[gobolinux-users] Internet problem in 013

Andy Feldman nereusren at gmail.com
Sat Dec 16 04:51:23 UTC 2006


I split this into a new thread since it doesn't have anything to do
with the Scripts release.

I'm having internet trouble with 013. The DHCP client connects to my
Linux router and get the IP it's supposed to during boot, and then the
internet works for exactly 1 minute. After that, the entire network is
unreachable (besides the loopback to the IP that was assigned).
ifconfig still shows the same info on the interface as if it's
connected fine. Running "StopTask Network; StartTask Network" results
in a DHCP timeout error.

If I unplug my ethernet cable and plug it back in, it "resets" it so I
get one more minute of internet by running StartTask Network... and
then the same thing happens again. The minute starts from when it
recognizes that I've plugged it back in, rather than when the DHCP
lease is acquired.

I copied some possibly relevant logs and info below. Please let me
know what other relevant information I can get, or any tests I can
try. Thanks!

(Incidentally, the script /P/DHCP/Settings/dhcpc/dhcp.exe didn't have
execute permissions by default, but DHCP worked fine without it. The
only difference I got after chmod +x was more debug messages when
running dhcpcd -d, but the same ultimate behavior. Nevertheless, that
could be checked in the package. The lack of debug messages confused
me for a while.)

-Andy


Here's the more detailed info:

Under my 012 setup (which works), the ipv6 driver loads by default,
where in 013 it doesn't. However, modprobe ipv6 in 013 doesn't change
the behavior.

dmesg in both 012 and 013 shows that it picks up my ethernet card as
"eth0: Davicom DM9102/DM9102A rev 49 at e0a7cc00, 00:08:A1:04:4F:7E,
IRQ 11."

Attempting "StartTask Network" and having it fail as described above
adds these lines to dmesg:
ADDRCONF(NETDEV_UP): eth0: link is not ready
0000:02:00.0: tulip_stop_rxtx() failed
0000:02:00.0: tulip_stop_rxtx() failed

Unplugging it and plugging it back in before retrying adds
"ADDRCONF(NETDEV_CHANGE): eth0: link becomes ready" to dmesg. This is
the indicator that I have 1 minute of internet. I measured it by
issuing "StartTask Network; ping 192.168.10.1". dhcpcd -d (I added the
debug flag to DHCPNetworkInterface) echoes the "link becomes ready"
message, and then gets the lease almost instantly and lets the ping
command run. ping issues packets once per second and gets through 59
packets before they start getting dropped.

If I wait for a while after plugging it in before issuing StartTask
Network, the "link becomes ready" message is never echoed (it probably
already picked up on it), and ping doesn't last a full minute.

During the minute I can successfully bring the interface up and down
repeatedly, which cancels and renews the lease each time as expected.


In 013, ifconfig eth0 returns:

eth0      Link encap:Ethernet  HWaddr 00:08:A1:04:4F:7E
          inet addr:192.168.10.10  Bcast:192.168.10.255  Mask:255.255.255.0
          UP BROADCAST MULTICAST  MTU:1500  Metric:1
          RX packets:0 errors:0 dropped:0 overruns:0 frame:0
          TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:0 (0.0 b)  TX bytes:0 (0.0 b)
          Interrupt:11 Base address:0xcc00

Which all seems to be correct. (192.168.10.10 is the address this MAC
in supposed to receive from the dhcp server on my router.) The only
real differences in my 012 ifconfig are "NOTRAILERS" in the "UP" line,
and an ipv6-specific line (since ipv6 driver isn't loaded in 013):
"inet6 addr: fe80::208:a1ff:fe04:4f7e/64 Scope:Link"


More information about the gobolinux-users mailing list