Error from Network driver

QNX Newsgroups (Archives) qnx.qnx4
1

I found quite a lot of these errors in the tracefile, coming from the
network driver.

It is a bit strange that multiple errors complaing “timeout” have the
same timestamp.

This node is attempting to communicate to another node where we have
special hardware drivers running at very high priority (29).

Any idea about what is happening ?

(QNX4 - Latest version and patches)

Jun 03 10:50:03.848 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:03.848 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:03.848 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:03.848 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:03.848 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)

Jun 03 10:50:11.147 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:11.147 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:11.147 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:11.147 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:11.147 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)

Jun 03 10:50:14.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:14.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:14.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:14.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:14.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)

Jun 03 10:50:16.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:16.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:16.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:16.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:16.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)

Jun 03 10:50:18.247 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:18.247 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:18.247 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:18.247 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:18.247 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)

Thanks in advance,
Enrico

2

It means that a transmit timed out waiting for an ACK or a NACK from the
receiving station. The reason you see 5 messages with the same time stamp is
that the driver outputs 5 messages for each timeout. If you look at the
output from netinfo you will see the differences.

“Enrico Bendinelli” <enricob@prisma-eng.it> wrote in message
news:ca4ceb$hsc$1@inn.qnx.com…

I found quite a lot of these errors in the tracefile, coming from the
network driver.

It is a bit strange that multiple errors complaing “timeout” have the
same timestamp.

This node is attempting to communicate to another node where we have
special hardware drivers running at very high priority (29).

Any idea about what is happening ?

(QNX4 - Latest version and patches)

Jun 03 10:50:03.848 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:03.848 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:03.848 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:03.848 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:03.848 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)

Jun 03 10:50:11.147 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:11.147 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:11.147 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:11.147 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:11.147 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)

Jun 03 10:50:14.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:14.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:14.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:14.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:14.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)

Jun 03 10:50:16.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:16.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:16.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:16.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:16.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)

Jun 03 10:50:18.247 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:18.247 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:18.247 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:18.247 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:18.247 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)

Thanks in advance,
Enrico

3

Our driver has higher priority than the network driver, maybe it delays
the activation of the network driver too much.

Is the timeout related to the fleet protocol ? How long is it ?

After these errors we note that the communication between the two nodes
looks to be sleeping for a while then it restarts. I would like to
minimize this sleeping period.

There are some command line options for the driver (Net.ether82557) that
set the number and the period of retries. Are they referring to the same
issue ?

Thanks in advance,
Enrico

Hugh Brown wrote:

It means that a transmit timed out waiting for an ACK or a NACK from the
receiving station. The reason you see 5 messages with the same time stamp is
that the driver outputs 5 messages for each timeout. If you look at the
output from netinfo you will see the differences.

“Enrico Bendinelli” <> enricob@prisma-eng.it> > wrote in message
news:ca4ceb$hsc$> 1@inn.qnx.com> …

I found quite a lot of these errors in the tracefile, coming from the
network driver.

It is a bit strange that multiple errors complaing “timeout” have the
same timestamp.

This node is attempting to communicate to another node where we have
special hardware drivers running at very high priority (29).

Any idea about what is happening ?

(QNX4 - Latest version and patches)

Jun 03 10:50:03.848 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:03.848 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:03.848 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:03.848 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:03.848 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)

Jun 03 10:50:11.147 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:11.147 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:11.147 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:11.147 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:11.147 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)

Jun 03 10:50:14.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:14.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:14.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:14.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:14.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)

Jun 03 10:50:16.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:16.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:16.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:16.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:16.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)

Jun 03 10:50:18.247 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:18.247 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:18.247 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:18.247 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:18.247 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)

Thanks in advance,
Enrico

\

4

You can either use the -n option to increase the number of ticks before the
message times out (the default is 3), or you can increase the transmit tick
timer with the -t option.
(the default is 20 ticks).

“Enrico Bendinelli” <enricob@prisma-eng.it> wrote in message
news:ca56m3$838$1@inn.qnx.com…

Our driver has higher priority than the network driver, maybe it delays
the activation of the network driver too much.

Is the timeout related to the fleet protocol ? How long is it ?

After these errors we note that the communication between the two nodes
looks to be sleeping for a while then it restarts. I would like to
minimize this sleeping period.

There are some command line options for the driver (Net.ether82557) that
set the number and the period of retries. Are they referring to the same
issue ?

Thanks in advance,
Enrico

Hugh Brown wrote:
It means that a transmit timed out waiting for an ACK or a NACK from the
receiving station. The reason you see 5 messages with the same time
stamp is
that the driver outputs 5 messages for each timeout. If you look at the
output from netinfo you will see the differences.

“Enrico Bendinelli” <> enricob@prisma-eng.it> > wrote in message
news:ca4ceb$hsc$> 1@inn.qnx.com> …

I found quite a lot of these errors in the tracefile, coming from the
network driver.

It is a bit strange that multiple errors complaing “timeout” have the
same timestamp.

This node is attempting to communicate to another node where we have
special hardware drivers running at very high priority (29).

Any idea about what is happening ?

(QNX4 - Latest version and patches)

Jun 03 10:50:03.848 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:03.848 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:03.848 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:03.848 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:03.848 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)

Jun 03 10:50:11.147 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:11.147 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:11.147 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:11.147 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:11.147 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)

Jun 03 10:50:14.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:14.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:14.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:14.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:14.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)

Jun 03 10:50:16.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:16.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:16.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:16.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:16.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)

Jun 03 10:50:18.247 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:18.247 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:18.247 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:18.247 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:18.247 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)

Thanks in advance,
Enrico

\

5

The output of “use Net.ether82557” shows a different description
concerning “-n” option, that is also aligned with the “Helpviewer”.

Could you please confirm that “-n” set the timeout period in terms of OS
ticks (e.g. 3 → 30 ms if OS tick is 10 ms as usual).

Net.ether82557
[-p IoBase] [-i IrqNum] [-m MAC_ADDR] [-S Port Speed] [-M]
[-n TxMaxRetries]
[-N TxMaxReplyRetries] [-P] [-r MediaRate] [-t TxRetryTicks]
[-s Speed]
[-F] [-l NetId] [-L NetManagerName] [-I PciIndex] [-V VendorId]
[-D DeviceId] [-v] [-?]
Where:
-p IoBase I/O Port Base Address
-i IrqNum Interrupt Number
-m MAC_ADDR Ethernet Physical Address
-S Port Speed Output to Serial Port (Debugging)
-M Multicast Mode Enabled
-n TxMaxRetries Max No. of retries after Tx Timeout
(Default : 3)
-N TxMaxReplyRetries
-P Promiscuous Mode Enabled
-r MediaRate Media Rate
-t TxRetryTicks No. of 50ms ticks between Tx retries
(Default : 20)
-s Speed Speed setting to be forced (10 / 100)
-F Full Duplex Mode Forced.
-l NetId Logical Network ID (Default : 1)
-L NetManagerName Default : “qnx/net”
-I PciIndex PCI Index to use
-V VendorId PCI Vendor Id to use (Default : 0x8086 - INTEL)
-D DeviceId PCI Device Id (Default : 0 - ignored)
-v Verbose Output

Thank you for your help,
Enrico

Hugh Brown wrote:

You can either use the -n option to increase the number of ticks before the
message times out (the default is 3), or you can increase the transmit tick
timer with the -t option.
(the default is 20 ticks).

“Enrico Bendinelli” <> enricob@prisma-eng.it> > wrote in message
news:ca56m3$838$> 1@inn.qnx.com> …

Our driver has higher priority than the network driver, maybe it delays
the activation of the network driver too much.

Is the timeout related to the fleet protocol ? How long is it ?

After these errors we note that the communication between the two nodes
looks to be sleeping for a while then it restarts. I would like to
minimize this sleeping period.

There are some command line options for the driver (Net.ether82557) that
set the number and the period of retries. Are they referring to the same
issue ?

Thanks in advance,
Enrico

Hugh Brown wrote:

It means that a transmit timed out waiting for an ACK or a NACK from the
receiving station. The reason you see 5 messages with the same time

stamp is

that the driver outputs 5 messages for each timeout. If you look at the
output from netinfo you will see the differences.

“Enrico Bendinelli” <> enricob@prisma-eng.it> > wrote in message
news:ca4ceb$hsc$> 1@inn.qnx.com> …


I found quite a lot of these errors in the tracefile, coming from the
network driver.

It is a bit strange that multiple errors complaing “timeout” have the
same timestamp.

This node is attempting to communicate to another node where we have
special hardware drivers running at very high priority (29).

Any idea about what is happening ?

(QNX4 - Latest version and patches)

Jun 03 10:50:03.848 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:03.848 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:03.848 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:03.848 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:03.848 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)

Jun 03 10:50:11.147 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:11.147 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:11.147 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:11.147 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:11.147 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)

Jun 03 10:50:14.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:14.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:14.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:14.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:14.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)

Jun 03 10:50:16.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:16.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:16.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:16.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:16.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)

Jun 03 10:50:18.247 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:18.247 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:18.247 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:18.247 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:18.247 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)

Thanks in advance,
Enrico



\

6

Sorry, my mistake! The -n is the TxMaxRetries, not the ticks. You can try
increasing this number or the TxRetryTicks.

“Enrico Bendinelli” <enricob@prisma-eng.it> wrote in message
news:ca70gq$km4$1@inn.qnx.com…

The output of “use Net.ether82557” shows a different description
concerning “-n” option, that is also aligned with the “Helpviewer”.

Could you please confirm that “-n” set the timeout period in terms of OS
ticks (e.g. 3 → 30 ms if OS tick is 10 ms as usual).

Net.ether82557
[-p IoBase] [-i IrqNum] [-m MAC_ADDR] [-S Port Speed] [-M]
[-n TxMaxRetries]
[-N TxMaxReplyRetries] [-P] [-r MediaRate] [-t TxRetryTicks]
[-s Speed]
[-F] [-l NetId] [-L NetManagerName] [-I PciIndex] [-V VendorId]
[-D DeviceId] [-v] [-?]
Where:
-p IoBase I/O Port Base Address
-i IrqNum Interrupt Number
-m MAC_ADDR Ethernet Physical Address
-S Port Speed Output to Serial Port (Debugging)
-M Multicast Mode Enabled
-n TxMaxRetries Max No. of retries after Tx Timeout
(Default : 3)
-N TxMaxReplyRetries
-P Promiscuous Mode Enabled
-r MediaRate Media Rate
-t TxRetryTicks No. of 50ms ticks between Tx retries
(Default : 20)
-s Speed Speed setting to be forced (10 / 100)
-F Full Duplex Mode Forced.
-l NetId Logical Network ID (Default : 1)
-L NetManagerName Default : “qnx/net”
-I PciIndex PCI Index to use
-V VendorId PCI Vendor Id to use (Default : 0x8086 - INTEL)
-D DeviceId PCI Device Id (Default : 0 - ignored)
-v Verbose Output

Thank you for your help,
Enrico

Hugh Brown wrote:

You can either use the -n option to increase the number of ticks before
the
message times out (the default is 3), or you can increase the transmit
tick
timer with the -t option.
(the default is 20 ticks).

“Enrico Bendinelli” <> enricob@prisma-eng.it> > wrote in message
news:ca56m3$838$> 1@inn.qnx.com> …

Our driver has higher priority than the network driver, maybe it delays
the activation of the network driver too much.

Is the timeout related to the fleet protocol ? How long is it ?

After these errors we note that the communication between the two nodes
looks to be sleeping for a while then it restarts. I would like to
minimize this sleeping period.

There are some command line options for the driver (Net.ether82557) that
set the number and the period of retries. Are they referring to the same
issue ?

Thanks in advance,
Enrico

Hugh Brown wrote:

It means that a transmit timed out waiting for an ACK or a NACK from
the
receiving station. The reason you see 5 messages with the same time

stamp is

that the driver outputs 5 messages for each timeout. If you look at the
output from netinfo you will see the differences.

“Enrico Bendinelli” <> enricob@prisma-eng.it> > wrote in message
news:ca4ceb$hsc$> 1@inn.qnx.com> …


I found quite a lot of these errors in the tracefile, coming from the
network driver.

It is a bit strange that multiple errors complaing “timeout” have the
same timestamp.

This node is attempting to communicate to another node where we have
special hardware drivers running at very high priority (29).

Any idea about what is happening ?

(QNX4 - Latest version and patches)

Jun 03 10:50:03.848 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:03.848 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:03.848 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:03.848 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:03.848 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)

Jun 03 10:50:11.147 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:11.147 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:11.147 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:11.147 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:11.147 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)

Jun 03 10:50:14.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:14.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:14.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:14.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:14.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)

Jun 03 10:50:16.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:16.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:16.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:16.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:16.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)

Jun 03 10:50:18.247 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:18.247 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:18.247 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:18.247 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:18.247 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)

Thanks in advance,
Enrico



\

7

Finally, it looks there is no way to increase the timeout that causes:

Jun 03 10:50:03.848 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)

Please could you please let me know the duration of this timeout ?

Thanks again,
Enrico

Hugh Brown wrote:

Sorry, my mistake! The -n is the TxMaxRetries, not the ticks. You can try
increasing this number or the TxRetryTicks.

“Enrico Bendinelli” <> enricob@prisma-eng.it> > wrote in message
news:ca70gq$km4$> 1@inn.qnx.com> …

The output of “use Net.ether82557” shows a different description
concerning “-n” option, that is also aligned with the “Helpviewer”.

Could you please confirm that “-n” set the timeout period in terms of OS
ticks (e.g. 3 → 30 ms if OS tick is 10 ms as usual).

Net.ether82557
[-p IoBase] [-i IrqNum] [-m MAC_ADDR] [-S Port Speed] [-M]
[-n TxMaxRetries]
[-N TxMaxReplyRetries] [-P] [-r MediaRate] [-t TxRetryTicks]
[-s Speed]
[-F] [-l NetId] [-L NetManagerName] [-I PciIndex] [-V VendorId]
[-D DeviceId] [-v] [-?]
Where:
-p IoBase I/O Port Base Address
-i IrqNum Interrupt Number
-m MAC_ADDR Ethernet Physical Address
-S Port Speed Output to Serial Port (Debugging)
-M Multicast Mode Enabled
-n TxMaxRetries Max No. of retries after Tx Timeout
(Default : 3)
-N TxMaxReplyRetries
-P Promiscuous Mode Enabled
-r MediaRate Media Rate
-t TxRetryTicks No. of 50ms ticks between Tx retries
(Default : 20)
-s Speed Speed setting to be forced (10 / 100)
-F Full Duplex Mode Forced.
-l NetId Logical Network ID (Default : 1)
-L NetManagerName Default : “qnx/net”
-I PciIndex PCI Index to use
-V VendorId PCI Vendor Id to use (Default : 0x8086 - INTEL)
-D DeviceId PCI Device Id (Default : 0 - ignored)
-v Verbose Output

Thank you for your help,
Enrico

Hugh Brown wrote:


You can either use the -n option to increase the number of ticks before

the

message times out (the default is 3), or you can increase the transmit

tick

timer with the -t option.
(the default is 20 ticks).

“Enrico Bendinelli” <> enricob@prisma-eng.it> > wrote in message
news:ca56m3$838$> 1@inn.qnx.com> …


Our driver has higher priority than the network driver, maybe it delays
the activation of the network driver too much.

Is the timeout related to the fleet protocol ? How long is it ?

After these errors we note that the communication between the two nodes
looks to be sleeping for a while then it restarts. I would like to
minimize this sleeping period.

There are some command line options for the driver (Net.ether82557) that
set the number and the period of retries. Are they referring to the same
issue ?

Thanks in advance,
Enrico

Hugh Brown wrote:


It means that a transmit timed out waiting for an ACK or a NACK from

the

receiving station. The reason you see 5 messages with the same time

stamp is


that the driver outputs 5 messages for each timeout. If you look at the
output from netinfo you will see the differences.

“Enrico Bendinelli” <> enricob@prisma-eng.it> > wrote in message
news:ca4ceb$hsc$> 1@inn.qnx.com> …



I found quite a lot of these errors in the tracefile, coming from the
network driver.

It is a bit strange that multiple errors complaing “timeout” have the
same timestamp.

This node is attempting to communicate to another node where we have
special hardware drivers running at very high priority (29).

Any idea about what is happening ?

(QNX4 - Latest version and patches)

Jun 03 10:50:03.848 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:03.848 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:03.848 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:03.848 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:03.848 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)

Jun 03 10:50:11.147 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:11.147 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:11.147 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:11.147 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:11.147 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)

Jun 03 10:50:14.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:14.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:14.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:14.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:14.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)

Jun 03 10:50:16.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:16.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:16.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:16.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:16.197 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)

Jun 03 10:50:18.247 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:18.247 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:18.247 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:18.247 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)
Jun 03 10:50:18.247 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)

Thanks in advance,
Enrico




\

8

“Enrico Bendinelli” <enricob@prisma-eng.it> wrote in message
news:ca7lh5$7e6$1@inn.qnx.com…

Finally, it looks there is no way to increase the timeout that causes:

Jun 03 10:50:03.848 2 00004903 2 (2307) 82557 ( tx) timeout (no nack)

Please could you please let me know the duration of this timeout ?

The timer tick is every 50ms and the tx_retry_ticks is set to 20, so this
would give a 1 second delay. If your application is doing a lot of work at a
higher priority, then Net might not get much time to run.


Thanks again,
Enrico

Hugh Brown wrote:

Sorry, my mistake! The -n is the TxMaxRetries, not the ticks. You can
try
increasing this number or the TxRetryTicks.

“Enrico Bendinelli” <> enricob@prisma-eng.it> > wrote in message
news:ca70gq$km4$> 1@inn.qnx.com> …

The output of “use Net.ether82557” shows a different description
concerning “-n” option, that is also aligned with the “Helpviewer”.

Could you please confirm that “-n” set the timeout period in terms of OS
ticks (e.g. 3 → 30 ms if OS tick is 10 ms as usual).

Net.ether82557
[-p IoBase] [-i IrqNum] [-m MAC_ADDR] [-S Port Speed] [-M]
[-n TxMaxRetries]
[-N TxMaxReplyRetries] [-P] [-r MediaRate] [-t TxRetryTicks]
[-s Speed]
[-F] [-l NetId] [-L NetManagerName] [-I PciIndex] [-V VendorId]
[-D DeviceId] [-v] [-?]
Where:
-p IoBase I/O Port Base Address
-i IrqNum Interrupt Number
-m MAC_ADDR Ethernet Physical Address
-S Port Speed Output to Serial Port (Debugging)
-M Multicast Mode Enabled
-n TxMaxRetries Max No. of retries after Tx Timeout
(Default : 3)
-N TxMaxReplyRetries
-P Promiscuous Mode Enabled
-r MediaRate Media Rate
-t TxRetryTicks No. of 50ms ticks between Tx retries
(Default : 20)
-s Speed Speed setting to be forced (10 / 100)
-F Full Duplex Mode Forced.
-l NetId Logical Network ID (Default : 1)
-L NetManagerName Default : “qnx/net”
-I PciIndex PCI Index to use
-V VendorId PCI Vendor Id to use (Default : 0x8086 - INTEL)
-D DeviceId PCI Device Id (Default : 0 - ignored)
-v Verbose Output

Thank you for your help,
Enrico

Hugh Brown wrote:


You can either use the -n option to increase the number of ticks before

the

message times out (the default is 3), or you can increase the transmit

tick

timer with the -t option.
(the default is 20 ticks).

“Enrico Bendinelli” <> enricob@prisma-eng.it> > wrote in message
news:ca56m3$838$> 1@inn.qnx.com> …


Our driver has higher priority than the network driver, maybe it
delays
the activation of the network driver too much.

Is the timeout related to the fleet protocol ? How long is it ?

After these errors we note that the communication between the two
nodes
looks to be sleeping for a while then it restarts. I would like to
minimize this sleeping period.

There are some command line options for the driver (Net.ether82557)
that
set the number and the period of retries. Are they referring to the
same
issue ?

Thanks in advance,
Enrico

Hugh Brown wrote:


It means that a transmit timed out waiting for an ACK or a NACK from

the

receiving station. The reason you see 5 messages with the same time

stamp is


that the driver outputs 5 messages for each timeout. If you look at
the
output from netinfo you will see the differences.

“Enrico Bendinelli” <> enricob@prisma-eng.it> > wrote in message
news:ca4ceb$hsc$> 1@inn.qnx.com> …



I found quite a lot of these errors in the tracefile, coming from
the
network driver.

It is a bit strange that multiple errors complaing “timeout” have
the
same timestamp.

This node is attempting to communicate to another node where we have
special hardware drivers running at very high priority (29).

Any idea about what is happening ?

(QNX4 - Latest version and patches)

Jun 03 10:50:03.848 2 00004903 2 (2307) 82557 ( tx) timeout (no
nack)
Jun 03 10:50:03.848 2 00004903 2 (2307) 82557 ( tx) timeout (no
nack)
Jun 03 10:50:03.848 2 00004903 2 (2307) 82557 ( tx) timeout (no
nack)
Jun 03 10:50:03.848 2 00004903 2 (2307) 82557 ( tx) timeout (no
nack)
Jun 03 10:50:03.848 2 00004903 2 (2307) 82557 ( tx) timeout (no
nack)

Jun 03 10:50:11.147 2 00004903 2 (2307) 82557 ( tx) timeout (no
nack)
Jun 03 10:50:11.147 2 00004903 2 (2307) 82557 ( tx) timeout (no
nack)
Jun 03 10:50:11.147 2 00004903 2 (2307) 82557 ( tx) timeout (no
nack)
Jun 03 10:50:11.147 2 00004903 2 (2307) 82557 ( tx) timeout (no
nack)
Jun 03 10:50:11.147 2 00004903 2 (2307) 82557 ( tx) timeout (no
nack)

Jun 03 10:50:14.197 2 00004903 2 (2307) 82557 ( tx) timeout (no
nack)
Jun 03 10:50:14.197 2 00004903 2 (2307) 82557 ( tx) timeout (no
nack)
Jun 03 10:50:14.197 2 00004903 2 (2307) 82557 ( tx) timeout (no
nack)
Jun 03 10:50:14.197 2 00004903 2 (2307) 82557 ( tx) timeout (no
nack)
Jun 03 10:50:14.197 2 00004903 2 (2307) 82557 ( tx) timeout (no
nack)

Jun 03 10:50:16.197 2 00004903 2 (2307) 82557 ( tx) timeout (no
nack)
Jun 03 10:50:16.197 2 00004903 2 (2307) 82557 ( tx) timeout (no
nack)
Jun 03 10:50:16.197 2 00004903 2 (2307) 82557 ( tx) timeout (no
nack)
Jun 03 10:50:16.197 2 00004903 2 (2307) 82557 ( tx) timeout (no
nack)
Jun 03 10:50:16.197 2 00004903 2 (2307) 82557 ( tx) timeout (no
nack)

Jun 03 10:50:18.247 2 00004903 2 (2307) 82557 ( tx) timeout (no
nack)
Jun 03 10:50:18.247 2 00004903 2 (2307) 82557 ( tx) timeout (no
nack)
Jun 03 10:50:18.247 2 00004903 2 (2307) 82557 ( tx) timeout (no
nack)
Jun 03 10:50:18.247 2 00004903 2 (2307) 82557 ( tx) timeout (no
nack)
Jun 03 10:50:18.247 2 00004903 2 (2307) 82557 ( tx) timeout (no
nack)

Thanks in advance,
Enrico




\