[RFC V3 20/20] Documentation/rv: Add watchdog-monitor documentation
From: Daniel Bristot de Oliveira
Date: Thu May 05 2022 - 12:09:47 EST
Adds documentation about the safe_wtd and safe_wtd_nwo RV monitors,
and their usage via a safety application.
Cc: Wim Van Sebroeck <wim@xxxxxxxxxxxxxxxxxx>
Cc: Guenter Roeck <linux@xxxxxxxxxxxx>
Cc: Jonathan Corbet <corbet@xxxxxxx>
Cc: Steven Rostedt <rostedt@xxxxxxxxxxx>
Cc: Ingo Molnar <mingo@xxxxxxxxxx>
Cc: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
Cc: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
Cc: Will Deacon <will@xxxxxxxxxx>
Cc: Catalin Marinas <catalin.marinas@xxxxxxx>
Cc: Marco Elver <elver@xxxxxxxxxx>
Cc: Dmitry Vyukov <dvyukov@xxxxxxxxxx>
Cc: "Paul E. McKenney" <paulmck@xxxxxxxxxx>
Cc: Shuah Khan <skhan@xxxxxxxxxxxxxxxxxxx>
Cc: Gabriele Paoloni <gpaoloni@xxxxxxxxxx>
Cc: Juri Lelli <juri.lelli@xxxxxxxxxx>
Cc: Clark Williams <williams@xxxxxxxxxx>
Cc: linux-doc@xxxxxxxxxxxxxxx
Cc: linux-kernel@xxxxxxxxxxxxxxx
Cc: linux-trace-devel@xxxxxxxxxxxxxxx
Signed-off-by: Daniel Bristot de Oliveira <bristot@xxxxxxxxxx>
---
Documentation/trace/rv/watchdog-monitor.rst | 250 ++++++++++++++++++++
1 file changed, 250 insertions(+)
create mode 100644 Documentation/trace/rv/watchdog-monitor.rst
diff --git a/Documentation/trace/rv/watchdog-monitor.rst b/Documentation/trace/rv/watchdog-monitor.rst
new file mode 100644
index 000000000000..6d98fc3af3f4
--- /dev/null
+++ b/Documentation/trace/rv/watchdog-monitor.rst
@@ -0,0 +1,250 @@
+Watchdog monitor
+----------------
+
+The watchdog is an essential building block for the usage of Linux in
+safety-critical systems because it allows the system to be monitored from
+an external element - the watchdog hardware, acting as a safety-monitor.
+
+A user-space application controls the watchdog device via the watchdog
+interface. This application, hereafter safety_app, enables the watchdog
+and periodically pets the watchdog upon correct completion of the safety
+related processing.
+
+If the safety_app, for any reason, stops pinging the watchdog,
+the watchdog hardware can set the system in a fail-safe state. For
+example, shutting the system down.
+
+Given the importance of the safety_app / watchdog hardware couple,
+the interaction between these software pieces also needs some
+sort of monitoring. In other words, "who monitors the monitor?"
+
+The safe watchdog (safe_wtd) RV monitor monitors the interaction between
+the safety_app and the watchdog device, enforcing the correct sequence of
+events that leads the system to a safe state.
+
+Furthermore, the safety_app can monitor the RV monitor by collecting the
+events generated by the RV monitor itself via tracing interface. In this way,
+closing the monitoring loop with the safety_app.
+
+A diagram of the components and their interactions is::
+
+ user-space:
+ +--------------------------------+
+ | safety_app |-----------+
+ +--------------------------------+ |
+ | ^ |
+ | Configure | Enable and |
+ | | check data |
+ ===================+====================+=============== |
+ kernel-space: | | |
+ v v |
+ +----------+ instr. +-------------+ |
+ | watchdog | ----------->| RV Monitor |----+ |
+ | device | +-------------+ | |
+ +----------+ | |
+ | | |
+ | | |
+ ================+====================================== | |
+ hardware: | | |
+ v | +-> Bring the system
+ +--------------------+ +----> to a safe state,
+ | watchdog hardware |---------------------------> e.g., halt.
+ +--------------------+
+
+Sample safety_app
+-----------------
+
+The user-space safety_app sample code in ``tools/tracing/rv/safety_app/``
+serves to illustrate the usage of the RV monitors for this use-case, as
+well as the starting point to the development of a user-specific safety_app.
+
+Watchdog events
+---------------
+
+The RV monitor observes the watchdog by using instrumentation to
+process the events generated by the interaction between the
+safety_app and the watchdog device layer in kernel.
+
+The monitored events are:
+
+ - watchdog:watchdog_open: open the watchdog device;
+ - watchdog:watchdog_close: close the watchdog device;
+ - watchdog:watchdog_start: start the watchdog;
+ - watchdog:watchdog_stop: stop the watchdog;
+ - watchdog:watchdog_set_timeout: set the watchdog timeout;
+ - watchdog:watchdog_ping: reprogram the watchdog with the previously set
+ timeout;
+ - watchdog:watchdog_nowayout: prevents the watchdog from stopping;
+ - watchdog:watchdog_set_keep_alive: set an intermediary ping to overcome
+ the limitation of a hardware watchdog maximum timeout being shorter than
+ the timeout set by the user-space tool;
+ - watchdog:watchdog_keep_alive: the execution of the function that runs the
+ intermediary keep alive ping;
+
+RV monitor events
+-----------------
+
+The RV monitor monitors the relevant events as an outside observer,
+interpreting all the components (the hardware; the watchdog device
+interface; and the safety monitor) as an integrated component.
+
+The events selected for the monitor are:
+
+ - other_threads: an event generated by any thread other than the
+ one that set nowayout or open the watchdog the last time.
+ - open: a thread opens the watchdog to manipulate it;
+ - close: a thread closes the watchdog;
+ - start: starts the watchdog countdown;
+ - stop: stops the watchdog;
+ - set_safe_timeout: configures the watchdog with a given timeout;
+ - ping: resets the watchdog countdown with the previously configured timeout;
+ - nowayout: prevents the watchdog to be stopped until the system's shutdown;
+ - sched_keep_alive: schedules a kernel worker to ping the watchdog if the
+ timeout is longer than the watchdog hardware can handle.
+ - keep_alive: executes the previously scheduled watchdog ping;
+
+Noting that the events that does not appear in the automata models are
+considered blocked events, and their execution will always cause the
+RV monitor to react to an unexpected event.
+
+RV monitor specification
+------------------------
+
+The monitor's goal is to assess a set of specifications that conducts the
+system to a safe state.
+
+These specifications are:
+
+ - 1: Once open, only one process manipulates the watchdog;
+ - 2: Following 1, the keep-alive mechanisms will not be used;
+ - 3: If required, nowayout will be set before opening the watchdog;
+ - 4: A safe timeout must be set;
+ - 5: At least one ping must be made before entering the safe/safe_nwo states
+ - 6: The RV monitor does not react if the watchdog is closed without stopping.
+ But the hardware watchdog is expected to react.
+
+Deterministic automata monitors
+-------------------------------
+
+Following the specifications, a deterministic automata monitor
+was developed. The monitor is modeled as Deterministic Automata model.
+
+The deterministic automata model for safe_wtd is::
+
+ #==================================# other_threads
+ H H ----------------+
+ -----------> H init H |
+ H H <---------------+
+ #==================================#
+ | | ^
+ | | | close
+ | | +----------------------------------------------------+
+ | | |
+ | | open |
+ | +------------------------------------------------------+ |
+ | | |
+ | nowayout | |
+ v | |
+ nowayout +-------------------+ | |
+ other_threads | | nowayout | |
+ +---------------- | nwo |<-------------------------------------+ | |
+ | | | | | |
+ +---------------> | | <+ | | |
+ +-------------------+ | | | |
+ | | | | |
+ | open | close | | |
+ v | | | |
+ +-------------------+ | | | |
+ | opened_nwo | -+ | | |
+ +-------------------+ | | |
+ | | | |
+ | start | | |
+ v | | |
+ +-------------------+ | | |
+ +---------------> | started_nwo | -+ | | |
+ | +-------------------+ | | | |
+ | | | | | |
+ | open | set_safe_timeout | | | |
+ | v | | | |
+ | +-------------------+ | | | |
+ | | set_nwo | | | | |
+ | +-------------------+ | | | |
+ | | | | | |
+ | +-------------+ | ping | | | |
+ | | | | | | | |
+ | | ping v v | | | |
+ | | +-------------------+ | | | |
+ | +-----------| safe_nwo | | | | |
+ | +-------------------+ | | | |
+ | | | | | |
+ | | close | close | | |
+ | v v | | |
+ | +----------------------------------+ nowayout | | |
+ | | | other_threads | | |
+ | | closed_running_nwo | ----------------+ | | |
+ | | | | | | |
+ +---------------- | | <---------------+ | | |
+ +----------------------------------+ | | |
+ | nowayout ^ | | |
+ +-----------------------------+ | | |
+ | | |
+ | | |
+ +-------------------+ +--------+ | | |
+ | | | |------+---+ |
+ | started | start | opened | | |
+ +---------------- | | <-------- | |>-----+-------+
+ | +-------------------+ +--------+ | ^
+ | | | |
+ | | set_safe_timeout +-------------+-------+
+ | v | |
+ | +-------------------+ | |
+ | | | | |
+ | | set | | |
+ +----------+---------------> | | | |
+ | | +-------------------+ | |
+ | | | | |
+ | | | ping | |
+ | | v | |
+ | | +-------------------+ ping | |
+ | | | | -------+ | |
+ | | +---- | safe | | | |
+ | | | | | <------+ | |
+ | | | +-------------------+ | |
+ | | | | | |
+ | | stop | | stop | |
+ | | | v | |
+ | | | +-------------------+ close | |
+ | +-----------+---> | stopped |-------------+ |
+ | | +-------------------+ |
+ | +---+ |
+ | | close |
+ | v |
+ | other_threads +----------------------------------------+ |
+ | +--------------> | | |
+ | | | closed_running | |
+ | +--------------- | |--------------+
+ | +----------------------------------------+
+ | | ^
+ | open | | close
+ | v |
+ | set_safe_timeout +-------------------+
+ +-------------------------> | reopened |
+ +-------------------+
+
+It is important to note that the events sched_keep_alive and keep_alive
+are not allowed in the monitor (they are said to be blocked events).
+The execution of any blocked events leads the RV monitor to react.
+
+Additional options
+------------------
+
+The RV monitor also has a set of options enabled via kernel command
+line/module options. They are:
+
+ - watchdog_id: the device id to monitor (default 0);
+ - dont_stop: once enabled, do not allow the RV monitor to be stopped (default off);
+ - safe_timeout: define a maximum safe value that a user-space application can
+ set as the watchdog timeout (default unlimited);
+ - check_timeout: After every ping, check if the time left in the watchdog is less
+ than or equal to the last timeout set for the watchdog. It only works for watchdog
+ devices that provide the get_timeleft() function (default off);
--
2.35.1