Hello again,
Is it normal for io-net to be using 700 usecs of processor time every
3 seconds at priority 21? (I noticed someone was stealing the
processor from me and upped my process’s priority until it
stopped. After checking pidin for likely troublemakers at
that priority and then killing io-net, the 700 usec delays
disappeared.)
So, is 700usecs a normal amount of time for io-net to be running
(on a PIII667) or do you think I’m getting ping flooded or maybe
the driver is broken and doing some nasty faulting or something?
Is there a log file or utility to monitor what io-net is doing?
Thanks,
-Arthur
Arthur Q <3_arthur@my-deja.com> wrote:
Hello again,
Is it normal for io-net to be using 700 usecs of processor time every
3 seconds at priority 21? (I noticed someone was stealing the
processor from me and upped my process’s priority until it
stopped. After checking pidin for likely troublemakers at
that priority and then killing io-net, the 700 usec delays
disappeared.)
What ethernet card are you using? In terms of checking things,
you can use nicinfo to track the number of packets being sent
and received on a given network interface.
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:
: Arthur Q <3_arthur@my-deja.com> wrote:
:> Hello again,
:>
:> Is it normal for io-net to be using 700 usecs of processor time every
:> 3 seconds at priority 21? (I noticed someone was stealing the
:> processor from me and upped my process’s priority until it
:> stopped. After checking pidin for likely troublemakers at
:> that priority and then killing io-net, the 700 usec delays
:> disappeared.)
:>
: What ethernet card are you using? In terms of checking things,
: you can use nicinfo to track the number of packets being sent
: and received on a given network interface.
: chris
Also, which stack are you using? Do you have any tcp connections
when you see the blip? If so and you close them down, does it stop?
-seanb
“Chris McKillop” <cdm@qnx.com> wrote in message
news:8top0h$mbd$1@nntp.qnx.com…
Arthur Q <> 3_arthur@my-deja.com> > wrote:
Hello again,
Is it normal for io-net to be using 700 usecs of processor time every
3 seconds at priority 21? (I noticed someone was stealing the
processor from me and upped my process’s priority until it
stopped. After checking pidin for likely troublemakers at
that priority and then killing io-net, the 700 usec delays
disappeared.)
What ethernet card are you using? In terms of checking things,
you can use nicinfo to track the number of packets being sent
and received on a given network interface.
chris
It’s a 3com 3c905c fast etherlink xl pci card. Running nicinfo shows very
little network traffic. Unplugging the network cable (to eliminate all
packets),
doesn’t stop the 700 usec delays. Killing io-net does stop the delays.
-Arthur
“Sean Boudreau” <seanb@qnx.com> wrote in message
news:8tp8lc$1uu$1@nntp.qnx.com…
Chris McKillop <> cdm@qnx.com> > wrote:
: Arthur Q <> 3_arthur@my-deja.com> > wrote:
:> Hello again,
:
:> Is it normal for io-net to be using 700 usecs of processor time every
:> 3 seconds at priority 21? (I noticed someone was stealing the
:> processor from me and upped my process’s priority until it
:> stopped. After checking pidin for likely troublemakers at
:> that priority and then killing io-net, the 700 usec delays
:> disappeared.)
:
: What ethernet card are you using? In terms of checking things,
: you can use nicinfo to track the number of packets being sent
: and received on a given network interface.
: chris
Also, which stack are you using? Do you have any tcp connections
when you see the blip? If so and you close them down, does it stop?
-seanb
It’s the ttcpip stack, and there’s no tcp connections that I know of…if I
unplug
the network cable, I still see the blips.
I tried killing the io-net and restarting with the full tcpip stack, and
still saw
the blips, but then ls and ps commands slowed way down after I did that.
How can you get the full tcpip stack to startup on reboot?
-Arthur
Arthur Q <3_arthur@my-deja.com> wrote:
So, is 700usecs a normal amount of time for io-net to be running
(on a PIII667) or do you think I’m getting ping flooded or maybe
the driver is broken and doing some nasty faulting or something?
Is there a log file or utility to monitor what io-net is doing?
Try this for me…
slay io-net
io-net -d el900 speed=10,duplex=0,verbose -p ttcpip
Now, the speed and duplex are assuming that you are on a 10Mbit,
1/2 duplex hub. If you are connecting to something else, change
the values appropriatly. Let me know if the delay is still present
when you give these values.
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:8tpfuh$5la$1@nntp.qnx.com…
Arthur Q <> 3_arthur@my-deja.com> > wrote:
So, is 700usecs a normal amount of time for io-net to be running
(on a PIII667) or do you think I’m getting ping flooded or maybe
the driver is broken and doing some nasty faulting or something?
Is there a log file or utility to monitor what io-net is doing?
Try this for me…
slay io-net
io-net -d el900 speed=10,duplex=0,verbose -p ttcpip
Now, the speed and duplex are assuming that you are on a 10Mbit,
1/2 duplex hub. If you are connecting to something else, change
the values appropriatly. Let me know if the delay is still present
when you give these values.
chris
OK…I had to do a kill -9 to get rid of the io-net, and have a
100 Mbit hub, so I did:
/sbin/io-net -d el900 speed=100,duplex=0,verbose -p ttcpip
Found device: (10b7, 9200 at index 0)
Caps (72a2). Card is a 90XB
media_options: a
el900_init_phy: speed: 100000, duplex: 0
PHY found at address 24.
Media Options: a
Internal Config: 1800000
100BTHD
Link down 0
Link up 0 (100bT)
I still get the delays, even before enabling the device with phlip.
Note that networking was working for both the startup io-net
(with options -pttcpip -ppppmgr -del900 pci=0) and the above
io-net call (after enabling the device with phlip).
-Arthur
Arthur Q <3_arthur@my-deja.com> wrote:
OK…I had to do a kill -9 to get rid of the io-net, and have a
100 Mbit hub, so I did:
Okay, I need to move this to email. Can you send me email from an address
that you actually use? Not sure if you use your @my-deja.com address.
chris
–
cdm@qnx.com > “The faster I go, the behinder I get.”
Chris McKillop – Lewis Carroll –
Software Engineer, QSSL
<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<
We are looking at the smm (system management mode) as a possible cause can you give me as much hardware details as you can? That would include make model revision level on both bios and the board. does this board have wake on lan capability??? if so can you try to disable it
Previously, Arthur Q wrote in qdn.public.qnxrtp.os:
{ Hello again,
{
{ Is it normal for io-net to be using 700 usecs of processor time every
{ 3 seconds at priority 21? (I noticed someone was stealing the
{ processor from me and upped my process’s priority until it
{ stopped. After checking pidin for likely troublemakers at
{ that priority and then killing io-net, the 700 usec delays
{ disappeared.)
{
{ So, is 700usecs a normal amount of time for io-net to be running
{ (on a PIII667) or do you think I’m getting ping flooded or maybe
{ the driver is broken and doing some nasty faulting or something?
{ Is there a log file or utility to monitor what io-net is doing?
{
{ Thanks,
{ -Arthur
{
{
{
{
–
Pat Ford email: pford@qnx.com
QNX Software Systems, Ltd. WWW: http://www.qnx.com
(613) 591-0931 (voice) mail: 175 Terence Matthews
(613) 591-3579 (fax) Kanata, Ontario, Canada K2M 1W8