On 2013-05-02 08:59, Xu, Xinhao wrote: > Hi, Jan > Can you provide details of your test environment? Pretty simple: Run a Linux 3.9(-rc5) kernel as L2 on top of the same kernel with QEMU (git head at that time). I've attached the config. Jan > > -----Original Message----- > From: Jan Kiszka [mailto:jan.kiszka@xxxxxx] > Sent: Sunday, April 28, 2013 5:36 PM > To: Xu, Xinhao > Cc: Nakajima, Jun; kvm@xxxxxxxxxxxxxxx > Subject: Re: [PATCH 11/11] nEPT: Provide the correct exit qualification upon EPT > > On 2013-04-27 08:42, Xu, Xinhao wrote: >> From 2df72c1e8e3b167a6008ab11e2a68d734c68e425 Mon Sep 17 00:00:00 2001 >> From: Xinhao Xu <xinhao.xu@xxxxxxxxx> >> Date: Sat, 27 Apr 2013 05:45:49 +0800 >> Subject: [PATCH] nEPT: Check EPT misconfiguration while walking addr & >> move pte check code to vmx.c > > Please fix up the subject when posting a patch in reply to another one. > >> >> I add code to detect EPT misconfiguration and inject it to L1. >> Now L1 can correctly go to ept_misconfig handler(instead of wrongly >> going to fast_page_fault), it will try to handle mmio page fault, if >> failed, it is a real ept misconfiguration. >> >> For scalability, Xiantao suggests me moving vendor specific code out >> from common code. In order to do this, I add new ops in kvm_mmu >> struct, check_tdp_pte, to provide an interface to check fault while >> walking address. The better way is to do fault detecting at here, but >> so far I just have checked ept misconfiguration. More patches will be >> added in future. > > Seems there are some issues remaining. I can boot Linux as L2 when I remove this patch. When it's applied, L2 becomes pretty slow and eventually resets during kernel boot of L2. L1 remains stable. > > Jan > > -- Siemens AG, Corporate Technology, CT RTC ITP SDP-DE Corporate Competence Center Embedded Linux
Attachment:
.config.bz2
Description: application/bzip
