Project Home
Project Home
Wiki
Wiki
Discussion Forums
Discussions
Project Information
Project Info
Forum Topic - intel 82754 network driver problem: (18 Items)
   
intel 82754 network driver problem  
I'm using a motherboard with 2 82754 chips. The first run fine, but the second doesn't work.
The driver seems to be devn-e1000.so.
Is there an updated release of this driver ?
Where can I download it ?
Thank you
Paolo
Re: intel 82754 network driver problem  
What version of the O/S are you running? Also, please post the output of
'pci -v'.

Thanks.




On 2012-12-14 6:00 AM, "Paolo Gussago" <community-noreply@qnx.com> wrote:

>I'm using a motherboard with 2 82754 chips. The first run fine, but the
>second doesn't work.
>The driver seems to be devn-e1000.so.
>Is there an updated release of this driver ?
>Where can I download it ?
>Thank you
>Paolo
>
>
>
>
>_______________________________________________
>
>Networking Drivers
>http://community.qnx.com/sf/go/post98114
>To cancel your subscription to this discussion, please e-mail
>drivers-networking-unsubscribe@community.qnx.com

Re: intel 82754 network driver problem  
The O/S version is 6.5.0
Here is the output of pci -v :

PCI version    = 3.00

Class          = Bridge (Host/PCI)
Vendor ID      = 8086h, Intel Corporation 
Device ID      = 104h,  Sandy Bridge DRAM Controller
PCI index      = 0h
Class Codes    = 060000h
Revision ID    = 9h
Bus number     = 0
Device number  = 0
Function num   = 0
Status Reg     = 2090h
Command Reg    = 6h
Header type    = 0h Single-function
BIST           = 0h Build-in-self-test not supported
Latency Timer  = 0h
Cache Line Size= 0h 
Subsystem Vendor ID = 8086h
Subsystem ID        = 104h
Max Lat        = 0ns
Min Gnt        = 0ns
PCI Int Pin    = NC
Interrupt line = 0
CPU Interrupt  = 0h
Capabilities Pointer = e0h
Capability ID        = 9h - Vendor Specific
Capabilities         = 10ch - a0806196h 

Class          = Bridge (PCI/PCI)
Vendor ID      = 8086h, Intel Corporation 
Device ID      = 101h,  Sandy Bridge PCI Express Root Port
PCI index      = 0h
Class Codes    = 060400h
Revision ID    = 9h
Bus number     = 0
Device number  = 1
Function num   = 0
Status Reg     = 10h
Command Reg    = 7h
Header type    = 1h Multi-function
BIST           = 0h Build-in-self-test not supported
Latency Timer  = 0h
Cache Line Size= 10h un-cacheable
Primary Bus Number       = 0
Secondary Bus Number     = 1
Subordinate Bus Number   = 1
Secondary Latency Timer  = 0h
I/O Base                 = f0h
I/O Limit                = 0h
Secondary Status         = 0h
Memory Base              = fff0h
Memory Limit             = 0h
Prefetchable Memory Base = fff1h
Prefetchable Memory Limit= 1h
Prefetchable Base Upper 32 Bits  = 0h
Prefetchable Limit Upper 32 Bits = 0h
I/O Base Upper 16 Bits   = 0h
I/O Limit Upper 16 Bits  = 0h
Bridge Control           = 10h
PCI Int Pin              = INT A
Interrupt line           = 11
CPU Interrupt            = bh
Capabilities Pointer = 88h
Capability ID        = dh - PCI Bridge Subsystem Vendor ID
Capabilities         = 0h - 1018086h
Capability ID        = 1h - Power Management
Capabilities         = c803h - 8h
Capability ID        = 5h - Message Signaled Interrupts
Capabilities         = 0h - 0h
Capability ID        = 10h - PCI Express
Capabilities         = 142h - 8000h

Class          = Bridge (PCI/PCI)
Vendor ID      = 8086h, Intel Corporation 
Device ID      = 109h,  Sandy Bridge PCI Express Root Port
PCI index      = 0h
Class Codes    = 060400h
Revision ID    = 9h
Bus number     = 0
Device number  = 1
Function num   = 2
Status Reg     = 10h
Command Reg    = 7h
Header type    = 1h Multi-function
BIST           = 0h Build-in-self-test not supported
Latency Timer  = 0h
Cache Line Size= 10h un-cacheable
Primary Bus Number       = 0
Secondary Bus Number     = 2
Subordinate Bus Number   = 2
Secondary Latency Timer  = 0h
I/O Base                 = f0h
I/O Limit                = 0h
Secondary Status         = 0h
Memory Base              = fff0h
Memory Limit             = 0h
Prefetchable Memory Base = fff1h
Prefetchable Memory Limit= 1h
Prefetchable Base Upper 32 Bits  = 0h
Prefetchable Limit Upper 32 Bits = 0h
I/O Base Upper 16 Bits   = 0h
I/O Limit Upper 16 Bits  = 0h
Bridge Control           = 10h
PCI Int Pin              = INT A
Interrupt line           = 11
CPU Interrupt            = bh
Capabilities Pointer = 88h
Capability ID        = dh - PCI Bridge Subsystem Vendor ID
Capabilities         = 0h - 1098086h
Capability ID        = 1h - Power Management
Capabilities         = c803h - 8h
Capability ID        = 5h - Message Signaled Interrupts
Capabilities         = 0h - 0h
Capability ID        = 10h - PCI Express
Capabilities         = 142h - 8000h

Class          = Display (VGA)
Vendor ID      = 8086h, Intel Corporation 
Device ID      = 106h,  Sandy Bridge Integrated Graphics Controller
PCI index      = 0h
Class Codes    = 030000h
Revision ID    = 9h
Bus number     = 0
Device number  = 2
Function num   = 0
Status Reg     = 90h
Command...
View Full Message
Re: intel 82754 network driver problem  
The latest driver and enumeration file can be found here
http://community.qnx.com/sf/wiki/do/viewPage/projects.bsp/wiki/Experimental
DriversAndUtilities




On 2012-12-14 9:23 AM, "Paolo Gussago" <community-noreply@qnx.com> wrote:

>The O/S version is 6.5.0
>Here is the output of pci -v :
>
>PCI version    = 3.00
>
>Class          = Bridge (Host/PCI)
>Vendor ID      = 8086h, Intel Corporation
>Device ID      = 104h,  Sandy Bridge DRAM Controller
>PCI index      = 0h
>Class Codes    = 060000h
>Revision ID    = 9h
>Bus number     = 0
>Device number  = 0
>Function num   = 0
>Status Reg     = 2090h
>Command Reg    = 6h
>Header type    = 0h Single-function
>BIST           = 0h Build-in-self-test not supported
>Latency Timer  = 0h
>Cache Line Size= 0h
>Subsystem Vendor ID = 8086h
>Subsystem ID        = 104h
>Max Lat        = 0ns
>Min Gnt        = 0ns
>PCI Int Pin    = NC
>Interrupt line = 0
>CPU Interrupt  = 0h
>Capabilities Pointer = e0h
>Capability ID        = 9h - Vendor Specific
>Capabilities         = 10ch - a0806196h
>
>Class          = Bridge (PCI/PCI)
>Vendor ID      = 8086h, Intel Corporation
>Device ID      = 101h,  Sandy Bridge PCI Express Root Port
>PCI index      = 0h
>Class Codes    = 060400h
>Revision ID    = 9h
>Bus number     = 0
>Device number  = 1
>Function num   = 0
>Status Reg     = 10h
>Command Reg    = 7h
>Header type    = 1h Multi-function
>BIST           = 0h Build-in-self-test not supported
>Latency Timer  = 0h
>Cache Line Size= 10h un-cacheable
>Primary Bus Number       = 0
>Secondary Bus Number     = 1
>Subordinate Bus Number   = 1
>Secondary Latency Timer  = 0h
>I/O Base                 = f0h
>I/O Limit                = 0h
>Secondary Status         = 0h
>Memory Base              = fff0h
>Memory Limit             = 0h
>Prefetchable Memory Base = fff1h
>Prefetchable Memory Limit= 1h
>Prefetchable Base Upper 32 Bits  = 0h
>Prefetchable Limit Upper 32 Bits = 0h
>I/O Base Upper 16 Bits   = 0h
>I/O Limit Upper 16 Bits  = 0h
>Bridge Control           = 10h
>PCI Int Pin              = INT A
>Interrupt line           = 11
>CPU Interrupt            = bh
>Capabilities Pointer = 88h
>Capability ID        = dh - PCI Bridge Subsystem Vendor ID
>Capabilities         = 0h - 1018086h
>Capability ID        = 1h - Power Management
>Capabilities         = c803h - 8h
>Capability ID        = 5h - Message Signaled Interrupts
>Capabilities         = 0h - 0h
>Capability ID        = 10h - PCI Express
>Capabilities         = 142h - 8000h
>
>Class          = Bridge (PCI/PCI)
>Vendor ID      = 8086h, Intel Corporation
>Device ID      = 109h,  Sandy Bridge PCI Express Root Port
>PCI index      = 0h
>Class Codes    = 060400h
>Revision ID    = 9h
>Bus number     = 0
>Device number  = 1
>Function num   = 2
>Status Reg     = 10h
>Command Reg    = 7h
>Header type    = 1h Multi-function
>BIST           = 0h Build-in-self-test not supported
>Latency Timer  = 0h
>Cache Line Size= 10h un-cacheable
>Primary Bus Number       = 0
>Secondary Bus Number     = 2
>Subordinate Bus Number   = 2
>Secondary Latency Timer  = 0h
>I/O Base                 = f0h
>I/O Limit                = 0h
>Secondary Status         = 0h
>Memory Base              = fff0h
>Memory Limit             = 0h
>Prefetchable Memory Base = fff1h
>Prefetchable Memory Limit= 1h
>Prefetchable Base Upper 32 Bits  = 0h
>Prefetchable Limit Upper 32 Bits = 0h
>I/O Base Upper 16 Bits   = 0h
>I/O Limit Upper 16 Bits  = 0h
>Bridge Control           = 10h
>PCI Int Pin              = INT A
>Interrupt line           = 11
>CPU Interrupt            =...
View Full Message
Re: intel 82754 network driver problem  
Hi all, 

I have the same problem using a motherboard with 2 Intel 82574L ethernet ports.
I've already downloaded and installed the last devnp-e1000.so driver release (found on the Experimentals linked page) 
but it doesn't solve the problem.
Attached to this post you'll find a ZIP archive containing some LOGs, obtained from QNX 6.5.0.
More in details, the ZIP archive contains the following files:
- pci.txt
pci device list (both ethernet are listed)

- ifconfig.txt
"ifconfig" command's output

- net.txt
"/etc/system/enum/include/net" file contents

- routeShow.txt
"route show" command's output  (you can see that the address "10.3.2.200" appears associated to the  wm0 interface (see 
the MAC Address in ifconfig for confirmation), while the address "10.3.2.201" - related to the wm1 interface, is not 
listed in the routing table

- pingWM0.txt
"ping 10.3.2.200" command's output

- pingWM1.txt
"ping 10.3.2.201" command's output

Is there a solution for this problem? What can be related to?
Attachment: Compressed file NETLogs.zip 2.2 KB
Re: intel 82754 network driver problem  
The issue you are seeing is because you have configured two interfaces 
on the same network within a single instance of io-pkt. Note that each 
instance of io-pkt has one routing table and it will discard redundant 
routes. In this case the route entry for wm1 is redundant since the 
entry for wm0 already points to the network. This is why you see pings 
working with wm0 but not wm1. In fact wm1 should be able to receive 
incoming packets but will be unable to transmit any since the packets 
will be sent of via wm0 (again because of the way the routing table is 
setup). This is by design. Normally it is not recommended to configure 
two interfaces on the network. You could either use a separate instance 
of io-pkt for each interface (which gives each interface its own routing 
table) or you might want to look into using a bridge.

On 2014-11-03 01:39 PM, Loris Colasante wrote:
> Hi all,
>
> I have the same problem using a motherboard with 2 Intel 82574L ethernet ports.
> I've already downloaded and installed the last devnp-e1000.so driver release (found on the Experimentals linked page) 
but it doesn't solve the problem.
> Attached to this post you'll find a ZIP archive containing some LOGs, obtained from QNX 6.5.0.
> More in details, the ZIP archive contains the following files:
> - pci.txt
> pci device list (both ethernet are listed)
>
> - ifconfig.txt
> "ifconfig" command's output
>
> - net.txt
> "/etc/system/enum/include/net" file contents
>
> - routeShow.txt
> "route show" command's output  (you can see that the address "10.3.2.200" appears associated to the  wm0 interface 
(see the MAC Address in ifconfig for confirmation), while the address "10.3.2.201" - related to the wm1 interface, is 
not listed in the routing table
>
> - pingWM0.txt
> "ping 10.3.2.200" command's output
>
> - pingWM1.txt
> "ping 10.3.2.201" command's output
>
> Is there a solution for this problem? What can be related to?
>
>
>
> _______________________________________________
>
> Networking Drivers
> http://community.qnx.com/sf/go/post112240
> To cancel your subscription to this discussion, please e-mail drivers-networking-unsubscribe@community.qnx.com

Re: intel 82754 network driver problem  
First of all thank you for your quick reply. Now it's more clear what this behaviour is related to.
Anyway, because I need to use both interfaces, I'm trying to start 2 different instances of io-pkt but with no success.
I've modified the "etc/system/enum/include/net" file as reported below:

-------------------------------------------------------------------------------------------
#
# macro definitions for network
#
	
all
	set(IOPKT_CMD, io-pkt-v4-hc -i1 unit=0 -ptcpip)
	set(IOPKT_CMD, io-pkt-v4-hc -i2 unit=1 -ptcpip)

-------------------------------------------------------------------------------------------

but in this way the following errors occurs at the startup:

-------------------------------------------------------
mount: Can't mount / (type io-pkt)
mount: Possible reason: Invalid argument
mount: Can't mount / (type io-pkt)
mount: Possible reason: Invalid argument
-------------------------------------------------------

By specifying only the '-i1' and -i2' option (without 'unit=xx') the same error occurs.

So ,which is the right way to create 2 separate instances of io-pkt, one for each interface 'wm0' and 'wm1', running at 
startup?
Could you help me, please?

Thank you in advance

> The issue you are seeing is because you have configured two interfaces 
> on the same network within a single instance of io-pkt. Note that each 
> instance of io-pkt has one routing table and it will discard redundant 
> routes. In this case the route entry for wm1 is redundant since the 
> entry for wm0 already points to the network. This is why you see pings 
> working with wm0 but not wm1. In fact wm1 should be able to receive 
> incoming packets but will be unable to transmit any since the packets 
> will be sent of via wm0 (again because of the way the routing table is 
> setup). This is by design. Normally it is not recommended to configure 
> two interfaces on the network. You could either use a separate instance 
> of io-pkt for each interface (which gives each interface its own routing 
> table) or you might want to look into using a bridge.
> 
> On 2014-11-03 01:39 PM, Loris Colasante wrote:
> > Hi all,
> >
> > I have the same problem using a motherboard with 2 Intel 82574L ethernet 
> ports.
> > I've already downloaded and installed the last devnp-e1000.so driver release
>  (found on the Experimentals linked page) but it doesn't solve the problem.
> > Attached to this post you'll find a ZIP archive containing some LOGs, 
> obtained from QNX 6.5.0.
> > More in details, the ZIP archive contains the following files:
> > - pci.txt
> > pci device list (both ethernet are listed)
> >
> > - ifconfig.txt
> > "ifconfig" command's output
> >
> > - net.txt
> > "/etc/system/enum/include/net" file contents
> >
> > - routeShow.txt
> > "route show" command's output  (you can see that the address "10.3.2.200" 
> appears associated to the  wm0 interface (see the MAC Address in ifconfig for 
> confirmation), while the address "10.3.2.201" - related to the wm1 interface, 
> is not listed in the routing table
> >
> > - pingWM0.txt
> > "ping 10.3.2.200" command's output
> >
> > - pingWM1.txt
> > "ping 10.3.2.201" command's output
> >
> > Is there a solution for this problem? What can be related to?
> >
> >
> >
> > _______________________________________________
> >
> > Networking Drivers
> > http://community.qnx.com/sf/go/post112240
> > To cancel your subscription to this discussion, please e-mail drivers-
> networking-unsubscribe@community.qnx.com
> 


Re: intel 82754 network driver problem  
Not only do you need to identify each instance of io-pkt via -iX you 
also need to specify an alternative prefix for each additional instance 
of io-pkt otherwise it will overlap the first one causing some problems.

Try:

io-pkt-v4-hc -i1 -ptcpip
io-pkt-v4-hc -i2 -ptcpip prefix=/alt1
io-pkt-v4-hc -i3 -ptcpip prefix=/alt2

Now when you want to access the default instance:

ifconfig

For the second instance you need to use the SOCK environment variable to 
tell the socket layer which instance you are "talking" to:

SOCK=/alt1 ifconfig

For the third instance:

SOCK=/alt2 ifconfig


Basically you need to use SOCK to tell any application that uses the 
socket layer which tcpip stack to connect with (except for the default 
one which in my example was the first one).

On 2014-11-04 09:26 AM, Loris Colasante wrote:
> First of all thank you for your quick reply. Now it's more clear what this behaviour is related to.
> Anyway, because I need to use both interfaces, I'm trying to start 2 different instances of io-pkt but with no success
.
> I've modified the "etc/system/enum/include/net" file as reported below:
>
> -------------------------------------------------------------------------------------------
> #
> # macro definitions for network
> #
> 	
> all
> 	set(IOPKT_CMD, io-pkt-v4-hc -i1 unit=0 -ptcpip)
> 	set(IOPKT_CMD, io-pkt-v4-hc -i2 unit=1 -ptcpip)
>
> -------------------------------------------------------------------------------------------
>
> but in this way the following errors occurs at the startup:
>
> -------------------------------------------------------
> mount: Can't mount / (type io-pkt)
> mount: Possible reason: Invalid argument
> mount: Can't mount / (type io-pkt)
> mount: Possible reason: Invalid argument
> -------------------------------------------------------
>
> By specifying only the '-i1' and -i2' option (without 'unit=xx') the same error occurs.
>
> So ,which is the right way to create 2 separate instances of io-pkt, one for each interface 'wm0' and 'wm1', running 
at startup?
> Could you help me, please?
>
> Thank you in advance
>
>> The issue you are seeing is because you have configured two interfaces
>> on the same network within a single instance of io-pkt. Note that each
>> instance of io-pkt has one routing table and it will discard redundant
>> routes. In this case the route entry for wm1 is redundant since the
>> entry for wm0 already points to the network. This is why you see pings
>> working with wm0 but not wm1. In fact wm1 should be able to receive
>> incoming packets but will be unable to transmit any since the packets
>> will be sent of via wm0 (again because of the way the routing table is
>> setup). This is by design. Normally it is not recommended to configure
>> two interfaces on the network. You could either use a separate instance
>> of io-pkt for each interface (which gives each interface its own routing
>> table) or you might want to look into using a bridge.
>>
>> On 2014-11-03 01:39 PM, Loris Colasante wrote:
>>> Hi all,
>>>
>>> I have the same problem using a motherboard with 2 Intel 82574L ethernet
>> ports.
>>> I've already downloaded and installed the last devnp-e1000.so driver release
>>   (found on the Experimentals linked page) but it doesn't solve the problem.
>>> Attached to this post you'll find a ZIP archive containing some LOGs,
>> obtained from QNX 6.5.0.
>>> More in details, the ZIP archive contains the following files:
>>> - pci.txt
>>> pci device list (both ethernet are listed)
>>>
>>> - ifconfig.txt
>>> "ifconfig" command's output
>>>
>>> - net.txt
>>>...
View Full Message
Re: intel 82754 network driver problem  
I'm sorry, but there are still some problem:
Attached to this post you'll find my "/etc/system/enum/include/net" file.

Using this file, the following error occurs at startup:
----------------------------------------------------------------------
mount: Can't mount / (type io-pkt)
mount: Possible reason: Invalid argument
mount: Can't mount / (type io-pkt)
mount: Possible reason: Invalid argument
netmanager error: Could not open socket: Address family not supported by protocol family
netmanager error: Could not open socket: Address family not supported by protocol family
error creating socket
bad addr or host: <NULL> (No address associated with hostname)
Memory fault
----------------------------------------------------------------------

So, the first thing should be to solve this problem, then I'd like to have a way that allow both interfaces  -wm0 and 
wm1 - to be running without using any specific variables (like suggested SOCK).

Is there a way to do this? Is there a way to "instruct" QNX to add both the wm0 and wm1 interfaces to the routing table 
and give them both reacheable (and so, 'pingable')?
In the past, I've tried to install Windows 7 on the same machine and  both interfaces was working fine, without doing 
nothing special...
I'd like to add some instructions (somewhere in QNX files...) ONLY ONCE in order to have both the interfaces running in 
the future, without adding specific code to the applications running on the machine: is that possible?

Thank you.

> Not only do you need to identify each instance of io-pkt via -iX you 
> also need to specify an alternative prefix for each additional instance 
> of io-pkt otherwise it will overlap the first one causing some problems.
> 
> Try:
> 
> io-pkt-v4-hc -i1 -ptcpip
> io-pkt-v4-hc -i2 -ptcpip prefix=/alt1
> io-pkt-v4-hc -i3 -ptcpip prefix=/alt2
> 
> Now when you want to access the default instance:
> 
> ifconfig
> 
> For the second instance you need to use the SOCK environment variable to 
> tell the socket layer which instance you are "talking" to:
> 
> SOCK=/alt1 ifconfig
> 
> For the third instance:
> 
> SOCK=/alt2 ifconfig
> 
> 
> Basically you need to use SOCK to tell any application that uses the 
> socket layer which tcpip stack to connect with (except for the default 
> one which in my example was the first one).
> 
> On 2014-11-04 09:26 AM, Loris Colasante wrote:
> > First of all thank you for your quick reply. Now it's more clear what this 
> behaviour is related to.
> > Anyway, because I need to use both interfaces, I'm trying to start 2 
> different instances of io-pkt but with no success.
> > I've modified the "etc/system/enum/include/net" file as reported below:
> >
> > ----------------------------------------------------------------------------
> ---------------
> > #
> > # macro definitions for network
> > #
> > 	
> > all
> > 	set(IOPKT_CMD, io-pkt-v4-hc -i1 unit=0 -ptcpip)
> > 	set(IOPKT_CMD, io-pkt-v4-hc -i2 unit=1 -ptcpip)
> >
> > ----------------------------------------------------------------------------
> ---------------
> >
> > but in this way the following errors occurs at the startup:
> >
> > -------------------------------------------------------
> > mount: Can't mount / (type io-pkt)
> > mount: Possible reason: Invalid argument
> > mount: Can't mount / (type io-pkt)
> > mount: Possible reason: Invalid argument
> > -------------------------------------------------------
> >
> > By specifying only the '-i1' and -i2' option (without 'unit=xx') the same 
> error occurs.
> >
> > So ,which is the right way to create 2 separate instances of io-pkt, one for
>  each interface 'wm0' and 'wm1', running...
View Full Message
Attachment: Text net 195 bytes
Re: intel 82754 network driver problem  
-i instance
     The stack instance number, which is useful if you're running 
multiple instances of io-pkt. The io-pkt manager will service mount 
requests of type io-pktX, where X is the instance number. For example:

     io-pkt-v4 -i1 -ptcpip prefix=/alt
     mount -Tio-pkt1 /lib/dll/devnp-i82544.so


This is taken from the io-pkt documentation we have online. You need to 
tell 'mount' which instance is going to load the driver.

On 2014-11-04 10:05 AM, Loris Colasante wrote:
> I'm sorry, but there are still some problem:
> Attached to this post you'll find my "/etc/system/enum/include/net" file.
>
> Using this file, the following error occurs at startup:
> ----------------------------------------------------------------------
> mount: Can't mount / (type io-pkt)
> mount: Possible reason: Invalid argument
> mount: Can't mount / (type io-pkt)
> mount: Possible reason: Invalid argument
> netmanager error: Could not open socket: Address family not supported by protocol family
> netmanager error: Could not open socket: Address family not supported by protocol family
> error creating socket
> bad addr or host: <NULL> (No address associated with hostname)
> Memory fault
> ----------------------------------------------------------------------
>
> So, the first thing should be to solve this problem, then I'd like to have a way that allow both interfaces  -wm0 and 
wm1 - to be running without using any specific variables (like suggested SOCK).
>
> Is there a way to do this? Is there a way to "instruct" QNX to add both the wm0 and wm1 interfaces to the routing 
table and give them both reacheable (and so, 'pingable')?
> In the past, I've tried to install Windows 7 on the same machine and  both interfaces was working fine, without doing 
nothing special...
> I'd like to add some instructions (somewhere in QNX files...) ONLY ONCE in order to have both the interfaces running 
in the future, without adding specific code to the applications running on the machine: is that possible?
>
> Thank you.
>
>> Not only do you need to identify each instance of io-pkt via -iX you
>> also need to specify an alternative prefix for each additional instance
>> of io-pkt otherwise it will overlap the first one causing some problems.
>>
>> Try:
>>
>> io-pkt-v4-hc -i1 -ptcpip
>> io-pkt-v4-hc -i2 -ptcpip prefix=/alt1
>> io-pkt-v4-hc -i3 -ptcpip prefix=/alt2
>>
>> Now when you want to access the default instance:
>>
>> ifconfig
>>
>> For the second instance you need to use the SOCK environment variable to
>> tell the socket layer which instance you are "talking" to:
>>
>> SOCK=/alt1 ifconfig
>>
>> For the third instance:
>>
>> SOCK=/alt2 ifconfig
>>
>>
>> Basically you need to use SOCK to tell any application that uses the
>> socket layer which tcpip stack to connect with (except for the default
>> one which in my example was the first one).
>>
>> On 2014-11-04 09:26 AM, Loris Colasante wrote:
>>> First of all thank you for your quick reply. Now it's more clear what this
>> behaviour is related to.
>>> Anyway, because I need to use both interfaces, I'm trying to start 2
>> different instances of io-pkt but with no success.
>>> I've modified the "etc/system/enum/include/net" file as reported below:
>>>
>>> ----------------------------------------------------------------------------
>> ---------------
>>> #
>>> # macro definitions for network
>>> #
>>> 	
>>> all
>>> 	set(IOPKT_CMD, io-pkt-v4-hc -i1 unit=0 -ptcpip)
>>> 	set(IOPKT_CMD, io-pkt-v4-hc -i2 unit=1 -ptcpip)
>>>
>>>...
View Full Message
Re: intel 82754 network driver problem  
I've already read the io-pkt online documentation but, as I told you before, using the "-i" option it seems to cause an 
error mounting io-pkt (please refers to my previous post)

Also, I'm not sure about the right syntax to use inside the 'net' file.
Please, could you send me back a 'net' file with the full right commands? It could be more helpful for me in order to 
better understand the commands I have to use and the right syntax.

Thank you in advance (and sorry but I'm not a QNX expert, as you can imagine...)

> -i instance
>      The stack instance number, which is useful if you're running 
> multiple instances of io-pkt. The io-pkt manager will service mount 
> requests of type io-pktX, where X is the instance number. For example:
> 
>      io-pkt-v4 -i1 -ptcpip prefix=/alt
>      mount -Tio-pkt1 /lib/dll/devnp-i82544.so
> 
> 
> This is taken from the io-pkt documentation we have online. You need to 
> tell 'mount' which instance is going to load the driver.
> 
> On 2014-11-04 10:05 AM, Loris Colasante wrote:
> > I'm sorry, but there are still some problem:
> > Attached to this post you'll find my "/etc/system/enum/include/net" file.
> >
> > Using this file, the following error occurs at startup:
> > ----------------------------------------------------------------------
> > mount: Can't mount / (type io-pkt)
> > mount: Possible reason: Invalid argument
> > mount: Can't mount / (type io-pkt)
> > mount: Possible reason: Invalid argument
> > netmanager error: Could not open socket: Address family not supported by 
> protocol family
> > netmanager error: Could not open socket: Address family not supported by 
> protocol family
> > error creating socket
> > bad addr or host: <NULL> (No address associated with hostname)
> > Memory fault
> > ----------------------------------------------------------------------
> >
> > So, the first thing should be to solve this problem, then I'd like to have a
>  way that allow both interfaces  -wm0 and wm1 - to be running without using 
> any specific variables (like suggested SOCK).
> >
> > Is there a way to do this? Is there a way to "instruct" QNX to add both the 
> wm0 and wm1 interfaces to the routing table and give them both reacheable (and
>  so, 'pingable')?
> > In the past, I've tried to install Windows 7 on the same machine and  both 
> interfaces was working fine, without doing nothing special...
> > I'd like to add some instructions (somewhere in QNX files...) ONLY ONCE in 
> order to have both the interfaces running in the future, without adding 
> specific code to the applications running on the machine: is that possible?
> >
> > Thank you.
> >
> >> Not only do you need to identify each instance of io-pkt via -iX you
> >> also need to specify an alternative prefix for each additional instance
> >> of io-pkt otherwise it will overlap the first one causing some problems.
> >>
> >> Try:
> >>
> >> io-pkt-v4-hc -i1 -ptcpip
> >> io-pkt-v4-hc -i2 -ptcpip prefix=/alt1
> >> io-pkt-v4-hc -i3 -ptcpip prefix=/alt2
> >>
> >> Now when you want to access the default instance:
> >>
> >> ifconfig
> >>
> >> For the second instance you need to use the SOCK environment variable to
> >> tell the socket layer which instance you are "talking" to:
> >>
> >> SOCK=/alt1 ifconfig
> >>
> >> For the third instance:
> >>
> >> SOCK=/alt2 ifconfig
> >>
> >>
> >> Basically you need to use SOCK to tell any application that uses the
> >> socket layer which tcpip stack to connect with (except for...
View Full Message
Re: intel 82754 network driver problem  
# mount -t io-pkt1

The above will mount into the instance that was passed -i1

On Tue, Nov 04, 2014 at 10:23:53AM -0500, Loris Colasante wrote:
> I've already read the io-pkt online documentation but, as I told you before, using the "-i" option it seems to cause 
an error mounting io-pkt (please refers to my previous post)
> 
> Also, I'm not sure about the right syntax to use inside the 'net' file.
> Please, could you send me back a 'net' file with the full right commands? It could be more helpful for me in order to 
better understand the commands I have to use and the right syntax.
> 
> Thank you in advance (and sorry but I'm not a QNX expert, as you can imagine...)
> 
> > -i instance
> >      The stack instance number, which is useful if you're running 
> > multiple instances of io-pkt. The io-pkt manager will service mount 
> > requests of type io-pktX, where X is the instance number. For example:
> > 
> >      io-pkt-v4 -i1 -ptcpip prefix=/alt
> >      mount -Tio-pkt1 /lib/dll/devnp-i82544.so
> > 
> > 
> > This is taken from the io-pkt documentation we have online. You need to 
> > tell 'mount' which instance is going to load the driver.
> > 
> > On 2014-11-04 10:05 AM, Loris Colasante wrote:
> > > I'm sorry, but there are still some problem:
> > > Attached to this post you'll find my "/etc/system/enum/include/net" file.
> > >
> > > Using this file, the following error occurs at startup:
> > > ----------------------------------------------------------------------
> > > mount: Can't mount / (type io-pkt)
> > > mount: Possible reason: Invalid argument
> > > mount: Can't mount / (type io-pkt)
> > > mount: Possible reason: Invalid argument
> > > netmanager error: Could not open socket: Address family not supported by 
> > protocol family
> > > netmanager error: Could not open socket: Address family not supported by 
> > protocol family
> > > error creating socket
> > > bad addr or host: <NULL> (No address associated with hostname)
> > > Memory fault
> > > ----------------------------------------------------------------------
> > >
> > > So, the first thing should be to solve this problem, then I'd like to have a
> >  way that allow both interfaces  -wm0 and wm1 - to be running without using 
> > any specific variables (like suggested SOCK).
> > >
> > > Is there a way to do this? Is there a way to "instruct" QNX to add both the 
> > wm0 and wm1 interfaces to the routing table and give them both reacheable (and
> >  so, 'pingable')?
> > > In the past, I've tried to install Windows 7 on the same machine and  both 
> > interfaces was working fine, without doing nothing special...
> > > I'd like to add some instructions (somewhere in QNX files...) ONLY ONCE in 
> > order to have both the interfaces running in the future, without adding 
> > specific code to the applications running on the machine: is that possible?
> > >
> > > Thank you.
> > >
> > >> Not only do you need to identify each instance of io-pkt via -iX you
> > >> also need to specify an alternative prefix for each additional instance
> > >> of io-pkt otherwise it will overlap the first one causing some problems.
> > >>
> > >> Try:
> > >>
> > >> io-pkt-v4-hc -i1 -ptcpip
> > >> io-pkt-v4-hc -i2 -ptcpip prefix=/alt1
> > >> io-pkt-v4-hc -i3 -ptcpip prefix=/alt2
> > >>
> > >> Now when you want to access the default instance:
> > >>
> > >> ifconfig
> > >>
> >...
View Full Message
Re: intel 82754 network driver problem  
Well... maybe I'm a bit stupid (or I have a really bad english...) but:

1. Trying to write the "net" file adding what suggested by Gervais Mulongoy (the 'net' file is attached), at the startup
 a non-finite list of "Expected token at net(8)" appears (now I'm using a live-CD to recover the file in order to make 
QNX runing again)

2. Trying what suggested by Sean Boudreau directly at command prompt, a different error occurs (I'm sorry but at this 
moment I don't have the exact description)

In conclusion, by considering that I'm not an expert, if you really want to help me, please give me all the instructions
, step-by-step, or send me a 'net' file containing the right commands.

Otherwise, I think I have no possibility to solve the issue...
Again... thank you in advance
> 
> # mount -t io-pkt1
> 
> The above will mount into the instance that was passed -i1
> 
> On Tue, Nov 04, 2014 at 10:23:53AM -0500, Loris Colasante wrote:
> > I've already read the io-pkt online documentation but, as I told you before,
>  using the "-i" option it seems to cause an error mounting io-pkt (please 
> refers to my previous post)
> > 
> > Also, I'm not sure about the right syntax to use inside the 'net' file.
> > Please, could you send me back a 'net' file with the full right commands? It
>  could be more helpful for me in order to better understand the commands I 
> have to use and the right syntax.
> > 
> > Thank you in advance (and sorry but I'm not a QNX expert, as you can imagine
> ...)
> > 
> > > -i instance
> > >      The stack instance number, which is useful if you're running 
> > > multiple instances of io-pkt. The io-pkt manager will service mount 
> > > requests of type io-pktX, where X is the instance number. For example:
> > > 
> > >      io-pkt-v4 -i1 -ptcpip prefix=/alt
> > >      mount -Tio-pkt1 /lib/dll/devnp-i82544.so
> > > 
> > > 
> > > This is taken from the io-pkt documentation we have online. You need to 
> > > tell 'mount' which instance is going to load the driver.
> > > 
> > > On 2014-11-04 10:05 AM, Loris Colasante wrote:
> > > > I'm sorry, but there are still some problem:
> > > > Attached to this post you'll find my "/etc/system/enum/include/net" file
> .
> > > >
> > > > Using this file, the following error occurs at startup:
> > > > ----------------------------------------------------------------------
> > > > mount: Can't mount / (type io-pkt)
> > > > mount: Possible reason: Invalid argument
> > > > mount: Can't mount / (type io-pkt)
> > > > mount: Possible reason: Invalid argument
> > > > netmanager error: Could not open socket: Address family not supported by
>  
> > > protocol family
> > > > netmanager error: Could not open socket: Address family not supported by
>  
> > > protocol family
> > > > error creating socket
> > > > bad addr or host: <NULL> (No address associated with hostname)
> > > > Memory fault
> > > > ----------------------------------------------------------------------
> > > >
> > > > So, the first thing should be to solve this problem, then I'd like to 
> have a
> > >  way that allow both interfaces  -wm0 and wm1 - to be running without 
> using 
> > > any specific variables (like suggested SOCK).
> > > >
> > > > Is there a way to do this? Is there a way to "instruct" QNX to add both 
> the 
> > > wm0 and wm1 interfaces to the routing table and give them both reacheable 
> (and
> > >  so, 'pingable')?
> > > > In the past,...
View Full Message
Re: intel 82754 network driver problem  
Try this from the command line:

1) slay -f io-pkt-v4
2) io-pkt-v4 -i1 -ptcpip prefix=/alt1
3) mount -Tio-pkt1 -opci=0 /lib/dll/devnp-e1000.so
4) SOCK=/alt1 ifconfig
5) io-pkt-v4 -i2 -ptcpip prefix=/alt2
6) mount -Tio-pkt2 -opci=1 /lib/dll/devnp-e1000.so
7) SOCK=/alt2 ifconfig


The ifconfig utility called in step 4 should list lo0 and wm0, the one 
called in step 7 should also list lo0 and wm0 but compare the MAC 
addresses which should be different. Now take those commands and add 
them to your /etc/rc.d/rc.local and remove the entries from your 'net' 
file since you are now manually starting the networking stack instead of 
leaving it to the enumerator which doesn't understand SOCK (it wasn't 
designed for those types of use cases).


On 2014-11-04 10:47 AM, Loris Colasante wrote:
> > > >      io-pkt-v4 -i1 -ptcpip prefix=/alt
> > > >      mount -Tio-pkt1 /lib/dll/devnp-i82544.so
> > > >

Re: intel 82754 network driver problem  
Well, thank you for the informations but unfortunately also using these commands the 2nd ethernet doesn't work.

I've added all the commands to the /etc/rc.d/rc.local" file (except the slay command), then:

1. If I simply comment all the lines inside the 'net' file, a lot of errors occurs at the startup
2. If I rename the 'net' file to 'net.old', there are no errors at the startup but trying to ping the wm0 interface (10.
3.2.200) it works fine, while truing to ping the wm1 interface (10.3.2.201) I receive an "Host is down" message.

I don't know... maybe there is still something wrong but it seems to be 'impossible' to have both interfaces working... 
it seems there is no way to have 'wm1' working...

> Try this from the command line:
> 
> 1) slay -f io-pkt-v4
> 2) io-pkt-v4 -i1 -ptcpip prefix=/alt1
> 3) mount -Tio-pkt1 -opci=0 /lib/dll/devnp-e1000.so
> 4) SOCK=/alt1 ifconfig
> 5) io-pkt-v4 -i2 -ptcpip prefix=/alt2
> 6) mount -Tio-pkt2 -opci=1 /lib/dll/devnp-e1000.so
> 7) SOCK=/alt2 ifconfig
> 
> 
> The ifconfig utility called in step 4 should list lo0 and wm0, the one 
> called in step 7 should also list lo0 and wm0 but compare the MAC 
> addresses which should be different. Now take those commands and add 
> them to your /etc/rc.d/rc.local and remove the entries from your 'net' 
> file since you are now manually starting the networking stack instead of 
> leaving it to the enumerator which doesn't understand SOCK (it wasn't 
> designed for those types of use cases).
> 
> 
> On 2014-11-04 10:47 AM, Loris Colasante wrote:
> > > > >      io-pkt-v4 -i1 -ptcpip prefix=/alt
> > > > >      mount -Tio-pkt1 /lib/dll/devnp-i82544.so
> > > > >
> 


Re: intel 82754 network driver problem  
It isn't impossible as I have done this myself. You just need to 
troubleshoot your setup - try doing this step by step making sure that 
you have confirmed behavior before proceeding. First check the output of 
pci -v to make sure that the NICs have distinct "PCI Index" values. Then 
try each instance on its own - so one at a time. Configure the interface 
and make sure it is pingable. Only once you have confirm that both 
instances work should you proceed to testing both at the same time.

On 2014-11-04 12:07 PM, Loris Colasante wrote:
> Well, thank you for the informations but unfortunately also using these commands the 2nd ethernet doesn't work.
>
> I've added all the commands to the /etc/rc.d/rc.local" file (except the slay command), then:
>
> 1. If I simply comment all the lines inside the 'net' file, a lot of errors occurs at the startup
> 2. If I rename the 'net' file to 'net.old', there are no errors at the startup but trying to ping the wm0 interface 
(10.3.2.200) it works fine, while truing to ping the wm1 interface (10.3.2.201) I receive an "Host is down" message.
>
> I don't know... maybe there is still something wrong but it seems to be 'impossible' to have both interfaces working..
. it seems there is no way to have 'wm1' working...
>
>> Try this from the command line:
>>
>> 1) slay -f io-pkt-v4
>> 2) io-pkt-v4 -i1 -ptcpip prefix=/alt1
>> 3) mount -Tio-pkt1 -opci=0 /lib/dll/devnp-e1000.so
>> 4) SOCK=/alt1 ifconfig
>> 5) io-pkt-v4 -i2 -ptcpip prefix=/alt2
>> 6) mount -Tio-pkt2 -opci=1 /lib/dll/devnp-e1000.so
>> 7) SOCK=/alt2 ifconfig
>>
>>
>> The ifconfig utility called in step 4 should list lo0 and wm0, the one
>> called in step 7 should also list lo0 and wm0 but compare the MAC
>> addresses which should be different. Now take those commands and add
>> them to your /etc/rc.d/rc.local and remove the entries from your 'net'
>> file since you are now manually starting the networking stack instead of
>> leaving it to the enumerator which doesn't understand SOCK (it wasn't
>> designed for those types of use cases).
>>
>>
>> On 2014-11-04 10:47 AM, Loris Colasante wrote:
>>>>>>       io-pkt-v4 -i1 -ptcpip prefix=/alt
>>>>>>       mount -Tio-pkt1 /lib/dll/devnp-i82544.so
>>>>>>
>
>
>
>
>
> _______________________________________________
>
> Networking Drivers
> http://community.qnx.com/sf/go/post112253
> To cancel your subscription to this discussion, please e-mail drivers-networking-unsubscribe@community.qnx.com

Re: intel 82754 network driver problem  
Note the interface will be named 'wm0' in each instance.  Not
sure if you're confused looking explicitly for 'wm1'.

On Tue, Nov 04, 2014 at 02:12:55PM -0500, Gervais Mulongoy wrote:
> It isn't impossible as I have done this myself. You just need to 
> troubleshoot your setup - try doing this step by step making sure that 
> you have confirmed behavior before proceeding. First check the output of 
> pci -v to make sure that the NICs have distinct "PCI Index" values. Then 
> try each instance on its own - so one at a time. Configure the interface 
> and make sure it is pingable. Only once you have confirm that both 
> instances work should you proceed to testing both at the same time.
> 
> On 2014-11-04 12:07 PM, Loris Colasante wrote:
> > Well, thank you for the informations but unfortunately also using these commands the 2nd ethernet doesn't work.
> >
> > I've added all the commands to the /etc/rc.d/rc.local" file (except the slay command), then:
> >
> > 1. If I simply comment all the lines inside the 'net' file, a lot of errors occurs at the startup
> > 2. If I rename the 'net' file to 'net.old', there are no errors at the startup but trying to ping the wm0 interface 
(10.3.2.200) it works fine, while truing to ping the wm1 interface (10.3.2.201) I receive an "Host is down" message.
> >
> > I don't know... maybe there is still something wrong but it seems to be 'impossible' to have both interfaces working
... it seems there is no way to have 'wm1' working...
> >
> >> Try this from the command line:
> >>
> >> 1) slay -f io-pkt-v4
> >> 2) io-pkt-v4 -i1 -ptcpip prefix=/alt1
> >> 3) mount -Tio-pkt1 -opci=0 /lib/dll/devnp-e1000.so
> >> 4) SOCK=/alt1 ifconfig
> >> 5) io-pkt-v4 -i2 -ptcpip prefix=/alt2
> >> 6) mount -Tio-pkt2 -opci=1 /lib/dll/devnp-e1000.so
> >> 7) SOCK=/alt2 ifconfig
> >>
> >>
> >> The ifconfig utility called in step 4 should list lo0 and wm0, the one
> >> called in step 7 should also list lo0 and wm0 but compare the MAC
> >> addresses which should be different. Now take those commands and add
> >> them to your /etc/rc.d/rc.local and remove the entries from your 'net'
> >> file since you are now manually starting the networking stack instead of
> >> leaving it to the enumerator which doesn't understand SOCK (it wasn't
> >> designed for those types of use cases).
> >>
> >>
> >> On 2014-11-04 10:47 AM, Loris Colasante wrote:
> >>>>>>       io-pkt-v4 -i1 -ptcpip prefix=/alt
> >>>>>>       mount -Tio-pkt1 /lib/dll/devnp-i82544.so
> >>>>>>
> >
> >
> >
> >
> >
> > _______________________________________________
> >
> > Networking Drivers
> > http://community.qnx.com/sf/go/post112253
> > To cancel your subscription to this discussion, please e-mail drivers-networking-unsubscribe@community.qnx.com
> 
> 
> 
> 
> 
> _______________________________________________
> 
> Networking Drivers
> http://community.qnx.com/sf/go/post112257
> To cancel your subscription to this discussion, please e-mail drivers-networking-unsubscribe@community.qnx.com
Re: intel 82754 network driver problem  
I suspect that the problem here is the same as in topc25682:

The PCI driver does not reserve and allocate memory correctly.  Please look at that topic as well as topc25206.

The problem is that the scan_windows() function in server.c adds up memory based on *sane* memory allocation methods but
 pci_allocate_resources() and the E1000 driver force all PCI memory windows to be aligned on 1 MB boundaries.

You will end up running out of PCI memory with additional Gigabit ethernet parts.

You should look in your sloginfo print out to be sure that you do not see an error from addup_resources() in the PCI 
system.