From mboxrd@z Thu Jan  1 00:00:00 1970
From: erik quanstrom <quanstro@coraid.com>
Date: Mon, 23 Feb 2009 17:20:45 -0500
To: me@acm.jhu.edu, 9fans@9fans.net
Message-ID: <9ad6c896cc1ad905714b3c2255a1dbf4@coraid.com>
In-Reply-To: <20090223215942.GA30413@unknown>
References: <138575260902230741t4672eacaw1f3f9f928638e0ab@mail.gmail.com>
	<14989d6e0902231344h1a5cf501ge8afb8024d0992a4@mail.gmail.com>
	<b6bac952aaa147aa078ad79996517760@coraid.com>
	<20090223215942.GA30413@unknown>
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
Subject: Re: [9fans] spreding the word
Topicbox-Message-UUID: a720cb62-ead4-11e9-9d60-3106f5b1d025

On Mon Feb 23 17:02:29 EST 2009, me@acm.jhu.edu wrote:
> >if you're after the historical progression of how the
> >structure of the kernel evolved, the file server kernel
> >is much more interesting.
>
> Any bits in particular? Any reason why?

port/proc.c is very interesting, as are pc/lock.c
and pc/trap.c.

they are all very interesting as they illustrate
the same concepts the pc kernel deals with, but
they are substatially simplier.  i fear i've complicated
things a bit.  hardware is more complicated these
days.

naturally, the fs kernel is less capable.  it lacks
dynamic memory, for example.  but the beauty
is that it doesn't need those things.  i think it's
an interesting study in tradeoffs.

einstein said "make everything as simple as
possible, but not simpler".  i think ken's take
is that if it's not simple enough, you're solving
the wrong problem.

- erik