[Cryptography] randomness +- entropy

[Bear]

On Tue, 2013-11-12 at 17:54 -0800, Bill Stewart wrote:
> At 12:23 AM 11/12/2013, James A. Donald wrote:
> >On 2013-11-12 16:44, John Denker wrote:
> >>The fact is, there are some applications that cannot make do with
> >>low-quality randomness *and* cannot afford to wait.
> >
> >I don't think so.
> 
> Most applications can wait.  Some of them could wait, but currently don't.
> But what applications are there that really do need to run early?
> 
> The one potential example I can think of is hard drive encryption -
> it definitely needs good (pseudo)randomness,
> and needs to start pretty early in the boot process
> so other applications can have a file system to write to,

I think I'm not buying it.  Hard drive encryption doesn't 
need *randomness* early in the boot process; it needs *A KEY* 
early in the boot process.  

A machine with an encrypted hard drive has to be able to read
and write sectors encrypted with an existing key before boot 
can proceed.  IMO that means it either halts during boot and 
the BIOS asks for someone to type in the passkey (the option 
I'd prefer on a "secure" machine) or it has the key stored 
unencrypted somewhere (obviously less secure but probably more 
manageable).  Randomness for keying material is needed when 
creating a *new* key, but does not help us get the existing 
key/s we need to read the boot sector.  

Why would an encrypted drive really need a *new* key during 
bootup?  

				Bear