On 5 May 2015 at 15:42, Farley, Peter x23353 <[email protected]> wrote: > I suspect that "CPU-determined" number of bytes probably is not documented. > If IBM still published "Functional Specifications" manuals for each CPU model > it would probably be there, but they don't publish those any more.
It's documented (and presumably by definition architected) for some instructions. SRST and a small number of others will process at least 256 bytes, some promise at least 1 byte, and some others don't even guarantee more than 0. This might well be good material for one of the instruction summaries. For instance I find use of SRST without the branch back to be handy when I know that my string is well-formed and shorter than 256 bytes, but without a quick reference I have to read the prose accompanying other instructions to be sure, even when I otherwise know the instruction well enough. Tony H.
