optimize memory initialization for payload

[GitMensch]

This applies to the "heavy" chunkdata, but also to others.

Instead of setting multiple parts of the payload to ALL x"FF" with multiple statements, do it "once" for the compete payload fields (if not dynamically), then only store the payload that is not of this value.

As the payload-size likely differs it is likely only necessary to set the whole buffer to ALL x"FF" (possibly directly with VALUE) and when the code is re-executed only set the previous payload size back to x"FF" (not the maximum buffer).