blob: cc7bef673af19193ce20d8b180058bc6b37c681c [file] [log] [blame]
# Kconfig - IP stack config
#
# Copyright (c) 2016 Intel Corporation.
#
# SPDX-License-Identifier: Apache-2.0
#
menu "IP stack"
config NET_NATIVE
bool "Enable native IP stack"
default y
help
Enables Zephyr native IP stack. If you disable this, then
you need to enable the offloading support if you want to
have IP connectivity.
# Hidden options for enabling native IPv6/IPv4. Using these options
# avoids having "defined(CONFIG_NET_IPV6) && defined(CONFIG_NET_NATIVE)"
# in the code as we can have "defined(CONFIG_NET_NATIVE_IPV6)" instead.
config NET_NATIVE_IPV6
bool
depends on NET_NATIVE
default y if NET_IPV6
config NET_NATIVE_IPV4
bool
depends on NET_NATIVE
default y if NET_IPV4
config NET_NATIVE_TCP
bool
depends on NET_NATIVE
default y if NET_TCP
config NET_NATIVE_UDP
bool
depends on NET_NATIVE
default y if NET_UDP
config NET_OFFLOAD
bool "Offload IP stack [EXPERIMENTAL]"
help
Enables TCP/IP stack to be offload to a co-processor.
if NET_OFFLOAD
module = NET_OFFLOAD
module-dep = NET_LOG
module-str = Log level for offload layer
module-help = Enables offload layer to output debug messages.
source "subsys/net/Kconfig.template.log_config.net"
endif # NET_OFFLOAD
# Hidden option
config NET_RAW_MODE
bool
help
This is a very specific option used to built only the very minimal
part of the net stack in order to get network drivers working without
any net stack above: core, L2 etc... Basically this will build only
net_pkt part. It is currently used only by IEEE 802.15.4 drivers,
though any type of net drivers could use it.
if !NET_RAW_MODE
choice
prompt "Qemu networking"
default NET_QEMU_PPP if NET_PPP
default NET_QEMU_SLIP
depends on QEMU_TARGET
help
Can be used to select how the network connectivity is established
from inside qemu to host system. This can be done either via
serial connection (SLIP) or via Qemu ethernet driver.
config NET_QEMU_SLIP
bool "SLIP"
help
Connect to host or to another Qemu via SLIP.
config NET_QEMU_PPP
bool "PPP"
help
Connect to host via PPP.
config NET_QEMU_ETHERNET
bool "Ethernet"
help
Connect to host system via Qemu ethernet driver support. One such
driver that Zephyr supports is Intel e1000 ethernet driver.
endchoice
config NET_INIT_PRIO
int
default 90
help
Network initialization priority level. This number tells how
early in the boot the network stack is initialized.
source "subsys/net/ip/Kconfig.ipv6"
source "subsys/net/ip/Kconfig.ipv4"
config NET_SHELL
bool "Enable network shell utilities"
select SHELL
help
Activate shell module that provides network commands like
ping to the console.
config NET_SHELL_DYN_CMD_COMPLETION
bool "Enable network shell dynamic command completion"
depends on NET_SHELL
default y
help
Enable various net-shell command to support dynamic command
completion. This means that for example the nbr command can
automatically complete the neighboring IPv6 address and user
does not need to type it manually.
Please note that this uses more memory in order to save the
dynamic command strings. For example for nbr command the
increase is 320 bytes (8 neighbors * 40 bytes for IPv6 address
length) by default. Other dynamic completion commands in
net-shell require also some smaller amount of memory.
config NET_TC_TX_COUNT
int "How many Tx traffic classes to have for each network device"
default 1
range 1 8
help
Define how many Tx traffic classes (queues) the system should have
when sending a network packet. The network packet priority can then
be mapped to this traffic class so that higher prioritized packets
can be processed before lower prioritized ones. Each queue is handled
by a separate thread which will need RAM for stack space.
Only increase the value from 1 if you really need this feature.
The default value is 1 which means that all the network traffic is
handled equally. In this implementation, the higher traffic class
value corresponds to lower thread priority.
config NET_TC_RX_COUNT
int "How many Rx traffic classes to have for each network device"
default 1
range 1 8
help
Define how many Rx traffic classes (queues) the system should have
when receiving a network packet. The network packet priority can then
be mapped to this traffic class so that higher prioritized packets
can be processed before lower prioritized ones. Each queue is handled
by a separate thread which will need RAM for stack space.
Only increase the value from 1 if you really need this feature.
The default value is 1 which means that all the network traffic is
handled equally. In this implementation, the higher traffic class
value corresponds to lower thread priority.
choice
prompt "Priority to traffic class mapping"
help
Select mapping to use to map network packet priorities to traffic
classes.
config NET_TC_MAPPING_STRICT
bool "Strict priority mapping"
help
This is the recommended default priority to traffic class mapping.
Use it for implementations that do not support the credit-based
shaper transmission selection algorithm.
See 802.1Q, chapter 8.6.6 for more information.
config NET_TC_MAPPING_SR_CLASS_A_AND_B
bool "SR class A and class B mapping"
depends on NET_TC_TX_COUNT >= 2
depends on NET_TC_RX_COUNT >= 2
help
This is the recommended priority to traffic class mapping for a
system that supports SR (Stream Reservation) class A and SR class B.
See 802.1Q, chapter 34.5 for more information.
config NET_TC_MAPPING_SR_CLASS_B_ONLY
bool "SR class B only mapping"
depends on NET_TC_TX_COUNT >= 2
depends on NET_TC_RX_COUNT >= 2
help
This is the recommended priority to traffic class mapping for a
system that supports SR (Stream Reservation) class B only.
See 802.1Q, chapter 34.5 for more information.
endchoice
config NET_TX_DEFAULT_PRIORITY
int "Default network packet priority if none have been set"
default 1
range 0 7
help
What is the default network packet priority if user has not specified
one. The value 0 means lowest priority and 7 is the highest.
config NET_IP_ADDR_CHECK
bool "Check IP address validity before sending IP packet"
default y
help
Check that either the source or destination address is
correct before sending either IPv4 or IPv6 network packet.
config NET_MAX_ROUTERS
int "How many routers are supported"
default 2 if NET_IPV4 && NET_IPV6
default 1 if NET_IPV4 && !NET_IPV6
default 1 if !NET_IPV4 && NET_IPV6
range 1 254
help
The value depends on your network needs.
# Normally the route support is enabled by RPL or similar technology
# that needs to use the routing infrastructure.
config NET_ROUTE
bool
depends on NET_IPV6_NBR_CACHE
default y if NET_IPV6_NBR_CACHE
# Temporarily hide the routing option as we do not have RPL in the system
# that used to populate the routing table.
config NET_ROUTING
bool
depends on NET_ROUTE
help
Allow IPv6 routing between different network interfaces and
technologies. Currently this has limited use as some entity
would need to populate the routing table. RPL used to do that
earlier but currently there is no RPL support in Zephyr.
config NET_MAX_ROUTES
int "Max number of routing entries stored."
default NET_IPV6_MAX_NEIGHBORS
depends on NET_ROUTE
help
This determines how many entries can be stored in routing table.
config NET_MAX_NEXTHOPS
int "Max number of next hop entries stored."
default NET_MAX_ROUTES
depends on NET_ROUTE
help
This determines how many entries can be stored in nexthop table.
config NET_ROUTE_MCAST
bool
depends on NET_ROUTE
config NET_MAX_MCAST_ROUTES
int "Max number of multicast routing entries stored."
default 1
depends on NET_ROUTE_MCAST
help
This determines how many entries can be stored in multicast
routing table.
config NET_TCP
bool "Enable TCP"
help
The value depends on your network needs.
config NET_TCP_CHECKSUM
bool "Check TCP checksum"
default y
depends on NET_TCP
help
Enables TCP handler to check TCP checksum. If the checksum is invalid,
then the packet is discarded.
if NET_TCP
module = NET_TCP
module-dep = NET_LOG
module-str = Log level for TCP
module-help = Enables TCP handler output debug messages
source "subsys/net/Kconfig.template.log_config.net"
endif # NET_TCP
config NET_TCP_BACKLOG_SIZE
int "Number of simultaneous incoming TCP connections"
depends on NET_TCP
default 1
range 1 128
help
The number of simultaneous TCP connection attempts, i.e. outstanding
TCP connections waiting for initial ACK.
config NET_TCP_TIME_WAIT_DELAY
int "How long to wait in TIME_WAIT state (in milliseconds)"
depends on NET_TCP
default 250
help
To avoid a (low-probability) issue when delayed packets from
previous connection get delivered to next connection reusing
the same local/remote ports, RFC 793 (TCP) suggests to keep
an old, closed connection in a special "TIME_WAIT" state for
the duration of 2*MSL (Maximum Segment Lifetime). The RFC
suggests to use MSL of 2 minutes, but notes "This is an
engineering choice, and may be changed if experience indicates
it is desirable to do so." For low-resource systems, having
large MSL may lead to quick resource exhaustion (and related
DoS attacks). At the same time, the issue of packet misdelivery
is largely alleviated in the modern TCP stacks by using random,
non-repeating port numbers and initial sequence numbers. Due
to this, Zephyr uses much lower value of 250ms by default.
Value of 0 disables TIME_WAIT state completely.
config NET_TCP_ACK_TIMEOUT
int "How long to wait for ACK (in milliseconds)"
depends on NET_TCP
default 1000
range 1 2147483647
help
This value affects the timeout when waiting ACK to arrive in
various TCP states. The value is in milliseconds. Note that
having a very low value here could prevent connectivity.
config NET_TCP_INIT_RETRANSMISSION_TIMEOUT
int "Initial value of Retransmission Timeout (RTO) (in milliseconds)"
depends on NET_TCP
default 200
range 100 60000
help
This value affects the timeout between initial retransmission
of TCP data packets. The value is in milliseconds.
config NET_TCP_RETRY_COUNT
int "Maximum number of TCP segment retransmissions"
depends on NET_TCP
default 9
help
The following formula can be used to determine the time (in ms)
that a segment will be be buffered awaiting retransmission:
n=NET_TCP_RETRY_COUNT
Sum((1<<n) * NET_TCP_INIT_RETRANSMISSION_TIMEOUT)
n=0
With the default value of 9, the IP stack will try to
retransmit for up to 1:42 minutes. This is as close as possible
to the minimum value recommended by RFC1122 (1:40 minutes).
Only 5 bits are dedicated for the retransmission count, so accepted
values are in the 0-31 range. It's highly recommended to not go
below 9, though.
Should a retransmission timeout occur, the receive callback is
called with -ECONNRESET error code and the context is dereferenced.
config NET_UDP
bool "Enable UDP"
default y
help
The value depends on your network needs.
config NET_UDP_CHECKSUM
bool "Check UDP checksum"
default y
depends on NET_UDP
help
Enables UDP handler to check UDP checksum. If the checksum is invalid,
then the packet is discarded.
config NET_UDP_MISSING_CHECKSUM
bool "Accept missing checksum (IPv4 only)"
depends on NET_UDP && NET_IPV4
help
RFC 768 states the possibility to have a missing checksum, for
debugging purposes for instance. That feature is however valid only
for IPv4 and on reception only, since Zephyr will always compute the
UDP checksum in transmission path.
if NET_UDP
module = NET_UDP
module-dep = NET_LOG
module-str = Log level for UDP
module-help = Enables UDP handler output debug messages
source "subsys/net/Kconfig.template.log_config.net"
endif # NET_UDP
config NET_MAX_CONN
int "How many network connections are supported"
depends on NET_UDP || NET_TCP || NET_SOCKETS_PACKET || NET_SOCKETS_CAN
default 8 if NET_IPV6 && NET_IPV4
default 4
help
The value depends on your network needs. The value
should include both UDP and TCP connections.
config NET_MAX_CONTEXTS
int "Number of network contexts to allocate"
default 6
help
Each network context is used to describe a network 5-tuple that
is used when listening or sending network traffic. This is very
similar as one could call a network socket in some other systems.
config NET_CONTEXT_NET_PKT_POOL
bool "Enable net_buf TX pool / context"
default y if NET_TCP && NET_6LO
help
If enabled, then it is possible to fine-tune network packet pool
for each context when sending network data. If this setting is
enabled, then you should define the context pools in your application
using NET_PKT_TX_POOL_DEFINE() and NET_PKT_DATA_POOL_DEFINE()
macros and tie these pools to desired context using the
net_context_setup_pools() function.
config NET_CONTEXT_SYNC_RECV
bool "Support synchronous functionality in net_context_recv() API"
default y
help
You can disable sync support to save some memory if you are calling
net_context_recv() in async way only when timeout is set to 0.
config NET_CONTEXT_CHECK
bool "Check options when calling various net_context functions"
default y
help
If you know that the options passed to net_context...() functions
are ok, then you can disable the checks to save some memory.
config NET_CONTEXT_PRIORITY
bool "Add priority support to net_context"
help
It is possible to prioritize network traffic. This requires
also traffic class support to work as expected.
config NET_CONTEXT_TIMESTAMP
bool "Add timestamp support to net_context"
select NET_PKT_TIMESTAMP
help
It is possible to timestamp outgoing packets and get information
about these timestamps.
config NET_CONTEXT_TXTIME
bool "Add TXTIME support to net_context"
select NET_PKT_TXTIME
help
It is possible to add information when the outgoing network packet
should be sent. The TX time information should be placed into
ancillary data field in sendmsg call.
config NET_TEST
bool "Network Testing"
help
Used for self-contained networking tests that do not require a
network device.
config NET_SLIP_TAP
bool "TAP SLIP driver"
depends on NET_QEMU_SLIP
depends on NET_NATIVE
select SLIP
select UART_PIPE
select UART_INTERRUPT_DRIVEN
select SLIP_TAP
default y if (QEMU_TARGET && !NET_TEST && !NET_L2_BT)
help
SLIP TAP support is necessary when testing with QEMU. The host
needs to have tunslip6 with TAP support running in order to
communicate via the SLIP driver. See net-tools project at
https://github.com/zephyrproject-rtos/net-tools for more details.
config NET_TRICKLE
bool "Enable Trickle library"
help
Normally this is enabled automatically if needed,
so say 'n' if unsure.
if NET_TRICKLE
module = NET_TRICKLE
module-dep = NET_LOG
module-str = Log level for Trickle algorithm
module-help = Enables Trickle library output debug messages
source "subsys/net/Kconfig.template.log_config.net"
endif # NET_TRICKLE
endif # NET_RAW_MODE
config NET_PKT_RX_COUNT
int "How many packet receives can be pending at the same time"
default 14 if NET_L2_ETHERNET
default 4
help
Each RX buffer will occupy smallish amount of memory.
See include/net/net_pkt.h and the sizeof(struct net_pkt)
config NET_PKT_TX_COUNT
int "How many packet sends can be pending at the same time"
default 14 if NET_L2_ETHERNET
default 4
help
Each TX buffer will occupy smallish amount of memory.
See include/net/net_pkt.h and the sizeof(struct net_pkt)
config NET_BUF_RX_COUNT
int "How many network buffers are allocated for receiving data"
default 36 if NET_L2_ETHERNET
default 16
help
Each data buffer will occupy CONFIG_NET_BUF_DATA_SIZE + smallish
header (sizeof(struct net_buf)) amount of data.
config NET_BUF_TX_COUNT
int "How many network buffers are allocated for sending data"
default 36 if NET_L2_ETHERNET
default 16
help
Each data buffer will occupy CONFIG_NET_BUF_DATA_SIZE + smallish
header (sizeof(struct net_buf)) amount of data.
choice
prompt "Network packet data allocator type"
default NET_BUF_FIXED_DATA_SIZE
help
Select the memory allocator for the network buffers that hold the
packet data.
config NET_BUF_FIXED_DATA_SIZE
bool "Fixed data size buffer"
help
Each buffer comes with a built time configured size. If runtime
requested is bigger than that, it will allocate as many net_buf
as necessary to reach that request.
config NET_BUF_VARIABLE_DATA_SIZE
bool "Variable data size buffer [EXPERIMENTAL]"
help
The buffer is dynamically allocated from runtime requested size.
endchoice
config NET_BUF_DATA_SIZE
int "Size of each network data fragment"
default 128
depends on NET_BUF_FIXED_DATA_SIZE
help
This value tells what is the fixed size of each network buffer.
config NET_BUF_DATA_POOL_SIZE
int "Size of the memory pool where buffers are allocated from"
default 4096 if NET_L2_ETHERNET
default 2048
depends on NET_BUF_VARIABLE_DATA_SIZE
help
This value tell what is the size of the memory pool where each
network buffer is allocated from.
config NET_HEADERS_ALWAYS_CONTIGUOUS
bool
help
This a hidden option, which one should use with a lot of care.
NO bug reports will be accepted if that option is enabled!
You are warned.
If you are 100% sure the headers memory space is always in a
contiguous space, this will save stack usage and ROM in net core.
This is a possible case when using IPv4 only, with
NET_BUF_FIXED_DATA_SIZE enabled and NET_BUF_DATA_SIZE of 128 for
instance.
config NET_IF_USERSPACE_ACCESS
bool "Allow userspace app to manipulate network interface"
depends on USERSPACE
help
Only enable this if you want a userspace application to manipulate
network interface. Currently this is limited to add or remove
IP addresses. By default this is not allowed.
choice
prompt "Default Network Interface"
default NET_DEFAULT_IF_FIRST
help
If system has multiple interfaces enabled, then user shall be able
to choose default interface. Otherwise first interface will be the
default interface.
config NET_DEFAULT_IF_FIRST
bool "First available interface"
config NET_DEFAULT_IF_ETHERNET
bool "Ethernet"
depends on NET_L2_ETHERNET
config NET_DEFAULT_IF_BLUETOOTH
bool "Bluetooth"
depends on NET_L2_BT
config NET_DEFAULT_IF_IEEE802154
bool "IEEE 802.15.4"
depends on NET_L2_IEEE802154
config NET_DEFAULT_IF_OFFLOAD
bool "Offloaded interface"
depends on NET_OFFLOAD
config NET_DEFAULT_IF_DUMMY
bool "Dummy testing interface"
depends on NET_L2_DUMMY
config NET_DEFAULT_IF_CANBUS
bool "6LoCAN (IPv6 over CAN) interface"
depends on NET_L2_CANBUS
config NET_DEFAULT_IF_CANBUS_RAW
bool "Socket CAN interface"
depends on NET_L2_CANBUS_RAW
endchoice
config NET_PKT_TIMESTAMP
bool "Enable network packet timestamp support"
help
Enable network packet timestamp support. This is needed for
example in gPTP which needs to know how long it takes to send
a network packet.
config NET_PKT_TIMESTAMP_THREAD
bool "Create TX timestamp thread"
default y if NET_GPTP
depends on NET_PKT_TIMESTAMP
help
Create a TX timestamp thread that will pass the timestamped network
packets to some other module like gPTP for further processing.
If you just want to timestamp network packets and get information
how long the network packets flow in the system, you can disable
the thread support.
config NET_PKT_TIMESTAMP_STACK_SIZE
int "Timestamp thread stack size"
default 1024
depends on NET_PKT_TIMESTAMP_THREAD
help
Set the timestamp thread stack size in bytes. The timestamp
thread waits for timestamped TX frames and calls registered
callbacks.
config NET_PKT_TXTIME
bool "Enable network packet TX time support"
help
Enable network packet TX time support. This is needed for
when the application wants to set the exact time when the network
packet should be sent.
config NET_PROMISCUOUS_MODE
bool "Enable promiscuous mode support [EXPERIMENTAL]"
select NET_MGMT
select NET_MGMT_EVENT
select NET_L2_ETHERNET_MGMT if NET_L2_ETHERNET
help
Enable promiscuous mode support. This only works if the network
device driver supports promiscuous mode. The user application
also needs to read the promiscuous mode data.
if NET_PROMISCUOUS_MODE
module = NET_PROMISC
module-dep = NET_LOG
module-str = Log level for promiscuous mode
module-help = Enables promiscuous mode to output debug messages.
source "subsys/net/Kconfig.template.log_config.net"
endif # NET_PROMISCUOUS_MODE
source "subsys/net/ip/Kconfig.stack"
source "subsys/net/ip/Kconfig.mgmt"
source "subsys/net/ip/Kconfig.stats"
source "subsys/net/ip/Kconfig.debug"
endmenu