Red Hat Bugzilla – Bug 137740
Nvidia ethernet requires quirk to fix invalid PCI class
Last modified: 2015-01-04 17:11:28 EST
From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.7.3) Gecko/20040921 Description of problem: When using the x86_64 FC3 RC1 diskboot.img installer on an nForce3 chipset motherboard, anaconda fails to list the forcedeth network module for manual selection. The install image contains forcedeth.ko in modules.cgz and a forcedeth entry is present in the loader (module-info). Version-Release number of selected component (if applicable): 10.1.0.2-1 How reproducible: Always Steps to Reproduce: 1. Boot from USB pen drive using FC3 RC1 diskboot.img 2. Select a network install method. 3. When asked to select driver, browse the list of 59 entries. Actual Results: Forcedeth entry does not exist, network install using onboard Ethernet not possible. Expected Results: A Forcedeth entry should have been listed and when selected, network install should have proceeded normally. Additional info: I have not tried this using the FC3 RC1 i386 tree, only x86_64.
that means the module already got loaded but no network devices are found. Can you provide lspci -v of your system?
Created attachment 106021 [details] lspci -v command output from a running FC2, kernel 2.6.9 system (MSI K8N Neo-FSR) Requested information for [Bug 137740]
Forcedeth does load on a running 2.6.9 kernel and is able to connect to the network with this motherboard. I guess I don't understand why forcedeth would load in the boot image, but be unable to find any network device. I would have thought that the module would exit and unload if it did not detect supported hardware.
Created attachment 106058 [details] /etc/sysconfig/hwconf Some more system info. 7th entry most relevant.
The installer (like the PCI spec) believes that ethernet devices should be network class devices not claim to be bridges.
Makes sense, so where does the change from bridge to network class need to be made? Is this a BIOS/firmware issue with MSI or a table update in the PCI spec? Where should the bits be flipped?
I can't comment on what facilities Nviidia may have for BIOS/firmware fixing this. Either way Linux can potentially lie to userspace (and does for a couple of other stupidities), so it would want dropping as a header fixup in the PCI quirks
An update has been released for Fedora Core 3 (kernel-2.6.12-1.1372_FC3) which may contain a fix for your problem. Please update to this new kernel, and report whether or not it fixes your problem. If you have updated to Fedora Core 4 since this bug was opened, and the problem still occurs with the latest updates for that release, please change the version field of this bug to 'fc4'. Thank you.
This bug has been automatically closed as part of a mass update. It had been in NEEDINFO state since July 2005. If this bug still exists in current errata kernels, please reopen this bug. There are a large number of inactive bugs in the database, and this is the only way to purge them. Thank you.