[snip an exec'ing parent]
Well, most programs are not exec'ed with a child already existing. inetd, for
example, isn't. It can trust that its parent, which was an rc script, did not
do anything wacky like that. It should not, however, trust its children to
not use CLONE_PPID maliciously. (Its children are sometimes run with different
uids precisely because they are not fully trusted.) Is inetd robust in the
face of an unknown child exiting? Hopefully it will just syslog a warning and
go on, but if the programmer didn't feel like catering to this CLONE_PPID
case which he's never heard of, maybe it will fall off the end of an array
or linked list and crash.
Who volunteers to audit all programs for CLONE_PPID awareness?
-
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/