VERSION: Disable git snapshots for the 4.1.5 release.

Signed-off-by: Karolin Seeger <kseeger@samba.org>

WHATSNEW: Add release notes for Samba 4.1.5.

Signed-off-by: Karolin Seeger <kseeger@samba.org>

s3:smb2_notify: fix use after free on long living notify requests

This is a hack, but it should fix the bug:

   change_notify_add_request() talloc moves smb_request away,
   which is not expected by the smb2_notify.c code...

   smbd_smb2_notify_reply() uses tevent_req_defer_callback()
   (in older versions an immediate event) to defer the response.
   This is needed as change_notify_reply() will do more things
   after calling reply_fn() (smbd_smb2_notify_reply is this case)
   and often change_notify_remove_request() is called after
   change_notify_reply().

   change_notify_remove_request() implicitly free's the smb_request
   that was passed to change_notify_add_request().

   smbd_smb2_fake_smb_request() added the smb_request as smb2req->smb1req,
   which is expected to be available after smbd_smb2_notify_recv() returned.

The long term solution would be the following interface:

struct tevent_req *change_notify_request_send(TALLOC_CTX *mem_ctx,
                                              struct tevent_context *ev,
                                              struct files_struct *fsp,
                                              uint32_t max_length,
                                              uint32_t filter,
                                              bool recursive);
NTSTATUS change_notify_request_recv(struct tevent_req *req,
                                    TALLOC_CTX *mem_ctx,
                                    DATA_BLOB *buffer);

Bug: https://bugzilla.samba.org/show_bug.cgi?id=10442

Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Stefan Metzmacher <metze@samba.org>
Autobuild-Date(master): Fri Feb 14 11:18:15 CET 2014 on sn-devel-104
(cherry picked from commit e0bf930f23fe20ee00d0006a5f6c2ba1a8f592a0)

Autobuild-User(v4-1-test): Karolin Seeger <kseeger@samba.org>
Autobuild-Date(v4-1-test): Sun Feb 16 19:14:12 CET 2014 on sn-devel-104

s3: modules: streaminfo: As we have no VFS function SMB_VFS_LLISTXATTR we can't cope with a symlink when lp_posix_pathnames() is true.

Fix bug : Bug 10429 - samba returns STATUS_OBJECT_NAME_NOT_FOUND when attempting to remove dangling symlink

https://bugzilla.samba.org/show_bug.cgi?id=10429

Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Jeff Layton <jlayton@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Sat Feb  8 00:01:16 CET 2014 on sn-devel-104
(cherry picked from commit 17adbbcad7e401dd544dfa76f7ec9aeb6a847381)

Autobuild-User(v4-1-test): Karolin Seeger <kseeger@samba.org>
Autobuild-Date(v4-1-test): Fri Feb 14 22:30:55 CET 2014 on sn-devel-104

s3: vfs_dirsort module.

Add raw.search torture test on a share definition with:

vfs objects = dirsort

https://bugzilla.samba.org/show_bug.cgi?id=10406

Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
(cherry picked from commit 30cc8f41da58b87ff575860b7cde640520829923)

s3: vfs_dirsort module.

Allow dirsort to work when multiple simultaneous
directories are open. The old code only keeps one
active private data pointer on the connection struct, opening
a second directory on the same connection will overwrite it.

This modification turns the private data pointer
into a linked list of open directories on the
connection struct, and finds the correct one by searching
on the passed in DIR *.

With this code in place, smbd passes raw.search
torture test on a share definition with:

vfs objects = dirsort

https://bugzilla.samba.org/show_bug.cgi?id=10406

Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
(cherry picked from commit fd79652b0e598882e0c4d156bd897c29dae8ec04)

smbd: Fix an ancient oplock bug

If we get an oplock break response, we forgot to remove the oplock break
timeout.

Found by stopping raw.oplock.exclusive5 after the 2nd open and watching a debug
level 10 log. This amends 08a9de89 from 2007.

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
(cherry picked from commit 0670975261c5f29394f9e9d25f899a7de948dad1)

Fix bug #10436 - cancel fsp->oplock_timeout in downgrade_file_oplock().

vfs_btrfs: pass-through copy-chunk(len=0) requests

Never map copy-chunk(len=0) requests to BTRFS_IOC_CLONE_RANGE ioctls. A
BTRFS_IOC_CLONE_RANGE with @src_length=0 results in a clone of all data
from @src_offset->EOF!

BUG: https://bugzilla.samba.org/show_bug.cgi?id=10424

Signed-off-by: David Disseldorp <ddiss@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
(cherry picked from commit 3be664969d4de41ebb4778caabce8bcf5e303064)

smbd/smb2_ioctl: fail zero length copy chunk requests

As documented in MS-SMB2 3.3.5.15.6 Handling a Server-Side Data Copy
Request, an invalid parameter response should be sent when:

The Length value in a single chunk is greater than
ServerSideCopyMaxChunkSize or *equal to zero*.

We do not currently abide by the latter part of this clause.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=10424

Signed-off-by: David Disseldorp <ddiss@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
(cherry picked from commit 00906f9604ad3e633e3d3cbc8d9dc4e2e305a455)

torture: add zero length FSCTL_SRV_COPYCHUNK test

Windows Server 2012 returns NT_STATUS_INVALID_PARAMETER for
FSCTL_SRV_COPYCHUNK requests that include a server-side copy length of
zero, in line with MS-SMB2 3.3.5.15.6.

We should match this behaviour, so test for it.

Signed-off-by: David Disseldorp <ddiss@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
(cherry picked from commit 54d07da81e181072b530e88b42d0d0d17fe60df0)

kdc: Add belts-and-braces check that we fail if the hdb version changes

This checks both if host system run-time Heimdal has changed version,
and that the build-time version is supported.

Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-By: Jelmer Vernooij <jelmer@samba.org>
Autobuild-User(master): Andrew Bartlett <abartlet@samba.org>
Autobuild-Date(master): Mon Jan 20 22:26:49 CET 2014 on sn-devel-104

(cherry picked from commit e758f4111356fafce5b5c8393648c9ea5c400601)

The last 3 patches address bug #10418 - INTERNAL ERROR: Signal 11 in the kdc
pid.

Support for Heimdal's unified krb5 and hdb plugin system.

Fixes exportkeytab and a kdc crash when building against heimdal master.

Bug-Debian: http://bugs.debian.org/732342

Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Jelmer Vernooij <jelmer@samba.org>
(cherry picked from commit 368208069e33a6d9b78a90f7326d5ac1bdd204f7)

Cope with first element in hdb_method having a different name in different heimdal versions.

It's called `interface_version` in older Heimdal versions and
`version` in newer versions.

Signed-off-by: Jelmer Vernooij <jelmer@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Autobuild-User(master): Andrew Bartlett <abartlet@samba.org>
Autobuild-Date(master): Thu Nov 28 04:17:55 CET 2013 on sn-devel-104

(cherry picked from commit 92489bfed4e6350b6858d5e39d538f41768912ae)

smbd: Fix memory overwrites

SIVAL writes 32 bit, not 16

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Michael Adam <obnox@samba.org>
Autobuild-User(master): Michael Adam <obnox@samba.org>
Autobuild-Date(master): Fri Feb  7 20:07:23 CET 2014 on sn-devel-104

(cherry picked from commit 9088bde059e93a84745ec2158e2e640b5bb13844)

Fix bug #10415 - *** glibc detected *** /usr/sbin/smbd: free(): invalid next
size (fast).

s3-winbind: Improve performance of wb_fill_pwent_sid2uid_done().

BUG: https://bugzilla.samba.org/show_bug.cgi?id=2191

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
Autobuild-User(master): Andreas Schneider <asn@cryptomilk.org>
Autobuild-Date(master): Thu Jan 16 20:17:24 CET 2014 on sn-devel-104

(cherry picked from commit 1a43778433934530d77791edd1af538de8b1d8a3)
Signed-off-by: Andreas Schneider <asn@samba.org>
Autobuild-User(v4-1-test): Karolin Seeger <kseeger@samba.org>
Autobuild-Date(v4-1-test): Thu Feb  6 13:59:54 CET 2014 on sn-devel-104

Stop use after free

Fixes bug #10087

Thanks to Man Min Yan for their analysis and providing a solution to the issue.

Reviewed-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Fri Sep 27 14:29:46 CEST 2013 on sn-devel-104
(cherry picked from commit 6bf9a774718917c3429fa1492f5b0268ae5e01c3)

Autobuild-User(v4-1-test): Karolin Seeger <kseeger@samba.org>
Autobuild-Date(v4-1-test): Wed Feb  5 13:12:49 CET 2014 on sn-devel-104

s3: smbpasswd - fix crashes on invalid input.

get_pass can return NULL on error. Ensure that
this is always the case and fix all callers to cope
(some already did).

Reported by Joonas Kuorilehto <joneskoo@codenomicon.com>

BUG: https://bugzilla.samba.org/show_bug.cgi?id=10320

Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
Autobuild-User(master): Volker Lendecke <vl@samba.org>
Autobuild-Date(master): Mon Dec 16 15:17:58 CET 2013 on sn-devel-104

(cherry picked from commit ef5a3bedab74420baf0c653cf8e304fe6c2a13b4)

Autobuild-User(v4-1-test): Karolin Seeger <kseeger@samba.org>
Autobuild-Date(v4-1-test): Sat Jan 25 18:16:55 CET 2014 on sn-devel-104

s3:dir - We now pass the previously spinning directory tests on ext4.

https://bugzilla.samba.org/show_bug.cgi?id=2662

Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
Autobuild-User(master): Andreas Schneider <asn@cryptomilk.org>
Autobuild-Date(master): Wed Jan 15 11:39:12 CET 2014 on sn-devel-104

(cherry picked from commit 0f9a189e36d8e30dfd40e42130329a0984938ddd)

Autobuild-User(v4-1-test): Karolin Seeger <kseeger@samba.org>
Autobuild-Date(v4-1-test): Fri Jan 17 11:10:43 CET 2014 on sn-devel-104

s3:dir - Introduce a 64-bit directory offset <-> 32 bit wire offset map using memcache.

Should fix the DOS clients against 64-bit smbd's bug.

https://bugzilla.samba.org/show_bug.cgi?id=2662

Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
(cherry picked from commit 97cd9c6729a3821faa2dbc1588a40c5e03b9fd4f)

s3:dir - Add a new memcache type (non-talloc) - SMB1_SEARCH_OFFSET_MAP.

We will use this in mapping 64-bit directory offset
cookies to a 32-bit counter.

https://bugzilla.samba.org/show_bug.cgi?id=2662

Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
(cherry picked from commit 4e0c41a321b2683610748c8c176fc46aaa8d114d)

s3:dir - Map wire offsets to native directory cookies.

Take care of the special offsets.

https://bugzilla.samba.org/show_bug.cgi?id=2662

Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
(cherry picked from commit 42c80358c83dca65cdde78f442056ec0f55ecbb1)

s3:dir - Cope with fixed mapping of 'special' values.

https://bugzilla.samba.org/show_bug.cgi?id=2662

Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
(cherry picked from commit 81df4123ca6fae6e9d901c59a12407f3f89dc335)

s3: dir - Introduce 32-bit wire versions of the 'special' values.

https://bugzilla.samba.org/show_bug.cgi?id=2662

Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
(cherry picked from commit 51a115b62048735b4c8ec79211ce45600cfa5c01)

s3:dir - Introduce a function to map a directory cookie to a 32-bit wire cookie.

Make this an identity for now.

https://bugzilla.samba.org/show_bug.cgi?id=2662

Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
(cherry picked from commit 5afc25eceb0c0e031bbe162617309178f3bcc425)

s3:dir - In the old SMB1 search code, rename offset to wire_offset to distinguish between wire and native offsets.

Rename uint32 type to correct uint32_t.

https://bugzilla.samba.org/show_bug.cgi?id=2662

Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
(cherry picked from commit 584de2078d3320ba2e232e5f504191616347d0d7)

vfs/glusterfs: in case atime is not passed, set it to the current atime

The Linux CIFS client does not pass an updated atime when a write() is
done. This causes the vfs/glusterfs module to set the atime to -1 on the
Gluster backend, resulting in an atime far in the future (year 2106).

Signed-off-by: Niels de Vos <ndevos@redhat.com>
Reviewed-by: Ira Cooper <ira@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Wed Jan 15 21:31:30 CET 2014 on sn-devel-104

Fix bug #10384 - [BACKPORT REQUEST] vfs/glusterfs: in case atime is not passed,
set it to the current atime.

s3-passdb: Fix string duplication to pointers.

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Günther Deschner <gd@samba.org>
Autobuild-User(master): Andreas Schneider <asn@cryptomilk.org>
Autobuild-Date(master): Thu Jan  9 22:35:25 CET 2014 on sn-devel-104

(cherry picked from commit bff3ac250e9d4e7d91820eb53c28257aa38fff88)

The last 4 patches address bug #10367 - Fix several memory leaks.

Autobuild-User(v4-1-test): Karolin Seeger <kseeger@samba.org>
Autobuild-Date(v4-1-test): Mon Jan 13 12:09:26 CET 2014 on sn-devel-104

wbinfo: Fix a memory leak in wbinfo_ping_dc().

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Günther Deschner <gd@samba.org>
(cherry picked from commit 541164d47a86bab90ef96a9be40b8c0997abdd61)

s3-libads: Fix memory leaks in ads_build_path().

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Günther Deschner <gd@samba.org>
(cherry picked from commit c8371b4ec12f2dea6ce18724de59a23e04826c1d)

lib: Fix strict-aliasing warning in md5 code.

If the compiler detects strict aliasing problems it isn't able to
optimize the code.

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Günther Deschner <gd@samba.org>
(cherry picked from commit 615efa4ae84373ae8aefb36fcf7583338665429a)

shadow_copy2: add a comment explaining why we don't talloc_zero_array().

Since I stumbled over this slighly sublte point, I thought it is
worthwile to point it our in a comment.

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
Autobuild-User(master): Michael Adam <obnox@samba.org>
Autobuild-Date(master): Wed Dec 11 19:55:20 CET 2013 on sn-devel-104

(cherry picked from commit 27baff0ec96cded0446ecd7739e9d31aaeb90868)

The last 36 patches address bug #10259 - shadow_copy2 module "Previous Version"
not working in Windows 7.

shadow_copy2: revert expensive and unnecessary zero-initialization

I was being overly cautious. This is initialization is not
necessary, since further down in the for-loop, the memory
always gets fully initialized because the insert string is
inserted at various slash positions.

So this talloc_zero_array can be skipped: this an expensive
thing to do in virtually every VFS call.

This essentially reverts commit 249e9b4a34d8959bd94735c1921ecfc24d6a2705.

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
(cherry picked from commit 71318d7d685f8e1e112076f84ef2d788acb83547)

docs: Fix typos in vfs_shadow_copy2.8.xml.

Signed-off-by: Karolin Seeger <kseeger@samba.org>
Reviewed-by: David Disseldorp <ddiss@samba.org>
Autobuild-User(master): David Disseldorp <ddiss@samba.org>
Autobuild-Date(master): Tue Dec 10 18:49:21 CET 2013 on sn-devel-104

(cherry picked from commit c16afddeef2938e218df63b1c669caa5fb65d39c)

docs: update the manpage of vfs_shadow_copy2

Document the configuration and all the options
available for the shadow_copy2 module.

Pair-Programmed-With: Björn Baumbach <bb@sernet.de>

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Autobuild-User(master): Andrew Bartlett <abartlet@samba.org>
Autobuild-Date(master): Fri Dec  6 22:26:31 CET 2013 on sn-devel-104
(cherry picked from commit f6ac6f20540f81257e1f180454b6a2c1239e85fa)

s3:modules:shadow_copy2: remove redundant documentation comment block

and refer to the manual page instead

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
(cherry picked from commit 6685e6512e03d9420439a730a40fcca3411a48db)

s3:modules:shadow_copy2: improve headline comment

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
(cherry picked from commit bffaf17d482cb1ad88698ea1a00aa7e4ddb0c0e7)

s3:module:shadow_copy2: add my (C)

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
(cherry picked from commit b5b5674287c935bf923cf570cf218ffb0ae7ce78)

shadow_copy2: use stored mount_point instead of recalculating.

In the case of snapdirseverywhere but NOT crossmountpoints.

This spares stat calls.
And is the only correct thing to do if the mount point was
specified in the configuration.

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
(cherry picked from commit 6ccbc1347d3240fc3c874a1957b654456fb6234c)

shadow_copy2: improve debug in shadow_copy2_convert() in snapdirseverywhere mode

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
(cherry picked from commit 935120dbc01890e313e2902b13141d4a13fe96d5)

shadow_copy2: fix shadow_copy2_convert() in the classical case.

I.e. the non-snapdirseverywhere case.
This in particular fixes the case of a snapdir hierarchy
that is parallel to the share or mountpoint and not subordinate.

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
(cherry picked from commit 9ab89371c8eddad2f274736b508866e2a92b74a3)

shadow_copy2: add some blank lines for visual separation to shadow_copy2_convert()

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
(cherry picked from commit 86988db1f0ebd170d2bc91b6ed78f8845bfd270c)

shadow_copy2: initialize "converted" string to null in shadow_copy2_convert()

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
(cherry picked from commit 249e9b4a34d8959bd94735c1921ecfc24d6a2705)

shadow_copy2: fix shadow_copy2_strip_snapshot() in the classical case

I.e., fix detection of already converted names.

This is done by using the shadow_copy2_snapshot_path() function
and comparing if the input string starts with that.

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
(cherry picked from commit 4b8d9c65f45db426716763629f1e22b0eb859a99)

shadow_copy2: add some debug to shadow_copy2_strip_snapshot()

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
(cherry picked from commit afcb84e69f46671030710bdda1c8798235b9ace7)

shadow_copy2: add comments explaining decisions in shadow_copy2_strip_snapshot()

This should make it more easy to understand what the cases are.

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
(cherry picked from commit 3d053b1ab3e0b918509e06086a54834a9ae9cdb7)

shadow_copy2: introduce shadow_copy2_snapshot_path()

This builds the posix snapshot path for the connection
at the provided timestamp. For the non-snapdirseverywhere case.

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
(cherry picked from commit 38e108563d7e9e14203bf4dabfda09bd1408e980)

shadow_copy2: factor shadow_copy2_posix_gmt_string() out of shadow_copy2_insert_string()

for re-use..

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
(cherry picked from commit 4cc5140bbe207f08d9b4fb0d119e74cc839e55dd)

shadow_copy2: shadow_copy2_insert_string(): do not prepend a "/" in absolute mode

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
(cherry picked from commit 304a0f531caa5f33f205739470f17e983d25a6b5)

shadow_copy2: make shadow_copy2_find_snapdir() return const char *

instead of char *. This eliminates compiler warnings.
snapdir is a const string in all occasions.

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
(cherry picked from commit e86923eb52633c5b6133c45678355ce69bb43a54)

shadow_copy2: in the classical case, use configured path in shadow_copy2_find_snapdir()

There is no point in searching for snapdir if not in snapdirseverywhere mode.

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
(cherry picked from commit dbdb436a80e2fb75d9fd8ae17192702123c3b530)

shadow_copy2: implement disk_free

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
(cherry picked from commit ea898ea1ac1cc9364c5b7396db3902aeb114cfb8)

shadow_copy2: log resulting config at the end of shadow_copy2_connect()

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
(cherry picked from commit c4f9954ebb04da94a5bcd2cb328fb2fbaf9fa062)

shadow_copy2: add snapshot_basepath to the config.

This is the absolute version of snapdir.

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
(cherry picked from commit a7ca55c8da6fbe6452a7a0bfc3c84b5941b8aa27)

shadow_copy2: add rel_connectpath to config.

This is the share root, relative to the basedir.

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
(cherry picked from commit d34dc1b0025d18afc9ce638c7000b702f98b5d03)

shadow_copy2: introduce "shadow:mountpoint" option

Possiblity to explicitly set the share's mount point.
This is useful mainly for debugging and testing purposes.

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
(cherry picked from commit 2d5a3af4bc44d13877a26fa1857b3ceafe138de8)

shadow_copy2: re-add the basedir option.

Disable basedir if it is not an absolute path or if
snapdirseverywhere or crossmountpoints is enabled.

Pair-Programmed-With: Björn Baumbach <bb@sernet.de>

Signed-off-by: Michael Adam <obnox@samba.org>
Signed-off-by: Björn Baumbach <bb@sernet.de>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
(cherry picked from commit ed751b9ee49d8e4a319759640321e8b49be4f154)

shadow_copy2: disable "snapdir:crossmountpoints" if the snapdir is absolute.

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
(cherry picked from commit 1e887fcda097b93a879df726f2b7c2c8d3c4cf54)

shadow_copy2: introduce the bool "snapdir_absolute" in the config.

Not exposed but to be used internally.

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
(cherry picked from commit 843954989cbec6640d2565d0d23a48f296740a23)

shadow_copy2: introduce config struct and function shadow_copy2_connect()

This moves the parsing of the config to a central place.
So users of configuation don't need to call lp_parm_... all the time.

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
(cherry picked from commit 1ecef5743583cf617f5506bc2fca3baa70cfb9b3)

shadow_copy2: add comment explaining the SMB level GMT format pattern

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
(cherry picked from commit 6da7375cd881f85f2873578db7fcfb368deab94f)

shadow_copy2: add comment block explaining shadow_copy2_convert()

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
(cherry picked from commit 5c900fd930edd45e9f23b36c1e68e5c2d8b96867)

shadow_copy2: add comment block explaining shadow_copy2_insert_string()

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
(cherry picked from commit 5da5512985cf65c09abb33abaf5e8dc28167dac3)

shadow_copy2: add comment block explaining shadow_copy2_find_snapdir()

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
(cherry picked from commit b90d1e6ac06fd4c1aaaceadcb711800499334117)

shadow_copy2: add header comment explaining have_snapdir()

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
(cherry picked from commit 5b494b3dea559f632d57c9d33172e46e459e852f)

shadow_copy2: add comment header describing shadow_copy2_strip_snapshot()

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
(cherry picked from commit 9361824ddd291cb0e543a5a0829246831fcb9e84)

shadow_copy2: break overly long lines in shadow_copy2_snapshot_to_gmt()

According to coding guidelines.

Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
(cherry picked from commit 01cb88977da5bc44443407b100345531d047c77c)

s3: winbindd: Move calling setup_domain_child() into add_trusted_domain().

Ensure it only gets called when a new domain is allocated
and added to the list.

This should fix problems with the previous logic where
setup_domain_child() was called in places where an existing
domain was returned.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=10358

Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Wed Jan  8 20:46:55 CET 2014 on sn-devel-104

(cherry picked from commit ca931e460460ffe46735f98b31db47220772d566)

Autobuild-User(v4-1-test): Karolin Seeger <kseeger@samba.org>
Autobuild-Date(v4-1-test): Fri Jan 10 11:42:44 CET 2014 on sn-devel-104

s3: winbindd: Move the logic of whether to set 'domain->primary' into add_trusted_domain().

BUG: https://bugzilla.samba.org/show_bug.cgi?id=10358

Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
(cherry picked from commit cfa6a36467f5679a88e49419e8af32b724c242bd)

VERSION: Bump version number up to 4.1.5...

and re-enable git snapshots.

Signed-off-by: Karolin Seeger <kseeger@samba.org>

VERSION: Disable git snapshots for the 4.1.4 release.

Signed-off-by: Karolin Seeger <kseeger@samba.org>

WHATSNEW: Add release notes for Samba 4.1.4.

Signed-off-by: Karolin Seeger <kseeger@samba.org>

s3:winbindd fix use of uninitialized variables

Bug: https://bugzilla.samba.org/show_bug.cgi?id=10280

Signed-off-by: Christian Ambach <ambi@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
(cherry picked from commit 7393781a57891687b464762b0954e6c936f750bb)

Autobuild-User(v4-1-test): Karolin Seeger <kseeger@samba.org>
Autobuild-Date(v4-1-test): Thu Jan  2 21:49:28 CET 2014 on sn-devel-104

vfs_glusterfs: Enable per client log file

In Samba configuration file, one of the options of gluster type is
log file, the value of this option was not allowed to contain any
variables, as a result all the clients would have a single log file,
which complicated debugging.
In this patch, variable substitution is performed for gluster log file.
Hence allowing user to customise the gluster log file name.

Signed-off-by: Poornima Gurusiddaiah <pgurusid@redhat.com>
Reviewed-by: Ira Cooper <ira@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Tue Dec 17 23:44:16 CET 2013 on sn-devel-104

Fix bug #10337 - vfs_glusterfs: Enable per client log file.

Autobuild-User(v4-1-test): Karolin Seeger <kseeger@samba.org>
Autobuild-Date(v4-1-test): Thu Dec 26 21:21:07 CET 2013 on sn-devel-104

ldb: bad if test in ldb_comparison_fold()

Found by David Binderman <dcb314@hotmail.com>

BUG: https://bugzilla.samba.org/show_bug.cgi?id=10305

Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Michael Adam <obnox@samba.org>
Autobuild-User(master): Michael Adam <obnox@samba.org>
Autobuild-Date(master): Sat Dec  7 11:10:47 CET 2013 on sn-devel-104

s3-lib: Fix %G substitution for domain users in smbd

BUG: https://bugzilla.samba.org/show_bug.cgi?id=10286
Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Christian Ambach <ambi@samba.org>
Autobuild-User(master): Christian Ambach <ambi@samba.org>
Autobuild-Date(master): Tue Dec 10 16:39:43 CET 2013 on sn-devel-104

(cherry picked from commit 8eef4ab79ec5fb7e96ad2f2ad6c9bf30db13a50d)
Signed-off-by: Andreas Schneider <asn@samba.org>
Autobuild-User(v4-1-test): Karolin Seeger <kseeger@samba.org>
Autobuild-Date(v4-1-test): Sat Dec 14 23:47:20 CET 2013 on sn-devel-104

smbtorture: New torture test for bug #9870.

Not fetching the latest modification time on a folder if we have read locks on it.

Prove we should just rely on the mtime value from the underlying
filesystem, even with an open handle.

  BUG: https://bugzilla.samba.org/show_bug.cgi?id=9870

Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
Autobuild-User(master): Volker Lendecke <vl@samba.org>
Autobuild-Date(master): Thu Dec  5 10:05:06 CET 2013 on sn-devel-104

BUG: https://bugzilla.samba.org/show_bug.cgi?id=9870
Not fetching the latest modification time on a folder if we have read locks on
it.

Autobuild-User(v4-1-test): Karolin Seeger <kseeger@samba.org>
Autobuild-Date(v4-1-test): Tue Dec 10 20:19:08 CET 2013 on sn-devel-104

smbd - allow updates on directory write times on open handles.

If we set a non-null 'old timestamp' in the share mode database
when creating a directory handle, this prevents mtime (write time)
updates from being seen by clients, as we will always return the
timestamp stored in the database whilst the handle is open.

For files this is ok, as we update the stored timestamp
ourselves when we write to the handle. For directories
we should just rely on the mtime value from the underlying
filesystem.

Torture test to follow.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=9870

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

s3-winbindd: Fix DEBUG statement in winbind_msg_offline().

Guenther

BUG: https://bugzilla.samba.org/show_bug.cgi?id=10285

Reviewed-by: Günther Deschner <gd@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
Autobuild-User(master): Günther Deschner <gd@samba.org>
Autobuild-Date(master): Mon Dec  9 17:28:37 CET 2013 on sn-devel-104

Autobuild-User(v4-1-test): Karolin Seeger <kseeger@samba.org>
Autobuild-Date(v4-1-test): Tue Dec 10 15:58:13 CET 2013 on sn-devel-104

smbd: Fix a panic when a smb2 brlock times out

Found by Peter Somogyi.

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Thu Dec  5 21:21:35 CET 2013 on sn-devel-104

Fix bug #10311 - SMB2 server can panic when a smb2 brlock times out.

selftest: Remove samba3.smb2.lock.*.rw-exclusive from flapping file

This test demonstrates a problem with byte range locks and AIO.

Signed-off-by: Christof Schmitt <cs@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Fri Dec  6 05:19:37 CET 2013 on sn-devel-104
(cherry picked from commit 8c3bf7b84950fbb0305bcccd49ecfc202e08901a)

The last 5 patches address bug #10310 - Fix AIO with SMB2 and locks.

selftest: Run smb2.lock tests also against AIO share

Signed-off-by: Christof Schmitt <cs@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
(cherry picked from commit d551d5256f9b1ca57b8018d816ea665c9b847ced)

selftest: Introduce share for testing AIO

Signed-off-by: Christof Schmitt <cs@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
(cherry picked from commit 63727c15450e1db2be49ade758c369aa4599657a)

s3: Return correct error code from SMB2 AIO read failure

This is similar to commit 27e20d5d60ea8aa526bcb7c2dfc18dd2de0bb97b which
fixed the same case for SMB2 writes: When sending the AIO read fails,
return the real error instead of mapping it to NT_STATUS_FILE_CLOSED.

Signed-off-by: Christof Schmitt <cs@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
(cherry picked from commit eadb2a54d1733a482999eb770182156dad1e184d)

s3-aio: Use correct locking context for SMB2

The synchronous SMB2 reads and writes use open_persistent_id. The AIO
codepathes have to use the same, otherwise a write will conflict with a
lock on the same open file.

Signed-off-by: Christof Schmitt <cs@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
(cherry picked from commit dfef0701c398982226dde8a8e15ff97bba0fef53)

s3:smb2_server: avoid calling set_current_user_info() for each request

Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: David Disseldorp <ddiss@samba.org>
Autobuild-User(master): Stefan Metzmacher <metze@samba.org>
Autobuild-Date(master): Wed Nov 27 16:31:44 CET 2013 on sn-devel-104
(cherry picked from commit 3cc0651d9feda00b6a04f84b76744b2acc3a0446)

The last 6 patches address bug #10298 - smb2_server processing overhead.

s3:smb2_server: generate a header blob for the sendfile path

We need to pass the NBT header, SMB2 header and SMB2 Read header
as header blob to SMB_VFS_SENDFILE(). This allows the usage
of MSG_SEND or other tricks to avoid multiple TCP packets
on the wire.

Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: David Disseldorp <ddiss@samba.org>
(cherry picked from commit 36efaac2597d2d36826c02f23be15e7323b09784)

s3:smb2_server: allocate smbd_smb2_request on talloc_tos()

This matches the behavior for smb1 requests
and avoids an additional malloc() per request.

Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: David Disseldorp <ddiss@samba.org>
(cherry picked from commit 9d33a3f3e814e2924a423496ccc133c6c73fcd12)

s3:smb2_server: use tevent_req_notify_callback() in smbd_smb2_request_pending_queue()

If the request is already done we can avoid one iteration
of tevent_loop_once(), which means we avoids one
talloc_stackframe_pool/talloc_free pair.

Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: David Disseldorp <ddiss@samba.org>
(cherry picked from commit acfd4b068a5b99ac1d3fe716afff34cb7d2a0147)

s3:smb2_server: for performance reasons we use tevent_fd and readv/writev directly

Going via tevent_req_create/talloc_free at multiple layer costs
too much cpu cycles per request.

I tested downloading a 16GB (sparse) file with smbclient -b1 -mNT1,
and -mSMB2_02. Using smb2 max read = 64512, which means smb1 and smb2
will use the same read size.

I build with -O3 -g and compared the results with valgrind --tool=callgrind.

With -mNT1 the server uses about 2.000.000.000 cpu cycles.

This patch reduces the userspace cpu cycles for -mSMB2_02
from about ~ 8.000.000.000 down to ~ 4.000.000.000.

Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: David Disseldorp <ddiss@samba.org>
(cherry picked from commit 4244a2686cddcdc754c284df884ae497afa4053a)

s3:smb2_server: fix drain_socket error handling

smbd_smb2_request_error_ex() should return NTSTATUS and the caller
will terminate the connection.

Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: David Disseldorp <ddiss@samba.org>
(cherry picked from commit 9393e28df59954414313bfae70ffb796d3e332fe)

smbd: Always use UCF_PREP_CREATEFILE for filename_convert calls to resolve a path for open.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=10297

Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Mon Dec  9 21:02:21 CET 2013 on sn-devel-104

(cherry picked from commit f98d10af2a05f0261611f4cabdfe274cd9fe91c0)

smbd: change flag name from UCF_CREATING_FILE to UCF_PREP_CREATEFILE

In preparation to using it for all open calls.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=10297

Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
(cherry picked from commit 874318a97868e08837a1febb1be8e8a167b5ae0f)

smbd: Fix regression for the dropbox case.

We need to allow to save a file to a directory with perm -wx.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=10297

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
(cherry picked from commit 5b49fe24c906cbae12beff7a1b45de6809258cab)

lib/util: use proper include for struct stat

use <sys/stat.h> instead of forward declaring it
This fixes the build on AIX

Bug: https://bugzilla.samba.org/show_bug.cgi?id=10276

Signed-off-by: Christian Ambach <ambi@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
Autobuild-User(master): Christian Ambach <ambi@samba.org>
Autobuild-Date(master): Thu Nov 21 22:12:45 CET 2013 on sn-devel-104
(cherry picked from commit 4ea92841b72a7b701a7c0fadf63430a7d0a3fac8)

VERSION: Bump version up to 4.1.4.

Signed-off-by: Karolin Seeger <kseeger@samba.org>

Merge tag 'samba-4.1.3' into v4-1-test

samba: tag release samba-4.1.3

VERSION: Disable git snapshots for the 4.1.3 release.

Bug 10185 - CVE-2013-4408: DCERPC frag_len not checked
BUG: https://bugzilla.samba.org/show_bug.cgi?id=10185

Bug 10306 - CVE-2012-6150: Fail authentication if user isn't member of *any*
require_membership_of specified groups
BUG: https://bugzilla.samba.org/show_bug.cgi?id=10306
(BUG: https://bugzilla.samba.org/show_bug.cgi?id=10300)

Signed-off-by: Karolin Seeger <kseeger@samba.org>

WHATSNEW: Add release notes for Samba 4.1.3.

Bug 10185 - CVE-2013-4408: DCERPC frag_len not checked
BUG: https://bugzilla.samba.org/show_bug.cgi?id=10185

Bug 10306 - CVE-2012-6150: Fail authentication if user isn't member of *any*
require_membership_of specified groups
BUG: https://bugzilla.samba.org/show_bug.cgi?id=10306
(BUG: https://bugzilla.samba.org/show_bug.cgi?id=10300)

Signed-off-by: Karolin Seeger <kseeger@samba.org>

CVE-2012-6150: Fail authentication for single group name which cannot be converted to sid

furthermore if more than one name is supplied and no sid is converted
then also fail.

Bug: https://bugzilla.samba.org/show_bug.cgi?id=10300
Bug: https://bugzilla.samba.org/show_bug.cgi?id=10306

Signed-off-by: Noel Power <noel.power@suse.com>
Reviewed-by: Andreas Schneider <asn@samba.org>
Reviewed-by: David Disseldorp <ddiss@samba.org>
[ddiss@samba.org: fixed incorrect bugzilla tag I added to master commit]

CVE-2013-4408:s3:Ensure LookupRids() replies arrays are range checked.

Bug: https://bugzilla.samba.org/show_bug.cgi?id=10185

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