Re: cant login after make installworld: pam_opie.so.6 not found

2023-01-06 Thread Ivan Quitschal

Hi Dag-Erling

no problem at all , i just wanted to pass thru the login thats it.
its working fine with this new pam_login_access.so / pam_unix.so (not sure which 
one we are using now)


thanks

--tzk


On Fri, 6 Jan 2023, Dag-Erling Smørgrav wrote:


Ivan Quitschal  writes:

I had to "install" the new temporary /etc/pam.d files with mergemaster. I think 
the problematic file was
the "system" config file. Now its all fine


If you wish to continue to use OPIE (which I don't recommend), there is
now a port (security/opie).  You will have to manually add back the
pam_opie and pam_opieaccess lines in the relevant PAM policies.

DES
--
Dag-Erling Smørgrav - d...@freebsd.org


Re: cant login after make installworld: pam_opie.so.6 not found

2023-01-06 Thread Dag-Erling Smørgrav
Ivan Quitschal  writes:
> I had to "install" the new temporary /etc/pam.d files with mergemaster. I 
> think the problematic file was
> the "system" config file. Now its all fine

If you wish to continue to use OPIE (which I don't recommend), there is
now a port (security/opie).  You will have to manually add back the
pam_opie and pam_opieaccess lines in the relevant PAM policies.

DES
-- 
Dag-Erling Smørgrav - d...@freebsd.org



Re: cant login after make installworld: pam_opie.so.6 not found

2023-01-06 Thread grarpamp
On 1/6/23, Xin Li  wrote:
> Security team has discussed this a decade ago.  See
> https://www.miknet.net/security/skey-dungeon-attack/
> for technical details.

That would mean that FreeBSD knowingly left users exploitable without
doing even the "easy fix" in that article to the opie code for over a decade.
And further left opie vulnerable and present since the commit in all RELENG,
STABLE, and handbook. And did not issue a SA on it since the commit, nor
ever since the article. If attempting to claim security as reason to delete,
then FreeBSD might appear to be faulty of this. Which would present good
opportunity to consider any potential improvements to that process too.

> And this could have been avoided if user have followed source upgrade

Lockout avoided... yes maybe if users wanted to quit their opie forever
at that moment, but if not, then opie code module hasn't yet been
moved to ports for anyone to use and or update as they wish.
The nature of port security in every unix OS is 3rd-party and un-dedicated,
so that wouldn't be reason not to port such things either.

Onward :)