From:   Usama Arif <usama.arif@bytedance.com>
To:     dwmw2@infradead.org, tglx@linutronix.de, kim.phillips@amd.com,
        brgerst@gmail.com
Cc:     piotrgorski@cachyos.org, oleksandr@natalenko.name,
        arjan@linux.intel.com, mingo@redhat.com, bp@alien8.de,
        dave.hansen@linux.intel.com, hpa@zytor.com, x86@kernel.org,
        pbonzini@redhat.com, paulmck@kernel.org,
        linux-kernel@vger.kernel.org, kvm@vger.kernel.org,
        rcu@vger.kernel.org, mimoja@mimoja.de, hewenliang4@huawei.com,
        thomas.lendacky@amd.com, seanjc@google.com, pmenzel@molgen.mpg.de,
        fam.zheng@bytedance.com, punit.agrawal@bytedance.com,
        simon.evans@bytedance.com, liangma@liangbit.com,
        David Woodhouse <dwmw@amazon.co.uk>,
        Sabin Rapan <sabrapan@amazon.com>,
        Usama Arif <usama.arif@bytedance.com>
Subject: [PATCH v14 12/12] x86/smpboot: Allow parallel bringup for SEV-ES
Date:   Wed,  8 Mar 2023 17:13:28 +0000
Message-Id: <20230308171328.1562857-13-usama.arif@bytedance.com>
In-Reply-To: <20230308171328.1562857-1-usama.arif@bytedance.com>
References: <20230308171328.1562857-1-usama.arif@bytedance.com>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Precedence: bulk

From: David Woodhouse <dwmw@amazon.co.uk>

Enable parallel bringup for SEV-ES guests. The APs can't actually
execute the CPUID instruction directly during early startup, but they
can make the GHCB call directly instead, just as the VC trap handler
would do.

Thanks to Sabin for talking me through the way this works.

Suggested-by: Sabin Rapan <sabrapan@amazon.com>
Signed-off-by: David Woodhouse <dwmw@amazon.co.uk>
Signed-off-by: Usama Arif <usama.arif@bytedance.com>
---
 arch/x86/include/asm/sev-common.h |  3 +++
 arch/x86/include/asm/sev.h        |  5 +++++
 arch/x86/include/asm/smp.h        |  5 ++++-
 arch/x86/kernel/head_64.S         | 30 ++++++++++++++++++++++++++++++
 arch/x86/kernel/smpboot.c         | 31 +++++++++++++++++++++++++------
 5 files changed, 67 insertions(+), 7 deletions(-)

diff --git a/arch/x86/include/asm/sev-common.h b/arch/x86/include/asm/sev-common.h
index b8357d6ecd47..f25df4bd318e 100644
--- a/arch/x86/include/asm/sev-common.h
+++ b/arch/x86/include/asm/sev-common.h
@@ -70,6 +70,7 @@
 	/* GHCBData[63:12] */				\
 	(((u64)(v) & GENMASK_ULL(63, 12)) >> 12)
 
+#ifndef __ASSEMBLY__
 /*
  * SNP Page State Change Operation
  *
@@ -160,6 +161,8 @@ struct snp_psc_desc {
 
 #define GHCB_RESP_CODE(v)		((v) & GHCB_MSR_INFO_MASK)
 
+#endif /* __ASSEMBLY__ */
+
 /*
  * Error codes related to GHCB input that can be communicated back to the guest
  * by setting the lower 32-bits of the GHCB SW_EXITINFO1 field to 2.
diff --git a/arch/x86/include/asm/sev.h b/arch/x86/include/asm/sev.h
index ebc271bb6d8e..d687a586cafa 100644
--- a/arch/x86/include/asm/sev.h
+++ b/arch/x86/include/asm/sev.h
@@ -135,6 +135,10 @@ struct snp_secrets_page_layout {
 
 #ifdef CONFIG_AMD_MEM_ENCRYPT
 extern struct static_key_false sev_es_enable_key;
+static inline bool sev_es_active(void)
+{
+	return static_branch_unlikely(&sev_es_enable_key);
+}
 extern void __sev_es_ist_enter(struct pt_regs *regs);
 extern void __sev_es_ist_exit(void);
 static __always_inline void sev_es_ist_enter(struct pt_regs *regs)
@@ -198,6 +202,7 @@ bool snp_init(struct boot_params *bp);
 void __init __noreturn snp_abort(void);
 int snp_issue_guest_request(u64 exit_code, struct snp_req_data *input, unsigned long *fw_err);
 #else
+static inline bool sev_es_active(void) { return false; }
 static inline void sev_es_ist_enter(struct pt_regs *regs) { }
 static inline void sev_es_ist_exit(void) { }
 static inline int sev_es_setup_ap_jump_table(struct real_mode_header *rmh) { return 0; }
diff --git a/arch/x86/include/asm/smp.h b/arch/x86/include/asm/smp.h
index defe76ee9e64..1584f04a7007 100644
--- a/arch/x86/include/asm/smp.h
+++ b/arch/x86/include/asm/smp.h
@@ -204,7 +204,10 @@ extern unsigned int smpboot_control;
 /* Control bits for startup_64 */
 #define STARTUP_APICID_CPUID_0B	0x80000000
 #define STARTUP_APICID_CPUID_01	0x40000000
+#define STARTUP_APICID_SEV_ES	0x20000000
 
-#define STARTUP_PARALLEL_MASK (STARTUP_APICID_CPUID_01 | STARTUP_APICID_CPUID_0B)
+#define STARTUP_PARALLEL_MASK (STARTUP_APICID_CPUID_01 | \
+			       STARTUP_APICID_CPUID_0B | \
+			       STARTUP_APICID_SEV_ES)
 
 #endif /* _ASM_X86_SMP_H */
diff --git a/arch/x86/kernel/head_64.S b/arch/x86/kernel/head_64.S
index c35f7c173832..714c2afdbd9a 100644
--- a/arch/x86/kernel/head_64.S
+++ b/arch/x86/kernel/head_64.S
@@ -26,6 +26,7 @@
 #include <asm/nospec-branch.h>
 #include <asm/fixmap.h>
 #include <asm/smp.h>
+#include <asm/sev-common.h>
 
 /*
  * We are not able to switch in one step to the final KERNEL ADDRESS SPACE
@@ -242,6 +243,7 @@ SYM_INNER_LABEL(secondary_startup_64_no_verify, SYM_L_GLOBAL)
 	 *
 	 * Bit 31	STARTUP_APICID_CPUID_0B flag (use CPUID 0x0b)
 	 * Bit 30	STARTUP_APICID_CPUID_01 flag (use CPUID 0x01)
+	 * Bit 29	STARTUP_APICID_SEV_ES flag (CPUID 0x0b via GHCB MSR)
 	 * Bit 0-24	CPU# if STARTUP_APICID_CPUID_xx flags are not set
 	 */
 	movl	smpboot_control(%rip), %ecx
@@ -249,6 +251,10 @@ SYM_INNER_LABEL(secondary_startup_64_no_verify, SYM_L_GLOBAL)
 	jnz	.Luse_cpuid_0b
 	testl	$STARTUP_APICID_CPUID_01, %ecx
 	jnz	.Luse_cpuid_01
+#ifdef CONFIG_AMD_MEM_ENCRYPT
+	testl	$STARTUP_APICID_SEV_ES, %ecx
+	jnz	.Luse_sev_cpuid_0b
+#endif
 	andl	$0x0FFFFFFF, %ecx
 	jmp	.Lsetup_cpu
 
@@ -259,6 +265,30 @@ SYM_INNER_LABEL(secondary_startup_64_no_verify, SYM_L_GLOBAL)
 	shr	$24, %edx
 	jmp	.Lsetup_AP
 
+#ifdef CONFIG_AMD_MEM_ENCRYPT
+.Luse_sev_cpuid_0b:
+	/* Set the GHCB MSR to request CPUID 0xB_EDX */
+	movl	$MSR_AMD64_SEV_ES_GHCB, %ecx
+	movl	$(GHCB_CPUID_REQ_EDX << 30) | GHCB_MSR_CPUID_REQ, %eax
+	movl	$0x0B, %edx
+	wrmsr
+
+	/* Perform GHCB MSR protocol */
+	rep; vmmcall		/* vmgexit */
+
+	/*
+	 * Get the result. After the RDMSR:
+	 *   EAX should be 0xc0000005
+	 *   EDX should have the CPUID register value and since EDX
+	 *   is the target register, no need to move the result.
+	 */
+	rdmsr
+	andl	$GHCB_MSR_INFO_MASK, %eax
+	cmpl	$GHCB_MSR_CPUID_RESP, %eax
+	jne	1f
+	jmp	.Lsetup_AP
+#endif
+
 .Luse_cpuid_0b:
 	mov	$0x0B, %eax
 	xorl	%ecx, %ecx
diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c
index 0dc123a536ab..b4265c5b46da 100644
--- a/arch/x86/kernel/smpboot.c
+++ b/arch/x86/kernel/smpboot.c
@@ -1515,15 +1515,29 @@ void __init smp_prepare_cpus_common(void)
  * We can do 64-bit AP bringup in parallel if the CPU reports its APIC
  * ID in CPUID (either leaf 0x0B if we need the full APIC ID in X2APIC
  * mode, or leaf 0x01 if 8 bits are sufficient). Otherwise it's too
- * hard. And not for SEV-ES guests because they can't use CPUID that
- * early.
+ * hard.
  */
 static bool prepare_parallel_bringup(void)
 {
-	if (IS_ENABLED(CONFIG_X86_32) || cc_platform_has(CC_ATTR_GUEST_STATE_ENCRYPT))
+	bool has_sev_es = sev_es_active();
+
+	if (IS_ENABLED(CONFIG_X86_32))
 		return false;
 
-	if (x2apic_mode) {
+	/*
+	 * Encrypted guests other than SEV-ES (in the future) will need to
+	 * implement an early way of finding the APIC ID, since they will
+	 * presumably block direct CPUID too. Be kind to our future selves
+	 * by warning here instead of just letting them break. Parallel
+	 * startup doesn't have to be in the first round of enabling patches
+	 * for any such technology.
+	 */
+	if (cc_platform_has(CC_ATTR_GUEST_STATE_ENCRYPT) && !has_sev_es) {
+		pr_info("Disabling parallel bringup due to guest memory encryption\n");
+		return false;
+	}
+
+	if (x2apic_mode || has_sev_es) {
 		if (boot_cpu_data.cpuid_level < 0x0b)
 			return false;
 
@@ -1532,8 +1546,13 @@ static bool prepare_parallel_bringup(void)
 			return false;
 		}
 
-		pr_debug("Using CPUID 0xb for parallel CPU startup\n");
-		smpboot_control = STARTUP_APICID_CPUID_0B;
+		if (has_sev_es) {
+			pr_debug("Using SEV-ES CPUID 0xb for parallel CPU startup\n");
+			smpboot_control = STARTUP_APICID_SEV_ES;
+		} else {
+			pr_debug("Using CPUID 0xb for parallel CPU startup\n");
+			smpboot_control = STARTUP_APICID_CPUID_0B;
+		}
 	} else {
 		/* Without X2APIC, what's in CPUID 0x01 should suffice. */
 		if (boot_cpu_data.cpuid_level < 0x01)
-- 
2.25.1