lima-march-sad happens for only 14 people in total, most of which on a very small amount of different servers.

The code around the common crash offset doesn’t make much sense either - if this function were to crash, this should have happened a few lines above, not suddenly at this point.

I’m not sure how relevant it is, but all the users reporting this issue were using an Ivy Bridge processor (3570k/3770k/… for example, or lower-end versions of the same). Might be some microcode/CPU bug?

Have any affected users tried updating their BIOS?