Fix function_prologue example

[fmayer]

We don't need to (x)or the tag into the sp. That makes returning from
the function harder, and serves no benefit because we add the tag
separately into each of the objects anyway.

[fmayer]

@deepak0414

[cmuellner]

It would be great if the example also showed that the tag needs to be set (with SETTAG).
Otherwise, this example produces tagged pointers without tagged memory.

I propose to move the example into an appendix (like many other specifications have done) and make it more complete (what is needed in the epilogue if the prologue tags pointers and memory?)

Further, using OR to set a pointer tag only works, if there is a guarantee that the pointer, that is being tagged, is not tagged yet. In the example a comment like "# We assume that SP holds a non-tagged pointer" might be sufficient.

[fmayer]

It would be great if the example also showed that the tag needs to be set (with SETTAG). Otherwise, this example produces tagged pointers without tagged memory.

I propose to move the example into an appendix (like many other specifications have done) and make it more complete (what is needed in the epilogue if the prologue tags pointers and memory?)

Further, using OR to set a pointer tag only works, if there is a guarantee that the pointer, that is being tagged, is not tagged yet. In the example a comment like "# We assume that SP holds a non-tagged pointer" might be sufficient.

Thanks for the comment, done.