RE: as/400 / linux / lpar

["Andy Nolen-Parkhouse" <aparkhouse@xxxxxxxxx>]

I've had a long and successful career dealing with the IBM midrange and
have never heard of an MSR-bit.  I would not be qualified to answer your
questions.  It strikes me as intuitively correct that some provision
should be made within silicon when two operating systems are sharing the
same processor, whether there are new instructions or not.  My
understanding is that Linux is talking directly to the hardware, not
riding on top of the MI level of OS/400.  I realize you are looking for
more than non-technical intuition and I hope you get your answers.

Your point on fractional LPAR is valid.  Yes, it isn't just Linux and
yes a customer could save some money if they had partition requirements
for partial processor allocation.  I've partitioned a few boxes and it
was very frustrating (and expensive) to allocate an entire processor to
a partition which did not need that much power.

Regards,
Andy

> What would IBM rather have: 1) a customer running Linux (and remember
> that the issue is not Linux but fractional LPARs) on an older box, or
2)
> a customer buying a new box because fractional LPARs (according to
> Glen saves you a LOT of money)?

> Subject: Re: as/400 / linux / lpar
>
> From: Andy Nolen-Parkhouse <aparkhouse@attbi.com>
> > Those are my thoughts.  Aside from a basic distrust of large
> > corporations, why do you think that IBM would intentionally disable
> > support for Linux on partial processors in older machines?
> >
>
> I didn't say (or meant to say) that. To me Linux was not the issue.
> The issue was: do you NEED special hardware to run fractional
> partitions. What does hardware provide for this that can't be done
> in software? This question has not been answered. All I hear is that
> people "believe" this or that, or "have heard" this or that, or "IBM
> says" this or that. No valid technical reason (to my satisfaction)
> has been given. So, when the claim is fractional partitions won't
> work on older processors, it means to mean that some software
> must prevent that since the hardware does not. All it takes for
> me to shut up is that someone tells me specifically what NEW/different
> instruction(s) or MSR-bit(s) is/are used by/for fractional LPARs.