Bug 127195
Summary: | nforce2 (shuttle sn41g2) ethernet doesnt work on latest kernels | ||
---|---|---|---|
Product: | [Fedora] Fedora | Reporter: | Ed Price <edp> |
Component: | kernel | Assignee: | Arjan van de Ven <arjanv> |
Status: | CLOSED CURRENTRELEASE | QA Contact: | |
Severity: | medium | Docs Contact: | |
Priority: | medium | ||
Version: | 2 | CC: | barryn |
Target Milestone: | --- | ||
Target Release: | --- | ||
Hardware: | athlon | ||
OS: | Linux | ||
Whiteboard: | |||
Fixed In Version: | 2.6.7-1.494.2.2 | Doc Type: | Bug Fix |
Doc Text: | Story Points: | --- | |
Clone Of: | Environment: | ||
Last Closed: | 2004-08-03 23:00:02 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
Ed Price
2004-07-03 18:45:59 UTC
Created attachment 101614 [details]
dmesg for working kernel
Created attachment 101615 [details]
dmesg for non-working kernel
vanilla kernel.org linux-2.6.7 does not work either. also tried 2.6.7-bk19 and that fails too. WORKAROUND: disable USB 2.0 in BIOS (set "on-chip USB" to "1.1" instead of "1.1+2.0"). after this eth0 works in kernel-2.6.6-1.435.2.3. I have the same problem with a Shuttle MN31N and onboard ethernet. kernel 2.6.5-1.358 works fine but none of the 2.6.6 kernels (through the now current 2.6.6-1.435.2.3) work. I'm attaching two files with concatenated outputs from "/sbin/lspci -vxx", "cat /proc/interrupts", and "dmesg" for working and non-working kernels. Created attachment 101791 [details]
lspci, /proc/interrupts, and dmesg for working kernel
Created attachment 101792 [details]
lspci, /proc/interrupts, and dmesg for non-working kernel
There seems to be some similar bug tracking activity on the kernel bugzilla, but there kernel 2.6.6 works fine and it is 2.6.7 release candidates that are causing problems. http://bugzilla.kernel.org/show_bug.cgi?id=2920 http://bugzilla.kernel.org/show_bug.cgi?id=2846 http://bugzilla.kernel.org/show_bug.cgi?id=2799 There's even a thread going on in linux.kernel complex enough to include Linus Torvalds in its contributors regarding the 2.6.6->2.6.7 nforce issue. It looks like this is a pretty subtle bug involving interrupt sharing. Link to the thread attached below: http://groups.google.com/groups?hl=en&lr=&ie=UTF-8&c2coff=1&threadm=23T5k-2MZ-5%40gated-at.bofh.it&rnum=4&prev=/groups%3Fq%3Dnforce%2B%2B2.6.6%2Bethernet%26hl%3Den%26lr%3D%26ie%3DUTF-8%26c2coff%3D1%26selm%3D23T5k-2MZ-5%2540gated-at.bofh.it%26rnum%3D4 kernel.org linux-2.6.8-rc1 works. fedora "development" (FC3-test1 i guess) kernel-2.6.7-1.478 works. (there was some strange hex in the boot messages though.) Created attachment 101864 [details]
dmesg for linux-2.6.8-rc1 (works)
Created attachment 101865 [details]
dmesg for kernel-2.6.7-1.478 (works)
2.6.7-1.478 has a patch for GPG signing of kernel modules. I think that's where the hex output is coming from. I have the same issue with the 2.6.6-1.435.2.3 kernel on a Biostar M7NCG400 Nforce2 MB. The change from USB 2.0 to USB 1.1 in BIOS worked for me as well. latest FC2 kernel (2.6.7-1.494.2.2) resolves this issue for me. thx. closing. Works for me too (2.6.7-1.494.2.2) FYI: Self-compiled 2.6.7 vanilla kernel worked too, as dida 2.6.7-1.492 kernel from arjan. hmm, vanilla 2.6.7 worked? maybe i should try that again, just to double-check. actually, FYI ..... at some point, it seemed that i could not reproduce the original problem. ie, even with USB 2.0 enabled i could boot from previously non-working kernels, and eth0 worked! i looked at my history of up2date activity, thought maybe there was some other component involved... never figured it out. just recently i reinstalled FC2 from scratch, verified the original bug, then up2dated things, except kernel, gradually, checking every time, but eth0 never worked until i finally upgraded to kernel 2.6.7-1.494.2.2. so i'm still confused as to what happened on my previous install... PS realized later that i probably should have closed with resolution of "ERRATA" instead of "CURRENTRELEASE". oops sorry. |