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.
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>
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
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>
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>
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>
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>
Removes accessor functions as now this parameter is set
under user control in smb.conf. Default is 100.
Note that this doesn't limit the number of outstanding
aio requests, it just causes them to go onto the
pthreadpool queue.
Now we need to prioritize pthreadpool pipe replies
ahead of incoming SMB2 requests, but that's a patch
for another day.
Based on ideas from Volker.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Christof Schmitt <cs@samba.org>
Reviewed-by: Michael Adam <obnox@samba.org>
Also handle a review comment by Metze.
Signed-off-by: Richard Sharpe <rsharpe@samba.org>
Reviewed-by: Ira Cooper <ira@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Fri Aug 28 03:19:36 CEST 2015 on sn-devel-104
sid_parse takes a binary blob, uint8_t reflects this a bit
better than char * does
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
The "mode = " from a very casual view looked as if it was part of the
if-condition
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Ira Cooper <ira@samba.org>
With a central notifyd, we can't do this per share anymore. Notifyd will
only look at absolute paths, not shares.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Use a dynamically allocated copy_chunk buffer size with an upper bound
of 8 MB for now.
The previous size of 64 KB has proven to really hurt performance,
especially with "strict locking = yes".
The SMB2 protocol level maximum allowed copy_chunk size is 1 MB, that's
what will be used as buffer size in the typical case.
With the AAPL copyfile extension the requested copy_chunk size is the
size whole file, which would then make use of a larger buffer up to the
limit of 8 MB.
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
There are still many files to touch before we can remove the #define but this
gets the last include files in source3/includes.
Signed-off-by: Richard Sharpe <rsharpe@samba.org>
Reviewed-by: Jeremy Allson <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Fri May 1 21:52:25 CEST 2015 on sn-devel-104
This change adds three new VFS hooks covering snapshot manipulation:
- snap_check_path
Check whether a path supports snapshots.
- snap_create
Request the creation of a snapshot of the provided path.
- snap_delete
Request the deletion of a snapshot.
These VFS call-outs will be used in future by Samba's File Server Remote
VSS Protocol (FSRVP) server.
MS-FSVRP states:
At any given time, Windows servers allow only one shadow copy set to
be going through the creation process.
Therefore, only provide synchronous hooks for now, which can be
converted to asynchronous _send/_recv functions when the corresponding
DCE/RPC server infrastructure is in place.
Signed-off-by: David Disseldorp <ddiss@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
The Linux fallocate syscall offers a mode parameter which can take the
following flags:
FALLOC_FL_KEEP_SIZE
FALLOC_FL_PUNCH_HOLE (since 2.6.38)
FALLOC_FL_COLLAPSE_RANGE (since 3.15)
FALLOC_FL_ZERO_RANGE (since 3.14)
The flags are not exclusive, e.g. FALLOC_FL_PUNCH_HOLE must be specified
alongside FALLOC_FL_KEEP_SIZE.
Samba currently takes a vfs_fallocate_mode enum parameter for the VFS
fallocate hook, taking either an EXTEND_SIZE or KEEP_SIZE value. This
commit changes the fallocate hook such that it accepts a uint32_t flags
parameter, in preparation for PUNCH_HOLE and ZERO_RANGE support.
Signed-off-by: David Disseldorp <ddiss@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
The small_query parameter for SMB_VFS_DISK_FREE() was, prior to the
previous commit, used to obtain 16-bit wide free-space information for
the deprecated dskattr SMB_COM_QUERY_INFORMATION_DISK command.
With the dskattr handler now performing the 16-bit collapse directly,
the small_query parameter can be removed from the entire code path.
Signed-off-by: David Disseldorp <ddiss@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Tue Feb 17 05:37:20 CET 2015 on sn-devel-104
Signed-of-by: Richard Sharpe <rsharpe@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
Autobuild-User(master): Volker Lendecke <vl@samba.org>
Autobuild-Date(master): Sun Jan 25 12:58:08 CET 2015 on sn-devel-104
This will make it easier to integrate into proper memory hierarchies.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
More like a cleanup, but I want to use inotify_watch in notifyd
that I would like to keep as light as possible
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Notifyd should be as independent of Samba infrastructure as possible,
and it will call notify_inotify
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
vfswrap_fallocate() is broken in that it can call posix_fallocate()
which returns an int error (and doesn't set errno) but can also
call Linux fallocate() which returns -1 and sets errno.
Standardize on the -1,errno convention.
Reported by Jones <jones.kstw@gmail.com> who provided the
initial patch. This patch tested and confirmed working
by him as well.
https://bugzilla.samba.org/show_bug.cgi?id=10982
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: David Disseldorp <ddiss@suse.de>
SMB_VFS_READDIR_ATTR is a last minute hook to fetch additional metadata
for a directory entry when we're already marshalling the SMB reply
buffer.
This would be used, when there's a need to repurpose some fields in the
the reply, like it's done with Apple's SMB2 extension "AAPL".
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
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 Nov 19 23:13:10 CET 2014 on sn-devel-104
srvstr_push_fn() now returns an NTSTATUS reporting any
string conversion failure.
We need to get serious about returning character set conversion errors
inside smbd.
Bug 10775 - smbd crashes when accessing garbage filenames
https://bugzilla.samba.org/show_bug.cgi?id=10775
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: David Disseldorp <ddiss@samba.org>
This doesn't belong to a connection, it's state used within the VFS stack.
And smbd_server_connection is the legacy structure we're using for
global VFS state.
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Michael Adam <obnox@samba.org>
In changes to come this will be possible for an INTERNAL_OPEN_ONLY.
The protection was already in place for some code paths, this
makes the coverage compete.
Bug 10564 - Lock order violation and file lost
https://bugzilla.samba.org/show_bug.cgi?id=10564
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
Some file systems do not allocate a block for very
small files. But for non-empty file should report a
positive size.
Pair-Programmed-With: Michael Adam <obnox@samba.org>
Signed-off-by: Björn Baumbach <bb@sernet.de>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Sat Apr 5 03:09:00 CEST 2014 on sn-devel-104
When multiple aio requests finish simultaneously, this saves a few syscalls
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 Mar 27 08:05:46 CET 2014 on sn-devel-104
This makes use of C99 dynamic arrays. In this performance-sensitive code, I
would like to avoid malloc/free, and I think 15 years after the standard we
might be able to use this feature. Alternatively, we could use the "results"
memory area and store the jobids in the upper range, playing some cast-tricks.
Should work as well.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
The current FSCTL_GET_SHADOW_COPY_DATA/FSCTL_SRV_ENUMERATE_SNAPSHOTS
handler squashes all non-ENOSYS errors into an NT_STATUS_UNSUCCESSFUL
response.
This commit ensures that all errors are propagated up to the client, to
aid debugging.
The Windows Explorer "Previous Versions" dialogue doesn't appear to
distinguish between error codes. It displays "There are no previous
versions available".
Signed-off-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 Feb 27 22:42:50 CET 2014 on sn-devel-104
We have to ship 64 bytes, and we have to initialize the whole thing.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Michael Adam <obnox@samba.org>
Signed-off-by: Christian Ambach <ambi@samba.org>
Reviewed-by: Michael Adam <obnox@samba.org>
Autobuild-User(master): Michael Adam <obnox@samba.org>
Autobuild-Date(master): Sat Dec 7 14:20:43 CET 2013 on sn-devel-104
The VFS interfaces are sychronous, as the operations only modify
meta-data.
These hooks are dependent on support for transparent compression by the
underlying filesystem - vfs_default returns INVALID_DEVICE_REQUEST.
Support for other filesystems providing transparent comression, such as
Btrfs and ZFS, can be added in future.
The get_compression function takes fsp and smb_fname arguments. The
smb_fname argument is needed due to the current dosmode() code-path.
Signed-off-by: David Disseldorp <ddiss@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
labels_data_count already accounts for the unicode null character at the
end of the array. There is no need in adding space for it again.
Signed-off-by: Christof Schmitt <christof.schmitt@us.ibm.com>
Reviewed-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Simo Sorce <idra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Tue Aug 6 04:03:17 CEST 2013 on sn-devel-104
Otherwise num_volumes and the end marker can return uninitialized data
to the client.
Signed-off-by: Christof Schmitt <christof.schmitt@us.ibm.com>
Reviewed-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Simo Sorce <idra@samba.org>
Tests show significant speedup in directory listings
by using fstatat instead of a full pathname walk.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
Autobuild-User(master): Andreas Schneider <asn@cryptomilk.org>
Autobuild-Date(master): Mon Jun 10 20:14:12 CEST 2013 on sn-devel-104
Eliminates the need to hold locks across the
entire lifetime of the call.
Next commit will remove these.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: David Disseldorp <ddiss@samba.org>
copy_chunk copies n bytes from a source file at a specific offset to a
destination file at a given offset. This interface will be used in
handling smb2 FSCTL_SRV_COPYCHUNK ioctl requests.
Use a pread/pwrite loop in vfs_default, so that requests referring to
the same src and dest file are possible.
Provide send and receive hooks for copy chunk VFS interface, allowing
asynchronous behaviour.
Check whether the request source offset + length exceeds the current
size. Return STATUS_INVALID_VIEW_SIZE under such a condition, matching
Windows server behaviour.
Reviewed by: Jeremy Allison <jra@samba.org>
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Autobuild-User(master): Volker Lendecke <vl@samba.org>
Autobuild-Date(master): Mon Nov 5 19:01:13 CET 2012 on sn-devel-104
This makes it clear which context the returned SD is allocated on, as
a number of callers do not want it on talloc_tos().
As the ACL transformation allocates and then no longer needs a great
deal of memory, a talloc_stackframe() call is used to contain the
memory that is not returned further up the stack.
Andrew Bartlett
This changes from allocation on NULL to allocation on the supplied
memory context.
Currently that supplied context is talloc_tos() at the the final consumer of
the ACL.
Andrew Bartlett
We only grant durable handles for CIFS/SMB2 only access,
that means "kernel oplocks", "kernel share modes" and "posix locking"
need to be set to "no".
For now we also don't grant durable handles if delete on close
is active on the handle.
Pair-Programmed-With: Stefan Metzmacher <metze@samba.org>
Pair-Programmed-With: Volker Lendecke <vl@samba.org>
We no longer do struct smb_acl_t manipuations via the VFS layer,
which is now reduced to handling the get/set functions.
The only backend that implemented these functions (aside from audit)
was the vfs_default module calling the sys_acl code. The various ACL
implementation modules either worked on the fully initilaised
smb_acl_t object or on NT ACLs.
This not only makes the operation of the posix ACL code more efficient
(as allocation and free is not put via the VFS), it makes it easier to
test and removes the fantasy that a module could safely redefine this
structure or the behaviour here.
The smb_acls.idl now defines the structure, and it is now allocated
with talloc.
These operations were originally added to the VFS in commit
3bb219161a.
Andrew Bartlett
They use talloc_tos() internally: hoist that up to the callers, some
of whom don't want to us talloc_tos().
A simple patch, but hits a lot of files.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Based on a fix from "Etienne Dechamps " <e-t172@akegroup.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Thu Jun 21 21:53:58 CEST 2012 on sn-devel-104
Note: it is actually wrong to access the fsp->fnum at all here,
since the fnum is part of the smb layer that should not be used
in the vfs layer. But this is subject be separated more cleanly
in later commits. This change only unifies the logging of fsp->fnum.
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Signed-off-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Tue Jun 12 02:26:31 CEST 2012 on sn-devel-104
On platforms that don't have an RT signal space, signal initialization
fails. aio_fork and aio_pthread don't need the signal, so this would
block them from running as well.
on a 32-bit system and defined as a long, then inside vfswrap_get_alloc_size()
we cast to a uint64_t. This sign-extends when converting to unsigned,
so if the high bit of st_ex_blksize is set we return insane values to clients.
Now that we always require a 64 bit off_t, we no longer need SMB_OFF_T.
Andrew Bartlett
Autobuild-User: Andrew Bartlett <abartlet@samba.org>
Autobuild-Date: Fri Apr 6 01:47:43 CEST 2012 on sn-devel-104
This removes a dependency on "struct notify_entry" and makes the nature of the
API more explicit. We depend upon the VFS module to mask out elements from
e->filter and e->subdir_filter that it took over to handle.
Autobuild-User: Volker Lendecke <vl@samba.org>
Autobuild-Date: Mon Mar 26 17:45:44 CEST 2012 on sn-devel-104
This should finally fix the AIX build and allow to remove AIX specific ifdefs.
Guenther
Signed-off-by: Jeremy Allison <jra@samba.org>
Autobuild-User: Jeremy Allison <jra@samba.org>
Autobuild-Date: Thu Apr 21 02:01:20 CEST 2011 on sn-devel-104
There are systems where ./configure has detected advanced utimes calls which
are then not available on other kernels. We should do a proper fallback.
Autobuild-User: Volker Lendecke <vlendec@samba.org>
Autobuild-Date: Tue Feb 1 22:37:35 CET 2011 on sn-devel-104
strict allocation on sparse files. Files opened as POSIX opens are always
sparse.
Autobuild-User: Jeremy Allison <jra@samba.org>
Autobuild-Date: Tue Dec 21 04:12:22 CET 2010 on sn-devel-104
allocation extent without changing end-of-file size.
Autobuild-User: Jeremy Allison <jra@samba.org>
Autobuild-Date: Tue Dec 21 02:41:24 CET 2010 on sn-devel-104
It turns out we need the fallocate operations to be able to both
allocate and extend filesize, and to allocate and not extend
filesize, and posix_fallocate can only do the former. So by defining
the vfs op as posix_fallocate we lose the opportunity to use any
underlying syscalls (like Linux fallocate) that can do the latter
as well.
We don't currently use the non-extending filesize call, but now
I've changed the vfs op definition we can in the future. For the
moment simply map the fallocate op onto posix_fallocate for the
VFS_FALLOCATE_EXTEND_SIZE case and return ENOSYS for the
VFS_FALLOCATE_KEEP_SIZE case.
Jeremy.
Autobuild-User: Jeremy Allison <jra@samba.org>
Autobuild-Date: Sat Dec 18 08:59:27 CET 2010 on sn-devel-104
module, change the signature of VFS_REALPATH to always return a
malloc'ed string.
Needed to make some privileges work I plan on doing shortly
easier to code.
Jeremy.
Autobuild-User: Jeremy Allison <jra@samba.org>
Autobuild-Date: Sat Nov 20 02:15:50 CET 2010 on sn-devel-104
cross-device rename support has some major limitations:
- on huge files clients will timeout or hang
- ACLs and EA information is not retained
Usually a client will have to handle this. A Windows Server with a reparse
point will also just return NT_STATUS_NOT_SAME_DEVICE. We will now by default
do the same.
I will add a vfs module which will restore the old cross-device renames.
Ensure we don't use any of the create_options for Samba private
use. Add a new parameter to the VFS_CREATE call (private_flags)
which is only used internally. Renumber NTCREATEX_OPTIONS_PRIVATE_DENY_DOS
and NTCREATEX_OPTIONS_PRIVATE_DENY_FCB to match the S4 code).
Rev. the VFS interface to version 28.
Jeremy.
posix_fallocate is more efficient than manual zero'ing the file. When
preallocation in kernel space is supported it's extremely fast. Support for
preallocation at fs layer via posix_fallocate and fallocate at kernel site
can be found in Linux kernel 2.6.23/glibc 2.10 with ext4, XFS and OCFS2. Other
systems that I know of which support fast preallocation in kernel space are
AIX 6.1 with JFS2 and recent Solaris versions with ZFS maybe UFS2, too.
People who have a system with preallocation in kernel space might want to set
"strict allocate = yes". This reduces file fragentation and it's also safer for
setups with quota being turned on.
As of today most systems still don't have preallocation in kernel space, and
that's why "strict allocate = no" will stay the default for now.
