Topic: Intel e1000 'not found'

[Paul Isaacs]

Hi all,

I am trying to install Intels PRO/1000 MT desktop doard using the Intel e1000 drivers in 5.6u4 without success.

I have read several posts describing a 'bug' with the location of the drivers, but my problem appears to be a little different.

I have downloaded and installed the contrib:
e1000-4.4.19-sme56.i386.rpm

After doing this, the configuration panel correctly identifies the chipset as:
e1000        Intel Corp.|82540EM Gigabit Ethernet Controller
and allows me to save this as the configured driver (both eth0 and eth1).

Unfortunately, neither of the cards are detected when the system boots.
The following appears in the logfile:
===============================================
Sep 20 15:48:51 server1 kernel: ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx
Sep 20 15:48:51 server1 kernel: VP_IDE: VIA vt8233 (rev 00) IDE UDMA100 controller on pci00:11.1
Sep 20 15:48:51 server1 network: Bringing up interface eth0:  
Sep 20 15:48:51 server1 kernel:     ide0: BM-DMA at 0xfc00-0xfc07, BIOS settings: hda:DMA, hdb:DMA
Sep 20 15:48:51 server1 kernel:     ide1: BM-DMA at 0xfc08-0xfc0f, BIOS settings: hdc:pio, hdd:DMA
Sep 20 15:48:51 server1 kernel: hda: IC35L040AVER07-0, ATA DISK drive
Sep 20 15:48:51 server1 kernel: hdb: CDU5211, ATAPI CD/DVD-ROM drive
Sep 20 15:48:51 server1 kernel: hdd: IC35L040AVER07-0, ATA DISK drive
Sep 20 15:48:52 server1 kernel: ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
Sep 20 15:48:52 server1 kernel: ide1 at 0x170-0x177,0x376 on irq 15
Sep 20 15:48:52 server1 kernel: hda: 80418240 sectors (41174 MB) w/1916KiB Cache, CHS=5005/255/63, UDMA(100)
Sep 20 15:48:52 server1 kernel: hdd: 80418240 sectors (41174 MB) w/1916KiB Cache, CHS=79780/16/63, UDMA(100)
Sep 20 15:48:52 server1 kernel: ide-floppy driver 0.99.newide
Sep 20 15:48:52 server1 kernel: Partition check:
Sep 20 15:48:52 server1 ifup: e1000 device does not seem to be present, delaying eth0 initialization.
Sep 20 15:48:52 server1 kernel:  hda: hda1 hda2 < hda5 hda6 >
Sep 20 15:48:52 server1 kernel:  hdd: [PTBL] [5005/255/63] hdd1 hdd2 < hdd5 hdd6 >
Sep 20 15:48:52 server1 kernel: Floppy drive(s): fd0 is 1.44M
Sep 20 15:48:52 server1 kernel: FDC 0 is a post-1991 82077
Sep 20 15:48:52 server1 kernel: RAMDISK driver initialized: 16 RAM disks of 4096K size 1024 blocksize
Sep 20 15:48:52 server1 kernel: ide-floppy driver 0.99.newide
Sep 20 15:48:52 server1 kernel: md: md driver 0.90.0 MAX_MD_DEVS=256, MD_SB_DISKS=27
Sep 20 15:48:52 server1 kernel: md: Autodetecting RAID arrays.
Sep 20 15:48:52 server1 kernel:  [events: 000000ad]
Sep 20 15:48:52 server1 kernel:  [events: 000000a9]
Sep 20 15:48:52 server1 kernel:  [events: 000000a9]
Sep 20 15:48:52 server1 kernel:  [events: 000000ad]
Sep 20 15:48:52 server1 network: e1000 device does not seem to be present, delaying eth0 initialization.
Sep 20 15:48:52 server1 network: Bringing up interface eth0:  failed
Sep 20 15:48:52 server1 kernel:  [events: 000000a9]
Sep 20 15:48:52 server1 kernel:  [events: 000000a9]
Sep 20 15:48:52 server1 kernel: md: autorun ...
Sep 20 15:48:52 server1 kernel: md: considering hdd6 ...
===============================================

The Bootup BIOS screen shows the two etherenet cards being assigned interrupts 10 and 12.


I am really not experienced enough to debug this and would appreciate any assistance available,


   regards

      Paul

[Lloyd Keen]

Are you able to use 6.0beta3? It automagically detects the Gigabit Intel NIC. Asking the obvious, but have you checked they are enabled in the bios. BTW If you are trying to setup software raid the drives should both be setup as the master ie hda and hdc. You appear to have the second drive setup as a slave (hdd)

[Paul Isaacs]

Hi Lloyd,
Thanks for the reply:

1. This is a production server, hence I don't wish to use 6.0 while its in beta (I'd get flamed by everyone).
2. The NICs are plug in PCI cards, not onboard, so I'm not sure what (if anything) should be enabled in BIOS.
3. I have only just updated to 5.6u4 (actually u5 now), and the server used to have an IDE tape as master device on the secondary IDE, hence the hard disk was the slave. I have a philosophy of "If it ain't broke", so I have left it alone and it seems to be functioning fine. The server now has a DLT SCSI tape with the IDE drive removed.

 regards
   Paul

[Lloyd Keen]

Forget that last post, I'm getting a bit confused with scsi adapters. Have you tried doing a #modprobe e1000 after bootup? If so does it load the module or what errors does it throw up. Check to see if the module has loaded using #lsmod.

[Paul Isaacs]

Hi Lloyd...
Thankls for your help, I really appreciate the assistance.

re:
#modprobe e1000

OUCH!!!! (results below).
=======================================
/lib/modules/2.4.18-5smp/kernel/drivers/net/e1000.o: unresolved symbol skb_under
_panic_Rsmp_8cf74b1a
/lib/modules/2.4.18-5smp/kernel/drivers/net/e1000.o: unresolved symbol __kfree_s
kb_Rsmp_99c012ad
/lib/modules/2.4.18-5smp/kernel/drivers/net/e1000.o: unresolved symbol alloc_skb
_Rsmp_56c40618
/lib/modules/2.4.18-5smp/kernel/drivers/net/e1000.o: unresolved symbol __netdev_
watchdog_up_Rsmp_26a85e68
/lib/modules/2.4.18-5smp/kernel/drivers/net/e1000.o: unresolved symbol qdisc_res
et_Rsmp_2dfa69ab
/lib/modules/2.4.18-5smp/kernel/drivers/net/e1000.o: unresolved symbol eth_type_
trans_Rsmp_5cb3db4e
/lib/modules/2.4.18-5smp/kernel/drivers/net/e1000.o: unresolved symbol irq_stat_
Rsmp_624c800f
/lib/modules/2.4.18-5smp/kernel/drivers/net/e1000.o: unresolved symbol skb_reall
oc_headroom_Rsmp_178f84df
/lib/modules/2.4.18-5smp/kernel/drivers/net/e1000.o: unresolved symbol softnet_d
ata_Rsmp_839eab9e
/lib/modules/2.4.18-5smp/kernel/drivers/net/e1000.o: unresolved symbol netif_rx_
Rsmp_0e1638e3
/lib/modules/2.4.18-5smp/kernel/drivers/net/e1000.o: unresolved symbol skb_over_
panic_Rsmp_9fc06b69
/lib/modules/2.4.18-5smp/kernel/drivers/net/e1000.o: insmod /lib/modules/2.4.18-
5smp/kernel/drivers/net/e1000.o failed
/lib/modules/2.4.18-5smp/kernel/drivers/net/e1000.o: insmod e1000 failed
============================================

I may have been a bad boy...
When I wnt to install the full e1000 contribs, it came up with some dependancy errors and didnt install.

The full contrib is:
e100-2.1.29-sme56.i386.rpm 129 1/25/2003 7:47 pm
e1000-4.4.19-sme56.i386.rpm 142 1/25/2003 7:49 pm
iANS-2.0.12-sme56.i386.rpm 245 1/25/2003 7:46 pm
inic-snmp-1.3.40-RH8x.i386.rpm 213 1/25/2003 7:44 pm
procfg-1.4.32-1.i386.rpm 821 1/25/2003 7:44 pm
procfgd-1.4.33-1.i386.rpm 2,170 1/25/2003 7:43 pm

rpm -Uvh produced:
error: failed dependencies:
        net-snmp is needed by inic-snmp-1.3.40-RH8x
        libcrypto.so.1 is needed by procfg-1.4.32-1
        libcrypto.so.1 is needed by procfgd-1.4.33-1

So I just installed
e1000-4.4.19-sme56.i386.rpm
based on the forums talking only about this module.

This may be a red herring.. but it may also be part of the problem.

regards

Paul


Lloyd Keen wrote:
>
> Forget
> that last post, I'm getting a bit confused with scsi
> adapters. Have you tried doing a #modprobe e1000 after
> bootup? If so does it load the module or what errors does it
> throw up. Check to see if the module has loaded using #lsmod.

[Lloyd Keen]

Paul,
Try the default drivers that ship with sme5.6 and see if that works. They are located in:
> /lib/modules/2.4.18-5smp/kernel/drivers/addon/e100/e100.o
> /lib/modules/2.4.18-5/kernel/drivers/addon/e100/e100.o
> /lib/modules/2.4.18-5smp/kernel/drivers/addon/e1000/e1000.o
> /lib/modules/2.4.18-5/kernel/drivers/addon/e1000/e1000.o
Uninstall the 4.4.19 drivers first:
#rpm -e e1000
Then try:
#modprobe e1000
and see what happens. You don't need all the other stuff as they are basically management utilities.

[Paul Isaacs]

Success.

You are quite right.

I uninstalled the e1000 and used:
/sbin/e-smith/db configuration set EthernetDriver1 e1000
/sbin/e-smith/signal-event console-save

rebooted and all is well.

Thank you very much for your help. I wouldn't have got through this without it.
Much appreciated,

Paul

[Lloyd Keen]

Just lucky you're from down south, if you had of been in QLD you would have been on your own after the Lions knocked off the swannies last night )