Fixes race condition for k_event_post_internal() in an
SMP environment while walking the waitq. Uses z_sched_waitq_walk()
to safely walk the waitq by using a sched_spinlock.
It should be noted that since walking the wait queue is an
operation of indeterminant length, there exists the possibility
that the sched_spinlock (which is a highly used and contended-for
lock) may be locked for an indeterminant amount of time. However,
it is expected that few threads will be waiting on any given kernel
event object, which should ameliorate this risk.
Fixes#54317
Signed-off-by: Aastha Grover <aastha.grover@intel.com>
Removed struct variables for storing states: sx, plt_rst, espi_rst.
While Sx & espi_rst state variables were totally redundant, plt_rst
variable had a use-case to prevent sending multiple callbacks to app.
Experimentally proven that use-case was not valid, as plt_rst isr
itself gets called only upon rising/falling edge.
Removing the condition to check plt_rst's current and previous state,
also solves the problem where global reset do not update plt_rst state.
Signed-off-by: Aditya Bhutada <aditya.bhutada@intel.com>
The current setting of `CONFIG_ISR_STACK_SIZE=256` in the
`nucleo_f031k6_defconfig` for the `nucleo_f031k6` creates a stack
overflow in the `ISR0` thread. The stack is simply too small.
This can be verified by increasing `CONFIG_ISR_STACK_SIZE=512` and
then build some of the sample applications (fx. `samples/basic/blinky`
and `samples/helloworld`).
As part of the study the `THREAD_ANALYZER` was enabled, to get some
insigt to what a proper new stack size for `ISR0` should be. The used
config is shown below.
```
CONFIG_THREAD_ANALYZER=y
CONFIG_THREAD_ANALYZER_USE_PRINTK=y
CONFIG_THREAD_ANALYZER_AUTO=y
CONFIG_THREAD_ANALYZER_AUTO_INTERVAL=5
CONFIG_THREAD_ANALYZER_AUTO_STACK_SIZE=540
CONFIG_THREAD_NAME=y
CONFIG_THREAD_RUNTIME_STATS=y
```
During the tests the following results were recorded for the
`samples/basic/blinky` and `samples/helloworld`. As seen the `ISR0`
stack usage is 80% for the two use cases, which leaves a little
headroom for other samples and applications.
**samples/basic/blinky**:
```
Thread analyze:
thread_analyzer : STACK: unused 96 usage 504 / 600 (84 %); CPU: 0 %
: Total CPU cycles used: 12472939
idle : STACK: unused 104 usage 48 / 152 (31 %); CPU: 99 %
: Total CPU cycles used: 1679934713
main : STACK: unused 284 usage 228 / 512 (44 %); CPU: 0 %
: Total CPU cycles used: 274737
ISR0 : STACK: unused 100 usage 412 / 512 (80 %)
```
**samples/helloworld**:
```
Thread analyze:
thread_analyzer : STACK: unused 40 usage 504 / 544 (92 %); CPU: 0 %
: Total CPU cycles used: 2634078
idle : STACK: unused 104 usage 48 / 152 (31 %); CPU: 99 %
: Total CPU cycles used: 480004798
ISR0 : STACK: unused 100 usage 412 / 512 (80 %)
```
As part of the test it was also checked if the stack usage of `ISR0`
could be lowered by setting `CONFIG_UART_INTERRUPT_DRIVEN=n` in the
`nucleo_f031k6_defconfig`. This didn't seem to have any effect on the
`ISR0` stack utilization.
Signed-off-by: Carsten Thuesen <cvt@thucon.dk>
Moving timeslice events to timeouts isn't quite enough on SMP, as it's
still possible for systems that don't broadcast their timer interrupts
to end up handling an expiration for a foreign CPU. There, we need an
IPI, and a symmetric call to z_time_slice() (which is itempotent and
fast) in the IPI ISR.
Signed-off-by: Andy Ross <andyross@google.com>
Rework the fragile and ad-hoc computation of timeslice expirations
into per-CPU struct _timeout objects with regular callbacks. The
expiration callbacks themselves simply set a per-cpu flag (they might
run on any CPU), which gets checked at the end of the timer ISR on
every CPU.
This simplifies logic and removes a bunch of code. It also fixes at
least three bugs:
1. As @npitre discovered: On SMP, the number of ticks announced on any
given CPU is going to be a subset of all expired ticks. This broke
the accounting of timeslice ticks, and effectively meant that
timeslicing only worked on SMP on systems where one CPU could hog all
the announcements, and only on that CPU.
2. The bootstrap path to arm the timer driver after setting the first
timeout in an empty list couldn't take into account
sys_clock_elapsed() ticks, as it didn't know whether it was being
called underneath an existing announce loop. Now this code is no
longer responsible for knowing anything about time slicing at all.
3. Also on SMP, there was a case where two CPUs timeslicing
simultaneously could stomp on each others' timeouts in
z_set_timeout_expiry(), as neither had a way of knowing what the
other's state was. CPUs could miss their own expiration and have to
wait for the slice expiration on the other CPU. Now, timeouts are
global objects with simple expiration times, and there's no need for
that function at all.
Signed-off-by: Andy Ross <andyross@google.com>
The z_set_timeout_expiry() function is a wrapper provided in the
timeout subsystem that has some corrections for timelicing. But the
suspend code doesn't care; by definition there are no active threads
that might be timeslicing at the point where the OS is suspending.
Use the lower-level timer driver API instead.
(z_set_timeout_expiry() is also about to disappear)
Signed-off-by: Andy Ross <andyross@google.com>
1. `zsock_socket()` gets the right packet familiy.
2. `inet_pton()` returns 1 on success.
This should address #55193.
Signed-off-by: Sebastian Arnd <sebastianarnd@gmail.com>
The Dead Battery resistors interfere with port partner
detection. So, Dead Battery is disabled after the system
starts and sets the Rd or Rp resistors on the CC lines.
Tested on b_g474e_dpow1 with JP5 set to USBC.
Tested on stm32g081b_eval with JP17 set to D5V.
Signed-off-by: Sam Hurst <sbh1187@gmail.com>
If west is installed, we have to build the Zephyr documentation with a
recent version of west. This is because the west API documentation is
part of the Zephyr documentation, and the west API documentation
contains Sphinx autodoc directives which pull API documentation out of
the west source code itself. This is similar to how we pull Doxygen
comments out of the Zephyr API headers using breathe. (If west is not
installed, we don't build the west API docs, so you can uninstall west
if you don't want this.)
The current API docs require west v1.0.0 or later to build, since that
version of west introduced new APIs and documentation for them not
available in prior versions.
Signed-off-by: Martí Bolívar <marti.bolivar@nordicsemi.no>
Different users of the Zephyr CMake package may have different minimum
required versions of west. One important in-tree example is the
documentation, which must either be built with the latest version of
west, or with no west installed at all.
Make the MIN_WEST_VERSION variable configurable to support use cases
like this.
Signed-off-by: Martí Bolívar <marti.bolivar@nordicsemi.no>
Fixes an issue where thread preemption was not being disabled
correctly, failing to taking into account MetaIRQ, which can be
used to preempt any cooperative thread.
The updated code sets the current thread to `K_HIGHEST_THREAD_PRIO`
before calling the secure function, and restores the thread priority
level once secure execution has terminated, allowing the thread
to be preempted once again.
Signed-off-by: Kevin Townsend <kevin.townsend@linaro.org>
Provide a means of declaring zero or more HTTP services, each
with zero or more static HTTP resources.
Static HTTP resources are those which have fixed paths[1] which
are known prior to system initialization. Some examples of
static http resources would be
* a forwarder from '/' to '/index.html'
* a REST endpoint with fixed path '/api/foo' and detail
pointing at some implementation-specific function
* a Javascript file in string form with fixed path '/js/util.js'
* a 'construction' image with path '/res/work.png'
* a gzip-compressed 'Hello' HTML file at '/hello.html'
Without describing in any detail how static HTTP resources are
organized or served by any given HTTP server, we can describe
what static resources exist on a system in a common way that
does not require any optional facilities (e.g. filesystem) and
relies only on addressable memory.
Additionally, for the purposes of simply allowing others
to implement custom HTTP servers in a consistent way, or
benchmarking implementations, or having a consistent testsuite
to use across multiple implementations, it is helpful to have
a common method to declare HTTP services and static resources
for Zephyr.
[1] https://en.wikipedia.org/wiki/URL
Signed-off-by: Chris Friedt <cfriedt@meta.com>
This make is easier to verify if proper use of BTP is maintained.
tester_rsp and tester_rsp_full will be removed eventually when
BTP is fully async.
Signed-off-by: Szymon Janc <szymon.janc@codecoup.pl>
We always have 0 index (and occasional NONE index) so there is no
point in validating those in each handler.
Signed-off-by: Szymon Janc <szymon.janc@codecoup.pl>
bt_addr_le_t matches BTP convention for addresses and it makes it
so much better to handle those directly from command structures.
Signed-off-by: Szymon Janc <szymon.janc@codecoup.pl>
This changes the way of BTP commmand handlers are called. Instead of
calling functions to handle service we now provide API for registering
callbacks per command. This allows to keep common checks (size etc)
in single place for (most) commands. Also provide common buffer for
reponses forcing callbacks to handle replay in common way.
In this commit only Core and GAP services are converted. This leaves
rest of services unfuntional (but code stil compiles correctly).
Other services are converted in following commits.
Signed-off-by: Szymon Janc <szymon.janc@codecoup.pl>
In the unicast client we should reset the stream before calling
the released callback. This is so that we can reuse the stream object
in the released callback.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
Adding an Openthread radio API `otPlatRadioSetMacFrameCounterIfLarger`
implementation and the corresponding call to the IEEEE802154 driver.
Signed-off-by: Przemyslaw Bida <przemyslaw.bida@nordicsemi.no>
In the adc_stm32_init() function, when adc_stm32_calib() is called,
the ADC is not yet enabled but still disabled.
This patch makes sure to Disable the ADC before its calibration.
Signed-off-by: Francois Ramu <francois.ramu@st.com>
USB Mass Storage Specification for Bootability requires MODE SENSE(10)
command. MODE SENSE(6) and MODE SENSE(10) generally access the same data
but differ in the maximum allocated length and LLBAA support. However
there is no point in extracting common handling because there no mode
pages are supported now.
Fail MODE SENSE requests if asking for anything other than supported
page codes (to which the essentially hardcoded response is valid).
Signed-off-by: Tomasz Moń <tomasz.mon@nordicsemi.no>
There are a couple of places where read access to uninitialized memory
happens in new mesh-1.1 tests. Valgrind highlights them.
PR fixes them.
Signed-off-by: Aleksandr Khromykh <aleksandr.khromykh@nordicsemi.no>
