summaryrefslogtreecommitdiffstats
path: root/src/ulogd.c
Commit message (Collapse)AuthorAgeFilesLines
* ulogd: raise error on unknown config keyCorubba Smith2025-03-121-0/+2
| | | | | | | | | | | | | | | | | | Until a6fbeb96e889 ("new configuration file syntax (Magnus Boden)") this was already caught, and the enum member is still present. Check if the for loop worked throught the whole array without hitting a matching config option, and return with the unknown key error code. Because there is no existing config_entry struct with that unknwon key to use with the established config_errce pointer, allocate a new struct. This potentially creates a memory leak if that config_entry is never freed again, but for me that is acceptable in this rare case. Since the memory allocation for the struct can fail, also reuse the old out-of-memory error to indicate that. Signed-off-by: Corubba Smith <corubba@gmx.de> Signed-off-by: Florian Westphal <fw@strlen.de>
* ulogd: provide default configure implementationCorubba Smith2025-03-121-8/+18
| | | | | | | | | | | Provide a default implementation for the configure hook which simply calls ulogd_parse_configfile(), so simple plugins only need to provide the config_keyset. This also triggers an "unknown key" error if a plugin defines no config_keyset (aka it has no options), but the config file contains directives for it. Signed-off-by: Corubba Smith <corubba@gmx.de> Signed-off-by: Florian Westphal <fw@strlen.de>
* ulogd: improve integer option parsingCorubba Smith2025-03-121-0/+10
| | | | | | | | | | | | | | | | | | | | | | | | | | The `value` union member in `struct config_entry` is declared as `int` since basically the beginning in e07722e46001 ("config stuff added"). The parsing was switched from the original `atoi()` in 015849995f7f ("Fix hexadecimal parsing in config file") to `strtoul()`. Switch the function for parsing to the signed `strtol()` variant since the result will be stored in a signed int, and it makes sense to support negative numbers. Detect when `strtol()` does not properly consume the whole argument and return a new format error. Also check the numerical value to make sure the signed int does not overflow, in which case a new range error is returned. Unfortunately there is no `strtoi()` which would do the proper range check itself, so the intermediate `long` and range-check for `int` is required. I also considered changing the `value` union member from `int` to `long`, which would make it possible to use the parsed value as-is. But since this is part of the api towards plugins (including third party) such a potentially breaking change felt unwarranted. This also means that still only 16bit integer values are *guaranteed* to work, although most platforms use bigger widths for int. Signed-off-by: Corubba Smith <corubba@gmx.de> Signed-off-by: Florian Westphal <fw@strlen.de>
* ulogd: add ulogd_parse_configfile public functionCorubba Smith2025-03-121-49/+48
| | | | | | | | | | | | | | | | | | Provide a new function `ulogd_parse_configfile()` in the public interface, which wraps `parse_config_file()` to parse a section of the config file and communicates found errors to the user. It can be used as a drop-in replacement because arguments and return value are compatible. This relieves plugins of the need to translate the individual error codes to human readable messages, and plugins are mostly interested if there is any error, not what specific error. This reuses the existing `parse_conffile()` function with slight adjustments. Signed-off-by: Corubba Smith <corubba@gmx.de> Signed-off-by: Florian Westphal <fw@strlen.de>
* ulogd: ignore private data on plugin stopCorubba Smith2025-03-111-3/+1
| | | | | | | | | | | | | | | When deciding whether to call the stop hook of a plugin instance, only two things are relevant: If the plugin actually has a stop hook defined, and if the plugin instance is still used in a different stack. The private data of a plugin instance is opaque to ulogd, so its size or content are irrelevant to the stop-hook decision. And in the same vein should ulogd never write to it. The one-null-byte write could previously lead to an out-of-bounds write on plugins with a stop hook and zero-size private data. Signed-off-by: Corubba Smith <corubba@gmx.de> Signed-off-by: Florian Westphal <fw@strlen.de>
* ulogd2: avoid use after free in unregister on global ulogd_fds linked listKyuwon Shim2023-03-201-0/+9
| | | | | | | | | | | | | | | | | | | | | | | | Invalid read of size 4 at 0x405F60: ulogd_unregister_fd (select.c:74) by 0x4E4E3DF: ??? (in /usr/lib/ulogd/ulogd_inppkt_NFLOG.so) by 0x405003: stop_pluginstances (ulogd.c:1335) by 0x405003: sigterm_handler_task (ulogd.c:1383) by 0x405153: call_signal_handler_tasks (ulogd.c:424) by 0x405153: signal_channel_callback (ulogd.c:443) by 0x406163: ulogd_select_main (select.c:105) by 0x403CF3: ulogd_main_loop (ulogd.c:1070) by 0x403CF3: main (ulogd.c:1649) Problem is that ulogd_inppkt_NFLOG.c::stop() calls ulogd_unregister_fd() which does llist_del(). This llist_del may touch ->prev pointer. As the list element is in private data, we cannot do this llist_del from stop_pluginstances(). Therefore, the free() process moved location after finishing ulogd_unregister_fd(). Signed-off-by: Kyuwon Shim <kyuwon.shim@alliedtelesis.co.nz> Signed-off-by: Florian Westphal <fw@strlen.de>
* ulogd: fix parse-error checkJeremy Sowden2022-12-081-1/+1
| | | | | | | | | If `config_parse_file` returns `-ERRTOOLONG`, `config_errce` may be `NULL`. However, the calling function checks whether `config_errce->key` is `NULL` instead. Signed-off-by: Jeremy Sowden <jeremy@azazel.net> Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
* Replace malloc+memset with callocJeremy Sowden2021-11-301-2/+1
| | | | | | | | There are a number of places where we `malloc` some memory and then `memset` it to zero. Use `calloc` instead. Signed-off-by: Jeremy Sowden <jeremy@azazel.net> Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
* ulogd: fix order of log argumentsJeremy Sowden2021-11-231-1/+1
| | | | | | | | | If `daemon` fails during start-up, ulogd attempts to print `errno` and `strerror(errno)` to the log. However, the arguments are the wrong way round. Swap them. Signed-off-by: Jeremy Sowden <jeremy@azazel.net> Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
* ulogd: remove empty log-lineJeremy Sowden2021-11-231-1/+0
| | | | | | | | | | There is a `strdup` at the beginning of `create_stack`. If it fails, an empty log-line is printed. It's not useful, so remove it. This is consistent with the error-handling of the `malloc` which immediately follows it. Signed-off-by: Jeremy Sowden <jeremy@azazel.net> Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
* ulogd: fix build with musl libcCameron Norman2018-11-011-0/+1
| | | | | | | | The attached patch fixes building ulogd2 with musl libc. It is being used on Void Linux right now. Closes: https://bugzilla.netfilter.org/show_bug.cgi?id=1278 Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
* ulogd: load all plugins by defaultArturo Borrero Gonzalez2018-01-081-1/+39
| | | | | | | | | | | | | | | | | | | | | | | | | | | This new configuration behaviour option eases a bit the configuration of ulogd2 by allowing to load all plugins in one go, without having to know their full path. Choosing concrete plugins and using full path for them is great for some environmnets, but I don't think it's a common case. The common case is to load all plugins, even ignoring where do they live in the filesystem. Even worse, the full path may be architecture-dependant, which makes copying the ulogd.conf file between machines unnecesarily complex. To experiment this new behaviour, don't put any 'plugin=' directive in the config file. Plugins will be loaded from a default directory, choosen at build/configure time (--with-ulogd2libdir). If no specified, this is something like '/usr/local/lib/ulogd/'. This new configuration option doesn't implement any special logic. We simply open the dir and try to load all files ending with '.so'. The log message level for plugins loading is increased so users can see by default which plugins are loaded. Signed-off-by: Arturo Borrero Gonzalez <arturo@netfilter.org> Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
* ulogd: use a RT scheduler by defaultArturo Borrero Gonzalez2017-09-201-0/+15
| | | | | | | | | | | | | | | | | | | | | | | | | | | | Is common that ulogd runs in scenarios where a lot of packets are to be logged. If there are more packets than ulogd can handle, users can start seing log messages like this: ulogd[556]: We are losing events. Please, consider using the clauses \ `netlink_socket_buffer_size' and `netlink_socket_buffer_maxsize' Which means that Netlink buffer overrun have happened. There are several approaches to prevent this situation: * in the ruleset, limit the amount of packet queued for log * in the ruleset, instruct the kernel to use a queue-threshold * from userspace, increment Netlink buffer sizes * from userspace, configure ulogd to run as high priority process The first 3 method can be configured by users at runtime. This patch deals with the last method. SCHED_RR is configured by default, with no associated configuration parameter for users, since I believe this is common enough, and should produce no harm. A similar approach is used in the conntrackd daemon. Signed-off-by: Arturo Borrero Gonzalez <arturo@netfilter.org> Acked-by: Eric Leblond <eric@regit.org> Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
* ulogd: fix crash when plugin version are incorrectEric Leblond2017-07-021-0/+1
| | | | | Format string in error message had more arguments than given and it was resulting in a crash at start.
* ulogd: use strncpy instead of memcpyEric Leblond2017-03-211-1/+1
| | | | | | | | | On some architecture, ulogd is not starting due to a crash in memcpy. This patch switches to strncpy to avoid the problem. Reported-by: Alexandru Ardelean <ardeleanalex@gmail.com> Signed-off-by: Eric Leblond <eric@regit.org>
* ulogd: add missing newline in log messageEric Leblond2016-02-051-1/+1
|
* ulogd: restructures signal handling by self-pipe trickHironobu Ishii2016-02-051-7/+125
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | ulogd had a critical bug that is calling Async-Signal-Unsafe functions in signal hander context. - Most of libc functions like fopen(), malloc() are Async-Signal-Unsafe. So you should not call these functions in signal handler context. - Calling pluginstances in signal handler context is danger. For implementer of pluginstances, it is very hard to recognize their functions are called in signal handler context. To solve the issue, I restructured signal handling by self-pipe trick. For more detail on self-pipe trick, please see the following. https://lwn.net/Articles/177897/ This patch will solve various symptoms like following. - Deadlock - Segmentation fault caused by libc management data corruption, - Other unpredictable behavior. Deadlock example ================ This bug was already filed at: https://bugzilla.netfilter.org/show_bug.cgi?id=1030 I also hit this bug. The backtrace of this issue is following. In this case, main thread was calling ctime(), and signal handler called localtime_r(). That caused the dead lock while getting tzset_lock in __tz_convert(). Because vsyslog() is Async-Signal-Unsafe function, we cannot call this function in signal handler context. (gdb) bt #0 __lll_lock_wait_private () at ../nptl/sysdeps/unix/sysv/linux/x86_64/lowlevellock.S:95 #1 0x00007f3c3fc7e4ac in _L_lock_2462 () at tzset.c:621 #2 0x00007f3c3fc7e2e7 in __tz_convert (timer=0x7f3c3ff8bf00 <tzset_lock>, timer@entry=0x7fffcfa923b8, use_localtime=use_localtime@entry=1, tp=tp@entry=0x7fffcfa92400) at tzset.c:624 #3 0x00007f3c3fc7c28d in __localtime_r (t=t@entry=0x7fffcfa923b8, tp=tp@entry=0x7fffcfa92400) at localtime.c:32 #4 0x00007f3c3fcbf1ba in __GI___vsyslog_chk (pri=<optimized out>, flag=1, fmt=0x406fa8 "signal received, calling pluginstances\n", ap=0x7fffcfa924a0) at ../misc/syslog.c:199 #5 0x00000000004037b5 in __ulogd_log () #6 0x00000000004047be in signal_handler () #7 <signal handler called> #8 0x00007f3c3fcb62f5 in __GI___xstat (vers=<optimized out>, name=0x7f3c3fd4b2c3 "/etc/localtime", buf=0x7fffcfa92c10) at ../sysdeps/unix/sysv/linux/wordsize-64/xstat.c:37 #9 0x00007f3c3fc7e5f6 in __tzfile_read (file=file@entry=0x7f3c3fd4b2c3 "/etc/localtime", extra=extra@entry=0, extrap=extrap@entry=0x0) at tzfile.c:170 #10 0x00007f3c3fc7d954 in tzset_internal (always=<optimized out>, explicit=explicit@entry=1) at tzset.c:444 #11 0x00007f3c3fc7e303 in __tz_convert (timer=0x7fffcfa92d50, use_localtime=use_localtime@entry=1, tp=tp@entry=0x7f3c3ff8ed80 <_tmbuf>) at tzset.c:629 #12 0x00007f3c3fc7c2a1 in __GI_localtime (t=<optimized out>) at localtime.c:42 #13 0x00007f3c3fc7c1f9 in ctime (t=<optimized out>) at ctime.c:27 #14 0x00007f3c3e180ec2 in ?? () #15 0x0000000056a100c2 in ?? () #16 0xf8570f79d4fc4200 in ?? () #17 0x000000000209bec0 in ?? () #18 0x00007f3c4059f1f8 in ?? () #19 0x000000000000003c in ?? () #20 0x0000000000404952 in ulogd_propagate_results () #21 0x00007f3c3f9cc203 in ?? () #22 0x0000000000000000 in ?? () Segmentation fault in free() ============================ >From my experience, I think this was caused by some routine called malloc()/free() in signal handler context. By that, malloc() management data became inconsistent. As a result, free() made a wrong dereference. Program terminated with signal SIGSEGV, Segmentation fault. #0 __GI___libc_free (mem=0x7f430f011000) at malloc.c:2903 2903 if (chunk_is_mmapped(p)) /* release mmapped memory. */ (gdb) bt #0 __GI___libc_free (mem=0x7f430f011000) at malloc.c:2903 #1 0x00007f430e68affa in __GI__IO_free_backup_area (fp=fp@entry=0x742500) at genops.c:210 #2 0x00007f430e68a795 in _IO_new_file_overflow (f=0x742500, ch=-1) at fileops.c:849 #3 0x00007f430e689511 in _IO_new_file_xsputn (f=0x742500, data=<optimized out>, n=15) at fileops.c:1372 #4 0x00007f430e65aa4d in _IO_vfprintf_internal (s=s@entry=0x742500, format=<optimized out>, format@entry=0x7f430cbc4008 "%.15s %s %s", ap=ap@entry=0x7fff456ece38) at vfprintf.c:1635 #5 0x00007f430e71d615 in ___fprintf_chk (fp=0x742500, flag=flag@entry=1, format=format@entry=0x7f430cbc4008 "%.15s %s %s") at fprintf_chk.c:36 #6 0x00007f430cbc3f04 in fprintf (__fmt=0x7f430cbc4008 "%.15s %s %s", __stream=<optimized out>) at /usr/include/bits/stdio2.h:97 #7 _output_logemu (upi=0x74e5a0) at ulogd_output_LOGEMU.c:102 #8 0x0000000000404952 in ulogd_propagate_results () #9 0x00007f430e40f203 in interp_packet (ldata=0x7fff456ed060, pf_family=2 '\002', upi=0x74a6b0) at ulogd_inppkt_NFLOG.c:400 #10 msg_cb (gh=<optimized out>, nfmsg=0x7f430efe2020, nfa=0x7fff456ed060, data=0x74a6b0) at ulogd_inppkt_NFLOG.c:483 #11 0x00007f430e20a307 in __nflog_rcv_pkt (nlh=<optimized out>, nfa=<optimized out>, data=<optimized out>) at libnetfilter_log.c:160 #12 0x00007f430e0056b7 in __nfnl_handle_msg (len=268, nlh=0x7f430efe2010, h=0x74e8e0) at libnfnetlink.c:1236 #13 nfnl_handle_packet (h=0x74e8e0, buf=0x7f430efe2010 "\f\001", len=<optimized out>) at libnfnetlink.c:1256 #14 0x00007f430e20a508 in nflog_handle_packet (h=<optimized out>, buf=<optimized out>, len=<optimized out>) at libnetfilter_log.c:323 #15 0x00007f430e40eaed in nful_read_cb (fd=<optimized out>, what=<optimized out>, param=0x74a6b0) at ulogd_inppkt_NFLOG.c:463 #16 0x0000000000404ee0 in ulogd_select_main () #17 0x0000000000402b17 in main () Signed-off-by: Hironobu Ishii <ishii.hironobu@jp.fujitsu.com>
* ulogd: Use /dev/null as dummy logfile when logging to syslogFelix Janda2015-06-231-5/+5
| | | | | | | | | Fixes compilation error with musl libc: ulogd.c:86:13: error: storage size of 'syslog_dummy' isn't known static FILE syslog_dummy; Signed-off-by: Felix Janda <felix.janda@posteo.de>
* ulogd: fix loglevel handlingKen-ichirou MATSUZAWA2014-03-071-1/+2
| | | | | | It was always default if not specified by command parameter. Signed-off-by: Ken-ichirou MATSUZAWA <chamas@h4.dion.ne.jp>
* ulogd: avoid potential double print of messageEric Leblond2014-02-041-2/+2
| | | | | In case there is no logfile, ulogd could possibly display each log message twice to stderr.
* ulogd: use daemon() functionEric Leblond2013-10-101-17/+13
| | | | | | | | This patches update the daemonization code. It is done earlier and it uses the daemon(à function which is used for daemonization by most projects. Signed-off-by: Eric Leblond <eric@regit.org>
* Improve pid file handling.Eric Leblond2013-05-211-12/+54
| | | | | | | | | | This patch improves latest patch by splitting in two part the pid file creation. This allows to display a message to stdout when ulogd can not be started. Another linked improvement is that the plugin initialization is not done if the pid file existence will result in a ulogd exit. Signed-off-by: Eric Leblond <eric@regit.org>
* ulogd: Implement PID file writingChris Boot2013-05-211-1/+146
| | | | | | | | The deamon currently does not have the ability to write a PID file to track its process ID. This is very useful to an init script and to ensure there is only one running instance. This patch implements this functionality. Signed-off-by: Chris Boot <bootc@bootc.net>
* ulogd: Perform nice() before giving up rootChris Boot2013-05-211-7/+7
| | | | | | | | The daemon code currently tries to nice(-1) just after having given up root privileges, which fails. This patch moves the nice(-1) call to just before the code that gives up the required privileges. Signed-off-by: Chris Boot <bootc@bootc.net>
* ulogd: display stack during configurationEric Leblond2013-04-201-1/+1
|
* Revert "ulogd: close logfile description in the exit path of parent process"Pablo Neira Ayuso2013-03-271-1/+0
| | | | | | | This reverts commit 3179bd4de89de7c2388849f5bc48e8f5aad9e5b9. Pointing to the wrong place. This is not the file descriptor that ulogd is leaking.
* ulogd: close logfile description in the exit path of parent processPablo Neira Ayuso2013-03-261-0/+1
| | | | | | | | | | | | | | | Joan Touzet reported that file descriptor 3 was not ever closed in the exit path of the parent process: open("ulogd.conf", O_RDONLY) = 3 That corresponds to the the file descriptor that was used to parse the configuration file was not closed. This closes: http://bugzilla.netfilter.org/show_bug.cgi?id=793 Reported-by: Joan Touzet <joant@cloudant.com> Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
* ulogd: change verbosity of a messageEric Leblond2013-03-201-1/+1
| | | | | The "registering plugin" message is not really useful as the message is really explicit if a plugin is missing.
* Add handling of too long line and arguments.Eric Leblond2013-02-181-0/+9
| | | | | When an argument or a line is too long, it can not be store into ulogd configuration and this must results in a error.
* Use access to ensure readability of config gileEric Leblond2013-02-181-1/+1
| | | | | This patch adds a call to access to check the readability of the configuration file.
* Suppress dead FIXME.Eric Leblond2013-02-181-1/+0
|
* Get rid of SVN tag in comment.Eric Leblond2013-01-181-3/+2
| | | | This patch also update some copyright and licence declaration.
* Add -l option to set log level from command lineEric Leblond2013-01-181-2/+13
| | | | | | This patch adds a '-l' option which can be used to setup ulogd loglevel. Command line option has precedence on the configuration file one.
* ulogd: add -v option to display message on stderr.Eric Leblond2013-01-061-8/+33
| | | | | If can be painful to have to check the logfile, so this patch adds a '-v' option which display logs message to stderr.
* src: update copyright and authors informationPablo Neira Ayuso2012-08-031-1/+3
| | | | | | | | Include Eric and myself in the copyright notice and the AUTHORS file since we're the most recurrent contributors (of course, after the original author of this software, Harald Welte). Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
* src: fix version that -V displaysPablo Neira Ayuso2012-08-031-4/+4
| | | | | | | It was wrong, use VERSION constant which uses the version information available in configure.ac. Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
* ulogd: fix segfault if syslog and SIGTERM is receivedSalih Gonullu2011-03-271-2/+2
| | | | Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
* ulogd: fix double call of stop for reused input pluginsPablo Neira Ayuso2011-01-071-2/+12
| | | | | | | | | | | This patch adds reference counting for plugins. This is used to fix a double stop for input plugins that are reused. This problem was reported by Salih Gonullu <sag@open.ch>: http://marc.info/?l=netfilter&m=129439584700693&w=2 Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
* fix wrong list addition in plugin re-use logicPablo Neira Ayuso2010-06-171-1/+1
| | | | | | | | | | This patch fixes a bug that makes ulogd loops forever while propagating inputs to the output plugin. It is reproducible if you re-use three or more plugin instances. The problem is that the parameters in the list addition are in incorrect order. Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
* fix incorrect negative EINTR checking in main loopPablo Neira Ayuso2010-06-131-1/+1
| | | | | | | | | This patch fixes the following error that is displayed if we send SIGHUP to reopen the logfile: ulogd.c:904 select says Interrupted system call Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
* ulogd: configure all plugins before resolving keys.Pierre Chifflier2009-04-171-7/+13
| | | | | | | | | | Split the 'resolve keys' step in two parts: first call the configure fonction for all plugins (in reverse order), then loop again to resolve the keys. This allows dynamic construction of the input and output keys, even for filter plugins. Signed-off-by: Pierre Chifflier <chifflier@inl.fr>
* Display which keys are optional in info mode.Eric Leblond2009-01-271-1/+6
| | | | | This patch adds a "optional" keyword to description of input key which are optional when --info is used to dump information about a plugin.
* Display logfile to check in case of error.Eric Leblond2009-01-131-2/+7
| | | | | This patch adds the display of the used logging file to look at if there is a critical error.
* Add valgrind compilation option.Eric Leblond2008-12-091-0/+4
| | | | | | Valgrind messages are obscur when the plugins are unloaded. This patch adds a macro that can be used to desactivate unloading. To use it, you have to specify 'CPPFLAGS=-DDEBUG_VALGRIND' on configure line.
* Free stacks when exiting.Eric Leblond2008-12-091-0/+12
| | | | This patch modifies ulogd2 to have it free the stacks when leaving.
* Introduce config_stop() functionEric Leblond2008-12-091-0/+2
| | | | | This patch adds the config_stop function which is in charge of releasing ressources allocated for configuration file parsing.
* Unload plugins when quitting.Eric Leblond2008-12-091-1/+20
| | | | | This patch adds unloading of plugins (call dlclose()) in ulogd2. This make valgrind happy and will be useful for daemon live reconfiguration.
* Add SIGINT to list of terminal signal.Eric Leblond2008-12-091-1/+2
| | | | | This patch modifies ulogd to intercept SIGINT signal and quit nicely when this signal is received.
* Call pluginstance stop function when exitingEric Leblond2008-12-091-1/+25
| | | | | | The stop function of plugin was not called when ulogd2 was preparing to quit. This patch adds a call to stop for all plugins in each stack and free pluginstance.
* Treat nice function return.Eric Leblond2008-12-091-1/+7
| | | | | gcc was warning that the return of the nice function should be treated. This patch adds an error message in case of failure.
generated by cgit v1.2.3 (git 2.25.0) at 2025-08-03 08:32:41 +0000