|
Bugzilla – Full Text Bug Listing |
| Summary: | Onboard network card using forcedeth module on NVidia CK804 chipset does not work | ||
|---|---|---|---|
| Product: | [openSUSE] SUSE LINUX 10.0 | Reporter: | Thomas Renninger <trenn> |
| Component: | Kernel | Assignee: | Carl-Daniel Hailfinger <kernel01> |
| Status: | RESOLVED FIXED | QA Contact: | E-mail List <qa-bugs> |
| Severity: | Major | ||
| Priority: | P5 - None | CC: | herbert, kernel01, updates |
| Version: | Final | ||
| Target Milestone: | --- | ||
| Hardware: | x86-64 | ||
| OS: | All | ||
| Whiteboard: | |||
| Found By: | Other | Services Priority: | |
| Business Priority: | Blocker: | --- | |
| Marketing QA Status: | --- | IT Deployment: | --- |
| Attachments: | Patch for forcedeth.c version 0.35 to bugfree version | ||
|
Description
Thomas Renninger
2005-09-27 15:03:38 UTC
I think these issues should be fixed with the latest release of the driver (which we also added to the SLES9 SP3 kernel). Could you please give it a try? You can find the source in bugzilla #117677 Please report back if it indeed fixes the problem. SLES9-SP2 worked fine. Updated to SLES9-SP3 still works fine. 10.0 does not work. Once booted 10.0 I need to unplug power adapter (reboot does not help) to even get it working on SLES9 again. Will this stay a won't fix or is there still hope for a YOU update for 10.0? Is it worth trying 10.1 Alpha? 10.1 Alpha will most probably not help. Updating the driver in 10.0 might make sense, but I'm not yet fully convinced. Will test forcedeth 0.42 on my local machine and proceed with this bug afterwards. Hello, this seems to be a known problem: http://bugzilla.kernel.org/attachment.cgi?id=5623&action=view I'm having the exact same problem (also CK804 chipset, dual Opteron, two Nforce network slots). Strangely, my card works fine with the included forcedeth driver 0.35 in SL 10.0-OSS, UNLESS I a) put it under load (eg. downloads) b) install the NVIDIA graphics driver via fetchnvidia.sh / Yast online update. In case a), the driver outputs the error message "kernel: nv_stop_tx: TransmitterStatus remained busy<7>eth0: no IPv6 routers present" and gradually stops working (strangely, first DNS requests don't work any more, then the rest). In case b), it just doesn't work at all. The driver loads fine, I can assign IP adresses etc, but it cannot send packets any more (nothing gets out / in). To enable the cards again, I have to run switch2nv.sh, and then physically power down the machine ("reboot" does not do it) and restart. IMHO this is quite a serious bug: Many people use NVIDIA cards with NForce chipsets and those should work together. Maybe forcedeth 0.42 fixes the problems? Jens *** Bug 117573 has been marked as a duplicate of this bug. *** Carl-Daniel did you the test (10.0 kernel + 0.42) and did it work ? I had the same problem and updated the kernel to 2.6.14.2-3 from SL10.1 factory using forcedeth 0.41. Everything works fine now except that I had to update gcc to 4.1 to get the nvidia graphics driver to work. I have a solution for this bug from Manfred Spraul <manfred@colorfullife.com>. He sent me a small patch for the SUSE kernel, which upgrades forcedeth.c from version 0.35 to a bugfree version. He also wrote me: The bug is trivial to trigger: # modprobe # ifup # ifdown # ifup And then the network will die after a few minutes. During my own tests, I have never tested that sequence, just modprobe+ifup. But something in the boot sequence does the ifup/ifdown/ifup, and thus the nic is not initialized properly :-( I'll attach the patch. Created attachment 74553 [details]
Patch for forcedeth.c version 0.35 to bugfree version
Fixes hangup of forcedeth driver/card
If this is marked Resolved/Fixed, can someone please get this incorporated into a kernel update? |