[PATCH 4/4] UML - Fix FP register corruption : fix for that FP register corruption bug in UML
UML kernel corrupts floating-point registers : manifests as occasional NaN values in running processes (via Mark Martinec)
#
Oh, what fun. Patch of the day, I think. (The last UML patch was a drag-out fight to try to come up with some way to stop UML going CPU-chewingly crazy forever whenever the host time moved backwards, say, because of NTP...)
Oh, no, wait, it's ancient, incorporated upstream in 2.6.25 and later.