Red Hat Bugzilla – Bug 163353
e1000 for ABIT IC7-G missing from pcitable, 0x8086 0x1075
Last modified: 2007-11-30 17:11:10 EST
Description of problem:
When trying out diskless stateless-linux 2 of our machines, two
of our machines failed to start the network card e1000 since
the vendor and device numbers are missing from:
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. Boot stateless linux with kernel-smp-2.6.11-1.27_FC3, and I would
imagine the latest kernel as well which I will try soon, in a diskless
mode via pxe, on an ABIT IC7-G motherboard.
It modprobes nfs but fails to modprobe e1000 so the kernel panics
trying to mount the root filesystem.
the initrd should contain a pcitable with the right numbers
so the kernel can load e1000 and boot.
I manually patched the initrd and added this line to pcitable:
0x8086 0x1075 "e1000"
After that it booted just fine.
I'll attach a patch to hwdata-0.145.
Here is the lspci -vvn info:
02:01.0 Class 0200: 8086:1075
Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr-
Stepping- SERR- FastB2B-
Status: Cap+ 66Mhz+ UDF- FastB2B- ParErr- DEVSEL=medium >TAbort-
<TAbort- <MAbort- >SERR- <PERR-
Latency: 0 (63750ns min), Cache Line Size 08
Interrupt: pin A routed to IRQ 11
Region 0: Memory at fd000000 (32-bit, non-prefetchable) [size=128K]
Region 2: I/O ports at a000 [size=32]
Capabilities: [dc] Power Management version 2
Flags: PMEClk- DSI+ D1- D2- AuxCurrent=0mA
Status: D0 PME-Enable- DSel=0 DScale=1 PME-
Here is a link to the motherboard:
Intel PRO/1000 CT Desktop Connection Gigabit LAN on board
is what they call the network card.
Created attachment 116798 [details]
A patch to pcitable to add the correct entry.
It's in the modules.pcimap for the driver, so this shouldn't be needed; I'd
suspect the problem is elsewhere.
Is the module actually on your initrd?
Yes it is. The 'linuxrc' file in the initrd from stateless linux
uses pcitable. That's how I found the solution. Patching that
fixes the boot problems for diskless stateless linux on those
two machines we have.
Maybe stateless-linux is using an old and cumbersome way of
determining which modules if should load?
Yes; it should be using some combination of modules.pcimap/pcitable.
Not sure what component to move the bug to, though
It should probably be assigned to havoc at least.
*** This bug has been marked as a duplicate of 166018 ***