So can #Spectre be worked around by flushing the TLB and also flushing the branch predictor with a long series of mispredicted branches?
-
-
Section 7 of https://spectreattack.com/spectre.pdf talks about inserting a CPUID at every conditional branch to prevent speculation. I was thinking of flushing the predictor once on entry/exit of kernel. I see now the paper suggests that too, but says there's no architecturally-defined method.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.