News for hackers. Stuff that matters.

A comment on Linux page table isolation is not needed on AMD processors. Parent comment

Re:Linux page table isolation is not needed on AMD processors by BeeOnRope on Tuesday January 02, 2018 @04:58PM (#16056149)
Yup, I think so.

In that case (PCID hardware on a PCID-enabled kernel), the performance effect should be more limited to the syscall itself. That said, why is the hit still so big with PCID? Surely just the CR3-swap by itself shouldn't be so slow?

[ Parent | on HN ]
Re:Linux page table isolation is not needed on AMD processors by bonzini on Wednesday January 03, 2018 @01:07PM (#16063364)
MOV-to-CR3 is pretty slow, yes. In the ballpark of a hundred clock cycles, and you have to do 2 of them. The cost of a system call was about 1000 cycles, maybe less on newer processors---both OSes and processors optimize the hell out of SYSCALL/SYSRET.
[ Parent | on HN ]