The forcedeth driver for Nvidia2 is seriously broken for gentoo-sources-2.4.26-r13 and earlier. I face for some time now really massive problems in my network for all clients. I even exchanged cables until I detected hundreds of entries in the kernel log: ===== %< ====== Dec 10 22:40:46 [kernel] eth1: received irq with unknown events 0x1. Please report - Last output repeated 88 times - Dec 10 22:41:25 [kernel] eth1: received irq with unknown events 0x111. Please report Dec 10 22:41:25 [kernel] eth1: received irq with unknown events 0x1. Please report - Last output repeated 49 times - Dec 10 22:41:43 [kernel] eth1: received irq with unknown events 0x21. Please report Dec 10 22:41:44 [kernel] eth1: received irq with unknown events 0x1. Please report - Last output repeated 142 times - Dec 10 22:42:36 [kernel] eth1: received irq with unknown events 0x21. Please report Dec 10 22:42:37 [kernel] eth1: received irq with unknown events 0x1. Please report - Last output repeated 64 times - Dec 10 22:42:57 [kernel] eth1: received irq with unknown events 0x111. Please report Dec 10 22:42:57 [kernel] eth1: received irq with unknown events 0x1. Please report - Last output repeated 301 times - Dec 10 22:44:44 [kernel] eth1: received irq with unknown events 0x111. Please report Dec 10 22:44:45 [kernel] eth1: received irq with unknown events 0x1. Please report - Last output repeated 15 times - Dec 10 22:44:51 [kernel] eth1: received irq with unknown events 0x111. Please report Dec 10 22:44:51 [kernel] eth1: received irq with unknown events 0x1. Please report - Last output repeated 21 times - Dec 10 22:45:00 [kernel] eth1: received irq with unknown events 0x111. Please report Dec 10 22:45:00 [kernel] eth1: received irq with unknown events 0x1. Please report - Last output repeated 20 times - Dec 10 22:45:08 [kernel] eth1: received irq with unknown events 0x111. Please report Dec 10 22:45:08 [kernel] eth1: received irq with unknown events 0x1. Please report - Last output repeated 74 times - Dec 10 22:45:37 [kernel] eth1: received irq with unknown events 0x111. Please report Dec 10 22:45:38 [kernel] eth1: received irq with unknown events 0x1. Please report - Last output repeated 127 times - Dec 10 22:46:23 [kernel] eth1: received irq with unknown events 0x111. Please report Dec 10 22:46:24 [kernel] eth1: received irq with unknown events 0x1. Please report - Last output repeated 46 times - Dec 10 22:46:45 [kernel] eth1: received irq with unknown events 0x111. Please report Dec 10 22:46:45 [kernel] eth1: received irq with unknown events 0x1. Please report - Last output repeated 38 times - Dec 10 22:47:03 [kernel] eth1: received irq with unknown events 0x111. Please report Dec 10 22:47:03 [kernel] eth1: received irq with unknown events 0x1. Please report - Last output repeated 270 times - Dec 10 22:48:36 [kernel] eth1: received irq with unknown events 0x111. Please report Dec 10 22:48:36 [kernel] eth1: received irq with unknown events 0x1. Please report - Last output repeated 69 times - Dec 10 22:49:03 [kernel] eth1: received irq with unknown events 0x111. Please report Dec 10 22:49:03 [kernel] eth1: received irq with unknown events 0x1. Please report - Last output repeated 352 times - Dec 10 22:51:19 [kernel] eth1: received irq with unknown events 0x21. Please report Dec 10 22:51:20 [kernel] eth1: received irq with unknown events 0x1. Please report - Last output repeated 2 times - Dec 10 22:51:21 [kernel] eth1: received irq with unknown events 0x21. Please report Dec 10 22:51:23 [kernel] eth1: received irq with unknown events 0x1. Please report ===== %< ====== The clients loose sporadic network connection, fail to get new DHCP leases (and Windows clients will then get one by autoconfig <grrr>), have massive problems to load and save the roaming profiles via samba (including lost data). Reproducible: Always Steps to Reproduce: 1. Boot server 2. Boot client 3. Wait for network access Actual Results: Sporadic irq events 0x1. Can get very massive as shown in the log above with more unknown events. Network fails then for clients. Expected Results: No kernel logs, no network breakdowns. Board: ASUS A7N8X Targa Edition (stripped deluxe version) Never had any issues with the nforce-net driver :(
Please test 2.4.28, I'm fairly sure its fixed there
Hmmm. Meanwhile I've used the patch from here http://article.gmane.org/gmane.linux.kernel.commits.2-4/8779 to update the forcedeth driver from verison 0.25 to 0.30. I had to modify that patch to apply it though ... I'll attach it. Works so far, but this was not long enough to say something final.
Created attachment 45761 [details, diff] forcedeth-0.30.patch patch to upgrade forcedeth 0.25 to 0.30 in 2.4.26 kernel
Unfortunately the update to 0.30 of forcedeth did not help. Same problems, but without entries in the kernel log :( Will try 2.4.28 next ...
Unfortunately gentoo-sources-2.4.28-r2 did not help either. They have still forcedeth-0.25 and after some time suddently the entries in the log appeared and the client was cut from the net. Will now apply patch for forcedeth and try again.
Created attachment 46009 [details, diff] forcedeth-0.30-2.4.28.patch Patch for gentoo-sources-2.4.28-r2 to upgrade to forcedeth-0.30. Did unfortunately not work either. Network nearly down (ping loses 80% data) with no entry in the kernel logs (same behaviour as 2.4.28).
And it's getting weird. I've resurrected nforce-net-1.0.0261-r2 and booted with this driver. Network still fails ... 80% data losses for ping. I am now tempted to resurrect last gs-sources. I am quite sure I never had this issue with that kernel. Hmmm. I was able to stop net.eth1, unload module nvnet, load forcedeth-0.30 instead and after restarting my dependent services, the network worked for some minutes normally, but then started to fail then again. Sigh.
OK, meanwhile I assume it must bge something else. Just tried last gs-sources with the nforce-net and the network fails also. :(
OK. Exchanged my switch and network seems operational at full speed again. Had hardly any connection in the last two days. At least you have now a patch to upgrade forcedeth for 2.4.26 and 2.4.28 :) Therefore I did not close this issue, since I don't know, if you'll gonna integrate it.
Patch no longer applies to latest 2.4.28 version in gentoo-sources.
