Signed-off-by: Volker Lendecke <vl@samba.org> Reviewed-by: Stefan Metzmacher <metze@samba.org> Reviewed-by: Ralph Boehme <slow@samba.org>
Autobuild-User(master): Ralph Böhme <slow@samba.org>
Autobuild-Date(master): Mon Mar 6 19:18:31 CET 2017 on sn-devel-144
s3:libads: remove unused fallback to gss_acquire_cred()
Heimdal and all supported versions of MIT krb5 prove gss_krb5_import_cred(),
so we don't need an #ifdef here.
Signed-off-by: Stefan Metzmacher <metze@samba.org> Reviewed-by: Alexander Bokovoy <ab@samba.org>
Autobuild-User(master): Stefan Metzmacher <metze@samba.org>
Autobuild-Date(master): Mon Mar 6 11:44:54 CET 2017 on sn-devel-144
Ralph Boehme [Mon, 27 Feb 2017 11:55:04 +0000 (12:55 +0100)]
s4/torture: add a creditting test skipping a SMB2 MID
This tests that skipping a SMB2 MID the client's usable MID window is
[unused mid, unused mid + 8192]
The test currently fails against Samba as we only grant up to 512
credits. It passes against Windows 2016 as that grants up to 8192
credits by default.
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 Mar 4 01:54:07 CET 2017 on sn-devel-144
Ralph Boehme [Mon, 27 Feb 2017 06:12:09 +0000 (07:12 +0100)]
s4/torture: add some SMB2 crediting tests
These tests verify that a server grants at least 8192 credits in a
successfull session setup and in a single SMB2 request. Both tests pass
against Windows 2016 Server but currently fail against Samba.
Signed-off-by: Ralph Boehme <slow@samba.org> Reviewed-by: Jeremy Allison <jra@samba.org>
Ralph Boehme [Mon, 27 Feb 2017 15:14:39 +0000 (16:14 +0100)]
libcli/smb: add max_credits arg to smbXcli_negprot_send()
This allows source4/torture code to set the option for tests by
preparing a struct smbcli_options with max_credits set to some value and
pass that to a torture_smb2_connection_ext().
This will be used in subsequent smbtorture test for SMB2 creditting.
Behaviour of existing upper layers is unchanged, they simply pass the
wanted max credits value to smbXcli_negprot_send() instead of
retrofitting it with a call to smb2cli_conn_set_max_credits().
Signed-off-by: Ralph Boehme <slow@samba.org> Reviewed-by: Jeremy Allison <jra@samba.org>
Signed-off-by: Andreas Schneider <asn@samba.org> Reviewed-by: Michael Adam <obnox@samba.org>
Autobuild-User(master): Andreas Schneider <asn@cryptomilk.org>
Autobuild-Date(master): Fri Mar 3 16:54:34 CET 2017 on sn-devel-144
Signed-off-by: Stefan Metzmacher <metze@samba.org> Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Autobuild-User(master): Andrew Bartlett <abartlet@samba.org>
Autobuild-Date(master): Fri Mar 3 12:57:06 CET 2017 on sn-devel-144
Signed-off-by: Volker Lendecke <vl@samba.org> Reviewed-by: Ralph Böhme <slow@samba.org> Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Fri Mar 3 06:20:50 CET 2017 on sn-devel-144
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): Fri Mar 3 02:23:20 CET 2017 on sn-devel-144
Uri Simchoni [Thu, 2 Mar 2017 06:39:56 +0000 (08:39 +0200)]
smbd: remove coupling between get_ea_names_from_file() and "ea support"
The "ea support" configuration variable determines whether smbd
should attempt to manipulate extended attributes via SMB protocol.
It does not pertain to the underlying storage and its support for
extended attributes.
get_ea_names_from_file() is being used also by vfs_streams_xattr -
a module which has nothing to do with client-visible extended
attributes. As such, vfs_streams_xattr should be able to operate
irrespective of the value of "ea support".
This patch moves the check for ea support to the callers.
Signed-off-by: Uri Simchoni <uri@samba.org> Reviewed-by: Ralph Boehme <slow@samba.org>
Uri Simchoni [Thu, 2 Mar 2017 06:46:44 +0000 (08:46 +0200)]
smbd: refuse_symlink() - do not fail if the file does not exist
If the file does not exist, it is not a symlink. Current callers
use this function to see if extended attributes can be set / fetched.
Allow them to try and leave the error code at the discretion of the
VFS.
Signed-off-by: Uri Simchoni <uri@samba.org> Reviewed-by: Ralph Boehme <slow@samba.org>
The fixes for https://bugzilla.samba.org/show_bug.cgi?id=11691
changed the way we fill domain->domain_flags for domains
in other forests.
Commit fffefe72fcc62d9688b45f53a5327667dc0b2fe6 readded the
ability to enumerate trusts of other forests again, in order to
fix https://bugzilla.samba.org/show_bug.cgi?id=11830
Now we have the problem that multiple domains
(even outside of our forest) are considert to be
our forest root, as they have the following flags:
NETR_TRUST_FLAG_TREEROOT and NETR_TRUST_FLAG_IN_FOREST.
Signed-off-by: Stefan Metzmacher <metze@samba.org> Reviewed-by: Ralph Boehme <slow@samba.org>
Autobuild-User(master): Ralph Böhme <slow@samba.org>
Autobuild-Date(master): Thu Mar 2 17:53:14 CET 2017 on sn-devel-144
Pair-Programmed-With: Stefan Metzmacher <metze@samba.org> Signed-off-by: Andreas Schneider <asn@samba.org> Signed-off-by: Stefan Metzmacher <metze@samba.org>
Autobuild-User(master): Andreas Schneider <asn@cryptomilk.org>
Autobuild-Date(master): Thu Mar 2 12:41:40 CET 2017 on sn-devel-144
Douglas Bagnall [Thu, 23 Feb 2017 22:58:33 +0000 (11:58 +1300)]
ndr: Use resizing array instead of linked lists (breaking ABI)
The ndr token code keeps a temporary store of tokens which are
referred to a small number of times (often once) before being
discarded. The access patterns are somewhat stack-like, with recently
placed tokens being accessed most often.
The old code kept these tokens in a linked list, which we replace with
a self-resizing array.
This keeps everything roughly the same in big-O terms, but makes it
all faster in practice by vastly reducing the amount of tallocing and
pointer-chasing.
The peak memory use is strictly reduced. On a 64 bit machine each core
token struct fits in 16 bytes (after padding) while the two pointers
used by the DLIST add another 16 bytes, so the overall list allocation
is the same as the peak 2n array allocation -- except in the list case
it is dwarfed by the talloc and malloc metadata overhead.
Before settling on the resized arrays, we tried red-black trees, which
are bound to be better for large ndr structures. As it happens, we
don't deal with large structures (the size of replication clumps is
limited to 400 objects) and the asymptotic benefits of the trees are
not realised in practice.
With luck you should find graphs comparing the performance of these
various techniques at:
This necessarily breaks the ABI because the linked list implementation
was publicly exposed.
Signed-off-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 Mar 2 08:38:22 CET 2017 on sn-devel-144
Ralph Boehme [Tue, 28 Feb 2017 08:39:37 +0000 (09:39 +0100)]
vfs_fruit: enabling AAPL extensions must be a global switch
Apple's SMB2 AAPL extension is enabled once per SMB2
connection. Unfortunately the (per se correct) fix for bug #12541
results in vfs_fruit checking a per tcon config state variable to
determine whether AAPL has been negotiated. This variable will be false
for all but the first tcon. We must make it a global variable.
Signed-off-by: Ralph Boehme <slow@samba.org> Reviewed-by: Uri Simchoni <uri@samba.org>
Autobuild-User(master): Uri Simchoni <uri@samba.org>
Autobuild-Date(master): Thu Mar 2 04:34:10 CET 2017 on sn-devel-144
Ralph Boehme [Thu, 19 Jan 2017 08:30:45 +0000 (09:30 +0100)]
vfs_fruit: only veto AppleDouble files with fruit:resource=file
vfs_fruit only creates AppleDouble files itself when "fruit:resource" is
set to "file" (the default). It is only then the these AppleDouble files
should be treated as an internal representation and should be
inaccessible from clients.
Ralph Boehme [Sun, 11 Dec 2016 18:02:37 +0000 (19:02 +0100)]
s4/torture: change shares in used torture_suite_add_2ns_smb2_test()
torture_suite_add_2ns_smb2_test wan't used, change it to use the default
share as share 1 and a second share taken from torture option
"torture:share2".
Ralph Boehme [Thu, 8 Dec 2016 18:12:32 +0000 (19:12 +0100)]
vfs_fruit: rework struct adouble API
o factor out ad_open(), opens an fd to be used with IO on the metadata
streams
o rename ad_write() to ad_set() as this aligns nicely with the existing
ad_get. This is the pathname based version used to solely set metadata
on a file, there's also a handle based version:
o add ad_fset(), a handle based version that can be used to set metadata
and to update the AppleDouble header in a ._ AppleDouble file
o remove fruit_fsp_recheck(), looking at it more closely, it is not
needed
This commit *compiles*, but all callers in the VFS ops must now be
updated to use the new semantics, this comes next...
Ralph Boehme [Thu, 8 Dec 2016 10:08:53 +0000 (11:08 +0100)]
vfs_fruit: fix fruit_check_access()
Applying fcntl read locks requires an fd opened for reading. This means
we have to check the open flags of an fd and if the fd is not opened for
reading, we can't use it to set shared fctnl locks.
Iow we won't be applying interop locks with Netatalk for files openend
write-only.
Ralph Boehme [Wed, 8 Feb 2017 18:15:38 +0000 (19:15 +0100)]
vfs_fruit: use SMB_VFS_NEXT_OPEN in two places
Using the SMB_VFS_OPEN leads to a logic recursion that is hard to follow
and debug. It may be called twice for the same fsp with two different
smb_fname's.
Changing this here to use the NEXT function instead should have no
change in behaviour otherwise, but it allows adding sane caching to
vfs_catia to handle based VFS ops.
Ralph Boehme [Fri, 2 Dec 2016 15:20:46 +0000 (16:20 +0100)]
vfs_fruit: refactor fruit_ftruncate() and fix stream case
The fruit:metadata=stream case wasn't handled previosly, this is now
done in fruit_ftruncate_rsrc_stream(). The rest is just refactoring and
no change in behaviour.
projects / thirdparty / samba.git / log
