Subject: Re: mac-on-darwin?
From: Terry Lynch (mtnweb@mac.com)
Date: Wed Apr 25 2001 - 19:26:03 MDT
on 25/4/01 2:15 PM, Samuel Rydh at samuel@ibrium.se wrote:
>
> On Mon, Apr 23, 2001 at 06:25:31PM -0700, Terry Lynch wrote:
>>> I wonder if there is any effort made or drive towards making a
>>> "mac-on-darwin" extension? (Sorry, I'm sure there is another word for
>>> "extension" in the linux world...) I got that impression that from
>>> one of your latest posts, Samuel, that there might be such a future?
>>
>> I am very interested in "mac-on-darwin," and I am willing to contribute what
>> I can in the porting effort. I have done some c work, but I don't know much
>> about MOL. Does anyone know actually what it would take to port MOL over to
>> darwin? Just a brief summary would be wonderful and perhaps an estimate of
>> what percentage of code would have to be changed.
>>
>> Terry
>>
>
> Actually, I've thought about porting MOL to *BSD and
> to Darwin. The non-trivial part is porting the kernel module -
> the rest should be more or less routine.
>
> Most of the Linux-dependent code is page table realted.
> Linux uses the PTE hash as a cache, while Darwin (and possibly
> *BSD) uses a more complicated model with cast-out lists.
> It is also necessary to find a way to detect swap-outs
> (MOL hooks a Linux function for this).
>
> I think the number of lines needed is quite limited. The difficulty
> is more technical (assembly skills and detailed knowledge
> about memory management and the MMU is probably what it takes).
> A good starting point is understanding the code in mol/kernel_module/.
This is great! I am going to start studying the kernel_module sources right
now.
Thanks Samuel!
This archive was generated by hypermail 2a24 : Wed Apr 25 2001 - 22:33:01 MDT