linux-kernel - Re: [PATCH v7 3/3] KVM: vmx: handle vm-exit for UMWAIT and TPAUSE

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:   Fri, 12 Jul 2019 09:03:52 -0700
From:   Sean Christopherson <sean.j.christopherson@...el.com>
To:     Tao Xu <tao3.xu@...el.com>
Cc:     pbonzini@...hat.com, rkrcmar@...hat.com, corbet@....net,
        tglx@...utronix.de, mingo@...hat.com, bp@...en8.de, hpa@...or.com,
        kvm@...r.kernel.org, linux-kernel@...r.kernel.org,
        fenghua.yu@...el.com, xiaoyao.li@...ux.intel.com,
        jingqi.liu@...el.com
Subject: Re: [PATCH v7 3/3] KVM: vmx: handle vm-exit for UMWAIT and TPAUSE

On Fri, Jul 12, 2019 at 04:29:07PM +0800, Tao Xu wrote:
> --- a/arch/x86/kvm/vmx/nested.c
> +++ b/arch/x86/kvm/vmx/nested.c
> @@ -5213,6 +5213,9 @@ bool nested_vmx_exit_reflected(struct kvm_vcpu *vcpu, u32 exit_reason)
>  	case EXIT_REASON_ENCLS:
>  		/* SGX is never exposed to L1 */
>  		return false;
> +	case EXIT_REASON_UMWAIT: case EXIT_REASON_TPAUSE:

Grouped case statements are usually stacked vertically, e.g.:

	case EXIT_REASON_UMWAIT:
	case EXIT_REASON_TPAUSE:

> +		return nested_cpu_has2(vmcs12,
> +			SECONDARY_EXEC_ENABLE_USR_WAIT_PAUSE);
>  	default:
>  		return true;
>  	}
> diff --git a/arch/x86/kvm/vmx/vmx.c b/arch/x86/kvm/vmx/vmx.c
> index 0787f140d155..e026b1313dc3 100644
> --- a/arch/x86/kvm/vmx/vmx.c
> +++ b/arch/x86/kvm/vmx/vmx.c
> @@ -5349,6 +5349,20 @@ static int handle_monitor(struct kvm_vcpu *vcpu)
>  	return handle_nop(vcpu);
>  }
>  
> +static int handle_umwait(struct kvm_vcpu *vcpu)
> +{
> +	kvm_skip_emulated_instruction(vcpu);
> +	WARN(1, "this should never happen\n");

Blech.  I'm guessing this code was copy-pasted from handle_xsaves() and
handle_xrstors().  The blurb of "this should never happen" isn't very
helpful, e.g. the WARN itself makes it pretty obvious that we don't expect
to reach this point.  WARN_ONCE would also be preferable, no need to spam
the log in the event things go completely haywire.

Rather than propagate ugly code, what about defining a common helper, e.g.

static int handle_unexpected_vmexit(struct kvm_vcpu *vcpu)
{
	kvm_skip_emulated_instruction(vcpu);
	WARN_ONCE(1, "Unexpected VM-Exit = 0x%x", vmcs_read32(VM_EXIT_REASON));
	return 1;
}

...
{
	[EXIT_REASON_XSAVES]                  = handle_unexpected_vmexit,
	[EXIT_REASON_XRSTORS]                 = handle_unexpected_vmexit,

	[EXIT_REASON_UMWAIT]                  = handle_unexpected_vmexit,
	[EXIT_REASON_TPAUSE]                  = handle_unexpected_vmexit,

}

> +	return 1;
> +}
> +
> +static int handle_tpause(struct kvm_vcpu *vcpu)
> +{
> +	kvm_skip_emulated_instruction(vcpu);
> +	WARN(1, "this should never happen\n");
> +	return 1;
> +}
> +
>  static int handle_invpcid(struct kvm_vcpu *vcpu)
>  {
>  	u32 vmx_instruction_info;
> @@ -5559,6 +5573,8 @@ static int (*kvm_vmx_exit_handlers[])(struct kvm_vcpu *vcpu) = {
>  	[EXIT_REASON_VMFUNC]		      = handle_vmx_instruction,
>  	[EXIT_REASON_PREEMPTION_TIMER]	      = handle_preemption_timer,
>  	[EXIT_REASON_ENCLS]		      = handle_encls,
> +	[EXIT_REASON_UMWAIT]                  = handle_umwait,
> +	[EXIT_REASON_TPAUSE]                  = handle_tpause,
>  };
>  
>  static const int kvm_vmx_max_exit_handlers =
> -- 
> 2.20.1
>