Merge branch 'kvm-arm64/nv-xnx-haf' into kvmarm/next

* kvm-arm64/nv-xnx-haf: (22 commits)
: Support for FEAT_XNX and FEAT_HAF in nested
:
: Add support for a couple of MMU-related features that weren't
: implemented by KVM's software page table walk:
:
: - FEAT_XNX: Allows the hypervisor to describe execute permissions
: separately for EL0 and EL1
:
: - FEAT_HAF: Hardware update of the Access Flag, which in the context of
: nested means software walkers must also set the Access Flag.
:
: The series also adds some basic support for testing KVM's emulation of
: the AT instruction, including the implementation detail that AT sets the
: Access Flag in KVM.
KVM: arm64: at: Update AF on software walk only if VM has FEAT_HAFDBS
KVM: arm64: at: Use correct HA bit in TCR_EL2 when regime is EL2
KVM: arm64: Document KVM_PGTABLE_PROT_{UX,PX}
KVM: arm64: Fix spelling mistake "Unexpeced" -> "Unexpected"
KVM: arm64: Add break to default case in kvm_pgtable_stage2_pte_prot()
KVM: arm64: Add endian casting to kvm_swap_s[12]_desc()
KVM: arm64: Fix compilation when CONFIG_ARM64_USE_LSE_ATOMICS=n
KVM: arm64: selftests: Add test for AT emulation
KVM: arm64: nv: Expose hardware access flag management to NV guests
KVM: arm64: nv: Implement HW access flag management in stage-2 SW PTW
KVM: arm64: Implement HW access flag management in stage-1 SW PTW
KVM: arm64: Propagate PTW errors up to AT emulation
KVM: arm64: Add helper for swapping guest descriptor
KVM: arm64: nv: Use pgtable definitions in stage-2 walk
KVM: arm64: Handle endianness in read helper for emulated PTW
KVM: arm64: nv: Stop passing vCPU through void ptr in S2 PTW
KVM: arm64: Call helper for reading descriptors directly
KVM: arm64: nv: Advertise support for FEAT_XNX
KVM: arm64: Teach ptdump about FEAT_XNX permissions
KVM: arm64: nv: Forward FEAT_XNX permissions to the shadow stage-2
...

Signed-off-by: Oliver Upton <[e-mail]>

arch/arm64/include/asm/kvm_arm.h

···
 #define TCR_EL2_DS		(1UL << 32)
 #define TCR_EL2_RES1		((1U << 31) | (1 << 23))
 #define TCR_EL2_HPD		(1 << 24)
+#define TCR_EL2_HA		(1 << 21)
 #define TCR_EL2_TBI		(1 << 20)
 #define TCR_EL2_PS_SHIFT	16
 #define TCR_EL2_PS_MASK		(7 << TCR_EL2_PS_SHIFT)

arch/arm64/include/asm/kvm_asm.h

···
 extern int __kvm_tlbi_s1e2(struct kvm_s2_mmu *mmu, u64 va, u64 sys_encoding);
 
 extern void __kvm_timer_set_cntvoff(u64 cntvoff);
-extern void __kvm_at_s1e01(struct kvm_vcpu *vcpu, u32 op, u64 vaddr);
-extern void __kvm_at_s1e2(struct kvm_vcpu *vcpu, u32 op, u64 vaddr);
-extern void __kvm_at_s12(struct kvm_vcpu *vcpu, u32 op, u64 vaddr);
+extern int __kvm_at_s1e01(struct kvm_vcpu *vcpu, u32 op, u64 vaddr);
+extern int __kvm_at_s1e2(struct kvm_vcpu *vcpu, u32 op, u64 vaddr);
+extern int __kvm_at_s12(struct kvm_vcpu *vcpu, u32 op, u64 vaddr);
 
 extern int __kvm_vcpu_run(struct kvm_vcpu *vcpu);
 

arch/arm64/include/asm/kvm_nested.h

···
 	return trans->writable;
 }
 
-static inline bool kvm_s2_trans_executable(struct kvm_s2_trans *trans)
+static inline bool kvm_has_xnx(struct kvm *kvm)
 {
-	return !(trans->desc & BIT(54));
+	return cpus_have_final_cap(ARM64_HAS_XNX) &&
+		kvm_has_feat(kvm, ID_AA64MMFR1_EL1, XNX, IMP);
+}
+
+static inline bool kvm_s2_trans_exec_el0(struct kvm *kvm, struct kvm_s2_trans *trans)
+{
+	u8 xn = FIELD_GET(KVM_PTE_LEAF_ATTR_HI_S2_XN, trans->desc);
+
+	if (!kvm_has_xnx(kvm))
+		xn &= FIELD_PREP(KVM_PTE_LEAF_ATTR_HI_S2_XN, 0b10);
+
+	switch (xn) {
+	case 0b00:
+	case 0b01:
+		return true;
+	default:
+		return false;
+	}
+}
+
+static inline bool kvm_s2_trans_exec_el1(struct kvm *kvm, struct kvm_s2_trans *trans)
+{
+	u8 xn = FIELD_GET(KVM_PTE_LEAF_ATTR_HI_S2_XN, trans->desc);
+
+	if (!kvm_has_xnx(kvm))
+		xn &= FIELD_PREP(KVM_PTE_LEAF_ATTR_HI_S2_XN, 0b10);
+
+	switch (xn) {
+	case 0b00:
+	case 0b11:
+		return true;
+	default:
+		return false;
+	}
 }
 
 extern int kvm_walk_nested_s2(struct kvm_vcpu *vcpu, phys_addr_t gipa,
···
 	bool	     		be;
 	bool	     		s2;
 	bool			pa52bit;
+	bool			ha;
 };
 
 struct s1_walk_result {
···
 		BUG_ON(__c >= NR_CPUS);				\
 		(FIX_VNCR - __c);				\
 	})
+
+int __kvm_at_swap_desc(struct kvm *kvm, gpa_t ipa, u64 old, u64 new);
 
 #endif /* __ARM64_KVM_NESTED_H */

arch/arm64/include/asm/kvm_pgtable.h

···
 
 #define KVM_PTE_LEAF_ATTR_HI_S1_XN	BIT(54)
 
-#define KVM_PTE_LEAF_ATTR_HI_S2_XN	BIT(54)
+#define KVM_PTE_LEAF_ATTR_HI_S2_XN	GENMASK(54, 53)
 
 #define KVM_PTE_LEAF_ATTR_HI_S1_GP	BIT(50)
 
···
 
 /**
  * enum kvm_pgtable_prot - Page-table permissions and attributes.
- * @KVM_PGTABLE_PROT_X:		Execute permission.
+ * @KVM_PGTABLE_PROT_UX:	Unprivileged execute permission.
+ * @KVM_PGTABLE_PROT_PX:	Privileged execute permission.
+ * @KVM_PGTABLE_PROT_X:		Privileged and unprivileged execute permission.
  * @KVM_PGTABLE_PROT_W:		Write permission.
  * @KVM_PGTABLE_PROT_R:		Read permission.
  * @KVM_PGTABLE_PROT_DEVICE:	Device attributes.
···
  * @KVM_PGTABLE_PROT_SW3:	Software bit 3.
  */
 enum kvm_pgtable_prot {
-	KVM_PGTABLE_PROT_X			= BIT(0),
-	KVM_PGTABLE_PROT_W			= BIT(1),
-	KVM_PGTABLE_PROT_R			= BIT(2),
+	KVM_PGTABLE_PROT_PX			= BIT(0),
+	KVM_PGTABLE_PROT_UX			= BIT(1),
+	KVM_PGTABLE_PROT_X			= KVM_PGTABLE_PROT_PX	|
+						  KVM_PGTABLE_PROT_UX,
+	KVM_PGTABLE_PROT_W			= BIT(2),
+	KVM_PGTABLE_PROT_R			= BIT(3),
 
-	KVM_PGTABLE_PROT_DEVICE			= BIT(3),
-	KVM_PGTABLE_PROT_NORMAL_NC		= BIT(4),
+	KVM_PGTABLE_PROT_DEVICE			= BIT(4),
+	KVM_PGTABLE_PROT_NORMAL_NC		= BIT(5),
 
 	KVM_PGTABLE_PROT_SW0			= BIT(55),
 	KVM_PGTABLE_PROT_SW1			= BIT(56),

arch/arm64/kernel/cpufeature.c

···
 		.capability = ARM64_HAS_GICV5_LEGACY,
 		.matches = test_has_gicv5_legacy,
 	},
+	{
+		.desc = "XNX",
+		.capability = ARM64_HAS_XNX,
+		.type = ARM64_CPUCAP_SYSTEM_FEATURE,
+		.matches = has_cpuid_feature,
+		ARM64_CPUID_FIELDS(ID_AA64MMFR1_EL1, XNX, IMP)
+	},
 	{},
 };
 

arch/arm64/kvm/at.c

···
 
 	wi->baddr &= GENMASK_ULL(wi->max_oa_bits - 1, x);
 
+	wi->ha  = kvm_has_feat(vcpu->kvm, ID_AA64MMFR1_EL1, HAFDBS, AF);
+	wi->ha &= (wi->regime == TR_EL2 ?
+		  FIELD_GET(TCR_EL2_HA, tcr) :
+		  FIELD_GET(TCR_HA, tcr));
+
 	return 0;
 
 addrsz:
···
 	return -EFAULT;
 }
 
+static int kvm_read_s1_desc(struct kvm_vcpu *vcpu, u64 pa, u64 *desc,
+			    struct s1_walk_info *wi)
+{
+	u64 val;
+	int r;
+
+	r = kvm_read_guest(vcpu->kvm, pa, &val, sizeof(val));
+	if (r)
+		return r;
+
+	if (wi->be)
+		*desc = be64_to_cpu((__force __be64)val);
+	else
+		*desc = le64_to_cpu((__force __le64)val);
+
+	return 0;
+}
+
+static int kvm_swap_s1_desc(struct kvm_vcpu *vcpu, u64 pa, u64 old, u64 new,
+			    struct s1_walk_info *wi)
+{
+	if (wi->be) {
+		old = (__force u64)cpu_to_be64(old);
+		new = (__force u64)cpu_to_be64(new);
+	} else {
+		old = (__force u64)cpu_to_le64(old);
+		new = (__force u64)cpu_to_le64(new);
+	}
+
+	return __kvm_at_swap_desc(vcpu->kvm, pa, old, new);
+}
+
 static int walk_s1(struct kvm_vcpu *vcpu, struct s1_walk_info *wi,
 		   struct s1_walk_result *wr, u64 va)
 {
-	u64 va_top, va_bottom, baddr, desc;
+	u64 va_top, va_bottom, baddr, desc, new_desc, ipa;
 	int level, stride, ret;
 
 	level = wi->sl;
···
 	va_top = get_ia_size(wi) - 1;
 
 	while (1) {
-		u64 index, ipa;
+		u64 index;
 
 		va_bottom = (3 - level) * stride + wi->pgshift;
 		index = (va & GENMASK_ULL(va_top, va_bottom)) >> (va_bottom - 3);
···
 				return ret;
 		}
 
-		ret = kvm_read_guest(vcpu->kvm, ipa, &desc, sizeof(desc));
+		ret = kvm_read_s1_desc(vcpu, ipa, &desc, wi);
 		if (ret) {
 			fail_s1_walk(wr, ESR_ELx_FSC_SEA_TTW(level), false);
 			return ret;
 		}
 
-		if (wi->be)
-			desc = be64_to_cpu((__force __be64)desc);
-		else
-			desc = le64_to_cpu((__force __le64)desc);
+		new_desc = desc;
 
 		/* Invalid descriptor */
 		if (!(desc & BIT(0)))
···
 	baddr = desc_to_oa(wi, desc);
 	if (check_output_size(baddr & GENMASK(52, va_bottom), wi))
 		goto addrsz;
+
+	if (wi->ha)
+		new_desc |= PTE_AF;
+
+	if (new_desc != desc) {
+		ret = kvm_swap_s1_desc(vcpu, ipa, desc, new_desc, wi);
+		if (ret)
+			return ret;
+
+		desc = new_desc;
+	}
 
 	if (!(desc & PTE_AF)) {
 		fail_s1_walk(wr, ESR_ELx_FSC_ACCESS_L(level), false);
···
 	wr->pr &= !pan;
 }
 
-static u64 handle_at_slow(struct kvm_vcpu *vcpu, u32 op, u64 vaddr)
+static int handle_at_slow(struct kvm_vcpu *vcpu, u32 op, u64 vaddr, u64 *par)
 {
 	struct s1_walk_result wr = {};
 	struct s1_walk_info wi = {};
···
 
 	srcu_read_unlock(&vcpu->kvm->srcu, idx);
 
+	/*
+	 * Race to update a descriptor -- restart the walk.
+	 */
+	if (ret == -EAGAIN)
+		return ret;
 	if (ret)
 		goto compute_par;
 
···
 		fail_s1_walk(&wr, ESR_ELx_FSC_PERM_L(wr.level), false);
 
 compute_par:
-	return compute_par_s1(vcpu, &wi, &wr);
+	*par = compute_par_s1(vcpu, &wi, &wr);
+	return 0;
 }
 
 /*
···
 		 !(par & SYS_PAR_EL1_S));
 }
 
-void __kvm_at_s1e01(struct kvm_vcpu *vcpu, u32 op, u64 vaddr)
+int __kvm_at_s1e01(struct kvm_vcpu *vcpu, u32 op, u64 vaddr)
 {
 	u64 par = __kvm_at_s1e01_fast(vcpu, op, vaddr);
+	int ret;
 
 	/*
 	 * If PAR_EL1 reports that AT failed on a S1 permission or access
···
 	 */
 	if ((par & SYS_PAR_EL1_F) &&
 	    !par_check_s1_perm_fault(par) &&
-	    !par_check_s1_access_fault(par))
-		par = handle_at_slow(vcpu, op, vaddr);
+	    !par_check_s1_access_fault(par)) {
+		ret = handle_at_slow(vcpu, op, vaddr, &par);
+		if (ret)
+			return ret;
+	}
 
 	vcpu_write_sys_reg(vcpu, par, PAR_EL1);
+	return 0;
 }
 
-void __kvm_at_s1e2(struct kvm_vcpu *vcpu, u32 op, u64 vaddr)
+int __kvm_at_s1e2(struct kvm_vcpu *vcpu, u32 op, u64 vaddr)
 {
 	u64 par;
+	int ret;
 
 	/*
 	 * We've trapped, so everything is live on the CPU. As we will be
···
 	}
 
 	/* We failed the translation, let's replay it in slow motion */
-	if ((par & SYS_PAR_EL1_F) && !par_check_s1_perm_fault(par))
-		par = handle_at_slow(vcpu, op, vaddr);
+	if ((par & SYS_PAR_EL1_F) && !par_check_s1_perm_fault(par)) {
+		ret = handle_at_slow(vcpu, op, vaddr, &par);
+		if (ret)
+			return ret;
+	}
 
 	vcpu_write_sys_reg(vcpu, par, PAR_EL1);
+	return 0;
 }
 
-void __kvm_at_s12(struct kvm_vcpu *vcpu, u32 op, u64 vaddr)
+int __kvm_at_s12(struct kvm_vcpu *vcpu, u32 op, u64 vaddr)
 {
 	struct kvm_s2_trans out = {};
 	u64 ipa, par;
···
 		break;
 	default:
 		WARN_ON_ONCE(1);
-		return;
+		return 0;
 	}
 
 	__kvm_at_s1e01(vcpu, op, vaddr);
 	par = vcpu_read_sys_reg(vcpu, PAR_EL1);
 	if (par & SYS_PAR_EL1_F)
-		return;
+		return 0;
 
 	/*
 	 * If we only have a single stage of translation (EL2&0), exit
···
 	 */
 	if (compute_translation_regime(vcpu, op) == TR_EL20 ||
 	    !(vcpu_read_sys_reg(vcpu, HCR_EL2) & (HCR_VM | HCR_DC)))
-		return;
+		return 0;
 
 	/* Do the stage-2 translation */
 	ipa = (par & GENMASK_ULL(47, 12)) | (vaddr & GENMASK_ULL(11, 0));
 	out.esr = 0;
 	ret = kvm_walk_nested_s2(vcpu, ipa, &out);
 	if (ret < 0)
-		return;
+		return ret;
 
 	/* Check the access permission */
 	if (!out.esr &&
···
 
 	par = compute_par_s12(vcpu, par, &out);
 	vcpu_write_sys_reg(vcpu, par, PAR_EL1);
+	return 0;
 }
 
 /*
···
 		/* Any other error... */
 		return ret;
 	}
+}
+
+#ifdef CONFIG_ARM64_LSE_ATOMICS
+static int __lse_swap_desc(u64 __user *ptep, u64 old, u64 new)
+{
+	u64 tmp = old;
+	int ret = 0;
+
+	uaccess_enable_privileged();
+
+	asm volatile(__LSE_PREAMBLE
+		     "1: cas	%[old], %[new], %[addr]\n"
+		     "2:\n"
+		     _ASM_EXTABLE_UACCESS_ERR(1b, 2b, %w[ret])
+		     : [old] "+r" (old), [addr] "+Q" (*ptep), [ret] "+r" (ret)
+		     : [new] "r" (new)
+		     : "memory");
+
+	uaccess_disable_privileged();
+
+	if (ret)
+		return ret;
+	if (tmp != old)
+		return -EAGAIN;
+
+	return ret;
+}
+#else
+static int __lse_swap_desc(u64 __user *ptep, u64 old, u64 new)
+{
+	return -EINVAL;
+}
+#endif
+
+static int __llsc_swap_desc(u64 __user *ptep, u64 old, u64 new)
+{
+	int ret = 1;
+	u64 tmp;
+
+	uaccess_enable_privileged();
+
+	asm volatile("prfm	pstl1strm, %[addr]\n"
+		     "1: ldxr	%[tmp], %[addr]\n"
+		     "sub	%[tmp], %[tmp], %[old]\n"
+		     "cbnz	%[tmp], 3f\n"
+		     "2: stlxr	%w[ret], %[new], %[addr]\n"
+		     "3:\n"
+		     _ASM_EXTABLE_UACCESS_ERR(1b, 3b, %w[ret])
+		     _ASM_EXTABLE_UACCESS_ERR(2b, 3b, %w[ret])
+		     : [ret] "+r" (ret), [addr] "+Q" (*ptep), [tmp] "=&r" (tmp)
+		     : [old] "r" (old), [new] "r" (new)
+		     : "memory");
+
+	uaccess_disable_privileged();
+
+	/* STLXR didn't update the descriptor, or the compare failed */
+	if (ret == 1)
+		return -EAGAIN;
+
+	return ret;
+}
+
+int __kvm_at_swap_desc(struct kvm *kvm, gpa_t ipa, u64 old, u64 new)
+{
+	struct kvm_memory_slot *slot;
+	unsigned long hva;
+	u64 __user *ptep;
+	bool writable;
+	int offset;
+	gfn_t gfn;
+	int r;
+
+	lockdep_assert(srcu_read_lock_held(&kvm->srcu));
+
+	gfn = ipa >> PAGE_SHIFT;
+	offset = offset_in_page(ipa);
+	slot = gfn_to_memslot(kvm, gfn);
+	hva = gfn_to_hva_memslot_prot(slot, gfn, &writable);
+	if (kvm_is_error_hva(hva))
+		return -EINVAL;
+	if (!writable)
+		return -EPERM;
+
+	ptep = (u64 __user *)hva + offset;
+	if (cpus_have_final_cap(ARM64_HAS_LSE_ATOMICS))
+		r = __lse_swap_desc(ptep, old, new);
+	else
+		r = __llsc_swap_desc(ptep, old, new);
+
+	if (r < 0)
+		return r;
+
+	mark_page_dirty_in_slot(kvm, slot, gfn);
+	return 0;
 }

arch/arm64/kvm/hyp/pgtable.c

···
 
 #define KVM_S2_MEMATTR(pgt, attr) PAGE_S2_MEMATTR(attr, stage2_has_fwb(pgt))
 
+static int stage2_set_xn_attr(enum kvm_pgtable_prot prot, kvm_pte_t *attr)
+{
+	bool px, ux;
+	u8 xn;
+
+	px = prot & KVM_PGTABLE_PROT_PX;
+	ux = prot & KVM_PGTABLE_PROT_UX;
+
+	if (!cpus_have_final_cap(ARM64_HAS_XNX) && px != ux)
+		return -EINVAL;
+
+	if (px && ux)
+		xn = 0b00;
+	else if (!px && ux)
+		xn = 0b01;
+	else if (!px && !ux)
+		xn = 0b10;
+	else
+		xn = 0b11;
+
+	*attr &= ~KVM_PTE_LEAF_ATTR_HI_S2_XN;
+	*attr |= FIELD_PREP(KVM_PTE_LEAF_ATTR_HI_S2_XN, xn);
+	return 0;
+}
+
 static int stage2_set_prot_attr(struct kvm_pgtable *pgt, enum kvm_pgtable_prot prot,
 				kvm_pte_t *ptep)
 {
 	kvm_pte_t attr;
 	u32 sh = KVM_PTE_LEAF_ATTR_LO_S2_SH_IS;
+	int r;
 
 	switch (prot & (KVM_PGTABLE_PROT_DEVICE |
 			KVM_PGTABLE_PROT_NORMAL_NC)) {
···
 		attr = KVM_S2_MEMATTR(pgt, NORMAL);
 	}
 
-	if (!(prot & KVM_PGTABLE_PROT_X))
-		attr |= KVM_PTE_LEAF_ATTR_HI_S2_XN;
+	r = stage2_set_xn_attr(prot, &attr);
+	if (r)
+		return r;
 
 	if (prot & KVM_PGTABLE_PROT_R)
 		attr |= KVM_PTE_LEAF_ATTR_LO_S2_S2AP_R;
···
 		prot |= KVM_PGTABLE_PROT_R;
 	if (pte & KVM_PTE_LEAF_ATTR_LO_S2_S2AP_W)
 		prot |= KVM_PGTABLE_PROT_W;
-	if (!(pte & KVM_PTE_LEAF_ATTR_HI_S2_XN))
-		prot |= KVM_PGTABLE_PROT_X;
+
+	switch (FIELD_GET(KVM_PTE_LEAF_ATTR_HI_S2_XN, pte)) {
+	case 0b00:
+		prot |= KVM_PGTABLE_PROT_PX | KVM_PGTABLE_PROT_UX;
+		break;
+	case 0b01:
+		prot |= KVM_PGTABLE_PROT_UX;
+		break;
+	case 0b11:
+		prot |= KVM_PGTABLE_PROT_PX;
+		break;
+	default:
+		break;
+	}
 
 	return prot;
 }
···
 int kvm_pgtable_stage2_relax_perms(struct kvm_pgtable *pgt, u64 addr,
 				   enum kvm_pgtable_prot prot, enum kvm_pgtable_walk_flags flags)
 {
-	int ret;
+	kvm_pte_t xn = 0, set = 0, clr = 0;
 	s8 level;
-	kvm_pte_t set = 0, clr = 0;
+	int ret;
 
 	if (prot & KVM_PTE_LEAF_ATTR_HI_SW)
 		return -EINVAL;
···
 	if (prot & KVM_PGTABLE_PROT_W)
 		set |= KVM_PTE_LEAF_ATTR_LO_S2_S2AP_W;
 
-	if (prot & KVM_PGTABLE_PROT_X)
-		clr |= KVM_PTE_LEAF_ATTR_HI_S2_XN;
+	ret = stage2_set_xn_attr(prot, &xn);
+	if (ret)
+		return ret;
+
+	set |= xn & KVM_PTE_LEAF_ATTR_HI_S2_XN;
+	clr |= ~xn & KVM_PTE_LEAF_ATTR_HI_S2_XN;
 
 	ret = stage2_update_leaf_attrs(pgt, addr, 1, set, clr, NULL, &level, flags);
 	if (!ret || ret == -EAGAIN)

arch/arm64/kvm/mmu.c

···
 	*prot |= kvm_encode_nested_level(nested);
 }
 
+static void adjust_nested_exec_perms(struct kvm *kvm,
+				     struct kvm_s2_trans *nested,
+				     enum kvm_pgtable_prot *prot)
+{
+	if (!kvm_s2_trans_exec_el0(kvm, nested))
+		*prot &= ~KVM_PGTABLE_PROT_UX;
+	if (!kvm_s2_trans_exec_el1(kvm, nested))
+		*prot &= ~KVM_PGTABLE_PROT_PX;
+}
+
 #define KVM_PGTABLE_WALK_MEMABORT_FLAGS (KVM_PGTABLE_WALK_HANDLE_FAULT | KVM_PGTABLE_WALK_SHARED)
 
 static int gmem_abort(struct kvm_vcpu *vcpu, phys_addr_t fault_ipa,
···
 	if (writable)
 		prot |= KVM_PGTABLE_PROT_W;
 
-	if (exec_fault ||
-	    (cpus_have_final_cap(ARM64_HAS_CACHE_DIC) &&
-	     (!nested || kvm_s2_trans_executable(nested))))
+	if (exec_fault || cpus_have_final_cap(ARM64_HAS_CACHE_DIC))
 		prot |= KVM_PGTABLE_PROT_X;
+
+	if (nested)
+		adjust_nested_exec_perms(kvm, nested, &prot);
 
 	kvm_fault_lock(kvm);
 	if (mmu_invalidate_retry(kvm, mmu_seq)) {
···
 			prot |= KVM_PGTABLE_PROT_NORMAL_NC;
 		else
 			prot |= KVM_PGTABLE_PROT_DEVICE;
-	} else if (cpus_have_final_cap(ARM64_HAS_CACHE_DIC) &&
-		   (!nested || kvm_s2_trans_executable(nested))) {
+	} else if (cpus_have_final_cap(ARM64_HAS_CACHE_DIC)) {
 		prot |= KVM_PGTABLE_PROT_X;
 	}
+
+	if (nested)
+		adjust_nested_exec_perms(kvm, nested, &prot);
 
 	/*
 	 * Under the premise of getting a FSC_PERM fault, we just need to relax
···
 		u32 esr;
 
 		ret = kvm_walk_nested_s2(vcpu, fault_ipa, &nested_trans);
+		if (ret == -EAGAIN) {
+			ret = 1;
+			goto out_unlock;
+		}
+
 		if (ret) {
 			esr = kvm_s2_trans_esr(&nested_trans);
 			kvm_inject_s2_fault(vcpu, esr);

arch/arm64/kvm/nested.c

···
 }
 
 struct s2_walk_info {
-	int	     (*read_desc)(phys_addr_t pa, u64 *desc, void *data);
-	void	     *data;
-	u64	     baddr;
-	unsigned int max_oa_bits;
-	unsigned int pgshift;
-	unsigned int sl;
-	unsigned int t0sz;
-	bool	     be;
+	u64		baddr;
+	unsigned int	max_oa_bits;
+	unsigned int	pgshift;
+	unsigned int	sl;
+	unsigned int	t0sz;
+	bool		be;
+	bool		ha;
 };
 
 static u32 compute_fsc(int level, u32 fsc)
···
 	return 0;
 }
 
+static int read_guest_s2_desc(struct kvm_vcpu *vcpu, phys_addr_t pa, u64 *desc,
+			      struct s2_walk_info *wi)
+{
+	u64 val;
+	int r;
+
+	r = kvm_read_guest(vcpu->kvm, pa, &val, sizeof(val));
+	if (r)
+		return r;
+
+	/*
+	 * Handle reversedescriptors if endianness differs between the
+	 * host and the guest hypervisor.
+	 */
+	if (wi->be)
+		*desc = be64_to_cpu((__force __be64)val);
+	else
+		*desc = le64_to_cpu((__force __le64)val);
+
+	return 0;
+}
+
+static int swap_guest_s2_desc(struct kvm_vcpu *vcpu, phys_addr_t pa, u64 old, u64 new,
+			      struct s2_walk_info *wi)
+{
+	if (wi->be) {
+		old = (__force u64)cpu_to_be64(old);
+		new = (__force u64)cpu_to_be64(new);
+	} else {
+		old = (__force u64)cpu_to_le64(old);
+		new = (__force u64)cpu_to_le64(new);
+	}
+
+	return __kvm_at_swap_desc(vcpu->kvm, pa, old, new);
+}
+
 /*
  * This is essentially a C-version of the pseudo code from the ARM ARM
  * AArch64.TranslationTableWalk  function.  I strongly recommend looking at
···
  *
  * Must be called with the kvm->srcu read lock held
  */
-static int walk_nested_s2_pgd(phys_addr_t ipa,
+static int walk_nested_s2_pgd(struct kvm_vcpu *vcpu, phys_addr_t ipa,
 			      struct s2_walk_info *wi, struct kvm_s2_trans *out)
 {
 	int first_block_level, level, stride, input_size, base_lower_bound;
 	phys_addr_t base_addr;
 	unsigned int addr_top, addr_bottom;
-	u64 desc;  /* page table entry */
+	u64 desc, new_desc;  /* page table entry */
 	int ret;
 	phys_addr_t paddr;
 
···
 			>> (addr_bottom - 3);
 
 		paddr = base_addr | index;
-		ret = wi->read_desc(paddr, &desc, wi->data);
+		ret = read_guest_s2_desc(vcpu, paddr, &desc, wi);
 		if (ret < 0)
 			return ret;
 
-		/*
-		 * Handle reversedescriptors if endianness differs between the
-		 * host and the guest hypervisor.
-		 */
-		if (wi->be)
-			desc = be64_to_cpu((__force __be64)desc);
-		else
-			desc = le64_to_cpu((__force __le64)desc);
+		new_desc = desc;
 
 		/* Check for valid descriptor at this point */
-		if (!(desc & 1) || ((desc & 3) == 1 && level == 3)) {
+		if (!(desc & KVM_PTE_VALID)) {
 			out->esr = compute_fsc(level, ESR_ELx_FSC_FAULT);
 			out->desc = desc;
 			return 1;
 		}
 
-		/* We're at the final level or block translation level */
-		if ((desc & 3) == 1 || level == 3)
+		if (FIELD_GET(KVM_PTE_TYPE, desc) == KVM_PTE_TYPE_BLOCK) {
+			if (level < 3)
+				break;
+
+			out->esr = compute_fsc(level, ESR_ELx_FSC_FAULT);
+			out->desc = desc;
+			return 1;
+		}
+
+		/* We're at the final level */
+		if (level == 3)
 			break;
 
 		if (check_output_size(wi, desc)) {
···
 		return 1;
 	}
 
-	if (!(desc & BIT(10))) {
+	if (wi->ha)
+		new_desc |= KVM_PTE_LEAF_ATTR_LO_S2_AF;
+
+	if (new_desc != desc) {
+		ret = swap_guest_s2_desc(vcpu, paddr, desc, new_desc, wi);
+		if (ret)
+			return ret;
+
+		desc = new_desc;
+	}
+
+	if (!(desc & KVM_PTE_LEAF_ATTR_LO_S2_AF)) {
 		out->esr = compute_fsc(level, ESR_ELx_FSC_ACCESS);
 		out->desc = desc;
 		return 1;
···
 		(ipa & GENMASK_ULL(addr_bottom - 1, 0));
 	out->output = paddr;
 	out->block_size = 1UL << ((3 - level) * stride + wi->pgshift);
-	out->readable = desc & (0b01 << 6);
-	out->writable = desc & (0b10 << 6);
+	out->readable = desc & KVM_PTE_LEAF_ATTR_LO_S2_S2AP_R;
+	out->writable = desc & KVM_PTE_LEAF_ATTR_LO_S2_S2AP_W;
 	out->level = level;
 	out->desc = desc;
 	return 0;
-}
-
-static int read_guest_s2_desc(phys_addr_t pa, u64 *desc, void *data)
-{
-	struct kvm_vcpu *vcpu = data;
-
-	return kvm_read_guest(vcpu->kvm, pa, desc, sizeof(*desc));
 }
 
 static void vtcr_to_walk_info(u64 vtcr, struct s2_walk_info *wi)
···
 	/* Global limit for now, should eventually be per-VM */
 	wi->max_oa_bits = min(get_kvm_ipa_limit(),
 			      ps_to_output_size(FIELD_GET(VTCR_EL2_PS_MASK, vtcr), false));
+
+	wi->ha = vtcr & VTCR_EL2_HA;
 }
 
 int kvm_walk_nested_s2(struct kvm_vcpu *vcpu, phys_addr_t gipa,
···
 	if (!vcpu_has_nv(vcpu))
 		return 0;
 
-	wi.read_desc = read_guest_s2_desc;
-	wi.data = vcpu;
 	wi.baddr = vcpu_read_sys_reg(vcpu, VTTBR_EL2);
 
 	vtcr_to_walk_info(vtcr, &wi);
 
 	wi.be = vcpu_read_sys_reg(vcpu, SCTLR_EL2) & SCTLR_ELx_EE;
 
-	ret = walk_nested_s2_pgd(gipa, &wi, result);
+	ret = walk_nested_s2_pgd(vcpu, gipa, &wi, result);
 	if (ret)
 		result->esr |= (kvm_vcpu_get_esr(vcpu) & ~ESR_ELx_FSC);
 
···
 		return 0;
 
 	if (kvm_vcpu_trap_is_iabt(vcpu)) {
-		forward_fault = !kvm_s2_trans_executable(trans);
+		if (vcpu_mode_priv(vcpu))
+			forward_fault = !kvm_s2_trans_exec_el1(vcpu->kvm, trans);
+		else
+			forward_fault = !kvm_s2_trans_exec_el0(vcpu->kvm, trans);
 	} else {
 		bool write_fault = kvm_is_write_fault(vcpu);
 
···
 	case SYS_ID_AA64MMFR1_EL1:
 		val &= ~(ID_AA64MMFR1_EL1_CMOW		|
 			 ID_AA64MMFR1_EL1_nTLBPA	|
-			 ID_AA64MMFR1_EL1_ETS		|
-			 ID_AA64MMFR1_EL1_XNX		|
-			 ID_AA64MMFR1_EL1_HAFDBS);
+			 ID_AA64MMFR1_EL1_ETS);
+
 		/* FEAT_E2H0 implies no VHE */
 		if (test_bit(KVM_ARM_VCPU_HAS_EL2_E2H0, kvm->arch.vcpu_features))
 			val &= ~ID_AA64MMFR1_EL1_VH;
+
+		val = ID_REG_LIMIT_FIELD_ENUM(val, ID_AA64MMFR1_EL1, HAFDBS, AF);
 		break;
 
 	case SYS_ID_AA64MMFR2_EL1:

arch/arm64/kvm/ptdump.c

···
 		.val	= PTE_VALID,
 		.set	= " ",
 		.clear	= "F",
-	}, {
+	},
+	{
 		.mask	= KVM_PTE_LEAF_ATTR_LO_S2_S2AP_R,
 		.val	= KVM_PTE_LEAF_ATTR_LO_S2_S2AP_R,
 		.set	= "R",
 		.clear	= " ",
-	}, {
+	},
+	{
 		.mask	= KVM_PTE_LEAF_ATTR_LO_S2_S2AP_W,
 		.val	= KVM_PTE_LEAF_ATTR_LO_S2_S2AP_W,
 		.set	= "W",
 		.clear	= " ",
-	}, {
+	},
+	{
 		.mask	= KVM_PTE_LEAF_ATTR_HI_S2_XN,
-		.val	= KVM_PTE_LEAF_ATTR_HI_S2_XN,
-		.set	= "NX",
-		.clear	= "x ",
-	}, {
+		.val	= 0b00UL << __bf_shf(KVM_PTE_LEAF_ATTR_HI_S2_XN),
+		.set	= "px ux ",
+	},
+	{
+		.mask	= KVM_PTE_LEAF_ATTR_HI_S2_XN,
+		.val	= 0b01UL << __bf_shf(KVM_PTE_LEAF_ATTR_HI_S2_XN),
+		.set	= "PXNux ",
+	},
+	{
+		.mask	= KVM_PTE_LEAF_ATTR_HI_S2_XN,
+		.val	= 0b10UL << __bf_shf(KVM_PTE_LEAF_ATTR_HI_S2_XN),
+		.set	= "PXNUXN",
+	},
+	{
+		.mask	= KVM_PTE_LEAF_ATTR_HI_S2_XN,
+		.val	= 0b11UL << __bf_shf(KVM_PTE_LEAF_ATTR_HI_S2_XN),
+		.set	= "px UXN",
+	},
+	{
 		.mask	= KVM_PTE_LEAF_ATTR_LO_S2_AF,
 		.val	= KVM_PTE_LEAF_ATTR_LO_S2_AF,
 		.set	= "AF",
 		.clear	= "  ",
-	}, {
+	},
+	{
 		.mask	= PMD_TYPE_MASK,
 		.val	= PMD_TYPE_SECT,
 		.set	= "BLK",

arch/arm64/kvm/sys_regs.c

···
 {
 	u32 op = sys_insn(p->Op0, p->Op1, p->CRn, p->CRm, p->Op2);
 
-	__kvm_at_s1e01(vcpu, op, p->regval);
+	if (__kvm_at_s1e01(vcpu, op, p->regval))
+		return false;
 
 	return true;
 }
···
 		return false;
 	}
 
-	__kvm_at_s1e2(vcpu, op, p->regval);
+	if (__kvm_at_s1e2(vcpu, op, p->regval))
+		return false;
 
 	return true;
 }
···
 {
 	u32 op = sys_insn(p->Op0, p->Op1, p->CRn, p->CRm, p->Op2);
 
-	__kvm_at_s12(vcpu, op, p->regval);
+	if (__kvm_at_s12(vcpu, op, p->regval))
+		return false;
 
 	return true;
 }

arch/arm64/tools/cpucaps

···
 HAS_VA52
 HAS_VIRT_HOST_EXTN
 HAS_WFXT
+HAS_XNX
 HAFT
 HW_DBM
 KVM_HVHE

tools/testing/selftests/kvm/Makefile.kvm

···
 TEST_GEN_PROGS_arm64 = $(TEST_GEN_PROGS_COMMON)
 TEST_GEN_PROGS_arm64 += arm64/aarch32_id_regs
 TEST_GEN_PROGS_arm64 += arm64/arch_timer_edge_cases
+TEST_GEN_PROGS_arm64 += arm64/at
 TEST_GEN_PROGS_arm64 += arm64/debug-exceptions
 TEST_GEN_PROGS_arm64 += arm64/hello_el2
 TEST_GEN_PROGS_arm64 += arm64/host_sve

tools/testing/selftests/kvm/arm64/at.c

···
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * at - Test for KVM's AT emulation in the EL2&0 and EL1&0 translation regimes.
+ */
+#include "kvm_util.h"
+#include "processor.h"
+#include "test_util.h"
+#include "ucall.h"
+
+#include <asm/sysreg.h>
+
+#define TEST_ADDR	0x80000000
+
+enum {
+	CLEAR_ACCESS_FLAG,
+	TEST_ACCESS_FLAG,
+};
+
+static u64 *ptep_hva;
+
+#define copy_el2_to_el1(reg)						\
+	write_sysreg_s(read_sysreg_s(SYS_##reg##_EL1), SYS_##reg##_EL12)
+
+/* Yes, this is an ugly hack */
+#define __at(op, addr)	write_sysreg_s(addr, op)
+
+#define test_at_insn(op, expect_fault)							\
+do {											\
+	u64 par, fsc;									\
+	bool fault;									\
+											\
+	GUEST_SYNC(CLEAR_ACCESS_FLAG);							\
+											\
+	__at(OP_AT_##op, TEST_ADDR);							\
+	isb();										\
+	par = read_sysreg(par_el1);							\
+											\
+	fault = par & SYS_PAR_EL1_F;							\
+	fsc = FIELD_GET(SYS_PAR_EL1_FST, par);						\
+											\
+	__GUEST_ASSERT((expect_fault) == fault,						\
+		       "AT "#op": %sexpected fault (par: %lx)1",			\
+		       (expect_fault) ? "" : "un", par);				\
+	if ((expect_fault)) {								\
+		__GUEST_ASSERT(fsc == ESR_ELx_FSC_ACCESS_L(3),				\
+			       "AT "#op": expected access flag fault (par: %lx)",	\
+			       par);							\
+	} else {									\
+		GUEST_ASSERT_EQ(FIELD_GET(SYS_PAR_EL1_ATTR, par), MAIR_ATTR_NORMAL);	\
+		GUEST_ASSERT_EQ(FIELD_GET(SYS_PAR_EL1_SH, par), PTE_SHARED >> 8);	\
+		GUEST_ASSERT_EQ(par & SYS_PAR_EL1_PA, TEST_ADDR);			\
+		GUEST_SYNC(TEST_ACCESS_FLAG);						\
+	}										\
+} while (0)
+
+static void test_at(bool expect_fault)
+{
+	test_at_insn(S1E2R, expect_fault);
+	test_at_insn(S1E2W, expect_fault);
+
+	/* Reuse the stage-1 MMU context from EL2 at EL1 */
+	copy_el2_to_el1(SCTLR);
+	copy_el2_to_el1(MAIR);
+	copy_el2_to_el1(TCR);
+	copy_el2_to_el1(TTBR0);
+	copy_el2_to_el1(TTBR1);
+
+	/* Disable stage-2 translation and enter a non-host context */
+	write_sysreg(0, vtcr_el2);
+	write_sysreg(0, vttbr_el2);
+	sysreg_clear_set(hcr_el2, HCR_EL2_TGE | HCR_EL2_VM, 0);
+	isb();
+
+	test_at_insn(S1E1R, expect_fault);
+	test_at_insn(S1E1W, expect_fault);
+}
+
+static void guest_code(void)
+{
+	sysreg_clear_set(tcr_el1, TCR_HA, 0);
+	isb();
+
+	test_at(true);
+
+	if (!SYS_FIELD_GET(ID_AA64MMFR1_EL1, HAFDBS, read_sysreg(id_aa64mmfr1_el1)))
+		GUEST_DONE();
+
+	/*
+	 * KVM's software PTW makes the implementation choice that the AT
+	 * instruction sets the access flag.
+	 */
+	sysreg_clear_set(tcr_el1, 0, TCR_HA);
+	isb();
+	test_at(false);
+
+	GUEST_DONE();
+}
+
+static void handle_sync(struct kvm_vcpu *vcpu, struct ucall *uc)
+{
+	switch (uc->args[1]) {
+	case CLEAR_ACCESS_FLAG:
+		/*
+		 * Delete + reinstall the memslot to invalidate stage-2
+		 * mappings of the stage-1 page tables, forcing KVM to
+		 * use the 'slow' AT emulation path.
+		 *
+		 * This and clearing the access flag from host userspace
+		 * ensures that the access flag cannot be set speculatively
+		 * and is reliably cleared at the time of the AT instruction.
+		 */
+		clear_bit(__ffs(PTE_AF), ptep_hva);
+		vm_mem_region_reload(vcpu->vm, vcpu->vm->memslots[MEM_REGION_PT]);
+		break;
+	case TEST_ACCESS_FLAG:
+		TEST_ASSERT(test_bit(__ffs(PTE_AF), ptep_hva),
+			    "Expected access flag to be set (desc: %lu)", *ptep_hva);
+		break;
+	default:
+		TEST_FAIL("Unexpected SYNC arg: %lu", uc->args[1]);
+	}
+}
+
+static void run_test(struct kvm_vcpu *vcpu)
+{
+	struct ucall uc;
+
+	while (true) {
+		vcpu_run(vcpu);
+		switch (get_ucall(vcpu, &uc)) {
+		case UCALL_DONE:
+			return;
+		case UCALL_SYNC:
+			handle_sync(vcpu, &uc);
+			continue;
+		case UCALL_ABORT:
+			REPORT_GUEST_ASSERT(uc);
+			return;
+		default:
+			TEST_FAIL("Unexpected ucall: %lu", uc.cmd);
+		}
+	}
+}
+
+int main(void)
+{
+	struct kvm_vcpu_init init;
+	struct kvm_vcpu *vcpu;
+	struct kvm_vm *vm;
+
+	TEST_REQUIRE(kvm_check_cap(KVM_CAP_ARM_EL2));
+
+	vm = vm_create(1);
+
+	kvm_get_default_vcpu_target(vm, &init);
+	init.features[0] |= BIT(KVM_ARM_VCPU_HAS_EL2);
+	vcpu = aarch64_vcpu_add(vm, 0, &init, guest_code);
+	kvm_arch_vm_finalize_vcpus(vm);
+
+	virt_map(vm, TEST_ADDR, TEST_ADDR, 1);
+	ptep_hva = virt_get_pte_hva_at_level(vm, TEST_ADDR, 3);
+	run_test(vcpu);
+
+	kvm_vm_free(vm);
+	return 0;
+}

tools/testing/selftests/kvm/include/kvm_util.h

···
 #endif
 
 void vm_mem_region_set_flags(struct kvm_vm *vm, uint32_t slot, uint32_t flags);
+void vm_mem_region_reload(struct kvm_vm *vm, uint32_t slot);
 void vm_mem_region_move(struct kvm_vm *vm, uint32_t slot, uint64_t new_gpa);
 void vm_mem_region_delete(struct kvm_vm *vm, uint32_t slot);
 struct kvm_vcpu *__vm_vcpu_add(struct kvm_vm *vm, uint32_t vcpu_id);

tools/testing/selftests/kvm/lib/kvm_util.c

···
 		ret, errno, slot, flags);
 }
 
+void vm_mem_region_reload(struct kvm_vm *vm, uint32_t slot)
+{
+	struct userspace_mem_region *region = memslot2region(vm, slot);
+	struct kvm_userspace_memory_region2 tmp = region->region;
+
+	tmp.memory_size = 0;
+	vm_ioctl(vm, KVM_SET_USER_MEMORY_REGION2, &tmp);
+	vm_ioctl(vm, KVM_SET_USER_MEMORY_REGION2, &region->region);
+}
+
 /*
  * VM Memory Region Move
  *