Re: [TLS] TLSrenego - possibilities, suggestion for SSLv3

[Yoav Nir <[e-mail]>]

I disagree that it's a waste to repurpose the entire 32-bit value.  Sure we could use only the top two bits (00 for initial, 01 for reneg - it's after 2004, so all times have MSB=1), but then you can't tell the difference between a patched TLS implementation and an implementation with its system clock not set properly.

A 32-bit magic value (the date of the first chartering of the TLS WG and the date of the vulnerability being published?) will reduce the likelihood of error to negligible. Two bits won't


________________________________________
From: [e-mail] [[e-mail]] On Behalf Of Martin Rex [[e-mail]]
Sent: Wednesday, November 11, 2009 22:15
To: Marsh Ray
Cc: [e-mail]
Subject: Re: [TLS] TLSrenego - possibilities, suggestion for SSLv3

Marsh Ray wrote:
>
> Yair Elharrar wrote:
> > Martin Rex wrote:
> >
> > It would probably be better if the entire 32-bit gmt_unix_time was
> > set to one of two predefined magic values:
>
> Do you know that nothing cares about the actual time encoded?
>
> Why burn 32 bits of entropy to represent one bit?

Yair's suggestion is actually 3 states (~1.5 bits)
   initial, renegotiation, unknown

but other than that, it also seems like waste to me to re-purpose
the _entire_ 32-bits.