home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!wupost!waikato.ac.nz!aukuni.ac.nz!kcbbs!kc
- Newsgroups: comp.unix.aix
- Subject: Re: process numbers
- Message-ID: <102939.78394.17730@kcbbs.gen.nz>
- From: Craig_Anderson@kcbbs.gen.nz (Craig Anderson)
- Date: 10 Jan 93 21:46:34 GMT
- Organization: Kappa Crucis Unix BBS, Auckland, New Zealand
- Lines: 26
-
- In article <C0Ly9M.nB6@austin.ibm.com> jfh@greenber.austin.ibm.com (John F Hau
- >The process ID is a covert storage channel. If AIX used sequential
- >pids two co-operating processes could pass information back and forth
- >to each other with the following little piece of code -
-
- But it's only a *covert* channel if you don't have auditing available
- for the fork() system call. It would have been much more pleasant to
- leave the PIDs alone and allow auditing of fork() for the security
- insane.
-
- Besides, regardless of the selection of PIDs i can still use use fork()
- [not to mention millions of other things] as a covert channel.
-
- If process A does a bunch of forks to send a 1, and doesn't do any
- fork()s to send a 0, then process B, which sits in a loop counting how
- many fork()s it can do per unit time, still knows what A is sending
- (since B will be able to do less forks() when A is also fork()ing).
-
- Regardless of how many other processes are out there doing their own
- fork()ing, the channel will still have a signal to noise ratio greater
- than zero and that is good enough for transmitting error free data.
-
- -Craig
- --
- Craig Anderson
- craig%abstract@kcbbs.gen.nz
-