From Bugzilla Helper:
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.2b) Gecko/20021010
Description of problem:
My NetGear FA-311 adapter didn't seem to work at first... I had no errors in
dmesg or /var/log/messages, but i couldn't ping to or from the machine... I
wanted to see if i got any other data, broadcasts or something, so I started
tcpdump -i eth0 and suddenly all pings and everything worked fine.
I have tried the 2.4.20-pre11 kernel and the natsemi driver in that seems to
work - i havn't tried the redhat errata kernel 184.108.40.206.8.0 yet, but i don't
think that it will make a difference regarding this problem as it seems to be
the same natsemi.c as in the 2.4.18-14 kernel
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. install netgear fa311 adapter and install rh8.0
2. after install ping anything and notice how it just doesn't work
3. run tcpdump -i eth0 on the machine and notice how everything changes :)
Actual Results: no data is transferred if not in promisc. mode
Expected Results: data should be transferrer even if not in promisc. mode
The failing configuration is an Dual-PPro 200 MHz machine - the error is seen
both with the smp and the uni kernel...
But I have an almost identical setup on a P-MMX 200MHz machine that seems to
work very well... (perhaps i'm missing something in the config-comparison, but
the seem to be very much alike) except that this on actually works without being
put into promisc. mode
Can you verify that the problem is fixed by the net driver in stock kernel
version 2.4.20-pre11? (i.e. 2.4.19 + 2.4.20-pre11 patch)
Yes - I have not seen any problems with the driver from stock kernel
2.4.20-pre11... I installed the 2.4.20-pre11 kernel on the system that failed
for me first(and still does) without altering any configuration on that
system... Booting the kernel 2.4.18-14(RedHat) will show the problem!
Booting 2.4.20-pre11 the problem is gone!
I would like to confirm this bug. Oddly enough, this is also on a Pro-200.
Perhaps it is specific to the processor or PPro chipset? I do not recall the
brandname of my ethernet card, but it is a natsemi chip. It worked fine in
RH7.3, does not in 8.0. Running tcpdump causes the network to suddenly start
I have not yet compiled a "stock" kernel, but I assume that this would work
around my problem, given the above information.
Even though the driver seems to have changed slightly in 2.4.19-0.pp.9(rawhide)
over 2.4.19(stock) the driver is still not nearly as updated as the 2.4.20-rc1
version... 1.0.15 vs 1.0.17 (looking inside the .c file) And it still doesn't work.
Actually, the problem might just be what Alan mentions in #75482 - And the
scenario descriped there seems to match some of the fixes in 1.0.17 of the driver
This is fixed in our rawhide kernel, which includes the updated natsemi driver
from kernel 2.4.20, and also our latest errata release.