On Sun, 2005-11-13 at 11:14 +0200, Oded Arbel wrote:
> On Saturday, 12 בNovember 2005 01:32, Gilboa Davara wrote:
> > > About buffer overflow: you are missing the point. You are not
> > > overflowing the host stack, but the VMs one. This actually is good
> > > thing from that point of view.
> >
> > Um.... and once I did that, what prevents me from generating a code
> > that will cause the vm to delete the c:\boot.ini file?
> 
> I'm not sure what the MS guys are doing, but if I were them I would run 
> the networking code with no file system permissions. The up side of 
> running everything in a VM(*) is that you don't have to link in file 
> system operations for a program that isn't supposed to use them, so 
> even if you buffer overflow the program you can't cause it to do stuff 
> that it isn't supposed to be doing.
> 
> (*) the .Net VM isn't really a VM. Its more like a virtual virtual 
> machine (the virtual machine itself is virtual) - the .Net spec call 
> for everything to be JITed and cached.
> 

Which means they are trying to mimic selinux without calling it selinux.

As for the .NET VM, I'll bow down and call it a "very extensive and
super heavy RT library that slows things down almost to the point of
JVM"... How's that?

Gilboa



=================================================================
To unsubscribe, send mail to [EMAIL PROTECTED] with
the word "unsubscribe" in the message body, e.g., run the command
echo unsubscribe | mail [EMAIL PROTECTED]

Reply via email to