| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
torture test
smbtorture4's BASE-DELETE:deltest17 was failing against win2k8,
win2k3, and winXPsp2 but passing against samba.
deltest17 does the following:
1. open file -> file is created
2. closes file
3. open file with DOC -> fnum1
4. check that DOC is not reported as being set from fnum1
5. opens file again Read Only -> fnum2
6. check that DOC is not reported as being set from either file handle
7. close fnum1 (the file handle that requested DOC to be set)
8. check if DOC is reported as being set from fnum2
* This is where windows and samba begin to diverge. Windows
reports that the DOC bit is set, while samba reports that it is not set.
9. close fnum2 (the last remaining open handle for the file)
10.See if the file has been deleted.
* On samba the file still exists. On windows the file was deleted.
The way open_file_ntcreate is written now, if an open has the DOC bit
set on the wire, DOC (fsp->initial_delete_on_close) is not set unless:
a. the open creates the file, or b. there is an open file handle with
a share_entry in the struct lck that has the
SHARE_MODE_ALLOW_INITIAL_DELETE_ON_CLOSE bit set (let's call it
SM_AIDOC).
My understanding of SM_AIDOC is that it was added to differentiate
between DOC being set on an open that creates a file vs an open that
opens an existing. As described in step 8/10 above, it appears that
windows does not make this differentiation.
To resolve this issue there are three patches. This first patch is a
simple proof of concept change that is sufficient to fix the bug. It
removes the differentiation in open_file_ntcreate, and updates
deltest17 to allow it to pass against win2k3/xp. This makes
open_file_ntcreate more closely match the semantics in open_directory
and rename_internals_fsp. This change also does not break any other
tests in BASE-DELETE or "make test". Specifically test deltest20b
which verifies the CIFSFS rename DOC semantics still passes :).
|
|
This is the first pass at extending the onefs vfs module to support
the CIFS-specific enhancements available on OneFS. Most of this patch
is massaging the sama open path to work with ifs_createfile.
ifs_createfile is a CIFS-specific syscall for opening/files and
directories. It adds support for:
- Full in-kernel access checks using a windows access_mask
- Cluster-coherent share mode locks
- Cluster-coherent oplocks
- Streams
- Setting security descriptors at create time
- Setting dos_attributes at create time
This patch does not implement the samba side of the streams support or
oplocks support. Tests that expect oplocks to be granted or streams
to be supported will fail. This will be remedied in upcoming patches.
|
|
This is an intermediate step that makes it much easier to see how the
OneFS SMB_VFS_CREATE_FILE implementation diverges from stock samba.
The goal is that more common code can be refactored into utility
functions.
|
|
Some implementations of SMB_VFS_CREATE_FILE implementations actually
keep an fd open for directories just as files. In this case it is
necessary to call fd_close when closing directories. This is safe
because fd_close is a no-op when fd == -1, which is true for directory
opens originating from open.c (the default SMB_VFS_CREATE_FILE
implementation).
|
|
|
|
|
|
This allows vfs modules that implement SMB_VFS_CREATE_FILE to access
some of the useful utility functions.
|
|
|
|
Guenther
(cherry picked from commit 26139344fd0fac4fdd2a6752628b252fbd9b7450)
(cherry picked from commit 866efa63a26f75bbf17cd4bebf639594e2feafba)
|
|
Guenther
|
|
Guenther
|
|
Guenther
|
|
|
|
Jeremy.
|
|
|
|
|
|
Introduce async_req_is_error() and async_req_simple_recv()
|
|
|
|
For example open&x and write&x needs the bytes to be aligned relative to the
SMB header. In particular for write&x we should not have to move stuff around.
|
|
|
|
|
|
|
|
Remove three pointless variables
|
|
|
|
|
|
Jeremy.
|
|
|
|
non-encrypted packet with the crypto state set.
|
|
Guenther
|
|
smbfs also would use the USER environment variable if the user was not
specified on the mount command and no credential file specified, and
mount.cifs man page says that we will use this environment variable
(in most cases this will not cause a behavior change, because
we were doing getuid of the current process which will usually be
the same name).
Fixes Samba bug #5934
Acked-by: Jeff Layton
|
|
Guenther
|
|
Guenther
|
|
Guenther
|
|
Guenther
|
|
Guenther
|
|
Guenther
|
|
Guenther
|
|
Guenther
|
|
Jeremy.
|
|
--with-cluster-support and GPFS
|
|
A somewhat more elegant fix than I could use for 3.2.x or 3.0.x.
Turns out the only part of check_user_ok() that needs to change
for share level security is the VUID cache pieces, so I can just
always use check_user_ok() for all lp_security() cases.
Jeremy
|
|
getgrsid_sid2gid_recv
to make code more readble
Michael
|
|
This is just to find the corresponding domain struct.
Actual connection is handled by the domain child.
Michael
|
|
This is just to find the corresponding domain struct.
Actual connection is handled by the domain child.
Michael
|
|
metze
|
|
|
|
This replaces the is_dos_path bool with a more future-proof argument.
The next step is to plumb INTERNAL_OPEN_ONLY through this flag instead
of overridding the oplock_request.
|
|
Also removed open_file_ntcreate/open_file from proto.h so they can no
longer be called directly.
|
|
|
|
SMB_VFS_CREATE_FILE
|
