linux-kernel - [PATCH v14 112/113] RFC: KVM: TDX: Make busy with S-EPT on entry bug

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:   Sun, 28 May 2023 21:20:34 -0700
From:   isaku.yamahata@...el.com
To:     kvm@...r.kernel.org, linux-kernel@...r.kernel.org
Cc:     isaku.yamahata@...el.com, isaku.yamahata@...il.com,
        Paolo Bonzini <pbonzini@...hat.com>, erdemaktas@...gle.com,
        Sean Christopherson <seanjc@...gle.com>,
        Sagi Shahar <sagis@...gle.com>,
        David Matlack <dmatlack@...gle.com>,
        Kai Huang <kai.huang@...el.com>,
        Zhi Wang <zhi.wang.linux@...il.com>, chen.bo@...el.com
Subject: [PATCH v14 112/113] RFC: KVM: TDX: Make busy with S-EPT on entry bug

From: Isaku Yamahata <isaku.yamahata@...el.com>

TDX module has mitigation against zero-step attacks or single-step attacks.
When the TDX module finds repeated EPT violations on the same guest RIP,
i.e. no advance in guest, it starts to suspect the attack.  The mitigation
logic on the next entry tries to take the lock of S-EPT.  It may result in
an error of TDX_OPERAND_BUSY | TDX_OPERAND_ID_SEPT.  As KVM shouldn't
spuriously zap private S-EPT so that guest can make progress, KVM
shouldn't cause the TDX module to trigger the mitigation.  Make
(TDX_OPERAND_BUSY | TDX_OPERAND_ID_SEPT) on entry KVM bug.

Suggested-by: Sean Christopherson <seanjc@...gle.com>
Signed-off-by: Isaku Yamahata <isaku.yamahata@...el.com>
---
- This patch is RFC because this is only lightly tested and stress test
  isn't done.
---
 arch/x86/kvm/vmx/tdx.c | 16 ++++++++++++++--
 1 file changed, 14 insertions(+), 2 deletions(-)

diff --git a/arch/x86/kvm/vmx/tdx.c b/arch/x86/kvm/vmx/tdx.c
index 12d6c9cacf6a..885fb1a7e324 100644
--- a/arch/x86/kvm/vmx/tdx.c
+++ b/arch/x86/kvm/vmx/tdx.c
@@ -1702,8 +1702,20 @@ int tdx_handle_exit(struct kvm_vcpu *vcpu, fastpath_t fastpath)
 {
 	union tdx_exit_reason exit_reason = to_tdx(vcpu)->exit_reason;
 
-	/* See the comment of tdh_sept_seamcall(). */
-	if (unlikely(exit_reason.full == (TDX_OPERAND_BUSY | TDX_OPERAND_ID_SEPT)))
+	/*
+	 * See the comment of tdh_sept_seamcall().
+	 * TDX module has mitigation against zero-step attacks or single-step
+	 * attacks.  When the TDX module finds repeated EPT violations on the
+	 * same guest RIP, i.e. no advance in guest, it starts to suspect the
+	 * attack.  The mitigation logic on the next entry tries to take the
+	 * lock of S-EPT.  It may result in an error of (TDX_OPERAND_BUSY |
+	 * TDX_OPERAND_ID_SEPT).  As KVM shouldn't spuriously zap private S-EPT
+	 * so that guest can make progress, KVM shouldn't cause the TDX module
+	 * to trigger the mitigation.  Make (TDX_OPERAND_BUSY |
+	 * TDX_OPERAND_ID_SEPT) on entry KVM bug.
+	 */
+	if (KVM_BUG_ON(exit_reason.full == (TDX_OPERAND_BUSY | TDX_OPERAND_ID_SEPT),
+		       vcpu->kvm))
 		return 1;
 
 	/*
-- 
2.25.1