Comment 9 for bug 331415

Just to add to my comment about e100 driver yesterday, I found discussion about this problem in the 'Intel Wired Ethernet' project in SourceForge (http://sourceforge.net/tracker/?func=detail&aid=2876241&group_id=42302&atid=447449), and also a fix to the problem in kernel.org git repositories (commit 7e15b0c9991dfe0bf05a2f8fab9154bb7827622e on October 28th).

The fix does not use request_firmware_nowait(), but instead it stores a copy of the firmware when it is first loaded from userspace (during boot typically), and then uses this copy when resuming from suspend or hibernate later. This approach seems feasible since the e100 firmware is very small, less than 200 bytes. With bigger firmwares it might not be feasible.

Hoping to see a kernel for Ubuntu Karmic with this fix included.