Subject: Re: MACHINE_NONCONTIG
To: None <firstname.lastname@example.org>
From: Gordon W. Ross <email@example.com>
Date: 11/15/1996 12:37:13
> From: firstname.lastname@example.org
> Date: Fri, 15 Nov 1996 09:47:25 -0500 (EST)
> [ MACHINE_NONCONTIG is an option only for the port maintainer ... ]
> Well, some macs have contiguous memory, others have holes... The macs
> have been using MACHINE_NONCONTIG for some time (0.9 or 1.0, I think).
> > Maybe this should be stated more prominently in order to at least encourage
> > new porters to code for the MACHINE_NONCONTIG case instead of the old one.
> If we want to do that, it would make sense to reverse the define to
> be MACHINE_CONTIG and define that on the ports where it isn't used.
> It does make some sense to move all ports to the MACHINE_NONCONTIG
> interface since it would mean one less conditional in the VM system
> and the MACHINE_CONTIG can be represented as a MACHINE_NONCONTIG
> configuration with one bank.
Yes, replacing this "option" with one of the opposite sense
would help discourage proliferation of the old pmap interface
by making it a little more difficult. (The port master would
have to consciously add MACHINE_CONTIG 8^) Maybe the opposite
sense name could be better -- how about PMAP_MEM_CONTIG instead?
So, I propose that we:
(1) Add the PMAP_MEM_CONTIG definition to all ports that do not
presently use the MACHINE_NONCONTIG option.
(2) Change the VM code conditionals from
#ifdef PMAP_MEM_CONTIG /* old, depreciated way */
(3) Remove the MACHINE_NONCONTIG definition from the
ports that use the preferred interface.
The intention would be to make it more convenient to use the
preferred interface than to use the depreciated interface.