Re: OpenMPE redux, executive view. - Hewlett Packard

This is a discussion on Re: OpenMPE redux, executive view. - Hewlett Packard ; On Thu, 30 Oct 2008 23:12:45 -0500, Paul Raulerson wrote: >On Oct 29, 2008, at 2:49 AM, James Hofmeister wrote: > >> Hello All, >> >> It would be very cool to see MPE running as a Xen virtual machine ...

+ Reply to Thread
Results 1 to 3 of 3

Thread: Re: OpenMPE redux, executive view.

  1. Re: OpenMPE redux, executive view.

    On Thu, 30 Oct 2008 23:12:45 -0500, Paul Raulerson
    wrote:

    >On Oct 29, 2008, at 2:49 AM, James Hofmeister wrote:
    >
    >> Hello All,
    >>
    >> It would be very cool to see MPE running as a Xen virtual machine on
    >> Linux.
    >> How to do this is something that donna and I have discussed over a
    >> glass of
    >> California wine more than once...
    >>
    >> My thoughts and discourses...
    >>
    >> - Emulation of the RISC instruction set is vitally important to avoid
    >> reworking a lot of Modcal (enhanced MPE Pascal) and SPL code.

    >
    >Well, this is already done on HP-UX, the Arial emulator will
    >seamlessly run HP-UX PA-RISC binaries.
    >That means processor emulation is indeed possible. SIMH would make a
    >really good starting point.
    >
    >Q. Is there a Principles of Operation for PA-RISC available somewhere?
    >


    It looks like some work may have already started to add PA-RISC support to
    QEMU: http://hppaqemu.sourceforge.net/

    Here are some sites that may have more information about the underlying
    architecture:

    http://www.openpa.net/
    http://www.parisc-linux.org/
    http://www.openbsd.org/hppa.html
    http://www.openbsd.org/hppa64.html
    http://www.netbsd.org/ports/hp700/

    --
    Edward Berner

    * To join/leave the list, search archives, change list settings, *
    * etc., please visit http://raven.utc.edu/archives/hp3000-l.html *


  2. Re: OpenMPE redux, executive view.

    Hardware Emulation

    As has been pointed out, QEMU has had a PA-RISC project. I don't know the
    details, but it seems to me it was harder than expected due to things like
    PA-RISC's ability to preload branches ahead of execution, and significant
    hardware variations in the CPU design. HP certainly has the talent and
    resources to add PA-RISC support to QEMU. So why not? Are the purported
    millions of dollars available, not available? Does it not fit in with HP's
    business plan? QEMU is already a highly reguarded hardware emulator for a
    variety of architectures, great and small.. Adding another architecture to
    it would greatly minimize the cost of developing one, and unlike a
    commercial emulator, there is no chance that you are going to be held
    hostage or have it abandoned due to some business plan.


    Code and Data Pages

    This is not an OS feature, it is a hardware feature! It is a feature of the
    HP3000 and other business servers and mainframes. The Intel CPU's and their
    clones did not support it for a long time, because their CPUs were designed
    for PERSONAL COMPUTERS. Linux has supported code and data page separation
    for years, though it did intially have some pieces that performed some
    "magic" by executing a data page. A bit of knowledge and/or understanding
    would be helpful here to have an intelligent discussion about this.


    Files

    Linux files are decent containers. How they are accessed is a function of
    the intrinsics. Much of the controls that MPE exerts is due to the system
    file label which I proposed would be just another Linux file with special
    handling by the MPE layer. This can be enforced using extended attributes
    in Linux for all of the special files in MPE. I think it is pretty secure,
    but not absolutely certain. Here is the description of how EAs should be
    used in Linux: http://www.freedesktop.org/wiki/Comm...ndedAttributes If
    for some reason the security model for EAs in Linux is not sufficient to
    keep hackers in check, I can assure you that SELinux is, but is a lot more
    bother and complexity.


    Where to Start?

    How about create sample Linux user and home directory on a Linux with a
    sample Linux directory structure to support MPE?

    Then how about a command interpreter (shell) for the MPE Linux user to run
    accessing the its Linux home directory, presenting an MPE view?

    --------------------

    James H. seems to have a grasp of the subject, though I may disagree with
    him on some points. Wirt is apparently keeping low and out of the line of
    fire (the smartest "Windows Weenie" ;-) I know by far). Alfredo apparently
    lost interest years ago. HP3000 lab people are probably under strict orders
    to not get involved. Is there anyone left capable, knowledgeable, and
    interested? I guess I may be getting tired of the boisterous arm waving,
    and looking for some intellectual meat. I do appreciate the honest
    questions and concerns, no matter at what technical level.

    Pete

    * To join/leave the list, search archives, change list settings, *
    * etc., please visit http://raven.utc.edu/archives/hp3000-l.html *


  3. Re: OpenMPE redux, executive view.

    Pete writes:

    [snip]

    > Code and Data Pages
    >
    >This is not an OS feature, it is a hardware feature! It is a
    >feature of the HP3000 and other business servers and
    >mainframes. The Intel CPU's and their clones did not support
    >it for a long time, because their CPUs were designed for
    >PERSONAL COMPUTERS. Linux has supported code and data page
    >separation for years, though it did intially have some pieces
    >that performed some "magic" by executing a data page. A bit
    >of knowledge and/or understanding would be helpful here to
    >have an intelligent discussion about this.


    I would argue this is 'both' as most implementations have the ability
    to limit execution of pages as a kernel enabled feature.

    Yes, first the h/w must support it in the native architecture, but the o/s must
    usually support the feature.

    Any intel/amd server cpu you can get these days has such support and it is
    the o/s that must work in concert with the feature to support it.

    duane

    * To join/leave the list, search archives, change list settings, *
    * etc., please visit http://raven.utc.edu/archives/hp3000-l.html *


+ Reply to Thread