Skip to contentSkip to conversation

Add Secure TSC support for SNP guests

PATCH Superseded Nikunj A Dadhania 3 participants 33 messages Mar 26, 2023 – Apr 14, 2023

Subsystems (4)

v2 of 16 Superseded Latest is v16 →

lore
Share & tools
lore
#0 lore raw

v2 cover Add Secure TSC support for SNP 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. 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 Security processor (PSP). This communication channel is encrypted between the security processor and the guest, hypervisor is just the conduit to deliver the guest messages to the 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 minimal GCM library to encrypt/decrypt SNP Guest messages to communicate with the PSP which is available at earlyboot.

SEV-guest driver has the implementation for guest and security coprocessor 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 as part to kernel/sev.c and provide well defined APIs to the sev-guest driver to use the interface to avoid code-duplication.

Patches:
01-05: Preparation and movement of sev-guest driver code
06: Adds generic guest initialization hook
07-11: SecureTSC enablement patches.

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

SecureTSC hypervisor patches based on top of SEV-SNP UPM series:
https://github.com/nikunjad/linux/tree/upmv10-snpv8_securetsc

QEMU changes:
https://github.com/nikunjad/qemu/tree/upmv10b-snpv3-wip-securetsc

QEMU commandline SEV-SNP-UPM with SecureTSC:

  qemu-system-x86_64 -cpu EPYC-Milan-v2 \
    -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:
----------
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 (11):
virt: sev-guest: Use AES GCM crypto library
virt: sev-guest: Move mutex to SNP guest device structure
virt: sev-guest: Add snp_guest_req structure
virt: sev-guest: Add simplified helper to assign vmpck x86/sev: Move and reorganize sev guest request api x86/mm: Add generic guest initialization hook x86/sev: Change TSC MSR behavior for Secure TSC enabled guests x86/sev: Add Secure TSC support for SNP guests x86/kvmclock: Use Secure TSC as clock if 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 +
 .../x86/include/asm}/sev-guest.h              |  46 ++
 arch/x86/include/asm/sev.h                    |  24 +
 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                         | 541 +++++++++++++++-
 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       | 592 ++----------------
 include/linux/cc_platform.h                   |   8 +
 17 files changed, 708 insertions(+), 551 deletions(-)
 rename {drivers/virt/coco/sev-guest => arch/x86/include/asm}/sev-guest.h (54%)

base-commit: e8d018dd0257f744ca50a729e3d042cf2ec9da65

Signature · 1 line
2.34.1
#1 lore raw

PATCH v2 1/11 virt: sev-guest: Use AES GCM crypto library

SEV-SNP guests with SecureTSC enabled need to send a TSC_INFO SNP Guest message to the AMD security processor before the smpboot phase starts. Details from the TSC_INFO response have to be programmed in the VMSA before the secondary CPUs are brought up.

Start using the AES GCM library implementation as the crypto API is not available yet this early.

---
 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
#2 lore raw

Re: [PATCH v2 01/11] virt: sev-guest: Use AES GCM crypto library

Quoted 7 lines · Nikunj A Dadhania

SEV-SNP guests with SecureTSC enabled need to send a TSC_INFO SNP Guest message to the AMD security processor before the smpboot phase starts. Details from the TSC_INFO response have to be programmed in the VMSA before the secondary CPUs are brought up.

Start using the AES GCM library implementation as the crypto API is not available yet this early.

This isn't quite true, yet. You should add that the encryption code will be moved out of sev-guest to support Secure TSC, but to make the diffs easier to review, convert the Crypto API usage over to AES GCM library usage before moving it.

Thanks,
Tom

Quoted 362 lines
---
  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
index 46f1a8d558b0..57af908bafba 100644
@@ -16,8 +16,7 @@
  #include <linux/miscdevice.h>
  #include <linux/set_memory.h>
  #include <linux/fs.h>
-#include <crypto/aead.h>
-#include <linux/scatterlist.h>
+#include <crypto/gcm.h>
  #include <linux/psp-sev.h>
  #include <uapi/linux/sev-guest.h>
  #include <uapi/linux/psp-sev.h>
@@ -28,24 +27,16 @@
  #include "sev-guest.h"
 
  #define DEVICE_NAME	"sev-guest"
-#define AAD_LEN		48
-#define MSG_HDR_VER	1
 
  #define SNP_REQ_MAX_RETRY_DURATION	(60*HZ)
  #define SNP_REQ_RETRY_DELAY		(2*HZ)
 
-struct snp_guest_crypto {
-	struct crypto_aead *tfm;
-	u8 *iv, *authtag;
-	int iv_len, a_len;
-};
-
  struct snp_guest_dev {
  	struct device *dev;
  	struct miscdevice misc;
 
  	void *certs_data;
-	struct snp_guest_crypto *crypto;
+	struct aesgcm_ctx *ctx;
  	struct snp_guest_msg *request, *response;
  	struct snp_secrets_page_layout *layout;
  	struct snp_req_data input;
@@ -60,6 +51,15 @@ 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)
+		return snp_dev->ctx->authsize;
+
+	WARN_ONCE(1, "Unable to get crypto authsize\n");
+	return 0;
+}
+
  static bool is_vmpck_empty(struct snp_guest_dev *snp_dev)
  {
  	char zero_key[VMPCK_KEY_LEN] = {0};
@@ -144,132 +144,59 @@ 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)) {
-			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");
+		kfree(ctx);
+		return NULL;
  	}
 
-	crypto->a_len = crypto_aead_authsize(crypto->tfm);
-	crypto->authtag = kmalloc(crypto->a_len, GFP_KERNEL_ACCOUNT);
-	if (!crypto->authtag)
-		goto e_free_iv;
-
-	return crypto;
-
-e_free_iv:
-	kfree(crypto->iv);
-e_free_crypto:
-	crypto_free_aead(crypto->tfm);
-e_free:
-	kfree(crypto);
-
-	return NULL;
+	return ctx;
  }
 
-static void deinit_crypto(struct snp_guest_crypto *crypto)
-{
-	crypto_free_aead(crypto->tfm);
-	kfree(crypto->iv);
-	kfree(crypto->authtag);
-	kfree(crypto);
-}
-
-static int enc_dec_message(struct snp_guest_crypto *crypto, struct snp_guest_msg *msg,
-			   u8 *src_buf, u8 *dst_buf, size_t len, bool enc)
-{
-	struct snp_guest_msg_hdr *hdr = &msg->hdr;
-	struct scatterlist src[3], dst[3];
-	DECLARE_CRYPTO_WAIT(wait);
-	struct aead_request *req;
-	int ret;
-
-	req = aead_request_alloc(crypto->tfm, GFP_KERNEL);
-	if (!req)
-		return -ENOMEM;
-
-	/*
-	 * AEAD memory operations:
-	 * +------ AAD -------+------- DATA -----+---- AUTHTAG----+
-	 * |  msg header      |  plaintext       |  hdr->authtag  |
-	 * | bytes 30h - 5Fh  |    or            |                |
-	 * |                  |   cipher         |                |
-	 * +------------------+------------------+----------------+
-	 */
-	sg_init_table(src, 3);
-	sg_set_buf(&src[0], &hdr->algo, AAD_LEN);
-	sg_set_buf(&src[1], src_buf, hdr->msg_sz);
-	sg_set_buf(&src[2], hdr->authtag, crypto->a_len);
-
-	sg_init_table(dst, 3);
-	sg_set_buf(&dst[0], &hdr->algo, AAD_LEN);
-	sg_set_buf(&dst[1], dst_buf, hdr->msg_sz);
-	sg_set_buf(&dst[2], hdr->authtag, crypto->a_len);
-
-	aead_request_set_ad(req, AAD_LEN);
-	aead_request_set_tfm(req, crypto->tfm);
-	aead_request_set_callback(req, 0, crypto_req_done, &wait);
-
-	aead_request_set_crypt(req, src, dst, len, crypto->iv);
-	ret = crypto_wait_req(enc ? crypto_aead_encrypt(req) : crypto_aead_decrypt(req), &wait);
-
-	aead_request_free(req);
-	return ret;
-}
-
-static int __enc_payload(struct snp_guest_dev *snp_dev, struct snp_guest_msg *msg,
+static int __enc_payload(struct aesgcm_ctx *ctx, struct snp_guest_msg *msg,
  			 void *plaintext, size_t len)
  {
-	struct snp_guest_crypto *crypto = snp_dev->crypto;
  	struct snp_guest_msg_hdr *hdr = &msg->hdr;
+	u8 iv[GCM_AES_IV_SIZE] = {};
 
-	memset(crypto->iv, 0, crypto->iv_len);
-	memcpy(crypto->iv, &hdr->msg_seqno, sizeof(hdr->msg_seqno));
+	if (WARN_ON((hdr->msg_sz + ctx->authsize) > sizeof(msg->payload)))
+		return -EBADMSG;
 
-	return enc_dec_message(crypto, msg, plaintext, msg->payload, len, true);
+	memcpy(iv, &hdr->msg_seqno, sizeof(hdr->msg_seqno));
+	aesgcm_encrypt(ctx, msg->payload, plaintext, len, &hdr->algo, AAD_LEN,
+		       iv, hdr->authtag);
+	return 0;
  }
 
-static int dec_payload(struct snp_guest_dev *snp_dev, struct snp_guest_msg *msg,
+static int dec_payload(struct aesgcm_ctx *ctx, struct snp_guest_msg *msg,
  		       void *plaintext, size_t len)
  {
-	struct snp_guest_crypto *crypto = snp_dev->crypto;
  	struct snp_guest_msg_hdr *hdr = &msg->hdr;
+	u8 iv[GCM_AES_IV_SIZE] = {};
 
-	/* Build IV with response buffer sequence number */
-	memset(crypto->iv, 0, crypto->iv_len);
-	memcpy(crypto->iv, &hdr->msg_seqno, sizeof(hdr->msg_seqno));
-
-	return enc_dec_message(crypto, msg, msg->payload, plaintext, len, false);
+	memcpy(iv, &hdr->msg_seqno, sizeof(hdr->msg_seqno));
+	if (aesgcm_decrypt(ctx, plaintext, msg->payload, len, &hdr->algo,
+			   AAD_LEN, iv, hdr->authtag))
+		return 0;
+	else
+		return -EBADMSG;
  }
 
  static int verify_and_dec_payload(struct snp_guest_dev *snp_dev, void *payload, u32 sz)
  {
-	struct snp_guest_crypto *crypto = snp_dev->crypto;
  	struct snp_guest_msg *resp = snp_dev->response;
  	struct snp_guest_msg *req = snp_dev->request;
  	struct snp_guest_msg_hdr *req_hdr = &req->hdr;
  	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);
@@ -287,11 +214,11 @@ static int verify_and_dec_payload(struct snp_guest_dev *snp_dev, void *payload,
  	 * If the message size is greater than our buffer length then return
  	 * an error.
  	 */
-	if (unlikely((resp_hdr->msg_sz + crypto->a_len) > sz))
+	if (unlikely((resp_hdr->msg_sz + ctx->authsize) > sz))
  		return -EBADMSG;
 
  	/* Decrypt the payload */
-	return dec_payload(snp_dev, resp, payload, resp_hdr->msg_sz + crypto->a_len);
+	return dec_payload(ctx, resp, payload, resp_hdr->msg_sz);
  }
 
  static int enc_payload(struct snp_guest_dev *snp_dev, u64 seqno, int version, u8 type,
@@ -318,7 +245,7 @@ static int enc_payload(struct snp_guest_dev *snp_dev, u64 seqno, int version, u8
  	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);
 
-	return __enc_payload(snp_dev, req, payload, sz);
+	return __enc_payload(snp_dev->ctx, req, payload, sz);
  }
 
  static int __handle_guest_request(struct snp_guest_dev *snp_dev, u64 exit_code, __u64 *fw_err)
@@ -446,7 +373,6 @@ static int handle_guest_request(struct snp_guest_dev *snp_dev, u64 exit_code, in
 
  static int get_report(struct snp_guest_dev *snp_dev, struct snp_guest_request_ioctl *arg)
  {
-	struct snp_guest_crypto *crypto = snp_dev->crypto;
  	struct snp_report_resp *resp;
  	struct snp_report_req req;
  	int rc, resp_len;
@@ -464,7 +390,7 @@ static int get_report(struct snp_guest_dev *snp_dev, struct snp_guest_request_io
  	 * response payload. Make sure that it has enough space to cover the
  	 * authtag.
  	 */
-	resp_len = sizeof(resp->data) + crypto->a_len;
+	resp_len = sizeof(resp->data) + get_ctx_authsize(snp_dev);
  	resp = kzalloc(resp_len, GFP_KERNEL_ACCOUNT);
  	if (!resp)
  		return -ENOMEM;
@@ -485,7 +411,6 @@ static int get_report(struct snp_guest_dev *snp_dev, struct snp_guest_request_io
 
  static int get_derived_key(struct snp_guest_dev *snp_dev, struct snp_guest_request_ioctl *arg)
  {
-	struct snp_guest_crypto *crypto = snp_dev->crypto;
  	struct snp_derived_key_resp resp = {0};
  	struct snp_derived_key_req req;
  	int rc, resp_len;
@@ -502,7 +427,7 @@ static int get_derived_key(struct snp_guest_dev *snp_dev, struct snp_guest_reque
  	 * response payload. Make sure that it has enough space to cover the
  	 * authtag.
  	 */
-	resp_len = sizeof(resp.data) + crypto->a_len;
+	resp_len = sizeof(resp.data) + get_ctx_authsize(snp_dev);
  	if (sizeof(buf) < resp_len)
  		return -ENOMEM;
 
@@ -527,7 +452,6 @@ static int get_derived_key(struct snp_guest_dev *snp_dev, struct snp_guest_reque
 
  static int get_ext_report(struct snp_guest_dev *snp_dev, struct snp_guest_request_ioctl *arg)
  {
-	struct snp_guest_crypto *crypto = snp_dev->crypto;
  	struct snp_ext_report_req req;
  	struct snp_report_resp *resp;
  	int ret, npages = 0, resp_len;
@@ -565,7 +489,7 @@ static int get_ext_report(struct snp_guest_dev *snp_dev, struct snp_guest_reques
  	 * response payload. Make sure that it has enough space to cover the
  	 * authtag.
  	 */
-	resp_len = sizeof(resp->data) + crypto->a_len;
+	resp_len = sizeof(resp->data) + get_ctx_authsize(snp_dev);
  	resp = kzalloc(resp_len, GFP_KERNEL_ACCOUNT);
  	if (!resp)
  		return -ENOMEM;
@@ -777,8 +701,8 @@ static int __init sev_guest_probe(struct platform_device *pdev)
  		goto e_free_response;
 
  	ret = -EIO;
-	snp_dev->crypto = init_crypto(snp_dev, snp_dev->vmpck, VMPCK_KEY_LEN);
-	if (!snp_dev->crypto)
+	snp_dev->ctx = snp_init_crypto(snp_dev->vmpck, VMPCK_KEY_LEN);
+	if (!snp_dev->ctx)
  		goto e_free_cert_data;
 
  	misc = &snp_dev->misc;
@@ -793,11 +717,13 @@ static int __init sev_guest_probe(struct platform_device *pdev)
 
  	ret =  misc_register(misc);
  	if (ret)
-		goto e_free_cert_data;
+		goto e_free_ctx;
 
  	dev_info(dev, "Initialized SEV guest driver (using vmpck_id %d)\n", vmpck_id);
  	return 0;
 
+e_free_ctx:
+	kfree(snp_dev->ctx);
  e_free_cert_data:
  	free_shared_pages(snp_dev->certs_data, SEV_FW_BLOB_MAX_SIZE);
  e_free_response:
@@ -816,7 +742,7 @@ static int __exit sev_guest_remove(struct platform_device *pdev)
  	free_shared_pages(snp_dev->certs_data, SEV_FW_BLOB_MAX_SIZE);
  	free_shared_pages(snp_dev->response, sizeof(struct snp_guest_msg));
  	free_shared_pages(snp_dev->request, sizeof(struct snp_guest_msg));
-	deinit_crypto(snp_dev->crypto);
+	kfree(snp_dev->ctx);
  	misc_deregister(&snp_dev->misc);
 
  	return 0;
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 {
#3 lore raw

Re: [PATCH v2 01/11] virt: sev-guest: Use AES GCM crypto library

Quoted 11 lines · Tom Lendacky
Quoted 7 lines · Nikunj A Dadhania

SEV-SNP guests with SecureTSC enabled need to send a TSC_INFO SNP Guest message to the AMD security processor before the smpboot phase starts. Details from the TSC_INFO response have to be programmed in the VMSA before the secondary CPUs are brought up.

Start using the AES GCM library implementation as the crypto API is not available yet this early.

This isn't quite true, yet. You should add that the encryption code will
be moved out of sev-guest to support Secure TSC, but to make the diffs> easier to review, convert the Crypto API usage over to AES GCM library > usage before moving it.

Yes, will change it accordingly.

Regards
Nikunj

#4 lore raw

PATCH v2 2/11 virt: sev-guest: Move mutex to SNP guest device structure

In reply to Nikunj A Dadhania

In preparation to provide new API for sending SNP guest message to the sev-guest driver, 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 57af908bafba..6ae197b57644 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;
 	struct snp_guest_msg *request, *response;
@@ -48,9 +51,6 @@ 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)
@@ -99,7 +99,7 @@ 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;
@@ -377,7 +377,7 @@ 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;
@@ -417,7 +417,7 @@ 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;
@@ -456,7 +456,7 @@ 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;
@@ -541,12 +541,12 @@ 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;
 	}
 
@@ -564,7 +564,7 @@ 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.fw_err && copy_to_user(argp, &input, sizeof(input)))
 		return -EFAULT;
@@ -683,6 +683,7 @@ 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
Reviewed-by on v2 2/11
#5 lore raw

Re: [PATCH v2 02/11] virt: sev-guest: Move mutex to SNP guest device structure

Quoted 3 lines · Nikunj A Dadhania

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

In preparation for providing a new API to the sev-guest driver for sending an SNP guest message, ...

Quoted 2 lines
Quoted 96 lines
---
  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 57af908bafba..6ae197b57644 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;
  	struct snp_guest_msg *request, *response;
@@ -48,9 +51,6 @@ 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)
@@ -99,7 +99,7 @@ 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;
@@ -377,7 +377,7 @@ 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;
@@ -417,7 +417,7 @@ 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;
@@ -456,7 +456,7 @@ 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;
@@ -541,12 +541,12 @@ 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;
  	}
 
@@ -564,7 +564,7 @@ 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.fw_err && copy_to_user(argp, &input, sizeof(input)))
  		return -EFAULT;
@@ -683,6 +683,7 @@ 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;
#6 lore raw

Re: [PATCH v2 02/11] virt: sev-guest: Move mutex to SNP guest device structure

Quoted 6 lines · Tom Lendacky
Quoted 3 lines · Nikunj A Dadhania

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

In preparation for providing a new API to the sev-guest driver for sending an SNP guest message, ...

Sure.

Quoted 5 lines
Quoted 2 lines

Regards
Nikunj

#7 lore raw

PATCH v2 3/11 virt: sev-guest: Add snp_guest_req 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.

---
 drivers/virt/coco/sev-guest/sev-guest.c | 87 ++++++++++++++-----------
 drivers/virt/coco/sev-guest/sev-guest.h | 19 ++++++
 2 files changed, 68 insertions(+), 38 deletions(-)
drivers/virt/coco/sev-guest/sev-guest.c +49 -38
index 6ae197b57644..ec93dee330f2 100644
@@ -60,16 +60,6 @@ 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)
-{
-	char zero_key[VMPCK_KEY_LEN] = {0};
-
-	if (snp_dev->vmpck)
-		return !memcmp(snp_dev->vmpck, zero_key, VMPCK_KEY_LEN);
-
-	return true;
-}
-
 /*
  * If an error is received from the host or AMD Secure Processor (ASP) there
  * are two options. Either retry the exact same encrypted request or discontinue
@@ -198,8 +188,9 @@ 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);
 
 	/* Verify that the sequence counter is incremented by 1 */
 	if (unlikely(resp_hdr->msg_seqno != (req_hdr->msg_seqno + 1)))
@@ -221,34 +212,34 @@ static int verify_and_dec_payload(struct snp_guest_dev *snp_dev, void *payload,
 	return dec_payload(ctx, resp, payload, resp_hdr->msg_sz);
 }
 
-static int enc_payload(struct snp_guest_dev *snp_dev, u64 seqno, int version, u8 type,
-			void *payload, size_t sz)
+static int enc_payload(struct snp_guest_dev *snp_dev, u64 seqno,
+		       struct snp_guest_req *req, u8 __vmpck_id)
 {
-	struct snp_guest_msg *req = snp_dev->request;
-	struct snp_guest_msg_hdr *hdr = &req->hdr;
+	struct snp_guest_msg *msg = snp_dev->request;
+	struct snp_guest_msg_hdr *hdr = &msg->hdr;
 
-	memset(req, 0, sizeof(*req));
+	memset(msg, 0, sizeof(*msg));
 
 	hdr->algo = SNP_AEAD_AES_256_GCM;
 	hdr->hdr_version = MSG_HDR_VER;
 	hdr->hdr_sz = sizeof(*hdr);
-	hdr->msg_type = type;
-	hdr->msg_version = version;
+	hdr->msg_type = req->msg_type;
+	hdr->msg_version = req->msg_version;
 	hdr->msg_seqno = seqno;
-	hdr->msg_vmpck = vmpck_id;
-	hdr->msg_sz = sz;
+	hdr->msg_vmpck = __vmpck_id;
+	hdr->msg_sz = req->req_sz;
 
 	/* Verify the sequence number is non-zero */
 	if (!hdr->msg_seqno)
 		return -ENOSR;
 
-	dev_dbg(snp_dev->dev, "request [seqno %lld type %d version %d sz %d]\n",
+	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);
+	return __enc_payload(snp_dev->ctx, msg, req->req_buf, req->req_sz);
 }
 
-static int __handle_guest_request(struct snp_guest_dev *snp_dev, u64 exit_code, __u64 *fw_err)
+static int __handle_guest_request(struct snp_guest_dev *snp_dev, struct snp_guest_req *req)
 {
 	unsigned long err = 0xff, override_err = 0;
 	unsigned long req_start = jiffies;
@@ -262,7 +253,7 @@ static int __handle_guest_request(struct snp_guest_dev *snp_dev, u64 exit_code,
 	 * sequence number must be incremented or the VMPCK must be deleted to
 	 * prevent reuse of the IV.
 	 */
-	rc = snp_issue_guest_request(exit_code, &snp_dev->input, &err);
+	rc = snp_issue_guest_request(req->exit_code, &snp_dev->input, &err);
 	switch (rc) {
 	case -ENOSPC:
 		/*
@@ -273,7 +264,7 @@ static int __handle_guest_request(struct snp_guest_dev *snp_dev, u64 exit_code,
 		 * IV reuse.
 		 */
 		override_npages = snp_dev->input.data_npages;
-		exit_code	= SVM_VMGEXIT_GUEST_REQUEST;
+		req->exit_code	= SVM_VMGEXIT_GUEST_REQUEST;
 
 		/*
 		 * Override the error to inform callers the given extended
@@ -314,8 +305,8 @@ static int __handle_guest_request(struct snp_guest_dev *snp_dev, u64 exit_code,
 	 */
 	snp_inc_msg_seqno(snp_dev);
 
-	if (fw_err)
-		*fw_err = override_err ?: err;
+	if (req->fw_err)
+		*req->fw_err = override_err ?: err;
 
 	if (override_npages)
 		snp_dev->input.data_npages = override_npages;
@@ -332,13 +323,14 @@ static int __handle_guest_request(struct snp_guest_dev *snp_dev, u64 exit_code,
 	return rc;
 }
 
-static int handle_guest_request(struct snp_guest_dev *snp_dev, u64 exit_code, int msg_ver,
-				u8 type, void *req_buf, size_t req_sz, void *resp_buf,
-				u32 resp_sz, __u64 *fw_err)
+static int snp_send_guest_request(struct snp_guest_dev *snp_dev, struct snp_guest_req *req)
 {
 	u64 seqno;
 	int rc;
 
+	if (!snp_dev || !req)
+		return -ENODEV;
+
 	/* Get message sequence and verify that its a non-zero */
 	seqno = snp_get_msg_seqno(snp_dev);
 	if (!seqno)
@@ -347,21 +339,22 @@ static int handle_guest_request(struct snp_guest_dev *snp_dev, u64 exit_code, in
 	memset(snp_dev->response, 0, sizeof(struct snp_guest_msg));
 
 	/* Encrypt the userspace provided payload */
-	rc = enc_payload(snp_dev, seqno, msg_ver, type, req_buf, req_sz);
+	rc = enc_payload(snp_dev, seqno, req, vmpck_id);
 	if (rc)
 		return rc;
 
-	rc = __handle_guest_request(snp_dev, exit_code, fw_err);
+	rc = __handle_guest_request(snp_dev, req);
 	if (rc) {
-		if (rc == -EIO && *fw_err == SNP_GUEST_REQ_INVALID_LEN)
+		if (rc == -EIO && *req->fw_err == SNP_GUEST_REQ_INVALID_LEN)
 			return rc;
 
-		dev_alert(snp_dev->dev, "Detected error from ASP request. rc: %d, fw_err: %llu\n", rc, *fw_err);
+		dev_alert(snp_dev->dev, "Detected error from ASP request. rc: %d, fw_err: %llu\n",
+			  rc, *req->fw_err);
 		snp_disable_vmpck(snp_dev);
 		return rc;
 	}
 
-	rc = verify_and_dec_payload(snp_dev, resp_buf, resp_sz);
+	rc = verify_and_dec_payload(snp_dev, req->resp_buf, req->resp_sz);
 	if (rc) {
 		dev_alert(snp_dev->dev, "Detected unexpected decode failure from ASP. rc: %d\n", rc);
 		snp_disable_vmpck(snp_dev);
@@ -371,6 +364,24 @@ static int handle_guest_request(struct snp_guest_dev *snp_dev, u64 exit_code, in
 	return 0;
 }
 
+
+static int handle_guest_request(struct snp_guest_dev *snp_dev, u64 exit_code, u8 msg_version,
+				u8 msg_type, void *req_buf, size_t req_sz, void *resp_buf,
+				u32 resp_sz, __u64 *fw_err)
+{
+	struct snp_guest_req guest_req = {
+		.msg_version = msg_version,
+		.msg_type = msg_type,
+		.req_buf = req_buf,
+		.req_sz = req_sz,
+		.resp_buf = resp_buf,
+		.resp_sz = resp_sz,
+		.fw_err = fw_err,
+		.exit_code = exit_code,
+	};
+	return snp_send_guest_request(snp_dev, &guest_req);
+}
+
 static int get_report(struct snp_guest_dev *snp_dev, struct snp_guest_request_ioctl *arg)
 {
 	struct snp_report_resp *resp;
@@ -544,7 +555,7 @@ 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 (is_vmpck_empty(snp_dev->vmpck)) {
 		dev_err_ratelimited(snp_dev->dev, "VMPCK is disabled\n");
 		mutex_unlock(&snp_dev->cmd_mutex);
 		return -ENOTTY;
@@ -678,7 +689,7 @@ static int __init sev_guest_probe(struct platform_device *pdev)
 	}
 
 	/* Verify that VMPCK is not zero. */
-	if (is_vmpck_empty(snp_dev)) {
+	if (is_vmpck_empty(snp_dev->vmpck)) {
 		dev_err(dev, "vmpck id %d is null\n", vmpck_id);
 		goto e_unmap;
 	}
drivers/virt/coco/sev-guest/sev-guest.h +19 -0
index ceb798a404d6..d245578d988e 100644
@@ -63,4 +63,23 @@ struct snp_guest_msg {
 	u8 payload[4000];
 } __packed;
 
+struct snp_guest_req {
+	void *req_buf, *resp_buf;
+	size_t req_sz, resp_sz;
+	u64 exit_code;
+	u64 *fw_err;
+	u8 msg_version;
+	u8 msg_type;
+};
+
+static inline bool is_vmpck_empty(u8 *vmpck)
+{
+	char zero_key[VMPCK_KEY_LEN] = {0};
+
+	if (vmpck)
+		return !memcmp(vmpck, zero_key, VMPCK_KEY_LEN);
+
+	return true;
+}
+
 #endif /* __VIRT_SEVGUEST_H__ */
Signature · 1 line
2.34.1
#8 lore raw

Re: [PATCH v2 03/11] virt: sev-guest: Add snp_guest_req structure

Quoted 18 lines · 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.

---
  drivers/virt/coco/sev-guest/sev-guest.c | 87 ++++++++++++++-----------
  drivers/virt/coco/sev-guest/sev-guest.h | 19 ++++++
  2 files changed, 68 insertions(+), 38 deletions(-)
drivers/virt/coco/sev-guest/sev-guest.c +0 -0
index 6ae197b57644..ec93dee330f2 100644
@@ -60,16 +60,6 @@ static inline unsigned int get_ctx_authsize(struct snp_guest_dev *snp_dev)
  	return 0;
  }
 

drivers/virt/coco/sev-guest/sev-guest.c static inline unsigned int get_ctx_authsize(struct snp_guest_dev *snp_dev) · line 72

Quoted 10 lines
drivers/virt/coco/sev-guest/sev-guest.c +0 -10
-static bool is_vmpck_empty(struct snp_guest_dev *snp_dev)
-{
-	char zero_key[VMPCK_KEY_LEN] = {0};
-
-	if (snp_dev->vmpck)
-		return !memcmp(snp_dev->vmpck, zero_key, VMPCK_KEY_LEN);
-
-	return true;
-}
-

This change seems separate from the changes for snp_guest_req.

drivers/virt/coco/sev-guest/sev-guest.c static int verify_and_dec_payload(struct snp_guest_dev *snp_dev, void *payload, · line 193

Quoted 12 lines
drivers/virt/coco/sev-guest/sev-guest.c +3 -2
  /*
   * If an error is received from the host or AMD Secure Processor (ASP) there
   * are two options. Either retry the exact same encrypted request or discontinue
@@ -198,8 +188,9 @@ 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);

Again, not related to the purpose of this patch.

drivers/virt/coco/sev-guest/sev-guest.c static int verify_and_dec_payload(struct snp_guest_dev *snp_dev, void *payload, · line 216

Quoted 11 lines
drivers/virt/coco/sev-guest/sev-guest.c +2 -2
  	/* Verify that the sequence counter is incremented by 1 */
  	if (unlikely(resp_hdr->msg_seqno != (req_hdr->msg_seqno + 1)))
@@ -221,34 +212,34 @@ static int verify_and_dec_payload(struct snp_guest_dev *snp_dev, void *payload,
  	return dec_payload(ctx, resp, payload, resp_hdr->msg_sz);
  }
 
-static int enc_payload(struct snp_guest_dev *snp_dev, u64 seqno, int version, u8 type,
-			void *payload, size_t sz)
+static int enc_payload(struct snp_guest_dev *snp_dev, u64 seqno,
+		       struct snp_guest_req *req, u8 __vmpck_id)

Can the vmpck_id be part of the snp_guest_req structure?

Quoted 18 lines
drivers/virt/coco/sev-guest/sev-guest.c +5 -6
  {
-	struct snp_guest_msg *req = snp_dev->request;
-	struct snp_guest_msg_hdr *hdr = &req->hdr;
+	struct snp_guest_msg *msg = snp_dev->request;
+	struct snp_guest_msg_hdr *hdr = &msg->hdr;
 
-	memset(req, 0, sizeof(*req));
+	memset(msg, 0, sizeof(*msg));
 
  	hdr->algo = SNP_AEAD_AES_256_GCM;
  	hdr->hdr_version = MSG_HDR_VER;
  	hdr->hdr_sz = sizeof(*hdr);
-	hdr->msg_type = type;
-	hdr->msg_version = version;
+	hdr->msg_type = req->msg_type;
+	hdr->msg_version = req->msg_version;
  	hdr->msg_seqno = seqno;
-	hdr->msg_vmpck = vmpck_id;

drivers/virt/coco/sev-guest/sev-guest.c static int verify_and_dec_payload(struct snp_guest_dev *snp_dev, void *payload, · line 236

Quoted 10 lines
drivers/virt/coco/sev-guest/sev-guest.c +3 -2
-	hdr->msg_sz = sz;
+	hdr->msg_vmpck = __vmpck_id;
+	hdr->msg_sz = req->req_sz;
 
  	/* Verify the sequence number is non-zero */
  	if (!hdr->msg_seqno)
  		return -ENOSR;
 
-	dev_dbg(snp_dev->dev, "request [seqno %lld type %d version %d sz %d]\n",
+	pr_debug("request [seqno %lld type %d version %d sz %d]\n",

Unrelated change.

Quoted 44 lines
drivers/virt/coco/sev-guest/sev-guest.c +6 -6
  		hdr->msg_seqno, hdr->msg_type, hdr->msg_version, hdr->msg_sz);
 
-	return __enc_payload(snp_dev->ctx, req, payload, sz);
+	return __enc_payload(snp_dev->ctx, msg, req->req_buf, req->req_sz);
  }
 
-static int __handle_guest_request(struct snp_guest_dev *snp_dev, u64 exit_code, __u64 *fw_err)
+static int __handle_guest_request(struct snp_guest_dev *snp_dev, struct snp_guest_req *req)
  {
  	unsigned long err = 0xff, override_err = 0;
  	unsigned long req_start = jiffies;
@@ -262,7 +253,7 @@ static int __handle_guest_request(struct snp_guest_dev *snp_dev, u64 exit_code,
  	 * sequence number must be incremented or the VMPCK must be deleted to
  	 * prevent reuse of the IV.
  	 */
-	rc = snp_issue_guest_request(exit_code, &snp_dev->input, &err);
+	rc = snp_issue_guest_request(req->exit_code, &snp_dev->input, &err);
  	switch (rc) {
  	case -ENOSPC:
  		/*
@@ -273,7 +264,7 @@ static int __handle_guest_request(struct snp_guest_dev *snp_dev, u64 exit_code,
  		 * IV reuse.
  		 */
  		override_npages = snp_dev->input.data_npages;
-		exit_code	= SVM_VMGEXIT_GUEST_REQUEST;
+		req->exit_code	= SVM_VMGEXIT_GUEST_REQUEST;
 
  		/*
  		 * Override the error to inform callers the given extended
@@ -314,8 +305,8 @@ static int __handle_guest_request(struct snp_guest_dev *snp_dev, u64 exit_code,
  	 */
  	snp_inc_msg_seqno(snp_dev);
 
-	if (fw_err)
-		*fw_err = override_err ?: err;
+	if (req->fw_err)
+		*req->fw_err = override_err ?: err;
 
  	if (override_npages)
  		snp_dev->input.data_npages = override_npages;
@@ -332,13 +323,14 @@ static int __handle_guest_request(struct snp_guest_dev *snp_dev, u64 exit_code,
  	return rc;
  }
 

drivers/virt/coco/sev-guest/sev-guest.c static int __handle_guest_request(struct snp_guest_dev *snp_dev, u64 exit_code, · line 332

Quoted 10 lines
drivers/virt/coco/sev-guest/sev-guest.c +3 -3
-static int handle_guest_request(struct snp_guest_dev *snp_dev, u64 exit_code, int msg_ver,
-				u8 type, void *req_buf, size_t req_sz, void *resp_buf,
-				u32 resp_sz, __u64 *fw_err)
+static int snp_send_guest_request(struct snp_guest_dev *snp_dev, struct snp_guest_req *req)
  {
  	u64 seqno;
  	int rc;
 
+	if (!snp_dev || !req)
+		return -ENODEV;

This seems unrelated, at least the check for snp_dev. And looking at the only caller, a guest request is always provided. So this seems unnecessary - at least at this point in the series.

Quoted 19 lines
drivers/virt/coco/sev-guest/sev-guest.c +4 -3
+
  	/* Get message sequence and verify that its a non-zero */
  	seqno = snp_get_msg_seqno(snp_dev);
  	if (!seqno)
@@ -347,21 +339,22 @@ static int handle_guest_request(struct snp_guest_dev *snp_dev, u64 exit_code, in
  	memset(snp_dev->response, 0, sizeof(struct snp_guest_msg));
 
  	/* Encrypt the userspace provided payload */
-	rc = enc_payload(snp_dev, seqno, msg_ver, type, req_buf, req_sz);
+	rc = enc_payload(snp_dev, seqno, req, vmpck_id);
  	if (rc)
  		return rc;
 
-	rc = __handle_guest_request(snp_dev, exit_code, fw_err);
+	rc = __handle_guest_request(snp_dev, req);
  	if (rc) {
-		if (rc == -EIO && *fw_err == SNP_GUEST_REQ_INVALID_LEN)
+		if (rc == -EIO && *req->fw_err == SNP_GUEST_REQ_INVALID_LEN)
  			return rc;

drivers/virt/coco/sev-guest/sev-guest.c static int handle_guest_request(struct snp_guest_dev *snp_dev, u64 exit_code, in · line 357

Quoted 10 lines
drivers/virt/coco/sev-guest/sev-guest.c +3 -2
 
-		dev_alert(snp_dev->dev, "Detected error from ASP request. rc: %d, fw_err: %llu\n", rc, *fw_err);
+		dev_alert(snp_dev->dev, "Detected error from ASP request. rc: %d, fw_err: %llu\n",
+			  rc, *req->fw_err);
  		snp_disable_vmpck(snp_dev);
  		return rc;
  	}
 
-	rc = verify_and_dec_payload(snp_dev, resp_buf, resp_sz);
+	rc = verify_and_dec_payload(snp_dev, req->resp_buf, req->resp_sz);

Can't you just pass req here?

Quoted 12 lines
drivers/virt/coco/sev-guest/sev-guest.c +5 -0
  	if (rc) {
  		dev_alert(snp_dev->dev, "Detected unexpected decode failure from ASP. rc: %d\n", rc);
  		snp_disable_vmpck(snp_dev);
@@ -371,6 +364,24 @@ static int handle_guest_request(struct snp_guest_dev *snp_dev, u64 exit_code, in
  	return 0;
  }
 
+
+static int handle_guest_request(struct snp_guest_dev *snp_dev, u64 exit_code, u8 msg_version,
+				u8 msg_type, void *req_buf, size_t req_sz, void *resp_buf,
+				u32 resp_sz, __u64 *fw_err)
+{

drivers/virt/coco/sev-guest/sev-guest.c static int handle_guest_request(struct snp_guest_dev *snp_dev, u64 exit_code, in · line 381

Quoted 10 lines
drivers/virt/coco/sev-guest/sev-guest.c +10 -0
+	struct snp_guest_req guest_req = {
+		.msg_version = msg_version,
+		.msg_type = msg_type,
+		.req_buf = req_buf,
+		.req_sz = req_sz,
+		.resp_buf = resp_buf,
+		.resp_sz = resp_sz,
+		.fw_err = fw_err,
+		.exit_code = exit_code,
+	};

Add a blank line here.

Thanks,
Tom

Quoted 52 lines
drivers/virt/coco/sev-guest/sev-guest.c +5 -2
+	return snp_send_guest_request(snp_dev, &guest_req);
+}
+
  static int get_report(struct snp_guest_dev *snp_dev, struct snp_guest_request_ioctl *arg)
  {
  	struct snp_report_resp *resp;
@@ -544,7 +555,7 @@ 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 (is_vmpck_empty(snp_dev->vmpck)) {
  		dev_err_ratelimited(snp_dev->dev, "VMPCK is disabled\n");
  		mutex_unlock(&snp_dev->cmd_mutex);
  		return -ENOTTY;
@@ -678,7 +689,7 @@ static int __init sev_guest_probe(struct platform_device *pdev)
  	}
 
  	/* Verify that VMPCK is not zero. */
-	if (is_vmpck_empty(snp_dev)) {
+	if (is_vmpck_empty(snp_dev->vmpck)) {
  		dev_err(dev, "vmpck id %d is null\n", vmpck_id);
  		goto e_unmap;
  	}
drivers/virt/coco/sev-guest/sev-guest.h +19 -0
index ceb798a404d6..d245578d988e 100644
@@ -63,4 +63,23 @@ struct snp_guest_msg {
  	u8 payload[4000];
  } __packed;
 
+struct snp_guest_req {
+	void *req_buf, *resp_buf;
+	size_t req_sz, resp_sz;
+	u64 exit_code;
+	u64 *fw_err;
+	u8 msg_version;
+	u8 msg_type;
+};
+
+static inline bool is_vmpck_empty(u8 *vmpck)
+{
+	char zero_key[VMPCK_KEY_LEN] = {0};
+
+	if (vmpck)
+		return !memcmp(vmpck, zero_key, VMPCK_KEY_LEN);
+
+	return true;
+}
+
  #endif /* __VIRT_SEVGUEST_H__ */
#9 lore raw

Re: [PATCH v2 03/11] virt: sev-guest: Add snp_guest_req structure

Quoted 22 lines · Tom Lendacky
Quoted 21 lines · 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.

---
  drivers/virt/coco/sev-guest/sev-guest.c | 87 ++++++++++++++----------- drivers/virt/coco/sev-guest/sev-guest.h | 19 ++++++

  2 files changed, 68 insertions(+), 38 deletions(-)
drivers/virt/coco/sev-guest/sev-guest.c +0 -3
index 6ae197b57644..ec93dee330f2 100644
@@ -60,16 +60,6 @@ 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)
-{
-    char zero_key[VMPCK_KEY_LEN] = {0};
-
Quoted 8 lines

drivers/virt/coco/sev-guest/sev-guest.c static inline unsigned int get_ctx_authsize(struct snp_guest_dev *snp_dev) · line 72

Quoted 6 lines
drivers/virt/coco/sev-guest/sev-guest.c +0 -6
-    if (snp_dev->vmpck)
-        return !memcmp(snp_dev->vmpck, zero_key, VMPCK_KEY_LEN);
-
-    return true;
-}
-

This change seems separate from the changes for snp_guest_req.

Sure, will create a separate patch.

Quoted 14 lines

drivers/virt/coco/sev-guest/sev-guest.c static int verify_and_dec_payload(struct snp_guest_dev *snp_dev, void *payload, · line 193

Quoted 12 lines
drivers/virt/coco/sev-guest/sev-guest.c +3 -1
   /*
    * If an error is received from the host or AMD Secure Processor (ASP) there
    * are two options. Either retry the exact same encrypted request or discontinue
@@ -198,8 +188,9 @@ 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);

Again, not related to the purpose of this patch.

The idea was to get rid of dev_dbg for the movement, will do in a separate patch.

Quoted 12 lines

drivers/virt/coco/sev-guest/sev-guest.c static int verify_and_dec_payload(struct snp_guest_dev *snp_dev, void *payload, · line 216

Quoted 10 lines
drivers/virt/coco/sev-guest/sev-guest.c +2 -1
         /* Verify that the sequence counter is incremented by 1 */
       if (unlikely(resp_hdr->msg_seqno != (req_hdr->msg_seqno + 1)))
@@ -221,34 +212,34 @@ static int verify_and_dec_payload(struct snp_guest_dev *snp_dev, void *payload,
       return dec_payload(ctx, resp, payload, resp_hdr->msg_sz);
   }
   -static int enc_payload(struct snp_guest_dev *snp_dev, u64 seqno, int version, u8 type,
-            void *payload, size_t sz)
+static int enc_payload(struct snp_guest_dev *snp_dev, u64 seqno,
+               struct snp_guest_req *req, u8 __vmpck_id)

Can the vmpck_id be part of the snp_guest_req structure?

Sure, that should be possible.

Quoted 26 lines
Quoted 15 lines
drivers/virt/coco/sev-guest/sev-guest.c +5 -4
   {
-    struct snp_guest_msg *req = snp_dev->request;
-    struct snp_guest_msg_hdr *hdr = &req->hdr;
+    struct snp_guest_msg *msg = snp_dev->request;
+    struct snp_guest_msg_hdr *hdr = &msg->hdr;
   -    memset(req, 0, sizeof(*req));
+    memset(msg, 0, sizeof(*msg));
         hdr->algo = SNP_AEAD_AES_256_GCM;
       hdr->hdr_version = MSG_HDR_VER;
       hdr->hdr_sz = sizeof(*hdr);
-    hdr->msg_type = type;
-    hdr->msg_version = version;
+    hdr->msg_type = req->msg_type;
+    hdr->msg_version = req->msg_version;

drivers/virt/coco/sev-guest/sev-guest.c static int verify_and_dec_payload(struct snp_guest_dev *snp_dev, void *payload, · line 236

Quoted 10 lines
drivers/virt/coco/sev-guest/sev-guest.c +3 -2
       hdr->msg_seqno = seqno;
-    hdr->msg_vmpck = vmpck_id;
-    hdr->msg_sz = sz;
+    hdr->msg_vmpck = __vmpck_id;
+    hdr->msg_sz = req->req_sz;
         /* Verify the sequence number is non-zero */
       if (!hdr->msg_seqno)
           return -ENOSR;
   -    dev_dbg(snp_dev->dev, "request [seqno %lld type %d version %d sz %d]\n",
+    pr_debug("request [seqno %lld type %d version %d sz %d]\n",
Quoted 1 line

Unrelated change.

Will move to a separate patch.

Quoted 42 lines
Quoted 42 lines
drivers/virt/coco/sev-guest/sev-guest.c +3 -1
           hdr->msg_seqno, hdr->msg_type, hdr->msg_version, hdr->msg_sz);
   -    return __enc_payload(snp_dev->ctx, req, payload, sz);
+    return __enc_payload(snp_dev->ctx, msg, req->req_buf, req->req_sz);
   }
   -static int __handle_guest_request(struct snp_guest_dev *snp_dev, u64 exit_code, __u64 *fw_err)
+static int __handle_guest_request(struct snp_guest_dev *snp_dev, struct snp_guest_req *req)
   {
       unsigned long err = 0xff, override_err = 0;
       unsigned long req_start = jiffies;
@@ -262,7 +253,7 @@ static int __handle_guest_request(struct snp_guest_dev *snp_dev, u64 exit_code,
        * sequence number must be incremented or the VMPCK must be deleted to
        * prevent reuse of the IV.
        */
-    rc = snp_issue_guest_request(exit_code, &snp_dev->input, &err);
+    rc = snp_issue_guest_request(req->exit_code, &snp_dev->input, &err);
       switch (rc) {

      case -ENOSPC:
          /*

drivers/virt/coco/sev-guest/sev-guest.c +3 -4
@@ -273,7 +264,7 @@ static int __handle_guest_request(struct snp_guest_dev *snp_dev, u64 exit_code,
            * IV reuse.
            */
           override_npages = snp_dev->input.data_npages;
-        exit_code    = SVM_VMGEXIT_GUEST_REQUEST;
+        req->exit_code    = SVM_VMGEXIT_GUEST_REQUEST;
             /*
            * Override the error to inform callers the given extended
@@ -314,8 +305,8 @@ static int __handle_guest_request(struct snp_guest_dev *snp_dev, u64 exit_code,
        */
       snp_inc_msg_seqno(snp_dev);
   -    if (fw_err)
-        *fw_err = override_err ?: err;
+    if (req->fw_err)
+        *req->fw_err = override_err ?: err;
         if (override_npages)
           snp_dev->input.data_npages = override_npages;
@@ -332,13 +323,14 @@ static int __handle_guest_request(struct snp_guest_dev *snp_dev, u64 exit_code,
       return rc;
   }
   -static int handle_guest_request(struct snp_guest_dev *snp_dev, u64 exit_code, int msg_ver,
-                u8 type, void *req_buf, size_t req_sz, void *resp_buf,
-                u32 resp_sz, __u64 *fw_err)
Quoted 8 lines

drivers/virt/coco/sev-guest/sev-guest.c static int __handle_guest_request(struct snp_guest_dev *snp_dev, u64 exit_code, · line 332

Quoted 6 lines
drivers/virt/coco/sev-guest/sev-guest.c +2 -0
+static int snp_send_guest_request(struct snp_guest_dev *snp_dev, struct snp_guest_req *req)
   {
       u64 seqno;
       int rc;
   +    if (!snp_dev || !req)
+        return -ENODEV;

This seems unrelated, at least the check for snp_dev. And looking at the only caller, a guest request is always provided. So this seems unnecessary - at least at this point in the series.

Right, not necessary here, but will be needed when sev-guest driver calls this after the movement to sev.c. Otherwise, I will need to add this in the movement patch 5/11.

Quoted 20 lines
Quoted 20 lines
drivers/virt/coco/sev-guest/sev-guest.c +5 -2
+
       /* Get message sequence and verify that its a non-zero */
       seqno = snp_get_msg_seqno(snp_dev);
       if (!seqno)
@@ -347,21 +339,22 @@ static int handle_guest_request(struct snp_guest_dev *snp_dev, u64 exit_code, in
       memset(snp_dev->response, 0, sizeof(struct snp_guest_msg));
         /* Encrypt the userspace provided payload */
-    rc = enc_payload(snp_dev, seqno, msg_ver, type, req_buf, req_sz);
+    rc = enc_payload(snp_dev, seqno, req, vmpck_id);
       if (rc)
           return rc;
   -    rc = __handle_guest_request(snp_dev, exit_code, fw_err);
+    rc = __handle_guest_request(snp_dev, req);
       if (rc) {
-        if (rc == -EIO && *fw_err == SNP_GUEST_REQ_INVALID_LEN)
+        if (rc == -EIO && *req->fw_err == SNP_GUEST_REQ_INVALID_LEN)
               return rc;
   -        dev_alert(snp_dev->dev, "Detected error from ASP request. rc: %d, fw_err: %llu\n", rc, *fw_err);
+        dev_alert(snp_dev->dev, "Detected error from ASP request. rc: %d, fw_err: %llu\n",
Quoted 8 lines

drivers/virt/coco/sev-guest/sev-guest.c static int handle_guest_request(struct snp_guest_dev *snp_dev, u64 exit_code, in · line 357

Quoted 6 lines
drivers/virt/coco/sev-guest/sev-guest.c +2 -0
+              rc, *req->fw_err);
           snp_disable_vmpck(snp_dev);
           return rc;
       }
   -    rc = verify_and_dec_payload(snp_dev, resp_buf, resp_sz);
+    rc = verify_and_dec_payload(snp_dev, req->resp_buf, req->resp_sz);

Can't you just pass req here?

Yes, can do that.

Quoted 16 lines
Quoted 16 lines
drivers/virt/coco/sev-guest/sev-guest.c +0 -0
       if (rc) {
           dev_alert(snp_dev->dev, "Detected unexpected decode failure from ASP. rc: %d\n", rc);
           snp_disable_vmpck(snp_dev);
@@ -371,6 +364,24 @@ static int handle_guest_request(struct snp_guest_dev *snp_dev, u64 exit_code, in
       return 0;
   }

  +

drivers/virt/coco/sev-guest/sev-guest.c +8 -0
+static int handle_guest_request(struct snp_guest_dev *snp_dev, u64 exit_code, u8 msg_version,
+                u8 msg_type, void *req_buf, size_t req_sz, void *resp_buf,
+                u32 resp_sz, __u64 *fw_err)
+{
+    struct snp_guest_req guest_req = {
+        .msg_version = msg_version,
+        .msg_type = msg_type,
+        .req_buf = req_buf,
Quoted 8 lines
Quoted 6 lines

+        .req_sz = req_sz,
+        .resp_buf = resp_buf,
+        .resp_sz = resp_sz,
+        .fw_err = fw_err,
+        .exit_code = exit_code,
+    };

Add a blank line here.

Sure.

Regards
Nikunj

#10 lore raw

PATCH v2 4/11 virt: sev-guest: Add simplified helper to assign vmpck

In reply to Nikunj A Dadhania

Simplify get_vmpck and prepare it to be used as an API. Update the snp_guest_dev structure in snp_assign_vmpck(). Added vmpck_id to the snp_guest_dev structure which can be used in SNP guest request API and will remove direct use of vmpck_id command line parameter.

---
 drivers/virt/coco/sev-guest/sev-guest.c | 41 ++++++++-----------------
 1 file changed, 12 insertions(+), 29 deletions(-)
drivers/virt/coco/sev-guest/sev-guest.c +12 -29
index ec93dee330f2..4901ebc8fa1a 100644
@@ -45,6 +45,7 @@ struct snp_guest_dev {
 	struct snp_req_data input;
 	u32 *os_area_msg_seqno;
 	u8 *vmpck;
+	u8 vmpck_id;
 };
 
 static u32 vmpck_id;
@@ -80,7 +81,7 @@ static inline unsigned int get_ctx_authsize(struct snp_guest_dev *snp_dev)
 static void snp_disable_vmpck(struct snp_guest_dev *snp_dev)
 {
 	dev_alert(snp_dev->dev, "Disabling vmpck_id %d to prevent IV reuse.\n",
-		  vmpck_id);
+		  snp_dev->vmpck_id);
 	memzero_explicit(snp_dev->vmpck, VMPCK_KEY_LEN);
 	snp_dev->vmpck = NULL;
 }
@@ -339,7 +340,7 @@ static int snp_send_guest_request(struct snp_guest_dev *snp_dev, struct snp_gues
 	memset(snp_dev->response, 0, sizeof(struct snp_guest_msg));
 
 	/* Encrypt the userspace provided payload */
-	rc = enc_payload(snp_dev, seqno, req, vmpck_id);
+	rc = enc_payload(snp_dev, seqno, req, snp_dev->vmpck_id);
 	if (rc)
 		return rc;
 
@@ -364,7 +365,6 @@ static int snp_send_guest_request(struct snp_guest_dev *snp_dev, struct snp_gues
 	return 0;
 }
 
-
 static int handle_guest_request(struct snp_guest_dev *snp_dev, u64 exit_code, u8 msg_version,
 				u8 msg_type, void *req_buf, size_t req_sz, void *resp_buf,
 				u32 resp_sz, __u64 *fw_err)
@@ -625,32 +625,16 @@ 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;
+	dev->vmpck = dev->layout->vmpck0 + vmpck_id * VMPCK_KEY_LEN;
+	dev->os_area_msg_seqno = &dev->layout->os_area.msg_seqno_0 + vmpck_id;
 
-	return key;
+	return true;
 }
 
 static int __init sev_guest_probe(struct platform_device *pdev)
@@ -682,8 +666,8 @@ 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;
 	}
@@ -697,7 +681,6 @@ 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));
Signature · 1 line
2.34.1
#11 lore raw

Re: [PATCH v2 04/11] virt: sev-guest: Add simplified helper to assign vmpck

Quoted 4 lines · Nikunj A Dadhania

Simplify get_vmpck and prepare it to be used as an API. Update the snp_guest_dev structure in snp_assign_vmpck(). Added vmpck_id to the snp_guest_dev structure which can be used in SNP guest request API and will remove direct use of vmpck_id command line parameter.

This needs to say more about why you're doing this. The first two sentences need more explanation as they don't really stand on their own.

The last sentence talks about removing direct use of vmpck_id, so you should probably do that in this patch since it seems to be the real reason for all this.

Thanks,
Tom

Quoted 103 lines
---
  drivers/virt/coco/sev-guest/sev-guest.c | 41 ++++++++-----------------
  1 file changed, 12 insertions(+), 29 deletions(-)
drivers/virt/coco/sev-guest/sev-guest.c +12 -29
index ec93dee330f2..4901ebc8fa1a 100644
@@ -45,6 +45,7 @@ struct snp_guest_dev {
  	struct snp_req_data input;
  	u32 *os_area_msg_seqno;
  	u8 *vmpck;
+	u8 vmpck_id;
  };
 
  static u32 vmpck_id;
@@ -80,7 +81,7 @@ static inline unsigned int get_ctx_authsize(struct snp_guest_dev *snp_dev)
  static void snp_disable_vmpck(struct snp_guest_dev *snp_dev)
  {
  	dev_alert(snp_dev->dev, "Disabling vmpck_id %d to prevent IV reuse.\n",
-		  vmpck_id);
+		  snp_dev->vmpck_id);
  	memzero_explicit(snp_dev->vmpck, VMPCK_KEY_LEN);
  	snp_dev->vmpck = NULL;
  }
@@ -339,7 +340,7 @@ static int snp_send_guest_request(struct snp_guest_dev *snp_dev, struct snp_gues
  	memset(snp_dev->response, 0, sizeof(struct snp_guest_msg));
 
  	/* Encrypt the userspace provided payload */
-	rc = enc_payload(snp_dev, seqno, req, vmpck_id);
+	rc = enc_payload(snp_dev, seqno, req, snp_dev->vmpck_id);
  	if (rc)
  		return rc;
 
@@ -364,7 +365,6 @@ static int snp_send_guest_request(struct snp_guest_dev *snp_dev, struct snp_gues
  	return 0;
  }
 
-
  static int handle_guest_request(struct snp_guest_dev *snp_dev, u64 exit_code, u8 msg_version,
  				u8 msg_type, void *req_buf, size_t req_sz, void *resp_buf,
  				u32 resp_sz, __u64 *fw_err)
@@ -625,32 +625,16 @@ 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;
+	dev->vmpck = dev->layout->vmpck0 + vmpck_id * VMPCK_KEY_LEN;
+	dev->os_area_msg_seqno = &dev->layout->os_area.msg_seqno_0 + vmpck_id;
 
-	return key;
+	return true;
  }
 
  static int __init sev_guest_probe(struct platform_device *pdev)
@@ -682,8 +666,8 @@ 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;
  	}
@@ -697,7 +681,6 @@ 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));
#12 lore raw

Re: [PATCH v2 04/11] virt: sev-guest: Add simplified helper to assign vmpck

Quoted 9 lines · Tom Lendacky
Quoted 4 lines · Nikunj A Dadhania

Simplify get_vmpck and prepare it to be used as an API. Update the snp_guest_dev structure in snp_assign_vmpck(). Added vmpck_id to the snp_guest_dev structure which can be used in SNP guest request API and will remove direct use of vmpck_id command line parameter.

This needs to say more about why you're doing this. The first two sentences need more explanation as they don't really stand on their own. How about the this:

Add vmpck_id to snp_guest_dev struct instead of directly using vmpck_id command line parameter. Each SNP guest has four VMPCK, which the firmware generates and provides to the guest in a special secrets page. Each VMPCK has a sequence number. Simplify assignment of key and sequence number instead of using a switch case.

Quoted 3 lines

The last sentence talks about removing direct use of vmpck_id, so you should probably do that in this patch since it seems to be the real reason for all this.

I did change all places using vmpck_id to snp_dev->vmpck_id. This is the only place the command line parameter is being used.

+ if (!snp_assign_vmpck(snp_dev, vmpck_id)) {

Regards
Nikunj

Quoted 97 lines

Thanks,
Tom

Quoted 93 lines

---
  drivers/virt/coco/sev-guest/sev-guest.c | 41 ++++++++-----------------

  1 file changed, 12 insertions(+), 29 deletions(-)
drivers/virt/coco/sev-guest/sev-guest.c +3 -2
index ec93dee330f2..4901ebc8fa1a 100644
@@ -45,6 +45,7 @@ struct snp_guest_dev {
       struct snp_req_data input;
       u32 *os_area_msg_seqno;
       u8 *vmpck;
+    u8 vmpck_id;
   };
     static u32 vmpck_id;
@@ -80,7 +81,7 @@ static inline unsigned int get_ctx_authsize(struct snp_guest_dev *snp_dev)
   static void snp_disable_vmpck(struct snp_guest_dev *snp_dev)
   {
       dev_alert(snp_dev->dev, "Disabling vmpck_id %d to prevent IV reuse.\n",
-          vmpck_id);
+          snp_dev->vmpck_id);
       memzero_explicit(snp_dev->vmpck, VMPCK_KEY_LEN);
       snp_dev->vmpck = NULL;
   }
@@ -339,7 +340,7 @@ static int snp_send_guest_request(struct snp_guest_dev *snp_dev, struct snp_gues
       memset(snp_dev->response, 0, sizeof(struct snp_guest_msg));
         /* Encrypt the userspace provided payload */
-    rc = enc_payload(snp_dev, seqno, req, vmpck_id);
+    rc = enc_payload(snp_dev, seqno, req, snp_dev->vmpck_id);
       if (rc)
           return rc;

  @@ -364,7 +365,6 @@ static int snp_send_guest_request(struct snp_guest_dev *snp_dev, struct snp_gues return 0;
  }
  -
  static int handle_guest_request(struct snp_guest_dev *snp_dev, u64 exit_code, u8 msg_version, u8 msg_type, void *req_buf, size_t req_sz, void *resp_buf,
                  u32 resp_sz, __u64 *fw_err)

drivers/virt/coco/sev-guest/sev-guest.c +9 -23
@@ -625,32 +625,16 @@ 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;
+    dev->vmpck = dev->layout->vmpck0 + vmpck_id * VMPCK_KEY_LEN;
+    dev->os_area_msg_seqno = &dev->layout->os_area.msg_seqno_0 + vmpck_id;
   -    return key;
+    return true;
   }
     static int __init sev_guest_probe(struct platform_device *pdev)
@@ -682,8 +666,8 @@ 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;
       }
@@ -697,7 +681,6 @@ 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));
#13 lore raw

Re: [PATCH v2 04/11] virt: sev-guest: Add simplified helper to assign vmpck

In reply to Nikunj A Dadhania

drivers/virt/coco/sev-guest/sev-guest.c static const struct file_operations snp_guest_fops = { · line 631

Quoted 7 lines
drivers/virt/coco/sev-guest/sev-guest.c +3 -2
-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;

Should vmpck_id be a uint to match the module parameter or should we check if vmpck_id < 0 ?

#14 lore raw

Re: [PATCH v2 04/11] virt: sev-guest: Add simplified helper to assign vmpck

Quoted 10 lines · Peter Gonda

drivers/virt/coco/sev-guest/sev-guest.c static const struct file_operations snp_guest_fops = { · line 631

Quoted 7 lines
drivers/virt/coco/sev-guest/sev-guest.c +3 -2
-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;

Should vmpck_id be a uint to match the module parameter or should we check if vmpck_id < 0 ?

Thanks for pointing that out, I will make vmpck_id as a uint.

Regards
Nikunj

#15 lore raw

PATCH v2 5/11 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 security coprocessor really early during boot. Lot of the required functions are implemented in the sev-guest driver. Move the required functions and provide API to the driver to assign VM communications key and send guest request.

---
 arch/x86/Kconfig                              |   1 +
 .../x86/include/asm}/sev-guest.h              |   6 +
 arch/x86/include/asm/sev.h                    |  22 +
 arch/x86/kernel/sev.c                         | 437 ++++++++++++++++-
 drivers/virt/coco/sev-guest/Kconfig           |   1 -
 drivers/virt/coco/sev-guest/sev-guest.c       | 455 ++----------------
 6 files changed, 487 insertions(+), 435 deletions(-)
 rename {drivers/virt/coco/sev-guest => arch/x86/include/asm}/sev-guest.h (86%)
arch/x86/Kconfig +1 -0
index a825bf031f49..42c52bfd2bac 100644
@@ -1545,6 +1545,7 @@ 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 renamed from drivers/virt/coco/sev-guest/sev-guest.h +6 -0
similarity index 86%
rename from drivers/virt/coco/sev-guest/sev-guest.h
rename to arch/x86/include/asm/sev-guest.h
index d245578d988e..834cdae302ad 100644
@@ -12,6 +12,9 @@
 
 #include <linux/types.h>
 
+#define SNP_REQ_MAX_RETRY_DURATION    (60*HZ)
+#define SNP_REQ_RETRY_DELAY           (2*HZ)
+
 #define MAX_AUTHTAG_LEN		32
 #define AUTHTAG_LEN		16
 #define AAD_LEN			48
@@ -72,6 +75,9 @@ struct snp_guest_req {
 	u8 msg_type;
 };
 
+int snp_send_guest_request(struct snp_guest_dev *dev, struct snp_guest_req *req);
+bool snp_assign_vmpck(struct snp_guest_dev *dev, int vmpck_id);
+
 static inline bool is_vmpck_empty(u8 *vmpck)
 {
 	char zero_key[VMPCK_KEY_LEN] = {0};
arch/x86/include/asm/sev.h +22 -0
index ebc271bb6d8e..36868e21c3e0 100644
@@ -9,6 +9,7 @@
 #define __ASM_ENCRYPTED_STATE_H
 
 #include <linux/types.h>
+#include <linux/miscdevice.h>
 #include <asm/insn.h>
 #include <asm/sev-common.h>
 #include <asm/bootparam.h>
@@ -96,6 +97,27 @@ struct snp_req_data {
 
 struct sev_guest_platform_data {
 	u64 secrets_gpa;
+
+	void *certs_data;
+	struct aesgcm_ctx *ctx;
+	struct snp_guest_msg *req, *resp;
+	struct snp_secrets_page_layout *layout;
+	struct snp_req_data input;
+	u8 *vmpck0;
+};
+
+struct snp_guest_dev {
+	struct device *dev;
+	struct miscdevice misc;
+
+	/* Mutex to serialize the shared buffer access and command handling. */
+	struct mutex cmd_mutex;
+
+	struct sev_guest_platform_data *pdata;
+
+	u32 *os_area_msg_seqno;
+	u8 *vmpck;
+	u8 vmpck_id;
 };
 
 /*
arch/x86/kernel/sev.c +430 -7

464 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 +28 -427

573 lines not on this page · raw message

Signature · 1 line
2.34.1
#16 lore raw

Re: [PATCH v2 05/11] x86/sev: Move and reorganize sev guest request api

Quoted 2 lines · Nikunj A Dadhania

For enabling Secure TSC, SEV-SNP guests need to communicate with the security coprocessor really early during boot. Lot of the required

s/security coprocessor really/AMD Secure Processor/

s/Lot/Many/

Quoted 1 line

functions are implemented in the sev-guest driver. Move the required

... in the sev-guest driver and therefore not available at early boot.

Quoted 1 line

functions and provide API to the driver to assign VM communications

s/provide API/provide an API/

s/to assign.*//

Quoted 65 lines

key and send guest request.

---
  arch/x86/Kconfig                              |   1 +
  .../x86/include/asm}/sev-guest.h              |   6 +
  arch/x86/include/asm/sev.h                    |  22 +
  arch/x86/kernel/sev.c                         | 437 ++++++++++++++++-
  drivers/virt/coco/sev-guest/Kconfig           |   1 -
  drivers/virt/coco/sev-guest/sev-guest.c       | 455 ++----------------
  6 files changed, 487 insertions(+), 435 deletions(-)

rename {drivers/virt/coco/sev-guest => arch/x86/include/asm}/sev-guest.h (86%)

arch/x86/Kconfig +1 -0
index a825bf031f49..42c52bfd2bac 100644
@@ -1545,6 +1545,7 @@ 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 renamed from drivers/virt/coco/sev-guest/sev-guest.h +6 -0
similarity index 86%
rename from drivers/virt/coco/sev-guest/sev-guest.h
rename to arch/x86/include/asm/sev-guest.h
index d245578d988e..834cdae302ad 100644
@@ -12,6 +12,9 @@
 
  #include <linux/types.h>
 
+#define SNP_REQ_MAX_RETRY_DURATION    (60*HZ)
+#define SNP_REQ_RETRY_DELAY           (2*HZ)
+
  #define MAX_AUTHTAG_LEN		32
  #define AUTHTAG_LEN		16
  #define AAD_LEN			48
@@ -72,6 +75,9 @@ struct snp_guest_req {
  	u8 msg_type;
  };
 
+int snp_send_guest_request(struct snp_guest_dev *dev, struct snp_guest_req *req);
+bool snp_assign_vmpck(struct snp_guest_dev *dev, int vmpck_id);
+
  static inline bool is_vmpck_empty(u8 *vmpck)
  {
  	char zero_key[VMPCK_KEY_LEN] = {0};
arch/x86/include/asm/sev.h +1 -0
index ebc271bb6d8e..36868e21c3e0 100644
@@ -9,6 +9,7 @@
  #define __ASM_ENCRYPTED_STATE_H
 
  #include <linux/types.h>
+#include <linux/miscdevice.h>
  #include <asm/insn.h>
  #include <asm/sev-common.h>
  #include <asm/bootparam.h>

arch/x86/include/asm/sev.h struct snp_req_data { · line 106

Quoted 10 lines
arch/x86/include/asm/sev.h +7 -0
@@ -96,6 +97,27 @@ struct snp_req_data {
 
  struct sev_guest_platform_data {
  	u64 secrets_gpa;
+
+	void *certs_data;
+	struct aesgcm_ctx *ctx;
+	struct snp_guest_msg *req, *resp;
+	struct snp_secrets_page_layout *layout;
+	struct snp_req_data input;
+	u8 *vmpck0;

Isn't this unneeded? You have the vmpck and vmpck_id in the snp_guest_dev struct which will be set based on the module parameter, so vmpck0 and associated checks shouldn't be needed.

Quoted 470 lines
arch/x86/include/asm/sev.h +14 -0
+};
+
+struct snp_guest_dev {
+	struct device *dev;
+	struct miscdevice misc;
+
+	/* Mutex to serialize the shared buffer access and command handling. */
+	struct mutex cmd_mutex;
+
+	struct sev_guest_platform_data *pdata;
+
+	u32 *os_area_msg_seqno;
+	u8 *vmpck;
+	u8 vmpck_id;
  };
 
  /*
arch/x86/kernel/sev.c +424 -3
index 3f664ab277c4..c5ca97aab8c1 100644
@@ -22,10 +22,13 @@
  #include <linux/efi.h>
  #include <linux/platform_device.h>
  #include <linux/io.h>
+#include <linux/psp-sev.h>
+#include <crypto/gcm.h>
 
  #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>
@@ -982,6 +985,427 @@ static void snp_cleanup_vmsa(struct sev_es_save_area *vmsa)
  		free_page((unsigned long)vmsa);
  }
 
+static struct sev_guest_platform_data *platform_data;
+
+/*
+ * If an error is received from the host or AMD Secure Processor (ASP) there
+ * are two options. Either retry the exact same encrypted request or discontinue
+ * using the VMPCK.
+ *
+ * This is because in the current encryption scheme GHCB v2 uses AES-GCM to
+ * encrypt the requests. The IV for this scheme is the sequence number. GCM
+ * cannot tolerate IV reuse.
+ *
+ * The ASP FW v1.51 only increments the sequence numbers on a successful
+ * guest<->ASP back and forth and only accepts messages at its exact sequence
+ * number.
+ *
+ * So if the sequence number were to be reused the encryption scheme is
+ * vulnerable. If the sequence number were incremented for a fresh IV the ASP
+ * will reject the request.
+ */
+static void snp_disable_vmpck(struct snp_guest_dev *snp_dev)
+{
+	pr_alert("Disabling vmpck_id %d to prevent IV reuse.\n", snp_dev->vmpck_id);
+	memzero_explicit(snp_dev->vmpck, VMPCK_KEY_LEN);
+	snp_dev->vmpck = NULL;
+}
+
+static inline u64 __snp_get_msg_seqno(struct snp_guest_dev *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;
+
+	return count + 1;
+}
+
+/* Return a non-zero on success */
+static u64 snp_get_msg_seqno(struct snp_guest_dev *snp_dev)
+{
+	u64 count = __snp_get_msg_seqno(snp_dev);
+
+	/*
+	 * The message sequence counter for the SNP guest request is a  64-bit
+	 * value but the version 2 of GHCB specification defines a 32-bit storage
+	 * for it. If the counter exceeds the 32-bit value then return zero.
+	 * The caller should check the return value, but if the caller happens to
+	 * not check the value and use it, then the firmware treats zero as an
+	 * invalid number and will fail the  message request.
+	 */
+	if (count >= UINT_MAX) {
+		pr_err("SNP request message sequence counter overflow\n");
+		return 0;
+	}
+
+	return count;
+}
+
+static void snp_inc_msg_seqno(struct snp_guest_dev *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;
+}
+
+static struct aesgcm_ctx *snp_init_crypto(u8 *key, size_t keylen)
+{
+	struct aesgcm_ctx *ctx;
+
+	ctx = kzalloc(sizeof(*ctx), GFP_KERNEL_ACCOUNT);
+	if (!ctx)
+		return NULL;
+
+	if (aesgcm_expandkey(ctx, key, keylen, AUTHTAG_LEN)) {
+		pr_err("SNP: crypto init failed\n");
+		kfree(ctx);
+		return NULL;
+	}
+
+	return ctx;
+}
+
+static void free_shared_pages(void *buf, size_t sz)
+{
+	unsigned int npages = PAGE_ALIGN(sz) >> PAGE_SHIFT;
+	int ret;
+
+	if (!buf)
+		return;
+
+	ret = set_memory_encrypted((unsigned long)buf, npages);
+	if (ret) {
+		WARN_ONCE(ret, "failed to restore encryption mask (leak it)\n");
+		return;
+	}
+
+	__free_pages(virt_to_page(buf), get_order(sz));
+}
+
+static void *alloc_shared_pages(size_t sz)
+{
+	unsigned int npages = PAGE_ALIGN(sz) >> PAGE_SHIFT;
+	struct page *page;
+	int ret;
+
+	page = alloc_pages(GFP_KERNEL_ACCOUNT, get_order(sz));
+	if (!page)
+		return NULL;
+
+	ret = set_memory_decrypted((unsigned long)page_address(page), npages);
+	if (ret) {
+		pr_err("%s: failed to mark page shared, ret=%d\n", __func__, ret);
+		__free_pages(page, get_order(sz));
+		return NULL;
+	}
+
+	return page_address(page);
+}
+
+static int snp_setup_psp_messaging(struct sev_guest_platform_data *pdata)
+{
+	u64 gpa;
+	int ret;
+
+	if (!cc_platform_has(CC_ATTR_GUEST_SEV_SNP))
+		return -ENODEV;
+
+	gpa = get_secrets_page();
+	if (!gpa)
+		return -ENODEV;
+
+	pdata->layout = (__force void *)ioremap_encrypted(gpa, PAGE_SIZE);
+	if (!pdata->layout) {
+		pr_err("Unable to locate AP jump table address: failed to map the SNP secrets page.\n");
+		return -ENODEV;
+	}
+
+	ret = -EINVAL;
+	pdata->vmpck0 = pdata->layout->vmpck0;
+
+	/* Verify that VMPCK is not zero. */
+	if (is_vmpck_empty(pdata->vmpck0)) {
+		pr_err("vmpck id 0 is null\n");
+		goto e_unmap;
+	}
+
+	ret = -ENOMEM;
+	/* Allocate the shared page used for the request and response message. */
+	pdata->req = alloc_shared_pages(sizeof(struct snp_guest_msg));
+	if (!pdata->req)
+		goto e_unmap;
+
+	pdata->resp = alloc_shared_pages(sizeof(struct snp_guest_msg));
+	if (!pdata->resp)
+		goto e_free_request;
+
+	pdata->certs_data = alloc_shared_pages(SEV_FW_BLOB_MAX_SIZE);
+	if (!pdata->certs_data)
+		goto e_free_response;
+
+	ret = -EIO;
+	pdata->ctx = snp_init_crypto(pdata->vmpck0, VMPCK_KEY_LEN);
+	if (!pdata->ctx)
+		goto e_free_cert_data;
+
+	/* initial the input address for guest request */
+	pdata->input.req_gpa = __pa(pdata->req);
+	pdata->input.resp_gpa = __pa(pdata->resp);
+	pdata->input.data_gpa = __pa(pdata->certs_data);
+
+	return 0;
+
+e_free_cert_data:
+	free_shared_pages(pdata->certs_data, SEV_FW_BLOB_MAX_SIZE);
+e_free_response:
+	free_shared_pages(pdata->resp, sizeof(struct snp_guest_msg));
+e_free_request:
+	free_shared_pages(pdata->req, sizeof(struct snp_guest_msg));
+e_unmap:
+	iounmap(pdata->layout);
+
+	return ret;
+}
+
+static int __enc_payload(struct aesgcm_ctx *ctx, struct snp_guest_msg *msg,
+			 void *plaintext, size_t len)
+{
+	struct snp_guest_msg_hdr *hdr = &msg->hdr;
+	u8 iv[GCM_AES_IV_SIZE] = {};
+
+	if (WARN_ON((hdr->msg_sz + ctx->authsize) > sizeof(msg->payload)))
+		return -EBADMSG;
+
+	memcpy(iv, &hdr->msg_seqno, sizeof(hdr->msg_seqno));
+	aesgcm_encrypt(ctx, msg->payload, plaintext, len, &hdr->algo, AAD_LEN,
+		       iv, hdr->authtag);
+	return 0;
+}
+
+static int dec_payload(struct aesgcm_ctx *ctx, struct snp_guest_msg *msg,
+		       void *plaintext, size_t len)
+{
+	struct snp_guest_msg_hdr *hdr = &msg->hdr;
+	u8 iv[GCM_AES_IV_SIZE] = {};
+
+	memcpy(iv, &hdr->msg_seqno, sizeof(hdr->msg_seqno));
+	if (aesgcm_decrypt(ctx, plaintext, msg->payload, len, &hdr->algo,
+			   AAD_LEN, iv, hdr->authtag))
+		return 0;
+	else
+		return -EBADMSG;
+}
+
+static int verify_and_dec_payload(struct sev_guest_platform_data *pdata,
+				  void *payload, u32 sz)
+{
+	struct snp_guest_msg *resp = pdata->resp;
+	struct snp_guest_msg *req = pdata->req;
+	struct snp_guest_msg_hdr *req_hdr = &req->hdr;
+	struct snp_guest_msg_hdr *resp_hdr = &resp->hdr;
+	struct aesgcm_ctx *ctx = pdata->ctx;
+
+	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);
+
+	/* Verify that the sequence counter is incremented by 1 */
+	if (unlikely(resp_hdr->msg_seqno != (req_hdr->msg_seqno + 1)))
+		return -EBADMSG;
+
+	/* Verify response message type and version number. */
+	if (resp_hdr->msg_type != (req_hdr->msg_type + 1) ||
+	    resp_hdr->msg_version != req_hdr->msg_version)
+		return -EBADMSG;
+
+	/*
+	 * If the message size is greater than our buffer length then return
+	 * an error.
+	 */
+	if (unlikely((resp_hdr->msg_sz + ctx->authsize) > sz))
+		return -EBADMSG;
+
+	return dec_payload(ctx, resp, payload, resp_hdr->msg_sz);
+}
+
+static int enc_payload(struct sev_guest_platform_data *pdata, u64 seqno,
+		       struct snp_guest_req *req, u8 vmpck_id)
+{
+	struct snp_guest_msg *msg = pdata->req;
+	struct snp_guest_msg_hdr *hdr = &msg->hdr;
+
+	memset(msg, 0, sizeof(*msg));
+
+	hdr->algo = SNP_AEAD_AES_256_GCM;
+	hdr->hdr_version = MSG_HDR_VER;
+	hdr->hdr_sz = sizeof(*hdr);
+	hdr->msg_type = req->msg_type;
+	hdr->msg_version = req->msg_version;
+	hdr->msg_seqno = seqno;
+	hdr->msg_vmpck = vmpck_id;
+	hdr->msg_sz = req->req_sz;
+
+	/* Verify the sequence number is non-zero */
+	if (!hdr->msg_seqno)
+		return -ENOSR;
+
+	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(pdata->ctx, msg, req->req_buf, req->req_sz);
+}
+
+static int __handle_guest_request(struct snp_guest_dev *snp_dev,
+				  struct sev_guest_platform_data *pdata,
+				  struct snp_guest_req *req)
+{
+	unsigned long err = 0xff, override_err = 0;
+	unsigned long req_start = jiffies;
+	unsigned int override_npages = 0;
+	int rc;
+
+retry_request:
+	/*
+	 * Call firmware to process the request. In this function the encrypted
+	 * message enters shared memory with the host. So after this call the
+	 * sequence number must be incremented or the VMPCK must be deleted to
+	 * prevent reuse of the IV.
+	 */
+	rc = snp_issue_guest_request(req->exit_code, &pdata->input, &err);
+	switch (rc) {
+	case -ENOSPC:
+		/*
+		 * If the extended guest request fails due to having too
+		 * small of a certificate data buffer, retry the same
+		 * guest request without the extended data request in
+		 * order to increment the sequence number and thus avoid
+		 * IV reuse.
+		 */
+		override_npages = pdata->input.data_npages;
+		req->exit_code	= SVM_VMGEXIT_GUEST_REQUEST;
+
+		/*
+		 * Override the error to inform callers the given extended
+		 * request buffer size was too small and give the caller the
+		 * required buffer size.
+		 */
+		override_err	= SNP_GUEST_REQ_INVALID_LEN;
+
+		/*
+		 * If this call to the firmware succeeds, the sequence number can
+		 * be incremented allowing for continued use of the VMPCK. If
+		 * there is an error reflected in the return value, this value
+		 * is checked further down and the result will be the deletion
+		 * of the VMPCK and the error code being propagated back to the
+		 * user as an ioctl() return code.
+		 */
+		goto retry_request;
+
+	/*
+	 * The host may return SNP_GUEST_REQ_ERR_EBUSY if the request has been
+	 * throttled. Retry in the driver to avoid returning and reusing the
+	 * message sequence number on a different message.
+	 */
+	case -EAGAIN:
+		if (jiffies - req_start > SNP_REQ_MAX_RETRY_DURATION) {
+			rc = -ETIMEDOUT;
+			break;
+		}
+		schedule_timeout_killable(SNP_REQ_RETRY_DELAY);
+		goto retry_request;
+	}
+
+	/*
+	 * Increment the message sequence number. There is no harm in doing
+	 * this now because decryption uses the value stored in the response
+	 * structure and any failure will wipe the VMPCK, preventing further
+	 * use anyway.
+	 */
+	snp_inc_msg_seqno(snp_dev);
+
+	if (req->fw_err)
+		*req->fw_err = override_err ?: err;
+
+	if (override_npages)
+		pdata->input.data_npages = override_npages;
+
+	/*
+	 * If an extended guest request was issued and the supplied certificate
+	 * buffer was not large enough, a standard guest request was issued to
+	 * prevent IV reuse. If the standard request was successful, return -EIO
+	 * back to the caller as would have originally been returned.
+	 */
+	if (!rc && override_err == SNP_GUEST_REQ_INVALID_LEN)
+		return -EIO;
+
+	return rc;
+}
+
+int snp_send_guest_request(struct snp_guest_dev *snp_dev, struct snp_guest_req *req)
+{
+	struct sev_guest_platform_data *pdata;
+	u64 seqno;
+	int rc;
+
+	if (!snp_dev || !snp_dev->pdata || !req)
+		return -ENODEV;
+
+	pdata = snp_dev->pdata;
+
+	/* Get message sequence and verify that its a non-zero */
+	seqno = snp_get_msg_seqno(snp_dev);
+	if (!seqno)
+		return -EIO;
+
+	memset(pdata->resp, 0, sizeof(struct snp_guest_msg));
+
+	/* Encrypt the provided payload */
+	rc = enc_payload(pdata, seqno, req, snp_dev->vmpck_id);
+	if (rc)
+		return rc;
+
+	rc = __handle_guest_request(snp_dev, pdata, req);
+	if (rc) {
+		if (rc == -EIO && *req->fw_err == SNP_GUEST_REQ_INVALID_LEN)
+			return rc;
+
+		pr_alert("Detected error from ASP request. rc: %d, fw_err: %llu\n",
+			 rc, *req->fw_err);
+		snp_disable_vmpck(snp_dev);
+		return rc;
+	}
+
+	rc = verify_and_dec_payload(pdata, req->resp_buf, req->resp_sz);
+	if (rc) {
+		pr_alert("Detected unexpected decode failure from ASP. rc: %d\n", rc);
+		snp_disable_vmpck(snp_dev);
+		return rc;
+	}
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(snp_send_guest_request);
+
+bool snp_assign_vmpck(struct snp_guest_dev *dev, int vmpck_id)
+{
+	struct sev_guest_platform_data *pdata = dev->pdata;
+
+	if (!pdata || WARN_ON(vmpck_id > 3))
+		return false;
+
+	dev->vmpck_id = vmpck_id;
+	dev->vmpck = pdata->layout->vmpck0 + vmpck_id * VMPCK_KEY_LEN;
+	dev->os_area_msg_seqno = &pdata->layout->os_area.msg_seqno_0 + vmpck_id;
+
+	return true;
+}
+EXPORT_SYMBOL_GPL(snp_assign_vmpck);
+
  static int wakeup_cpu_via_vmgexit(int apic_id, unsigned long start_ip)
  {
  	struct sev_es_save_area *cur_vmsa, *vmsa;
@@ -2247,18 +2671,17 @@ static struct platform_device sev_guest_device = {
 
  static int __init snp_init_platform_device(void)
  {
-	struct sev_guest_platform_data data;
-	u64 gpa;
-

arch/x86/kernel/sev.c static struct platform_device sev_guest_device = { · line 2681

Quoted 10 lines
arch/x86/kernel/sev.c +5 -2
  	if (!cc_platform_has(CC_ATTR_GUEST_SEV_SNP))
  		return -ENODEV;
 
-	gpa = get_secrets_page();
-	if (!gpa)
+	platform_data = kzalloc(sizeof(*platform_data), GFP_KERNEL);
+	if (!platform_data)
+		return -ENOMEM;
+
+	if (snp_setup_psp_messaging(platform_data))

This shouldn't be done here (or yet) since you only moving the routines. The sev-guest driver should call this regardless of the vmpck_id value.

Quoted 523 lines
arch/x86/kernel/sev.c +1 -2
  		return -ENODEV;
 
-	data.secrets_gpa = gpa;
-	if (platform_device_add_data(&sev_guest_device, &data, sizeof(data)))
+	if (platform_device_add_data(&sev_guest_device, platform_data, sizeof(*platform_data)))
  		return -ENODEV;
 
  	if (platform_device_register(&sev_guest_device))
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 +16 -387
index 4901ebc8fa1a..a011df21c03a 100644
@@ -24,110 +24,25 @@
  #include <asm/svm.h>
  #include <asm/sev.h>
 
-#include "sev-guest.h"
+#include <asm/sev-guest.h>
 
  #define DEVICE_NAME	"sev-guest"
 
-#define SNP_REQ_MAX_RETRY_DURATION	(60*HZ)
-#define SNP_REQ_RETRY_DELAY		(2*HZ)
-
-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;
-	struct snp_guest_msg *request, *response;
-	struct snp_secrets_page_layout *layout;
-	struct snp_req_data input;
-	u32 *os_area_msg_seqno;
-	u8 *vmpck;
-	u8 vmpck_id;
-};
-
  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.");
 
  static inline unsigned int get_ctx_authsize(struct snp_guest_dev *snp_dev)
  {
-	if (snp_dev && snp_dev->ctx)
-		return snp_dev->ctx->authsize;
+	struct sev_guest_platform_data *pdata = snp_dev->pdata;
+
+	if (pdata && pdata->ctx)
+		return pdata->ctx->authsize;
 
  	WARN_ONCE(1, "Unable to get crypto authsize\n");
  	return 0;
  }
 
-/*
- * If an error is received from the host or AMD Secure Processor (ASP) there
- * are two options. Either retry the exact same encrypted request or discontinue
- * using the VMPCK.
- *
- * This is because in the current encryption scheme GHCB v2 uses AES-GCM to
- * encrypt the requests. The IV for this scheme is the sequence number. GCM
- * cannot tolerate IV reuse.
- *
- * The ASP FW v1.51 only increments the sequence numbers on a successful
- * guest<->ASP back and forth and only accepts messages at its exact sequence
- * number.
- *
- * So if the sequence number were to be reused the encryption scheme is
- * vulnerable. If the sequence number were incremented for a fresh IV the ASP
- * will reject the request.
- */
-static void snp_disable_vmpck(struct snp_guest_dev *snp_dev)
-{
-	dev_alert(snp_dev->dev, "Disabling vmpck_id %d to prevent IV reuse.\n",
-		  snp_dev->vmpck_id);
-	memzero_explicit(snp_dev->vmpck, VMPCK_KEY_LEN);
-	snp_dev->vmpck = NULL;
-}
-
-static inline u64 __snp_get_msg_seqno(struct snp_guest_dev *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;
-
-	return count + 1;
-}
-
-/* Return a non-zero on success */
-static u64 snp_get_msg_seqno(struct snp_guest_dev *snp_dev)
-{
-	u64 count = __snp_get_msg_seqno(snp_dev);
-
-	/*
-	 * The message sequence counter for the SNP guest request is a  64-bit
-	 * value but the version 2 of GHCB specification defines a 32-bit storage
-	 * for it. If the counter exceeds the 32-bit value then return zero.
-	 * The caller should check the return value, but if the caller happens to
-	 * not check the value and use it, then the firmware treats zero as an
-	 * invalid number and will fail the  message request.
-	 */
-	if (count >= UINT_MAX) {
-		dev_err(snp_dev->dev, "request message sequence counter overflow\n");
-		return 0;
-	}
-
-	return count;
-}
-
-static void snp_inc_msg_seqno(struct snp_guest_dev *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;
-}
-
  static inline struct snp_guest_dev *to_snp_dev(struct file *file)
  {
  	struct miscdevice *dev = file->private_data;
@@ -135,236 +50,6 @@ 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)
-{
-	struct aesgcm_ctx *ctx;
-
-	ctx = kzalloc(sizeof(*ctx), GFP_KERNEL_ACCOUNT);
-	if (!ctx)
-		return NULL;
-
-	if (aesgcm_expandkey(ctx, key, keylen, AUTHTAG_LEN)) {
-		pr_err("SNP: crypto init failed\n");
-		kfree(ctx);
-		return NULL;
-	}
-
-	return ctx;
-}
-
-static int __enc_payload(struct aesgcm_ctx *ctx, struct snp_guest_msg *msg,
-			 void *plaintext, size_t len)
-{
-	struct snp_guest_msg_hdr *hdr = &msg->hdr;
-	u8 iv[GCM_AES_IV_SIZE] = {};
-
-	if (WARN_ON((hdr->msg_sz + ctx->authsize) > sizeof(msg->payload)))
-		return -EBADMSG;
-
-	memcpy(iv, &hdr->msg_seqno, sizeof(hdr->msg_seqno));
-	aesgcm_encrypt(ctx, msg->payload, plaintext, len, &hdr->algo, AAD_LEN,
-		       iv, hdr->authtag);
-	return 0;
-}
-
-static int dec_payload(struct aesgcm_ctx *ctx, struct snp_guest_msg *msg,
-		       void *plaintext, size_t len)
-{
-	struct snp_guest_msg_hdr *hdr = &msg->hdr;
-	u8 iv[GCM_AES_IV_SIZE] = {};
-
-	memcpy(iv, &hdr->msg_seqno, sizeof(hdr->msg_seqno));
-	if (aesgcm_decrypt(ctx, plaintext, msg->payload, len, &hdr->algo,
-			   AAD_LEN, iv, hdr->authtag))
-		return 0;
-	else
-		return -EBADMSG;
-}
-
-static int verify_and_dec_payload(struct snp_guest_dev *snp_dev, void *payload, u32 sz)
-{
-	struct snp_guest_msg *resp = snp_dev->response;
-	struct snp_guest_msg *req = snp_dev->request;
-	struct snp_guest_msg_hdr *req_hdr = &req->hdr;
-	struct snp_guest_msg_hdr *resp_hdr = &resp->hdr;
-	struct aesgcm_ctx *ctx = snp_dev->ctx;
-
-	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);
-
-	/* Verify that the sequence counter is incremented by 1 */
-	if (unlikely(resp_hdr->msg_seqno != (req_hdr->msg_seqno + 1)))
-		return -EBADMSG;
-
-	/* Verify response message type and version number. */
-	if (resp_hdr->msg_type != (req_hdr->msg_type + 1) ||
-	    resp_hdr->msg_version != req_hdr->msg_version)
-		return -EBADMSG;
-
-	/*
-	 * If the message size is greater than our buffer length then return
-	 * an error.
-	 */
-	if (unlikely((resp_hdr->msg_sz + ctx->authsize) > sz))
-		return -EBADMSG;
-
-	/* Decrypt the payload */
-	return dec_payload(ctx, resp, payload, resp_hdr->msg_sz);
-}
-
-static int enc_payload(struct snp_guest_dev *snp_dev, u64 seqno,
-		       struct snp_guest_req *req, u8 __vmpck_id)
-{
-	struct snp_guest_msg *msg = snp_dev->request;
-	struct snp_guest_msg_hdr *hdr = &msg->hdr;
-
-	memset(msg, 0, sizeof(*msg));
-
-	hdr->algo = SNP_AEAD_AES_256_GCM;
-	hdr->hdr_version = MSG_HDR_VER;
-	hdr->hdr_sz = sizeof(*hdr);
-	hdr->msg_type = req->msg_type;
-	hdr->msg_version = req->msg_version;
-	hdr->msg_seqno = seqno;
-	hdr->msg_vmpck = __vmpck_id;
-	hdr->msg_sz = req->req_sz;
-
-	/* Verify the sequence number is non-zero */
-	if (!hdr->msg_seqno)
-		return -ENOSR;
-
-	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, msg, req->req_buf, req->req_sz);
-}
-
-static int __handle_guest_request(struct snp_guest_dev *snp_dev, struct snp_guest_req *req)
-{
-	unsigned long err = 0xff, override_err = 0;
-	unsigned long req_start = jiffies;
-	unsigned int override_npages = 0;
-	int rc;
-
-retry_request:
-	/*
-	 * Call firmware to process the request. In this function the encrypted
-	 * message enters shared memory with the host. So after this call the
-	 * sequence number must be incremented or the VMPCK must be deleted to
-	 * prevent reuse of the IV.
-	 */
-	rc = snp_issue_guest_request(req->exit_code, &snp_dev->input, &err);
-	switch (rc) {
-	case -ENOSPC:
-		/*
-		 * If the extended guest request fails due to having too
-		 * small of a certificate data buffer, retry the same
-		 * guest request without the extended data request in
-		 * order to increment the sequence number and thus avoid
-		 * IV reuse.
-		 */
-		override_npages = snp_dev->input.data_npages;
-		req->exit_code	= SVM_VMGEXIT_GUEST_REQUEST;
-
-		/*
-		 * Override the error to inform callers the given extended
-		 * request buffer size was too small and give the caller the
-		 * required buffer size.
-		 */
-		override_err	= SNP_GUEST_REQ_INVALID_LEN;
-
-		/*
-		 * If this call to the firmware succeeds, the sequence number can
-		 * be incremented allowing for continued use of the VMPCK. If
-		 * there is an error reflected in the return value, this value
-		 * is checked further down and the result will be the deletion
-		 * of the VMPCK and the error code being propagated back to the
-		 * user as an ioctl() return code.
-		 */
-		goto retry_request;
-
-	/*
-	 * The host may return SNP_GUEST_REQ_ERR_EBUSY if the request has been
-	 * throttled. Retry in the driver to avoid returning and reusing the
-	 * message sequence number on a different message.
-	 */
-	case -EAGAIN:
-		if (jiffies - req_start > SNP_REQ_MAX_RETRY_DURATION) {
-			rc = -ETIMEDOUT;
-			break;
-		}
-		schedule_timeout_killable(SNP_REQ_RETRY_DELAY);
-		goto retry_request;
-	}
-
-	/*
-	 * Increment the message sequence number. There is no harm in doing
-	 * this now because decryption uses the value stored in the response
-	 * structure and any failure will wipe the VMPCK, preventing further
-	 * use anyway.
-	 */
-	snp_inc_msg_seqno(snp_dev);
-
-	if (req->fw_err)
-		*req->fw_err = override_err ?: err;
-
-	if (override_npages)
-		snp_dev->input.data_npages = override_npages;
-
-	/*
-	 * If an extended guest request was issued and the supplied certificate
-	 * buffer was not large enough, a standard guest request was issued to
-	 * prevent IV reuse. If the standard request was successful, return -EIO
-	 * back to the caller as would have originally been returned.
-	 */
-	if (!rc && override_err == SNP_GUEST_REQ_INVALID_LEN)
-		return -EIO;
-
-	return rc;
-}
-
-static int snp_send_guest_request(struct snp_guest_dev *snp_dev, struct snp_guest_req *req)
-{
-	u64 seqno;
-	int rc;
-
-	if (!snp_dev || !req)
-		return -ENODEV;
-
-	/* Get message sequence and verify that its a non-zero */
-	seqno = snp_get_msg_seqno(snp_dev);
-	if (!seqno)
-		return -EIO;
-
-	memset(snp_dev->response, 0, sizeof(struct snp_guest_msg));
-
-	/* Encrypt the userspace provided payload */
-	rc = enc_payload(snp_dev, seqno, req, snp_dev->vmpck_id);
-	if (rc)
-		return rc;
-
-	rc = __handle_guest_request(snp_dev, req);
-	if (rc) {
-		if (rc == -EIO && *req->fw_err == SNP_GUEST_REQ_INVALID_LEN)
-			return rc;
-
-		dev_alert(snp_dev->dev, "Detected error from ASP request. rc: %d, fw_err: %llu\n",
-			  rc, *req->fw_err);
-		snp_disable_vmpck(snp_dev);
-		return rc;
-	}
-
-	rc = verify_and_dec_payload(snp_dev, req->resp_buf, req->resp_sz);
-	if (rc) {
-		dev_alert(snp_dev->dev, "Detected unexpected decode failure from ASP. rc: %d\n", rc);
-		snp_disable_vmpck(snp_dev);
-		return rc;
-	}
-
-	return 0;
-}
-
  static int handle_guest_request(struct snp_guest_dev *snp_dev, u64 exit_code, u8 msg_version,
  				u8 msg_type, void *req_buf, size_t req_sz, void *resp_buf,
  				u32 resp_sz, __u64 *fw_err)
@@ -463,12 +148,14 @@ static int get_derived_key(struct snp_guest_dev *snp_dev, struct snp_guest_reque
 
  static int get_ext_report(struct snp_guest_dev *snp_dev, struct snp_guest_request_ioctl *arg)
  {
+	struct sev_guest_platform_data *pdata;
  	struct snp_ext_report_req req;
  	struct snp_report_resp *resp;
  	int ret, npages = 0, resp_len;
 
  	lockdep_assert_held(&snp_dev->cmd_mutex);
 
+	pdata = snp_dev->pdata;
  	if (!arg->req_data || !arg->resp_data)
  		return -EINVAL;
 
@@ -492,7 +179,7 @@ static int get_ext_report(struct snp_guest_dev *snp_dev, struct snp_guest_reques
  	 * the host. If host does not supply any certs in it, then copy
  	 * zeros to indicate that certificate data was not provided.
  	 */
-	memset(snp_dev->certs_data, 0, req.certs_len);
+	memset(pdata->certs_data, 0, req.certs_len);
  	npages = req.certs_len >> PAGE_SHIFT;
  cmd:
  	/*
@@ -505,14 +192,14 @@ static int get_ext_report(struct snp_guest_dev *snp_dev, struct snp_guest_reques
  	if (!resp)
  		return -ENOMEM;
 
-	snp_dev->input.data_npages = npages;
+	pdata->input.data_npages = npages;
  	ret = handle_guest_request(snp_dev, SVM_VMGEXIT_EXT_GUEST_REQUEST, arg->msg_version,
  				   SNP_MSG_REPORT_REQ, &req.data,
  				   sizeof(req.data), resp->data, resp_len, &arg->fw_err);
 
  	/* If certs length is invalid then copy the returned length */
  	if (arg->fw_err == SNP_GUEST_REQ_INVALID_LEN) {
-		req.certs_len = snp_dev->input.data_npages << PAGE_SHIFT;
+		req.certs_len = pdata->input.data_npages << PAGE_SHIFT;
 
  		if (copy_to_user((void __user *)arg->req_data, &req, sizeof(req)))
  			ret = -EFAULT;
@@ -522,7 +209,7 @@ static int get_ext_report(struct snp_guest_dev *snp_dev, struct snp_guest_reques
  		goto e_free;
 
  	if (npages &&
-	    copy_to_user((void __user *)req.certs_address, snp_dev->certs_data,
+	    copy_to_user((void __user *)req.certs_address, pdata->certs_data,
  			 req.certs_len)) {
  		ret = -EFAULT;
  		goto e_free;
@@ -583,68 +270,17 @@ static long snp_guest_ioctl(struct file *file, unsigned int ioctl, unsigned long
  	return ret;
  }
 
-static void free_shared_pages(void *buf, size_t sz)
-{
-	unsigned int npages = PAGE_ALIGN(sz) >> PAGE_SHIFT;
-	int ret;
-
-	if (!buf)
-		return;
-
-	ret = set_memory_encrypted((unsigned long)buf, npages);
-	if (ret) {
-		WARN_ONCE(ret, "failed to restore encryption mask (leak it)\n");
-		return;
-	}
-
-	__free_pages(virt_to_page(buf), get_order(sz));
-}
-
-static void *alloc_shared_pages(struct device *dev, size_t sz)
-{
-	unsigned int npages = PAGE_ALIGN(sz) >> PAGE_SHIFT;
-	struct page *page;
-	int ret;
-
-	page = alloc_pages(GFP_KERNEL_ACCOUNT, get_order(sz));
-	if (!page)
-		return NULL;
-
-	ret = set_memory_decrypted((unsigned long)page_address(page), npages);
-	if (ret) {
-		dev_err(dev, "failed to mark page shared, ret=%d\n", ret);
-		__free_pages(page, get_order(sz));
-		return NULL;
-	}
-
-	return page_address(page);
-}
-
  static const struct file_operations snp_guest_fops = {
  	.owner	= THIS_MODULE,
  	.unlocked_ioctl = snp_guest_ioctl,
  };
 
-bool snp_assign_vmpck(struct snp_guest_dev *dev, int vmpck_id)
-{
-	if (WARN_ON(vmpck_id > 3))
-		return false;
-
-	dev->vmpck_id = vmpck_id;
-	dev->vmpck = dev->layout->vmpck0 + vmpck_id * VMPCK_KEY_LEN;
-	dev->os_area_msg_seqno = &dev->layout->os_area.msg_seqno_0 + vmpck_id;
-
-	return true;
-}
-
  static int __init sev_guest_probe(struct platform_device *pdev)
  {
-	struct snp_secrets_page_layout *layout;
-	struct sev_guest_platform_data *data;
+	struct sev_guest_platform_data *pdata;
  	struct device *dev = &pdev->dev;
  	struct snp_guest_dev *snp_dev;
  	struct miscdevice *misc;
-	void __iomem *mapping;
  	int ret;
 
  	if (!cc_platform_has(CC_ATTR_GUEST_SEV_SNP))
@@ -653,80 +289,48 @@ static int __init sev_guest_probe(struct platform_device *pdev)
  	if (!dev->platform_data)
  		return -ENODEV;
 
-	data = (struct sev_guest_platform_data *)dev->platform_data;
-	mapping = ioremap_encrypted(data->secrets_gpa, PAGE_SIZE);
-	if (!mapping)
-		return -ENODEV;
-
-	layout = (__force void *)mapping;
-
-	ret = -ENOMEM;
+	pdata = (struct sev_guest_platform_data *)dev->platform_data;
  	snp_dev = devm_kzalloc(&pdev->dev, sizeof(struct snp_guest_dev), GFP_KERNEL);
  	if (!snp_dev)
-		goto e_unmap;
+		return -ENOMEM;
 
  	ret = -EINVAL;
-	snp_dev->layout = layout;
+	snp_dev->pdata = pdata;
  	if (!snp_assign_vmpck(snp_dev, vmpck_id)) {
  		dev_err(dev, "invalid vmpck id %d\n", vmpck_id);
-		goto e_unmap;
+		goto e_free_snpdev;
  	}

drivers/virt/coco/sev-guest/sev-guest.c static int __init sev_guest_probe(struct platform_device *pdev) · line 311

Quoted 10 lines
drivers/virt/coco/sev-guest/sev-guest.c +5 -1
 
  	/* Verify that VMPCK is not zero. */
  	if (is_vmpck_empty(snp_dev->vmpck)) {
  		dev_err(dev, "vmpck id %d is null\n", vmpck_id);
-		goto e_unmap;
+		goto e_free_snpdev;
+	}
+
+	/* Skip VMPCK0 initialization as the key is already initialized during early boot */
+	if (vmpck_id && aesgcm_expandkey(pdata->ctx, snp_dev->vmpck, VMPCK_KEY_LEN, AUTHTAG_LEN)) {

See previous comment. The sev-guest driver should be setting up everything private to it no matter the vmpck to be used.

Thanks,
Tom

Quoted 73 lines
drivers/virt/coco/sev-guest/sev-guest.c +7 -39
+		ret = -ENODEV;
+		goto e_free_snpdev;
  	}
 
  	mutex_init(&snp_dev->cmd_mutex);
  	platform_set_drvdata(pdev, snp_dev);
  	snp_dev->dev = dev;
 
-	/* Allocate the shared page used for the request and response message. */
-	snp_dev->request = alloc_shared_pages(dev, sizeof(struct snp_guest_msg));
-	if (!snp_dev->request)
-		goto e_unmap;
-
-	snp_dev->response = alloc_shared_pages(dev, sizeof(struct snp_guest_msg));
-	if (!snp_dev->response)
-		goto e_free_request;
-
-	snp_dev->certs_data = alloc_shared_pages(dev, SEV_FW_BLOB_MAX_SIZE);
-	if (!snp_dev->certs_data)
-		goto e_free_response;
-
-	ret = -EIO;
-	snp_dev->ctx = snp_init_crypto(snp_dev->vmpck, VMPCK_KEY_LEN);
-	if (!snp_dev->ctx)
-		goto e_free_cert_data;
-
  	misc = &snp_dev->misc;
  	misc->minor = MISC_DYNAMIC_MINOR;
  	misc->name = DEVICE_NAME;
  	misc->fops = &snp_guest_fops;
 
-	/* initial the input address for guest request */
-	snp_dev->input.req_gpa = __pa(snp_dev->request);
-	snp_dev->input.resp_gpa = __pa(snp_dev->response);
-	snp_dev->input.data_gpa = __pa(snp_dev->certs_data);
-
-	ret =  misc_register(misc);
+	ret = misc_register(misc);
  	if (ret)
-		goto e_free_ctx;
+		goto e_free_snpdev;
 
  	dev_info(dev, "Initialized SEV guest driver (using vmpck_id %d)\n", vmpck_id);
  	return 0;
 
-e_free_ctx:
-	kfree(snp_dev->ctx);
-e_free_cert_data:
-	free_shared_pages(snp_dev->certs_data, SEV_FW_BLOB_MAX_SIZE);
-e_free_response:
-	free_shared_pages(snp_dev->response, sizeof(struct snp_guest_msg));
-e_free_request:
-	free_shared_pages(snp_dev->request, sizeof(struct snp_guest_msg));
-e_unmap:
-	iounmap(mapping);
+e_free_snpdev:
+	kfree(snp_dev);
  	return ret;
  }
 
@@ -734,11 +338,8 @@ static int __exit sev_guest_remove(struct platform_device *pdev)
  {
  	struct snp_guest_dev *snp_dev = platform_get_drvdata(pdev);
 
-	free_shared_pages(snp_dev->certs_data, SEV_FW_BLOB_MAX_SIZE);
-	free_shared_pages(snp_dev->response, sizeof(struct snp_guest_msg));
-	free_shared_pages(snp_dev->request, sizeof(struct snp_guest_msg));
-	kfree(snp_dev->ctx);
  	misc_deregister(&snp_dev->misc);
+	kfree(snp_dev);
 
  	return 0;
  }
#17 lore raw

Re: [PATCH v2 05/11] x86/sev: Move and reorganize sev guest request api

Quoted 10 lines · Tom Lendacky
Quoted 2 lines · Nikunj A Dadhania

For enabling Secure TSC, SEV-SNP guests need to communicate with the security coprocessor really early during boot. Lot of the required

s/security coprocessor really/AMD Secure Processor/

s/Lot/Many/

Quoted 1 line

functions are implemented in the sev-guest driver. Move the required

Quoted 9 lines

... in the sev-guest driver and therefore not available at early boot.

Quoted 1 line

functions and provide API to the driver to assign VM communications

s/provide API/provide an API/

s/to assign.*//

Quoted 1 line

key and send guest request.

Sure will change.

Quoted 4 lines
Quoted 4 lines
arch/x86/include/asm/sev.h +1 -0
@@ -96,6 +97,27 @@ struct snp_req_data {
     struct sev_guest_platform_data {
       u64 secrets_gpa;
+
Quoted 8 lines

arch/x86/include/asm/sev.h struct snp_req_data { · line 106

Quoted 6 lines
arch/x86/include/asm/sev.h +6 -0
+    void *certs_data;
+    struct aesgcm_ctx *ctx;
+    struct snp_guest_msg *req, *resp;
+    struct snp_secrets_page_layout *layout;
+    struct snp_req_data input;
+    u8 *vmpck0;

Isn't this unneeded? You have the vmpck and vmpck_id in the snp_guest_dev struct which will be set based on the module parameter, so vmpck0 and associated checks shouldn't be needed.

Yes, this can be removed.

Quoted 7 lines

arch/x86/kernel/sev.c static struct platform_device sev_guest_device = { · line 2681

Quoted 5 lines
arch/x86/kernel/sev.c +5 -0
+    platform_data = kzalloc(sizeof(*platform_data), GFP_KERNEL);
+    if (!platform_data)
+        return -ENOMEM;
+
+    if (snp_setup_psp_messaging(platform_data))

This shouldn't be done here (or yet) since you only moving the routines. The sev-guest driver should call this regardless of the vmpck_id value.

I am moving the routines and also making sure that sev-guest driver works with this change. So sev-guest driver will not need to call snp_setup_psp_messaging().

Quoted 6 lines

drivers/virt/coco/sev-guest/sev-guest.c static int __init sev_guest_probe(struct platform_device *pdev) · line 311

Quoted 4 lines
drivers/virt/coco/sev-guest/sev-guest.c +4 -0
+    }
+
+    /* Skip VMPCK0 initialization as the key is already initialized during early boot */
+    if (vmpck_id && aesgcm_expandkey(pdata->ctx, snp_dev->vmpck, VMPCK_KEY_LEN, AUTHTAG_LEN)) {

See previous comment. The sev-guest driver should be setting up everything private to it no matter the vmpck to be used.

I will try this out, as for secure tsc the vmpck will be used once and the sequence number would have incremented. And VMPCK0 will be initialized twice.

Regards
Nikunj

#18 lore raw

PATCH v2 6/11 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 c1c8c581759d..4026ac30a79f 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 ef80d361b463..a71996e51a9c 100644
@@ -134,6 +134,7 @@ static void enc_status_change_prepare_noop(unsigned long vaddr, int npages, bool
 static bool enc_status_change_finish_noop(unsigned long vaddr, int npages, bool enc) { return false; }
 static bool enc_tlb_flush_required_noop(bool enc) { return false; }
 static bool enc_cache_flush_required_noop(void) { return false; }
+static void enc_init_noop(void) { }
 
 struct x86_platform_ops x86_platform __ro_after_init = {
 	.calibrate_cpu			= native_calibrate_cpu_early,
@@ -155,6 +156,7 @@ 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
#19 lore raw

PATCH v2 7/11 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 c5ca97aab8c1..3750e545d688 100644
@@ -1642,6 +1642,30 @@ 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 && (sev_status & MSR_AMD64_SNP_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
Reviewed-by on v2 7/11
#20 lore raw

Re: [PATCH v2 07/11] x86/sev: Change TSC MSR behavior for Secure TSC enabled guests

Quoted 10 lines · 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.

Quoted 39 lines
---
  arch/x86/kernel/sev.c | 24 ++++++++++++++++++++++++
  1 file changed, 24 insertions(+)
arch/x86/kernel/sev.c +24 -0
index c5ca97aab8c1..3750e545d688 100644
@@ -1642,6 +1642,30 @@ 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 && (sev_status & MSR_AMD64_SNP_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);
#21 lore raw

PATCH v2 8/11 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 Security processor (PSP). This communication channel is encrypted between the security processor and the guest, hypervisor is just the conduit to deliver the guest messages to the security processor. Each message is protected with an AEAD (AES-256 GCM). Use minimal GCM library to encrypt/decrypt SNP Guest messages to communicate with the PSP.

Moreover, 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/include/asm/sev-guest.h | 18 +++++++
 arch/x86/include/asm/sev.h       |  2 +
 arch/x86/include/asm/svm.h       |  6 ++-
 arch/x86/kernel/sev-shared.c     |  7 +++
 arch/x86/kernel/sev.c            | 92 +++++++++++++++++++++++++++++---
 arch/x86/mm/mem_encrypt_amd.c    |  6 +++
 include/linux/cc_platform.h      |  8 +++
 7 files changed, 131 insertions(+), 8 deletions(-)
arch/x86/include/asm/sev-guest.h +18 -0
index 834cdae302ad..d5ed041ce06b 100644
@@ -37,6 +37,8 @@ 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
 };
@@ -75,6 +77,22 @@ 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;
+	u64 tsc_factor;
+	u8 rsvd2[96];
+} __packed;
+
 int snp_send_guest_request(struct snp_guest_dev *dev, struct snp_guest_req *req);
 bool snp_assign_vmpck(struct snp_guest_dev *dev, int vmpck_id);
 
arch/x86/include/asm/sev.h +2 -0
index 36868e21c3e0..d05cbab5e9e0 100644
@@ -218,6 +218,7 @@ 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);
+bool __init snp_secure_tsc_prepare(void);
 int snp_issue_guest_request(u64 exit_code, struct snp_req_data *input, unsigned long *fw_err);
 #else
 static inline void sev_es_ist_enter(struct pt_regs *regs) { }
@@ -238,6 +239,7 @@ 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 bool __init snp_secure_tsc_prepare(void) { return false; }
 static inline int snp_issue_guest_request(u64 exit_code, struct snp_req_data *input,
 					  unsigned long *fw_err)
 {
arch/x86/include/asm/svm.h +4 -2
index 770dcf75eaa9..c781f8e0aae4 100644
@@ -404,7 +404,9 @@ 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;
@@ -536,7 +538,7 @@ 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-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;
arch/x86/kernel/sev.c +86 -6

122 lines not on this page · raw message

arch/x86/mm/mem_encrypt_amd.c +6 -0
index 9c4d8dbcb129..7d2388e52b8f 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
#22 lore raw

Re: [PATCH v2 08/11] x86/sev: Add Secure TSC support for SNP guests

Quoted 7 lines · 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 Security processor (PSP).

s/Security processor (PSP)/AMD Secure Processor/

Quoted 1 line

This communication channel is encrypted between the security

here as well.

Quoted 1 line

processor and the guest, hypervisor is just the conduit to

s/hypervisor/the hypervisor/

Quoted 9 lines

deliver the guest messages to the security processor. Each message is protected with an AEAD (AES-256 GCM). Use minimal GCM library to encrypt/decrypt SNP Guest messages to communicate with the PSP.

Moreover, 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.

This seems like a separate patch.

Quoted 32 lines
---
  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-shared.c     |  7 +++
  arch/x86/kernel/sev.c            | 92 +++++++++++++++++++++++++++++---
  arch/x86/mm/mem_encrypt_amd.c    |  6 +++
  include/linux/cc_platform.h      |  8 +++
  7 files changed, 131 insertions(+), 8 deletions(-)
arch/x86/include/asm/sev-guest.h +5 -0
index 834cdae302ad..d5ed041ce06b 100644
@@ -37,6 +37,8 @@ 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
  };
@@ -75,6 +77,22 @@ struct snp_guest_req {
  	u8 msg_type;
  };
 
+struct snp_tsc_info_req {
+#define SNP_TSC_INFO_REQ_SZ 128
+	/* Must be zero filled */

arch/x86/include/asm/sev-guest.h struct snp_guest_req { · line 92

Quoted 10 lines
arch/x86/include/asm/sev-guest.h +10 -0
+	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;
+	u64 tsc_factor;

This should be a u32 ...

arch/x86/include/asm/sev-guest.h struct snp_guest_req { · line 93

Quoted 1 line
arch/x86/include/asm/sev-guest.h +1 -0
+	u8 rsvd2[96];

Which then makes this 100.

Quoted 151 lines
arch/x86/include/asm/sev-guest.h +2 -0
+} __packed;
+
  int snp_send_guest_request(struct snp_guest_dev *dev, struct snp_guest_req *req);
  bool snp_assign_vmpck(struct snp_guest_dev *dev, int vmpck_id);
 
arch/x86/include/asm/sev.h +2 -0
index 36868e21c3e0..d05cbab5e9e0 100644
@@ -218,6 +218,7 @@ 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);
+bool __init snp_secure_tsc_prepare(void);
  int snp_issue_guest_request(u64 exit_code, struct snp_req_data *input, unsigned long *fw_err);
  #else
  static inline void sev_es_ist_enter(struct pt_regs *regs) { }
@@ -238,6 +239,7 @@ 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 bool __init snp_secure_tsc_prepare(void) { return false; }
  static inline int snp_issue_guest_request(u64 exit_code, struct snp_req_data *input,
  					  unsigned long *fw_err)
  {
arch/x86/include/asm/svm.h +4 -2
index 770dcf75eaa9..c781f8e0aae4 100644
@@ -404,7 +404,9 @@ 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;
@@ -536,7 +538,7 @@ 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-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;
arch/x86/kernel/sev.c +61 -1
index 3750e545d688..280aaa1e6aad 100644
@@ -72,6 +72,10 @@ static struct ghcb *boot_ghcb __section(".data");
  /* Bitmap of SEV features supported by the hypervisor */
  static u64 sev_hv_features __ro_after_init;
 
+/* Secure TSC values read using TSC_INFO SNP Guest request */
+static u64 guest_tsc_scale __ro_after_init;
+static u64 guest_tsc_offset __ro_after_init;
+
  /* #VC handler runtime per-CPU data */
  struct sev_es_runtime_data {
  	struct ghcb ghcb_page;
@@ -1107,7 +1111,7 @@ static void *alloc_shared_pages(size_t sz)
  	return page_address(page);
  }
 
-static int snp_setup_psp_messaging(struct sev_guest_platform_data *pdata)
+static int __init snp_setup_psp_messaging(struct sev_guest_platform_data *pdata)
  {
  	u64 gpa;
  	int ret;
@@ -1406,6 +1410,80 @@ bool snp_assign_vmpck(struct snp_guest_dev *dev, int vmpck_id)
  }
  EXPORT_SYMBOL_GPL(snp_assign_vmpck);
 
+static int __init snp_get_tsc_info(void)
+{
+	u8 buf[SNP_TSC_INFO_REQ_SZ + AUTHTAG_LEN];
+	struct snp_tsc_info_resp tsc_resp = {0};
+	struct snp_tsc_info_req tsc_req;
+	struct snp_guest_req req;
+	struct snp_guest_dev dev;
+	int rc, resp_len;
+
+	/*
+	 * The intermediate response buffer is used while decrypting the
+	 * response payload. Make sure that it has enough space to cover the
+	 * authtag.
+	 */
+	resp_len = sizeof(tsc_resp) + AUTHTAG_LEN;
+	if (sizeof(buf) < resp_len)
+		return -EINVAL;
+
+	/* Zero the tsc_info_req */
+	memzero_explicit(&tsc_req, sizeof(tsc_req));
+	memzero_explicit(&req, sizeof(req));
+
+	dev.pdata = platform_data;
+	if (!snp_assign_vmpck(&dev, 0))
+		return -EINVAL;
+
+	req.msg_version = MSG_HDR_VER;
+	req.msg_type = SNP_MSG_TSC_INFO_REQ;
+	req.req_buf = &tsc_req;
+	req.req_sz = sizeof(tsc_req);
+	req.resp_buf = buf;
+	req.resp_sz = resp_len;
+	req.fw_err = NULL;
+	req.exit_code = SVM_VMGEXIT_GUEST_REQUEST;
+	rc = snp_send_guest_request(&dev, &req);
+	if (rc)
+		goto err_req;
+
+	memcpy(&tsc_resp, buf, sizeof(tsc_resp));
+	pr_debug("%s: Valid response status %x scale %llx offset %llx factor %llx\n",
+		 __func__, tsc_resp.status, tsc_resp.tsc_scale, tsc_resp.tsc_offset,
+		 tsc_resp.tsc_factor);
+
+	guest_tsc_scale = tsc_resp.tsc_scale;
+	guest_tsc_offset = tsc_resp.tsc_offset;
+
+err_req:
+	/* The response buffer contains the sensitive data, explicitly clear it. */
+	memzero_explicit(buf, sizeof(buf));
+	memzero_explicit(&tsc_resp, sizeof(tsc_resp));
+	memzero_explicit(&req, sizeof(req));
+
+	return rc;
+}
+
+bool __init snp_secure_tsc_prepare(void)

arch/x86/kernel/sev.c bool snp_assign_vmpck(struct snp_guest_dev *dev, int vmpck_id) · line 1478

Quoted 10 lines
arch/x86/kernel/sev.c +10 -0
+{
+	platform_data = kzalloc(sizeof(*platform_data), GFP_KERNEL);
+	if (!platform_data)
+		return false;
+
+	/* Initialize the PSP channel to send snp messages */
+	if (snp_setup_psp_messaging(platform_data))
+		sev_es_terminate(SEV_TERM_SET_GEN, GHCB_SNP_UNSUPPORTED);
+
+	if (cc_platform_has(CC_ATTR_GUEST_SECURE_TSC)) {

Should this be checked before allocating memory and calling snp_setup_psp_messaging()?

Also, I notice here you use the cc_platform_has() function but in previous patches you check sev_status directly. And you don't implement support for CC_ATTR_GUEST_SECURE_TSC until the last patch instead of now.

You can't get here until SNP_FEATURES_PRESENT is updated.

arch/x86/kernel/sev.c bool snp_assign_vmpck(struct snp_guest_dev *dev, int vmpck_id) · line 1483

Quoted 5 lines
arch/x86/kernel/sev.c +5 -0
+		if (snp_get_tsc_info())
+			sev_es_terminate(SEV_TERM_SET_GEN, GHCB_SNP_UNSUPPORTED);
+
+		pr_info("SecureTSC enabled\n");
+	}

Blank line.

Thanks,
Tom

Quoted 75 lines
arch/x86/kernel/sev.c +10 -5
+	return true;
+}
+
  static int wakeup_cpu_via_vmgexit(int apic_id, unsigned long start_ip)
  {
  	struct sev_es_save_area *cur_vmsa, *vmsa;
@@ -1506,6 +1584,12 @@ static int wakeup_cpu_via_vmgexit(int apic_id, unsigned long start_ip)
  	vmsa->vmpl		= 0;
  	vmsa->sev_features	= sev_status >> 2;
 
+	/* Setting Secure TSC parameters */
+	if (cc_platform_has(CC_ATTR_GUEST_SECURE_TSC)) {
+		vmsa->tsc_scale = guest_tsc_scale;
+		vmsa->tsc_offset = guest_tsc_offset;
+	}
+
  	/* Switch the page over to a VMSA page now that it is initialized */
  	ret = snp_set_vmsa(vmsa, true);
  	if (ret) {
@@ -2698,11 +2782,7 @@ static int __init snp_init_platform_device(void)
  	if (!cc_platform_has(CC_ATTR_GUEST_SEV_SNP))
  		return -ENODEV;
 
-	platform_data = kzalloc(sizeof(*platform_data), GFP_KERNEL);
-	if (!platform_data)
-		return -ENOMEM;
-
-	if (snp_setup_psp_messaging(platform_data))
+	if (!platform_data->ctx)
  		return -ENODEV;
 
  	if (platform_device_add_data(&sev_guest_device, platform_data, sizeof(*platform_data)))
arch/x86/mm/mem_encrypt_amd.c +6 -0
index 9c4d8dbcb129..7d2388e52b8f 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
#23 lore raw

Re: [PATCH v2 08/11] x86/sev: Add Secure TSC support for SNP guests

Quoted 11 lines · Tom Lendacky
Quoted 7 lines · 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 Security processor (PSP).

s/Security processor (PSP)/AMD Secure Processor/

Quoted 8 lines
Quoted 1 line

This communication channel is encrypted between the security

here as well.

Quoted 1 line

processor and the guest, hypervisor is just the conduit to

s/hypervisor/the hypervisor/

Sure, will change commit message

Quoted 11 lines
Quoted 9 lines

deliver the guest messages to the security processor. Each message is protected with an AEAD (AES-256 GCM). Use minimal GCM library to encrypt/decrypt SNP Guest messages to communicate with the PSP.

Moreover, 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.

This seems like a separate patch.

Sure.

Quoted 9 lines
Quoted 8 lines

---

arch/x86/include/asm/sev-guest.h +0 -0
index 834cdae302ad..d5ed041ce06b 100644
Quoted 15 lines

arch/x86/include/asm/sev-guest.h struct snp_guest_req { · line 92

Quoted 12 lines
arch/x86/include/asm/sev-guest.h +12 -0
+#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;
+    u64 tsc_factor;

This should be a u32 ...

Right, will correct.

Quoted 3 lines
Quoted 1 line

+    u8 rsvd2[96];

Which then makes this 100.

Yes.

Quoted 5 lines
Quoted 4 lines
arch/x86/kernel/sev.c +0 -0
index 3750e545d688..280aaa1e6aad 100644
Quoted 13 lines

arch/x86/kernel/sev.c bool snp_assign_vmpck(struct snp_guest_dev *dev, int vmpck_id) · line 1478

Quoted 11 lines
arch/x86/kernel/sev.c +11 -0
+bool __init snp_secure_tsc_prepare(void)
+{
+    platform_data = kzalloc(sizeof(*platform_data), GFP_KERNEL);
+    if (!platform_data)
+        return false;
+
+    /* Initialize the PSP channel to send snp messages */
+    if (snp_setup_psp_messaging(platform_data))
+        sev_es_terminate(SEV_TERM_SET_GEN, GHCB_SNP_UNSUPPORTED);
+
+    if (cc_platform_has(CC_ATTR_GUEST_SECURE_TSC)) {

Should this be checked before allocating memory and calling snp_setup_psp_messaging()?

No, as we need guest messaging to work without Secure TSC as well.

Quoted 2 lines

Also, I notice here you use the cc_platform_has() function but in previous patches you check sev_status directly.

For sev-shared.c, cc_platform_has() is not available when compiling boot/compressed.

I will change the below call site to cc_platfrom_has() after testing.

arch/x86/kernel/sev.c: if (regs->cx == MSR_IA32_TSC && (sev_status & MSR_AMD64_SNP_SECURE_TSC)) {

Quoted 2 lines

And you don't implement support for CC_ATTR_GUEST_SECURE_TSC until the last patch instead of now.

This is to make sure that SECURE_TSC is enabled only after the last patch. As cc_platform_has() is being used at multiple places to enable the feature.

Quoted 9 lines

You can't get here until SNP_FEATURES_PRESENT is updated.

Quoted 5 lines

+        if (snp_get_tsc_info())
+            sev_es_terminate(SEV_TERM_SET_GEN, GHCB_SNP_UNSUPPORTED); +
+        pr_info("SecureTSC enabled\n");
+    }

Blank line.

Sure

Regards
Nikunj

#24 lore raw

Re: [PATCH v2 08/11] x86/sev: Add Secure TSC support for SNP guests

Quoted 89 lines · Nikunj A. Dadhania
Quoted 11 lines · Tom Lendacky
Quoted 7 lines · 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 Security processor (PSP).

s/Security processor (PSP)/AMD Secure Processor/

Quoted 8 lines
Quoted 1 line

This communication channel is encrypted between the security

here as well.

Quoted 1 line

processor and the guest, hypervisor is just the conduit to

s/hypervisor/the hypervisor/

Sure, will change commit message

Quoted 11 lines
Quoted 9 lines

deliver the guest messages to the security processor. Each message is protected with an AEAD (AES-256 GCM). Use minimal GCM library to encrypt/decrypt SNP Guest messages to communicate with the PSP.

Moreover, 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.

This seems like a separate patch.

Sure.

Quoted 9 lines
Quoted 8 lines

---

arch/x86/include/asm/sev-guest.h +0 -0
index 834cdae302ad..d5ed041ce06b 100644
Quoted 15 lines

arch/x86/include/asm/sev-guest.h struct snp_guest_req { · line 92

Quoted 12 lines
arch/x86/include/asm/sev-guest.h +12 -0
+#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;
+    u64 tsc_factor;

This should be a u32 ...

Right, will correct.

Quoted 3 lines
Quoted 1 line

+    u8 rsvd2[96];

Which then makes this 100.

Yes.

Quoted 5 lines
Quoted 4 lines
arch/x86/kernel/sev.c +0 -0
index 3750e545d688..280aaa1e6aad 100644
Quoted 13 lines

arch/x86/kernel/sev.c bool snp_assign_vmpck(struct snp_guest_dev *dev, int vmpck_id) · line 1478

Quoted 11 lines
arch/x86/kernel/sev.c +11 -0
+bool __init snp_secure_tsc_prepare(void)
+{
+    platform_data = kzalloc(sizeof(*platform_data), GFP_KERNEL);
+    if (!platform_data)
+        return false;
+
+    /* Initialize the PSP channel to send snp messages */
+    if (snp_setup_psp_messaging(platform_data))
+        sev_es_terminate(SEV_TERM_SET_GEN, GHCB_SNP_UNSUPPORTED);
+
+    if (cc_platform_has(CC_ATTR_GUEST_SECURE_TSC)) {

Should this be checked before allocating memory and calling snp_setup_psp_messaging()?

Quoted 1 line

No, as we need guest messaging to work without Secure TSC as well.

My take is that the sev-guest driver should establish the key it is going to use at the time the driver is loaded. snp_setup_psp_messaging() should be called by whatever is going to use guest messaging.

Having a generic routine that is accessed by both the kernel and the driver should be the goal. Maybe it is best to only have the vmpck_id be part of any structure and then the appropriate key and sequence number are used based on that id when the request is made.

The sev_guest_platform_data struct can just hold context information (it doesn't need the secrets_gpa any more since everything is now in sev.c which knows what that value is) for whatever is using guest messaging.

Quoted 9 lines
Quoted 4 lines

Also, I notice here you use the cc_platform_has() function but in previous patches you check sev_status directly.

For sev-shared.c, cc_platform_has() is not available when compiling boot/compressed.

I will change the below call site to cc_platfrom_has() after testing.

Quoted 7 lines

arch/x86/kernel/sev.c: if (regs->cx == MSR_IA32_TSC && (sev_status & MSR_AMD64_SNP_SECURE_TSC)) {

Quoted 2 lines

And you don't implement support for CC_ATTR_GUEST_SECURE_TSC until the last patch instead of now.

This is to make sure that SECURE_TSC is enabled only after the last patch. As cc_platform_has() is being used at multiple places to enable the feature.

But you terminate long before that in snp_check_features() since you don't update SNP_FEATURES_PRESENT with SECURE_TSC until the last patch, right?

Thanks,
Tom

Quoted 16 lines
Quoted 10 lines

You can't get here until SNP_FEATURES_PRESENT is updated.

Quoted 5 lines

+        if (snp_get_tsc_info())
+            sev_es_terminate(SEV_TERM_SET_GEN, GHCB_SNP_UNSUPPORTED); +
+        pr_info("SecureTSC enabled\n");
+    }

Blank line.

Sure

Regards
Nikunj

#25 lore raw

Re: [PATCH v2 08/11] x86/sev: Add Secure TSC support for SNP guests

Quoted 90 lines · Tom Lendacky
Quoted 89 lines · Nikunj A. Dadhania
Quoted 11 lines · Tom Lendacky
Quoted 7 lines · 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 Security processor (PSP).

s/Security processor (PSP)/AMD Secure Processor/

Quoted 8 lines
Quoted 1 line

This communication channel is encrypted between the security

here as well.

Quoted 1 line

processor and the guest, hypervisor is just the conduit to

s/hypervisor/the hypervisor/

Sure, will change commit message

Quoted 11 lines
Quoted 9 lines

deliver the guest messages to the security processor. Each message is protected with an AEAD (AES-256 GCM). Use minimal GCM library to encrypt/decrypt SNP Guest messages to communicate with the PSP.

Moreover, 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.

This seems like a separate patch.

Sure.

Quoted 9 lines
Quoted 8 lines

---

arch/x86/include/asm/sev-guest.h +0 -0
index 834cdae302ad..d5ed041ce06b 100644
Quoted 15 lines

arch/x86/include/asm/sev-guest.h struct snp_guest_req { · line 92

Quoted 12 lines
arch/x86/include/asm/sev-guest.h +12 -0
+#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;
+    u64 tsc_factor;

This should be a u32 ...

Right, will correct.

Quoted 3 lines
Quoted 1 line

+    u8 rsvd2[96];

Which then makes this 100.

Yes.

Quoted 18 lines
Quoted 6 lines
arch/x86/kernel/sev.c +1 -0
index 3750e545d688..280aaa1e6aad 100644
 
+bool __init snp_secure_tsc_prepare(void)

arch/x86/kernel/sev.c bool snp_assign_vmpck(struct snp_guest_dev *dev, int vmpck_id) · line 1478

Quoted 10 lines
arch/x86/kernel/sev.c +10 -0
+{
+    platform_data = kzalloc(sizeof(*platform_data), GFP_KERNEL);
+    if (!platform_data)
+        return false;
+
+    /* Initialize the PSP channel to send snp messages */
+    if (snp_setup_psp_messaging(platform_data))
+        sev_es_terminate(SEV_TERM_SET_GEN, GHCB_SNP_UNSUPPORTED);
+
+    if (cc_platform_has(CC_ATTR_GUEST_SECURE_TSC)) {

Should this be checked before allocating memory and calling snp_setup_psp_messaging()?

Quoted 3 lines
Quoted 1 line

No, as we need guest messaging to work without Secure TSC as well.

My take is that the sev-guest driver should establish the key it is going to use at the time the driver is loaded. > snp_setup_psp_messaging() should be called by whatever is going to use guest messaging.

Should sev-guest also call snp_setup_psp_messaging() ?

As per the current design:
* snp_secure_tsc_prepare populates platform_data (probably I should rename it snp_prepare_psp_messaging)
* In snp_init_platform_device() platform_data is copied using platform_device_add_data().

sev-guest driver in sev_guest_probe() uses this platform_data populated by snp_secure_tsc_prepare().

Quoted 2 lines

Having a generic routine that is accessed by both the kernel and the driver should be the goal. Maybe it is best to only have the vmpck_id be part of any structure and then the appropriate key and sequence number are used based on that id when the request is made.

I understand what you are suggesting, let me think over this and I will come back.

Quoted 1 line

The sev_guest_platform_data struct can just hold context information (it doesn't need the secrets_gpa any more since everything is now in sev.c which knows what that value is) for whatever is using guest messaging.

We are using snp_assign_vmpck() in the sev-guest driver that is dependent on secrets_gpa, we can get rid of that dependency I think.

Quoted 11 lines
Quoted 11 lines

Quoted 2 lines

Also, I notice here you use the cc_platform_has() function but in previous patches you check sev_status directly.

For sev-shared.c, cc_platform_has() is not available when compiling boot/compressed.

I will change the below call site to cc_platfrom_has() after testing.

arch/x86/kernel/sev.c:  if (regs->cx == MSR_IA32_TSC && (sev_status & MSR_AMD64_SNP_SECURE_TSC)) {

Quoted 7 lines
Quoted 5 lines
Quoted 2 lines

And you don't implement support for CC_ATTR_GUEST_SECURE_TSC until the last patch instead of now.

This is to make sure that SECURE_TSC is enabled only after the last patch. As cc_platform_has() is being used at multiple places to enable the feature.

But you terminate long before that in snp_check_features() since you don't update SNP_FEATURES_PRESENT with SECURE_TSC until the last patch, right?

Yes, that correct. I will implement full support for CC_ATTR_GUEST_SECURE_TSC in this patch.

Regards
Nikunj

#26 lore raw

Re: [PATCH v2 08/11] x86/sev: Add Secure TSC support for SNP guests

In reply to Nikunj A Dadhania
Quoted 28 lines
arch/x86/kernel/sev.c +13 -1
+
 /* #VC handler runtime per-CPU data */
 struct sev_es_runtime_data {
        struct ghcb ghcb_page;
@@ -1107,7 +1111,7 @@ static void *alloc_shared_pages(size_t sz)
        return page_address(page);
 }
 
-static int snp_setup_psp_messaging(struct sev_guest_platform_data *pdata)
+static int __init snp_setup_psp_messaging(struct sev_guest_platform_data *pdata)
 {
        u64 gpa;
        int ret;
@@ -1406,6 +1410,80 @@ bool snp_assign_vmpck(struct snp_guest_dev *dev, int vmpck_id)
 }
 EXPORT_SYMBOL_GPL(snp_assign_vmpck);
 
+static int __init snp_get_tsc_info(void)
+{
+       u8 buf[SNP_TSC_INFO_REQ_SZ + AUTHTAG_LEN];
+       struct snp_tsc_info_resp tsc_resp = {0};
+       struct snp_tsc_info_req tsc_req;
+       struct snp_guest_req req;
+       struct snp_guest_dev dev;
+       int rc, resp_len;
+
+       /*
+        * The intermediate response buffer is used while decrypting the

arch/x86/kernel/sev.c bool snp_assign_vmpck(struct snp_guest_dev *dev, int vmpck_id) · line 1433

Quoted 10 lines
arch/x86/kernel/sev.c +10 -0
+        * response payload. Make sure that it has enough space to cover the
+        * authtag.
+        */
+       resp_len = sizeof(tsc_resp) + AUTHTAG_LEN;
+       if (sizeof(buf) < resp_len)
+               return -EINVAL;
+
+       /* Zero the tsc_info_req */
+       memzero_explicit(&tsc_req, sizeof(tsc_req));
+       memzero_explicit(&req, sizeof(req));

Whats the guidance on when we should use memzero_explicit() vs just something like: snp_tsc_info_resp tsc_resp = {0};?

arch/x86/kernel/sev.c bool snp_assign_vmpck(struct snp_guest_dev *dev, int vmpck_id) · line 1445

Quoted 12 lines
arch/x86/kernel/sev.c +12 -0
+
+       dev.pdata = platform_data;
+       if (!snp_assign_vmpck(&dev, 0))
+               return -EINVAL;
+
+       req.msg_version = MSG_HDR_VER;
+       req.msg_type = SNP_MSG_TSC_INFO_REQ;
+       req.req_buf = &tsc_req;
+       req.req_sz = sizeof(tsc_req);
+       req.resp_buf = buf;
+       req.resp_sz = resp_len;
+       req.fw_err = NULL;

Why do we not want the FW error code?

Quoted 22 lines
arch/x86/kernel/sev.c +22 -0
+       req.exit_code = SVM_VMGEXIT_GUEST_REQUEST;
+       rc = snp_send_guest_request(&dev, &req);
+       if (rc)
+               goto err_req;
+
+       memcpy(&tsc_resp, buf, sizeof(tsc_resp));
+       pr_debug("%s: Valid response status %x scale %llx offset %llx factor %llx\n",
+                __func__, tsc_resp.status, tsc_resp.tsc_scale, tsc_resp.tsc_offset,
+                tsc_resp.tsc_factor);
+
+       guest_tsc_scale = tsc_resp.tsc_scale;
+       guest_tsc_offset = tsc_resp.tsc_offset;
+
+err_req:
+       /* The response buffer contains the sensitive data, explicitly clear it. */
+       memzero_explicit(buf, sizeof(buf));
+       memzero_explicit(&tsc_resp, sizeof(tsc_resp));
+       memzero_explicit(&req, sizeof(req));
+
+       return rc;
+}
+
#27 lore raw

Re: [PATCH v2 08/11] x86/sev: Add Secure TSC support for SNP guests

Quoted 33 lines · Peter Gonda
Quoted 33 lines
arch/x86/kernel/sev.c +18 -1
+
 /* #VC handler runtime per-CPU data */
 struct sev_es_runtime_data {
        struct ghcb ghcb_page;
@@ -1107,7 +1111,7 @@ static void *alloc_shared_pages(size_t sz)
        return page_address(page);
 }
 
-static int snp_setup_psp_messaging(struct sev_guest_platform_data *pdata)
+static int __init snp_setup_psp_messaging(struct sev_guest_platform_data *pdata)
 {
        u64 gpa;
        int ret;
@@ -1406,6 +1410,80 @@ bool snp_assign_vmpck(struct snp_guest_dev *dev, int vmpck_id)
 }
 EXPORT_SYMBOL_GPL(snp_assign_vmpck);
 
+static int __init snp_get_tsc_info(void)
+{
+       u8 buf[SNP_TSC_INFO_REQ_SZ + AUTHTAG_LEN];
+       struct snp_tsc_info_resp tsc_resp = {0};
+       struct snp_tsc_info_req tsc_req;
+       struct snp_guest_req req;
+       struct snp_guest_dev dev;
+       int rc, resp_len;
+
+       /*
+        * The intermediate response buffer is used while decrypting the
+        * response payload. Make sure that it has enough space to cover the
+        * authtag.
+        */
+       resp_len = sizeof(tsc_resp) + AUTHTAG_LEN;
+       if (sizeof(buf) < resp_len)
Quoted 8 lines

arch/x86/kernel/sev.c bool snp_assign_vmpck(struct snp_guest_dev *dev, int vmpck_id) · line 1433

Quoted 5 lines
arch/x86/kernel/sev.c +5 -0
+               return -EINVAL;
+
+       /* Zero the tsc_info_req */
+       memzero_explicit(&tsc_req, sizeof(tsc_req));
+       memzero_explicit(&req, sizeof(req));

Whats the guidance on when we should use memzero_explicit() vs just something like: snp_tsc_info_resp tsc_resp = {0};?

Going over the history of memzero_explicit, it seems it was introduce to explicitly zero sensitive information before the variable goes out of scope. GCC was optimizing out the memset in these cases:

d4c5efdb9777 ("random: add and use memzero_explicit() for clearing data")

https://bugzilla.kernel.org/show_bug.cgi?id=82041

With the above detail, IMHO, we do not need the memzero_explicit() for both case.

Quoted 15 lines

arch/x86/kernel/sev.c bool snp_assign_vmpck(struct snp_guest_dev *dev, int vmpck_id) · line 1445

Quoted 13 lines
arch/x86/kernel/sev.c +12 -0
+
+       dev.pdata = platform_data;
+       if (!snp_assign_vmpck(&dev, 0))
+               return -EINVAL;
+
+       req.msg_version = MSG_HDR_VER;
+       req.msg_type = SNP_MSG_TSC_INFO_REQ;
+       req.req_buf = &tsc_req;
+       req.req_sz = sizeof(tsc_req);
+       req.resp_buf = buf;
+       req.resp_sz = resp_len;
+       req.fw_err = NULL;

Why do we not want the FW error code?

I will add the FW error code.

Regards
Nikunj

#28 lore raw

PATCH v2 9/11 x86/kvmclock: Use Secure TSC as clock if available

In reply to Nikunj A Dadhania

For AMD SNP guests having Secure TSC enabled, use Secure TSC based clocksource instead of kvmclock.

---
 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
#29 lore raw

Re: [PATCH v2 09/11] x86/kvmclock: Use Secure TSC as clock if available

Quoted 2 lines · Nikunj A Dadhania

For AMD SNP guests having Secure TSC enabled, use Secure TSC based clocksource instead of kvmclock.

This patch only results in skipping the initialization of KVM clock, it doesn't cause Secure TSC to be used as the clocksource.

Thanks,
Tom

Quoted 19 lines
---
  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)) {
#30 lore raw

Re: [PATCH v2 09/11] x86/kvmclock: Use Secure TSC as clock if available

Quoted 6 lines · Tom Lendacky
Quoted 2 lines · Nikunj A Dadhania

For AMD SNP guests having Secure TSC enabled, use Secure TSC based clocksource instead of kvmclock.

This patch only results in skipping the initialization of KVM clock, it doesn't cause Secure TSC to be used as the clocksource.

Will update the commit message and reorder with 10/11. After making TSC reliable and skipping kvmclock, secure tsc is used as clocksource.

Regards
Nikunj

#31 lore raw

PATCH v2 10/11 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
#32 lore raw

PATCH v2 11/11 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/coco/core.c           |  3 +++
 arch/x86/mm/mem_encrypt.c      | 10 ++++++++--
 3 files changed, 12 insertions(+), 3 deletions(-)
arch/x86/boot/compressed/sev.c +1 -1
index d63ad8f99f83..13b6ff192c3e 100644
@@ -315,7 +315,7 @@ 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/coco/core.c +3 -0
index 49b44f881484..a3d93ed722c9 100644
@@ -68,6 +68,9 @@ 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/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