Re: Profiling a murderer

Hi Tim,

> In fact different levels of SysConfig behave differently in this regard.
> There may also be more of this undocumented (at least to any sufficient
> level of detail) special-casing of particular adapter id.s/types.

I agree, the conclusions are derived from the Mod. 77 experiment, and
should not be applied to every single case, unless proved.


> If an "adapter" (including on-planar psuedo-adapters/virtual slots) is
> DISABLED then the resources that it may have previously been configured
with
> SHOULD be made free for use. However, if the reason for this behaviour is
to
> compensate for inadequate hardware design/implementation and done as a
> "fix/workaround" then I would be somewhat mollified.

I concluded earlier that DISABLED does not necessarily imply REMOVED. To my
knowledge there is no state REMOVED defined, but it could have been handled
with DISABLED at the cost of higher complexity. Perhaps it is this cost that
influenced the decision to treat DISABLED as a sort of STAND-BY, or there
are other OS and drivers issues, or but REMOVED was dropped for a number of
more important reasons.

Deviations from specs, undocumented features are common. It could be that
IBM did not trust third party producers, or it was not always possible, to
deliver strictly "physically isolated" designs and adapters. Another
concern might have been security, once you introduce a software-driven
configuration as opposed to manual configuration with jumpers, it becomes
possible to perform "unauthorized" changes to the configuration.


> Otherwise there is NO sense to the concept of POS
> registers and the Hardware Interface and a large part of the MicroChannel
> ARCHITECTURE.

Seems like some MCA aspects have not been truly finalized, speaking of the
PS/2 line. Whereas they could have handled common cases as expected, that
has not become the strictly general behaviour. I have to think more about
it, a possibility could be that the problem does not have or need at all a
clean solution.


> This is the reason that I would take issue with IBM about their
> under-the-cover beaching of their own standard.

I guess, we'll have to keep on searching for more answers to more questions.






.

Relevant Pages

  • Re: How Alpha will save Itanium - must reading for Bill Todd
    ... SuperDome configuration linearly to 4 processors, which, of course, was the ... So even if IBM ... > Madison is considerably more powerful than: ... Haven't seen any Opteron numbers for SAP SD: ...
    (comp.os.vms)
  • Re: ZA activité & log
    ... I didn't ask for a different protection ... configuration. ... I've almost everything AND ZoneAlarm. ... I just asked the reason why the computer tried to answer to the massive ...
    (comp.security.firewalls)
  • [opensuse] Re: [opensuse-amd64] Some problems after upgrading to 10.2
    ... For some reason, rpm seems to consume over 99% of the CPU time. ... If I invoke yast from inside KDE as a normal user, ... from the command line, the very same password is recognized. ... you make sure you did an on-line update configuration. ...
    (SuSE)
  • Re: IBM 4247-001 4247-V03 Front and Rear feeding
    ... Since the configuration is a 3812 *LAN, then you are most likely using host print transform. ... You therefore probably need to retrieve the workstation customization object for the manufacturer type and model and modify values for the drawer selection. ... Managed to cinfigure the IBM 4247 to print using crtdevprt. ...
    (comp.sys.ibm.as400.misc)
  • Re: Exclude from GPO ..
    ... Non-domain members aren't going to be trying to process the GPO anyway. ... The reason you should only deny Apply Policy, ... Computer configuration runs as your computer ... Paul Williams ...
    (microsoft.public.windows.server.active_directory)