TrustedBSD Extensions Project

[David Collier-Brown - Sun Canada]

David Collier-Brown - Sun Canada <davecb at scot.canada.sun.com>
| > 	Actually the write never gets to the driver, having been
| > 	caught up in the permissions module that all the open
| > 	operations have to pass through.
 
stanislav shalunov <shalunov at att.com>| No.  You ask to transfer this data to 
SCSI ID 3, block 45467.
| The controller decides to write to SCSI ID 1, block 45467.

	That's interesting... if it was intentional, it would be a
	covert channel. [If it were unintentional and probabalistic, 
	it would crash your filesystem eventually, and so get caught
	by QC or an enraged user.]
	
	I think it would be considered a bug, though, and so not be
	specified in particular...
	

| > 	H-P sells their firewalls hosted on a B1 OS, and 
| 
| Why would I buy anything from HP?  

	Well, performance, price and reliability, for three (;-))
	(And no, I don't work for H-P: rather the opposite, in fact)


| I'm not talking about Solaris.  I'm talking about just one possible
| way to implement the (silly) requirement.

	With respect, that's a silly implementation, not a silly 
	requirement.
	
--dave
--
David Collier-Brown in Boston
Phone: (781) 442-0734, Room BUR03-3632

To Unsubscribe: send mail to majordomo at trustedbsd.org
with "unsubscribe trustedbsd-discuss" in the body of the message