Improve/clarify log reporting/tracking (ES System Log and EVS Event Log)

[skliper]

Is your feature request related to a problem? Please describe. EVS event logging reports LogFullFlag, LogMode, and LogOverflowCounter.

• The log is an array of CFE_EVS_LongEventTlm_t of size CFE_PLATFORM_EVS_LOG_MAX so each entry has a slot
• In either discard or overwrite, once the log is filled LogFullFlag is set to true, and LogOverflowCounter is incremented for each additional message (so you can tell how many messages were dropped or added after being full).
• Note LogCount is maintained internally but is not in tlm (counts number of entries in the log, clamps at max)

ES syslog reports SysLogBytesUsed, SysLogSize, SysLogEntries, and SysLogMode.

• The log is a more freeform char array SystemLog of size CFE_PLATFORM_ES_SYSTEM_LOG_SIZE
• In either mode the log is "full" when CFE_PLATFORM_ES_SYSTEM_LOG_SIZE - SysLogBytesUsed is less than CFE_TIME_PRINTED_STRING_SIZE
• SysLogBytesUsed is set to the internal SystemLogEndIdx which is the highest used element of the char array
• SysLogEntries increments whenever a message is written, so if in overwrite it's not actually the number of entries in the log but the number of entries that have ever been written to the log
• There are return codes (CFE_ES_ERR_SYS_LOG_TRUNCATED and CFE_ES_ERR_SYS_LOG_FULL), but they are only returned in discard mode, and not available in tlm
• SysLogSize is the full size of the log, CFE_PLATFORM_ES_SYSTEM_LOG_SIZE

Describe the solution you'd like Numerous misnomers and inconsistencies in log management should be fixed/clarified, within the context of the actual differences in these logs (char array vs event array).

Describe alternatives you've considered None

Additional context Code review

Requester Info Jacob Hageman - NASA/GSFC

ping @dmknutsen @olamoo13

[skliper]

@dmknutsen @olamoo13 - looks like there's a couple options to knowing when the system log is full based on the current implementation In either case:

• check for SysLogBytesUsed > CFE_PLATFORM_ES_SYSTEM_LOG_SIZE (or SysLogSize) - CFE_TIME_PRINTED_STRING_SIZE
• check that SysLogBytesUsed doesn't change between two sys log writes

In discard mode:

• you could create an event from the test app that gets sent if write to sys log returns CFE_ES_ERR_SYS_LOG_FULL
• If you could store values/do math, for a new log entry if SysLogEntries doesn't change it's full

EDIT - this had some incorrect descriptions, fixed

[skliper]

@dmknutsen @olamoo13 - if you wanted to get really fancy, the test app could subscribe to the ES housekeeping and do the math for you. Basically add a command to "fill the log", and it could continue creating syslog messages until the hk packet values show it's full. That would move the logic into the app. I'm not a huge fan of this approach (test app should be as simple as possible), but it's a way to get it done for now.