Signed-off-by: Jeremy Allison <jra@samba.org> Reviewed-by: Ralph Boehme <slow@samba.org>
Autobuild-User(master): Ralph Böhme <slow@samba.org>
Autobuild-Date(master): Sat Nov 11 20:12:26 CET 2017 on sn-devel-144
Jeremy Allison [Thu, 9 Nov 2017 20:48:15 +0000 (12:48 -0800)]
s3: smbd: kernel oplocks. Replace retry_open() with setup_kernel_oplock_poll_open().
If a O_NONBLOCK open fails with EWOULDBLOCK, this code changes smbd to
do a retry open every second, until either the timeout or we get a successful
open. If we're opening a file that has a kernel lease set by a non-smbd
process, this is the best we can do.
Prior to this, smbd would block on the second open on such a leased file
(not using O_NONBLOCK) which freezes active clients.
Joe Guo [Fri, 15 Sep 2017 04:13:26 +0000 (16:13 +1200)]
python: use communicate to fix Popen deadlock
`Popen.wait()` will deadlock when using stdout=PIPE and/or stderr=PIPE and the
child process generates large output to a pipe such that it blocks waiting for
the OS pipe buffer to accept more data. Use communicate() to avoid that.
Signed-off-by: Joe Guo <joeg@catalyst.net.nz> Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz> Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Autobuild-User(master): Andrew Bartlett <abartlet@samba.org>
Autobuild-Date(master): Thu Oct 19 09:27:16 CEST 2017 on sn-devel-144
Gary Lockyer [Wed, 16 Aug 2017 01:52:25 +0000 (13:52 +1200)]
blackbox tests: method to check specific exit codes
Signed-off-by: Gary Lockyer <gary@catalyst.net.nz> Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz> Reviewed-by: Garming Sam <garming@catalyst.net.nz>
(cherry picked from commit 74ebcf6dfc84b6aab6838fa99e12808eb6b913d9)
* Remove unused select backend
* Fix a race condition in tevent_threaded_schedule_immediate()
(bug #13130)
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Autobuild-User(master): Stefan Metzmacher <metze@samba.org>
Autobuild-Date(master): Mon Nov 13 18:02:46 CET 2017 on sn-devel-144
Volker Lendecke [Fri, 10 Nov 2017 20:22:26 +0000 (21:22 +0100)]
tevent: Fix a race condition
We can't rely on tctx to exist after we unlocked the mutex. It took a
while, but this does lead to data corruption. If *tctx is replaced with
something where tctx->wakeup_fd points to a real, existing file
descriptor, we're screwed. And by screwed, this means file corruption
on disk.
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): Sat Nov 11 03:20:09 CET 2017 on sn-devel-144
* make tevent_req_print() more robust against crashes
Signed-off-by: Stefan Metzmacher <metze@samba.org> Reviewed-by: Andreas Schneider <asn@samba.org>
(cherry picked from commit 7ad3a99d462e4a43c9f64b0877111d5e58f56e6e)
Autobuild-User(v4-7-test): Stefan Metzmacher <metze@samba.org>
Autobuild-Date(v4-7-test): Sun Jul 23 14:41:25 CEST 2017 on sn-devel-144
* Fix mutex locking in tevent_threaded_context_destructor().
* Fix a memleak on FreeBSD.
* Re-init threading in tevent_re_initialise().
* Include the finish location in tevent_req_default_print().
Signed-off-by: Stefan Metzmacher <metze@samba.org> Reviewed-by: Volker Lendecke <vl@samba.org>
Autobuild-User(master): Stefan Metzmacher <metze@samba.org>
Autobuild-Date(master): Thu Jun 22 17:17:33 CEST 2017 on sn-devel-144
which causes the immediate event to fire. This immediate
event is the pthread job terminate event, which was previously
set up in pthreadpool_tevent_job_signal() by:
198 if (state->tctx != NULL) {
199 /* with HAVE_PTHREAD */
200 tevent_threaded_schedule_immediate(state->tctx, state->im,
201 pthreadpool_tevent_job_done,
202 state);
So we now call pthreadpool_tevent_job_done() - which does:
225 TALLOC_FREE(state->tctx);
calling tevent_threaded_context_destructor():
384 ret = pthread_mutex_destroy(&tctx->event_ctx_mutex); <---------------- BOOM returns an error !
385 if (ret != 0) {
386 abort();
387 }
as we haven't gotten to line 481 above (the line after
HERE!!!!) so the tctx->event_ctx_mutex is still
locked when we try to destroy it.
So doing an additional:
ret = pthread_mutex_lock(&tctx->event_ctx_mutex);
ret = pthread_mutex_unlock(&tctx->event_ctx_mutex);
(error checking elided) forces tevent_threaded_context_destructor()
to wait until tctx->event_ctx_mutex is unlocked before it locks/unlocks
and then is guaranteed safe to destroy.
Signed-off-by: Volker Lendecke <vl@samba.org> Reviewed-by: Jeremy Allison <jra@samba.org> Reviewed-by: Stefan Metzmacher <metze@samba.org>
(cherry picked from commit 1fe7ec237a7036d76764ef1981de6b3000b2cfd3)
Volker Lendecke [Wed, 24 May 2017 14:22:34 +0000 (16:22 +0200)]
tevent: Fix a race condition in tevent context rundown
We protect setting tctx->event_ctx=NULL with tctx->event_ctx_mutex.
But in _tevent_threaded_schedule_immediate we have the classic
TOCTOU race: After we checked "ev==NULL", looking at
tevent_common_context_destructor the event context can go after
_tevent_threaded_schedule_immediate checked. We need to serialize
things a bit by keeping tctx->event_ctx_mutex locked while we
reference "ev", in particular in the
DLIST_ADD_END(ev->scheduled_immediates,im);
I think the locking hierarchy is still maintained, tevent_atfork_prepare()
first locks all the tctx locks, and then the scheduled_mutex. Also,
I don't think this will impact parallelism too badly: event_ctx_mutex
is only used to protect setting tctx->ev.
Found by staring at code while fixing the FreeBSD memleak due to
not destroying scheduled_mutex.
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): Fri Jun 9 00:45:26 CEST 2017 on sn-devel-144
Volker Lendecke [Fri, 10 Nov 2017 20:22:26 +0000 (21:22 +0100)]
tevent: Fix a race condition
We can't rely on tctx to exist after we unlocked the mutex. It took a
while, but this does lead to data corruption. If *tctx is replaced with
something where tctx->wakeup_fd points to a real, existing file
descriptor, we're screwed. And by screwed, this means file corruption
on disk.
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): Sat Nov 11 03:20:09 CET 2017 on sn-devel-144
Signed-off-by: Ralph Wuerthner <ralph.wuerthner@de.ibm.com> Reviewed-by: Jeremy Allison <jra@samba.org> Reviewed-by: Volker Lendecke <vl@samba.org>
Autobuild-User(master): Volker Lendecke <vl@samba.org>
Autobuild-Date(master): Sun Nov 5 12:31:12 CET 2017 on sn-devel-144
Ralph Wuerthner [Fri, 3 Nov 2017 22:33:28 +0000 (22:33 +0000)]
s3: smbd: Fix delete-on-close after smb2_find
Both dptr_create() and can_delete_directory_fsp() are calling OpenDir_fsp()
to get a directory handle. This causes an issue when delete-on-close is
set after smb2_find because both directory handle instances share the same
underlying file descriptor. In addition the SMB_ASSERT() in destructor
smb_Dir_destructor() gets triggered.
To avoid this use OpenDir() instead of OpenDir_fsp().
Jeremy Allison [Fri, 3 Nov 2017 19:02:17 +0000 (12:02 -0700)]
s4: torture: kernel_oplocks. Create a regression test case for bug #13058.
It implements the following test case:
1. client of smbd-1 opens the file and sets the oplock.
2. client of smbd-2 tries to open the file. open() fails(EAGAIN) and open is deferred.
3. client of smbd-1 sends oplock break request to the client.
4. client of smbd-1 closes the file.
5. client of smbd-1 opens the file and sets the oplock.
6. client of smbd-2 calls defer_open_done(), sees that the file lease was not changed
and does not reschedule open.
1. client of smbd-1 opens the file and sets the oplock.
2. client of smbd-2 tries to open the file. open() fails(EAGAIN) and open is deferred.
3. client of smbd-1 sends oplock break request to the client.
4. client of smbd-1 closes the file.
5. client of smbd-1 opens the file and sets the oplock.
6. client of smbd-2 calls defer_open_done(), sees that the file lease was not changed
and does not reschedule open.
and is no longer needed now vfs_streams_xattr.c no longer opens
the base file internally.
Signed-off-by: Andreas Schneider <asn@samba.org> Reviewed-by: Jeremy Allison <jra@samba.org>
(cherry picked from commit 16389bed0773952ca563b7bf1fecc2a737587257)
Signed-off-by: Andreas Schneider <asn@samba.org> Reviewed-by: Jeremy Allison <jra@samba.org>
(cherry picked from commit e56626e864492831a3dbbca2d4fb8f3281547a90)
Signed-off-by: Andreas Schneider <asn@samba.org> Reviewed-by: Jeremy Allison <jra@samba.org>
(cherry picked from commit 5274beba4cf722a34403dc07bf287815a6df6281)
Signed-off-by: Andreas Schneider <asn@samba.org> Reviewed-by: Jeremy Allison <jra@samba.org>
(cherry picked from commit d6a418c13f0a41851ecc0579765502e076a5cd3b)
Signed-off-by: Jeremy Allison <jra@samba.org> Reviewed-by: Andreas Schneider <asn@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Tue Oct 24 23:32:58 CEST 2017 on sn-devel-144
Jeremy Allison [Fri, 20 Oct 2017 22:09:38 +0000 (15:09 -0700)]
s3: smbclient: Ensure we call client_clean_name() before all operations on remote pathnames.
This allows names containing .. components to be resolved on the client side
before being sent to the server. Relative names work in SMB1 but not in SMB2.
Fix both client.c and clitar.c
Signed-off-by: Jeremy Allison <jra@samba.org> Reviewed-by: Andreas Schneider <asn@samba.org>
(cherry picked from commit f81c34c296f87127c6d1e4dd6ea74aa75660885d)
Jeremy Allison [Sat, 21 Oct 2017 00:08:08 +0000 (00:08 +0000)]
s3: client: Add new utility function client_clean_name().
Correctly canonicalizes a remote pathname removing '..'
elements before sending to a remote server. '..' elements
work in SMB1 pathnames, but not in SMB2.
Signed-off-by: Jeremy Allison <jra@samba.org> Reviewed-by: Andreas Schneider <asn@samba.org>
(cherry picked from commit d4d9d1941bdac9993968c34cf928c645e4152fd3)
Michael Adam [Fri, 20 Oct 2017 12:55:10 +0000 (14:55 +0200)]
vfs_glusterfs: Fix exporting subdirs with shadow_copy2
Since the glusterfs vfs module does not operate on a
locally mounted path, but on a "virtual" path starting
at the volume root, some assumptions of the code about
the vfs connect path fail. One example is the shadow_copy2
module which tries to detect the mount point from the
connectpath. In order to circumvent this problem, this
patch forces the "shadow:mountpoint" option to "/", which
skips the mount-point-detection code.
This patch will only have an effect if both the glusterfs
and the shadow_copy2 module are listed in vfs objects
in the right order, i.e. first shadow_copy2, and then
glusterfs.
Pair-Programmed-With: Anoop C S <anoopcs@redhat.com>
Signed-off-by: Michael Adam <obnox@samba.org> Signed-off-by: Anoop C S <anoopcs@redhat.com> Reviewed-by: Jeremy Allison <jra@samba.org>
(cherry picked from commit 1249801ec73c55449068dd72efde81ce6235ec8e)
Anoop C S [Fri, 13 Oct 2017 15:08:31 +0000 (20:38 +0530)]
vfs_fruit: Replace closedir() by SMB_VFS_CLOSEDIR
Pointer to directory 'dh' inside fruit_rmdir() is obtained using
SMB_VFS_OPENDIR. But this handle is closed directly by invoking
closedir() rather than SMB_VFS_CLOSEDIR. This will result in a
smbd crash if this handle was not obtained from local file system.
Therefore use SMB_VFS_CLOSEDIR corresponding to SMB_VFS_OPENDIR
to correctly close the directory handle.
Signed-off-by: Anoop C S <anoopcs@redhat.com> Reviewed-by: Ralph Boehme <slow@samba.org> Reviewed-by: Jeremy Allison <jra@samba.org> Reviewed-by: Guenther Deschner <gd@samba.org>
Autobuild-User(master): Günther Deschner <gd@samba.org>
Autobuild-Date(master): Mon Oct 16 19:56:55 CEST 2017 on sn-devel-144
Signed-off-by: Volker Lendecke <vl@samba.org> Reviewed-by: Ralph Boehme <slow@samba.org>
Autobuild-User(master): Ralph Böhme <slow@samba.org>
Autobuild-Date(master): Tue Oct 17 18:53:48 CEST 2017 on sn-devel-144
Signed-off-by: Stefan Metzmacher <metze@samba.org> Reviewed-by: Björn Jacke <bjacke@samba.org>
(cherry picked from commit 96e471eecce91e6cd9b92d854a3c6ca10e0634f3)
Signed-off-by: Stefan Metzmacher <metze@samba.org> Reviewed-by: Björn Jacke <bjacke@samba.org>
(cherry picked from commit 70146841272bc87c335bd24b736ba2c62efdfe06)
Signed-off-by: Ralph Boehme <slow@samba.org> Reviewed-by: Volker Lendecke <vl@samba.org>
Autobuild-User(master): Ralph Böhme <slow@samba.org>
Autobuild-Date(master): Fri Oct 13 21:44:02 CEST 2017 on sn-devel-144
Ralph Boehme [Wed, 11 Oct 2017 16:11:12 +0000 (18:11 +0200)]
vfs_fruit: fix ftruncating resource fork
fruit_ftruncate_rsrc_adouble() is called to effectively ftruncate() the
._ AppleDouble file to the requested size.
The VFS function SMB_VFS_NEXT_FTRUNCATE() otoh would attempt to truncate
to fsp *stream* in any way the next VFS module seems fit. As we know
we're stacked with a streams module, the module will attempt to truncate
the stream. So we're not truncating the ._ file.
This went unnoticed as the AppleDouble file header contains the
authorative resource fork size that was updated correctly.
Ralph Boehme [Wed, 11 Oct 2017 09:35:15 +0000 (11:35 +0200)]
vfs_catia: factor out mapping functions
This moves the core mapping functions to a seperate file and makes them
global.
string_replace_init_map() is called to parse a mapping in string and
produce a mapping object that can then be passed to
string_replace_allocate() to do the actual mapping of a string.
The only time a ctdb_queue is created without valid fd is when CTDB
is trying to establish connections with other nodes in the cluster.
All the other uses always create a ctdb_queue with valid fd.
This avoids queueing up packets for dead nodes or nodes that are not
running in the cluster and stops consuming memory.
Signed-off-by: Amitay Isaacs <amitay@gmail.com> Reviewed-by: Martin Schwenke <martin@meltin.net>
(cherry picked from commit ddd97553f0a8bfaada178ec4a7460d76fa21f079)
lib: gpo: Fixes issue with GPOPTIONS_BLOCK_INHERITANCE.
GP links with the GPOPTIONS_BLOCK_INHERITANCE option set
were blocking GPOs from the same link (i.e. an OU with
the flag set would block its own GPOs). This patch makes
sure the GPOs from the link are added to the list.
lib: gpo: Changes order to match GPO application order.
The order of GPOs in a gpo_list generated by ads_get_gpo_list
did not match the order of application. Since GPOs are pushed
to the FRONT of gpo_list, GPOs have to be pushed in the opposite
order of application. (Pushing to front is useful to get
inheritance blocking right).
Signed-off-by: Ralph Boehme <slow@samba.org> Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Ralph Böhme <slow@samba.org>
Autobuild-Date(master): Tue Oct 17 11:48:09 CEST 2017 on sn-devel-144
Ralph Boehme [Tue, 29 Aug 2017 13:55:19 +0000 (15:55 +0200)]
s3/smbd: fix access checks in set_ea_dos_attribute()
We wanted to set the DOS attributes and failed with permission denied
from the VFS/kernel/filesystem. Next thing we wanna do here is override
this if either
- "dos filemode = true" is set and the security descriptor gives the
user write access or if
- the stored security descriptor has FILE_WRITE_ATTRIBUTES
The former was working, but the latter was not implemented at all.
Signed-off-by: Samuel Cabrero <scabrero@suse.de> Reviewed-by: David Disseldorp <ddiss@samba.org> Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Fri Oct 6 05:21:25 CEST 2017 on sn-devel-144
Ralph Boehme [Sun, 27 Aug 2017 17:22:38 +0000 (19:22 +0200)]
s3/mdssvc: missing assignment in sl_pack_float
Spotted by -Werror=maybe-uninitialized:
../source3/rpc_server/mdssvc/marshalling.c: In function ‘sl_pack_float’:
../source3/rpc_server/mdssvc/marshalling.c:171:11: error:
‘ieee_fp_union.w’ may be used uninitialized in this function
[-Werror=maybe-uninitialized]
offset = sl_push_uint64_val(buf, offset, bufsize, ieee_fp_union.w);
Signed-off-by: Ralph Boehme <slow@samba.org> Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Sat Aug 26 05:05:08 CEST 2017 on sn-devel-144
Ralph Boehme [Thu, 13 Jul 2017 14:01:53 +0000 (16:01 +0200)]
selftest: enable kernel change notifications in the fileserver environment
This environment is currently not used for any test in the smb2
testsuite, so this change doesn't affect any existing test.
A subsequent commit will add a test for change notifications with kernel
change notify enabled. It verifies a bug (this one) that only crops up
in such a setup and causes rename events to get lost.
Signed-off-by: Ralph Boehme <slow@samba.org> Reviewed-by: Volker Lendecke <vl@samba.org> Reviewed-by: Andreas Schneider <asn@samba.org>
Autobuild-User(master): Ralph Böhme <slow@samba.org>
Autobuild-Date(master): Sat Oct 14 06:02:50 CEST 2017 on sn-devel-144
Signed-off-by: Ralph Boehme <slow@samba.org> Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Fri Oct 13 01:22:05 CEST 2017 on sn-devel-144
Ralph Boehme [Fri, 6 Oct 2017 13:25:54 +0000 (15:25 +0200)]
vfs_fake_acls: deny give-ownership
Windows doesn't allow giving ownership away unless the user has
SEC_PRIV_RESTORE privilege.
This follows from MS-FSA 2.1.5.1, so it's a property of the filesystem
layer, not the SMB layer. By implementing this restriction here, we can
now have test for this restriction.
Other filesystems may want to deliberately allow this behaviour --
although I'm not aware of any that does -- therefor I'm putting in this
restriction in the implementation of the chmod VFS function and not into
the caller.
projects / thirdparty / samba.git / log
