4.11. System Monitoring Module
4.11.2. Detailed Description
The system monitoring module allows to detect timing variations in the execution of the system tasks on millisecond level. This is done by capturing a timestamp on entry and exit of the task functions. A continuously running monitoring task compares the actual runtime of the tasks against a configured maximum allowed runtime. In the case that the predefined thresholds are surpassed, a diagnosis entry is created and, if configured, the violation is recorded in persistent memory. The current state of the system monitoring is communicated through a CAN message.
4.11.2.1. Configuration
The monitoring actions for each task can be configured in the array described in Current system monitoring configuration.
1SYSM_MONITORING_CFG_s sysm_ch_cfg[] = {
2 {SYSM_TASK_ID_ENGINE,
3 SYSM_ENABLED,
4 FTSK_TASK_ENGINE_CYCLE_TIME,
5 FTSK_TASK_ENGINE_MAXIMUM_JITTER,
6 SYSM_RECORDING_ENABLED,
7 SYSM_HANDLING_SWITCH_OFF_CONTACTOR,
8 SYSM_DummyCallback},
9 {SYSM_TASK_ID_CYCLIC_1ms,
10 SYSM_ENABLED,
11 FTSK_TASK_CYCLIC_1MS_CYCLE_TIME,
12 FTSK_TASK_CYCLIC_1MS_MAXIMUM_JITTER,
13 SYSM_RECORDING_ENABLED,
14 SYSM_HANDLING_SWITCH_OFF_CONTACTOR,
15 SYSM_DummyCallback},
16 {SYSM_TASK_ID_CYCLIC_10ms,
17 SYSM_ENABLED,
18 FTSK_TASK_CYCLIC_10MS_CYCLE_TIME,
19 FTSK_TASK_CYCLIC_10MS_MAXIMUM_JITTER,
20 SYSM_RECORDING_ENABLED,
21 SYSM_HANDLING_SWITCH_OFF_CONTACTOR,
22 SYSM_DummyCallback},
23 {SYSM_TASK_ID_CYCLIC_100ms,
24 SYSM_ENABLED,
25 FTSK_TASK_CYCLIC_100MS_CYCLE_TIME,
26 FTSK_TASK_CYCLIC_100MS_MAXIMUM_JITTER,
27 SYSM_RECORDING_ENABLED,
28 SYSM_HANDLING_SWITCH_OFF_CONTACTOR,
29 SYSM_DummyCallback},
30 {SYSM_TASK_ID_CYCLIC_ALGORITHM_100ms,
31 SYSM_ENABLED,
32 FTSK_TASK_CYCLIC_ALGORITHM_100MS_CYCLE_TIME,
33 FTSK_TASK_CYCLIC_ALGORITHM_100MS_MAXIMUM_JITTER,
34 SYSM_RECORDING_ENABLED,
35 SYSM_HANDLING_SWITCH_OFF_CONTACTOR,
36 SYSM_DummyCallback},
37};
In this configuration, every task must have an entry. For each task, the system monitoring can be enabled or disabled. This is for example useful, when the algorithm task may violate the timings from time to time. Care should be taken, when a higher priority task is not monitored. A timing violation on a higher priority task likely means that it takes up all remaining resources leading to a block of the remaining lower priority tasks.
Apart from settings such as the expected cycle time and timing variation, it is possible to configure whether a flag in persistent memory shall be set for each task. If this is disabled, the flag still exists, but will always return false (meaning no violation).
4.11.2.2. Diagnosis entries
When monitoring is enabled for a task and when the task violates its timings,
a diagnosis entry of type DIAG_ID_SYSTEM_MONITORING is created in the
Diagnosis Module.
The diagnosis handler for this ID sets the appropriate error flags in the
DATA_BLOCK_ERROR_STATE_s database table.
These flags will not be cleared automatically.
The only actions that clear these flags are a power on reset cycle or sending
the appropriate flag in the foxBMS_Command CAN message.
4.11.2.3. Error recording
If enabled, the system monitoring sets flags in persistent memory.
This is handled through the FRAM module.
Flags are written to the FRAM_SYS_MON_RECORD_s entry and committed to
persistent memory through a handler that is called from the 10 millisecond
task.
These flags will not be cleared automatically.
The only action that clears these flags is sending the appropriate flag in the
foxBMS_Command CAN message.
4.11.2.4. CAN interface
The system monitoring module communicates through the CAN interface.
Detected violations of timing expectations are communicated through two
messages.
A general flag indicating if any violation has occurred is sent as the
variable foxBMS_SysMonError in message foxBMS_State.
This flag is a sum value of all recorded and current entries in the database
and the persistent memory.
A detail information on the current state is transmitted in the
foxBMS_DetailState message.
This message sets flags for recorded and current violations of each task.
In order to reset all current and recorded flags, the foxBMS_resetFlags
must be set in the foxBMS_Command message.
This calls the function SYSM_ClearAllTimingViolations() which resets
both the current and recorded flags and commits the update directly to the
persistent memory.
