Do these Meltdown/Spectre fixes affect page fault handlers in addition to syscall codepaths?
-
-
Replying to @dakami
without looking at the code, conceptually the same steps need to be done for any user-to-kernel switch.
1 reply 1 retweet 1 like -
Every single trap from page faults to breakpoints to system calls to machine check aborts.
1 reply 0 retweets 5 likes -
I'm pursing the "we're just going to need to run more kernels" hypothesis and it's looking less heretical and more inevitable every day.
2 replies 0 retweets 2 likes -
pursuing
1 reply 0 retweets 0 likes -
I mean you pretty much now have a micro GS segment that UM can see, a micro stack, a micro set of interrupt and trap stubs, a micro page directory... it’s essentially a UM-visible shadow kernel that jumps into the real kernel.
1 reply 0 retweets 5 likes -
At the end of the day we're still pretending we have multiple security domains we're sharing resources across, only it's really always the same security domain (different stability domains though, that didn't change)
1 reply 0 retweets 0 likes -
That's funny though, you're right, the system is subsystem by subsystem evolving straight towards multi-kernel.
1 reply 0 retweets 2 likes
see also: Hyper-V and VBS :D
-
-
Yeh, I'm deep in KVM, but there's also LKL and various stunts I'm trying to pull with User Mode Linux
0 replies 0 retweets 2 likesThanks. 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.