On Thu, 18 Sep 2008, Lindy Mayfield wrote: > I understand that the first step in porting a new operating system to > MVS is to train the compiler to output MVS instructions. But how would > one compile the changes to the compiler?
Ah, MVS is an OS, not an architecture. Just to be a pendant, I think you are talking about porting from one architecture to another. Most likely you will need the original system. You take the source code to the compiler and change the "code generator" to generate the instructions for the new architecture. E.g. output zArch instructions instead of VAX or Alpha instruction. This creates what is called a "cross compiler". That is, a compiler which runs on one system, but creates an executable that runs on a different system. Now, you must look at all the architecture dependant code, likely coded in assembler, and rewrite it to be functionally compatable on the new architecture. Or perhaps even rewrite parts of the code entirely if the architectures are incompatable. As an example, most systems today are 32 or 64 bit. But there were many 36 bit systems in the past. There are likely a lot of assumptions in the low level code about how many bits are in a "word" or a register. That can make a huge difference in the port. > > > > Say for example I wanted to port OpenVMS to VM. > > > > Thanks > > Lindy > > > > > ---------------------------------------------------------------------- > For IBM-MAIN subscribe / signoff / archive access instructions, > send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO > Search the archives at http://bama.ua.edu/archives/ibm-main.html > > -- Q: What do theoretical physicists drink beer from? A: Ein Stein. Maranatha! John McKown ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
