linux-kernel - [RFC 08/12] KVM/VMX: Use the new host mapping API for mapping nested PML

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]

Message-Id: <1517856451-2932-9-git-send-email-karahmed@amazon.de>
Date:   Mon,  5 Feb 2018 19:47:27 +0100
From:   KarimAllah Ahmed <karahmed@...zon.de>
To:     linux-kernel@...r.kernel.org, kvm@...r.kernel.org
Cc:     KarimAllah Ahmed <karahmed@...zon.de>,
        Paolo Bonzini <pbonzini@...hat.com>,
        Radim Krčmář <rkrcmar@...hat.com>
Subject: [RFC 08/12] KVM/VMX: Use the new host mapping API for mapping nested PML

For nested guests the PML page was mapped to the host kernel using
kvm_vcpu_gpa_to_page which assumes that all guest memory is backed by a
"struct page". This breaks guests that have their memory outside the kernel
control.

Switch to the new host mapping API which takes care of this use-case as
well.

Cc: Paolo Bonzini <pbonzini@...hat.com>
Cc: Radim Krčmář <rkrcmar@...hat.com>
Cc: kvm@...r.kernel.org
Cc: linux-kernel@...r.kernel.org
Signed-off-by: KarimAllah Ahmed <karahmed@...zon.de>
---
 arch/x86/kvm/vmx.c | 11 +++++------
 1 file changed, 5 insertions(+), 6 deletions(-)

diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
index 7177176..620b4ff 100644
--- a/arch/x86/kvm/vmx.c
+++ b/arch/x86/kvm/vmx.c
@@ -11866,7 +11866,7 @@ static int vmx_write_pml_buffer(struct kvm_vcpu *vcpu)
 	struct vmcs12 *vmcs12;
 	struct vcpu_vmx *vmx = to_vmx(vcpu);
 	gpa_t gpa;
-	struct page *page = NULL;
+	struct kvm_host_mapping mapping;
 	u64 *pml_address;
 
 	if (is_guest_mode(vcpu)) {
@@ -11888,14 +11888,13 @@ static int vmx_write_pml_buffer(struct kvm_vcpu *vcpu)
 
 		gpa = vmcs_read64(GUEST_PHYSICAL_ADDRESS) & ~0xFFFull;
 
-		page = kvm_vcpu_gpa_to_page(vcpu, vmcs12->pml_address);
-		if (is_error_page(page))
+		if (!kvm_vcpu_gpa_to_host_mapping(vcpu, vmcs12->pml_address, &mapping, true))
 			return 0;
 
-		pml_address = kmap(page);
+		pml_address = mapping.kaddr;
 		pml_address[vmcs12->guest_pml_index--] = gpa;
-		kunmap(page);
-		kvm_release_page_clean(page);
+
+		kvm_release_host_mapping(&mapping, false);
 	}
 
 	return 0;
-- 
2.7.4