"pump -i eth0", now non-functional, huh ?
Under 2.4.17.x I was using "pump -i eth0" to put my NIC in dhcp mode and all
I have now upgraded to "testing", i.e. [2.6.17-2-k7] and now pump won't
function. In fact it
puts the network down, which I bring back up with "networking restart" and
I can ping out onto the web by dotted-quad-addresses but not by name, since
dhcp mode is
not implemented, yet.
Another anomaly in all this is: I can't find the driver module [de620.o]
the system, but the NIC [D-Link 620] works 'cause I can ping [by
websites and LAN members.
Also, when I "ifconfig eth0" I see that eth0 does not have my NIC's proper
hardware address, i.e.
aa:bb:cc:dd:ee:ff type addr., though it does have some kind of addr which is
way longer than a real
address and makes no mention of the de620.
de620 is listed in /etc/modules, but after bootup, "lsmod" reveals no loaded
module, but the NIC does work.
I did not compile the above described kernel but rather installed the
kernel-image... in which everything else
seems to work OK.
"dmesg" does not reveal that de620 is implemented, though there are musings
Question1 [MAIN]: how can I implement this box as a dhclient, which
conceivably could sidestep the de620 issue, since
I CAN ping dotted-quad-addresses ?
Question2: how is the NIC implemented, compiled-in in some way ? How,
since there's no bootup msg saying this, in dmesg ?
If the NIC is not compiled and there's no module, how's it
Question3: why is the ifconfig eth0 output so strange and uncorresponding to
what formerly appeared,... but the NIC works ?
I'll gladly publish all the formerly referenced output but would prefer
first establishing a dhclient and see if that resolves the inability
to browse the web, and then; if possibly profitable and necessary, publish.
:-) After all, I need dhclient anyway, and
this may avoid all the other issues.