[英]Does the VMX mode have the capability to detect previously non-trappable sensitive instructions?
VMX模式是否为Virtualisation Systems Architect提供了检测以前不可捕获的敏感指令的功能?
No, not in all cases. 不,不是在所有情况下。 Instead VMX mode changes the behavior of some of these instructions so that they are not sensitive.
相反,VMX模式会更改其中某些指令的行为,以使它们不敏感。
Here are a couple examples of how sensitive instructions behave in VMX non-root mode: 以下是一些示例,说明敏感指令在VMX非根模式下的行为:
SMSW
: it doesn't trap when in VMX mode, but it substitutes values specified by the VMM for the actual values of certain bits. SMSW
:它在VMX模式下不会陷阱,但它会将VMM指定的值替换为某些位的实际值。 POPF
: the behavior of the instruction doesn't change, but control of interrupts can be overridden by controls in the VMCS, to prevent guest changes to IF from affecting the VMM or other guests. POPF
:指令的行为不会改变,但VMCS中的控件可以覆盖对中断的控制,以防止guest虚拟机更改为影响VMM或其他guest虚拟机。 This reference has a list of sensitive x86 instructions on page 28: https://www.cs.cmu.edu/~410-s14/lectures/L30_Virtualization.pdf . 此参考有一个敏感的x86指令列表(第28页): https : //www.cs.cmu.edu/~410-s14/lectures/L30_Virtualization.pdf 。 Volume 3, chapter 25 of the Intel SDM has a detailed description of the change in behavior of every instruction.
英特尔SDM第3卷第25章详细描述了每条指令的行为变化。
I'd say yes. 我会说是的。 According to pages 13 through 17 of this Virtualization and Virtual Machines presentation , hardware virtualization seems to be an attempt to solve that problem.
根据此虚拟化和虚拟机演示文稿的第13页到第17页,硬件虚拟化似乎是尝试解决该问题。
The pdf @prl posted also says on slide 45: 发布的pdf @prl也在幻灯片45中说:
if a guest kernel runs a sensitive instruction, hardware does a 'VM exit' back to the VMM, indicates which instruction trapped
如果客户内核运行敏感指令,则硬件将“VM退出”返回VMM,指示捕获的指令
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.