>>> On 1/19/2015 at 04:10 PM, "Shmuel Metz (Seymour J.)" <[email protected]> wrote: > In <[email protected]>, on 01/19/2015 > at 09:22 AM, Mark Post <[email protected]> said: > >>Of course not, but the things you were talking about are not part of >>the hypervisor features of CP, but the OS itself. > > Untrue; CP includes code for error handling. and the port of, e.g, > Solaris, relied on it.
Yes it is true. Of course CP includes code for error handling. It's an operating system after all. You have to understand, "CP" is the nucleus of z/VM. There are pieces of it that do "normal" operating system things, and there are parts that do hypervisor functions. The team doing the Solaris port made a conscious decision to take advantage of z/VM facilities for doing I/O and the like, tightly coupling it to that particular OS. The people porting Linux had a bigger goal in mind, but they also provided the _option_ of using z/VM facilities to do I/O (known as DIAG 250), using DCSSes (xip2fs), IUCV (terminal server), and a bunch more. None of them are required, and if you don't ask to use them, z/VM won't be doing your error recovery for you. Mark Post ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
