Signed-off-by: Bartosz Golaszewski <bartekgola@gmail.com>

Add release notes for libgpiod v0.3.3.

Signed-off-by: Bartosz Golaszewski <bartekgola@gmail.com>

Starting with linux v4.16 the direction of gpio-mockup lines is input
by default as is the case for all real gpio drivers.

Modify the expected strings in test cases for gpioinfo and make v4.16
the minimum kernel version to run libgpiod tests.

Signed-off-by: Bartosz Golaszewski <bartekgola@gmail.com>

The attributes must be located next to function declarations in the
header file. Otherwise, the NORETURN attribute seems to have no effect
and GCC emits several Wimplicit-fallthrough warnings when compiling
libgpiod. This patch fixes those warnings.

Signed-off-by: Clemens Gruber <clemens.gruber@pqgruber.com>
Signed-off-by: Bartosz Golaszewski <bartekgola@gmail.com>

Doxyfile was missing from release tarballs. Include it.

Reported-by: SZ Lin <sz.lin@moxa.com>
Signed-off-by: Bartosz Golaszewski <bartekgola@gmail.com>

The license type of COPYING file is "LGPL-2.1+", but it is declared
as "LGPL-2.1" in the preamble of source files.

This patch fixes license type inconsistency issue in each source file.

Signed-off-by: SZ Lin <sz.lin@moxa.com>
[Bartosz: tweaked the commit message]
Signed-off-by: Bartosz Golaszewski <bartekgola@gmail.com>

We plan to change the way gpio-mockup platform devices are registered
in the kernel. This change will add an ID to the gpio-mockup platform
device name and will result in the devpath being changed to
'/devices/platform/gpio-mockup.X/gpiochip'.

When checking the devpath during the test case setup, it's enough to
verify that we got an event from the correct platform device - we
don't need to check the gpiochip name. Shorten the string against
which we check to skip whatever comes after the platform device name.

While we're at it: move the string definition into the function which
is the only user.

Signed-off-by: Bartosz Golaszewski <bartekgola@gmail.com>

Signed-off-by: Bartosz Golaszewski <bartekgola@gmail.com>

Add release notes for libgpiod v0.3.2.

Signed-off-by: Bartosz Golaszewski <bartekgola@gmail.com>

We're not using errnos in v0.3.x branch. Set the thread-local error
code variable instead. While we're at it: update the relevant test
case.

Signed-off-by: Bartosz Golaszewski <bartekgola@gmail.com>

If a GPIO chip with given label cannot be found but no other error
occurred, set errno to ENOENT before returning a NULL pointer.

Signed-off-by: Bartosz Golaszewski <bartekgola@gmail.com>

In some configurations ar prints the following warnings:

  ar: `u' modifier ignored since `D' is the default (see `U')

Silence it by setting the ar flags to 'cr' instead of 'cru' which is
the default.

Signed-off-by: Bartosz Golaszewski <bartekgola@gmail.com>

Setting the open-drain or open-source flags for input doesn't make any
sense as these flags are only valid for output mode. This should
actually be checked in the kernel. For now fix the test case.

Signed-off-by: Bartosz Golaszewski <bartekgola@gmail.com>

We set the consumer string to "gpioset" in the call to
gpiod_simple_get_value_multiple(). Make it "gpioget" as it should be.

Signed-off-by: Bartosz Golaszewski <bartekgola@gmail.com>

We call gpiod_version_string() from tools-common.c but adding
libgpiod.la to LDADD before libtools-common.la makes the linker
discard this symbol as unused.

Invert the LDADD order to fix the static build.

Signed-off-by: Bartosz Golaszewski <bartekgola@gmail.com>

We want to handle EINTR as a timeout event. Don't return here.

Signed-off-by: Bartosz Golaszewski <bartekgola@gmail.com>

Signed-off-by: Bartosz Golaszewski <bartekgola@gmail.com>

Add release notes for libgpiod v0.3.1.

Signed-off-by: Bartosz Golaszewski <bartekgola@gmail.com>

We now host libgpiod at kernel.org, so no need for the .md file.

This also fixes 'make dist' as README.md was not being packaged by
autotools.

Signed-off-by: Bartosz Golaszewski <bartekgola@gmail.com>

The test case in which we verify if falling-edge events are correctly
ignored was failing due to a bug in the kernel. This has now been
fixed with commit df1e76f28ffe ("gpiolib: skip unwanted events, don't
convert them to opposite edge") and released in linux v4.12.6.

Re-enable the test and make this version a hard requirement for
running the test suite.

Signed-off-by: Bartosz Golaszewski <bartekgola@gmail.com>

Previous check was incorrect and would break at major version change.

Signed-off-by: Bartosz Golaszewski <bartekgola@gmail.com>

Signed-off-by: Bartosz Golaszewski <bartekgola@gmail.com>

Add release notes for libgpiod v0.3.

Signed-off-by: Bartosz Golaszewski <bartekgola@gmail.com>

Verify that gpiomon bails out if the same line offset is given more
than once.

Signed-off-by: Bartosz Golaszewski <bartekgola@gmail.com>

Make sure gpiomon behaves correctly when the requested line offsets
are not in order.

Signed-off-by: Bartosz Golaszewski <bartekgola@gmail.com>

Switch from the simple API to low-level libgpiod routines and
implement watching multiple lines at the same time.

Update tests and README.

Signed-off-by: Bartosz Golaszewski <bartekgola@gmail.com>

We must use the GPIOEVENT_REQUEST_* flags, not GPIOEVENT_EVENT_* since
the latter are only meaningful when receiving event info.

Signed-off-by: Bartosz Golaszewski <bartekgola@gmail.com>

Say more than just the name of the failed function.

Signed-off-by: Bartosz Golaszewski <bartekgola@gmail.com>

We're using the same character for comments and the command prompt.
That's not very readable. Use '$' as prompt.

Signed-off-by: Bartosz Golaszewski <bartekgola@gmail.com>

In the examples section: show how to specify a custom, easily parsable
output in gpiomon.

Signed-off-by: Bartosz Golaszewski <bartekgola@gmail.com>

Verify that we actually ignore events we're not watching.

Signed-off-by: Bartosz Golaszewski <bartekgola@gmail.com>

Commit 6876721f ("build: use a more standard autogen.sh") made
the supplied autogen execute ./configure unless the NOCONFIGURE
environment variable is set. Reflect that in the README file.

Signed-off-by: Bartosz Golaszewski <bartekgola@gmail.com>

Instead of replacing the format specifiers in a separate buffer, just
make use of the output stream buffering.

Also: update the gpiomon tests with some corner cases.

Signed-off-by: Bartosz Golaszewski <bartekgola@gmail.com>

It's more efficient to just allocate it at testing suite's startup and
reuse it for all child processes.

While we're at it: verify that the output of subprocesses consists
of ASCII characters only.

Signed-off-by: Bartosz Golaszewski <bartekgola@gmail.com>

We previously used a character that is not interpreted by
printf()-like functions due to a bug in the testing framework.

This is now fixed, so just use %x so that we self-test the test suite
at the same time.

Signed-off-by: Bartosz Golaszewski <bartekgola@gmail.com>

We're currently using vasprintf() internally when reading the output
streams of terminated subprocesses. This routine interprets the buffer
as a printf() format string, so substrings like '%x' get converted
unintentionally.

Instead allocate a big enough buffer and just read the output without
any modifications.

Signed-off-by: Bartosz Golaszewski <bartekgola@gmail.com>

The '%%e' string should be printed as '%e' - the same as for printf()
format specifiers - but gpiomon prints '%%e' in this case. Fix it.

Signed-off-by: Bartosz Golaszewski <bartekgola@gmail.com>

When building tests: require libkmod and libudev versions to be at
least 18 & 215 respectively. These versions are shipped with debian
jessie and are known to work.

Signed-off-by: Bartosz Golaszewski <bartekgola@gmail.com>

Check if regexec() is available if we need to build the test suite.

Signed-off-by: Bartosz Golaszewski <bartekgola@gmail.com>

Reading output from child processes should be fixed at some point to
ignore printf() format specifiers.

Signed-off-by: Bartosz Golaszewski <bartekgola@gmail.com>