Fix autosave soft limits

[barnettwilliam]

From ppmac testing at I15, autosave only partially restores soft limits restoring only DLLM and DHLM and not LLM and HLM. This means that values from the ioc config are used for LLM and HLM while DLLM and DHLM come from the last autosaved values.

In addition the autosaved softlimit values are also restored after ioc record initialisation this means that they are not sent to the brick so it can be the case that the limits indicated by the ioc pv and synoptic do not match the limits the brick is using so an axis may stop in an unexpected place compared to the values indicated by the soft limit PVs.

This behaviour is particularly confusing when offsets are also used when there is a difference between D*LM and *LM.

This fix adds LLM and HLM to autosave so that all soft limits values are restored from autosave values and autosaved soft limit values are restored before ioc record initialisation (#% autosave 0) so they are sent to the brick on ioc startup.

[barnettwilliam]

The missing newlines from req files issue epics-containers/ibek-support#191 was a potential cause of the isssue. Testing an ibek ioc with fixed req files this soft limit sync issue still occurs so the missing new lines is not the cause.

Testing a geobrick BL15I STEP-09 with its builder based ioc , the soft limits are fully restored including synced to the brick with the original autosave config DHLM and DLLM under autosave 1.

Another version of the fix was tested that only archives DHLM and DLLM under autosave 0 and the soft limit synced correctly so LLM and HLM are not required.

[barnettwilliam]

Containerising STEP-09 the soft limit issue is present with the geobrick so it is due to the newer version of epics or other ibek difference