Comment 21 for bug 1176577

I believe I have the same issue with my Dell Optiplex 755 small form factor desktop. I just upgraded from 12.10 to 13.04 with kernel 3.8.0-19-generic amd64. I'm using it as a firewall/router, so it has 2 Intel Pro/1000 gigabit ethernet adapters installed (one onboard, one PCI). I get streams of the same MEI error message as the original report above, and the PCI ID from the error corresponds to the PCI GbE adapter; I also get frequent reports of TX Unit Hang Detected on the system's onboard GbE adapter, which causes an auto-reset of the adapter and loss of internet connectivity for 10-15sec. The MEI error on eth0 (PCI, internet side) occurs every few (3-10) seconds, while the TX Unit Hang error/reset on eth1 (onboard, LAN side) occurs randomly a few times a day so far.

The odd thing is that I have another Optiplex 755, a mini-tower unit, which has only the onboard GbE NIC. I just upgraded it from 12.10 to 13.04 as well, again with kernel 3.8.0-19-generic amd64, and I get neither the MEI errors nor the TX Unit Hang Detected messages. In both systems the onboard GbE NIC uses the e1000e driver, while on my firewall box the additional PCI GbE NIC uses the e1000 driver.

I just rebooted the firewall box to kernel 3.5.0-28-generic amd64, and so far I haven't seen the MEI errors appear. Time will tell whether the TX Unit Hang errors will appear. I think for now I will remove the rest of the kernels and keep 3.5.0-28.