[vortex] vortex_error(), status=0xe081, 3c905B

[Rune Mossige]

Hello,
I have one RedHat 7.0 with kernel 2.4.1 and a 3c905B, as in:

eth0: 3Com PCI 3c905B Cyclone 100baseTx at 0xe400,  00:50:04:9b:1c:4e,
IRQ 10

that give some very stranger network performance, as a simple ftp show:

ftp> get /unix /dev/null
local: /dev/null remote: /unix
227 Entering Passive Mode (136,250,62,21,226,254)
150 Opening data connection for /unix (3842678 bytes).
226 Transfer complete.
3842678 bytes received in 327 secs (11 Kbytes/sec)

The ftp is from an RS6000 F50 with a GigaBit interface.

The Linux box is connected to a 3COM SuperStack II 3300, and the
SuperStack have one 100Mbit/s full duplex uplink into a cabletron
SSR2000, where the F50 is connected.

I have tried to set the port on the SuperStack where the Linux box is
connected to: forced 100-FD, forced 100-HD, and auto negotiate.

I have tried the Linux box with the 3c59x compiled into the kernel, and
as a module. Using it as a module, I have tried with and without
'options=0x204' to force it 100-FD, but nothing appears to give me any
better performance....

When runing as a module, I also added 'debug=3', and that printed out
pages and pages of:
eth0: vortex_error(), status=0xe081
eth0: vortex_error(), status=0xe081
eth0: vortex_error(), status=0xe081
eth0: Media selection timer tick happened, Autonegotiate.
dev->watchdog_timeo=40
eth0: MII transceiver has status 786d.
eth0: Media selection timer finished, Autonegotiate.
eth0: vortex_error(), status=0xe081
eth0: vortex_error(), status=0xe081
eth0: vortex_error(), status=0xe081
 Rx error: status 08.
eth0: vortex_error(), status=0xe081
eth0: vortex_error(), status=0xe081
eth0: vortex_error(), status=0xe081
 Rx error: status 08.
 Rx error: status 08.
eth0: vortex_error(), status=0xe081
eth0: vortex_error(), status=0xe081
eth0: vortex_error(), status=0xe081

I have also tried to ftp to/from other hosts connected to the same
SuperStack switch, without any change in behaviour and speed.

I have also tried to connect the Linux box directly to the SSR2000, but
still no change.....

I have also replaced all drop cables and patch cables between the Linux
box and the switch, but no change....

Do You have any idea of what might be wrong?

-------------------------------------------------------------------
(-: Hiroshima 45, Chernobyl 86, Windows 95 :-)
Our ultimate goal is to make overloaded machines appear to be idle.
High performance, High reliability, Low cost -------- Pick any two.
-------------------------------------------------------------------
Rune Mossige, Systems Support, WesternGeco, Stavanger, Norway
Tel: (+47)51946869                         Mobile:(+47)90871024