Petr Špaček [Wed, 9 Feb 2022 17:42:54 +0000 (18:42 +0100)]
Reorder list of supported RFCs in more user-oriented manner
For users it's not really important if a RFC is Internet Standard,
Proposed Standard, or Experimental. RFCs are now regrouped by
"Protocol", Best Current Practice, and "catch all" category FYI.
Ondřej Surý [Mon, 14 Feb 2022 08:17:41 +0000 (09:17 +0100)]
Bump the required automake version to 1.14
After the build system refactoring, we no longer call AM_PROG_CC_C_O
because it is obsolescent macro. According to the automake manual the
`AC_PROG_CC` has been rewritten in automake 1.14 to not required the
call, thus we need to require at least automake version 1.14.
Ondřej Surý [Wed, 2 Feb 2022 09:50:45 +0000 (10:50 +0100)]
Change the bug reporting address from email to gitlab URI
In autoconf, the AC_INIT() accepts bugreport address for reporting
issues (f.e. when the test suite fails). Instead of providing generic
emails address, change this to the address where to report with the
default Bug template applied.
Ondřej Surý [Thu, 10 Feb 2022 08:21:26 +0000 (09:21 +0100)]
Log when starting and ending task exclusive mode
The task exclusive mode stops all processing (tasks and networking IO)
except the designated exclusive task events. This has impact on the
operation of the server. Add log messages indicating when we start the
exclusive mode, and when we end exclusive task mode.
Ondřej Surý [Mon, 24 Jan 2022 11:13:24 +0000 (12:13 +0100)]
Use compile-time paths in the manual pages
Replace the hard-coded paths for various BIND 9 files (configuration,
pid, etc.) in the man pages and ARM with compile-time values using the
sphinx-build replace system.
This is more complicated, because the restructured text specification
doesn't allow |substitions| inside ``code-blocks``, so for each specific
file we had to create own substition which is sub-optimal, but it is
only way how to do this without adding Sphinx extension.
Ondřej Surý [Wed, 9 Feb 2022 09:00:59 +0000 (10:00 +0100)]
Remove unused functions from isc_thread API
The isc_thread_setaffinity call was removed in !5265 and we are not
going to restore it because it was proven that the performance is better
without it. Additionally, remove the already disabled cpu system test.
The isc_thread_setconcurrency function is unused and also calling
pthread_setconcurrency() on Linux has no meaning, formerly it was
added because of Solaris in 2001 and it was removed when taskmgr was
refactored to run on top of netmgr in !4918.
Michal Nowak [Wed, 2 Feb 2022 09:39:53 +0000 (10:39 +0100)]
Fix path to suppr-lsan.txt file
The "suppr-lsan.txt" file needs to be referenced with GitLab-specific
variable, otherwise AddressSanitizer won't find it outside the
"isc-projects" project group.
Matthijs Mekking [Mon, 31 Jan 2022 15:53:40 +0000 (16:53 +0100)]
Fix keyfromlabel echo output
The 'id' variable is either keyfromlabel-ksk or keyfromlabel-zsk and is
set in the 'keygen' and 'keyfromlabel' functions. It should not be used
outside these functions.
Matthijs Mekking [Tue, 18 Jan 2022 08:36:59 +0000 (09:36 +0100)]
Add test for assertion failure in pk11_numbits
This test was originally in the pkcs11 system test. While this crash
happened in the native pkcs11 of BIND 9, and that code has been
removed in 9.17, there is no need for this test. Nevertheless, it
doesn't hurt having the test case persist.
Ondřej Surý [Sat, 23 Oct 2021 21:37:57 +0000 (23:37 +0200)]
Clarify effect of "directory" configuration option
The "directory" configuration options affects the configuration listed
after the directive but not before which may affect ``include``
directive with relative file paths.
Ondřej Surý [Tue, 1 Feb 2022 17:36:12 +0000 (18:36 +0100)]
Add log message when hard quota is reached in TCP accept
When isc_quota_attach_cb() API returns ISC_R_QUOTA (meaning hard quota
was reached) the accept_connection() would return without logging a
message about quota reached.
Change the connection callback to log the quota reached message.
Petr Špaček [Tue, 1 Feb 2022 12:04:21 +0000 (13:04 +0100)]
Add missing parental-agents grammar generation
Formerly parental-agents grammar was an exception and it did not
auto-generate itself from source code. From now on it is generated using
the same mechanism as other grammars.
For consistency with rest of the system, I've also renamed the grammar
file and the link anchors from "parentals" to "parental-agents".
Evan Hunt [Fri, 28 Jan 2022 23:41:04 +0000 (15:41 -0800)]
complete removal of !PTHREADS code from DLZ modules
DLZ modules no longer support being built without threads,
so the "#if PTHREADS" conditionals were no longer necessary,
and were also causing errors in some of the modules due to
PTHREADS no longer being defined in dlz_pthread.h.
Evan Hunt [Wed, 3 Nov 2021 05:38:45 +0000 (22:38 -0700)]
allow dns_clientinfo to store client ECS data
this brings DNS_CLIENTINFO_VERSION into line with the subscription
branch so that fixes applied to clientinfo processing can also be
applied to the main branch without diverging.
Michal Nowak [Wed, 26 Jan 2022 12:49:32 +0000 (13:49 +0100)]
Drop support for sequential system tests
System test used to have sequential system tests, which can't run in
parallel with the rest of system tests. As there are no such tests
anymore the underlying infrastructure can be dropped.
Michal Nowak [Thu, 20 Jan 2022 12:11:49 +0000 (13:11 +0100)]
Drop bin/tests/system/parallel.sh
"parallel.sh" script was used on Windows to run system tests in
parallel. Since Windows support was removed from BIND 9, the script is
not needed anymore.
Michal Nowak [Mon, 14 Dec 2020 19:40:58 +0000 (20:40 +0100)]
Drop testsummary.sh
testsummary.sh was not updated after build system rewrite to Autotools,
and needs to be fixed to produce test summary and core dump, assertion
failures, and ThreadSanitizer reports.
Given that all of this is provided by Autotools and run.sh already,
there's little use to testsummary.sh script and should be dropped.
Petr Menšík [Tue, 25 Jan 2022 22:26:27 +0000 (23:26 +0100)]
Use detected cache line size
IBM power architecture has L1 cache line size equal to 128. Take
advantage of that on that architecture, do not force more common value
of 64. When it is possible to detect higher value, use that value
instead. Keep the default to be 64.
Ondřej Surý [Thu, 27 Jan 2022 11:51:06 +0000 (12:51 +0100)]
Fix typo in qname-wait-recurse
In the RPZ documentation, there's a mistake where it states that the
default behavior will be disabled by setting `qname-wait-recurse yes;`
while in fact it's opposite `qname-wait-recurse no;`.
Michal Nowak [Tue, 18 Jan 2022 18:39:37 +0000 (19:39 +0100)]
Drop runall.sh
runall.sh was mainly used on Windows and as it's support was removed
from the "main" branch the script is not needed anymore.
Also, remove bin/tests/system/README text on running multiple system
test suites simultaneously with runall.sh as that support was not
present in the script anyway.
Michal Nowak [Thu, 10 Dec 2020 13:10:49 +0000 (14:10 +0100)]
Drop bin/tests/system/setup.sh
bin/tests/system/setup.sh just executes setup.sh script of a particular
system test in the directory of the system test. This does not seems to
be useful enough to maintain it.
Michal Nowak [Thu, 10 Dec 2020 12:53:44 +0000 (13:53 +0100)]
Drop stopall.sh
stopall.sh script takes almost 2 minutes to go thru all test
subdirectories (due to a sleep in stop.pl) and does not seems to be
efficient way to stop manually started tests.
Matthijs Mekking [Tue, 11 Jan 2022 08:04:55 +0000 (09:04 +0100)]
Create keys with pkcs11-tool --id
The keyfromlabel system ECDSA tests sometimes fail. When this happens
the ZSK and KSK key id values differ by 1, which is an indication that
the same key is used for both DNSKEY records.
When the private key is retrieved with 'ENGINE_load_private_key()', the
public key is already set. But sometimes that key differs from the key
which was retrieved with 'ENGINE_load_public_key()'.
The libp11 source code uses id to find the key and without IDs all the
keys are "equal", so it is returning the first key in the array of the
enumerated keys instead of the matching key. In our test we didn't use
'--id', just '--label'. With this change, the system test should no
longer fail intermittently.
Note this is only an issue for ECDSA keys, not RSA keys.
Matthijs Mekking [Mon, 15 Nov 2021 09:23:04 +0000 (10:23 +0100)]
Add system test for dnssec-keyfromlabel
Add missing system test for dnssec-keyfromlabel. Test for various
algorithms that we can generate key files from a key that is stored in a
HSM, and that those keys can be used for signing with dnssec-signzone.
Matthijs Mekking [Thu, 11 Nov 2021 08:35:46 +0000 (09:35 +0100)]
Update .gitlab-ci.yml with openssl setup
GitLab CI needs to know about some environment variables that will
tell where OpenSSL and SoftHSM2 is installed. This is done in the
image, making the prepare-softhsm2.sh script obsolete.
Evan Hunt [Fri, 15 Oct 2021 07:05:36 +0000 (00:05 -0700)]
Remove leftover test code for Windows
- Removed all code that only runs under CYGWIN, and made all
code that doesn't run under CYGWIN non-optional.
- Removed the $TP variable which was used to add optional
trailing dots to filenames; they're no longer optional.
- Removed references to pssuspend and dos2unix.
- No need to use environment variables for diff and kill.
- Removed uses of "tr -d '\r'"; this was a workaround for
a cygwin regex bug that is no longer needed.
Michał Kępień [Wed, 26 Jan 2022 14:18:43 +0000 (15:18 +0100)]
Fix waiting for lock file removal upon exit
Commit c787a539d2a931ba9023677c1c269ed191455512 fixed a certain class of
intermittent system test failures caused by named instances unable to
restart. The root cause was bin/tests/system/stop.pl returning without
waiting for a named instance to remove its lock file.
Later on, it turned out that the above change causes other issues on
Windows due to the way named handles signals on that platform. Commit 761ba4514f7eceab8019d71dc9cabd269d274597 intended to address those
issues by making the server_lock_file() subroutine in
bin/tests/system/stop.pl return an empty value on Windows, in order to
prevent the script for waiting for lock file cleanup on that platform.
Note, however, that Windows detection in that subroutine is limited to
checking whether the CYGWIN environment variable is set.
While that environment variable was not set on Unix-like systems before
commit 761ba4514f7eceab8019d71dc9cabd269d274597, another commit
(a33237f070c95480f581d85b0169f41ce5a12017, merged a few weeks later)
changed that by setting the CYGWIN environment variable to an empty
value on Unix-like systems. This made the defined($ENV{'CYGWIN'}) check
in server_lock_file() return true, inadvertently preventing
bin/tests/system/stop.pl from waiting for lock file removal before
exiting on Unix-like systems and therefore reintroducing the original
issue.
Fix by making server_lock_file() only return an empty value when the
CYGWIN environment variable is set to a non-empty value (which is what
bin/tests/system/conf.sh.win32 does). Adjust a similar check in the
pid_file_exists() subroutine in the same way for consistency.
Michał Kępień [Wed, 26 Jan 2022 14:18:43 +0000 (15:18 +0100)]
Do not strip leading whitespace from test output
The echo_*() and cat_*() functions in bin/tests/system/conf.sh.common
call the "read" builtin command without specifying the field separator
to use. This results in leading whitespace getting stripped from each
line of the texts passed to those functions, which mangles e.g. pytest
output, hindering test failure troubleshooting.
Address by setting IFS to an empty value for the "read" calls used in
the aforementioned helper functions.
Michał Kępień [Wed, 26 Jan 2022 14:18:43 +0000 (15:18 +0100)]
Retain all named.run files from each test run
The bin/tests/system/start.pl script truncates the named.run file for a
given named instance unless it is invoked with the --restart
command-line option. Ever since Python-based tests were introduced,
bin/tests/system/run.sh may start named instances used by a given system
test multiple times within a single run, causing the
bin/tests/system/start.pl script to truncate some of the log files
written during the test. This makes troubleshooting certain test
failures hard or even impossible.
Fix by calling bin/tests/system/start.pl with the --restart command-line
option for every start_servers() invocation except the first one.
projects / thirdparty / bind9.git / log
