From mboxrd@z Thu Jan  1 00:00:00 1970
Message-ID: <ee9e417a0607201510s795c4b58o123c184895709e0f@mail.gmail.com>
Date: Thu, 20 Jul 2006 15:10:44 -0700
From: "Russ Cox" <rsc@swtch.com>
To: David.Eckhardt@cs.cmu.edu,
	"Fans of the OS Plan 9 from Bell Labs" <9fans@cse.psu.edu>
Subject: Re: [9fans] assert(*norealmode || *nomp)?
In-Reply-To: <12728.1153429938@piper.nectar.cs.cmu.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
References: <12728.1153429938@piper.nectar.cs.cmu.edu>
Cc:
Topicbox-Message-UUID: 8698d1da-ead1-11e9-9d60-3106f5b1d025

> Am I right in believing that due to the
> proliferation of single-processor
> motherboards with APICs that it's generally
> advisable for x86 machines to have either
>   *norealmode=
> or
>   *nomp=
> in PLAN9.INI?

No.  The E820 scanner uses realmode and works
fine even on multiprocessors.

> Should the realmode code
> reject attempts to use it if the kernel
> has already gone down the APIC path?
> Are there other things (APM?) which should
> disable themselves?

Why would it be incompatible with APICs?
When we go to real mode we turn off interrupts
on the local processor.

Russ