From mboxrd@z Thu Jan  1 00:00:00 1970
From: Jan Kiszka <jan.kiszka@siemens.com>
Subject: Re: [PATCH] KVM: vmx: fix VPID is 0000H in non-root operation
Date: Wed, 16 Sep 2015 14:32:33 +0200
Message-ID: <55F96161.2090405@siemens.com>
References: <BLU436-SMTP979E5F0855A14EAB12C4A5805B0@phx.gbl>
Mime-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: QUOTED-PRINTABLE
Cc: Bandan Das <bsd@redhat.com>, Wincy Van <fanwenyi0529@gmail.com>,
	kvm@vger.kernel.org, linux-kernel@vger.kernel.org
To: Wanpeng Li <wanpeng.li@hotmail.com>,
	Paolo Bonzini <pbonzini@redhat.com>
Return-path: <linux-kernel-owner@vger.kernel.org>
In-Reply-To: <BLU436-SMTP979E5F0855A14EAB12C4A5805B0@phx.gbl>
Sender: linux-kernel-owner@vger.kernel.org
List-Id: kvm.vger.kernel.org

On 2015-09-16 13:31, Wanpeng Li wrote:
> Reference SDM 28.1:
>=20
> The current VPID is 0000H in the following situations:
> =E2=80=94 Outside VMX operation. (This includes operation in system-m=
anagement=20
>   mode under the default treatment of SMIs and SMM with VMX operation=
;=20
>   see Section 34.14.)
> =E2=80=94 In VMX root operation.
> =E2=80=94 In VMX non-root operation when the =E2=80=9Cenable VPID=E2=80=
=9D VM-execution control=20
>   is 0.
>=20
> The VPID should never be 0000H in non-root operation when "enable VPI=
D"=20
> VM-execution control is 1. However, commit (34a1cd60: 'kvm: x86: vmx:=
=20
> move some vmx setting from vmx_init() to hardware_setup()') remove th=
e=20
> codes which reserve 0000H for VMX root operation.=20
>=20
> This patch fix it by reintroducing reserve 0000H for VMX root operati=
on.
>=20
> Reported-by: Wincy Van <fanwenyi0529@gmail.com>
> Signed-off-by: Wanpeng Li <wanpeng.li@hotmail.com>
> ---
>  arch/x86/kvm/vmx.c | 2 ++
>  1 file changed, 2 insertions(+)
>=20
> diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
> index 9ff6a3f..a63b9ca 100644
> --- a/arch/x86/kvm/vmx.c
> +++ b/arch/x86/kvm/vmx.c
> @@ -6056,6 +6056,8 @@ static __init int hardware_setup(void)
>  	memcpy(vmx_msr_bitmap_longmode_x2apic,
>  			vmx_msr_bitmap_longmode, PAGE_SIZE);
> =20
> +	set_bit(0, vmx_vpid_bitmap); /* 0 is reserved for host */
> +
>  	if (enable_apicv) {
>  		for (msr =3D 0x800; msr <=3D 0x8ff; msr++)
>  			vmx_disable_intercept_msr_read_x2apic(msr);
>=20

Good point.

BTW, what will happen if allocate_vpid runs out of free slots and
returns 0? Will we always fail then...?

Jan

--=20
Siemens AG, Corporate Technology, CT RTC ITP SES-DE
Corporate Competence Center Embedded Linux