- The --gcov-tool argument now has a reasonable default
if the Zephyr SDK is in use.
- --coverage-platform, if unspecified, defaults to what
was passed to --platform
- --coverage implies --enable-slow, so that tests with
the 'slow' tag are built and run.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
We have a number of timing sensitive tests which run
correctly on a much more frequent basis if the system
is not so heavily loaded. Instead of squeezing a few
more crumbs of performance by doubling the CPU count,
just use the number of CPUs reported by the system.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
If GCOV coverage is enabled, the coverage dump happens after
"PROJECT EXECUTION SUCCESSFUL" is printed. In some cases,
the additional time added was not enough to capture all the
GCOV output on a heavily loaded system before the emulator
gets killed.
Ideally, the decision to kill the emulator needs to be smarter
and less race-prone, but that can wait for a future
enhancement.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
So the time used to run boards which use the BinaryHandler can be
reported, record the time used from spawning the process until
it finnishes or is killed.
The BinaryHandler is used by the "native" boards, unit tests,
nsim and Renode.
Signed-off-by: Alberto Escolar Piedras <alpi@oticon.com>
This patch adds generation of `*_FIXED_CLOCK_FREQUENCY` macros for
clock consuming nodes that are provided with a fixed rate clock.
Signed-off-by: Andrzej Głąbek <andrzej.glabek@nordicsemi.no>
This patch is needed in order to get compilable macros for compatibles
like "arm,cortex-m0+".
Signed-off-by: Andrzej Głąbek <andrzej.glabek@nordicsemi.no>
Commit 212ec9a29a / feature #14121 already ordered partitions by
decreasing size, however it was common in samples/userspace/shared_mem/
/sample.kernel.memory_protection.shared_mem for two partitions to have
the same size and be randomly ordered between them. This adds the
partition name as a second sort key.
Unlike previous attempt in commit 725abdf430 this doesn't use the
partition name as the first (and only) key and doesn't break the
decreasing size order. Huge thanks to Sigvart Hovland for spotting this
in a post-merge but prompt code review.
Signed-off-by: Marc Herbert <marc.herbert@intel.com>
This reverts commit 725abdf430 which did get rid of randomness in the
order of the partition _names_ as claimed but regressed commit
212ec9a29a / feature #14121 and broke the previous size order which I
missed. Huge thanks to Sigvart Hovland for spotting this in a post-merge
but prompt code review. Proper fix in the next commit.
Signed-off-by: Marc Herbert <marc.herbert@intel.com>
Dictionaries are not ordered in Python 3.5 and before, so building twice
in a row could lead to a different partition order, different
build/zephyr/include/generated/app_smem_*.ld files and different
binaries.
Fix with a minor change to the "for" loop in the output function:
make it iterate on sorted(partitions.items()) instead of the raw and
randomly ordered partitions dictionary.
It is easy to reproduce the issue even without downgrading to an
obsolete Python version; pick a test like samples/userspace/shared_mem/
and simply change the code to this:
--- a/scripts/gen_app_partitions.py
+++ b/scripts/gen_app_partitions.py
@@ -159,10 +159,12 @@ def parse_elf_file(partitions):
partitions[partition_name][SZ] += size
+import random
def generate_final_linker(linker_file, partitions):
string = linker_start_seq
size_string = ''
- for partition, item in sorted(partitions.items()):
+ for partition, item in sorted(partitions.items(),
+ key=lambda x: random.random()):
string += data_template.format(partition)
if LIB in item:
for lib in item[LIB]:
Signed-off-by: Marc Herbert <marc.herbert@intel.com>
Dictionaries are not ordered in Python 3.5 and before, so building twice
${ZEPHYR_BASE}/samples/application_development/code_relocation/
in a row could lead to a different sections order, different
build/zephyr/include/generated/linker_relocate.d and code_relocation.c
and different binaries.
Fix with a minor change to three "for" loops in the output functions:
make them iterate on sorted(list of sections) instead of the raw and
randomly ordered dictionaries of sections.
Signed-off-by: Marc Herbert <marc.herbert@intel.com>
When using a build folder format with build.dir-fmt that includes any
parameters that need resolving, the west runners cannot find the folder
since the required information (board, source dir or app) is not
available.
Add a very simple heuristic to support the case where a build folder
starts with a hardcoded prefix (for example 'build/') and a single build
is present under that prefix.
The heuristic is gated behind a new configuration option:
build.guess-dir
Signed-off-by: Carles Cufi <carles.cufi@nordicsemi.no>
Use Device Tree,and in particular a new 'bt-c2h-uart' to select which
UART is being used to communicate with an external BLE Host when acting
as a Controller.
Signed-off-by: Carles Cufi <carles.cufi@nordicsemi.no>
Print a friendlier error message on ValueError, but don't throw away
the stack trace.
Move another call to log.die().
Signed-off-by: Marti Bolivar <marti.bolivar@nordicsemi.no>
Follow along with changes made in west flash/debug/etc to make it
easier to see the output steps visually.
Signed-off-by: Marti Bolivar <marti.bolivar@nordicsemi.no>
Having common log handlers now lets us improve our logging output so
that info messages are prefixed with the runner they came from, and
doing something similar with the high level steps as we go, like this:
-- west <command>: using runners
-- runners.RUNNER_NAME: doing something
<output from RUNNER_NAME subprocesses go here>
-- runners.RUNNER_NAME: all done, bye
We can also colorize the west output to make it stand out better from
subprocesses, using the same output formatting style that west
commands like west list do.
Signed-off-by: Marti Bolivar <marti.bolivar@nordicsemi.no>
I've had some requests to be able to use code in the runners package
without having west installed.
It turns out to be pretty easy to make this happen, as west is
currently only used for west.log and some trivial helper methods:
- To replace west log, use the standard logging module
- Add an appropriate handler for each runner's logger in
run_common.py which delegates to west.log, to keep
output working as expected.
Signed-off-by: Marti Bolivar <marti.bolivar@nordicsemi.no>
Add support so that we can flag any "defines" associated with a call to
either extract_cells or extract_controller as deprecated.
Signed-off-by: Kumar Gala <kumar.gala@linaro.org>
We never set 'use-prop-name' on clock bindings so lets just always
use CLOCK_CONTROLLER as the define name we generate.
Signed-off-by: Kumar Gala <kumar.gala@linaro.org>
We have 'use-prop-name' flag in the bindings which is specifically used
for GPIO properties to control if we get "GPIO" or "GPIOS" as the
generated define name.
Lets remove the inconsistancy and use "GPIOS" as the preferred name as
this matches the DTS property name. Towards that we will generate both
forms and remove support for 'use-prop-name'.
This also impacts "PWM" generation. So we'll have "PWM" and "PWMS"
Signed-off-by: Kumar Gala <kumar.gala@linaro.org>
It is useful that the ptp_clock_get() function can be called from
the userspace. Create also unit test for calling that function
from userspace.
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
A k_futex is a lightweight mutual exclusion primitive designed
to minimize kernel involvement. Uncontended operation relies
only on atomic access to shared memory. k_futex structure lives
in application memory. And when using futexes, the majority of
the synchronization operations are performed in user mode. A
user-mode thread employs the futex wait system call only when
it is likely that the program has to block for a longer time
until the condition becomes true. When the condition comes true,
futex wake operation will be used to wake up one or more threads
waiting on that futex.
This patch implements two futex operations: k_futex_wait and
k_futex_wake. For k_futex_wait, the comparison with the expected
value, and starting to sleep are performed atomically to prevent
lost wake-ups. If different context changed futex's value after
the calling use-mode thread decided to block himself based on
the old value, the comparison will help observing the value
change and will not start to sleep. And for k_futex_wake, it
will wake at most num_waiters of the waiters that are sleeping
on that futex. But no guarantees are made on which threads are
woken, that means scheduling priority is not taken into
consideration.
Fixes: #14493.
Signed-off-by: Wentong Wu <wentong.wu@intel.com>
Similar deal to commit cc14c40a2d ("kconfiglib: Unclutter symbol
strings, avoid redundant writes, misc.").
Hide the direct dependencies in the defaults, selects, and implies
sections. Do the same in menuconfig/guiconfig as well.
This uses a new Kconfiglib API, so update Kconfiglib to upstream
revision 164ef007a8. This also includes some minor optimizations and
cleanups.
Signed-off-by: Ulf Magnusson <Ulf.Magnusson@nordicsemi.no>
The privilege stacks are not sandboxed inside an MPU region,
so they do not have to be aligned with the stack buffer size.
We fix this by using the PRIVILEGE_STACK_ALIGN macro, which
is defined in arch.h and reflects the minimum alignment
requirement for privilege stack buffers.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
If the type of property is a 'array' we should generate defines as
if its a list even if theres only a single element in the list.
Signed-off-by: Kumar Gala <kumar.gala@linaro.org>
If the type of property is a 'string-list' we should generate defines as
if its a list even if theres only a single element in the list.
Signed-off-by: Kumar Gala <kumar.gala@linaro.org>
We already have the info so let's show it. This helps spots intermittent
issues[*], gives an indication of the time --build-only saves, can help
spot an overloaded test system, highlights the most time-consuming tests
which may need a longer timeout in their config, shows the effective
timeout value when one occurs... all this for a dirt cheap screen estate
price and two extra lines of code.
Sample -v output:
32/81 board123 tests/testme PASSED (qemu 2.049s)
33/81 board456 samples/hello PASSED (build)
34/81 qemu_x3 tests/kernel.stack.usage FAILED: timeout (qemu 60.029s)
see: sanity-out/qemu_x3/tests/kernel.stack.usage/handler.log
35/81 board456 tests/testme PASSED (build)
36/81 qemu_x5 tests/kernel.queue FAILED: failed (qemu 2.191s)
see: sanity-out/qemu_x5/tests/kernel.queue/handler.log
[*] running qemu in heavily packed cloud virtual machines comes to mind,
also see #12553, #14173 etc.
Signed-off-by: Marc Herbert <marc.herbert@intel.com>
Several bindings have an expectation of sub-nodes that describe the
actual infomation. The sub-nodes don't have any compatiable so we can't
key on that.
So we can add the concept of a sub-node to the YAML to handle cases like
'gpio-keys', 'gpio-leds', 'pwm-leds', etc..
The sub-node in the YAML is effective the "binding" params that describe
what properties should exist in the sub-node.
Signed-off-by: Kumar Gala <kumar.gala@linaro.org>
The alias generation wasn't doing the right thing with regards to
keeping the names consistent. We would drop the index from the define
name for aliases.
So we'd get
DT_NXP_KINETIS_GPIO_GPIO_D_IRQ
which should be
DT_NXP_KINETIS_GPIO_GPIO_D_IRQ_0
Signed-off-by: Kumar Gala <kumar.gala@linaro.org>
Add a deprecate flag to add_prop_aliases so we can make the aliases it
generates as deprecated if needed.
Signed-off-by: Kumar Gala <kumar.gala@linaro.org>
Think I understand it now, and that was the goal.
- _extract_partition() adds index-based entries. extract_partition()
adds label-based entries.
Rename them to _add_partition_index_entries() and
_add_partition_label_entries(), and call them from a top-level
extract_partition() function.
This makes the logic clearer. It took me a long time to spot it.
- Generate indicies with a simple counter and remove the _flash_area
logic. This would break if partitions were extracted more than once,
but they aren't, and now you know that they can't be.
- Rename _create_legacy_label() to add_legacy_alias() and make it
global. It doesn't use 'self'.
- Break the logic for finding the flash controller into a separate
helper function
- Add doc-comments for the new functions
- Misc. other small clean-ups
generated_dts_board.conf and generated_dts_board_unfixed.h were verified
to be identical for disco_l475_iot1 and frdm_kw41z before and after the
cleanup.
Signed-off-by: Ulf Magnusson <Ulf.Magnusson@nordicsemi.no>
Found a few annoying typos and figured I better run script and
fix anything it can find, here are the results...
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
The way sanitycheck did its ordered regexes is that it would test
every regex against every line, and store the matching lines and their
regexes in an OrderedDict and check that they happened in the right
order.
That's wrong, because it disallows matching against a line that
previously appeared (and should have been ignored) in the input
stream. The watchdog sample is the best illustration: the first boot
will (by definition) contain all the output already, but the regex has
to match against a line from the SECOND boot and not the same one it
saw earlier.
Do this the simple way: keep a counter of which regex we're trying to
apply next and increment it on a match. This is faster too as we only
need to check one pattern per line.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
There is a case where using startswith to determine if a path is a
subdirectory of another path can erroneously match. When using a
testcase root outside of ZEPHYR_BASE, an erroneous match will cause the
relative path containing ".." to get prepended to the test output
directory.
Example:
$HOME/zephyr/zephyr # ZEPHYR_BASE
$HOME/zephyr/zephyr-rust/tests # testcase root
The relative path prepended to the testcase name is ../zephyr-rust/tests
and an example test output dir is
./sanity-out/qemu_x86/../zephyr-rust/tests/rust/rust.main
In this case, the build directory escapes the board directory and is no
longer unique. Parallel tests then clobber each other.
Use pathlib instead of string matching to cover this case.
Signed-off-by: Tyler Hall <tylerwhall@gmail.com>
The start of generated/cfb_font_dice.h looked like this:
/*
* This file was automatically generated using the following command:
* /home/john/zephyrproject/zephyr/scripts/gen_cfb_font_header.py
* --input fonts/dice.png --output
* /home/john/tmp/build/zephyr/include/generated//cfb_font_dice.h
* --width 32 --height 32 --first 49 --last 54
*/
For build reproduction and "privacy" reasons, change it to this:
/*
* This file was automatically generated using the following command:
* ${ZEPHYR_BASE}/scripts/gen_cfb_font_header.py
* --input fonts/dice.png --output
* zephyr/include/generated//cfb_font_dice.h
* --width 32 --height 32 --first 49 --last 54
*/
Test with:
sanitycheck -p reel_board \
-T $ZEPHYR_BASE/samples/display/cfb_custom_font/
Signed-off-by: Marc Herbert <marc.herbert@intel.com>
pyocd 0.21.0 provides pack support 'pack support' functionality,
as opposed to current 'buitlin support'.
This new feature enables the possibility to add pyocd support
for any chip that is documented in Keil database. Then one doesn't
need anymore to wait pyocd is updated with a new target to use
pyocd with his target, as long as it is populated in Keil database.
Signed-off-by: Erwan Gouriou <erwan.gouriou@linaro.org>
Add an option that only invokes the cmake phase of sanitycheck. This
can be useful for any testing that only needs to initial generation
phase of cmake, for example device tree. Also useful if we want to
just generate compile_commands.json files from cmake via:
./sanitycheck -xCMAKE_EXPORT_COMPILE_COMMANDS=1 --cmake-only
Signed-off-by: Kumar Gala <kumar.gala@linaro.org>
Add self.require() checks before running commands. Increase test
coverage, including for this feature, while we are here.
Signed-off-by: Marti Bolivar <marti.bolivar@nordicsemi.no>
The runners/jlink.py script has a mechanism for erroring out if a host
tool is not installed. Abstract it into runners/core.py and handle it
from run_common.py. This will let it be used in more places.
Signed-off-by: Marti Bolivar <marti.bolivar@nordicsemi.no>
Add new option --report-excluded to list all those tests with bad
filtering that never build or run. This option produces accurate results
with --all but can be used with default sanitycheck options to see what
does not run/build in CI for example. (limited coverage).
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
Check the CONFIG_BUILD_OUTPUT_HEX and CONFIG_BUILD_OUTPUT_BIN options
are enabled before attempting to build signed versions of these formats.
Signed-off-by: Marti Bolivar <marti.bolivar@nordicsemi.no>
Andrew Boie