[1] The notion of using a fixed size swap partition and then sizing it
by trial and error is primitive. Does it work well enough most of the
time or do we need something more adaptive?
[2] File system based swap files. Are these neccessarily much less
efficient than swap partition based swap spaces? They are much more
adaptable but you can always run out of disk. You could always implement
swap files as a file system (under the Virtual Filesystem) and keep similar
semantics to the current swap file semantics.
[3] Daemons that run every NNs checking for events usually run too early
or too late and should never do really important stuff. A better solution
is to have the kernel demand the services of a daemon when thresholds
(eg %age free buffers) are reached.
Dave
----------------------------------------------------------------------
David A Rusling Principal Engineer
European Semiconductor Applications Digital Equipment Co Ltd.,
Engineering PO Box 121,
Imperial Way,
Worton Grange
Reading RG2 0TU
Linux, Alpha, StrongArm, PCI Tel: UK-(0)1734-204380
Fax: UK-(0)1734-203133
----------------------------------------------------------------------