Re: final repost - MPPE incorrect REJECT/NAK behaviour (was Re: Windows

[Phil Mayers <p.mayers@imperial.ac.uk>]

James Cameron wrote:
> Looked fine to me, and so I began to test it, but it would not apply to
> PPP CVS, which appears to contain code that does something similar.
> 
> Is it already fixed in CVS?  Could you try CVS?  Or adapt your change to
> what is already there?

Oops. How embarrassing! Apologies for not checking CVS and for the noise.

I'll test it, but I guess that should resolve the issue - though the 
patch I wrote only NAKs with the preferred crypto bit which is what I 
observed win2k to do, and at a glance that code will NAK with all bits 
we support (i.e. 40 and 128 set, if you permit that)? For reference:

client sends:

Generic Routing Encapsulation (PPP)
  Protocol Type: PPP (0x880b)
Point-to-Point Protocol
  Protocol: Compression Control Protocol (0x80fd)
PPP Compression Control Protocol
  Code: Configuration Request (0x01)
  Identifier: 0x00
  Length: 10
  Options: (6 bytes)
   Microsoft PPC: Supported Bits: 0x00000001
    .... .... .... .... .... .... .... ...1 = Desire to negotiate MPPC
    .... .... .... .... .... .... ...0 .... = Obsolete
    .... .... .... .... .... .... ..0. .... = 40-bit encryption OFF
    .... .... .... .... .... .... .0.. .... = 128-bit encryption OFF
    .... .... .... .... .... .... 0... .... = 56-bit encryption OFF
    .... ...0 .... .... .... .... .... .... = Stateless mode OFF

server replies (server does permit 40 bit):

Generic Routing Encapsulation (PPP)
  Protocol Type: PPP (0x880b)
Point-to-Point Protocol
  Protocol: Compression Control Protocol (0x80fd)
PPP Compression Control Protocol
  Code: Configuration Nak (0x03)
  Identifier: 0x00
  Length: 10
  Options: (6 bytes)
   Microsoft PPC: Supported Bits: 0x00000041
    .... .... .... .... .... .... .... ...1 = Desire to negotiate MPPC
    .... .... .... .... .... .... ...0 .... = Obsolete
    .... .... .... .... .... .... ..0. .... = 40-bit encryption OFF
    .... .... .... .... .... .... .1.. .... = 128-bit encryption ON
    .... .... .... .... .... .... 0... .... = 56-bit encryption OFF
    .... ...0 .... .... .... .... .... .... = Stateless mode OFF


Having said that, I agree the version in CVS is "more correct", you 
should NAK with what you can permit.

Thanks again,
Phil