linux-kernel - [PATCH kernel v5] x86/sev: Do not handle #VC for DR7 read/write

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-next>] [day] [month] [year] [list]

Date:   Wed, 16 Aug 2023 12:21:22 +1000
From:   Alexey Kardashevskiy <aik@....com>
To:     <x86@...nel.org>
CC:     <linux-kernel@...r.kernel.org>, Borislav Petkov <bp@...en8.de>,
        "Alexey Kardashevskiy" <aik@....com>,
        Carlos Bilbao <carlos.bilbao@....com>,
        "Tom Lendacky" <thomas.lendacky@....com>,
        Pankaj Gupta <pankaj.gupta@....com>
Subject: [PATCH kernel v5] x86/sev: Do not handle #VC for DR7 read/write

With MSR_AMD64_SEV_DEBUG_SWAP enabled, the guest is not expected to
receive a #VC for reads or writes of DR7.

Update the SNP_FEATURES_PRESENT mask with MSR_AMD64_SNP_DEBUG_SWAP so
an SNP guest doesn't gracefully terminate during SNP feature negotiation
if MSR_AMD64_SEV_DEBUG_SWAP is enabled.

Since a guest is not expected to receive a #VC on DR7 accesses when
MSR_AMD64_SEV_DEBUG_SWAP is enabled, return an error from the #VC
handler in this situation.

Signed-off-by: Alexey Kardashevskiy <aik@....com>
Reviewed-by: Carlos Bilbao <carlos.bilbao@....com>
Reviewed-by: Tom Lendacky <thomas.lendacky@....com>
Reviewed-by: Pankaj Gupta <pankaj.gupta@....com>
---

The KVM part of DebugSwap went to kvm-x86 tree:
https://lore.kernel.org/all/169058576410.1024559.1052772292093755719.b4-ty@google.com/
but it is been suggested that this one can go via TIP:
https://lore.kernel.org/all/ZGv+Y60qvbjeTY57@google.com/
I am trying to bribe now by reviewing patches (internally), does this
work? :) Thanks,


---
Changes:
v5:
* rebased on latest TIP

v4:
* rebased on top of SNP feature negotiation

v2:
* use new bit definition
---
 arch/x86/boot/compressed/sev.c | 2 +-
 arch/x86/kernel/sev.c          | 6 ++++++
 2 files changed, 7 insertions(+), 1 deletion(-)

diff --git a/arch/x86/boot/compressed/sev.c b/arch/x86/boot/compressed/sev.c
index 199155b8af3b..dc8c876fbd8f 100644
--- a/arch/x86/boot/compressed/sev.c
+++ b/arch/x86/boot/compressed/sev.c
@@ -365,7 +365,7 @@ static void enforce_vmpl0(void)
  * by the guest kernel. As and when a new feature is implemented in the
  * guest kernel, a corresponding bit should be added to the mask.
  */
-#define SNP_FEATURES_PRESENT (0)
+#define SNP_FEATURES_PRESENT	MSR_AMD64_SNP_DEBUG_SWAP
 
 u64 snp_get_unsupported_features(u64 status)
 {
diff --git a/arch/x86/kernel/sev.c b/arch/x86/kernel/sev.c
index a2b50ae61b72..2787826d9f60 100644
--- a/arch/x86/kernel/sev.c
+++ b/arch/x86/kernel/sev.c
@@ -1575,6 +1575,9 @@ static enum es_result vc_handle_dr7_write(struct ghcb *ghcb,
 	long val, *reg = vc_insn_get_rm(ctxt);
 	enum es_result ret;
 
+	if (sev_status & MSR_AMD64_SNP_DEBUG_SWAP)
+		return ES_VMM_ERROR;
+
 	if (!reg)
 		return ES_DECODE_FAILED;
 
@@ -1612,6 +1615,9 @@ static enum es_result vc_handle_dr7_read(struct ghcb *ghcb,
 	struct sev_es_runtime_data *data = this_cpu_read(runtime_data);
 	long *reg = vc_insn_get_rm(ctxt);
 
+	if (sev_status & MSR_AMD64_SNP_DEBUG_SWAP)
+		return ES_VMM_ERROR;
+
 	if (!reg)
 		return ES_DECODE_FAILED;
 
-- 
2.41.0