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.
stat_cache_lookup() allocates its result on top of talloc_tos().
filename_convert_smb1_search_path() creates a talloc_stackframe(),
which makes the names which were supposed to be allocated on the "ctx"
parameter of filename_convert_smb1_search_path() go away too
early. Reparent the results from stat_cache_lookup() properly.
Bug: https://bugzilla.samba.org/show_bug.cgi?id=14989
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 Mar 1 20:59:55 UTC 2022 on sn-devel-184
(cherry picked from commit 8c97743511e4d53f795f2469a28aabfb96da0dfa)
This fixes a use-after-free in smb_full_audit_create_file() when
calling SMB_VFS_CREATE_FILE with fsp->fsp_name as smb_fname.
create_file_unixpath() has this comment:
* This is really subtle. If someone passes in an smb_fname
* where smb_fname actually is taken from fsp->fsp_name, then
* the lifetime of these objects is meant to be the same.
so it seems legitimate to call CREATE_FILE this way.
When CREATE_FILE runs into an error, create_file_unixpath() does a
file_free, which also takes fsp->fsp_name with
it. smb_full_audit_create_file() wants to log the failure including
the smb_fname after NEXT_CREATE_FILE has exited, but this will then
use the already free'ed data.
Fix by only doing the file_free() on an fsp that
create_file_unixpath() created itself.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14975
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 Feb 10 19:11:33 UTC 2022 on sn-devel-184
(cherry picked from commit 434e6d4b4b45757878642d229d26d146792a3878)
Autobuild-User(v4-16-test): Jule Anger <janger@samba.org>
Autobuild-Date(v4-16-test): Mon Feb 14 18:36:26 UTC 2022 on sn-devel-184
This does almost everything that close_file_free() does, but it leaves
the fsp around.
A normal close_file() now calls fsp_unbind_smb() twice. Functionally
this is not a problem, fsp_unbind_smb() is idempotent. The only
potential performance penalty might come from the loops in
remove_smb2_chained_fsp(), but those only are potentially large with
deeply queued smb2 requests. If that turns out to be a problem, we'll
cope with it later. The alternative would be to split up file_free()
into even more routines and make it more difficult to figure out which
of the "rundown/unbind/free" routines to call in any particular
situation.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14975
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
(cherry picked from commit e91b59c4dfb2b35661dbecbc5769584109e23571)
For example, remove our entry from smbXsrv_open_global.tdb
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14975
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
(cherry picked from commit e751c6237b750adb4cb59df4a42bb9f39354e7e4)
We are no longer called on base_fsp's in SHUTDOWN_CLOSE. That
simplifies the logic in the common case, we now have a linear flow for
the very often-called close_file()
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14975
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
(cherry picked from commit 93fe9c83145d31ea11a9cd25049ac527ad4a000d)
Same logic as the change for file_close_conn()
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14975
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
(cherry picked from commit 1fbd9877fead466a17d697c143cd370c0b27f610)
To be reused in file_close_user(). Deliberately a separate commit to
make the previous commit easier to understand.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14975
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
(cherry picked from commit 61f57ba24ee2e54abf224118f93bd0ccda44ec41)
close_file_free() needs to handle base fsps specially. This can be
simplified a lot if we pass the the open files a second time in case
we encountered base_fsps that we could not immediately delete.
file_close_conn() is not our hot code path, and also we don't expect
many thousand open files that we need to walk a second time.
A subsequent patch will simplify close_file_free(), the complicated
logic is now in files.c, where it IMHO belongs because
file_set_base_fsp() are here as well.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14975
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
(cherry picked from commit d1341d666af12965b4318f89b1d0e1e8769e861e)
Quite a few places already had this in the caller, but not all. Rename
close_file() to close_file_free() appropriately. We'll factor out
close_file_smb() doing only parts of close_file_free() later.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14975
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
(cherry picked from commit f5bc73a2ad97647f76143f7962c964f45aa6b1a0)
Centralize calling file_free(), but leave close_fake_file() in for API
symmetry reasons.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14975
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
(cherry picked from commit 244c5a7d31c3a37082b320680f2b71108d77bbd4)
Call file_free() just once
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14975
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
(cherry picked from commit 2293ca5b572178404273856f8d8989a5ee7de80c)
Call file_free() just once
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14975
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
(cherry picked from commit 9966b5e233ef2ff0368ba5860c824c7cd6420415)
Avoid the "needs_fsp_unlink" variable, describe the talloc hierarchy a
bit differently in the comments.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14975
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
(cherry picked from commit 1c1734974fcf1d060bc6bcdbe1858cba1b7e5a73)
We need to take SMB1+POSIX into account here and do an LSTAT if it's
a POSIX name.
Remove knownfail.d/posix_sylink_rename
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14911
Signed-off-by: Jeremy Allison <jra@samba.org>
If filename_convert() fails to convert the path, we never call
check_name(). This means we can return an incorrect error code
(NT_STATUS_ACCESS_DENIED) if we ran into a symlink that points
outside the share to a non-readable directory. We need to make
sure in this case we always call check_name().
Remove knownfail.d/symlink_traversal.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14911
Signed-off-by: Jeremy Allison <jra@samba.org>
NT_STATUS_OBJECT_PATH_NOT_FOUND for a path component failure.
NT_STATUS_OBJECT_NAME_NOT_FOUND for a terminal component failure.
Remove:
samba3.blackbox.test_symlink_traversal.SMB1.posix
samba3.blackbox.smbclient_s3.*.Ensure\ widelinks\ are\ restricted\(.*\)
samba3.blackbox.smbclient_s3.*.follow\ symlinks\ \=\ no\(.*\)
in knownfail.d/symlink_traversal as we now pass these. Only one more fix
remaining to get rid of knownfail.d/symlink_traversal completely.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14911
Signed-off-by: Jeremy Allison <jra@samba.org>
Matches the error return from openat_pathref_fsp().
NT_STATUS_OBJECT_PATH_NOT_FOUND is for a bad component in a path, not
a bad terminal symlink.
Remove knownfail.d/simple_posix_open, we now pass.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14911
Signed-off-by: Jeremy Allison <jra@samba.org>
This is an internal xattr that should not be user visible.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14914
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
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 Jan 18 21:17:43 UTC 2022 on sn-devel-184
The smb.conf is parsed in post mode of a popt callback. The smbd
--build-options parameter should be handled when first encountered
to avoid requiring smb.conf presence.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14945
Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: David Disseldorp <ddiss@samba.org>
If delete_all_streams() fails.
Found by Andrew Walker <awalker@ixsystems.com>
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14944
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Christof Schmitt <cs@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Fri Jan 14 03:34:47 UTC 2022 on sn-devel-184
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14942
Remove fsp which is always NULL and replace it with smb_fname->fsp.
Found by covscan.
Signed-off-by: Pavel Filipenský <pfilipen@redhat.com>
Reviewed-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
This first gets the clock_gettime_mono() value, converts to an NTTIME (as
this is what is stored in the dos attribute EA), then mixes in 8 bits of
randomness shifted up by 55 bits to cope with poor resolution clocks to
avoid duplicate inodes.
Using 8 bits of randomness on top of an NTTIME gives us around 114
years headroom. We can now guarentee returning a itime-based
fileid in a normal share (storing dos attributes in an EA).
Remove knownfail.d/fileid-unique
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14928
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Christof Schmitt <cs@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Sat Jan 8 06:35:22 UTC 2022 on sn-devel-184
We already stored this in struct smbd_lock_element.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Noel Power <npower@samba.org>
Autobuild-User(master): Noel Power <npower@samba.org>
Autobuild-Date(master): Thu Jan 6 16:03:28 UTC 2022 on sn-devel-184
We lookup the lock array from the state stored in the passed-in req
and all the locks in an array are always the same flavour, so this
isn't needed.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Noel Power <npower@samba.org>
And also inside struct smbd_smb1_do_locks_state.
All calls to this always (a) have one or more locks of the same type.
(the setup for smbd_smb1_do_locks_send() ensures there is always
at least one lock) and (b) always set locks[0].lock_flav correctly before calling.
lock_flav is thus a redundent argument. Removing it means
we can never drift out of sync with the lock_flav element
in the passed in locks array.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Noel Power <npower@samba.org>
When we remove the lock_flav parameter this will need to look into the array itself.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Noel Power <npower@samba.org>
This is now contained in the struct smbd_lock_element for
each lock.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Noel Power <npower@samba.org>
Make it explicit. When we add POSIX handles to SMB2 we will only
look at the handle type. lp_posix_cifsu_locktype() already does this,
but hidden inside init_strict_lock_struct() makes it hard to see.
No logic change.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Noel Power <npower@samba.org>
Initialized correctly but not yet used.
Will allow 'brl_flavour' to be removed from lock calls.
This will allow SMB2 POSIX handles to call with POSIX_LOCK
flavour instead of always using WINDOWS_LOCK (as now).
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Noel Power <npower@samba.org>
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Autobuild-User(master): Ralph Böhme <slow@samba.org>
Autobuild-Date(master): Thu Dec 30 11:54:17 UTC 2021 on sn-devel-184
fd_close() mostly wraps SMB_VFS_CLOSE() but also takes care of refcounting
fsp->fh properly and also makes sure that fsp->fh->fd is set to -1 after close.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
After openat_pathref_fsp() is successful we know we have a VALID_STAT().
It either returns NT_STATUS_OK or we look at the stat
struct for S_ISLNK so we know we have VALID_STAT().
If it's not successful we error out, so we don't need
another vfs_stat() here.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Thu Dec 16 07:33:09 UTC 2021 on sn-devel-184
We know at this point that we have VALID_STAT(smb_fname->st).
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Noel Power <npower@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Wed Dec 15 19:26:50 UTC 2021 on sn-devel-184
qpathinfo must operate on an existing file, so we
can exit early if filename_convert() returned the "new file"
NT_STATUS_OK, !VALID_STAT case.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Noel Power <npower@samba.org>
