linux-kernel - Re: [PATCH] KVM: x86: Don't snapshot "max" TSC if host TSC is constant

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: <7086443d5e1e21d72a3d5c386c16f0c07d37a0a8.camel@infradead.org>
Date:   Fri, 25 Feb 2022 12:20:56 +0000
From:   David Woodhouse <dwmw2@...radead.org>
To:     Paolo Bonzini <pbonzini@...hat.com>,
        Sean Christopherson <seanjc@...gle.com>
Cc:     Vitaly Kuznetsov <vkuznets@...hat.com>,
        Wanpeng Li <wanpengli@...cent.com>,
        Jim Mattson <jmattson@...gle.com>,
        Joerg Roedel <joro@...tes.org>, kvm@...r.kernel.org,
        linux-kernel@...r.kernel.org,
        Suleiman Souhlal <suleiman@...gle.com>,
        Anton Romanov <romanton@...gle.com>
Subject: Re: [PATCH] KVM: x86: Don't snapshot "max" TSC if host TSC is
 constant

On Fri, 2022-02-25 at 13:10 +0100, Paolo Bonzini wrote:
> 
> Queued, but I'd rather have a subject that calls out that max_tsc_khz 
> needs a replacement at vCPU creation time.  In fact, the real change 
> (and bug, and fix) is in kvm_arch_vcpu_create(), while the subject 
> mentions only the change in kvm_timer_init().

In
https://lore.kernel.org/kvm/e7be32b06676c7ebf415d9deea5faf50aa8c0785.camel@infradead.org/T/
last night I was coming round to the idea that we might want a KVM-wide 
default frequency which is settable from userspace and is used instead
of max_tsc_khz anyway.

I also have questions about the use case for the above patch.... if
this is a clean boot and you're just starting to host guests, surely we
can wait for the time it takes for the TSC synchronization to complete?

And if this is a live update scenario, where we pause the guests, kexec
into a new kernel, then resume the "migrated" guests again... why in
$DEITY's name isn't the precise TSC frequency being handed over from
kernel#1 to kernel#2 over the kexec so that it's known from the start?

Download attachment "smime.p7s" of type "application/pkcs7-signature" (5965 bytes)