kai/samba.git
7 years agobuild: Do not put a .distversion file into the GIT tree
Andrew Bartlett [Sat, 18 Aug 2012 11:25:30 +0000 (21:25 +1000)]
build: Do not put a .distversion file into the GIT tree

This places the file only in the tarball, and shows how to
auto-generate other files for placement in the tarball.

Andrew Bartlett

7 years agos3-build: Enable vfs_fake_acls when in developer mode or on the build farm
Andrew Bartlett [Sat, 18 Aug 2012 14:27:01 +0000 (00:27 +1000)]
s3-build: Enable vfs_fake_acls when in developer mode or on the build farm

7 years agos3-vfs: Add lstat and lchown hooks to the vfs_fake_acls module
Andrew Bartlett [Mon, 20 Aug 2012 11:28:57 +0000 (21:28 +1000)]
s3-vfs: Add lstat and lchown hooks to the vfs_fake_acls module

7 years agos3-vfs: Correct the implementation of fake_acls_sys_acl_delete_def_file()
Andrew Bartlett [Sat, 18 Aug 2012 14:01:57 +0000 (00:01 +1000)]
s3-vfs: Correct the implementation of fake_acls_sys_acl_delete_def_file()

7 years agos3-vfs: Use the system. namespace for fake ACLs
Andrew Bartlett [Mon, 20 Aug 2012 00:10:27 +0000 (10:10 +1000)]
s3-vfs: Use the system. namespace for fake ACLs

By using the system. namespace, we make sure this is only run on top of a TDB
based ACL store (ie in make test).

Andrew Bartlett

7 years agos3-smbd: ensure we give appropriate errors for EA requests on streams
Andrew Bartlett [Sun, 19 Aug 2012 12:46:58 +0000 (22:46 +1000)]
s3-smbd: ensure we give appropriate errors for EA requests on streams

7 years agos3-smbd: Do not look for EA information on a stream
Andrew Bartlett [Sun, 19 Aug 2012 11:58:56 +0000 (21:58 +1000)]
s3-smbd: Do not look for EA information on a stream

The estimated EA size needs to be of the main file.  However, the fsp
may point to the stream, so we need to ignore it if this is the case.

This may mean we estimate wrong if there has been a rename.

Andrew Bartlett

7 years agos3-smbd: Push smb_fname into estimate_ea_size
Andrew Bartlett [Sun, 19 Aug 2012 05:24:04 +0000 (15:24 +1000)]
s3-smbd: Push smb_fname into estimate_ea_size

This ensures that we return the ea size of the stream, not the overall file.

This is important as if there is an EA on the main file, the raw.streams
test was failing.

Andrew Bartlett

7 years agos4-ntvfs: Add TODO on ea_size
Andrew Bartlett [Sun, 19 Aug 2012 23:49:30 +0000 (09:49 +1000)]
s4-ntvfs: Add TODO on ea_size

This is almost certainly un-important.

Andrew Bartlett

7 years agos4-ntvfs: Ensure we do not attempt to write EAs on streams
Andrew Bartlett [Sun, 19 Aug 2012 23:47:43 +0000 (09:47 +1000)]
s4-ntvfs: Ensure we do not attempt to write EAs on streams

7 years agos3-vfs: Allow vfs_xattr_tdb to work without a connected share
Andrew Bartlett [Mon, 20 Aug 2012 09:20:19 +0000 (19:20 +1000)]
s3-vfs: Allow vfs_xattr_tdb to work without a connected share

This is needed to that get_nt_acl_no_snum() can work.

Andrew Bartlett

7 years agos4:torture:basic: add more delete test - variants of deltest16 and deltest17
Michael Adam [Wed, 8 Aug 2012 21:43:05 +0000 (23:43 +0200)]
s4:torture:basic: add more delete test - variants of deltest16 and deltest17

There seems to be a difference if the initial delete_on_close flag
was set on a handle that created the file or if the handle if was
for a file that already existed.

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

Signed-off-by: Stefan Metzmacher <metze@samba.org>
Autobuild-User(master): Stefan Metzmacher <metze@samba.org>
Autobuild-Date(master): Fri Aug 17 21:44:24 CEST 2012 on sn-devel-104

7 years agos3: add a debug message for failed execv in sys_popen()
Michael Adam [Wed, 1 Aug 2012 13:42:49 +0000 (15:42 +0200)]
s3: add a debug message for failed execv in sys_popen()

Signed-off-by: Stefan Metzmacher <metze@samba.org>
7 years agolib/util: add server_id_set_disconnected() and server_id_is_disconnected()
Michael Adam [Mon, 6 Aug 2012 14:35:46 +0000 (16:35 +0200)]
lib/util: add server_id_set_disconnected() and server_id_is_disconnected()

Utility functions for handling the special placeholder server-id value
for disconnected clients (to be used for durable handles).

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

7 years agolib/util: let server_id_str() skip the task_id if 0 in the cluster case too
Stefan Metzmacher [Fri, 17 Aug 2012 10:50:09 +0000 (12:50 +0200)]
lib/util: let server_id_str() skip the task_id if 0 in the cluster case too

server_id_from_string() already handles that case.

metze

7 years agos3:lib: implement interpret_pid() as wrapper of server_id_from_string()
Stefan Metzmacher [Fri, 17 Aug 2012 10:49:26 +0000 (12:49 +0200)]
s3:lib: implement interpret_pid() as wrapper of server_id_from_string()

metze

7 years agolib/util: add server_id_from_string()
Stefan Metzmacher [Fri, 17 Aug 2012 10:47:57 +0000 (12:47 +0200)]
lib/util: add server_id_from_string()

metze

7 years agos3:lib: implement serverid_equal() as macro of server_id_equal()
Stefan Metzmacher [Fri, 17 Aug 2012 10:45:33 +0000 (12:45 +0200)]
s3:lib: implement serverid_equal() as macro of server_id_equal()

metze

7 years agolib/util: add server_id_equal()
Stefan Metzmacher [Fri, 17 Aug 2012 10:41:02 +0000 (12:41 +0200)]
lib/util: add server_id_equal()

metze

7 years agos3:vfs_tsmsm only send notifications when file was offline before
Christian Ambach [Fri, 17 Aug 2012 15:40:24 +0000 (17:40 +0200)]
s3:vfs_tsmsm only send notifications when file was offline before

Autobuild-User(master): Christian Ambach <ambi@samba.org>
Autobuild-Date(master): Fri Aug 17 20:05:30 CEST 2012 on sn-devel-104

7 years agos3: Adapt the tsmsm module to the new aio routines
Volker Lendecke [Fri, 10 Aug 2012 09:33:54 +0000 (11:33 +0200)]
s3: Adapt the tsmsm module to the new aio routines

Signed-off-by: Christian Ambach <ambi@samba.org>
7 years agos3-vfs: Fix calls of lp_parm_talloc_string
Volker Lendecke [Wed, 8 Aug 2012 09:23:12 +0000 (11:23 +0200)]
s3-vfs: Fix calls of lp_parm_talloc_string

Signed-off-by: Christian Ambach <ambi@samba.org>
7 years agos3: Remove the gpfs_hsm_notify module
Volker Lendecke [Wed, 8 Aug 2012 08:38:36 +0000 (10:38 +0200)]
s3: Remove the gpfs_hsm_notify module

The functionality has been merged into vfs_gpfs

Signed-off-by: Christian Ambach <ambi@samba.org>
7 years agos3: Merge vfs_gpfs_hsm_notify into vfs_gpfs.c
Volker Lendecke [Tue, 7 Aug 2012 22:32:35 +0000 (00:32 +0200)]
s3: Merge vfs_gpfs_hsm_notify into vfs_gpfs.c

It was separated out because formerly our async I/O was not properly
stackable. aio_fork could for example catch aio and not get aio_return
get through to vfs_gpfs

Signed-off-by: Christian Ambach <ambi@samba.org>
7 years agos4-dsdb: Use tmp_ctx in kccsrv_check_deleted to avoid leaking memory onto part->dn
Andrew Bartlett [Fri, 17 Aug 2012 13:04:56 +0000 (23:04 +1000)]
s4-dsdb: Use tmp_ctx in kccsrv_check_deleted to avoid leaking memory onto part->dn

The confusing use of do_dn as a memory context while legitimate
created a bug when it was copied and modified to search on a DN from
long-term state.

By always using a temporary memory context it is clear what paramter
is the memory context.

This was found based on a log provided by Ricky Nance
<ricky.nance@weaubleau.k12.mo.us>.  Thanks Ricky!

Andrew Bartlett

Autobuild-User(master): Andrew Bartlett <abartlet@samba.org>
Autobuild-Date(master): Fri Aug 17 18:24:10 CEST 2012 on sn-devel-104

7 years agos4-kcc: Avoid use-after-free of dn and add tmp_ctx
Andrew Bartlett [Fri, 17 Aug 2012 12:47:44 +0000 (22:47 +1000)]
s4-kcc: Avoid use-after-free of dn and add tmp_ctx

By using a tmp_ctx we are clearer about allocating temporary memory.

Andrew Bartlett

7 years agos3:selftest: add some tests against a share the requires encryption
Stefan Metzmacher [Fri, 17 Aug 2012 06:50:21 +0000 (08:50 +0200)]
s3:selftest: add some tests against a share the requires encryption

metze

Autobuild-User(master): Stefan Metzmacher <metze@samba.org>
Autobuild-Date(master): Fri Aug 17 16:34:26 CEST 2012 on sn-devel-104

7 years agos3:smb2_negprot: annouce/negotiate SMB3 encryption support
Stefan Metzmacher [Wed, 8 Aug 2012 05:07:53 +0000 (07:07 +0200)]
s3:smb2_negprot: annouce/negotiate SMB3 encryption support

metze

7 years agos3:smb2_server: add SMB3 encryption support
Stefan Metzmacher [Wed, 8 Aug 2012 05:07:03 +0000 (07:07 +0200)]
s3:smb2_server: add SMB3 encryption support

metze

7 years agos3:smbd: don't disconnect the client when a share has "smb encrypt = required"
Stefan Metzmacher [Fri, 17 Aug 2012 07:42:27 +0000 (09:42 +0200)]
s3:smbd: don't disconnect the client when a share has "smb encrypt = required"

It's not the client fault, if he doesn't know that encryption is required.
We should just return ACCESS_DENIED and let the client work on other
shares and open files on the current SMB connection.

metze

7 years agos3:smbd: lp_smb_encrypt() returns SMB_SIGNING_* values
Stefan Metzmacher [Fri, 17 Aug 2012 07:40:52 +0000 (09:40 +0200)]
s3:smbd: lp_smb_encrypt() returns SMB_SIGNING_* values

metze

7 years agos3:smbd: make use of ENCRYPTION_REQUIRED()
Stefan Metzmacher [Fri, 17 Aug 2012 07:41:42 +0000 (09:41 +0200)]
s3:smbd: make use of ENCRYPTION_REQUIRED()

metze

7 years agolibcli/smb: make sure the SMB2_TRANSFORM pdu is complete
Stefan Metzmacher [Fri, 17 Aug 2012 06:31:52 +0000 (08:31 +0200)]
libcli/smb: make sure the SMB2_TRANSFORM pdu is complete

metze

7 years agos4:libcli/smb2: reset trsnport->compound.related when a compound chain is finished
Stefan Metzmacher [Fri, 17 Aug 2012 06:25:08 +0000 (08:25 +0200)]
s4:libcli/smb2: reset trsnport->compound.related when a compound chain is finished

metze

7 years agobuild: Ensure -Werror=format works with -Wformat=2 on NULL format strings
Andrew Bartlett [Fri, 17 Aug 2012 10:43:07 +0000 (20:43 +1000)]
build: Ensure -Werror=format works with -Wformat=2 on NULL format strings

This should fix the build on some gcc versions, (noticed on FreeBSD by Volker).

We want the protection of -Werror=format without the errors extending
to the format string itself being NULL, because that is valid for
ldb_search().

Andrew Bartlett

Autobuild-User(master): Andrew Bartlett <abartlet@samba.org>
Autobuild-Date(master): Fri Aug 17 14:50:16 CEST 2012 on sn-devel-104

7 years agos4-dsdb: Ensure we always free tmp_ctx in schema refresh check
Andrew Bartlett [Fri, 17 Aug 2012 02:42:15 +0000 (12:42 +1000)]
s4-dsdb: Ensure we always free tmp_ctx in schema refresh check

This was found based on a log provided by Ricky Nance
<ricky.nance@weaubleau.k12.mo.us>.  Thanks Ricky!

In that log, over 2.5 days this particular allocation was repeated:

1715099     talloc_new: ../source4/dsdb/samdb/ldb_modules/schema_load.c:120 contains      0 bytes in   1 blocks

Andrew Bartlett

Autobuild-User(master): Andrew Bartlett <abartlet@samba.org>
Autobuild-Date(master): Fri Aug 17 06:21:18 CEST 2012 on sn-devel-104

7 years agoFix bug #9085 - NMB registration for a duplicate workstation fails with registration...
S Hargagan [Thu, 16 Aug 2012 23:27:10 +0000 (16:27 -0700)]
Fix bug #9085 - NMB registration for a duplicate workstation fails with registration refuse

Signed-off-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Fri Aug 17 04:15:21 CEST 2012 on sn-devel-104

7 years agos3-torture: Allow vfstest to set ACLs on a directory
Andrew Bartlett [Thu, 16 Aug 2012 06:50:47 +0000 (16:50 +1000)]
s3-torture: Allow vfstest to set ACLs on a directory

Autobuild-User(master): Andrew Bartlett <abartlet@samba.org>
Autobuild-Date(master): Fri Aug 17 02:35:49 CEST 2012 on sn-devel-104

7 years agos4: Fix returns in py_check_dcerpc_type
Volker Lendecke [Thu, 16 Aug 2012 09:38:26 +0000 (11:38 +0200)]
s4: Fix returns in py_check_dcerpc_type

Signed-off-by: Andrew Bartlett <abartlet@samba.org>
7 years agos3:smb2_server: try to sign an error response if we have a signing key
Stefan Metzmacher [Thu, 16 Aug 2012 13:14:51 +0000 (15:14 +0200)]
s3:smb2_server: try to sign an error response if we have a signing key

metze

Autobuild-User(master): Stefan Metzmacher <metze@samba.org>
Autobuild-Date(master): Fri Aug 17 00:54:01 CEST 2012 on sn-devel-104

7 years agos3:smb2_server: verify the signature before the session_status
Stefan Metzmacher [Thu, 16 Aug 2012 13:08:40 +0000 (15:08 +0200)]
s3:smb2_server: verify the signature before the session_status

metze

7 years agos3:smb2_server: add some const to print_req_vectors()
Stefan Metzmacher [Thu, 16 Aug 2012 10:00:57 +0000 (12:00 +0200)]
s3:smb2_server: add some const to print_req_vectors()

metze

7 years agos4:cldap_server: only return DS_SERVER_*TIMESERV if "ntp_signd" is used
Stefan Metzmacher [Thu, 16 Aug 2012 11:32:14 +0000 (13:32 +0200)]
s4:cldap_server: only return DS_SERVER_*TIMESERV if "ntp_signd" is used

metze

7 years agos4:cldap_server: set DS_SERVER_SELECT_SECRET_DOMAIN_6 if we're a RODC
Stefan Metzmacher [Thu, 16 Aug 2012 11:31:33 +0000 (13:31 +0200)]
s4:cldap_server: set DS_SERVER_SELECT_SECRET_DOMAIN_6 if we're a RODC

metze

7 years agos3-libsmb: Add a simple test for python bindings
Volker Lendecke [Wed, 15 Aug 2012 12:08:45 +0000 (14:08 +0200)]
s3-libsmb: Add a simple test for python bindings

Signed-off-by: Stefan Metzmacher <metze@samba.org>
Autobuild-User(master): Stefan Metzmacher <metze@samba.org>
Autobuild-Date(master): Thu Aug 16 22:49:06 CEST 2012 on sn-devel-104

7 years agos3-libsmb: Add a python wrapper
Volker Lendecke [Thu, 2 Aug 2012 21:21:27 +0000 (23:21 +0200)]
s3-libsmb: Add a python wrapper

Please note that this is not finished and only for internal use.

Signed-off-by: Stefan Metzmacher <metze@samba.org>
7 years agotevent: change version to 0.9.17 after adding the "poll_mt" backend save-diskspace-tags/tevent-0.9.17
Stefan Metzmacher [Thu, 16 Aug 2012 19:06:45 +0000 (21:06 +0200)]
tevent: change version to 0.9.17 after adding the "poll_mt" backend

metze

7 years agotevent: Add threaded poll_mt testcase
Volker Lendecke [Mon, 30 Jul 2012 07:09:46 +0000 (09:09 +0200)]
tevent: Add threaded poll_mt testcase

Signed-off-by: Stefan Metzmacher <metze@samba.org>
7 years agolib/tevent: Add a thread-safe tevent backend
Volker Lendecke [Mon, 13 Aug 2012 14:06:01 +0000 (16:06 +0200)]
lib/tevent: Add a thread-safe tevent backend

Signed-off-by: Stefan Metzmacher <metze@samba.org>
7 years agotevent_poll: Decouple poll_ev->fds handling from adding/removing fds
Volker Lendecke [Sun, 29 Jul 2012 11:05:36 +0000 (13:05 +0200)]
tevent_poll: Decouple poll_ev->fds handling from adding/removing fds

Step 1 in a python backend for multiple threads

Signed-off-by: Stefan Metzmacher <metze@samba.org>
7 years agos3-g_lock: Make g_lock_lock more robust
Volker Lendecke [Fri, 10 Aug 2012 15:00:38 +0000 (17:00 +0200)]
s3-g_lock: Make g_lock_lock more robust

If for some reason the cleanup of dbwrap_watch_send does not work
properly, we might starve indefinitely. Make the lock routine more
robust by retrying every 5-10 seconds. g_lock_trylock will clean up
orphaned entries.

Signed-off-by: Christian Ambach <ambi@samba.org>
Autobuild-User(master): Christian Ambach <ambi@samba.org>
Autobuild-Date(master): Thu Aug 16 19:44:00 CEST 2012 on sn-devel-104

7 years agos3-msg: For msg_channel, correct the talloc hierarchy
Volker Lendecke [Fri, 10 Aug 2012 13:24:23 +0000 (15:24 +0200)]
s3-msg: For msg_channel, correct the talloc hierarchy

talloc_free() of a channel must free all immediate requests that have
piled up

Signed-off-by: Christian Ambach <ambi@samba.org>
7 years agos3-msg: Rename msg_channel_init_destructor
Volker Lendecke [Fri, 10 Aug 2012 12:54:37 +0000 (14:54 +0200)]
s3-msg: Rename msg_channel_init_destructor

Signed-off-by: Christian Ambach <ambi@samba.org>
7 years agos3-autoconf: Fix deps for dbwrap_torture
Volker Lendecke [Fri, 10 Aug 2012 12:29:44 +0000 (14:29 +0200)]
s3-autoconf: Fix deps for dbwrap_torture

Signed-off-by: Christian Ambach <ambi@samba.org>
7 years agos3-g_lock: Properly free "rec" on retry to avoid deadlock
Volker Lendecke [Fri, 10 Aug 2012 11:42:51 +0000 (13:42 +0200)]
s3-g_lock: Properly free "rec" on retry to avoid deadlock

Signed-off-by: Christian Ambach <ambi@samba.org>
7 years agos3:brlock: give traverse_fn a proper name
Christian Ambach [Wed, 8 Aug 2012 15:46:59 +0000 (17:46 +0200)]
s3:brlock: give traverse_fn a proper name

7 years agos3:vfs_gpfs: make sure parameters are set correctly for leases
Christian Ambach [Tue, 7 Aug 2012 16:42:39 +0000 (18:42 +0200)]
s3:vfs_gpfs: make sure parameters are set correctly for leases

gpfs:leases requires kernel oplocks = yes and level2 oplocks = no
to work properly

make sure those are set correctly for a share

7 years agos3:vfs_gpfs: Fix compile error in gpfs module
Christof Schmitt [Mon, 24 Oct 2011 23:52:50 +0000 (16:52 -0700)]
s3:vfs_gpfs: Fix compile error in gpfs module

Fix this compile error by adding fcntl.h as a include that defines F_RDLCK
and F_WRLCK:

modules/gpfs.c: In function ‘set_gpfs_lease’:
modules/gpfs.c:102: error: ‘F_RDLCK’ undeclared (first use in this function)
modules/gpfs.c:102: error: (Each undeclared identifier is reported only once
modules/gpfs.c:102: error: for each function it appears in.)
modules/gpfs.c:105: error: ‘F_WRLCK’ undeclared (first use in this function)

7 years agolibcli/smb: support broken OS/2 error responses bug #9096
Stefan Metzmacher [Thu, 16 Aug 2012 04:55:36 +0000 (06:55 +0200)]
libcli/smb: support broken OS/2 error responses bug #9096

OS/2 skips the DATA Block in SMB1 responses.

metze

Autobuild-User(master): Stefan Metzmacher <metze@samba.org>
Autobuild-Date(master): Thu Aug 16 13:16:49 CEST 2012 on sn-devel-104

7 years agos3-selftest: Add a seperate test for ACL tests using vfstest
Andrew Bartlett [Thu, 16 Aug 2012 03:46:02 +0000 (13:46 +1000)]
s3-selftest: Add a seperate test for ACL tests using vfstest

This does not check for consistency or correctness yet, that will be
done with python unit tests.  The purpose of this test is to ensure
that the vfstest wrapper doesn't crash.

Andrew Bartlett

Autobuild-User(master): Andrew Bartlett <abartlet@samba.org>
Autobuild-Date(master): Thu Aug 16 09:32:25 CEST 2012 on sn-devel-104

7 years agos3-torture: Add ACL commands to vfstest
Andrew Bartlett [Thu, 16 Aug 2012 03:44:48 +0000 (13:44 +1000)]
s3-torture: Add ACL commands to vfstest

This will allow easier investigation of our ACL layer.

Andrew Bartlett

7 years agos3-torture: Use talloc more in vfstest
Andrew Bartlett [Thu, 16 Aug 2012 05:16:01 +0000 (15:16 +1000)]
s3-torture: Use talloc more in vfstest

This matches the rest of Samba, which allocates many of these structures with talloc.

Andrew Bartlett

7 years agos3-torture: Initialise fsp fully in vfstest open
Andrew Bartlett [Thu, 16 Aug 2012 04:37:42 +0000 (14:37 +1000)]
s3-torture: Initialise fsp fully in vfstest open

7 years agos3-smbd: Do not check no_acl_syscall_error(errno) after sys_acl_init()
Andrew Bartlett [Thu, 16 Aug 2012 03:44:01 +0000 (13:44 +1000)]
s3-smbd: Do not check no_acl_syscall_error(errno) after sys_acl_init()

This is no longer a VFS call, so will no longer fail in this way.

Andrew Bartlett

7 years agoselftest: Extend xattr-tdb-1 vfstest to call stat
Andrew Bartlett [Thu, 16 Aug 2012 00:03:35 +0000 (10:03 +1000)]
selftest: Extend xattr-tdb-1 vfstest to call stat

7 years agos3-vfs: Continue to make vfs_xattr_tdb non-recursive
Andrew Bartlett [Thu, 16 Aug 2012 00:02:44 +0000 (10:02 +1000)]
s3-vfs: Continue to make vfs_xattr_tdb non-recursive

We now always use _NEXT_ VFS calls, even to build the file id.

Andrew Bartlett

7 years agos3-vfs: Add new VFS module to fake setting an ACL
Andrew Bartlett [Wed, 15 Aug 2012 10:34:41 +0000 (20:34 +1000)]
s3-vfs: Add new VFS module to fake setting an ACL

The purpose of this module is to remove the relience on the system having ACL support
to test NT ACLs.

Andrew Bartlett

7 years agolibrpc/idl: Fix acl array definition in smb_acl_t
Andrew Bartlett [Thu, 16 Aug 2012 04:13:00 +0000 (14:13 +1000)]
librpc/idl: Fix acl array definition in smb_acl_t

7 years agos3-selftest: convert xattr-tdb-1 vfstest driver into a subunit test
Andrew Bartlett [Wed, 15 Aug 2012 22:55:43 +0000 (08:55 +1000)]
s3-selftest: convert xattr-tdb-1 vfstest driver into a subunit test

We don't use the simple smb.conf because we need to override all the
paths for this to work as non-root without a panic, so we use the s3dc
environment, which already loads this module.

Andrew Bartlett

Autobuild-User(master): Andrew Bartlett <abartlet@samba.org>
Autobuild-Date(master): Thu Aug 16 02:55:19 CEST 2012 on sn-devel-104

7 years agos3-selftest: convert stream_depot vfstest driver into a subunit test
Andrew Bartlett [Wed, 15 Aug 2012 22:37:54 +0000 (08:37 +1000)]
s3-selftest: convert stream_depot vfstest driver into a subunit test

This gives us our first automated coverage of the vfstest binary.

We don't use the simple smb.conf because we need to override all the
paths for this to work as non-root without a panic, so we use the s3dc
environment, which already loads this module.

Andrew Bartlett

7 years agofix printf warning in net connections
Andrew Bartlett [Wed, 15 Aug 2012 21:47:57 +0000 (07:47 +1000)]
fix printf warning in net connections

7 years agos3:utils: remove standalone cclean tool
Gregor Beck [Fri, 13 Jul 2012 13:31:16 +0000 (15:31 +0200)]
s3:utils: remove standalone cclean tool

Signed-off-by: Andrew Bartlett <abartlet@samba.org>
7 years agos3:doc manpage for "net connections cleanup"
Gregor Beck [Mon, 16 Jul 2012 13:10:46 +0000 (15:10 +0200)]
s3:doc manpage for "net connections cleanup"

Signed-off-by: Andrew Bartlett <abartlet@samba.org>
7 years agos3:net add command "connections cleanup"
Gregor Beck [Mon, 16 Jul 2012 07:34:15 +0000 (09:34 +0200)]
s3:net add command "connections cleanup"

Signed-off-by: Andrew Bartlett <abartlet@samba.org>
7 years agos3-vfs: Set errno in xattr emulation
Andrew Bartlett [Wed, 15 Aug 2012 12:22:26 +0000 (22:22 +1000)]
s3-vfs: Set errno in xattr emulation

The caller may check this errno.

Andrew Bartlett

Autobuild-User(master): Andrew Bartlett <abartlet@samba.org>
Autobuild-Date(master): Wed Aug 15 18:05:33 CEST 2012 on sn-devel-104

7 years agos3-vfs: Avoid loops in VFS modules: call _NEXT functions in xattr emulation
Andrew Bartlett [Wed, 15 Aug 2012 12:21:48 +0000 (22:21 +1000)]
s3-vfs: Avoid loops in VFS modules: call _NEXT functions in xattr emulation

We need to call the next module in the stack otherwise we will loop if
the stat call is in turn implemented in terms of extended attribute
lookup.

Andrew Bartlett

7 years agos3-vfs: ensure we strictly free the talloc_stackframe
Andrew Bartlett [Wed, 15 Aug 2012 11:22:18 +0000 (21:22 +1000)]
s3-vfs: ensure we strictly free the talloc_stackframe

We must do this when leaving the function or else in development, we
will panic.

Andrew Bartlett

7 years agos4-selftest: Fix test name for samba.tests.dcerpc.bare
Andrew Bartlett [Wed, 15 Aug 2012 10:33:46 +0000 (20:33 +1000)]
s4-selftest: Fix test name for samba.tests.dcerpc.bare

7 years agolibrpc/idl: Make smb_acl_t public so we can pull/push it as a blob
Andrew Bartlett [Wed, 15 Aug 2012 10:33:27 +0000 (20:33 +1000)]
librpc/idl: Make smb_acl_t public so we can pull/push it as a blob

7 years agolibcli/smb: verify decrypted SMB2 pdus correctly
Stefan Metzmacher [Tue, 14 Aug 2012 07:35:59 +0000 (09:35 +0200)]
libcli/smb: verify decrypted SMB2 pdus correctly

We need to make sure we got a encrypted response if we asked
for it.

If we don't get a encrypted response, we use a similar logic
as with signing to propagated wellknown errors to the higher
layer and set state->smb2.signing_skipped = true.

metze

Autobuild-User(master): Stefan Metzmacher <metze@samba.org>
Autobuild-Date(master): Wed Aug 15 16:26:26 CEST 2012 on sn-devel-104

7 years agolibcli/smb: fix parsing of compounded messages within a SMB2_TRANSFORM pdu
Stefan Metzmacher [Tue, 14 Aug 2012 07:33:01 +0000 (09:33 +0200)]
libcli/smb: fix parsing of compounded messages within a SMB2_TRANSFORM pdu

One SMB2_TRANSFORM pdu wraps multiple SMB2 pdus.

We inject the SMB2_TRANSFORM header to each response which was wrapped
inside. This allows the next layer to verify if the SMB2 pdu was encrypted.

metze

7 years agolibcli/smb: fix smb2cli_req_compound_submit for multiple encrypted messages
Stefan Metzmacher [Tue, 14 Aug 2012 07:30:43 +0000 (09:30 +0200)]
libcli/smb: fix smb2cli_req_compound_submit for multiple encrypted messages

There should be only one SMB2_TRANSFORM header for all compound requests.

metze

7 years agos3:smb2_server: do calculations based on SMBD_SMB2_NUM_IOV_PER_REQ in smbd_smb2_reque...
Stefan Metzmacher [Wed, 15 Aug 2012 12:43:40 +0000 (14:43 +0200)]
s3:smb2_server: do calculations based on SMBD_SMB2_NUM_IOV_PER_REQ in smbd_smb2_request_validate()

metze

7 years agolibcli/smb: all flags except SMB2_HDR_FLAG_ASYNC should be cleared in a cancel request.
Stefan Metzmacher [Wed, 15 Aug 2012 12:17:25 +0000 (14:17 +0200)]
libcli/smb: all flags except SMB2_HDR_FLAG_ASYNC should be cleared in a cancel request.

metze

7 years agos3-sysacls: Remove sys_acl_free_qualifier() as it is a no-op
Andrew Bartlett [Mon, 13 Aug 2012 10:54:24 +0000 (20:54 +1000)]
s3-sysacls: Remove sys_acl_free_qualifier() as it is a no-op

Autobuild-User(master): Andrew Bartlett <abartlet@samba.org>
Autobuild-Date(master): Wed Aug 15 05:23:18 CEST 2012 on sn-devel-104

7 years agos3-sysacls: Remove sys_acl_free_acl() and replace with TALLOC_FREE()
Andrew Bartlett [Mon, 13 Aug 2012 10:51:41 +0000 (20:51 +1000)]
s3-sysacls: Remove sys_acl_free_acl() and replace with TALLOC_FREE()

7 years agos3-smbd: Remove sys_acl_*() VFS wrapper functions
Andrew Bartlett [Mon, 13 Aug 2012 10:03:48 +0000 (20:03 +1000)]
s3-smbd: Remove sys_acl_*() VFS wrapper functions

We no longer do struct smb_acl_t manipuations via the VFS layer,
which is now reduced to handling the get/set functions.

The only backend that implemented these functions (aside from audit)
was the vfs_default module calling the sys_acl code.  The various ACL
implementation modules either worked on the fully initilaised
smb_acl_t object or on NT ACLs.

This not only makes the operation of the posix ACL code more efficient
(as allocation and free is not put via the VFS), it makes it easier to
test and removes the fantasy that a module could safely redefine this
structure or the behaviour here.

The smb_acls.idl now defines the structure, and it is now allocated
with talloc.

These operations were originally added to the VFS in commit
3bb219161a270f12c27c3bc7e1220829c6e9f284.

Andrew Bartlett

7 years agos3-smbd: Remove unused conn argument from convert_permset_to_mode_t()
Andrew Bartlett [Mon, 13 Aug 2012 10:00:59 +0000 (20:00 +1000)]
s3-smbd: Remove unused conn argument from convert_permset_to_mode_t()

7 years agos3-smbd: Call sys_acl_set_permset() directly rather than via the VFS
Andrew Bartlett [Mon, 13 Aug 2012 10:00:21 +0000 (20:00 +1000)]
s3-smbd: Call sys_acl_set_permset() directly rather than via the VFS

This will allow us to remove the struct smb_acl_t manipuations from the VFS layer,
which will be reduced to handling the get/set functions.

Andrew Bartlett

7 years agos3-smbd: Call sys_acl_set_qualifier() directly rather than via the VFS
Andrew Bartlett [Mon, 13 Aug 2012 09:59:54 +0000 (19:59 +1000)]
s3-smbd: Call sys_acl_set_qualifier() directly rather than via the VFS

This will allow us to remove the struct smb_acl_t manipuations from the VFS layer,
which will be reduced to handling the get/set functions.

Andrew Bartlett

7 years agos3-smbd: Call sys_acl_set_tag_type() directly rather than via the VFS
Andrew Bartlett [Mon, 13 Aug 2012 09:59:33 +0000 (19:59 +1000)]
s3-smbd: Call sys_acl_set_tag_type() directly rather than via the VFS

This will allow us to remove the struct smb_acl_t manipuations from the VFS layer,
which will be reduced to handling the get/set functions.

Andrew Bartlett

7 years agos3-smbd: Call sys_acl_create_entry() directly rather than via the VFS
Andrew Bartlett [Mon, 13 Aug 2012 09:59:03 +0000 (19:59 +1000)]
s3-smbd: Call sys_acl_create_entry() directly rather than via the VFS

This will allow us to remove the struct smb_acl_t manipuations from the VFS layer,
which will be reduced to handling the get/set functions.

Andrew Bartlett

7 years agos3-smbd: Call sys_acl_add_perm() directly rather than via the VFS
Andrew Bartlett [Mon, 13 Aug 2012 09:58:17 +0000 (19:58 +1000)]
s3-smbd: Call sys_acl_add_perm() directly rather than via the VFS

This will allow us to remove the struct smb_acl_t manipuations from the VFS layer,
which will be reduced to handling the get/set functions.

Andrew Bartlett

7 years agos3-smbd: Call sys_acl_clear_perms() directly rather than via the VFS
Andrew Bartlett [Mon, 13 Aug 2012 09:57:34 +0000 (19:57 +1000)]
s3-smbd: Call sys_acl_clear_perms() directly rather than via the VFS

This will allow us to remove the struct smb_acl_t manipuations from the VFS layer,
which will be reduced to handling the get/set functions.

Andrew Bartlett

7 years agos3-smbd: Call sys_acl_init() directly rather than via the VFS
Andrew Bartlett [Mon, 13 Aug 2012 09:56:31 +0000 (19:56 +1000)]
s3-smbd: Call sys_acl_init() directly rather than via the VFS

This will allow us to remove the struct smb_acl_t manipuations from the VFS layer,
which will be reduced to handling the get/set functions.

Andrew Bartlett

7 years agos3-smbd: Call sys_acl_free_acl() directly rather than via the VFS
Andrew Bartlett [Mon, 13 Aug 2012 09:55:25 +0000 (19:55 +1000)]
s3-smbd: Call sys_acl_free_acl() directly rather than via the VFS

This will allow us to remove the struct smb_acl_t manipuations from the VFS layer,
which will be reduced to handling the get/set functions.

Andrew Bartlett

7 years agos3-smbd: Call sys_acl_free_qualifier() directly rather than via the VFS
Andrew Bartlett [Mon, 13 Aug 2012 09:53:41 +0000 (19:53 +1000)]
s3-smbd: Call sys_acl_free_qualifier() directly rather than via the VFS

This will allow us to remove the struct smb_acl_t manipuations from the VFS layer,
which will be reduced to handling the get/set functions.

Andrew Bartlett

7 years agos3-smbd: Call sys_acl_get_entry() directly rather than via the VFS
Andrew Bartlett [Mon, 13 Aug 2012 09:52:23 +0000 (19:52 +1000)]
s3-smbd: Call sys_acl_get_entry() directly rather than via the VFS

This will allow us to remove the struct smb_acl_t manipuations from the VFS layer,
which will be reduced to handling the get/set functions.

Andrew Bartlett

7 years agos3-smbd: Call sys_acl_free_qualifier() directly rather than via the VFS
Andrew Bartlett [Mon, 13 Aug 2012 09:47:16 +0000 (19:47 +1000)]
s3-smbd: Call sys_acl_free_qualifier() directly rather than via the VFS

This will allow us to remove the struct smb_acl_t manipuations from the VFS layer,
which will be reduced to handling the get/set functions.

Andrew Bartlett

7 years agos3-smbd: Call sys_acl_get_qualifier() directly rather than via the VFS
Andrew Bartlett [Mon, 13 Aug 2012 09:43:23 +0000 (19:43 +1000)]
s3-smbd: Call sys_acl_get_qualifier() directly rather than via the VFS

This will allow us to remove the struct smb_acl_t manipuations from the VFS layer,
which will be reduced to handling the get/set functions.

Andrew Bartlett