linux-kernel - Re: [PATCH] KVM: VMX: CPU frequency scaling for intel x86

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

Message-ID: <YpYaYK7a28DFT5Ne@hirez.programming.kicks-ass.net>
Date:   Tue, 31 May 2022 15:38:40 +0200
From:   Peter Zijlstra <peterz@...radead.org>
To:     Jack Allister <jalliste@...zon.com>
Cc:     diapop@...zon.co.uk, metikaya@...zon.co.uk,
        Paolo Bonzini <pbonzini@...hat.com>,
        Radim Krčmář <rkrcmar@...hat.com>,
        Sean Christopherson <sean.j.christopherson@...el.com>,
        Vitaly Kuznetsov <vkuznets@...hat.com>,
        Wanpeng Li <wanpengli@...cent.com>,
        Jim Mattson <jmattson@...gle.com>,
        Joerg Roedel <joro@...tes.org>,
        Thomas Gleixner <tglx@...utronix.de>,
        Ingo Molnar <mingo@...hat.com>, Borislav Petkov <bp@...en8.de>,
        "H. Peter Anvin" <hpa@...or.com>, x86@...nel.org,
        kvm@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] KVM: VMX: CPU frequency scaling for intel x86_64 KVM
 guests

On Tue, May 31, 2022 at 10:59:25AM +0000, Jack Allister wrote:
> A VMM can control a vCPU's CPU frequency by interfacing with KVM via
> the vCPU file descriptor to enable/set CPU frequency scaling for a
> guest. Instead of creating a separate IOCTL to this this, KVM capabil-
> ities are extended to include a capability called
> KVM_CAP_CPU_FREQ_SCALING.
> 
> A generic set_cpu_freq interface is added to kvm_x86_ops
> to allow for architecture (AMD/Intel) independent CPU frequency
> scaling setting.
> 
> For Intel platforms, Hardware-Controlled Performance States (HWP) are
> used to implement CPU scaling within the guest. Further information on
> this mechanism can be seen in Intel SDM Vol 3B (section 14.4). The CPU
> frequency is set as soon as this function is called and is kept running
> until explicitly reset or set again.
> 
> Currently the AMD frequency setting interface is left unimplemented.
> 
> Please note that CPU frequency scaling will have an effect on host
> processing in it's current form. To change back to full performance
> when running in host context an IOCTL with a frequency value of 0
> is needed to run back at uncapped speed.

Nowhere does this explain *WHY* we would want to do this?