zephyr/subsys/net/buf.c

856 lines
19 KiB
C

/* buf.c - Buffer management */
/*
* Copyright (c) 2015 Intel Corporation
*
* SPDX-License-Identifier: Apache-2.0
*/
#include <stdio.h>
#include <errno.h>
#include <stddef.h>
#include <string.h>
#include <misc/byteorder.h>
#include <net/buf.h>
#if defined(CONFIG_NET_BUF_LOG)
#define SYS_LOG_DOMAIN "net/buf"
#define SYS_LOG_LEVEL CONFIG_SYS_LOG_NET_BUF_LEVEL
#include <logging/sys_log.h>
#define NET_BUF_DBG(fmt, ...) SYS_LOG_DBG("(%p) " fmt, k_current_get(), \
##__VA_ARGS__)
#define NET_BUF_ERR(fmt, ...) SYS_LOG_ERR(fmt, ##__VA_ARGS__)
#define NET_BUF_WARN(fmt, ...) SYS_LOG_WRN(fmt, ##__VA_ARGS__)
#define NET_BUF_INFO(fmt, ...) SYS_LOG_INF(fmt, ##__VA_ARGS__)
#define NET_BUF_ASSERT(cond) do { if (!(cond)) { \
NET_BUF_ERR("assert: '" #cond "' failed"); \
} } while (0)
#else
#define NET_BUF_DBG(fmt, ...)
#define NET_BUF_ERR(fmt, ...)
#define NET_BUF_WARN(fmt, ...)
#define NET_BUF_INFO(fmt, ...)
#define NET_BUF_ASSERT(cond)
#endif /* CONFIG_NET_BUF_LOG */
#if CONFIG_NET_BUF_WARN_ALLOC_INTERVAL > 0
#define WARN_ALLOC_INTERVAL K_SECONDS(CONFIG_NET_BUF_WARN_ALLOC_INTERVAL)
#else
#define WARN_ALLOC_INTERVAL K_FOREVER
#endif
/* Linker-defined symbol bound to the static pool structs */
extern struct net_buf_pool _net_buf_pool_list[];
struct net_buf_pool *net_buf_pool_get(int id)
{
return &_net_buf_pool_list[id];
}
static int pool_id(struct net_buf_pool *pool)
{
return pool - _net_buf_pool_list;
}
int net_buf_id(struct net_buf *buf)
{
struct net_buf_pool *pool = net_buf_pool_get(buf->pool_id);
return buf - pool->__bufs;
}
static inline struct net_buf *pool_get_uninit(struct net_buf_pool *pool,
u16_t uninit_count)
{
struct net_buf *buf;
buf = &pool->__bufs[pool->buf_count - uninit_count];
buf->pool_id = pool_id(pool);
return buf;
}
void net_buf_reset(struct net_buf *buf)
{
NET_BUF_ASSERT(buf->flags == 0);
NET_BUF_ASSERT(buf->frags == NULL);
net_buf_simple_reset(&buf->b);
}
static u8_t *generic_data_ref(struct net_buf *buf, u8_t *data)
{
u8_t *ref_count;
ref_count = data - 1;
(*ref_count)++;
return data;
}
static u8_t *mem_pool_data_alloc(struct net_buf *buf, size_t *size,
s32_t timeout)
{
struct net_buf_pool *buf_pool = net_buf_pool_get(buf->pool_id);
struct k_mem_pool *pool = buf_pool->alloc->alloc_data;
struct k_mem_block block;
u8_t *ref_count;
/* Reserve extra space for k_mem_block_id and ref-count (u8_t) */
if (k_mem_pool_alloc(pool, &block,
sizeof(struct k_mem_block_id) + 1 + *size,
timeout)) {
return NULL;
}
/* save the block descriptor info at the start of the actual block */
memcpy(block.data, &block.id, sizeof(block.id));
ref_count = (u8_t *)block.data + sizeof(block.id);
*ref_count = 1;
/* Return pointer to the byte following the ref count */
return ref_count + 1;
}
static void mem_pool_data_unref(struct net_buf *buf, u8_t *data)
{
struct k_mem_block_id id;
u8_t *ref_count;
ref_count = data - 1;
if (--(*ref_count)) {
return;
}
/* Need to copy to local variable due to alignment */
memcpy(&id, ref_count - sizeof(id), sizeof(id));
k_mem_pool_free_id(&id);
}
const struct net_buf_data_cb net_buf_var_cb = {
.alloc = mem_pool_data_alloc,
.ref = generic_data_ref,
.unref = mem_pool_data_unref,
};
static u8_t *fixed_data_alloc(struct net_buf *buf, size_t *size, s32_t timeout)
{
struct net_buf_pool *pool = net_buf_pool_get(buf->pool_id);
const struct net_buf_pool_fixed *fixed = pool->alloc->alloc_data;
*size = min(fixed->data_size, *size);
return fixed->data_pool + fixed->data_size * net_buf_id(buf);
}
static void fixed_data_unref(struct net_buf *buf, u8_t *data)
{
/* Nothing needed for fixed-size data pools */
}
const struct net_buf_data_cb net_buf_fixed_cb = {
.alloc = fixed_data_alloc,
.unref = fixed_data_unref,
};
#if (CONFIG_HEAP_MEM_POOL_SIZE > 0)
static u8_t *heap_data_alloc(struct net_buf *buf, size_t *size, s32_t timeout)
{
u8_t *ref_count;
ref_count = k_malloc(1 + *size);
if (!ref_count) {
return NULL;
}
*ref_count = 1;
return ref_count + 1;
}
static void heap_data_unref(struct net_buf *buf, u8_t *data)
{
u8_t *ref_count;
ref_count = data - 1;
if (--(*ref_count)) {
return;
}
k_free(ref_count);
}
static const struct net_buf_data_cb net_buf_heap_cb = {
.alloc = heap_data_alloc,
.ref = generic_data_ref,
.unref = heap_data_unref,
};
const struct net_buf_data_alloc net_buf_heap_alloc = {
.cb = &net_buf_heap_cb,
};
#endif /* CONFIG_HEAP_MEM_POOL_SIZE > 0 */
static u8_t *data_alloc(struct net_buf *buf, size_t *size, s32_t timeout)
{
struct net_buf_pool *pool = net_buf_pool_get(buf->pool_id);
return pool->alloc->cb->alloc(buf, size, timeout);
}
static u8_t *data_ref(struct net_buf *buf, u8_t *data)
{
struct net_buf_pool *pool = net_buf_pool_get(buf->pool_id);
return pool->alloc->cb->ref(buf, data);
}
static void data_unref(struct net_buf *buf, u8_t *data)
{
struct net_buf_pool *pool = net_buf_pool_get(buf->pool_id);
if (buf->flags & NET_BUF_EXTERNAL_DATA) {
return;
}
pool->alloc->cb->unref(buf, data);
}
#if defined(CONFIG_NET_BUF_LOG)
struct net_buf *net_buf_alloc_len_debug(struct net_buf_pool *pool, size_t size,
s32_t timeout, const char *func,
int line)
#else
struct net_buf *net_buf_alloc_len(struct net_buf_pool *pool, size_t size,
s32_t timeout)
#endif
{
u32_t alloc_start = k_uptime_get_32();
struct net_buf *buf;
unsigned int key;
NET_BUF_ASSERT(pool);
NET_BUF_DBG("%s():%d: pool %p size %zu timeout %d", func, line, pool,
size, timeout);
/* We need to lock interrupts temporarily to prevent race conditions
* when accessing pool->uninit_count.
*/
key = irq_lock();
/* If there are uninitialized buffers we're guaranteed to succeed
* with the allocation one way or another.
*/
if (pool->uninit_count) {
u16_t uninit_count;
/* If this is not the first access to the pool, we can
* be opportunistic and try to fetch a previously used
* buffer from the LIFO with K_NO_WAIT.
*/
if (pool->uninit_count < pool->buf_count) {
buf = k_lifo_get(&pool->free, K_NO_WAIT);
if (buf) {
irq_unlock(key);
goto success;
}
}
uninit_count = pool->uninit_count--;
irq_unlock(key);
buf = pool_get_uninit(pool, uninit_count);
goto success;
}
irq_unlock(key);
#if defined(CONFIG_NET_BUF_LOG) && SYS_LOG_LEVEL >= SYS_LOG_LEVEL_WARNING
if (timeout == K_FOREVER) {
u32_t ref = k_uptime_get_32();
buf = k_lifo_get(&pool->free, K_NO_WAIT);
while (!buf) {
#if defined(CONFIG_NET_BUF_POOL_USAGE)
NET_BUF_WARN("%s():%d: Pool %s low on buffers.",
func, line, pool->name);
#else
NET_BUF_WARN("%s():%d: Pool %p low on buffers.",
func, line, pool);
#endif
buf = k_lifo_get(&pool->free, WARN_ALLOC_INTERVAL);
#if defined(CONFIG_NET_BUF_POOL_USAGE)
NET_BUF_WARN("%s():%d: Pool %s blocked for %u secs",
func, line, pool->name,
(k_uptime_get_32() - ref) / MSEC_PER_SEC);
#else
NET_BUF_WARN("%s():%d: Pool %p blocked for %u secs",
func, line, pool,
(k_uptime_get_32() - ref) / MSEC_PER_SEC);
#endif
}
} else {
buf = k_lifo_get(&pool->free, timeout);
}
#else
buf = k_lifo_get(&pool->free, timeout);
#endif
if (!buf) {
NET_BUF_ERR("%s():%d: Failed to get free buffer", func, line);
return NULL;
}
success:
NET_BUF_DBG("allocated buf %p", buf);
if (size) {
if (timeout != K_NO_WAIT && timeout != K_FOREVER) {
u32_t diff = k_uptime_get_32() - alloc_start;
timeout -= min(timeout, diff);
}
buf->__buf = data_alloc(buf, &size, timeout);
if (!buf->__buf) {
NET_BUF_ERR("%s():%d: Failed to allocate data",
func, line);
net_buf_destroy(buf);
return NULL;
}
} else {
buf->__buf = NULL;
}
buf->ref = 1;
buf->flags = 0;
buf->frags = NULL;
buf->size = size;
net_buf_reset(buf);
#if defined(CONFIG_NET_BUF_POOL_USAGE)
pool->avail_count--;
NET_BUF_ASSERT(pool->avail_count >= 0);
#endif
return buf;
}
#if defined(CONFIG_NET_BUF_LOG)
struct net_buf *net_buf_alloc_fixed_debug(struct net_buf_pool *pool,
s32_t timeout, const char *func,
int line)
{
const struct net_buf_pool_fixed *fixed = pool->alloc->alloc_data;
return net_buf_alloc_len_debug(pool, fixed->data_size, timeout, func,
line);
}
#else
struct net_buf *net_buf_alloc_fixed(struct net_buf_pool *pool, s32_t timeout)
{
const struct net_buf_pool_fixed *fixed = pool->alloc->alloc_data;
return net_buf_alloc_len(pool, fixed->data_size, timeout);
}
#endif
#if defined(CONFIG_NET_BUF_LOG)
struct net_buf *net_buf_alloc_with_data_debug(struct net_buf_pool *pool,
void *data, size_t size,
s32_t timeout, const char *func,
int line)
#else
struct net_buf *net_buf_alloc_with_data(struct net_buf_pool *pool,
void *data, size_t size,
s32_t timeout)
#endif
{
struct net_buf *buf;
#if defined(CONFIG_NET_BUF_LOG)
buf = net_buf_alloc_len_debug(pool, 0, timeout, func, line);
#else
buf = net_buf_alloc_len(pool, 0, timeout);
#endif
if (!buf) {
return NULL;
}
buf->__buf = data;
buf->data = data;
buf->size = size;
buf->len = size;
buf->flags = NET_BUF_EXTERNAL_DATA;
return buf;
}
#if defined(CONFIG_NET_BUF_LOG)
struct net_buf *net_buf_get_debug(struct k_fifo *fifo, s32_t timeout,
const char *func, int line)
#else
struct net_buf *net_buf_get(struct k_fifo *fifo, s32_t timeout)
#endif
{
struct net_buf *buf, *frag;
NET_BUF_DBG("%s():%d: fifo %p timeout %d", func, line, fifo, timeout);
buf = k_fifo_get(fifo, timeout);
if (!buf) {
return NULL;
}
NET_BUF_DBG("%s():%d: buf %p fifo %p", func, line, buf, fifo);
/* Get any fragments belonging to this buffer */
for (frag = buf; (frag->flags & NET_BUF_FRAGS); frag = frag->frags) {
frag->frags = k_fifo_get(fifo, K_NO_WAIT);
NET_BUF_ASSERT(frag->frags);
/* The fragments flag is only for FIFO-internal usage */
frag->flags &= ~NET_BUF_FRAGS;
}
/* Mark the end of the fragment list */
frag->frags = NULL;
return buf;
}
void net_buf_simple_reserve(struct net_buf_simple *buf, size_t reserve)
{
NET_BUF_ASSERT(buf);
NET_BUF_ASSERT(buf->len == 0);
NET_BUF_DBG("buf %p reserve %zu", buf, reserve);
buf->data = buf->__buf + reserve;
}
void net_buf_slist_put(sys_slist_t *list, struct net_buf *buf)
{
struct net_buf *tail;
unsigned int key;
NET_BUF_ASSERT(list);
NET_BUF_ASSERT(buf);
for (tail = buf; tail->frags; tail = tail->frags) {
tail->flags |= NET_BUF_FRAGS;
}
key = irq_lock();
sys_slist_append_list(list, &buf->node, &tail->node);
irq_unlock(key);
}
struct net_buf *net_buf_slist_get(sys_slist_t *list)
{
struct net_buf *buf, *frag;
unsigned int key;
NET_BUF_ASSERT(list);
key = irq_lock();
buf = (void *)sys_slist_get(list);
irq_unlock(key);
if (!buf) {
return NULL;
}
/* Get any fragments belonging to this buffer */
for (frag = buf; (frag->flags & NET_BUF_FRAGS); frag = frag->frags) {
key = irq_lock();
frag->frags = (void *)sys_slist_get(list);
irq_unlock(key);
NET_BUF_ASSERT(frag->frags);
/* The fragments flag is only for list-internal usage */
frag->flags &= ~NET_BUF_FRAGS;
}
/* Mark the end of the fragment list */
frag->frags = NULL;
return buf;
}
void net_buf_put(struct k_fifo *fifo, struct net_buf *buf)
{
struct net_buf *tail;
NET_BUF_ASSERT(fifo);
NET_BUF_ASSERT(buf);
for (tail = buf; tail->frags; tail = tail->frags) {
tail->flags |= NET_BUF_FRAGS;
}
k_fifo_put_list(fifo, buf, tail);
}
#if defined(CONFIG_NET_BUF_LOG)
void net_buf_unref_debug(struct net_buf *buf, const char *func, int line)
#else
void net_buf_unref(struct net_buf *buf)
#endif
{
NET_BUF_ASSERT(buf);
while (buf) {
struct net_buf *frags = buf->frags;
struct net_buf_pool *pool;
#if defined(CONFIG_NET_BUF_LOG)
if (!buf->ref) {
NET_BUF_ERR("%s():%d: buf %p double free", func, line,
buf);
return;
}
#endif
NET_BUF_DBG("buf %p ref %u pool_id %u frags %p", buf, buf->ref,
buf->pool_id, buf->frags);
if (--buf->ref > 0) {
return;
}
if (buf->__buf) {
data_unref(buf, buf->__buf);
buf->__buf = NULL;
}
buf->data = NULL;
buf->frags = NULL;
pool = net_buf_pool_get(buf->pool_id);
#if defined(CONFIG_NET_BUF_POOL_USAGE)
pool->avail_count++;
NET_BUF_ASSERT(pool->avail_count <= pool->buf_count);
#endif
if (pool->destroy) {
pool->destroy(buf);
} else {
net_buf_destroy(buf);
}
buf = frags;
}
}
struct net_buf *net_buf_ref(struct net_buf *buf)
{
NET_BUF_ASSERT(buf);
NET_BUF_DBG("buf %p (old) ref %u pool_id %u",
buf, buf->ref, buf->pool_id);
buf->ref++;
return buf;
}
struct net_buf *net_buf_clone(struct net_buf *buf, s32_t timeout)
{
u32_t alloc_start = k_uptime_get_32();
struct net_buf_pool *pool;
struct net_buf *clone;
NET_BUF_ASSERT(buf);
pool = net_buf_pool_get(buf->pool_id);
clone = net_buf_alloc_len(pool, 0, timeout);
if (!clone) {
return NULL;
}
/* If the pool supports data referencing use that. Otherwise
* we need to allocate new data and make a copy.
*/
if (pool->alloc->cb->ref && !(buf->flags & NET_BUF_EXTERNAL_DATA)) {
clone->__buf = data_ref(buf, buf->__buf);
clone->data = buf->data;
clone->len = buf->len;
clone->size = buf->size;
} else {
size_t size = buf->size;
if (timeout != K_NO_WAIT && timeout != K_FOREVER) {
u32_t diff = k_uptime_get_32() - alloc_start;
timeout -= min(timeout, diff);
}
clone->__buf = data_alloc(clone, &size, timeout);
if (!clone->__buf || size < buf->size) {
net_buf_destroy(clone);
return NULL;
}
clone->size = size;
clone->data = clone->__buf + net_buf_headroom(buf);
net_buf_add_mem(clone, buf->data, buf->len);
}
return clone;
}
struct net_buf *net_buf_frag_last(struct net_buf *buf)
{
NET_BUF_ASSERT(buf);
while (buf->frags) {
buf = buf->frags;
}
return buf;
}
void net_buf_frag_insert(struct net_buf *parent, struct net_buf *frag)
{
NET_BUF_ASSERT(parent);
NET_BUF_ASSERT(frag);
if (parent->frags) {
net_buf_frag_last(frag)->frags = parent->frags;
}
/* Take ownership of the fragment reference */
parent->frags = frag;
}
struct net_buf *net_buf_frag_add(struct net_buf *head, struct net_buf *frag)
{
NET_BUF_ASSERT(frag);
if (!head) {
return net_buf_ref(frag);
}
net_buf_frag_insert(net_buf_frag_last(head), frag);
return head;
}
#if defined(CONFIG_NET_BUF_LOG)
struct net_buf *net_buf_frag_del_debug(struct net_buf *parent,
struct net_buf *frag,
const char *func, int line)
#else
struct net_buf *net_buf_frag_del(struct net_buf *parent, struct net_buf *frag)
#endif
{
struct net_buf *next_frag;
NET_BUF_ASSERT(frag);
if (parent) {
NET_BUF_ASSERT(parent->frags);
NET_BUF_ASSERT(parent->frags == frag);
parent->frags = frag->frags;
}
next_frag = frag->frags;
frag->frags = NULL;
#if defined(CONFIG_NET_BUF_LOG)
net_buf_unref_debug(frag, func, line);
#else
net_buf_unref(frag);
#endif
return next_frag;
}
#if defined(CONFIG_NET_BUF_SIMPLE_LOG)
#define NET_BUF_SIMPLE_DBG(fmt, ...) NET_BUF_DBG(fmt, ##__VA_ARGS__)
#define NET_BUF_SIMPLE_ERR(fmt, ...) NET_BUF_ERR(fmt, ##__VA_ARGS__)
#define NET_BUF_SIMPLE_WARN(fmt, ...) NET_BUF_WARN(fmt, ##__VA_ARGS__)
#define NET_BUF_SIMPLE_INFO(fmt, ...) NET_BUF_INFO(fmt, ##__VA_ARGS__)
#define NET_BUF_SIMPLE_ASSERT(cond) NET_BUF_ASSERT(cond)
#else
#define NET_BUF_SIMPLE_DBG(fmt, ...)
#define NET_BUF_SIMPLE_ERR(fmt, ...)
#define NET_BUF_SIMPLE_WARN(fmt, ...)
#define NET_BUF_SIMPLE_INFO(fmt, ...)
#define NET_BUF_SIMPLE_ASSERT(cond)
#endif /* CONFIG_NET_BUF_SIMPLE_LOG */
void *net_buf_simple_add(struct net_buf_simple *buf, size_t len)
{
u8_t *tail = net_buf_simple_tail(buf);
NET_BUF_SIMPLE_DBG("buf %p len %zu", buf, len);
NET_BUF_SIMPLE_ASSERT(net_buf_simple_tailroom(buf) >= len);
buf->len += len;
return tail;
}
void *net_buf_simple_add_mem(struct net_buf_simple *buf, const void *mem,
size_t len)
{
NET_BUF_SIMPLE_DBG("buf %p len %zu", buf, len);
return memcpy(net_buf_simple_add(buf, len), mem, len);
}
u8_t *net_buf_simple_add_u8(struct net_buf_simple *buf, u8_t val)
{
u8_t *u8;
NET_BUF_SIMPLE_DBG("buf %p val 0x%02x", buf, val);
u8 = net_buf_simple_add(buf, 1);
*u8 = val;
return u8;
}
void net_buf_simple_add_le16(struct net_buf_simple *buf, u16_t val)
{
NET_BUF_SIMPLE_DBG("buf %p val %u", buf, val);
val = sys_cpu_to_le16(val);
memcpy(net_buf_simple_add(buf, sizeof(val)), &val, sizeof(val));
}
void net_buf_simple_add_be16(struct net_buf_simple *buf, u16_t val)
{
NET_BUF_SIMPLE_DBG("buf %p val %u", buf, val);
val = sys_cpu_to_be16(val);
memcpy(net_buf_simple_add(buf, sizeof(val)), &val, sizeof(val));
}
void net_buf_simple_add_le32(struct net_buf_simple *buf, u32_t val)
{
NET_BUF_SIMPLE_DBG("buf %p val %u", buf, val);
val = sys_cpu_to_le32(val);
memcpy(net_buf_simple_add(buf, sizeof(val)), &val, sizeof(val));
}
void net_buf_simple_add_be32(struct net_buf_simple *buf, u32_t val)
{
NET_BUF_SIMPLE_DBG("buf %p val %u", buf, val);
val = sys_cpu_to_be32(val);
memcpy(net_buf_simple_add(buf, sizeof(val)), &val, sizeof(val));
}
void *net_buf_simple_push(struct net_buf_simple *buf, size_t len)
{
NET_BUF_SIMPLE_DBG("buf %p len %zu", buf, len);
NET_BUF_SIMPLE_ASSERT(net_buf_simple_headroom(buf) >= len);
buf->data -= len;
buf->len += len;
return buf->data;
}
void net_buf_simple_push_le16(struct net_buf_simple *buf, u16_t val)
{
NET_BUF_SIMPLE_DBG("buf %p val %u", buf, val);
val = sys_cpu_to_le16(val);
memcpy(net_buf_simple_push(buf, sizeof(val)), &val, sizeof(val));
}
void net_buf_simple_push_be16(struct net_buf_simple *buf, u16_t val)
{
NET_BUF_SIMPLE_DBG("buf %p val %u", buf, val);
val = sys_cpu_to_be16(val);
memcpy(net_buf_simple_push(buf, sizeof(val)), &val, sizeof(val));
}
void net_buf_simple_push_u8(struct net_buf_simple *buf, u8_t val)
{
u8_t *data = net_buf_simple_push(buf, 1);
*data = val;
}
void *net_buf_simple_pull(struct net_buf_simple *buf, size_t len)
{
NET_BUF_SIMPLE_DBG("buf %p len %zu", buf, len);
NET_BUF_SIMPLE_ASSERT(buf->len >= len);
buf->len -= len;
return buf->data += len;
}
u8_t net_buf_simple_pull_u8(struct net_buf_simple *buf)
{
u8_t val;
val = buf->data[0];
net_buf_simple_pull(buf, 1);
return val;
}
u16_t net_buf_simple_pull_le16(struct net_buf_simple *buf)
{
u16_t val;
val = UNALIGNED_GET((u16_t *)buf->data);
net_buf_simple_pull(buf, sizeof(val));
return sys_le16_to_cpu(val);
}
u16_t net_buf_simple_pull_be16(struct net_buf_simple *buf)
{
u16_t val;
val = UNALIGNED_GET((u16_t *)buf->data);
net_buf_simple_pull(buf, sizeof(val));
return sys_be16_to_cpu(val);
}
u32_t net_buf_simple_pull_le32(struct net_buf_simple *buf)
{
u32_t val;
val = UNALIGNED_GET((u32_t *)buf->data);
net_buf_simple_pull(buf, sizeof(val));
return sys_le32_to_cpu(val);
}
u32_t net_buf_simple_pull_be32(struct net_buf_simple *buf)
{
u32_t val;
val = UNALIGNED_GET((u32_t *)buf->data);
net_buf_simple_pull(buf, sizeof(val));
return sys_be32_to_cpu(val);
}
size_t net_buf_simple_headroom(struct net_buf_simple *buf)
{
return buf->data - buf->__buf;
}
size_t net_buf_simple_tailroom(struct net_buf_simple *buf)
{
return buf->size - net_buf_simple_headroom(buf) - buf->len;
}