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.
The winbindd_dual_pam_auth_crap() function will be converted to a local
RPC call handler and it won't receive a winbindd_cli_state struct. Move
the checks accessing this struct to the parent.
Signed-off-by: Samuel Cabrero <scabrero@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Next commits will covert the winbindd_dual_pam_auth_crap() function to a
local RPC call handler receiving the challenge as a DATA_BLOB in the 'r'
struct.
Signed-off-by: Samuel Cabrero <scabrero@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Remove knownfail.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15042
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Autobuild-User(master): Ralph Böhme <slow@samba.org>
Autobuild-Date(master): Fri Apr 29 15:50:21 UTC 2022 on sn-devel-184
Passes against Windows. Shows that Windows allows a durable handle
on a leased open for READ_ATTRUBUTES only (a stat open).
Mark as knownfail for now.
NB. Not sure why we are testing smb2.durable-open against ad_dc
as that provisioning has "smb2 leases = no" which precludes
granting durable handles. Not changing for this bug but this
should be looked at in future.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15042
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
The 'git clone' used by autobuild.py fails if
GIT_DIR and GIT_WORK_TREE are already defined in the
environment.
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Autobuild-User(master): Stefan Metzmacher <metze@samba.org>
Autobuild-Date(master): Thu Apr 28 15:23:24 UTC 2022 on sn-devel-184
As part of the filename_convert() process, keep a pathref dirfsp of
the containing directory for later use. This avoids having to do
another non_widelink_open() on every SMB2_CREATE and ntcreate&x in
later patches.
Future work will be to go through other filename_convert() calls and
make them use filename_convert_dirfsp(). If we manage to convert all
of them except the one in filename_convert_dirfsp() itself, we can
simplify filename_convert() and unix_convert() significantly.
Too large a patch, but I don't know how to split this up into smaller
logic pieces.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
parent_pathref() is expensive, and we should avoid it if possible.
Not effective at this point, we always pass in NULL, but will be used
soon.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
This ensures we detect the runtime system and release, not the ones
when Samba was build. It's necessary to detect the correct kernel
version we're running on because for kernels before 5.3.1 O_PATH opens
unnecessarily broke kernel oplocks, which breaks our tests. And in
gitlab it can happen that we build on kernels after 5.3.1 and later
run on older kernels. In this situation we can't run kernel oplock
tests.
Signed-off-by: Ralph Boehme <slow@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>
gluster seems not to implement O_PATH, so it should be possible to do
a glfs_fgetxattr() on the pathref dirfsp.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Copy the logic from ceph_snap_gmt_get_real_filename(). This is
untested in autobuild, but as ceph is broken anyway due to
812cb602e3be, we need to talk to the ceph developers before 4.17.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Make get_real_filename() a wrapper.
Right now shadow_copy2 does a fallback to do get_real_filename() on
the twrp=0 tree in case of snapdirseverywhere because snapdirs can be
somewhere deep in the tree, and doing that correctly would be a
full-tree walk. I'd say that snapdirseverywhere is impossible to
implement if you want symlink safety, i.e. careful top-down tree
traversal together with snapdirseverywhere. If you have
snapdirseverywhere you need to pass down the full path very deep down,
which contradicts our fd-based approach we want to take.
Also, I believe that our test does not 100% correctly reflect what
actually is there: My understanding is that if you activate
snapdirseverywhere for example in GPFS, you see all snapshots at every
level (this would need to be verified). Our test does something more
nasty: It creates and tests a specific snapshot only at one place deep
in the directory hierarchy, which makes it impossible to find without
the full path.
This is all a big mess, but for now we need to deal with it. This adds
the twrp=0 fallback to core smbd, but I don't see any other way to do
that properly. And I do want a fd-based getrealfilename....
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
If we have a directory pathref fsp, do an openat(dirfd, ".", O_RDONLY)
to cheaply get a real directory handle.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
In a patchset that I'm working on right now there's the need to call
getrealfilename while the code does have a pathref fsp already
around. Doing the name-based call including non_widelink_open is not
necessary in this case. Start by adding the _at based call to the VFS.
For now, fall back to the name-based call. glusterfs-fuse will in a
future patch be converted to fgetxattr.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
If openat_pathref_fsp() fails, callers might want to inspect the stat
info. If we really failed on STOPPED_ON_SYMLINK, the caller might need
to know this, although openat_pathref_fsp() masked this error.
As there is no smb_fname->fsp returned from openat_pathref_fsp() on
error, we need to pass this up in smb_fname itself.
This essentially reverts de439cd03047, which does basically the same
thing but is too specific. We need to cover the general !O_PATH case
more broadly.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Setting gpfs:recalls=no should prevent data access to offline files.
Since Samba 4.14, the VFS openat function is also called with O_PATH to
get a reference to the path. These accesses should not be blocked,
otherwise this would prevent offline files from being included in
directory listings.
Fix this by skipping the check for pathref fds.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15055
Signed-off-by: Christof Schmitt <cs@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Autobuild-User(master): Ralph Böhme <slow@samba.org>
Autobuild-Date(master): Thu Apr 28 07:59:47 UTC 2022 on sn-devel-184
Pass down "fsp" and "info_level", no need to parse this inside
call_trans2qpipeinfo() when the caller also has to do it.
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): Tue Apr 26 22:38:59 UTC 2022 on sn-devel-184
Move includes.h for struct files_struct to fd_handle.c. Both
printing.c and smb1_utils.c depended on fd_handle.h to include the
prototypes. Do that explicitly in those files.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
