Re: Kernel cpu selection + other platforms?
david parsons (o.r.c@p.e.l.l.p.o.r.t.l.a.n.d.o.r.u.s)
23 Jul 1997 00:53:39 -0700
In article <linux.kernel.Pine.LNX.3.96.970722223623.1839D-100000@sigil.csc.com>,
Teunis Peters <teunis@usa.net> wrote:
>On 22 Jul 1997, david parsons wrote:
>
>> In article <linux.kernel.33D3BDE2.404EB0D1@netzblick.com>,
>> Roland Steinbach <roland@netzblick.de> wrote:
>> >Mike Jagdis wrote:
>> >
>> >> Why? It is pretty easy to recognise all x86 varieties (other than the
>> >> early 486 clones) at run time and simply Do The Right Thing without
>> >> asking the user daft questions about whether they have microcode
>> >> bug #564 or working frozzle optimizations.
>> >
>> >_I_ don't want my settings be controlled by an OS,
>>
>> Whyever not? Resource management is, after all, why you've got an OS
>> in the first place.
>
>CPU settings that drastically change how the OS acts don't count as
>resources....
>
>Not unless the CPU can be readjusted without affecting running code.
>[as an example : try changing page-sizing on the fly... say to 8K pages
>from 4K pages....]
You don't need to do that on the fly; you do it just once, when the
system is initializing. Unless it's a build option that changes
the instructions that the machine execute, or an optimization for
the particular processor (adding waitstates so you won't drain a
particular pipeline and the like), it's more useful for a general-
purpose kernel to have it try and properly set up the cpu at
runtime, and have it so you can turn things off manually if they
give you problems.
____
david parsons \bi/ I like autoconfiguring systems.
\/