SPCI is now called PSA FF-A.

Change-Id: Iaa10e0449edf5f6493ab21e648219392b17cc5ec
diff --git a/src/ffa_memory.c b/src/ffa_memory.c
new file mode 100644
index 0000000..8def572
--- /dev/null
+++ b/src/ffa_memory.c
@@ -0,0 +1,1787 @@
+/*
+ * Copyright 2019 The Hafnium Authors.
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *     https://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+#include "hf/ffa_memory.h"
+
+#include "hf/arch/tee.h"
+
+#include "hf/api.h"
+#include "hf/check.h"
+#include "hf/dlog.h"
+#include "hf/ffa_internal.h"
+#include "hf/mpool.h"
+#include "hf/std.h"
+#include "hf/vm.h"
+
+/** The maximum number of recipients a memory region may be sent to. */
+#define MAX_MEM_SHARE_RECIPIENTS 1
+
+/**
+ * The maximum number of memory sharing handles which may be active at once. A
+ * DONATE handle is active from when it is sent to when it is retrieved; a SHARE
+ * or LEND handle is active from when it is sent to when it is reclaimed.
+ */
+#define MAX_MEM_SHARES 100
+
+static_assert(sizeof(struct ffa_memory_region_constituent) % 16 == 0,
+	      "struct ffa_memory_region_constituent must be a multiple of 16 "
+	      "bytes long.");
+static_assert(sizeof(struct ffa_composite_memory_region) % 16 == 0,
+	      "struct ffa_composite_memory_region must be a multiple of 16 "
+	      "bytes long.");
+static_assert(sizeof(struct ffa_memory_region_attributes) == 4,
+	      "struct ffa_memory_region_attributes must be 4bytes long.");
+static_assert(sizeof(struct ffa_memory_access) % 16 == 0,
+	      "struct ffa_memory_access must be a multiple of 16 bytes long.");
+static_assert(sizeof(struct ffa_memory_region) % 16 == 0,
+	      "struct ffa_memory_region must be a multiple of 16 bytes long.");
+static_assert(sizeof(struct ffa_mem_relinquish) % 16 == 0,
+	      "struct ffa_mem_relinquish must be a multiple of 16 "
+	      "bytes long.");
+
+struct ffa_memory_share_state {
+	/**
+	 * The memory region being shared, or NULL if this share state is
+	 * unallocated.
+	 */
+	struct ffa_memory_region *memory_region;
+
+	/**
+	 * The FF-A function used for sharing the memory. Must be one of
+	 * FFA_MEM_DONATE_32, FFA_MEM_LEND_32 or FFA_MEM_SHARE_32 if the
+	 * share state is allocated, or 0.
+	 */
+	uint32_t share_func;
+
+	/**
+	 * Whether each recipient has retrieved the memory region yet. The order
+	 * of this array matches the order of the attribute descriptors in the
+	 * memory region descriptor. Any entries beyond the attribute_count will
+	 * always be false.
+	 */
+	bool retrieved[MAX_MEM_SHARE_RECIPIENTS];
+};
+
+/**
+ * Encapsulates the set of share states while the `share_states_lock` is held.
+ */
+struct share_states_locked {
+	struct ffa_memory_share_state *share_states;
+};
+
+/**
+ * All access to members of a `struct ffa_memory_share_state` must be guarded
+ * by this lock.
+ */
+static struct spinlock share_states_lock_instance = SPINLOCK_INIT;
+static struct ffa_memory_share_state share_states[MAX_MEM_SHARES];
+
+/**
+ * Initialises the next available `struct ffa_memory_share_state` and sets
+ * `handle` to its handle. Returns true on succes or false if none are
+ * available.
+ */
+static bool allocate_share_state(uint32_t share_func,
+				 struct ffa_memory_region *memory_region,
+				 ffa_memory_handle_t *handle)
+{
+	uint64_t i;
+
+	CHECK(memory_region != NULL);
+
+	sl_lock(&share_states_lock_instance);
+	for (i = 0; i < MAX_MEM_SHARES; ++i) {
+		if (share_states[i].share_func == 0) {
+			uint32_t j;
+			struct ffa_memory_share_state *allocated_state =
+				&share_states[i];
+			allocated_state->share_func = share_func;
+			allocated_state->memory_region = memory_region;
+			for (j = 0; j < MAX_MEM_SHARE_RECIPIENTS; ++j) {
+				allocated_state->retrieved[j] = false;
+			}
+			*handle = i | FFA_MEMORY_HANDLE_ALLOCATOR_HYPERVISOR;
+			sl_unlock(&share_states_lock_instance);
+			return true;
+		}
+	}
+
+	sl_unlock(&share_states_lock_instance);
+	return false;
+}
+
+/** Locks the share states lock. */
+struct share_states_locked share_states_lock(void)
+{
+	sl_lock(&share_states_lock_instance);
+
+	return (struct share_states_locked){.share_states = share_states};
+}
+
+/** Unlocks the share states lock. */
+static void share_states_unlock(struct share_states_locked *share_states)
+{
+	CHECK(share_states->share_states != NULL);
+	share_states->share_states = NULL;
+	sl_unlock(&share_states_lock_instance);
+}
+
+/**
+ * If the given handle is a valid handle for an allocated share state then takes
+ * the lock, initialises `share_state_locked` to point to the share state and
+ * returns true. Otherwise returns false and doesn't take the lock.
+ */
+static bool get_share_state(struct share_states_locked share_states,
+			    ffa_memory_handle_t handle,
+			    struct ffa_memory_share_state **share_state_ret)
+{
+	struct ffa_memory_share_state *share_state;
+	uint32_t index = handle & ~FFA_MEMORY_HANDLE_ALLOCATOR_MASK;
+
+	if (index >= MAX_MEM_SHARES) {
+		return false;
+	}
+
+	share_state = &share_states.share_states[index];
+
+	if (share_state->share_func == 0) {
+		return false;
+	}
+
+	*share_state_ret = share_state;
+	return true;
+}
+
+/** Marks a share state as unallocated. */
+static void share_state_free(struct share_states_locked share_states,
+			     struct ffa_memory_share_state *share_state,
+			     struct mpool *page_pool)
+{
+	CHECK(share_states.share_states != NULL);
+	share_state->share_func = 0;
+	mpool_free(page_pool, share_state->memory_region);
+	share_state->memory_region = NULL;
+}
+
+/**
+ * Marks the share state with the given handle as unallocated, or returns false
+ * if the handle was invalid.
+ */
+static bool share_state_free_handle(ffa_memory_handle_t handle,
+				    struct mpool *page_pool)
+{
+	struct share_states_locked share_states = share_states_lock();
+	struct ffa_memory_share_state *share_state;
+
+	if (!get_share_state(share_states, handle, &share_state)) {
+		share_states_unlock(&share_states);
+		return false;
+	}
+
+	share_state_free(share_states, share_state, page_pool);
+	share_states_unlock(&share_states);
+
+	return true;
+}
+
+static void dump_memory_region(struct ffa_memory_region *memory_region)
+{
+	uint32_t i;
+
+	if (LOG_LEVEL < LOG_LEVEL_VERBOSE) {
+		return;
+	}
+
+	dlog("from VM %d, attributes %#x, flags %#x, handle %#x, tag %d, to %d "
+	     "recipients [",
+	     memory_region->sender, memory_region->attributes,
+	     memory_region->flags, memory_region->handle, memory_region->tag,
+	     memory_region->receiver_count);
+	for (i = 0; i < memory_region->receiver_count; ++i) {
+		if (i != 0) {
+			dlog(", ");
+		}
+		dlog("VM %d: %#x (offset %d)",
+		     memory_region->receivers[i].receiver_permissions.receiver,
+		     memory_region->receivers[i]
+			     .receiver_permissions.permissions,
+		     memory_region->receivers[i]
+			     .composite_memory_region_offset);
+	}
+	dlog("]");
+}
+
+static void dump_share_states(void)
+{
+	uint32_t i;
+
+	if (LOG_LEVEL < LOG_LEVEL_VERBOSE) {
+		return;
+	}
+
+	dlog("Current share states:\n");
+	sl_lock(&share_states_lock_instance);
+	for (i = 0; i < MAX_MEM_SHARES; ++i) {
+		if (share_states[i].share_func != 0) {
+			dlog("%d: ", i);
+			switch (share_states[i].share_func) {
+			case FFA_MEM_SHARE_32:
+				dlog("SHARE");
+				break;
+			case FFA_MEM_LEND_32:
+				dlog("LEND");
+				break;
+			case FFA_MEM_DONATE_32:
+				dlog("DONATE");
+				break;
+			default:
+				dlog("invalid share_func %#x",
+				     share_states[i].share_func);
+			}
+			dlog(" (");
+			dump_memory_region(share_states[i].memory_region);
+			if (share_states[i].retrieved[0]) {
+				dlog("): retrieved\n");
+			} else {
+				dlog("): not retrieved\n");
+			}
+			break;
+		}
+	}
+	sl_unlock(&share_states_lock_instance);
+}
+
+/* TODO: Add device attributes: GRE, cacheability, shareability. */
+static inline uint32_t ffa_memory_permissions_to_mode(
+	ffa_memory_access_permissions_t permissions)
+{
+	uint32_t mode = 0;
+
+	switch (ffa_get_data_access_attr(permissions)) {
+	case FFA_DATA_ACCESS_RO:
+		mode = MM_MODE_R;
+		break;
+	case FFA_DATA_ACCESS_RW:
+	case FFA_DATA_ACCESS_NOT_SPECIFIED:
+		mode = MM_MODE_R | MM_MODE_W;
+		break;
+	case FFA_DATA_ACCESS_RESERVED:
+		panic("Tried to convert FFA_DATA_ACCESS_RESERVED.");
+	}
+
+	switch (ffa_get_instruction_access_attr(permissions)) {
+	case FFA_INSTRUCTION_ACCESS_NX:
+		break;
+	case FFA_INSTRUCTION_ACCESS_X:
+	case FFA_INSTRUCTION_ACCESS_NOT_SPECIFIED:
+		mode |= MM_MODE_X;
+		break;
+	case FFA_INSTRUCTION_ACCESS_RESERVED:
+		panic("Tried to convert FFA_INSTRUCTION_ACCESS_RESVERVED.");
+	}
+
+	return mode;
+}
+
+/**
+ * Get the current mode in the stage-2 page table of the given vm of all the
+ * pages in the given constituents, if they all have the same mode, or return
+ * an appropriate FF-A error if not.
+ */
+static struct ffa_value constituents_get_mode(
+	struct vm_locked vm, uint32_t *orig_mode,
+	struct ffa_memory_region_constituent *constituents,
+	uint32_t constituent_count)
+{
+	uint32_t i;
+
+	if (constituent_count == 0) {
+		/*
+		 * Fail if there are no constituents. Otherwise we would get an
+		 * uninitialised *orig_mode.
+		 */
+		return ffa_error(FFA_INVALID_PARAMETERS);
+	}
+
+	for (i = 0; i < constituent_count; ++i) {
+		ipaddr_t begin = ipa_init(constituents[i].address);
+		size_t size = constituents[i].page_count * PAGE_SIZE;
+		ipaddr_t end = ipa_add(begin, size);
+		uint32_t current_mode;
+
+		/* Fail if addresses are not page-aligned. */
+		if (!is_aligned(ipa_addr(begin), PAGE_SIZE) ||
+		    !is_aligned(ipa_addr(end), PAGE_SIZE)) {
+			return ffa_error(FFA_INVALID_PARAMETERS);
+		}
+
+		/*
+		 * Ensure that this constituent memory range is all mapped with
+		 * the same mode.
+		 */
+		if (!mm_vm_get_mode(&vm.vm->ptable, begin, end,
+				    &current_mode)) {
+			return ffa_error(FFA_DENIED);
+		}
+
+		/*
+		 * Ensure that all constituents are mapped with the same mode.
+		 */
+		if (i == 0) {
+			*orig_mode = current_mode;
+		} else if (current_mode != *orig_mode) {
+			return ffa_error(FFA_DENIED);
+		}
+	}
+
+	return (struct ffa_value){.func = FFA_SUCCESS_32};
+}
+
+/**
+ * Verify that all pages have the same mode, that the starting mode
+ * constitutes a valid state and obtain the next mode to apply
+ * to the sending VM.
+ *
+ * Returns:
+ *   1) FFA_DENIED if a state transition was not found;
+ *   2) FFA_DENIED if the pages being shared do not have the same mode within
+ *     the <from> VM;
+ *   3) FFA_INVALID_PARAMETERS if the beginning and end IPAs are not page
+ *     aligned;
+ *   4) FFA_INVALID_PARAMETERS if the requested share type was not handled.
+ *  Or FFA_SUCCESS on success.
+ */
+static struct ffa_value ffa_send_check_transition(
+	struct vm_locked from, uint32_t share_func,
+	ffa_memory_access_permissions_t permissions, uint32_t *orig_from_mode,
+	struct ffa_memory_region_constituent *constituents,
+	uint32_t constituent_count, uint32_t *from_mode)
+{
+	const uint32_t state_mask =
+		MM_MODE_INVALID | MM_MODE_UNOWNED | MM_MODE_SHARED;
+	const uint32_t required_from_mode =
+		ffa_memory_permissions_to_mode(permissions);
+	struct ffa_value ret;
+
+	ret = constituents_get_mode(from, orig_from_mode, constituents,
+				    constituent_count);
+	if (ret.func != FFA_SUCCESS_32) {
+		return ret;
+	}
+
+	/* Ensure the address range is normal memory and not a device. */
+	if (*orig_from_mode & MM_MODE_D) {
+		dlog_verbose("Can't share device memory (mode is %#x).\n",
+			     *orig_from_mode);
+		return ffa_error(FFA_DENIED);
+	}
+
+	/*
+	 * Ensure the sender is the owner and has exclusive access to the
+	 * memory.
+	 */
+	if ((*orig_from_mode & state_mask) != 0) {
+		return ffa_error(FFA_DENIED);
+	}
+
+	if ((*orig_from_mode & required_from_mode) != required_from_mode) {
+		dlog_verbose(
+			"Sender tried to send memory with permissions which "
+			"required mode %#x but only had %#x itself.\n",
+			required_from_mode, *orig_from_mode);
+		return ffa_error(FFA_DENIED);
+	}
+
+	/* Find the appropriate new mode. */
+	*from_mode = ~state_mask & *orig_from_mode;
+	switch (share_func) {
+	case FFA_MEM_DONATE_32:
+		*from_mode |= MM_MODE_INVALID | MM_MODE_UNOWNED;
+		break;
+
+	case FFA_MEM_LEND_32:
+		*from_mode |= MM_MODE_INVALID;
+		break;
+
+	case FFA_MEM_SHARE_32:
+		*from_mode |= MM_MODE_SHARED;
+		break;
+
+	default:
+		return ffa_error(FFA_INVALID_PARAMETERS);
+	}
+
+	return (struct ffa_value){.func = FFA_SUCCESS_32};
+}
+
+static struct ffa_value ffa_relinquish_check_transition(
+	struct vm_locked from, uint32_t *orig_from_mode,
+	struct ffa_memory_region_constituent *constituents,
+	uint32_t constituent_count, uint32_t *from_mode)
+{
+	const uint32_t state_mask =
+		MM_MODE_INVALID | MM_MODE_UNOWNED | MM_MODE_SHARED;
+	uint32_t orig_from_state;
+	struct ffa_value ret;
+
+	ret = constituents_get_mode(from, orig_from_mode, constituents,
+				    constituent_count);
+	if (ret.func != FFA_SUCCESS_32) {
+		return ret;
+	}
+
+	/* Ensure the address range is normal memory and not a device. */
+	if (*orig_from_mode & MM_MODE_D) {
+		dlog_verbose("Can't relinquish device memory (mode is %#x).\n",
+			     *orig_from_mode);
+		return ffa_error(FFA_DENIED);
+	}
+
+	/*
+	 * Ensure the relinquishing VM is not the owner but has access to the
+	 * memory.
+	 */
+	orig_from_state = *orig_from_mode & state_mask;
+	if ((orig_from_state & ~MM_MODE_SHARED) != MM_MODE_UNOWNED) {
+		dlog_verbose(
+			"Tried to relinquish memory in state %#x (masked %#x "
+			"but "
+			"should be %#x).\n",
+			*orig_from_mode, orig_from_state, MM_MODE_UNOWNED);
+		return ffa_error(FFA_DENIED);
+	}
+
+	/* Find the appropriate new mode. */
+	*from_mode = (~state_mask & *orig_from_mode) | MM_MODE_UNMAPPED_MASK;
+
+	return (struct ffa_value){.func = FFA_SUCCESS_32};
+}
+
+/**
+ * Verify that all pages have the same mode, that the starting mode
+ * constitutes a valid state and obtain the next mode to apply
+ * to the retrieving VM.
+ *
+ * Returns:
+ *   1) FFA_DENIED if a state transition was not found;
+ *   2) FFA_DENIED if the pages being shared do not have the same mode within
+ *     the <to> VM;
+ *   3) FFA_INVALID_PARAMETERS if the beginning and end IPAs are not page
+ *     aligned;
+ *   4) FFA_INVALID_PARAMETERS if the requested share type was not handled.
+ *  Or FFA_SUCCESS on success.
+ */
+static struct ffa_value ffa_retrieve_check_transition(
+	struct vm_locked to, uint32_t share_func,
+	struct ffa_memory_region_constituent *constituents,
+	uint32_t constituent_count, uint32_t memory_to_attributes,
+	uint32_t *to_mode)
+{
+	uint32_t orig_to_mode;
+	struct ffa_value ret;
+
+	ret = constituents_get_mode(to, &orig_to_mode, constituents,
+				    constituent_count);
+	if (ret.func != FFA_SUCCESS_32) {
+		return ret;
+	}
+
+	if (share_func == FFA_MEM_RECLAIM_32) {
+		const uint32_t state_mask =
+			MM_MODE_INVALID | MM_MODE_UNOWNED | MM_MODE_SHARED;
+		uint32_t orig_to_state = orig_to_mode & state_mask;
+
+		if (orig_to_state != MM_MODE_INVALID &&
+		    orig_to_state != MM_MODE_SHARED) {
+			return ffa_error(FFA_DENIED);
+		}
+	} else {
+		/*
+		 * Ensure the retriever has the expected state. We don't care
+		 * about the MM_MODE_SHARED bit; either with or without it set
+		 * are both valid representations of the !O-NA state.
+		 */
+		if ((orig_to_mode & MM_MODE_UNMAPPED_MASK) !=
+		    MM_MODE_UNMAPPED_MASK) {
+			return ffa_error(FFA_DENIED);
+		}
+	}
+
+	/* Find the appropriate new mode. */
+	*to_mode = memory_to_attributes;
+	switch (share_func) {
+	case FFA_MEM_DONATE_32:
+		*to_mode |= 0;
+		break;
+
+	case FFA_MEM_LEND_32:
+		*to_mode |= MM_MODE_UNOWNED;
+		break;
+
+	case FFA_MEM_SHARE_32:
+		*to_mode |= MM_MODE_UNOWNED | MM_MODE_SHARED;
+		break;
+
+	case FFA_MEM_RECLAIM_32:
+		*to_mode |= 0;
+		break;
+
+	default:
+		return ffa_error(FFA_INVALID_PARAMETERS);
+	}
+
+	return (struct ffa_value){.func = FFA_SUCCESS_32};
+}
+
+/**
+ * Updates a VM's page table such that the given set of physical address ranges
+ * are mapped in the address space at the corresponding address ranges, in the
+ * mode provided.
+ *
+ * If commit is false, the page tables will be allocated from the mpool but no
+ * mappings will actually be updated. This function must always be called first
+ * with commit false to check that it will succeed before calling with commit
+ * true, to avoid leaving the page table in a half-updated state. To make a
+ * series of changes atomically you can call them all with commit false before
+ * calling them all with commit true.
+ *
+ * mm_vm_defrag should always be called after a series of page table updates,
+ * whether they succeed or fail.
+ *
+ * Returns true on success, or false if the update failed and no changes were
+ * made to memory mappings.
+ */
+static bool ffa_region_group_identity_map(
+	struct vm_locked vm_locked,
+	struct ffa_memory_region_constituent *constituents,
+	uint32_t constituent_count, int mode, struct mpool *ppool, bool commit)
+{
+	/* Iterate over the memory region constituents. */
+	for (uint32_t index = 0; index < constituent_count; index++) {
+		size_t size = constituents[index].page_count * PAGE_SIZE;
+		paddr_t pa_begin =
+			pa_from_ipa(ipa_init(constituents[index].address));
+		paddr_t pa_end = pa_add(pa_begin, size);
+
+		if (commit) {
+			vm_identity_commit(vm_locked, pa_begin, pa_end, mode,
+					   ppool, NULL);
+		} else if (!vm_identity_prepare(vm_locked, pa_begin, pa_end,
+						mode, ppool)) {
+			return false;
+		}
+	}
+
+	return true;
+}
+
+/**
+ * Clears a region of physical memory by overwriting it with zeros. The data is
+ * flushed from the cache so the memory has been cleared across the system.
+ */
+static bool clear_memory(paddr_t begin, paddr_t end, struct mpool *ppool)
+{
+	/*
+	 * TODO: change this to a CPU local single page window rather than a
+	 *       global mapping of the whole range. Such an approach will limit
+	 *       the changes to stage-1 tables and will allow only local
+	 *       invalidation.
+	 */
+	bool ret;
+	struct mm_stage1_locked stage1_locked = mm_lock_stage1();
+	void *ptr =
+		mm_identity_map(stage1_locked, begin, end, MM_MODE_W, ppool);
+	size_t size = pa_difference(begin, end);
+
+	if (!ptr) {
+		/* TODO: partial defrag of failed range. */
+		/* Recover any memory consumed in failed mapping. */
+		mm_defrag(stage1_locked, ppool);
+		goto fail;
+	}
+
+	memset_s(ptr, size, 0, size);
+	arch_mm_flush_dcache(ptr, size);
+	mm_unmap(stage1_locked, begin, end, ppool);
+
+	ret = true;
+	goto out;
+
+fail:
+	ret = false;
+
+out:
+	mm_unlock_stage1(&stage1_locked);
+
+	return ret;
+}
+
+/**
+ * Clears a region of physical memory by overwriting it with zeros. The data is
+ * flushed from the cache so the memory has been cleared across the system.
+ */
+static bool ffa_clear_memory_constituents(
+	struct ffa_memory_region_constituent *constituents,
+	uint32_t constituent_count, struct mpool *page_pool)
+{
+	struct mpool local_page_pool;
+	struct mm_stage1_locked stage1_locked;
+	bool ret = false;
+
+	/*
+	 * Create a local pool so any freed memory can't be used by another
+	 * thread. This is to ensure each constituent that is mapped can be
+	 * unmapped again afterwards.
+	 */
+	mpool_init_with_fallback(&local_page_pool, page_pool);
+
+	/* Iterate over the memory region constituents. */
+	for (uint32_t i = 0; i < constituent_count; ++i) {
+		size_t size = constituents[i].page_count * PAGE_SIZE;
+		paddr_t begin = pa_from_ipa(ipa_init(constituents[i].address));
+		paddr_t end = pa_add(begin, size);
+
+		if (!clear_memory(begin, end, &local_page_pool)) {
+			/*
+			 * api_clear_memory will defrag on failure, so no need
+			 * to do it here.
+			 */
+			goto out;
+		}
+	}
+
+	/*
+	 * Need to defrag after clearing, as it may have added extra mappings to
+	 * the stage 1 page table.
+	 */
+	stage1_locked = mm_lock_stage1();
+	mm_defrag(stage1_locked, &local_page_pool);
+	mm_unlock_stage1(&stage1_locked);
+
+	ret = true;
+
+out:
+	mpool_fini(&local_page_pool);
+	return ret;
+}
+
+/**
+ * Validates and prepares memory to be sent from the calling VM to another.
+ *
+ * This function requires the calling context to hold the <from> VM lock.
+ *
+ * Returns:
+ *  In case of error, one of the following values is returned:
+ *   1) FFA_INVALID_PARAMETERS - The endpoint provided parameters were
+ *     erroneous;
+ *   2) FFA_NO_MEMORY - Hafnium did not have sufficient memory to complete
+ *     the request.
+ *   3) FFA_DENIED - The sender doesn't have sufficient access to send the
+ *     memory with the given permissions.
+ *  Success is indicated by FFA_SUCCESS.
+ */
+static struct ffa_value ffa_send_memory(
+	struct vm_locked from_locked,
+	struct ffa_memory_region_constituent *constituents,
+	uint32_t constituent_count, uint32_t share_func,
+	ffa_memory_access_permissions_t permissions, struct mpool *page_pool,
+	bool clear)
+{
+	struct vm *from = from_locked.vm;
+	uint32_t orig_from_mode;
+	uint32_t from_mode;
+	struct mpool local_page_pool;
+	struct ffa_value ret;
+
+	/*
+	 * Make sure constituents are properly aligned to a 64-bit boundary. If
+	 * not we would get alignment faults trying to read (64-bit) values.
+	 */
+	if (!is_aligned(constituents, 8)) {
+		return ffa_error(FFA_INVALID_PARAMETERS);
+	}
+
+	/*
+	 * Check if the state transition is lawful for the sender, ensure that
+	 * all constituents of a memory region being shared are at the same
+	 * state.
+	 */
+	ret = ffa_send_check_transition(from_locked, share_func, permissions,
+					&orig_from_mode, constituents,
+					constituent_count, &from_mode);
+	if (ret.func != FFA_SUCCESS_32) {
+		return ret;
+	}
+
+	/*
+	 * Create a local pool so any freed memory can't be used by another
+	 * thread. This is to ensure the original mapping can be restored if the
+	 * clear fails.
+	 */
+	mpool_init_with_fallback(&local_page_pool, page_pool);
+
+	/*
+	 * First reserve all required memory for the new page table entries
+	 * without committing, to make sure the entire operation will succeed
+	 * without exhausting the page pool.
+	 */
+	if (!ffa_region_group_identity_map(from_locked, constituents,
+					   constituent_count, from_mode,
+					   page_pool, false)) {
+		/* TODO: partial defrag of failed range. */
+		ret = ffa_error(FFA_NO_MEMORY);
+		goto out;
+	}
+
+	/*
+	 * Update the mapping for the sender. This won't allocate because the
+	 * transaction was already prepared above, but may free pages in the
+	 * case that a whole block is being unmapped that was previously
+	 * partially mapped.
+	 */
+	CHECK(ffa_region_group_identity_map(from_locked, constituents,
+					    constituent_count, from_mode,
+					    &local_page_pool, true));
+
+	/* Clear the memory so no VM or device can see the previous contents. */
+	if (clear && !ffa_clear_memory_constituents(
+			     constituents, constituent_count, page_pool)) {
+		/*
+		 * On failure, roll back by returning memory to the sender. This
+		 * may allocate pages which were previously freed into
+		 * `local_page_pool` by the call above, but will never allocate
+		 * more pages than that so can never fail.
+		 */
+		CHECK(ffa_region_group_identity_map(
+			from_locked, constituents, constituent_count,
+			orig_from_mode, &local_page_pool, true));
+
+		ret = ffa_error(FFA_NO_MEMORY);
+		goto out;
+	}
+
+	ret = (struct ffa_value){.func = FFA_SUCCESS_32};
+
+out:
+	mpool_fini(&local_page_pool);
+
+	/*
+	 * Tidy up the page table by reclaiming failed mappings (if there was an
+	 * error) or merging entries into blocks where possible (on success).
+	 */
+	mm_vm_defrag(&from->ptable, page_pool);
+
+	return ret;
+}
+
+/**
+ * Validates and maps memory shared from one VM to another.
+ *
+ * This function requires the calling context to hold the <to> lock.
+ *
+ * Returns:
+ *  In case of error, one of the following values is returned:
+ *   1) FFA_INVALID_PARAMETERS - The endpoint provided parameters were
+ *     erroneous;
+ *   2) FFA_NO_MEMORY - Hafnium did not have sufficient memory to complete
+ *     the request.
+ *  Success is indicated by FFA_SUCCESS.
+ */
+static struct ffa_value ffa_retrieve_memory(
+	struct vm_locked to_locked,
+	struct ffa_memory_region_constituent *constituents,
+	uint32_t constituent_count, uint32_t memory_to_attributes,
+	uint32_t share_func, bool clear, struct mpool *page_pool)
+{
+	struct vm *to = to_locked.vm;
+	uint32_t to_mode;
+	struct mpool local_page_pool;
+	struct ffa_value ret;
+
+	/*
+	 * Make sure constituents are properly aligned to a 32-bit boundary. If
+	 * not we would get alignment faults trying to read (32-bit) values.
+	 */
+	if (!is_aligned(constituents, 4)) {
+		dlog_verbose("Constituents not aligned.\n");
+		return ffa_error(FFA_INVALID_PARAMETERS);
+	}
+
+	/*
+	 * Check if the state transition is lawful for the recipient, and ensure
+	 * that all constituents of the memory region being retrieved are at the
+	 * same state.
+	 */
+	ret = ffa_retrieve_check_transition(to_locked, share_func, constituents,
+					    constituent_count,
+					    memory_to_attributes, &to_mode);
+	if (ret.func != FFA_SUCCESS_32) {
+		dlog_verbose("Invalid transition.\n");
+		return ret;
+	}
+
+	/*
+	 * Create a local pool so any freed memory can't be used by another
+	 * thread. This is to ensure the original mapping can be restored if the
+	 * clear fails.
+	 */
+	mpool_init_with_fallback(&local_page_pool, page_pool);
+
+	/*
+	 * First reserve all required memory for the new page table entries in
+	 * the recipient page tables without committing, to make sure the entire
+	 * operation will succeed without exhausting the page pool.
+	 */
+	if (!ffa_region_group_identity_map(to_locked, constituents,
+					   constituent_count, to_mode,
+					   page_pool, false)) {
+		/* TODO: partial defrag of failed range. */
+		dlog_verbose(
+			"Insufficient memory to update recipient page "
+			"table.\n");
+		ret = ffa_error(FFA_NO_MEMORY);
+		goto out;
+	}
+
+	/* Clear the memory so no VM or device can see the previous contents. */
+	if (clear && !ffa_clear_memory_constituents(
+			     constituents, constituent_count, page_pool)) {
+		ret = ffa_error(FFA_NO_MEMORY);
+		goto out;
+	}
+
+	/*
+	 * Complete the transfer by mapping the memory into the recipient. This
+	 * won't allocate because the transaction was already prepared above, so
+	 * it doesn't need to use the `local_page_pool`.
+	 */
+	CHECK(ffa_region_group_identity_map(to_locked, constituents,
+					    constituent_count, to_mode,
+					    page_pool, true));
+
+	ret = (struct ffa_value){.func = FFA_SUCCESS_32};
+
+out:
+	mpool_fini(&local_page_pool);
+
+	/*
+	 * Tidy up the page table by reclaiming failed mappings (if there was
+	 * an error) or merging entries into blocks where possible (on success).
+	 */
+	mm_vm_defrag(&to->ptable, page_pool);
+
+	return ret;
+}
+
+/**
+ * Reclaims the given memory from the TEE. To do this space is first reserved in
+ * the <to> VM's page table, then the reclaim request is sent on to the TEE,
+ * then (if that is successful) the memory is mapped back into the <to> VM's
+ * page table.
+ *
+ * This function requires the calling context to hold the <to> lock.
+ *
+ * Returns:
+ *  In case of error, one of the following values is returned:
+ *   1) FFA_INVALID_PARAMETERS - The endpoint provided parameters were
+ *     erroneous;
+ *   2) FFA_NO_MEMORY - Hafnium did not have sufficient memory to complete
+ *     the request.
+ *  Success is indicated by FFA_SUCCESS.
+ */
+static struct ffa_value ffa_tee_reclaim_memory(
+	struct vm_locked to_locked, ffa_memory_handle_t handle,
+	struct ffa_memory_region_constituent *constituents,
+	uint32_t constituent_count, uint32_t memory_to_attributes, bool clear,
+	struct mpool *page_pool)
+{
+	struct vm *to = to_locked.vm;
+	uint32_t to_mode;
+	struct mpool local_page_pool;
+	struct ffa_value ret;
+	ffa_memory_region_flags_t tee_flags;
+
+	/*
+	 * Make sure constituents are properly aligned to a 32-bit boundary. If
+	 * not we would get alignment faults trying to read (32-bit) values.
+	 */
+	if (!is_aligned(constituents, 4)) {
+		dlog_verbose("Constituents not aligned.\n");
+		return ffa_error(FFA_INVALID_PARAMETERS);
+	}
+
+	/*
+	 * Check if the state transition is lawful for the recipient, and ensure
+	 * that all constituents of the memory region being retrieved are at the
+	 * same state.
+	 */
+	ret = ffa_retrieve_check_transition(to_locked, FFA_MEM_RECLAIM_32,
+					    constituents, constituent_count,
+					    memory_to_attributes, &to_mode);
+	if (ret.func != FFA_SUCCESS_32) {
+		dlog_verbose("Invalid transition.\n");
+		return ret;
+	}
+
+	/*
+	 * Create a local pool so any freed memory can't be used by another
+	 * thread. This is to ensure the original mapping can be restored if the
+	 * clear fails.
+	 */
+	mpool_init_with_fallback(&local_page_pool, page_pool);
+
+	/*
+	 * First reserve all required memory for the new page table entries in
+	 * the recipient page tables without committing, to make sure the entire
+	 * operation will succeed without exhausting the page pool.
+	 */
+	if (!ffa_region_group_identity_map(to_locked, constituents,
+					   constituent_count, to_mode,
+					   page_pool, false)) {
+		/* TODO: partial defrag of failed range. */
+		dlog_verbose(
+			"Insufficient memory to update recipient page "
+			"table.\n");
+		ret = ffa_error(FFA_NO_MEMORY);
+		goto out;
+	}
+
+	/*
+	 * Forward the request to the TEE and see what happens.
+	 */
+	tee_flags = 0;
+	if (clear) {
+		tee_flags |= FFA_MEMORY_REGION_FLAG_CLEAR;
+	}
+	ret = arch_tee_call((struct ffa_value){.func = FFA_MEM_RECLAIM_32,
+					       .arg1 = (uint32_t)handle,
+					       .arg2 = (uint32_t)(handle >> 32),
+					       .arg3 = tee_flags});
+
+	if (ret.func != FFA_SUCCESS_32) {
+		dlog_verbose(
+			"Got %#x (%d) from EL3 in response to "
+			"FFA_MEM_RECLAIM_32, expected FFA_SUCCESS_32.\n",
+			ret.func, ret.arg2);
+		goto out;
+	}
+
+	/*
+	 * The TEE was happy with it, so complete the reclaim by mapping the
+	 * memory into the recipient. This won't allocate because the
+	 * transaction was already prepared above, so it doesn't need to use the
+	 * `local_page_pool`.
+	 */
+	CHECK(ffa_region_group_identity_map(to_locked, constituents,
+					    constituent_count, to_mode,
+					    page_pool, true));
+
+	ret = (struct ffa_value){.func = FFA_SUCCESS_32};
+
+out:
+	mpool_fini(&local_page_pool);
+
+	/*
+	 * Tidy up the page table by reclaiming failed mappings (if there was
+	 * an error) or merging entries into blocks where possible (on success).
+	 */
+	mm_vm_defrag(&to->ptable, page_pool);
+
+	return ret;
+}
+
+static struct ffa_value ffa_relinquish_memory(
+	struct vm_locked from_locked,
+	struct ffa_memory_region_constituent *constituents,
+	uint32_t constituent_count, struct mpool *page_pool, bool clear)
+{
+	uint32_t orig_from_mode;
+	uint32_t from_mode;
+	struct mpool local_page_pool;
+	struct ffa_value ret;
+
+	ret = ffa_relinquish_check_transition(from_locked, &orig_from_mode,
+					      constituents, constituent_count,
+					      &from_mode);
+	if (ret.func != FFA_SUCCESS_32) {
+		dlog_verbose("Invalid transition.\n");
+		return ret;
+	}
+
+	/*
+	 * Create a local pool so any freed memory can't be used by another
+	 * thread. This is to ensure the original mapping can be restored if the
+	 * clear fails.
+	 */
+	mpool_init_with_fallback(&local_page_pool, page_pool);
+
+	/*
+	 * First reserve all required memory for the new page table entries
+	 * without committing, to make sure the entire operation will succeed
+	 * without exhausting the page pool.
+	 */
+	if (!ffa_region_group_identity_map(from_locked, constituents,
+					   constituent_count, from_mode,
+					   page_pool, false)) {
+		/* TODO: partial defrag of failed range. */
+		ret = ffa_error(FFA_NO_MEMORY);
+		goto out;
+	}
+
+	/*
+	 * Update the mapping for the sender. This won't allocate because the
+	 * transaction was already prepared above, but may free pages in the
+	 * case that a whole block is being unmapped that was previously
+	 * partially mapped.
+	 */
+	CHECK(ffa_region_group_identity_map(from_locked, constituents,
+					    constituent_count, from_mode,
+					    &local_page_pool, true));
+
+	/* Clear the memory so no VM or device can see the previous contents. */
+	if (clear && !ffa_clear_memory_constituents(
+			     constituents, constituent_count, page_pool)) {
+		/*
+		 * On failure, roll back by returning memory to the sender. This
+		 * may allocate pages which were previously freed into
+		 * `local_page_pool` by the call above, but will never allocate
+		 * more pages than that so can never fail.
+		 */
+		CHECK(ffa_region_group_identity_map(
+			from_locked, constituents, constituent_count,
+			orig_from_mode, &local_page_pool, true));
+
+		ret = ffa_error(FFA_NO_MEMORY);
+		goto out;
+	}
+
+	ret = (struct ffa_value){.func = FFA_SUCCESS_32};
+
+out:
+	mpool_fini(&local_page_pool);
+
+	/*
+	 * Tidy up the page table by reclaiming failed mappings (if there was an
+	 * error) or merging entries into blocks where possible (on success).
+	 */
+	mm_vm_defrag(&from_locked.vm->ptable, page_pool);
+
+	return ret;
+}
+
+/**
+ * Check that the given `memory_region` represents a valid memory send request
+ * of the given `share_func` type, return the clear flag and permissions via the
+ * respective output parameters, and update the permissions if necessary.
+ * Returns FFA_SUCCESS if the request was valid, or the relevant FFA_ERROR if
+ * not.
+ */
+static struct ffa_value ffa_memory_send_validate(
+	struct vm *to, struct vm_locked from_locked,
+	struct ffa_memory_region *memory_region, uint32_t memory_share_size,
+	uint32_t share_func, bool *clear,
+	ffa_memory_access_permissions_t *permissions)
+{
+	struct ffa_composite_memory_region *composite;
+	uint32_t receivers_size;
+	uint32_t constituents_size;
+	enum ffa_data_access data_access;
+	enum ffa_instruction_access instruction_access;
+
+	CHECK(clear != NULL);
+	CHECK(permissions != NULL);
+
+	/* The sender must match the message sender. */
+	if (memory_region->sender != from_locked.vm->id) {
+		dlog_verbose("Invalid sender %d.\n", memory_region->sender);
+		return ffa_error(FFA_INVALID_PARAMETERS);
+	}
+
+	/* We only support a single recipient. */
+	if (memory_region->receiver_count != 1) {
+		dlog_verbose("Multiple recipients not supported.\n");
+		return ffa_error(FFA_INVALID_PARAMETERS);
+	}
+
+	/*
+	 * Ensure that the composite header is within the memory bounds and
+	 * doesn't overlap the first part of the message.
+	 */
+	receivers_size = sizeof(struct ffa_memory_access) *
+			 memory_region->receiver_count;
+	if (memory_region->receivers[0].composite_memory_region_offset <
+		    sizeof(struct ffa_memory_region) + receivers_size ||
+	    memory_region->receivers[0].composite_memory_region_offset +
+			    sizeof(struct ffa_composite_memory_region) >=
+		    memory_share_size) {
+		dlog_verbose(
+			"Invalid composite memory region descriptor offset.\n");
+		return ffa_error(FFA_INVALID_PARAMETERS);
+	}
+
+	composite = ffa_memory_region_get_composite(memory_region, 0);
+
+	/*
+	 * Ensure the number of constituents are within the memory
+	 * bounds.
+	 */
+	constituents_size = sizeof(struct ffa_memory_region_constituent) *
+			    composite->constituent_count;
+	if (memory_share_size !=
+	    memory_region->receivers[0].composite_memory_region_offset +
+		    sizeof(struct ffa_composite_memory_region) +
+		    constituents_size) {
+		dlog_verbose("Invalid size %d or constituent offset %d.\n",
+			     memory_share_size,
+			     memory_region->receivers[0]
+				     .composite_memory_region_offset);
+		return ffa_error(FFA_INVALID_PARAMETERS);
+	}
+
+	/* The recipient must match the message recipient. */
+	if (memory_region->receivers[0].receiver_permissions.receiver !=
+	    to->id) {
+		return ffa_error(FFA_INVALID_PARAMETERS);
+	}
+
+	*clear = memory_region->flags & FFA_MEMORY_REGION_FLAG_CLEAR;
+	/*
+	 * Clear is not allowed for memory sharing, as the sender still has
+	 * access to the memory.
+	 */
+	if (*clear && share_func == FFA_MEM_SHARE_32) {
+		dlog_verbose("Memory can't be cleared while being shared.\n");
+		return ffa_error(FFA_INVALID_PARAMETERS);
+	}
+
+	/* No other flags are allowed/supported here. */
+	if (memory_region->flags & ~FFA_MEMORY_REGION_FLAG_CLEAR) {
+		dlog_verbose("Invalid flags %#x.\n", memory_region->flags);
+		return ffa_error(FFA_INVALID_PARAMETERS);
+	}
+
+	/* Check that the permissions are valid. */
+	*permissions =
+		memory_region->receivers[0].receiver_permissions.permissions;
+	data_access = ffa_get_data_access_attr(*permissions);
+	instruction_access = ffa_get_instruction_access_attr(*permissions);
+	if (data_access == FFA_DATA_ACCESS_RESERVED ||
+	    instruction_access == FFA_INSTRUCTION_ACCESS_RESERVED) {
+		dlog_verbose("Reserved value for receiver permissions %#x.\n",
+			     *permissions);
+		return ffa_error(FFA_INVALID_PARAMETERS);
+	}
+	if (instruction_access != FFA_INSTRUCTION_ACCESS_NOT_SPECIFIED) {
+		dlog_verbose(
+			"Invalid instruction access permissions %#x for "
+			"sending memory.\n",
+			*permissions);
+		return ffa_error(FFA_INVALID_PARAMETERS);
+	}
+	if (share_func == FFA_MEM_SHARE_32) {
+		if (data_access == FFA_DATA_ACCESS_NOT_SPECIFIED) {
+			dlog_verbose(
+				"Invalid data access permissions %#x for "
+				"sharing memory.\n",
+				*permissions);
+			return ffa_error(FFA_INVALID_PARAMETERS);
+		}
+		/*
+		 * According to section 6.11.3 of the FF-A spec NX is required
+		 * for share operations (but must not be specified by the
+		 * sender) so set it in the copy that we store, ready to be
+		 * returned to the retriever.
+		 */
+		ffa_set_instruction_access_attr(permissions,
+						FFA_INSTRUCTION_ACCESS_NX);
+		memory_region->receivers[0].receiver_permissions.permissions =
+			*permissions;
+	}
+	if (share_func == FFA_MEM_LEND_32 &&
+	    data_access == FFA_DATA_ACCESS_NOT_SPECIFIED) {
+		dlog_verbose(
+			"Invalid data access permissions %#x for lending "
+			"memory.\n",
+			*permissions);
+		return ffa_error(FFA_INVALID_PARAMETERS);
+	}
+	if (share_func == FFA_MEM_DONATE_32 &&
+	    data_access != FFA_DATA_ACCESS_NOT_SPECIFIED) {
+		dlog_verbose(
+			"Invalid data access permissions %#x for donating "
+			"memory.\n",
+			*permissions);
+		return ffa_error(FFA_INVALID_PARAMETERS);
+	}
+
+	return (struct ffa_value){.func = FFA_SUCCESS_32};
+}
+
+/**
+ * Validates a call to donate, lend or share memory and then updates the stage-2
+ * page tables. Specifically, check if the message length and number of memory
+ * region constituents match, and if the transition is valid for the type of
+ * memory sending operation.
+ *
+ * Assumes that the caller has already found and locked both VMs and ensured
+ * that the destination RX buffer is available, and copied the memory region
+ * descriptor from the sender's TX buffer to a freshly allocated page from
+ * Hafnium's internal pool.
+ *
+ * This function takes ownership of the `memory_region` passed in; it must not
+ * be freed by the caller.
+ */
+struct ffa_value ffa_memory_send(struct vm *to, struct vm_locked from_locked,
+				 struct ffa_memory_region *memory_region,
+				 uint32_t memory_share_size,
+				 uint32_t share_func, struct mpool *page_pool)
+{
+	struct ffa_composite_memory_region *composite;
+	bool clear;
+	ffa_memory_access_permissions_t permissions;
+	struct ffa_value ret;
+	ffa_memory_handle_t handle;
+
+	/*
+	 * If there is an error validating the `memory_region` then we need to
+	 * free it because we own it but we won't be storing it in a share state
+	 * after all.
+	 */
+	ret = ffa_memory_send_validate(to, from_locked, memory_region,
+				       memory_share_size, share_func, &clear,
+				       &permissions);
+	if (ret.func != FFA_SUCCESS_32) {
+		mpool_free(page_pool, memory_region);
+		return ret;
+	}
+
+	/* Set flag for share function, ready to be retrieved later. */
+	switch (share_func) {
+	case FFA_MEM_SHARE_32:
+		memory_region->flags |=
+			FFA_MEMORY_REGION_TRANSACTION_TYPE_SHARE;
+		break;
+	case FFA_MEM_LEND_32:
+		memory_region->flags |= FFA_MEMORY_REGION_TRANSACTION_TYPE_LEND;
+		break;
+	case FFA_MEM_DONATE_32:
+		memory_region->flags |=
+			FFA_MEMORY_REGION_TRANSACTION_TYPE_DONATE;
+		break;
+	}
+
+	/*
+	 * Allocate a share state before updating the page table. Otherwise if
+	 * updating the page table succeeded but allocating the share state
+	 * failed then it would leave the memory in a state where nobody could
+	 * get it back.
+	 */
+	if (to->id != HF_TEE_VM_ID &&
+	    !allocate_share_state(share_func, memory_region, &handle)) {
+		dlog_verbose("Failed to allocate share state.\n");
+		mpool_free(page_pool, memory_region);
+		return ffa_error(FFA_NO_MEMORY);
+	}
+
+	dump_share_states();
+
+	/* Check that state is valid in sender page table and update. */
+	composite = ffa_memory_region_get_composite(memory_region, 0);
+	ret = ffa_send_memory(from_locked, composite->constituents,
+			      composite->constituent_count, share_func,
+			      permissions, page_pool, clear);
+	if (ret.func != FFA_SUCCESS_32) {
+		if (to->id != HF_TEE_VM_ID) {
+			/* Free share state. */
+			bool freed = share_state_free_handle(handle, page_pool);
+
+			CHECK(freed);
+		}
+
+		return ret;
+	}
+
+	if (to->id == HF_TEE_VM_ID) {
+		/* No share state allocated here so no handle to return. */
+		return (struct ffa_value){.func = FFA_SUCCESS_32};
+	}
+
+	return (struct ffa_value){.func = FFA_SUCCESS_32, .arg2 = handle};
+}
+
+struct ffa_value ffa_memory_retrieve(struct vm_locked to_locked,
+				     struct ffa_memory_region *retrieve_request,
+				     uint32_t retrieve_request_size,
+				     struct mpool *page_pool)
+{
+	uint32_t expected_retrieve_request_size =
+		sizeof(struct ffa_memory_region) +
+		retrieve_request->receiver_count *
+			sizeof(struct ffa_memory_access);
+	ffa_memory_handle_t handle = retrieve_request->handle;
+	ffa_memory_region_flags_t transaction_type =
+		retrieve_request->flags &
+		FFA_MEMORY_REGION_TRANSACTION_TYPE_MASK;
+	struct ffa_memory_region *memory_region;
+	ffa_memory_access_permissions_t sent_permissions;
+	enum ffa_data_access sent_data_access;
+	enum ffa_instruction_access sent_instruction_access;
+	ffa_memory_access_permissions_t requested_permissions;
+	enum ffa_data_access requested_data_access;
+	enum ffa_instruction_access requested_instruction_access;
+	ffa_memory_access_permissions_t permissions;
+	uint32_t memory_to_attributes;
+	struct ffa_composite_memory_region *composite;
+	struct share_states_locked share_states;
+	struct ffa_memory_share_state *share_state;
+	struct ffa_value ret;
+	uint32_t response_size;
+
+	dump_share_states();
+
+	if (retrieve_request_size != expected_retrieve_request_size) {
+		dlog_verbose(
+			"Invalid length for FFA_MEM_RETRIEVE_REQ, expected %d "
+			"but was %d.\n",
+			expected_retrieve_request_size, retrieve_request_size);
+		return ffa_error(FFA_INVALID_PARAMETERS);
+	}
+
+	if (retrieve_request->receiver_count != 1) {
+		dlog_verbose(
+			"Multi-way memory sharing not supported (got %d "
+			"receivers descriptors on FFA_MEM_RETRIEVE_REQ, "
+			"expected 1).\n",
+			retrieve_request->receiver_count);
+		return ffa_error(FFA_INVALID_PARAMETERS);
+	}
+
+	share_states = share_states_lock();
+	if (!get_share_state(share_states, handle, &share_state)) {
+		dlog_verbose("Invalid handle %#x for FFA_MEM_RETRIEVE_REQ.\n",
+			     handle);
+		ret = ffa_error(FFA_INVALID_PARAMETERS);
+		goto out;
+	}
+
+	memory_region = share_state->memory_region;
+	CHECK(memory_region != NULL);
+
+	/*
+	 * Check that the transaction type expected by the receiver is correct,
+	 * if it has been specified.
+	 */
+	if (transaction_type !=
+		    FFA_MEMORY_REGION_TRANSACTION_TYPE_UNSPECIFIED &&
+	    transaction_type != (memory_region->flags &
+				 FFA_MEMORY_REGION_TRANSACTION_TYPE_MASK)) {
+		dlog_verbose(
+			"Incorrect transaction type %#x for "
+			"FFA_MEM_RETRIEVE_REQ, expected %#x for handle %#x.\n",
+			transaction_type,
+			memory_region->flags &
+				FFA_MEMORY_REGION_TRANSACTION_TYPE_MASK,
+			handle);
+		ret = ffa_error(FFA_INVALID_PARAMETERS);
+		goto out;
+	}
+
+	if (retrieve_request->sender != memory_region->sender) {
+		dlog_verbose(
+			"Incorrect sender ID %d for FFA_MEM_RETRIEVE_REQ, "
+			"expected %d for handle %#x.\n",
+			retrieve_request->sender, memory_region->sender,
+			handle);
+		ret = ffa_error(FFA_INVALID_PARAMETERS);
+		goto out;
+	}
+
+	if (retrieve_request->tag != memory_region->tag) {
+		dlog_verbose(
+			"Incorrect tag %d for FFA_MEM_RETRIEVE_REQ, expected "
+			"%d for handle %#x.\n",
+			retrieve_request->tag, memory_region->tag, handle);
+		ret = ffa_error(FFA_INVALID_PARAMETERS);
+		goto out;
+	}
+
+	if (retrieve_request->receivers[0].receiver_permissions.receiver !=
+	    to_locked.vm->id) {
+		dlog_verbose(
+			"Retrieve request receiver VM ID %d didn't match "
+			"caller of FFA_MEM_RETRIEVE_REQ.\n",
+			retrieve_request->receivers[0]
+				.receiver_permissions.receiver);
+		ret = ffa_error(FFA_INVALID_PARAMETERS);
+		goto out;
+	}
+
+	if (memory_region->receivers[0].receiver_permissions.receiver !=
+	    to_locked.vm->id) {
+		dlog_verbose(
+			"Incorrect receiver VM ID %d for "
+			"FFA_MEM_RETRIEVE_REQ, expected %d for handle %#x.\n",
+			to_locked.vm->id,
+			memory_region->receivers[0]
+				.receiver_permissions.receiver,
+			handle);
+		ret = ffa_error(FFA_INVALID_PARAMETERS);
+		goto out;
+	}
+
+	if (share_state->retrieved[0]) {
+		dlog_verbose("Memory with handle %#x already retrieved.\n",
+			     handle);
+		ret = ffa_error(FFA_DENIED);
+		goto out;
+	}
+
+	if (retrieve_request->receivers[0].composite_memory_region_offset !=
+	    0) {
+		dlog_verbose(
+			"Retriever specified address ranges not supported (got "
+			"offset"
+			"%d).\n",
+			retrieve_request->receivers[0]
+				.composite_memory_region_offset);
+		ret = ffa_error(FFA_INVALID_PARAMETERS);
+		goto out;
+	}
+
+	/*
+	 * Check permissions from sender against permissions requested by
+	 * receiver.
+	 */
+	/* TODO: Check attributes too. */
+	sent_permissions =
+		memory_region->receivers[0].receiver_permissions.permissions;
+	sent_data_access = ffa_get_data_access_attr(sent_permissions);
+	sent_instruction_access =
+		ffa_get_instruction_access_attr(sent_permissions);
+	requested_permissions =
+		retrieve_request->receivers[0].receiver_permissions.permissions;
+	requested_data_access = ffa_get_data_access_attr(requested_permissions);
+	requested_instruction_access =
+		ffa_get_instruction_access_attr(requested_permissions);
+	permissions = 0;
+	switch (sent_data_access) {
+	case FFA_DATA_ACCESS_NOT_SPECIFIED:
+	case FFA_DATA_ACCESS_RW:
+		if (requested_data_access == FFA_DATA_ACCESS_NOT_SPECIFIED ||
+		    requested_data_access == FFA_DATA_ACCESS_RW) {
+			ffa_set_data_access_attr(&permissions,
+						 FFA_DATA_ACCESS_RW);
+			break;
+		}
+		/* Intentional fall-through. */
+	case FFA_DATA_ACCESS_RO:
+		if (requested_data_access == FFA_DATA_ACCESS_NOT_SPECIFIED ||
+		    requested_data_access == FFA_DATA_ACCESS_RO) {
+			ffa_set_data_access_attr(&permissions,
+						 FFA_DATA_ACCESS_RO);
+			break;
+		}
+		dlog_verbose(
+			"Invalid data access requested; sender specified "
+			"permissions %#x but receiver requested %#x.\n",
+			sent_permissions, requested_permissions);
+		ret = ffa_error(FFA_DENIED);
+		goto out;
+	case FFA_DATA_ACCESS_RESERVED:
+		panic("Got unexpected FFA_DATA_ACCESS_RESERVED. Should be "
+		      "checked before this point.");
+	}
+	switch (sent_instruction_access) {
+	case FFA_INSTRUCTION_ACCESS_NOT_SPECIFIED:
+	case FFA_INSTRUCTION_ACCESS_X:
+		if (requested_instruction_access ==
+			    FFA_INSTRUCTION_ACCESS_NOT_SPECIFIED ||
+		    requested_instruction_access == FFA_INSTRUCTION_ACCESS_X) {
+			ffa_set_instruction_access_attr(
+				&permissions, FFA_INSTRUCTION_ACCESS_X);
+			break;
+		}
+	case FFA_INSTRUCTION_ACCESS_NX:
+		if (requested_instruction_access ==
+			    FFA_INSTRUCTION_ACCESS_NOT_SPECIFIED ||
+		    requested_instruction_access == FFA_INSTRUCTION_ACCESS_NX) {
+			ffa_set_instruction_access_attr(
+				&permissions, FFA_INSTRUCTION_ACCESS_NX);
+			break;
+		}
+		dlog_verbose(
+			"Invalid instruction access requested; sender "
+			"specified "
+			"permissions %#x but receiver requested %#x.\n",
+			sent_permissions, requested_permissions);
+		ret = ffa_error(FFA_DENIED);
+		goto out;
+	case FFA_INSTRUCTION_ACCESS_RESERVED:
+		panic("Got unexpected FFA_INSTRUCTION_ACCESS_RESERVED. Should "
+		      "be checked before this point.");
+	}
+	memory_to_attributes = ffa_memory_permissions_to_mode(permissions);
+
+	composite = ffa_memory_region_get_composite(memory_region, 0);
+	ret = ffa_retrieve_memory(to_locked, composite->constituents,
+				  composite->constituent_count,
+				  memory_to_attributes, share_state->share_func,
+				  false, page_pool);
+	if (ret.func != FFA_SUCCESS_32) {
+		goto out;
+	}
+
+	/*
+	 * Copy response to RX buffer of caller and deliver the message. This
+	 * must be done before the share_state is (possibly) freed.
+	 */
+	/* TODO: combine attributes from sender and request. */
+	response_size = ffa_retrieved_memory_region_init(
+		to_locked.vm->mailbox.recv, HF_MAILBOX_SIZE,
+		memory_region->sender, memory_region->attributes,
+		memory_region->flags, handle, to_locked.vm->id, permissions,
+		composite->constituents, composite->constituent_count);
+	to_locked.vm->mailbox.recv_size = response_size;
+	to_locked.vm->mailbox.recv_sender = HF_HYPERVISOR_VM_ID;
+	to_locked.vm->mailbox.recv_func = FFA_MEM_RETRIEVE_RESP_32;
+	to_locked.vm->mailbox.state = MAILBOX_STATE_READ;
+
+	if (share_state->share_func == FFA_MEM_DONATE_32) {
+		/*
+		 * Memory that has been donated can't be relinquished, so no
+		 * need to keep the share state around.
+		 */
+		share_state_free(share_states, share_state, page_pool);
+		dlog_verbose("Freed share state for donate.\n");
+	} else {
+		share_state->retrieved[0] = true;
+	}
+
+	ret = (struct ffa_value){.func = FFA_MEM_RETRIEVE_RESP_32,
+				 .arg1 = response_size,
+				 .arg2 = response_size};
+
+out:
+	share_states_unlock(&share_states);
+	dump_share_states();
+	return ret;
+}
+
+struct ffa_value ffa_memory_relinquish(
+	struct vm_locked from_locked,
+	struct ffa_mem_relinquish *relinquish_request, struct mpool *page_pool)
+{
+	ffa_memory_handle_t handle = relinquish_request->handle;
+	struct share_states_locked share_states;
+	struct ffa_memory_share_state *share_state;
+	struct ffa_memory_region *memory_region;
+	bool clear;
+	struct ffa_composite_memory_region *composite;
+	struct ffa_value ret;
+
+	if (relinquish_request->endpoint_count != 1) {
+		dlog_verbose(
+			"Stream endpoints not supported (got %d endpoints on "
+			"FFA_MEM_RELINQUISH, expected 1).\n",
+			relinquish_request->endpoint_count);
+		return ffa_error(FFA_INVALID_PARAMETERS);
+	}
+
+	if (relinquish_request->endpoints[0] != from_locked.vm->id) {
+		dlog_verbose(
+			"VM ID %d in relinquish message doesn't match calling "
+			"VM ID %d.\n",
+			relinquish_request->endpoints[0], from_locked.vm->id);
+		return ffa_error(FFA_INVALID_PARAMETERS);
+	}
+
+	dump_share_states();
+
+	share_states = share_states_lock();
+	if (!get_share_state(share_states, handle, &share_state)) {
+		dlog_verbose("Invalid handle %#x for FFA_MEM_RELINQUISH.\n",
+			     handle);
+		ret = ffa_error(FFA_INVALID_PARAMETERS);
+		goto out;
+	}
+
+	memory_region = share_state->memory_region;
+	CHECK(memory_region != NULL);
+
+	if (memory_region->receivers[0].receiver_permissions.receiver !=
+	    from_locked.vm->id) {
+		dlog_verbose(
+			"VM ID %d tried to relinquish memory region with "
+			"handle %#x but receiver was %d.\n",
+			from_locked.vm->id, handle,
+			memory_region->receivers[0]
+				.receiver_permissions.receiver);
+		ret = ffa_error(FFA_INVALID_PARAMETERS);
+		goto out;
+	}
+
+	if (!share_state->retrieved[0]) {
+		dlog_verbose(
+			"Memory with handle %#x not yet retrieved, can't "
+			"relinquish.\n",
+			handle);
+		ret = ffa_error(FFA_INVALID_PARAMETERS);
+		goto out;
+	}
+
+	clear = relinquish_request->flags & FFA_MEMORY_REGION_FLAG_CLEAR;
+
+	/*
+	 * Clear is not allowed for memory that was shared, as the original
+	 * sender still has access to the memory.
+	 */
+	if (clear && share_state->share_func == FFA_MEM_SHARE_32) {
+		dlog_verbose("Memory which was shared can't be cleared.\n");
+		ret = ffa_error(FFA_INVALID_PARAMETERS);
+		goto out;
+	}
+
+	composite = ffa_memory_region_get_composite(memory_region, 0);
+	ret = ffa_relinquish_memory(from_locked, composite->constituents,
+				    composite->constituent_count, page_pool,
+				    clear);
+
+	if (ret.func == FFA_SUCCESS_32) {
+		/*
+		 * Mark memory handle as not retrieved, so it can be reclaimed
+		 * (or retrieved again).
+		 */
+		share_state->retrieved[0] = false;
+	}
+
+out:
+	share_states_unlock(&share_states);
+	dump_share_states();
+	return ret;
+}
+
+/**
+ * Validates that the reclaim transition is allowed for the given handle,
+ * updates the page table of the reclaiming VM, and frees the internal state
+ * associated with the handle.
+ */
+struct ffa_value ffa_memory_reclaim(struct vm_locked to_locked,
+				    ffa_memory_handle_t handle, bool clear,
+				    struct mpool *page_pool)
+{
+	struct share_states_locked share_states;
+	struct ffa_memory_share_state *share_state;
+	struct ffa_memory_region *memory_region;
+	struct ffa_composite_memory_region *composite;
+	uint32_t memory_to_attributes = MM_MODE_R | MM_MODE_W | MM_MODE_X;
+	struct ffa_value ret;
+
+	dump_share_states();
+
+	share_states = share_states_lock();
+	if (!get_share_state(share_states, handle, &share_state)) {
+		dlog_verbose("Invalid handle %#x for FFA_MEM_RECLAIM.\n",
+			     handle);
+		ret = ffa_error(FFA_INVALID_PARAMETERS);
+		goto out;
+	}
+
+	memory_region = share_state->memory_region;
+	CHECK(memory_region != NULL);
+
+	if (to_locked.vm->id != memory_region->sender) {
+		dlog_verbose(
+			"VM %d attempted to reclaim memory handle %#x "
+			"originally sent by VM %d.\n",
+			to_locked.vm->id, handle, memory_region->sender);
+		ret = ffa_error(FFA_INVALID_PARAMETERS);
+		goto out;
+	}
+
+	if (share_state->retrieved[0]) {
+		dlog_verbose(
+			"Tried to reclaim memory handle %#x that has not been "
+			"relinquished.\n",
+			handle);
+		ret = ffa_error(FFA_DENIED);
+		goto out;
+	}
+
+	composite = ffa_memory_region_get_composite(memory_region, 0);
+	ret = ffa_retrieve_memory(to_locked, composite->constituents,
+				  composite->constituent_count,
+				  memory_to_attributes, FFA_MEM_RECLAIM_32,
+				  clear, page_pool);
+
+	if (ret.func == FFA_SUCCESS_32) {
+		share_state_free(share_states, share_state, page_pool);
+		dlog_verbose("Freed share state after successful reclaim.\n");
+	}
+
+out:
+	share_states_unlock(&share_states);
+	return ret;
+}
+
+/**
+ * Validates that the reclaim transition is allowed for the given memory region
+ * and updates the page table of the reclaiming VM.
+ */
+struct ffa_value ffa_memory_tee_reclaim(struct vm_locked to_locked,
+					ffa_memory_handle_t handle,
+					struct ffa_memory_region *memory_region,
+					bool clear, struct mpool *page_pool)
+{
+	uint32_t memory_to_attributes = MM_MODE_R | MM_MODE_W | MM_MODE_X;
+	struct ffa_composite_memory_region *composite;
+
+	if (memory_region->receiver_count != 1) {
+		/* Only one receiver supported by Hafnium for now. */
+		dlog_verbose(
+			"Multiple recipients not supported (got %d, expected "
+			"1).\n",
+			memory_region->receiver_count);
+		return ffa_error(FFA_NOT_SUPPORTED);
+	}
+
+	if (memory_region->handle != handle) {
+		dlog_verbose(
+			"Got memory region handle %#x from TEE but requested "
+			"handle %#x.\n",
+			memory_region->handle, handle);
+		return ffa_error(FFA_INVALID_PARAMETERS);
+	}
+
+	/* The original sender must match the caller. */
+	if (to_locked.vm->id != memory_region->sender) {
+		dlog_verbose(
+			"VM %d attempted to reclaim memory handle %#x "
+			"originally sent by VM %d.\n",
+			to_locked.vm->id, handle, memory_region->sender);
+		return ffa_error(FFA_INVALID_PARAMETERS);
+	}
+
+	composite = ffa_memory_region_get_composite(memory_region, 0);
+
+	/*
+	 * Forward the request to the TEE and then map the memory back into the
+	 * caller's stage-2 page table.
+	 */
+	return ffa_tee_reclaim_memory(to_locked, handle,
+				      composite->constituents,
+				      composite->constituent_count,
+				      memory_to_attributes, clear, page_pool);
+}