I'm trying to get an Intel X520-DA2 to work with RHEL 5.2x64, after running make and make install on the driver, the system kernel-panics on bootup. I spoke to L1 support for about an hour and a half and they weren't able to get it going. I called back, and have been on hold for almost 2 hours so far without having spoken to a person. This is urgent as this is a system being used to ingest video for a feature film being shot starting tomorrow.
Thank you for posting.
I understand you are using Intel X520-DA2. What type of modules are installed on the network card?
Please visit https://downloadcenter.intel.com/product/69658/Intel-Ethernet-Server-Adapter-X520-Series Intel® Download Center to find the adapter's drivers. If you select Linux as your operating system, you'll see these versions:
1. Network Adapter Driver for PCI-E 10 Gigabit Network Connections under Linux*
Release Date: 2/23/2015
Version: 22.214.171.124 (latest version)
2. Network Adapter Driver for PCI-E 10 Gigabit Network Connections under Linux*
Release Date: 2/26/2014
Version: 3.19.1 (previously released)
In case you continue to encounter the same issue, please provide us error logs/messages.
I understand that you were able to talk to our support team. Were you given a service ticket number? If yes, please send it to me via private message (PM), so I can further look into the issue.
No modules installed yet - I don't know yet what the end-user is using. They are hard-wired copper cables, I believe. I will know later today. Version 126.96.36.199 is the one which crashed the system. Note that I am using a mainstream Linux with an unpatched kernel. I can't provide error logs because the system won't boot with the card/driver installed. I sent a screenshot of the kernel panic to support - case/ticker number is 8001109796.
I flew to Atlanta to look at this problem, and am in Atlanta now, and would greatly appreciate help today.
http://sourceforge.net/p/e1000/bugs/465/ Intel Ethernet Drivers and Utilities / Bugs / # 465 oops in ixgbe_probe on 2.6.18 kernel
This sounds like the issue you are running into. It also covers the proposed solution:
The correct fix is to move pci_save_state() after the call to pci_set_drvdata(), since the remove function relies on the adapter (derived from pci_get_drvdata) check to avoid double freenig resources in case of an error in ixgbe_probe().
Try that as see if that helps.