When retrieving the payload length of the payload, there was a mistake,
what was already in the buffer (buf->len) was being considered twice,
net_buf_tailroom() already takes buf->len into account, resulting in
incorrect values.
Change-Id: I3b78d55abca4a448cab5e035b9d66352a3c59110
Signed-off-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>
This patch introduces a k_sem that will block until:
- data is received, or
- the user-provided timeout expires
This change allows us to simplify our previous DNS client
implementation.
This change is related to ZEP-1357 because we are refactoring the
DNS client API removing the continuous update of the net_context_recv
routine that seems to be causing issues after the kernel's update.
Jira: ZEP-1357
Change-Id: If01c9274ac8f096f0095a2872f86be2e007212ee
Signed-off-by: Flavio Santes <flavio.santes@intel.com>
This commit updates some DNS client private routines to simplify
function signatures and reduce stack usage. Updated routines now
use the dns_context structure.
Change-Id: Ic0d347ea9d610e85eed4179aee08955cf2302436
Signed-off-by: Flavio Santes <flavio.santes@intel.com>
This commit introduces the dns_context structure.
This new structure will reduce stack overhead due to
the simplication of the dns_resolve routine signature.
Furthermore, the timeout parameter is now int32_t
instead of uint32_t.
The dns sample application is also updated.
Change-Id: I5d789656bacbd23c4654edce5d116a88dc42c354
Signed-off-by: Flavio Santes <flavio.santes@intel.com>
Moved all libc Kconfigs to where the code is and remove the default
Kconfig for selecting the minimal libc. Minimal libc is now the default
if nothing else is configured in.
Removed the options for extended libc, this obviously was restricting
features in the minimal libc without a good reason, most of the
functions are available directly when using newlib, so there is no
reason why we need to restrict those in minimal libc.
Jira: ZEP-1440
Change-Id: If0a3adf4314e2ebdf0e139dee3eb4f47ce07aa89
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
In order to keep the initialization process light-weight, remove
net_buf_pool_init() and instead perform the initialization of the pool
and buffers in a "lazy" manner. This means storing more information
in the pool, and removing any 'const' members from net_buf. Since
there are no more const members in net_buf the buffer array can be
declared with __noinit, which further reduces initialization overhead.
Change-Id: Ia126af101c2727c130651b697dcba99d159a1c76
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
If receiving a malformed MQTT packet with less than 5 bytes it's
possible to get a read one byte behind buf.
Change-Id: I34425add57c937c8fd9df5bf7b72af092d6f5f32
Signed-off-by: Stig Bjørlykke <stig@bjorlykke.org>
Until now it has been necessary to separately define a k_fifo and
an array of buffers when creating net_buf pools. This has been a bit
of an inconvenience as well as blurred the line of what exactly
constitutes the "pool".
This patch removes the NET_BUF_POOL() macro and replaces it with a
NET_BUF_POOL_DEFINE() macro that internally expands into the buffer
array and new net_buf_pool struct with a given name:
NET_BUF_POOL_DEFINE(pool_name, ...);
Having a dedicated context struct for the pool has the added benefit
that we can start moving there net_buf members that have the same
value for all buffers from the same pool. The first such member that
gets moved is the destroy callback, thus shrinking net_buf by four
bytes. Another potential candidate is the user_data_size, however
right not that's left out since it would just leave 2 bytes of padding
in net_buf (i.e. not influence its size). Another common value is
buf->size, however that one is also used by net_buf_simple and can
therefore not be moved.
This patch also splits getting buffers from a FIFO and allocating a
new buffer from a pool into two separate APIs: net_buf_get and
net_buf_alloc, thus simplifying the APIs and their usage. There is no
separate 'reserve_head' parameter anymore when allocating, rather the
user is expected to call net_buf_reserve() afterwards if something
else than 0 headroom is desired.
Change-Id: Id91b1e5c2be2deb1274dde47f5edebfe29af383a
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
This commit adds HTTP message handling support for Zephyr.
So, no network routines are involved at this level.
To add HTTP message handling support for Zephyr, we explored the
following options:
1. Importing an external project and perhaps adapting it to fit our
requirements.
The criteria to pick one codebase among all the available projects
are: licensing, correctness and performance.
2. Writing our own implementation from scratch.
We decided to import an external project instead of implementing our
own parser, mainly due to code maturity and correctness. It could take
more time to obtain a production-ready parser from scratch than adapting
a state-of-art library.
The following is a list of some projects offering similar functionality.
lighttpd (many files)
* License: revised BSD license
* Supported: active
* Comments: this parser can't be integrated to Zephyr due to
dependencies that currently are not satisficed by our SDK.
nginx (src/http/ngx_http_parse.c)
* License: 2-clause BSD-like
* Supported: very active
* Comments: this parser can't be integrated to Zephyr due to
dependencies that currently are not satisficed by our SDK.
wget (src/http-parse.c)
* License: GPL 3.0
* Supported: very active
* Comments: this code can't be included in Zephyr due to
licensing issues
curl (lib/http.c)
* License: MIT/X derivate, see:
https://curl.haxx.se/docs/copyright.html
* Supported: very active
* Comment: it must be forked and adapted to run in Zephyr.
It is not optimized for low-power devices.
nodejs http-parser (http_parser.c)
* License: nginx license (2-clause BSD-like) and MIT license
* Supported: very active
* Comments: optimized with performance in mind.
From https://github.com/nodejs/http-parser: "It does
not make any syscalls nor allocations, it does not buffer data,
it can be interrupted at anytime. It only requires about 40
bytes of data per message stream."
So, nodejs/http-parser looks a very good choice for Zephyr. In this
commit, we integrate nodejs' parser to Zephyr.
Origin: https://github.com/nodejs/http-parser/releases/tag/v2.7.1https://github.com/nodejs/http-parser/archive/v2.7.1.tar.gz
NOTE:
This patch reformats the http_parser files to reduce checkpatch
warnings. Changes made in this refactoring are available at:
Repo: https://gitlab.com/santes/http_parser/commits/refactoring1
Commit: 9ccfaa23f1c8438855211fa902ec8e7236b702b1
Jira: ZEP-346
Jira: ZEP-776
Change-Id: I29b1d47f323a5841cd4d0a2afbc2cc83a0f576f0
Signed-off-by: Flavio Santes <flavio.santes@intel.com>
To avoid the risk of overflowing when dealing with retransmission
timeouts, increase the size of their representation to 32-bits.
Change-Id: I7c9c1e00c41f5ba75e19b0fd4527f273852eb85f
Signed-off-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>
This commit adds the MQTT high-level API with Quality-of-Service
support. The following MQTT messages are covered by this commit:
CONNECT (tx), DISCONNECT (tx), PUBACK (tx, rx), PUBCOMP (tx, rx),
PUBREC (tx, rx), PUBREL (tx, rx), PUBLISH (tx), PINGREQ (tx),
SUBSCRIBE (tx), UNSUBSCRIBE (tx), CONNACK (rx), PINGRESP (rx),
SUBACK (rx) and UNSUBACK (rx).
Where 'tx' stands for transmission: routines that create and send
messages. 'rx' stands for reception: routines that receive an RX
buffer from the IP stack and parse the MQTT mesage contained in
that buffer.
Jira: ZEP-365
Jira: ZEP-591
Jira: ZEP-856
Change-Id: Ibee701a298127eb713aa3fde5aaf7d089ecd1b9d
Signed-off-by: Flavio Santes <flavio.santes@intel.com>
These cflags were only needed when using the old network stack.
Change-Id: I660e397b2648137450c45d4f2edc8ec6ec7ae774
Signed-off-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>
RFC6690[1] defines a lightweight format for listing and querying
resources and their relationships.
The RFC defines an '.well-known/core' resource that will list the
resources and their associated metadata. It also allows resources to
filtered by their attributes.
The implementation uses the fact that resources are organized in an
array: only the resources present in the array after the
"ZOAP_WELL_KNOWN_CORE_RESOURCE" will be considered, this allows a
primitive form of visibility control for the attributes.
[1] https://tools.ietf.org/html/rfc6690
Change-Id: I2bc21ddf45f20e1f749d8ac36d247474fdaa9d64
Signed-off-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>
Add an alternative to zoap_packet_get_payload(), that instead of
returning a byte array, returns the net_buf (with the COAP_MARKER
added, if needed) associated with the packet, positioned so the
application can add more data.
Change-Id: I7c955ef42f5ef8406d77da994d1673e6a69b0b6b
Signed-off-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>
This commit adds support for the MQTT protocol v3.1.1.
Specifically, this commit allows a Zephyr application to create
the following MQTT messages:
CONNACK, CONNECT, PUBLISH, PUBACK, PUBREC, PUBREL, PUBCOMP, UNSUBACK,
SUBSCRIBE, SUBACK, UNSUBSCRIBE, PINGREQ, PINGRESP, and DISCONNECT.
Furthermore, the following messages can be parsed by the routines
provided by this commit:
CONNACK, CONNECT, PUBLISH, PUBACK, PUBREC, PUBREL, PUBCOMP, UNSUBACK,
SUBSCRIBE, SUBACK, PINGREQ, PINGRESP and DISCONNECT.
NOTE: client behavior (routines with network access) and QoS will be
integrated in future patches.
The MQTT v3.1.1 specification can be found at:
http://docs.oasis-open.org/mqtt/mqtt/v3.1.1/mqtt-v3.1.1.html
Origin: Original
Jira: ZEP-365
Jira: ZEP-591
Jira: ZEP-856
Change-Id: Ie0c179370cea22f7554564692bc426a8d5c419d2
Signed-off-by: Flavio Santes <flavio.santes@intel.com>
Current net_nbuf_write() api just appends data to last fragment. And
doesn't write data based on offset. That's why renaming this api.
New net_nbuf_write() apis based on offset will be coming soon.
Change-Id: Ie8e13e5f6091a279b62b6d8b0b3928a5187e75b0
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Signed-off-by: Ravi kumar Veeramally <ravikumar.veeramally@linux.intel.com>
This commit adds the '+' operator that keeps symmetry with the line
below.
Change-Id: Icb2a5f22b5202597e3e0cc252436b422ce7b44a2
Signed-off-by: Flavio Santes <flavio.santes@intel.com>
Most applications will want to use randomly generated tokens, add a
helper for that.
Change-Id: If2a6b1d96596024afd2d2ce8e3632900adfe9c0f
Signed-off-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>
This ports zoap to the native stack.
Just for reference, and totally not scientific, here are the numbers
using the old stack:
$ size outdir/qemu_x86/zephyr.elf
text data bss dec hex filename
34657 10316 16916 61889 f1c1 outdir/qemu_x86/zephyr.elf
With yaip:
$ size outdir/qemu_x86/zephyr.elf
text data bss dec hex filename
30575 9148 6164 45887 b33f outdir/qemu_x86/zephyr.elf
Jira: ZEP-818
Change-Id: I7992a3e2af7d419081ee5a64d7cc2d49fb628ead
Signed-off-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>
The "Unsupported Content-Format" error was missing from the list of
supported errors.
Change-Id: I208d79f8949838187b877eaa0a53597d8a5bc6cb
Signed-off-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>
This commit adds support the DNS client API on top of the new
native IP stack. Some features of this implementation are:
- Support for IPv4 and IPv6
- Support for multiple concurrent queries. A net_buf structure is
required per context. See the DNS_RESOLVER_ADDITIONAL_BUF_CTR
configuration variable
Origin: Original
Jira: ZEP-793
Jira: ZEP-855
Jira: ZEP-975
Change-Id: I351a636462a1b78a412c9bce1ef3cd0fa6223a52
Signed-off-by: Flavio Santes <flavio.santes@intel.com>
snprintf() implements the ability to foce a negative value through the
(unsigned) size_t len parameter to allow the formatter to use a
maximum size string.
This is point less, we don't have as much memory and this is a recipe
for all kinds of vulnerabilities.
Kill the whole thing, the testcase it represents and thank Coverity
for finding this thing. Whatever use it had before, it has no more.
Change-Id: If422246548664699d8aa328a1b9304ef13cab7ea
Coverity-ID: 131625
Coverity-ID: 131626
Signed-off-by: Inaky Perez-Gonzalez <inaky.perez-gonzalez@intel.com>
Coverity reported a block of deadcode in _prf() that seems to be a
leftover carcass from a previous time. Replaced with a comment in case
someone decides it was needed back.
Change-Id: Id97e84f3279f807e6188371f27f6af157e6d5038
Coverity-ID: 131631
Signed-off-by: Inaky Perez-Gonzalez <inaky.perez-gonzalez@intel.com>
When an option code or length representation is encoded in a 16-bit
value, the access was wrong.
Coverity-CID: 151963
Change-Id: Ie7741998cbde348ccf490a6686e68a1ace99920e
Signed-off-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>
This is a tentative fix for CID 93807, a valid Coverity warning;
however the code is valid too.
We can choose to silence the warning or rewrite the code in a way that
makes it more verbose but keeps Coverity happy.
Coverity complains about doing an implicit fallthrough in switch case
statement. I prefer patching the code to make it explicit, as the
compiler will optimize out, to avoid having to constantly filter out
the checker's warnings.
Coverity-CID: 93807
Change-Id: I7be334d48567bf52fc2b21de043310e0f73b72db
Signed-off-by: Inaky Perez-Gonzalez <inaky.perez-gonzalez@intel.com>
If newlib is configured with --enable-newlib-nano-formatted-io,
floating-point support is split out of the formatted I/O code into
weak functions which are not linked by default. This leads to a smaller
code by about 16~20k when using newlib "printf" and/or "sscanf" but not
using floating point I/O.
Programs that need floating-point I/O support must explicitly request
linking of one or both of the floating-point functions:
_printf_float or _scanf_float.
This can be done at link time using the -u option which can be passed
to either gcc or ld.
Implemented via new configuration options:
CONFIG_NEWLIB_LIBC_FLOAT_PRINTF
CONFIG_NEWLIB_LIBC_FLOAT_SCANF
Change-Id: I57f9d9f02e6d21d6011d14de7153b1d3ba6f6e32
Signed-off-by: Juro Bystricky <juro.bystricky@intel.com>
This will add basic support for sending bodies of data that exceed the
size of a single UDP packet.
Block-wise transfers are defined in the recently adopted RFC 7959[1].
The RFC defines four options, two negotiating the block transfer, and
two for informing the size of the transfer. Depending whether the packet
is a request or a response, each option is defined as
"control" (informative) or descriptive (describes what's in the
payload).
Change-Id: Ic71275558c4afed0298d20e8712f76d53904f89f
Signed-off-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>
strncmp() expects pointers of type 'char *', we could cast option.value
to 'char *', but we can use memcmp() instead, which has the benefit of
being simpler.
Change-Id: I8c4ee4401712f3617c134d8e5b6d84e8f5cc4e1d
Signed-off-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>
EWOULDBLOCK is an old error code from BSD. Not much uses it, and on basically
all modern systems it's defined to the same value as EAGAIN, which is the
System V value for the same condition.
Jira: ZEP-982
Change-Id: I5435ce55fa47f5bd2fac5d881b5b195b025f48a2
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
A good number of option values represent numbers, so add these helpers
so applications can avoid repeating this.
Change-Id: Ied2a844c51808dcafe751a77492d00f2063de7d6
Signed-off-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>
uIP keeps the port separated from the IP addresses, so if the
application wants to communicate with a remote endpoint we must also
have the port information available.
Change-Id: I8e2b01fe5717166e1f9cebcc74b2056325b8ccc3
Signed-off-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>
This adds support for observing resources, for client and server side.
For the server side we augment the resource struct so it can hold
information about each interested observer.
For client side, observing a resource abstracted as an reply that
can be called multiple times.
Change-Id: If3f0b41e302cff357ab891e6e91ec2d41579fb92
Signed-off-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>
As there is only one response token, we can do this only once at the
beginning of the function, instead of every iteration.
Change-Id: Ibb324a1189929227bd1eb9837c5d330ff8c8dac2
Signed-off-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>
There were a few lines where the alignment was wrong.
Change-Id: I474976c218ac28564dfe9dcda0e0687ed110834e
Signed-off-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>
According to commit b71a8a4591 and Jira: ZEP-733
(libc: remove stddef.h which is provided by the compiler),
the stddef.h in zephyr code is removed and it's provided
by the compiler. The original stddef.h includes two head
files, null.h and size_t.h, which are also useless now.
So remove these two deprecated files, in case conflicting
definition.
Change-Id: Ie7163fdbd23c32759425b50f3deff2a57cc051a9
Signed-off-by: Qiu Peiyang <peiyangx.qiu@intel.com>
This is a CoAP implementation tailored for Zephyr's requirements, it
tries to use memory predictably by exposing its control structures to
the user (so the user only pays for the features that are used) and
having a lightweight packet representation (the trade-off is that
adding options is not as efficient as it could be, if there were more
information available).
Change-Id: I6f74146c4626a0c554f50b42f163a076e82805ba
Signed-off-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>
Remove the function wrappers around the 64 bit math and just use C
syntax natively, combining ops where appropriate (e.g. there was a
sequence implementing "(x<<2+x)<<2" to do "multiply by 10"). The
_ldiv5 and _rlrshift routines are non-standard ops that provide useful
abstraction, so they remain as separate functions.
Change-Id: I4d83847348fdd7be09887b833c8ccbd2aa1e4182
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
The _to_float() implementation had a somewhat kludgey hand-written 64
bit math implementation, which is unhelpful on Zephyr as all our
toolchains provide a working uint64_t runtime. This is at best just
dupicated code from libgcc, and at worst less efficient.
This patch replaces the existing 64 bit minilibrary but keeps the
uint32_t[2] API as is for ease of validation and review.
One exception is _ldiv5, a specialized divide-by-five implementation.
The 64 bit division routines are large on some architectures (ARM and
ARC in particular), not pulled in by a default Zephyr build, and will
swamp the benefit from this patch. So this includes a
refactored/improved _ldiv5 which leverages libgcc for multiword shifts
instead of just using raw division.
Note also the "noinline" attribute on _ladd(). This is a workaround
for an apparent compiler bug when built with -Og or -Os (hand-hacking
the Makefiles to build with -O0 works), perhaps due to my aliasing the
int array with a long long. This will go away in phase 2.
Change-Id: I63e8c82dabe2bfaa75b63ddb59e5f11d51be538e
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
The _to_float routine was apparently written to be able to take a 32
bit float bit representationa as well as a 64 bit double. But in a
printf routine, that can never happen per the C standard (where floats
are always promoted to doubles in varargs context).
This was just hard-configured to 1 at the top of the file, and nothing
else in the project sets "DOUBLE" to try to change it. Just remove
it. If we ever want code to convert a float to a double in memory so
we can use this routine, we have it in libgcc. Or even in hardware on
the FPU where available.
Change-Id: I796814c0fce3ce96faa34fde8da411a28c826699
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Hex, octal and decimal all had separately implemented reduction loops
to generate strings. With only a little work these can all be unified
to a single implementation that works with an arbitrary base.
Performance is probably a little lower owing to the fact that
hex/octal now requires a division per character, and the extra
"reverse the string" trick at the end of the conversion. But code
size savings are substantial.
Change-Id: I11ff376aeca1483f974d328271e19918221b2a41
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
The 'z' length specifier is the appropriate one to be used with size_t
(%zu) and ssize_t (%zd) types. Having support for this in our libc
means that we can utilize the compiler format string checks
(__printf_like) without getting warnings of incorrect format
specifiers for size_t and ssize_t variables.
Change-Id: I73fec0145692e0a59934cab548caf24c1c16a3df
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
This implements a subset of the standard inttypes.h, based on what
other functionality exists in minimal libc.
Change-Id: Ib5685a6da13768ee46acbfca734d145f7018b9e0
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
The default offsetof() implementation generates a warning
(variably modified <variable> at file scope) when used in
to define the size of an array. By using this builtin with
GCC we avoid the warning and make sure no variable-size
arrays are used.
Change-Id: Iae215f777241f7daaa061067230086dffaa8311d
Signed-off-by: Carles Cufi <carles.cufi@nordicsemi.no>
This is an adaptation of
net/ip/tinydtls/platform-specific/config-zephyr.h to allow external
imports such as KSDK to include assert.h.
Change-Id: I0afee37deb79447363037ba6b4bb9cdbc629be3a
Signed-off-by: Maureen Helm <maureen.helm@nxp.com>
The strstr function finds the first occurrence of a substring
in another string, null terminated strings are not compared, this
function is added for compability for porting other libraries (like
mbedtls)
Jira: ZEP-327
Origin: http://www.leidinger.net/freebsd/dox/libkern/html/d3/d29/
strstr_8c_source.html
Change-Id: I52aac218ce0bd86373ec60f5afc49a92c85f6319
Signed-off-by: Sergio Rodriguez <sergio.sf.rodriguez@intel.com>
We get these out of system.h instead. A clause in libc-hooks.c
for newlib added since we don't get RAM size from
CONFIG_SRAM_BASE_ADDRESS.
Change-Id: Ic35113395b951f625e8e29658afe19c525037964
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
Flavio Santes