[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Bug#765632: ForwardX11Trusted set to yes over a decade ago, for release reasons?



On Wed, 2015-08-19 at 20:01 -0700, Nikolaus Rath wrote:
> Until now, I did not know how much trust I'm actually putting into
> the
> remote server when using -X (on Debian). I'll probably continue to 
> use
> it in the majority of cases (because the alternatively seems rather
> useless), but in my opinion it would be great if it could be somehow
> communicated to the user what -X really implies.

Hmm the best thing would be if one could make X forwarding actually
secure.
I once proposed the idea[0] to e.g. use something like Xephyr as
destination for any forwardings; probably at least one instance per
host/user, possibly even per connection.

The idea was that the client automatically spawns Xephyr as necessary
(with options that e.g. forbid fullscreen, focus stealing, clipboard
stealing, keylogging and that like).

But at least to me it's not even clear whether Xephyr would really add
security so that my idea works, or whether it basically just passes
everything on (as X protocol) to the actual X server.
So perhaps one would need something else,... like VNC... X forwarding
drawn to jpegs ;-)


> I believe sudo prints an extensive warning on first invocation (and 
> uses
> a flag file  ~/.sudo_as_admin_successfull to be less verbose after
> that) - maybe ssh could do something similar?

If, then something like this should go upstream... and they may argue
that they ship OpenSSH anyway with safe defaults.


Cheers,
Chris.


[0] https://bugzilla.mindrot.org/show_bug.cgi?id=1926

Attachment: smime.p7s
Description: S/MIME cryptographic signature


Reply to: