Re: Ideas for memory management hackers (2)

Evan Jeffrey (
Sun, 16 Nov 1997 16:43:40 -0600

>There could be a mechanism to "hide" this (1*PGSIZE-10)*8 page
>limitation by allowing bigger swap partitions, and embedding into
>them several sequential swap areas.
>At the mm/swapfile.c: sys_swapon(), the system would try to see of the
>partition is bigger than the magic limit, and if it is, sees if there
>is another swap-signature with associated swap pages, and takes that
>new area into use too, etc. until the end of the partition.

I think it would be better to create a "new" swapfile type for > 128 (or
512) MB swap areas, where the swap area signature (SWAP-SPACE ?) would be
replaced by something like "NEW-SWAPXX" where XX is a short int indicating
the number of bitmap blocks at the beginning of this swap area. This give
us an ~8 TB theoretical limit, which is pretty big.