[Cryptography] RISC-V branch predicting

Bill Frantz frantz at pwpconsult.com
Tue Feb 13 00:10:08 EST 2018


On 2/10/18 at 12:32 PM, leichter at lrw.com (Jerry Leichter) wrote:

>It turns out that the Unix approach - in which process creation 
>is assumed to be very inexpensive - is probably better than the 
>approach of other OS's, where processes are more expensive to 
>create, thus longer-lived and more likely to be subdivided into 
>software-enforced security domains.  If even Unix processes are 
>too expensive - which will likely be the response of browser 
>makers to the notion that each individual piece of Javascript 
>should be segmented off into its own process - then perhaps we 
>should look at hardware and software models of very cheap 
>hardware isolation.

There are two worked examples of using Unix processes for 
isolation, Postfix and Qmail. In both cases, the solution to 
mail server vulnerability has been addressed by isolating the 
parts of the server which parse user input.

One security rule I was able to articulate after reading 
Bernstein's Qmail paper 
<http://cr.yp.to/qmail/qmailsec-20071101.pdf> is: Give the parts 
of the program most likely to have bugs the fewest possible 
privileges. Bernstein reduces the privileges of some code to 
reading it's input stream and writing it's output stream. I say, 
any part of the program which parses arbitrary strings is a good candidate.

This kind of isolation is a bit klunky in Unix. It is a lot 
easier in some object oriented languages, but using languages to 
implement isolation requires including the language runtime in 
the security kernel which, in practice, defeats most of the 
benefit. Capability operating systems give OS levels of 
assurance with the ease of use of object languages. 
Unfortunately, these OSes have been experimental for the last 35 
or so years.

Cheers - Bill

---------------------------------------------------------------------------
Bill Frantz        |"After all, if the conventional wisdom was 
working, the
408-356-8506       | rate of systems being compromised would be 
going down,
www.pwpconsult.com | wouldn't it?" -- Marcus Ranum



More information about the cryptography mailing list