3c59x problems

jj ve1jot at hfx.eastlink.ca
Tue Jan 13 15:19:54 UTC 2004


Marina Buitrago wrote:

> Price, Jason wrote:
>
>> I have just installed Fedora for the first time, and am running into
>> problems with my NIC.  The box has a 3com 905B card in it, which is 
>> capable
>> of 100full.  The OS seems to recognize this capability, and also 
>> recognizes
>> the link partner also has this capability.  However, the link auto
>> negotiates to 10full.  I have tried forcing it via mii-tool, but it 
>> won't
>> take the change.  I added these lines to modules.conf:
>>
>> "options 3c59x options=4 full_duplex=1"
>>
>> But, that failed to obtain a link at all.  Below is the output of 
>> mii-tool
>> with no driver options:
>>
>> eth0: 10 Mbit, full duplex, link ok
>>   product info: vendor 01:e1:c1, model 56 rev 7
>>   basic mode:   isolate, collision test, 10 Mbit, full duplex
>>   basic status: autonegotiation restarted, link ok
>>   capabilities:
>>   advertising:  100baseT4 100baseTx-FD 100baseTx-HD flow-control
>>   link partner: 100baseT4 100baseTx-FD 100baseTx-HD flow-control
>>
>>
>> Any ideas how I can force this to 100full?
>>
>> Thanks.
>>
>> Jason
>>
>>
>
> I use the 3c59x module too and I couldn't make my card work properly 
> with the default autonegotiation, so I forced 100-FD both at the 
> switch and my computer, and now the errors are gone.
>
> These are the relevant lines in "/etc/modules.conf":
>
>
> ######
> # Tarjeta de red (3Com 3c905C-TX/TX-M [Tornado])
> #
> # 0x200 = 10BASET, FD
> # 0x204 = 100BASETX, FD
> # 0x008 = Auto
> #
> alias eth0 3c59x
> options 3c59x "options=0x204"
>
>
>
> Hope this helps.
>
I have a similar problem with same card, but my hub uses 10baseT 1/2 
duplex, and no matter what I
seem to do with mii-tool, the card stubbornly stays at full-duplex, 
10baseT and the outcome is horribly slow data rates
with ton's of frame errors..best it'll do is slower than dialup, lol! 
Card works well in XP, and correctly is set to 10baseT HD
Cheers,
John





More information about the fedora-list mailing list