I've tagged v3.5 of my paranoid stack code. It's in a strange state right now:
Patch 1 (the actual stack switching) is standalone. It's not really a good base for the memory failure stuff, because all of the ist_enter stuff is missing. It should *work*, but the result will be ugly. OTOH, it should be okay as a base for fsgsbase. Patch 2 is from Paul McKenney. It fixes RCU NMI nesting, which is a prerequisite for ist_enter. I think that the current plan is for that to happen in 3.20. Speak up if you want it sooner. Patch 4 is purely for testing. It forces "NMIs" to nest. "x86, mce: Improve timeout error messages" is in there so I don't forget about it. Tony and Borislav, if you like it, want to pick it up? It really has very little to do with the rest of this code. The remainder of the non-atomic IST stuff is in there in a plausibly final form, but it's oddly based. --Andy -- Andy Lutomirski AMA Capital Management, LLC -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
