Merge branch 'kvm-arm64/hyp-pcpu' into kvmarm-master/next

Signed-off-by: Marc Zyngier <maz@kernel.org>

arch/arm64/Kconfig

@@ -1165,32 +1165,6 @@ config UNMAP_KERNEL_AT_EL0
 
 	  If unsure, say Y.
 
-config HARDEN_BRANCH_PREDICTOR
-	bool "Harden the branch predictor against aliasing attacks" if EXPERT
-	default y
-	help
-	  Speculation attacks against some high-performance processors rely on
-	  being able to manipulate the branch predictor for a victim context by
-	  executing aliasing branches in the attacker context.  Such attacks
-	  can be partially mitigated against by clearing internal branch
-	  predictor state and limiting the prediction logic in some situations.
-
-	  This config option will take CPU-specific actions to harden the
-	  branch predictor against aliasing attacks and may rely on specific
-	  instruction sequences or control bits being set by the system
-	  firmware.
-
-	  If unsure, say Y.
-
-config ARM64_SSBD
-	bool "Speculative Store Bypass Disable" if EXPERT
-	default y
-	help
-	  This enables mitigation of the bypassing of previous stores
-	  by speculative loads.
-
-	  If unsure, say Y.
-
 config RODATA_FULL_DEFAULT_ENABLED
 	bool "Apply r/o permissions of VM areas also to their linear aliases"
 	default y

arch/arm64/include/asm/assembler.h

@@ -218,6 +218,23 @@ lr	.req	x30		// link register
 	str	\src, [\tmp, :lo12:\sym]
 	.endm
 
+	/*
+	 * @dst: destination register
+	 */
+#if defined(__KVM_NVHE_HYPERVISOR__) || defined(__KVM_VHE_HYPERVISOR__)
+	.macro	this_cpu_offset, dst
+	mrs	\dst, tpidr_el2
+	.endm
+#else
+	.macro	this_cpu_offset, dst
+alternative_if_not ARM64_HAS_VIRT_HOST_EXTN
+	mrs	\dst, tpidr_el1
+alternative_else
+	mrs	\dst, tpidr_el2
+alternative_endif
+	.endm
+#endif
+
 	/*
 	 * @dst: Result of per_cpu(sym, smp_processor_id()) (can be SP)
 	 * @sym: The name of the per-cpu variable
@@ -226,11 +243,7 @@ lr	.req	x30		// link register
 	.macro adr_this_cpu, dst, sym, tmp
 	adrp	\tmp, \sym
 	add	\dst, \tmp, #:lo12:\sym
-alternative_if_not ARM64_HAS_VIRT_HOST_EXTN
-	mrs	\tmp, tpidr_el1
-alternative_else
-	mrs	\tmp, tpidr_el2
-alternative_endif
+	this_cpu_offset \tmp
 	add	\dst, \dst, \tmp
 	.endm
 
@@ -241,11 +254,7 @@ alternative_endif
 	 */
 	.macro ldr_this_cpu dst, sym, tmp
 	adr_l	\dst, \sym
-alternative_if_not ARM64_HAS_VIRT_HOST_EXTN
-	mrs	\tmp, tpidr_el1
-alternative_else
-	mrs	\tmp, tpidr_el2
-alternative_endif
+	this_cpu_offset \tmp
 	ldr	\dst, [\dst, \tmp]
 	.endm
 

arch/arm64/include/asm/cpucaps.h

@@ -31,13 +31,13 @@
 #define ARM64_HAS_DCPOP				21
 #define ARM64_SVE				22
 #define ARM64_UNMAP_KERNEL_AT_EL0		23
-#define ARM64_HARDEN_BRANCH_PREDICTOR		24
+#define ARM64_SPECTRE_V2			24
 #define ARM64_HAS_RAS_EXTN			25
 #define ARM64_WORKAROUND_843419			26
 #define ARM64_HAS_CACHE_IDC			27
 #define ARM64_HAS_CACHE_DIC			28
 #define ARM64_HW_DBM				29
-#define ARM64_SSBD				30
+#define ARM64_SPECTRE_V4			30
 #define ARM64_MISMATCHED_CACHE_TYPE		31
 #define ARM64_HAS_STAGE2_FWB			32
 #define ARM64_HAS_CRC32				33

arch/arm64/include/asm/cpufeature.h

@@ -698,30 +698,6 @@ static inline bool system_supports_tlb_range(void)
 		cpus_have_const_cap(ARM64_HAS_TLB_RANGE);
 }
 
-#define ARM64_BP_HARDEN_UNKNOWN		-1
-#define ARM64_BP_HARDEN_WA_NEEDED	0
-#define ARM64_BP_HARDEN_NOT_REQUIRED	1
-
-int get_spectre_v2_workaround_state(void);
-
-#define ARM64_SSBD_UNKNOWN		-1
-#define ARM64_SSBD_FORCE_DISABLE	0
-#define ARM64_SSBD_KERNEL		1
-#define ARM64_SSBD_FORCE_ENABLE		2
-#define ARM64_SSBD_MITIGATED		3
-
-static inline int arm64_get_ssbd_state(void)
-{
-#ifdef CONFIG_ARM64_SSBD
-	extern int ssbd_state;
-	return ssbd_state;
-#else
-	return ARM64_SSBD_UNKNOWN;
-#endif
-}
-
-void arm64_set_ssbd_mitigation(bool state);
-
 extern int do_emulate_mrs(struct pt_regs *regs, u32 sys_reg, u32 rt);
 
 static inline u32 id_aa64mmfr0_parange_to_phys_shift(int parange)

arch/arm64/include/asm/hyp_image.h

+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Copyright (C) 2020 Google LLC.
+ * Written by David Brazdil <dbrazdil@google.com>
+ */
+
+#ifndef __ARM64_HYP_IMAGE_H__
+#define __ARM64_HYP_IMAGE_H__
+
+/*
+ * KVM nVHE code has its own symbol namespace prefixed with __kvm_nvhe_,
+ * to separate it from the kernel proper.
+ */
+#define kvm_nvhe_sym(sym)	__kvm_nvhe_##sym
+
+#ifdef LINKER_SCRIPT
+
+/*
+ * KVM nVHE ELF section names are prefixed with .hyp, to separate them
+ * from the kernel proper.
+ */
+#define HYP_SECTION_NAME(NAME)	.hyp##NAME
+
+/* Defines an ELF hyp section from input section @NAME and its subsections. */
+#define HYP_SECTION(NAME) \
+	HYP_SECTION_NAME(NAME) : { *(NAME NAME##.*) }
+
+/*
+ * Defines a linker script alias of a kernel-proper symbol referenced by
+ * KVM nVHE hyp code.
+ */
+#define KVM_NVHE_ALIAS(sym)	kvm_nvhe_sym(sym) = sym;
+
+#endif /* LINKER_SCRIPT */
+
+#endif /* __ARM64_HYP_IMAGE_H__ */

arch/arm64/include/asm/kvm_asm.h

@@ -7,11 +7,9 @@
 #ifndef __ARM_KVM_ASM_H__
 #define __ARM_KVM_ASM_H__
 
+#include <asm/hyp_image.h>
 #include <asm/virt.h>
 
-#define	VCPU_WORKAROUND_2_FLAG_SHIFT	0
-#define	VCPU_WORKAROUND_2_FLAG		(_AC(1, UL) << VCPU_WORKAROUND_2_FLAG_SHIFT)
-
 #define ARM_EXIT_WITH_SERROR_BIT  31
 #define ARM_EXCEPTION_CODE(x)	  ((x) & ~(1U << ARM_EXIT_WITH_SERROR_BIT))
 #define ARM_EXCEPTION_IS_TRAP(x)  (ARM_EXCEPTION_CODE((x)) == ARM_EXCEPTION_TRAP)
@@ -66,13 +64,6 @@
 
 #include <linux/mm.h>
 
-/*
- * Translate name of a symbol defined in nVHE hyp to the name seen
- * by kernel proper. All nVHE symbols are prefixed by the build system
- * to avoid clashes with the VHE variants.
- */
-#define kvm_nvhe_sym(sym)	__kvm_nvhe_##sym
-
 #define DECLARE_KVM_VHE_SYM(sym)	extern char sym[]
 #define DECLARE_KVM_NVHE_SYM(sym)	extern char kvm_nvhe_sym(sym)[]
 
@@ -84,21 +75,50 @@
 	DECLARE_KVM_VHE_SYM(sym);		\
 	DECLARE_KVM_NVHE_SYM(sym)
 
+#define DECLARE_KVM_VHE_PER_CPU(type, sym)	\
+	DECLARE_PER_CPU(type, sym)
+#define DECLARE_KVM_NVHE_PER_CPU(type, sym)	\
+	DECLARE_PER_CPU(type, kvm_nvhe_sym(sym))
+
+#define DECLARE_KVM_HYP_PER_CPU(type, sym)	\
+	DECLARE_KVM_VHE_PER_CPU(type, sym);	\
+	DECLARE_KVM_NVHE_PER_CPU(type, sym)
+
+/*
+ * Compute pointer to a symbol defined in nVHE percpu region.
+ * Returns NULL if percpu memory has not been allocated yet.
+ */
+#define this_cpu_ptr_nvhe_sym(sym)	per_cpu_ptr_nvhe_sym(sym, smp_processor_id())
+#define per_cpu_ptr_nvhe_sym(sym, cpu)						\
+	({									\
+		unsigned long base, off;					\
+		base = kvm_arm_hyp_percpu_base[cpu];				\
+		off = (unsigned long)&CHOOSE_NVHE_SYM(sym) -			\
+		      (unsigned long)&CHOOSE_NVHE_SYM(__per_cpu_start);		\
+		base ? (typeof(CHOOSE_NVHE_SYM(sym))*)(base + off) : NULL;	\
+	})
+
 #if defined(__KVM_NVHE_HYPERVISOR__)
 
-#define CHOOSE_HYP_SYM(sym)	CHOOSE_NVHE_SYM(sym)
 #define CHOOSE_NVHE_SYM(sym)	sym
+#define CHOOSE_HYP_SYM(sym)	CHOOSE_NVHE_SYM(sym)
+
 /* The nVHE hypervisor shouldn't even try to access VHE symbols */
 extern void *__nvhe_undefined_symbol;
 #define CHOOSE_VHE_SYM(sym)		__nvhe_undefined_symbol
+#define this_cpu_ptr_hyp_sym(sym)	(&__nvhe_undefined_symbol)
+#define per_cpu_ptr_hyp_sym(sym, cpu)	(&__nvhe_undefined_symbol)
 
-#elif defined(__KVM_VHE_HYPERVISOR)
+#elif defined(__KVM_VHE_HYPERVISOR__)
 
-#define CHOOSE_HYP_SYM(sym)	CHOOSE_VHE_SYM(sym)
 #define CHOOSE_VHE_SYM(sym)	sym
+#define CHOOSE_HYP_SYM(sym)	CHOOSE_VHE_SYM(sym)
+
 /* The VHE hypervisor shouldn't even try to access nVHE symbols */
 extern void *__vhe_undefined_symbol;
 #define CHOOSE_NVHE_SYM(sym)		__vhe_undefined_symbol
+#define this_cpu_ptr_hyp_sym(sym)	(&__vhe_undefined_symbol)
+#define per_cpu_ptr_hyp_sym(sym, cpu)	(&__vhe_undefined_symbol)
 
 #else
 
@@ -113,8 +133,18 @@ extern void *__vhe_undefined_symbol;
  * - Don't let the nVHE hypervisor have access to this, as it will
  *   pick the *wrong* symbol (yes, it runs at EL2...).
  */
-#define CHOOSE_HYP_SYM(sym)	(is_kernel_in_hyp_mode() ? CHOOSE_VHE_SYM(sym) \
+#define CHOOSE_HYP_SYM(sym)		(is_kernel_in_hyp_mode()	\
+					   ? CHOOSE_VHE_SYM(sym)	\
 					   : CHOOSE_NVHE_SYM(sym))
+
+#define this_cpu_ptr_hyp_sym(sym)	(is_kernel_in_hyp_mode()	\
+					   ? this_cpu_ptr(&sym)		\
+					   : this_cpu_ptr_nvhe_sym(sym))
+
+#define per_cpu_ptr_hyp_sym(sym, cpu)	(is_kernel_in_hyp_mode()	\
+					   ? per_cpu_ptr(&sym, cpu)	\
+					   : per_cpu_ptr_nvhe_sym(sym, cpu))
+
 #define CHOOSE_VHE_SYM(sym)	sym
 #define CHOOSE_NVHE_SYM(sym)	kvm_nvhe_sym(sym)
 
@@ -141,11 +171,13 @@ DECLARE_KVM_HYP_SYM(__kvm_hyp_vector);
 #define __kvm_hyp_host_vector	CHOOSE_NVHE_SYM(__kvm_hyp_host_vector)
 #define __kvm_hyp_vector	CHOOSE_HYP_SYM(__kvm_hyp_vector)
 
-#ifdef CONFIG_KVM_INDIRECT_VECTORS
+extern unsigned long kvm_arm_hyp_percpu_base[NR_CPUS];
+DECLARE_KVM_NVHE_SYM(__per_cpu_start);
+DECLARE_KVM_NVHE_SYM(__per_cpu_end);
+
 extern atomic_t arm64_el2_vector_last_slot;
 DECLARE_KVM_HYP_SYM(__bp_harden_hyp_vecs);
 #define __bp_harden_hyp_vecs	CHOOSE_HYP_SYM(__bp_harden_hyp_vecs)
-#endif
 
 extern void __kvm_flush_vm_context(void);
 extern void __kvm_tlb_flush_vmid_ipa(struct kvm_s2_mmu *mmu, phys_addr_t ipa,
@@ -188,26 +220,6 @@ extern char __smccc_workaround_1_smc[__SMCCC_WORKAROUND_1_SMC_SZ];
 		addr;							\
 	})
 
-/*
- * Home-grown __this_cpu_{ptr,read} variants that always work at HYP,
- * provided that sym is really a *symbol* and not a pointer obtained from
- * a data structure. As for SHIFT_PERCPU_PTR(), the creative casting keeps
- * sparse quiet.
- */
-#define __hyp_this_cpu_ptr(sym)						\
-	({								\
-		void *__ptr;						\
-		__verify_pcpu_ptr(&sym);				\
-		__ptr = hyp_symbol_addr(sym);				\
-		__ptr += read_sysreg(tpidr_el2);			\
-		(typeof(sym) __kernel __force *)__ptr;			\
-	 })
-
-#define __hyp_this_cpu_read(sym)					\
-	({								\
-		*__hyp_this_cpu_ptr(sym);				\
-	 })
-
 #define __KVM_EXTABLE(from, to)						\
 	"	.pushsection	__kvm_ex_table, \"a\"\n"		\
 	"	.align		3\n"					\
@@ -238,20 +250,8 @@ extern char __smccc_workaround_1_smc[__SMCCC_WORKAROUND_1_SMC_SZ];
 
 #else /* __ASSEMBLY__ */
 
-.macro hyp_adr_this_cpu reg, sym, tmp
-	adr_l	\reg, \sym
-	mrs	\tmp, tpidr_el2
-	add	\reg, \reg, \tmp
-.endm
-
-.macro hyp_ldr_this_cpu reg, sym, tmp
-	adr_l	\reg, \sym
-	mrs	\tmp, tpidr_el2
-	ldr	\reg,  [\reg, \tmp]
-.endm
-
 .macro get_host_ctxt reg, tmp
-	hyp_adr_this_cpu \reg, kvm_host_data, \tmp
+	adr_this_cpu \reg, kvm_host_data, \tmp
 	add	\reg, \reg, #HOST_DATA_CONTEXT
 .endm
 
@@ -261,12 +261,12 @@ extern char __smccc_workaround_1_smc[__SMCCC_WORKAROUND_1_SMC_SZ];
 .endm
 
 .macro get_loaded_vcpu vcpu, ctxt
-	hyp_adr_this_cpu \ctxt, kvm_hyp_ctxt, \vcpu
+	adr_this_cpu \ctxt, kvm_hyp_ctxt, \vcpu
 	ldr	\vcpu, [\ctxt, #HOST_CONTEXT_VCPU]
 .endm
 
 .macro set_loaded_vcpu vcpu, ctxt, tmp
-	hyp_adr_this_cpu \ctxt, kvm_hyp_ctxt, \tmp
+	adr_this_cpu \ctxt, kvm_hyp_ctxt, \tmp
 	str	\vcpu, [\ctxt, #HOST_CONTEXT_VCPU]
 .endm
 

arch/arm64/include/asm/kvm_emulate.h

@@ -383,20 +383,6 @@ static inline unsigned long kvm_vcpu_get_mpidr_aff(struct kvm_vcpu *vcpu)
 	return vcpu_read_sys_reg(vcpu, MPIDR_EL1) & MPIDR_HWID_BITMASK;
 }
 
-static inline bool kvm_arm_get_vcpu_workaround_2_flag(struct kvm_vcpu *vcpu)
-{
-	return vcpu->arch.workaround_flags & VCPU_WORKAROUND_2_FLAG;
-}
-
-static inline void kvm_arm_set_vcpu_workaround_2_flag(struct kvm_vcpu *vcpu,
-						      bool flag)
-{
-	if (flag)
-		vcpu->arch.workaround_flags |= VCPU_WORKAROUND_2_FLAG;
-	else
-		vcpu->arch.workaround_flags &= ~VCPU_WORKAROUND_2_FLAG;
-}
-
 static inline void kvm_vcpu_set_be(struct kvm_vcpu *vcpu)
 {
 	if (vcpu_mode_is_32bit(vcpu)) {

arch/arm64/include/asm/kvm_host.h

@@ -568,7 +568,7 @@ void kvm_set_sei_esr(struct kvm_vcpu *vcpu, u64 syndrome);
 
 struct kvm_vcpu *kvm_mpidr_to_vcpu(struct kvm *kvm, unsigned long mpidr);
 
-DECLARE_PER_CPU(struct kvm_host_data, kvm_host_data);
+DECLARE_KVM_HYP_PER_CPU(struct kvm_host_data, kvm_host_data);
 
 static inline void kvm_init_host_cpu_context(struct kvm_cpu_context *cpu_ctxt)
 {
@@ -634,46 +634,6 @@ static inline void kvm_set_pmu_events(u32 set, struct perf_event_attr *attr) {}
 static inline void kvm_clr_pmu_events(u32 clr) {}
 #endif
 
-#define KVM_BP_HARDEN_UNKNOWN		-1
-#define KVM_BP_HARDEN_WA_NEEDED		0
-#define KVM_BP_HARDEN_NOT_REQUIRED	1
-
-static inline int kvm_arm_harden_branch_predictor(void)
-{
-	switch (get_spectre_v2_workaround_state()) {
-	case ARM64_BP_HARDEN_WA_NEEDED:
-		return KVM_BP_HARDEN_WA_NEEDED;
-	case ARM64_BP_HARDEN_NOT_REQUIRED:
-		return KVM_BP_HARDEN_NOT_REQUIRED;
-	case ARM64_BP_HARDEN_UNKNOWN:
-	default:
-		return KVM_BP_HARDEN_UNKNOWN;
-	}
-}
-
-#define KVM_SSBD_UNKNOWN		-1
-#define KVM_SSBD_FORCE_DISABLE		0
-#define KVM_SSBD_KERNEL		1
-#define KVM_SSBD_FORCE_ENABLE		2
-#define KVM_SSBD_MITIGATED		3
-
-static inline int kvm_arm_have_ssbd(void)
-{
-	switch (arm64_get_ssbd_state()) {
-	case ARM64_SSBD_FORCE_DISABLE:
-		return KVM_SSBD_FORCE_DISABLE;
-	case ARM64_SSBD_KERNEL:
-		return KVM_SSBD_KERNEL;
-	case ARM64_SSBD_FORCE_ENABLE:
-		return KVM_SSBD_FORCE_ENABLE;
-	case ARM64_SSBD_MITIGATED:
-		return KVM_SSBD_MITIGATED;
-	case ARM64_SSBD_UNKNOWN:
-	default:
-		return KVM_SSBD_UNKNOWN;
-	}
-}
-
 void kvm_vcpu_load_sysregs_vhe(struct kvm_vcpu *vcpu);
 void kvm_vcpu_put_sysregs_vhe(struct kvm_vcpu *vcpu);
 

arch/arm64/include/asm/kvm_mmu.h

@@ -9,6 +9,7 @@
 
 #include <asm/page.h>
 #include <asm/memory.h>
+#include <asm/mmu.h>
 #include <asm/cpufeature.h>
 
 /*
@@ -207,19 +208,17 @@ static inline int kvm_write_guest_lock(struct kvm *kvm, gpa_t gpa,
 	return ret;
 }
 
-#ifdef CONFIG_KVM_INDIRECT_VECTORS
 /*
  * EL2 vectors can be mapped and rerouted in a number of ways,
  * depending on the kernel configuration and CPU present:
  *
- * - If the CPU has the ARM64_HARDEN_BRANCH_PREDICTOR cap, the
- *   hardening sequence is placed in one of the vector slots, which is
- *   executed before jumping to the real vectors.
+ * - If the CPU is affected by Spectre-v2, the hardening sequence is
+ *   placed in one of the vector slots, which is executed before jumping
+ *   to the real vectors.
  *
- * - If the CPU has both the ARM64_HARDEN_EL2_VECTORS cap and the
- *   ARM64_HARDEN_BRANCH_PREDICTOR cap, the slot containing the
- *   hardening sequence is mapped next to the idmap page, and executed
- *   before jumping to the real vectors.
+ * - If the CPU also has the ARM64_HARDEN_EL2_VECTORS cap, the slot
+ *   containing the hardening sequence is mapped next to the idmap page,
+ *   and executed before jumping to the real vectors.
  *
  * - If the CPU only has the ARM64_HARDEN_EL2_VECTORS cap, then an
  *   empty slot is selected, mapped next to the idmap page, and
@@ -229,19 +228,16 @@ static inline int kvm_write_guest_lock(struct kvm *kvm, gpa_t gpa,
  * VHE, as we don't have hypervisor-specific mappings. If the system
  * is VHE and yet selects this capability, it will be ignored.
  */
-#include <asm/mmu.h>
-
 extern void *__kvm_bp_vect_base;
 extern int __kvm_harden_el2_vector_slot;
 
-/*  This is called on both VHE and !VHE systems */
 static inline void *kvm_get_hyp_vector(void)
 {
 	struct bp_hardening_data *data = arm64_get_bp_hardening_data();
 	void *vect = kern_hyp_va(kvm_ksym_ref(__kvm_hyp_vector));
 	int slot = -1;
 
-	if (cpus_have_const_cap(ARM64_HARDEN_BRANCH_PREDICTOR) && data->fn) {
+	if (cpus_have_const_cap(ARM64_SPECTRE_V2) && data->fn) {
 		vect = kern_hyp_va(kvm_ksym_ref(__bp_harden_hyp_vecs));
 		slot = data->hyp_vectors_slot;
 	}
@@ -258,76 +254,6 @@ static inline void *kvm_get_hyp_vector(void)
 	return vect;
 }
 
-/*  This is only called on a !VHE system */
-static inline int kvm_map_vectors(void)
-{
-	/*
-	 * HBP  = ARM64_HARDEN_BRANCH_PREDICTOR
-	 * HEL2 = ARM64_HARDEN_EL2_VECTORS
-	 *
-	 * !HBP + !HEL2 -> use direct vectors
-	 *  HBP + !HEL2 -> use hardened vectors in place
-	 * !HBP +  HEL2 -> allocate one vector slot and use exec mapping
-	 *  HBP +  HEL2 -> use hardened vertors and use exec mapping
-	 */
-	if (cpus_have_const_cap(ARM64_HARDEN_BRANCH_PREDICTOR)) {
-		__kvm_bp_vect_base = kvm_ksym_ref(__bp_harden_hyp_vecs);
-		__kvm_bp_vect_base = kern_hyp_va(__kvm_bp_vect_base);
-	}
-
-	if (cpus_have_const_cap(ARM64_HARDEN_EL2_VECTORS)) {
-		phys_addr_t vect_pa = __pa_symbol(__bp_harden_hyp_vecs);
-		unsigned long size = __BP_HARDEN_HYP_VECS_SZ;
-
-		/*
-		 * Always allocate a spare vector slot, as we don't
-		 * know yet which CPUs have a BP hardening slot that
-		 * we can reuse.
-		 */
-		__kvm_harden_el2_vector_slot = atomic_inc_return(&arm64_el2_vector_last_slot);
-		BUG_ON(__kvm_harden_el2_vector_slot >= BP_HARDEN_EL2_SLOTS);
-		return create_hyp_exec_mappings(vect_pa, size,
-						&__kvm_bp_vect_base);
-	}
-
-	return 0;
-}
-#else
-static inline void *kvm_get_hyp_vector(void)
-{
-	return kern_hyp_va(kvm_ksym_ref(__kvm_hyp_vector));
-}
-
-static inline int kvm_map_vectors(void)
-{
-	return 0;
-}
-#endif
-
-#ifdef CONFIG_ARM64_SSBD
-DECLARE_PER_CPU_READ_MOSTLY(u64, arm64_ssbd_callback_required);
-
-static inline int hyp_map_aux_data(void)
-{
-	int cpu, err;
-
-	for_each_possible_cpu(cpu) {
-		u64 *ptr;
-
-		ptr = per_cpu_ptr(&arm64_ssbd_callback_required, cpu);
-		err = create_hyp_mappings(ptr, ptr + 1, PAGE_HYP);
-		if (err)
-			return err;
-	}
-	return 0;
-}
-#else
-static inline int hyp_map_aux_data(void)
-{
-	return 0;
-}
-#endif
-
 #define kvm_phys_to_vttbr(addr)		phys_to_ttbr(addr)
 
 static __always_inline u64 kvm_get_vttbr(struct kvm_s2_mmu *mmu)

arch/arm64/include/asm/mmu.h

@@ -45,7 +45,6 @@ struct bp_hardening_data {
 	bp_hardening_cb_t	fn;
 };
 
-#ifdef CONFIG_HARDEN_BRANCH_PREDICTOR
 DECLARE_PER_CPU_READ_MOSTLY(struct bp_hardening_data, bp_hardening_data);
 
 static inline struct bp_hardening_data *arm64_get_bp_hardening_data(void)
@@ -57,21 +56,13 @@ static inline void arm64_apply_bp_hardening(void)
 {
 	struct bp_hardening_data *d;
 
-	if (!cpus_have_const_cap(ARM64_HARDEN_BRANCH_PREDICTOR))
+	if (!cpus_have_const_cap(ARM64_SPECTRE_V2))
 		return;
 
 	d = arm64_get_bp_hardening_data();
 	if (d->fn)
 		d->fn();
 }
-#else
-static inline struct bp_hardening_data *arm64_get_bp_hardening_data(void)
-{
-	return NULL;
-}
-
-static inline void arm64_apply_bp_hardening(void)	{ }
-#endif	/* CONFIG_HARDEN_BRANCH_PREDICTOR */
 
 extern void arm64_memblock_init(void);
 extern void paging_init(void);

arch/arm64/include/asm/percpu.h

@@ -19,7 +19,16 @@ static inline void set_my_cpu_offset(unsigned long off)
 			:: "r" (off) : "memory");
 }
 
-static inline unsigned long __my_cpu_offset(void)
+static inline unsigned long __hyp_my_cpu_offset(void)
+{
+	/*
+	 * Non-VHE hyp code runs with preemption disabled. No need to hazard
+	 * the register access against barrier() as in __kern_my_cpu_offset.
+	 */
+	return read_sysreg(tpidr_el2);
+}
+
+static inline unsigned long __kern_my_cpu_offset(void)
 {
 	unsigned long off;
 
@@ -35,7 +44,12 @@ static inline unsigned long __my_cpu_offset(void)
 
 	return off;
 }
-#define __my_cpu_offset __my_cpu_offset()
+
+#ifdef __KVM_NVHE_HYPERVISOR__
+#define __my_cpu_offset __hyp_my_cpu_offset()
+#else
+#define __my_cpu_offset __kern_my_cpu_offset()
+#endif
 
 #define PERCPU_RW_OPS(sz)						\
 static inline unsigned long __percpu_read_##sz(void *ptr)		\
@@ -227,4 +241,14 @@ PERCPU_RET_OP(add, add, ldadd)
 
 #include <asm-generic/percpu.h>
 
+/* Redefine macros for nVHE hyp under DEBUG_PREEMPT to avoid its dependencies. */
+#if defined(__KVM_NVHE_HYPERVISOR__) && defined(CONFIG_DEBUG_PREEMPT)
+#undef	this_cpu_ptr
+#define	this_cpu_ptr		raw_cpu_ptr
+#undef	__this_cpu_read
+#define	__this_cpu_read		raw_cpu_read
+#undef	__this_cpu_write
+#define	__this_cpu_write	raw_cpu_write
+#endif
+
 #endif /* __ASM_PERCPU_H */

arch/arm64/include/asm/processor.h

@@ -38,6 +38,7 @@
 #include <asm/pgtable-hwdef.h>
 #include <asm/pointer_auth.h>
 #include <asm/ptrace.h>
+#include <asm/spectre.h>
 #include <asm/types.h>
 
 /*
@@ -197,40 +198,15 @@ static inline void start_thread_common(struct pt_regs *regs, unsigned long pc)
 		regs->pmr_save = GIC_PRIO_IRQON;
 }
 
-static inline void set_ssbs_bit(struct pt_regs *regs)
-{
-	regs->pstate |= PSR_SSBS_BIT;
-}
-
-static inline void set_compat_ssbs_bit(struct pt_regs *regs)
-{
-	regs->pstate |= PSR_AA32_SSBS_BIT;
-}
-
 static inline void start_thread(struct pt_regs *regs, unsigned long pc,
 				unsigned long sp)
 {
 	start_thread_common(regs, pc);
 	regs->pstate = PSR_MODE_EL0t;
-
-	if (arm64_get_ssbd_state() != ARM64_SSBD_FORCE_ENABLE)
-		set_ssbs_bit(regs);
-
+	spectre_v4_enable_task_mitigation(current);
 	regs->sp = sp;
 }
 
-static inline bool is_ttbr0_addr(unsigned long addr)
-{
-	/* entry assembly clears tags for TTBR0 addrs */
-	return addr < TASK_SIZE;
-}
-
-static inline bool is_ttbr1_addr(unsigned long addr)
-{
-	/* TTBR1 addresses may have a tag if KASAN_SW_TAGS is in use */
-	return arch_kasan_reset_tag(addr) >= PAGE_OFFSET;
-}
-
 #ifdef CONFIG_COMPAT
 static inline void compat_start_thread(struct pt_regs *regs, unsigned long pc,
 				       unsigned long sp)
@@ -244,13 +220,23 @@ static inline void compat_start_thread(struct pt_regs *regs, unsigned long pc,
 	regs->pstate |= PSR_AA32_E_BIT;
 #endif
 
-	if (arm64_get_ssbd_state() != ARM64_SSBD_FORCE_ENABLE)
-		set_compat_ssbs_bit(regs);
-
+	spectre_v4_enable_task_mitigation(current);
 	regs->compat_sp = sp;
 }
 #endif
 
+static inline bool is_ttbr0_addr(unsigned long addr)
+{
+	/* entry assembly clears tags for TTBR0 addrs */
+	return addr < TASK_SIZE;
+}
+
+static inline bool is_ttbr1_addr(unsigned long addr)
+{
+	/* TTBR1 addresses may have a tag if KASAN_SW_TAGS is in use */
+	return arch_kasan_reset_tag(addr) >= PAGE_OFFSET;
+}
+
 /* Forward declaration, a strange C thing */
 struct task_struct;
 

arch/arm64/include/asm/spectre.h

+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * Interface for managing mitigations for Spectre vulnerabilities.
+ *
+ * Copyright (C) 2020 Google LLC
+ * Author: Will Deacon <will@kernel.org>
+ */
+
+#ifndef __ASM_SPECTRE_H
+#define __ASM_SPECTRE_H
+
+#include <asm/cpufeature.h>
+
+/* Watch out, ordering is important here. */
+enum mitigation_state {
+	SPECTRE_UNAFFECTED,
+	SPECTRE_MITIGATED,
+	SPECTRE_VULNERABLE,
+};
+
+struct task_struct;
+
+enum mitigation_state arm64_get_spectre_v2_state(void);
+bool has_spectre_v2(const struct arm64_cpu_capabilities *cap, int scope);
+void spectre_v2_enable_mitigation(const struct arm64_cpu_capabilities *__unused);
+
+enum mitigation_state arm64_get_spectre_v4_state(void);
+bool has_spectre_v4(const struct arm64_cpu_capabilities *cap, int scope);
+void spectre_v4_enable_mitigation(const struct arm64_cpu_capabilities *__unused);
+void spectre_v4_enable_task_mitigation(struct task_struct *tsk);
+
+#endif	/* __ASM_SPECTRE_H */

arch/arm64/include/uapi/asm/kvm.h

@@ -257,6 +257,15 @@ struct kvm_vcpu_events {
 #define KVM_REG_ARM_SMCCC_ARCH_WORKAROUND_1_NOT_AVAIL		0
 #define KVM_REG_ARM_SMCCC_ARCH_WORKAROUND_1_AVAIL		1
 #define KVM_REG_ARM_SMCCC_ARCH_WORKAROUND_1_NOT_REQUIRED	2
+
+/*
+ * Only two states can be presented by the host kernel:
+ * - NOT_REQUIRED: the guest doesn't need to do anything
+ * - NOT_AVAIL: the guest isn't mitigated (it can still use SSBS if available)
+ *
+ * All the other values are deprecated. The host still accepts all
+ * values (they are ABI), but will narrow them to the above two.
+ */
 #define KVM_REG_ARM_SMCCC_ARCH_WORKAROUND_2	KVM_REG_ARM_FW_REG(2)
 #define KVM_REG_ARM_SMCCC_ARCH_WORKAROUND_2_NOT_AVAIL		0
 #define KVM_REG_ARM_SMCCC_ARCH_WORKAROUND_2_UNKNOWN		1

arch/arm64/kernel/Makefile

@@ -19,7 +19,7 @@ obj-y			:= debug-monitors.o entry.o irq.o fpsimd.o		\
 			   return_address.o cpuinfo.o cpu_errata.o		\
 			   cpufeature.o alternative.o cacheinfo.o		\
 			   smp.o smp_spin_table.o topology.o smccc-call.o	\
-			   syscall.o
+			   syscall.o proton-pack.o
 
 targets			+= efi-entry.o
 
@@ -59,7 +59,6 @@ arm64-reloc-test-y := reloc_test_core.o reloc_test_syms.o
 obj-$(CONFIG_CRASH_DUMP)		+= crash_dump.o
 obj-$(CONFIG_CRASH_CORE)		+= crash_core.o
 obj-$(CONFIG_ARM_SDE_INTERFACE)		+= sdei.o
-obj-$(CONFIG_ARM64_SSBD)		+= ssbd.o
 obj-$(CONFIG_ARM64_PTR_AUTH)		+= pointer_auth.o
 obj-$(CONFIG_SHADOW_CALL_STACK)		+= scs.o
 

arch/arm64/kernel/cpu_errata.c

@@ -106,365 +106,6 @@ cpu_enable_trap_ctr_access(const struct arm64_cpu_capabilities *cap)
 		sysreg_clear_set(sctlr_el1, SCTLR_EL1_UCT, 0);
 }
 
-atomic_t arm64_el2_vector_last_slot = ATOMIC_INIT(-1);
-
-#include <asm/mmu_context.h>
-#include <asm/cacheflush.h>
-
-DEFINE_PER_CPU_READ_MOSTLY(struct bp_hardening_data, bp_hardening_data);
-
-#ifdef CONFIG_KVM_INDIRECT_VECTORS
-static void __copy_hyp_vect_bpi(int slot, const char *hyp_vecs_start,
-				const char *hyp_vecs_end)
-{
-	void *dst = lm_alias(__bp_harden_hyp_vecs + slot * SZ_2K);
-	int i;
-
-	for (i = 0; i < SZ_2K; i += 0x80)
-		memcpy(dst + i, hyp_vecs_start, hyp_vecs_end - hyp_vecs_start);
-
-	__flush_icache_range((uintptr_t)dst, (uintptr_t)dst + SZ_2K);
-}
-
-static void install_bp_hardening_cb(bp_hardening_cb_t fn,
-				    const char *hyp_vecs_start,
-				    const char *hyp_vecs_end)
-{
-	static DEFINE_RAW_SPINLOCK(bp_lock);
-	int cpu, slot = -1;
-
-	/*
-	 * detect_harden_bp_fw() passes NULL for the hyp_vecs start/end if
-	 * we're a guest. Skip the hyp-vectors work.
-	 */
-	if (!hyp_vecs_start) {
-		__this_cpu_write(bp_hardening_data.fn, fn);
-		return;
-	}
-
-	raw_spin_lock(&bp_lock);
-	for_each_possible_cpu(cpu) {
-		if (per_cpu(bp_hardening_data.fn, cpu) == fn) {
-			slot = per_cpu(bp_hardening_data.hyp_vectors_slot, cpu);
-			break;
-		}
-	}
-
-	if (slot == -1) {
-		slot = atomic_inc_return(&arm64_el2_vector_last_slot);
-		BUG_ON(slot >= BP_HARDEN_EL2_SLOTS);
-		__copy_hyp_vect_bpi(slot, hyp_vecs_start, hyp_vecs_end);
-	}
-
-	__this_cpu_write(bp_hardening_data.hyp_vectors_slot, slot);
-	__this_cpu_write(bp_hardening_data.fn, fn);
-	raw_spin_unlock(&bp_lock);
-}
-#else
-static void install_bp_hardening_cb(bp_hardening_cb_t fn,
-				      const char *hyp_vecs_start,
-				      const char *hyp_vecs_end)
-{
-	__this_cpu_write(bp_hardening_data.fn, fn);
-}
-#endif	/* CONFIG_KVM_INDIRECT_VECTORS */
-
-#include <linux/arm-smccc.h>
-
-static void __maybe_unused call_smc_arch_workaround_1(void)
-{
-	arm_smccc_1_1_smc(ARM_SMCCC_ARCH_WORKAROUND_1, NULL);
-}
-
-static void call_hvc_arch_workaround_1(void)
-{
-	arm_smccc_1_1_hvc(ARM_SMCCC_ARCH_WORKAROUND_1, NULL);
-}
-
-static void qcom_link_stack_sanitization(void)
-{
-	u64 tmp;
-
-	asm volatile("mov	%0, x30		\n"
-		     ".rept	16		\n"
-		     "bl	. + 4		\n"
-		     ".endr			\n"
-		     "mov	x30, %0		\n"
-		     : "=&r" (tmp));
-}
-
-static bool __nospectre_v2;
-static int __init parse_nospectre_v2(char *str)
-{
-	__nospectre_v2 = true;
-	return 0;
-}
-early_param("nospectre_v2", parse_nospectre_v2);
-
-/*
- * -1: No workaround
- *  0: No workaround required
- *  1: Workaround installed
- */
-static int detect_harden_bp_fw(void)
-{
-	bp_hardening_cb_t cb;
-	void *smccc_start, *smccc_end;
-	struct arm_smccc_res res;
-	u32 midr = read_cpuid_id();
-
-	arm_smccc_1_1_invoke(ARM_SMCCC_ARCH_FEATURES_FUNC_ID,
-			     ARM_SMCCC_ARCH_WORKAROUND_1, &res);
-
-	switch ((int)res.a0) {
-	case 1:
-		/* Firmware says we're just fine */
-		return 0;
-	case 0:
-		break;
-	default:
-		return -1;
-	}
-
-	switch (arm_smccc_1_1_get_conduit()) {
-	case SMCCC_CONDUIT_HVC:
-		cb = call_hvc_arch_workaround_1;
-		/* This is a guest, no need to patch KVM vectors */
-		smccc_start = NULL;
-		smccc_end = NULL;
-		break;
-
-#if IS_ENABLED(CONFIG_KVM)
-	case SMCCC_CONDUIT_SMC:
-		cb = call_smc_arch_workaround_1;
-		smccc_start = __smccc_workaround_1_smc;
-		smccc_end = __smccc_workaround_1_smc +
-			__SMCCC_WORKAROUND_1_SMC_SZ;
-		break;
-#endif
-
-	default:
-		return -1;
-	}
-
-	if (((midr & MIDR_CPU_MODEL_MASK) == MIDR_QCOM_FALKOR) ||
-	    ((midr & MIDR_CPU_MODEL_MASK) == MIDR_QCOM_FALKOR_V1))
-		cb = qcom_link_stack_sanitization;
-
-	if (IS_ENABLED(CONFIG_HARDEN_BRANCH_PREDICTOR))
-		install_bp_hardening_cb(cb, smccc_start, smccc_end);
-
-	return 1;
-}
-
-DEFINE_PER_CPU_READ_MOSTLY(u64, arm64_ssbd_callback_required);
-
-int ssbd_state __read_mostly = ARM64_SSBD_KERNEL;
-static bool __ssb_safe = true;
-
-static const struct ssbd_options {
-	const char	*str;
-	int		state;
-} ssbd_options[] = {
-	{ "force-on",	ARM64_SSBD_FORCE_ENABLE, },
-	{ "force-off",	ARM64_SSBD_FORCE_DISABLE, },
-	{ "kernel",	ARM64_SSBD_KERNEL, },
-};
-
-static int __init ssbd_cfg(char *buf)
-{
-	int i;
-
-	if (!buf || !buf[0])
-		return -EINVAL;
-
-	for (i = 0; i < ARRAY_SIZE(ssbd_options); i++) {
-		int len = strlen(ssbd_options[i].str);
-
-		if (strncmp(buf, ssbd_options[i].str, len))
-			continue;
-
-		ssbd_state = ssbd_options[i].state;
-		return 0;
-	}
-
-	return -EINVAL;
-}
-early_param("ssbd", ssbd_cfg);
-
-void __init arm64_update_smccc_conduit(struct alt_instr *alt,
-				       __le32 *origptr, __le32 *updptr,
-				       int nr_inst)
-{
-	u32 insn;
-
-	BUG_ON(nr_inst != 1);
-
-	switch (arm_smccc_1_1_get_conduit()) {
-	case SMCCC_CONDUIT_HVC:
-		insn = aarch64_insn_get_hvc_value();
-		break;
-	case SMCCC_CONDUIT_SMC:
-		insn = aarch64_insn_get_smc_value();
-		break;
-	default:
-		return;
-	}
-
-	*updptr = cpu_to_le32(insn);
-}
-
-void __init arm64_enable_wa2_handling(struct alt_instr *alt,
-				      __le32 *origptr, __le32 *updptr,
-				      int nr_inst)
-{
-	BUG_ON(nr_inst != 1);
-	/*
-	 * Only allow mitigation on EL1 entry/exit and guest
-	 * ARCH_WORKAROUND_2 handling if the SSBD state allows it to
-	 * be flipped.
-	 */
-	if (arm64_get_ssbd_state() == ARM64_SSBD_KERNEL)
-		*updptr = cpu_to_le32(aarch64_insn_gen_nop());
-}
-
-void arm64_set_ssbd_mitigation(bool state)
-{
-	int conduit;
-
-	if (!IS_ENABLED(CONFIG_ARM64_SSBD)) {
-		pr_info_once("SSBD disabled by kernel configuration\n");
-		return;
-	}
-
-	if (this_cpu_has_cap(ARM64_SSBS)) {
-		if (state)
-			asm volatile(SET_PSTATE_SSBS(0));
-		else
-			asm volatile(SET_PSTATE_SSBS(1));
-		return;
-	}
-
-	conduit = arm_smccc_1_1_invoke(ARM_SMCCC_ARCH_WORKAROUND_2, state,
-				       NULL);
-
-	WARN_ON_ONCE(conduit == SMCCC_CONDUIT_NONE);
-}
-
-static bool has_ssbd_mitigation(const struct arm64_cpu_capabilities *entry,
-				    int scope)
-{
-	struct arm_smccc_res res;
-	bool required = true;
-	s32 val;
-	bool this_cpu_safe = false;
-	int conduit;
-
-	WARN_ON(scope != SCOPE_LOCAL_CPU || preemptible());
-
-	if (cpu_mitigations_off())
-		ssbd_state = ARM64_SSBD_FORCE_DISABLE;
-
-	/* delay setting __ssb_safe until we get a firmware response */
-	if (is_midr_in_range_list(read_cpuid_id(), entry->midr_range_list))
-		this_cpu_safe = true;
-
-	if (this_cpu_has_cap(ARM64_SSBS)) {
-		if (!this_cpu_safe)
-			__ssb_safe = false;
-		required = false;
-		goto out_printmsg;
-	}
-
-	conduit = arm_smccc_1_1_invoke(ARM_SMCCC_ARCH_FEATURES_FUNC_ID,
-				       ARM_SMCCC_ARCH_WORKAROUND_2, &res);
-
-	if (conduit == SMCCC_CONDUIT_NONE) {
-		ssbd_state = ARM64_SSBD_UNKNOWN;
-		if (!this_cpu_safe)
-			__ssb_safe = false;
-		return false;
-	}
-
-	val = (s32)res.a0;
-
-	switch (val) {
-	case SMCCC_RET_NOT_SUPPORTED:
-		ssbd_state = ARM64_SSBD_UNKNOWN;
-		if (!this_cpu_safe)
-			__ssb_safe = false;
-		return false;
-
-	/* machines with mixed mitigation requirements must not return this */
-	case SMCCC_RET_NOT_REQUIRED:
-		pr_info_once("%s mitigation not required\n", entry->desc);
-		ssbd_state = ARM64_SSBD_MITIGATED;
-		return false;
-
-	case SMCCC_RET_SUCCESS:
-		__ssb_safe = false;
-		required = true;
-		break;
-
-	case 1:	/* Mitigation not required on this CPU */
-		required = false;
-		break;
-
-	default:
-		WARN_ON(1);
-		if (!this_cpu_safe)
-			__ssb_safe = false;
-		return false;
-	}
-
-	switch (ssbd_state) {
-	case ARM64_SSBD_FORCE_DISABLE:
-		arm64_set_ssbd_mitigation(false);
-		required = false;
-		break;
-
-	case ARM64_SSBD_KERNEL:
-		if (required) {
-			__this_cpu_write(arm64_ssbd_callback_required, 1);
-			arm64_set_ssbd_mitigation(true);
-		}
-		break;
-
-	case ARM64_SSBD_FORCE_ENABLE:
-		arm64_set_ssbd_mitigation(true);
-		required = true;
-		break;
-
-	default:
-		WARN_ON(1);
-		break;
-	}
-
-out_printmsg:
-	switch (ssbd_state) {
-	case ARM64_SSBD_FORCE_DISABLE:
-		pr_info_once("%s disabled from command-line\n", entry->desc);
-		break;
-
-	case ARM64_SSBD_FORCE_ENABLE:
-		pr_info_once("%s forced from command-line\n", entry->desc);
-		break;
-	}
-
-	return required;
-}
-
-/* known invulnerable cores */
-static const struct midr_range arm64_ssb_cpus[] = {
-	MIDR_ALL_VERSIONS(MIDR_CORTEX_A35),
-	MIDR_ALL_VERSIONS(MIDR_CORTEX_A53),
-	MIDR_ALL_VERSIONS(MIDR_CORTEX_A55),
-	MIDR_ALL_VERSIONS(MIDR_BRAHMA_B53),
-	MIDR_ALL_VERSIONS(MIDR_QCOM_KRYO_3XX_SILVER),
-	MIDR_ALL_VERSIONS(MIDR_QCOM_KRYO_4XX_SILVER),
-	{},
-};
-
 #ifdef CONFIG_ARM64_ERRATUM_1463225
 DEFINE_PER_CPU(int, __in_cortex_a76_erratum_1463225_wa);
 
@@ -519,83 +160,6 @@ cpu_enable_cache_maint_trap(const struct arm64_cpu_capabilities *__unused)
 	.type = ARM64_CPUCAP_LOCAL_CPU_ERRATUM,			\
 	CAP_MIDR_RANGE_LIST(midr_list)
 
-/* Track overall mitigation state. We are only mitigated if all cores are ok */
-static bool __hardenbp_enab = true;
-static bool __spectrev2_safe = true;
-
-int get_spectre_v2_workaround_state(void)
-{
-	if (__spectrev2_safe)
-		return ARM64_BP_HARDEN_NOT_REQUIRED;
-
-	if (!__hardenbp_enab)
-		return ARM64_BP_HARDEN_UNKNOWN;
-
-	return ARM64_BP_HARDEN_WA_NEEDED;
-}
-
-/*
- * List of CPUs that do not need any Spectre-v2 mitigation at all.
- */
-static const struct midr_range spectre_v2_safe_list[] = {
-	MIDR_ALL_VERSIONS(MIDR_CORTEX_A35),
-	MIDR_ALL_VERSIONS(MIDR_CORTEX_A53),
-	MIDR_ALL_VERSIONS(MIDR_CORTEX_A55),
-	MIDR_ALL_VERSIONS(MIDR_BRAHMA_B53),
-	MIDR_ALL_VERSIONS(MIDR_HISI_TSV110),
-	MIDR_ALL_VERSIONS(MIDR_QCOM_KRYO_3XX_SILVER),
-	MIDR_ALL_VERSIONS(MIDR_QCOM_KRYO_4XX_SILVER),
-	{ /* sentinel */ }
-};
-
-/*
- * Track overall bp hardening for all heterogeneous cores in the machine.
- * We are only considered "safe" if all booted cores are known safe.
- */
-static bool __maybe_unused
-check_branch_predictor(const struct arm64_cpu_capabilities *entry, int scope)
-{
-	int need_wa;
-
-	WARN_ON(scope != SCOPE_LOCAL_CPU || preemptible());
-
-	/* If the CPU has CSV2 set, we're safe */
-	if (cpuid_feature_extract_unsigned_field(read_cpuid(ID_AA64PFR0_EL1),
-						 ID_AA64PFR0_CSV2_SHIFT))
-		return false;
-
-	/* Alternatively, we have a list of unaffected CPUs */
-	if (is_midr_in_range_list(read_cpuid_id(), spectre_v2_safe_list))
-		return false;
-
-	/* Fallback to firmware detection */
-	need_wa = detect_harden_bp_fw();
-	if (!need_wa)
-		return false;
-
-	__spectrev2_safe = false;
-
-	if (!IS_ENABLED(CONFIG_HARDEN_BRANCH_PREDICTOR)) {
-		pr_warn_once("spectrev2 mitigation disabled by kernel configuration\n");
-		__hardenbp_enab = false;
-		return false;
-	}
-
-	/* forced off */
-	if (__nospectre_v2 || cpu_mitigations_off()) {
-		pr_info_once("spectrev2 mitigation disabled by command line option\n");
-		__hardenbp_enab = false;
-		return false;
-	}
-
-	if (need_wa < 0) {
-		pr_warn_once("ARM_SMCCC_ARCH_WORKAROUND_1 missing from firmware\n");
-		__hardenbp_enab = false;
-	}
-
-	return (need_wa > 0);
-}
-
 static const __maybe_unused struct midr_range tx2_family_cpus[] = {
 	MIDR_ALL_VERSIONS(MIDR_BRCM_VULCAN),
 	MIDR_ALL_VERSIONS(MIDR_CAVIUM_THUNDERX2),
@@ -887,9 +451,11 @@ const struct arm64_cpu_capabilities arm64_errata[] = {
 	},
 #endif
 	{
-		.capability = ARM64_HARDEN_BRANCH_PREDICTOR,
+		.desc = "Spectre-v2",
+		.capability = ARM64_SPECTRE_V2,
 		.type = ARM64_CPUCAP_LOCAL_CPU_ERRATUM,
-		.matches = check_branch_predictor,
+		.matches = has_spectre_v2,
+		.cpu_enable = spectre_v2_enable_mitigation,
 	},
 #ifdef CONFIG_RANDOMIZE_BASE
 	{
@@ -899,11 +465,11 @@ const struct arm64_cpu_capabilities arm64_errata[] = {
 	},
 #endif
 	{
-		.desc = "Speculative Store Bypass Disable",
-		.capability = ARM64_SSBD,
+		.desc = "Spectre-v4",
+		.capability = ARM64_SPECTRE_V4,
 		.type = ARM64_CPUCAP_LOCAL_CPU_ERRATUM,
-		.matches = has_ssbd_mitigation,
-		.midr_range_list = arm64_ssb_cpus,
+		.matches = has_spectre_v4,
+		.cpu_enable = spectre_v4_enable_mitigation,
 	},
 #ifdef CONFIG_ARM64_ERRATUM_1418040
 	{
@@ -956,40 +522,3 @@ const struct arm64_cpu_capabilities arm64_errata[] = {
 	{
 	}
 };
-
-ssize_t cpu_show_spectre_v1(struct device *dev, struct device_attribute *attr,
-			    char *buf)
-{
-	return sprintf(buf, "Mitigation: __user pointer sanitization\n");
-}
-
-ssize_t cpu_show_spectre_v2(struct device *dev, struct device_attribute *attr,
-		char *buf)
-{
-	switch (get_spectre_v2_workaround_state()) {
-	case ARM64_BP_HARDEN_NOT_REQUIRED:
-		return sprintf(buf, "Not affected\n");
-        case ARM64_BP_HARDEN_WA_NEEDED:
-		return sprintf(buf, "Mitigation: Branch predictor hardening\n");
-        case ARM64_BP_HARDEN_UNKNOWN:
-	default:
-		return sprintf(buf, "Vulnerable\n");
-	}
-}
-
-ssize_t cpu_show_spec_store_bypass(struct device *dev,
-		struct device_attribute *attr, char *buf)
-{
-	if (__ssb_safe)
-		return sprintf(buf, "Not affected\n");
-
-	switch (ssbd_state) {
-	case ARM64_SSBD_KERNEL:
-	case ARM64_SSBD_FORCE_ENABLE:
-		if (IS_ENABLED(CONFIG_ARM64_SSBD))
-			return sprintf(buf,
-			    "Mitigation: Speculative Store Bypass disabled via prctl\n");
-	}
-
-	return sprintf(buf, "Vulnerable\n");
-}

arch/arm64/kernel/cpufeature.c

@@ -227,7 +227,7 @@ static const struct arm64_ftr_bits ftr_id_aa64pfr0[] = {
 static const struct arm64_ftr_bits ftr_id_aa64pfr1[] = {
 	ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_LOWER_SAFE, ID_AA64PFR1_MPAMFRAC_SHIFT, 4, 0),
 	ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_LOWER_SAFE, ID_AA64PFR1_RASFRAC_SHIFT, 4, 0),
-	ARM64_FTR_BITS(FTR_VISIBLE, FTR_STRICT, FTR_LOWER_SAFE, ID_AA64PFR1_SSBS_SHIFT, 4, ID_AA64PFR1_SSBS_PSTATE_NI),
+	ARM64_FTR_BITS(FTR_VISIBLE, FTR_NONSTRICT, FTR_LOWER_SAFE, ID_AA64PFR1_SSBS_SHIFT, 4, ID_AA64PFR1_SSBS_PSTATE_NI),
 	ARM64_FTR_BITS(FTR_VISIBLE_IF_IS_ENABLED(CONFIG_ARM64_BTI),
 				    FTR_STRICT, FTR_LOWER_SAFE, ID_AA64PFR1_BT_SHIFT, 4, 0),
 	ARM64_FTR_END,
@@ -487,7 +487,7 @@ static const struct arm64_ftr_bits ftr_id_pfr1[] = {
 };
 
 static const struct arm64_ftr_bits ftr_id_pfr2[] = {
-	ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_LOWER_SAFE, ID_PFR2_SSBS_SHIFT, 4, 0),
+	ARM64_FTR_BITS(FTR_HIDDEN, FTR_NONSTRICT, FTR_LOWER_SAFE, ID_PFR2_SSBS_SHIFT, 4, 0),
 	ARM64_FTR_BITS(FTR_HIDDEN, FTR_NONSTRICT, FTR_LOWER_SAFE, ID_PFR2_CSV3_SHIFT, 4, 0),
 	ARM64_FTR_END,
 };
@@ -1583,48 +1583,6 @@ static void cpu_has_fwb(const struct arm64_cpu_capabilities *__unused)
 	WARN_ON(val & (7 << 27 | 7 << 21));
 }
 
-#ifdef CONFIG_ARM64_SSBD
-static int ssbs_emulation_handler(struct pt_regs *regs, u32 instr)
-{
-	if (user_mode(regs))
-		return 1;
-
-	if (instr & BIT(PSTATE_Imm_shift))
-		regs->pstate |= PSR_SSBS_BIT;
-	else
-		regs->pstate &= ~PSR_SSBS_BIT;
-
-	arm64_skip_faulting_instruction(regs, 4);
-	return 0;
-}
-
-static struct undef_hook ssbs_emulation_hook = {
-	.instr_mask	= ~(1U << PSTATE_Imm_shift),
-	.instr_val	= 0xd500401f | PSTATE_SSBS,
-	.fn		= ssbs_emulation_handler,
-};
-
-static void cpu_enable_ssbs(const struct arm64_cpu_capabilities *__unused)
-{
-	static bool undef_hook_registered = false;
-	static DEFINE_RAW_SPINLOCK(hook_lock);
-
-	raw_spin_lock(&hook_lock);
-	if (!undef_hook_registered) {
-		register_undef_hook(&ssbs_emulation_hook);
-		undef_hook_registered = true;
-	}
-	raw_spin_unlock(&hook_lock);
-
-	if (arm64_get_ssbd_state() == ARM64_SSBD_FORCE_DISABLE) {
-		sysreg_clear_set(sctlr_el1, 0, SCTLR_ELx_DSSBS);
-		arm64_set_ssbd_mitigation(false);
-	} else {
-		arm64_set_ssbd_mitigation(true);
-	}
-}
-#endif /* CONFIG_ARM64_SSBD */
-
 #ifdef CONFIG_ARM64_PAN
 static void cpu_enable_pan(const struct arm64_cpu_capabilities *__unused)
 {
@@ -1976,19 +1934,16 @@ static const struct arm64_cpu_capabilities arm64_features[] = {
 		.field_pos = ID_AA64ISAR0_CRC32_SHIFT,
 		.min_field_value = 1,
 	},
-#ifdef CONFIG_ARM64_SSBD
 	{
 		.desc = "Speculative Store Bypassing Safe (SSBS)",
 		.capability = ARM64_SSBS,
-		.type = ARM64_CPUCAP_WEAK_LOCAL_CPU_FEATURE,
+		.type = ARM64_CPUCAP_SYSTEM_FEATURE,
 		.matches = has_cpuid_feature,
 		.sys_reg = SYS_ID_AA64PFR1_EL1,
 		.field_pos = ID_AA64PFR1_SSBS_SHIFT,
 		.sign = FTR_UNSIGNED,
 		.min_field_value = ID_AA64PFR1_SSBS_PSTATE_ONLY,
-		.cpu_enable = cpu_enable_ssbs,
 	},
-#endif
 #ifdef CONFIG_ARM64_CNP
 	{
 		.desc = "Common not Private translations",

arch/arm64/kernel/entry.S

@@ -132,9 +132,8 @@ alternative_else_nop_endif
 	 * them if required.
 	 */
 	.macro	apply_ssbd, state, tmp1, tmp2
-#ifdef CONFIG_ARM64_SSBD
-alternative_cb	arm64_enable_wa2_handling
-	b	.L__asm_ssbd_skip\@
+alternative_cb	spectre_v4_patch_fw_mitigation_enable
+	b	.L__asm_ssbd_skip\@		// Patched to NOP
 alternative_cb_end
 	ldr_this_cpu	\tmp2, arm64_ssbd_callback_required, \tmp1
 	cbz	\tmp2,	.L__asm_ssbd_skip\@
@@ -142,11 +141,10 @@ alternative_cb_end
 	tbnz	\tmp2, #TIF_SSBD, .L__asm_ssbd_skip\@
 	mov	w0, #ARM_SMCCC_ARCH_WORKAROUND_2
 	mov	w1, #\state
-alternative_cb	arm64_update_smccc_conduit
+alternative_cb	spectre_v4_patch_fw_mitigation_conduit
 	nop					// Patched to SMC/HVC #0
 alternative_cb_end
 .L__asm_ssbd_skip\@:
-#endif
 	.endm
 
 	.macro	kernel_entry, el, regsize = 64
@@ -697,11 +695,9 @@ el0_irq_naked:
 	bl	trace_hardirqs_off
 #endif
 
-#ifdef CONFIG_HARDEN_BRANCH_PREDICTOR
 	tbz	x22, #55, 1f
 	bl	do_el0_irq_bp_hardening
 1:
-#endif
 	irq_handler
 
 #ifdef CONFIG_TRACE_IRQFLAGS

arch/arm64/kernel/hibernate.c

@@ -332,11 +332,7 @@ int swsusp_arch_suspend(void)
 		 * mitigation off behind our back, let's set the state
 		 * to what we expect it to be.
 		 */
-		switch (arm64_get_ssbd_state()) {
-		case ARM64_SSBD_FORCE_ENABLE:
-		case ARM64_SSBD_KERNEL:
-			arm64_set_ssbd_mitigation(true);
-		}
+		spectre_v4_enable_mitigation(NULL);
 	}
 
 	local_daif_restore(flags);

arch/arm64/kernel/image-vars.h

@@ -61,18 +61,11 @@ __efistub__ctype		= _ctype;
  * memory mappings.
  */
 
-#define KVM_NVHE_ALIAS(sym) __kvm_nvhe_##sym = sym;
-
 /* Alternative callbacks for init-time patching of nVHE hyp code. */
-KVM_NVHE_ALIAS(arm64_enable_wa2_handling);
 KVM_NVHE_ALIAS(kvm_patch_vector_branch);
 KVM_NVHE_ALIAS(kvm_update_va_mask);
 
 /* Global kernel state accessed by nVHE hyp code. */
-KVM_NVHE_ALIAS(arm64_ssbd_callback_required);
-KVM_NVHE_ALIAS(kvm_host_data);
-KVM_NVHE_ALIAS(kvm_hyp_ctxt);
-KVM_NVHE_ALIAS(kvm_hyp_vector);
 KVM_NVHE_ALIAS(kvm_vgic_global_state);
 
 /* Kernel constant needed to compute idmap addresses. */

arch/arm64/kernel/process.c

@@ -21,6 +21,7 @@
 #include <linux/lockdep.h>
 #include <linux/mman.h>
 #include <linux/mm.h>
+#include <linux/nospec.h>
 #include <linux/stddef.h>
 #include <linux/sysctl.h>
 #include <linux/unistd.h>
@@ -421,8 +422,7 @@ int copy_thread(unsigned long clone_flags, unsigned long stack_start,
 		    cpus_have_const_cap(ARM64_HAS_UAO))
 			childregs->pstate |= PSR_UAO_BIT;
 
-		if (arm64_get_ssbd_state() == ARM64_SSBD_FORCE_DISABLE)
-			set_ssbs_bit(childregs);
+		spectre_v4_enable_task_mitigation(p);
 
 		if (system_uses_irq_prio_masking())
 			childregs->pmr_save = GIC_PRIO_IRQON;
@@ -472,8 +472,6 @@ void uao_thread_switch(struct task_struct *next)
  */
 static void ssbs_thread_switch(struct task_struct *next)
 {
-	struct pt_regs *regs = task_pt_regs(next);
-
 	/*
 	 * Nothing to do for kernel threads, but 'regs' may be junk
 	 * (e.g. idle task) so check the flags and bail early.
@@ -485,18 +483,10 @@ static void ssbs_thread_switch(struct task_struct *next)
 	 * If all CPUs implement the SSBS extension, then we just need to
 	 * context-switch the PSTATE field.
 	 */
-	if (cpu_have_feature(cpu_feature(SSBS)))
-		return;
-
-	/* If the mitigation is enabled, then we leave SSBS clear. */
-	if ((arm64_get_ssbd_state() == ARM64_SSBD_FORCE_ENABLE) ||
-	    test_tsk_thread_flag(next, TIF_SSBD))
+	if (cpus_have_const_cap(ARM64_SSBS))
 		return;
 
-	if (compat_user_mode(regs))
-		set_compat_ssbs_bit(regs);
-	else if (user_mode(regs))
-		set_ssbs_bit(regs);
+	spectre_v4_enable_task_mitigation(next);
 }
 
 /*
@@ -620,6 +610,11 @@ void arch_setup_new_exec(void)
 	current->mm->context.flags = is_compat_task() ? MMCF_AARCH32 : 0;
 
 	ptrauth_thread_init_user(current);
+
+	if (task_spec_ssb_noexec(current)) {
+		arch_prctl_spec_ctrl_set(current, PR_SPEC_STORE_BYPASS,
+					 PR_SPEC_ENABLE);
+	}
 }
 
 #ifdef CONFIG_ARM64_TAGGED_ADDR_ABI

arch/arm64/kernel/proton-pack.c

+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Handle detection, reporting and mitigation of Spectre v1, v2 and v4, as
+ * detailed at:
+ *
+ *   https://developer.arm.com/support/arm-security-updates/speculative-processor-vulnerability
+ *
+ * This code was originally written hastily under an awful lot of stress and so
+ * aspects of it are somewhat hacky. Unfortunately, changing anything in here
+ * instantly makes me feel ill. Thanks, Jann. Thann.
+ *
+ * Copyright (C) 2018 ARM Ltd, All Rights Reserved.
+ * Copyright (C) 2020 Google LLC
+ *
+ * "If there's something strange in your neighbourhood, who you gonna call?"
+ *
+ * Authors: Will Deacon <will@kernel.org> and Marc Zyngier <maz@kernel.org>
+ */
+
+#include <linux/arm-smccc.h>
+#include <linux/cpu.h>
+#include <linux/device.h>
+#include <linux/nospec.h>
+#include <linux/prctl.h>
+#include <linux/sched/task_stack.h>
+
+#include <asm/spectre.h>
+#include <asm/traps.h>
+
+/*
+ * We try to ensure that the mitigation state can never change as the result of
+ * onlining a late CPU.
+ */
+static void update_mitigation_state(enum mitigation_state *oldp,
+				    enum mitigation_state new)
+{
+	enum mitigation_state state;
+
+	do {
+		state = READ_ONCE(*oldp);
+		if (new <= state)
+			break;
+
+		/* Userspace almost certainly can't deal with this. */
+		if (WARN_ON(system_capabilities_finalized()))
+			break;
+	} while (cmpxchg_relaxed(oldp, state, new) != state);
+}
+
+/*
+ * Spectre v1.
+ *
+ * The kernel can't protect userspace for this one: it's each person for
+ * themselves. Advertise what we're doing and be done with it.
+ */
+ssize_t cpu_show_spectre_v1(struct device *dev, struct device_attribute *attr,
+			    char *buf)
+{
+	return sprintf(buf, "Mitigation: __user pointer sanitization\n");
+}
+
+/*
+ * Spectre v2.
+ *
+ * This one sucks. A CPU is either:
+ *
+ * - Mitigated in hardware and advertised by ID_AA64PFR0_EL1.CSV2.
+ * - Mitigated in hardware and listed in our "safe list".
+ * - Mitigated in software by firmware.
+ * - Mitigated in software by a CPU-specific dance in the kernel.
+ * - Vulnerable.
+ *
+ * It's not unlikely for different CPUs in a big.LITTLE system to fall into
+ * different camps.
+ */
+static enum mitigation_state spectre_v2_state;
+
+static bool __read_mostly __nospectre_v2;
+static int __init parse_spectre_v2_param(char *str)
+{
+	__nospectre_v2 = true;
+	return 0;
+}
+early_param("nospectre_v2", parse_spectre_v2_param);
+
+static bool spectre_v2_mitigations_off(void)
+{
+	bool ret = __nospectre_v2 || cpu_mitigations_off();
+
+	if (ret)
+		pr_info_once("spectre-v2 mitigation disabled by command line option\n");
+
+	return ret;
+}
+
+ssize_t cpu_show_spectre_v2(struct device *dev, struct device_attribute *attr,
+			    char *buf)
+{
+	switch (spectre_v2_state) {
+	case SPECTRE_UNAFFECTED:
+		return sprintf(buf, "Not affected\n");
+	case SPECTRE_MITIGATED:
+		return sprintf(buf, "Mitigation: Branch predictor hardening\n");
+	case SPECTRE_VULNERABLE:
+		fallthrough;
+	default:
+		return sprintf(buf, "Vulnerable\n");
+	}
+}
+
+static enum mitigation_state spectre_v2_get_cpu_hw_mitigation_state(void)
+{
+	u64 pfr0;
+	static const struct midr_range spectre_v2_safe_list[] = {
+		MIDR_ALL_VERSIONS(MIDR_CORTEX_A35),
+		MIDR_ALL_VERSIONS(MIDR_CORTEX_A53),
+		MIDR_ALL_VERSIONS(MIDR_CORTEX_A55),
+		MIDR_ALL_VERSIONS(MIDR_BRAHMA_B53),
+		MIDR_ALL_VERSIONS(MIDR_HISI_TSV110),
+		MIDR_ALL_VERSIONS(MIDR_QCOM_KRYO_3XX_SILVER),
+		MIDR_ALL_VERSIONS(MIDR_QCOM_KRYO_4XX_SILVER),
+		{ /* sentinel */ }
+	};
+
+	/* If the CPU has CSV2 set, we're safe */
+	pfr0 = read_cpuid(ID_AA64PFR0_EL1);
+	if (cpuid_feature_extract_unsigned_field(pfr0, ID_AA64PFR0_CSV2_SHIFT))
+		return SPECTRE_UNAFFECTED;
+
+	/* Alternatively, we have a list of unaffected CPUs */
+	if (is_midr_in_range_list(read_cpuid_id(), spectre_v2_safe_list))
+		return SPECTRE_UNAFFECTED;
+
+	return SPECTRE_VULNERABLE;
+}
+
+#define SMCCC_ARCH_WORKAROUND_RET_UNAFFECTED	(1)
+
+static enum mitigation_state spectre_v2_get_cpu_fw_mitigation_state(void)
+{
+	int ret;
+	struct arm_smccc_res res;
+
+	arm_smccc_1_1_invoke(ARM_SMCCC_ARCH_FEATURES_FUNC_ID,
+			     ARM_SMCCC_ARCH_WORKAROUND_1, &res);
+
+	ret = res.a0;
+	switch (ret) {
+	case SMCCC_RET_SUCCESS:
+		return SPECTRE_MITIGATED;
+	case SMCCC_ARCH_WORKAROUND_RET_UNAFFECTED:
+		return SPECTRE_UNAFFECTED;
+	default:
+		fallthrough;
+	case SMCCC_RET_NOT_SUPPORTED:
+		return SPECTRE_VULNERABLE;
+	}
+}
+
+bool has_spectre_v2(const struct arm64_cpu_capabilities *entry, int scope)
+{
+	WARN_ON(scope != SCOPE_LOCAL_CPU || preemptible());
+
+	if (spectre_v2_get_cpu_hw_mitigation_state() == SPECTRE_UNAFFECTED)
+		return false;
+
+	if (spectre_v2_get_cpu_fw_mitigation_state() == SPECTRE_UNAFFECTED)
+		return false;
+
+	return true;
+}
+
+DEFINE_PER_CPU_READ_MOSTLY(struct bp_hardening_data, bp_hardening_data);
+
+enum mitigation_state arm64_get_spectre_v2_state(void)
+{
+	return spectre_v2_state;
+}
+
+#ifdef CONFIG_KVM
+#include <asm/cacheflush.h>
+#include <asm/kvm_asm.h>
+
+atomic_t arm64_el2_vector_last_slot = ATOMIC_INIT(-1);
+
+static void __copy_hyp_vect_bpi(int slot, const char *hyp_vecs_start,
+				const char *hyp_vecs_end)
+{
+	void *dst = lm_alias(__bp_harden_hyp_vecs + slot * SZ_2K);
+	int i;
+
+	for (i = 0; i < SZ_2K; i += 0x80)
+		memcpy(dst + i, hyp_vecs_start, hyp_vecs_end - hyp_vecs_start);
+
+	__flush_icache_range((uintptr_t)dst, (uintptr_t)dst + SZ_2K);
+}
+
+static void install_bp_hardening_cb(bp_hardening_cb_t fn)
+{
+	static DEFINE_RAW_SPINLOCK(bp_lock);
+	int cpu, slot = -1;
+	const char *hyp_vecs_start = __smccc_workaround_1_smc;
+	const char *hyp_vecs_end = __smccc_workaround_1_smc +
+				   __SMCCC_WORKAROUND_1_SMC_SZ;
+
+	/*
+	 * detect_harden_bp_fw() passes NULL for the hyp_vecs start/end if
+	 * we're a guest. Skip the hyp-vectors work.
+	 */
+	if (!is_hyp_mode_available()) {
+		__this_cpu_write(bp_hardening_data.fn, fn);
+		return;
+	}
+
+	raw_spin_lock(&bp_lock);
+	for_each_possible_cpu(cpu) {
+		if (per_cpu(bp_hardening_data.fn, cpu) == fn) {
+			slot = per_cpu(bp_hardening_data.hyp_vectors_slot, cpu);
+			break;
+		}
+	}
+
+	if (slot == -1) {
+		slot = atomic_inc_return(&arm64_el2_vector_last_slot);
+		BUG_ON(slot >= BP_HARDEN_EL2_SLOTS);
+		__copy_hyp_vect_bpi(slot, hyp_vecs_start, hyp_vecs_end);
+	}
+
+	__this_cpu_write(bp_hardening_data.hyp_vectors_slot, slot);
+	__this_cpu_write(bp_hardening_data.fn, fn);
+	raw_spin_unlock(&bp_lock);
+}
+#else
+static void install_bp_hardening_cb(bp_hardening_cb_t fn)
+{
+	__this_cpu_write(bp_hardening_data.fn, fn);
+}
+#endif	/* CONFIG_KVM */
+
+static void call_smc_arch_workaround_1(void)
+{
+	arm_smccc_1_1_smc(ARM_SMCCC_ARCH_WORKAROUND_1, NULL);
+}
+
+static void call_hvc_arch_workaround_1(void)
+{
+	arm_smccc_1_1_hvc(ARM_SMCCC_ARCH_WORKAROUND_1, NULL);
+}
+
+static void qcom_link_stack_sanitisation(void)
+{
+	u64 tmp;
+
+	asm volatile("mov	%0, x30		\n"
+		     ".rept	16		\n"
+		     "bl	. + 4		\n"
+		     ".endr			\n"
+		     "mov	x30, %0		\n"
+		     : "=&r" (tmp));
+}
+
+static enum mitigation_state spectre_v2_enable_fw_mitigation(void)
+{
+	bp_hardening_cb_t cb;
+	enum mitigation_state state;
+
+	state = spectre_v2_get_cpu_fw_mitigation_state();
+	if (state != SPECTRE_MITIGATED)
+		return state;
+
+	if (spectre_v2_mitigations_off())
+		return SPECTRE_VULNERABLE;
+
+	switch (arm_smccc_1_1_get_conduit()) {
+	case SMCCC_CONDUIT_HVC:
+		cb = call_hvc_arch_workaround_1;
+		break;
+
+	case SMCCC_CONDUIT_SMC:
+		cb = call_smc_arch_workaround_1;
+		break;
+
+	default:
+		return SPECTRE_VULNERABLE;
+	}
+
+	install_bp_hardening_cb(cb);
+	return SPECTRE_MITIGATED;
+}
+
+static enum mitigation_state spectre_v2_enable_sw_mitigation(void)
+{
+	u32 midr;
+
+	if (spectre_v2_mitigations_off())
+		return SPECTRE_VULNERABLE;
+
+	midr = read_cpuid_id();
+	if (((midr & MIDR_CPU_MODEL_MASK) != MIDR_QCOM_FALKOR) &&
+	    ((midr & MIDR_CPU_MODEL_MASK) != MIDR_QCOM_FALKOR_V1))
+		return SPECTRE_VULNERABLE;
+
+	install_bp_hardening_cb(qcom_link_stack_sanitisation);
+	return SPECTRE_MITIGATED;
+}
+
+void spectre_v2_enable_mitigation(const struct arm64_cpu_capabilities *__unused)
+{
+	enum mitigation_state state;
+
+	WARN_ON(preemptible());
+
+	state = spectre_v2_get_cpu_hw_mitigation_state();
+	if (state == SPECTRE_VULNERABLE)
+		state = spectre_v2_enable_fw_mitigation();
+	if (state == SPECTRE_VULNERABLE)
+		state = spectre_v2_enable_sw_mitigation();
+
+	update_mitigation_state(&spectre_v2_state, state);
+}
+
+/*
+ * Spectre v4.
+ *
+ * If you thought Spectre v2 was nasty, wait until you see this mess. A CPU is
+ * either:
+ *
+ * - Mitigated in hardware and listed in our "safe list".
+ * - Mitigated in hardware via PSTATE.SSBS.
+ * - Mitigated in software by firmware (sometimes referred to as SSBD).
+ *
+ * Wait, that doesn't sound so bad, does it? Keep reading...
+ *
+ * A major source of headaches is that the software mitigation is enabled both
+ * on a per-task basis, but can also be forced on for the kernel, necessitating
+ * both context-switch *and* entry/exit hooks. To make it even worse, some CPUs
+ * allow EL0 to toggle SSBS directly, which can end up with the prctl() state
+ * being stale when re-entering the kernel. The usual big.LITTLE caveats apply,
+ * so you can have systems that have both firmware and SSBS mitigations. This
+ * means we actually have to reject late onlining of CPUs with mitigations if
+ * all of the currently onlined CPUs are safelisted, as the mitigation tends to
+ * be opt-in for userspace. Yes, really, the cure is worse than the disease.
+ *
+ * The only good part is that if the firmware mitigation is present, then it is
+ * present for all CPUs, meaning we don't have to worry about late onlining of a
+ * vulnerable CPU if one of the boot CPUs is using the firmware mitigation.
+ *
+ * Give me a VAX-11/780 any day of the week...
+ */
+static enum mitigation_state spectre_v4_state;
+
+/* This is the per-cpu state tracking whether we need to talk to firmware */
+DEFINE_PER_CPU_READ_MOSTLY(u64, arm64_ssbd_callback_required);
+
+enum spectre_v4_policy {
+	SPECTRE_V4_POLICY_MITIGATION_DYNAMIC,
+	SPECTRE_V4_POLICY_MITIGATION_ENABLED,
+	SPECTRE_V4_POLICY_MITIGATION_DISABLED,
+};
+
+static enum spectre_v4_policy __read_mostly __spectre_v4_policy;
+
+static const struct spectre_v4_param {
+	const char		*str;
+	enum spectre_v4_policy	policy;
+} spectre_v4_params[] = {
+	{ "force-on",	SPECTRE_V4_POLICY_MITIGATION_ENABLED, },
+	{ "force-off",	SPECTRE_V4_POLICY_MITIGATION_DISABLED, },
+	{ "kernel",	SPECTRE_V4_POLICY_MITIGATION_DYNAMIC, },
+};
+static int __init parse_spectre_v4_param(char *str)
+{
+	int i;
+
+	if (!str || !str[0])
+		return -EINVAL;
+
+	for (i = 0; i < ARRAY_SIZE(spectre_v4_params); i++) {
+		const struct spectre_v4_param *param = &spectre_v4_params[i];
+
+		if (strncmp(str, param->str, strlen(param->str)))
+			continue;
+
+		__spectre_v4_policy = param->policy;
+		return 0;
+	}
+
+	return -EINVAL;
+}
+early_param("ssbd", parse_spectre_v4_param);
+
+/*
+ * Because this was all written in a rush by people working in different silos,
+ * we've ended up with multiple command line options to control the same thing.
+ * Wrap these up in some helpers, which prefer disabling the mitigation if faced
+ * with contradictory parameters. The mitigation is always either "off",
+ * "dynamic" or "on".
+ */
+static bool spectre_v4_mitigations_off(void)
+{
+	bool ret = cpu_mitigations_off() ||
+		   __spectre_v4_policy == SPECTRE_V4_POLICY_MITIGATION_DISABLED;
+
+	if (ret)
+		pr_info_once("spectre-v4 mitigation disabled by command-line option\n");
+
+	return ret;
+}
+
+/* Do we need to toggle the mitigation state on entry to/exit from the kernel? */
+static bool spectre_v4_mitigations_dynamic(void)
+{
+	return !spectre_v4_mitigations_off() &&
+	       __spectre_v4_policy == SPECTRE_V4_POLICY_MITIGATION_DYNAMIC;
+}
+
+static bool spectre_v4_mitigations_on(void)
+{
+	return !spectre_v4_mitigations_off() &&
+	       __spectre_v4_policy == SPECTRE_V4_POLICY_MITIGATION_ENABLED;
+}
+
+ssize_t cpu_show_spec_store_bypass(struct device *dev,
+				   struct device_attribute *attr, char *buf)
+{
+	switch (spectre_v4_state) {
+	case SPECTRE_UNAFFECTED:
+		return sprintf(buf, "Not affected\n");
+	case SPECTRE_MITIGATED:
+		return sprintf(buf, "Mitigation: Speculative Store Bypass disabled via prctl\n");
+	case SPECTRE_VULNERABLE:
+		fallthrough;
+	default:
+		return sprintf(buf, "Vulnerable\n");
+	}
+}
+
+enum mitigation_state arm64_get_spectre_v4_state(void)
+{
+	return spectre_v4_state;
+}
+
+static enum mitigation_state spectre_v4_get_cpu_hw_mitigation_state(void)
+{
+	static const struct midr_range spectre_v4_safe_list[] = {
+		MIDR_ALL_VERSIONS(MIDR_CORTEX_A35),
+		MIDR_ALL_VERSIONS(MIDR_CORTEX_A53),
+		MIDR_ALL_VERSIONS(MIDR_CORTEX_A55),
+		MIDR_ALL_VERSIONS(MIDR_BRAHMA_B53),
+		MIDR_ALL_VERSIONS(MIDR_QCOM_KRYO_3XX_SILVER),
+		MIDR_ALL_VERSIONS(MIDR_QCOM_KRYO_4XX_SILVER),
+		{ /* sentinel */ },
+	};
+
+	if (is_midr_in_range_list(read_cpuid_id(), spectre_v4_safe_list))
+		return SPECTRE_UNAFFECTED;
+
+	/* CPU features are detected first */
+	if (this_cpu_has_cap(ARM64_SSBS))
+		return SPECTRE_MITIGATED;
+
+	return SPECTRE_VULNERABLE;
+}
+
+static enum mitigation_state spectre_v4_get_cpu_fw_mitigation_state(void)
+{
+	int ret;
+	struct arm_smccc_res res;
+
+	arm_smccc_1_1_invoke(ARM_SMCCC_ARCH_FEATURES_FUNC_ID,
+			     ARM_SMCCC_ARCH_WORKAROUND_2, &res);
+
+	ret = res.a0;
+	switch (ret) {
+	case SMCCC_RET_SUCCESS:
+		return SPECTRE_MITIGATED;
+	case SMCCC_ARCH_WORKAROUND_RET_UNAFFECTED:
+		fallthrough;
+	case SMCCC_RET_NOT_REQUIRED:
+		return SPECTRE_UNAFFECTED;
+	default:
+		fallthrough;
+	case SMCCC_RET_NOT_SUPPORTED:
+		return SPECTRE_VULNERABLE;
+	}
+}
+
+bool has_spectre_v4(const struct arm64_cpu_capabilities *cap, int scope)
+{
+	enum mitigation_state state;
+
+	WARN_ON(scope != SCOPE_LOCAL_CPU || preemptible());
+
+	state = spectre_v4_get_cpu_hw_mitigation_state();
+	if (state == SPECTRE_VULNERABLE)
+		state = spectre_v4_get_cpu_fw_mitigation_state();
+
+	return state != SPECTRE_UNAFFECTED;
+}
+
+static int ssbs_emulation_handler(struct pt_regs *regs, u32 instr)
+{
+	if (user_mode(regs))
+		return 1;
+
+	if (instr & BIT(PSTATE_Imm_shift))
+		regs->pstate |= PSR_SSBS_BIT;
+	else
+		regs->pstate &= ~PSR_SSBS_BIT;
+
+	arm64_skip_faulting_instruction(regs, 4);
+	return 0;
+}
+
+static struct undef_hook ssbs_emulation_hook = {
+	.instr_mask	= ~(1U << PSTATE_Imm_shift),
+	.instr_val	= 0xd500401f | PSTATE_SSBS,
+	.fn		= ssbs_emulation_handler,
+};
+
+static enum mitigation_state spectre_v4_enable_hw_mitigation(void)
+{
+	static bool undef_hook_registered = false;
+	static DEFINE_RAW_SPINLOCK(hook_lock);
+	enum mitigation_state state;
+
+	/*
+	 * If the system is mitigated but this CPU doesn't have SSBS, then
+	 * we must be on the safelist and there's nothing more to do.
+	 */
+	state = spectre_v4_get_cpu_hw_mitigation_state();
+	if (state != SPECTRE_MITIGATED || !this_cpu_has_cap(ARM64_SSBS))
+		return state;
+
+	raw_spin_lock(&hook_lock);
+	if (!undef_hook_registered) {
+		register_undef_hook(&ssbs_emulation_hook);
+		undef_hook_registered = true;
+	}
+	raw_spin_unlock(&hook_lock);
+
+	if (spectre_v4_mitigations_off()) {
+		sysreg_clear_set(sctlr_el1, 0, SCTLR_ELx_DSSBS);
+		asm volatile(SET_PSTATE_SSBS(1));
+		return SPECTRE_VULNERABLE;
+	}
+
+	/* SCTLR_EL1.DSSBS was initialised to 0 during boot */
+	asm volatile(SET_PSTATE_SSBS(0));
+	return SPECTRE_MITIGATED;
+}
+
+/*
+ * Patch a branch over the Spectre-v4 mitigation code with a NOP so that
+ * we fallthrough and check whether firmware needs to be called on this CPU.
+ */
+void __init spectre_v4_patch_fw_mitigation_enable(struct alt_instr *alt,
+						  __le32 *origptr,
+						  __le32 *updptr, int nr_inst)
+{
+	BUG_ON(nr_inst != 1); /* Branch -> NOP */
+
+	if (spectre_v4_mitigations_off())
+		return;
+
+	if (cpus_have_final_cap(ARM64_SSBS))
+		return;
+
+	if (spectre_v4_mitigations_dynamic())
+		*updptr = cpu_to_le32(aarch64_insn_gen_nop());
+}
+
+/*
+ * Patch a NOP in the Spectre-v4 mitigation code with an SMC/HVC instruction
+ * to call into firmware to adjust the mitigation state.
+ */
+void __init spectre_v4_patch_fw_mitigation_conduit(struct alt_instr *alt,
+						   __le32 *origptr,
+						   __le32 *updptr, int nr_inst)
+{
+	u32 insn;
+
+	BUG_ON(nr_inst != 1); /* NOP -> HVC/SMC */
+
+	switch (arm_smccc_1_1_get_conduit()) {
+	case SMCCC_CONDUIT_HVC:
+		insn = aarch64_insn_get_hvc_value();
+		break;
+	case SMCCC_CONDUIT_SMC:
+		insn = aarch64_insn_get_smc_value();
+		break;
+	default:
+		return;
+	}
+
+	*updptr = cpu_to_le32(insn);
+}
+
+static enum mitigation_state spectre_v4_enable_fw_mitigation(void)
+{
+	enum mitigation_state state;
+
+	state = spectre_v4_get_cpu_fw_mitigation_state();
+	if (state != SPECTRE_MITIGATED)
+		return state;
+
+	if (spectre_v4_mitigations_off()) {
+		arm_smccc_1_1_invoke(ARM_SMCCC_ARCH_WORKAROUND_2, false, NULL);
+		return SPECTRE_VULNERABLE;
+	}
+
+	arm_smccc_1_1_invoke(ARM_SMCCC_ARCH_WORKAROUND_2, true, NULL);
+
+	if (spectre_v4_mitigations_dynamic())
+		__this_cpu_write(arm64_ssbd_callback_required, 1);
+
+	return SPECTRE_MITIGATED;
+}
+
+void spectre_v4_enable_mitigation(const struct arm64_cpu_capabilities *__unused)
+{
+	enum mitigation_state state;
+
+	WARN_ON(preemptible());
+
+	state = spectre_v4_enable_hw_mitigation();
+	if (state == SPECTRE_VULNERABLE)
+		state = spectre_v4_enable_fw_mitigation();
+
+	update_mitigation_state(&spectre_v4_state, state);
+}
+
+static void __update_pstate_ssbs(struct pt_regs *regs, bool state)
+{
+	u64 bit = compat_user_mode(regs) ? PSR_AA32_SSBS_BIT : PSR_SSBS_BIT;
+
+	if (state)
+		regs->pstate |= bit;
+	else
+		regs->pstate &= ~bit;
+}
+
+void spectre_v4_enable_task_mitigation(struct task_struct *tsk)
+{
+	struct pt_regs *regs = task_pt_regs(tsk);
+	bool ssbs = false, kthread = tsk->flags & PF_KTHREAD;
+
+	if (spectre_v4_mitigations_off())
+		ssbs = true;
+	else if (spectre_v4_mitigations_dynamic() && !kthread)
+		ssbs = !test_tsk_thread_flag(tsk, TIF_SSBD);
+
+	__update_pstate_ssbs(regs, ssbs);
+}
+
+/*
+ * The Spectre-v4 mitigation can be controlled via a prctl() from userspace.
+ * This is interesting because the "speculation disabled" behaviour can be
+ * configured so that it is preserved across exec(), which means that the
+ * prctl() may be necessary even when PSTATE.SSBS can be toggled directly
+ * from userspace.
+ */
+static void ssbd_prctl_enable_mitigation(struct task_struct *task)
+{
+	task_clear_spec_ssb_noexec(task);
+	task_set_spec_ssb_disable(task);
+	set_tsk_thread_flag(task, TIF_SSBD);
+}
+
+static void ssbd_prctl_disable_mitigation(struct task_struct *task)
+{
+	task_clear_spec_ssb_noexec(task);
+	task_clear_spec_ssb_disable(task);
+	clear_tsk_thread_flag(task, TIF_SSBD);
+}
+
+static int ssbd_prctl_set(struct task_struct *task, unsigned long ctrl)
+{
+	switch (ctrl) {
+	case PR_SPEC_ENABLE:
+		/* Enable speculation: disable mitigation */
+		/*
+		 * Force disabled speculation prevents it from being
+		 * re-enabled.
+		 */
+		if (task_spec_ssb_force_disable(task))
+			return -EPERM;
+
+		/*
+		 * If the mitigation is forced on, then speculation is forced
+		 * off and we again prevent it from being re-enabled.
+		 */
+		if (spectre_v4_mitigations_on())
+			return -EPERM;
+
+		ssbd_prctl_disable_mitigation(task);
+		break;
+	case PR_SPEC_FORCE_DISABLE:
+		/* Force disable speculation: force enable mitigation */
+		/*
+		 * If the mitigation is forced off, then speculation is forced
+		 * on and we prevent it from being disabled.
+		 */
+		if (spectre_v4_mitigations_off())
+			return -EPERM;
+
+		task_set_spec_ssb_force_disable(task);
+		fallthrough;
+	case PR_SPEC_DISABLE:
+		/* Disable speculation: enable mitigation */
+		/* Same as PR_SPEC_FORCE_DISABLE */
+		if (spectre_v4_mitigations_off())
+			return -EPERM;
+
+		ssbd_prctl_enable_mitigation(task);
+		break;
+	case PR_SPEC_DISABLE_NOEXEC:
+		/* Disable speculation until execve(): enable mitigation */
+		/*
+		 * If the mitigation state is forced one way or the other, then
+		 * we must fail now before we try to toggle it on execve().
+		 */
+		if (task_spec_ssb_force_disable(task) ||
+		    spectre_v4_mitigations_off() ||
+		    spectre_v4_mitigations_on()) {
+			return -EPERM;
+		}
+
+		ssbd_prctl_enable_mitigation(task);
+		task_set_spec_ssb_noexec(task);
+		break;
+	default:
+		return -ERANGE;
+	}
+
+	spectre_v4_enable_task_mitigation(task);
+	return 0;
+}
+
+int arch_prctl_spec_ctrl_set(struct task_struct *task, unsigned long which,
+			     unsigned long ctrl)
+{
+	switch (which) {
+	case PR_SPEC_STORE_BYPASS:
+		return ssbd_prctl_set(task, ctrl);
+	default:
+		return -ENODEV;
+	}
+}
+
+static int ssbd_prctl_get(struct task_struct *task)
+{
+	switch (spectre_v4_state) {
+	case SPECTRE_UNAFFECTED:
+		return PR_SPEC_NOT_AFFECTED;
+	case SPECTRE_MITIGATED:
+		if (spectre_v4_mitigations_on())
+			return PR_SPEC_NOT_AFFECTED;
+
+		if (spectre_v4_mitigations_dynamic())
+			break;
+
+		/* Mitigations are disabled, so we're vulnerable. */
+		fallthrough;
+	case SPECTRE_VULNERABLE:
+		fallthrough;
+	default:
+		return PR_SPEC_ENABLE;
+	}
+
+	/* Check the mitigation state for this task */
+	if (task_spec_ssb_force_disable(task))
+		return PR_SPEC_PRCTL | PR_SPEC_FORCE_DISABLE;
+
+	if (task_spec_ssb_noexec(task))
+		return PR_SPEC_PRCTL | PR_SPEC_DISABLE_NOEXEC;
+
+	if (task_spec_ssb_disable(task))
+		return PR_SPEC_PRCTL | PR_SPEC_DISABLE;
+
+	return PR_SPEC_PRCTL | PR_SPEC_ENABLE;
+}
+
+int arch_prctl_spec_ctrl_get(struct task_struct *task, unsigned long which)
+{
+	switch (which) {
+	case PR_SPEC_STORE_BYPASS:
+		return ssbd_prctl_get(task);
+	default:
+		return -ENODEV;
+	}
+}

arch/arm64/kernel/ssbd.c

-// SPDX-License-Identifier: GPL-2.0
-/*
- * Copyright (C) 2018 ARM Ltd, All Rights Reserved.
- */
-
-#include <linux/compat.h>
-#include <linux/errno.h>
-#include <linux/prctl.h>
-#include <linux/sched.h>
-#include <linux/sched/task_stack.h>
-#include <linux/thread_info.h>
-
-#include <asm/cpufeature.h>
-
-static void ssbd_ssbs_enable(struct task_struct *task)
-{
-	u64 val = is_compat_thread(task_thread_info(task)) ?
-		  PSR_AA32_SSBS_BIT : PSR_SSBS_BIT;
-
-	task_pt_regs(task)->pstate |= val;
-}
-
-static void ssbd_ssbs_disable(struct task_struct *task)
-{
-	u64 val = is_compat_thread(task_thread_info(task)) ?
-		  PSR_AA32_SSBS_BIT : PSR_SSBS_BIT;
-
-	task_pt_regs(task)->pstate &= ~val;
-}
-
-/*
- * prctl interface for SSBD
- */
-static int ssbd_prctl_set(struct task_struct *task, unsigned long ctrl)
-{
-	int state = arm64_get_ssbd_state();
-
-	/* Unsupported */
-	if (state == ARM64_SSBD_UNKNOWN)
-		return -ENODEV;
-
-	/* Treat the unaffected/mitigated state separately */
-	if (state == ARM64_SSBD_MITIGATED) {
-		switch (ctrl) {
-		case PR_SPEC_ENABLE:
-			return -EPERM;
-		case PR_SPEC_DISABLE:
-		case PR_SPEC_FORCE_DISABLE:
-			return 0;
-		}
-	}
-
-	/*
-	 * Things are a bit backward here: the arm64 internal API
-	 * *enables the mitigation* when the userspace API *disables
-	 * speculation*. So much fun.
-	 */
-	switch (ctrl) {
-	case PR_SPEC_ENABLE:
-		/* If speculation is force disabled, enable is not allowed */
-		if (state == ARM64_SSBD_FORCE_ENABLE ||
-		    task_spec_ssb_force_disable(task))
-			return -EPERM;
-		task_clear_spec_ssb_disable(task);
-		clear_tsk_thread_flag(task, TIF_SSBD);
-		ssbd_ssbs_enable(task);
-		break;
-	case PR_SPEC_DISABLE:
-		if (state == ARM64_SSBD_FORCE_DISABLE)
-			return -EPERM;
-		task_set_spec_ssb_disable(task);
-		set_tsk_thread_flag(task, TIF_SSBD);
-		ssbd_ssbs_disable(task);
-		break;
-	case PR_SPEC_FORCE_DISABLE:
-		if (state == ARM64_SSBD_FORCE_DISABLE)
-			return -EPERM;
-		task_set_spec_ssb_disable(task);
-		task_set_spec_ssb_force_disable(task);
-		set_tsk_thread_flag(task, TIF_SSBD);
-		ssbd_ssbs_disable(task);
-		break;
-	default:
-		return -ERANGE;
-	}
-
-	return 0;
-}
-
-int arch_prctl_spec_ctrl_set(struct task_struct *task, unsigned long which,
-			     unsigned long ctrl)
-{
-	switch (which) {
-	case PR_SPEC_STORE_BYPASS:
-		return ssbd_prctl_set(task, ctrl);
-	default:
-		return -ENODEV;
-	}
-}
-
-static int ssbd_prctl_get(struct task_struct *task)
-{
-	switch (arm64_get_ssbd_state()) {
-	case ARM64_SSBD_UNKNOWN:
-		return -ENODEV;
-	case ARM64_SSBD_FORCE_ENABLE:
-		return PR_SPEC_DISABLE;
-	case ARM64_SSBD_KERNEL:
-		if (task_spec_ssb_force_disable(task))
-			return PR_SPEC_PRCTL | PR_SPEC_FORCE_DISABLE;
-		if (task_spec_ssb_disable(task))
-			return PR_SPEC_PRCTL | PR_SPEC_DISABLE;
-		return PR_SPEC_PRCTL | PR_SPEC_ENABLE;
-	case ARM64_SSBD_FORCE_DISABLE:
-		return PR_SPEC_ENABLE;
-	default:
-		return PR_SPEC_NOT_AFFECTED;
-	}
-}
-
-int arch_prctl_spec_ctrl_get(struct task_struct *task, unsigned long which)
-{
-	switch (which) {
-	case PR_SPEC_STORE_BYPASS:
-		return ssbd_prctl_get(task);
-	default:
-		return -ENODEV;
-	}
-}

arch/arm64/kernel/suspend.c

@@ -72,8 +72,7 @@ void notrace __cpu_suspend_exit(void)
 	 * have turned the mitigation on. If the user has forcefully
 	 * disabled it, make sure their wishes are obeyed.
 	 */
-	if (arm64_get_ssbd_state() == ARM64_SSBD_FORCE_DISABLE)
-		arm64_set_ssbd_mitigation(false);
+	spectre_v4_enable_mitigation(NULL);
 }
 
 /*

arch/arm64/kernel/vmlinux.lds.S

@@ -9,6 +9,7 @@
 
 #include <asm-generic/vmlinux.lds.h>
 #include <asm/cache.h>
+#include <asm/hyp_image.h>
 #include <asm/kernel-pgtable.h>
 #include <asm/memory.h>
 #include <asm/page.h>
@@ -21,12 +22,23 @@ ENTRY(_text)
 jiffies = jiffies_64;
 
 
+#ifdef CONFIG_KVM
 #define HYPERVISOR_EXTABLE					\
 	. = ALIGN(SZ_8);					\
 	__start___kvm_ex_table = .;				\
 	*(__kvm_ex_table)					\
 	__stop___kvm_ex_table = .;
 
+#define HYPERVISOR_PERCPU_SECTION				\
+	. = ALIGN(PAGE_SIZE);					\
+	HYP_SECTION_NAME(.data..percpu) : {			\
+		*(HYP_SECTION_NAME(.data..percpu))		\
+	}
+#else /* CONFIG_KVM */
+#define HYPERVISOR_EXTABLE
+#define HYPERVISOR_PERCPU_SECTION
+#endif
+
 #define HYPERVISOR_TEXT					\
 	/*						\
 	 * Align to 4 KB so that			\
@@ -190,6 +202,7 @@ SECTIONS
 	}
 
 	PERCPU_SECTION(L1_CACHE_BYTES)
+	HYPERVISOR_PERCPU_SECTION
 
 	.rela.dyn : ALIGN(8) {
 		*(.rela .rela*)

arch/arm64/kvm/Kconfig

@@ -57,9 +57,6 @@ config KVM_ARM_PMU
 	  Adds support for a virtual Performance Monitoring Unit (PMU) in
 	  virtual machines.
 
-config KVM_INDIRECT_VECTORS
-	def_bool HARDEN_BRANCH_PREDICTOR || RANDOMIZE_BASE
-
 endif # KVM
 
 endif # VIRTUALIZATION

arch/arm64/kvm/arm.c

@@ -46,10 +46,10 @@
 __asm__(".arch_extension	virt");
 #endif
 
-DEFINE_PER_CPU(struct kvm_host_data, kvm_host_data);
-DEFINE_PER_CPU(struct kvm_cpu_context, kvm_hyp_ctxt);
-DEFINE_PER_CPU(unsigned long, kvm_hyp_vector);
+DECLARE_KVM_HYP_PER_CPU(unsigned long, kvm_hyp_vector);
+
 static DEFINE_PER_CPU(unsigned long, kvm_arm_hyp_stack_page);
+unsigned long kvm_arm_hyp_percpu_base[NR_CPUS];
 
 /* The VMID used in the VTTBR */
 static atomic64_t kvm_vmid_gen = ATOMIC64_INIT(1);
@@ -1263,6 +1263,53 @@ long kvm_arch_vm_ioctl(struct file *filp,
 	}
 }
 
+static unsigned long nvhe_percpu_size(void)
+{
+	return (unsigned long)CHOOSE_NVHE_SYM(__per_cpu_end) -
+		(unsigned long)CHOOSE_NVHE_SYM(__per_cpu_start);
+}
+
+static unsigned long nvhe_percpu_order(void)
+{
+	unsigned long size = nvhe_percpu_size();
+
+	return size ? get_order(size) : 0;
+}
+
+static int kvm_map_vectors(void)
+{
+	/*
+	 * SV2  = ARM64_SPECTRE_V2
+	 * HEL2 = ARM64_HARDEN_EL2_VECTORS
+	 *
+	 * !SV2 + !HEL2 -> use direct vectors
+	 *  SV2 + !HEL2 -> use hardened vectors in place
+	 * !SV2 +  HEL2 -> allocate one vector slot and use exec mapping
+	 *  SV2 +  HEL2 -> use hardened vectors and use exec mapping
+	 */
+	if (cpus_have_const_cap(ARM64_SPECTRE_V2)) {
+		__kvm_bp_vect_base = kvm_ksym_ref(__bp_harden_hyp_vecs);
+		__kvm_bp_vect_base = kern_hyp_va(__kvm_bp_vect_base);
+	}
+
+	if (cpus_have_const_cap(ARM64_HARDEN_EL2_VECTORS)) {
+		phys_addr_t vect_pa = __pa_symbol(__bp_harden_hyp_vecs);
+		unsigned long size = __BP_HARDEN_HYP_VECS_SZ;
+
+		/*
+		 * Always allocate a spare vector slot, as we don't
+		 * know yet which CPUs have a BP hardening slot that
+		 * we can reuse.
+		 */
+		__kvm_harden_el2_vector_slot = atomic_inc_return(&arm64_el2_vector_last_slot);
+		BUG_ON(__kvm_harden_el2_vector_slot >= BP_HARDEN_EL2_SLOTS);
+		return create_hyp_exec_mappings(vect_pa, size,
+						&__kvm_bp_vect_base);
+	}
+
+	return 0;
+}
+
 static void cpu_init_hyp_mode(void)
 {
 	phys_addr_t pgd_ptr;
@@ -1279,8 +1326,8 @@ static void cpu_init_hyp_mode(void)
 	 * kernel's mapping to the linear mapping, and store it in tpidr_el2
 	 * so that we can use adr_l to access per-cpu variables in EL2.
 	 */
-	tpidr_el2 = ((unsigned long)this_cpu_ptr(&kvm_host_data) -
-		     (unsigned long)kvm_ksym_ref(&kvm_host_data));
+	tpidr_el2 = (unsigned long)this_cpu_ptr_nvhe_sym(__per_cpu_start) -
+		    (unsigned long)kvm_ksym_ref(CHOOSE_NVHE_SYM(__per_cpu_start));
 
 	pgd_ptr = kvm_mmu_get_httbr();
 	hyp_stack_ptr = __this_cpu_read(kvm_arm_hyp_stack_page) + PAGE_SIZE;
@@ -1303,7 +1350,7 @@ static void cpu_init_hyp_mode(void)
 	 * at EL2.
 	 */
 	if (this_cpu_has_cap(ARM64_SSBS) &&
-	    arm64_get_ssbd_state() == ARM64_SSBD_FORCE_DISABLE) {
+	    arm64_get_spectre_v4_state() == SPECTRE_VULNERABLE) {
 		kvm_call_hyp_nvhe(__kvm_enable_ssbs);
 	}
 }
@@ -1316,11 +1363,11 @@ static void cpu_hyp_reset(void)
 
 static void cpu_hyp_reinit(void)
 {
-	kvm_init_host_cpu_context(&this_cpu_ptr(&kvm_host_data)->host_ctxt);
+	kvm_init_host_cpu_context(&this_cpu_ptr_hyp_sym(kvm_host_data)->host_ctxt);
 
 	cpu_hyp_reset();
 
-	__this_cpu_write(kvm_hyp_vector, (unsigned long)kvm_get_hyp_vector());
+	*this_cpu_ptr_hyp_sym(kvm_hyp_vector) = (unsigned long)kvm_get_hyp_vector();
 
 	if (is_kernel_in_hyp_mode())
 		kvm_timer_init_vhe();
@@ -1472,8 +1519,10 @@ static void teardown_hyp_mode(void)
 	int cpu;
 
 	free_hyp_pgds();
-	for_each_possible_cpu(cpu)
+	for_each_possible_cpu(cpu) {
 		free_page(per_cpu(kvm_arm_hyp_stack_page, cpu));
+		free_pages(kvm_arm_hyp_percpu_base[cpu], nvhe_percpu_order());
+	}
 }
 
 /**
@@ -1506,6 +1555,24 @@ static int init_hyp_mode(void)
 		per_cpu(kvm_arm_hyp_stack_page, cpu) = stack_page;
 	}
 
+	/*
+	 * Allocate and initialize pages for Hypervisor-mode percpu regions.
+	 */
+	for_each_possible_cpu(cpu) {
+		struct page *page;
+		void *page_addr;
+
+		page = alloc_pages(GFP_KERNEL, nvhe_percpu_order());
+		if (!page) {
+			err = -ENOMEM;
+			goto out_err;
+		}
+
+		page_addr = page_address(page);
+		memcpy(page_addr, CHOOSE_NVHE_SYM(__per_cpu_start), nvhe_percpu_size());
+		kvm_arm_hyp_percpu_base[cpu] = (unsigned long)page_addr;
+	}
+
 	/*
 	 * Map the Hyp-code called directly from the host
 	 */
@@ -1550,39 +1617,20 @@ static int init_hyp_mode(void)
 		}
 	}
 
+	/*
+	 * Map Hyp percpu pages
+	 */
 	for_each_possible_cpu(cpu) {
-		struct kvm_host_data *cpu_data;
-		struct kvm_cpu_context *hyp_ctxt;
-		unsigned long *vector;
+		char *percpu_begin = (char *)kvm_arm_hyp_percpu_base[cpu];
+		char *percpu_end = percpu_begin + nvhe_percpu_size();
 
-		cpu_data = per_cpu_ptr(&kvm_host_data, cpu);
-		err = create_hyp_mappings(cpu_data, cpu_data + 1, PAGE_HYP);
+		err = create_hyp_mappings(percpu_begin, percpu_end, PAGE_HYP);
 
 		if (err) {
-			kvm_err("Cannot map host CPU state: %d\n", err);
+			kvm_err("Cannot map hyp percpu region\n");
 			goto out_err;
 		}
-
-		hyp_ctxt = per_cpu_ptr(&kvm_hyp_ctxt, cpu);
-		err = create_hyp_mappings(hyp_ctxt, hyp_ctxt + 1, PAGE_HYP);
-
-		if (err) {
-			kvm_err("Cannot map hyp context: %d\n", err);
-			goto out_err;
-		}
-
-		vector = per_cpu_ptr(&kvm_hyp_vector, cpu);
-		err = create_hyp_mappings(vector, vector + 1, PAGE_HYP);
-
-		if (err) {
-			kvm_err("Cannot map hyp guest vector address\n");
-			goto out_err;
 	}
-	}
-
-	err = hyp_map_aux_data();
-	if (err)
-		kvm_err("Cannot map host auxiliary data: %d\n", err);
 
 	return 0;
 

arch/arm64/kvm/hyp/Makefile

@@ -10,5 +10,4 @@ subdir-ccflags-y := -I$(incdir)				\
 		    -DDISABLE_BRANCH_PROFILING		\
 		    $(DISABLE_STACKLEAK_PLUGIN)
 
-obj-$(CONFIG_KVM) += vhe/ nvhe/ pgtable.o
-obj-$(CONFIG_KVM_INDIRECT_VECTORS) += smccc_wa.o
+obj-$(CONFIG_KVM) += vhe/ nvhe/ pgtable.o smccc_wa.o

arch/arm64/kvm/hyp/entry.S

@@ -25,7 +25,7 @@ SYM_FUNC_START(__guest_enter)
 	// x1-x17: clobbered by macros
 	// x29: guest context
 
-	hyp_adr_this_cpu x1, kvm_hyp_ctxt, x2
+	adr_this_cpu x1, kvm_hyp_ctxt, x2
 
 	// Store the hyp regs
 	save_callee_saved_regs x1
@@ -93,7 +93,7 @@ SYM_INNER_LABEL(__guest_exit_panic, SYM_L_GLOBAL)
 	// return address to tail call into hyp_panic. As a side effect, the
 	// current state is saved to the guest context but it will only be
 	// accurate if the guest had been completely restored.
-	hyp_adr_this_cpu x0, kvm_hyp_ctxt, x1
+	adr_this_cpu x0, kvm_hyp_ctxt, x1
 	adr	x1, hyp_panic
 	str	x1, [x0, #CPU_XREG_OFFSET(30)]
 
@@ -131,7 +131,7 @@ SYM_INNER_LABEL(__guest_exit, SYM_L_GLOBAL)
 	// Store the guest's sp_el0
 	save_sp_el0	x1, x2
 
-	hyp_adr_this_cpu x2, kvm_hyp_ctxt, x3
+	adr_this_cpu x2, kvm_hyp_ctxt, x3
 
 	// Macro ptrauth_switch_to_hyp format:
 	// 	ptrauth_switch_to_hyp(guest cxt, host cxt, tmp1, tmp2, tmp3)

arch/arm64/kvm/hyp/hyp-entry.S

@@ -63,35 +63,6 @@ el1_sync:				// Guest trapped into EL2
 			  ARM_SMCCC_ARCH_WORKAROUND_2)
 	cbnz	w1, el1_trap
 
-#ifdef CONFIG_ARM64_SSBD
-alternative_cb	arm64_enable_wa2_handling
-	b	wa2_end
-alternative_cb_end
-	get_vcpu_ptr	x2, x0
-	ldr	x0, [x2, #VCPU_WORKAROUND_FLAGS]
-
-	// Sanitize the argument and update the guest flags
-	ldr	x1, [sp, #8]			// Guest's x1
-	clz	w1, w1				// Murphy's device:
-	lsr	w1, w1, #5			// w1 = !!w1 without using
-	eor	w1, w1, #1			// the flags...
-	bfi	x0, x1, #VCPU_WORKAROUND_2_FLAG_SHIFT, #1
-	str	x0, [x2, #VCPU_WORKAROUND_FLAGS]
-
-	/* Check that we actually need to perform the call */
-	hyp_ldr_this_cpu x0, arm64_ssbd_callback_required, x2
-	cbz	x0, wa2_end
-
-	mov	w0, #ARM_SMCCC_ARCH_WORKAROUND_2
-	smc	#0
-
-	/* Don't leak data from the SMC call */
-	mov	x3, xzr
-wa2_end:
-	mov	x2, xzr
-	mov	x1, xzr
-#endif
-
 wa_epilogue:
 	mov	x0, xzr
 	add	sp, sp, #16
@@ -216,7 +187,6 @@ SYM_CODE_START(__kvm_hyp_vector)
 	valid_vect	el1_error		// Error 32-bit EL1
 SYM_CODE_END(__kvm_hyp_vector)
 
-#ifdef CONFIG_KVM_INDIRECT_VECTORS
 .macro hyp_ventry
 	.align 7
 1:	esb
@@ -266,4 +236,3 @@ SYM_CODE_START(__bp_harden_hyp_vecs)
 1:	.org __bp_harden_hyp_vecs + __BP_HARDEN_HYP_VECS_SZ
 	.org 1b
 SYM_CODE_END(__bp_harden_hyp_vecs)
-#endif

arch/arm64/kvm/hyp/include/hyp/debug-sr.h

@@ -135,7 +135,7 @@ static inline void __debug_switch_to_guest_common(struct kvm_vcpu *vcpu)
 	if (!(vcpu->arch.flags & KVM_ARM64_DEBUG_DIRTY))
 		return;
 
-	host_ctxt = &__hyp_this_cpu_ptr(kvm_host_data)->host_ctxt;
+	host_ctxt = &this_cpu_ptr(&kvm_host_data)->host_ctxt;
 	guest_ctxt = &vcpu->arch.ctxt;
 	host_dbg = &vcpu->arch.host_debug_state.regs;
 	guest_dbg = kern_hyp_va(vcpu->arch.debug_ptr);
@@ -154,7 +154,7 @@ static inline void __debug_switch_to_host_common(struct kvm_vcpu *vcpu)
 	if (!(vcpu->arch.flags & KVM_ARM64_DEBUG_DIRTY))
 		return;
 
-	host_ctxt = &__hyp_this_cpu_ptr(kvm_host_data)->host_ctxt;
+	host_ctxt = &this_cpu_ptr(&kvm_host_data)->host_ctxt;
 	guest_ctxt = &vcpu->arch.ctxt;
 	host_dbg = &vcpu->arch.host_debug_state.regs;
 	guest_dbg = kern_hyp_va(vcpu->arch.debug_ptr);

arch/arm64/kvm/hyp/include/hyp/switch.h

@@ -383,7 +383,7 @@ static inline bool __hyp_handle_ptrauth(struct kvm_vcpu *vcpu)
 	    !esr_is_ptrauth_trap(kvm_vcpu_get_esr(vcpu)))
 		return false;
 
-	ctxt = __hyp_this_cpu_ptr(kvm_hyp_ctxt);
+	ctxt = this_cpu_ptr(&kvm_hyp_ctxt);
 	__ptrauth_save_key(ctxt, APIA);
 	__ptrauth_save_key(ctxt, APIB);
 	__ptrauth_save_key(ctxt, APDA);
@@ -476,39 +476,6 @@ static inline bool fixup_guest_exit(struct kvm_vcpu *vcpu, u64 *exit_code)
 	return false;
 }
 
-static inline bool __needs_ssbd_off(struct kvm_vcpu *vcpu)
-{
-	if (!cpus_have_final_cap(ARM64_SSBD))
-		return false;
-
-	return !(vcpu->arch.workaround_flags & VCPU_WORKAROUND_2_FLAG);
-}
-
-static inline void __set_guest_arch_workaround_state(struct kvm_vcpu *vcpu)
-{
-#ifdef CONFIG_ARM64_SSBD
-	/*
-	 * The host runs with the workaround always present. If the
-	 * guest wants it disabled, so be it...
-	 */
-	if (__needs_ssbd_off(vcpu) &&
-	    __hyp_this_cpu_read(arm64_ssbd_callback_required))
-		arm_smccc_1_1_smc(ARM_SMCCC_ARCH_WORKAROUND_2, 0, NULL);
-#endif
-}
-
-static inline void __set_host_arch_workaround_state(struct kvm_vcpu *vcpu)
-{
-#ifdef CONFIG_ARM64_SSBD
-	/*
-	 * If the guest has disabled the workaround, bring it back on.
-	 */
-	if (__needs_ssbd_off(vcpu) &&
-	    __hyp_this_cpu_read(arm64_ssbd_callback_required))
-		arm_smccc_1_1_smc(ARM_SMCCC_ARCH_WORKAROUND_2, 1, NULL);
-#endif
-}
-
 static inline void __kvm_unexpected_el2_exception(void)
 {
 	extern char __guest_exit_panic[];

arch/arm64/kvm/hyp/nvhe/.gitignore

+# SPDX-License-Identifier: GPL-2.0-only
+hyp.lds

arch/arm64/kvm/hyp/nvhe/Makefile

@@ -10,40 +10,46 @@ obj-y := timer-sr.o sysreg-sr.o debug-sr.o switch.o tlb.o hyp-init.o host.o hyp-
 obj-y += ../vgic-v3-sr.o ../aarch32.o ../vgic-v2-cpuif-proxy.o ../entry.o \
 	 ../fpsimd.o ../hyp-entry.o
 
-obj-y := $(patsubst %.o,%.hyp.o,$(obj-y))
-extra-y := $(patsubst %.hyp.o,%.hyp.tmp.o,$(obj-y))
+##
+## Build rules for compiling nVHE hyp code
+## Output of this folder is `kvm_nvhe.o`, a partially linked object
+## file containing all nVHE hyp code and data.
+##
 
-$(obj)/%.hyp.tmp.o: $(src)/%.c FORCE
+hyp-obj := $(patsubst %.o,%.nvhe.o,$(obj-y))
+obj-y := kvm_nvhe.o
+extra-y := $(hyp-obj) kvm_nvhe.tmp.o hyp.lds
+
+# 1) Compile all source files to `.nvhe.o` object files. The file extension
+#    avoids file name clashes for files shared with VHE.
+$(obj)/%.nvhe.o: $(src)/%.c FORCE
 	$(call if_changed_rule,cc_o_c)
-$(obj)/%.hyp.tmp.o: $(src)/%.S FORCE
+$(obj)/%.nvhe.o: $(src)/%.S FORCE
 	$(call if_changed_rule,as_o_S)
-$(obj)/%.hyp.o: $(obj)/%.hyp.tmp.o FORCE
-	$(call if_changed,hypcopy)
 
-# Disable reordering functions by GCC (enabled at -O2).
-# This pass puts functions into '.text.*' sections to aid the linker
-# in optimizing ELF layout. See HYPCOPY comment below for more info.
-ccflags-y += $(call cc-option,-fno-reorder-functions)
+# 2) Compile linker script.
+$(obj)/hyp.lds: $(src)/hyp.lds.S FORCE
+	$(call if_changed_dep,cpp_lds_S)
+
+# 3) Partially link all '.nvhe.o' files and apply the linker script.
+#    Prefixes names of ELF sections with '.hyp', eg. '.hyp.text'.
+#    Note: The following rule assumes that the 'ld' rule puts LDFLAGS before
+#          the list of dependencies to form '-T $(obj)/hyp.lds'. This is to
+#          keep the dependency on the target while avoiding an error from
+#          GNU ld if the linker script is passed to it twice.
+LDFLAGS_kvm_nvhe.tmp.o := -r -T
+$(obj)/kvm_nvhe.tmp.o: $(obj)/hyp.lds $(addprefix $(obj)/,$(hyp-obj)) FORCE
+	$(call if_changed,ld)
+
+# 4) Produce the final 'kvm_nvhe.o', ready to be linked into 'vmlinux'.
+#    Prefixes names of ELF symbols with '__kvm_nvhe_'.
+$(obj)/kvm_nvhe.o: $(obj)/kvm_nvhe.tmp.o FORCE
+	$(call if_changed,hypcopy)
 
 # The HYPCOPY command uses `objcopy` to prefix all ELF symbol names
-# and relevant ELF section names to avoid clashes with VHE code/data.
-#
-# Hyp code is assumed to be in the '.text' section of the input object
-# files (with the exception of specialized sections such as
-# '.hyp.idmap.text'). This assumption may be broken by a compiler that
-# divides code into sections like '.text.unlikely' so as to optimize
-# ELF layout. HYPCOPY checks that no such sections exist in the input
-# using `objdump`, otherwise they would be linked together with other
-# kernel code and not memory-mapped correctly at runtime.
+# to avoid clashes with VHE code/data.
 quiet_cmd_hypcopy = HYPCOPY $@
-      cmd_hypcopy =							\
-	if $(OBJDUMP) -h $< | grep -F '.text.'; then			\
-		echo "$@: function reordering not supported in nVHE hyp code" >&2; \
-		/bin/false;						\
-	fi;								\
-	$(OBJCOPY) --prefix-symbols=__kvm_nvhe_				\
-		   --rename-section=.text=.hyp.text			\
-		   $< $@
+      cmd_hypcopy = $(OBJCOPY) --prefix-symbols=__kvm_nvhe_ $< $@
 
 # Remove ftrace and Shadow Call Stack CFLAGS.
 # This is equivalent to the 'notrace' and '__noscs' annotations.

arch/arm64/kvm/hyp/nvhe/hyp.lds.S

+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Copyright (C) 2020 Google LLC.
+ * Written by David Brazdil <dbrazdil@google.com>
+ *
+ * Linker script used for partial linking of nVHE EL2 object files.
+ */
+
+#include <asm/hyp_image.h>
+#include <asm-generic/vmlinux.lds.h>
+#include <asm/cache.h>
+#include <asm/memory.h>
+
+SECTIONS {
+	HYP_SECTION(.text)
+	HYP_SECTION_NAME(.data..percpu) : {
+		PERCPU_INPUT(L1_CACHE_BYTES)
+	}
+}

arch/arm64/kvm/hyp/nvhe/switch.c

@@ -27,6 +27,11 @@
 #include <asm/processor.h>
 #include <asm/thread_info.h>
 
+/* Non-VHE specific context */
+DEFINE_PER_CPU(struct kvm_host_data, kvm_host_data);
+DEFINE_PER_CPU(struct kvm_cpu_context, kvm_hyp_ctxt);
+DEFINE_PER_CPU(unsigned long, kvm_hyp_vector);
+
 static void __activate_traps(struct kvm_vcpu *vcpu)
 {
 	u64 val;
@@ -42,7 +47,7 @@ static void __activate_traps(struct kvm_vcpu *vcpu)
 	}
 
 	write_sysreg(val, cptr_el2);
-	write_sysreg(__hyp_this_cpu_read(kvm_hyp_vector), vbar_el2);
+	write_sysreg(__this_cpu_read(kvm_hyp_vector), vbar_el2);
 
 	if (cpus_have_final_cap(ARM64_WORKAROUND_SPECULATIVE_AT)) {
 		struct kvm_cpu_context *ctxt = &vcpu->arch.ctxt;
@@ -176,7 +181,7 @@ int __kvm_vcpu_run(struct kvm_vcpu *vcpu)
 		pmr_sync();
 	}
 
-	host_ctxt = &__hyp_this_cpu_ptr(kvm_host_data)->host_ctxt;
+	host_ctxt = &this_cpu_ptr(&kvm_host_data)->host_ctxt;
 	host_ctxt->__hyp_running_vcpu = vcpu;
 	guest_ctxt = &vcpu->arch.ctxt;
 
@@ -203,8 +208,6 @@ int __kvm_vcpu_run(struct kvm_vcpu *vcpu)
 
 	__debug_switch_to_guest(vcpu);
 
-	__set_guest_arch_workaround_state(vcpu);
-
 	do {
 		/* Jump in the fire! */
 		exit_code = __guest_enter(vcpu);
@@ -212,8 +215,6 @@ int __kvm_vcpu_run(struct kvm_vcpu *vcpu)
 		/* And we're baaack! */
 	} while (fixup_guest_exit(vcpu, &exit_code));
 
-	__set_host_arch_workaround_state(vcpu);
-
 	__sysreg_save_state_nvhe(guest_ctxt);
 	__sysreg32_save_state(vcpu);
 	__timer_disable_traps(vcpu);
@@ -254,7 +255,7 @@ void __noreturn hyp_panic(void)
 	struct kvm_cpu_context *host_ctxt;
 	struct kvm_vcpu *vcpu;
 
-	host_ctxt = &__hyp_this_cpu_ptr(kvm_host_data)->host_ctxt;
+	host_ctxt = &this_cpu_ptr(&kvm_host_data)->host_ctxt;
 	vcpu = host_ctxt->__hyp_running_vcpu;
 
 	if (vcpu) {

arch/arm64/kvm/hyp/vhe/switch.c

@@ -28,6 +28,11 @@
 
 const char __hyp_panic_string[] = "HYP panic:\nPS:%08llx PC:%016llx ESR:%08llx\nFAR:%016llx HPFAR:%016llx PAR:%016llx\nVCPU:%p\n";
 
+/* VHE specific context */
+DEFINE_PER_CPU(struct kvm_host_data, kvm_host_data);
+DEFINE_PER_CPU(struct kvm_cpu_context, kvm_hyp_ctxt);
+DEFINE_PER_CPU(unsigned long, kvm_hyp_vector);
+
 static void __activate_traps(struct kvm_vcpu *vcpu)
 {
 	u64 val;
@@ -108,7 +113,7 @@ static int __kvm_vcpu_run_vhe(struct kvm_vcpu *vcpu)
 	struct kvm_cpu_context *guest_ctxt;
 	u64 exit_code;
 
-	host_ctxt = &__hyp_this_cpu_ptr(kvm_host_data)->host_ctxt;
+	host_ctxt = &this_cpu_ptr(&kvm_host_data)->host_ctxt;
 	host_ctxt->__hyp_running_vcpu = vcpu;
 	guest_ctxt = &vcpu->arch.ctxt;
 
@@ -131,8 +136,6 @@ static int __kvm_vcpu_run_vhe(struct kvm_vcpu *vcpu)
 	sysreg_restore_guest_state_vhe(guest_ctxt);
 	__debug_switch_to_guest(vcpu);
 
-	__set_guest_arch_workaround_state(vcpu);
-
 	do {
 		/* Jump in the fire! */
 		exit_code = __guest_enter(vcpu);
@@ -140,8 +143,6 @@ static int __kvm_vcpu_run_vhe(struct kvm_vcpu *vcpu)
 		/* And we're baaack! */
 	} while (fixup_guest_exit(vcpu, &exit_code));
 
-	__set_host_arch_workaround_state(vcpu);
-
 	sysreg_save_guest_state_vhe(guest_ctxt);
 
 	__deactivate_traps(vcpu);
@@ -197,7 +198,7 @@ static void __hyp_call_panic(u64 spsr, u64 elr, u64 par)
 	struct kvm_cpu_context *host_ctxt;
 	struct kvm_vcpu *vcpu;
 
-	host_ctxt = &__hyp_this_cpu_ptr(kvm_host_data)->host_ctxt;
+	host_ctxt = &this_cpu_ptr(&kvm_host_data)->host_ctxt;
 	vcpu = host_ctxt->__hyp_running_vcpu;
 
 	__deactivate_traps(vcpu);

arch/arm64/kvm/hyp/vhe/sysreg-sr.c

@@ -66,7 +66,7 @@ void kvm_vcpu_load_sysregs_vhe(struct kvm_vcpu *vcpu)
 	struct kvm_cpu_context *guest_ctxt = &vcpu->arch.ctxt;
 	struct kvm_cpu_context *host_ctxt;
 
-	host_ctxt = &__hyp_this_cpu_ptr(kvm_host_data)->host_ctxt;
+	host_ctxt = &this_cpu_ptr(&kvm_host_data)->host_ctxt;
 	__sysreg_save_user_state(host_ctxt);
 
 	/*
@@ -100,7 +100,7 @@ void kvm_vcpu_put_sysregs_vhe(struct kvm_vcpu *vcpu)
 	struct kvm_cpu_context *guest_ctxt = &vcpu->arch.ctxt;
 	struct kvm_cpu_context *host_ctxt;
 
-	host_ctxt = &__hyp_this_cpu_ptr(kvm_host_data)->host_ctxt;
+	host_ctxt = &this_cpu_ptr(&kvm_host_data)->host_ctxt;
 	deactivate_traps_vhe_put();
 
 	__sysreg_save_el1_state(guest_ctxt);

arch/arm64/kvm/hypercalls.c

@@ -24,27 +24,36 @@ int kvm_hvc_call_handler(struct kvm_vcpu *vcpu)
 		feature = smccc_get_arg1(vcpu);
 		switch (feature) {
 		case ARM_SMCCC_ARCH_WORKAROUND_1:
-			switch (kvm_arm_harden_branch_predictor()) {
-			case KVM_BP_HARDEN_UNKNOWN:
+			switch (arm64_get_spectre_v2_state()) {
+			case SPECTRE_VULNERABLE:
 				break;
-			case KVM_BP_HARDEN_WA_NEEDED:
+			case SPECTRE_MITIGATED:
 				val = SMCCC_RET_SUCCESS;
 				break;
-			case KVM_BP_HARDEN_NOT_REQUIRED:
+			case SPECTRE_UNAFFECTED:
 				val = SMCCC_RET_NOT_REQUIRED;
 				break;
 			}
 			break;
 		case ARM_SMCCC_ARCH_WORKAROUND_2:
-			switch (kvm_arm_have_ssbd()) {
-			case KVM_SSBD_FORCE_DISABLE:
-			case KVM_SSBD_UNKNOWN:
+			switch (arm64_get_spectre_v4_state()) {
+			case SPECTRE_VULNERABLE:
 				break;
-			case KVM_SSBD_KERNEL:
-				val = SMCCC_RET_SUCCESS;
+			case SPECTRE_MITIGATED:
+				/*
+				 * SSBS everywhere: Indicate no firmware
+				 * support, as the SSBS support will be
+				 * indicated to the guest and the default is
+				 * safe.
+				 *
+				 * Otherwise, expose a permanent mitigation
+				 * to the guest, and hide SSBS so that the
+				 * guest stays protected.
+				 */
+				if (cpus_have_final_cap(ARM64_SSBS))
 					break;
-			case KVM_SSBD_FORCE_ENABLE:
-			case KVM_SSBD_MITIGATED:
+				fallthrough;
+			case SPECTRE_UNAFFECTED:
 				val = SMCCC_RET_NOT_REQUIRED;
 				break;
 			}

arch/arm64/kvm/pmu.c

@@ -31,9 +31,9 @@ static bool kvm_pmu_switch_needed(struct perf_event_attr *attr)
  */
 void kvm_set_pmu_events(u32 set, struct perf_event_attr *attr)
 {
-	struct kvm_host_data *ctx = this_cpu_ptr(&kvm_host_data);
+	struct kvm_host_data *ctx = this_cpu_ptr_hyp_sym(kvm_host_data);
 
-	if (!kvm_pmu_switch_needed(attr))
+	if (!ctx || !kvm_pmu_switch_needed(attr))
 		return;
 
 	if (!attr->exclude_host)
@@ -47,7 +47,10 @@ void kvm_set_pmu_events(u32 set, struct perf_event_attr *attr)
  */
 void kvm_clr_pmu_events(u32 clr)
 {
-	struct kvm_host_data *ctx = this_cpu_ptr(&kvm_host_data);
+	struct kvm_host_data *ctx = this_cpu_ptr_hyp_sym(kvm_host_data);
+
+	if (!ctx)
+		return;
 
 	ctx->pmu_events.events_host &= ~clr;
 	ctx->pmu_events.events_guest &= ~clr;
@@ -173,7 +176,7 @@ void kvm_vcpu_pmu_restore_guest(struct kvm_vcpu *vcpu)
 		return;
 
 	preempt_disable();
-	host = this_cpu_ptr(&kvm_host_data);
+	host = this_cpu_ptr_hyp_sym(kvm_host_data);
 	events_guest = host->pmu_events.events_guest;
 	events_host = host->pmu_events.events_host;
 
@@ -193,7 +196,7 @@ void kvm_vcpu_pmu_restore_host(struct kvm_vcpu *vcpu)
 	if (!has_vhe())
 		return;
 
-	host = this_cpu_ptr(&kvm_host_data);
+	host = this_cpu_ptr_hyp_sym(kvm_host_data);
 	events_guest = host->pmu_events.events_guest;
 	events_host = host->pmu_events.events_host;
 

arch/arm64/kvm/psci.c

@@ -425,27 +425,30 @@ static int get_kernel_wa_level(u64 regid)
 {
 	switch (regid) {
 	case KVM_REG_ARM_SMCCC_ARCH_WORKAROUND_1:
-		switch (kvm_arm_harden_branch_predictor()) {
-		case KVM_BP_HARDEN_UNKNOWN:
+		switch (arm64_get_spectre_v2_state()) {
+		case SPECTRE_VULNERABLE:
 			return KVM_REG_ARM_SMCCC_ARCH_WORKAROUND_1_NOT_AVAIL;
-		case KVM_BP_HARDEN_WA_NEEDED:
+		case SPECTRE_MITIGATED:
 			return KVM_REG_ARM_SMCCC_ARCH_WORKAROUND_1_AVAIL;
-		case KVM_BP_HARDEN_NOT_REQUIRED:
+		case SPECTRE_UNAFFECTED:
 			return KVM_REG_ARM_SMCCC_ARCH_WORKAROUND_1_NOT_REQUIRED;
 		}
 		return KVM_REG_ARM_SMCCC_ARCH_WORKAROUND_1_NOT_AVAIL;
 	case KVM_REG_ARM_SMCCC_ARCH_WORKAROUND_2:
-		switch (kvm_arm_have_ssbd()) {
-		case KVM_SSBD_FORCE_DISABLE:
+		switch (arm64_get_spectre_v4_state()) {
+		case SPECTRE_MITIGATED:
+			/*
+			 * As for the hypercall discovery, we pretend we
+			 * don't have any FW mitigation if SSBS is there at
+			 * all times.
+			 */
+			if (cpus_have_final_cap(ARM64_SSBS))
 				return KVM_REG_ARM_SMCCC_ARCH_WORKAROUND_2_NOT_AVAIL;
-		case KVM_SSBD_KERNEL:
-			return KVM_REG_ARM_SMCCC_ARCH_WORKAROUND_2_AVAIL;
-		case KVM_SSBD_FORCE_ENABLE:
-		case KVM_SSBD_MITIGATED:
+			fallthrough;
+		case SPECTRE_UNAFFECTED:
 			return KVM_REG_ARM_SMCCC_ARCH_WORKAROUND_2_NOT_REQUIRED;
-		case KVM_SSBD_UNKNOWN:
-		default:
-			return KVM_REG_ARM_SMCCC_ARCH_WORKAROUND_2_UNKNOWN;
+		case SPECTRE_VULNERABLE:
+			return KVM_REG_ARM_SMCCC_ARCH_WORKAROUND_2_NOT_AVAIL;
 		}
 	}
 
@@ -462,14 +465,8 @@ int kvm_arm_get_fw_reg(struct kvm_vcpu *vcpu, const struct kvm_one_reg *reg)
 		val = kvm_psci_version(vcpu, vcpu->kvm);
 		break;
 	case KVM_REG_ARM_SMCCC_ARCH_WORKAROUND_1:
-		val = get_kernel_wa_level(reg->id) & KVM_REG_FEATURE_LEVEL_MASK;
-		break;
 	case KVM_REG_ARM_SMCCC_ARCH_WORKAROUND_2:
 		val = get_kernel_wa_level(reg->id) & KVM_REG_FEATURE_LEVEL_MASK;
-
-		if (val == KVM_REG_ARM_SMCCC_ARCH_WORKAROUND_2_AVAIL &&
-		    kvm_arm_get_vcpu_workaround_2_flag(vcpu))
-			val |= KVM_REG_ARM_SMCCC_ARCH_WORKAROUND_2_ENABLED;
 		break;
 	default:
 		return -ENOENT;
@@ -527,34 +524,35 @@ int kvm_arm_set_fw_reg(struct kvm_vcpu *vcpu, const struct kvm_one_reg *reg)
 			    KVM_REG_ARM_SMCCC_ARCH_WORKAROUND_2_ENABLED))
 			return -EINVAL;
 
-		wa_level = val & KVM_REG_FEATURE_LEVEL_MASK;
-
-		if (get_kernel_wa_level(reg->id) < wa_level)
-			return -EINVAL;
-
 		/* The enabled bit must not be set unless the level is AVAIL. */
-		if (wa_level != KVM_REG_ARM_SMCCC_ARCH_WORKAROUND_2_AVAIL &&
-		    wa_level != val)
+		if ((val & KVM_REG_ARM_SMCCC_ARCH_WORKAROUND_2_ENABLED) &&
+		    (val & KVM_REG_FEATURE_LEVEL_MASK) != KVM_REG_ARM_SMCCC_ARCH_WORKAROUND_2_AVAIL)
 			return -EINVAL;
 
-		/* Are we finished or do we need to check the enable bit ? */
-		if (kvm_arm_have_ssbd() != KVM_SSBD_KERNEL)
-			return 0;
-
 		/*
-		 * If this kernel supports the workaround to be switched on
-		 * or off, make sure it matches the requested setting.
+		 * Map all the possible incoming states to the only two we
+		 * really want to deal with.
 		 */
-		switch (wa_level) {
-		case KVM_REG_ARM_SMCCC_ARCH_WORKAROUND_2_AVAIL:
-			kvm_arm_set_vcpu_workaround_2_flag(vcpu,
-			    val & KVM_REG_ARM_SMCCC_ARCH_WORKAROUND_2_ENABLED);
+		switch (val & KVM_REG_FEATURE_LEVEL_MASK) {
+		case KVM_REG_ARM_SMCCC_ARCH_WORKAROUND_2_NOT_AVAIL:
+		case KVM_REG_ARM_SMCCC_ARCH_WORKAROUND_2_UNKNOWN:
+			wa_level = KVM_REG_ARM_SMCCC_ARCH_WORKAROUND_2_NOT_AVAIL;
 			break;
+		case KVM_REG_ARM_SMCCC_ARCH_WORKAROUND_2_AVAIL:
 		case KVM_REG_ARM_SMCCC_ARCH_WORKAROUND_2_NOT_REQUIRED:
-			kvm_arm_set_vcpu_workaround_2_flag(vcpu, true);
+			wa_level = KVM_REG_ARM_SMCCC_ARCH_WORKAROUND_2_NOT_REQUIRED;
 			break;
+		default:
+			return -EINVAL;
 		}
 
+		/*
+		 * We can deal with NOT_AVAIL on NOT_REQUIRED, but not the
+		 * other way around.
+		 */
+		if (get_kernel_wa_level(reg->id) < wa_level)
+			return -EINVAL;
+
 		return 0;
 	default:
 		return -ENOENT;

arch/arm64/kvm/reset.c

@@ -319,10 +319,6 @@ int kvm_reset_vcpu(struct kvm_vcpu *vcpu)
 		vcpu->arch.reset_state.reset = false;
 	}
 
-	/* Default workaround setup is enabled (if supported) */
-	if (kvm_arm_have_ssbd() == KVM_SSBD_KERNEL)
-		vcpu->arch.workaround_flags |= VCPU_WORKAROUND_2_FLAG;
-
 	/* Reset timer */
 	ret = kvm_timer_vcpu_reset(vcpu);
 out:

arch/arm64/kvm/sys_regs.c

@@ -1128,6 +1128,9 @@ static u64 read_id_reg(const struct kvm_vcpu *vcpu,
 		if (!vcpu_has_sve(vcpu))
 			val &= ~(0xfUL << ID_AA64PFR0_SVE_SHIFT);
 		val &= ~(0xfUL << ID_AA64PFR0_AMU_SHIFT);
+		if (!(val & (0xfUL << ID_AA64PFR0_CSV2_SHIFT)) &&
+		    arm64_get_spectre_v2_state() == SPECTRE_UNAFFECTED)
+			val |= (1UL << ID_AA64PFR0_CSV2_SHIFT);
 	} else if (id == SYS_ID_AA64ISAR1_EL1 && !vcpu_has_ptrauth(vcpu)) {
 		val &= ~((0xfUL << ID_AA64ISAR1_APA_SHIFT) |
 			 (0xfUL << ID_AA64ISAR1_API_SHIFT) |