Received-SPF: pass (google.com: domain of linux-kernel+bounces-72034-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.80.249 as permitted sender) client-ip=147.75.80.249;
Date: Mon, 19 Feb 2024 13:53:07 -0800
In-Reply-To: <20240215152916.1158-13-paul@xen.org>
Precedence: bulk
Mime-Version: 1.0
References: <20240215152916.1158-1-paul@xen.org> <20240215152916.1158-13-paul@xen.org>
Message-ID: <ZdPNw9evnWw2JzG-@google.com>
Subject: Re: [PATCH v13 12/21] KVM: x86/xen: allow shared_info to be mapped by
 fixed HVA
From: Sean Christopherson <seanjc@google.com>
To: Paul Durrant <paul@xen.org>
Cc: Paolo Bonzini <pbonzini@redhat.com>, Jonathan Corbet <corbet@lwn.net>, 
	Christian Borntraeger <borntraeger@linux.ibm.com>, Janosch Frank <frankja@linux.ibm.com>, 
	Claudio Imbrenda <imbrenda@linux.ibm.com>, David Hildenbrand <david@redhat.com>, 
	Heiko Carstens <hca@linux.ibm.com>, Vasily Gorbik <gor@linux.ibm.com>, 
	Alexander Gordeev <agordeev@linux.ibm.com>, Sven Schnelle <svens@linux.ibm.com>, 
	Thomas Gleixner <tglx@linutronix.de>, Ingo Molnar <mingo@redhat.com>, Borislav Petkov <bp@alien8.de>, 
	Dave Hansen <dave.hansen@linux.intel.com>, x86@kernel.org, 
	"H. Peter Anvin" <hpa@zytor.com>, David Woodhouse <dwmw2@infradead.org>, Shuah Khan <shuah@kernel.org>, 
	kvm@vger.kernel.org, linux-doc@vger.kernel.org, linux-kernel@vger.kernel.org, 
	linux-s390@vger.kernel.org, linux-kselftest@vger.kernel.org
Content-Type: text/plain; charset="us-ascii"

On Thu, Feb 15, 2024, Paul Durrant wrote:
> @@ -715,13 +731,23 @@ int kvm_xen_hvm_get_attr(struct kvm *kvm, struct kvm_xen_hvm_attr *data)
>  		break;
>  
>  	case KVM_XEN_ATTR_TYPE_SHARED_INFO:
> -		if (kvm->arch.xen.shinfo_cache.active)
> +		if (kvm->arch.xen.shinfo_cache.active &&
> +		    kvm->arch.xen.shinfo_cache.gpa != KVM_XEN_INVALID_GPA)

This should really use INVALID_GPA when checking internal gpc state.  Mostly to
help clarify what is/isn't KVM Xen ABI, but also because I don't like the subtle
assumption that KVM_XEN_INVALID_GPA == INVALID_GPA.

Even better, if we slot in two helpers when the HVA-based GPC support is added,
then the Xen code doesn't need to to make assumptions about how the GPC code
manages HVA vs. GPA internally.  E.g. if we ever refactor the code again and use
a dedicated flag instead of gpc->gpa as the implicit flag.

  static inline bool kvm_gpc_is_gpa_active(struct gfn_to_pfn_cache *gpc)
  {
	return gpc->active && !kvm_is_error_gpa(gpc->gpa);
  }

  static inline bool kvm_gpc_is_hva_active(struct gfn_to_pfn_cache *gpc)
  {
	return gpc->active && kvm_is_error_gpa(gpc->gpa);
  }