CVE-2022-38023 s4:rpc_server/netlogon: split out dcesrv_netr_check_schannel() function

This will allow us to reuse the function in other places.
As it will also get some additional checks soon.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15240

Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
(cherry picked from commit f43dc4f0bd60d4e127b714565147f82435aa4f07)

CVE-2022-38023 selftest:Samba4: avoid global 'allow nt4 crypto = yes' and 'reject md5 clients = no'

Instead of using the generic deprecated option use the specific
allow nt4 crypto:COMPUTERACCOUNT = yes and
server reject md5 schannel:COMPUTERACCOUNT = no
in order to allow legacy tests for pass.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15240

Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
(backported from commit 7ae3735810c2db32fa50f309f8af3c76ffa29768)

CVE-2022-38023 s4:rpc_server/netlogon: debug 'reject md5 servers' and 'allow nt4 crypto' misconfigurations

This allows the admin to notice what's wrong in order to adjust the
configuration if required.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15240

Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
(cherry picked from commit 43df4be35950f491864ae8ada05d51b42a556381)

CVE-2022-38023 docs-xml/smbdotconf: document "server reject md5 schannel:COMPUTERACCOUNT"

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15240

Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
(cherry picked from commit 2ad302b42254e3c2800aaf11669fe2e6d55fa8a1)

CVE-2022-38023 docs-xml/smbdotconf: document "allow nt4 crypto:COMPUTERACCOUNT = no"

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15240

Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
(cherry picked from commit bd429d025981b445bf63935063e8e302bfab3f9b)

CVE-2022-38023 s4:rpc_server/netlogon: add 'server reject md5 schannel:COMPUTERACCOUNT = no' and 'allow nt4 crypto:COMPUTERACCOUNT = yes'

This makes it more flexible when we change the global default to
'reject md5 servers = yes'.

'allow nt4 crypto = no' is already the default.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15240

Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
(cherry picked from commit 69b36541606d7064de9648cd54b35adfdf8f0e8f)

CVE-2022-38023 s4:rpc_server/netlogon: defer downgrade check until we found the account in our SAM

We'll soon make it possible to use 'reject md5 servers:CLIENTACCOUNT$ = no',
which means we'll need use the account name from our SAM.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15240

Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
(cherry picked from commit b09f51eefc311bbb1525efd1dc7b9a837f7ec3c2)

CVE-2022-38023 docs-xml/smbdotconf: change 'reject md5 clients' default to yes

AES is supported by Windows Server >= 2008R2, Windows (Client) >= 7 and Samba >= 4.0,
so there's no reason to allow md5 clients by default.
However some third party domain members may need it.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15240

Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
(cherry picked from commit c8e53394b98b128ed460a6111faf05dfbad980d1)

CVE-2022-38023 s4:rpc_server/netlogon: split out dcesrv_netr_ServerAuthenticate3_check_downgrade()

We'll soon make it possible to use 'reject md5 servers:CLIENTACCOUNT$ = no',
which means we'll need the downgrade detection in more places.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15240

Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
(cherry picked from commit b6339fd1dcbe903e73efeea074ab0bd04ef83561)

CVE-2022-38023 s4:torture: use NETLOGON_NEG_SUPPORTS_AES by default

For generic tests we should use the best available features.

And AES will be required by default soon.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15240

Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
(backported from commit cfd55a22cda113fbb2bfa373b54091dde1ea6e66)

CVE-2022-38023 selftest:Samba4: avoid global 'server schannel = auto'

Instead of using the generic deprecated option use the specific
server require schannel:COMPUTERACCOUNT = no in order to allow
legacy tests for pass.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15240

Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
(cherry picked from commit 63c96ea6c02981795e67336401143f2a8836992c)

CVE-2022-38023 s4:rpc_server/netlogon: improve CVE-2020-1472(ZeroLogon) debug messages

In order to avoid generating useless debug messages during make test,
we will use 'CVE_2020_1472:warn_about_unused_debug_level = 3'
and 'CVE_2020_1472:error_debug_level = 2' in order to avoid schannel warnings.

Review with: git show -w

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15240

Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
(cherry picked from commit 16ee03efc194d9c1c2c746f63236b977a419918d)

CVE-2022-38023 s4:rpc_server/netlogon: re-order checking in dcesrv_netr_creds_server_step_check()

This will simplify the following changes.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15240

Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
(cherry picked from commit ec62151a2fb49ecbeaa3bf924f49a956832b735e)

CVE-2022-38023 s4:rpc_server/netlogon: add talloc_stackframe() to dcesrv_netr_creds_server_step_check()

This will simplify the following changes.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15240

Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
(cherry picked from commit 0e6a2ba83ef1be3c6a0f5514c21395121621a145)

CVE-2022-38023 s4:rpc_server/netlogon: add a lp_ctx variable to dcesrv_netr_creds_server_step_check()

This will simplify the following changes.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15240

Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
(cherry picked from commit 7baabbe9819cd5a2714e7ea4e57a0c23062c0150)

CVE-2022-38023 s4:rpc_server/netlogon: 'server schannel != yes' warning to dcesrv_interface_netlogon_bind

This will simplify the following changes.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15240

Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
(cherry picked from commit e060ea5b3edbe3cba492062c9605f88fae212ee0)

CVE-2022-38023 docs-xml/smbdotconf: change 'reject md5 servers' default to yes

AES is supported by Windows >= 2008R2 and Samba >= 4.0 so there's no
reason to allow md5 servers by default.

Note the change in netlogon_creds_cli_context_global() is only cosmetic,
but avoids confusion while reading the code. Check with:

 git show -U35 libcli/auth/netlogon_creds_cli.c

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15240

Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
(cherry picked from commit 1c6c1129905d0c7a60018e7bf0f17a0fd198a584)

CVE-2022-38023 s3:winbindd: also allow per domain "winbind sealed pipes:DOMAIN" and "require strong key:DOMAIN"

This avoids advising insecure defaults for the global options.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15240

Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
(backported from commit d60828f6391307a59abaa02b72b6a8acf66b2fef)

CVE-2022-38023 s3:net: add and use net_warn_member_options() helper

This makes sure domain member related 'net' commands print warnings
about unsecure smb.conf options.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15240

Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
(backported from commit 1fdf1d55a5dd550bdb16d037b5dc995c33c1a67a)

CVE-2022-38023 libcli/auth: add/use netlogon_creds_cli_warn_options()

This warns the admin about insecure options

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15240

Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
(similar to commit 7e7adf86e59e8a673fbe87de46cef0d62221e800)
[jsutton@samba.org Replaced call to tevent_cached_getpid() with one to
 getpid()]

(cherry picked from commit 1fe8857b4d90809ff5b160c73e3d2c28acd3ce65)

CVE-2022-38023 libcli/auth: pass lp_ctx to netlogon_creds_cli_set_global_db()

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15240

Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
(cherry picked from commit 992f39a2c8a58301ceeb965f401e29cd64c5a209)

CVE-2022-38023 docs-xml: improve wording for several options: "yields precedence" -> "is over-riden"

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15240

Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
(cherry picked from commit 830e865ba5648f6520bc552ffd71b61f754b8251)

CVE-2022-38023 docs-xml: improve wording for several options: "takes precedence" -> "overrides"

BUG: https://bugzilla.samba.org/show_bug.cgi?id=15240

Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
(backported from commit 8ec62694a94c346e6ba8f3144a417c9984a1c8b9)

CVE-2020-1472(ZeroLogon): s4 torture rpc: repeated bytes in client challenge

Ensure that client challenges with the first 5 bytes identical are
rejected.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14497

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
[abartlet@samba.org: backported from master as test order was flipped]

CVE-2020-1472(ZeroLogon): s4 torture rpc: Test empty machine acct pwd

Ensure that an empty machine account password can't be set by
netr_ServerPasswordSet2

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14497

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>

CVE-2020-1472(ZeroLogon): docs-xml: document 'server require schannel:COMPUTERACCOUNT'

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14497

Signed-off-by: Stefan Metzmacher <metze@samba.org>

CVE-2020-1472(ZeroLogon): s3:rpc_server/netlogon: log warnings about unsecure configurations

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14497

Pair-Programmed-With: Stefan Metzmacher <metze@samba.org>

Signed-off-by: Günther Deschner <gd@samba.org>
Signed-off-by: Stefan Metzmacher <metze@samba.org>

CVE-2020-1472(ZeroLogon): s3:rpc_server/netlogon: support "server require schannel:WORKSTATION$ = no"

This allows to add expections for individual workstations, when using "server schannel = yes".
"server schannel = auto" is very insecure and will be removed soon.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14497

Pair-Programmed-With: Stefan Metzmacher <metze@samba.org>

Signed-off-by: Günther Deschner <gd@samba.org>
Signed-off-by: Stefan Metzmacher <metze@samba.org>

CVE-2020-1472(ZeroLogon): s3:rpc_server/netlogon: refactor dcesrv_netr_creds_server_step_check()

We should debug more details about the failing request.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14497

Pair-Programmed-With: Stefan Metzmacher <metze@samba.org>

Signed-off-by: Günther Deschner <gd@samba.org>
Signed-off-by: Stefan Metzmacher <metze@samba.org>

CVE-2020-1472(ZeroLogon): s4:rpc_server/netlogon: log warnings about unsecure configurations

This should give admins wawrnings until they have a secure
configuration.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14497

Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Günther Deschner <gd@samba.org>

CVE-2020-1472(ZeroLogon): s4:rpc_server/netlogon: support "server require schannel:WORKSTATION$ = no"

This allows to add expections for individual workstations, when using "server schannel = yes".
"server schannel = auto" is very insecure and will be removed soon.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14497

Signed-off-by: Stefan Metzmacher <metze@samba.org>

CVE-2020-1472(ZeroLogon): s4:rpc_server/netlogon: refactor dcesrv_netr_creds_server_step_check()

We should debug more details about the failing request.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14497

Signed-off-by: Stefan Metzmacher <metze@samba.org>

CVE-2020-1472(ZeroLogon): s3:rpc_server/netlogon: protect netr_ServerPasswordSet2 against unencrypted passwords

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14497

Pair-Programmed-With: Stefan Metzmacher <metze@samba.org>

Signed-off-by: Jeremy Allison <jra@samba.org>
Signed-off-by: Stefan Metzmacher <metze@samba.org>

CVE-2020-1472(ZeroLogon): s3:rpc_server/netlogon: Fix mem leak onto p->mem_ctx in error path of _netr_ServerPasswordSet2().

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14497

Signed-off-by: Jeremy Allison <jra@samba.org>

CVE-2020-1472(ZeroLogon): s4:rpc_server/netlogon: protect netr_ServerPasswordSet2 against unencrypted passwords

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14497

Signed-off-by: Stefan Metzmacher <metze@samba.org>

CVE-2020-1472(ZeroLogon): libcli/auth: reject weak client challenges in netlogon_creds_server_init()

This implements the note from MS-NRPC 3.1.4.1 Session-Key Negotiation:

 7. If none of the first 5 bytes of the client challenge is unique, the
    server MUST fail session-key negotiation without further processing of
    the following steps.

It lets ./zerologon_tester.py from
https://github.com/SecuraBV/CVE-2020-1472.git
report: "Attack failed. Target is probably patched."

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14497

Signed-off-by: Stefan Metzmacher <metze@samba.org>

CVE-2020-1472(ZeroLogon): libcli/auth: add netlogon_creds_is_random_challenge() to avoid weak values

This is the check Windows is using, so we won't generate challenges,
which are rejected by Windows DCs (and future Samba DCs).

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14497

Signed-off-by: Stefan Metzmacher <metze@samba.org>

CVE-2020-1472(ZeroLogon): s4:rpc_server:netlogon: make use of netlogon_creds_random_challenge()

This is not strictly needed, but makes things more clear.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14497

Signed-off-by: Stefan Metzmacher <metze@samba.org>

CVE-2020-1472(ZeroLogon): s3:rpc_server:netlogon: make use of netlogon_creds_random_challenge()

This is not strictly needed, but makes things more clear.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14497

Signed-off-by: Stefan Metzmacher <metze@samba.org>

CVE-2020-1472(ZeroLogon): libcli/auth: make use of netlogon_creds_random_challenge() in netlogon_creds_cli.c

This will avoid getting rejected by the server if we generate
a weak challenge.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14497

Signed-off-by: Stefan Metzmacher <metze@samba.org>

CVE-2020-1472(ZeroLogon): s4:torture/rpc: make use of netlogon_creds_random_challenge()

This will avoid getting flakey tests once our server starts to
reject weak challenges.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14497

Signed-off-by: Stefan Metzmacher <metze@samba.org>

CVE-2020-1472(ZeroLogon): libcli/auth: add netlogon_creds_random_challenge()

It's good to have just a single isolated function that will generate
random challenges, in future we can add some logic in order to
avoid weak values, which are likely to be rejected by a server.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14497

Signed-off-by: Stefan Metzmacher <metze@samba.org>

CVE-2021-44142: libadouble: harden parsing code

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14914

Signed-off-by: Ralph Boehme <slow@samba.org>

CVE-2021-44142: libadouble: add basic cmocka tests

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14914

Signed-off-by: Ralph Boehme <slow@samba.org>
[slow@samba.org: conflict due to missing test in selftest/tests.py]

CVE-2021-44142: libadouble: harden ad_unpack_xattrs()

This ensures ad_unpack_xattrs() is only called for an ad_type of ADOUBLE_RSRC,
which is used for parsing ._ AppleDouble sidecar files, and the buffer
ad->ad_data is AD_XATTR_MAX_HDR_SIZE bytes large which is a prerequisite for all
buffer out-of-bounds access checks in ad_unpack_xattrs().

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14914

Signed-off-by: Ralph Boehme <slow@samba.org>

CVE-2021-44142: smbd: add Netatalk xattr used by vfs_fruit to the list of private Samba xattrs

This is an internal xattr that should not be user visible.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14914

Signed-off-by: Ralph Boehme <slow@samba.org>
[slow@samba.org: conflict due to changed includes in source3/smbd/trans2.c]

CVE-2021-44142: libadouble: add defines for icon lengths

From https://www.ietf.org/rfc/rfc1740.txt

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14914

Signed-off-by: Ralph Boehme <slow@samba.org>

IPA DC: add missing checks

When introducing FreeIPA support, two places were forgotten:

 - schannel gensec module needs to be aware of IPA DC
 - _lsa_QueryInfoPolicy should treat IPA DC as PDC

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14903

Signed-off-by: Alexander Bokovoy <ab@samba.org>
Reviewed-by: Guenther Deschner <gd@samba.org>
Autobuild-User(master): Alexander Bokovoy <ab@samba.org>
Autobuild-Date(master): Sat Nov 13 07:01:26 UTC 2021 on sn-devel-184

(cherry picked from commit c69b66f649c1d47a7367f7efe25b8df32369a3a5)

CVE-2020-25721 krb5pac: Add new buffers for samAccountName and objectSID

These appear when PAC_UPN_DNS_FLAG_HAS_SAM_NAME_AND_SID is set.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14835

Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Joseph Sutton <josephsutton@catalyst.net.nz>

CVE-2020-25719 krb5pac.idl: Add PAC_REQUESTER_SID PAC buffer type

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14561

Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

CVE-2020-25719 krb5pac.idl: Add PAC_ATTRIBUTES_INFO PAC buffer type

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14561

Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

krb5pac.idl: Add missing buffer type values

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14642

Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Backported-by: Andreas Schneider <asn@samba.org>

security.idl: Add well-known SIDs for FAST

Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Isaac Boukris <iboukris@samba.org>
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14881
(cherry picked from commit 0092b4a3ed58b2c256d4dd9117cce927a3edde12)

krb5pac.idl: Add ticket checksum PAC buffer type

Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Isaac Boukris <iboukris@samba.org>
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14881
(cherry picked from commit ff2f38fae79220e16765e17671972f9a55eb7cce)

CVE-2020-25717: s3:auth: simplify make_session_info_krb5() by removing unused arguments

This is only ever be called in standalone mode with an MIT realm,
so we don't have a PAC/info3 structure.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14556

Signed-off-by: Stefan Metzmacher <metze@samba.org>

CVE-2020-25717: s3:auth: simplify get_user_from_kerberos_info() by removing the unused logon_info argument

This code is only every called in standalone mode on a MIT realm,
it means we never have a PAC and we also don't have winbindd arround.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14556

Signed-off-by: Stefan Metzmacher <metze@samba.org>

CVE-2020-25717: s3:auth: let auth3_generate_session_info_pac() reject a PAC in standalone mode

We should be strict in standalone mode, that we only support MIT realms
without a PAC in order to keep the code sane.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14556

Signed-off-by: Stefan Metzmacher <metze@samba.org>
[abartlet@samba.org Backported to Samba 4.12 has conflcits
 as the share reload code is in a different spot]

CVE-2020-25717: selftest: configure 'ktest' env with winbindd and idmap_autorid

The 'ktest' environment was/is designed to test kerberos in an active
directory member setup. It was created at a time we wanted to test
smbd/winbindd with kerberos without having the source4 ad dc available.

This still applies to testing the build with system krb5 libraries
but without relying on a running ad dc.

As a domain member setup requires a running winbindd, we should test it
that way, in order to reflect a valid setup.

As a side effect it provides a way to demonstrate that we can accept
smb connections authenticated via kerberos, but no connection to
a domain controller! In order get this working offline, we need an
idmap backend with ID_TYPE_BOTH support, so we use 'autorid', which
should be the default choice.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14646
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14556

Signed-off-by: Stefan Metzmacher <metze@samba.org>
[scabrero@samba.org Backported to 4.11 Run winbindd in offline mode
 but keep the user name mapping to avoid having to backport fixes
 for bso#14539]

CVE-2020-25717: s3:auth: let auth3_generate_session_info_pac() delegate everything to make_server_info_wbcAuthUserInfo()

This consolidates the code paths used for NTLMSSP and Kerberos!

I checked what we were already doing for NTLMSSP, which is this:

a) source3/auth/auth_winbind.c calls wbcAuthenticateUserEx()
b) as a domain member we require a valid response from winbindd,
   otherwise we'll return NT_STATUS_NO_LOGON_SERVERS
c) we call make_server_info_wbcAuthUserInfo(), which internally
   calls make_server_info_info3()
d) auth_check_ntlm_password() calls
   smb_pam_accountcheck(unix_username, rhost), where rhost
   is only an ipv4 or ipv6 address (without reverse dns lookup)
e) from auth3_check_password_send/auth3_check_password_recv()
   server_returned_info will be passed to auth3_generate_session_info(),
   triggered by gensec_session_info(), which means we'll call into
   create_local_token() in order to transform auth_serversupplied_info
   into auth_session_info.

For Kerberos gensec_session_info() will call
auth3_generate_session_info_pac() via the gensec_generate_session_info_pac()
helper function. The current logic is this:

a) gensec_generate_session_info_pac() is the function that
   evaluates the 'gensec:require_pac', which defaulted to 'no'
   before.
b) auth3_generate_session_info_pac() called
   wbcAuthenticateUserEx() in order to pass the PAC blob
   to winbindd, but only to prime its cache, e.g. netsamlogon cache
   and others. Most failures were just ignored.
c) If the PAC blob is available, it extracted the PAC_LOGON_INFO
   from it.
d) Then we called the horrible get_user_from_kerberos_info() function:
   - It uses a first part of the tickets principal name (before the @)
     as username and combines that with the 'logon_info->base.logon_domain'
     if the logon_info (PAC) is present.
   - As a fallback without a PAC it's tries to ask winbindd for a mapping
     from realm to netbios domain name.
   - Finally is falls back to using the realm as netbios domain name
   With this information is builds 'userdomain+winbind_separator+useraccount'
   and calls map_username() followed by smb_getpwnam() with create=true,
   Note this is similar to the make_server_info_info3() => check_account()
   => smb_getpwnam() logic under 3.
   - It also calls smb_pam_accountcheck(), but may pass the reverse DNS lookup name
     instead of the ip address as rhost.
   - It does some MAP_TO_GUEST_ON_BAD_UID logic and auto creates the
     guest account.
e) We called create_info3_from_pac_logon_info()
f) make_session_info_krb5() calls gets called and triggers this:
   - If get_user_from_kerberos_info() mapped to guest, it calls
     make_server_info_guest()
   - If create_info3_from_pac_logon_info() created a info3 from logon_info,
     it calls make_server_info_info3()
   - Without a PAC it tries pdb_getsampwnam()/make_server_info_sam() with
     a fallback to make_server_info_pw()
   From there it calls create_local_token()

I tried to change auth3_generate_session_info_pac() to behave similar
to auth_winbind.c together with auth3_generate_session_info() as
a domain member, as we now rely on a PAC:

a) As domain member we require a PAC and always call wbcAuthenticateUserEx()
   and require a valid response!
b) we call make_server_info_wbcAuthUserInfo(), which internally
   calls make_server_info_info3(). Note make_server_info_info3()
   handles MAP_TO_GUEST_ON_BAD_UID and make_server_info_guest()
   internally.
c) Similar to auth_check_ntlm_password() we now call
   smb_pam_accountcheck(unix_username, rhost), where rhost
   is only an ipv4 or ipv6 address (without reverse dns lookup)
d) From there it calls create_local_token()

As standalone server (in an MIT realm) we continue
with the already existing code logic, which works without a PAC:
a) we keep smb_getpwnam() with create=true logic as it
   also requires an explicit 'add user script' option.
b) In the following commits we assert that there's
   actually no PAC in this mode, which means we can
   remove unused and confusing code.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14646
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14556

Signed-off-by: Stefan Metzmacher <metze@samba.org>
[abartlet@samba.org Backported due to change in structure
 initialization with { 0 } to zero ]
[abartlet@samba.org backported to 4.12 due to conflict
 with code not present to reload shared on krb5 login]

CVE-2020-25717: s3:ntlm_auth: let ntlm_auth_generate_session_info_pac() base the name on the PAC LOGON_INFO only

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14801
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14556

Signed-off-by: Stefan Metzmacher <metze@samba.org>

CVE-2020-25717: s3:ntlm_auth: fix memory leaks in ntlm_auth_generate_session_info_pac()

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14801
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14556

Signed-off-by: Stefan Metzmacher <metze@samba.org>

CVE-2020-25717: s4:auth: remove unused auth_generate_session_info_principal()

We'll require a PAC at the main gensec layer already.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14801
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14556

Signed-off-by: Stefan Metzmacher <metze@samba.org>
[abartlet@samba.org Backported from master/4.15 as
 check_password is sync in 4.14]

CVE-2020-25717: auth/gensec: always require a PAC in domain mode (DC or member)

AD domains always provide a PAC unless UF_NO_AUTH_DATA_REQUIRED is set
on the service account, which can only be explicitly configured,
but that's an invalid configuration!

We still try to support standalone servers in an MIT realm,
as legacy setup.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14801
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14556

Signed-off-by: Stefan Metzmacher <metze@samba.org>

CVE-2020-25717: Add FreeIPA domain controller role

As we want to reduce use of 'classic domain controller' role but FreeIPA
relies on it internally, add a separate role to mark FreeIPA domain
controller role.

It means that role won't result in ROLE_STANDALONE.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14801
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14556

Pair-Programmed-With: Stefan Metzmacher <metze@samba.org>

Signed-off-by: Alexander Bokovoy <ab@samba.org>
Signed-off-by: Stefan Metzmacher <metze@samba.org>
[abartlet@samba.org Backported due to conflict with DEBUG
 statements and IPA branding changes in comments]

CVE-2020-25717: s3:auth: don't let create_local_token depend on !winbind_ping()

We always require a running winbindd on a domain member, so
we should better fail a request instead of silently alter
the behaviour, which results in a different unix token, just
because winbindd might be restarted.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14801
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14556

Signed-off-by: Stefan Metzmacher <metze@samba.org>

CVE-2020-25717: s3:auth: remove fallbacks in smb_getpwnam()

So far we tried getpwnam("DOMAIN\account") first and
always did a fallback to getpwnam("account") completely
ignoring the domain part, this just causes problems
as we mix "DOMAIN1\account", "DOMAIN2\account",
and "account"!

As we require a running winbindd for domain member setups
we should no longer do a fallback to just "account" for
users served by winbindd!

For users of the local SAM don't use this code path,
as check_sam_security() doesn't call check_account().

The only case where smb_getpwnam("account") happens is
when map_username() via ("username map [script]")  mapped
"DOMAIN\account" to something without '\', but that is
explicitly desired by the admin.

Note: use 'git show -w'

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14801
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14556

Pair-Programmed-With: Stefan Metzmacher <metze@samba.org>

Signed-off-by: Ralph Boehme <slow@samba.org>
Signed-off-by: Stefan Metzmacher <metze@samba.org>
[scabrero@samba.org Removed selftest/knownfail.d/ktest file after fixing
 user map in previous commit]

CVE-2020-25717: s3:auth: no longer let check_account() autocreate local users

So far we autocreated local user accounts based on just the
account_name (just ignoring any domain part).

This only happens via a possible 'add user script',
which is not typically defined on domain members
and on NT4 DCs local users already exist in the
local passdb anyway.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14556
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14801

Signed-off-by: Stefan Metzmacher <metze@samba.org>

CVE-2020-25717: s3:auth: we should not try to autocreate the guest account

We should avoid autocreation of users as much as possible.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14801
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14556

Signed-off-by: Stefan Metzmacher <metze@samba.org>

CVE-2020-25717: s3:auth: Check minimum domain uid

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14801
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14556

Pair-Programmed-With: Stefan Metzmacher <metze@samba.org>

Signed-off-by: Samuel Cabrero <scabrero@samba.org>
Signed-off-by: Stefan Metzmacher <metze@samba.org>

CVE-2020-25717: s3:auth: let auth3_generate_session_info_pac() forward the low level errors

Mapping everything to ACCESS_DENIED makes it hard to debug problems,
which may happen because of our more restrictive behaviour in future.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14801
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14556

Signed-off-by: Stefan Metzmacher <metze@samba.org>

CVE-2020-25717: loadparm: Add new parameter "min domain uid"

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14801
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14556

Pair-Programmed-With: Stefan Metzmacher <metze@samba.org>

Signed-off-by: Samuel Cabrero <scabrero@samba.org>
Signed-off-by: Stefan Metzmacher <metze@samba.org>
[abartlet@samba.org Backported from master/4.15 due to
 conflicts with other new parameters]

CVE-2020-25717: auth/ntlmssp: start with authoritative = 1

This is not strictly needed, but makes it easier to audit
that we don't miss important places.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14556

Signed-off-by: Stefan Metzmacher <metze@samba.org>

CVE-2020-25717: s3:auth: start with authoritative = 1

This is not strictly needed, but makes it easier to audit
that we don't miss important places.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14556

Signed-off-by: Stefan Metzmacher <metze@samba.org>
[scabrero@samba.org Backported to 4.11 due to missing commits
 7f75dec865256049e99f7fcf46317cd2d53e95d1 and
 434030ba711e677fdd167a255d05c1cd4db943b7]

CVE-2020-25717: s3:rpcclient: start with authoritative = 1

This is not strictly needed, but makes it easier to audit
that we don't miss important places.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14556

Signed-off-by: Stefan Metzmacher <metze@samba.org>

CVE-2020-25717: s3:torture: start with authoritative = 1

This is not strictly needed, but makes it easier to audit
that we don't miss important places.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14556

Signed-off-by: Stefan Metzmacher <metze@samba.org>
[scabrero@samba.org Backported to 4.11 due to missing commit
 a5548af018643f2e78c482e33ef0e6073db149e4 to check return value
 of SMBOWFencrypt()]

CVE-2020-25717: s3:ntlm_auth: start with authoritative = 1

This is not strictly needed, but makes it easier to audit
that we don't miss important places.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14556

Signed-off-by: Stefan Metzmacher <metze@samba.org>

CVE-2020-25717: s4:auth_simple: start with authoritative = 1

This is not strictly needed, but makes it easier to audit
that we don't miss important places.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14556

Signed-off-by: Stefan Metzmacher <metze@samba.org>

CVE-2020-25717: s4:smb_server: start with authoritative = 1

This is not strictly needed, but makes it easier to audit
that we don't miss important places.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14556

Signed-off-by: Stefan Metzmacher <metze@samba.org>

CVE-2020-25717: s4:torture: start with authoritative = 1

This is not strictly needed, but makes it easier to audit
that we don't miss important places.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14556

Signed-off-by: Stefan Metzmacher <metze@samba.org>

CVE-2020-25717: s4:auth/ntlm: make sure auth_check_password() defaults to r->out.authoritative = true

We need to make sure that temporary failures don't trigger a fallback
to the local SAM that silently ignores the domain name part for users.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14556

Signed-off-by: Stefan Metzmacher <metze@samba.org>

CVE-2020-25717: s3:winbindd: make sure we default to r->out.authoritative = true

We need to make sure that temporary failures don't trigger a fallback
to the local SAM that silently ignores the domain name part for users.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14556

Signed-off-by: Stefan Metzmacher <metze@samba.org>

CVE-2020-25719 CVE-2020-25717: selftest: remove "gensec:require_pac" settings

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14801
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14556

Signed-off-by: Stefan Metzmacher <metze@samba.org>

selftest/Samba3: replace (winbindd => "yes", skip_wait => 1) with (winbindd => "offline")

This is much more flexible and concentrates the logic in a single place.

We'll use winbindd => "offline" in other places soon.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14870
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14881

Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
(cherry picked from commit 4dc3c68c9a28f71888e3d6dd3b1f0bcdb8fa45de)
(cherry picked from commit 89b9cb8b786c3e4eb8691b5363390b68d8228a2d)

[scabrero@samba.org Backported to 4.11, skip unnecessary chunk just
 changing indentation]

selftest: Fix ktest usermap file

The user was not mapped:

user_in_list: checking user |KTEST/administrator| against |KTEST\Administrator|
The user 'KTEST/administrator' has no mapping. Skip it next time.

Signed-off-by: Samuel Cabrero <scabrero@samba.org>
[scabrero@samba.org Once smb_getpswnam() fallbacks are removed the user
 has to be mapped]

s3/auth: use set_current_user_info() in auth3_generate_session_info_pac()

This delays reloading config slightly, but I don't see how could affect
observable behaviour other then log messages coming from the functions in
between the different locations for lp_load_with_shares() like
make_session_info_krb5() are sent to a different logfile if "log file" uses %U.

Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
(cherry picked from commit dc4b1e39ce1f2201a2d6ae2d4cffef2448f69a62)

[scabrero@samba.org Prerequisite for CVE-2020-25717 backport]

CVE-2016-2124: s3:libsmb: don't fallback to non spnego authentication if we require kerberos

We should not send NTLM[v2] nor plaintext data on the wire if the user
asked for kerberos only.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=12444

Signed-off-by: Stefan Metzmacher <metze@samba.org>

CVE-2016-2124: s4:libcli/sesssetup: don't fallback to non spnego authentication if we require kerberos

We should not send NTLM[v2] data on the wire if the user asked for kerberos
only.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=12444

Signed-off-by: Stefan Metzmacher <metze@samba.org>

passdb: Simplify sids_to_unixids()

Best reviewed with "git show -b", there's a "continue" statement that
changes subsequent indentation.

Decouple lookup status of ids from ID_TYPE_NOT_SPECIFIED

Bug: https://bugzilla.samba.org/show_bug.cgi?id=14571

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

libnet_join: add SPNs for additional-dns-hostnames entries

and set msDS-AdditionalDnsHostName to the specified list.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14116

Signed-off-by: Isaac Boukris <iboukris@redhat.com>
Reviewed-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Alexander Bokovoy <ab@samba.org>
Autobuild-User(master): Ralph Böhme <slow@samba.org>
Autobuild-Date(master): Fri Oct 25 10:43:08 UTC 2019 on sn-devel-184

docs-xml: add "additional dns hostnames" smb.conf option

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14116

Signed-off-by: Isaac Boukris <iboukris@redhat.com>
Reviewed-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Alexander Bokovoy <ab@samba.org>

libnet_join_set_machine_spn: simplify adding uniq spn to array

and do not skip adding a fully qualified spn to netbios-aliases
in case a short spn already existed.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14116

Signed-off-by: Isaac Boukris <iboukris@redhat.com>
Reviewed-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Alexander Bokovoy <ab@samba.org>

libnet_join_set_machine_spn: simplify memory handling

and avoid a possible memory leak when passing null to
add_string_to_array() as mem_ctx.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14116

Signed-off-by: Isaac Boukris <iboukris@redhat.com>
Reviewed-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Alexander Bokovoy <ab@samba.org>

libnet_join_set_machine_spn: improve style and make a bit room for indentation

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14116

Signed-off-by: Isaac Boukris <iboukris@redhat.com>
Reviewed-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Alexander Bokovoy <ab@samba.org>

libnet_join: build dnsHostName from netbios name and lp_dnsdomain()

This make the join process much more reliable, and avoids "Constraint
violation" error when the fqdn returned from getaddrinfo has already
got assigned an SPN.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14116

Signed-off-by: Isaac Boukris <iboukris@redhat.com>
Reviewed-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Alexander Bokovoy <ab@samba.org>

smbd: Fix the build with clang

clang correctly complains that "close_fsp" is used uninitialized if
"get_posix_fsp" fails and we end up in "goto out;".

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14251

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
(cherry picked from commit a8a1ca3f83dce6d725392989cbc97271cbf52f4a)

s3: torture: Ensure SMB1 cli_qpathinfo2() doesn't return an inode number.

Piggyback on existing tests, ensure we don't regress on:

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14161

Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
(cherry picked from commit 8e55a8562951924e4b1aad5a6d67fc8b309590c1)

s3: libsmb: Ensure SMB1 cli_qpathinfo2() doesn't return an inode number.

The info level it uses doesn't return that, previously we
were using the field that is returned as the EA size as
the inode number (which is usually zero, so the code in
libsmbclient would then synthesize an inode number from
a hash of the pathname, which is all it can do for SMB1).

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14161

Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
(cherry picked from commit d495074ee27a5f528d5156a69800ee58d799b1eb)

lib:util: Log mkdir error on correct debug levels

For smbd we want an error and for smbclient we only want it in NOTICE
debug level.
The default log level of smbclient is log level 1 so we need notice to
not spam the user.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14253

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Guenther Deschner <gd@samba.org>
Autobuild-User(master): Günther Deschner <gd@samba.org>
Autobuild-Date(master): Mon Jan 27 15:55:24 UTC 2020 on sn-devel-184

(cherry picked from commit 0ad6a243b259d284064c0c5abcc7d430d55be7e1)

heimdal: do not compile weak crypto

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14202

Signed-off-by: Isaac Boukris <iboukris@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

selftest: allow any kdc error in mitm-s4u2self test

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14202

Signed-off-by: Isaac Boukris <iboukris@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>