linux-kernel - Re: [PATCH 11/12] perf intel-pt: Add VM Time Correlation to decoder

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Date:   Fri, 30 Apr 2021 08:02:14 +0300
From:   Adrian Hunter <adrian.hunter@...el.com>
To:     Andi Kleen <ak@...ux.intel.com>
Cc:     Arnaldo Carvalho de Melo <acme@...nel.org>,
        Jiri Olsa <jolsa@...hat.com>, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 11/12] perf intel-pt: Add VM Time Correlation to decoder

On 30/04/21 12:21 am, Andi Kleen wrote:
>> unchanging (during perf record) VMX TSC Offset and no VMX TSC scaling.
> 
> For TSC scaling we would need to add side band to KVM, correct?
> 
> I suppose it shouldn't be that hard to do that.
> 

Probably we just need to know a single TSC scaling value per VM i.e. it could just be exported via sysfs or something.

There is a reasonable reason that TSC Offset might be changing i.e. to match Guest time, although that seems to be done by paravirtualization instead at the moment.
At the moment TSC Offset does change if Host time changes e.g. Guest has state saved before a Host reboot, and is resumed after the Host reboot

However, I can't think of a reason for changing the TSC Scaling value, except maybe if the VM is taken to a Host with a different TSC frequency.