[2490] in cryptography@c2.net mail archive
Re: NYT Article on Groat Spy Case
daemon@ATHENA.MIT.EDU (Yoav Yerushalmi)
Mon Apr 13 19:11:03 1998
To: Phil Karn <karn@qualcomm.com>
Cc: reinhold@world.std.com, cryptography@c2.net
In-Reply-To: Your message of "Mon, 13 Apr 1998 14:17:31 PDT."
<199804132117.OAA00450@servo.qualcomm.com>
Date: Mon, 13 Apr 1998 17:54:02 EDT
From: Yoav Yerushalmi <yoav@MIT.EDU>
FYI, I'm looking at the manual pages here (on NetBSD) and apparently
(manpage follows) mprotect can be modified to achieve this.
It doesn't appear to have a bit specifying 'don't write to disk',
just controls on reading/writing pages. I'm not exactly sure how
useful having that would be, however, since if you assume the person
can read swap, why not just assume he can hack the kernel? Presumably,
after a crash, the swap file/partition may still contain the
interesting data, but then, it's also possible to halt the machine
and just examine memory regions.
This whole discussion of trojan horses can get really hairy if
you don't presume some level of security somewhere. A user
can't seriously be expected to verify every single component of his
system.
----------
NAME
mprotect - control the protection of pages
SYNOPSIS
#include <sys/types.h>
#include <sys/mman.h>
int
mprotect(void *addr, size_t len, int prot);
DESCRIPTION
The mprotect() system call changes the specified pages to have protection
prot. Not all implementations will guarantee protection on a page basis;
the granularity of protection changes may be as large as an entire re-
gion.
SEE ALSO
madvise(2), mincore(2), msync(2), munmap(2)
HISTORY
The mprotect() function first appeared in 4.4BSD.
4.4BSD June 9, 1993 1
------------------------------------------------------------
| Yoav Yerushalmi | My opinions are mine.. |
| M.I.T student at large | (I hope.. don't sue) |
| http://www.mit.edu/people/yoav/homepage.html |
------------------------------------------------------------
