[tulip] FA511 not passing data

Donald Becker becker@scyld.com
Mon Jan 6 15:49:02 2003


On Mon, 6 Jan 2003, Robert W March wrote:

> I installed a NetGear FA511 PCBus card on a Thinkpad 600E running RH7.3 and 
> the card shows up as:
> 05:00.0 Ethernet controller: Linksys Fast Ethernet 10/100 (rev 11)
>         Subsystem: Netgear: Unknown device 511a

That's strange -- it's a NetGear card, right?  It might be just the name
list, which is frequently wrong.  What are the numeric values for IDs?
(Use the '-n' option.)

>         Region 0: I/O ports at 4800 [size=256]
>         Region 1: Memory at 12c00000 (32-bit, non-prefetchable) [size=1K]
..
> I built and installed netdrivers-3.3.  All installed OK except:
> depmod: *** Unresolved symbols in 
> /lib/modules/2.4.18-19.7.x/kernel/drivers/pcmcia/cb_shim.o
> 
> Maybe cb_shim.o is not needed, or is it?

It is only needed if you are using the original PCMCIA code.
Unfortunately, there is no good way to detect that at compile time.  So
the Makefile unconditionally builds the module.

  The card is not working.  ping says:
> PING cougar.localdomain (192.168.0.252) from 192.168.0.10 : 56(84) bytes of 
> data.
> 64 bytes from cougar.localdomain (192.168.0.252): icmp_seq=1 ttl=64 time=0.570 
> ms
...
> 14 packets transmitted, 3 received, 78% loss, time 13064ms
> rtt min/avg/max/mdev = 0.303/0.400/0.570/0.121 ms

It's work a little bit... that narrows down the problem.

> dmesg says:
> pci-scan.c:v1.11 8/31/2002  Donald Becker <becker@scyld.com> 
> http://www.scyld.com/linux/drivers.html
> tulip.c:v0.95f 11/17/2002  Written by Donald Becker <becker@scyld.com>
>   http://www.scyld.com/network/tulip.html
>   The PCI BIOS has not enabled the device at 5/0!  Updating PCI command 
> 0003->0007.

That's curious!  I'm guessing that there is an error message from the
CardBus code, likely earlier in the boot sequence.

> eth0: ADMtek Centaur-C rev 17 at 0xd29ee000, EEPROM not present, 
> 00:4C:69:6E:75:79, IRQ 11.

This is a Bad Thing.  The driver had to make up a station address.  If
the EEPROM is really erased or missing, card-specific NIC parameters
have not been loaded.

> eth0:  MII transceiver #1 config 3100 status 7869 advertising 05e1.

What does 'mii-diag' report?

> tulip-diag -eee reports:
> tulip-diag.c:v2.16 12/17/2002 Donald Becker (becker@scyld.com)
>  http://www.scyld.com/diag/index.html
> Index #1: Found a ADMtek AL985 Centaur-C adapter at 0x4800.
>  Comet duplex is reported in the MII status registers.
>  Transmit started, Receive started.
>   The Rx process state is 'Waiting for packets'.
>   The Tx process state is 'Idle'.
>   The transmit threshold is 128.

Normal operation so far.

>   Comet MAC address registers 6e694c00 ffff7975
>   Comet multicast filter 0000000040000000.
> EEPROM 256 words, 8 address bits.
> WARNING: The EEPROM is missing or erased!
>   Ethernet MAC Station Address ff:ff:ff:ff:ff:ff.
>   Default connection type 'Default'.
>   PCI IDs Vendor ffff Device ffff  Subsystem ffff ffff

Hmmm, something is wrong.  The chip did manage to load the subsystem ID,
so part of the EEPROM is programmed.

> I see similar issues posted on this list but I have not seen a
> solution.

What similar issues?  I haven't noticed another error report that matches
this one.

-- 
Donald Becker				becker@scyld.com
Scyld Computing Corporation		http://www.scyld.com
410 Severn Ave. Suite 210		Scyld Beowulf cluster system
Annapolis MD 21403			410-990-9993