←back to thread

286 points joegibbs | 2 comments | | HN request time: 0.42s | source
Show context
dcow ◴[] No.42144029[source]
Why not just go all the way and separate the program instruction memory from the data memory, physically? I know there’s an approximation of this at the page level, but why even let the kernel modify its own memory in the first place? Memory unit that only loads signed pages. etc.
replies(3): >>42144066 #>>42144559 #>>42145502 #
1. astrange ◴[] No.42144559[source]
What is this a response to? This already happens though - the kernel can't modify its own code pages. But there are plenty remaining attacks like overwriting function pointers or other data.
replies(1): >>42156794 #
2. dcow ◴[] No.42156794[source]
If you can’t modify kernel memory then how do you overwrite function pointers? (I understand how rop works.) At some level you have to get privileged execution and doing so would be incredibly difficult if you can’t trick the kernel at some point after whatever amount of ropping you’ve done.