User callbacks allow code to respond to incoming messages without blocking
or polling. User callbacks are optional, and if no callback is registered
the code has no effect.
The mechanism is the same as for TCP: when a connection is accepted or a
TCP packet is received, a user callback is executed.
Fixes issue #3594.
Signed-off-by: Jack Whitham <jack.d.whitham@gmail.com>
Because socket objects have a finaliser they must be created carefully, in
case an exception is raised during the population of their members, eg
invalid input argument or out-of-memory when allocating additional arrays.
Prior to the fix in this commit, the finaliser would crash due to
`incoming.udp_raw.array` being an invalid pointer in the following cases:
- if a SOCK_RAW was created with a proto argument that was not an integer
- if a SOCK_DGRAM or SOCK_RAW was created where the allocation of
`lwip_incoming_packet_t` failed
- if an integer was passed in for the socket type but it was not one of
SOCK_STREAM, SOCK_DGRAM or SOCK_RAW
Furthermore, if the allocation of `lwip_incoming_packet_t` failed then it
may have led to corruption within lwIP when freeing `socket->pcb.raw`
because that PCB was not fully set up with its callbacks.
This commit fixes all of these issues by ensuring:
- `pcb.tcp` and `incoming.udp_raw.array` are initialised to NULL early on
- the proto argument is parsed before allocating the PCB
- the allocation of `lwip_incoming_packet_t` occurs befor allocating the
PCB
- `incoming.udp_raw.array` is checked for NULL in the finaliser code
The corresponding test (which already checked most of these causes of
failure) has been updated to include a previously-uncovered scenario.
Signed-off-by: Damien George <damien@micropython.org>
Fixes rp2 issue where socket.getaddrinfo() could block indefinitely if an
interface goes down and still has a DNS server configured, as the LWIP
timer stops running and can't time out the DNS query.
Adds a regression test under multi_wlan that times out on rp2 without this
fix.
Fixes issue #18797.
This work was funded through GitHub Sponsors.
Signed-off-by: Angus Gratton <angus@redyak.com.au>
If the return value is set only when certain condition are true,
the compiler sometimes raises an error.
Signed-off-by: robert-hh <robert@hammelrath.com>
This commit fixes a long standing bug/deficiency in the lwIP socket code,
whereby it would abandon all incoming TCP data if the remote sent a TCP
RST.
This behaviour it tested by the existing
`tests/multi_net/tcp_client_rst.py` and
`tests/multi_net/asyncio_tcp_client_rst.py` tests, and they both fail on
boards like PYBD_SFx and RPI_PICO_W due to the deficiency.
With the fix here, both of those tests now pass on lwIP targets, along with
all existing socket tests.
Signed-off-by: Damien George <damien@micropython.org>
The lwIP includes are now port-specific.
The `sys_arch_{,un}protect` functions are not used according to line 34
of `extmod/lwip-include/lwipopts_common.h`.
Neither have been touched in nearly a decade.
Signed-off-by: Thomas Watson <twatson52@icloud.com>
`socket.getaddrinfo()` supports the specification of an address family;
however, the LwIP implementation does not use it. This change allows the
application to specify the address family request in DNS resolution. If
no family is specified, it falls back to the default preference
configured with `network.ipconfig()`.
Signed-off-by: Jared Hancock <jared.hancock@centeredsolutions.com>
As timeout is of type `mp_uint_t`, it must be printed with UINT_FMT.
Before, the compiler plugin produced an error in the PYBD_SF6 build, which
is a nanboxing build with 64-bit ints.
Signed-off-by: Jeff Epler <jepler@gmail.com>
Add check to prevent calling recv on a socket in the listening state. This
prevents a crash/hard fault when user code mistakenly tries to recv on the
listening socket instead of on the accepted connection.
Add corresponding test case to demonstrate the bug.
Signed-off-by: Andrew Leech <andrew.leech@planetinnovation.com.au>
The bare-metal lwIP socket interface is currently quite limited when used
for UDP streams, because it only allows one outstanding incoming UDP
packet. If one UDP packet is waiting to be socket.recv'd and another one
comes along, then the second one is simply dropped.
This commit implements a queue for incoming UDP and raw packets. The queue
depth is fixed at compile time, and is currently 4.
This allows better use of UDP connections, eg more efficient. It also
makes DTLS work better which sometimes has a queue of UDP packets (eg
during the connection phase).
Signed-off-by: Damien George <damien@micropython.org>
This commit upgrades from codespell==2.2.6 to the current codespell==2.4.1,
adding emac to the ignore-words-list.
Signed-off-by: Christian Clauss <cclauss@me.com>
For IPv6 connections, the peer address was previously defined as only the
first four bytes of the IP address. For IPv6 addresses, this resulted in an
incorrect IPv4 address. For instance, receiving a packet via `::recvfrom`
from `'fe80::87:e7ff:fe48:629a'` is returned as having a peer address of
`'254.128.0.0'`
Signed-off-by: Jared Hancock <jared.hancock@centeredsolutions.com>
This was missed in 628abf8f25. The the bug
was that, when IPv6 is enabled, the `sizeof(ip_addr_t)` is much larger than
IPv4 size, which is what's needed for IGMP addressing.
Fixes issue #16100.
Signed-off-by: Damien George <damien@micropython.org>
Fixes a compile error if STM32 port is compiled with:
make BOARD=(..) MICROPY_PY_LWIP=1 MICROPY_PY_LWIP_SLIP=1
`sio_send()` and `sio_tryread()` now use `mp_get_stream`.
Signed-off-by: Carl Treudler <cjt@users.sf.net>
Use new function mp_obj_new_str_from_cstr() where appropriate. It
simplifies the code, and makes it smaller too.
Signed-off-by: Jon Foster <jon@jon-foster.co.uk>
This makes the code a bit simpler to understand for the three cases of
timeout behaviour (-1, 0, non-zero), and eliminates a dependency on the
(slow) `mp_hal_delay_ms(100)` call.
Signed-off-by: Damien George <damien@micropython.org>
If the socket timeout is 0 then a failed socket.connect() raises
EINPROGRESS (which is what the lwIP bindings already did), but if the
socket timeout is non-zero then a failed socket.connect() should raise
ETIMEDOUT. The latter is fixed in this commit.
A test is added for these timeout cases.
Signed-off-by: Damien George <damien@micropython.org>
This adds support to use the Nagle algorithm implemented already in lwIP to
determine when TCP data should be sent.
As currently written, MicroPython will only create packets if there is <25%
remaining in the send buffer. Using it, sending a small message of ~50
bytes will not trigger output of the message on the network. So it will
remained queued until the TCP interval timer expires, which can be up to
500ms.
Using Nagle's algorithm, the first write, no matter how small, will
generate a packet on the network. And sending lots of data still makes
efficient use of the link.
In addition to this, an application designer may choose to always create
packets for every write by setting the TCP_NODELAY socket option. That's
also implemented in this commit.
This commit implements a new <AbstractNIC>.ipconfig() function for the NIC
classes that use lwIP, to set and retrieve network configuration for a NIC.
Currently this method supports:
- ipconfig("addr4"): obtain a tuple (addr, netmask) of the currently
configured ipv4 address
- ipconfig("addr6"): obtain a list of tuples (addr, state,
prefered_lifetime, valid_lifetime) of all currently active ipv6
addresses; this includes static, slaac and link-local addresses
- ipconfig("has_dhcp4"): whether ipv4 dhcp has supplied an address
- ipconfig("has_autoconf6"): if there is a valid, non-static ipv6 address
- ipconfig(addr4="1.2.3.4/24"): to set the ipv4 address and netmask
- ipconfig(addr6="2a01::2"): to set a static ipv6 address; note that this
does not configure an interface route, as this does not seem supported by
lwIP
- ipconfig(autoconf6=True): to enable ipv6 network configuration with slaac
- ipconfig(gw4="1.2.3.1"): to set the ipv4 gateway
- ipconfig(dhcp4=True): enable ipv4 dhcp; this sets ipv4 address, netmask,
gateway and a dns server
- ipconfig(dhcp4=False): stops dhcp, releases the ip, and clears the
configured ipv4 address.
- ipconfig(dhcp6=True): enable stateless dhcpv6 to obtain a dns server
There is also a new global configuration function network.ipconfig() that
supports the following:
- network.ipconfig(dns="2a01::2"): set the primary dns server (can be a
ipv4 or ipv6 address)
- network.ipconfig(prefer=6): to prefer ipv6 addresses to be returned as
dns responses (falling back to ipv4 if the host does not have an ipv6
address); note that this does not flush the dns cache, so if a host is
already in the dns cache with its v4 address, subsequent lookups will
return that address even if prefer=6 is set
This interface replaces NIC.ifconfig() completely, and ifconfig() should be
marked as deprecated and removed in a future version.
Signed-off-by: Felix Dörre <felix@dogcraft.de>
The STATIC macro was introduced a very long time ago in commit
d5df6cd44a. The original reason for this was
to have the option to define it to nothing so that all static functions
become global functions and therefore visible to certain debug tools, so
one could do function size comparison and other things.
This STATIC feature is rarely (if ever) used. And with the use of LTO and
heavy inline optimisation, analysing the size of individual functions when
they are not static is not a good representation of the size of code when
fully optimised.
So the macro does not have much use and it's simpler to just remove it.
Then you know exactly what it's doing. For example, newcomers don't have
to learn what the STATIC macro is and why it exists. Reading the code is
also less "loud" with a lowercase static.
One other minor point in favour of removing it, is that it stops bugs with
`STATIC inline`, which should always be `static inline`.
Methodology for this commit was:
1) git ls-files | egrep '\.[ch]$' | \
xargs sed -Ei "s/(^| )STATIC($| )/\1static\2/"
2) Do some manual cleanup in the diff by searching for the word STATIC in
comments and changing those back.
3) "git-grep STATIC docs/", manually fixed those cases.
4) "rg -t python STATIC", manually fixed codegen lines that used STATIC.
This work was funded through GitHub Sponsors.
Signed-off-by: Angus Gratton <angus@redyak.com.au>
Prior to commit 628abf8f25 which added IPv6
support, binding a socket with
s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
s.bind(("", PORT))
was possible. But, the empty string is not regarded as a valid IP address
by lwip. This commit adds a special case for the empty IP string,
restoring the previous CPython-compatible behaviour.
Signed-off-by: Felix Dörre <felix@dogcraft.de>
With these changes IPv6 works on the rp2 port (and possibly others that use
the lwIP socket implementation).
Things that have been tested and work:
- Neighbour solicitation for v6 link local address.
- Ping of v6 link-local address.
- Receiving a SLAAC address via router advertisement.
- Ping a v6 address allocated via SLAAC.
- Perform an outgoing connection to a routed v6-address (via default
gateway).
- Create a listening IPv6 wildcard socked bound to ::, and trying to access
it via link-local, SLAAC, and IPv4 (to ensure the dual-stack binding
works).
Things that could be improved:
- socket.socket().getaddrinfo only returns the v4 address. It could also
return v6 addresses (getaddrinfo is actively programmed to only return a
single address, and this is the v4-address by default, with fallback to
the v6 address if both are enabled).
Signed-off-by: Felix Dörre <felix@dogcraft.de>
See previous commit for details of these functions. As of this commit,
these still call the old hook macros on all ports.
This work was funded through GitHub Sponsors.
Signed-off-by: Angus Gratton <angus@redyak.com.au>
Updates any includes, and references from Makefiles/CMake.
This essentially reverts what was done long ago in commit
136b5cbd76
This work was funded through GitHub Sponsors.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
This renames the builtin-modules, such that help('modules') and printing
the module object will show "module" rather than "umodule".
This work was funded through GitHub Sponsors.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
The actual underlying error number raised from the lwIP subsystem when
attempting to listen on a socket is swallowed and replaced with an
out-of-memory error which is confusing.
This commit passes the underlying error message from the lwIP subsystem to
the appropriate OSError exception.
Instead of being an explicit field, it's now a slot like all the other
methods.
This is a marginal code size improvement because most types have a make_new
(100/138 on PYBV11), however it improves consistency in how types are
declared, removing the special case for make_new.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
Previously the desired output type was specified. Now make the type part
of the function name. Because this function is used in a few places this
saves code size due to smaller call-site.
This makes `mp_obj_new_str_type_from_vstr` a private function of objstr.c
(which is almost the only place where the output type isn't a compile-time
constant).
This saves ~140 bytes on PYBV11.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
This uses MP_REGISTER_ROOT_POINTER() to register lwip_slip_stream
instead of using a conditional inside of mp_state_vm_t.
Signed-off-by: David Lechner <david@pybricks.com>
The following changes are made:
- Guard entire file with MICROPY_PY_LWIP, so it can be included in the
build while still being disabled (for consistency with other extmod
modules).
- Add modlwip.c to list of all extmod source in py/py.mk and
extmod/extmod.cmake so all ports can easily use it.
- Move generic modlwip GIT_SUBMODULES build configuration code from
ports/rp2/CMakeLists.txt to extmod/extmod.cmake, so it can be reused by
other ports.
- Remove now unnecessary inclusion of modlwip.c in EXTMOD_SRC_C in esp8266
port, and in SRC_QSTR in mimxrt port.
Signed-off-by: Damien George <damien@micropython.org>
It's no longer needed because this macro is now processed after
preprocessing the source code via cpp (in the qstr extraction stage), which
means unused MP_REGISTER_MODULE's are filtered out by the preprocessor.
Signed-off-by: Damien George <damien@micropython.org>
Jack Whitham
Damien George
Angus Gratton
robert-hh
Daniël van de Giessen
Thomas Watson
Jared Hancock
Jeff Epler
Andrew Leech
Christian Clauss
cajt
Jon Foster
Felix Dörre
Wang Xuancong
Jim Mussared
MrJake222
David Lechner