IRC channel logs
2019-11-06.log
back to list of logs
<oriansj>janneke: why does init_symbol consume 41 cells? <oriansj>tazjin: achieves nothing more than what you could get with coreboot+grub+write protect screw for bios <stikonas>yeah, some parts of that graph they have in that article are still marked as proprietary... <oriansj>because physical access ==> game over (Anyone can replace chips) <oriansj>probably a short way of saying "wildly unsecure shit here" <oriansj>because everytime a vendor has said the words "proprietary ..." and we did everything required to get the details; it is always ends up being "horrible crap that doesn't work" <oriansj>"proprietary encryption" => 32bit DES <oriansj>"proprietary fraud detection" => randomly select individuals <oriansj>"proprietary binary hardening" => we turned on a couple compiler flags to drop ELF data <oriansj>"proprietary recycling method" => we throw that shit in the trash <oriansj>"proprietary document destruction" => we just put them in a cargo container and ship them to China and then we don't care what happens to it <oriansj>I really wish that last one was a joke <janneke>oriansj: init_symbol should only consume 1 + <length of symbol-name>/cell-size cells <oriansj>well I will be pushing a commit that makes it easier to walk vectors and structs in gdb and when you walk it; it is a mess <janneke>great! another list of things to clean up :) <oriansj>well I did promise to make mes.c clean enough to be put into textbooks <oriansj>now vector->list simply becomes changing TREF->TPAIR and list->vector simply becomes changing TPAIR->TREF <oriansj>I probably broke structs hard doing this but I don't have good tests for structs to avoid breaking it <oriansj>now we can really simplify the symbol initalization and make symbol lookup much faster