Ian Johnston wrote:

> Dear People,
>
> I have recently added WiFi to my Thinkpad, but I can't get NETBeui
> stuff to work over the WiFi link.
>
> Under "Adaptors and protocols" I have
>
> Broadcom Ethernet
> 0 - IBM OS/2 NETBIOS OVER TCP/IP
> 0 - IBM OS/2 TCP/IP
> Generic Prism
> 1 - IBM OS/2 NETBIOS OVER TCP/IP
> 1 - IBM OS/2 TCP/IP
>
> and in c:\ibmlan\ibmlan.ini I have (much trimmed)
>
> [networks]
>
> net1 = TCPBEUI$,0,LM10,34,100,14
> net2 = TCPBEUI$,1,LM10,34,100,14
>
> [requester]
>
> wrkservices = MESSENGER
> wrknets = NET1,NET2
>
> [peer]
>
> SRVSERVICES =
> srvnets = NET1,NET2
>
> When the wired connection is running "Start File and Print Client"
> (NET START PEER) runs happily though it does seem to wait around a
> long time for something or other. When the wifi connection is running
> it just times out at start up.

Try it when the wired connection is up as well. You may find that when the
ethernet port is down, that NetBIOS/IP chokes, because that adapter is
listed first. In another thread, I recommended binding NetBIOS over TCP/IP
to the NULL MAC adapter and only TCP/IP to the real adapters.

As an experiment, try changing the wrknets and srvnets order. If wireless
now works, you've identified the cause of the problem. However, now the
wired connection may not work, if the wifi adapter is down. Binding to the
NULL MAC adapter avoids the problem, because it's never down. You'll also
have to bind TCP/IP to it.

On Thu, 10 Nov 2005 23:27:45 UTC, "Bob Eager"
wrote:

> On Thu, 10 Nov 2005 23:10:12 UTC, "Ian Johnston"
> wrote:
>
> > I have recently added WiFi to my Thinkpad, but I can't get NETBeui
> > stuff to work over the WiFi link.
>
> Let's see PROTOCOL.INI too.

'Ere we go. Sorry about the length, not sure what bits matter ...

Ian

[PROT_MAN]

DRIVERNAME = PROTMAN$

[IBMLXCFG]

tcpbeui_nif = tcpbeui.nif
tcpip_nif = tcpip.nif
B57_nif = B57.nif
GENPRISM_nif = genprism.nif

[NETBIOS]

DriverName = netbios$
ADAPTER0 = tcpbeui$,0
ADAPTER1 = tcpbeui$,1

[tcpbeui_nif]

DriverName = tcpbeui$
Bindings = B57_nif,GENPRISM_nif
NODETYPE = "B-Node"
OS2TRACEMASK = 0x0
SESSIONS = 130
NCBS = 225
NAMES = 21
SELECTORS = 15
USEMAXDATAGRAM = "NO"
NETBIOSTIMEOUT = 500
NETBIOSRETRIES = 2
NAMECACHE = 1000
PURGECACHE = 0
PRELOADCACHE = "NO"
NAMESFILE = 0
DATAGRAMPACKETS = 20
PACKETS = 50
ENABLEDNS = 0
INTERFACERATE = 300

[tcpip_nif]

DriverName = TCPIP$
Bindings = B57_nif,GENPRISM_nif

[B57_nif]

DriverName = B57$

[GENPRISM_nif]

DriverName = PRISMC$
HW_TYPE = "PCCARD"
NETADDRESS = "00022D1D6A33"
DESIRED_SSID = "IBM"
STA_MODE = "BSS"
CREATE_IBSS = "NO"
PREAMBLE = "BOTH"
MW_ROBUSTNESS = "NO"
PSP_MODE = "DISABLED"
WEP = "DISABLED"
AUTH_TYPE = "OPEN_SYSTEM"
SUPPLY_VOLTAGE = "3.3V"
DEBUG_LEVEL = 0xFF1F



On Fri, 11 Nov 2005 01:58:26 UTC, James Knott
wrote:

> Ian Johnston wrote:

> > When the wired connection is running "Start File and Print Client"
> > (NET START PEER) runs happily though it does seem to wait around a
> > long time for something or other. When the wifi connection is running
> > it just times out at start up.
>
> Try it when the wired connection is up as well. You may find that when the
> ethernet port is down, that NetBIOS/IP chokes, because that adapter is
> listed first. In another thread, I recommended binding NetBIOS over TCP/IP
> to the NULL MAC adapter and only TCP/IP to the real adapters.

Thanks. Sorry if this sounds a stupid question, but could you advise
on the adaptor numbering in that case? I haven't yet been able to work
out when two instances of the same thing can have, or must have,
different or the same numbers.

> As an experiment, try changing the wrknets and srvnets order. If wireless
> now works, you've identified the cause of the problem. However, now the
> wired connection may not work, if the wifi adapter is down. Binding to the
> NULL MAC adapter avoids the problem, because it's never down. You'll also
> have to bind TCP/IP to it.

As an experiment, I just tried NET START PEER (I'm on WiFi) and to my
great surprise it worked. So why would it fail at startup but not
later - and what is it waiting for at startup anyway?

Apologies for all the questions - hope I don't sound too ignorant or
lazy.

Ian

The only suggestion I have right now is that the NAMES value is too low.
The IBMLAN.INI entries in [networks] are each calling for 14 names, and
you have only 21 available. Bump that up to (say) 40, and reboot.

On Fri, 11 Nov 2005 08:53:22 UTC, "Bob Eager"
wrote:

> The only suggestion I have right now is that the NAMES value is too low.
> The IBMLAN.INI entries in [networks] are each calling for 14 names, and
> you have only 21 available. Bump that up to (say) 40, and reboot.

Thanks. I'll try that ...

Ian

On Fri, 11 Nov 2005 08:13:26 UTC, "Ian Johnston"
wrote:

> As an experiment, I just tried NET START PEER (I'm on WiFi) and to my
> great surprise it worked. So why would it fail at startup but not
> later - and what is it waiting for at startup anyway?

And now it doesn't work again!

[C:\]net start peer
The REQUESTER service is not started.
The REQUESTER service is
starting.......................................... .....
..............
The REQUESTER service could not be started.
NET3060: The service did not respond to a control signal and was
stopped with
the DosKillProc function.

For more information, type HELP NET3060.

Ian

[A complimentary Cc of this posting was sent to
Ian Johnston
], who wrote in article :
> [C:\]net start peer
> The REQUESTER service is not started.
> The REQUESTER service is
> starting.......................................... .....
> .............
> The REQUESTER service could not be started.
> NET3060: The service did not respond to a control signal and was
> stopped with
> the DosKillProc function.
>
> For more information, type HELP NET3060.

For more information, see LANTRANS.LOG. I could fix all my net30??
errors by looking there (docs were not helpful, but the log is
descriptive enough).

You must understand that the damage could already be done *before* the
"net start" - some drivers read the configuration files too...

Hope this helps,
Ilya

On Sat, 12 Nov 2005 19:56:38 UTC, Ilya Zakharevich
wrote:

> > The REQUESTER service could not be started.
> > NET3060: The service did not respond to a control signal and was
> > stopped with
> > the DosKillProc function.
> >
> > For more information, type HELP NET3060.
>
> For more information, see LANTRANS.LOG. I could fix all my net30??
> errors by looking there (docs were not helpful, but the log is
> descriptive enough).
>
> You must understand that the damage could already be done *before* the
> "net start" - some drivers read the configuration files too...

Thanks. My LANTRAN.LOG contains

LT00073: FFST/2 is installed but is not started. LANTRAN.LOG is being
created.
IBM OS/2 LANMSGDD [05/21/04] 6.00 is loaded and operational.
IBM OS/2 TCPBEUI 6.00.0
TCPBEUI: Using a 32-bit data segment.
Installing NETWKSTA.200 Version 6.0. IBM LAN Redirector ( Oct 10,
2001)

IBM OS/2 NETBIOS 4.0
Adapter 0 has 123 NCBs, 96 sessions, and 25 names available to NETBIOS
applications.
Adapter 1 has 123 NCBs, 96 sessions, and 25 names available to NETBIOS
applications.
NETBIOS 4.0 is loaded and operational.
SYS0318: Message file GENPRISM.MSG cannot be found for message 0042.
.... (repeated many times)
IBM LANVDD is loaded and operational.
IBM OS/2 LAN Netbind
SYS0318: Message file GENPRISM.MSG cannot be found for message 0042.
TCPBEUI: Logical adapter 0 is trying to bind to TCP/IP interface lan0.
TCPBEUI: Logical adapter 1 is trying to bind to TCP/IP interface lan1.
.... (repeated many times)
TCPBEUI: Logical adapter 1 is bound to TCP/IP interface lan1.

genprism.msg is in ibmcom\macs, so I guess the above means that, being
just a RC, it doesn't have message 0042 in it.

Ian

[A complimentary Cc of this posting was sent to
Ian Johnston
], who wrote in article :
> > > For more information, type HELP NET3060.

> > For more information, see LANTRANS.LOG. I could fix all my net30??
> > errors by looking there (docs were not helpful, but the log is
> > descriptive enough).

Well, your situation looks different...

> SYS0318: Message file GENPRISM.MSG cannot be found for message 0042.

> genprism.msg is in ibmcom\macs, so I guess the above means that, being
> just a RC, it doesn't have message 0042 in it.

One thought: Is ibmcom/macs on DATA= path?

Yours,
Ilya

On Sun, 13 Nov 2005 11:31:33 UTC, Ilya Zakharevich
wrote:

> One thought: Is ibmcom/macs on DATA= path?

I don't have one of them! Is that the equivalent of DPATH? Various
networking things are there, but not ibmcom/macs

Ian

Ian Johnston wrote:
> On Sun, 13 Nov 2005 11:31:33 UTC, Ilya Zakharevich
> wrote:
>
>
>>One thought: Is ibmcom/macs on DATA= path?
>
>
> I don't have one of them! Is that the equivalent of DPATH? Various
> networking things are there, but not ibmcom/macs

I think Ilya meant DPATH=, not DATA= .

Cheers,
Martin

>
> Ian
>

[A complimentary Cc of this posting was sent to
MMI
], who wrote in article <6f664$43773bed$542ac208$23852@news.chello.cz>:

> I think Ilya meant DPATH=, not DATA= .

Right.

Sorry,
Ilya

On Sat, 12 Nov 2005 09:08:41 UTC, "Ian Johnston"
wrote:

> On Fri, 11 Nov 2005 08:13:26 UTC, "Ian Johnston"
> wrote:
>
> > As an experiment, I just tried NET START PEER (I'm on WiFi) and to my
> > great surprise it worked. So why would it fail at startup but not
> > later - and what is it waiting for at startup anyway?
>
> And now it doesn't work again!

And now it does. I'm getting really, really confused here.

Just for the record, I have now ...

a) added \ibmcom\macs to dpath

b) bound IBM OS/2 TCP/IP to Ethernet (0), Wifi (1) and
No-network-adaptor (2) and IBM OS/2 NETBIOS OVER TCP/IP to
No-network-adaptor (1) only

c) modified ibmlan.ini to contain net1 = TCPBEUI$,2,LM10,34,100,14 and
arranged for srvnets and wrknets to refer to this only

d) looked a protocol.ini which contains

[NETBIOS]

DriverName = netbios$
ADAPTER2 = tcpbeui$,2

[tcpbeui_nif]

DriverName = tcpbeui$
Bindings = ,,NONETADP_nif

and worried slightly about those two commas in the Bindings entry, but
not known whether to do anything about them

And the result is that about 9 times out of 10 NET START PEER fails
and the remaining time it works perfectly. Time since reboot and
presence of other systems (an ASUS combined NAS/WAP device and -
earlier - an XP laptop) seem to have no effect either way.

Help! This is looking random, and computers shouldn't do random
things!

Ian

[A complimentary Cc of this posting was sent to
Ian Johnston
], who wrote in article :

> And now it doesn't work again!

Got into the same situation today. But the changes were not very
substantial, so I could deduce the contributing factor for my net3060:

I moved MPTSTART entry from CONFIG.SYS to STARTUP.CMD

Doing binary search resulted in:

ifconfig.exe lan0 ip.addr.ess

is needed for NETBEUI to initialize. Here is how my CONFIG.SYS looks now:

RUN=D:\MPTN\BIN\AFNBINI.EXE
RUN=D:\MPTN\BIN\CNTRL.EXE

rem This call is NEEDED to make TCPBEUE initialize correctly; use a temporary value
CALL=D:\MPTN\BIN\ifconfig.exe lan0 192.168.1.37
rem CALL=D:\OS2\CMD.EXE /Q /C D:\MPTN\BIN\MPTSTART.CMD >NUL

RUN=D:\IBMCOM\PROTOCOL\NBTCP.EXE
RUN=D:\MPTN\BIN\VDOSCTL.EXE
DEVICE=D:\IBMCOM\PROTOCOL\NETBEUI.OS2
DEVICE=D:\IBMCOM\PROTOCOL\TCPBEUI.OS2

Hope this helps,
Ilya

On Fri, 11 Nov 2005 01:58:26 UTC, James Knott
wrote:

> Try it when the wired connection is up as well. You may find that when the
> ethernet port is down, that NetBIOS/IP chokes, because that adapter is
> listed first. In another thread, I recommended binding NetBIOS over TCP/IP
> to the NULL MAC adapter and only TCP/IP to the real adapters.

I have finally got NetBIOS working over my WiFi link. What I had to do
was bind TCP/IP, NetBIOS and NetBIOS over TCP/IP to the wireless
adaptor.

Without NetBIOS, NET START PEER timed out (I still don't know what
it's waiting for) unless the laptop had been running for more than a
few hours, in which case it occasionally worked.

Without NetBIOS over TCP/IP I can't get a response out of the only
other thing on my network at the moment, an ASUS WL-HDD25 combined WAP
and NAS box (and a very neat little bit of equipment too, I add).

Binding to the NULL adaptor didn't seem to work (I tried NetBIOS over
TCP/IP and TCP/IP bound to NULL, TCP/IP bound to wireless and NetBIOS
bound to either or neither).

All most odd, but it seems to be working at the moment...

Ian