Bug 120289
| Summary: | Installer does not detect ne2k-pci | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Product: | [Fedora] Fedora | Reporter: | Marek Kassur <cmarco> | ||||||||
| Component: | kudzu | Assignee: | Bill Nottingham <notting> | ||||||||
| Status: | CLOSED RAWHIDE | QA Contact: | |||||||||
| Severity: | medium | Docs Contact: | |||||||||
| Priority: | medium | ||||||||||
| Version: | rawhide | CC: | dolik, guinan, katzj, kmilos, leonard-rh-bugzilla, notting, rvokal | ||||||||
| Target Milestone: | --- | ||||||||||
| Target Release: | --- | ||||||||||
| Hardware: | i386 | ||||||||||
| OS: | Linux | ||||||||||
| Whiteboard: | |||||||||||
| Fixed In Version: | 1.1.60-1 | Doc Type: | Bug Fix | ||||||||
| Doc Text: | Story Points: | --- | |||||||||
| Clone Of: | Environment: | ||||||||||
| Last Closed: | 2004-05-10 16:35:01 UTC | Type: | --- | ||||||||
| Regression: | --- | Mount Type: | --- | ||||||||
| Documentation: | --- | CRM: | |||||||||
| Verified Versions: | Category: | --- | |||||||||
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |||||||||
| Cloudforms Team: | --- | Target Upstream Version: | |||||||||
| Embargoed: | |||||||||||
| Attachments: |
|
||||||||||
|
Description
Marek Kassur
2004-04-07 16:05:59 UTC
I have the same problem here with an RTL8129 card, tried with both test2 and 04/04 development images. So you have a different card in there? Can you get to the second stage of the install and switch to tty2 and grab /tmp/anaconda.log, /tmp/syslog and the output of # /usr/bin/python -c 'import kudzu; print kudzu.probe(kudzu.CLASS_NETWORK, kudzu.BUS_UNSPEC, kudzu.PROBE_ALL)' Created attachment 99223 [details]
/tmp/anaconda.log
Created attachment 99224 [details]
/tmp/syslog
Created attachment 99225 [details]
output of python -c 'import kudzu...
Sorry, my card is RTL8029 (uses ne2k-pci as well), not 8129. I can't get to stage2 as I don't have another card in my machine but I suspect it would be more or less the same output. I do see though in stage1 in tty3 and tty4 that ne2k-pci is loaded and eth0 is up. Still no dice with 4/18 pxeboot images, but I hope I'll be able to at least give test3 a spin :) Same thing with Test3. Its on the list this time, but still no go. How about you Milos ? No success here either with test3's boot.iso A wired thing happeded when I add an ne2k-pci to my other test box (i810 with e100 on-board). It mess up my network settings (re-oreder ne2k-pci as eth0 and e100 as eth1, but preserve MAC settings). I can manually setup ne2k-pci using ifconfig, but not e100, don't know why as e100 module is loaded reporting itself as eth1. I also discovered that my USB mouse stopped working with an ne2k-pci controller. I am not so sure this is a kudzu bug. My RTL-8029 gets detected fine according to the output in tty4, it just is not brougth up. Manually inserting an ISA 3Com Etherlink III also fails. The card is also recognized, but not brought up. Maybe ifconfig does not get run? I can't do it manually as there is no prompt in tty2 yet. I'm seeing the same here (see http://www.redhat.com/archives/fedora-test-list/2004-April/msg02882.html) Indeed the card is detected and driver loaded, it's just that anaconda doesn't think there's anything to install from. FC2 test2 installed over the network just fine on the same box. See also bug 120360. And forget about the ISA card ;-) Running "linux rescue" from CD 1 does detect my RTL-8029 and Winbond ne2k-pci cards and configures them. I just want to add to my earlier #10 comment that it's normally working already installed system using e100 on-board (eth0). It's messed up by just plug-in an RL2000. Kudzu detect it and ask for configuration, but then ne2k-pci came up with e100 settings (IP, MAC). On-board e100 is partially configured as eth1 but refuses to came up: #ifup eth1 cannot change name of eth0 to eth1: Device or resource busy interface 'eth0' not found Device eth1 has different MAC address than expected, ignoring *** Bug 120360 has been marked as a duplicate of this bug. *** *** Bug 121955 has been marked as a duplicate of this bug. *** Fixed in kudzu-1.1.60-1. The kernel translates the - in the module name to an _ and thus confuses kudzu's isLoaded() logic. There is still something wrong with ne2k-pci, not sure about other
modules with dash in the name. My problem only appear with two or more
NICs, when one of them is ne2k-pci. I tested my system with 3 NICs
adding one by one (e100 is on-board). When I add pcnet32 everything
seems to be ok.
modprobe.conf:
alias eth0 e100
alias eth1 pcnet32
dmesg:
e100: eth0: e100_probe: addr 0xff8fe000, irq 11, MAC addr
00:03:47:12:35:AA
e100: eth0: e100_watchdog: link up, 100Mbps, full-duplex
^^^
pcnet32: PCnet/PCI II 79C970A at 0xdf80, warning: CSR address invalid,
using instead PROM address of 08 00 06 08 47 37 assigned IRQ 11.
eth1: registered as PCnet/PCI II 79C970A
pcnet32: 1 cards_found.
Then I plug-in ne2k-pci (not connected to hub).
modprobe.conf looks sane for me as well:
alias eth0 e100
alias eth1 pcnet32
alias eth2 ne2k-pci
But dmesg:
eth0: Compex RL2000 found at 0xdf80, IRQ 10, 00:80:48:C2:70:32.
eth1: registered as PCnet/PCI II 79C970A
e100: eth2: e100_probe: addr 0xff8f6000, irq 11, MAC addr
00:03:47:12:35:AA
...
NETDEV WATCHDOG: eth0: transmit timed out
eth0: Tx timed out, cable problem? TSR=0x0, ISR=0x0, t=397.
NETDEV WATCHDOG: eth0: transmit timed out
eth0: Tx timed out, cable problem? TSR=0x0, ISR=0x0, t=553.
NETDEV WATCHDOG: eth0: transmit timed out
May be related to Bug 121860 ?
Could someone else verify this behaviour.
Additional info:
kudzu-1.1.62-1
kernel-2.6.5-1.349
That's not related to this bug. It's instead something on the kernel side. Please open a separate bug for tracking this. |