ESP-IDF uses the open source lwIP lightweight TCP/IP stack. The ESP-IDF version of lwIP (esp-lwip) has some modifications and additions compared to the upstream project.

ESP-IDF supports the following lwIP TCP/IP stack functions:

is enabled but not officially supported for ESP-IDF applications

Warning

When using any lwIP API other than the , please make sure that the API is thread-safe. To check if a given API call is thread-safe, enable the configuration option and run the application. This enables lwIP to assert the correct access of the TCP/IP core functionality. If the API is not accessed or locked properly from the appropriate , the execution will be aborted. The general recommendation is to use the ESP-NETIF component to interact with lwIP.

Some common lwIP app APIs are supported indirectly by ESP-IDF:

Note

DNS server configuration in lwIP is global, not interface-specific. If you are using multiple network interfaces with distinct DNS servers, exercise caution to prevent inadvertent overwrites of one interface's DNS settings when acquiring a DHCP lease from another interface.

The BSD Sockets API is a common cross-platform TCP/IP sockets API that originated in the Berkeley Standard Distribution of UNIX but is now standardized in a section of the POSIX specification. BSD Sockets are sometimes called POSIX Sockets or Berkeley Sockets.

As implemented in ESP-IDF, lwIP supports all of the common usages of the BSD Sockets API. However, not all operations are fully thread-safe, and simultaneous reads and writes from multiple threads may require additional synchronization mechanisms, see for more details.

A wide range of BSD Sockets reference materials are available, including:

A number of ESP-IDF examples show how to use the BSD Sockets APIs:

The following BSD socket API functions are supported. For full details, see lwip/lwip/src/include/lwip/sockets.h.

Non-standard functions:

ioctl(): see

Note

Some lwIP application sample code uses prefixed versions of BSD APIs, e.g., lwip_socket(), instead of the standard socket(). Both forms can be used with ESP-IDF, but using standard names is recommended.

BSD Socket error handling code is very important for robust socket applications. Normally, socket error handling involves the following aspects:

Detecting the error

Getting the error reason code

Handling the error according to the reason code

In lwIP, we have two different scenarios for handling socket errors:

Socket API returns an error. For more information, see .

select(int maxfdp1, fd_set *readset, fd_set *writeset, fd_set *exceptset, struct timeval *timeout) has an exception descriptor indicating that the socket has an error. For more information, see .

Error detection

We can know that the socket API fails according to its return value.

Get the error reason code

When socket API fails, the return value does not contain the failure reason and the application can get the error reason code by accessing errno. Different values indicate different meanings. For more information, see .

Example:

int err; int sockfd; if (sockfd = socket(AF_INET,SOCK_STREAM,0) < 0) { // the error code is obtained from errno err = errno; return err; }

Error detection

Socket error when select() has exception descriptor.

Get the error reason code

If the select() indicates that the socket fails, we can not get the error reason code by accessing errno, instead we should call getsockopt() to get the failure reason code. Since select() has exception descriptor, the error code is not given to errno.

Note

The getsockopt() function has the following prototype: int getsockopt(int s, int level, int optname, void *optval, socklen_t *optlen). Its purpose is to get the current value of the option of any type, any state socket, and store the result in optval. For example, when you get the error code on a socket, you can get it by getsockopt(sockfd, SOL_SOCKET, SO_ERROR, &err, &optlen).

Example:

int err; if (select(sockfd + 1, NULL, NULL, &exfds, &tval) <= 0) { err = errno; return err; } else { if (FD_ISSET(sockfd, &exfds)) { // select() exception set using getsockopt() int optlen = sizeof(int); getsockopt(sockfd, SOL_SOCKET, SO_ERROR, &err, &optlen); return err; } }

Below is a list of common error codes. For a more detailed list of standard POSIX/C error codes, please see newlib errno.h and the platform-specific extensions newlib/platform_include/sys/errno.h.

Error code

Description

The getsockopt() and setsockopt() functions allow getting and setting per-socket options respectively.

Not all standard socket options are supported by lwIP in ESP-IDF. The following socket options are supported:

Used with level argument SOL_SOCKET.

SO_REUSEADDR: available if is set, whose behavior can be customized by setting

SO_KEEPALIVE

SO_BROADCAST

SO_ACCEPTCONN

SO_RCVBUF: available if is set

SO_SNDTIMEO / SO_RCVTIMEO

SO_ERROR: only used with select(), see

SO_TYPE

SO_NO_CHECK: for UDP sockets only

Used with level argument IPPROTO_IP.

IP_TOS

IP_TTL

IP_PKTINFO: available if is set

For multicast UDP sockets:

IP_MULTICAST_IF

IP_MULTICAST_LOOP

IP_MULTICAST_TTL

IP_ADD_MEMBERSHIP

IP_DROP_MEMBERSHIP

TCP sockets only. Used with level argument IPPROTO_TCP.

TCP_NODELAY

Options relating to TCP keepalive probes:

TCP_KEEPALIVE: int value, TCP keepalive period in milliseconds

TCP_KEEPIDLE: same as TCP_KEEPALIVE, but the value is in seconds

TCP_KEEPINTVL: int value, the interval between keepalive probes in seconds

TCP_KEEPCNT: int value, number of keepalive probes before timing out

IPv6 sockets only. Used with level argument IPPROTO_IPV6.

IPV6_CHECKSUM

IPV6_V6ONLY

For multicast IPv6 UDP sockets:

IPV6_JOIN_GROUP / IPV6_ADD_MEMBERSHIP

IPV6_LEAVE_GROUP / IPV6_DROP_MEMBERSHIP

IPV6_MULTICAST_IF

IPV6_MULTICAST_HOPS

IPV6_MULTICAST_LOOP

The fcntl() function is a standard API for manipulating options related to a file descriptor. In ESP-IDF, the Virtual Filesystem Component layer is used to implement this function.

When the file descriptor is a socket, only the following fcntl() values are supported:

O_NONBLOCK to set or clear non-blocking I/O mode. Also supports O_NDELAY, which is identical to O_NONBLOCK.

O_RDONLY, O_WRONLY, O_RDWR flags for different read or write modes. These flags can only be read using F_GETFL, and cannot be set using F_SETFL. A TCP socket returns a different mode depending on whether the connection has been closed at either end or is still open at both ends. UDP sockets always return O_RDWR.

The ioctl() function provides a semi-standard way to access some internal features of the TCP/IP stack. In ESP-IDF, the Virtual Filesystem Component layer is used to implement this function.

When the file descriptor is a socket, only the following ioctl() values are supported:

FIONREAD returns the number of bytes of the pending data already received in the socket's network buffer.

FIONBIO is an alternative way to set/clear non-blocking I/O status for a socket, equivalent to fcntl(fd, F_SETFL, O_NONBLOCK, ...).

lwIP supports two lower-level APIs as well as the BSD Sockets API: the Netconn API and the Raw API.

The lwIP Raw API is designed for single-threaded devices and is not supported in ESP-IDF.

The Netconn API is used to implement the BSD Sockets API inside lwIP, and it can also be called directly from ESP-IDF apps. This API has lower resource usage than the BSD Sockets API. In particular, it can send and receive data without firstly copying it into internal lwIP buffers.

Important

Espressif does not test the Netconn API in ESP-IDF. As such, this functionality is enabled but not supported. Some functionality may only work correctly when used from the BSD Sockets API.

For more information about the Netconn API, consult lwip/lwip/src/include/lwip/api.h and part of the **unofficial** lwIP Application Developers Manual.

lwIP creates a dedicated TCP/IP FreeRTOS task to handle socket API requests from other tasks.

A number of configuration items are available to modify the task and the queues (mailboxes) used to send data to/from the TCP/IP task:

Both IPv4 and IPv6 are supported in a dual-stack configuration and are enabled by default. Both IPv6 and IPv4 may be disabled separately if they are not needed, see .

IPv6 support is limited to Stateless Autoconfiguration only. Stateful configuration is not supported in ESP-IDF, nor in upstream lwIP.

IPv6 Address configuration is defined by means of these protocols or services:

SLAAC IPv6 Stateless Address Autoconfiguration (RFC-2462)

DHCPv6 Dynamic Host Configuration Protocol for IPv6 (RFC-8415)

None of these two types of address configuration is enabled by default, so the device uses only Link Local addresses or statically-defined addresses.

To enable address autoconfiguration using the Router Advertisement protocol, please enable:

This configuration option enables IPv6 autoconfiguration for all network interfaces, which differs from the upstream lwIP behavior, where the autoconfiguration needs to be explicitly enabled for each netif with netif->ip6_autoconfig_enabled=1.

DHCPv6 in lwIP is very simple and supports only stateless configuration. It could be enabled using:

Since the DHCPv6 works only in its stateless configuration, the has to be enabled as well via .

Moreover, the DHCPv6 needs to be explicitly enabled from the application code using:

dhcp6_enable_stateless(netif);

In order to autoconfigure DNS server(s), especially in IPv6-only networks, we have these two options:

Recursive Domain Name System (DNS): this belongs to the Neighbor Discovery Protocol (NDP) and uses .

The number of servers must be set , this option is disabled by default, i.e., set to 0.

DHCPv6 stateless configuration, uses to configure DNS servers. Note that this configuration assumes IPv6 Router Advertisement Flags (RFC-5175) to be set to

Managed Address Configuration Flag = 0

Other Configuration Flag = 1

The following code is added, which is not present in the upstream lwIP release:

It is possible to close() a socket from a different thread than the one that created it. The close() call blocks, until any function calls currently using that socket from other tasks have returned.

It is, however, not possible to delete a task while it is actively waiting on select() or poll() APIs. It is always necessary that these APIs exit before destroying the task, as this might corrupt internal structures and cause subsequent crashes of the lwIP. These APIs allocate globally referenced callback pointers on the stack so that when the task gets destroyed before unrolling the stack, the lwIP could still hold pointers to the deleted stack.

lwIP IGMP and MLD6 feature both initialize a timer in order to trigger timeout events at certain times.

The default lwIP implementation is to have these timers enabled all the time, even if no timeout events are active. This increases CPU usage and power consumption when using automatic Light-sleep mode. ESP-lwIP default behavior is to set each timer on demand, so it is only enabled when an event is pending.

To return to the default lwIP behavior, which is always-on timers, disable .

When not using Wi-Fi, the lwIP timer can be turned off via the API to reduce power consumption.

The following API functions are supported. For full details, see lwip/lwip/src/include/lwip/timeouts.h.

sys_timeouts_init()

sys_timeouts_deinit()

Some standard IPV4 and IPV6 multicast socket options are implemented, see .

Possible to set IPV6-only UDP and TCP sockets with IPV6_V6ONLY socket option, while normal lwIP is TCP-only.

IPV4-source-based routing implementation is different

IPV4-mapped IPV6 addresses are supported

IPV4 network address port translation (NAPT) and port forwarding are supported. However, the enabling of NAPT is limited to a single interface.

ESP-IDF port layer provides a default hook file, which is included in the lwIP build process. This file is located at lwip/port/include/lwip_default_hooks.h and defines several hooks that implement default ESP-IDF behavior of lwIP stack. These hooks could be further modified by choosing one of these options:

None: No hook is declared.

Default: Default IDF implementation is provided (declared as weak in most cases and can be overridden).

Custom: Only the hook declaration is provided, and the application must implement it.

DHCP Extra Option Hook

ESP-IDF allows applications to define a hook for handling extra DHCP options. This can be useful for implementing custom DHCP-based behaviors, such as retrieving specific vendor options. To enable this feature, configure either to Default (weak implementation is provided which could be replaced by custom implementation) or Custom (you will have to implement the hook and define its link dependency to lwip)

Example Usage

An application can define the following function to handle a specific DHCP option (e.g., captive portal URI):

#include "esp_netif.h" #include "lwip/dhcp.h" void lwip_dhcp_on_extra_option(struct dhcp *dhcp, uint8_t state, uint8_t option, uint8_t len, struct pbuf* p, uint16_t offset) { if (option == ESP_NETIF_CAPTIVEPORTAL_URI) { char *uri = (char *)p->payload + offset; ESP_LOGI(TAG, "Captive Portal URI: %s", uri); } }

Other Default Hooks

ESP-IDF provides additional lwIP hooks that can be overridden. These include:

TCP ISN Hook (): Allows custom randomization of TCP Initial Sequence Numbers (ESP-IDF provided implementation is the default option, set to Custom to provide custom implementation or None to use lwIP implementation).

IPv6 Route Hook (): Enables custom route selection for IPv6 packets (No hook by default, use Default or Custom to override).

IPv6 Get Gateway Hook (): Enables defining a custom gateway selection logic (No hook by default, use Default or Custom to override).

IPv6 Source Address Selection Hook (): Allows customization of source address selection (No hook by default, use Default or Custom to override).

Netconn External Resolve Hook (): Allows overriding the DNS resolution logic for network connections (No hook by default, use Default or Custom to override).

DNS External Resolve Hook (): Provides a hook for custom DNS resolution logic with callbacks (No hook by default, but could be selected by an external component to prefer the custom option; use Default or Custom to override).

IPv6 Packet Input Hook (): Provides filtering or modification of incoming IPv6 packets (ESP-IDF provided weak implementation is the default option; use Custom or provide a strong definition to override the Default option; choose None to disable IPv6 packet input filtering)

Each of these hooks can be configured in menuconfig, allowing selection of default, custom, or no implementation.

The original lwIP supports implementing custom compile-time modifications via LWIP_HOOK_FILENAME. This file is already used by the ESP-IDF port layer, but ESP-IDF users could still include and implement any custom additions via a header file defined by the macro ESP_IDF_LWIP_HOOK_FILENAME. Here is an example of adding a custom hook file to the build process, and the hook is called my_hook.h, located in the project's main folder:

idf_component_get_property(lwiplwipCOMPONENT_LIB) target_compile_options(${lwip}PRIVATE"-I${PROJECT_DIR}/main") target_compile_definitions(${lwip}PRIVATE"-DESP_IDF_LWIP_HOOK_FILENAME=\"my_hook.h\"")

The most common lwIP options are configurable through the component configuration menu. However, certain definitions need to be injected from the command line. The CMake function target_compile_definitions() can be employed to define macros, as illustrated below:

idf_component_get_property(lwiplwipCOMPONENT_LIB) target_compile_definitions(${lwip}PRIVATE"-DETHARP_SUPPORT_VLAN=1")

This approach may not work for function-like macros, as there is no guarantee that the definition will be accepted by all compilers, although it is supported in GCC. To address this limitation, the add_definitions() function can be utilized to define the macro for the entire project, for example: add_definitions("-DFALLBACK_DNS_SERVER_ADDRESS(addr)=\"IP_ADDR4((addr), 8,8,8,8)\"").

Alternatively, you can define your function-like macro in a header file which will be pre-included as an lwIP hook file, see .

lwIP in ESP-IDF supports thread safety in certain scenarios, but with limitations. It is possible to perform read, write, and close operations from different threads on the same socket simultaneously. However, performing multiple reads or multiple writes from more than one thread on the same socket at the same time is not supported. Applications that require simultaneous reads or writes from multiple threads on the same socket must implement additional synchronization mechanisms, such as locking around socket operations.

ESP-IDF additions to lwIP still suffer from the global DNS limitation, described in . To address this limitation from application code, the FALLBACK_DNS_SERVER_ADDRESS() macro can be utilized to define a global DNS fallback server accessible from all interfaces. Alternatively, you have the option to maintain per-interface DNS servers and reconfigure them whenever the default interface changes.

The number of IP addresses returned by network database APIs such as getaddrinfo() and gethostbyname() is restricted by the macro DNS_MAX_HOST_IP. By default, the value of this macro is set to 1.

In the implementation of getaddrinfo(), the canonical name is not available. Therefore, the ai_canonname field of the first returned addrinfo structure will always refer to the nodename argument or a string with the same contents.

The getaddrinfo() system call in lwIP within ESP-IDF has a limitation when using AF_UNSPEC, as it defaults to returning only an IPv4 address in dual stack mode. This can cause issues in IPv6-only networks. To handle this, a workaround involves making two sequential calls to getaddrinfo(): the first with AF_INET to query for IPv4 addresses, and the second with AF_INET6 to retrieve IPv6 addresses. To provide a more robust solution, the custom esp_getaddrinfo() function has been added to the lwIP port layer to handle both IPv4 and IPv6 addresses when AF_UNSPEC is used. The option, available when both IPv4 and IPv6 are enabled, controls whether esp_getaddrinfo() or getaddrinfo() is used. It is disabled by default.

Calling send() or sendto() repeatedly on a UDP socket may eventually fail with errno equal to ENOMEM. This failure occurs due to the limitations of buffer sizes in the lower-layer network interface drivers. If all driver transmit buffers are full, the UDP transmission will fail. For applications that transmit a high volume of UDP datagrams and aim to avoid any dropped datagrams by the sender, it is advisable to implement error code checking and employ a retransmission mechanism with a short delay.

Increasing the number of TX buffers in the or project configuration as applicable may also help.

TCP/IP performance is a complex subject, and performance can be optimized toward multiple goals. The default settings of ESP-IDF are tuned for a compromise between throughput, latency, and moderate memory usage.

Espressif tests ESP-IDF TCP/IP throughput using the iperf test application: https://iperf.fr/, please refer to for more details about the actual testing and using the optimized configuration.

Important

Suggest applying changes a few at a time and checking the performance each time with a particular application workload.

If a lot of tasks are competing for CPU time on the system, consider that the lwIP task has configurable CPU affinity () and runs at fixed priority (18, ESP_TASK_TCPIP_PRIO). To optimize CPU utilization, consider assigning competing tasks to different cores or adjusting their priorities to lower values. For additional details on built-in task priorities, please refer to .

If using select() function with socket arguments only, disabling will make select() calls faster.

If there is enough free IRAM, select and to improve TX/RX throughput.

If using a Wi-Fi network interface, please also refer to .

Except for increasing buffer sizes, most changes that increase throughput also decrease latency by reducing the amount of CPU time spent in lwIP functions.

For TCP sockets, lwIP supports setting the standard TCP_NODELAY flag to disable Nagle's algorithm.

Most lwIP RAM usage is on-demand, as RAM is allocated from the heap as needed. Therefore, changing lwIP settings to reduce RAM usage may not change RAM usage at idle, but can change it at peak.

Reducing reduces the maximum number of sockets in the system. This also causes TCP sockets in the WAIT_CLOSE state to be closed and recycled more rapidly when needed to open a new socket, further reducing peak RAM usage.

Reducing , and reduce RAM usage at the expense of throughput, depending on usage.

Reducing reduce RAM usage by limiting concurrent accepted connections.

Reducing and reduces the maximum segment lifetime in the system. This also causes TCP sockets in the TIME_WAIT and FIN_WAIT_2 states to be closed and recycled more rapidly.

Disabling can save about 39 KB for firmware size and 2 KB RAM when the system is powered up and 7 KB RAM when the TCP/IP stack is running. If there is no requirement for supporting IPV6, it can be disabled to save flash and RAM footprint.

Disabling can save about 26 KB of firmware size and 600 B RAM on power up and 6 KB RAM when the TCP/IP stack is running. If the local network supports IPv6-only configuration, IPv4 can be disabled to save flash and RAM footprint.

If using Wi-Fi, please also refer to .

The peak heap memory that lwIP consumes is the theoretically-maximum memory that the lwIP driver consumes. Generally, the peak heap memory that lwIP consumes depends on:

the memory required to create a UDP connection: lwip_udp_conn

the memory required to create a TCP connection: lwip_tcp_conn

the number of UDP connections that the application has: lwip_udp_con_num

the number of TCP connections that the application has: lwip_tcp_con_num

the TCP TX window size: lwip_tcp_tx_win_size

the TCP RX window size: lwip_tcp_rx_win_size

Some TCP-based applications need only one TCP connection. However, they may choose to close this TCP connection and create a new one when an error occurs (e.g., a sending failure). This may result in multiple TCP connections existing in the system simultaneously, because it may take a long time for a TCP connection to close, according to the TCP state machine, refer to RFC793.