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.
shadow_copy2_strip_snapshot() will happily return without modifying the passed
timestamp=0 if the path is already converted and refers to an object in a
snapshot, eg (first debug line from extra debugging patch [1]):
[10 2021/07/02 08:19:28.811424 pid=738290 ../../source3/modules/vfs_shadow_copy2.c:1303 shadow_copy2_fstat]
shadow_copy2_fstat: fsp [test.txt {@GMT-2000.01.02-03.04.05}]
[10 2021/07/02 08:19:28.811449 pid=738290 ../../source3/modules/vfs_shadow_copy2.c:607 _shadow_copy2_strip_snapshot_internal]
_shadow_copy2_strip_snapshot_internal: [from shadow_copy2_fstat()] Path 'test.txt {@GMT-2000.01.02-03.04.05}'
[10 2021/07/02 08:19:28.811474 pid=738290 ../../source3/modules/vfs_shadow_copy2.c:619 _shadow_copy2_strip_snapshot_internal]
_shadow_copy2_strip_snapshot_internal: abs path '/gpfs0/smb_snapshots2/filesetone/.snapshots/@GMT-2000.01.02-03.04.05/test.txt'
[10 2021/07/02 08:19:28.811496 pid=738290 ../../source3/modules/vfs_shadow_copy2.c:1924 shadow_copy2_snapshot_to_gmt]
shadow_copy2_snapshot_to_gmt: match @GMT-%Y.%m.%d-%H.%M.%S: @GMT-2000.01.02-03.04.05
[10 2021/07/02 08:19:28.811536 pid=738290 ../../source3/modules/vfs_shadow_copy2.c:566 check_for_converted_path]
check_for_converted_path: path |/gpfs0/smb_snapshots2/filesetone/.snapshots/@GMT-2000.01.02-03.04.05/test.txt| is already converted. connect path = |/gpfs0/smb_snapshots2/filesetone/.snapshots/@GMT-2000.01.02-03.04.05|
As check_for_converted_path() detects an "already converted path",
_shadow_copy2_strip_snapshot_internal() just returns without modifying the value
of the timestamp.
By using shadow_copy2_strip_snapshot_converted() instead of
shadow_copy2_strip_snapshot() we can check if the path is in fact referring to a
VSS object by checking the "converted" bool.
An alternative way would have been directly checking fsp->fsp_name->twrp != 0,
but that would be a new semantic in the module, I'll leave this excersize for
the future when we clean up the usage of shadow_copy2_strip_snapshot() in the
whole module.
This change also switches to using the absolute paths in both place where
convert_sbuf() is called.
[1]
@@ -1309,8 +1348,16 @@ static int shadow_copy2_fstat(vfs_handle_struct *handle, files_struct *fsp,
saved_errno = errno;
}
+ DBG_DEBUG("fsp [%s]\n", fsp_str_dbg(fsp));
RN: vfs_shadow_copy2 fixinodes not correctly updating inode numbers
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14756
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
This will fail with
Failed to open file \@GMT-2015.10.31-19.40.30\subdir\hardlink. NT_STATUS_ACCESS_DENIED
The open is failing in openat_pathref_fsp():
[2021/07/06 04:58:17.677104, 10, pid=95070, effective(1000, 1000), real(1000, 0)] ../../source3/smbd/files.c:541(openat_pathref_fsp)
openat_pathref_fsp: file [subdir/hardlink {@GMT-2015.10.31-19.40.30}] - dev/ino mismatch. Old (dev=64770, ino=3826943444). New (dev=64770, ino=1746568660).
[2021/07/06 04:58:17.677114, 10, pid=95070, effective(1000, 1000), real(1000, 0)] ../../source3/smbd/files.c:568(openat_pathref_fsp)
openat_pathref_fsp: Opening pathref for [subdir/hardlink {@GMT-2015.10.31-19.40.30}] failed: NT_STATUS_ACCESS_DENIED
The reason is subtle:
shadow_copy2 calculates inode numbers of snapshot files based on the path of the
file. The result of that when doing a path based stat() from filename_convert()
was
[2021/07/06 04:58:17.676159, 10, pid=95070, effective(1000, 1000), real(1000, 0)] ../../source3/smbd/filename.c:1945(filename_convert_internal)
filename_convert_internal: XXX smb_fname [subdir/hardlink {@GMT-2015.10.31-19.40.30}] (dev=64770, ino=3826943444).
which is the "Old" inode shown above.
Later in the open code called from openat_pathref_fsp() -> fd_openat() ->
non_widelink_open() since 4.14 we call SMB_VFS_FSTAT() where fsp->fsp_name will
be set to the new relative *basename* of the file:
[2021/07/06 04:58:17.676917, 10, pid=95070, effective(1000, 1000), real(1000, 0), class=vfs] ../../source3/modules/vfs_default.c:1302(vfswrap_fstat)
vfswrap_fstat: XXX fsp [hardlink {@GMT-2015.10.31-19.40.30}] (dev=64770, ino=3826943444)
So for stat() the hash function in called with the full path relative to the share
root:
subdir/hardlink
while for fstat() the hash function will used
hardlink
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14756
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
fd_openat() has done an FSTAT on the handle so update the smb_fname stat info
with "truth". from the handle.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14756
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
This reverts commit a6df051dd5e8c63f2fdfdb20ee01169d2bdb97dd:
"s3: smbd: In openat_pathref_fsp(), just check we're opening the same file type, not dev and inode."
The prior changes mean we can go back to checking dev/ino
matches.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14756
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
This helps code inside any module implementing fstat() looking at
fsp->fsp_name->st instead of the passed in stat buf.
I only ran afoul of this in a DEBUG message I added while debugging some inode
related problem.
No change in behaviour.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14756
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Remove my poor imitation of synthetic_pathref(), just call the real thing.
We need to go through the full VFS stack here to get
the ino correct to get the fsp handle.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14756
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
We'll be calling SMB_VFS_FGETXATTR() on the base fsp anyway.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14756
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
We actually don't need this to get the right semantics, as the open or the
set_sd() code catches the correct cases and returns ACCESS_DENIED, but it
makes me much happier to see the prerequisites needed expressed in code here
right at the point of use.
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): Tue Jul 13 08:11:36 UTC 2021 on sn-devel-184
Now map_nt_error_from_unix() returns the same value as
map_nt_error_from_unix_common().
Signed-off-by: Samuel Cabrero <scabrero@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
If libtdb is used from the system, we should use those tools by default.
Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
I saw systems with locking.tdb records being part of:
ctdb catdb smbXsrv_tcon_global.tdb
It's yet unknown how that happened, but we should not panic in srvsvc_*
calls because the info0 pointer was NULL.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14752
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
Autobuild-User(master): Stefan Metzmacher <metze@samba.org>
Autobuild-Date(master): Tue Jul 6 11:08:43 UTC 2021 on sn-devel-184
If the string is too shhort we don't want to atoi() whatever is beyond
the end of it.
Found using Honggfuzz and the fuzz_parse_lpq_entry fuzzer.
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Autobuild-User(master): Andrew Bartlett <abartlet@samba.org>
Autobuild-Date(master): Mon Jul 5 05:07:13 UTC 2021 on sn-devel-184
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): Sun Jul 4 18:01:16 UTC 2021 on sn-devel-184
Not yet used. We will use this to replace calls
to ad_get(..., ADOUBLE_META). It uses openat_pathref_fsp()
to get a handle before calling into ad_get(..., ADOUBLE_META).
Uses the recursion guard to prevent recursion into openat_pathref_fsp()
from stat calls within.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Not yet used.
Same technique as used to prevent recursion in stat calls in vfs_fake_acls.c
This will go away once SMB_VFS_STATX() is added and we can select exactly
what fields we are calling stat() to get.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Both must succeed for a valid return, and we're next going
to replace ad_get() with a wrapper that calls openat_pathref_fsp(),
which needs a VALID_STAT().
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Eventually we will guarantee a valid smb_fname->fsp here and
will remove SMB_VFS_GETXATTR().
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Currently in master when we call into openat() in the VFS
we violate the invarient:
fsp->base_fsp->fsp_name->fsp == fsp->base_fsp.
The reason for this is subtle. Inside open.c:non_widelink_open()
we change the fsp->base_fsp to be relative to the new $cwd.
We do this by the following code in open.c:non_widelink_open():
/* Also setup base_fsp to be relative to the new cwd */
if (fsp->base_fsp != NULL) {
base_smb_fname_rel = (struct smb_filename) {
.base_name = smb_fname_rel->base_name,
};
orig_base_fsp_name = fsp->base_fsp->fsp_name;
fsp->base_fsp->fsp_name = &base_smb_fname_rel;
}
Note that fsp->base_fsp->fsp_name now points at a
stack variable struct smb_filename, with smb_fname->fsp == NULL.
This fixes that problem by removing the horrid
stack based smb_filename and changing to use a
talloc'ed fsp->base_fsp->fsp_name, with
correctly linked fsp->base_fsp->fsp_name-> pointer.
Remove the selftest/knownfail.d/fruit_vfs_invariant
file as all vfs_fruit tests now pass again.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
vfs_fruit isn't the bad guy here. It's just a convenient
place to show that non_widelink_open() violates:
fsp->base_fsp->fsp_name->fsp == fsp->base_fsp invarient
Add selftest/knownfail.d/fruit_vfs_invariant to show
what this breaks. Next patch will fix the non_widelink_open()
code and remove the knownfail.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
net ads keytab list prints extra new line for uknown encoding types,
so it spans over two lines, instead over a single line:
1 AES-128 CTS mode with 96-bit SHA-1 HMAC ADDC$@ADDOM.SAMBA.EXAMPLE.COM
1 UNKNOWN: 3
ADDC$@ADDOM.SAMBA.EXAMPLE.COM
Signed-off-by: Pavel Filipenský <pfilipen@redhat.com>
Reviewed-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
Autobuild-User(master): Ralph Böhme <slow@samba.org>
Autobuild-Date(master): Fri Jul 2 11:45:48 UTC 2021 on sn-devel-184
Unintentionally used fsp_get_pathref_fd() in the initial patchset.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=12033
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Thu Jul 1 17:45:49 UTC 2021 on sn-devel-184
The following are the default values:
preopen:nomatch_log_level = 5
preopen:match_log_level = 5
preopen:nodigits_log_level = 1
preopen:founddigits_log_level = 3
preopen:reset_log_level = 5
preopen:push_log_level = 3
preopen:queue_log_level = 10
This gives admins a way to debug/audit the preopen usage.
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
This will allow the usage of patterns as
'POSIX Basic Regular Expression'
vfs objects = preopen
preopen:posix-basic-regex = yes
preopen:names = /Re7599Ex\([0-9]\).*\.txt/test\([0-9]*\)\.dat/
The key is that exactly one 'subexpression' starting with '\(' and
ending with '\)' is specified in order to select the position where
the digits are searched.
E.g. given a file name 'Re7599Ex01234.txt' will actually preopen:
Re7599Ex01234.txt
Re7599Ex11234.txt
Re7599Ex21234.txt
Re7599Ex31234.txt
Re7599Ex41234.txt
As '\([0-9]\)' will only match the first digit after 'Re7599Ex'.
It also means it's now possible to have digits in the fixed part of the
filename, which was the actual motivation for this patchset.
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
samba_path_matching_check_last_component() may return the start and end
offset of a submatch (for us the bytes where the digits are expected).
We use that in order to allow preopen_parse_fname() to just
look at these bytes and ignore any trailing digits after the submatch.
For the current use of samba_path_matching_mswild_create(),
there's no difference as we'll always get replace_start=-1 and
replace_end=-1. But the next commit will make optional use of
samba_path_matching_regex_sub1_create(), which will change the situation
and allow to return hints we got from regexec().
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
If we have a single digit we only replace up to '9', which also fits
into a single digit.
We operate on numbers from 0 to 9999999999999999999 independent of the
architecture.
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Calculating the start_idx and num_digits at the first possible place
will make the following commits much easier.
At the end we just want to assign the return values without any logic.
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
There's no point in trying to check if the current number is part
of the existing queue. This makes the logic at least more unstandable
to me.
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
