Provide a test that checks that ulib operates as expected.
Add a .gitignore file for the executables thus created
There is a strange interaction with PLATFORM_LIBS which can cause the
examples to fail to build via 'make qcheck':
- CI runs 'make qcheck'
- the main Makefile sets PLATFORM_LIBS
- test/run calls test.py
- at some point test_ulib_demos() starts, with PLATFORM_LIBS set
- the test calls 'make' on examples/ulib/Makefile
- PLATFORM_LIBS is left alone, since it already has a value
- lSDL ends up not being in the link line
Thank you to Claude for helping to debug this and figure out the
PLATFORM_LIBS interaction.
Series-to: concept
Series-cc: heinrich
Cover-letter:
ulib: Complete initial U-Boot library
This series completes support for building U-Boot as a shared or static
library, enabling reuse of U-Boot functionality in external programs and
test suites.
The U-Boot library (ulib) allows developers to:
- Link against U-Boot functionality without a full U-Boot image
- Use U-Boot's OS abstraction layer, drivers, and utility functions
- Build test programs that can exercise U-Boot code in isolation
- Create applications that benefit from U-Boot's hardware support
Key features:
- Builds both shared (libu-boot.so) and static (libu-boot.a) libraries
- Preserves U-Boot linker lists for proper driver/subsystem init
- Configurable symbol renaming to avoid conflicts with system libraries
- Generated API headers with renamed function declarations
- Documentation and working examples
- Currently only supports sandbox architecture
The series includes:
- More build-infrastructure and Makefile integration
- Python-based mechanism for symbol renaming and API generation
- Test programs demonstrating basic library usage
- A simple example program showing real-world usage patterns
Symbol renaming ensures that U-Boot functions don't conflict with system
libraries. For example, printf() remains the standard library function
while ub_printf() provides access to U-Boot's printf implementation.
This is handled automatically during the build process.
The library excludes main() to allow external programs to provide their own
entry points while still accessing U-Boot functionality through ulib_init()
and ulib_uninit().
For example:
#include <u-boot-lib.h>
#include <u-boot-api.h>
int main(int argc, char *argv[])
{
if (ulib_init(argv[0]) < 0)
return 1;
ub_printf("Hello from U-Boot library!\n");
ulib_uninit();
return 0;
}
License implications are documented - the GPL-2.0+ license applies to
any programs linked with the library, requiring source code distribution
for compliant usage.
Future work will look at expanding support to other architectures.
END
Co-developed-by: Claude <noreply@anthropic.com>
Signed-off-by: Simon Glass <sjg@chromium.org>
Binman needs access to tools like mkimage and cbfstool to run all the
tests. Set up the TOOLS_DIR earlier and allow use of a TOOLS_DIR
environment variable to specify where the tools are. This can be used
from CI, for example.
Signed-off-by: Simon Glass <sjg@chromium.org>
Pytest shows about 10 lines of output of limited merit at the top. Use
the -q flag to request that it doesn't so that the actionable things are
more visible.
Signed-off-by: Simon Glass <sjg@chromium.org>
The original purpose of 'make tcheck' was to run the tools and skip any
pytests. This has been lost by the addition of sandbox_noinst and
sandbox_vpl
Restore the intended behaviour by skipping all tests with sandbox_spl
(since it still needs to be built) and skipping the other builds
entirely.
Signed-off-by: Simon Glass <sjg@chromium.org>
The current directory may be changed by a tool run from this script,
so use an absolute path when setting the DTC variable.
Signed-off-by: Simon Glass <sjg@chromium.org>
When LTO is used it seems to miss some line-number information, so the
event_dump.py tool does not fully work, thus causing test_event_dump to
fail.
Disable LTO to prevent this problem. It speeds up the build in any case.
Signed-off-by: Simon Glass <sjg@chromium.org>
The new Kbuild changes have broken this script, so 'make qcheck' and
'make pcheck' don't work anymore.
Unset some environment variables to resolve this.
Signed-off-by: Simon Glass <sjg@chromium.org>
Fixes: bf030467d2 (Merge patch series "Update kbuild")
The patman directory has a number of modules which are used by other tools
in U-Boot. This makes it hard to package the tools using pypi since the
common files must be copied along with the tool that uses them.
To address this, move these files into a new u_boot_pylib library. This
can be packaged separately and listed as a dependency of each tool.
Signed-off-by: Simon Glass <sjg@chromium.org>
For other sandbox tests the printed test name corresponds to the
configuration except for this one.
Signed-off-by: Michal Suchanek <msuchanek@suse.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Simon Glass <sjg@chromium.org>
Running tests in parallel is much faster, e.g. 15 seconds to run the tests
on sandbox (only), instead of 100 seconds (on a 16-core machine). Add a
'make pcheck' option to access this feature.
Note that the tools/ tests still run each tool's tests once after the
other, although within that, they do run in parallel. So for example,
the buildman tests run in parallel, then the binman tests run in
parallel. There would be a signiificant advantage to running them all
in parallel together, but that would require a large amount of
refactoring, e.g. with more use of pytest fixtures.
Update the documentation to represent the current state.
Signed-off-by: Simon Glass <sjg@chromium.org>
At present we use --test to indicate that tests should be run. It is
better to use a subcommand for list, like binman. Change it and adjust
the existing code to fit under a 'send' subcommand, the default.
Give this subcommand the same default arguments as the others.
Signed-off-by: Simon Glass <sjg@chromium.org>
At present we use --test to indicate that tests should be run. It is
better to use a subcommand for list, like binman. Change it and adjust
the existing code to fit under a 'send' subcommand, the default.
Give this subcommand the same default arguments as the others.
Signed-off-by: Simon Glass <sjg@chromium.org>
Running all the unit tests takes a while and is not useful when you are
just modifying the tools. Add an option to run only the tools tests.
Signed-off-by: Simon Glass <sjg@chromium.org>
This class is the new way to handle arguments in Python. Convert binman
over to use it. At the same time, introduce commands so that we can
separate out the different parts of binman functionality.
Signed-off-by: Simon Glass <sjg@chromium.org>
Tools like ifwitool may not be available in the PATH, but are available in
the build. These tools may be needed by tests, so allow tests to use the
--toolpath flag.
Also use this flag with travis.
Signed-off-by: Simon Glass <sjg@chromium.org>
At present tests are quite slow to run, over a minute on my machine. This
presents a considerable barrier to bisecting for failures.
The slowest tests are the filesystem ones and the buildman --fetch-arch
test. Add a new 'qcheck' target that skips these tests. This reduces test
time down to about 40 second, still too long, but bearable.
Signed-off-by: Simon Glass <sjg@chromium.org>
There is some basic informaton that SPL normally wants to pass through to
U-Boot, such as the SDRAM size and bank information.
Mkae use of the new bloblist structure for this. Add a new 'handoff' blob
which is set up in SPL and passed to U-Boot proper. Also adda test for
sandbox_spl that checks that this works correctly and a new 'sb' command
to show the information passed from SPL.
Reviewed-by: Tom Rini <trini@konsulko.com>
Signed-off-by: Simon Glass <sjg@chromium.org>
We only need to run driver-model tests with this config, since this is the
only thing that is different when CONFIG_OF_LIVE is not defined. Filter
out the other tests to same time.
Signed-off-by: Simon Glass <sjg@chromium.org>
The 'result' variable counts the number of failures in running the tests.
Rename it to 'failures' to make this more obvious. Also tidy up a few
comments.
Signed-off-by: Simon Glass <sjg@chromium.org>
At present the tests are run without any indication of what is running.
For the tests which start with a build this is pretty obvious, but for
tools it is not.
Add a name for each test we run, and print it before starting the test.
Signed-off-by: Simon Glass <sjg@chromium.org>
The most portable way to get access to coverage is to invoke it as
'python-coverage'.
Cc: Simon Glass <sjg@chromium.org>
Signed-off-by: Tom Rini <trini@konsulko.com>
Signed-off-by: Simon Glass <sjg@chromium.org>
Binman has 100% test coverage for the code as it is at present. To
encourage it to stay that way, run the code-coverage test as part of the
normal U-Boot tests.
This is RFC because it requires the Python code coverage tools to be
available.
Signed-off-by: Simon Glass <sjg@chromium.org>
This script runs the tests but does not report failure. Also it always
returns an exit code of 0 even on failure.
Fix these problems by checking the result of each test.
Signed-off-by: Simon Glass <sjg@chromium.org>
The standard sandbox board cannot run the of-platdata test since it needs
SPL. Also, we should test the flat tree version of sandbox.
Add these tests to the default test script.
Signed-off-by: Simon Glass <sjg@chromium.org>
A common check before sending patches is to run all available tests on
sandbox. But everytime I do this I have to look up the README. This presents
quite a barrier to actually doing this.
Add a shell script to help. To run the tests, type:
test/run
in the U-Boot directory, which should be easy to remember.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Teddy Reed <teddy.reed@gmail.com>
