11:10rhyskidd: mooch2: thanks for your comments on the falcon envytools series
11:11rhyskidd: i've got a much better understanding of them over the last few weeks
11:20RSpliet: rhyskidd: is there anything in particular you're trying to get done with falcons atm?
11:20rhyskidd: two things: banging on the trust barriers between LS and HS, and just learning for learning's sake
11:21rhyskidd: i've also got one of the PRIME laptops that suffers some secboot related runpm problems, so that would be nice to fix too
11:21RSpliet: I'm... unfortunately too busy to offer structural help or contributions (sorry, I saw your request for reviews...), but I've played around with both PMU, FECS and GPCCS
11:21RSpliet: Oh hmm... yeah, I've never investigated the secure boot faff
11:21rhyskidd: sure, that's fine
11:21RSpliet:shakes a fist about and grunts *NIVIDAAAA*
11:21rhyskidd: i read your paper -- interesting stuff
11:22RSpliet: Oh thanks. I merged a stripped version of the tool I used for data extraction and monitoring (to at least perform the context size data extraction... the other bit requires firmware changes that don't make sense upstream :-P)
11:22RSpliet: ... Did I just summon NVIDIA with my fist?
11:24RSpliet: If I have free time I'd love to see if I can also extract the exact registers that are swapped out from FRCS/GPCCS. Would make it much easier to check for problems in this area...
11:24RSpliet: The current workflow of tracing is quite involved
15:12karolherbst: that nvammiotracereplay isn't really implemented fully, right?
15:12karolherbst: like it does all writes into bar0
15:12karolherbst: and doesn't check for waits