gmane.org.netlabs.genmac.user

Updated GenmU_TEST

Doug Bissett — 2012-04-02T23:18:02

I have updated the test version of GENMU at:


(use the word "unsupported" - without the quotes, and in lower case - 
if you are asked for a password). The devices that are waiting for 
confirmation that they work (or don't) are:
RALINK Wlan [1814:2790]
Atheros Wlan [168C:0023]
3COM Megahertz 10/100 LAN CardBus PC Card [10B7:5157]
Atheros Wlan [168C:002B]
Realtek 8172 Wlan [10EC:8172]
Realtek 8199 Wlan [10EC:8199]
Ralink Wlan [1814:0781]
Centrino Ultimate-N 6300 3x3 AGN [8086:4238]
Broadcom NetXtreme Gigabit [14E4:1648]
Broadcom NetXtreme Gigabit [14E4:16B7]
Atheros Wlan [168C:0024]

The main change (other than a couple of devices) is that the installer 

now performs the patching of \IBMCOM\MACS\genm32w.os2 for you. It does 

expect to find an unaltered version of genm32w.os2 in that location, or 

the patch can silently fail to happen. PLEASE READ the 
GENMU_TEST_README.TXT file.

In fact, the 168C:0024 has been made to work, which probably means that 

the 168C:0023 and 168C:002B devices will also work, but I left it here 

for testing purposes.

NOTE: that the GenMU-0-1-6.wpi installer has not, yet, been changed to 

use the new install method (which only applies to Wireless Lan 
Adapters).

If you have one of the above devices, please test this, and let me know 

if it works. You should UNINSTALL any previously installed GENMU 
devices before trying these ones.

Thanks...

Re: MAC addresses

Eirik Romstad — 2012-03-14T01:32:27

Thanks to Doug for his answer to my question.  I tried your suggestion, but it
did not work (the MAC address is still 000000:000000 in lantran.log).  As Doug I
was concerned that the restriction on one card would be a problem, but I think
it is not as long as there is only one card that uses GenMac (my wired
connection is the MultiMac Intel E1000E driver that works excellent).

Eirik

Re: MAC addresses

Doug Bissett — 2012-03-13T01:17:52

I have never tried it, but you might have some luck with:

http://hobbes.nmsu.edu/download/pub/os2/system/drivers/network/genmac_ma
c.zip

As I understand it, the MAC address of all zeros is, in fact, in the 
specs (as long as there is only one in any given network), but it seems 
that some routers will accept that, while others won't. Whether that is 
what your problem is, or not, I have no idea.

MAC addresses

Eirik Romstad — 2012-03-12T21:53:29

My quest for making the Intel Centrion 6300 a/g/n/ card work seemed to have
ended when nobody were able to explain the error message

WRND32$  WRAPPER ErrorLog 0x40001b7c

in Lantran.log.  One possible explanation for this error could be that the MAC
address is 

WRND32$  WRAPPER hardware init successfull MAC : 000000:000000.

On the eComStation.ru FAQ pages for unsupported GenMac installation it says that
if the MAC address is 000000:000000, one needs to set the MAC address manually.
 Unfortunately, it does not say how this is done.  I've searched the OS/2 help
pages where it says:

Use NET RIPLMACH to create, change, display, or delete a remote IPL (RIPL)
requester definition using parameters that are specified on the command line.
Use this command only with WorkSpace On-Demand requester definitions

with the following specific command: NET RIPLMAC /MA(C) mac:adr

Anyone with experience here, or with another suggestion?

Re: Broadcom BCM5751PKFB (integrated): No Valid Hardware Found

Doug Bissett — 2012-03-06T17:21:17

I could make a few other suggestions, that have nothing to do with the 
problem, but for now, I will suggest that you add a blank line at the 
end, right after the GENMAC driver line. Some people say it makes no 
difference, but others find otherwise, especially when the last line 
follows a non blank line, and contains an EOF (which is what you 
have).

Also, REM the powerman.exe line. That may be causing your choppy 
activity (it needs to be removed when you update to the ACPI 3.19.15 
release, which I highly recommend, anyway).

You could also try adding:
BASEDEV=BOOTDLY.SYS /D:2
immediatly following the
PSD=ACPI.PSD
line. That fixed (actually worked around) a problem on my IBM ThinkPad 
T43 where it couldn't read LANMSGEX.EXE during boot.

You could also try my LCSS (Logical Config.Sys Sort) package, to make 
it much easier to actually read your CONFIG.SYS (human reading, the 
machine doesn't care). It will also correct the non blank last line, 
and get rid of the EOF.


Be sure to READ the instructions.

Re: Broadcom BCM5751PKFB (integrated): No Valid Hardware Found

Al Savage — 2012-03-06T15:32:55

On Mon, 5 Mar 2012 21:39:02 UTC, "Doug Bissett" <dougb007< at >telus.net> 
wrote:


About 20 lines in:
RUN=C:\ECS\SYSTEM\GenMac\DRIVER\HELPERW.EXE


Yes. I have four of these DC7100s:
*  I'm running FreeNAS 0.7x on one of these boxes for months, no problem
with networking.
* WinXP SP3 on another, for almost a year.
* And, I can remove the GenMAC driver and use the native Broadcom 
Gigabit Extreme driver on this eCS installation, and it does work.  
NETBEUI over TCP/IP. But it's choppy: copying multi-gigabytes of data, 
the networking will disappear in the middle of copying.  I might get a 
SYS0054 or a different error. Most of the time, re-running 
\MPTN\BIN\SETUP.CMD (ie flushing the routes and arp tables) will get it 
going again, but it's a mess.  As this eCS box is supposed to replace my
aging W4-over-Warp Server Advanced box with HPFS386 that had been 
rock-solid for a decade, choppy networking isn't going to cut it.  But, 
these are problems for a later day: for now, I'd like to figure out why 
GenMAC isn't working for a "supported" adapter.  Might help someone 
else, too.


Have done so.

GENMAC UNSUPPORTED updates

Doug Bissett — 2012-03-06T03:07:38

Thanks to Lewis Rosenthal, I have added two devices to the GENMU 
package, and one to the GENMU_TEST package.

The GENMU package now contains:
< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >
UNOFFICIAL DEVICES supported as UNSUPPORTED:
=====================
Atheros Wlan AR5007EG/AR242X [168C:001C]
Attansic Fast Ethernet [1969:2048]
Broadcom Gigabit [14E4:1601]
Broadcom Gigabit [14E4:165A]
Broadcom Gigabit [14E4:1693]
Broadcom GigaBit [14E4:16A7]
Broadcom GigaBit [14E4:16C7]


OFFICIAL DEVICES UPDATED:
====================
Intel 2200 Wlan [8086:4220]
Intel 2915abg Wlan [8086:4224]
< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >

The Broadcom GigaBit [14E4:16A7] and Broadcom GigaBit [14E4:16C7] are 
new.

The GENMU_TEST package now contains:
< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >
TEST devices included:
RALINK [1814:2790]
Atheros Wlan [168C:0023]
3COM Megahertz 10/100 LAN CardBus PC Card [10B7:5157]
Atheros Wlan [168C:002B]
Realtek Wlan [10EC:8172]
Realtek Wlan [10EC:8199]
Ralink Wlan [1814:0781]
Centrino Ultimate-N 6300 3x3 AGN [8086:4238]
Broadcom NetXtreme Gigabit [14E4:1648]
Broadcom NetXtreme Gigabit [14E4:16B7]
< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >< at >

The Broadcom NetXtreme Gigabit [14E4:16B7] is new, and will probably 
work, but it is untested.

NOTE: none of the GENMU_TEST devices have been reported to work, or not 
work (except the Centrino Ultimate-N 6300 3x3 AGN [8086:4238] and the 
Realtek Wlan [10EC:8172], which have been tried, without success). If 
you have one of them please see what you can do to test them.  Be sure 
to READ the README. Some of the test devices also have alternate 
drivers, so you should try all of them. Just change the appropriate 
file names in the \IBMCOM\PROTOCOL.INI file (it is a plain text file, 
and it should be obvious what to change).

You can get both packages, and the READMEs at:


Use the password "unsupported" (without the quotes, and lower case) if 
you are asked.

Re: What devices (wired or wireless) do people require support for?

Paul Smedley — 2012-03-06T01:03:22

Hi Steven,

On 06/03/12 02:57, Steven Levine wrote:
Yes - Pasha clarified the need for os4 kernel - in the readme it 
suggests otherwise....


I suspect it will be a GPL violation - but then, I haven't seen the SDK 
yet - although I have asked for it.

Cheers,

Paul

Re: Broadcom BCM5751PKFB (integrated): No Valid Hardware Found

Paul Smedley — 2012-03-06T01:02:27

Hi Guys,

On 06/03/12 08:09, Doug Bissett wrote:

FWIW - the tigon3 broadcom driver looks like it may be _relatively_ easy 
to get working with multimac - at least it's source code is made up on 
only two files :)

Cheers,

Paul

Re: Broadcom BCM5751PKFB (integrated): No Valid Hardware Found

Doug Bissett — 2012-03-05T21:39:02

Yeah. I must cut off that finger  :-)


Okay, all of that says that the driver is not loading, for some 
reason.

...snip...

Okay, that says that it is a [14E4:1677] device.


PCI.EXE says it is a 1677.

The only thing I forgot to ask (that I can think of): 
Is the line:
RUN=x:\ecs\system\genmac\DRIVER\HELPERW.EXE
near the top of your CONFIG.SYS? (x: is your boot drive), and your path 
may be different if you installed GENMAC elsewhere).

I assume that this thing actually works in that other OS.

I doubt if it will help, but send me your CONFIG.SYS, and I will see if 
I can see anything there that might cause trouble. (send it to my 
private address).

Re: Broadcom BCM5751PKFB (integrated): No Valid Hardware Found

Al Savage — 2012-03-05T21:04:48

On Mon, 5 Mar 2012 03:14:24 UTC, "Doug Bissett" <dougb007< at >telus.net> 
wrote:


Without a DEBUGLEVEL line at all:

[C:\]type wrnddb$

[C:\]

Same with  DEBUGLEVEL="NONE".


I assume the redirect was a typo, but just in case I tried it anyway:

[C:\]copy wrnd32$ > mylogfile.log
SYS0002: The system cannot find the file specified.

[C:\]copy wrnd32$  mylogfile.log
wrnd32$
SYS0002: The system cannot find the file specified.
        0 file(s) copied.


IBM OS/2 LANMSGDD [05/21/04] 6.00 is loaded and operational.
IBM OS/2 LAN Netbind
PRO0025: An error occurred when the program tried to bind TCPIP_NIF to 
W14E4x1677_nif.

 

Bus 64 (PCI Express), Device Number 0, Device Function 0
Vendor 14E4h Broadcom Corporation
Device 1677h NetXtreme BCM5751 Gigabit Ethernet PCI Express
Command 0106h (Memory Access, BusMaster, System Errors)
Status 0010h (Has Capabilities List, Fast Timing)
Revision 01h, Header Type 00h, Bus Latency Timer 00h
Self test 00h (Self test not supported)
Cache line size 64 Bytes (16 DWords)
PCI Class Network, type Ethernet
Subsystem ID 3005103Ch DC7100 SFF(DX878AV) (Guess Only!)
Subsystem Vendor 103Ch Hewlett-Packard Company
Address 0 is a Memory Address (64-bit) : F0500000h
CIS can read from expansion ROM image 0, offset 00h
System IRQ 17, INT# A
New Capabilities List Present:
  Power Management Capability, Version 1.1
    Does not support low power State D1 or D2
    Supports PME# signalling from mode(s) D3hot, D3cold
    PME# signalling is currently disabled
    3.3v AUX Current required : 0 mA (Self powered)
    Current Power State : D0 (Device operational, no power saving)
    10000 mW D0 power consume
    10000 mW D0 power dissipate
        0 mW D1 power consume
        0 mW D1 power dissipate
        0 mW D2 power consume
        0 mW D2 power dissipate
     1000 mW D3 power consume
     1000 mW D3 power dissipate
        0 mW common logic power consume
   Vital Product Data Capability
     Identifier          Broadcom NetXtreme Gigabit Ethernet Controller 
     Part Number         BCM95751
     EC Level (Version)  106679-15
     Serial Number       0123456789
     Manufacturer ID     14e4
     Checksum/End        27h 28*00h
     System Specific     XYZ01234567
     Remaining RW Area   107*00h
     End Tag
   Message Signalled Interrupt Capability
     MSI is disabled
     MSI function can generate 64-bit addresses
   PCI Express Capability, Version 1
     Device/Port Type :
       PCI Express Endpoint Device
     Device Capabilities :
       Unsupported Request Severity is Non-Fatal
     Device Status :
       Non-Fatal Error Detected
       Unsupported Request Detected
       AUX Power Detected
     Link Capabilities :
       Maximum Link speed : 2.5Gb/s
       Maximum Link Width : x1
       Link Port Number   : 0
     Link Control :
       Common Clock Configuration In Use
     Link Status :
       Current Link speed : 2.5Gb/s
       Current Link Width : x1

[ . . . ]
Shared IRQs: IRQ 17 is shared by 3 PCI Devices
IRQ control is currently managed by the system APIC controller - IRQ 
info is
not actual hardware settings...



Is there a method to determine the device ID?

Re: Re: What devices (wired or wireless) do people require support for?

Steven Levine — 2012-03-05T16:27:03

In <jj1u7i$o0n$1< at >dough.gmane.org>, on 03/05/12
   at 07:11 PM, Paul Smedley <paul< at >smedley.id.au> said:

Hi Paul,


Unimac is effectively part of the os4 project.  The driver requires the
os4 kernel according to Jerry, but I suspect Jerry misread Pasha's post. 
As I read the post, the os4 kernel is needed only for debugging.

We shall see about the source code.  Pasha's note implies he plans to keep
part of the sources closed.  I don't know enough about the architecture of
unimac to have an opinion as to whether or not this will be a GPL
violation.

Steven

Re: What devices (wired or wireless) do people require support for?

Paul Smedley — 2012-03-05T08:41:22

Hi Steven,

On 05/03/12 18:12, Steven Levine wrote:



Sounds like Pasha's UniMac - except Pasha doesn't release source code :P

Cheers,

Paul

Re: Re: What devices (wired or wireless) do people require support for?

Steven Levine — 2012-03-05T07:42:23

In <jiv895$ogi$1< at >dough.gmane.org>, on 03/04/12
   at 06:44 PM, Paul Smedley <paul< at >smedley.id.au> said:

Hi,


He did the r8110 first.  The previous two developers did not have the
skill set needed to create a working driver.

P:\IBMCOM\MACS\r8110.os2
 2-25-12  16:33          31,705      0  r8110.os2

Build Level Display Facility Version 6.12.675 Sep 25 2001
(C) Copyright IBM Corporation 1993-2001
Signature:       < at >#Mensys BV:0.1.7#< at >##1## 25 Feb 2012 16:33:01     DAZAR1   
::::7::Beta< at >< at >Realtek 8110/1Gb PCIe MAC Driver Vendor:          Mensys BV
Revision:        0.01.7
Date/Time:       25 Feb 2012 16:33:01   
Build Machine:   DAZAR1    
FixPak Version:  Beta
File Version:    0.1.7
Description:     Realtek 8110/1Gb PCIe MAC Driver

eda2489eb494d739a657835ccacb77be *r8110.os2

David tells me he is not done with the sources for this driver so it too
will probably get a refresh.


It did and it still does.  However, what we have now is much closer to the
Linux code.


Absolutely.  The goal is to just be able to drop the linux code into it's
directory and build.

I've been thinking a bit about a support driver or possibly a DLL that
would provide the glue between the Linux code and the eCS/OS2 kernel.

Steven

Re: Broadcom BCM5751PKFB (integrated): No Valid Hardware Found

Doug Bissett — 2012-03-05T03:14:24

The version that comes with eCS 2.1 has a few devices added in 
UNSUPPORTED mode, and the Atheros windows drivers have been ripped out 
(licencing concerns). Otherwise, it is the same as the one from 
netlabs.


Hmmm. That one should be supported.


Cosmetic. There is room in the name for "WRND322$" (the second genmac 
driver, if used).


Try without that as well as with it. There may be problems with 
DEBUGLEVEL="ALL".


what does
copy wrnd32$ > mylogfile.log
give you? (I expect nothing since the driver isn't loading).


What does \IBMCOM\LANTRAN.LOG have to say?

What does PCI.EXE have to say about the device? What you are seeing 
indicates that the device may not be a [14E4:1677], but it is likely 
similar enough to make it work in UNSUPPORTED mode once you figure out 
the proper device ID.

Broadcom BCM5751PKFB (integrated): No Valid Hardware Found

Al Savage — 2012-03-04T19:03:13

I have four HP DC7100 desktop PCs, circa 2005. Single processor, single 
core 3.0Ghz Intel.
I have installed eCS 2.1GA to one of them.  Single OS, single HDD, new 
blank HDD (WD20EARS, 2TiB) using Promise TX4 SATA 300 PIC host adapter.

MEMTEST86+ 4.20 tested the 2GiB RAM OK.

Kernel is 105_SMP

ACPI is 3.19.15, installed via the WarpIn pkg. (This hardware will hang 
at boot if installing "Modern Hardware, safe mode". Hangs during boot if
the PSD= line is REMd.)

GenMAC 2.2 (I did download the "Full Installation" from Netlabs and 
installed that one too, even though it appears to be the same as shipped
with eCS 2.1GA: no change before or after.)

The NIC is integrated, Broadcom BCM5751PKFB (14E4:1677).  The mainboard 
has only two expansion slots, so I am very interested in using this 
integrated NIC (rather than adding a NIC to a valuable PCI slot).

During boot, I see the GenMAC wrapper driver stating, "No Valid Hardware
Found for "WRND32$ " (note the space between '$' & the ending 
double-quote).

In PROTOCOL.INI, I have
   DEBUGLEVEL = "ALL"

At boot completion to desktop, at a command prompt:
C:\> type wrnddb$ > mylogfile.log

The file MYLOGFILE.LOG is zero-bytes (empty).

The file C:\ECS\SYSTEM\GENMAC\DRIVER\WRND32.LOG contains only one line:
//------------ Version 2.1 ------------------


The entire PROTOCOL.INI:
================================
[PROT_MAN]

   DRIVERNAME = PROTMAN$

[IBMLXCFG]

   tcpip_nif = tcpip.nif
   W14E4x1677_nif = W14E4x1677.nif

[tcpip_nif]

   DriverName = TCPIP$
   Bindings = W14E4x1677_nif

[W14E4x1677_nif]

   DriverName = WRND32$
   VENDOR = 0x14E4
   DEVICE = 0x1677
   NDIS_SYS = "B57XP32.SYS"
   NDIS_INF = "B57XP32.INF"
   DEBUGLEVEL = "ALL"
   OPTIONS = "NO"
==================================

What is my next step?  Can I provide any additional information?

Re: What devices (wired or wireless) do people require support for?

Steven Levine — 2012-03-04T17:34:06

In <100.d83b05000647524f.004< at >telus.net>, on 03/03/12
   at 09:29 AM, "Doug Bissett" <dougb007< at >telus.net> said:

Hi Doug,


Pretty much.  It's been done before.  For example there's lxapi32.  A less
generic implemenation is uniaud.


There always will be.  ACPI continues to evolve and, even now, not all
vendors do fully complaint implementations.  This is not an issue for the
vendors, but it obviously is for eCS.

Steven

Re: What devices (wired or wireless) do people require support for?

Paul Smedley — 2012-03-04T08:14:27

Hi Steven,

On 03/03/12 16:18, Steven Levine wrote:
Yup I've been talking to David a little via email. I think you mean the 
e1000e driver above? That one he just resynced against the 3.2.4 linux 
kernel.

The e1000e driver appears fairly 'pure' in this respect, in that there 
are basically only diffs to make it compile with Watcom.

The realtek driver is a different story, as the base code substantially 
differs from the linux code - David reported he is working on getting 
this one to the same state as the e1000e driver - ie so that it has 
minimal changes from the linux code.

Agree - to me, this means moving definitions into the appropriate linux 
header - like uniaud, saves having to modify #include lines from the 
network card source files.

Cheers,

Paul

Re: What devices (wired or wireless) do people require support for?

Doug Bissett — 2012-03-03T16:29:58

,,,snip...

Sounds like you are reinventing GENMAC, but using Linux drivers, rather 
than Windows drivers (nothing wrong with that).


It has done that, but there is still work to be done.


Many thanks to you, and all others who are working to advance the 
cause.

Re: [genmac-user] What devices (wired or wireless) do people require support for?

Doug Bissett — 2012-03-03T06:28:31

This one is already on the supported by Multimac list. You can get it 
with Software subscription, or with eCS 2.1 (although the SS version is 
newer).

I use the Intel driver on my Lenovo ThinkPad T510 (8086:10EA), and it 
works great. Unfortunately, my Realtek wireless won't go with GENMAC 
(reported to Paul at os2world.com), but I suspect that part of the 
problem might be ACPI related (even with ACPI 3.19.15). I am waiting 
for Multimac to support wireless. I am even considering buying an Intel 
device (they aren't all that expensive) if that will mean that I can 
get support sooner.


I expect that there will be some overlap, which is good, but I agree 
that it is not worth the effort to make a specific Multimac driver for 
devices already supported by GENMAC. Most of those are already 
considered to be "old", and their life expectancy goes down by the day. 
Perhaps, after the technique is developed, and understood better, it 
will only  take a few minutes to put the parts together, and somebody 
might decide to create drivers for older devices. That would be great, 
but for now, trying to get newer devices to work is the main idea.

Re: What devices (wired or wireless) do people require support for?

Steven Levine — 2012-03-03T05:48:39

In <jirkmp$vv1$1< at >dough.gmane.org>, on 03/03/12
   at 09:52 AM, Paul Smedley <paul< at >smedley.id.au> said:

Hi Paul,


I recommend you get in contact with David Azarewicz and me.  David just
finished reworking the r8110 driver based on the Linux sources.  The
driver works fine on the systems we have tested it with.  He could
probably use some help cleaning up the code which he tells me is a tad
dirty.

We are looking at what can be done to move more of the common code out of
the realtek part of the driver.  We are also considering converting the
multimac sources from C++ to C.  This is more to accomodate the
limitiations of the OpenWatcom C++ implementation than any reluctance to
use C++ from driver development.


What we believe is needed for efficient wireless driver support is a clean
set of wrappers designed with the goal allowing Linux code to be dropped
in with the minumum number of modifications.

For example the ACPI codebase has been refactored to provide clear
separation between the Intel supplied ACPICA source code and the eCS/OS2
specific code.  This has already provided significant benefits.

Regards,

Steven