IF YOU WOULD LIKE TO GET AN ACCOUNT, please write an
email to Administrator. User accounts are meant only to access repo
and report issues and/or generate pull requests.
This is a purpose-specific Git hosting for
BaseALT
projects. Thank you for your understanding!
Только зарегистрированные пользователи имеют доступ к сервису!
Для получения аккаунта, обратитесь к администратору.
Handle all settings of fruit:metadata and fruit:resource in helper
functions.
Resource fork streams of 0 bytes must be filtered out, this wasn't done
previously for the fruit:resource=stream and xattr case.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12427
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
This fixes fruit_stat_rsrc for the fruit:resource=xattr case.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12427
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
This fixes fruit_stat_rsrc for the fruit:resource=stream case.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12427
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
Use helper functions for the fruit:resource cases. No change in
behaveour.
The next patch will add the proper helper functions for
fruit:resource=xattr and fruit:resource=stream.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12427
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
This just splits up fruit_open_rsrc() to use helper functions for each
config->rsrc setting. No change in behaviour.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12427
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
This ensures we only delete valid AppleDouble files whose names begin
with "._", not just *any* file that matches "^._*".
Also use proper VFS functions instead of direclty calling the C library
functions.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12427
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
The following code must only be executed for the fruit:resource=file
case. Otherwise no change in behaviour.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12427
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
The following code must only be executed for the fruit:resource=file
case.
While at it, remove an unnecessary lstat, use the stat info from
smb_fname.
Otherwise no change in behaviour for the fruit:resource=file case (the
default).
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12427
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
The following code must only be executed for the fruit:resource=file
case.
While at it, remove an unnecessary lstat, use the stat info from
smb_fname.
Otherwise no change in behaviour for the fruit:resource=file case (the
default).
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12427
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
Refactor fruit_unlink() addin helper functions for all fruit:metadata
and fruit:resource settings.
No change in behaviour for fruit:metadata=netatalk and
fruit:resource=file (both the default), but fixes behaviour for the
other cases.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12427
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
o fix the config check, we must only run following code for the
fruit:resource=file
o properly call SMB_VFS_NEXT_RENAME() instead diretly calling rename()
o bail out if we get an invalid stat
Otherwise, no change in behaviour.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12427
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
This gets correct behaviour in readdir_attr_meta_finderi for the
metadata=stream case.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12427
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
Move the FinderInfo stuff to helper functions, no change in behaviour so
far.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12427
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
Handle config->meta in helper functions. No change in behaviour. The next step
will add the correct implementation of fruit_stat_meta_stream().
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12427
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
This needs to create and write a metadata blob when the stream is
created.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12427
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
Just split out the fruit:metadata=stream case into a helper function, no
change in behaviour.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12427
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
This makes the test "delete AFP_AfpInfo by writing all 0" pass with
fruit:metadata=stream.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12427
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
No change in behaviour, but ad_entry(ad, ADEID_FINDERI) can in theory
return NULL.
The next commit will add the same function for a AfpInfo type, so rename
this function that works on struct adouble to ad_empty_finderinfo().
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12427
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
With this the test "setinfo eof AFP_AfpInfo" passed with
fruit:metadata=stream.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12427
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
This make the test "fix offset and len handling for AFP_AfpInfo stream"
pass with fruit:metadata=stream.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12427
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
Also run translation of fsp->fsp_name and optionally
fsp->base_fsp->fsp_name if set for all handle based VFS functions.
This is necessary because some modules might use the path for something
in the handle based VFS ops.
An existing example is vfs_fruit. A later commit will remove the
translation that was done there because catia didn't before this commit.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12427
Pair-Programmed-With: Stefan Metzmacher <metze@samba.org>
Signed-off-by: Ralph Boehme <slow@samba.org>
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
Using the SMB_VFS_OPEN leads to a recursion in the VFS that is hard to
follow and debug. It's called twice for the same fsp with two different
smb_fname's which makes it at least hard to debug, it may even be an
invalid usage.
Changing this here to use the NEXT function instead should have no
change in behaviour otherwise.
Bug: https://bugzilla.samba.org/show_bug.cgi?id=12565
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
In case an SMB_VFS_OPEN() on a stream basename fails with EISDIR, we
retry the open as O_RDONLY. The retry should be done with the
smb_fname_base as well.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12427
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
Signed-off-by: Chris Lamb <chris@chris-lamb.co.uk>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
The base_fsp's fd is always -1 as it's closed after being openend in
create_file_unixpath().
Additionally in streams_xattr_open force using of SMB_VFS_FSETXATTR() by
sticking the just created fd into the fsp (and removing it afterwards).
Bug: https://bugzilla.samba.org/show_bug.cgi?id=12591
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): Wed Feb 22 08:25:46 CET 2017 on sn-devel-144
The "ceph: user_id" parameter can be specified in smb.conf to explicitly
set the Ceph client ID used when creating the mount handle.
Signed-off-by: David Disseldorp <ddiss@samba.org>
Reviewed-by: Ralph Böhme <slow@samba.org>
ceph_shutdown() is the equivalent to ceph_unmount() + ceph_release()
without error handling.
Signed-off-by: David Disseldorp <ddiss@samba.org>
Reviewed-by: Ralph Böhme <slow@samba.org>
Fix resource fork xattr name broken in
e4d1f8354f.
Bug: https://bugzilla.samba.org/show_bug.cgi?id=12490
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Autobuild-User(master): Ralph Böhme <slow@samba.org>
Autobuild-Date(master): Tue Feb 14 21:26:01 CET 2017 on sn-devel-144
Just some cleanup, no change in behaviour. This also removes the hokey
tag. :)
Bug: https://bugzilla.samba.org/show_bug.cgi?id=12490
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
GCC 7 warns about snprintf truncating a dirent d_name (potentially 255 bytes) to 25 bytes,
even though we have checked that it is 25 long in shadow_copy_match_name().
Using strlcpy instead of snprintf lets us check it again, JUST TO BE SURE.
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
On a slow filesystem or network filesystem this can make a huge
difference.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12571
Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
If the open is changing directories, fsp->fsp_name->base_name
will be the full path from the share root, whilst
smb_fname will be relative to the $cwd.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12546
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Ralph Böhme <slow@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Thu Feb 2 01:55:42 CET 2017 on sn-devel-144
Snapshot paths are a read-only filesystem.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12531
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Mon Jan 30 22:26:29 CET 2017 on sn-devel-144
Rationale:
VFS calls must act like their POSIX equivalents, and the POSIX versions
*only* set errno on a failure. There is actually code in the upper smbd
layers that depends on errno being correct on a fail return from a VFS call.
For a compound VFS module like this, a common pattern is :
SMB_VFS_CALL_X()
{
int ret;
syscall1();
ret = syscall2();
syscall3();
return ret;
}
Where if *any* of the contained syscallX()'s fail, they'll set errno.
However, the actual errno we should return is *only* the one returned
if syscall2() fails (the others are lstat's checking for existence etc.).
So what we should do to correctly return only the errno from syscall2() is:
SMB_VFS_CALL_X()
{
int ret;
int saved_errno = 0;
syscall1()
ret = syscall2();
if (ret == -1) {
saved_errno = errno;
}
syscall3()
if (saved_errno != 0) {
errno = saved_errno;
}
return ret;
}
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12531
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
Completely cleans up the horrible shadow_copy2_strip_snapshot()
and adds an explaination of what it's actually trying to do.
* This function does two things.
*
* 1). Checks if an incoming filename is already a
* snapshot converted pathname.
* If so, it returns the pathname truncated
* at the snapshot point which will be used
* as the connectpath, and then does an early return.
*
* 2). Checks if an incoming filename contains an
* SMB-layer @GMT- style timestamp.
* If so, it strips the timestamp, and returns
* both the timestamp and the stripped path
* (making it cwd-relative).
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12531
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12531
This is not yet used, the users of this will be added later.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12531
Allows an extra (currently unused) parameter to be added.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
Allow the called functions to be fixed to not touch them on error.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12531
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
readdirattr should only be enabled if the client enables it via AAPL
negotitiation, not for all clients when vfs_fruit is loaded.
Unfortunately the check in fruit_readdir_attr() is
if (!config->use_aapl) {
return SMB_VFS_NEXT_READDIR_ATTR(handle, fname, mem_ctx, pattr_data);
}
This uses the wrong config state "use_aapl" which is always true by
default (config option "fruit:aapl").
We must use "nego_aapl" instead which is only true if the client
really negotiated this feature.
Bug: https://bugzilla.samba.org/show_bug.cgi?id=12541
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 Jan 28 01:49:11 CET 2017 on sn-devel-144
Signed-off-by: Bjoern Jacke <bj@sernet.de>
Reviewed-by: David Disseldorp <ddiss@samba.org>
Autobuild-User(master): Björn Jacke <bj@sernet.de>
Autobuild-Date(master): Sat Jan 21 17:00:54 CET 2017 on sn-devel-144
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Christof Schmitt <cs@samba.org>
Autobuild-User(master): Ralph Böhme <slow@samba.org>
Autobuild-Date(master): Sat Dec 17 12:58:07 CET 2016 on sn-devel-144
This is now handled by the vfs_gpfs_(f)get_dos_attributes. Getting rid
of this in the stat VFS functions is a huge performance saver. perf
report found that in a kernel copy workload smbd was spending
considerable CPU time in vfs_gpfs_(f|l)stat -> gpfs_get_winattrs.
Most of the time the VFS stat caller is not interested in the btime. The
SMB frontend processing around btime is designed to fetch btime together
with DOS attributes via dos_mode() in all places that need these
attributes. That's the way it is implemented in the default VFS module
and that's what vfs_gpfs now does as well for performance reasons.
This makes vfs_gpfs_fstat a null op and I'm therefor removing it.
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Christof Schmitt <cs@samba.org>
This paves the way for removing btime updates from the stat VFS
functions.
This way we behave like the default VFS module where DOS attributes and
btime are fetched from the same backing store and the frontend is
designed around using dos_mode() -> SMB_VFS_GET_ATTRIBUTES to update
both attributes as necessary in the SMB processing.
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Christof Schmitt <cs@samba.org>
Fix all occurences of bad spelling of "resource" as "res*s*ource" (two
s).
One of the places where this was wrong was when parsing parametric
options in the VFS connect() function in the module. As a result any
setting of
fruit:resource=something
in smb.conf was silently ignored and the default ("file") was active.
In Samba 4.6 we accept both the wrong and the correct spelling, in Samba
4.7 the bad spelling will be removed.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12412
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
...per Jeremy's recommendation.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Mon Nov 21 03:34:26 CET 2016 on sn-devel-144
Add a configure test for the ceph_statx function, and use that to
determine whether to compile in new functions that use it and its
variants, or whether to use a the older code that fetches birthtimes
from an xattr.
For cephwrap_lstat, we can use ceph_statx with the AT_SYMLINK_NOFOLLOW
flag to get the right lookup semantics.
For setting the times via cephwrap_ntimes, We can just use ceph_setattrx
and pass them all in at the same time.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Reviewed-by: Jeremy Allison <jra@samba.org>
This makes it more obvious where this legacy code is used
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): Sun Nov 20 06:23:19 CET 2016 on sn-devel-144
conn->cwd can change over the life of the connection,
conn->connectpath remains static.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12387
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
Autobuild-User(master): Uri Simchoni <uri@samba.org>
Autobuild-Date(master): Mon Oct 24 23:52:48 CEST 2016 on sn-devel-144
This makes us independent of the allocation
method used inside glfs_realpath.
Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Ira Cooper <ira@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Sat Oct 22 00:28:41 CEST 2016 on sn-devel-144
By the time we get to SMB_VFS_UNLINK/SMB_VFS_RMDIR the ACL
checks have already been done.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12384
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
The patchset `git log -2 3031815f982e365be50148564d47d7d5afab46e0`
missed a change to vfs_gpfs_is_offline() which is now merely a helper
function that returns true or false and mustn't call into the VFS.
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: David Disseldorp <ddiss@samba.org>
Autobuild-User(master): David Disseldorp <ddiss@samba.org>
Autobuild-Date(master): Mon Oct 17 21:28:12 CEST 2016 on sn-devel-144
glusterfs:volfile_server option can be used in smb.conf to define where
to fetch the volfile from. Currently it supports only a single IP or a
hostname. The default is 'localhost'.
glfs_set_volfile_server() has been enhanced in gfapi to support
multiple invocations. A list is maintained in libgfapi which gets
appended on every invocation. When glfs_init is performed, libgfapi
would first try to fetch the volfile from glusterd on that node.
However, on failure to fetch the volfile, it would proceed to contact
glusterd on every node in the list until it gets the volfile or
exhausts the list. This enhacement was done in Gluster commit [2].
This commit is available in 3.6, 3.7, 3.8 versions of Gluster.
As we cannot have multiple lines having the same key of
glusterfs:volfile_server in a share definition in smb.conf, we propose
a scheme like this:
where value of glusterfs:volfile_server could be list of white space seperated
elements where each element could be unix+/path/to/socket/file or
[tcp+]IP|hostname|\[IPv6\][:port].
Note the restriction on naming a IPv6 host, it follows the same
restriction that is based on IPv6 naming in URL as per RFC 2732[1].
[1] http://www.ietf.org/rfc/rfc2732.txt
[2] 0c1d78f5c52c69268ec3a1d8d5fcb1a1bf15f243
Signed-off-by: Raghavendra Talur <rtalur@redhat.com>
Reviewed-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Michael Adam <obnox@samba.org>
Reviewed-by: Guenther Deschner <gd@samba.org>
Autobuild-User(master): Günther Deschner <gd@samba.org>
Autobuild-Date(master): Fri Oct 14 17:09:24 CEST 2016 on sn-devel-144
Signed-off-by: Trever L. Adams <trever.adams@gmail.com>
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): Thu Oct 13 04:26:26 CEST 2016 on sn-devel-144
The previous commit removed all callers of this, so lets remove it.
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
Autobuild-User(master): Volker Lendecke <vl@samba.org>
Autobuild-Date(master): Tue Oct 11 14:44:03 CEST 2016 on sn-devel-144
The offline VFS functions predate the SMB_VFS_{GET|SET}_DOS_ATTRIBUTES()
functions, now that we have these, we can use them for the offline
attribute as well.
The primary reason for this is: performance. Merging both functions has
the benefit that in VFS modules that use same backing store bits for
both offline attribute and DOS attributes (like gpfs), we avoid calling
the backing store twice in dos_mode() and file_set_dosmode().
This commit modifies all existing users of the offline attribute to
adapt to the change, the next commit will then remove the obsolete
offline functions.
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
ACE4_SYSTEM_ALARM_ACE_TYPE does not exist
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 11 02:17:10 CEST 2016 on sn-devel-144
Early return in case of failure to set snapdir-entry-path xlator option
leaks talloced tmp_ctx.
Signed-off-by: Anoop C S <anoopcs@redhat.com>
Reviewed-by: Guenther Deschner <gd@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
Autobuild-User(master): Günther Deschner <gd@samba.org>
Autobuild-Date(master): Fri Oct 7 19:39:11 CEST 2016 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): Tue Sep 27 04:03:03 CEST 2016 on sn-devel-144
This fixes a corner case when using NFS4 ACLs with ID_TYPE_BOTH. Before
this patch, the owner entry in the ACL would be mapped to a gid entry in
the NFSv4 ACL, and not the expected special owner entry. This is caused
by the id mapping returning a valid gid and the nfs4 mapping assumed
that this was actually a group.
Fix this by asking for the uid first, and explicitly checking if the
mapped uid matches the owner. That creates a uid entry in the NFSv4 ACL
that can be changed later in smbacl4_substitute_{simple,special} to the
expected special owner entry.
Signed-off-by: Christof Schmitt <cs@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
When "ignore system acls" is set to "yes, we need to ensure filesystem
permission always grant access so that when doing our own access checks
we don't run into situations where we grant access but the filesystem
doesn't.
Bug: https://bugzilla.samba.org/show_bug.cgi?id=12181
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): Wed Aug 31 18:41:20 CEST 2016 on sn-devel-144
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Ralph Böhme <slow@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Wed Aug 31 01:06:43 CEST 2016 on sn-devel-144
Existing behaviour is "posix" style. Next commit will (re)add the
"windows" style. This commit doesn't change behaviour in any way.
Bug: https://bugzilla.samba.org/show_bug.cgi?id=12177
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
If ignore_system_acls is set and we're synthesizing a default ACL, we
were fetching the filesystem ACL just to free it again. This change
avoids this.
Bug: https://bugzilla.samba.org/show_bug.cgi?id=12177
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Will be reused in the next commit when moving the
make_default_filesystem_acl() stuff to a different place.
Bug: https://bugzilla.samba.org/show_bug.cgi?id=12177
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Better for performance and a subsequent commit will add one more option
where this will pay off.
Bug: https://bugzilla.samba.org/show_bug.cgi?id=12177
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
No change in behaviour (hopefully! :-). This paves the way for moving
the ACL blob validation to a helper function in the next commit.
Bug: https://bugzilla.samba.org/show_bug.cgi?id=12177
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
The variables are already set to NULL by TALLOC_FREE.
Bug: https://bugzilla.samba.org/show_bug.cgi?id=12177
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
In most realistic cases the "next" VFS op will return the permissions
from the filesystem. This rename makes it explicit where the SD is
originating from. No change in behaviour.
This just paves the way for a later change that will simplify the whole
logic and talloc hierarchy.
Bug: https://bugzilla.samba.org/show_bug.cgi?id=12177
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
This makes it explicit where the SD is originating from. No change in
behaviour.
This just paves the way for a later change that will simplify the whole
logic and talloc hierarchy, therefor this also strictly renames the
occurences after the out label.
Logically, behind the out label, we're dealing with a variable that
points to what we're going to return, so the name psd_blob is
misleading, but I'm desperately trying to avoid logic changes in this
commit and therefor I'm just strictly renaming.
Bug: https://bugzilla.samba.org/show_bug.cgi?id=12177
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
This reverts commit 961c4b591b.
Subsequent commits will add the same functionality as an optional
feature.
Bug: https://bugzilla.samba.org/show_bug.cgi?id=12177
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
During path checking, the vfs connectpath_fn is called to
determine the share's root, relative to the file being
queried (for example, in snapshot file this may be other
than the share's "usual" root directory). connectpath_fn
must be able to answer this question even if the path does
not exist and its parent does exist. The convention in this
case is that this refers to a yet-uncreated file under the parent
and all queries are relative to the parent.
This also serves as a workaround for the case where connectpath_fn
has to handle wildcards, as with the case of SMB1 trans2 findfirst.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12172
Signed-off-by: Uri Simchoni <uri@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Thu Aug 25 05:35:29 CEST 2016 on sn-devel-144
No longer needed after conversion to pthreadpool_tevent
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Prerequisite to convert the vfs _send/recv functions
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12150
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Christof Schmitt <cs@samba.org>
Reviewed-by: David Disseldorp <ddiss@samba.org>
Autobuild-User(master): David Disseldorp <ddiss@samba.org>
Autobuild-Date(master): Thu Aug 18 06:43:02 CEST 2016 on sn-devel-144
Make the code closer to the same functionality in shadow_copy2.c:shadow_copy2_strip_snapshot().
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12150
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Christof Schmitt <cs@samba.org>
Reviewed-by: David Disseldorp <ddiss@samba.org>
p and name don't change, and we've already calculated this length.
Part of the effort to make the code inside vfs_snapper.c that does
the same thing more similar (we can't make these functions identical
due to the 'snapdir' use case).
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12150
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Christof Schmitt <cs@samba.org>
Reviewed-by: David Disseldorp <ddiss@samba.org>
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12150
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Christof Schmitt <cs@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Wed Aug 17 05:25:30 CEST 2016 on sn-devel-144
Most calls to disk_quotas originate at a state with an
open file descriptor. Pass the file's stat info down to
disk_quota, so that we can avoid extra stat's and the related
error handling.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12145
Signed-off-by: Uri Simchoni <uri@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
This reduces user-space CPU for metadata intensive workloads, lp_parm_* is
expensive doing lots of strwicmp.
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): Fri Aug 12 21:03:36 CEST 2016 on sn-devel-144
vfs_gpfs & others will be able to embed this structure in their special config.
We could have gone with an anonymous struct and a talloc'ed object, but for my
taste this is specialized and hidden enough that it's not worth the
indirection.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Pure placeholder right now, this will allow vfs modules to load the params in
advance
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Pure placeholder right now, this will allow vfs modules to load the params in
advance
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Pure placeholder right now, this will allow vfs modules to load the params
in advance. nfs4 acl parameters should not change while a tcon is live,
and lp_parm_get_* show up in profiles. Loading the parameters once at
tcon time will remove this.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
This will disable the vfs_glusterfs and vfs_cephfs modules if
you provide --without-acl-support.
In addition it makes compiling in POSIXACL_XATTR conditional
on vfs_glusterfs or vfs_cephfs being built in.
Signed-off-by: Ira Cooper <ira@samba.org>
Reviewed-by: Steve French <sfrench@samba.org>
Tested-by: Steve French <sfrench@samba.org>
Autobuild-User(master): Ira Cooper <ira@samba.org>
Autobuild-Date(master): Mon Aug 8 21:27:46 CEST 2016 on sn-devel-144
Memory was freed in most of the failure cases. It is always better
to free the memory at the end of the function so that all exit path
of the function will free the memory. Otherwise chances are that
you might miss some cleanup.
Signed-off-by: Rajesh Joseph <rjoseph@redhat.com>
Reviewed-by: Uri Simchoni <uri@samba.org>
Reviewed-by: Michael Adam <obnox@samba.org>
With growing number of snapshots file-systems need some mechanism
to differentiate one set of snapshots from other, e.g. monthly, weekly,
manual, special events, etc. Therefore these file-systems provide
different ways to tag snapshots, e.g. provide a configurable way to
name snapshots, which is not just based on time. With only shadow:format
it is very difficult to filter these snapshots.
As part of this change added two new options, shadow:snapprefix and
shadow:delimiter, in shadow_copy2 config. This option will accept regular
expression (BRE) as input. With this optional parameter, one can specify a
variable prefix component for names of the snapshot directories in the
file-system. If this parameter is set, together with the shadow:format and
shadow:delimiter parameters it determines the possible names of snapshot
directories in the file-system.
e.g.
shadow:snapprefix = [a-z]*[0-9]
When this option is provided then shadow:format option should always
start with <delimiter> string. This delimiter is configurable via a new option,
i.e. shadow:delimiter. Default value for this is "_GMT",
e.g. _GMT-%Y.%m.%d-%H.%M.%S
Signed-off-by: Rajesh Joseph <rjoseph@redhat.com>
Reviewed-by: Uri Simchoni <uri@samba.org>
Reviewed-by: Michael Adam <obnox@samba.org>
Create a separate structure to store module specific information. Currently
only config values are saved. As of now there is no cleaner way to store run-time
information or other module specific information in shadow_copy2 module.
Therefore created a new structure to store all module specific information
including config.
Signed-off-by: Rajesh Joseph <rjoseph@redhat.com>
Reviewed-by: Uri Simchoni <uri@samba.org>
Reviewed-by: Michael Adam <obnox@samba.org>
This function returns -1 on error but the return type is
size_t which is unsigned.
Signed-off-by: Rajesh Joseph <rjoseph@redhat.com>
Reviewed-by: Uri Simchoni <uri@samba.org>
Reviewed-by: Michael Adam <obnox@samba.org>
Samba always tries to read the ACL of a file and checks it internally.
If the READ_ACL permission is missing in GPFS, then then reading the ACL
for Samba internal evaluation will be denied and opening the file or
directory fails. Change this by retrying reading the ACL with the DAC
capability if access was denied.
Signed-off-by: Christof Schmitt <cs@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Christof Schmitt <cs@samba.org>
Autobuild-Date(master): Mon Jul 25 10:30:02 CEST 2016 on sn-devel-144
Signed-off-by: Yan, Zheng <zyan@redhat.com>
Signed-off-by: Ira Cooper <ira@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Ira Cooper <ira@samba.org>
Autobuild-Date(master): Sun Jul 24 04:08:23 CEST 2016 on sn-devel-144
Move posix acl related code in vfs_glusterfs.c to a seperate module.
Signed-off-by: Yan, Zheng <zyan@redhat.com>
Signed-off-by: Ira Cooper <ira@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Pair-programmed-with: Andrew Bartlett <abartlet@samba.org>
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Even with "ignore system acls" set to "yes", for objects without NT ACL
xattr we use the underlying filesystem permissions to construct an NT
ACL. This can result in *very* unexpected permissions, eg:
- a directory with the following ACL:
$ ./bin/smbcacls -Uslow%pass //localhost/normal ""
REVISION:1
CONTROL:SR|DP
OWNER:SLOW\slow
GROUP:Unix Group\root
ACL:SLOW\slow:ALLOWED/0x0/FULL
So only one non-inheritable(!) ACE.
- creating a subdirectory:
$ ./bin/smbclient -Uslow%pass //localhost/normal -c "mkdir dir1"
- checking whether there's an ACL xattr:
$ getfattr -m "" /Volumes/normal/dir1
getfattr: Removing leading '/' from absolute path names
system.posix_acl_access
system.posix_acl_default
user.DOSATTRIB
So there isn't an ACL xattr, because there where no inheritable ACEs on
the parent folder.
- reading the new subdirectories ACL:
$ ./bin/smbcacls -Uslow%pass //localhost/normal "dir1"
REVISION:1
CONTROL:SR|DP
OWNER:SLOW\slow
GROUP:Unix Group\slow
ACL:SLOW\slow:ALLOWED/0x0/FULL
ACL:Unix Group\slow:ALLOWED/0x0/READ
ACL:Everyone:ALLOWED/0x0/READ
ACL:NT Authority\SYSTEM:ALLOWED/0x0/FULL
The ACES for "SLOW\slow", "Unix Group\slow" and "Everyone" are coming
from the underlying filesystem. This is the problem.
- Windows assigns the following ACL in this situation:
$ ./bin/smbcacls -UAdministrator%Passw0rd //10.10.10.14/data "dir"
REVISION:1
CONTROL:SR|PD|DI|DP
OWNER:VORDEFINIERT\Administratoren
GROUP:WIN2008R2\Domänen-Benutzer
ACL:WIN2008R2\Administrator:ALLOWED/0x0/FULL
$ ./bin/smbclient -UAdministrator%Passw0rd //10.10.10.14/data -c "mkdir dir\dir1"
$ ./bin/smbcacls -UAdministrator%Passw0rd //10.10.10.14/data "dir\dir1"
REVISION:1
CONTROL:SR|DI|DP
OWNER:VORDEFINIERT\Administratoren
GROUP:WIN2008R2\Domänen-Benutzer
ACL:VORDEFINIERT\Administratoren:ALLOWED/0x0/FULL
ACL:NT-AUTORITÄT\SYSTEM:ALLOWED/0x0/FULL
By changing make_default_filesystem_acl() to only adds user and system
ACE to the ACL of objects that lack an ACL xattr, we match Windows
behaviour:
$ ./bin/smbclient -Uslow%pass //localhost/normal -c "mkdir dir2"
$ ./bin/smbcacls -Uslow%pass //localhost/normal "dir2"
REVISION:1
CONTROL:SR|DP
OWNER:SLOW\slow
GROUP:Unix Group\slow
ACL:SLOW\slow:ALLOWED/0x0/FULL
ACL:NT Authority\SYSTEM:ALLOWED/0x0/FULL
Bug: https://bugzilla.samba.org/show_bug.cgi?id=12028
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 Jul 19 10:22:05 CEST 2016 on sn-devel-144
This function is only used in vfs_acl_common.c and will be modified in
the next commit.
Bug: https://bugzilla.samba.org/show_bug.cgi?id=12028
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Reviewed-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 3 07:26:34 CEST 2016 on sn-devel-144
Remove the option to retrieve valid user/group quota while
returning -1 and EDQUOT errno - this is no longer part of the
protocol between the quota backend and smbd.
Signed-off-by: Uri Simchoni <uri@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Add support for mocking FS user/group quotas (default quota and
quota flags).
Make the default block size 4096 instead of 0. This
turns the default into "no quota" instead of "punt to
lower VFS module" (that is, if the mock module is asked
to retrieve quota of a user/group/default for which there
is no config).
Add support for ENOSYS error
BUG: https://bugzilla.samba.org/show_bug.cgi?id=11937
Signed-off-by: Uri Simchoni <uri@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Only in the default VFS. Gpfs, Ceph, Gluster and other modern
backend VFS filesystems might want to do the same.
Allow tuneable "smbd:force process locks = true" to turn
off OFD locks if in use and the kernel doesn't support them.
Display debug message showing admins what to do in this case.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Jeff Layton <jlayton@samba.org>
Will allow us to move to open file description locks
from process-associated locks.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Jeff Layton <jlayton@samba.org>
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 May 19 23:53:47 CEST 2016 on sn-devel-144
https://bugzilla.samba.org/show_bug.cgi?id=11721
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): Thu May 5 00:04:50 CEST 2016 on sn-devel-144
Usually off_t is the same as size_t, but not on a 32 bit machine where
_FILE_OFFSET_BITS is set to 64. In that case size_t stays 32 bit, and
off_t goes to 64.
This only affects debug messages.
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
add_srt should add the mappings to the linked list even if
mappings==NULL (the default)
Bug: https://bugzilla.samba.org/show_bug.cgi?id=11827
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): Mon Apr 11 14:25:59 CEST 2016 on sn-devel-144
The xattr functions intercepted only the calls from dosmode. With the
implementation of the dos_attribute interface, the xattr codepaths never
get called and can be removed.
Signed-off-by: Christof Schmitt <cs@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Signed-off-by: Uri Simchoni <uri@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Tue Mar 29 00:12:14 CEST 2016 on sn-devel-144
This will make it easier to support those systems and file systems that
can store DOS attributes. It should retain the original functionality if
VFS functions providing these things are not provided.
Signed-off-by: Richard Sharpe <rsharpe@samba.org>
Signed-off-by: Jeremy Allison <jra@samba.org>
Check the smb_filename->flags field, or req->posix_pathnames
instead, depending on what is available.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
We don't have access to a struct smb_filename here and
can't get one, so simply always set AT_SYMLINK_NOFOLLOW
and remove the optimization if we ended up fstatat()'ing
a symlink, as we don't know if the caller wanted a link
stat or not.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
Get it from parent/deriving smb_filename if present.
Use 0 (as usually this a Windows-style lookup) if
not.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
Only one defined flag for now, SMB_FILENAME_POSIX_PATH.
Define as the same as FSP_POSIX_FLAGS_PATHNAMES to keep
the value consistent.
Set this inside unix_convert() when a posix path parse
is selected.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
For a posix acl call on a symlink, we've already refused it.
For a Windows acl mapped call on a symlink, we want to follow
it.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
Missed conversion of get_nt_acl_fn from const char *
to const struct smb_filename *.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
When "ignore system acls" is set, do not mess at all with POSIX ACLS,
do not even calculate the would-be POSIX-ACL-based security descriptor
(for performance reasons).
Instead, just store a V3 blob with zero hash. This means that if we
later read the ACL without ignoring system ACLs, the NT ACL shall be
reset to the info derivable from the POSIX ACL.
File ownership is still modified as it has bearing on disk quotas.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=11806
Signed-off-by: Uri Simchoni <uri@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
At present, substitutions in the streams directory path are ignored. Fix it
by modifying 'stream_dir' function to call 'lp_parm_talloc_string' which
internally calls 'lp_string' on the path.
Signed-off-by: Shyamsunder Rathi <shyam.rathi@nutanix.com>
Reviewed-by: Uri Simchoni <uri@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
Autobuild-User(master): Uri Simchoni <uri@samba.org>
Autobuild-Date(master): Tue Mar 15 11:45:19 CET 2016 on sn-devel-144
For a posix acl call on a symlink, we've already refused it.
For a Windows acl mapped call on a symlink, we want to follow
it.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
For a posix acl call on a symlink, we've already refused it.
For a Windows acl mapped call on a symlink, we want to follow
it.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
Push the struct further down closer to places that use
lp_posix_pathname() functions.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Ralph Boehme <rb@sernet.de>
Missed conversion of get_nt_acl_fn from const char *
to const struct smb_filename *.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Ralph Boehme <rb@sernet.de>
Prepare for changing vfs_streaminfo to do the same.
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): Mon Mar 7 21:12:56 CET 2016 on sn-devel-144
The wrapper->state pointer is not getting NULLed during free
allowing use of freed memory, causing a crash.
Thanks to Red Hat for discovering this issue.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=11774
Signed-off-by: Ira Copper <ira@samba.org>
Reviewed-by: Poornima G <pgurusid@redhat.com>
Tested-by: Christopher Blum <cblum@redhat.com>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Sat Mar 5 09:08:53 CET 2016 on sn-devel-144
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 Mar 4 02:09:15 CET 2016 on sn-devel-144
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
Autobuild-User(master): Uri Simchoni <uri@samba.org>
Autobuild-Date(master): Wed Mar 2 10:50:58 CET 2016 on sn-devel-144
Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Wed Mar 2 04:34:15 CET 2016 on sn-devel-144
Finally! The previous commits changed the VFS and the async backend to
pass the duration of an aync operation up the stack.
We now can use this value instead of doing our own sampling which avoids
the following problem:
1. SMB2 read request received, added to the async queue
2. SMB2 create_file request comes in, is processed and blocks for N
seconds in open()
3. async read completes in the dispatcher thread, completion callback
will be called when we enter the main tevent loop
4. open() completes after N seconds
5. main tevent event loop is entered, async results are processed
6. async read result is processed, time sampling will include the N
seconds blocked in open()
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Subsequent commits that are going to track aio request duration in the
aio backends will use this.
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Preparing to reduce use of lp_posix_pathnames().
Uses the same techniques as commit 616d068f0c
(synthetic_smb_fname()) to cope with modules that
modify the incoming pathname.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
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 Feb 26 01:21:21 CET 2016 on sn-devel-144
The `unused` member being set to zero is an uint64_t array. We want the
byte size here.
Signed-off-by: Aurelien Aptel <aaptel@suse.com>
Reviewed-by: David Disseldorp <ddiss@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
According to discussions with dochelp@microsoft.com, an SMB
server should ignore an SVHDX_OPEN_DEVICE_CONTEXT or
SVHDX_OPEN_DEVICE_CONTEXT_V2 create context if it does not
support the RSVD protocol. This is contrary to [MS-SMB2] rev 48.0
which states (3.3.5.9.14) that the open should fail in this case.
Failing the create fails Windows backup if the SMB dialect is
SMB3.0.2 or higher.
Hopefully a new revision of MS-SMB2 will clear this up in the future.
Meanwhile, this patch modifies smbd to ignore the
SVHDX_OPEN_DEVICE_CONTEXT by default. This can be overriden by a VFS
module if a VFS module adds support for RSVD.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=11753
Signed-off-by: Uri Simchoni <uri@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Richard Sharpe <realrichardsharpe@gmail.com>
Preparing to reduce use of lp_posix_pathnames().
Uses the same techniques as commit 616d068f0c
(synthetic_smb_fname()) to cope with modules that
modify the incoming pathname.
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): Thu Feb 25 20:46:49 CET 2016 on sn-devel-144
Preparing to reduce use of lp_posix_pathnames().
Uses the same techniques as commit 616d068f0c
(synthetic_smb_fname()) to cope with modules that
modify the incoming pathname.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
Autobuild-User(master): Uri Simchoni <uri@samba.org>
Autobuild-Date(master): Wed Feb 24 16:05:55 CET 2016 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): Tue Feb 16 23:26:48 CET 2016 on sn-devel-144
Bumps VFS version to 35.
Preparing to reduce use of lp_posix_pathnames().
Most of this is boilerplate, the only subtleties are in
the modules:
vfs_catia.c
vfs_media_harmony.c
vfs_shadow_copy2.c
vfs_unityed_media.c
Where the path is modified then passed to SMB_VFS_NEXT_GET_NT_ACL().
In these cases the change uses synthetic_smb_fname() to
create a new struct smb_filename from the modified path.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Ralph Boehme <rb@sernet.de>
Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
Autobuild-User(master): Uri Simchoni <uri@samba.org>
Autobuild-Date(master): Mon Feb 1 22:09:18 CET 2016 on sn-devel-144
Add sys_realpath() function that captures the OS variations
on realpath().
Signed-off-by: Uri Simchoni <uri@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
add implementation of get_quota_fn to vfs_gpfs. The implemetation
returns ENOSYS for the case of user and group quota, to make sure
the default VFS does not accidentally succeed (and return wrong
values which would alter the disk-free calculation)
For other quota types the function calls the underlying VFS as
before.
Signed-off-by: Uri Simchoni <uri@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
When mocking disk-free, do not take quota into
account since this is now done in the SMB layer.
Signed-off-by: Uri Simchoni <uri@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
Move most of the logic that handles determination of
disk size and free space from default VFS handler to
the SMB layer - letting the VFS handle the basic task
of querying the file system for general stats and
quota.
Signed-off-by: Uri Simchoni <uri@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
This is in preparation for handling the quota part
of disk_free via the VFS - each module with a
disk_free_fn should also have a get_quota_fn.
Signed-off-by: Uri Simchoni <uri@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
This is in preparation for handling the quota part
of disk_free via the VFS - each module with a
disk_free_fn should also have a get_quota_fn.
Signed-off-by: Uri Simchoni <uri@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
This is in preparation for handling the quota part
of disk_free via the VFS - each module with a
disk_free_fn should also have a get_quota_fn.
Signed-off-by: Uri Simchoni <uri@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
Adding a path parameter would allow the VFS get_quota
function to be used for determining the quota/usage
when calculating size and free spacei.
Signed-off-by: Uri Simchoni <uri@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
Add a vfs module "vfs_fake_dfq" for mocking disk-free
and user/group quota functions.
In this commit there are quota considerations in
disk_free calculations, based on the mocking of quota.
Those considerations will later be removed once we
refactor the server code to weigh the disk-free and quota in
the smb layer and not in individual vfs implementations.
Signed-off-by: Uri Simchoni <uri@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
This new parameter defines how to get from the snapshot's
root directory to the share's root directory. It is an
alternative to the "basedir" parameter, but functionally
is a superset of basedir.
Signed-off-by: Uri Simchoni <uri@samba.org>
Reviewed-by: Michael Adam <obnox@samba.org>
Remove the basedir state variable from the module-specific data
of vfs_shadow_copy2 - this variable is not being used.
Signed-off-by: Uri Simchoni <uri@samba.org>
Reviewed-by: Michael Adam <obnox@samba.org>
If crossmountpoints is enabled, verify that snapdirseverywhere is
enabled too, since crossmountpoints has no meaning otherwise.
This obviates the check of crossmountpoints against other config
variables.
Signed-off-by: Uri Simchoni <uri@samba.org>
Reviewed-by: Michael Adam <obnox@samba.org>
Autobuild-User(master): Michael Adam <obnox@samba.org>
Autobuild-Date(master): Fri Jan 22 01:54:06 CET 2016 on sn-devel-144
Adjust the connect path to the root of the share in the snapshot,
or to the root of the snapshot if the snapshot is "inside" the
share.
This way snapshot symlink regarded as "wide links" if and only if
they point outside the snapshot or they were wide links when
the snapshot was taken.
This allows mounting the snapshots outside the share's root.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=11580
Signed-off-by: Uri Simchoni <uri@samba.org>
Reviewed-by: Michael Adam <obnox@samba.org>
Add a new routine shadow_copy2_do_convert() which is
like shadow_copy2_convert(), but beside calculating the
path of the snapshot file, it also returns the minimum length
of the subpath which is both inside the share and inside the
same snapshot as the file in question, i.e. (at least in the
classical case) the subdirectory of the the snapshot file's
snapshot directory that corresponds to the file's share root.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=11580
Signed-off-by: Uri Simchoni <uri@samba.org>
Reviewed-by: Michael Adam <obnox@samba.org>
The debug message for the mountpoint option was the
wrong way around. This fixes it, also converting
the message to use DBG_WARNING() while we're touching it.
Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Uri Simchoni <uri@samba.org>
Autobuild-User(master): Uri Simchoni <uri@samba.org>
Autobuild-Date(master): Wed Jan 13 10:56:42 CET 2016 on sn-devel-144
OS X ignores deletes on the AFP_Resource stream. This was discovered by
torture tests against OS X SMB server.
Bug: https://bugzilla.samba.org/show_bug.cgi?id=11347
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
When reading from the AFP_AfpInfo stream, OS X ignores the offset from
the request and always reads from offset=0.
The offset bounds check has a off-by-1 bug in OS X, so a request
offset=60 (AFP_AfpInfo stream has a ficed size of 60 bytes), len=1
returns 1 byte from offset 0 insteaf of returning 0.
Bug: https://bugzilla.samba.org/show_bug.cgi?id=11347
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
When writing all 0 to AFP_AfpInfo stream we can remove the underlying
storage object. This beaviour of OS X SMB server was found with a
torture test.
Bug: https://bugzilla.samba.org/show_bug.cgi?id=11347
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
With help of some torture tests I verified the following behaviour of OS
X SMB server:
* ftruncate AFP_AfpInfo stream > 60 bytes results in an error
NT_STATUS_ALLOTTED_SPACE_EXCEEDED
* ftruncate AFP_AfpInfo stream <=60 returns success but has no effect
Bug: https://bugzilla.samba.org/show_bug.cgi?id=11347
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Signed-off-by: Jose A. Rivera <jarrpa@samba.org>
Reviewed-by: Ira Cooper <ira@samba.org>
Autobuild-User(master): Ira Cooper <ira@samba.org>
Autobuild-Date(master): Thu Dec 17 20:55:22 CET 2015 on sn-devel-144
This activates the new AIO code's cancellation logic.
Signed-off-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 15 23:33:12 CET 2015 on sn-devel-104
Move the functions to a more logical location.
Signed-off-by: Ira Cooper <ira@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Sat Dec 12 01:03:40 CET 2015 on sn-devel-104
These two functions are basically the same thing, so they should be
formatted the same.
Signed-off-by: Ira Cooper <ira@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
