Jeremy Allison [Fri, 28 Feb 2020 23:59:16 +0000 (15:59 -0800)]
s3: VFS: vfs_glusterfs. Pass in struct vfs_gluster_fsync_state as the callback data to the subreq.
Find the req we're finishing off by looking inside vfs_gluster_fsync_state.
In a shutdown close the caller calls talloc_free(req), so we can't
access it directly as callback data.
The next commit will NULL out the vfs_gluster_fsync_state->req pointer
when a caller calls talloc_free(req), and the request is still in
flight.
Jeremy Allison [Fri, 28 Feb 2020 23:55:36 +0000 (15:55 -0800)]
s3: VFS: vfs_glusterfs. Protect vfs_gluster_pwrite_done() from accessing a freed req pointer.
If the fsp is forced closed by a SHUTDOWN_CLOSE whilst the
request is in flight (share forced closed by smbcontrol),
then we set state->req = NULL in the state destructor.
The existing state destructor prevents the state memory
from being freed, so when the thread completes and calls
vfs_gluster_pwrite_done(), just throw away the result if
state->req == NULL.
Jeremy Allison [Fri, 28 Feb 2020 23:53:19 +0000 (15:53 -0800)]
s3: VFS: vfs_glusterfs. Pass in struct vfs_gluster_pwrite_state as the callback data to the subreq.
Find the req we're finishing off by looking inside vfs_gluster_pwrite_state.
In a shutdown close the caller calls talloc_free(req), so we can't
access it directly as callback data.
The next commit will NULL out the vfs_gluster_pwrite_state->req pointer
when a caller calls talloc_free(req), and the request is still in
flight.
Jeremy Allison [Fri, 28 Feb 2020 23:38:04 +0000 (15:38 -0800)]
s3: VFS: vfs_glusterfs. Protect vfs_gluster_pread_done() from accessing a freed req pointer.
If the fsp is forced closed by a SHUTDOWN_CLOSE whilst the
request is in flight (share forced closed by smbcontrol),
then we set state->req = NULL in the state destructor.
The existing state destructor prevents the state memory
from being freed, so when the thread completes and calls
vfs_gluster_pread_done(), just throw away the result if
state->req == NULL.
Jeremy Allison [Fri, 28 Feb 2020 23:35:46 +0000 (15:35 -0800)]
s3: VFS: vfs_glusterfs. Pass in struct vfs_gluster_pread_state as the callback data to the subreq.
Find the req we're finishing off by looking inside vfs_gluster_pread_state.
In a shutdown close the caller calls talloc_free(req), so we can't
access it directly as callback data.
The next commit will NULL out the vfs_gluster_pread_state->req pointer
when a caller calls talloc_free(req), and the request is still in
flight.
Jeremy Allison [Fri, 28 Feb 2020 00:56:41 +0000 (16:56 -0800)]
s3: VFS: vfs_default. Protect vfs_fsync_done() from accessing a freed req pointer.
If the fsp is forced closed by a SHUTDOWN_CLOSE whilst the
request is in flight (share forced closed by smbcontrol),
then we set state->req = NULL in the state destructor.
The existing state destructor prevents the state memory
from being freed, so when the thread completes and calls
vfs_fsync_done(), just throw away the result if
state->req == NULL.
Jeremy Allison [Fri, 28 Feb 2020 00:54:47 +0000 (16:54 -0800)]
s3: VFS: vfs_default. Pass in struct vfswrap_fsync_state as the callback data to the subreq.
Find the req we're finishing off by looking inside vfswrap_fsync_state.
In a shutdown close the caller calls talloc_free(req), so we can't
access it directly as callback data.
The next commit will NULL out the vfswrap_fsync_state->req pointer
when a caller calls talloc_free(req), and the request is still in
flight.
Jeremy Allison [Fri, 28 Feb 2020 00:51:35 +0000 (16:51 -0800)]
s3: VFS: vfs_default. Protect vfs_pwrite_done() from accessing a freed req pointer.
If the fsp is forced closed by a SHUTDOWN_CLOSE whilst the
request is in flight (share forced closed by smbcontrol),
then we set state->req = NULL in the state destructor.
The existing state destructor prevents the state memory
from being freed, so when the thread completes and calls
vfs_pwrite_done(), just throw away the result if
state->req == NULL.
Jeremy Allison [Fri, 28 Feb 2020 00:49:38 +0000 (16:49 -0800)]
s3: VFS: vfs_default. Pass in struct vfswrap_pwrite_state as the callback data to the subreq.
Find the req we're finishing off by looking inside vfswrap_pwrite_state.
In a shutdown close the caller calls talloc_free(req), so we can't
access it directly as callback data.
The next commit will NULL out the vfswrap_pwrite_state->req pointer
when a caller calls talloc_free(req), and the request is still in
flight.
Jeremy Allison [Fri, 28 Feb 2020 00:40:46 +0000 (16:40 -0800)]
s3: VFS: vfs_default. Protect vfs_pread_done() from accessing a freed req pointer.
If the fsp is forced closed by a SHUTDOWN_CLOSE whilst the
request is in flight (share forced closed by smbcontrol),
then we set state->req = NULL in the state destructor.
The existing state destructor prevents the state memory
from being freed, so when the thread completes and calls
vfs_pread_done(), just throw away the result if
state->req == NULL.
Jeremy Allison [Fri, 28 Feb 2020 00:34:51 +0000 (16:34 -0800)]
s3: VFS: vfs_default. Pass in struct vfswrap_pread_state as the callback data to the subreq.
Find the req we're finishing off by looking inside vfswrap_pread_state.
In a shutdown close the caller calls talloc_free(req), so we can't
access it directly as callback data.
The next commit will NULL out the vfswrap_pread_state->req pointer
when a caller calls talloc_free(req), and the request is still in
flight.
Gary Lockyer [Mon, 2 Mar 2020 21:44:47 +0000 (10:44 +1300)]
audit_logging tests: Fix timezone validation
test_audit_get_timestamp used the "%Z" format specifier in strptime,
this is non-portable. Updated tests now explicitly set the time zone to
"UTC".
Signed-off-by: Gary Lockyer <gary@catalyst.net.nz> Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Autobuild-User(master): Andrew Bartlett <abartlet@samba.org>
Autobuild-Date(master): Sat Mar 7 06:37:09 UTC 2020 on sn-devel-184
vfs_fruit: tmsize prevent overflow Force the type during arithmetic in order to prevent overflow when summing the Time Machine folder size. Increase the precision to off_t (used for file sizes), leave the overflow error traps but with more precise wording.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=13622 Signed-off-by: Art M. Gallagher <smblock@artmg.org> Reviewed-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 Mar 7 01:37:31 UTC 2020 on sn-devel-184
The talloc_steal() in dsdb_enum_group_mem() is unnecessary, because
members was already allocated from the same mem_ctx.
The talloc_steal() in pdb_samba_dsdb_enum_aliasmem() is also unnecessary
for the same reason, but also incorrect, because it should be
dereferencing pmembers:
talloc_steal(mem_ctx, *pmembers);
Furthermore, we should only assign to *pnum_members on success; otherwise
num_members is used uninitialized.
Signed-off-by: Jonathon Reinhart <Jonathon.Reinhart@gmail.com> Reviewed-by: Andreas Schneider <asn@samba.org> Reviewed-by: Guenther Deschner <gd@samba.org>
Autobuild-User(master): Andreas Schneider <asn@cryptomilk.org>
Autobuild-Date(master): Thu Mar 5 18:40:16 UTC 2020 on sn-devel-184
python: Fix userPrincipalName in GetPasswordCommand.get_account_attributes()
Signed-off-by: Jonathon Reinhart <Jonathon.Reinhart@gmail.com> Reviewed-by: Andreas Schneider <asn@samba.org> Reviewed-by: Alexander Bokovoy <ab@samba.org>
Autobuild-User(master): Andreas Schneider <asn@cryptomilk.org>
Autobuild-Date(master): Thu Mar 5 16:27:50 UTC 2020 on sn-devel-184
Björn Baumbach [Mon, 10 Feb 2020 18:19:44 +0000 (19:19 +0100)]
s3-libads: use dns name to open a ldap session
Required for working certificate verification.
Bug: https://bugzilla.samba.org/show_bug.cgi?id=13124 Signed-off-by: Björn Baumbach <bb@sernet.de> Reviewed-by: Bjoern Jacke <bjacke@samba.org>
Autobuild-User(master): Björn Baumbach <bb@sernet.de>
Autobuild-Date(master): Thu Mar 5 12:29:26 UTC 2020 on sn-devel-184
Swen Schillig [Wed, 22 Jan 2020 10:55:46 +0000 (11:55 +0100)]
[s4] possible memleak in torture vfs-fruit
The allocated memory for "full_name" must be free'd
before returning to caller.
Signed-off-by: Swen Schillig <swen@linux.ibm.com> Reviewed-by: Ralph Boehme <slow@samba.org> Reviewed-by: David Disseldorp <ddiss@samba.org>
Autobuild-User(master): Ralph Böhme <slow@samba.org>
Autobuild-Date(master): Wed Mar 4 10:43:54 UTC 2020 on sn-devel-184
Ralph Boehme [Thu, 27 Feb 2020 16:01:10 +0000 (17:01 +0100)]
smbd: ignore set NTACL requests which contain S-1-5-88 NFS ACEs
We apply the same "ignore" logic already in the POSIX ACL code and in the
vfs_acl_xattr|tdb VFS modules to smb_set_nt_acl_nfs4() in the nfs4_acl helper
subsystem which is common to a bunch of VFS modules: GPFS, ZFS, NFS4_xattr and
aixacl2.
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): Tue Mar 3 19:15:10 UTC 2020 on sn-devel-184
Volker Lendecke [Sun, 1 Mar 2020 17:41:15 +0000 (18:41 +0100)]
selftest: Adapt libsmbclient.readdirplus2 to unix extensions
A few lines above the mode check we created a file with mode
0666. With unix exensions we expect this back 1:1, without them the
server changes them on the fly.
Signed-off-by: Volker Lendecke <vl@samba.org> Reviewed-by: Jeremy Allison <jra@samba.org>
Volker Lendecke [Fri, 28 Feb 2020 10:10:47 +0000 (11:10 +0100)]
libsmb: Implement SMB_FIND_FILE_UNIX_INFO2 dir listing
This keeps the original SMB_STRUCT_STAT coming from posix as part of
struct file_info. It is a slight waste of space, as the timestamps are
kept twice, but having a full SMB_STRUCT_STAT with the nlink!=0
validity check makes thinking about which mode/size/etc is the correct
one a no-brainer. We can save space later by referencing only one set
of time stamps for example.
This for the time fixes readdirplus2, but for the wrong reason: We don't yet
create files the "proper" way using posix create.
Signed-off-by: Volker Lendecke <vl@samba.org> Reviewed-by: Jeremy Allison <jra@samba.org>
Volker Lendecke [Mon, 2 Mar 2020 20:34:37 +0000 (21:34 +0100)]
selftest: Run smbtorture3 SHORTNAME-TEST without unix extensions
Soon cli_list() will change to SMB_FIND_FILE_UNIX_INFO2 which does not
provide a shortname. For now we lose that as an encrypted test, as for
now it's a SMB1 test which requires unix extensions for
encryption. Hopefully we don't forget to reenable this once the test
is converted to SMB2
Signed-off-by: Volker Lendecke <vl@samba.org> Reviewed-by: Jeremy Allison <jra@samba.org>
nonunix
Martin Schwenke [Fri, 21 Feb 2020 09:35:24 +0000 (20:35 +1100)]
ctdb-tests: Use a local "ctdb shutdown" command to avoid a race
When "ctdb shutdown" is run with -n <N> it does not wait for the node
<N>'s ctdbd to go down but exits immediately. This means that the
local_daemons.sh shutdown command can find the PID file still present
and then attempt the shutdown, but the daemon can have exited between
the check and the shutdown. Although the test waits until the node is
disconnected, the transport is taken down just before the exit, so
this does not guarantee the daemon has exited.
A local shutdown command (no -n <N>) waits until the socket
disconnects and this happens *after* the PID file is gone, so this is
safe to use with the local_daemons.sh shutdown command.
Signed-off-by: Martin Schwenke <martin@meltin.net> Reviewed-by: Amitay Isaacs <amitay@gmail.com>
Autobuild-User(master): Amitay Isaacs <amitay@samba.org>
Autobuild-Date(master): Mon Mar 2 10:39:28 UTC 2020 on sn-devel-184
Andrew Bartlett [Fri, 28 Feb 2020 20:30:15 +0000 (09:30 +1300)]
dsdb: Add debugging for a contrived situation where a non-schema attribute is on the record
I had to modify the backend DB to produce this error, but
I would like a clear error anyway.
Signed-off-by: Andrew Bartlett <abartlet@samba.org> Reviewed-by: Gary Lockyer <gary@catalyst.net.nz>
Autobuild-User(master): Andrew Bartlett <abartlet@samba.org>
Autobuild-Date(master): Mon Mar 2 04:14:22 UTC 2020 on sn-devel-184
Andrew Bartlett [Sat, 29 Feb 2020 03:03:00 +0000 (16:03 +1300)]
ldb: ensure that ldbedit operates like ldbmodify: set DONT_CREATE_DB
ldb* tools, when passed a raw filename assume tdb://
By default, ldb_tdb will call tdb with O_CREAT.
TDB, when passed O_CREAT and a not-tdb file, will wipe the file.
This means that if you run ldbedit <path to mdb-format-ldb file> the file
will be wiped, which is unexpected. I noticed this while trying to
corrupt a sam.ldb backend file (for testing), but instead I wiped it!
Ideally tdb would not do that, but the behaviour has been this way for
decades. Ideally ldb would have had a "create db" command, but this
has been the job of ldbadd for over a decade.
Martin Schwenke [Sat, 29 Feb 2020 09:53:28 +0000 (20:53 +1100)]
ctdb-tests: Silence a ShellCheck warning
Signed-off-by: Martin Schwenke <martin@meltin.net> Reviewed-by: Ralph Boehme <slow@samba.org>
Autobuild-User(master): Ralph Böhme <slow@samba.org>
Autobuild-Date(master): Sat Feb 29 11:53:42 UTC 2020 on sn-devel-184
Ralph Boehme [Fri, 28 Feb 2020 12:02:43 +0000 (13:02 +0100)]
ctdb: add tail logs option to local_daemons.sh
Signed-off-by: Ralph Boehme <slow@samba.org> Reviewed-by: Martin Schwenke <martin@meltin.net>
Autobuild-User(master): Martin Schwenke <martins@samba.org>
Autobuild-Date(master): Sat Feb 29 08:02:50 UTC 2020 on sn-devel-184
Andrew Bartlett [Mon, 24 Feb 2020 03:57:35 +0000 (16:57 +1300)]
dsdb: Rewrite comment to remove refernece to LDAP backends
This is required despire the demise of the LDAP backend.
Signed-off-by: Andrew Bartlett <abartlet@samba.org> Reviewed-by: Gary Lockyer <gary@catalyst.net.nz>
Autobuild-User(master): Andrew Bartlett <abartlet@samba.org>
Autobuild-Date(master): Fri Feb 28 04:42:23 UTC 2020 on sn-devel-184
Andrew Bartlett [Mon, 24 Feb 2020 03:22:24 +0000 (16:22 +1300)]
dsdb: Do not use ldb_save_controls() in partitions module for domain_scope
The LDAP backend is long-removed so we do not need this workaround
for a confused server any longer.
This avoids references to old (but valid) memory after a new ldb_control array is
allocated in ldb_save_controls() and keeps the controls pointer as
constant as possible given the multiple ldb_request structures it
will appear in.
Signed-off-by: Andrew Bartlett <abartlet@samba.org> Reviewed-by: Gary Lockyer <gary@catalyst.net.nz>
Swen Schillig [Wed, 22 Jan 2020 10:28:28 +0000 (11:28 +0100)]
prevent NULL reference from being used as '%s' argument.
The two string arguments to torture_comment() can be NULL
as the succeeding checks suggest. This is not wanted because a compile
with --enable-developer throws an error of [-Werror=format-overflow=]
in those situations.
Signed-off-by: Swen Schillig <swen@linux.ibm.com> Reviewed-by: Gary Lockyer <gary@catalyst.net.nz> Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Peter Eriksson [Thu, 27 Feb 2020 19:54:52 +0000 (20:54 +0100)]
Undef ARRAY_SIZE, if defined, before define to avoid compilation warnings
Signed-off-by: Peter Eriksson <pen@lysator.liu.se> Reviewed-by: Jeremy Allison <jra@samba.org> Reviewed-by: Gary Lockyer <<gary@catalyst.net.nz>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Thu Feb 27 21:43:29 UTC 2020 on sn-devel-184
Volker Lendecke [Thu, 27 Feb 2020 08:52:30 +0000 (00:52 -0800)]
selftest: Run libsmbclient with and without unix extensions
The libsmbclient readdir tests are broken just for the unix extension
case. For example they assume our "map archive" behaviour. This will
have to be parameterized once unix extensions become better
implemented in libsmbclient
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 Feb 27 19:34:36 UTC 2020 on sn-devel-184
Volker Lendecke [Wed, 26 Feb 2020 22:07:54 +0000 (14:07 -0800)]
selftest: Assemble smbtorture arguments as a list
I'm planning to mess with libsmbclient tests calling into
smbtorture4. For this it will be much more convenient to have the
arguments available as a higher-level data structure than just a
string.
Checked by "diff" on the testlist before and after -- no change.
Signed-off-by: Volker Lendecke <vl@samba.org> Reviewed-by: Jeremy Allison <jra@samba.org>
Anoop C S [Tue, 25 Feb 2020 12:38:35 +0000 (18:08 +0530)]
ctdb-docs: Move CTDB_SERVICE_NMB to new 48.netbios section
Signed-off-by: Anoop C S <anoopcs@redhat.com> Reviewed-by: Guenther Deschner <gd@samba.org> Reviewed-by: Martin Schwenke <martin@meltin.net>
Autobuild-User(master): Martin Schwenke <martins@samba.org>
Autobuild-Date(master): Thu Feb 27 07:34:53 UTC 2020 on sn-devel-184
Anoop C S [Wed, 26 Feb 2020 05:40:46 +0000 (11:10 +0530)]
ctdb-scripts: Change CTDB_SERVICE_NMB default value to 'nmb'
Till now 50.samba script was based on RHEL versions <=6 where we didn't
have separate start up script for nmb and smbd used to start nmbd when
required. Now that nmbd has its own start up script named "nmb" it is
reasonable to have "nmb" as default value for CTDB_SERVICE_NMB inside
new 48.netbios ctdb script.
Signed-off-by: Anoop C S <anoopcs@redhat.com> Reviewed-by: Guenther Deschner <gd@samba.org> Reviewed-by: Martin Schwenke <martin@meltin.net>
Günther Deschner [Tue, 25 Feb 2020 14:04:41 +0000 (15:04 +0100)]
ctdb-scripts: add new 48.netbios script for starting nmbd
This change basically moves out nmbd references from 50.samba script to
a new 48.netbios script. Accordingly ctdb test scripts are tweaked to
cope with newly added script.
Signed-off-by: Guenther Deschner <gd@samba.org> Reviewed-by: Martin Schwenke <martin@meltin.net>
Andrew Bartlett [Tue, 25 Feb 2020 23:02:36 +0000 (12:02 +1300)]
lib/util/fault.c: Unify printing of the stack trace with the INTERNAL ERROR string
We can diverged into the s3 and lib/fault panic action stuff once
we have printed the backtrace.
Our tests require we use the word PANIC, and some administrative scripts
might look for similar things, so keep those words.
The use of DEBUG(0, ... is deliberate to keep the output
of the PANIC line as consistent as possible with the historical smbd output.
Signed-off-by: Andrew Bartlett <abartlet@samba.org> Reviewed-by: Gary Lockyer <gary@catalyst.net.nz>
Autobuild-User(master): Andrew Bartlett <abartlet@samba.org>
Autobuild-Date(master): Thu Feb 27 05:09:44 UTC 2020 on sn-devel-184
Andrew Bartlett [Tue, 25 Feb 2020 22:58:48 +0000 (11:58 +1300)]
Move INTERNAL ERROR... printout into smb_panic() and improve "why" string
The "why" string is now eg "Signal 11: Segmentation fault" and
so more descriptive, and hopefully this will encourage the
Samba version to be included in more error reports.
Signed-off-by: Andrew Bartlett <abartlet@samba.org> Reviewed-by: Gary Lockyer <gary@catalyst.net.nz>
Andrew Bartlett [Fri, 21 Feb 2020 20:45:39 +0000 (09:45 +1300)]
lib/tfork: Improve process titles for the tfork waiter process
This makes it easier to understand the process graph as there is
not a duplicate (eg) ldap[master] process for each tfork waiter.
when useing "ps -ef -o pid,comm"
Gary Lockyer [Fri, 7 Feb 2020 02:18:45 +0000 (15:18 +1300)]
idl: limit recurion on recursive elements
Limit the max_recursion on self recursive definitions in the idl to
20,000. This value is hopefully large eneough to not impact normal
operation, but small eneough to prevent stack over flow issues.
Signed-off-by: Gary Lockyer <gary@catalyst.net.nz> Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Autobuild-User(master): Andrew Bartlett <abartlet@samba.org>
Autobuild-Date(master): Thu Feb 27 02:29:21 UTC 2020 on sn-devel-184
Gary Lockyer [Wed, 29 Jan 2020 19:51:47 +0000 (08:51 +1300)]
pidl: Add recursive depth checks.
Add new parameter to elements "max_recursion" and modify pidl to call
NDR_RECURSION_CHECK and NDR_RECURSION_UNWIND for element tagged with
that attribute.
Flex adds "#line nnn <stdout>" to the generated source, which then causes
issues with lcov. This commit adds a SAMBA_GENERATOR for es_lexer.l and
sparql_lexer.l that strips out the offending lines.
Signed-off-by: Gary Lockyer <gary@catalyst.net.nz> Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Autobuild-User(master): Stefan Metzmacher <metze@samba.org>
Autobuild-Date(master): Wed Feb 26 23:11:39 UTC 2020 on sn-devel-184
s3:smb2_tcon: pass down in_flags to smbd_smb2_tree_connect_send()
SMB 3.1.1 converted the reserved field to a flags field.
Signed-off-by: Stefan Metzmacher <metze@samba.org> Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Wed Feb 26 21:11:59 UTC 2020 on sn-devel-184
lib/async_req: remove useless "reentrant" support from async_connect_send()
Now that open_socket_out*() doesn't do the strange reentrant calls,
we can remove support for this in async_connect_send(), so that we'll
never get any new users of this.
Signed-off-by: Stefan Metzmacher <metze@samba.org> Reviewed-by: Jeremy Allison <jra@samba.org>
s3:lib: avoid calling async_connect_send() again and again in open_socket_out*()
There's really no need to do that!
Once connect() is called and returned EINPROGRESS, the kernel
knowns what to do and reports any state change via
TEVENT_FD_READ or TEVENT_FD_WRITE.
The actual success or failure is available via
getsockopt(.., SOL_SOCKET, SO_ERROR, ...).
Before this commit we called connect() (via async_connect_send()) again
and again until we reached the final caller provided timeout,
even if the kernel already found out that the destination is
unreachable.
Signed-off-by: Stefan Metzmacher <metze@samba.org> Reviewed-by: Jeremy Allison <jra@samba.org>
lib/async_req: make sure we return errors early from async_connect_send/recv
While it is true that [e]poll() only needs POLLOUT
and POLLERR/POLLHUP are added implicitly.
For tevent we need TEVENT_FD_READ in order to see POLLERR/POLLHUP.
The socket becomes only readable when we hit an error.
Waiting for TEVENT_FD_WRITE is needed for getting success,
while TEVENT_FD_READ is required to get failures.
This matches what we have in tstream_bsd_connect_send().
Signed-off-by: Stefan Metzmacher <metze@samba.org> Reviewed-by: Jeremy Allison <jra@samba.org>
projects / thirdparty / samba.git / log
