GitLab

We may not get back to the test sitter upon signal (SIGINT/SIGALRM),
so move the code parsing the last data bulk behind the sigwait() call
instead.
Signed-off-by: Philippe Gerum <rpm@xenomai.org>

Signed-off-by: Philippe Gerum <rpm@xenomai.org>

Using -C bypasses the CPU checks, allowing to set the CPU affinity to
a non-isolated core.
Signed-off-by: Philippe Gerum <rpm@xenomai.org>

This mode measures the delay between the moment a synthetic interrupt
is posted from the oob stage and when it is eventually received by its
in-band handler. When measured under significant pressure, this gives
the typical interrupt latency experienced by the in-band kernel due to
local interrupt disabling.

Therefore, this is an in-band only test which measures IRQ latency
experienced by the common kernel infrastructure, _NOT_ by EVL.

Measurement is requested with '-s' option.
Signed-off-by: Philippe Gerum <rpm@xenomai.org>

An ugly work-around to pass the test on armv6l (e.g. Raspberry PI Zero
targeting the softfp floating-point ABI).
Signed-off-by: Philippe Gerum <rpm@xenomai.org>

Using [SCHED_FIFO, 98] for running the responder thread will work for
both EVL and native preemption configurations (i.e. GPIO mode). On the
latter, it won't compete with high-priority housekeeping threads
running in kernel space (prio 99) which is recommended and saner. On
both, this leaves lower priority levels for common application-level
threads.
Signed-off-by: Philippe Gerum <rpm@xenomai.org>

Signed-off-by: Philippe Gerum <rpm@xenomai.org>

We have been dealing properly with cells having no sample logged for
many moons. There is no reason to increment all cell counts anymore.
This makes the overall sample count consistent with the sum of all
cell counts.
Signed-off-by: Philippe Gerum <rpm@xenomai.org>

Signed-off-by: Philippe Gerum <rpm@xenomai.org>

Signed-off-by: Philippe Gerum <rpm@xenomai.org>

Signed-off-by: Philippe Gerum <rpm@xenomai.org>

Signed-off-by: Philippe Gerum <rpm@xenomai.org>

This is a general fix addressing several issues in the retrieval of
histogram data from latmon, which reverts #6b2425d8

 in the same move:

- use a safe socket send loop in order to cope with partial writes on
  the latmon side.

- drop the very notion of warmup time, the system must be ready to
  respond with no delay or preparation, and the average figures won't
  be affected over long enough runs anyway.
Signed-off-by: Philippe Gerum <rpm@xenomai.org>

Losing the whole result log upon network error communicating with
latmon is very annoying, especially when this happens at the very end
of an overnight test. Since we have a TCP connection, we can trust the
results received so far, so dump them to the result file regardless.
Signed-off-by: Philippe Gerum <rpm@xenomai.org>

Signed-off-by: Philippe Gerum <rpm@xenomai.org>

Signed-off-by: Philippe Gerum <rpm@xenomai.org>

Signed-off-by: Philippe Gerum <rpm@xenomai.org>

Signed-off-by: Philippe Gerum <rpm@xenomai.org>

Since ABI 24, we can read from the target file connected to a
proxy. The EVL_CLONE_INPUT, EVL_CLONE_OUTPUT creation flags select the
direction (may be combined for a single proxy instance).

The evl_new_proxy() short call form sticks with the original behavior,
creating an output-only proxy. There is actually no change needed to
libevl in order to enable this feature. Just add a test for peace of
mind.
Signed-off-by: Philippe Gerum <rpm@xenomai.org>

Signed-off-by: Philippe Gerum <rpm@xenomai.org>

Do build them unconditionally though.
Signed-off-by: Philippe Gerum <rpm@xenomai.org>

Signed-off-by: Philippe Gerum <rpm@xenomai.org>

Signed-off-by: Philippe Gerum <rpm@xenomai.org>

Signed-off-by: Philippe Gerum <rpm@xenomai.org>

Signed-off-by: Philippe Gerum <rpm@xenomai.org>

Since ABI 23, the core is able to channel T_WOSS, T_WOLI and T_WOSX
error notifications through the offender's observable component if
present.

Convert all SIGDEBUG_xxx cause codes to the new EVL_HMDIAG_xxx naming,
so that we have a single nomenclature for these errors regardless of
whether threads are notified via SIGDEBUG or their observable
component.

The API rev. is bumped to #17 as a result of these changes.
Signed-off-by: Philippe Gerum <rpm@xenomai.org>

Since ABI 23, the core provides the new observable element, which
enables the observer design pattern. Any EVL thread is in and of
itself an observable which can be monitored for events too.

As a by-product, the poll interface can now be given a user-defined
opaque data when subscribing file descriptors to poll elements, which
the core passes back on return to evl_poll().
Signed-off-by: Philippe Gerum <rpm@xenomai.org>

Signed-off-by: Philippe Gerum <rpm@xenomai.org>

Signed-off-by: Philippe Gerum <rpm@xenomai.org>

Signed-off-by: Philippe Gerum <rpm@xenomai.org>

Open-coding oob_ioctl() calls to set/clear mode bits in apps is
unhandy and fairly ugly. Let's provide sanctioned services for these
requests, namely evl_set_thread_mode() and evl_clear_thread_mode()
respectively.
Signed-off-by: Philippe Gerum <rpm@xenomai.org>

Signed-off-by: Philippe Gerum <rpm@xenomai.org>

Signed-off-by: Philippe Gerum <rpm@xenomai.org>

Signed-off-by: Philippe Gerum <rpm@xenomai.org>

Useful to pick the most appropriate CPU available for running a test,
accounting for the initial hint given by the caller.
Signed-off-by: Philippe Gerum <rpm@xenomai.org>

Starting with ABI 22, we can ask for the core to unblock a thread from
a wait state, which may include forcing it out of any real-time
scheduling class by demoting it to SCHED_WEAK in the same move.

Export evl_unblock_thread() and evl_demote_thread() as the
corresponding wrappers.
Signed-off-by: Philippe Gerum <rpm@xenomai.org>

Signed-off-by: Philippe Gerum <rpm@xenomai.org>

An equivalent of sched_yield().
Signed-off-by: Philippe Gerum <rpm@xenomai.org>

Signed-off-by: Philippe Gerum <rpm@xenomai.org>

Signed-off-by: Philippe Gerum <rpm@xenomai.org>