Re: [PATCH 0/3] KVM: x86: Fixes for kvm/queue

[Sean Christopherson]

On Fri, Feb 11, 2022, Paolo Bonzini wrote:
> On 2/11/22 17:31, Sean Christopherson wrote:
> > > Maybe the patch "Revert "KVM: VMX: Save HOST_CR3 in
> > > vmx_prepare_switch_to_guest()"" is still missing in the latest
> > > kvm/queue, I saw the same warning.
> > 
> > It hasn't made it way to Linus either.
> 
> This was supposed to fix the buggy patch, too:
> 
>     commit a9f2705ec84449e3b8d70c804766f8e97e23080d
>     Author: Lai Jiangshan <laijs@xxxxxxxxxxxxxxxxx>
>     Date:   Thu Dec 16 10:19:36 2021 +0800
> 
>     KVM: VMX: Save HOST_CR3 in vmx_set_host_fs_gs()
>     The host CR3 in the vcpu thread can only be changed when scheduling,
>     so commit 15ad9762d69f ("KVM: VMX: Save HOST_CR3 in vmx_prepare_switch_to_guest()")
>     changed vmx.c to only save it in vmx_prepare_switch_to_guest().
>     However, it also has to be synced in vmx_sync_vmcs_host_state() when switching VMCS.
>     vmx_set_host_fs_gs() is called in both places, so rename it to
>     vmx_set_vmcs_host_state() and make it update HOST_CR3.
>     Fixes: 15ad9762d69f ("KVM: VMX: Save HOST_CR3 in vmx_prepare_switch_to_guest()")
>     Signed-off-by: Lai Jiangshan <laijs@xxxxxxxxxxxxxxxxx>
>     Message-Id: <20211216021938.11752-2-jiangshanlai@xxxxxxxxx>
>     Signed-off-by: Paolo Bonzini <pbonzini@xxxxxxxxxx>

The underlying premise that CR3 can change only when scheduling is wrong, reverts
incoming...