Re: Re: Wait State 00050064
| Newsgroups | gmane.comp.emulators.hercules390.mvs |
|---|---|
| Message-ID | <[email protected]> |
In my mind the Architecture and Model are independent. The Model usually indicated number of CPUs and Memory installed, features, etc. which are, I think, outside the scope of the purpose of Hercules. I would pretty much allow any value in there and not react to it all. Why should Hercules care? I can't image that many SCP's even care about it beyond putting it into some messages or applications like some SMF records. On the other hand, the architecture is very important since it defines which instructions are valid and how they behave. For instance, the ESA/390 systems had vector instructions, but they were totally different from Z's. Or SIO vs SSCH. My vote on STDIP is that for ARCHLVL S/370 it should return a MECL of 0 in 6/7. As for all the other architectures I'd leave what we already have alone. Its pretty clear to me that an ESA/390 system running an LPAR in S/370 mode should return an ESA/390 STIDP (with F bit) despite the fact that the subsystem is in S/370 mode. Note that I already showed how MVS 3.8 is not supported in that config and this is one of the reasons why. bob On 9/23/2019 5:03 PM, ''Fish' (David B. Trout)' [email protected] [H390-MVS] wrote: > > (partial piggyback; killing two birds with one stone) > > Dave Wade wrote: > > Harold Grovesteen wrote: > > [...] > > From what I recall it (Hercules) has always tried to > > separate CPU model and architecture. i.e. changing > > the CPU model does not affect the operation of Hercules. > > Perhaps we need to change that. > > I believe we (Hercules developers) may have (BRIEFLY) discussed this > subject many years ago and decided trying to do that would only end up > being a nightmare. There are simply too many CPU models to deal with, > some of which have features which we (Hercules emulator) do not > support (and may not be able to support). > > We decided that attempting to do this would only end up being, from a > software maintenance point of view, a complete nightmare. > > > > I certainly encourage the GUI to factor that into > > > what it generates. > > And it will. > > (Hercules too! https://github.com/SDL-Hercules-390/hyperion/issues/256) > > -- > "Fish" (David B. Trout) > Software Development Laboratories > http://www.softdevlabs.com > mail: [email protected] > >