----------------------------------------------------------------------------------------

Scheduled Release 8.2406.0 (aka 2024.06) 2024-06-??

- 2024-04-12: bugfixes related to time_t on 32 bit platforms (y38k isue)

Thanks to Michael Biebl for the patch.

- 2024-04-18: [o|i]mprog/mmexternal bugfix: invalid command line parameter passing

Quoted command line parameters were incorrectly passed to the called program. This

resulted in unexpected behaviour. For example, in

bash -c "echo test"

bash got passed parameters '-c" and '"echo test"'. Note the double quotes in

'"echo test"'. These needed to be removed by Unix/Linux standards but were instead

passed to bash. That in turn resulted in invalid execution. With the fix, the

behaviour is now correct. Now, bash gets passed '-c' and 'echo test'.

This actually is a BREAKING CHANGE. However, it was outright wrong behaviour from

the beginning. We assume that people either never noticed it (because they did not

use quoted parameters) or used some workaroud, likely a "starter script", which

performed the right task. As such, we do expect that only a very limited set of

installations might be affected by the change.

Even more important, we would need to change the default behaviour in any case,

because the previous handling was obviosuly unacceptable. As such, there was no

way to keep rsyslog perform the previous action.

Thanks to Paul Fertser for the patch.

----------------------------------------------------------------------------------------

Scheduled Release 8.2404.0 (aka 2024.04) 2024-04-02

- 2024-04-02: omhttp patches and enhancement

New omhttp plugin configuration parameters added:

* restpathtimeout - configures a timeout value for an omhttp restpath, and retry again

* httpretrycodes (list) - configurable list of HTTP status codes that should be

retried by omhttp plugin. The default behavior is to retry any non 2xx status code.

* httpignorablecodes (list) that can specify HTTP status codes that should be ignored

(as failures). status codes specified as ignorable code will not be retried.

* proxyhost, proxyport - configures a proxy, for which omhttp can use to set up a

HTTP proxy tunnel connection. Also works by reading environment variable if

configured HTTP_PROXY. For more details see libcurl docs on CURLOPT_PROXY

thanks to @erenwh for the contribution!

* retry.addmetadata - when this option is enabled, omhttp will add the

response metadata to: $!omhttp!response. There are 3 response metadata added:

code, body, batch_index.

More statistics omhttp curl statistic counters:

* requests.count

* requests.status.0xx

* requests.status.1xx

* requests.status.2xx

* requests.status.3xx

* requests.status.4xx

* requests.status.5xx

* requests.bytes

* requests.time_ms

No longer discard 3xx, 4xx errors

see also https://github.com/rsyslog/rsyslog/issues/4636

Thanks to github user n2yen for the patches.

- 2024-04-02: remove CAP_IPC_LOCK capability

Does not seem to be necessary.

Thanks to Michael Biebl for the patch.

- 2024-03-28: TLS bugfix: TLS drivers did not properly load under some conditions

This was a regeression from the DTLS implementation

- 2024-03-28: mmdblookup bugfix: do not suspend on incompatible IP lookup

Mmdblookup module used to suspend after erroring on an IPv6 IP lookup on IPv4-only

DBs. The suspension of the module is now replaced by a simple log message,

allowing it to keep working for future lookups.

Thanks to Théo Bertin (frikilax) for the patch.

- 2024-03-28: tarball fix: dist tarball was missing a textbench file

This made the testbench fail.

----------------------------------------------------------------------------------------

Scheduled Release 8.2402.0 (aka 2024.02) 2024-02-27

- 2024-02-26: add DTLS support

This version comes with the initial implementation of imdtls and omdtls.

These modules permit secure message exchange over UDP.

- 2024-02-26: testbench: make omusrmsg-noabort test more reliable

The previous test did not always detect an abort of rsyslog/omusrmsg.

The detection method has now been improved, so it is far more

probable that an abort is detected.

While doing this, we noticed that the omusrmsg-noabort-legacy test was

now a 100% duplicate. There is no need any longer to check pure legacy

syntax, and so that test has been removed.

We also added a valgrind-based test ofr omusrmsg-noabort, which furthers

strengthens bug detection. Most importantly, it helps us to detect

potentially new memory leaks on all CI platforms (in case the lib

behaves differently depending on os/distro).

see also https://github.com/rsyslog/rsyslog/issues/5294

- 2024-02-26: omusrmsg bugfix: potential double free, which can cause segfault

omusrmsg frees a string which points to OS/system library memory. When

the os/libs clean up, it frees the memory as well. This results in a

double free. This bug interestingly seems to go unnoticed in many cases.

But it can cause a segfault or hard-to-trace memory corruptions which

could lead to other problems later on. The outcome of this bug most

probably depdns on os/library versions.

closes https://github.com/rsyslog/rsyslog/issues/5294

- 2024-02-26: ommysql bugfix: potential segfault on database error

Due to an invalid code path, ommysql may cause a segfault if database

transactions fail into a specific way. The main trigger is a totally

irrecoverrable database error which can lead to premature connection

close, which is not checked for in all recover code.

This was detected in a setting where a stored procedure is called that

rolls back a transaction in itself.

This patch fixes the issue.

closes https://github.com/rsyslog/rsyslog/issues/5288

- 2024-02-26: omfile: do not carry out actual action when writing to /dev/null

In some use cases omfile is configured to write to /dev/null. This seems

primarily be done because of statistics gathering but maybe some other

scenarios. We now add conditional logic to not do any actual omfile

action when the target file is /dev/null.

Note: this check only works on static file names. When /dev/null is

evaluated as part of dynafile, it will be handled just in the regular

case like before this patch.

----------------------------------------------------------------------------------------

Scheduled Release 8.2312.0 (aka 2023.12) 2023-12-12

- 2023-12-11: imjournal: Add new input module parameter 'defaulttag'

The DefaultTag option specifies the default value for the tag field.

In imjournal, this can happen when one of the following is missing:

* identifier string provided by the application (SYSLOG_IDENTIFIER)

* name of the process the journal entry originates from (_COMM)

Thanks to Attila Lakatos for the patch.

- 2023-12-08: core bugfix: rsyslog messages may not always have FQDN

Even if hostname FQDN is configured, rsyslog internal messages generated

after rsyslog startup and before the first HUP will not necessarily have

FQDN but instead only the shortname of the local host. This commit

fixes the situation.

Special thanks to github user eciii for doing a great bug analysis

and helping us considerably to fix the issue.

closes https://github.com/rsyslog/rsyslog/issues/5218

- 2023-12-08: omlibdbi regression fix: database path was not properly used

Commit 4a072d6c93015a63716c49a6c7756df22750086a caused a regression that made

the database path unreliable to use. Depending on platform/libc version the

basename was improperly extracted, which made access to the database of sqllite

impossible.

Thanks to Flávio Tapajós for the patch.

closes: https://github.com/rsyslog/rsyslog/issues/5282

- 2023-12-06: mazureeventhubs: Corrected handling of transport closed failures

- Added test for connection interrupts (requires root)

- Corrected handling of PN_TRANSPORT_CLOSED.

- Make sure Connection is being reestablished trough tryResume

- Enhanced Debug log output

closes: https://github.com/rsyslog/rsyslog/issues/5269

- 2023-11-24: imkmsg: add params "readMode" and "expectedBootCompleteSeconds"

These parameters permit to control when imkmsg reads the full

kernel log upon startup.

Parameter "readMode" provides the following options:

* full-boot - (default) read full klog, but only "immediately" after

boot. "Immediately" is hereby meant in seconds of system

uptime given in "expectedBootCompleteSeconds"

* full-always - read full klog on every rsyslog startup. Most

probably causes messag duplication

* new-only - never emit existing kernel log message, read only

new ones.

Note that some message loss can happen if rsyslog is stopped

in "full-boot" and "new-only" read mode. The longer rsyslog is

inactive, the higher the message loss probability and potential

number of messages lost. For typical restart scenarios, this

should be minimal. On HUP, no message loss occurs as rsyslog

is not actually stopped.

The default value for "expectedBootCompleteSeconds" is 90.

see also https://github.com/rsyslog/rsyslog/issues/5161

- 2023-11-10: imkmsg: add module param parseKernelTimestamp

The parameter permits to select whether or not and when kernel

timestamps shall parsed, that is be used as the actual time a

log message occurs.

This permits to work around problems with the way kernel

timestamps are represented. The reasoning is given in a sysklogd

commit by Joachim Wiberg, which we reproduce below ("QUOTE") to

have a stable reference.

The commit itself can be found for example at:

https://github.com/troglobit/sysklogd/commit/9f6fbb3301e571d8af95f8d771469291384e9e95

The new parameter parseKernelTimestamp has three possible modes:

"startup" - uses the kernel time stamp during the initial read

loop of /dev/kmsg, but replaced it later ignores it for later reads.

This is the DEFAULT setting.

"on" - kernel timestamps are always used and no correction is tried

"off" - kernel timestamps are never used, system time is used instead

Note that there this is a slightly breaking change. Previously, imkmsg

reported similar to "off" mode, now it reports by default in "startup"

mode. We consider this acceptable, as "off" mode timestamps are not

correct for startup. After startup, the behaviour is correct. All in

all, the new default is kind of a bugfix.

============== QUOTE ===============

The spec[1] says the /dev/kmsg timestamp is a monotonic clock and in

microseconds. After a while you realize it's also relative to the boot

of the system, that fact was probably too obvious to be put in the spec.

However, what's *not* in the spec, and what takes a while to realize, is

that this monotonic time is *not* adjusted for suspend/resume cycles ...

On a frequently used laptop this can manifest itself as follows. The

kernel is stuck on Nov 15, and for the life of me I cannot find any to

adjust for this offset:

$ dmesg -T |tail -1; date

[Mon Nov 15 01:42:08 2021] wlan0: Limiting TX power to 23 (23 - 0) dBm as advertised by 18:e8:29:55:b0:62

Tue 23 Nov 2021 05:20:53 PM CET

Hence this patch. After initial "emptying" of /dev/kmsg when syslogd

starts up, we raise a flag (denoting done with backlog), and after this

point we ignore the kernel's idea of time and replace it with the actual

time we have now, the same that userspace messages are logged with.

Sure, there will be occasions where there's a LOT of kernel messages to

read and we won't be able to keep track. Yet, this patch is better than

the current state (where we log Nov 15).

[1]: https://www.kernel.org/doc/Documentation/ABI/testing/dev-kmsg

===========END QUOTE ===============

closes https://github.com/rsyslog/rsyslog/issues/4561

closes https://github.com/rsyslog/rsyslog/issues/5161

- 2023-11-07: imfile bugfix: remove state file on file delete

The state file would remain in the working directory

after shutdown, even though deleteStateOnfileDelete is

set to "on" and the monitored file was removed.

closes https://github.com/rsyslog/rsyslog/issues/5258

Thanks to Attila Lakatos for the patch.

- 2023-10-31: TLS subsystem: fix small memory leak on startup

This was a one-time leak of the file name that hapened if a certificate file

was not accessible. It had no operational issues, but could confuse automatted

testing. As not only a side-effect, certificate load failures are now somewhat

more verbosely reported, which we consider helpful to the user.

Thanks to Attila Lakatos for the patch.

- 2023-10-31: imklog bugfix: keepKernelTimestamp=off config param did not work

... at least not as expected. It was only honored for kernel-level

messages and only when parseKernelTimestamp was "on". Otherwise, the

kernel timestamp was always kept inside the message.

closes https://github.com/rsyslog/rsyslog/issues/5160

- 2023-10-26: TLS subsystem: add remote hostname to error reporting

This provides richer and easier to process logs for error and warning

cases. One goal is to enable automatic operations without the need

to consolidate multiple message to a single information.

This improves one situation in gtls driver and provides a more

generic approach in ossl driver for OpenSSL error reporting.

There is probably still room for improvement, however this patch

is at least a good starting point for further work. Please

provide feedback if you need more!

closes https://github.com/rsyslog/rsyslog/issues/5244

- 2023-10-24: imjournal: add the ability to run multiple journal inputs

This may be useful to de-couple journal processing.

Thanks to Willy Tu for the patch.

- 2023-10-24: regression fix: forking rsyslogd on BSD did not work

Actually, this was an issue for all platforms that do not provide open file handle

detection via the /proc file system.

Tech details: After fork if the child process uses close_range to close open file

descriptors it has no way to exempt the parentPipeFD causing a failure to signal

successful startup to the parent process. This causes failures on all systems that

aren't Linux that implement close_range.

Thanks to Nathan Huff for the patch.

- 2023-10-24: omusrmsg: use logind instead of utmp for wall messages with systemd

Future SUSE versions will get rid of utmp due to a 32bit time_t counter

overflow in 2038.

See details at:

https://github.com/thkukuk/utmpx/blob/main/Y2038.md

On systemd based systems logind is an alternative to utmp.

Thanks to github user tblume for the patch.

- 2023-10-24: cleanup: rm no longer used --with-systemdsystemunitdir configure switch

This is a clean up following the removal of the service unit in

cfd07503ba055100a84d75d1a78a5c6cceb9fdab

- 2023-10-23: testbench: bump zookeeper version to match current offering

Older version can no longer be downloaded. It also makes sense to

test with mainstream version.

----------------------------------------------------------------------------------------

Scheduled Release 8.2310.0 (aka 2023.10) 2023-10-10

- 2023-10-04: Add CAP_NET_RAW capability due to the omudpspoof module

The CAP_NET_RAW ensures the use of RAW and PACKET sockets,

which is utilized by the omudpspoof module, more precisely

the libnet_init function.

Thanks to Attila Lakatos for the patch.

- 2023-10-04: Add new global config option "libcapng.enable"

Defines whether rsyslog should drop capabilities at startup or not.

By default, it is set to "on". Until this point, if the project was

compiled with --enable-libcap-ng option, capabilities were

automatically dropped. This is configurable now.

Thanks to Attila Lakatos for the patch.

- 2023-10-04: tcp net subsystem: handle data race gracefully

It may happen that a socket file descriptor has been closed either

while setting up poll() et al or while being inside the system call.

This was previously treated as error and caused abort in debug

builds. However, it was essentially ignored in production builds.

This has now been fixed and now is always gracefully ignored. This

most importantly fixes some flakes in CI runs (which were caused

by this situation).

- 2023-09-29: imrelp bufgifx: avoid crash on restart in imrelp SIGTTIN handler

While existing, if at specific time rsyslog receives a SIGTTIN, it

crashes due to 2 issues.

1. debug.unloadModules="off" a double free of pRelpEngine

2. debug.unloadModules="on" it crashes because the signal handler has

been unmapped from memory.

This patch covers both issues.

Thanks to Ali Abdallah for the patch.

- 2023-09-28: fix startup issue on modern systemd systems

When we startup AND are told to auto-background ourselfs, we must

close all unneeded file descriptors. Not doing this has some

security implications. Traditionally, we do this by iterating

over all possible file descriptor values. This is fairly compatible,

because we need no OS-specific method. However, modern systemd configs

tend to not limit the number of fds, so there are potentially 2^30(*)

fds to close. While this is OKish, it takes some time and makes

systemd think that rsyslog did not properly start up.

We have now solved this by using the /proc filesystem to obtain our

currently open fds. This works for Linux, as well as Cygwin, NetBSD,

FreeBDS and MacOS. Where not available,and close_range() is available

on the (build) platform, we try to use it. If that fails as well, we

fall back to the traditional method. In our opionion, this fallback

is unproblematic, as on these platforms there is no systemd and in

almost all cases a decent number of fds to close.

Very special thanks go out to Brennan Kinney, who clearly described

the issue to us on github and also provided ample ways to solve it.

What we did is just implement what we think is the best fit from

rsyslog's PoV.

(*) Some details below on the number of potentially to close fds.

This is directly from a github posting from Brennan Kinney.

Just to clarify, by default since systemd v240 (2018Q4), that

should be `1024:524288` limit. As in the soft limit is the expected

`1024`.

The problem is other software shipping misconfiguration in systemd

services that overrides this to something silly like

`LimitNOFILE=infinity`.

- Which will map to the sysctl `fs.nr_open` (_a value systemd

v240 also raises from `2^20` to 2^30`, some distro like Debian are

known to opt-out via patch for the `fs.nr_open` change_).

- With the biggest issue there being that the soft limit was also

set to `infinity` instead of their software requesting to raise

the soft limit to a higher value that the hard limit permits.

`infinity` isn't at all sane though.

- The known source of this misconfiguration is container software such

as Docker and `containerd` (_which would often sync with the

systemd `.service` config from the Docker daemon `dockerd.service`_).

closes https://github.com/rsyslog/rsyslog/issues/5158

- 2023-09-13: Add the 'batchsize' parameter to imhiredis

Parameter set to allow configuring the amount of entries imhiredis debatches at once.

Default value of '10' has been kept to avoid any side effect on existing

configurations.

Thanks to Jérémie Jourdin for the patch.

- 2023-09-13: omprog bugfix: Add CAP_DAC_OVERRIDE to the bounding set

The omprog module uses the execve() function to execute

a third party program. Some required capabilities were not

preserved in the bounding set [1]. This caused problems, e.g.

the program could not write to files even if rsyslog was

executed as root and privileges were not dropped. As of now,

only the CAP_DAC_OVERRIDE capability is added to the bounding

set. Others could be added later, if there is justification

behind that.

[1] The capability bounding set is a security mechanism that

can be used to limit the capabilities that can be gained

during an execve(2). During an execve, the capability

bounding set is ANDed with the file permitted capability

set, and the result of this operation is assigned to the

thread's permitted capability set. The capability

bounding set thus places a limit on the permitted

capabilities that may be granted by an executable file.

Thanks to Attila Lakatos for the patch.

- 2023-09-13: tcpflood bugfix: plain tcp send error not properly reported

The error code when plain tcp sending failed was improperly returned,

resulting in no meaningful error message.

Note: tcpflood is a testbench tool, not part of production rsyslog.

----------------------------------------------------------------------------------------

Scheduled Release 8.2308.0 (aka 2023.08) 2023-08-15

- 2023-08-07: crypto subsystem bugfix: potential undefined behaviour

The is some potential undefined behaviour when initializting the IV for locally

encrypting log files. The issue cancels itself out, but at least causes

some confusion when using undefined behaviour sanitizer (UBSAN). However,

UBSAN seems not to detect the issue on all platforms and/or in all versions

(we were not able to reproduce this issue in our CI).

Please also note that the functionality where this can happen is extremely

rarely being used.

Thanks to Jeffrey Walton for providing the patch.

- 2023-08-02: lookup tables: fix static analyzer issue

If something goes really wrong, a lookup table's name would not

be set. That could lead to a NULL pointer access. HOWEVER, this

would require serious bugs in config parameter parsing, as the

lookup table name is a required parameter and the parser will

error out if not set.

So the bug is mostly cosmetic - but it does not hurt to handle

this case, of course.

- 2023-08-02: lookup tables bugfix: reload on HUP did not work when backgrounded

Lookup tables were only reloaded on HUP if the -n option was given

and rsyslog no backgrounded. This patch fixes the issue.

closes: https://github.com/rsyslog/rsyslog/issues/4813

- 2023-07-30: testbench: make test more reliable

There was a race between tcpflood and rsyslog in imptpc_maxsessions.sh.

We now use the new -A tcpflood option to make the timing more

predictable, hopefully fixing test flakiness.

Note: if that does not help, we need to introduce a wait on the number

of error messages and maybe a delay before tcpflood termination. The

theory behind the latter is that rsyslog possibly does not fully

iniaitlize session which are quickly aborted before rsyslog receives

the related OS notification! We just record this info in case we

need it and are positive that this change will fix the situation.

- 2023-07-28: openssl: make connection setup more reliable by use of newer lib feature

Replaced depreceated method SSLv23_method with TLS_method.

In OpenSSL 1.1.0 and higher, SSLv23_method causes some errors

in TLS handshake from time to time. As this method is depreceated

since 1.1.0, I have replaced it with the follow up method

TLS_method which is the most generic one.

It fixes the random test failures in tests like

- sndrcv_tls_ossl_anon_rebind.sh

Also added some debug output in OpenSSL error handling, which is

useful when analysing debug files.

closes: https://github.com/rsyslog/rsyslog/issues/5201

- 2023-07-28: testbench improvement: define state file directories for imfile tests

Not all imfile tests have state file directories or a global working

directory defined. This results in usage of the default location.

While state file names should be sufficiently different, there is still

some riks of using the same name in different tests. That becomes

problematic if tests are run in parallel (and they are run in

parallel inside the regular CI).

NOTE: NOT YET COMPLETED FOR ALL TESTS! We are considering if it makes

sense to deliberately keep some as-is.

- 2023-07-28: tcpflood bugfix: TCP sending was not implemented properly

Note: tcpflood is a testbench tool. This bug could lead to testbench

false positives. No way it can affect production deployments.

The tcpflood tool did improperly assume that a TCP sendto() call

would send messages of any size in a single shot. This is not the

case. It has now been corrected to proper behavior.

As a side-activity, some int variables which acutally needed to be

size_t have been fixed as well.

- 2023-07-28: testbench: make waiting for HUP processing more reliable

The previous approach was more or less delay based. We have now

changed the code to enable imdiag to detect if HUP is underway

and wait until it is completed. The new method still employs some

kind of timeout, but is now quite reliable. Most importantly,

it works great with long-running HUP processing, which can happen

e.g. when querying the system name takes long or some actions need

longer time to persist their HUP processing.

The new approach will most likely reduce CI flakes and also speed

up testbench runs. The speedup happens from not having to wait a

full delay in cases where we detect HUP is completed (plus reduced

timeout when we cannot clearly detect this - see code comments why

the new method is still considered more reliable than the old one).

Code note: we needed to slightly re-structure the way actual HUP

processing and the "HUP mutex" is handled. After best analysis,

this does not affect the reliability or speed in production

settings.

closes https://github.com/rsyslog/rsyslog/issues/5192

- 2023-07-27: build system: make rsyslogd execute when --disable-inet is configured

This option is mostly useless, as network functionality depends on the

modules loaded by the config. The only real, and important, effect it

has is to control auto-load of omfwd - a feature almost all installations

depend in (backward compatibility).

This has been clarified in ./configure -help

Also, when --disable-inet is given, rsyslog now executes successfully.

The reason for the abort was that previously building of the lmnet

component was prevented, but that component is also needed by rsyslog

startup itself to query its own (correct) hostname.

Note that --disable-inet still does not compile some networking

libraries. So do not use it if you intend to load standard networking

modules like omfwd, imtcp or imudp.

closes https://github.com/rsyslog/rsyslog/issues/5188

- 2023-07-26: testbench/CI: update zookeper download to newer version

Old version is no longer available.

- 2023-07-24: openssl: add support for new-version init function

- 2023-07-07: add CRL support for network (TLS) drivers

Thanks to Darren J Moffat for implementing the OpenSSL part.

- 2023-07-07: omazureeventhubs: Initial implementation of new output module

The output module uses Apache "Qpid Proton C API" which is a solid

AMQP protocol library implementation that can be integrated

very well into the rsyslog dev environment.

- Implemented Delivery with submitted and accepted state checking

- saving of failed messages in a failed list with support of saving

and restoring.

- Add testcases (requires ENV variables) to testbench

- Using application/octect-stream (binary) to send messages based on

Microsoft Code Sample:

https://docs.microsoft.com/en-us/azure/event-hubs/event-hubs-c-getstarted-send

* Note original Microsoft Samplecode is not working anymore, we are using

* QPID Proton Proactor based on

https://github.com/apache/qpid-proton/blob/main/c/examples/send.c

- requires QPID-PROTON Version 0.13 or higher because of the proactor API

- Add EventProperties configuration parameters

- Slow down when sender credit reaches zero (10ns).

- Add support for static library linking of qpid-proton

This is needed to build the module from source and remove

library package dependencies.

- adjusted valgrind suppressions

- 2023-07-04: core bugfix: action.resumeintervalmax parameter was not respected

Unfortunately, defining action.resumeintervalmax in the configration

did not have any effect at all. Instead, the default value was used,

which is 1800. This was caused by not having all the letters in

lower-case.

Fixes https://github.com/rsyslog/rsyslog/issues/5132

Thanks to Attila Lakatos for the patch.

- 2023-06-29: core bugfix: do not try to drop capabilities when we don't have any

In case the process does not have any capabilities, e.g. running as regular user then

we do not have to force capability dropping. The capng_have_capabilities() returns

none if that's the case.

Fixes https://github.com/rsyslog/rsyslog/issues/5091

Thanks to Attila Lakatos for the patch.

- 2023-06-29: imhiredis bugfix: Restore compatiblity with hiredis < v1.0.0

RESP3 protocol wasn't implemented yet, some types weren't

available (REDIS_REPLY_DOUBLE)

Thanks to Théo Bertin (frikilax) for the patch.

- 2023-06-23: testbench: use newer zookeeper version in tests

- 2023-06-23: build system: more precise error message on too-old lib

When libcap-ng was enabled, the lib was present but did not meet the minimum version

dependency during configure, it was reported as "missing". We now emit a message

telling that it is present, but the version too old.

----------------------------------------------------------------------------------------

Scheduled Release 8.2306.0 (aka 2023.06) 2023-06-20

- 2023-06-19: mmnormalize bugfix: if msg cannot be parsed, parser chain is stopped

When an parser is not able to parse a message, it should indicate this

to rsyslog core, which then activates the next parser(s) inside the

configured parser chain.

Unfortunatley, mmnormalize always tells core "success", and so no

other parsers are activated.

closes https://github.com/rsyslog/rsyslog/issues/5148

- 2023-06-19: [i/o]mhiredis: various fixes and enhancements

please see the change log for details. Among others, suspending of the modules

has been fixed. Also a new "stream" mode has been added.

Thanks to Théo Bertin (frikilax) for the patch.

- 2023-06-19: testbench/bug: mmexternal-SegFault-empty-jroot-vg.sh fails due to typo

Fix the typo that makes the test fail.

Thanks to Paul Fertser for the patch.

- 2023-06-16: imjournal: Add FileCreateMode module parameter

FileCreateMode allows to set the default file mode bits

when creating new files. As of now, it has only impact on the state file.

Add test suite as well.

Minor indentation fix in run_journal.yml

Thanks to Attila Lakatos for the patch.

- 2023-06-16: core bugfix: potential segfault on busy systems

This was discovered by Konstantin J. Chernov in a practicaly deployment.

Here, msg object tag processing caused sporadic segfaults. We did not

hear from similiar cases, but there clearly is potential for problems

because a mutex lock had insufficient range, thus leading to a potential

race.

The patch is directly from Konstantin J. Chernov, thanks for that.

Please note that the mutex lock could be minimized as it is not strictly

needed for the pM == NULL case, but this cause is extremely exotic

and the resulting code would be harder to understand. Thus we opt

to do the locking on funtion level (as usual).

Descriptiond edited by Rainer Gerhards

closes: https://github.com/rsyslog/rsyslog/issues/5110

- 2023-06-16: Add new global config option "libcapng.default"

Defines how rsyslog should behave in case something went wrong

when capabilities were to be dropped. Default value is "on",

in which case rsyslog exits on a libcapng related error.

Thanks to Attila Lakatos for the patch.

Closes https://github.com/rsyslog/rsyslog/issues/5096

- 2023-06-05: imfile bugfix: file handle leak, primarily in kubernetes context

At this point there is a code imfile.c#L919 that adds an inotify observer to the

parent of the symbolic link target. But there is no such code that removes this

observer in the case when inotify events do not occur in the directory tree above.

This may be if the directory tree of the symbolic link target and the directory tree

of the symbolic link itself are divided into different subtrees somewhere at the levels

above.

For example, in the rsyslog configuration, an imfile with the

template /var/log/containers/*.log is configured and there is the following directory

tree:

/var/log/pods/pod-1/a/0.log

/var/log/containers/pod-1-a-0.log -> /var/log/pods/pod-1/a/0.log

In this example, kubernetes cron jobs will permanently delete directories at the

/var/log/pods/pod-* level. And thus, inotify observer on the parent object of the

symbolic link target (/var/log/pods/pod-1/a/0.log) looking at the directory

/var/log/pods/pod-1/a will constantly leak.

This is due to the fact that the list of active objects in the edge with path

/var/log/containers, where the parent object of the target symbolic link is added,

is not checked. Verification and deletion will occur only in the case of an inotify

event in the upper nodes of the directory tree, in /var/log and above.

Thanks to Sergey Kacheev for the patch!

- 2023-06-05: GNUTls Driver: Fix memory leaks in gtlsInitCred

Missing CA Certificate or multiple Connections caused

a memory leak in pThis->xcred as it was allocated each time in

gtlsInitCred by gnutls_certificate_allocate_credentials

closes: https://github.com/rsyslog/rsyslog/issues/5135

- 2023-05-24: CI: update base ubuntu image for github actions

----------------------------------------------------------------------------------------

Scheduled Release 8.2304.0 (aka 2023.04) 2023-04-18

- 2023-04-17: imptcp bugfix: spam log on oversize message

If an oversize message was received by imptcp, imptcp reported

one error message for EACH oversize character. This could

result in a potentially very large number of similar (and

useless) messages.

This is a regression from commit f052717178.

closes https://github.com/rsyslog/rsyslog/issues/5078

- 2023-04-17: core/bugfix: using $uuid msg prop can deadlock rsyslog on shutdown

This problem can occur if a large number of threads is used and rsyslog

cannot shut down all queues etc within the regular time interval. In this

case, it cancels some threads. That can leave the mutex guarding libuuid

calls locked and thus prevents other, not yet cancelled threads from

progressing. Assuming pthread_mutex_lock() is not a cancellation point,

this will case these other threads to hang forever and thus create a

deadlock situation.

closes https://github.com/rsyslog/rsyslog/issues/5104

- 2023-04-17: Do not preserve capabilities when changing credentials

In configurations where $PrivDropToGroup or $PrivDropToUser are used,

rsyslogd changes uid/gid to a non-privileged user. As part of that

change, all capabilities should be lost. However, if rsyslog is

compiled with --enable-libcap-ng option, some capabilities are

preserved due to using capng_change_id() instead of setgid()and

setuid(). https://linux.die.net/man/3/capng_change_id:

This function preserves capabilities while changing uid/gid, causing

rsyslogd to run as non-root user, but with some root capabilities.

Unfortunately, rsyslogd will run with higher privileges than before.

The patch also removes CAP_SETPCAP, because the capability set does

not need to be altered at a later phase.

Thanks to Attila Lakatos for the patch.

----------------------------------------------------------------------------------------

Scheduled Release 8.2302.0 (aka 2023.02) 2023-02-21

- 2023-01-27: core/template: implement negative position.to

This will easily permit to drop the last n characters from a property

without the need to know the exact length of the string. This is

especially useful as the exact length is most often not known

beforehand.

- 2023-01-18: Introduce --enable-libcap-ng configure option

The option allows to drop the capabilities to only

the necessary set, to minimize security exposure in

case there was ever a mistake in a networking

plugin or some other input resource. Moreover, it adds

ability to change uid and gid while retaining the

previously specified capabilities.

Add ability to change uid and gid while retaining the

capabilities previously specified.

closes https://github.com/rsyslog/rsyslog/issues/4986

Thanks to Attila Lakatos for the patch.

- 2023-01-16:

- omfile: add action parameters "rotation.*"

Add new action parameters

- rotation.sizeLimit

- rotation.sizeLimitCommand

provide automatic output file rotation functionality feature-wise

equivalent to legacy $outchannel. This finally permits to use

this feature set in rscript.

- core substring function: enhancement and hardening

Now, length can have a negative value -n to denote that the

substring should be build between startpos and the character

-n chars from the end. This is a shortcut for stripping charactes

on "both ends" of the string. See doc for details on the enhanced

semantics.

Also, some hardening against invalid startpos and length has

been added.

- core bugfix: wrong type conversion in internal string class could lead to segfault

This could only happen with very unusually large strings

Thanks to Flos Lonicerae for the patch.

- QA: changed to CodeQL scanning on github as LGTM replacement

- bugfix: wrong version number on daily stable builds

- CI: use newer version of zookeeper (needed modernization)

- ffaup bugfix : memory corruption with concurrent workers

The ffaup function fails to work properly when it is used with multiple workers.

The faup_handler_t struct is not supposed to be shared between threads.

This may have caused memory corruptions and race conditions when used

inside of actions.

Thanks to Thibaud Cartegnie for the fix.

- openssl bugfix: undefined reference error on OpenSSL 1.1 or higher.

This could have prevented ossl components from being loaded/used.

- 2023-01-02: core bugfix: template system may generate invalid json

When

- a list template

- is created with option.jsonf="on"

- and the last list element is a property with onEmpty="skip"

- and that property is actually empty

invalid JSON is generated.

The JSON string in this case ends with ", " instead of "}\n". This

patch fixes the issue.

closes https://github.com/rsyslog/rsyslog/issues/5050

----------------------------------------------------------------------------------------

Scheduled Release 8.2212.0 (aka 2022.12) 2022-12-06

- 2022-12-05: testbench: make python http server based tests more reliable

Harden them against races during server port assignment. Prevents

testbench flakes.

- 2022-12-05: omprog bugfix: invalid status handling at called program startup

There is a bug when external program *startup* does not return "OK". This

can also lead to a misadressing with potentially a segfault (very unlikely).

Note that no problem exists once the initializiation phase of the external

program is finished and regular message transfer runs.

The problem basically is that for a startup failure, the control data for

that external program instance is freed on error. Unfortunately, that state

data is needed later on to detect a suspended instance. We now keep the control

data even on init failure (as we then need to do normal control options).

closes https://github.com/rsyslog/rsyslog/issues/4967

- 2022-11-29: testbench bugfix: wrong message injection object of instance 1

In some client-server test cases, messages are supposed to be injected into

the instance 2(client), but they are actually injected into instance 1(server),

which may lead to false negative results. This patch fixed it by replacing

'injectmsg' with 'injectmsg2', and dealt with some minor issues.

Thanks to Guodong Zhu for the patch.

- 2022-11-21: rsyslog.conf man page bugfix: description of selectors

Document historic difference to BSD syslog selectors.

- 2022-11-18: imtcp bugfix: legacy config directives did no longer work

Many "$InputTCPServer..." config directives did no longer work

and were completely ignored (e.g. "$InputTCPServerStreamDriverMode").

This was a regression from a08591be5d9 (May, 5th 2021).

closes https://github.com/rsyslog/rsyslog/issues/5021

- 2022-11-16: ksi bugfix: sending of too many signing requests fixed.

As there is a bug in libksi where too many signing requests may have bene sent

out the amount of signing requests will be limited by KSI module until the fix

is implemented.

Thanks to Taavi Valjaots for the patch.

- 2022-11-14: bugfix: prevent potential segfault when switchung to queue emergency mode

When switching to Disk queue emergency mode, we destructed the in-memory

queue object. Practice has shown that this MAY cause races during

destruction which themselfs can lead to segfault. For that reason, we

now keep the disk queueu object. This will keep some ressources,

including disk space, allocated. But we prefer that over a segfault.

After all, it only happens after a serious queue error when we are

already at the edge of hard problems.

see also: https://github.com/rsyslog/rsyslog/issues/4963

- 2022-11-08: ksi bugfix: Segmentation fault in async mode fixed

Thanks to Taavi Valjaots for the patch.

- 2022-11-02: imjournal: add second fallback to _COMM

If SYSLOG_IDENTIFIER is not present in the journal message,

then lookup the _COMM field, which stands for the name

of the process the journal entry originates from. This is

needed in order to be in compliance with the journalctl

output.

Thanks to Attila Lakatos for the patch.

- 2022-10-25: core bugfix: local hostname invalid if no global() config object given

The local hostname is invalidly set to "[localhost]" on rsyslog startup

if no global() config object is present in rsyslog.conf. Sending a HUP

corrects the hostname.

This is a regression from ba00a9f25293f

closes https://github.com/rsyslog/rsyslog/issues/4975

closes https://github.com/rsyslog/rsyslog/issues/4825

- 2022-10-25: testbench bugfix: fixed timing issue that sometimes lead to test failure

Timing caused a race in test tool sync and could lead to premature termination of

tools, which in turn caused test failure

----------------------------------------------------------------------------------------

Scheduled Release 8.2210.0 (aka 2022.10) 2022-10-18

- 2022-10-13: fix NetBSD build issue

On NetBSD, time_t has for a long time now been __int64_t.

On 32-bit CPUs, the compiler is not obliged to define

__sync_bool_compare_and_swap_8, so instead this ends up

as an undefined symbol when linking rsyslog. This makes

the code fall back to the pthread / locking method on these

systems, but at least lets the program build.

Thanks to Havard Eidnes for the patch.

- 2022-10-12: omrabbitmq: Add TLS support

Thanks to github user 21stcavenan for the patch.

- 2022-09-14: config: add "abortOnFailedQueueStartup" global config parameter

similiar to "abortONUncleanConfig", this parameter aborts rsyslog

when a queue has problems during startup. Some users perfer rsyslog

to terminate in this case. By default, nothing changes.

closes https://github.com/rsyslog/rsyslog/issues/4902

- 2022-09-07: cor bugfix: leak in helper function SetString

A part of rsyslog runtime, SetString(), had a small memory leak when a value was

assigned multiple times. While this could potentially consume larger amounts of

memory, this did not happen in practice. The reason is that multiple assignments

to the same object occur very seldom.

Thanks to github user seuzw930 for the patch.

closes: https://github.com/rsyslog/rsyslog/issues/4961

- 2022-09-07: core bugfix: correct local host name after config processing

rsyslog.conf may affect the host's local name. These changes were

so far only activated after the first HUP. This patch now ensures

that the configured local host name is applied correctly throughout

all processing, including early startup.

This patch causes a slight change of behaviour. However, the behaviour

was inconsitent before. Now it is consistent and according to the config.

Please note: this patch also exposes a global entry point via "regular"

dynamic loading as this makes things much easier to do. This is in-line

with ongoing simplification effort.

Finally, we also remove a CI test that we do no longer need because

the problem covered is now addressed differently and the original issue

can no longer occur.

closes https://github.com/rsyslog/rsyslog/issues/4975

- 2022-08-31: imtcp: add option notifyonconnectionopen

Add this both as module an input parameter. Complements already-existing

config param notifyonconnectionclose and mirrors the similar feature from

imptcp.

The module parameter acts as default, similarly to notifyonconnectionclose.

Note that in contrast to imptcp, we emit IP addresses and not host

names. This sticks with the traditional semantics of imtcp.

Note that we also fixed a mislading error message in the case when a

disallowed sender tried to connect.

Thanks to John Chivian for suggesting the addition.

- 2022-08-26: openssl TLS driver: add mechanism to include extra CA files parameter

This change allows to include extra CA files so that no "unable to get issuer

certificates" issue is obtained when using chained cert files. New parameter name is

"NetstreamDriverCAExtraFiles".

Thanks to Sergio Arroutbi for the patch.

closes: https://github.com/rsyslog/rsyslog/issues/4851

- 2022-08-19: fix compile issue with older gcc compilers

Thanks to Julien Thomas for the contribution.

----------------------------------------------------------------------------------------

Scheduled Release 8.2208.0 (aka 2022.08) 2022-08-09

- 2022-08-09: ksi bugfix: request cache size and send timeout issue fixed.

Async service send timeout is not configurable and request cache size is too

small to handle large amount of signing requests with small amount of permitted

requests per aggregation round. For example user with max_requests = 4 results

cache size 5 * max_requests or at least 256. When signing 300 log files cache

will be too small resulting several unsigned blocks. When signing 200 log file

cache will be adequate, but with rate of 4 signatures per second, it is only

possible to sign 4 * 10 blocks before all requests that are not sent out will

timeout.

Fix for the issue is to make send timeout configurable and make the size of the

cache depend on the value of send timeout. New configuration value

sig.block.signtimeout="time, s" introduced that defines the time window wherein

the block has to be signed. The size of the request cache is increased to

3 * max_requests * sign_timeout or at least 256.

Thanks to Taavi Valjaots for the patch.

- 2022-08-09: imjournal bugfix: segmentation fault in close journal

Thanks to github user t-feng for the patch.

- 2022-08-09: net subsystem: support sha256 for StreamDriverAuthMode="x509/fingerprint"

Thanks to github user codemaker219 for the patch.

- 2022-08-05: imfile bugfix: message loss/duplication when monitored file is rotated

When a to-be-monitored file is being rotated, some messages may be lost or

duplicated. In case of duplication, many file lines may be duplicated

depending on actual timing. The whole bug was primarily timing depenedent

in general. It most often was visible in practice when the monitored

file was very frequently rotated (we had some report with every few

seconds).

Note that while we try hard to not lose any messages, input file

rotation always has some loss potential. This is inevitable if

the monitored file is being truncated.

Also note that this bugfix affects imfile, only. It has nothing to do

and no relation to rsyslog output files being rotated on HUP.

closes: https://github.com/rsyslog/rsyslog/issues/4797

- 2022-08-05: ksi bugfix: optimize processing of signer queue to fix delays.

There is a worker queue where rsyslog KSI module collects events and signing

requests. When queue is processed thread is periodically put to sleep. Previous

implementation handles signature requests well but sleeps every time after

handling new file open / close event. When several log files are opened or

closed simultaneously process is significantly slowed down. Another issue is

that thread always sleeps 1000ms that may be 2x longer than aggregation round.

This slows down overall signing process.

Fix for the issue is to simply not sleep after file open / close event if there

are next items to be processed. To speed up the signing process, rsyslog uses

KSI aggregator conf. to obtain the aggregation period that is used for the sleep

time configuration.

Thanks to Taavi Valjaots for the patch.

- 2022-08-04: ksi bugfix: possible crash fixed when several log files are opened.

KSI module in async mode used to request aggregator conf. every time a log

file was opened. When several log files were opened simultaneously

corresponding amount of pointless concurrent conf. requests were posted.

Concurrent conf. requests lead to a bug in libksi, where internal count of

pending requests was not decremented correctly causing system to crash.

Fix for the issue is to optimize the frequency of conf. requests so that only

one conf. requests is handled at once. Instead of checking conf. every time

log file is opened, conf is requested periodically after conf timeout. This will

affect both sync and async mode.

New option for KSI module introduced - sig.confinterval="time, s".

Thanks to Taavi Valjaots for the patch.

- 2022-08-04: openssl: add support to split tls commands by semicolon

- Add support to split tls commands by semicolon.

- Changed one test with multiple tls commands to use semicolon as

separator instead of newline.

closes: https://github.com/rsyslog/rsyslog/issues/4852

- 2022-08-04: openssl subsystem bugfix: build issue on Solaris

Needed header file was added. Platforms other than Solaris did not actually need it,

so this bug was discovered late.

Thanks to Jakub Kulík for the patch.

Import <strings.h> when index() is used.

- 2022-08-04: openssl: add more details to error messages

- Avoid LogMsg outputs osslEndSess on successfull terminated

connection. Only LogMsg if the connection was terminated

unsuccessfully.

- Handle SSL_ERROR_SYSCALL in both Send / osslRecordRecv,

do not log as error if underlaying socket was terminated

(ECONNRESET). Log as information instead.

closes: https://github.com/rsyslog/rsyslog/issues/4946

- 2022-08-04: omclickhouse: capture additional exceptions

- DB::NetException

- DB::ParsingExceptions

Thanks to Victor Kustov for the patch.

- 2022-08-04: mmanon bugfix: Simplified and fixed IPv4 digit detection.

- Fixed an issue with numbers above int64 in syntax_ipv4.

Numbers that were up to 256 above the max of an int64

could incorrectly be detected as valid ipv4 digit.

- Simplified the IPv4 digit detection function and renamed

to isPosByte.

- added testcasse for malformed IPvc4 addresses

closes: https://github.com/rsyslog/rsyslog/issues/4940

- 2022-07-21: imptcp: slight tuning

- reduce indirect addressing to obtain more speed

- also a fix for an annoying typo

- minor other optimizations

- modernization of one test

- 2022-07-20: template procesing/json: performance optimization

- 2022-07-19: core bugfix: memory leak when free action worker data table

During free action worker data table when action destruct, worker instance in worker

data table were not null. It resulted in memory leak.

Thanks to github user seuzw930 for the patch.

- 2022-07-13: omfile: support for zstd compression

The zstd library provides better and faster compression than zlib.

This patch integrates zstd as a dynamically-loadable functionality.

As such, no further dependencies need to be added to the rsyslog

base package.

Due to the increased performance, usage of zstd is highly recommended

for high-volume use cases.

This patch also refactor zlib compression in order to unify handling

in both compression cases.

- 2022-07-07: stream cleanup: move error message to debug log, only

This error message is most probably rooted in a kernel problem. At

least knowbody knows how it can happen. It's definitely not a

rsyslog issue. We also can recover from it for a long time now

so there is no reason to irritate users by emitteing this

"error" message.

- 2022-07-04: mmdblookup bugfix: Don't crash Rsyslog on mmdb file errors

Thanks to Théo Bertin (frikilax) for the patch.

- 2022-06-28: build error fix: libbson requires out-of-date language constructs

- 2022-06-27: OpenSSL: fix depreacted API issues for OpenSSL 3.x

- OpenSSL error strings are loaded automatically now

- Debug Callback has changed

- See for more:

https://www.openssl.org/docs/manmaster/man7/migration_guide.html

closes: https://github.com/rsyslog/rsyslog/issues/4912

----------------------------------------------------------------------------------------

Scheduled Release 8.2206.0 (aka 2022.06) 2022-06-14

- 2022-05-25: omelastisearch: allow omitting _type field

Allow omitting the _type field by setting it to an empty string.

Setting this field has been deprecated since 6.0, and support will

be removed in 8.0

Also add testbench test for empty searchType with ES 7.0

This checks for messages in the deprecation log and also provides

avoids deprecation messages from usage of transport.tcp.port in the

test configuration

Thanks to Jarkko Oranen for the patch.

- 2022-05-18: tcpsrv/imtcp: slight performance improvements

This change slightly improves performance for tcpsrv-based servers.

This affects imtcp and imgssapi as well as some helpers.

No other functional change is included in this change.

- 2022-05-12: imptcp bugfix: worker thread starvation on extreme traffic

When connectes were totally busy, without any pause, the assigened worker

did never terminate its reading loop. As such, it could not service any

other conenctions. If this happened multiple time and to all configured

workers, all other connections could not be processed at all. This extreme

scenario is very unlikely, as the whole issue is relatively unlikely.

In practice, the issue could lead to somewhat degraded performance and

resolved itself after some time (in practice no connection is 100% busy

for an extended period of time).

Note that this patch sets a fixed limit of 16 iterations for very busy

connections. This sounds like a good compromise between non-starvation

and performance. The exact number may be made configurable if there

is really need to.

- 2022-05-11: omelasticsearch: several support option for ElasticSearch 8

- config params searchIndex and documentType can be empty

- support for Data Stream API

Thanks to github user EHerzog76 for these changes.

- new config param esVersion.major

- 2022-05-09: tcp receiver bugfix: delay/potential hang on some error conditions

Error were not correctly handled in some cases for imtcp and imgssapi. This could

lead to a temporary stall of some connections. For ultry-low traffic systems, this

stall could stay for a long period of time. In most cases, it was resolved very quickly.

Note that imptcp was not affected.

Thanks to Iwan Timmer for the fix.

- 2022-05-05: net bugfix: potential buffer overrun

there is heap buffer overflow vulnerability in rsyslog tcp reception components.

This can only happen in octet-counted mode, which is enabled by default.

Affected components: imtcp, imptcp, imhttp, imgssapi, imdiag when octet-counted

framing was enabled.

If the receiver ports are exposed to the public Internet AND are used

without authentication, this can lead to remote DoS and potentially to

remote code execution. It is unclear if remote code execution is

actually possible. If so, it needs a very sophisticated attack.

When syslog best practices with proper firewalling and authentication

is used, thean attack can only be carried out from within the Intranet

and authorized systems. This limits the severity of the vulnerability

considerably (it would obviously require an attacker already to be

present inside the internal network).

Credits to Peter Agten for initially reporting the issue and working

with us on the resolution.

fixes CVE-2022-24903

Advisory:

https://github.com/rsyslog/rsyslog/security/advisories/GHSA-ggw7-xr6h-mmr8#advisory-comment-72243

- 2022-05-05: imptcp: set OS worker thread name

We now set the worker thread names to "imptcp/<thrd nbr>" where

<thrd nbr> is the numerical index (0, 1, ...) of the worker thread.

This enables to distinguish individual worker threads in OS tools like

htop. That is useful for performance testing and system monitoring.

The choosen name format is consistant with other similar thread

names inside rsyslog. For imptcp, worker threads were not yet

given individual names.

Note: "in:imptcp" is imptcp's "main" thread, which also is used

as a worker in some scenarios. This name was not modified.

- 2022-04-26: mmanon bugfix: shortened IPv6 form not always anonymized

If the IPv6 is in non-recommended form followed by a 5 digit port number, it

is not anonymized.

A reproducer for this is: 1a00:c820:1180:c84c::ad3f:d991:ec2e:49255

closes https://github.com/rsyslog/rsyslog/issues/4856

- 2022-04-22: mmdblookup fix: wrong copy of buffer

...following parse of libmaxminddb's return after a successful search sometimes

failed to return specific field from data.

Thanks to Théo Bertin for the patch.

- 2022-04-22: mmdblookup: several enhancements

- support arrays in MMDB entry

- support escaped quotes '"' in MMDB entry

- support '<' characters in MMDB entry, when in a field

- support '}' characters in MMDB entry, when in a field

Thanks to Théo Bertin for the patch.

----------------------------------------------------------------------------------------

Scheduled Release 8.2204.1 (aka 2022.04) 2021-05-05

- security bugfix: potential buffer overrun in imptcp, imtcp, imgssapi and others

This addresses CVE-2022-24903

see also https://github.com/rsyslog/rsyslog/security/advisories/GHSA-ggw7-xr6h-mmr8

----------------------------------------------------------------------------------------

Scheduled Release 8.2204.0 (aka 2022.04) 2021-04-19

- 2022-04-18: gnutls bugfix: possibility of infinite loop

There was a rare possibility that the E_AGAIN/E_INTERRUPT handling

could cause an infinite loop (100% CPU Usage), for example when a TLS

handshake is interrupted at a certain stage.

* After gnutls_record_recv is called, and E_AGAIN/E_INTERRUPT error

occurs, we need to do additional read/write direction handling

with gnutls_record_get_direction.

* After the second call of gnutls_record_recv (Expand buffer)

we needed to also check the eror codes for E_AGAIN/E_INTERRUPT

to do propper errorhandling.

* Add extra debug output based on ossl driver.

* Potential fix for 100% CPU Loop Receiveloop after gtlsRecordRecv

in doRetry call.

closes https://github.com/rsyslog/rsyslog/issues/4834

closes https://github.com/rsyslog/rsyslog/issues/4818

closes https://github.com/rsyslog/rsyslog/issues/4638

- 2022-04-17: core/bugfix: errorfile could grow over max configures size

When action.errorfile.maxsize configuration option is enabled and error file

already has a certain size smaller than max size configured, it is increasing

higher than configured max size as the error file is considered to be zero in code.

This fix reads current error file size and limits the size to the maximum

size configured.

Thanks to Sergio Arroutbi for the patch.

fixes https://github.com/rsyslog/rsyslog/issues/4821

- 2022-04-17: omkafka bugfix: potential misadressing

The `failedmsg_entry` expects a null-terminated string in `key`, but

here we allocate with malloc and copy a string-with-length-n into only

the first n bytes. If the final byte is null, this is by coincidence

only.

This was observed by means of seeing random binary data appended to