Re: [PATCHES] Solaris ident authentication using unix domain sockets
От
Florian G. Pflug
Тема
Re: [PATCHES] Solaris ident authentication using unix domain
sockets
Дата
Msg-id
4873BC28.3070507@phlo.org
Ответ на
Список
Дерево обсуждения
Solaris ident authentication using unix domain sockets Garick Hamlin <ghamlin@isc.upenn.edu>
Re: Solaris ident authentication using unix domain sockets Peter Eisentraut <peter_e@gmx.net>
Re: Solaris ident authentication using unix domain sockets Tom Lane <tgl@sss.pgh.pa.us>
Re: Solaris ident authentication using unix domain
sockets Garick Hamlin <ghamlin@isc.upenn.edu>
Re: Solaris ident authentication using unix domain
sockets Bruce Momjian <bruce@momjian.us>
Re: Solaris ident authentication using unix domain sockets Robert Treat <xzilla@users.sourceforge.net>
Re: [PATCHES] Solaris ident authentication using unix domain
sockets Andrew Dunstan <andrew@dunslane.net>
Re: [PATCHES] Solaris ident authentication using unix domain sockets Tom Lane <tgl@sss.pgh.pa.us>
Re: [PATCHES] Solaris ident authentication using unix domain sockets Josh Berkus <josh@agliodbs.com>
Re: [PATCHES] Solaris ident authentication using unix domain
sockets Andrew Dunstan <andrew@dunslane.net>
Re: [PATCHES] Solaris ident authentication using unix domain
sockets "Florian G. Pflug" <fgp@phlo.org>
Re: [PATCHES] Solaris ident authentication using unix domain
sockets Josh Berkus <josh@agliodbs.com>
Josh Berkus wrote: > Tom, > >> Indeed. If the Solaris folk feel that getupeercred() is insecure, >> they had better explain why their kernel is that broken. This is >> entirely unrelated to the known shortcomings of the "ident" IP >> protocol. > > The Solaris security & kernel folks do, actually. However, there's > no question that TRUST is inherently insecure, and that's what people > are going to use if they can't get IDENT to work. I'd be *very* interested in how they come to that assessment. I'd have thought that the only alternative to getpeereid/getupeercred is password-based or certificate-based authenticated - which seem *less* secure because a) they also rely on the client having the correct uid or gid (to read the password/private key), plus b) the risk of the password/private key getting into the wrong hands. How is that sort of authenticated handled by services shipping with solaris? regards, Florian Pflug, hoping to be enlightened beyond his limited posix-ish view of the world...
В списке pgsql-hackers по дате отправления