You know, he's not wrong. This is, in impact, way bigger than Intel's FDIV fiasco and that ended up in recalls.
The question is what are they replacing the processors WITH?
Speculative execution and OOO execution and branch preduction goes back nearly 30 years (Intel Pentium), and it's likely been bugged ever since then.
And it's not like Intel will be able to fix the bug even in the next generation of processors they release this year - at the very earliest it would be 2019 or even 2020 before the bug is fixe
There's no simple solution for Spectre, as is it much more widespread and affects pretty much every modern CPU. The only viable way is some sort of software-based mitigation.
De Raadt's rant was about Meltdown though, and he's absolutely right. Meltdown is a Intel-only fuckup; someone decided that protection domains should not apply to execution speculation in order to boost performance.
"I want repaired processors for free" (Score:5, Insightful)
You know, he's not wrong. This is, in impact, way bigger than Intel's FDIV fiasco and that ended up in recalls.
Re: (Score:2)
The question is what are they replacing the processors WITH?
Speculative execution and OOO execution and branch preduction goes back nearly 30 years (Intel Pentium), and it's likely been bugged ever since then.
And it's not like Intel will be able to fix the bug even in the next generation of processors they release this year - at the very earliest it would be 2019 or even 2020 before the bug is fixe
Re:"I want repaired processors for free" (Score:3)
There's no simple solution for Spectre, as is it much more widespread and affects pretty much every modern CPU. The only viable way is some sort of software-based mitigation.
De Raadt's rant was about Meltdown though, and he's absolutely right. Meltdown is a Intel-only fuckup; someone decided that protection domains should not apply to execution speculation in order to boost performance.