I think maybe one partial possibility is to have a userspace OOM kill manager which decides what is to be killed when OOM is reached and it is activated by the kernel. It would allow a few tasks to be unkillable (syslogd, perhaps, and other important crash-tracers), a few to be killable only when they themselves reach a certain unreasonable size (sendmail, inetd), root processes generally to be killable only when normal-user processes are dead (don't know how this works with capabilities). There could be special kill-procedures for certain processes, etc.
Just an idea, may not be possible (i.e. may not function in OOM conditions, may be insecure, etc....)
David Feuer
dfeuer@his.com
-
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/
This archive was generated by hypermail 2b29 : Fri Jul 07 2000 - 21:00:20 EST