linux-kernel - Re: [PATCH 17/30] KVM: nSVM: synchronize VMCB controls updated by the processor on every vmexit

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]

Date:   Sat, 30 May 2020 07:10:44 +0200
From:   Paolo Bonzini <pbonzini@...hat.com>
To:     Krish Sadhukhan <krish.sadhukhan@...cle.com>,
        linux-kernel@...r.kernel.org, kvm@...r.kernel.org
Subject: Re: [PATCH 17/30] KVM: nSVM: synchronize VMCB controls updated by the
 processor on every vmexit

On 30/05/20 04:06, Krish Sadhukhan wrote:
>>
>>   -    nested_vmcb->control.int_ctl           = vmcb->control.int_ctl;
>> -    nested_vmcb->control.int_vector        = vmcb->control.int_vector;
> 
> 
> While it's not related to this patch, I am wondering if we need the
> following line in enter_svm_guest_mode():
> 
>     svm->vmcb->control.int_vector = nested_vmcb->control.int_vector;
> 
> If int_vector is used only to trigger a #VMEXIT after we have decided to
> inject a virtual interrupt, we probably don't the vector value from the
> nested vmcb.

KVM does not use int_vector, but other hypervisor sometimes trigger
virtual interrupts using int_ctl+int_vector instead of eventinj.
(Unlike KVM they don't intercept EXIT_VINTR).  Hyper-V operates like that.

Paolo