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.
Looks larger than it is, this just adds a parameter and while there
adapts long lines to README.Coding
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Uses non-DFS names and DFS-names against a DFS share, shows that Windows
looks correctly at the DFS flag when SMB2 requests are
made on a DFS share. Passes against Windows 2022.
Mark as knownfail for smbd.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Noel Power <npower@samba.org>
Autobuild-User(master): Noel Power <npower@samba.org>
Autobuild-Date(master): Wed Sep 28 19:34:29 UTC 2022 on sn-devel-184
There is only one difference between Windows 2022 and Windows 2008.
Opening an empty ("") DFS path succeeds in opening the share
root on Windows 2008 but fails with NT_STATUS_INVALID_PARAMETER
on Windows 2022. Allow the test to cope with both.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Noel Power <npower@samba.org>
Passes against Windows.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Noel Power <npower@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Wed Sep 14 18:37:06 UTC 2022 on sn-devel-184
NB. This passes against Windows, but SMBctemp is broken on a Windows DFS
share and always returns NT_STATUS_FILE_IS_A_DIRECTORY.
When we fix the Samba server to correctly process DFS
pathnames we'll have to change this test to understand
it's running against smbd and modify the expected behavior
to match a working server.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Noel Power <npower@samba.org>
Only tests SMB1unlink for now, but I will add other operations
later.
smbtorture3 test is: SMB1-DFS-OPERATIONS.
Passes fully against Windows. Adds knownfail for smbd.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Noel Power <npower@samba.org>
This one is tricky. It sends SMB2 DFS pathnames to a non-DFS
share, and sets the SMB2 flag FLAGS2_DFS_PATHNAMES in the SMB2
packet.
Windows will have non of it and (correctly) treats the pathnames
as local paths (they're going to a non-DFS share). Samba fails.
This proves the server looks as the share DFS capability to
override the flag in the SMB2 packet.
Passes against Windows. Added knownfail for Samba.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Noel Power <npower@samba.org>
Ensures crtime of the root of the share and a newly created
file crtime are different. Should help avoid mistakes like the
error fixed by the previous commit.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Noel Power <npower@samba.org>
Autobuild-User(master): Noel Power <npower@samba.org>
Autobuild-Date(master): Mon Sep 12 16:21:23 UTC 2022 on sn-devel-184
The test SMB1-DFS-PATHS was using the file ino number
to check for file identity, fetching it using cli_qfileinfo_basic().
This works for SMB2, but the info level used by this for SMB1
(SMB_QUERY_FILE_ALL_INFO) doesn't return the ino number, so
all comparisons were succeeding as zero.
Change to using crtime (create time) for identity comparison
instead. This fix is mostly a rename of ino -> crtime, with
some changes around the tests and printf on error, but it
is easier to do in one go.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Noel Power <npower@samba.org>
smbtorture3 test is: SMB1-DFS-PATHS
Tests open, and then all 4 methods of renaming/hardlinking
files:
1). SMBmv
2). SMBtrans2 SETPATHINFO
3). SMBtrans2 SETFILEINFO
4). SMBntrename
Also added a test for SMB1findfirst.
smbtorture3 test is: SMB1-DFS-SEARCH-PATHS.
What this shows is that Windows strips off the
SMB1findfirst mask *before* calling the DFS path
parser (smbd currently does not).
Added so we know how to fix the server code to match Windows
behavior in parsing DFS paths in different calls going forward.
Passes fully against Windows. Adds knownfails for smbd.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Noel Power <noel.power@suse.com>
Passes fully against Windows.
This shows that DFS paths on Windows on SMB2 must
be of the form:
SERVER\SHARE\PATH
but the actual contents of the strings SERVER and
SHARE don't need to match the given server or share.
The algorithm the Windows server uses is the following:
Look for a '\\' character, and assign anything before
that to the SERVER component. The characters in this
component are not checked for validity.
Look for a second '\\' character and assign anything
between the first and second '\\' characters to the
SHARE component. The characters in the share component
are checked for validity, but only ':' is flagged as
an illegal sharename character despite what:
[MS-FSCC] https://docs.microsoft.com/en-us/openspecs/windows_protocols/ms-fscc/dc9978d7-6299-4c5a-a22d-a039cdc716ea
says.
Anything after the second '\\' character is assigned
to the PATH component and becomes the share-relative
path.
If there aren't two '\\' characters it removes
everything and ends up with the empty string as
the share relative path.
To give some examples, the following pathnames all map
to the directory at the root of the DFS share:
SERVER\SHARE
SERVER
""
ANY\NAME
ANY
::::\NAME
the name:
SERVER\:
is illegal (sharename contains ':') and the name:
ANY\NAME\file
maps to a share-relative pathname of "file",
despite "ANY" not being the server name, and
"NAME" not being the DFS share name we are
connected to.
Adds a knownfail for smbd as our current code
in parse_dfs_path() is completely incorrect
here and tries to map "incorrect" DFS names
into local paths. I will work on fixing this
later, but we should be able to remove parse_dfs_path()
entirely and move the DFS pathname logic before
the call to filename_convert_dirfsp() in the
same way Volker suggested and was able to achieve
for extract_snapshot_token() and the @GMT pathname
processing.
Also proves the "target" paths for SMB2_SETINFO
rename and hardlink must *not* be DFS-paths.
Next I will work on a torture tester for SMB1
DFS paths.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reivewed-by: Noel Power <npower@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Tue Aug 30 17:10:33 UTC 2022 on sn-devel-184
We already don't allow setting max_credits in the sync wrapper, so
omit the contexts there as well.
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 Aug 26 19:54:03 UTC 2022 on sn-devel-184
The Linux prototype for openat2 looks like this:
long openat2(int dirfd, const char *pathname,
struct open_how *how, size_t size);
where "struct open_how" is defined in "linux/openat2.h". It is
designed to be extensible with further flags.
The "size" parameter is required because there is no type checking
between userland and kernelspace, so the way for Linux to find which
version of open_how is being passed in is looking at the size:
"open_how" is expected to only every grow with additional fields,
should a change be necessary in the future.
Samba does not have this problem, we can typecheck the struct and
pointers, we expect all VFS modules to be compiled against the current
vfs.h.
For now this adds no functionality, but it will make further patches
much smaller.
Pair-programmed-with: Stefan Metzmacher <metze@samba.org>
Signed-off-by: Volker Lendecke <vl@samba.org>
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
The destructor triggered by dbwrap_watched_watch_recv() will
remove the watcher instance via a dedicated dbwrap_do_locked(),
just calling dbwrap_watched_watch_remove_instance() inside.
But the typical caller triggers a dbwrap_do_locked() again after
dbwrap_watched_watch_recv() returned. Which means we call
dbwrap_do_locked() twice.
We now allow dbwrap_watched_watch_recv() to return the existing
instance id (if it still exists) and removes the destructor.
That way the caller can pass the given instance id to
dbwrap_watched_watch_remove_instance() from within its own dbwrap_do_locked(),
when it decides to leave the queue, because it's happy with the new
state of the record. In order to get the best performance
dbwrap_watched_watch_remove_instance() should be called before any
dbwrap_record_storev() or dbwrap_record_delete(),
because that will only trigger a single low level storev/delete.
If the caller found out that the state of the record doesn't meet the
expectations and the callers wants to continue watching the
record (from its current position, most likely the first one),
dbwrap_watched_watch_remove_instance() can be skipped and the
instance id can be passed to dbwrap_watched_watch_send() again,
in order to resume waiting on the existing instance.
Currently the watcher instance were always removed (most likely from
the first position) and re-added (to the last position), which may
cause unfair latencies.
In order to improve the overhead of adding a new watcher instance
the caller can call dbwrap_watched_watch_add_instance() before
any dbwrap_record_storev() or dbwrap_record_delete(), which
will only result in a single low level storev/delete.
The returned instance id is then passed to dbwrap_watched_watch_send(),
within the same dbwrap_do_locked() run.
It also adds a way to avoid alerting any callers during
the current dbwrap_do_locked() run.
Layers above may only want to wake up watchers
during specific situations and while it's useless to wake
others in other situations.
This will soon be used to add more fairness to the g_lock code.
Note that this commit only prepares the api for the above to be useful,
the instance returned by dbwrap_watched_watch_recv() is most likely 0,
which means the watcher entry was already removed, but that will change
in the following commits.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15125
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
The main optimization is to avoid non_widelink_open() for streams
opens based on the fact that all streams opens are relative to
fsp->base_fsp, which is a pathref fsp already.
Neither streams_xattr nor streams_depot referenced dirfsp for the
streams case. Make this more obvious in the callers by passing NULL
and asserting this: non-streams opens and streams opens are just
different things, streams-opens can and do reference a base fsp and
don't need the non_widelink_open logic.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
This reverts commit 322574834f1e71bc01f21be9059ca4d386517c84.
Not strictly a revert anymore, but for future work we do need "dirfsp"
in create_file_default() passed through the VFS.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
This will enable a simplification in the stream-handling openat vfs
routines.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
If doing an SMB_VFS_FSTAT() returning onto the stat struct stored in the fsp,
we must call vfs_stat_fsp() as this preserves the iflags.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15022
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
