This patch replaces the pinmux APIs that require users to look up an arbitrary
function number for the desired function of each pin. The replacement API
functions have intuitive names and permit users to pass board-level IO port
numbers. The API functions internally convert those to CPU-level port numbers
when necessary. Furthermore, when configuring a pin to be a digital input or
output, those API functions also perform the corresponding configuration
operation on the CPU-level GPIO port. The revised APIs halt when users attempt
to configure a currently-unsupported GPIO, specifically those in the GPIO_SUS
port range and those implemented on the expander chip EXP2. This also means
that such ports are left unconfigured during initialization, whereas the
pinmuxing for them was setup by the old implementation.
This PR simply adds a packet sent callback to the unicast connection used in the example. Every time a packet is sent the callback is called and prints the linkaddr_t dest, the MAC status of the message sent, and the link layer number of transmissions of the packet. This can be used to compute link quality estimations.
This test could sometimes fail because of a lack of free contiguous
pages in the file system. Fix this by removing the created files at the
end of each test. Besides, the test files do not have to be removed at
the beginning of each test since the file system is initially formatted.
Signed-off-by: Benoît Thébaudeau <benoit.thebaudeau.dev@gmail.com>
The r variable was used instead of i to fill the buffer, resulting in
the end of the test loop after only a single iteration. The file was not
even closed at the end of each iteration although it is opened at the
beginning of each iteration, so the available file descriptors would
very quickly be exhausted.
Signed-off-by: Benoît Thébaudeau <benoit.thebaudeau.dev@gmail.com>
The filenames were mixed up between some of the tests, thus breaking the
purpose of these tests.
Signed-off-by: Benoît Thébaudeau <benoit.thebaudeau.dev@gmail.com>
Remove the file at the beginning of the test, before opening it for
writing, in order to start the test with an empty file system, not only
after flashing the test, but also following every reboot.
Signed-off-by: Benoît Thébaudeau <benoit.thebaudeau.dev@gmail.com>
Two errors have been spotted, when IPv6 is enabled in the ravenusb
Project-Makefile:
#CONTIKI_NO_NET=1
CONTIKI_WITH_IPV6=1
The compile error results from a variable name mismatch in cdc_task.c
The variable 'r' is undeclared and should be renamed to 'route'
The linker also fails with 'undefined references'
This has been mediated by adding 'core/net' to Modules in the
Project-Makefile.
Rename guhRF platform to osd-merkur-256, previous osd-merkur platform is
now osd-merkur-128. Also check that everything is consistent.
Add both platforms to the regression tests.
Move redundant files in platform dev directory of both platforms to
cpu/avr/dev. Note that this probably needs some rework. Already
discovered some inconsistency in io definitions of both devices in the
avr/io.h includes. Added a workaround in the obvious cases.
The platform makefiles now set correct parameters for bootloader and for
reading mac-address from flash memory.
Factor the flash programming into cpu/avr and platform/osd-merkur* and
rework *all* osd example makefiles to use the new settings. Also update
all the flash.sh and run.sh to use the new settings.
The suli ledstrip modules (and osd example) have also been removed.
This patch adds a simple non-driver protection domain sample to serve
as an example for defining other non-driver protection domains. It
simply performs a ping-pong test of protection domain switching
latency during boot, including optional accesses to a private metadata
region, and prints out the results.
Now that the CBM PFS supports file removal (and a file seek stub) it is possible to have the Telnet server leverage the IDE64 support of the CBM PFS.
Note: Using the CBM PFS for the Telnet server does _not_ reduce the code size since the POSIX I/O functions are additionally still linked in because the POSIX directory functions internally use the POSIX I/O functions. And that's the very reason why the CBM PFS is _not_ activated for the C128 Telnet server: The CBM PFS for the C128 doesn't bring IDE64 support but is supposed to be used to reduce code size - but this isn't possible for the Telnet server.
This patch implements a simple, lightweight form of protection domains
using a pluggable framework. Currently, the following plugin is
available:
- Flat memory model with paging.
The overall goal of a protection domain implementation within this
framework is to define a set of resources that should be accessible to
each protection domain and to prevent that protection domain from
accessing other resources. The details of each implementation of
protection domains may differ substantially, but they should all be
guided by the principle of least privilege. However, that idealized
principle is balanced against the practical objectives of limiting the
number of relatively time-consuming context switches and minimizing
changes to existing code.
For additional information, please refer to cpu/x86/mm/README.md.
This patch also causes the C compiler to be used as the default linker
and assembler.
This patch adds an example program to print out information about the
configuration of the Intel Quark X1000 SoC Isolated Memory Regions
(IMRs), the Host System Management Mode Controls register, and the
Host Memory I/O Boundary register.
Now callback functions get the URI of the request, this allows to use a
single resource for multiple different URIs.
The is_json flag is now gone for the to-string function, instead the
macro has an is_str flag. If set this automagically produces quotes
around the string for json output.
Now from-string functions can return an error-code, 0 for success, -1
for error.
Otherwise a crash results with a bootloader compiled with a newer AVR
toolchain (e.g. Debian Jessie). If you still have an ages-old bootloader
without a jump-table, as a short-term measure you can revert this change
in your run.sh. As a long-term fix we recommend you get your bootloader
updated!
Now we manage a timezone and daylight-savings aware version of
localtime. We parse UNIX timezone strings. The default (active after the
first call to localtime or localtime_r) is CET/CEST, the timezone of
Europe/Vienna. The wallclock-time osd-example demonstrates how to set a
different timezone via the timezone resource.
Note: After startup no timezone is set. So in this state querying the
timezone resource will return an empty string. After first call to
localtime (if not timezone has been set via the timezone resource) a
query to timezone will return the default timezone string for CET/CEST.
The string returned by the localtime and utc timezones now also includes
the timezone name.
New fields tm_gmtoff and tm_zone were added to the tm structure. These
are available in BSD systems and when setting special compiler
definitions on Linux.
Note: the timezone offset information in the tm structure (tm_gmtoff)
as well as in the tz structure returned by gettimeofday (tz_minuteswest)
may be wrong sign, this code is largely untested.
This patch adds an example for I2C callbacks that is very similar to
the i2c-LSM9DS0 example except that it uses a PWM device that is built
into the platform.
This patch revises the I2C and GPIO initialization code to always be
run during platform boot rather than within each process that requires
it.
This patch also revises the gpio-output example to use a pin that is
set as an output by the default pinmux configuration. Previously, it
used a pin that was set as an output by the pinmux configuration that
is in effect when the OS does not change the pinmux configuration.
Also added a 2-minutes average wind direction value, replaced the 240 bytes buffer from the reference example, but probably there's room for efficiency improvement