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 main reason for this change was the use of
iov_advance() in the next commits in
source3/smbd/smb2_server.c
And the function calls to iov_advance() showed up
in profiling with callgrind.
While there iov_buf() and iov_buflen() are moved as
well, as they are also used there.
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Commit 83fe7a0316 converted the
stdout/stderr logging types to DEBUG_FILE to get a header when using
DEBUG_SYSLOG_FORMAT_ALWAYS. However, this causes all configured
backends to be invoked. When syslog is one of those backends then
this is almost certainly not what is intended.
Instead, call debug_file_log() directly in that special case and
revert the parts of the above commit that convert to file logging.
Most of the changes to debughdrclass() still seem necessary, since
they handle the change of debug_syslog_format from a bool to an enum.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15460
Signed-off-by: Martin Schwenke <mschwenke@ddn.com>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Autobuild-User(master): Martin Schwenke <martins@samba.org>
Autobuild-Date(master): Mon Aug 28 01:21:07 UTC 2023 on atb-devel-224
CH_DISPLAY was removed in commit
125a2ff262, but NUM_CHARSETS was not
updated to match.
By assigning to NUM_CHARSETS the last enumeration value in charset_t, we
guard against its falling out of sync again.
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
The purpose of this is to make it clear which part of the AD DC (in particular)
has faulted without having to deduce it from the stacktrace.
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
This means samba_tevent_debug() is only called when needed.
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
We should not skip all of close_low_fd() just because we
detected valgrind headers at build time.
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
Prefer 'size_t' over 'int' in generate_random_buffer(),
generate_secret_buffer() and generate_nonce_buffer() to
match an underlying gnutls_rnd() calls.
Signed-off-by: Dmitry Antipov <dantipov@cloudlinux.com>
Reviewed-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
runtime error: left shift of 65535 by 16 places cannot be represented in type 'int'
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
This avoids unnecessary attempts to load libgpfs.so when it is not
needed.
Signed-off-by: Christof Schmitt <cs@samba.org>
Reviewed-by: Michael Tokarev <mjt@tls.msk.ru>
Reviewed-by: Volker Lendecke <vl@samba.org>
Autobuild-User(master): Volker Lendecke <vl@samba.org>
Autobuild-Date(master): Fri Apr 14 12:28:23 UTC 2023 on atb-devel-224
Prefer mallinfo2() with 'size_t' fields over deprecated
mallinfo() (with 'int' fields which may wrap around zero
and so be inaccurate on a 64-bit system) and move relevant
checks to lib/util/wscript_configure because mallinfo()
is not used beyond 'samba-util'.
Suggested-by: Andreas Schneider <asn@samba.org>
Signed-off-by: Dmitry Antipov <dantipov@cloudlinux.com>
Reviewed-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
In file_load()/file_lines_load(), the file's fd is obtained using
open(), and in fd_load() the fd is converted to a FILE* using
fdopen(). However, after fclose(), the fd is closed again using
close().
Bug: https://bugzilla.samba.org/show_bug.cgi?id=15311
Signed-off-by: baixiangcpp baixiangcpp@gmail.com
Reviewed-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Autobuild-User(master): Volker Lendecke <vl@samba.org>
Autobuild-Date(master): Thu Feb 16 12:13:05 UTC 2023 on atb-devel-224
Removes need for external stat() code when checking for timechange.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
Autobuild-User(master): Andreas Schneider <asn@cryptomilk.org>
Autobuild-Date(master): Fri Jan 27 08:30:35 UTC 2023 on atb-devel-224
One of changes is somewhat interesting, it is "tfork waiter proces"
process title in tfork.c. I wonder why no one noticed this before.
There's another similar process title in there, "tfork waiter process(%d)".
Hopefully no one does grep for "proces$" (and there's no reason to).
Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>
Reviewed-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Rowland Penny <rpenny@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Thu Jan 26 20:46:11 UTC 2023 on atb-devel-224
The cleanup phase of tfork_create() saves errno prior to calling
functions that might modify it, with the intention of restoring it
afterwards. However, the value of 'ret' is accidentally overwritten. It
will always be equal to 0, and hence errno will not be restored.
Fix this by introducing a new variable, ret2, for calling functions in
the cleanup phase.
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Jeremy Allison <jra@samba.org>
No need to recompile the world when only a few files need this.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
To be used in smbXsrv_open.c, for this we need a lower bound.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
With the proper NULL checks we don't need the stackframe,
use a passed in context instead.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
This adds intermediate NULL checks via talloc_asprintf_addbuf()
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
I wanted to use this in debug.c, but this would have meant to pollute
debug's deps with a lot of stuff. Also, looking through uses of
talloc_asprint_append(), very many of those don't do NULL checks
properly and could benefit from the _addbuf() flavor. We can add a
vasprintf variant later if the need shows up.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Sometimes (e.g. in lzxpress Huffman encoding, and in some of our
tests: c.f. https://lists.samba.org/archive/samba-technical/2018-March/126010.html)
we want a stable sort algorithm (meaning one that retains the previous
order of items that compare equal).
The GNU libc qsort() is *usually* stable, in that it first tries to
use a mergesort but reverts to quicksort if the necessary allocations
fail. That has led Samba developers to unthinkingly assume qsort() is
stable which is not the case on many platforms, and might not always
be on GNU/Linuxes either.
This adds four functions. stable_sort() sorts an array, and requires
an auxiliary working array of the same size. stable_sort_talloc()
takes a talloc context so it ca create a working array and call
stable_sort(). stable_sort_r() takes an opaque context blob that gets
passed to the compare function, like qsort_r() and ldb_qsort(). And
stable_sort_talloc_r() rounds out the quadrant.
These are LGPL so that the can be used in ldb, which has problems with
unstable sort.
The tests are borrowed and extended from test_ldb_qsort.c.
When sorting non-trivial structs this is roughly as fast as GNU qsort,
but GNU qsort has optimisations for small items, using direct
assignments of rather than memcpy where the size allows the item to be
cast as some kind of int.
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Joseph Sutton <josephsutton@catalyst.net.nz>
We need "struct iovec", which comes in via sys/uio.h, incuded by
system/filesys.h
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Panic if memset_s() fails.
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Checking for presence of both netgroup and innetgr. INNETGR is not
defined on libc's such as musl so not checking results in a build error.
Signed-off-by: listout <brahmajit.xyz@gmail.com>
Reviewed-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
Autobuild-User(master): Volker Lendecke <vl@samba.org>
Autobuild-Date(master): Mon Aug 8 07:28:31 UTC 2022 on sn-devel-184
In this case it is probably better to crash out.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15103
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andreas Schneider <asn@samba.org>
This avoids wasting getpid() calls in a lot of places...
Signed-off-by: Stefan Metzmacher <metze@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): Sun Jul 24 00:25:49 UTC 2022 on sn-devel-184
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15073
Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
Autobuild-User(master): Andreas Schneider <asn@cryptomilk.org>
Autobuild-Date(master): Tue Jul 19 12:17:35 UTC 2022 on sn-devel-184
The executables generated from lib/util/tests/test_logging.c are used
by the samba.tests.logfiles tests to test logging with various
smb.confs that assign classes to various files at different levels
etc.
Previously test_logging.c had its own version of the table; now it
shares one with debug.c
We put the table in a sub-directory (lib/util/debug-classes/), because
adding local_include=True to the wscript_build stanza causes the
compiler confusion between <time.h> and lib/util/time.h.
Note: there are still two other lists of the class names, in
python/samba/tests/logfiles.py and
docs-xml/smbdotconf/logging/loglevel.xml.
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
This can be a useful macro when you are trying to track the behaviour
of one process out of the dozens that samba starts up, and when your
interest is in following it over time, not necessarily in a single
stack.
In DEVELOPER mode, if you call 'debug_developer_enable()' in the
process you're following, then any instances of DBG_DEV() will work
like DBG_ERR(), also adding ":DEV:12345:" where "12345" is the pid of
th current process.
Within debug.c itself, the macro always writes to stderr, because the
debug.c functions are not all reentrant.
When not in DEVELOPER MODE, the macro evaluates to nothing.
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
There is one knownfail, where it seems an smb.conf like
log file = foo
log level = 2 tdb:2@baa ldb:3
will send the ldb logs to 'baa' instead of 'foo' (i.e., the last
opened log file, rather than the default log file).
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
The 'log level' line in smb.conf allows messages from different log
classes to be sent to different places, but we have not tested that
this works. Now we do, somewhat.
The test involves running a special binary based on a stripped down
source4/samba/server.c that just starts up, parses the command line
and a given smb.conf, then logs messages from multiple classes and
exits.
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
CTDB doesn't want this redirection of stderr to the log file. It
expects to be able to capture stderr of subprocesses and log them with
a header. This redirection stops that from happening.
Unfortunately this has to be a negative option (i.e. "no" in the name)
so that the default of 0/false maintains existing behaviour.
Note that the default behaviour is sub-optimal because it causes raw
data (i.e. debug data without a header) to appear in the log.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15090
Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Volker Lendecke <vl@samba.org>
gnutls_memcmp() is mostly identical to our own implementation, except
that ours will not break if supplied with 4 GiB or more of data.
However, using an external function permits us to disclaim
responsibility if some CPU/compiler combination happens to invalidate
our constant-time guarantee.
For reference, gnutls_memcmp() implementation:
78d9820de0/lib/safe-memfuncs.c (L41-67)
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Ensure that it gives the correct results for comparing two memory
regions.
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Ensure that it gives the correct results for comparing two data blobs.
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Since memcmp_const_time() doesn't act as an exact replacement for
memcmp(), and its return value is only ever compared with zero, simplify
it and emphasize the intention of checking equality by returning a bool
instead.
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Since data_blob_cmp_const_time() doesn't act as an exact replacement for
data_blob_cmp(), and its return value is only ever compared with zero,
simplify it and emphasize the intention of checking equality by
returning a bool instead.
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
This helps to avoid timing attacks.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15010
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
This allows it to be used in more places without needing to introduce
more dependencies.
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Backtraces when Samba is in PANIC state are better with
backtrace_symbols() than with libunwind on Ubuntu 20.04 x86_64
so move libunwind to a off-by-default option, prompted for
if backtrace_symbols() is not available.
Based on a request by Fco Javier Felix <ffelix@inode64.com>
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Joseph Sutton <josephsutton@catalyst.net.nz>
We look for backtrace_symbols() in this library, so we should link against
it if we find it.
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Previously this would leave static functions unused, which the compiler will
not allow for a developer build.
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Joseph Sutton <josephsutton@catalyst.net.nz>
This partly reverts commit 620de975f1.
Drop chunk including system/nis.h, drop wscript_build modifications,
use getdomainname() from glibc instead of yp_get_default_domain() from
libnsl.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15087
Signed-off-by: Samuel Cabrero <scabrero@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15069
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Christof Schmitt <cs@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Fri Jun 3 21:53:31 UTC 2022 on sn-devel-184
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
Autobuild-User(master): Andreas Schneider <asn@cryptomilk.org>
Autobuild-Date(master): Mon May 30 11:03:47 UTC 2022 on sn-devel-184
This just prints a warning for:
ms_fnmatch.c:95:51: error: array subscript 0 is outside array bounds of
‘struct max_n[0]’ [-Werror=array-bounds]
95 | if (max_n != NULL && max_n->predot &&
|
Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Alexander Bokovoy <ab@samba.org>
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
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>
This is an alternative patch for MR2339: It seems that Windows AD in
turkish locale is ASCII-compatible with 'i'. Björn tells me that the
turkish locale is the only one where upper/lower casing letters in the
ASCII range is not compatible to ASCII.
Simplify our code by not calling the locale-specific standard
toupper/tolower for the ASCII range but rely on our tables.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Alexander Bokovoy <ab@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
Autobuild-User(master): Volker Lendecke <vl@samba.org>
Autobuild-Date(master): Mon Apr 4 11:45:24 UTC 2022 on sn-devel-184
This is quite bizarre:
*** CID 1499409: Memory - corruptions (OVERLAPPING_COPY)
/lib/util/debug.c: 1742 in dbghdrclass()
1736 sizeof(tvbuf.buf),
1737 "%ld seconds since the Epoch", (long)t);
1738 }
1739 }
1740
1741 ensure_hostname();
>>> CID 1499409: Memory - corruptions (OVERLAPPING_COPY)
>>> In the call to function "snprintf", the object pointed to by argument "state.hostname" may overlap with the object pointed to by argument "state.header_str".
1742 state.hs_len = snprintf(state.header_str,
1743 sizeof(state.header_str),
1744 "%s %s %s[%u]: ",
1745 tvbuf.buf,
1746 state.hostname,
1747 state.prog_name,
Coverity doesn't explicitly say so but the only way this can happen is
if state.hostname is not NUL-terminated within its declared length.
ensure_hostname() and debug_set_hostname() ensure NUL-termination, but
the caching effect of ensure_hostname() probably stops Coverity from
being certain about anything.
Try making Coverity happy by using a precision to limit the number of
characters from hostname that can be used.
Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Autobuild-User(master): Martin Schwenke <martins@samba.org>
Autobuild-Date(master): Tue Feb 22 00:17:12 UTC 2022 on sn-devel-184
clang complains:
../../lib/util/genrand_util.c:99:9: error: variable 'num_chars' set but not used [-Werror,-Wunused-but-set-variable]
size_t num_chars = 0;
^
That is, the variable is initialised and incremented but the value is
never used.
Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Volker Lendecke <vl@samba.org>
No need to duplicate the fopen/fclose
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): Mon Feb 7 19:58:57 UTC 2022 on sn-devel-184
That will make it easy to see the difference
between two memory buffers.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14956
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
This simplifies the logic a lot for me.
It also fixes some corner cases regarding whitespaces in the
output, that's why we have to mark a few tests as knownfail,
they will be fixed in the next commit.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14956
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
Cleanup to eliminate duplicate code.
The high check is now done against ret.tv_sec,
not 'd', as after calling nt_time_to_unix_timespec_raw()
this is identical to the previous intermediate 'd'
variable.
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): Tue Jan 11 01:36:51 UTC 2022 on sn-devel-184
Cleanup to eliminate duplicate code.
The low/high checks are now done against ret.tv_sec,
not 'd', as after calling nt_time_to_unix_timespec_raw()
this is identical to the previous intermediate 'd'
variable.
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Christof Schmitt <cs@samba.org>
Not yet used. Does no checks on the converted values.
A later cleanup will allow us to move nt_time_to_unix_timespec()
and nt_time_to_full_timespec() to use common code.
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>
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): Wed Jan 5 01:02:38 UTC 2022 on sn-devel-184
Padding this buffer to a multiple of 8 bytes allows the PAC buffer
padding to match Windows.
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Fixes build error with samba-4.15.3 and uClibc:
../../source3/printing/samba-bgqd.c: In function ‘main’:
../../source3/printing/samba-bgqd.c:340:21: error: ‘SIGPIPE’ undeclared (first use in this function); did you mean ‘EPIPE’?
../../source3/printing/samba-bgqd.c:384:14: error: ‘SIGTERM’ undeclared (first use in this function)
Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Reviewed-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Günther Deschner <gd@samba.org>
Autobuild-User(master): Volker Lendecke <vl@samba.org>
Autobuild-Date(master): Mon Dec 13 16:22:28 UTC 2021 on sn-devel-184
Over the last month I got more and more reports,
that it's not possible to use a custom Samba version
on systems with sssd being installed, which depends on some
specific samba libraries installed in the system.
One major problem is that the custom libnss_winbind.so.2
depends on the libreplace-samba4.so of the custom build
and also injects an RPATH into the running process.
When sssd uses any nss library call it will get this,
when it then tries to load some of its plugins via dlopen(),
e.g.
ldd /usr/lib64/sssd/libsss_ad.so| grep samba
libsamba-util.so.0 => /lib64/libsamba-util.so.0
libreplace-samba4.so => /usr/lib64/samba/libreplace-samba4.so
libsamba-security-samba4.so => /usr/lib64/samba/libsamba-security-samba4.so
libsamba-errors.so.1 => /lib64/libsamba-errors.so.1
libsamba-debug-samba4.so => /usr/lib64/samba/libsamba-debug-samba4.so
libgenrand-samba4.so => /usr/lib64/samba/libgenrand-samba4.so
libsocket-blocking-samba4.so => /usr/lib64/samba/libsocket-blocking-samba4.so
libtime-basic-samba4.so => /usr/lib64/samba/libtime-basic-samba4.so
libsys-rw-samba4.so => /usr/lib64/samba/libsys-rw-samba4.so
libiov-buf-samba4.so => /usr/lib64/samba/libiov-buf-samba4.so
When that loads dlopen() will fail as a soname libreplace-samba4.so is
already loaded, but the symbol version within the other one don't match, as the
contain the exact version, e.g. replace_dummy@@SAMBA_4.13.3.
This is just an example and similar things can happen in all situations
where we provide libraries, which are potentially injected into every
process of the running system. These should only depend on libc.so and
related basic system libraries in order to avoid the problem.
We have the following libraries, which are in the that category:
- libnss_winbind.so.2
- libnss_wins.so.2
- pam_winbind.so
- winbind_krb5_locator.so
- async_dns_krb5_locator.so
The rules of library loading are really complex and symbol versioning
is not enough to solve it, only the combination of unique soname and
unique symbol version suffix seem to solve the problem, but injecting
an RPATH is still a problem.
In order to solve the problem I experimented with adding SAMBA_SUBSYSTEM()
definitions with 'hide_symbols=True' in order to do some static linking
of selected components, e.g.
bld.SAMBA_SUBSYSTEM('replace-hidden',
source=REPLACE_SOURCE,
group='base_libraries',
hide_symbols=True,
deps='dl attr' + extra_libs)
It's relatively simple to get to the point where the following are
completely static:
- libnss_winbind.so.2
- libnss_wins.so.2
- pam_winbind.so
- winbind_krb5_locator.so
But 'async_dns_krb5_locator.so' links in almost everything!
It seems we install the krb5 plugins into our own $MODULESDIR/krb5/,
so it may not be so critical, as long it's the admin who created
the desired symlinks into the location the kerberos libraries search
for plugins. Note the at least the locator plugins are always loaded
without any configuration, every .so in a special path are loaded with dlopen().
This is done by every application using kerberos, so we load a lot of samba libraries
into them.
Packagers should not put async_dns_krb5_locator.so (nor a symlink) into
the path that's reachable by libkrb5.so.
As a longterm solution we may want to change async_dns_krb5_locator.so
to use a helper process with posix_spawn() instead of doing everything
within the process.
Note I added hiden_symbols=True to the nss modules for Linux and
FreeBSD only, because these are the only platforms I'm able to test
on. We most likely should do the same on other platforms, but some
with access to the platform should provide a tested patch.
In order to avoid manual definitions of SAMBA_SUBSYSTEMS() with
'-hidden', I added the 'provide_builtin_linking=True' option,
as the logic is very similar to what we already have with the
'--builtin-libraries=BUILTIN_LIBRARIES' configure option.
SAMBA_PLUGIN() is used in order to use SAMBA_LIBRARY() in order
to make it more strict that these plugins can't be used as
normal depedency by other subsystems and libraries.
While being there it was easy enough to make libwbclient.so
also standalone without dependecies to other samba libraries.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14780
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
This simplifies the logging API for callers that typically would want to set
logging by just setup_logging() once without bothering that typically
configuration is loaded (via some lpcfg_load*() or lp_load*() varient) which
will only then pick up the configured logfile from smb.conf without actually
applying the new logifle to the logging subsytem.
Therefor our daemons will additionally call reopen_logs() explicitly in their
startup code after config is loaded, eg
setup_logging(getprogname(), DEBUG_FILE);
...
lpcfg_load(lp_ctx, config_file);
...
reopen_logs();
By calling reopen_logs_internal() implicitly from debug_set_logfile() there's no
need to call reopen_logs() explicitly anymore to apply the logfile.
As reopen_logs() will also apply other logging configuration options, we have to
keep the explicit calls in the daemon code. But at least this allows consistent
logging setup wrt to the logfile in the new cmdline library.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14897
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
Before I added per-class logfile and we had only one fd for the logfile the code
looked like this:
/* Take over stderr to catch output into logs */
if (state.fd > 0) {
if (dup2(state.fd, 2) == -1) {
/* Close stderr too, if dup2 can't point it -
at the logfile. There really isn't much
that can be done on such a fundamental
failure... */
close_low_fd(2);
}
}
In the current code the equivalent to state.fd is dbgc_config[DBGC_ALL].fd.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14897
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Ralph Boehme <slow@samba.org>
Autobuild-User(master): Ralph Böhme <slow@samba.org>
Autobuild-Date(master): Mon Nov 1 07:29:47 UTC 2021 on sn-devel-184
Without debug_hires_timestamp this produces a syslog style header
containing:
"MON DD HH:MM:SS HOSTNAME PROGNAME[PID] "
With debug_hires_timestamp this produces a syslog style header
containing:
"RFC5424-TIMESTAMP HOSTNAME PROGNAME[PID] "
All other settings are ignored.
This will be made visible via smb.conf in a subsequent commit.
This commit adds some simple hostname handling. It avoids using
get_myname() from util.c because using that potentially pulls in all
manner of dependencies. No real error handling is done. In the worst
case debug_set_hostname() sets the hostname to a truncated version of
the given string. Similarly, in an even weirder world,
ensure_hostname() sets the hostname to a truncation of "unknown".
Both of these are unlikely in all reasonable cases.
Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Ralph Boehme <slow@samba.org>
Preliminary handling of NTTIME_THAW to avoid NTTIME_THAW is passed as some
mangled value down to the VFS set timestamps function.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14127
RN: Avoid storing NTTIME_THAW (-2) as value on disk
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
NTTIME_FREEZE is not a nil sentinel value, instead it implies special, yet
unimplemented semantics. Callers must deal with those values specifically and
null_nttime() must not lie about their nature.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14127
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
The test was checking -1 twice:
torture_assert(tctx, null_nttime(-1), "-1");
torture_assert(tctx, null_nttime(-1), "-1");
The first line was likely supposed to test the value "0".
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14127
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
If it isn't used then it isn't copied.
Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Volker Lendecke <vl@samba.org>
Autobuild-User(master): Volker Lendecke <vl@samba.org>
Autobuild-Date(master): Thu Oct 14 11:10:40 UTC 2021 on sn-devel-184
The existing check is for truncation, not whether the buffer is full.
However, if the buffer is full (i.e. hs_len == sizeof(header_str) - 1)
then there's no use trying subsequent snprintf() calls because there
will be one byte available that already contains the NUL-terminator.
A subsequent call will just do a no-op truncation.
Check for full buffer instead.
This might be confusing because it isn't the standard check that is
done after snprintf() calls. Is it worth it for a rare corner case?
Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Volker Lendecke <vl@samba.org>
strlcat() needs to walk to the end of its first argument. However,
but the length of state.header_str is already known, so optimise by
manually appending the extra characters if they will fit.
Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Volker Lendecke <vl@samba.org>
This the first step in avoiding potentially repeated length
calculations in the backends. The length is known at call time for
most usual callers, so pass it down.
Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Volker Lendecke <vl@samba.org>
Currently the file backend can produce something like:
HEADER1 HEADER2 TEXT2
TEXT1
when different processes try to log at the same time.
Avoid this by writing the header and text at the same time using
writev(). This means that the header always has to be written by the
backend, so update all backends to do this.
The non-file backends should behave as before when they were invoked
separately to render the header. It might be possible to optimise
some of them (e.g. via sd_journal_sendv) but this requires more
investigation (e.g. sd_journal_sendv()'s handling of newlines) and is
beyond the scope of this change.
state.header_str_no_nl takes the place of msg_no_nl for the header,
since some of the backends need the no-newline version. It is handled
the same was as msg_no_nl: produce the no_nl version exactly once,
whether or not it is needed, since this is better than repeating it in
several backends.
Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Volker Lendecke <vl@samba.org>
Add an internal function to do the work and call it. It will be
called again in a subsequent commit.
Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Volker Lendecke <vl@samba.org>
They'll need to be accessible by the backends.
Note that the snprintf() and strlcat() calls can result in
state.hs_len >= sizeof(state.header_str), so state.hs_len needs to be
sanitised before any potential use. Previously this wasn't necessary
because this value was on the stack, so it couldn't be used after
dbghdrclass() returned.
Signed-off-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Volker Lendecke <vl@samba.org>
The DEBUG macro was missing and the CFStringGetBytes() was triggering a
-Werror,-Wpointer-sign build failure.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14862
Signed-off-by: Alex Richardson <Alexander.Richardson@cl.cam.ac.uk>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
This implements MS-FSA algorithms BlockAlign() and BlockAlignTruncate().
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Simplifies building up a string step by step, see next commit
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
The only writer to this variable left with c377845d27. The
closest match for override_logfile is is_default_dyn_LOGFILEBASE()
with the opposite logic.
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): Sat Sep 18 00:53:28 UTC 2021 on sn-devel-184
Don't directly overwrite the pointer for a realloc. On failure, the
original pointer is still valid.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
debug classes registered by vfs modules are not available immediately.
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Fri Jun 18 04:27:17 UTC 2021 on sn-devel-184
If we allow a string that encodes say '\0' as a multi-byte sequence,
we are open to confusion where we mix NUL terminated strings with
sized data blobs, which is to say EVERYWHERE.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14684
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Jeremy Allison <jra@samba.org>
