Comment 185 for bug 263555

> So in the interests of adding some closure to this bug. The issue turns out to
> have never been the e1000e driver's fault.

Just to clarify. there were two bugs here. Yes the ftrace code should have been more careful in using cmpxchg, and tried harder to not write into code that might have swapped out (note, 2.6.28 has this fixed).

But the e1000e driver absolutely did have a bug. The driver should never had left open that a random write into it could brick the board. I'm actually glad that ftrace was the culprit. Because it allowed for a consistent reproducer. Just imagine if ftrace did not cause this. Any little bug in the kernel could have brick you card. And guess what? You would be out of luck, because it would be extremely hard to ever reproduce it again.

I'm not denying that ftrace did not have a bug. I just want the record to state, that ftrace was not the only one at fault here.