Porting to 4MB pages

Craig Milo Rogers (rogers@ISI.EDU)
Thu, 29 Oct 1998 09:58:00 -0800


>>but 4M real-memory mappings raise other serious issues like how do we
>>allocate/deallocate (swap?) those monsters.
...
>2. swap the entire 4M page. Not very efficient.

Extending this thought a little, what would it take to create
an i386-4M "port" of Linux, by which I mean, one in which *all* pages
were 4M?

Let's see: bootstrap, libc, ld.so, binutils, swapper, kernel
mm, there are potentially quie a few areas of code that might need to
be changed. Yet, it should be a subset of the list of things that
need to be done to port Linux to an entirely new architecture, and
there's lots of experience doing that, right?

Of course, I don't even know if the underlying Intel
architecture *allows* an all-4M machine (my apologies for not doing my
homework before opening my mouth). There might be exceptional cases,
such as, say, mapping PCI memory blocks, which would still require 4K
pages.

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/