DEFCON 16 and Hacking OpenVMS - VMS

This is a discussion on DEFCON 16 and Hacking OpenVMS - VMS ; Main, Kerry wrote: > No - all vendors (not just VMS) are only responding to what Cust's say > they want. Oh come on now. This is like supermarkets. Supermarkets don't carry what customers say they want. They carry what ...

+ Reply to Thread
Page 25 of 35 FirstFirst ... 15 23 24 25 26 27 ... LastLast
Results 481 to 500 of 691

Thread: DEFCON 16 and Hacking OpenVMS

  1. Re: Loose Cannon-dian (was: Re: DEFCON 16 and Hacking OpenVMS)

    Main, Kerry wrote:

    > No - all vendors (not just VMS) are only responding to what Cust's say
    > they want.



    Oh come on now. This is like supermarkets. Supermarkets don't carry what
    customers say they want. They carry what manufacturers tell them to
    carry (and pay them to carry).

    HP doesn't respond to customers, they identify potential additional
    profit sources and then make pretty speeches and powerpoints to try to
    set new trends that will get the clueless CIOs to say "we need to do
    that too".

    Carly was especially good at that, with lots of pretty speeches that
    trying to convince CIOs it was necessary to adopt her new philosophy to
    survice. (I use "philosophy" here because stuff like "Adaptive
    enterprise" were more a question of a marketing than tangible products.

  2. RE: Loose Cannon-dian (was: Re: DEFCON 16 and Hacking OpenVMS)

    > -----Original Message-----
    > From: JF Mezei [mailto:jfmezei.spamnot@vaxination.ca]
    > Sent: August 31, 2008 3:02 PM
    > To: Info-VAX@Mvb.Saic.Com
    > Subject: Re: Loose Cannon-dian (was: Re: DEFCON 16 and Hacking OpenVMS)
    >
    > Main, Kerry wrote:
    >
    > > No - all vendors (not just VMS) are only responding to what Cust's

    > say
    > > they want.

    >
    >
    > Oh come on now. This is like supermarkets. Supermarkets don't carry
    > what
    > customers say they want. They carry what manufacturers tell them to
    > carry (and pay them to carry).
    >
    > HP doesn't respond to customers, they identify potential additional
    > profit sources and then make pretty speeches and powerpoints to try to
    > set new trends that will get the clueless CIOs to say "we need to do
    > that too".
    >


    Geeeez, you seem to have a very low opinion of Cust IT capabilities.

    Cust: I want a blue car.
    Vendor: All we sell are red trucks.
    Cust: Ok. I'll take a red truck.

    Not my idea of today's typical Sales discussion. I have a somewhat higher
    view of Cust IT depts. skills than you do.

    While CIO's might discuss going in a certain direction, they typically
    will not boldly go in a new direction without support from some groups
    within their IT groups and the business.


    > Carly was especially good at that, with lots of pretty speeches that
    > trying to convince CIOs it was necessary to adopt her new philosophy to
    > survice. (I use "philosophy" here because stuff like "Adaptive
    > enterprise" were more a question of a marketing than tangible products.


    Oh come on, these phrases and terms have been going on since the first
    Marketing campaign of the first computer series.

    Remember these?
    - "The Network is the Computer" (Sun)

    - "Real Time Enterprise" (Gartner)

    - "Autonomic Computing" (project eLiza - IBM)
    http://www.ibm.com/developerworks/autonomic (see retirement notices)

    - "IT as a Utility" (don't know source, but that was internal IT buzz
    phrase even back in the Digital days)

    Btw, the concept of the "Adaptive Enterprise" was also The subject of a
    book from the Meta Group:
    http://www.intel.com/intelpress/sum_book2.htm
    http://www.intel.com/intelpress/toc-book2.pdf
    (one does not have to agree with everything in this book, but it does
    have good points)

    And to HP's credit, the AE was never position as a set of products that
    you buy off the shelf, but rather an approach that mixed services
    and products (and it did not have to be HP products, but of course
    Sales preferred HP products) to develop a 2-3 project roadmap to upgrade
    your companies IT to help it support the companies business's to be
    more competitive.


    Regards

    Kerry Main
    Senior Consultant
    HP Services Canada
    Voice: 613-254-8911
    Fax: 613-591-4477
    kerryDOTmainAThpDOTcom
    (remove the DOT's and AT)

    OpenVMS - the secure, multi-site OS that just works.





  3. Re: Loose Cannon-dian (was: Re: DEFCON 16 and Hacking OpenVMS)

    On Sep 1, 9:55*am, "Main, Kerry" wrote:
    > > -----Original Message-----
    > > From: JF Mezei [mailto:jfmezei.spam...@vaxination.ca]
    > > Sent: August 31, 2008 3:02 PM
    > > To: Info-...@Mvb.Saic.Com
    > > Subject: Re: Loose Cannon-dian (was: Re: DEFCON 16 and Hacking OpenVMS)

    >
    > > Main, Kerry wrote:

    >
    > > > No - all vendors (not just VMS) are only responding to what Cust's

    > > say
    > > > they want.

    >
    > > Oh come on now. This is like supermarkets. Supermarkets don't carry
    > > what
    > > customers say they want. They carry what manufacturers tell them to
    > > carry (and pay them to carry).

    >
    > > HP doesn't respond to customers, they identify potential additional
    > > profit sources and then make pretty speeches and powerpoints to try to
    > > set new trends that will get the clueless CIOs to say "we need to do
    > > that too".

    >
    > Geeeez, you seem to have a very low opinion of Cust IT capabilities.
    >
    > Cust: I want a blue car.
    > Vendor: All we sell are red trucks.
    > Cust: Ok. I'll take a red truck.
    >
    > Not my idea of today's typical Sales discussion. I have a somewhat higher
    > view of Cust IT depts. skills than you do.
    >
    > While CIO's might discuss going in a certain direction, they typically
    > will not boldly go in a new direction without support from some groups
    > within their IT groups and the business.
    >
    > > Carly was especially good at that, with lots of pretty speeches that
    > > trying to convince CIOs it was necessary to adopt her new philosophy to
    > > survice. (I use "philosophy" here because *stuff like "Adaptive
    > > enterprise" were more a question of a marketing than tangible products.

    >
    > Oh come on, these phrases and terms have been going on since the first
    > Marketing campaign of the first computer series.
    >
    > Remember these?
    > - "The Network is the Computer" (Sun)
    >
    > - "Real Time Enterprise" (Gartner)
    >
    > - "Autonomic Computing" (project eLiza - IBM)http://www.ibm.com/developerworks/autonomic(see retirement notices)
    >
    > - "IT as a Utility" (don't know source, but that was internal IT buzz
    > phrase even back in the Digital days)
    >
    > Btw, the concept of the "Adaptive Enterprise" was also The subject of a
    > book from the Meta Group:http://www.intel.com/intelpress/sum_.../toc-book2.pdf
    > (one does not have to agree with everything in this book, but it does
    > have good points)
    >
    > And to HP's credit, the AE was never position as a set of products that
    > you buy off the shelf, but rather an approach that mixed services
    > and products (and it did not have to be HP products, but of course
    > Sales preferred HP products) to develop a 2-3 project roadmap to upgrade
    > your companies IT to help it support the companies business's to be
    > more competitive.
    >
    > Regards
    >
    > Kerry Main
    > Senior Consultant
    > HP Services Canada
    > Voice: 613-254-8911
    > Fax: 613-591-4477
    > kerryDOTmainAThpDOTcom
    > (remove the DOT's and AT)
    >
    > OpenVMS - the secure, multi-site OS that just works.



    The most successful vendors listen to what the marketplace is saying:
    what people are buying and what people might want to buy based upon
    technical and social trends.

    Existing customers are either happy, not happy or a combination
    thereof. Service and support are most important to them. Vendor-
    customer lines of communication develop and solidify over time. If a
    need arises that is outside of the established lines of communication,
    or can't be met by the current primary vendor, the customer will find
    something that works and use it. Often, the primary vendor doesn't
    even know they've done this until the tide has changed and it's too
    late. Remember why PDP was named PDP?

    The most successful vendors attract new customers and have a marketing
    strategy to do so. Some vendors don't care to grow certain product
    lines because they've set their sights in a different direction.

    Reality sometimes sucks, but reality doesn't really care what anyone
    thinks about it.

  4. Re: Loose Cannon-dian (was: Re: DEFCON 16 and Hacking OpenVMS)

    Doug Phillips wrote:

    > The most successful vendors listen to what the marketplace is saying:
    > what people are buying and what people might want to buy based upon
    > technical and social trends.



    Unfortunatly, many vendors (especially HP under Carly) wanted to set
    trends by announcing newfangled philosophies. Their real goal was to get
    customers to buy consulting services from HP to help them adopt that new
    trend that everyone *had* to adopt to stay "modern".

    So it isn't just "listening to customers" anymore. Many vendors are
    telling customers what they need, and customers then adopt those
    strategies because they are conditioned to adopt whatever new trends
    appear on the horizon.

    The is not too different from your local supermarket. It doesn't cater
    to its retail customers. It caters to suppliers who decide what goes on
    what shelf and how much space it takes. If you like product X, but
    product Y buys the whole shelf space for its soaps, then you will not be
    able to buy product X. And if you tell the store manager, there is
    nothing he can do because those deals are signed by headquarters with
    proctor and gamble or whatever.

  5. Get Mr Griswald back his car!

    Hi Kerry,

    > Cust: I want a blue car.
    > Vendor: All we sell are red trucks.
    > Cust: Ok. I'll take a red truck.


    No, more like: -

    Cust: I want the Arctic-Blue Sports-Wagon
    VMS: What you really want is the Wagon Queen Family Truxter in Metallic-Pea.

    And if you don't take the WSIT-Family-Truxter, you're stuck with a
    flattened-out DECforms :-(

    Cheers Richard Maher

    "Main, Kerry" wrote in message
    news:9D02E14BC0A2AE43A5D16A4CD8EC5A593ED5EBD946@GV W1158EXB.americas.hpqcorp.net...
    > -----Original Message-----
    > From: JF Mezei [mailto:jfmezei.spamnot@vaxination.ca]
    > Sent: August 31, 2008 3:02 PM
    > To: Info-VAX@Mvb.Saic.Com
    > Subject: Re: Loose Cannon-dian (was: Re: DEFCON 16 and Hacking OpenVMS)
    >
    > Main, Kerry wrote:
    >
    > > No - all vendors (not just VMS) are only responding to what Cust's

    > say
    > > they want.

    >
    >
    > Oh come on now. This is like supermarkets. Supermarkets don't carry
    > what
    > customers say they want. They carry what manufacturers tell them to
    > carry (and pay them to carry).
    >
    > HP doesn't respond to customers, they identify potential additional
    > profit sources and then make pretty speeches and powerpoints to try to
    > set new trends that will get the clueless CIOs to say "we need to do
    > that too".
    >


    Geeeez, you seem to have a very low opinion of Cust IT capabilities.

    Cust: I want a blue car.
    Vendor: All we sell are red trucks.
    Cust: Ok. I'll take a red truck.

    Not my idea of today's typical Sales discussion. I have a somewhat higher
    view of Cust IT depts. skills than you do.

    While CIO's might discuss going in a certain direction, they typically
    will not boldly go in a new direction without support from some groups
    within their IT groups and the business.


    > Carly was especially good at that, with lots of pretty speeches that
    > trying to convince CIOs it was necessary to adopt her new philosophy to
    > survice. (I use "philosophy" here because stuff like "Adaptive
    > enterprise" were more a question of a marketing than tangible products.


    Oh come on, these phrases and terms have been going on since the first
    Marketing campaign of the first computer series.

    Remember these?
    - "The Network is the Computer" (Sun)

    - "Real Time Enterprise" (Gartner)

    - "Autonomic Computing" (project eLiza - IBM)
    http://www.ibm.com/developerworks/autonomic (see retirement notices)

    - "IT as a Utility" (don't know source, but that was internal IT buzz
    phrase even back in the Digital days)

    Btw, the concept of the "Adaptive Enterprise" was also The subject of a
    book from the Meta Group:
    http://www.intel.com/intelpress/sum_book2.htm
    http://www.intel.com/intelpress/toc-book2.pdf
    (one does not have to agree with everything in this book, but it does
    have good points)

    And to HP's credit, the AE was never position as a set of products that
    you buy off the shelf, but rather an approach that mixed services
    and products (and it did not have to be HP products, but of course
    Sales preferred HP products) to develop a 2-3 project roadmap to upgrade
    your companies IT to help it support the companies business's to be
    more competitive.


    Regards

    Kerry Main
    Senior Consultant
    HP Services Canada
    Voice: 613-254-8911
    Fax: 613-591-4477
    kerryDOTmainAThpDOTcom
    (remove the DOT's and AT)

    OpenVMS - the secure, multi-site OS that just works.






  6. Re: OH YES HE IS! (Re: More VMS pantomime :-( Re: DEFCON 16 andHacking OpenVMS)

    On Aug 31, 11:02*am, "Richard Maher"
    wrote:
    > Hi Bugs,
    >
    > [
    > If consultancy dollars was the reason to do this we could have picked
    > far better "targets" than VMS. I very much doubt people will be lining
    > up to pay us to look at VMS security, don't you?
    > ]
    >
    > You really have no idea about VMS, or what it does, do you?
    >


    Well like we been saying from the start, we don't know much about VMS
    apart from how to break it. That being said we know it is being used
    in important places and that is why we think it should be looked at.
    What I mean but maybe wasn't clear about, is that a lot (I'm not
    saying all) VMS users do have very unrealistic expectations on their
    OS security and therefore would be very reluctant to have someone take
    a serious look at it... If someone would like us to take a serious
    look at VMS security, then hopefully they know where to find us. I
    think we'll leave being an annoying sales person and forcing stuff on
    people to persons better suited for that sort of thing and focus on
    what we do instead


    > You "killed the pig" with this! "Bagged an elephant" perhaps? Anyway, "Did
    > something extraordinary". OTOH, if you'd rather get your jollies from
    > gleaning credit-card details from your local library then that's up to you.
    >


    Glad you like it.

    > [
    > I'm not familiar with OM-click Users-Group, who are they? We would be
    > interested in getting in touch with anyone that can lend us some
    > weight reporting problems if we continue to look at VMS.
    > ]
    >
    > You can find out about Click-XT atwww.nasdaqomx.com(and many other
    > places). It is a highly-rated and popular settlments, matching and clearing,
    > trading system used all over the world. (I'm guessing here that it has a
    > Users Group). The next time you come up against one of the many smug,
    > arrogant, narcissistic, pricks at HP/VMS that is too bloated with
    > self-importance to listen to you about your silly vulnerability, I suggest
    > THAT you contact someone such as these. They are tasked specifically with
    > maintaining stability, transparency, and accountability in global trading
    > markets, and would not take kindly to market manipulation or having all
    > western economies plunged into an abyss by a simple DoS attack.
    >


    Ahh OMX, I know someone there... Thanks for the tip!

    > Regards Richard Maher


  7. Re: Loose Cannon-dian (was: Re: DEFCON 16 and Hacking OpenVMS)

    On Sep 1, 9:50*pm, B...@rabbit.turquoisewitch.com (Brad Hamilton)
    wrote:
    > In article <98806cee-a3cd-4fd4-8a3c-74e312e3d...@z72g2000hsb.googlegroups..com>,*b...@signedness.org wrote:

    :
    > >stumbled upon what I think are two new security bugs in VMS ( It is

    :
    > >I was hoping someone could tell us if there is a better place to
    > >report them at HP than the security-alert email address since they


    > Although I'm not privy to your interactions with HP, I still think that it was
    > the interaction here that quickly got HP's (VMS Engineering) attention last time.


    Right. Please at least give them the benefit of the doubt.
    From my modest interactions with them, I got the distinct impression
    that a patch was triggered thanks to your report. and a basic patch
    kit was available well before the C.O.V. reporting. If you were not
    properly thanked for that, then I am a little dissapointed, but do not
    knwo the full context.

    > If I were you, I would still report through "normal" channels, and then return
    > "here" to find a way to communicate "off-line" with folks here who may have
    > the ability to raise the visibility of your findings with the "proper" VMS
    > Engineering folks.


    Right. For example, send me (I'm not HP) or Kerry Main or John Reagan
    an Email and one of us can poke folks,
    or try to connect you more directly if deemed appropriate/useful. No
    need to send details, unless you want a quick sanity check. Other
    readers/replies know names to contact as well.

    fwiw,
    Hein.

    >
    > All that being said, please realize that a resolution may not happen quickly -
    > I believe that regression testing and other vetting must take place before an
    > ECO or MUP is released to customers.


    Right. Allthough not much of an excuse, this SMG report happened just
    while office and systems where being move. That did not help.

    Cheers,
    Hein.



  8. Re: Loose Cannon-dian

    Hein RMS van den Heuvel wrote:
    > On Sep 1, 9:50 pm, B...@rabbit.turquoisewitch.com (Brad Hamilton)
    > wrote:
    >
    >>In article <98806cee-a3cd-4fd4-8a3c-74e312e3d...@z72g2000hsb.googlegroups.com>, b...@signedness.org wrote:

    >
    > :
    >
    >>>stumbled upon what I think are two new security bugs in VMS ( It is

    >
    > :
    >
    >>>I was hoping someone could tell us if there is a better place to
    >>>report them at HP than the security-alert email address since they

    >
    >
    >>Although I'm not privy to your interactions with HP, I still think that it was
    >>the interaction here that quickly got HP's (VMS Engineering) attention last time.

    >
    >
    > Right. Please at least give them the benefit of the doubt.
    > From my modest interactions with them, I got the distinct impression
    > that a patch was triggered thanks to your report. and a basic patch
    > kit was available well before the C.O.V. reporting. If you were not
    > properly thanked for that, then I am a little dissapointed, but do not
    > knwo the full context.
    >
    >
    >>If I were you, I would still report through "normal" channels, and then return
    >>"here" to find a way to communicate "off-line" with folks here who may have
    >>the ability to raise the visibility of your findings with the "proper" VMS
    >>Engineering folks.

    >
    >
    > Right. For example, send me (I'm not HP) or Kerry Main or John Reagan
    > an Email and one of us can poke folks,
    > or try to connect you more directly if deemed appropriate/useful. No
    > need to send details, unless you want a quick sanity check. Other
    > readers/replies know names to contact as well.
    >
    > fwiw,
    > Hein.
    >
    >
    >>All that being said, please realize that a resolution may not happen quickly -
    >>I believe that regression testing and other vetting must take place before an
    >>ECO or MUP is released to customers.

    >
    >
    > Right. Allthough not much of an excuse, this SMG report happened just
    > while office and systems where being move. That did not help.
    >
    > Cheers,
    > Hein.


    There's a family story/legend, probably garbled that one of my aunts was
    vacationing or visiting somewhere on the coast of Maine during World War II,
    and spotted what she thought was a submarine. She reported it and later there
    was some kind of activity (Navy or Coast Guard ships sailing around where she
    saw it), but nothing was reported in the newspapers, and she heard nothing
    more about it until about 6 months later she got a cryptic letter from the
    Coast Guard thanking her for her diligence. Whether there really was a
    submarine out there, no one knows... (It wasn't an unlikely place to see
    one; she may have been visiting my Dad who worked one or two summers during
    high school at the Bath Iron Works, which was a major naval shipyard at the
    time. It's also fairly close to Portsmouth NH which was an American sub base.)

    So HP may have appreciated and may have acted on your information without
    acknowledging it, though some kind of "thank you" would certainly encourge
    others to come forward.

    I've seen somewhere that anyone can file a bug report (security-related or
    not) and they'll take it seriously, but they won't necessarily respond unless
    they need more information or if you have a support contract. It looks
    just like a black hole even if it isn't.

    --
    John Santos
    Evans Griffiths & Hart, Inc.
    781-861-0670 ext 539

  9. Re: Loose Cannon-dian

    In article <7F4vk.215$Af3.5@trnddc06>, John Santos writes:
    >Hein RMS van den Heuvel wrote:
    >> On Sep 1, 9:50 pm, B...@rabbit.turquoisewitch.com (Brad Hamilton)
    >> wrote:
    >>
    >>>In article <98806cee-a3cd-4fd4-8a3c-74e312e3d...@z72g2000hsb.googlegroups.com>, b...@signedness.org wrote:

    >>
    >> :
    >>

    >So HP may have appreciated and may have acted on your information without
    >acknowledging it, though some kind of "thank you" would certainly encourge
    >others to come forward.
    >
    >I've seen somewhere that anyone can file a bug report (security-related or
    >not) and they'll take it seriously, but they won't necessarily respond unless
    >they need more information or if you have a support contract. It looks
    >just like a black hole even if it isn't.
    >


    Feedback to those who report bugs is only polite. However with security bugs it
    is an essential part of the process since there is a long history of those
    discovering such bugs and not being kept informed by the vendor resorting to
    publication of the details or even demo exploits in order to pressure the
    vendor into providing a fix.

    David Webb
    Security team leader
    CCSS
    Middlesex University

    >--
    >John Santos
    >Evans Griffiths & Hart, Inc.
    >781-861-0670 ext 539


  10. Re: Loose Cannon-dian (was: Re: DEFCON 16 and Hacking OpenVMS)

    > bugs@signedness.org writes:
    >>
    >> ... Anyway I was preparing
    >> updated slides for a presentation we are doing in Stockholm,


    When and where ?
    On the "VMS Techical Update"

    Jan-Erik.

  11. Re: DEFCON 16 and Hacking OpenVMS

    In article <48b85f07$0$9629$c3e8da3@news.astraweb.com>, JF Mezei writes:
    > re: memory protection.
    >
    > Can anyone explain to me why I was under the impression that under VMS,
    > you could not branch to data because data blocks were marked no-execute
    > (or whatever) ?


    Yes, data has long been marked no-execute. But that was a linker
    optimization only on VAX and Alpha.

    > Was this ever the case ?
    >
    > This argument had been brought up, (perhaps in my dreams) to show how
    > VMS was far mroe immune to buffer overflows than other operating
    > systems. (with the corrolary that you could not write to execute
    > sections and overwrite code).


    The reason VMS itself is more immune stems more from using langauges
    where creating such a situation is not trivial instead of langauges
    where it's the default, and overall quality in design and
    implementation.

    > Recent discussions seem to have completely shattered these
    > misconceptions I had, but i am really puzzled as to why I would have had
    > them to begin with.


    Because you worked in Macro, or PSECT controls via linker options and
    saw the NOEXE flag without understanding what that meant.


  12. Re: DEFCON 16 and Hacking OpenVMS

    In article <3cde3b31-285e-4be0-98e1-287219571280@l42g2000hsc.googlegroups.com>, johnwallace4@yahoo.co.uk writes:
    >
    > I'd be interested in that too. For now, I'm thinking that I've assumed
    > that since the capability existed on PDPs (in hardware and in
    > compilers and in linker and OS), and that the same capability existed
    > in "bigger, better" VMS (at least in VMS compilers and linker (?)),
    > there's been an assumption on my part that the hardware and OS did the
    > sensible thing. As I mentioned earlier in a post which refers to the I
    > +DS manual, the PTE on VMS does seem to have a "fault on execute" bit,
    > which also supports my (apparently incorrect) assumption that data
    > sections are capable of being non-executable.


    If you used compatability mode on the VAX-11 series, or the emulator
    built into later versions of the RSX AME, you probably could take
    advantage of those PDP-11 features. But even on PDP-11 only the
    systems with mor ethan 64KB RAM installed actually could do anthing
    for you with those features.


  13. Re: DEFCON 16 and Hacking OpenVMS

    In article , Johnny Billquist writes:
    >
    > That would be equally true if the physical address space was only 16 bits as well.
    > The point is that your virtual addresses always goes through an address
    > translation, which both tells where the virtual address maps to in the physical
    > address space, and can also declare that some addresses aren't even valid.
    >


    No. The PDP-11 did not do virtual address translation on a page by
    page basis. It just took the virtual address and added the base
    physical address in the corresponding APR. If you had 64K or less
    RAM all physical addresses were within reach no matter what was in
    the APR. The mapping for I and D space might be different, but it
    was a calculable fixed offset for the duration of a task.

    And if you had less than 64KB RAM, no APR might be used. Lots of
    PDP-11 didn't even have APR.


  14. Re: DEFCON 16 and Hacking OpenVMS

    In article , Johnny Billquist writes:
    >
    > Well, that is assuming the OS don't allow the program to do that.
    > Since the page tables lives in I/O space, it requires that the user program have
    > access to the I/O page in order to manipulate this by itself.
    > Normal, timesharing OSes don't allow user programs to have access to the I/O
    > page, so correct, no program have the ability to change it's mapping like that.


    The memory space protection on the page table may be similar to the
    memory space protection on I/O space, but it is not necessarily in I/O
    space. In fact on VMS the page table is in P1 or P2 space and I/O
    space is in S0, S1, or S2.

    >> Can I assume things like the VAX and Alpha are strictly flat address space
    >> with all memory visible?

    >
    > Well, not really, no.


    The use of address tables does not change the fact that the
    architecture presents a flat address space.


  15. Re: DEFCON 16 and Hacking OpenVMS

    In article , koehler@eisner.nospam.encompasserve.org (Bob Koehler) writes:
    {...snip...}
    >
    > Because you worked in Macro, or PSECT controls via linker options and
    > saw the NOEXE flag without understanding what that meant.



    --
    VAXman- A Bored Certified VMS Kernel Mode Hacker VAXman(at)TMESIS(dot)COM

    .... pejorative statements of opinion are entitled to constitutional protection
    no matter how extreme, vituperous, or vigorously expressed they may be. (NJSC)

    Copr. 2008 Brian Schenkenberger. Publication of _this_ usenet article outside
    of usenet _must_ include its contents in its entirety including this copyright
    notice, disclaimer and quotations.

  16. Re: DEFCON 16 and Hacking OpenVMS

    Bob Koehler skrev:
    > In article , Johnny Billquist writes:
    >> Well, that is assuming the OS don't allow the program to do that.
    >> Since the page tables lives in I/O space, it requires that the user program have
    >> access to the I/O page in order to manipulate this by itself.
    >> Normal, timesharing OSes don't allow user programs to have access to the I/O
    >> page, so correct, no program have the ability to change it's mapping like that.

    >
    > The memory space protection on the page table may be similar to the
    > memory space protection on I/O space, but it is not necessarily in I/O
    > space. In fact on VMS the page table is in P1 or P2 space and I/O
    > space is in S0, S1, or S2.


    Now, I'm definitely not entirely up to speed on the Alpha, but for the VAX, it's
    P0 and P1 space, and S0 space, while S1 was reserved. I don't even think there
    is a way to setup any map for S1 space. S2 must definitely be some Alpha
    extension, right?

    However, the VAX also is very different from the PDP-11 in this aspect. In the
    PDP-11, the page table is at a fixed address in I/O space, while on the VAX, the
    page tables are pointed at by some special registers in the CPU. So, on the VAX,
    you need to be able to write to those registers, or else have the (normal)
    memory where the page table is located accessible, in order to modify the page
    table. On the PDP-11, all you need is access to the I/O page.
    The page table on the PDP-11 isn't in normal memory, but it's actually an area
    of RAM inside the MMU, which is mapped into I/O space. If the MMU isn't enabled,
    you can even use the page table as a small scratch area of really fast memory,
    or even write small diagnostic programs that execute from there, without you
    needing to even have any working normal memory on a PDP-11 (sometimes *very*
    useful when you try to diagnose a sick machine).

    >>> Can I assume things like the VAX and Alpha are strictly flat address space
    >>> with all memory visible?

    >> Well, not really, no.

    >
    > The use of address tables does not change the fact that the
    > architecture presents a flat address space.


    At a hardware level? Sure. The same does a PDP-11. The fact that you can't from
    a normal program get access to the full 4 megs of memory don't change that. :-)
    The CPU and MMU forms a 22-bit address for every memory access on a (large)
    PDP-11. And from the front panel all of that memory is acessible, using 22-bit
    addresses, the whole time.

    But as for virtual addresses, no, all memory is not always visible, nor is it
    addressed in a linear fashion, when looking at what each virtual address boils
    down to. But the same is true for the PDP-11. There isn't really that much
    difference between a PDP-11 and VAX. The fact that the physical address space is
    larger than the virtual on a PDP-11 doens't really change anything. Anyone who
    thinks it does needs to go back and study things a little more.

    Johnny

    --
    Johnny Billquist || "I'm on a bus
    || on a psychedelic trip
    email: bqt@softjar.se || Reading murder books
    pdp is alive! || tryin' to stay hip" - B. Idol

  17. Re: DEFCON 16 and Hacking OpenVMS

    Bob Koehler skrev:
    > In article , Johnny Billquist writes:
    >> That would be equally true if the physical address space was only 16 bits as well.
    >> The point is that your virtual addresses always goes through an address
    >> translation, which both tells where the virtual address maps to in the physical
    >> address space, and can also declare that some addresses aren't even valid.
    >>

    >
    > No. The PDP-11 did not do virtual address translation on a page by
    > page basis.


    Yes it did. You are totally confusing the PDP-11 with some Intel designs.

    > It just took the virtual address and added the base
    > physical address in the corresponding APR. If you had 64K or less
    > RAM all physical addresses were within reach no matter what was in
    > the APR. The mapping for I and D space might be different, but it
    > was a calculable fixed offset for the duration of a task.


    Sorry, but no. You are totally wrong. Now, if you insist, I'll be happy to teach
    you how the PDP-11 works in all the gory details, down to individual signals, if
    you want to. But maybe this would be regarded as slightly off-topc for
    comp.os.vms, so we should perhaps take in in another forum, such as
    alt.sys.pdp11, or vmsnet.pdp-11, or mail or something?

    Your idea, however, on how a PDP-11 works would make a lot of things an OS like
    RSX do impossible. You couldn't even have shared libraries if your view were
    correct, since that is pieces of memory that several processes have mapped into
    their address space, even at different virtual addresses (if it's position
    independent), and the rest of the address space is still their own, or used for
    even more shared libraries.

    > And if you had less than 64KB RAM, no APR might be used. Lots of
    > PDP-11 didn't even have APR.


    Yes, there were PDP-11s that didn't have an MMU. The same is true of any PDP-11
    that have an MMU, when the MMU isn't enabled. And that's the way things are at
    boot time for all of them. But once you enable the MMU, you play around with the
    address translations, and you'll have 8 pages, even if your machine only have 8K
    of physical memory. And all of those pages can be fooled around with, making it
    look like you have lots more of memory than you actually have.
    Virtual memory, it's called.

    Johnny

    --
    Johnny Billquist || "I'm on a bus
    || on a psychedelic trip
    email: bqt@softjar.se || Reading murder books
    pdp is alive! || tryin' to stay hip" - B. Idol

  18. Re: DEFCON 16 and Hacking OpenVMS

    Bob Koehler skrev:
    > In article <3cde3b31-285e-4be0-98e1-287219571280@l42g2000hsc.googlegroups.com>, johnwallace4@yahoo.co.uk writes:
    >> I'd be interested in that too. For now, I'm thinking that I've assumed
    >> that since the capability existed on PDPs (in hardware and in
    >> compilers and in linker and OS), and that the same capability existed
    >> in "bigger, better" VMS (at least in VMS compilers and linker (?)),
    >> there's been an assumption on my part that the hardware and OS did the
    >> sensible thing. As I mentioned earlier in a post which refers to the I
    >> +DS manual, the PTE on VMS does seem to have a "fault on execute" bit,
    >> which also supports my (apparently incorrect) assumption that data
    >> sections are capable of being non-executable.

    >
    > If you used compatability mode on the VAX-11 series, or the emulator
    > built into later versions of the RSX AME, you probably could take
    > advantage of those PDP-11 features. But even on PDP-11 only the
    > systems with mor ethan 64KB RAM installed actually could do anthing
    > for you with those features.


    No, you make assumptions that just aren't correct.
    First of all, the VAX-11 PDP-11 emulation only extended to user mode, so the MMU
    stuff of a PDP-11 was never emulated by a VAX. So nothing of this can be adopted
    from knowing how the VAX-11 worked.

    Second, you are still confusing the existance, and capabilities of an MMU, with
    physical memory. The MMU is used to play with virtual memory, more or less
    making physical memory an irrelevant parameter. All more memory will do is make
    it possible to get things done faster.
    Admittedly there are limits on how little memory you can have and still make
    something meaningful possible. For the PDP-11, I'd say the low limit of physical
    memory is probably four pages.
    You need atleast one page for the minimal kernel to exist and do something.
    And then one instruction can mean references to three pages, and since
    instructions can't be trapped without needing to restart, all three memory
    references must be able to complete without a fault for the instruction to be
    executed. (The three memory references are the instruction fetch itself, the
    source argument and the destination argument.)

    As long as you have that much, you can fool the user program that it actually
    have 64K of D-space and 64K of I-space. That's what virtual memory is all about.

    If the code reference a memory location that isn't currently valid, your "OS"
    will trap, where it can check the reason, the memory referenced, page it in,
    update the page table, and restart the instruction.
    Yes, all that is perfectly doable on a PDP-11.

    Just because no OS (that I know of) implemented demand paging don't mean the
    hardware can't do it.

    Johnny

    --
    Johnny Billquist || "I'm on a bus
    || on a psychedelic trip
    email: bqt@softjar.se || Reading murder books
    pdp is alive! || tryin' to stay hip" - B. Idol

  19. Re: Loose Cannon-dian (was: Re: DEFCON 16 and Hacking OpenVMS)

    In article <64f54c97-3f32-44f9-89ac-35d326e4734b@y38g2000hsy.googlegroups.com>, bugs@signedness.org writes:
    >I still haven't figured out how to quote multiple posts in this google
    >user interface but I'll try to answer several people here.


    Bugs, Google groups suck. Get a real newsreader and a news feed.

    --
    VAXman- A Bored Certified VMS Kernel Mode Hacker VAXman(at)TMESIS(dot)COM

    .... pejorative statements of opinion are entitled to constitutional protection
    no matter how extreme, vituperous, or vigorously expressed they may be. (NJSC)

    Copr. 2008 Brian Schenkenberger. Publication of _this_ usenet article outside
    of usenet _must_ include its contents in its entirety including this copyright
    notice, disclaimer and quotations.

  20. Re: Loose Cannon-dian (was: Re: DEFCON 16 and Hacking OpenVMS)

    On Tue, 02 Sep 2008 07:39:15 -0700, VAXman- <@SendSpamHere.ORG> wrote:

    > In article
    > <64f54c97-3f32-44f9-89ac-35d326e4734b@y38g2000hsy.googlegroups.com>,
    > bugs@signedness.org writes:
    >> I still haven't figured out how to quote multiple posts in this google
    >> user interface but I'll try to answer several people here.

    >
    > Bugs, Google groups suck. Get a real newsreader and a news feed.
    >


    Opera is pretty good as a newsreader.

    --
    PL/I for OpenVMS
    www.kednos.com

+ Reply to Thread
Page 25 of 35 FirstFirst ... 15 23 24 25 26 27 ... LastLast