linux-kernel - Re: [PATCH 0/3] KVM: x86: Clean up RESET "emulation"

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: <YUTRwNT/O5Ny0MOQ@google.com>
Date:   Fri, 17 Sep 2021 17:34:56 +0000
From:   Sean Christopherson <seanjc@...gle.com>
To:     Paolo Bonzini <pbonzini@...hat.com>
Cc:     Vitaly Kuznetsov <vkuznets@...hat.com>,
        Wanpeng Li <wanpengli@...cent.com>,
        Jim Mattson <jmattson@...gle.com>,
        Joerg Roedel <joro@...tes.org>, kvm <kvm@...r.kernel.org>,
        "Kernel Mailing List, Linux" <linux-kernel@...r.kernel.org>,
        Reiji Watanabe <reijiw@...gle.com>
Subject: Re: [PATCH 0/3] KVM: x86: Clean up RESET "emulation"

On Fri, Sep 17, 2021, Paolo Bonzini wrote:
> On Wed, Sep 15, 2021 at 1:08 AM Sean Christopherson <seanjc@...gle.com> wrote:
> > Add dedicated helpers to emulate RESET instead of having the relevant code
> > scattered through vcpu_create() and vcpu_reset().  Paolo, I think this is
> > what you meant by "have init_vmcb/svm_vcpu_reset look more like the VMX
> > code"[*].
> >
> > [*] https://lore.kernel.org/all/c3563870-62c3-897d-3148-e48bb755310c@redhat.com/
> 
> That assumes that I remember what I meant :)

Ha!  That's why I write changelogs with --verbose :-)

> but I do like it so yes, that was it. Especially the fact that init_vmcb now
> has a single caller. I would further consider moving save area initialization
> to *_vcpu_reset, and keeping the control fields in init_vmcb/vmcs. That would
> make it easier to relate the two functions to separate parts of the manuals.

I like the idea, but I think I'd prefer to tackle that at the same time as generic
support for handling MSRs at RESET/INIT.  E.g. instead of manually writing
vmcs.GUEST_SYSENTER_* at RESET, provide infrastruture to automagically run through
all emulated/virtualized at RESET and/or INIT as appropriate to initialize the
guest value.