[TAKE3] [PATCH] non-overcommit memory, improved OOM handling, safetymargin (was Re: Prevent OOM from killing init)

Ugh, something was going screwy. Trying from a different machine. -- The attached patch is against 2.4.1 and incorporates the following: - More optimistic OOM checking, and slightly improved OOM-kill algorithm, as per my previous patch. - Accounting of reserved memory, allowing for... -

[PATCH] non-overcommit memory, improved OOM handling, safetymargin (was Re: Prevent OOM from killing init)

The attached patch is against 2.4.1 and incorporates the following: - More optimistic OOM checking, and slightly improved OOM-kill algorithm, as per my previous patch. - Accounting of reserved memory, allowing for... - Non-overcommittal of memory if sysctl_overcommit_memory < 0, enforced even f