devn-speedo - not fixed (for us, anyway)

We have downloaded the new devn-speedo from developers.qnx.com but we are
still having problems with woeful network performance.

What we see is that (according to nicinfo) the driver/card negotiates a
half-duplex connection, even though the switch at the other end thinks it is
full-duplex. After this mis-negotiation has occurred, we get - not
surprisingly - about 1k/s throughput.

And yes, we have tried with two different brands of switch.

The only way we can get correct duplex-ness negotiated is to physically
remove the network cable from the back of the PC, then plug it back in, then
re-start io-net.

Just re-starting io-net (without first disconnecting the cable) doesn’t fix
it.

We have tried forcing full duplex using the command-line option duplex=1 but
this doesn’t seem to make any difference.

Is this the same problem as seen before, or a different one? Any
suggestions?

Rob Rutherford
Ruzz Technology

Robert Rutherford <ruzz@nospamplease.ruzz.com> wrote:

What we see is that (according to nicinfo) the driver/card negotiates a
half-duplex connection, even though the switch at the other end thinks it is
full-duplex. After this mis-negotiation has occurred, we get - not
surprisingly - about 1k/s throughput.

And yes, we have tried with two different brands of switch.

Have you tried giving all the options you want?

io-net -d speedo nomulticast,speed=100,duplex=1

chris

cdm@qnx.com > “The faster I go, the behinder I get.”

Chris McKillop – Lewis Carroll –
Software Engineer, QSSL
<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<

“Chris McKillop” <cdm@qnx.com> wrote in message
news:9tbdnv$bjq$2@nntp.qnx.com

Have you tried giving all the options you want?

io-net -d speedo nomulticast,speed=100,duplex=1

Yes. Below is a capture of exactly what we see. Any other ideas? Do you need
any other info?

slay io-net

io-net -d speedo nomulticast,speed=100,duplex=1 -p qnet -p tcpip

nicinfo

INTEL 82558 Ethernet Controller
Physical Node ID … 000347 881606
Current Physical Node ID … 000347 881606
Media Rate … 100.00 Mb/s half-duplex UTP
MTU … 1514
Lan … 0
I/O Port Range … 0xC400 → 0xC43F
Hardware Interrupt … 0xA
Promiscuous … Disabled
Multicast … Disabled

Total Packets Txd OK … 2
Total Packets Txd Bad … 0
Total Packets Rxd OK … 0
Total Rx Errors … 0

Total Bytes Txd … 162
Total Bytes Rxd … 0

Tx Collision Errors … 0
Tx Collisions Errors (aborted) … 0
Carrier Sense Lost on Tx … 0
FIFO Underruns During Tx … 0
Tx deferred … 0
Out of Window Collisions … 0
FIFO Overruns During Rx … 0
Alignment errors … 0
CRC errors … 0

ls -l /lib/dll/devn-speedo.so

-rwxrwxr-x 1 root bin 47698 Aug 10 10:19
/lib/dll/devn-speedo.so

uname -a

QNX neutrino3 6.1.0 2001/06/25-15:31:48edt x86pc x86

“Robert Rutherford” <ruzz@NoSpamPlease.ruzz.com> wrote in message
news:9t9lpm$gqc$1@inn.qnx.com

We have downloaded the new devn-speedo from developers.qnx.com but we are
still having problems with woeful network performance.

Is this the same problem as seen before, or a different one? Any
suggestions?

Have you tried specifing the nomulticast option with the new driver (from
developers.qnx.com). Perhaps the hardware group will have more insight into
this issue?

-Adam

There is a bug in setting duplex manually. and it is fixed recently. So you
need to get the latest binary and try “io-net -d speedo
nomulticast,speed=100,duplex=1”.

But I am not sure who is responsible for updating developers.qnx.com. Maybe
Chris knows.

lihong cao


“Robert Rutherford” <ruzz@NoSpamPlease.ruzz.com> wrote in message
news:9tc3k4$fu$1@inn.qnx.com

“Chris McKillop” <> cdm@qnx.com> > wrote in message
news:9tbdnv$bjq$> 2@nntp.qnx.com> …

Have you tried giving all the options you want?

io-net -d speedo nomulticast,speed=100,duplex=1

Yes. Below is a capture of exactly what we see. Any other ideas? Do you
need
any other info?

slay io-net

io-net -d speedo nomulticast,speed=100,duplex=1 -p qnet -p tcpip

nicinfo

INTEL 82558 Ethernet Controller
Physical Node ID … 000347 881606
Current Physical Node ID … 000347 881606
Media Rate … 100.00 Mb/s half-duplex UTP
MTU … 1514
Lan … 0
I/O Port Range … 0xC400 → 0xC43F
Hardware Interrupt … 0xA
Promiscuous … Disabled
Multicast … Disabled

Total Packets Txd OK … 2
Total Packets Txd Bad … 0
Total Packets Rxd OK … 0
Total Rx Errors … 0

Total Bytes Txd … 162
Total Bytes Rxd … 0

Tx Collision Errors … 0
Tx Collisions Errors (aborted) … 0
Carrier Sense Lost on Tx … 0
FIFO Underruns During Tx … 0
Tx deferred … 0
Out of Window Collisions … 0
FIFO Overruns During Rx … 0
Alignment errors … 0
CRC errors … 0

ls -l /lib/dll/devn-speedo.so

-rwxrwxr-x 1 root bin 47698 Aug 10 10:19
/lib/dll/devn-speedo.so

uname -a

QNX neutrino3 6.1.0 2001/06/25-15:31:48edt x86pc x86

And, in fact, once I get developers.qnx.com updated you don’t need to give
any options. You may still need speed/duplex but you will not need to use
the nomulticast.

chris


lihong cao <lcao@qnx.com> wrote:

There is a bug in setting duplex manually. and it is fixed recently. So you
need to get the latest binary and try “io-net -d speedo
nomulticast,speed=100,duplex=1”.

But I am not sure who is responsible for updating developers.qnx.com. Maybe
Chris knows.

lihong cao


“Robert Rutherford” <> ruzz@NoSpamPlease.ruzz.com> > wrote in message
news:9tc3k4$fu$> 1@inn.qnx.com> …

“Chris McKillop” <> cdm@qnx.com> > wrote in message
news:9tbdnv$bjq$> 2@nntp.qnx.com> …

Have you tried giving all the options you want?

io-net -d speedo nomulticast,speed=100,duplex=1

Yes. Below is a capture of exactly what we see. Any other ideas? Do you
need
any other info?

slay io-net

io-net -d speedo nomulticast,speed=100,duplex=1 -p qnet -p tcpip

nicinfo

INTEL 82558 Ethernet Controller
Physical Node ID … 000347 881606
Current Physical Node ID … 000347 881606
Media Rate … 100.00 Mb/s half-duplex UTP
MTU … 1514
Lan … 0
I/O Port Range … 0xC400 → 0xC43F
Hardware Interrupt … 0xA
Promiscuous … Disabled
Multicast … Disabled

Total Packets Txd OK … 2
Total Packets Txd Bad … 0
Total Packets Rxd OK … 0
Total Rx Errors … 0

Total Bytes Txd … 162
Total Bytes Rxd … 0

Tx Collision Errors … 0
Tx Collisions Errors (aborted) … 0
Carrier Sense Lost on Tx … 0
FIFO Underruns During Tx … 0
Tx deferred … 0
Out of Window Collisions … 0
FIFO Overruns During Rx … 0
Alignment errors … 0
CRC errors … 0

ls -l /lib/dll/devn-speedo.so

-rwxrwxr-x 1 root bin 47698 Aug 10 10:19
/lib/dll/devn-speedo.so

uname -a

QNX neutrino3 6.1.0 2001/06/25-15:31:48edt x86pc x86



\

cdm@qnx.com > “The faster I go, the behinder I get.”

Chris McKillop – Lewis Carroll –
Software Engineer, QSSL
<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<

“Chris McKillop” <cdm@qnx.com> wrote in message
news:9tdu7d$t86$2@nntp.qnx.com

And, in fact, once I get developers.qnx.com updated you don’t need to give
any options.

Any idea when this might be?

Robert

“Chris McKillop” <> cdm@qnx.com> > wrote in message
news:9tdu7d$t86$> 2@nntp.qnx.com> …
And, in fact, once I get developers.qnx.com updated you don’t need to give
any options.

Any idea when this might be?

Look for it possibly today (22nd).

chris

cdm@qnx.com > “The faster I go, the behinder I get.”

Chris McKillop – Lewis Carroll –
Software Engineer, QSSL
<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<

Chris,

Sorry, I’ve been on vacation for a week - but I still don’t see this file on
developers.qnx.com ??

Robert

“Chris McKillop” <cdm@qnx.com> wrote in message
news:9ti9us$lhd$4@nntp.qnx.com

“Chris McKillop” <> cdm@qnx.com> > wrote in message
news:9tdu7d$t86$> 2@nntp.qnx.com> …
And, in fact, once I get developers.qnx.com updated you don’t need to
give
any options.

Any idea when this might be?


Look for it possibly today (22nd).

chris

cdm@qnx.com > “The faster I go, the behinder I get.”
Chris McKillop – Lewis Carroll –
Software Engineer, QSSL