Merge branch kvm-arm64/pkvm-hyp-sharing into kvmarm-master/next

* kvm-arm64/pkvm-hyp-sharing:
  : .
  : Series from Quentin Perret, implementing HYP page share/unshare:
  :
  : This series implements an unshare hypercall at EL2 in nVHE
  : protected mode, and makes use of it to unmmap guest-specific
  : data-structures from EL2 stage-1 during guest tear-down.
  : Crucially, the implementation of the share and unshare
  : routines use page refcounts in the host kernel to avoid
  : accidentally unmapping data-structures that overlap a common
  : page.
  : [...]
  : .
  KVM: arm64: pkvm: Unshare guest structs during teardown
  KVM: arm64: Expose unshare hypercall to the host
  KVM: arm64: Implement do_unshare() helper for unsharing memory
  KVM: arm64: Implement __pkvm_host_share_hyp() using do_share()
  KVM: arm64: Implement do_share() helper for sharing memory
  KVM: arm64: Introduce wrappers for host and hyp spin lock accessors
  KVM: arm64: Extend pkvm_page_state enumeration to handle absent pages
  KVM: arm64: pkvm: Refcount the pages shared with EL2
  KVM: arm64: Introduce kvm_share_hyp()
  KVM: arm64: Implement kvm_pgtable_hyp_unmap() at EL2
  KVM: arm64: Hook up ->page_count() for hypervisor stage-1 page-table
  KVM: arm64: Fixup hyp stage-1 refcount
  KVM: arm64: Refcount hyp stage-1 pgtable pages
  KVM: arm64: Provide {get,put}_page() stubs for early hyp allocator

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

1 files changed, 425 insertions, 75 deletions

diff --git a/arch/arm64/kvm/hyp/nvhe/mem_protect.c b/arch/arm64/kvm/hyp/nvhe/mem_protect.c
index 92262e89672d..16776d1d6151 100644
--- a/arch/arm64/kvm/hyp/nvhe/mem_protect.c
+++ b/arch/arm64/kvm/hyp/nvhe/mem_protect.c
@@ -28,6 +28,26 @@ static struct hyp_pool host_s2_pool;
 
 const u8 pkvm_hyp_id = 1;
 
+static void host_lock_component(void)
+{
+	hyp_spin_lock(&host_kvm.lock);
+}
+
+static void host_unlock_component(void)
+{
+	hyp_spin_unlock(&host_kvm.lock);
+}
+
+static void hyp_lock_component(void)
+{
+	hyp_spin_lock(&pkvm_pgd_lock);
+}
+
+static void hyp_unlock_component(void)
+{
+	hyp_spin_unlock(&pkvm_pgd_lock);
+}
+
 static void *host_s2_zalloc_pages_exact(size_t size)
 {
 	void *addr = hyp_alloc_pages(&host_s2_pool, get_order(size));
@@ -339,116 +359,446 @@ static int host_stage2_idmap(u64 addr)
 
 	prot = is_memory ? PKVM_HOST_MEM_PROT : PKVM_HOST_MMIO_PROT;
 
-	hyp_spin_lock(&host_kvm.lock);
+	host_lock_component();
 	ret = host_stage2_adjust_range(addr, &range);
 	if (ret)
 		goto unlock;
 
 	ret = host_stage2_idmap_locked(range.start, range.end - range.start, prot);
 unlock:
-	hyp_spin_unlock(&host_kvm.lock);
+	host_unlock_component();
 
 	return ret;
 }
 
-static inline bool check_prot(enum kvm_pgtable_prot prot,
-			      enum kvm_pgtable_prot required,
-			      enum kvm_pgtable_prot denied)
+void handle_host_mem_abort(struct kvm_cpu_context *host_ctxt)
 {
-	return (prot & (required | denied)) == required;
+	struct kvm_vcpu_fault_info fault;
+	u64 esr, addr;
+	int ret = 0;
+
+	esr = read_sysreg_el2(SYS_ESR);
+	BUG_ON(!__get_fault_info(esr, &fault));
+
+	addr = (fault.hpfar_el2 & HPFAR_MASK) << 8;
+	ret = host_stage2_idmap(addr);
+	BUG_ON(ret && ret != -EAGAIN);
 }
 
-int __pkvm_host_share_hyp(u64 pfn)
+/* This corresponds to locking order */
+enum pkvm_component_id {
+	PKVM_ID_HOST,
+	PKVM_ID_HYP,
+};
+
+struct pkvm_mem_transition {
+	u64				nr_pages;
+
+	struct {
+		enum pkvm_component_id	id;
+		/* Address in the initiator's address space */
+		u64			addr;
+
+		union {
+			struct {
+				/* Address in the completer's address space */
+				u64	completer_addr;
+			} host;
+		};
+	} initiator;
+
+	struct {
+		enum pkvm_component_id	id;
+	} completer;
+};
+
+struct pkvm_mem_share {
+	const struct pkvm_mem_transition	tx;
+	const enum kvm_pgtable_prot		completer_prot;
+};
+
+struct check_walk_data {
+	enum pkvm_page_state	desired;
+	enum pkvm_page_state	(*get_page_state)(kvm_pte_t pte);
+};
+
+static int __check_page_state_visitor(u64 addr, u64 end, u32 level,
+				      kvm_pte_t *ptep,
+				      enum kvm_pgtable_walk_flags flag,
+				      void * const arg)
 {
-	phys_addr_t addr = hyp_pfn_to_phys(pfn);
-	enum kvm_pgtable_prot prot, cur;
-	void *virt = __hyp_va(addr);
-	enum pkvm_page_state state;
-	kvm_pte_t pte;
-	int ret;
+	struct check_walk_data *d = arg;
+	kvm_pte_t pte = *ptep;
 
-	if (!addr_is_memory(addr))
+	if (kvm_pte_valid(pte) && !addr_is_memory(kvm_pte_to_phys(pte)))
 		return -EINVAL;
 
-	hyp_spin_lock(&host_kvm.lock);
-	hyp_spin_lock(&pkvm_pgd_lock);
+	return d->get_page_state(pte) == d->desired ? 0 : -EPERM;
+}
+
+static int check_page_state_range(struct kvm_pgtable *pgt, u64 addr, u64 size,
+				  struct check_walk_data *data)
+{
+	struct kvm_pgtable_walker walker = {
+		.cb	= __check_page_state_visitor,
+		.arg	= data,
+		.flags	= KVM_PGTABLE_WALK_LEAF,
+	};
+
+	return kvm_pgtable_walk(pgt, addr, size, &walker);
+}
+
+static enum pkvm_page_state host_get_page_state(kvm_pte_t pte)
+{
+	if (!kvm_pte_valid(pte) && pte)
+		return PKVM_NOPAGE;
+
+	return pkvm_getstate(kvm_pgtable_stage2_pte_prot(pte));
+}
+
+static int __host_check_page_state_range(u64 addr, u64 size,
+					 enum pkvm_page_state state)
+{
+	struct check_walk_data d = {
+		.desired	= state,
+		.get_page_state	= host_get_page_state,
+	};
+
+	hyp_assert_lock_held(&host_kvm.lock);
+	return check_page_state_range(&host_kvm.pgt, addr, size, &d);
+}
+
+static int __host_set_page_state_range(u64 addr, u64 size,
+				       enum pkvm_page_state state)
+{
+	enum kvm_pgtable_prot prot = pkvm_mkstate(PKVM_HOST_MEM_PROT, state);
+
+	return host_stage2_idmap_locked(addr, size, prot);
+}
+
+static int host_request_owned_transition(u64 *completer_addr,
+					 const struct pkvm_mem_transition *tx)
+{
+	u64 size = tx->nr_pages * PAGE_SIZE;
+	u64 addr = tx->initiator.addr;
+
+	*completer_addr = tx->initiator.host.completer_addr;
+	return __host_check_page_state_range(addr, size, PKVM_PAGE_OWNED);
+}
+
+static int host_request_unshare(u64 *completer_addr,
+				const struct pkvm_mem_transition *tx)
+{
+	u64 size = tx->nr_pages * PAGE_SIZE;
+	u64 addr = tx->initiator.addr;
+
+	*completer_addr = tx->initiator.host.completer_addr;
+	return __host_check_page_state_range(addr, size, PKVM_PAGE_SHARED_OWNED);
+}
+
+static int host_initiate_share(u64 *completer_addr,
+			       const struct pkvm_mem_transition *tx)
+{
+	u64 size = tx->nr_pages * PAGE_SIZE;
+	u64 addr = tx->initiator.addr;
+
+	*completer_addr = tx->initiator.host.completer_addr;
+	return __host_set_page_state_range(addr, size, PKVM_PAGE_SHARED_OWNED);
+}
+
+static int host_initiate_unshare(u64 *completer_addr,
+				 const struct pkvm_mem_transition *tx)
+{
+	u64 size = tx->nr_pages * PAGE_SIZE;
+	u64 addr = tx->initiator.addr;
+
+	*completer_addr = tx->initiator.host.completer_addr;
+	return __host_set_page_state_range(addr, size, PKVM_PAGE_OWNED);
+}
+
+static enum pkvm_page_state hyp_get_page_state(kvm_pte_t pte)
+{
+	if (!kvm_pte_valid(pte))
+		return PKVM_NOPAGE;
+
+	return pkvm_getstate(kvm_pgtable_stage2_pte_prot(pte));
+}
+
+static int __hyp_check_page_state_range(u64 addr, u64 size,
+					enum pkvm_page_state state)
+{
+	struct check_walk_data d = {
+		.desired	= state,
+		.get_page_state	= hyp_get_page_state,
+	};
+
+	hyp_assert_lock_held(&pkvm_pgd_lock);
+	return check_page_state_range(&pkvm_pgtable, addr, size, &d);
+}
+
+static bool __hyp_ack_skip_pgtable_check(const struct pkvm_mem_transition *tx)
+{
+	return !(IS_ENABLED(CONFIG_NVHE_EL2_DEBUG) ||
+		 tx->initiator.id != PKVM_ID_HOST);
+}
+
+static int hyp_ack_share(u64 addr, const struct pkvm_mem_transition *tx,
+			 enum kvm_pgtable_prot perms)
+{
+	u64 size = tx->nr_pages * PAGE_SIZE;
+
+	if (perms != PAGE_HYP)
+		return -EPERM;
+
+	if (__hyp_ack_skip_pgtable_check(tx))
+		return 0;
+
+	return __hyp_check_page_state_range(addr, size, PKVM_NOPAGE);
+}
+
+static int hyp_ack_unshare(u64 addr, const struct pkvm_mem_transition *tx)
+{
+	u64 size = tx->nr_pages * PAGE_SIZE;
+
+	if (__hyp_ack_skip_pgtable_check(tx))
+		return 0;
+
+	return __hyp_check_page_state_range(addr, size,
+					    PKVM_PAGE_SHARED_BORROWED);
+}
+
+static int hyp_complete_share(u64 addr, const struct pkvm_mem_transition *tx,
+			      enum kvm_pgtable_prot perms)
+{
+	void *start = (void *)addr, *end = start + (tx->nr_pages * PAGE_SIZE);
+	enum kvm_pgtable_prot prot;
+
+	prot = pkvm_mkstate(perms, PKVM_PAGE_SHARED_BORROWED);
+	return pkvm_create_mappings_locked(start, end, prot);
+}
+
+static int hyp_complete_unshare(u64 addr, const struct pkvm_mem_transition *tx)
+{
+	u64 size = tx->nr_pages * PAGE_SIZE;
+	int ret = kvm_pgtable_hyp_unmap(&pkvm_pgtable, addr, size);
+
+	return (ret != size) ? -EFAULT : 0;
+}
+
+static int check_share(struct pkvm_mem_share *share)
+{
+	const struct pkvm_mem_transition *tx = &share->tx;
+	u64 completer_addr;
+	int ret;
+
+	switch (tx->initiator.id) {
+	case PKVM_ID_HOST:
+		ret = host_request_owned_transition(&completer_addr, tx);
+		break;
+	default:
+		ret = -EINVAL;
+	}
 
-	ret = kvm_pgtable_get_leaf(&host_kvm.pgt, addr, &pte, NULL);
 	if (ret)
-		goto unlock;
-	if (!pte)
-		goto map_shared;
+		return ret;
 
-	/*
-	 * Check attributes in the host stage-2 PTE. We need the page to be:
-	 *  - mapped RWX as we're sharing memory;
-	 *  - not borrowed, as that implies absence of ownership.
-	 * Otherwise, we can't let it got through
-	 */
-	cur = kvm_pgtable_stage2_pte_prot(pte);
-	prot = pkvm_mkstate(0, PKVM_PAGE_SHARED_BORROWED);
-	if (!check_prot(cur, PKVM_HOST_MEM_PROT, prot)) {
-		ret = -EPERM;
-		goto unlock;
+	switch (tx->completer.id) {
+	case PKVM_ID_HYP:
+		ret = hyp_ack_share(completer_addr, tx, share->completer_prot);
+		break;
+	default:
+		ret = -EINVAL;
 	}
 
-	state = pkvm_getstate(cur);
-	if (state == PKVM_PAGE_OWNED)
-		goto map_shared;
+	return ret;
+}
 
-	/*
-	 * Tolerate double-sharing the same page, but this requires
-	 * cross-checking the hypervisor stage-1.
-	 */
-	if (state != PKVM_PAGE_SHARED_OWNED) {
-		ret = -EPERM;
-		goto unlock;
+static int __do_share(struct pkvm_mem_share *share)
+{
+	const struct pkvm_mem_transition *tx = &share->tx;
+	u64 completer_addr;
+	int ret;
+
+	switch (tx->initiator.id) {
+	case PKVM_ID_HOST:
+		ret = host_initiate_share(&completer_addr, tx);
+		break;
+	default:
+		ret = -EINVAL;
 	}
 
-	ret = kvm_pgtable_get_leaf(&pkvm_pgtable, (u64)virt, &pte, NULL);
 	if (ret)
-		goto unlock;
+		return ret;
 
-	/*
-	 * If the page has been shared with the hypervisor, it must be
-	 * already mapped as SHARED_BORROWED in its stage-1.
-	 */
-	cur = kvm_pgtable_hyp_pte_prot(pte);
-	prot = pkvm_mkstate(PAGE_HYP, PKVM_PAGE_SHARED_BORROWED);
-	if (!check_prot(cur, prot, ~prot))
-		ret = -EPERM;
-	goto unlock;
+	switch (tx->completer.id) {
+	case PKVM_ID_HYP:
+		ret = hyp_complete_share(completer_addr, tx, share->completer_prot);
+		break;
+	default:
+		ret = -EINVAL;
+	}
 
-map_shared:
-	/*
-	 * If the page is not yet shared, adjust mappings in both page-tables
-	 * while both locks are held.
-	 */
-	prot = pkvm_mkstate(PAGE_HYP, PKVM_PAGE_SHARED_BORROWED);
-	ret = pkvm_create_mappings_locked(virt, virt + PAGE_SIZE, prot);
-	BUG_ON(ret);
+	return ret;
+}
 
-	prot = pkvm_mkstate(PKVM_HOST_MEM_PROT, PKVM_PAGE_SHARED_OWNED);
-	ret = host_stage2_idmap_locked(addr, PAGE_SIZE, prot);
-	BUG_ON(ret);
+/*
+ * do_share():
+ *
+ * The page owner grants access to another component with a given set
+ * of permissions.
+ *
+ * Initiator: OWNED	=> SHARED_OWNED
+ * Completer: NOPAGE	=> SHARED_BORROWED
+ */
+static int do_share(struct pkvm_mem_share *share)
+{
+	int ret;
 
-unlock:
-	hyp_spin_unlock(&pkvm_pgd_lock);
-	hyp_spin_unlock(&host_kvm.lock);
+	ret = check_share(share);
+	if (ret)
+		return ret;
+
+	return WARN_ON(__do_share(share));
+}
+
+static int check_unshare(struct pkvm_mem_share *share)
+{
+	const struct pkvm_mem_transition *tx = &share->tx;
+	u64 completer_addr;
+	int ret;
+
+	switch (tx->initiator.id) {
+	case PKVM_ID_HOST:
+		ret = host_request_unshare(&completer_addr, tx);
+		break;
+	default:
+		ret = -EINVAL;
+	}
+
+	if (ret)
+		return ret;
+
+	switch (tx->completer.id) {
+	case PKVM_ID_HYP:
+		ret = hyp_ack_unshare(completer_addr, tx);
+		break;
+	default:
+		ret = -EINVAL;
+	}
 
 	return ret;
 }
 
-void handle_host_mem_abort(struct kvm_cpu_context *host_ctxt)
+static int __do_unshare(struct pkvm_mem_share *share)
 {
-	struct kvm_vcpu_fault_info fault;
-	u64 esr, addr;
-	int ret = 0;
+	const struct pkvm_mem_transition *tx = &share->tx;
+	u64 completer_addr;
+	int ret;
 
-	esr = read_sysreg_el2(SYS_ESR);
-	BUG_ON(!__get_fault_info(esr, &fault));
+	switch (tx->initiator.id) {
+	case PKVM_ID_HOST:
+		ret = host_initiate_unshare(&completer_addr, tx);
+		break;
+	default:
+		ret = -EINVAL;
+	}
 
-	addr = (fault.hpfar_el2 & HPFAR_MASK) << 8;
-	ret = host_stage2_idmap(addr);
-	BUG_ON(ret && ret != -EAGAIN);
+	if (ret)
+		return ret;
+
+	switch (tx->completer.id) {
+	case PKVM_ID_HYP:
+		ret = hyp_complete_unshare(completer_addr, tx);
+		break;
+	default:
+		ret = -EINVAL;
+	}
+
+	return ret;
+}
+
+/*
+ * do_unshare():
+ *
+ * The page owner revokes access from another component for a range of
+ * pages which were previously shared using do_share().
+ *
+ * Initiator: SHARED_OWNED	=> OWNED
+ * Completer: SHARED_BORROWED	=> NOPAGE
+ */
+static int do_unshare(struct pkvm_mem_share *share)
+{
+	int ret;
+
+	ret = check_unshare(share);
+	if (ret)
+		return ret;
+
+	return WARN_ON(__do_unshare(share));
+}
+
+int __pkvm_host_share_hyp(u64 pfn)
+{
+	int ret;
+	u64 host_addr = hyp_pfn_to_phys(pfn);
+	u64 hyp_addr = (u64)__hyp_va(host_addr);
+	struct pkvm_mem_share share = {
+		.tx	= {
+			.nr_pages	= 1,
+			.initiator	= {
+				.id	= PKVM_ID_HOST,
+				.addr	= host_addr,
+				.host	= {
+					.completer_addr = hyp_addr,
+				},
+			},
+			.completer	= {
+				.id	= PKVM_ID_HYP,
+			},
+		},
+		.completer_prot	= PAGE_HYP,
+	};
+
+	host_lock_component();
+	hyp_lock_component();
+
+	ret = do_share(&share);
+
+	hyp_unlock_component();
+	host_unlock_component();
+
+	return ret;
+}
+
+int __pkvm_host_unshare_hyp(u64 pfn)
+{
+	int ret;
+	u64 host_addr = hyp_pfn_to_phys(pfn);
+	u64 hyp_addr = (u64)__hyp_va(host_addr);
+	struct pkvm_mem_share share = {
+		.tx	= {
+			.nr_pages	= 1,
+			.initiator	= {
+				.id	= PKVM_ID_HOST,
+				.addr	= host_addr,
+				.host	= {
+					.completer_addr = hyp_addr,
+				},
+			},
+			.completer	= {
+				.id	= PKVM_ID_HYP,
+			},
+		},
+		.completer_prot	= PAGE_HYP,
+	};
+
+	host_lock_component();
+	hyp_lock_component();
+
+	ret = do_unshare(&share);
+
+	hyp_unlock_component();
+	host_unlock_component();
+
+	return ret;
 }