Control-flow integrity in 5.13

Posted May 22, 2021 5:00 UTC (Sat) by Cyberax (✭ supporter ✭, #52523)
In reply to: Control-flow integrity in 5.13 by Paf
Parent article: Control-flow integrity in 5.13

It's not a bad idea, actually. In many places you might only have just a few targets.

to post comments

Control-flow integrity in 5.13

Posted May 22, 2021 14:45 UTC (Sat) by Paf (subscriber, #91811) [Link] (1 responses)

Ok, but - numbers? I’m struggling to see how multiple jumps is better than a single mispredicted execution branch.

Control-flow integrity in 5.13

Posted May 25, 2021 18:00 UTC (Tue) by andresfreund (subscriber, #69562) [Link]

I can't imagine a binary search working well, but it's not hard to believe a few hot cases checked linearly could work out. A mispredicted call is pretty expensive.

I'm pretty sure that several compilers use profile guided "optimistic" devirtualization, which basically ends up with code like
if (call_target == very_common_target) very_common_target() else if (call_target == also_common_target) also_common_target() else *call_target(). And I've seen code like that manually written in plenty places, with good success.