e1000 device does not seem to be present
I have two ethernet cards attached to my linux desktop.
1. Intel 82545EM Gigabit Ethernet Controller
2. Intel 82543GC Gigabit Ethernet Controller
I have e1000 module installed that works fine with 82543GC.
but, eth1 - 82545EM does not show up.
When I tried,
$ service network restart,
following message came up:
Bringing up interface eth2: e1000 device eth2 does not seem to be present, delaying initialization.
Eventhough I have the device present.
It seems to me the problem with using multiple NIC cards using same driver. But do not know how to resolve this issue.
If any body have came across similar problem, or know the solution, please let me know.
uname -r : 2.6.9-1.667smp (fedora core 3)
this is eth2 I am talking about. I wrote eth1 in previous post, that was a typo.
I have similar problem with my system (FC4 and MSI 9218 - 1U rackmount server), and I found out that it could be a problem of incorrect checksum / checksum calculation in intel drivers.
with command: dmesg | grep e1000 I would get:
e1000: 0000:04:00.0: e1000_probe: The EEPROM Checksum Is Not Valid
e1000: probe of 0000:04:00.0 failed with error -5
So I downloaded drivers from intel site for my device (Linux* Gigabit Adapter Base Driver [e1000-6.3.9.tar.gz]) for Intel® 82573L Gigabit Ethernet Controller. After a slight modification I managed to get the device up and running.
Modification is to change function e1000_validate_eeprom_checksum and make it return E1000_SUCCESS always.
Here is patch
diff -Naur original/e1000-6.3.9/src/e1000_hw.c patched/e1000-6.3.9/src/e1000_hw.c
--- original/e1000-6.3.9/src/e1000_hw.c 2005-12-16 20:16:01.000000000 +0100
+++ patched/e1000-6.3.9/src/e1000_hw.c 2006-03-08 09:47:48.000000000 +0100
@@ -4119,7 +4119,9 @@
DEBUGOUT("EEPROM Checksum Invalid\n");
- return -E1000_ERR_EEPROM;
+ // return success because it would fail anyway
+ return E1000_SUCCESS;
+ // return -E1000_ERR_EEPROM;
I hope it helps!
I've had similar problems with Intel 82545EM-based NICs on my linuxbox. Even more, I've tried two different Intel 82545EM-based NICs, and I've got similar problems.
But when I tried third NIC, it was absolutely OK both with e1000 driver version 6.0.54 and e1000-6.3.9.
So, I suggest, that's not driver but NIC problem.
You are right. We have around dozen of these MSI servers in company. We found out that problem occurs only on some of them, and these boxes also report error on boot ROM of ethernet devices. It seems that problem appeared after BIOS update (but only on some of boxes).
Anyway the patch I posted seems to workaround the problem, on troubled boxes. I still don't see problem in behavior of devices with incorrect checksum.
2 Ethernet Ports Hosed
Looking at some of the posts on this and similar threads, my colleagues and I tried executing the statements provided in some of the posts (i.e., the 'dmesg' command, 'network.restart', etc.).
Please find attached two text files depicting what we have tried. Just to let you know, we have a total of 6 Ethernet ports, 4 of them being on 2 Pro/1000 Pt Dual Ethernet Port Server Adapters (2 ports/card), and 2 attached to the motherboard (along with the SVGA, Serial, Printer, USB, and PS/2--Keyboard, Mouse--ports); these 2 Ethernet ports are the ones that fail.
I'm hoping that it can be resolved through the Red Hat Enterprise Linux drivers and not have to go through the trouble of replacing the hardware/motherboard.
Please advise. Thanks.
|All times are GMT -5. The time now is 06:26 PM.|