+static void
+mempool_add_elem(struct rte_mempool *mp, void *obj, phys_addr_t physaddr)
+{
+ struct rte_mempool_objhdr *hdr;
+ struct rte_mempool_objtlr *tlr __rte_unused;
+
+ /* set mempool ptr in header */
+ hdr = RTE_PTR_SUB(obj, sizeof(*hdr));
+ hdr->mp = mp;
+ hdr->physaddr = physaddr;
+ STAILQ_INSERT_TAIL(&mp->elt_list, hdr, next);
+ mp->populated_size++;
+
+#ifdef RTE_LIBRTE_MEMPOOL_DEBUG
+ hdr->cookie = RTE_MEMPOOL_HEADER_COOKIE2;
+ tlr = __mempool_get_trailer(obj);
+ tlr->cookie = RTE_MEMPOOL_TRAILER_COOKIE;
+#endif
+
+ /* enqueue in ring */
+ rte_mempool_ops_enqueue_bulk(mp, &obj, 1);
+}
+
+/* call obj_cb() for each mempool element */
+uint32_t
+rte_mempool_obj_iter(struct rte_mempool *mp,
+ rte_mempool_obj_cb_t *obj_cb, void *obj_cb_arg)
+{
+ struct rte_mempool_objhdr *hdr;
+ void *obj;
+ unsigned n = 0;
+
+ STAILQ_FOREACH(hdr, &mp->elt_list, next) {
+ obj = (char *)hdr + sizeof(*hdr);
+ obj_cb(mp, obj_cb_arg, obj, n);
+ n++;
+ }
+
+ return n;
+}
+
+/* call mem_cb() for each mempool memory chunk */
+uint32_t
+rte_mempool_mem_iter(struct rte_mempool *mp,
+ rte_mempool_mem_cb_t *mem_cb, void *mem_cb_arg)
+{
+ struct rte_mempool_memhdr *hdr;
+ unsigned n = 0;
+
+ STAILQ_FOREACH(hdr, &mp->mem_list, next) {
+ mem_cb(mp, mem_cb_arg, hdr, n);
+ n++;
+ }
+
+ return n;
+}
+
+/* get the header, trailer and total size of a mempool element. */
+uint32_t
+rte_mempool_calc_obj_size(uint32_t elt_size, uint32_t flags,
+ struct rte_mempool_objsz *sz)
+{
+ struct rte_mempool_objsz lsz;
+
+ sz = (sz != NULL) ? sz : &lsz;
+
+ sz->header_size = sizeof(struct rte_mempool_objhdr);
+ if ((flags & MEMPOOL_F_NO_CACHE_ALIGN) == 0)
+ sz->header_size = RTE_ALIGN_CEIL(sz->header_size,
+ RTE_MEMPOOL_ALIGN);
+
+ sz->trailer_size = sizeof(struct rte_mempool_objtlr);
+
+ /* element size is 8 bytes-aligned at least */
+ sz->elt_size = RTE_ALIGN_CEIL(elt_size, sizeof(uint64_t));
+
+ /* expand trailer to next cache line */
+ if ((flags & MEMPOOL_F_NO_CACHE_ALIGN) == 0) {
+ sz->total_size = sz->header_size + sz->elt_size +
+ sz->trailer_size;
+ sz->trailer_size += ((RTE_MEMPOOL_ALIGN -
+ (sz->total_size & RTE_MEMPOOL_ALIGN_MASK)) &
+ RTE_MEMPOOL_ALIGN_MASK);
+ }
+
+ /*
+ * increase trailer to add padding between objects in order to
+ * spread them across memory channels/ranks
+ */
+ if ((flags & MEMPOOL_F_NO_SPREAD) == 0) {
+ unsigned new_size;
+ new_size = optimize_object_size(sz->header_size + sz->elt_size +
+ sz->trailer_size);
+ sz->trailer_size = new_size - sz->header_size - sz->elt_size;
+ }
+
+ /* this is the size of an object, including header and trailer */
+ sz->total_size = sz->header_size + sz->elt_size + sz->trailer_size;
+
+ return sz->total_size;
+}
+
+
+/*
+ * Calculate maximum amount of memory required to store given number of objects.
+ */
+size_t
+rte_mempool_xmem_size(uint32_t elt_num, size_t total_elt_sz, uint32_t pg_shift)
+{
+ size_t obj_per_page, pg_num, pg_sz;
+
+ if (total_elt_sz == 0)
+ return 0;
+
+ if (pg_shift == 0)
+ return total_elt_sz * elt_num;
+
+ pg_sz = (size_t)1 << pg_shift;
+ obj_per_page = pg_sz / total_elt_sz;
+ if (obj_per_page == 0)
+ return RTE_ALIGN_CEIL(total_elt_sz, pg_sz) * elt_num;
+
+ pg_num = (elt_num + obj_per_page - 1) / obj_per_page;
+ return pg_num << pg_shift;
+}
+
+/*
+ * Calculate how much memory would be actually required with the
+ * given memory footprint to store required number of elements.
+ */
+ssize_t
+rte_mempool_xmem_usage(__rte_unused void *vaddr, uint32_t elt_num,
+ size_t total_elt_sz, const phys_addr_t paddr[], uint32_t pg_num,
+ uint32_t pg_shift)
+{
+ uint32_t elt_cnt = 0;
+ phys_addr_t start, end;
+ uint32_t paddr_idx;
+ size_t pg_sz = (size_t)1 << pg_shift;
+
+ /* if paddr is NULL, assume contiguous memory */
+ if (paddr == NULL) {
+ start = 0;
+ end = pg_sz * pg_num;
+ paddr_idx = pg_num;
+ } else {
+ start = paddr[0];
+ end = paddr[0] + pg_sz;
+ paddr_idx = 1;
+ }
+ while (elt_cnt < elt_num) {
+
+ if (end - start >= total_elt_sz) {
+ /* enough contiguous memory, add an object */
+ start += total_elt_sz;
+ elt_cnt++;
+ } else if (paddr_idx < pg_num) {
+ /* no room to store one obj, add a page */
+ if (end == paddr[paddr_idx]) {
+ end += pg_sz;
+ } else {
+ start = paddr[paddr_idx];
+ end = paddr[paddr_idx] + pg_sz;
+ }
+ paddr_idx++;
+
+ } else {
+ /* no more page, return how many elements fit */
+ return -(size_t)elt_cnt;
+ }
+ }
+
+ return (size_t)paddr_idx << pg_shift;
+}
+
+/* free a memchunk allocated with rte_memzone_reserve() */
+static void
+rte_mempool_memchunk_mz_free(__rte_unused struct rte_mempool_memhdr *memhdr,
+ void *opaque)
+{
+ const struct rte_memzone *mz = opaque;
+ rte_memzone_free(mz);
+}
+
+/* Free memory chunks used by a mempool. Objects must be in pool */
+static void
+rte_mempool_free_memchunks(struct rte_mempool *mp)
+{
+ struct rte_mempool_memhdr *memhdr;
+ void *elt;
+
+ while (!STAILQ_EMPTY(&mp->elt_list)) {
+ rte_mempool_ops_dequeue_bulk(mp, &elt, 1);
+ (void)elt;
+ STAILQ_REMOVE_HEAD(&mp->elt_list, next);
+ mp->populated_size--;
+ }
+
+ while (!STAILQ_EMPTY(&mp->mem_list)) {
+ memhdr = STAILQ_FIRST(&mp->mem_list);
+ STAILQ_REMOVE_HEAD(&mp->mem_list, next);
+ if (memhdr->free_cb != NULL)
+ memhdr->free_cb(memhdr, memhdr->opaque);
+ rte_free(memhdr);
+ mp->nb_mem_chunks--;
+ }
+}
+
+/* Add objects in the pool, using a physically contiguous memory
+ * zone. Return the number of objects added, or a negative value
+ * on error.
+ */
+int
+rte_mempool_populate_phys(struct rte_mempool *mp, char *vaddr,
+ phys_addr_t paddr, size_t len, rte_mempool_memchunk_free_cb_t *free_cb,
+ void *opaque)
+{
+ unsigned total_elt_sz;
+ unsigned i = 0;
+ size_t off;
+ struct rte_mempool_memhdr *memhdr;
+ int ret;
+
+ /* create the internal ring if not already done */
+ if ((mp->flags & MEMPOOL_F_POOL_CREATED) == 0) {
+ ret = rte_mempool_ops_alloc(mp);
+ if (ret != 0)
+ return ret;
+ mp->flags |= MEMPOOL_F_POOL_CREATED;
+ }
+
+ /* mempool is already populated */
+ if (mp->populated_size >= mp->size)
+ return -ENOSPC;
+
+ total_elt_sz = mp->header_size + mp->elt_size + mp->trailer_size;
+
+ memhdr = rte_zmalloc("MEMPOOL_MEMHDR", sizeof(*memhdr), 0);
+ if (memhdr == NULL)
+ return -ENOMEM;
+
+ memhdr->mp = mp;
+ memhdr->addr = vaddr;
+ memhdr->phys_addr = paddr;
+ memhdr->len = len;
+ memhdr->free_cb = free_cb;
+ memhdr->opaque = opaque;
+
+ if (mp->flags & MEMPOOL_F_NO_CACHE_ALIGN)
+ off = RTE_PTR_ALIGN_CEIL(vaddr, 8) - vaddr;
+ else
+ off = RTE_PTR_ALIGN_CEIL(vaddr, RTE_CACHE_LINE_SIZE) - vaddr;
+
+ while (off + total_elt_sz <= len && mp->populated_size < mp->size) {
+ off += mp->header_size;
+ if (paddr == RTE_BAD_PHYS_ADDR)
+ mempool_add_elem(mp, (char *)vaddr + off,
+ RTE_BAD_PHYS_ADDR);
+ else
+ mempool_add_elem(mp, (char *)vaddr + off, paddr + off);
+ off += mp->elt_size + mp->trailer_size;
+ i++;
+ }
+
+ /* not enough room to store one object */
+ if (i == 0)
+ return -EINVAL;
+
+ STAILQ_INSERT_TAIL(&mp->mem_list, memhdr, next);
+ mp->nb_mem_chunks++;
+ return i;
+}
+
+/* Add objects in the pool, using a table of physical pages. Return the
+ * number of objects added, or a negative value on error.
+ */
+int
+rte_mempool_populate_phys_tab(struct rte_mempool *mp, char *vaddr,
+ const phys_addr_t paddr[], uint32_t pg_num, uint32_t pg_shift,
+ rte_mempool_memchunk_free_cb_t *free_cb, void *opaque)
+{
+ uint32_t i, n;
+ int ret, cnt = 0;
+ size_t pg_sz = (size_t)1 << pg_shift;
+
+ /* mempool must not be populated */
+ if (mp->nb_mem_chunks != 0)
+ return -EEXIST;
+
+ if (mp->flags & MEMPOOL_F_NO_PHYS_CONTIG)
+ return rte_mempool_populate_phys(mp, vaddr, RTE_BAD_PHYS_ADDR,
+ pg_num * pg_sz, free_cb, opaque);
+
+ for (i = 0; i < pg_num && mp->populated_size < mp->size; i += n) {
+
+ /* populate with the largest group of contiguous pages */
+ for (n = 1; (i + n) < pg_num &&
+ paddr[i] + pg_sz == paddr[i+n]; n++)
+ ;
+
+ ret = rte_mempool_populate_phys(mp, vaddr + i * pg_sz,
+ paddr[i], n * pg_sz, free_cb, opaque);
+ if (ret < 0) {
+ rte_mempool_free_memchunks(mp);
+ return ret;
+ }
+ /* no need to call the free callback for next chunks */
+ free_cb = NULL;
+ cnt += ret;
+ }
+ return cnt;
+}
+
+/* Populate the mempool with a virtual area. Return the number of
+ * objects added, or a negative value on error.
+ */
+int
+rte_mempool_populate_virt(struct rte_mempool *mp, char *addr,
+ size_t len, size_t pg_sz, rte_mempool_memchunk_free_cb_t *free_cb,
+ void *opaque)
+{
+ phys_addr_t paddr;
+ size_t off, phys_len;
+ int ret, cnt = 0;
+
+ /* mempool must not be populated */
+ if (mp->nb_mem_chunks != 0)
+ return -EEXIST;
+ /* address and len must be page-aligned */
+ if (RTE_PTR_ALIGN_CEIL(addr, pg_sz) != addr)
+ return -EINVAL;
+ if (RTE_ALIGN_CEIL(len, pg_sz) != len)
+ return -EINVAL;
+
+ if (mp->flags & MEMPOOL_F_NO_PHYS_CONTIG)
+ return rte_mempool_populate_phys(mp, addr, RTE_BAD_PHYS_ADDR,
+ len, free_cb, opaque);
+
+ for (off = 0; off + pg_sz <= len &&
+ mp->populated_size < mp->size; off += phys_len) {
+
+ paddr = rte_mem_virt2phy(addr + off);
+ /* required for xen_dom0 to get the machine address */
+ paddr = rte_mem_phy2mch(-1, paddr);
+
+ if (paddr == RTE_BAD_PHYS_ADDR) {
+ ret = -EINVAL;
+ goto fail;
+ }
+
+ /* populate with the largest group of contiguous pages */
+ for (phys_len = pg_sz; off + phys_len < len; phys_len += pg_sz) {
+ phys_addr_t paddr_tmp;
+
+ paddr_tmp = rte_mem_virt2phy(addr + off + phys_len);
+ paddr_tmp = rte_mem_phy2mch(-1, paddr_tmp);
+
+ if (paddr_tmp != paddr + phys_len)
+ break;
+ }
+
+ ret = rte_mempool_populate_phys(mp, addr + off, paddr,
+ phys_len, free_cb, opaque);
+ if (ret < 0)
+ goto fail;
+ /* no need to call the free callback for next chunks */
+ free_cb = NULL;
+ cnt += ret;
+ }
+
+ return cnt;
+
+ fail:
+ rte_mempool_free_memchunks(mp);
+ return ret;
+}
+
+/* Default function to populate the mempool: allocate memory in memzones,
+ * and populate them. Return the number of objects added, or a negative
+ * value on error.
+ */
+int
+rte_mempool_populate_default(struct rte_mempool *mp)
+{
+ int mz_flags = RTE_MEMZONE_1GB|RTE_MEMZONE_SIZE_HINT_ONLY;
+ char mz_name[RTE_MEMZONE_NAMESIZE];
+ const struct rte_memzone *mz;
+ size_t size, total_elt_sz, align, pg_sz, pg_shift;
+ phys_addr_t paddr;
+ unsigned mz_id, n;
+ int ret;
+
+ /* mempool must not be populated */
+ if (mp->nb_mem_chunks != 0)
+ return -EEXIST;
+
+ if (rte_eal_has_hugepages()) {
+ pg_shift = 0; /* not needed, zone is physically contiguous */
+ pg_sz = 0;
+ align = RTE_CACHE_LINE_SIZE;
+ } else {
+ pg_sz = getpagesize();
+ pg_shift = rte_bsf32(pg_sz);
+ align = pg_sz;
+ }
+
+ total_elt_sz = mp->header_size + mp->elt_size + mp->trailer_size;
+ for (mz_id = 0, n = mp->size; n > 0; mz_id++, n -= ret) {
+ size = rte_mempool_xmem_size(n, total_elt_sz, pg_shift);
+
+ ret = snprintf(mz_name, sizeof(mz_name),
+ RTE_MEMPOOL_MZ_FORMAT "_%d", mp->name, mz_id);
+ if (ret < 0 || ret >= (int)sizeof(mz_name)) {
+ ret = -ENAMETOOLONG;
+ goto fail;
+ }
+
+ mz = rte_memzone_reserve_aligned(mz_name, size,
+ mp->socket_id, mz_flags, align);
+ /* not enough memory, retry with the biggest zone we have */
+ if (mz == NULL)
+ mz = rte_memzone_reserve_aligned(mz_name, 0,
+ mp->socket_id, mz_flags, align);
+ if (mz == NULL) {
+ ret = -rte_errno;
+ goto fail;
+ }
+
+ if (mp->flags & MEMPOOL_F_NO_PHYS_CONTIG)
+ paddr = RTE_BAD_PHYS_ADDR;
+ else
+ paddr = mz->phys_addr;
+
+ if (rte_eal_has_hugepages() && !rte_xen_dom0_supported())
+ ret = rte_mempool_populate_phys(mp, mz->addr,
+ paddr, mz->len,
+ rte_mempool_memchunk_mz_free,
+ (void *)(uintptr_t)mz);
+ else
+ ret = rte_mempool_populate_virt(mp, mz->addr,
+ mz->len, pg_sz,
+ rte_mempool_memchunk_mz_free,
+ (void *)(uintptr_t)mz);
+ if (ret < 0)
+ goto fail;
+ }
+
+ return mp->size;
+
+ fail:
+ rte_mempool_free_memchunks(mp);
+ return ret;
+}
+
+/* return the memory size required for mempool objects in anonymous mem */
+static size_t
+get_anon_size(const struct rte_mempool *mp)
+{
+ size_t size, total_elt_sz, pg_sz, pg_shift;
+
+ pg_sz = getpagesize();
+ pg_shift = rte_bsf32(pg_sz);
+ total_elt_sz = mp->header_size + mp->elt_size + mp->trailer_size;
+ size = rte_mempool_xmem_size(mp->size, total_elt_sz, pg_shift);
+
+ return size;
+}
+
+/* unmap a memory zone mapped by rte_mempool_populate_anon() */
+static void
+rte_mempool_memchunk_anon_free(struct rte_mempool_memhdr *memhdr,
+ void *opaque)
+{
+ munmap(opaque, get_anon_size(memhdr->mp));
+}
+
+/* populate the mempool with an anonymous mapping */
+int
+rte_mempool_populate_anon(struct rte_mempool *mp)
+{
+ size_t size;
+ int ret;
+ char *addr;
+
+ /* mempool is already populated, error */
+ if (!STAILQ_EMPTY(&mp->mem_list)) {
+ rte_errno = EINVAL;
+ return 0;
+ }
+
+ /* get chunk of virtually continuous memory */
+ size = get_anon_size(mp);
+ addr = mmap(NULL, size, PROT_READ | PROT_WRITE,
+ MAP_SHARED | MAP_ANONYMOUS, -1, 0);
+ if (addr == MAP_FAILED) {
+ rte_errno = errno;
+ return 0;
+ }
+ /* can't use MMAP_LOCKED, it does not exist on BSD */
+ if (mlock(addr, size) < 0) {
+ rte_errno = errno;
+ munmap(addr, size);
+ return 0;
+ }
+
+ ret = rte_mempool_populate_virt(mp, addr, size, getpagesize(),
+ rte_mempool_memchunk_anon_free, addr);
+ if (ret == 0)
+ goto fail;
+
+ return mp->populated_size;
+
+ fail:
+ rte_mempool_free_memchunks(mp);
+ return 0;
+}
+
+/* free a mempool */
+void
+rte_mempool_free(struct rte_mempool *mp)
+{
+ struct rte_mempool_list *mempool_list = NULL;
+ struct rte_tailq_entry *te;
+
+ if (mp == NULL)
+ return;
+
+ mempool_list = RTE_TAILQ_CAST(rte_mempool_tailq.head, rte_mempool_list);
+ rte_rwlock_write_lock(RTE_EAL_TAILQ_RWLOCK);
+ /* find out tailq entry */
+ TAILQ_FOREACH(te, mempool_list, next) {
+ if (te->data == (void *)mp)
+ break;
+ }
+
+ if (te != NULL) {
+ TAILQ_REMOVE(mempool_list, te, next);
+ rte_free(te);
+ }
+ rte_rwlock_write_unlock(RTE_EAL_TAILQ_RWLOCK);
+
+ rte_mempool_free_memchunks(mp);
+ rte_mempool_ops_free(mp);
+ rte_memzone_free(mp->mz);
+}
+
+/* create an empty mempool */