Jared L CTR USAF AFMC 46 SK/CCI <jared.ctr@eg...> - 2008-12-15 18:01:16
Smartcard support in rdesktop is suboptimal. Alexey Guzev's patch
enabled smartcard login from at least 32-bit PCs running Linux; mine of
a year ago added 64-bit PC/Linux support. But it still doesn't work from
Macs (neither from PowerPC nor Intel Macs), and even where smartcard
login works, use of the smart card within the Windows session after
login frequently fails.
Furthermore, as someone has noted in the tracker on SourceForge, the
smartcard support in rdesktop does not work with pcscd versions 1.4.100
and later (1.4.100 came out five or six months ago; 1.5.0 came out a few
There exists a publicly readable RDP smartcard extension protocol
<http://msdn.microsoft.com/en-us/library/cc242596(PROT.10).aspx>. In its
preamble it says that the protocol "may be covered by Microsoft's Open
Specifications Promise;" it says (RDPESC section 1.4) that it's an
extension of the filesystem virtual channel extension, which (RDPEFS
section 1.4) is embedded in a static virtual channel specified by basic
RDP, which (RDPBCGR section 1.4) is "based on the ITU T.120 series of
protocols." The Open Specifications Promise says that it covers "T.120."
0. Being as the promise covers T.120 but says nothing about the various
extensions which have been made to it, and given rdesktop's past
relationship with Microsoft (of which I don't really know anything),
how big is the intellectual property problem already for rdesktop?
1. Neither reading the RDPESC document nor guesssing at the protocol by
reading network traffic grants the implementor a license of whatever
software patents may cover the implementation; is there then a legal or
ethical difference between the two?
2. If someone were to try to update the smartcard support, and consulted
the RDPESC protocol document in doing so, is it likely that the
resultant work would be unwise to fold into rdesktop?