sorry bout that double posting. this intermitent turnaround time from this
list screws me up.
anyone ever thought of doing a /dev/swap ?
the way i see it, you could add a superblock to swap, and dump the entries
in /etc/fstab.
kernel scans and finds type fd on disk, reads superblock, sets up devices,
mounts root, etc.
kernel scans, finds type fe at the same time, reads superblock, sets up a
raid1 swap array. ( say raid one cause i would rather avoid the
computation of raid5 parity, dont know though). Could work?
then we mount /dev/swap as type swap.
problems i see are still code forking, this means two different places in
the kernel will be using similar code, just one with possible ram
allocation, and one without. this seems like a bad idea.
so, how about, is it possible to modify at least the raid1 code to never
do anything that will use up ram? perhaps setting aside a small space at
boot time, to do our calculations in? (can you tell i am not a c
programmer :-) then make sure the swap code is ok with this?
comments?
al
"so don't tell us it can't be done, putting down what you don't know.
money isn't our god, integrity will free our souls" - Max Cavalera
