On 3 November 2014 15:17, Mark Boonie <[email protected]> wrote: > However, I thought it would be useful to limit the USING range so that > attempts to > address beyond the end of a DSECT would be flagged. After making the > changes suggested by John E. (thanks, John), I find that an L instruction > does get flagged as I suspected. However, neither L nor LY can be used as > a substitute, since they only deal with 32-bit registers. What I need in > this (contrived) example is a non-long-displacement 64-bit LOAD, which > doesn't exist.
Depending on what you want to do with that 64-bit value (and just how contrived an example you want) you could use LD. Of course it'll be in the "wrong" register, but you can then store it somewhere else with STD. How about LMD? That has 12-bit displacements (two of them!), and you can jig it to effectively load a 64-bit general register from 8 bytes in storage. Then there's CSST. Not the most efficient way to load a register, but it too has 12-bit displacements. Unlike LMD, it does set the condition code. I think I should get back to work... Tony H.
