when looking for a possible wildcard match in the RPZ summary database,
use an rbtnodechain to walk up label by label, rather than using the
node's parent pointer.
(cherry picked from commit 6e9be9a952)
There's a deadlock in BIND 9 code where (dns_view_t){ .lock } and
(dns_resolver_t){ .buckets[i].lock } gets locked in different order. When
view->weakrefs gets converted to a reference counting we can reduce the locking
in dns_view_weakdetach only to cases where it's the last instance of the
dns_view_t object.
Make sure all unit tests include headers in a similar order:
1. Three headers which must be included before <cmocka.h>.
2. System headers.
3. UNIT_TESTING definition, followed by the <cmocka.h> header.
4. libisc headers.
5. Headers from other BIND libraries.
6. Local headers.
Also make sure header file names are sorted alphabetically within each
block of #include directives.
(cherry picked from commit 5381ac0fcc)
All unit tests define the UNIT_TESTING macro, which causes <cmocka.h> to
replace malloc(), calloc(), realloc(), and free() with its own functions
tracking memory allocations. In order for this not to break
compilation, the system header declaring the prototypes for these
standard functions must be included before <cmocka.h>.
Normally, these prototypes are only present in <stdlib.h>, so we make
sure it is included before <cmocka.h>. However, musl libc also defines
the prototypes for calloc() and free() in <sched.h>, which is included
by <pthread.h>, which is included e.g. by <isc/mutex.h>. Thus, unit
tests including "dnstest.h" (which includes <isc/mem.h>, which includes
<isc/mutex.h>) after <cmocka.h> will not compile with musl libc as for
these programs, <sched.h> will be included after <cmocka.h>.
Always including <cmocka.h> after all other header files is not a
feasible solution as that causes the mock assertion macros defined in
<isc/util.h> to mangle the contents of <cmocka.h>, thus breaking
compilation. We cannot really use the __noreturn__ or analyzer_noreturn
attributes with cmocka assertion functions because they do return if the
tested condition is true. The problem is that what BIND unit tests do
is incompatible with Clang Static Analyzer's assumptions: since we use
cmocka, our custom assertion handlers are present in a shared library
(i.e. it is the cmocka library that checks the assertion condition, not
a macro in unit test code). Redefining cmocka's assertion macros in
<isc/util.h> is an ugly hack to overcome that problem - unfortunately,
this is the only way we can think of to make Clang Static Analyzer
properly process unit test code. Giving up on Clang Static Analyzer
being able to properly process unit test code is not a satisfactory
solution.
Undefining _GNU_SOURCE for unit test code could work around the problem
(musl libc's <sched.h> only defines the prototypes for calloc() and
free() when _GNU_SOURCE is defined), but doing that could introduce
discrepancies for unit tests including entire *.c files, so it is also
not a good solution.
All in all, including <sched.h> before <cmocka.h> for all affected unit
tests seems to be the most benign way of working around this musl libc
quirk. While quite an ugly solution, it achieves our goals here, which
are to keep the benefit of proper static analysis of unit test code and
to fix compilation against musl libc.
(cherry picked from commit 59528d0e9d)
This commit changes the BIND cookie algorithms to match
draft-sury-toorop-dnsop-server-cookies-00. Namely, it changes the Client Cookie
algorithm to use SipHash 2-4, adds the new Server Cookie algorithm using SipHash
2-4. The change doesn't make the SipHash 2-4 to be the default algorithm, this
is up to the operator.
(cherry picked from commit 196b342bc9)
Each individual test opened GeoIP databases but the database handles were never
closed. This commit moves the open/close from the individual unit tests into
the _setup and _teardown methods where they really belong.
(cherry picked from commit d1c7b79183)
(cherry picked from commit aad88d2c00)
The MSVS C compiler requires every struct to have at least one member.
The dns_geoip_databases_t structure had one set of members for
HAVE_GEOIP and a different set for HAVE_GEOIP2, and none when neither
API is in use.
This commit silences the compiler error by moving the declaration of
dns_geoip_databases_t to types.h as an opaque reference, and commenting
out the contents of geoip.h when neither version of GeoIP is enabled.
(cherry picked from commit 81fcde5953)
- revise mapping of search terms to database types to match the
GeoIP2 schemas.
- open GeoIP2 databases when starting up; close when shutting down.
- clarify the logged error message when an unknown database type
is configured.
- add new geoip ACL subtypes to support searching for continent in
country databases.
- map geoip ACL subtypes to specific MMDB database queries.
- perform MMDB lookups based on subtype, saving state between
queries so repeated lookups for the same address aren't necessary.
(cherry picked from commit 6e0b93e5a0)
(cherry picked from commit 0283ab7512)
- "--with-geoip" is used to enable the legacy GeoIP library.
- "--with-geoip2" is used to enable the new GeoIP2 library
(libmaxminddb), and is on by default if the library is found.
- using both "--with-geoip" and "--with-geoip2" at the same time
is an error.
- an attempt is made to determine the default GeoIP2 database path at
compile time if pkg-config is able to report the module prefix. if
this fails, it will be necessary to set the path in named.conf with
geoip-directory
- Makefiles have been updated, and a stub lib/dns/geoip2.c has been
added for the eventual GeoIP2 search implementation.
(cherry picked from commit fea6b5bf10)
(cherry picked from commit 6a7e805796)
if "rndc reload" fails, the result code is supposed to be passed to
zone_postload, but for inline-signing zones, the result can be
overwritten first by a call to the ZONE_TRYLOCK macro. this can lead
to the partially-loaded unsigned zone being synced over to the signed
zone instead of being rejected.
(cherry picked from commit 0b792bd37b)
BIND 9.11.0 has bumped DNS_CLIENTINFOMETHODS_VERSION and _AGE to
version 2 and 1 in the dlz_minimal.h because a member was addet to the
dnsclientinfo struct. It was found out that the new member is not
used anywhere and there are no accessor functions therefore the change
was reverted.
Later on, it was found out that the revert caused some problems to the
users of BIND 9, and thus this changes takes a different approach by
syncing the values other way around.
(cherry picked from commit 39344dfb3e)
Compiling with -O3 triggers the following warnings with GCC 9.1:
task.c: In function ‘isc__taskmgr_create’:
task.c:1456:44: warning: ‘%04u’ directive output may be truncated writing between 4 and 10 bytes into a region of size 6 [-Wformat-truncation=]
1456 | snprintf(name, sizeof(name), "isc-worker%04u", i);
| ^~~~
task.c:1456:33: note: directive argument in the range [0, 4294967294]
1456 | snprintf(name, sizeof(name), "isc-worker%04u", i);
| ^~~~~~~~~~~~~~~~
task.c:1456:4: note: ‘snprintf’ output between 15 and 21 bytes into a destination of size 16
1456 | snprintf(name, sizeof(name), "isc-worker%04u", i);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
rrl.c: In function ‘debit_rrl_entry’:
rrl.c:602:35: error: ‘%d’ directive output may be truncated writing between 1 and 10 bytes into a region of size 9 [-Werror=format-truncation=]
602 | snprintf(buf, sizeof(buf), "age=%d", age);
| ^~
rrl.c:602:30: note: directive argument in the range [0, 2147483647]
602 | snprintf(buf, sizeof(buf), "age=%d", age);
| ^~~~~~~~
rrl.c:602:3: note: ‘snprintf’ output between 6 and 15 bytes into a destination of size 13
602 | snprintf(buf, sizeof(buf), "age=%d", age);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
rrl.c:602:35: error: ‘%d’ directive output may be truncated writing between 1 and 10 bytes into a region of size 9 [-Werror=format-truncation=]
602 | snprintf(buf, sizeof(buf), "age=%d", age);
| ^~
rrl.c:602:30: note: directive argument in the range [0, 2147483647]
602 | snprintf(buf, sizeof(buf), "age=%d", age);
| ^~~~~~~~
rrl.c:602:3: note: ‘snprintf’ output between 6 and 15 bytes into a destination of size 13
602 | snprintf(buf, sizeof(buf), "age=%d", age);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
rrl.c:602:35: error: ‘%d’ directive output may be truncated writing between 1 and 10 bytes into a region of size 9 [-Werror=format-truncation=]
602 | snprintf(buf, sizeof(buf), "age=%d", age);
| ^~
rrl.c:602:30: note: directive argument in the range [0, 2147483647]
602 | snprintf(buf, sizeof(buf), "age=%d", age);
| ^~~~~~~~
rrl.c:602:3: note: ‘snprintf’ output between 6 and 15 bytes into a destination of size 13
602 | snprintf(buf, sizeof(buf), "age=%d", age);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
private_test.c: In function ‘private_nsec3_totext_test’:
private_test.c:114:9: warning: array subscript 4 is outside array bounds of ‘uint32_t[1]’ {aka ‘unsigned int[1]’} [-Warray-bounds]
114 | while (*sp == '\0' && slen > 0) {
| ^~~
private_test.c:107:11: note: while referencing ‘salt’
107 | uint32_t salt;
| ^~~~
Prevent these warnings from being triggered by increasing the size of
the relevant arrays (task.c, rrl.c) and reordering conditions
(private_test.c).
(cherry picked from commit ce796ac1f4)
On some systems (namely Debian buster armhf) the readdir() call fails
with `Value too large for defined data type` unless the
_FILE_OFFSET_BITS=64 is defined. The correct way to fix this is to
get the appropriate compilation parameters from getconf system
interface.
(cherry picked from commit 4c7345bcb6)
If named is configured to perform DNSSEC validation and also forwards
all queries ("forward only;") to validating resolvers, negative trust
anchors do not work properly because the CD bit is not set in queries
sent to the forwarders. As a result, instead of retrieving bogus DNSSEC
material and making validation decisions based on its configuration,
named is only receiving SERVFAIL responses to queries for bogus data.
Fix by ensuring the CD bit is always set in queries sent to forwarders
if the query name is covered by an NTA.
(cherry picked from commit 5e80488270)
The option `update-check-ksk` will look if both KSK and ZSK are
available before signing records. It will make sure the keys are
active and available. However, for operational practices keys may
be offline. This commit relaxes the update-check-ksk check and will
mark a key that is offline to be available when adding signature
tasks.
(cherry picked from commit 3cb8c49c73)
(cherry picked from commit b508cffeee3bfb8bc7dcf39db59ec3782a5d9e4c)
