Transmit timed out

Peter Hanecak hany@megaloman.sk
Mon Feb 15 18:48:57 1999


On 7 Feb 1999, Osma Ahvenlampi wrote:

> Peter Hanecak <hany@megaloman.sk> writes:
> > i'm using kernel 2.2.1 (eepro100.c:v1.06 10/16/98) and QDI Legend IV
> > mother board with on-board Intel 82557 MAC + NS83840A 10/100Mbps LAN
> > controller.
> > eth0: Transmit timed out: status 6050  0000 at 0/15 command 0001a000.
> > eth0: Trying to restart the transmitter...

i did some more "testing" with setting you sugested.
since last report i updated net-tools from 1.33-6 to 1.50-1 and disabled
automatinc startup of named and sendmail to prevent IP trafic before some
tests.

> Add 
> 
> options eepro100 multicast_filter_limit=3

i did that and it makes no diference. i tried multicast_filter_limit=0
too. dmesg content is at the bottom of this email.

i make some testing with two linux boxes (one is OK - call it IT, one is
with our eepro net-card -call it THIS) running ping on THIS and watched on
both machines using IPTRAF.

THIS:	172.16.0.10
IT:	172.16.0.1

after running 'ping 172.16.0.1' on THIS, "ARP from THIS to ffffffffffff on
eth0" has been received at IT and IT produces responses "ARP from IT to
THIS on eth0". about 5 seconds after running ping, "eth0: Transmit timed
out" occurs on THIS. after restarting eth0 interface (ifconfig eth0 down;
ifconfig eth0 up), ping produce no more results (no "Transmit timed" on
THIS, no "ARP from THIS to ffffffffffff on eth0" on IT). i'm able to
repeat that after reboot of THIS.

- here is dmesg output (with multicast_filter_limit=3 or
multicast_filter_limit=64 - it's same):

<ping>
eth0: Transmit timed out: status 6050  0000 at 0/16 command 0001a000.
eth0: Trying to restart the transmitter...
eth0: Transmit timed out: status 6050  0000 at 0/16 command 0001a000.
eth0: Trying to restart the transmitter...
<and so on till interface reset>

- dmesg (with multicast_filter_limit=0):

<ping>
eth0: Transmit timed out: status 6050  0000 at 0/14 command 0001a000.
eth0: Trying to restart the transmitter...
eth0: Transmit timed out: status 7048  0000 at 0/14 command 0001a000.
eth0: Trying to restart the transmitter...
eth0: Transmit timed out: status 7048  0000 at 0/14 command 0001a000.
eth0: Trying to restart the transmitter...
<and so on till interface reset> 

- ifconfig report after running ping (with multicast_filter_limit=64 - it
is default value):

eth0	Link encap:Ethernet  HWaddr 00:10:5C:AA:1E:78
	inet addr:172.16.0.10  Bcast:172.16.255.255  Mask:255.255.0.0
	UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
	RX packets:0 errors:0 dropped:0 overruns:0 frame:0
	TX packets:0 errors:6 dropped:0 overruns:0 carrier:0
	collisions:0 txqueuelen:100
	Interrupt:11 Base address:0x6900

- ifconfig report after running ping (with multicast_filter_limit=0):

eth0	Link encap:Ethernet  HWaddr 00:10:5C:AA:1E:78
	inet addr:172.16.0.10  Bcast:172.16.255.255  Mask:255.255.0.0
	UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
	RX packets:0 errors:0 dropped:0 overruns:0 frame:22
	TX packets:0 errors:3 dropped:0 overruns:0 carrier:0
	collisions:13 txqueuelen:100
	Interrupt:11 Base address:0x6900

- ifconfig report after running ping (with multicast_filter_limit=3):

eth0	Link encap:Ethernet  HWaddr 00:10:5C:AA:1E:78
	inet addr:172.16.0.10  Bcast:172.16.255.255  Mask:255.255.0.0
	UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
	RX packets:0 errors:0 dropped:0 overruns:0 frame:0
	TX packets:0 errors:5 dropped:0 overruns:0 carrier:0
	collisions:0 txqueuelen:100
	Interrupt:11 Base address:0x6900

for me it looks like another bug than reported by Bret Martin 
<bam@student.net> and/or Geoffry Meek <geoff@meek.com> (as you mentioned 
to me) or like another manifestation of the  same bug. but i'm not expert
in that.

> to your /etc/conf.modules. There's a bug in the eepro100 driver's
> multicast support that makes it crash when it receives multicast
> traffic. The bug might also affect SMP systems on a busy network in
> general. If the above line doesn't help, try setting the limit to 0 to 
> disable the hardware multicast support completely.
> 
> This is the most common FAQ regarding the driver. I wish something
> would be done about it. :/
> 
> -- 
> Let's split up. We can do more damage that way.
> Osma Ahvenlampi <oa@spray.fi>

just to be complete: i set multicast_filter_limit value directly in source
(line 44) because i'm not using modules (for now).

peter hanecak