On (13/06/16 16:22), Jakub Hrozek wrote:
On Mon, Jun 13, 2016 at 03:16:25PM +0200, Lukas Slebodnik wrote:
> On (29/02/16 11:17), Jakub Hrozek wrote:
> >On Thu, Feb 25, 2016 at 01:37:27PM +0100, Sumit Bose wrote:
> >> On Thu, Feb 25, 2016 at 12:50:55PM +0100, Jakub Hrozek wrote:
> >> > On Tue, Feb 23, 2016 at 12:53:25PM +0100, Sumit Bose wrote:
> >> > > Hi,
> >> > >
> >> > > this patch fixes and issue during initgroups in AD forests. Please
see
> >> > > the commit message for details.
> >> > >
> >> > > To reproduce this you can create a new user outside of CN=Users on
the
> >> > > forest root. The new user can be created in an existing container
or in
> >> > > a new OU container. Most important is that it is not a child of
> >> > > CN=Users. In a child domain (it must be a child, domains with a
> >> > > different base won't trigger the issue) create a user with the
same
> >> > > name. With this setup 'id user(a)forest.root' will not
return the complete
> >> > > list of group the user is a member of and the patch should fix
this.
> >> > >
> >> > > bye,
> >> > > Sumit
> >> > >
> >> >
> >> > Hi,
> >> >
> >> > the patch works fine, tested with a user named the same in the same OU
> >> > in different domains. Coverity found some warnings:
> >> >
> >> > Error: COMPILER_WARNING:
> >> > sssd-1.13.90/src/providers/ldap/sdap_async_initgroups.c:2835:12:
warning: unused variable 'dn_len' [-Wunused-variable]
> >> > # size_t dn_len;
> >> > # ^
> >> > # 2833| const char *cname;
> >> > # 2834| bool in_transaction = false;
> >> > # 2835|-> size_t dn_len;
> >> > # 2836| size_t c = 0;
> >> > # 2837|
> >> >
> >> > Error: COMPILER_WARNING:
> >> > sssd-1.13.90/src/providers/ldap/sdap_async_initgroups.c: scope_hint: In
function 'sdap_get_initgr_user'
> >> > sssd-1.13.90/src/providers/ldap/sdap_async_initgroups.c:2836:12:
warning: unused variable 'c' [-Wunused-variable]
> >> > # size_t c = 0;
> >> > # ^
> >> > # 2834| bool in_transaction = false;
> >> > # 2835| size_t dn_len;
> >> > # 2836|-> size_t c = 0;
> >> > # 2837|
> >> > # 2838| DEBUG(SSSDBG_TRACE_ALL, "Receiving info for the
user\n");
> >> >
> >> > Error: COMPILER_WARNING:
> >> > sssd-1.13.90/src/db/sysdb_subdomains.c:23: included_from: Included from
here.
> >> > sssd-1.13.90/src/db/sysdb_subdomains.c: scope_hint: In function
'try_to_find_expected_dn'
> >> > sssd-1.13.90/src/util/util.h:144:9: warning: 'result_dn_str'
may be used uninitialized in this function [-Wmaybe-uninitialized]
> >> > # sss_debug_fn(__FILE__, __LINE__, __FUNCTION__, \
> >> > # ^
> >> > sssd-1.13.90/src/db/sysdb_subdomains.c:1075:17: note:
'result_dn_str' was declared here
> >> > # const char *result_dn_str;
> >> > # ^
> >> > # 142| int __debug_macro_level = level; \
> >> > # 143| if (DEBUG_IS_SET(__debug_macro_level)) { \
> >> > # 144|-> sss_debug_fn(__FILE__, __LINE__, __FUNCTION__, \
> >> > # 145| __debug_macro_level, \
> >> > # 146| format, ##__VA_ARGS__); \
> >> >
> >> > Additionally, I wonder if we should namespace the function?
> >>
> >> Thank you for the review, new version attached.
> >>
> >> bye,
> >> Sumit
> >
> >ACK.
> >
> >There were some downstream tests failing, but the same tests kept
> >failing even with a vanilla RHEL-7.2 package, so I think it's a fluke in
> >the test and not a regression.
> Was it a ldap_rfc2307_fallback_to_local_users related test?
>
> Because I finally get to state when I was able to look
> into failing tests and git bisect found out that
> this patch (5ff7a765434ed0b4d37564ade26d7761d06f81c3)
> broke option ldap_rfc2307_fallback_to_local_users
I really don't remember, but I think I only ran AD tests, not LDAP
tests.
>
> ::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::
> :: [ LOG ] :: fallback_to_local_users_002: bz948263 Value set to true
> ::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::
>
> :: [ PASS ] :: File '/var/log/sssd/sssd_LDAP.log' should contain
'Option ldap_rfc2307_fallback_to_local_users is TRUE'
> :: [ PASS ] :: Command 'getent group ldap_group | grep local_user | grep
ldap_user' (Expected 0, got 0)
> :: [ FAIL ] :: Command 'id local_user | grep ldap_group' (Expected 0, got
1)
> :: [ PASS ] :: Command 'id non_existant_user' (Expected 1, got 1)
> :: [ PASS ] :: Command 'getent group ldap_group | grep local_user | grep
ldap_user' (Expected 0, got 0)
> :: [ FAIL ] :: Command 'ldbsearch -H /var/lib/sss/db/cache_LDAP.ldb
name=ldap_group member | grep local_user' (Expected 0, got 1)
> :: [ FAIL ] :: Command 'ldbsearch -H /var/lib/sss/db/cache_LDAP.ldb
name=local_user memberof | grep ldap_group' (Expected 0, got 1)
> :: [ PASS ] :: Command 'ldbsearch -H /var/lib/sss/db/cache_LDAP.ldb
name=ldap_group dataExpireTimestamp | grep "dataExpireTimestamp: 1"'
(Expected 0, got 0)
> :: [ FAIL ] :: Command 'ldbsearch -H /var/lib/sss/db/cache_LDAP.ldb
name=local_user dataExpireTimestamp | grep "dataExpireTimestamp: 1"'
(Expected 0, got 1)
> :: [ FAIL ] :: Command 'id local_user | grep ldap_group' (Expected 0, got
1)
> :: [ PASS ] :: Command 'getent group ldap_group | grep local_user | grep
ldap_user' (Expected 0, got 0)
> :: [ LOG ] :: Duration: 16s
> :: [ LOG ] :: Assertions: 6 good, 5 bad
> :: [ FAIL ] :: RESULT: fallback_to_local_users_002: bz948263 Value set to true
>
> I can provide access to machine if you need a reproducer.
> But you should be able reproduce yourself with minimal rfc2307 config
>
> [sssd]
> config_file_version = 2
> services = nss, pam
> domains = LDAP
>
> [domain/LDAP]
> debug_level = 0xFFF0
> id_provider = ldap
> ldap_uri = ldap://$SERVER
> ldap_search_base = $DS_BASE_DN
> ldap_rfc2307_fallback_to_local_users = True
Please file a bug.
https://fedorahosted.org/sssd/ticket/3045
git bisect confirmed problematic commit also in 1.13 branch
So it's not a coincidence :-)
LS