xref: /freebsd/contrib/pam-krb5/TODO (revision bf6873c5786e333d679a7838d28812febf479a8a)
1                           pam-krb5 To-Do List
2
3PAM API:
4
5 * Support PAM_CHANGE_EXPIRED_AUTHTOK properly in pam_chauthtok.  This
6   will require prompting for the current password (if it's not already
7   available in the PAM data) and trying a regular authentication first to
8   see if the account is expired.
9
10 * Tighter verification that all of our flags are valid might be a good
11   idea.
12
13 * For informational messages followed by a prompt, find a way to combine
14   these into one PAM conversation call for better GUI presentation
15   behavior.
16
17Functionality:
18
19 * Change the authentication flow so that both Heimdal and MIT use the
20   same logic for attempting PKINIT first and then falling back to
21   password.  This will fix failure to store passwords in the PAM data
22   with try_pkinit and MIT Kerberos on password fallback and will allow
23   implementation of use_pkinit for MIT.  Based on discussion with MIT
24   Kerberos upstream, the best approach is probably to configure a custom
25   prompter that refuses to reply to any prompt.
26
27 * Add a daemon that can be used to verify TGTs that can be used when
28   pam-krb5 is run as a non-root user and hence doesn't have access to the
29   system keytab.  Jeff Hutzelman has a daemon and protocol for doing this
30   developed for a different PAM authentication module, and it would be
31   good to stay consistent with that protocol if possible.  (Debian
32   Bug#399001)
33
34 * The alt_auth_map parsing to find realms doesn't take into account
35   escaped @-signs and doesn't do proper principal parsing.
36
37 * Fix password expiration handling for the search_k5login and
38   alt_auth_map cases.  Right now, we may return expired password errors
39   that would trigger password expiration handling, which probably isn't
40   correct.
41
42 * Support authentication from a keytab.
43
44 * Support disabling of user canonicalization so that the PAM user is
45   retained even if the module did an aname to lname mapping.
46
47 * Use set_out_ccache to write the resulting ticket cache, if it is
48   available.  This ensures the correct flags are set in the ticket cache.
49   This poses some challenges due to the two-step ticket cache mechanism
50   currently used.  Perhaps there's a cache copying API?
51
52 * Use krb5_chpw_message to parse password change messages from Active
53   Directory.
54
55 * Consider exposing the Kerberos principal in the password prompt for a
56   password change.  (Debian Bug#667928)
57
58Code Cleanup:
59
60 * The PKINIT code for Heimdal involves too many #ifdefs right now for my
61   taste.  Find a way to restructure it to only wrap the main PKINIT
62   function for Heimdal.
63
64 * The current handling of error return codes is a mess.  We need to find
65   a way to return a rich set of error codes from the underlying functions
66   and then map error codes appropriately in the interface functions.
67   Helpful for this would be improved documentation of what error codes
68   are permitted and where.
69
70 * Tracking when to free the Kerberos context and other things stored in
71   the PAM context is currently too complicated.  It should be possible to
72   simplify it with a reference counting scheme.
73
74Documentation:
75
76 * Document PKINIT configuration with MIT in krb5.conf.  It looks like the
77   library supports configuration in [realms] with similar names to the
78   PAM module configuration.
79
80Portability:
81
82 * If pam_modutil_getpwnam is not available but getpwnam_r is, roll our
83   own using getpwnam_r.
84
85Logging:
86
87 * Log the information that the Kerberos library asks us to display, or at
88   least the info and error messages.
89
90 * Log unknown PAM flags on module entry.  Currently, only the symbolic
91   flags we know about will be logged.
92
93Test suite:
94
95 * Ensure that the test suite covers all possible PAM options.
96
97 * Figure out why the pin-mit script for module/pkinit prompts twice and
98   check if it's a bug in the module.
99
100 * Find a way of testing the PKINIT identity selection for MIT Kerberos
101   with use_pkinit enabled.
102