-----Original Message-----
From: Todd L. Miller <[EMAIL PROTECTED]>
To: [EMAIL PROTECTED] <[EMAIL PROTECTED]>
Cc: [EMAIL PROTECTED] <[EMAIL PROTECTED]>
Date: Tuesday, August 10, 1999 5:02 PM
Subject: Re: [JOS-Kernel] Priority List
>> Here follows the tentative list, pending updates based upon feedback.
>> Please reply to the list, so that everybody knows what's going on.
>>
>> (1) Memory management.
>> (2) Performance.
>
>(3) Deciding what to tell the architecture list, now that they've decided
>to try and design a driver system without involving the kernel list.
>
>-_Quinn
Ok, I'm at fault for this. I moved the discussion over to "Arch" since the
driver "architecture" doesn't directly relate to the kernel (but, my! the
implications it has!).
Actually, for Device architecture (from a nearly nieve perspective), I see
only two major hooks needed to join the kernel to the drivers:
(1) Pat down the hardware resource interface to a point where somebody is
truly satisfied. This is critical for both groups, but doesn't affect many
people beyond us.
(2) Create a hook where the devices can detect and install themselves.
This is basically the "main" method of the OS, I feel (though it could be
something else too). This is also a link into the registry and other cans
of worms. But still, the kernel group (I feel) need only be conserned with
the "main" routine link. If I'm not mistaken, this already exists.
Anyone see any oversights, or simplification of the real problem?
-Matt
_______________________________________________
Kernel maillist - [EMAIL PROTECTED]
http://jos.org/mailman/listinfo/kernel
