Neutrino 2.0c and Intel82559ER

We have 2 identical PC/104 boards, except one
has an Intel 82559 ethernet controller and the other
has an Intel 82559ER ethernet controller.

The devn-speedo.so file included with Neutrino 2.0c
works fine with the 82559, but says “netio: unable to init speedo”
with the 82559ER. The intel web site hints that these
parts should be driver compatible.

Any thoughts? I think earlier beta versions of this
speedo driver print “NOT GONNA WORK”, but actually
appears to work.

Upgrading to RTP isn’t an option yet…

Thanks!
Kevin Radke

Hi Kevin,

The version of Neutrino you are currently using is
older, you can get a newer version of the OS that doesn’t
come as a package (RTP). The current version of Neutrino
is 2.1

It will likly have a version of the driver that will function
properly.

P.S. The RTP runs on Neutrino 2.11


Kevin Radke <kradke@dsi-inc.net> wrote:

We have 2 identical PC/104 boards, except one
has an Intel 82559 ethernet controller and the other
has an Intel 82559ER ethernet controller.

The devn-speedo.so file included with Neutrino 2.0c
works fine with the 82559, but says “netio: unable to init speedo”
with the 82559ER. The intel web site hints that these
parts should be driver compatible.

Any thoughts? I think earlier beta versions of this
speedo driver print “NOT GONNA WORK”, but actually
appears to work.

Upgrading to RTP isn’t an option yet…

Thanks!
Kevin Radke

“Hardware Support Account” <hw@qnx.com> wrote in message
news:969dk2$15h$1@nntp.qnx.com

Hi Kevin,

The version of Neutrino you are currently using is
older, you can get a newer version of the OS that doesn’t
come as a package (RTP). The current version of Neutrino
is 2.1

It will likly have a version of the driver that will function
properly.

P.S. The RTP runs on Neutrino 2.11

Is Neutrino 2.1 compatible with the Windows hosted
cross compiler, and if so, why was 2.0c shipped with the “released”
version that shipped in January?

Thanks!
Kevin Radke

Hi Kevin,

Hmmm I talked to QA yesterday about what the released version of
Neutrino was and they informed me that it was 2.1, however I checked
the CD and you are correct it is 2.00C. My appologies.

Is there a particular reason for not going to RTP?



Kevin Radke <kradke@dsi-inc.net> wrote:

“Hardware Support Account” <> hw@qnx.com> > wrote in message
news:969dk2$15h$> 1@nntp.qnx.com> …
Hi Kevin,

The version of Neutrino you are currently using is
older, you can get a newer version of the OS that doesn’t
come as a package (RTP). The current version of Neutrino
is 2.1

It will likly have a version of the driver that will function
properly.

P.S. The RTP runs on Neutrino 2.11

Is Neutrino 2.1 compatible with the Windows hosted
cross compiler, and if so, why was 2.0c shipped with the “released”
version that shipped in January?

Thanks!
Kevin Radke

“Hardware Support Account” <hw@qnx.com> wrote in message
news:96bjkk$9n0$1@nntp.qnx.com

Hi Kevin,

Hmmm I talked to QA yesterday about what the released version of
Neutrino was and they informed me that it was 2.1, however I checked
the CD and you are correct it is 2.00C. My appologies.

Is there a particular reason for not going to RTP?

We currently require cross compilers that run under Windows, since
they are executed by our tool after it auto generates C++ code. This
application is heavily COM/ActiveX/MFC based so moving it to RTP
is not likely to happen any time soon.

Other options such as compiling on a remote RTP machine are possible,
but make it difficult to easily do demos with a single laptop and a small
embedded device.

Thanks!
Kevin

Hi Kevin,

You can use the cross compiler that Windows is using (Code Warrior I
am guessing) with the RTP without any problems.

Kevin Radke <kradke@dsi-inc.net> wrote:

“Hardware Support Account” <> hw@qnx.com> > wrote in message
news:96bjkk$9n0$> 1@nntp.qnx.com> …
Hi Kevin,

Hmmm I talked to QA yesterday about what the released version of
Neutrino was and they informed me that it was 2.1, however I checked
the CD and you are correct it is 2.00C. My appologies.

Is there a particular reason for not going to RTP?

We currently require cross compilers that run under Windows, since
they are executed by our tool after it auto generates C++ code. This
application is heavily COM/ActiveX/MFC based so moving it to RTP
is not likely to happen any time soon.

Other options such as compiling on a remote RTP machine are possible,
but make it difficult to easily do demos with a single laptop and a small
embedded device.

Thanks!
Kevin

Try to use the “pci” utility to verify the following information:

On a bad machine:

Vendor ID: 8086h
Device ID: 1209h

On good machine:

Vendor ID: 8086h
Device ID: 1229h

Have you tried forcing the vendor/device id on the command line to the
driver?

io-net -dspeedo vid=0x8086,did=0x1209,verbose=1 -pttcpip …

(worked for me on Neutrino and RTP 6)

NOTE:

Don’t place a “space” character between “,” and “did” !!!

“Otto Nielsen” <otnie@wmdata.com> wrote in message
news:96eggn$7ul$1@inn.qnx.com

Try to use the “pci” utility to verify the following information:

On a bad machine:

Vendor ID: 8086h
Device ID: 1209h

On good machine:

Vendor ID: 8086h
Device ID: 1229h

Have you tried forcing the vendor/device id on the command line to the
driver?

io-net -dspeedo vid=0x8086,did=0x1209,verbose=1 -pttcpip …

(worked for me on Neutrino and RTP 6)

NOTE:

Don’t place a “space” character between “,” and “did” !!!

YES! Specifying the “did” appears to make it work. Not sure why I
didn’t try this…

Thanks!
Kevin

Hi, I have 2 MBX2000 EBX systems that we have installed QNX 2.6 build C.
These boards use the 82559ER chip set which has been causing some
difficulties keeping our network setup information. I tried the manual
installation using io-net below but received errors but through persistance
I found using just,

io-net -dspeedo verbose -pttcpip verbose

worked. From here I was able to setup the IP and route information using
the standard GUI interface. This is where the problems escalated. Each
time I reset the machine(s) I loose all the network information. I have
created some scripts to reset them up but haven’t gotten the network to
consistantly work each reboot. After a reboot I receive these errors, I
believe from pci-bios:

Range Check Failed (MEM) Dev 1209 vend 8086 Class 20000 Addr fc02 0000 size
1000
Alloc failed fc02 0000 size 1000

Range Check Failed (MEM) Dev 1209 vend 8086 Class 20000 Addr fc00 0000 size
20000
Alloc failed fc00 0000 size 20000

From reading these newsgroups, I have tried slaying pci-bios and restarting
using pci-bios -vvv but have been unable to capture the output to a file
because the machine locks up. Here is what I copied down by hand.

Device 6 - Function 0 vendor 8086 Device 1209 Index 0 class code 20000
find_bus 0
Min_Gnt=8-Max_Lat=56-Bandwidth=115500000
Mem fc020000 1000
IO 1440 40
Mem fc000000 20000
Range failed 0-10000
Range check failed (ROM) Dev 1209 Vend 8086 class 20000 Addr 0 size 10000
Rom 0-10000
Mem fc020000 1000
IO 1440 40
Mem fc000000 20000
reserve IRQ 11

This is all the information I saw regarding the 82559ER. Is there an
updated pci-bios that will eliminate these errors or is there some
configuration file that can be edited?

Joe Barnes


Kevin Radke wrote:

“Otto Nielsen” <> otnie@wmdata.com> > wrote in message
news:96eggn$7ul$> 1@inn.qnx.com> …
Try to use the “pci” utility to verify the following information:

On a bad machine:

Vendor ID: 8086h
Device ID: 1209h

On good machine:

Vendor ID: 8086h
Device ID: 1229h

Have you tried forcing the vendor/device id on the command line to the
driver?

io-net -dspeedo vid=0x8086,did=0x1209,verbose=1 -pttcpip …

(worked for me on Neutrino and RTP 6)

NOTE:

Don’t place a “space” character between “,” and “did” !!!

YES! Specifying the “did” appears to make it work. Not sure why I
didn’t try this…

Thanks!
Kevin

Previously, Joseph Barnes wrote in qdn.public.qnxrtp.installation:

Hi, I have 2 MBX2000 EBX systems that we have installed QNX 2.6 build C.
These boards use the 82559ER chip set which has been causing some
difficulties keeping our network setup information. I tried the manual
installation using io-net below but received errors but through persistance
I found using just,

io-net -dspeedo verbose -pttcpip verbose

worked. From here I was able to setup the IP and route information using
the standard GUI interface. This is where the problems escalated. Each
time I reset the machine(s) I loose all the network information. I have
created some scripts to reset them up but haven’t gotten the network to
consistantly work each reboot. After a reboot I receive these errors, I
believe from pci-bios:

Range Check Failed (MEM) Dev 1209 vend 8086 Class 20000 Addr fc02 0000 size
1000
Alloc failed fc02 0000 size 1000

Range Check Failed (MEM) Dev 1209 vend 8086 Class 20000 Addr fc00 0000 size
20000
Alloc failed fc00 0000 size 20000

From reading these newsgroups, I have tried slaying pci-bios and restarting
using pci-bios -vvv but have been unable to capture the output to a file
because the machine locks up. Here is what I copied down by hand.

Device 6 - Function 0 vendor 8086 Device 1209 Index 0 class code 20000
find_bus 0
Min_Gnt=8-Max_Lat=56-Bandwidth=115500000
Mem fc020000 1000
IO 1440 40
Mem fc000000 20000
Range failed 0-10000
Range check failed (ROM) Dev 1209 Vend 8086 class 20000 Addr 0 size 10000
Rom 0-10000
Mem fc020000 1000
IO 1440 40
Mem fc000000 20000
reserve IRQ 11

This is all the information I saw regarding the 82559ER. Is there an
updated pci-bios that will eliminate these errors or is there some
configuration file that can be edited?

There is an updated pci-bios with RTP patch B that does not display this
information to the console. The range check failed on the ROM address and
this is not serious.

Hugh.

Joe Barnes


Kevin Radke wrote:

“Otto Nielsen” <> otnie@wmdata.com> > wrote in message
news:96eggn$7ul$> 1@inn.qnx.com> …
Try to use the “pci” utility to verify the following information:

On a bad machine:

Vendor ID: 8086h
Device ID: 1209h

On good machine:

Vendor ID: 8086h
Device ID: 1229h

Have you tried forcing the vendor/device id on the command line to the
driver?

io-net -dspeedo vid=0x8086,did=0x1209,verbose=1 -pttcpip …

(worked for me on Neutrino and RTP 6)

NOTE:

Don’t place a “space” character between “,” and “did” !!!

YES! Specifying the “did” appears to make it work. Not sure why I
didn’t try this…

Thanks!
Kevin
begin:vcard
n:Barnes;Joseph
tel;pager:(770)525-0916
tel;fax:(770)449-7740 or (800) 998-8421
tel;work:(770)441-2100 ext 127
x-mozilla-html:TRUE
org:TGA Technologies, Inc.;Engineering
version:2.1
email;internet:> jbarnes@tga.com
title:Sr. Software Engineer
adr;quoted-printable:;;100 Pinacle Way=0D=0ASuite 140;Norcross;Ga;30071-3633;USA
fn:Joseph Barnes
end:vcard

I have now modified my script to use the same command as suggested:
io-net -dspeedo vid=0x8086,did=0x1209,verbose=1 -pttcpip
if=en0:216.162.37.xxx:255.255.255.xxx
where the xxx can be any number within my network. The first time I tried it I had
added these parameters with different spacing which gave me the errors. This seams to
work fine because I can ping other systems on my network but I still cannot set the
route tables by any other means than the phlib GUI. So each time I reset the machine
I still have to retype in the route table. I have tried saving a copy of the net.cfg
that is written after I have configured the network via phlib. I copied this file to
net.cfg.bak and then try to use this file to configure netmanager. Here is the
configuration script that I wrote:

slay io-net
io-net -dspeedo vid=0x8086,did=0x1209,verbose=1 -pttcpip
if=en0:216.162.37.xxx:255.255.255.xxx
cp -p /etc/net.cfg.bak /etc/net.cfg
netmanager -r en0 -f /etc/net.cfg

Now when I run phlib I now see the information but when I try to ping systems on my
network I receive host down message. I have read in some other posts how to make
netstat, route, and ifconfig work, by using the full tcpip stack. Is there any other
way to configure the route information, using the tiny stack, besides the phlib GUI?
I plan to call a script from rc.local to configure my network during boot up.

Joe Barnes



Hugh Brown wrote:

Previously, Joseph Barnes wrote in qdn.public.qnxrtp.installation:
Hi, I have 2 MBX2000 EBX systems that we have installed QNX 2.6 build C.
These boards use the 82559ER chip set which has been causing some
difficulties keeping our network setup information. I tried the manual
installation using io-net below but received errors but through persistance
I found using just,

io-net -dspeedo verbose -pttcpip verbose

worked. From here I was able to setup the IP and route information using
the standard GUI interface. This is where the problems escalated. Each
time I reset the machine(s) I loose all the network information. I have
created some scripts to reset them up but haven’t gotten the network to
consistantly work each reboot. After a reboot I receive these errors, I
believe from pci-bios:

Range Check Failed (MEM) Dev 1209 vend 8086 Class 20000 Addr fc02 0000 size
1000
Alloc failed fc02 0000 size 1000

Range Check Failed (MEM) Dev 1209 vend 8086 Class 20000 Addr fc00 0000 size
20000
Alloc failed fc00 0000 size 20000

From reading these newsgroups, I have tried slaying pci-bios and restarting
using pci-bios -vvv but have been unable to capture the output to a file
because the machine locks up. Here is what I copied down by hand.

Device 6 - Function 0 vendor 8086 Device 1209 Index 0 class code 20000
find_bus 0
Min_Gnt=8-Max_Lat=56-Bandwidth=115500000
Mem fc020000 1000
IO 1440 40
Mem fc000000 20000
Range failed 0-10000
Range check failed (ROM) Dev 1209 Vend 8086 class 20000 Addr 0 size 10000
Rom 0-10000
Mem fc020000 1000
IO 1440 40
Mem fc000000 20000
reserve IRQ 11

This is all the information I saw regarding the 82559ER. Is there an
updated pci-bios that will eliminate these errors or is there some
configuration file that can be edited?


There is an updated pci-bios with RTP patch B that does not display this
information to the console. The range check failed on the ROM address and
this is not serious.

Hugh.

Joe Barnes


Kevin Radke wrote:

“Otto Nielsen” <> otnie@wmdata.com> > wrote in message
news:96eggn$7ul$> 1@inn.qnx.com> …
Try to use the “pci” utility to verify the following information:

On a bad machine:

Vendor ID: 8086h
Device ID: 1209h

On good machine:

Vendor ID: 8086h
Device ID: 1229h

Have you tried forcing the vendor/device id on the command line to the
driver?

io-net -dspeedo vid=0x8086,did=0x1209,verbose=1 -pttcpip …

(worked for me on Neutrino and RTP 6)

NOTE:

Don’t place a “space” character between “,” and “did” !!!

YES! Specifying the “did” appears to make it work. Not sure why I
didn’t try this…

Thanks!
Kevin
begin:vcard
n:Barnes;Joseph
tel;pager:(770)525-0916
tel;fax:(770)449-7740 or (800) 998-8421
tel;work:(770)441-2100 ext 127
x-mozilla-html:TRUE
org:TGA Technologies, Inc.;Engineering
version:2.1
email;internet:> jbarnes@tga.com
title:Sr. Software Engineer
adr;quoted-printable:;;100 Pinacle Way=0D=0ASuite 140;Norcross;Ga;30071-3633;USA
fn:Joseph Barnes
end:vcard

This should be documented. See the ‘route’ and ‘default’
options to the tiny stack (npm-ttcpip.so in the utilities
reference).

-seanb

Joseph Barnes <jtbarnes@mindspring.com> wrote:
: This is a multi-part message in MIME format.
: --------------89216C772BEA9BCF8FB241CF
: Content-Type: text/plain; charset=us-ascii
: Content-Transfer-Encoding: 7bit

: I have now modified my script to use the same command as suggested:
: io-net -dspeedo vid=0x8086,did=0x1209,verbose=1 -pttcpip
: if=en0:216.162.37.xxx:255.255.255.xxx
: where the xxx can be any number within my network. The first time I tried it I had
: added these parameters with different spacing which gave me the errors. This seams to
: work fine because I can ping other systems on my network but I still cannot set the
: route tables by any other means than the phlib GUI. So each time I reset the machine
: I still have to retype in the route table. I have tried saving a copy of the net.cfg
: that is written after I have configured the network via phlib. I copied this file to
: net.cfg.bak and then try to use this file to configure netmanager. Here is the
: configuration script that I wrote:

: slay io-net
: io-net -dspeedo vid=0x8086,did=0x1209,verbose=1 -pttcpip
: if=en0:216.162.37.xxx:255.255.255.xxx
: cp -p /etc/net.cfg.bak /etc/net.cfg
: netmanager -r en0 -f /etc/net.cfg

: Now when I run phlib I now see the information but when I try to ping systems on my
: network I receive host down message. I have read in some other posts how to make
: netstat, route, and ifconfig work, by using the full tcpip stack. Is there any other
: way to configure the route information, using the tiny stack, besides the phlib GUI?
: I plan to call a script from rc.local to configure my network during boot up.

: Joe Barnes

Also, make sure the PnP OS support is OFF in the BIOS. Our
newer JumpTec boards (with a slightly older version 2.10 of Neutrino
and the 82559ER chip) do not work with PnP on. The older Intel 82559
boards don’t really care.

Kevin