Skip to contentSkip to conversation

Add Secure TSC support for SNP guests

PATCH Superseded Nikunj A Dadhania 2 participants 24 messages Aug 14, 2023 – Oct 12, 2023

Subsystems (4)

v4 of 16 Superseded Latest is v16 →

lore
Share & tools
lore
#0 lore raw

v4 cover Add Secure TSC support for SNP guests

Secure TSC allows guests to securely use RDTSC/RDTSCP instructions as the parameters being used cannot be changed by hypervisor once the guest is launched. More details in the AMD64 APM Vol 2, Section "Secure TSC".

During the boot-up of the secondary cpus, SecureTSC enabled guests need to query TSC info from AMD Security Processor. This communication channel is encrypted between the AMD Security Processor and the guest, the hypervisor is just the conduit to deliver the guest messages to the AMD Security Processor. Each message is protected with an AEAD (AES-256 GCM). See "SEV Secure Nested Paging Firmware ABI Specification" document (currently at
https://www.amd.com/system/files/TechDocs/56860.pdf) section "TSC Info"

Use a minimal GCM library to encrypt/decrypt SNP Guest messages to communicate with the AMD Security Processor which is available at early boot.

SEV-guest driver has the implementation for guest and AMD Security Processor communication. As the TSC_INFO needs to be initialized during early boot before smp cpus are started, move most of the sev-guest driver code to kernel/sev.c and provide well defined APIs to the sev-guest driver to use the interface to avoid code-duplication.

Patches:
01-07: Preparation and movement of sev-guest driver code 08-14: SecureTSC enablement patches.

Testing SecureTSC
-----------------

SecureTSC hypervisor patches based on top of SEV-SNP UPM series:
https://github.com/nikunjad/linux/tree/snp-host-latest-securetsc

QEMU changes:
https://github.com/nikunjad/qemu/tree/snp_securetsc

QEMU commandline SEV-SNP-UPM with SecureTSC:

  qemu-system-x86_64 -cpu EPYC-Milan-v2,+secure-tsc \
    -object memory-backend-memfd-private,id=ram1,size=1G,share=true \
    -object sev-snp-guest,id=sev0,cbitpos=51,reduced-phys-bits=1,secure-tsc=on \
    -machine q35,confidential-guest-support=sev0,memory-backend=ram1,kvm-type=protected \
    ...

Changelog:
----------
v4:
* Drop handle_guest_request() and handle_guest_request_ext()
* Drop NULL check for key
* Corrected commit subject
* Added Reviewed-by from Tom

v3:
* Updated commit messages
* Made snp_setup_psp_messaging() generic that is accessed by both the kernel and the driver
* Moved most of the context information to sev.c, sev-guest driver does not need to know the secrets page layout anymore
* Add CC_ATTR_GUEST_SECURE_TSC early in the series therefore it can be used in later patches.
* Removed data_gpa and data_npages from struct snp_req_data, as certs_data and its size is passed to handle_guest_request_ext()
* Make vmpck_id as unsigned int
* Dropped unnecessary usage of memzero_explicit()
* Cache secrets_pa instead of remapping the cc_blob always
* Rebase on top of v6.4 kernel
https://lore.kernel.org/lkml/[email protected]/

v2:
* Rebased on top of v6.3-rc3 that has Boris's sev-guest cleanup series
https://lore.kernel.org/r/[email protected]/

v1: https://lore.kernel.org/r/[email protected]/

Nikunj A Dadhania (14):
virt: sev-guest: Use AES GCM crypto library
virt: sev-guest: Move mutex to SNP guest device structure
virt: sev-guest: Replace dev_dbg with pr_debug
virt: sev-guest: Add SNP guest request structure
virt: sev-guest: Add vmpck_id to snp_guest_dev struct x86/sev: Cache the secrets page address
x86/sev: Move and reorganize sev guest request api x86/mm: Add generic guest initialization hook
x86/sev: Add Secure TSC support for SNP guests x86/sev: Change TSC MSR behavior for Secure TSC enabled guests x86/sev: Prevent RDTSC/RDTSCP interception for Secure TSC enabled
guests
x86/kvmclock: Skip kvmclock when Secure TSC is available x86/tsc: Mark Secure TSC as reliable clocksource x86/sev: Enable Secure TSC for SNP guests

 arch/x86/Kconfig                        |   1 +
 arch/x86/boot/compressed/sev.c          |   2 +-
 arch/x86/coco/core.c                    |   3 +
 arch/x86/include/asm/sev-guest.h        | 175 +++++++
 arch/x86/include/asm/sev.h              |  19 +-
 arch/x86/include/asm/svm.h              |   6 +-
 arch/x86/include/asm/x86_init.h         |   2 +
 arch/x86/kernel/kvmclock.c              |   2 +-
 arch/x86/kernel/sev-shared.c            |   7 +
 arch/x86/kernel/sev.c                   | 631 +++++++++++++++++++++--
 arch/x86/kernel/tsc.c                   |   2 +-
 arch/x86/kernel/x86_init.c              |   2 +
 arch/x86/mm/mem_encrypt.c               |  13 +-
 arch/x86/mm/mem_encrypt_amd.c           |   6 +
 drivers/virt/coco/sev-guest/Kconfig     |   2 -
 drivers/virt/coco/sev-guest/sev-guest.c | 650 +++---------------------
 drivers/virt/coco/sev-guest/sev-guest.h |  63 ---
 include/linux/cc_platform.h             |   8 +
 18 files changed, 881 insertions(+), 713 deletions(-)
 create mode 100644 arch/x86/include/asm/sev-guest.h
 delete mode 100644 drivers/virt/coco/sev-guest/sev-guest.h

base-commit: 6995e2de6891c724bfeb2db33d7b87775f913ad1

Signature · 1 line
2.34.1
#1 lore raw

PATCH v4 1/14 virt: sev-guest: Use AES GCM crypto library

The sev-guest driver encryption code uses Crypto API for SNP guest messaging to interact with AMD Security processor. For enabling SecureTSC, SEV-SNP guests need to send a TSC_INFO request guest message before the smpboot phase starts. Details from the TSC_INFO response will be used to program the VMSA before the secondary CPUs are brought up. The Crypto API is not available this early in the boot phase.

In preparation of moving the encryption code out of sev-guest driver to support SecureTSC and make reviewing the diff easier, start using AES GCM library implementation instead of Crypto API.

---
 drivers/virt/coco/sev-guest/Kconfig     |   3 +-
 drivers/virt/coco/sev-guest/sev-guest.c | 172 +++++++-----------------
 drivers/virt/coco/sev-guest/sev-guest.h |   3 +
 3 files changed, 53 insertions(+), 125 deletions(-)
drivers/virt/coco/sev-guest/Kconfig +1 -2
index f9db0799ae67..bcc760bfb468 100644
@@ -2,8 +2,7 @@ config SEV_GUEST
 	tristate "AMD SEV Guest driver"
 	default m
 	depends on AMD_MEM_ENCRYPT
-	select CRYPTO_AEAD2
-	select CRYPTO_GCM
+	select CRYPTO_LIB_AESGCM
 	help
 	  SEV-SNP firmware provides the guest a mechanism to communicate with
 	  the PSP without risk from a malicious hypervisor who wishes to read,
drivers/virt/coco/sev-guest/sev-guest.c +49 -123

305 lines not on this page · raw message

drivers/virt/coco/sev-guest/sev-guest.h +3 -0
index 21bda26fdb95..ceb798a404d6 100644
@@ -13,6 +13,9 @@
 #include <linux/types.h>
 
 #define MAX_AUTHTAG_LEN		32
+#define AUTHTAG_LEN		16
+#define AAD_LEN			48
+#define MSG_HDR_VER		1
 
 /* See SNP spec SNP_GUEST_REQUEST section for the structure */
 enum msg_type {
Signature · 1 line
2.34.1
maintainer
#2 lore raw

Re: [PATCH v4 01/14] virt: sev-guest: Use AES GCM crypto library

Quoted 7 lines · Nikunj A Dadhania

The sev-guest driver encryption code uses Crypto API for SNP guest messaging to interact with AMD Security processor. For enabling SecureTSC, SEV-SNP guests need to send a TSC_INFO request guest message before the smpboot phase starts. Details from the TSC_INFO response will be used to program the VMSA before the secondary CPUs are brought up. The Crypto API is not available this early in the boot phase.

Quoted 5 lines

In preparation of moving the encryption code out of sev-guest driver to support SecureTSC and make reviewing the diff easier, start using AES GCM library implementation instead of Crypto API.

Why is that Link pointing to Ard's lib?

Link tags are used to point to relevant threads regarding *this* code
- not the lib you're using...

drivers/virt/coco/sev-guest/sev-guest.c MODULE_PARM_DESC(vmpck_id, "The VMPCK ID to use when communicating with the PSP. · line 67

Quoted 6 lines
drivers/virt/coco/sev-guest/sev-guest.c +6 -0
+static inline unsigned int get_ctx_authsize(struct snp_guest_dev *snp_dev)
+{
+	if (snp_dev && snp_dev->ctx)
+		return snp_dev->ctx->authsize;
+
+	WARN_ONCE(1, "Unable to get crypto authsize\n");

What's the point of this?

You either fail the whole process or you succeed. What's the point of warning and still returning 0?

What do you do when no one is looking at dmesg?

Quoted 31 lines
drivers/virt/coco/sev-guest/sev-guest.c +7 -13
+	return 0;
+}
+
 static bool is_vmpck_empty(struct snp_guest_dev *snp_dev)
 {
 	char zero_key[VMPCK_KEY_LEN] = {0};
@@ -152,132 +152,59 @@ static inline struct snp_guest_dev *to_snp_dev(struct file *file)
 	return container_of(dev, struct snp_guest_dev, misc);
 }
 
-static struct snp_guest_crypto *init_crypto(struct snp_guest_dev *snp_dev, u8 *key, size_t keylen)
+static struct aesgcm_ctx *snp_init_crypto(u8 *key, size_t keylen)
 {
-	struct snp_guest_crypto *crypto;
+	struct aesgcm_ctx *ctx;
 
-	crypto = kzalloc(sizeof(*crypto), GFP_KERNEL_ACCOUNT);
-	if (!crypto)
+	ctx = kzalloc(sizeof(*ctx), GFP_KERNEL_ACCOUNT);
+	if (!ctx)
 		return NULL;
 
-	crypto->tfm = crypto_alloc_aead("gcm(aes)", 0, 0);
-	if (IS_ERR(crypto->tfm))
-		goto e_free;
-
-	if (crypto_aead_setkey(crypto->tfm, key, keylen))
-		goto e_free_crypto;
-
-	crypto->iv_len = crypto_aead_ivsize(crypto->tfm);
-	crypto->iv = kmalloc(crypto->iv_len, GFP_KERNEL_ACCOUNT);

drivers/virt/coco/sev-guest/sev-guest.c static inline struct snp_guest_dev *to_snp_dev(struct file *file) · line 164

Quoted 10 lines
drivers/virt/coco/sev-guest/sev-guest.c +2 -8
-	if (!crypto->iv)
-		goto e_free_crypto;
-
-	if (crypto_aead_authsize(crypto->tfm) > MAX_AUTHTAG_LEN) {
-		if (crypto_aead_setauthsize(crypto->tfm, MAX_AUTHTAG_LEN)) {
-			dev_err(snp_dev->dev, "failed to set authsize to %d\n", MAX_AUTHTAG_LEN);
-			goto e_free_iv;
-		}
+	if (aesgcm_expandkey(ctx, key, keylen, AUTHTAG_LEN)) {
+		pr_err("SNP: crypto init failed\n");

This driver should already be printing with "sev-guest:" prefix - no need for "SNP:" too.

Quoted 3 lines
drivers/virt/coco/sev-guest/sev-guest.c +2 -0
+		kfree(ctx);
+		return NULL;
 	}

...

Signature · 4 lines
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette
#3 lore raw

Re: [PATCH v4 01/14] virt: sev-guest: Use AES GCM crypto library

Hi Boris,

Quoted 12 lines · Borislav Petkov
Quoted 11 lines · Nikunj A Dadhania

The sev-guest driver encryption code uses Crypto API for SNP guest messaging to interact with AMD Security processor. For enabling SecureTSC, SEV-SNP guests need to send a TSC_INFO request guest message before the smpboot phase starts. Details from the TSC_INFO response will be used to program the VMSA before the secondary CPUs are brought up. The Crypto API is not available this early in the boot phase.

In preparation of moving the encryption code out of sev-guest driver to support SecureTSC and make reviewing the diff easier, start using AES GCM library implementation instead of Crypto API.

Quoted 6 lines
Quoted 1 line

Why is that Link pointing to Ard's lib?

Link tags are used to point to relevant threads regarding *this* code
- not the lib you're using...

Ard had added the library specifically for this use case, this was the first discussion thread: https://lore.kernel.org/all/[email protected]/

Should I add the above link instead ?

Quoted 10 lines

drivers/virt/coco/sev-guest/sev-guest.c MODULE_PARM_DESC(vmpck_id, "The VMPCK ID to use when communicating with the PSP. · line 67

Quoted 7 lines
drivers/virt/coco/sev-guest/sev-guest.c +6 -0
+static inline unsigned int get_ctx_authsize(struct snp_guest_dev *snp_dev)
+{
+	if (snp_dev && snp_dev->ctx)
+		return snp_dev->ctx->authsize;
+
+	WARN_ONCE(1, "Unable to get crypto authsize\n");

What's the point of this?

Quoted 4 lines

You either fail the whole process or you succeed. What's the point of warning and still returning 0?

What do you do when no one is looking at dmesg?

Thinking more about this snp_dev and snp_dev->ctx will not be null, because the snp_init_crypto() would have failed resulting in sev-guest driver load failure. The below should suffice or I can drop the helper.

static inline unsigned int get_ctx_authsize(struct snp_guest_dev *snp_dev)

{
	return snp_dev->ctx->authsize;
}
Quoted 37 lines
Quoted 37 lines
drivers/virt/coco/sev-guest/sev-guest.c +7 -18
+	return 0;
+}
+
 static bool is_vmpck_empty(struct snp_guest_dev *snp_dev)
 {
 	char zero_key[VMPCK_KEY_LEN] = {0};
@@ -152,132 +152,59 @@ static inline struct snp_guest_dev *to_snp_dev(struct file *file)
 	return container_of(dev, struct snp_guest_dev, misc);
 }
 
-static struct snp_guest_crypto *init_crypto(struct snp_guest_dev *snp_dev, u8 *key, size_t keylen)
+static struct aesgcm_ctx *snp_init_crypto(u8 *key, size_t keylen)
 {
-	struct snp_guest_crypto *crypto;
+	struct aesgcm_ctx *ctx;
 
-	crypto = kzalloc(sizeof(*crypto), GFP_KERNEL_ACCOUNT);
-	if (!crypto)
+	ctx = kzalloc(sizeof(*ctx), GFP_KERNEL_ACCOUNT);
+	if (!ctx)
 		return NULL;
 
-	crypto->tfm = crypto_alloc_aead("gcm(aes)", 0, 0);
-	if (IS_ERR(crypto->tfm))
-		goto e_free;
-
-	if (crypto_aead_setkey(crypto->tfm, key, keylen))
-		goto e_free_crypto;
-
-	crypto->iv_len = crypto_aead_ivsize(crypto->tfm);
-	crypto->iv = kmalloc(crypto->iv_len, GFP_KERNEL_ACCOUNT);
-	if (!crypto->iv)
-		goto e_free_crypto;
-
-	if (crypto_aead_authsize(crypto->tfm) > MAX_AUTHTAG_LEN) {
-		if (crypto_aead_setauthsize(crypto->tfm, MAX_AUTHTAG_LEN)) {
Quoted 8 lines

drivers/virt/coco/sev-guest/sev-guest.c static inline struct snp_guest_dev *to_snp_dev(struct file *file) · line 164

Quoted 5 lines
drivers/virt/coco/sev-guest/sev-guest.c +2 -3
-			dev_err(snp_dev->dev, "failed to set authsize to %d\n", MAX_AUTHTAG_LEN);
-			goto e_free_iv;
-		}
+	if (aesgcm_expandkey(ctx, key, keylen, AUTHTAG_LEN)) {
+		pr_err("SNP: crypto init failed\n");

This driver should already be printing with "sev-guest:" prefix - no need for "SNP:" too.

Sure, I will change this.

Regards
Nikunj

maintainer
#4 lore raw

Re: [PATCH v4 01/14] virt: sev-guest: Use AES GCM crypto library

Quoted 4 lines · Nikunj A. Dadhania

Ard had added the library specifically for this use case, this was the first discussion thread: https://lore.kernel.org/all/[email protected]/

Should I add the above link instead ?

Please read up on what a Link: tag is used for.

Quoted 4 lines

static inline unsigned int get_ctx_authsize(struct snp_guest_dev *snp_dev)

{
	return snp_dev->ctx->authsize;
}

Or simply not use a one-liner function at all.

Thx.

Signature · 4 lines
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette
#5 lore raw

Re: [PATCH v4 01/14] virt: sev-guest: Use AES GCM crypto library

Quoted 7 lines · Borislav Petkov

On Thu, Oct 12, 2023 at 10:10:43AM +0530, Nikunj A. Dadhania wrote:>

Quoted 4 lines

static inline unsigned int get_ctx_authsize(struct snp_guest_dev *snp_dev)

{
	return snp_dev->ctx->authsize;
}

Or simply not use a one-liner function at all.

Sure, this is what I meant when I said "I can drop the helper" earlier.

Thanks
Nikunj

#6 lore raw

PATCH v4 2/14 virt: sev-guest: Move mutex to SNP guest device structure

In reply to Nikunj A Dadhania

In preparation for providing a new API to the sev-guest driver for sending an SNP guest message, move the SNP command mutex to the snp_guest_dev structure.

---
 drivers/virt/coco/sev-guest/sev-guest.c | 21 +++++++++++----------
 1 file changed, 11 insertions(+), 10 deletions(-)
drivers/virt/coco/sev-guest/sev-guest.c +11 -10
index 520e2b6613a7..8ba624088d73 100644
@@ -35,6 +35,9 @@ struct snp_guest_dev {
 	struct device *dev;
 	struct miscdevice misc;
 
+	/* Mutex to serialize the shared buffer access and command handling. */
+	struct mutex cmd_mutex;
+
 	void *certs_data;
 	struct aesgcm_ctx *ctx;
 	/* request and response are in unencrypted memory */
@@ -56,9 +59,6 @@ static u32 vmpck_id;
 module_param(vmpck_id, uint, 0444);
 MODULE_PARM_DESC(vmpck_id, "The VMPCK ID to use when communicating with the PSP.");
 
-/* Mutex to serialize the shared buffer access and command handling. */
-static DEFINE_MUTEX(snp_cmd_mutex);
-
 static inline unsigned int get_ctx_authsize(struct snp_guest_dev *snp_dev)
 {
 	if (snp_dev && snp_dev->ctx)
@@ -107,7 +107,7 @@ static inline u64 __snp_get_msg_seqno(struct snp_guest_dev *snp_dev)
 {
 	u64 count;
 
-	lockdep_assert_held(&snp_cmd_mutex);
+	lockdep_assert_held(&snp_dev->cmd_mutex);
 
 	/* Read the current message sequence counter from secrets pages */
 	count = *snp_dev->os_area_msg_seqno;
@@ -403,7 +403,7 @@ static int get_report(struct snp_guest_dev *snp_dev, struct snp_guest_request_io
 	struct snp_report_req req;
 	int rc, resp_len;
 
-	lockdep_assert_held(&snp_cmd_mutex);
+	lockdep_assert_held(&snp_dev->cmd_mutex);
 
 	if (!arg->req_data || !arg->resp_data)
 		return -EINVAL;
@@ -443,7 +443,7 @@ static int get_derived_key(struct snp_guest_dev *snp_dev, struct snp_guest_reque
 	/* Response data is 64 bytes and max authsize for GCM is 16 bytes. */
 	u8 buf[64 + 16];
 
-	lockdep_assert_held(&snp_cmd_mutex);
+	lockdep_assert_held(&snp_dev->cmd_mutex);
 
 	if (!arg->req_data || !arg->resp_data)
 		return -EINVAL;
@@ -481,7 +481,7 @@ static int get_ext_report(struct snp_guest_dev *snp_dev, struct snp_guest_reques
 	struct snp_report_resp *resp;
 	int ret, npages = 0, resp_len;
 
-	lockdep_assert_held(&snp_cmd_mutex);
+	lockdep_assert_held(&snp_dev->cmd_mutex);
 
 	if (!arg->req_data || !arg->resp_data)
 		return -EINVAL;
@@ -566,12 +566,12 @@ static long snp_guest_ioctl(struct file *file, unsigned int ioctl, unsigned long
 	if (!input.msg_version)
 		return -EINVAL;
 
-	mutex_lock(&snp_cmd_mutex);
+	mutex_lock(&snp_dev->cmd_mutex);
 
 	/* Check if the VMPCK is not empty */
 	if (is_vmpck_empty(snp_dev)) {
 		dev_err_ratelimited(snp_dev->dev, "VMPCK is disabled\n");
-		mutex_unlock(&snp_cmd_mutex);
+		mutex_unlock(&snp_dev->cmd_mutex);
 		return -ENOTTY;
 	}
 
@@ -589,7 +589,7 @@ static long snp_guest_ioctl(struct file *file, unsigned int ioctl, unsigned long
 		break;
 	}
 
-	mutex_unlock(&snp_cmd_mutex);
+	mutex_unlock(&snp_dev->cmd_mutex);
 
 	if (input.exitinfo2 && copy_to_user(argp, &input, sizeof(input)))
 		return -EFAULT;
@@ -708,6 +708,7 @@ static int __init sev_guest_probe(struct platform_device *pdev)
 		goto e_unmap;
 	}
 
+	mutex_init(&snp_dev->cmd_mutex);
 	platform_set_drvdata(pdev, snp_dev);
 	snp_dev->dev = dev;
 	snp_dev->layout = layout;
Signature · 1 line
2.34.1
#7 lore raw

PATCH v4 3/14 virt: sev-guest: Replace dev_dbg with pr_debug

In reply to Nikunj A Dadhania

In preparation of moving code to arch/x86/kernel/sev.c, replace dev_dbg with pr_debug.

---
 drivers/virt/coco/sev-guest/sev-guest.c | 9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)
drivers/virt/coco/sev-guest/sev-guest.c +5 -4
index 8ba624088d73..538c42e64baa 100644
@@ -206,8 +206,9 @@ static int verify_and_dec_payload(struct snp_guest_dev *snp_dev, void *payload,
 	struct snp_guest_msg_hdr *resp_hdr = &resp->hdr;
 	struct aesgcm_ctx *ctx = snp_dev->ctx;
 
-	dev_dbg(snp_dev->dev, "response [seqno %lld type %d version %d sz %d]\n",
-		resp_hdr->msg_seqno, resp_hdr->msg_type, resp_hdr->msg_version, resp_hdr->msg_sz);
+	pr_debug("response [seqno %lld type %d version %d sz %d]\n",
+		 resp_hdr->msg_seqno, resp_hdr->msg_type, resp_hdr->msg_version,
+		 resp_hdr->msg_sz);
 
 	/* Copy response from shared memory to encrypted memory. */
 	memcpy(resp, snp_dev->response, sizeof(*resp));
@@ -253,8 +254,8 @@ static int enc_payload(struct snp_guest_dev *snp_dev, u64 seqno, int version, u8
 	if (!hdr->msg_seqno)
 		return -ENOSR;
 
-	dev_dbg(snp_dev->dev, "request [seqno %lld type %d version %d sz %d]\n",
-		hdr->msg_seqno, hdr->msg_type, hdr->msg_version, hdr->msg_sz);
+	pr_debug("request [seqno %lld type %d version %d sz %d]\n",
+		 hdr->msg_seqno, hdr->msg_type, hdr->msg_version, hdr->msg_sz);
 
 	return __enc_payload(snp_dev->ctx, req, payload, sz);
 }
Signature · 1 line
2.34.1
#8 lore raw

PATCH v4 4/14 virt: sev-guest: Add SNP guest request structure

In reply to Nikunj A Dadhania

Add a snp_guest_req structure to simplify the function arguments. The structure will be used to call the SNP Guest message request API instead of passing a long list of parameters.

---
 .../x86/include/asm}/sev-guest.h              |  11 ++
 arch/x86/include/asm/sev.h                    |   7 --
 arch/x86/kernel/sev.c                         |  15 ++-
 drivers/virt/coco/sev-guest/sev-guest.c       | 103 +++++++++++-------
 4 files changed, 84 insertions(+), 52 deletions(-)
 rename {drivers/virt/coco/sev-guest => arch/x86/include/asm}/sev-guest.h (80%)
arch/x86/include/asm/sev-guest.h renamed from drivers/virt/coco/sev-guest/sev-guest.h +11 -0
similarity index 80%
rename from drivers/virt/coco/sev-guest/sev-guest.h
rename to arch/x86/include/asm/sev-guest.h
index ceb798a404d6..22ef97b55069 100644
@@ -63,4 +63,15 @@ struct snp_guest_msg {
 	u8 payload[4000];
 } __packed;
 
+struct snp_guest_req {
+	void *req_buf, *resp_buf, *data;
+	size_t req_sz, resp_sz, *data_npages;
+	u64 exit_code;
+	unsigned int vmpck_id;
+	u8 msg_version;
+	u8 msg_type;
+};
+
+int snp_issue_guest_request(struct snp_guest_req *req, struct snp_req_data *input,
+			    struct snp_guest_request_ioctl *rio);
 #endif /* __VIRT_SEVGUEST_H__ */
arch/x86/include/asm/sev.h +0 -7
index 13dc2a9d23c1..d8e972aeb22c 100644
@@ -92,8 +92,6 @@ extern bool handle_vc_boot_ghcb(struct pt_regs *regs);
 struct snp_req_data {
 	unsigned long req_gpa;
 	unsigned long resp_gpa;
-	unsigned long data_gpa;
-	unsigned int data_npages;
 };
 
 struct sev_guest_platform_data {
@@ -201,7 +199,6 @@ void snp_set_memory_private(unsigned long vaddr, unsigned int npages);
 void snp_set_wakeup_secondary_cpu(void);
 bool snp_init(struct boot_params *bp);
 void __init __noreturn snp_abort(void);
-int snp_issue_guest_request(u64 exit_code, struct snp_req_data *input, struct snp_guest_request_ioctl *rio);
 #else
 static inline void sev_es_ist_enter(struct pt_regs *regs) { }
 static inline void sev_es_ist_exit(void) { }
@@ -221,10 +218,6 @@ static inline void snp_set_memory_private(unsigned long vaddr, unsigned int npag
 static inline void snp_set_wakeup_secondary_cpu(void) { }
 static inline bool snp_init(struct boot_params *bp) { return false; }
 static inline void snp_abort(void) { }
-static inline int snp_issue_guest_request(u64 exit_code, struct snp_req_data *input, struct snp_guest_request_ioctl *rio)
-{
-	return -ENOTTY;
-}
 #endif
 
 #endif
arch/x86/kernel/sev.c +11 -4
index b031244d6d2d..a10cf9f21e22 100644
@@ -28,6 +28,7 @@
 #include <asm/cpu_entry_area.h>
 #include <asm/stacktrace.h>
 #include <asm/sev.h>
+#include <asm/sev-guest.h>
 #include <asm/insn-eval.h>
 #include <asm/fpu/xcr.h>
 #include <asm/processor.h>
@@ -2177,15 +2178,21 @@ static int __init init_sev_config(char *str)
 }
 __setup("sev=", init_sev_config);
 
-int snp_issue_guest_request(u64 exit_code, struct snp_req_data *input, struct snp_guest_request_ioctl *rio)
+int snp_issue_guest_request(struct snp_guest_req *req, struct snp_req_data *input,
+			    struct snp_guest_request_ioctl *rio)
 {
 	struct ghcb_state state;
 	struct es_em_ctxt ctxt;
 	unsigned long flags;
 	struct ghcb *ghcb;
+	u64 exit_code;
 	int ret;
 
 	rio->exitinfo2 = SEV_RET_NO_FW_CALL;
+	if (!req)
+		return -EINVAL;
+
+	exit_code = req->exit_code;
 
 	/*
 	 * __sev_get_ghcb() needs to run with IRQs disabled because it is using
@@ -2202,8 +2209,8 @@ int snp_issue_guest_request(u64 exit_code, struct snp_req_data *input, struct sn
 	vc_ghcb_invalidate(ghcb);
 
 	if (exit_code == SVM_VMGEXIT_EXT_GUEST_REQUEST) {
-		ghcb_set_rax(ghcb, input->data_gpa);
-		ghcb_set_rbx(ghcb, input->data_npages);
+		ghcb_set_rax(ghcb, __pa(req->data));
+		ghcb_set_rbx(ghcb, *req->data_npages);
 	}
 
 	ret = sev_es_ghcb_hv_call(ghcb, &ctxt, exit_code, input->req_gpa, input->resp_gpa);
@@ -2222,7 +2229,7 @@ int snp_issue_guest_request(u64 exit_code, struct snp_req_data *input, struct sn
 	case SNP_GUEST_VMM_ERR(SNP_GUEST_VMM_ERR_INVALID_LEN):
 		/* Number of expected pages are returned in RBX */
 		if (exit_code == SVM_VMGEXIT_EXT_GUEST_REQUEST) {
-			input->data_npages = ghcb_get_rbx(ghcb);
+			*req->data_npages = ghcb_get_rbx(ghcb);
 			ret = -ENOSPC;
 			break;
 		}
drivers/virt/coco/sev-guest/sev-guest.c +62 -41

238 lines not on this page · raw message

Signature · 1 line
2.34.1
#9 lore raw

PATCH v4 5/14 virt: sev-guest: Add vmpck_id to snp_guest_dev struct

In reply to Nikunj A Dadhania

Drop vmpck and os_area_msg_seqno pointers so that secret page layout does not need to be exposed to the sev-guest driver after the rework. Instead, add helper APIs to access vmpck and os_area_msg_seqno when needed.

Also, change function is_vmpck_empty() to snp_is_vmpck_empty() in preparation for moving to sev.c.

---
 drivers/virt/coco/sev-guest/sev-guest.c | 85 ++++++++++++-------------
 1 file changed, 42 insertions(+), 43 deletions(-)
drivers/virt/coco/sev-guest/sev-guest.c +42 -43
index 74b796409da9..0f519b855169 100644
@@ -50,8 +50,7 @@ struct snp_guest_dev {
 
 	struct snp_secrets_page_layout *layout;
 	struct snp_req_data input;
-	u32 *os_area_msg_seqno;
-	u8 *vmpck;
+	unsigned int vmpck_id;
 };
 
 static u32 vmpck_id;
@@ -67,14 +66,22 @@ static inline unsigned int get_ctx_authsize(struct snp_guest_dev *snp_dev)
 	return 0;
 }
 
-static bool is_vmpck_empty(struct snp_guest_dev *snp_dev)
+static inline u8 *snp_get_vmpck(struct snp_guest_dev *snp_dev)
 {
-	char zero_key[VMPCK_KEY_LEN] = {0};
+	return snp_dev->layout->vmpck0 + snp_dev->vmpck_id * VMPCK_KEY_LEN;
+}
 
-	if (snp_dev->vmpck)
-		return !memcmp(snp_dev->vmpck, zero_key, VMPCK_KEY_LEN);
+static inline u32 *snp_get_os_area_msg_seqno(struct snp_guest_dev *snp_dev)
+{
+	return &snp_dev->layout->os_area.msg_seqno_0 + snp_dev->vmpck_id;
+}
 
-	return true;
+static bool snp_is_vmpck_empty(struct snp_guest_dev *snp_dev)
+{
+	char zero_key[VMPCK_KEY_LEN] = {0};
+	u8 *key = snp_get_vmpck(snp_dev);
+
+	return !memcmp(key, zero_key, VMPCK_KEY_LEN);
 }
 
 /*
@@ -96,20 +103,22 @@ static bool is_vmpck_empty(struct snp_guest_dev *snp_dev)
  */
 static void snp_disable_vmpck(struct snp_guest_dev *snp_dev)
 {
+	u8 *key = snp_get_vmpck(snp_dev);
+
 	dev_alert(snp_dev->dev, "Disabling vmpck_id %d to prevent IV reuse.\n",
-		  vmpck_id);
-	memzero_explicit(snp_dev->vmpck, VMPCK_KEY_LEN);
-	snp_dev->vmpck = NULL;
+		  snp_dev->vmpck_id);
+	memzero_explicit(key, VMPCK_KEY_LEN);
 }
 
 static inline u64 __snp_get_msg_seqno(struct snp_guest_dev *snp_dev)
 {
+	u32 *os_area_msg_seqno = snp_get_os_area_msg_seqno(snp_dev);
 	u64 count;
 
 	lockdep_assert_held(&snp_dev->cmd_mutex);
 
 	/* Read the current message sequence counter from secrets pages */
-	count = *snp_dev->os_area_msg_seqno;
+	count = *os_area_msg_seqno;
 
 	return count + 1;
 }
@@ -137,11 +146,13 @@ static u64 snp_get_msg_seqno(struct snp_guest_dev *snp_dev)
 
 static void snp_inc_msg_seqno(struct snp_guest_dev *snp_dev)
 {
+	u32 *os_area_msg_seqno = snp_get_os_area_msg_seqno(snp_dev);
+
 	/*
 	 * The counter is also incremented by the PSP, so increment it by 2
 	 * and save in secrets page.
 	 */
-	*snp_dev->os_area_msg_seqno += 2;
+	*os_area_msg_seqno += 2;
 }
 
 static inline struct snp_guest_dev *to_snp_dev(struct file *file)
@@ -151,15 +162,22 @@ static inline struct snp_guest_dev *to_snp_dev(struct file *file)
 	return container_of(dev, struct snp_guest_dev, misc);
 }
 
-static struct aesgcm_ctx *snp_init_crypto(u8 *key, size_t keylen)
+static struct aesgcm_ctx *snp_init_crypto(struct snp_guest_dev *snp_dev)
 {
 	struct aesgcm_ctx *ctx;
+	u8 *key;
+
+	if (snp_is_vmpck_empty(snp_dev)) {
+		pr_err("SNP: vmpck id %d is null\n", snp_dev->vmpck_id);
+		return NULL;
+	}
 
 	ctx = kzalloc(sizeof(*ctx), GFP_KERNEL_ACCOUNT);
 	if (!ctx)
 		return NULL;
 
-	if (aesgcm_expandkey(ctx, key, keylen, AUTHTAG_LEN)) {
+	key = snp_get_vmpck(snp_dev);
+	if (aesgcm_expandkey(ctx, key, VMPCK_KEY_LEN, AUTHTAG_LEN)) {
 		pr_err("SNP: crypto init failed\n");
 		kfree(ctx);
 		return NULL;
@@ -592,7 +610,7 @@ static long snp_guest_ioctl(struct file *file, unsigned int ioctl, unsigned long
 	mutex_lock(&snp_dev->cmd_mutex);
 
 	/* Check if the VMPCK is not empty */
-	if (is_vmpck_empty(snp_dev)) {
+	if (snp_is_vmpck_empty(snp_dev)) {
 		dev_err_ratelimited(snp_dev->dev, "VMPCK is disabled\n");
 		mutex_unlock(&snp_dev->cmd_mutex);
 		return -ENOTTY;
@@ -662,32 +680,14 @@ static const struct file_operations snp_guest_fops = {
 	.unlocked_ioctl = snp_guest_ioctl,
 };
 
-static u8 *get_vmpck(int id, struct snp_secrets_page_layout *layout, u32 **seqno)
+bool snp_assign_vmpck(struct snp_guest_dev *dev, int vmpck_id)
 {
-	u8 *key = NULL;
+	if (WARN_ON(vmpck_id > 3))
+		return false;
 
-	switch (id) {
-	case 0:
-		*seqno = &layout->os_area.msg_seqno_0;
-		key = layout->vmpck0;
-		break;
-	case 1:
-		*seqno = &layout->os_area.msg_seqno_1;
-		key = layout->vmpck1;
-		break;
-	case 2:
-		*seqno = &layout->os_area.msg_seqno_2;
-		key = layout->vmpck2;
-		break;
-	case 3:
-		*seqno = &layout->os_area.msg_seqno_3;
-		key = layout->vmpck3;
-		break;
-	default:
-		break;
-	}
+	dev->vmpck_id = vmpck_id;
 
-	return key;
+	return true;
 }
 
 static int __init sev_guest_probe(struct platform_device *pdev)
@@ -719,14 +719,14 @@ static int __init sev_guest_probe(struct platform_device *pdev)
 		goto e_unmap;
 
 	ret = -EINVAL;
-	snp_dev->vmpck = get_vmpck(vmpck_id, layout, &snp_dev->os_area_msg_seqno);
-	if (!snp_dev->vmpck) {
+	snp_dev->layout = layout;
+	if (!snp_assign_vmpck(snp_dev, vmpck_id)) {
 		dev_err(dev, "invalid vmpck id %d\n", vmpck_id);
 		goto e_unmap;
 	}
 
 	/* Verify that VMPCK is not zero. */
-	if (is_vmpck_empty(snp_dev)) {
+	if (snp_is_vmpck_empty(snp_dev)) {
 		dev_err(dev, "vmpck id %d is null\n", vmpck_id);
 		goto e_unmap;
 	}
@@ -734,7 +734,6 @@ static int __init sev_guest_probe(struct platform_device *pdev)
 	mutex_init(&snp_dev->cmd_mutex);
 	platform_set_drvdata(pdev, snp_dev);
 	snp_dev->dev = dev;
-	snp_dev->layout = layout;
 
 	/* Allocate the shared page used for the request and response message. */
 	snp_dev->request = alloc_shared_pages(dev, sizeof(struct snp_guest_msg));
@@ -750,7 +749,7 @@ static int __init sev_guest_probe(struct platform_device *pdev)
 		goto e_free_response;
 
 	ret = -EIO;
-	snp_dev->ctx = snp_init_crypto(snp_dev->vmpck, VMPCK_KEY_LEN);
+	snp_dev->ctx = snp_init_crypto(snp_dev);
 	if (!snp_dev->ctx)
 		goto e_free_cert_data;
 
Signature · 1 line
2.34.1
#10 lore raw

PATCH v4 6/14 x86/sev: Cache the secrets page address

In reply to Nikunj A Dadhania

Save the secrets page address during snp_init() from the CC blob. Use secrets_pa instead of calling get_secrets_page() that remaps the CC blob for getting the secrets page every time.

---
 arch/x86/kernel/sev.c | 52 +++++++++++++------------------------------
 1 file changed, 16 insertions(+), 36 deletions(-)
arch/x86/kernel/sev.c +16 -36
index a10cf9f21e22..a63a3edf13d0 100644
@@ -72,6 +72,9 @@ static struct ghcb *boot_ghcb __section(".data");
 /* Bitmap of SEV features supported by the hypervisor */
 static u64 sev_hv_features __ro_after_init;
 
+/* Secrets page physical address from the CC blob */
+static u64 secrets_pa __ro_after_init;
+
 /* #VC handler runtime per-CPU data */
 struct sev_es_runtime_data {
 	struct ghcb ghcb_page;
@@ -561,45 +564,16 @@ void noinstr __sev_es_nmi_complete(void)
 	__sev_put_ghcb(&state);
 }
 
-static u64 __init get_secrets_page(void)
-{
-	u64 pa_data = boot_params.cc_blob_address;
-	struct cc_blob_sev_info info;
-	void *map;
-
-	/*
-	 * The CC blob contains the address of the secrets page, check if the
-	 * blob is present.
-	 */
-	if (!pa_data)
-		return 0;
-
-	map = early_memremap(pa_data, sizeof(info));
-	if (!map) {
-		pr_err("Unable to locate SNP secrets page: failed to map the Confidential Computing blob.\n");
-		return 0;
-	}
-	memcpy(&info, map, sizeof(info));
-	early_memunmap(map, sizeof(info));
-
-	/* smoke-test the secrets page passed */
-	if (!info.secrets_phys || info.secrets_len != PAGE_SIZE)
-		return 0;
-
-	return info.secrets_phys;
-}
-
 static u64 __init get_snp_jump_table_addr(void)
 {
 	struct snp_secrets_page_layout *layout;
 	void __iomem *mem;
-	u64 pa, addr;
+	u64 addr;
 
-	pa = get_secrets_page();
-	if (!pa)
+	if (!secrets_pa)
 		return 0;
 
-	mem = ioremap_encrypted(pa, PAGE_SIZE);
+	mem = ioremap_encrypted(secrets_pa, PAGE_SIZE);
 	if (!mem) {
 		pr_err("Unable to locate AP jump table address: failed to map the SNP secrets page.\n");
 		return 0;
@@ -2093,6 +2067,12 @@ static __init struct cc_blob_sev_info *find_cc_blob(struct boot_params *bp)
 	return cc_info;
 }
 
+static void __init set_secrets_pa(const struct cc_blob_sev_info *cc_info)
+{
+	if (cc_info && cc_info->secrets_phys && cc_info->secrets_len == PAGE_SIZE)
+		secrets_pa = cc_info->secrets_phys;
+}
+
 bool __init snp_init(struct boot_params *bp)
 {
 	struct cc_blob_sev_info *cc_info;
@@ -2104,6 +2084,8 @@ bool __init snp_init(struct boot_params *bp)
 	if (!cc_info)
 		return false;
 
+	set_secrets_pa(cc_info);
+
 	setup_cpuid_table(cc_info);
 
 	/*
@@ -2256,16 +2238,14 @@ static struct platform_device sev_guest_device = {
 static int __init snp_init_platform_device(void)
 {
 	struct sev_guest_platform_data data;
-	u64 gpa;
 
 	if (!cc_platform_has(CC_ATTR_GUEST_SEV_SNP))
 		return -ENODEV;
 
-	gpa = get_secrets_page();
-	if (!gpa)
+	if (!secrets_pa)
 		return -ENODEV;
 
-	data.secrets_gpa = gpa;
+	data.secrets_gpa = secrets_pa;
 	if (platform_device_add_data(&sev_guest_device, &data, sizeof(data)))
 		return -ENODEV;
 
Signature · 1 line
2.34.1
#11 lore raw

PATCH v4 7/14 x86/sev: Move and reorganize sev guest request api

In reply to Nikunj A Dadhania

For enabling Secure TSC, SEV-SNP guests need to communicate with the AMD Security Processor early during boot. Many of the required functions are implemented in the sev-guest driver and therefore not available at early boot. Move the required functions and provide an API to the driver to assign key and send guest request.

---
 arch/x86/Kconfig                        |   1 +
 arch/x86/include/asm/sev-guest.h        |  84 +++-
 arch/x86/include/asm/sev.h              |  10 -
 arch/x86/kernel/sev.c                   | 466 ++++++++++++++++++++++-
 drivers/virt/coco/sev-guest/Kconfig     |   1 -
 drivers/virt/coco/sev-guest/sev-guest.c | 486 +-----------------------
 6 files changed, 555 insertions(+), 493 deletions(-)
arch/x86/Kconfig +1 -0
index 53bab123a8ee..7502903cbc6b 100644
@@ -1546,6 +1546,7 @@ config AMD_MEM_ENCRYPT
 	select INSTRUCTION_DECODER
 	select ARCH_HAS_CC_PLATFORM
 	select X86_MEM_ENCRYPT
+	select CRYPTO_LIB_AESGCM
 	help
 	  Say yes to enable support for the encryption of system memory.
 	  This requires an AMD processor that supports Secure Memory
arch/x86/include/asm/sev-guest.h +82 -2

105 lines not on this page · raw message

arch/x86/include/asm/sev.h +0 -10
index d8e972aeb22c..807f85f8014c 100644
@@ -88,16 +88,6 @@ extern bool handle_vc_boot_ghcb(struct pt_regs *regs);
 
 #define RMPADJUST_VMSA_PAGE_BIT		BIT(16)
 
-/* SNP Guest message request */
-struct snp_req_data {
-	unsigned long req_gpa;
-	unsigned long resp_gpa;
-};
-
-struct sev_guest_platform_data {
-	u64 secrets_gpa;
-};
-
 /*
  * The secrets page contains 96-bytes of reserved field that can be used by
  * the guest OS. The guest OS uses the area to save the message sequence
arch/x86/kernel/sev.c +454 -12

499 lines not on this page · raw message

drivers/virt/coco/sev-guest/Kconfig +0 -1
index bcc760bfb468..c130456ad401 100644
@@ -2,7 +2,6 @@ config SEV_GUEST
 	tristate "AMD SEV Guest driver"
 	default m
 	depends on AMD_MEM_ENCRYPT
-	select CRYPTO_LIB_AESGCM
 	help
 	  SEV-SNP firmware provides the guest a mechanism to communicate with
 	  the PSP without risk from a malicious hypervisor who wishes to read,
drivers/virt/coco/sev-guest/sev-guest.c +18 -468

565 lines not on this page · raw message

Signature · 1 line
2.34.1
#12 lore raw

PATCH v4 8/14 x86/mm: Add generic guest initialization hook

In reply to Nikunj A Dadhania

Add generic enc_init guest hook for performing any type of initialization that is vendor specific.

---
 arch/x86/include/asm/x86_init.h | 2 ++
 arch/x86/kernel/x86_init.c      | 2 ++
 arch/x86/mm/mem_encrypt.c       | 3 +++
 3 files changed, 7 insertions(+)
arch/x86/include/asm/x86_init.h +2 -0
index 88085f369ff6..5bca02769074 100644
@@ -148,12 +148,14 @@ struct x86_init_acpi {
  * @enc_status_change_finish	Notify HV after the encryption status of a range is changed
  * @enc_tlb_flush_required	Returns true if a TLB flush is needed before changing page encryption status
  * @enc_cache_flush_required	Returns true if a cache flush is needed before changing page encryption status
+ * @enc_init			Prepare and initialize encryption features
  */
 struct x86_guest {
 	void (*enc_status_change_prepare)(unsigned long vaddr, int npages, bool enc);
 	bool (*enc_status_change_finish)(unsigned long vaddr, int npages, bool enc);
 	bool (*enc_tlb_flush_required)(bool enc);
 	bool (*enc_cache_flush_required)(void);
+	void (*enc_init)(void);
 };
 
 /**
arch/x86/kernel/x86_init.c +2 -0
index d82f4fa2f1bf..451e0f39d053 100644
@@ -135,6 +135,7 @@ static bool enc_status_change_finish_noop(unsigned long vaddr, int npages, bool
 static bool enc_tlb_flush_required_noop(bool enc) { return false; }
 static bool enc_cache_flush_required_noop(void) { return false; }
 static bool is_private_mmio_noop(u64 addr) {return false; }
+static void enc_init_noop(void) { }
 
 struct x86_platform_ops x86_platform __ro_after_init = {
 	.calibrate_cpu			= native_calibrate_cpu_early,
@@ -157,6 +158,7 @@ struct x86_platform_ops x86_platform __ro_after_init = {
 		.enc_status_change_finish  = enc_status_change_finish_noop,
 		.enc_tlb_flush_required	   = enc_tlb_flush_required_noop,
 		.enc_cache_flush_required  = enc_cache_flush_required_noop,
+		.enc_init		   = enc_init_noop,
 	},
 };
 
arch/x86/mm/mem_encrypt.c +3 -0
index 9f27e14e185f..01abecc9a774 100644
@@ -84,5 +84,8 @@ void __init mem_encrypt_init(void)
 	/* Call into SWIOTLB to update the SWIOTLB DMA buffers */
 	swiotlb_update_mem_attributes();
 
+	if (x86_platform.guest.enc_init)
+		x86_platform.guest.enc_init();
+
 	print_mem_encrypt_feature_info();
 }
Signature · 1 line
2.34.1
#13 lore raw

PATCH v4 9/14 x86/sev: Add Secure TSC support for SNP guests

In reply to Nikunj A Dadhania

Add support for Secure TSC in SNP enabled guests. Secure TSC allows guest to securely use RDTSC/RDTSCP instructions as the parameters being used cannot be changed by hypervisor once the guest is launched.

During the boot-up of the secondary cpus, SecureTSC enabled guests need to query TSC info from AMD Security Processor. This communication channel is encrypted between the AMD Security Processor and the guest, the hypervisor is just the conduit to deliver the guest messages to the AMD Security Processor. Each message is protected with an AEAD (AES-256 GCM). Use minimal AES GCM library to encrypt/decrypt SNP Guest messages to communicate with the PSP.

---
 arch/x86/coco/core.c             |  3 ++
 arch/x86/include/asm/sev-guest.h | 18 +++++++
 arch/x86/include/asm/sev.h       |  2 +
 arch/x86/include/asm/svm.h       |  6 ++-
 arch/x86/kernel/sev.c            | 82 ++++++++++++++++++++++++++++++++
 arch/x86/mm/mem_encrypt_amd.c    |  6 +++
 include/linux/cc_platform.h      |  8 ++++
 7 files changed, 123 insertions(+), 2 deletions(-)
arch/x86/coco/core.c +3 -0
index 73f83233d25d..1cfb86c6bd78 100644
@@ -89,6 +89,9 @@ static bool amd_cc_platform_has(enum cc_attr attr)
 	case CC_ATTR_GUEST_SEV_SNP:
 		return sev_status & MSR_AMD64_SEV_SNP_ENABLED;
 
+	case CC_ATTR_GUEST_SECURE_TSC:
+		return sev_status & MSR_AMD64_SNP_SECURE_TSC;
+
 	default:
 		return false;
 	}
arch/x86/include/asm/sev-guest.h +18 -0
index e6f94208173d..58739173eba9 100644
@@ -39,6 +39,8 @@ enum msg_type {
 	SNP_MSG_ABSORB_RSP,
 	SNP_MSG_VMRK_REQ,
 	SNP_MSG_VMRK_RSP,
+	SNP_MSG_TSC_INFO_REQ = 17,
+	SNP_MSG_TSC_INFO_RSP,
 
 	SNP_MSG_TYPE_MAX
 };
@@ -111,6 +113,22 @@ struct snp_guest_req {
 	u8 msg_type;
 };
 
+struct snp_tsc_info_req {
+#define SNP_TSC_INFO_REQ_SZ 128
+	/* Must be zero filled */
+	u8 rsvd[SNP_TSC_INFO_REQ_SZ];
+} __packed;
+
+struct snp_tsc_info_resp {
+	/* Status of TSC_INFO message */
+	u32 status;
+	u32 rsvd1;
+	u64 tsc_scale;
+	u64 tsc_offset;
+	u32 tsc_factor;
+	u8 rsvd2[100];
+} __packed;
+
 int snp_setup_psp_messaging(struct snp_guest_dev *snp_dev);
 int snp_send_guest_request(struct snp_guest_dev *dev, struct snp_guest_req *req,
 			   struct snp_guest_request_ioctl *rio);
arch/x86/include/asm/sev.h +2 -0
index 807f85f8014c..d5b35da1b583 100644
@@ -189,6 +189,7 @@ void snp_set_memory_private(unsigned long vaddr, unsigned int npages);
 void snp_set_wakeup_secondary_cpu(void);
 bool snp_init(struct boot_params *bp);
 void __init __noreturn snp_abort(void);
+void __init snp_secure_tsc_prepare(void);
 #else
 static inline void sev_es_ist_enter(struct pt_regs *regs) { }
 static inline void sev_es_ist_exit(void) { }
@@ -208,6 +209,7 @@ static inline void snp_set_memory_private(unsigned long vaddr, unsigned int npag
 static inline void snp_set_wakeup_secondary_cpu(void) { }
 static inline bool snp_init(struct boot_params *bp) { return false; }
 static inline void snp_abort(void) { }
+static inline void __init snp_secure_tsc_prepare(void) { }
 #endif
 
 #endif
arch/x86/include/asm/svm.h +4 -2
index e7c7379d6ac7..3956c5095109 100644
@@ -412,7 +412,9 @@ struct sev_es_save_area {
 	u8 reserved_0x298[80];
 	u32 pkru;
 	u32 tsc_aux;
-	u8 reserved_0x2f0[24];
+	u64 tsc_scale;
+	u64 tsc_offset;
+	u8 reserved_0x300[8];
 	u64 rcx;
 	u64 rdx;
 	u64 rbx;
@@ -544,7 +546,7 @@ static inline void __unused_size_checks(void)
 	BUILD_BUG_RESERVED_OFFSET(sev_es_save_area, 0x1c0);
 	BUILD_BUG_RESERVED_OFFSET(sev_es_save_area, 0x248);
 	BUILD_BUG_RESERVED_OFFSET(sev_es_save_area, 0x298);
-	BUILD_BUG_RESERVED_OFFSET(sev_es_save_area, 0x2f0);
+	BUILD_BUG_RESERVED_OFFSET(sev_es_save_area, 0x300);
 	BUILD_BUG_RESERVED_OFFSET(sev_es_save_area, 0x320);
 	BUILD_BUG_RESERVED_OFFSET(sev_es_save_area, 0x380);
 	BUILD_BUG_RESERVED_OFFSET(sev_es_save_area, 0x3f0);
arch/x86/kernel/sev.c +82 -0

100 lines not on this page · raw message

arch/x86/mm/mem_encrypt_amd.c +6 -0
index e0b51c09109f..fc25749fb2e5 100644
@@ -215,6 +215,11 @@ void __init sme_map_bootdata(char *real_mode_data)
 	__sme_early_map_unmap_mem(__va(cmdline_paddr), COMMAND_LINE_SIZE, true);
 }
 
+void __init amd_enc_init(void)
+{
+	snp_secure_tsc_prepare();
+}
+
 void __init sev_setup_arch(void)
 {
 	phys_addr_t total_mem = memblock_phys_mem_size();
@@ -501,6 +506,7 @@ void __init sme_early_init(void)
 	x86_platform.guest.enc_status_change_finish  = amd_enc_status_change_finish;
 	x86_platform.guest.enc_tlb_flush_required    = amd_enc_tlb_flush_required;
 	x86_platform.guest.enc_cache_flush_required  = amd_enc_cache_flush_required;
+	x86_platform.guest.enc_init		     = amd_enc_init;
 }
 
 void __init mem_encrypt_free_decrypted_mem(void)
include/linux/cc_platform.h +8 -0
index cb0d6cd1c12f..e081ca4d5da2 100644
@@ -90,6 +90,14 @@ enum cc_attr {
 	 * Examples include TDX Guest.
 	 */
 	CC_ATTR_HOTPLUG_DISABLED,
+
+	/**
+	 * @CC_ATTR_GUEST_SECURE_TSC: Secure TSC is active.
+	 *
+	 * The platform/OS is running as a guest/virtual machine and actively
+	 * using AMD SEV-SNP Secure TSC feature.
+	 */
+	CC_ATTR_GUEST_SECURE_TSC,
 };
 
 #ifdef CONFIG_ARCH_HAS_CC_PLATFORM
Signature · 1 line
2.34.1
#14 lore raw

PATCH v4 10/14 x86/sev: Change TSC MSR behavior for Secure TSC enabled guests

In reply to Nikunj A Dadhania

Secure TSC enabled guests should not write MSR_IA32_TSC(10H) register as the subsequent TSC value reads are undefined. MSR_IA32_TSC related accesses should not exit to the hypervisor for such guests.

Accesses to MSR_IA32_TSC needs special handling in the #VC handler for the guests with Secure TSC enabled. Writes to MSR_IA32_TSC should be ignored, and reads of MSR_IA32_TSC should return the result of the RDTSC instruction.

---
 arch/x86/kernel/sev.c | 24 ++++++++++++++++++++++++
 1 file changed, 24 insertions(+)
arch/x86/kernel/sev.c +24 -0
index d55562cd395d..2d42822fa01c 100644
@@ -1729,6 +1729,30 @@ static enum es_result vc_handle_msr(struct ghcb *ghcb, struct es_em_ctxt *ctxt)
 	/* Is it a WRMSR? */
 	exit_info_1 = (ctxt->insn.opcode.bytes[1] == 0x30) ? 1 : 0;
 
+	/*
+	 * TSC related accesses should not exit to the hypervisor when a
+	 * guest is executing with SecureTSC enabled, so special handling
+	 * is required for accesses of MSR_IA32_TSC:
+	 *
+	 * Writes: Writing to MSR_IA32_TSC can cause subsequent reads
+	 *         of the TSC to return undefined values, so ignore all
+	 *         writes.
+	 * Reads:  Reads of MSR_IA32_TSC should return the current TSC
+	 *         value, use the value returned by RDTSC.
+	 */
+	if (regs->cx == MSR_IA32_TSC && cc_platform_has(CC_ATTR_GUEST_SECURE_TSC)) {
+		u64 tsc;
+
+		if (exit_info_1)
+			return ES_OK;
+
+		tsc = rdtsc();
+		regs->ax = UINT_MAX & tsc;
+		regs->dx = UINT_MAX & (tsc >> 32);
+
+		return ES_OK;
+	}
+
 	ghcb_set_rcx(ghcb, regs->cx);
 	if (exit_info_1) {
 		ghcb_set_rax(ghcb, regs->ax);
Signature · 1 line
2.34.1
#15 lore raw

PATCH v4 11/14 x86/sev: Prevent RDTSC/RDTSCP interception for Secure TSC enabled guests

In reply to Nikunj A Dadhania

The hypervisor should not be intercepting RDTSC/RDTSCP when Secure TSC is enabled. A #VC exception will be generated if the RDTSC/RDTSCP instructions are being intercepted. If this should occur and Secure TSC is enabled, terminate guest execution.

---
 arch/x86/kernel/sev-shared.c | 7 +++++++
 1 file changed, 7 insertions(+)
arch/x86/kernel/sev-shared.c +7 -0
index 3a5b0c9c4fcc..1c22025b298f 100644
@@ -912,6 +912,13 @@ static enum es_result vc_handle_rdtsc(struct ghcb *ghcb,
 	bool rdtscp = (exit_code == SVM_EXIT_RDTSCP);
 	enum es_result ret;
 
+	/*
+	 * RDTSC and RDTSCP should not be intercepted when Secure TSC is
+	 * enabled. Terminate the SNP guest when the interception is enabled.
+	 */
+	if (sev_status & MSR_AMD64_SNP_SECURE_TSC)
+		return ES_VMM_ERROR;
+
 	ret = sev_es_ghcb_hv_call(ghcb, ctxt, exit_code, 0, 0);
 	if (ret != ES_OK)
 		return ret;
Signature · 1 line
2.34.1
#16 lore raw

PATCH v4 12/14 x86/kvmclock: Skip kvmclock when Secure TSC is available

In reply to Nikunj A Dadhania

For AMD SNP guests having Secure TSC enabled, skip using the kvmclock. The guest kernel will fallback and use Secure TSC based clocksource.

---
 arch/x86/kernel/kvmclock.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/x86/kernel/kvmclock.c +1 -1
index 0f35d44c56fe..1be342064851 100644
@@ -288,7 +288,7 @@ void __init kvmclock_init(void)
 {
 	u8 flags;
 
-	if (!kvm_para_available() || !kvmclock)
+	if (!kvm_para_available() || !kvmclock || cc_platform_has(CC_ATTR_GUEST_SECURE_TSC))
 		return;
 
 	if (kvm_para_has_feature(KVM_FEATURE_CLOCKSOURCE2)) {
Signature · 1 line
2.34.1
#17 lore raw

PATCH v4 13/14 x86/tsc: Mark Secure TSC as reliable clocksource

In reply to Nikunj A Dadhania

AMD SNP guests may have Secure TSC feature enabled. Secure TSC as clocksource is wrongly marked as unstable, mark Secure TSC as reliable.

---
 arch/x86/kernel/tsc.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/x86/kernel/tsc.c +1 -1
index 344698852146..5f1e2b51ae3b 100644
@@ -1221,7 +1221,7 @@ static void __init check_system_tsc_reliable(void)
 			tsc_clocksource_reliable = 1;
 	}
 #endif
-	if (boot_cpu_has(X86_FEATURE_TSC_RELIABLE))
+	if (boot_cpu_has(X86_FEATURE_TSC_RELIABLE) || cc_platform_has(CC_ATTR_GUEST_SECURE_TSC))
 		tsc_clocksource_reliable = 1;
 
 	/*
Signature · 1 line
2.34.1
#18 lore raw

PATCH v4 14/14 x86/sev: Enable Secure TSC for SNP guests

In reply to Nikunj A Dadhania

Now that all the required plumbing is done for enabling SNP Secure TSC feature, add Secure TSC to snp features present list.

The CC_ATTR_GUEST_SECURE_TSC can be used by the guest to query whether the SNP guest has Secure TSC feature active.

---
 arch/x86/boot/compressed/sev.c |  2 +-
 arch/x86/mm/mem_encrypt.c      | 10 ++++++++--
 2 files changed, 9 insertions(+), 3 deletions(-)
arch/x86/boot/compressed/sev.c +1 -1
index 014b89c89088..11f951caf2de 100644
@@ -313,7 +313,7 @@ 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_SECURE_TSC)
 
 void snp_check_features(void)
 {
arch/x86/mm/mem_encrypt.c +8 -2
index 01abecc9a774..26608b9f2ca7 100644
@@ -69,8 +69,14 @@ static void print_mem_encrypt_feature_info(void)
 		pr_cont(" SEV-ES");
 
 	/* Secure Nested Paging */
-	if (cc_platform_has(CC_ATTR_GUEST_SEV_SNP))
-		pr_cont(" SEV-SNP");
+	if (cc_platform_has(CC_ATTR_GUEST_SEV_SNP)) {
+		pr_cont(" SEV-SNP\n");
+		pr_cont("SNP Features active: ");
+
+		/* SNP Secure TSC */
+		if (cc_platform_has(CC_ATTR_GUEST_SECURE_TSC))
+			pr_cont(" SECURE-TSC");
+	}
 
 	pr_cont("\n");
 }
Signature · 1 line
2.34.1
#19 lore raw
In reply to Nikunj A Dadhania
Quoted 42 lines · Nikunj A Dadhania

Secure TSC allows guests to securely use RDTSC/RDTSCP instructions as the parameters being used cannot be changed by hypervisor once the guest is launched. More details in the AMD64 APM Vol 2, Section "Secure TSC".

During the boot-up of the secondary cpus, SecureTSC enabled guests need to query TSC info from AMD Security Processor. This communication channel is encrypted between the AMD Security Processor and the guest, the hypervisor is just the conduit to deliver the guest messages to the AMD Security Processor. Each message is protected with an AEAD (AES-256 GCM). See "SEV Secure Nested Paging Firmware ABI Specification" document (currently at
https://www.amd.com/system/files/TechDocs/56860.pdf) section "TSC Info"

Use a minimal GCM library to encrypt/decrypt SNP Guest messages to communicate with the AMD Security Processor which is available at early boot.

SEV-guest driver has the implementation for guest and AMD Security Processor communication. As the TSC_INFO needs to be initialized during early boot before smp cpus are started, move most of the sev-guest driver code to kernel/sev.c and provide well defined APIs to the sev-guest driver to use the interface to avoid code-duplication.

Patches:
01-07: Preparation and movement of sev-guest driver code 08-14: SecureTSC enablement patches.

Testing SecureTSC
-----------------

SecureTSC hypervisor patches based on top of SEV-SNP UPM series:
https://github.com/nikunjad/linux/tree/snp-host-latest-securetsc

QEMU changes:
https://github.com/nikunjad/qemu/tree/snp_securetsc

QEMU commandline SEV-SNP-UPM with SecureTSC:

  qemu-system-x86_64 -cpu EPYC-Milan-v2,+secure-tsc \
    -object memory-backend-memfd-private,id=ram1,size=1G,share=true \
    -object sev-snp-guest,id=sev0,cbitpos=51,reduced-phys-bits=1,secure-tsc=on \
    -machine q35,confidential-guest-support=sev0,memory-backend=ram1,kvm-type=protected \
    ...
Quoted 7 lines

Changelog:
----------
v4:
* Drop handle_guest_request() and handle_guest_request_ext()
* Drop NULL check for key
* Corrected commit subject
* Added Reviewed-by from Tom

A gentle reminder.

Regards
Nikunj

#20 lore raw
Quoted 5 lines · Nikunj A. Dadhania
Quoted 4 lines · Nikunj A Dadhania

Secure TSC allows guests to securely use RDTSC/RDTSCP instructions as the parameters being used cannot be changed by hypervisor once the guest is launched. More details in the AMD64 APM Vol 2, Section "Secure TSC".

...

Quoted 10 lines
Quoted 8 lines

Changelog:
----------
v4:
* Drop handle_guest_request() and handle_guest_request_ext()
* Drop NULL check for key
* Corrected commit subject
* Added Reviewed-by from Tom

A gentle reminder.

Ping ?

Regards
Nikunj

maintainer
#21 lore raw
Quoted 16 lines · Nikunj A. Dadhania
Quoted 5 lines · Nikunj A. Dadhania
Quoted 4 lines · Nikunj A Dadhania

Secure TSC allows guests to securely use RDTSC/RDTSCP instructions as the parameters being used cannot be changed by hypervisor once the guest is launched. More details in the AMD64 APM Vol 2, Section "Secure TSC".

...

Quoted 9 lines
Quoted 8 lines

Changelog:
----------
v4:
* Drop handle_guest_request() and handle_guest_request_ext()
* Drop NULL check for key
* Corrected commit subject
* Added Reviewed-by from Tom

Quoted 4 lines
Quoted 1 line

A gentle reminder.

Ping ?

Pong.

How about you find other SEV patches on LKML which are waiting for review and review them, while waiting for yours?

Thx.

Signature · 4 lines
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette
#22 lore raw
Quoted 20 lines · Borislav Petkov
Quoted 19 lines · Nikunj A. Dadhania
Quoted 5 lines · Nikunj A. Dadhania
Quoted 4 lines · Nikunj A Dadhania

Secure TSC allows guests to securely use RDTSC/RDTSCP instructions as the parameters being used cannot be changed by hypervisor once the guest is launched. More details in the AMD64 APM Vol 2, Section "Secure TSC".

...

Quoted 10 lines
Quoted 8 lines

Changelog:
----------
v4:
* Drop handle_guest_request() and handle_guest_request_ext()
* Drop NULL check for key
* Corrected commit subject
* Added Reviewed-by from Tom

A gentle reminder.

Quoted 6 lines
Quoted 1 line

Ping ?

Pong.

How about you find other SEV patches on LKML which are waiting for review and review them, while waiting for yours?

That is a good idea and I wasn't aware that this may work as a good bribe [1] :-)

Regards,
Nikunj

[1] https://lore.kernel.org/all/20230816081942.GBZNyGnk%2FK+nDuBzsC@fat_crate.local/

maintainer
#23 lore raw
Quoted 1 line · Nikunj A. Dadhania

That is a good idea and I wasn't aware that this may work as a good bribe [1] :-)

Whatever it takes to get you folks to review each other's patches and get more involved in the community.

:-)

Signature · 4 lines
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette