I don't know why people are so focused on intel here. Its not like other processors aren't affected, even ARM's latest core (A75 https://developer.arm.com/suppo

I don't know why people are so focused on intel here. Its not like other processors aren't affected, even ARM's latest core (A75 https://developer.arm.com/support/security-update https://developer.arm.com/support/security-update) is susceptible to meltdown. I suspect that if they had been building aggressive OoO CPU's with deep ROB rather than focusing on perf/W there would be a lot more cores on that list. Similarly IBM's high end POWER processors are also vulnerable (although they don't break it down as nicely as ARM https://www.ibm.com/blogs/psirt/potential-impact-processors-power-family/ https://www.ibm.com/blogs/psirt/potential-impact-processors-...) So, everyone is like Intel is so bad, when they just happen to be the one that people are running the most untrused code on. Let me quote IBM "This vulnerability doesn’t allow an external unauthorized party to gain access to a machine, but it could allow a party that has access to the system to access unauthorized data." People on this board are upset, because after spending the last decade+, making promises about how secure "cloud computing" is, once again the naysayers were proven correct. This time the flaw is so fundamental that in order to fix it you have the OS vendors making changes the destroy system performance for most applications that are I/O or just syscall intensive. This likely won't be the last time either if history is to be believed. But javascript you cry, again I'm going to say that you shouldn't be running random code from random people on the internet. I'm not a big fan of intel, but in a way I applaud them for pushing back against what I view as the crazy extent people go to in order to allow native code execution from untrused sources. I would much prefer this change be isolated to hypervisors, and have chrome/ff/etc detune their JIT's a bit to keep people from running cache timing attacks. So, I likely will be turning the kpti off on most of my machines the same way I run them with the iommu's disabled because I'm not running VM's with untrused code. (BTW, once the dust settles, i'm guessing a pretty large number of other aggressive OoO processors are vulnerable as well (old Alpha/PA-RISC/SPARC/etc)).