Received-SPF: pass (google.com: domain of linux-kernel+bounces-126911-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:40f1:3f00::1 as permitted sender) client-ip=2604:1380:40f1:3f00::1;
Date: Mon, 1 Apr 2024 10:34:05 -0700
In-Reply-To: <f6a80dd212e8c3fd14b40049eed33187008cf35a.1708933498.git.isaku.yamahata@intel.com>
Precedence: bulk
Mime-Version: 1.0
References: <cover.1708933498.git.isaku.yamahata@intel.com> <f6a80dd212e8c3fd14b40049eed33187008cf35a.1708933498.git.isaku.yamahata@intel.com>
Message-ID: <ZgrwDbZLFqiyS_5e@google.com>
Subject: Re: [PATCH v19 069/130] KVM: TDX: Require TDP MMU and mmio caching
 for TDX
From: Sean Christopherson <seanjc@google.com>
To: isaku.yamahata@intel.com
Cc: kvm@vger.kernel.org, linux-kernel@vger.kernel.org, 
	isaku.yamahata@gmail.com, Paolo Bonzini <pbonzini@redhat.com>, erdemaktas@google.com, 
	Sagi Shahar <sagis@google.com>, Kai Huang <kai.huang@intel.com>, chen.bo@intel.com, 
	hang.yuan@intel.com, tina.zhang@intel.com
Content-Type: text/plain; charset="us-ascii"

On Mon, Feb 26, 2024, isaku.yamahata@intel.com wrote:
> From: Isaku Yamahata <isaku.yamahata@intel.com>
> 
> As TDP MMU is becoming main stream than the legacy MMU, the legacy MMU
> support for TDX isn't implemented.  TDX requires KVM mmio caching.  Disable
> TDX support when TDP MMU or mmio caching aren't supported.
> 
> Signed-off-by: Isaku Yamahata <isaku.yamahata@intel.com>
> ---
>  arch/x86/kvm/mmu/mmu.c  |  1 +
>  arch/x86/kvm/vmx/main.c | 13 +++++++++++++
>  2 files changed, 14 insertions(+)
> 
> diff --git a/arch/x86/kvm/mmu/mmu.c b/arch/x86/kvm/mmu/mmu.c
> index 0e0321ad9ca2..b8d6ce02e66d 100644
> --- a/arch/x86/kvm/mmu/mmu.c
> +++ b/arch/x86/kvm/mmu/mmu.c
> @@ -104,6 +104,7 @@ module_param_named(flush_on_reuse, force_flush_and_sync_on_reuse, bool, 0644);
>   * If the hardware supports that we don't need to do shadow paging.
>   */
>  bool tdp_enabled = false;
> +EXPORT_SYMBOL_GPL(tdp_enabled);

I haven't looked at the rest of the series, but this should be unnecessary.  Just
use enable_ept.  Ah, the code is wrong.

>  static bool __ro_after_init tdp_mmu_allowed;
>  
> diff --git a/arch/x86/kvm/vmx/main.c b/arch/x86/kvm/vmx/main.c
> index 076a471d9aea..54df6653193e 100644
> --- a/arch/x86/kvm/vmx/main.c
> +++ b/arch/x86/kvm/vmx/main.c
> @@ -3,6 +3,7 @@
>  
>  #include "x86_ops.h"
>  #include "vmx.h"
> +#include "mmu.h"
>  #include "nested.h"
>  #include "pmu.h"
>  #include "tdx.h"
> @@ -36,6 +37,18 @@ static __init int vt_hardware_setup(void)
>  	if (ret)
>  		return ret;
>  
> +	/* TDX requires KVM TDP MMU. */

This is a useless comment (assuming the code is correctly written), it's quite
obvious from:

	if (!tdp_mmu_enabled)
		enable_tdx = false;

that the TDP MMU is required.  Explaining *why* could be useful, but I'd probably
just omit a comment.  In the not too distant future, it will likely be common
knowledge that the shadow MMU doesn't support newfangled features, and it _should_
be very easy for someone to get the info from the changelog.

> +	if (enable_tdx && !tdp_enabled) {

tdp_enabled can be true without the TDP MMU, you want tdp_mmu_enabled.

> +		enable_tdx = false;
> +		pr_warn_ratelimited("TDX requires TDP MMU.  Please enable TDP MMU for TDX.\n");

Drop the pr_warn(), TDX will presumably be on by default at some point, I don't
want to get spam every time I test with TDP disabled.

Also, ratelimiting this code is pointless (as is _once()), it should only ever
be called once per module module, and the limiting/once protections are tied to
the module, i.e. effectively get reset when a module is reloaded.

> +	}
> +
> +	/* TDX requires MMIO caching. */
> +	if (enable_tdx && !enable_mmio_caching) {
> +		enable_tdx = false;
> +		pr_warn_ratelimited("TDX requires mmio caching.  Please enable mmio caching for TDX.\n");

Same comments here.

> +	}
> +
>  	enable_tdx = enable_tdx && !tdx_hardware_setup(&vt_x86_ops);

All of the above code belongs in tdx_hardware_setup(), especially since you need
tdp_mmu_enabled, not enable_ept.

>  	if (enable_tdx)
>  		vt_x86_ops.vm_size = max_t(unsigned int, vt_x86_ops.vm_size,
> -- 
> 2.25.1
>