configured for use with IP? netinfo -l can show you the MAC address
of the card.

Is the MAC address for the card associated with IP address you desire and
the interface you desire? netstat -in can show this information.

Can you ping localhost?

  • Richard

Previously, Daniel Szymanski wrote in qdn.public.qnx4.devtools:

Mitchell,

I was only trying to run the cross wired cable from the QNX NIC to the windows
PC at this time. I do not have any information at this time about how the
server’s network is configured at this time.


Mitchell Schoenbrun wrote:

Previously, Daniel A. Szymanski wrote in qdn.public.qnx4.devtools:
I have experienced the following network problem under QNX4. Any
solutions are greatly appreciated.

I attempted to connect to an AIX based server from a panel pc with
touchscreen using QNX 4. I could not successfully ping the server.
Parameters are as follows:

Server IP: 152.116.28.152
Gateway IP: 152.116.28.129
QNX IP: 152.116.28.157

What netmask does QNX use? Post the output of ‘ifconfig en1’. Does
this netmask agree with the netmask that windows uses? QNX TCP/IP
will not accept a netmask that creates a supernet, and will re-write
it. Does this netmask agree with the server?

What is the output of ‘netstat -rn’?

Which network card are you using?

Have you tried forcing half or full duplex when you start Net.?

Did the QNX driver get the media rate correct? This is available with
netstat -L or netinfo -L (can’t remember which).

When you say that you cannot ping, what is the error that ping
returns? Does it say “host unreachable”, “no route to host”, “network
unreachable” or anything like that?

Cheers,
Andrew

Previously, Daniel A. Szymanski wrote in qdn.public.qnx4.devtools:

The results of ifconfig en1 are as follows:

en1: flags=8863<UP,BROADCAST,NOTRAILERS,RUNNING,SIMPLEX,MULTICAST
inet 152.116.28.157 netmask ffff0000 broadcast 152.116.255.255

This matches the mask of windows (255.255.0.0).

Ok, that’s not it.

What is the output of ‘netstat -rn’?

I do not know, but I do have the output of netinfo -l and netstat -in, they are as
follows:

Total Number Of Net Driver Slots: 2

Driver Slot 0: Driver Pid 55 Logical Net 1 Network Card: Ethernet/
rtl : RealTek 8139 Ethernet Controller
Vendor ID … 0x10ec
Device ID … 0x8139
Subsystem ID … 0x8139
Subsystem Vendor ID … 0x10ec
Revision … 0x2000010
Physical Node ID … 00D0C9 102726
Media Rate … 10Mb/s
Mtu … 1514
Hardware Interrupt … 11

Total Packets Txd OK … 4
Total Packets Txd Bad … 0
Tx Collision Errors … 0
Tx Collision Errors (aborted) … 0
Carrier Sense Lost on Tx … 0
FIFO Underruns During Tx … 0
Tx deferred … 0
Out of Window Collisions … 0

Total Packets Rxd OK … 0
Total Rx Errors … 0
FIFO Overruns During Rx … 0
Alignment errors … 0
CRC errors … 0


Driver Slot 1: Unused

Have you tried forcing half or full duplex when you start Net.?

No, is this an option when starting Net.rtl? What are the benefits
of forcing each of these?

Not much, as far as I can tell. However, all of your equipment has to
agree on full vs half duplex, or it won’t work. It is possible that
the auto-detection is not working in the QNX driver, or that the card
doesn’t auto-detect and the defaults are different in the QNX driver
vs the Windows driver. Check to see if the Net.rtl supports this, and
try forcing both.

Did the QNX driver get the media rate correct? This is available with
netstat -L or netinfo -L (can’t remember which).

I believe that the media rate is correct (10mbps). I will confirm it.
(see netinfo -l above).

Same deal - it’s supposed to auto-detect, but maybe not. Try forcing
the correct rate if that output is wrong.

I am suspicious of the netinfo output:

Hardware Interrupt … 11

Total Packets Txd OK … 4
Total Packets Rxd OK … 0

This is a PCI card, so the interrupt is assigned for you. The PCI
BIOS is perfectly happy to assign the same IRQ to more than one device
if the device advertises that it can handle it. However, the drivers
also have to be able to handle shared IRQs, so there might be a
problem if the interrupt is being shared with another device that
thinks it has the IRQ all to itself. You can see IRQ assignments with
show_pci, I think. The fact that your card has transmitted data, but
received nothing is a little funny - it means that it is not
listening. Which leads us to…

This card comes in multiple flavours, I suspect, with different
connectors. If the QNX driver supports choosing physical medium (in
QNX6 it is connector=[0-3]), then try forcing it to UTP or BNC or
whatever. It might not be auto-detecting properly.

Does your card have lights on it that indicate connection, tx or rx?
Do they flash appropriately when you initiate network activity? If
you try to ping the QNX machine from elsewhere, you should see rx
activity (well, actually any network activity should flash rx), and if
you try to ping out, you should see tx activity. If it has a
connection light, it should turn on when you plug in the network
cable.

Good luck,
Andrew

Andrew,

Thank You for the reply (and everyone else), I have answered your questions the best
that I could

Andrew Thomas wrote:

Previously, Daniel Szymanski wrote in qdn.public.qnx4.devtools:
Mitchell,

I was only trying to run the cross wired cable from the QNX NIC to the windows
PC at this time. I do not have any information at this time about how the
server’s network is configured at this time.


Mitchell Schoenbrun wrote:

Previously, Daniel A. Szymanski wrote in qdn.public.qnx4.devtools:
I have experienced the following network problem under QNX4. Any
solutions are greatly appreciated.

I attempted to connect to an AIX based server from a panel pc with
touchscreen using QNX 4. I could not successfully ping the server.
Parameters are as follows:

Server IP: 152.116.28.152
Gateway IP: 152.116.28.129
QNX IP: 152.116.28.157

What netmask does QNX use? Post the output of ‘ifconfig en1’. Does
this netmask agree with the netmask that windows uses? QNX TCP/IP
will not accept a netmask that creates a supernet, and will re-write
it. Does this netmask agree with the server?

The results of ifconfig en1 are as follows:

en1: flags=8863<UP,BROADCAST,NOTRAILERS,RUNNING,SIMPLEX,MULTICAST>
inet 152.116.28.157 netmask ffff0000 broadcast 152.116.255.255

This matches the mask of windows (255.255.0.0).

What is the output of ‘netstat -rn’?

I do not know, but I do have the output of netinfo -l and netstat -in, they are as
follows:

Total Number Of Net Driver Slots: 2

Driver Slot 0: Driver Pid 55 Logical Net 1 Network Card: Ethernet/
rtl : RealTek 8139 Ethernet Controller
Vendor ID … 0x10ec
Device ID … 0x8139
Subsystem ID … 0x8139
Subsystem Vendor ID … 0x10ec
Revision … 0x2000010
Physical Node ID … 00D0C9 102726
Media Rate … 10Mb/s
Mtu … 1514
Hardware Interrupt … 11

Total Packets Txd OK … 4
Total Packets Txd Bad … 0
Tx Collision Errors … 0
Tx Collision Errors (aborted) … 0
Carrier Sense Lost on Tx … 0
FIFO Underruns During Tx … 0
Tx deferred … 0
Out of Window Collisions … 0

Total Packets Rxd OK … 0
Total Rx Errors … 0
FIFO Overruns During Rx … 0
Alignment errors … 0
CRC errors … 0


Driver Slot 1: Unused


Which network card are you using?

See above.

Have you tried forcing half or full duplex when you start Net.?

No, is this an option when starting Net.rtl? What are the benefits of forcing each
of these?

Did the QNX driver get the media rate correct? This is available with
netstat -L or netinfo -L (can’t remember which).

I believe that the media rate is correct (10mbps). I will confirm it.
(see netinfo -l above).

When you say that you cannot ping, what is the error that ping
returns? Does it say “host unreachable”, “no route to host”, “network
unreachable” or anything like that?

No, it does not give any specific message (I forgot to try with the verbose
option). This machine is now at the customer site. It only stops after trying to
send the first packet (I must use control-C to escape from it).

The results of netstat -in is as follows:

Name Mtu Network Address Ipkts Ierrs Opkts Oerrs Coll
en1 1500 0.d0.c9.10.27.26 0 0 5 0 0
en1 1500 152.116 152.116.28.157 0 0 5 0 0
ppp0* 1500 0 0 0 0 0
lo0 1536 28 0 28 0 0
lo0 1536 127 127.0.0.1 28 0 28 0 0

I am able to ping the local host (127.0.0.1) successfully.

Cheers,
Andrew

Regards,

Dan Szymanski

Additional information:

The realtek hardware is on the Motherboard of a panel PC (not a PCI bus).

I tried to connect to the network with entirely different hardware (still QNX 4 with
on-board ethernet-Ethernet NE1000/200 read from ifconfig en1). The problem was
IDENTICAL with this new hardware. Any suggestions?

Thank You

Dan Szymanski


Andrew Thomas wrote:

Previously, Daniel A. Szymanski wrote in qdn.public.qnx4.devtools:
The results of ifconfig en1 are as follows:

en1: flags=8863<UP,BROADCAST,NOTRAILERS,RUNNING,SIMPLEX,MULTICAST
inet 152.116.28.157 netmask ffff0000 broadcast 152.116.255.255

This matches the mask of windows (255.255.0.0).

Ok, that’s not it.

What is the output of ‘netstat -rn’?

I do not know, but I do have the output of netinfo -l and netstat -in, they are as
follows:

Total Number Of Net Driver Slots: 2

Driver Slot 0: Driver Pid 55 Logical Net 1 Network Card: Ethernet/
rtl : RealTek 8139 Ethernet Controller
Vendor ID … 0x10ec
Device ID … 0x8139
Subsystem ID … 0x8139
Subsystem Vendor ID … 0x10ec
Revision … 0x2000010
Physical Node ID … 00D0C9 102726
Media Rate … 10Mb/s
Mtu … 1514
Hardware Interrupt … 11

Total Packets Txd OK … 4
Total Packets Txd Bad … 0
Tx Collision Errors … 0
Tx Collision Errors (aborted) … 0
Carrier Sense Lost on Tx … 0
FIFO Underruns During Tx … 0
Tx deferred … 0
Out of Window Collisions … 0

Total Packets Rxd OK … 0
Total Rx Errors … 0
FIFO Overruns During Rx … 0
Alignment errors … 0
CRC errors … 0


Driver Slot 1: Unused

Have you tried forcing half or full duplex when you start Net.?

No, is this an option when starting Net.rtl? What are the benefits
of forcing each of these?

Not much, as far as I can tell. However, all of your equipment has to
agree on full vs half duplex, or it won’t work. It is possible that
the auto-detection is not working in the QNX driver, or that the card
doesn’t auto-detect and the defaults are different in the QNX driver
vs the Windows driver. Check to see if the Net.rtl supports this, and
try forcing both.

Did the QNX driver get the media rate correct? This is available with
netstat -L or netinfo -L (can’t remember which).

I believe that the media rate is correct (10mbps). I will confirm it.
(see netinfo -l above).

Same deal - it’s supposed to auto-detect, but maybe not. Try forcing
the correct rate if that output is wrong.

I am suspicious of the netinfo output:

Hardware Interrupt … 11

Total Packets Txd OK … 4
Total Packets Rxd OK … 0

This is a PCI card, so the interrupt is assigned for you. The PCI
BIOS is perfectly happy to assign the same IRQ to more than one device
if the device advertises that it can handle it. However, the drivers
also have to be able to handle shared IRQs, so there might be a
problem if the interrupt is being shared with another device that
thinks it has the IRQ all to itself. You can see IRQ assignments with
show_pci, I think. The fact that your card has transmitted data, but
received nothing is a little funny - it means that it is not
listening. Which leads us to…

This card comes in multiple flavours, I suspect, with different
connectors. If the QNX driver supports choosing physical medium (in
QNX6 it is connector=[0-3]), then try forcing it to UTP or BNC or
whatever. It might not be auto-detecting properly.

Does your card have lights on it that indicate connection, tx or rx?
Do they flash appropriately when you initiate network activity? If
you try to ping the QNX machine from elsewhere, you should see rx
activity (well, actually any network activity should flash rx), and if
you try to ping out, you should see tx activity. If it has a
connection light, it should turn on when you plug in the network
cable.

Good luck,
Andrew

Previously, Daniel A. Szymanski wrote in qdn.public.qnx4.devtools:

Additional information:

The realtek hardware is on the Motherboard of a panel PC (not a PCI bus).

Vendor ID … 0x10ec
Device ID … 0x8139
Subsystem ID … 0x8139
Subsystem Vendor ID … 0x10ec
Revision … 0x2000010

I’ve never seen this information come from anywhere but a PCI
controller. It may not be a PCI backplane, but I suspect that it is
still PCI (e.g., PC/104+ is PCI).

I tried to connect to the network with entirely different hardware
(still QNX 4 with on-board ethernet-Ethernet NE1000/200 read from
ifconfig en1). The problem was IDENTICAL with this new hardware.
Any suggestions?

Not really. You will have to post some of the other information I
requested in earlier postings before I can really try to add anything
more. When you get access to the machine again, then let us know.

What is the output of ‘netstat -rn’?

Have you tried forcing half or full duplex when you start Net.?

Did the QNX driver get the media rate correct? This is available with
netstat -L or netinfo -L (can’t remember which).

This card comes in multiple flavours, I suspect, with different
connectors. If the QNX driver supports choosing physical medium (in
QNX6 it is connector=[0-3]), then try forcing it to UTP or BNC or
whatever. It might not be auto-detecting properly.

Does your card have lights on it that indicate connection, tx or rx?
Do they flash appropriately when you initiate network activity? If
you try to ping the QNX machine from elsewhere, you should see rx
activity (well, actually any network activity should flash rx), and if
you try to ping out, you should see tx activity. If it has a
connection light, it should turn on when you plug in the network
cable.

One more thing - if you get no network activity on any card, check
your network cable. Watch your hub or switch if you have one to see
if it sees the card being plugged in.

Cheers,
Andrew

I have received hardware and connection information from the customer. It is as
follows:

The manufacturer of the hub is ODS. The hardware is encased in a box similar to a
desktop PC with cards that slide into the box. One of the cards is #294-9 and
designated as the Network interface controller. The other eight cards are of
#294-TFR and are labeled 10 base T ethernet. I am told that this equipment is
around 7-8 years old. Also, I am told that both the server and the QNX machine
are connected to the same box (hub), but I could not visually confirm this myself.

Let me know if any additional information is required. Thank You.

Dan Szymanski


Andrew Thomas wrote:

Previously, Daniel A. Szymanski wrote in qdn.public.qnx4.devtools:
Additional information:

The realtek hardware is on the Motherboard of a panel PC (not a PCI bus).

Vendor ID … 0x10ec
Device ID … 0x8139
Subsystem ID … 0x8139
Subsystem Vendor ID … 0x10ec
Revision … 0x2000010

I’ve never seen this information come from anywhere but a PCI
controller. It may not be a PCI backplane, but I suspect that it is
still PCI (e.g., PC/104+ is PCI).

I tried to connect to the network with entirely different hardware
(still QNX 4 with on-board ethernet-Ethernet NE1000/200 read from
ifconfig en1). The problem was IDENTICAL with this new hardware.
Any suggestions?

Not really. You will have to post some of the other information I
requested in earlier postings before I can really try to add anything
more. When you get access to the machine again, then let us know.

What is the output of ‘netstat -rn’?

Have you tried forcing half or full duplex when you start Net.?

Did the QNX driver get the media rate correct? This is available with
netstat -L or netinfo -L (can’t remember which).

This card comes in multiple flavours, I suspect, with different
connectors. If the QNX driver supports choosing physical medium (in
QNX6 it is connector=[0-3]), then try forcing it to UTP or BNC or
whatever. It might not be auto-detecting properly.

Does your card have lights on it that indicate connection, tx or rx?
Do they flash appropriately when you initiate network activity? If
you try to ping the QNX machine from elsewhere, you should see rx
activity (well, actually any network activity should flash rx), and if
you try to ping out, you should see tx activity. If it has a
connection light, it should turn on when you plug in the network
cable.

One more thing - if you get no network activity on any card, check
your network cable. Watch your hub or switch if you have one to see
if it sees the card being plugged in.

Cheers,
Andrew

In addition, I was informed that the entire network is wired with CAT 3 cable. Since
the windows notebook is able to ping, I do not think that this is the problem however.

Thank You.

“Daniel A. Szymanski” wrote:

I have received hardware and connection information from the customer. It is as
follows:

The manufacturer of the hub is ODS. The hardware is encased in a box similar to a
desktop PC with cards that slide into the box. One of the cards is #294-9 and
designated as the Network interface controller. The other eight cards are of
#294-TFR and are labeled 10 base T ethernet. I am told that this equipment is
around 7-8 years old. Also, I am told that both the server and the QNX machine
are connected to the same box (hub), but I could not visually confirm this myself.

Let me know if any additional information is required. Thank You.

Dan Szymanski

Andrew Thomas wrote:

Previously, Daniel A. Szymanski wrote in qdn.public.qnx4.devtools:
Additional information:

The realtek hardware is on the Motherboard of a panel PC (not a PCI bus).

Vendor ID … 0x10ec
Device ID … 0x8139
Subsystem ID … 0x8139
Subsystem Vendor ID … 0x10ec
Revision … 0x2000010

I’ve never seen this information come from anywhere but a PCI
controller. It may not be a PCI backplane, but I suspect that it is
still PCI (e.g., PC/104+ is PCI).

I tried to connect to the network with entirely different hardware
(still QNX 4 with on-board ethernet-Ethernet NE1000/200 read from
ifconfig en1). The problem was IDENTICAL with this new hardware.
Any suggestions?

Not really. You will have to post some of the other information I
requested in earlier postings before I can really try to add anything
more. When you get access to the machine again, then let us know.

What is the output of ‘netstat -rn’?

Have you tried forcing half or full duplex when you start Net.?

Did the QNX driver get the media rate correct? This is available with
netstat -L or netinfo -L (can’t remember which).

This card comes in multiple flavours, I suspect, with different
connectors. If the QNX driver supports choosing physical medium (in
QNX6 it is connector=[0-3]), then try forcing it to UTP or BNC or
whatever. It might not be auto-detecting properly.

Does your card have lights on it that indicate connection, tx or rx?
Do they flash appropriately when you initiate network activity? If
you try to ping the QNX machine from elsewhere, you should see rx
activity (well, actually any network activity should flash rx), and if
you try to ping out, you should see tx activity. If it has a
connection light, it should turn on when you plug in the network
cable.

One more thing - if you get no network activity on any card, check
your network cable. Watch your hub or switch if you have one to see
if it sees the card being plugged in.

Cheers,
Andrew