| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
In the refactoring around filename_convert, the split between the functions
resolve_dfspath() and resolve_dfspath_wcard() was lost, leaving us only with
resolve_dfspath_wcard().
Internally resolve_dfspath_wcard() calls dfs_redirect() only with a
"allow_wcards" flag of true, wheras the old resolve_dfspath() would call with a
value of false. The loss of this case causes dfs_redirect to always masquerade
DFS links as directories, even when they are being queried directly by a trans2
QPATHINFO call. We should only masquerade DFS links as directories when called
from a SMBsearch or trans2 findfirst/findnext - which was the intent of the
"allow_wcards" flag.
This patch adds back an allow_wcards bool parameter to
resolve_dfspath_wcard(). This bool is set from the state of the ucf_flags when
filename_convert() is called.
I will follow this up with a new smbclient-based torture test that will prevent
us from ever regressing our DFS support again.
Jeremy.
|
|
Volker.
Create widelinks_warning(int snum) to cover the message needed in make_connection.
Jeremy.
|
|
This way we avoid any chance that a configuration reload may turn
back on wide links when unix extensions are enabled.
|
|
I know Volker will look at this closely so here's the explaination :-).
Originally on a case-sensitive share we simply did a stat (or lstat)
call and returned success of fail based on the result. However this
failed to take account of incoming paths with a wildcard (which must
always fail, and with different error messages depending on whether
the wildcard is the last component or in the path). Also it failed
to take account of a stat fail with ENOENT due to a missing component
of the path as the last component (which is ok as it could be a new
file) or if the ENOENT was due to the missing component within
the path (not the last component) - which must return the correct
error. What this means is that with "case sensitive = yes" we do
one more talloc call (to get the parent directory) and one more
stat call (on the parent directory) in the case where the stat
call fails. I think this is an acceptable overhead to enable
case sensitive shares to return the correct error messages for
applications. Volker please examine carefully :-).
Jeremy.
|
|
RAW-CHKPATH when case sensitive = yes, but isn't the correct way to do
it. I'm testing a larger patch to smbd/filename.c that should fix this
correctly, and will add a torture test to ensure RAW-CHKPATH is run against
a case sensitive share once this is done.
Jeremy.
|
|
wide links = no.
Jeremy.
|
|
out by Volker.Restores the pathname handling for FILE_FLAG_POSIX_SEMANTICS but still prevents the O_APPEND problems. Jeremy.
|
|
This reverts commit f7b4151a64d8c6851e62255a7139fd00a5fc63a3.
|
|
This reverts commit c85a4c9ba4a7de65a7850f6f9708df66bd24deea.
|
|
|
|
|
|
This is a hot codepath (called from the stat cache)
|
|
|
|
This vop is designed to work in tandem with SMB_VFS_READDIR to allow
vfs modules to make modifications to arbitrary filenames before
they're consumed by callers. Subsequently the core directory
enumeration code in smbd is now changed to free the memory that may be
allocated in a module. This vop enables the new version of catia in
the following patch.
Signed-off-by: Tim Prouty <tprouty@samba.org>
|
|
metze
|
|
This patch also changes the unix convert flags to make sure the
correct semantics are preservered for allowing/disallowing wildcards
in the last component of the path.
|
|
caller know if the path has a wildcard
This also eliminates the need for resolve_dfspath().
|
|
This allows SMB_VFS_[L]STAT to be called directly. Additionally, I
changed NTSTATUS result to be named status for consistency.
I also removed the stat_cache_add() from build_stream_path() because
stat_cache_lookup() is never actually called on a file with a stream.
There is no reason why the stat cache couldn't be consulted for
streams in the future.
Jeremy/Volker, please take a look at this one when you get a chance.
|
|
|
|
This is to ease the linking pain of everything that links LOCKING_OBJ
|
|
|
|
The goal is to eventually remove the need for the char **fname argument
once all callers have standardized on smb_filename.
|
|
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.
|
|
In copy_internals(), if the !CAN_WRITE(conn) kicks in, we end up
dereferencing a NULL smb_filename.
This adds a simple protection around it.
Tim, please check!
Volker
|
|
resolve_dfspath() -> unix_convert() -> get_full_smb_filename() -> check_name()
with a new function filename_convert().
This restores the check_name() calls that had gone missing
since the default create_file was changed. All "standard"
pathname processing now goes through filename_convert().
I'll take a look at the non-standard pathname processing
next. As a benefit, fixed a missing resolve_dfspath()
in the trans2 mkdir call.
Jeremy.
|
|
This was a little messy because of all of the vfs modules I had to
touch. Most of them were pretty straight forward, but the streams
modules required a little attention to handle smb_filename. Since the
use of smb_filename enables the vfs modules to access the raw,
over-the-wire stream, a little bit of the handling that was being done
by split_ntfs_stream_name has now been shifted into the individual
stream modules. It may be a little more code, but overall it gives
more flexibility to the streams modules, while also allowing correct
stream handling.
|
|
|
|
|
|
|
|
|
|
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 is the first of a series of patches that change path based
operations to operate on a struct smb_filename instead of a char *.
This same concept already exists in source4.
My goals for this series of patches are to eventually:
1) Solve the stream vs. posix filename that contains a colon ambiguity
that currently exists.
2) Make unix_convert the only function that parses the stream name.
3) Clean up the unix_convert API.
4) Change all path based vfs operation to take a struct smb_filename.
5) Make is_ntfs_stream_name() a constant operation that can simply
check the state of struct smb_filename rather than re-parse the
filename.
6) Eliminate the need for split_ntfs_stream_name() to exist.
My strategy is to start from the inside at unix_convert() and work my
way out through the vfs layer, call by call. This first patch does
just that, by changing unix_convert and all of its callers to operate
on struct smb_filename. Since this is such a large change, I plan on
pushing the patches in phases, where each phase keeps full
compatibility and passes make test.
The API of unix_convert has been simplified from:
NTSTATUS unix_convert(TALLOC_CTX *ctx,
connection_struct *conn,
const char *orig_path,
bool allow_wcard_last_component,
char **pp_conv_path,
char **pp_saved_last_component,
SMB_STRUCT_STAT *pst)
to:
NTSTATUS unix_convert(TALLOC_CTX *ctx,
connection_struct *conn,
const char *orig_path,
struct smb_filename *smb_fname,
uint32_t ucf_flags)
Currently the smb_filename struct looks like:
struct smb_filename {
char *base_name;
char *stream_name;
char *original_lcomp;
SMB_STRUCT_STAT st;
};
One key point here is the decision to break up the base_name and
stream_name. I have introduced a helper function called
get_full_smb_filename() that takes an smb_filename struct and
allocates the full_name. I changed the callers of unix_convert() to
subsequently call get_full_smb_filename() for the time being, but I
plan to eventually eliminate get_full_smb_filename().
|
|
The GPFS get_real_file name does not know about mangled names. Tim, if onefs
does not either, you need this bugfix :-)
In case onefs does 8.3 names, we need to pass the mangled flag down to
SMB_VFS_GET_REAL_FILENAME to give GPFS a chance say ENOTSUPP and do the
fallback.
|
|
Often times before creating a file, a client will first query to see
if it already exists. Since some systems have a case-insensitive stat
that is called from unix_convert, we can definitively return
STATUS_NO_SUCH_FILE to the client without scanning the whole
directory.
This code path is taken from trans2querypathinfo, but trans2findfirst
still does a full directory scan even though the get_real_filename
(the case-insensitive stat vfs call) can prevent this.
This patch adds the get_real_filename call to the trans2find* path,
and also changes the vfs_default behavior for
SMB_VFS_GET_REAL_FILENAME. Previously, in the absence of a
get_real_filename implementation, we would fallback to the full
directory scan. The default behavior now returns -1 and sets errno to
EOPNOTSUPP. This allows SMB_VFS_GET_REALFILENAME to be called from
trans2* and unix_convert.
|
|
This reverts commit b5b6ecb58aeba196c28370cde21a0826f81d8da6.
|
|
|
|
* VFS_OP_READDIR can now provide stat information, take advantage of it
if it's available
* is_visible_file(): optimistically expect the provided stat buffer is
already valid
* dptr_ReadDirName(): refactor code for easier readability, functionality
is the same
|
|
failing.
Reported by Kukks. Make sure we correctly use LSTAT in all cases where
POSIX pathnames are being used. This matters when dealing with symlinks
pointing to invalid paths being renamed or deleted not all deletes and
renames are done via an nt_create open.
Jeremy.
|
|
This is done to help file systems that can tell us about the real upper/lower
case combination given a case-insensitive file name. The sample I will soon
push is the gpfs module (recent gpfs has a get_real_filename function), others
might have a similar function to help alleviate the 1million files in a single
directory problem.
Jeremy, please comment!
Thanks,
Volker
|
|
It sets errno, so it might as well return 0/-1.
|
|
|
|
|
|
optimization". Makes the change clearer.
Jeremy.
|
|
With
case sensitive = yes
preserve case = no
short preserve case = no
default case = upper
a "dir FOO.txt" would not find "FOO.TXT" because FOO.txt ends up unconverted in
the mask for mask_match.
Jeremy, please check!
Volker
|
|
That was an old and subtle bug.
Jeremy.
|
|
map_nt_error_from_unix() now assumes that it is called in
an error path and returns an error even for a given errno == 0.
The original behaviour of unix_convert() used the mapping
of errno == 0 ==> NT_STATUS_OK to return success through
an error path.
I think this must have been an oversight, and unix_convert() worked
only by coincidence (or because explicitly using the knowledge
of the conceptually wrong working of map_nt_error_from_unix().
This patch puts this straight by not interpreting errno == 0
as an error condition and proceeding in that case.
Jeremy - please check!
Michael
(This used to be commit ec5956ab0df1b3f567470b2481b73da9c3c67371)
|
|
get_file_infos()
This means we need to fetch the record only once.
metze
(This used to be commit 4130b873291d39e363184fe4e38dc1f24ebe5056)
|
|
new stream code. (1) In smbd/filename, don't split the name at
':' if we know it's a posix path (this should be parameterized....).
(2). When calling posix_mkdir, we get the flag FILE_FLAG_POSIX_SEMANTICS
passed to open_directory(). I know for a posix client lp_posix_pathnames
should be true (which is checked for in is_ntfs_stream_name() but we
have an explicit flag here, so let's use it.
Jeremy.
(This used to be commit 7bb7a0def6518784befa75e5303289d2b4d36dd4)
|
|
This is the core of the streams support. The main change is that in
files_struct there is now a base_fsp pointer that holds the main file open
while a stream is open. This is necessary to get the rather strange delete
semantics right: You can't delete the main file while a stream is open without
FILE_SHARE_DELETE, and while a stream is open a successful unlink of the main
file leads to DELETE_PENDING for all further access on the main file or any
stream.
(This used to be commit 6022873cc155bdbbd3fb620689715f07a24d6ed1)
|
|
This cuts some mallocs on NtCreate&X
(This used to be commit 8e64107b7846d8f9cce71aabc95b28b7488d01ce)
|
