Distribution:Debian testing (Sarge) Hardware Environment:A PC with AMD XP Software Environment: Problem Description:With kernel 2.6.5 ethernet interface (VT6102 [Rhine-II] (rev 78) on MB (Epox 8K9A7I) works OK. With 2.6.6 - 2.6.8 ethernet is completely dead. Not even ping works. With 2.6.8 acpi=off boot parameter ethernet interface works perfectly. I tested with kernels from ftp.funet.fi, without Debian patches. I used make-kpkg to compile and dpkg to instal kernels. Steps to reproduce:
Created attachment 3522 [details] lspci -vvv with 2.6.8
Created attachment 3523 [details] dmidecode ( 2.6.8 apic)
Created attachment 3524 [details] dmesg
Created attachment 3526 [details] interrupts
Does this workaround help? (comment one line 'link->irq.active = 0;' in drivers/acpi/pci_link.c) /* * forget active IRQ that is not in possible list */ if (i == link->irq.possible_count) { if (acpi_strict) printk(KERN_WARNING PREFIX "_CRS %d not found" " in _PRS\n", link->irq.active); +// link->irq.active = 0; }
Created attachment 3629 [details] /var/log/messages after NFS client crash I removed line 'link->irq.active = 0;' from drivers/acpi/pci_link.c. Now kernel works exactly as before fix with acpi=off. Everything except NSF client seems to work. I can cd to NFS mounted dir and see contents with ls. But trying to access a file with less or cat, causes a crash. kill -9 does not kill less process! var/log/messages included. This could be a separate bug. NFS works well with 2.6.5
i assume this is the same in 2.6.9?
Yes! Network interface of 2.6.9 works only with acpi=off
please re-test with an un-patched 2.6.10, as we made some updates there.
Unfortunately 2.6.10 behaves exactly like earlier versions (6 - 9): Network (Ethernet) seems to be completely dead. With acpi=off I didn't find any problems.
I had a similar problem on a similar computer. My Rhine also has IRQ 11 like many other devices (same situation as yours). The Rhine would not work (Though it'd modprobe and everything) if I used DRI for the graphics card. Now my graphics card is Unichrome not nVidia, but we both have VIA boards so it might be related. Are you using DRI? Try disabling it and testing. Also check your dmesg to see anything about "IRQ 11: nobody cared' See http://averatecforums.com/viewtopic.php?t=694 for some more information. Good luck!
I also have Via-rhine-II NIC Bug http://bugzilla.kernel.org/show_bug.cgi?id=6100 is the same but describing more general situation with all PCI devices.
Eugene, re: comment #12 The issue in bug #6100 is different -- lets work that one in that bug report. Jukka, What is the answer to Kevin's question about DRI? Can you still reproduce this issue with 2.6.22.stable?
bugme-daemon@bugzilla.kernel.org writes: > ------- Comment #13 from len.brown@intel.com 2007-08-18 11:57 ------- > > Jukka, > What is the answer to Kevin's question about DRI? > Can you still reproduce this issue with 2.6.22.stable? I had forgotten this! Situation is much better now. No problems in quick test with 2.6.22. Debian stable kernel had problems. Next one I tried only had strange problems with NFS (configuration?). I'll test more soon. - Jukka
bugme-daemon@bugzilla.kernel.org writes: > Can you still reproduce this issue with 2.6.22.stable? I compiled 2.6.22.3 with Debian configs, but without any Debian or other patches. I cannot reproduce this bug any more. So it seems to be fixed. (My config: http://www.iki.fi/~jkn/config-2.6.22.3k7jkn) Thank you! - Jukka