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.
Now that the vfs_private cache is never read in vfs_gpfs, there is
no need any more to write it.
With this change, vfs_gpfs does not use vfs_private any more.
Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Christof Schmitt <cs@samba.org>
This was used as a cache for offline-info in the stat buffer.
But as the implementation of gpfs_is_offline() showed, this cache
does not always carry valid information when the stat itself is valid
(since at least one call goes to fstatat() directly, circumventing
the vfs).
So the correct thing is to always call SMB_VFS_IS_OFFLINE()
when checking whether a file is offline. For the pread and pwrite
calls, we need to call IS_OFFLINE before the actual read
and check afterwards if the file was offline before (as a basis
whether to send notifications).
Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Christof Schmitt <cs@samba.org>
This reverts commit 8f44883db94314c007c197927a9dd0809076754d.
The next commits will be removing all access to stat_ex.vfs_private from
vfs_gpfs. This revert of the last addition is a preparation.
Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Christof Schmitt <cs@samba.org>
In is_offline(), check whether the winAttrs are filled with bits
outside 0xFFFF and log it prominently: Since GPFS only
fills 0xFFFF, this could be due to an uninitialized buffer
(or another vfs module filling vfs_private? ...).
Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Christof Schmitt <cs@samba.org>
There is a problem of flapping offline due to uninitialized
stat buffers. Due to a optimization in vfswrap_readdir which
directly calling fastatat (i.e. not through vfs), marking the
stat buffer valid, there is nothing this module can do about
it and hence can not currently not rely on the vaildity of
the stat buffer.
By always calling out to GPFS even when the stat buffer is
flagged valid, we can always return correct offline information,
thereby sacrificing the readdir optimization.
Pair-Programmed-With: Volker Lendecke <vl@samba.org>
Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Christof Schmitt <cs@samba.org>
This may (e.g.) have lead to some occurrences of flapping offline bits.
Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Christof Schmitt <cs@samba.org>
Signed-off-by: Ralph Wuerthner <ralph.wuerthner@de.ibm.com>
Reviewed-by: Christof Schmitt <cs@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Thu May 15 02:19:41 CEST 2014 on sn-devel-104
Remove an unused variable and use discard_const_p.
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 Apr 18 22:25:25 CEST 2014 on sn-devel-104
Use literals to allow a compile and execution on gpfs 3.4.
Signed-off-by: Alexander Werth <alexander.werth@de.ibm.com>
Reviewed-by: Christof Schmitt <cs@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
When this option is set to no, an attempt to open an offline file will
be rejected with access denied. This helps preventing recall storms
triggered by careless applications like Finder and Explorer.
Signed-off-by: Christian Ambach <ambi@samba.org>
Reviewed-by: Christof Schmitt <cs@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
The gpfs_set_times API call allows setting timestamps directly in GPFS
without going through the utime() call. Using this API call fixes an
unecessary oplock break when a client sends a SET_FILE_ALLOCATION_INFO
request and no other client has opened the file. The call to utime()
triggers the oplock break through the Linux kernel. Using the
gpfs_set_times call for updating the timestamp avoids the call to
utime() and the oplock break.
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): Thu Jan 9 00:04:48 CET 2014 on sn-devel-104
Ensures correct lease owner for signal delivery.
Signed-off-by: Ralph Wuerthner <ralphw@de.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): Thu Aug 1 03:57:11 CEST 2013 on sn-devel-104
DISCONNECT can be arbitrarily complex, TALLOC_FREE of a simple struct
is easier.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Christian Ambach <ambi@samba.org>
Autobuild-User(master): Volker Lendecke <vl@samba.org>
Autobuild-Date(master): Tue May 14 18:11:29 CEST 2013 on sn-devel-104
Since the smb4acl is now correctly allocated on mem_ctx and not
the talloc stack frame we can free the stack frame correctly.
And the chmod emulation code now needs the vfs handle since
that is now required by the callback function to set the smb4acl.
Reviewed-By: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
This ensures the caller knows exactly what the memory lifetime of this
returned object is. This makes the NFSv4 ACL code consistent with the
POSIX and NT ACL code, to avoid supprising developers who have worked
on those other parts of the ACL code.
Most of this patch is adding a memory context to the callers and passing it in.
Andrew Bartlett
Reviewed-by: Jeremy Allison <jra@samba.org>
last caller has gone
Signed-off-by: Christian Ambach <ambi@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Autobuild-User(master): Andrew Bartlett <abartlet@samba.org>
Autobuild-Date(master): Mon Feb 4 14:10:08 CET 2013 on sn-devel-104
use the helper functions to return the blob based on the
raw GPFS ACL blob (if it is a NFSv4 ACL). If not, fall back
to the POSIX ACL code
Signed-off-by: Christian Ambach <ambi@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
in contrast to gpfs_getacl_alloc which always puts the
ACL on talloc_tos(), this one allows to specify the memory
context and if the caller is interested in the raw ACL blob
or a structured version
Signed-off-by: Christian Ambach <ambi@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Michael Adam <obnox@samba.org>
Autobuild-User(master): Michael Adam <obnox@samba.org>
Autobuild-Date(master): Wed Nov 28 14:06:27 CET 2012 on sn-devel-104
sys_acl_init returns a SMB_ACL_T with zero entries in the acl array
reallocate the array to proper size before filling it, otherwise we overwrite memory
This one is a result of a improper fixing in 7a61829629
Signed-off-by: Christian Ambach <ambi@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
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
The query of the fileset quota needs to determine the file set id first.
With the currently available interface, this requires opening the file
to get a file descriptor. For files, this open can fail when a share
mode is set.
Workaround this by querying the fileset id on the directory instead.
The proper solution would be getting an interface for getting the
fileset id that does not require opening the file.
Autobuild-User(master): Christian Ambach <ambi@samba.org>
Autobuild-Date(master): Wed Aug 29 18:58:34 CEST 2012 on sn-devel-104
It was separated out because formerly our async I/O was not properly
stackable. aio_fork could for example catch aio and not get aio_return
get through to vfs_gpfs
Signed-off-by: Christian Ambach <ambi@samba.org>
The acl element is changed to be a talloc child, and is no longer one element
longer than requested by virtue of the acl[1] base pointer.
This also avoids one of the few remaining cases of over-allocation of a structure.
Andrew Bartlett
GPFS 3.5 introduces ACL enhancements which are breaking our ACL length
calculations.
Signed-off-by: Ralph Wuerthner <ralph.wuerthner@de.ibm.com>
Autobuild-User(master): Christian Ambach <ambi@samba.org>
Autobuild-Date(master): Wed Jul 11 21:28:23 CEST 2012 on sn-devel-104
gpfs_quotactl can return a non-zero softquota gracetime even when no
softquota has been set. This could lead to "disk full" being reported to
a client. The easiest fix is to check for a valid softquota before
checking the softquota gracetime.
gpfs2smb_acl can leave errno!=0 around even if it returned a correct
result!=NULL. We can only rely on errno being set if another error
condition (in this case result==NULL) indicates an error. If
result!=NULL, errno is undefined and can be anything. This leads to
SAFE_FREE(result) further down even in the success case.
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Autobuild-User(master): Stefan Metzmacher <metze@samba.org>
Autobuild-Date(master): Fri Jun 22 19:27:39 CEST 2012 on sn-devel-104
With "gpfs:acl=no" you can pass the acl calls to the next SMB_VFS module.
Based on a patch from Hans-Dieter Schuster <hans-dieter.schuster@ts.fujitsu.com>
Pair-Programmed-With: Stefan Metzmacher <metze@samba.org>
Don't use the mode for the get_acl call that surpresses
inherited deny ACE's. This is now possible since
the inherited ACE flag exists now in GPFS and Samba.
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
if setting the lease in GPFS failed, there is not much sense in trying
to set a lease just locally that would not inform us of openers
on other cluster nodes
GPFS provides the gpfs_prealloc call. Implement the fallocate
callback with mode VFS_FALLOCATE_EXTEND_SIZE using this call.
There is no support for VFS_FALLOCATE_KEEP_SIZE, so report
this mode as not available.
When a client requests the information about free space and space used,
adjust the reported values according to quotas in the GPFS file system:
- Retrieve quotas for the current user, current group and fileset for
the top level of the share.
- If the soft block quota grace time has expired, report disk as full.
- If a hard block quota has been exceeded, report disk as full.
- If none of the hard block quotas been exceeded, report
share size and free space according to the lowest limits found in
the quotas.
- If no applicable hard block quota has been set, report the
information from the statfs call.
This feature is disabled by default and has to be enabled by setting the
option gpfs:dfreequota.
Setting the creation time through SetFileTime on a GPFS file system and
querying it with GetFileTime shows a mismatch.
The vfs_gpfs module first retrieves the information from the operating
system and the flag st_ex_calculated_birthtime is set to false. When
vfs_gpfs retrieves the birthtime from GPFS the flag
st_ex_calculated_birthtime has to be set to true. Otherwise the birth
time will get overwritten by a call to update_stat_ex_mtime, reporting
the wrong time to a client system.
Signed-off-by: Christian Ambach <ambi@samba.org>
With newer GPFS releases, libgpfs contains a function
gpfs_lib_init(). It marks the pseudo-device the lib uses
as interface to the GPFS kernel module to not be to
destroyed after each call to GPFS.
This saves us some microseconds creating and removing the device
each time we do a call to libgpfs.
Autobuild-User: Volker Lendecke <vlendec@samba.org>
Autobuild-Date: Tue Apr 26 20:52:36 CEST 2011 on sn-devel-104
On my autocluster setup, it's not set. Maybe it should be? Otherwise
smbclient and some Windows client programs will get errors like:
# smbclient //localhost/data -Uadministrator%XXX
Domain=[VSOFS1] OS=[Unix] Server=[Samba 3.4.2-ctdb-10]
smb: \> put /etc/resolv.conf resolv.conf
NT_STATUS_ACCESS_DENIED closing remote file \resolv.conf
smb: \>
Caused by attempting to update the time on close.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Michael Adam <obnox@samba.org>
ported from the v3-4-ctdb branch to master
This used to be commit 1f138cc9f4a
Autobuild-User: Volker Lendecke <vlendec@samba.org>
Autobuild-Date: Wed Oct 13 13:34:25 UTC 2010 on sn-devel-104
This adds a parameter "gpfs:refuse_dacl_protected" that defaults to false.
GPFS has no place to store the SEC_DESC_DACL_PROTECTED ACL bit. With this
parameter we give customers an option to either ignore this bit or refuse
setting an ACL with it.
as GPFS does not support the ACE4_FLAG_NO_PROPAGATE NFSv4 flag (which would be the mapping for the DESC_DACL_PROTECTED flag), the status of this flag is currently silently ignored by Samba. That means that if you deselect the "Allow inheritable permissions..." checkbox in Windows' ACL dialog and then apply the ACL, the flag will be back immediately.
To make sure that automatic migration with e.g. robocopy does not lead to ACLs silently (and unintentionally) changed, this patch adds an explicit check for this flag and if set, it will return NT_STATUS_NOT_SUPPORTED so errors are shown up on the Windows side and the Administrator is aware of the ACLs not being settable like intended
Signed-off-by: Christian Ambach <christian.ambach@de.ibm.com>
This patch introduces two new temporary helper functions
vfs_stat_smb_fname and vfs_lstat_smb_fname. They basically allowed me
to call the new smb_filename version of stat, while avoiding plumbing
it through callers that are still too inconvenient. As the conversion
moves along, I will be able to remove callers of this, with the goal
being to remove all callers.
There was also a bug in create_synthetic_smb_fname_split (also a
temporary utility function) that caused it to incorrectly handle
filenames with ':'s in them when in posix mode. This is now fixed.
This patch introduces
struct stat_ex {
dev_t st_ex_dev;
ino_t st_ex_ino;
mode_t st_ex_mode;
nlink_t st_ex_nlink;
uid_t st_ex_uid;
gid_t st_ex_gid;
dev_t st_ex_rdev;
off_t st_ex_size;
struct timespec st_ex_atime;
struct timespec st_ex_mtime;
struct timespec st_ex_ctime;
struct timespec st_ex_btime; /* birthtime */
blksize_t st_ex_blksize;
blkcnt_t st_ex_blocks;
};
typedef struct stat_ex SMB_STRUCT_STAT;
It is really large because due to the friendly libc headers playing macro
tricks with fields like st_ino, so I renamed them to st_ex_xxx.
Why this change? To support birthtime, we already have quite a few #ifdef's at
places where it does not really belong. With a stat struct that we control, we
can consolidate the nanosecond timestamps and the birthtime deep in the VFS
stat calls.
At this moment it is triggered by a request to support the birthtime field for
GPFS. GPFS does not extend the system level struct stat, but instead has a
separate call that gets us the additional information beyond posix. Without
being able to do that within the VFS stat calls, that support would have to be
scattered around the main smbd code.
It will very likely break all the onefs modules, but I think the changes will
be reasonably easy to do.
this can only be done via fset_nt_acl() using an open
file/directory handle. I'd like to do the same with
get_nt_acl() but am concerned about efficiency
problems with "hide unreadable/hide unwritable" when
doing a directory listing (this would mean opening
every file in the dir on list).
Moving closer to rationalizing the ACL model and
maybe moving the POSIX calls into a posix_acl VFS
module rather than having them as first class citizens
of the VFS.
Jeremy.
(This used to be commit f487f742cb)
Up to now, get_nt_acl() took a files_struct pointer (fsp) and
a file name. All the underlying functions should need and now
do need (after the previous preparatory work), is a connection_struct
and a file name. The connection_struct is already there in the
vfs_handle passed to the vfs functions. So the files_struct
argument can be eliminated.
This eliminates the need of calling open_file_stat in a couple
of places to produce the fsp needed.
Michael
(This used to be commit b5f600fab5)
This moves functionality from gpfsacl_get_nt_acl_common()
back to gpfsacl_get_nt_acl() and gpfsacl_fget_nt_acl(),
making both these functions more specific (calling the
corresponding fsp- and non-fsp functions).
gpfsacl_get_nt_acl_common(). is removed.
Michael
(This used to be commit d6043c1066)
Replace smbd/posix_acls.c:get_nt_acl() by two funcions:
posix_get_nt_acl() and posix_fget_nt_acl(). The first
takes a connection struct and a file name instead of a
files_struct pointer. This is in preparation of changing
the vfs api for SMB_VFS_GET_NT_ACL.
Michael
(This used to be commit 50c82cc145)
bugs in various places whilst doing this (places that assumed
BOOL == int). I also need to fix the Samba4 pidl generation
(next checkin).
Jeremy.
(This used to be commit f35a266b3c)
This adds the two functions talloc_stackframe() and talloc_tos().
* When a new talloc stackframe is allocated with talloc_stackframe(), then
* the TALLOC_CTX returned with talloc_tos() is reset to that new
* frame. Whenever that stack frame is TALLOC_FREE()'ed, then the reverse
* happens: The previous talloc_tos() is restored.
*
* This API is designed to be robust in the sense that if someone forgets to
* TALLOC_FREE() a stackframe, then the next outer one correctly cleans up and
* resets the talloc_tos().
The original motivation for this patch was to get rid of the
sid_string_static & friends buffers. Explicitly passing talloc context
everywhere clutters code too much for my taste, so an implicit
talloc_tos() is introduced here. Many of these static buffers are
replaced by a single static pointer.
The intended use would thus be that low-level functions can rather
freely push stuff to talloc_tos, the upper layers clean up by freeing
the stackframe. The more of these stackframes are used and correctly
freed the more exact the memory cleanup happens.
This patch removes the main_loop_talloc_ctx, tmp_talloc_ctx and
lp_talloc_ctx (did I forget any?)
So, never do a
tmp_ctx = talloc_init("foo");
anymore, instead, use
tmp_ctx = talloc_stackframe()
:-)
Volker
(This used to be commit 6585ea2cb7)
return to correctly return NT_STATUS_INVALID_OWNER if it
should be disallowed. Matches better what W2K3R3 does.
NFSv4 ACL module owners, please examine these changes.
Jeremy.
(This used to be commit fc6899a550)
it should be abstracted a little higher up so other os'es can have an
entry, but it will take a bit more work. Thanks to Chetan Shringarpure
and Mathias Dietz.
I didn't increment the vfs number again because the kernel change notify
stuff hasn't been released yet anyway.
(This used to be commit 9463211bf3)
