The race condition is the timer elapses before isc__timer_create()
returns the pointer to the caller. Assigning the return pointer before
enabling the timer will fix it.
(cherry picked from commit 21966423cd)
Zone loading happens in a different task (zone->loadtask) than other
zone actions (zone->task). Thus, when zone_postload() is called in the
context of zone->loadtask, it may cause zone maintenance to be queued in
zone->task and another thread can then execute zone_maintenance() before
zone_postload() gets a chance to finish its work in the first thread.
This would not be a problem if zone_maintenance() accounted for this
possibility by locking the zone before checking the state of its
DNS_ZONEFLG_LOADPENDING flag. However, the zone is currently not locked
before the state of that flag is checked, which may prevent zone
maintenance from happening despite zone_postload() scheduling it. Fix
by locking the zone in zone_maintenance() before checking the state of
the zone's DNS_ZONEFLG_LOADPENDING flag.
(cherry picked from commit 56003e9f9f)
While "rndc reload" causes dns_zone_asyncload() to be called for the
signed version of an inline-signed zone, the subsequent zone_load() call
causes the raw version to be reloaded from storage. This means that
DNS_ZONEFLG_LOADPENDING gets set for the signed version of the zone by
dns_zone_asyncload() before the reload is attempted, but zone_postload()
is only called for the raw version and thus DNS_ZONEFLG_LOADPENDING is
cleared for the raw version, but not for the signed version. This in
turn prevents zone maintenance from happening for the signed version of
the zone.
Until commit 749b3cacfc, this problem
remained dormant because DNS_ZONEFLG_LOADPENDING was previously
immediately, unconditionally cleared after zone loading was started
(whereas it should only be cleared when zone loading is finished or an
error occurs). This behavior caused other issues [1] and thus had to be
changed.
Fix reloading inline-signed zones by clearing DNS_ZONEFLG_LOADPENDING
for the signed version of the zone once the raw version reload
completes. Take care not to clear it prematurely during initial zone
load. Also make sure that DNS_ZONEFLG_LOADPENDING gets cleared when
zone_postload() encounters an error or returns early, to prevent other
scenarios from resulting in the same problem. Add comments aiming to
help explain code flow.
[1] see RT #47076
(cherry picked from commit 5431583971)
When an inline-signed zone is loaded, the master file for its signed
version is loaded and then a rollforward of the journal for the signed
version of the zone is performed. If DNS_JOURNALOPT_RESIGN is not set
during the latter phase, signatures loaded from the journal for the
signed version of the zone will not be scheduled for refresh. Fix the
conditional expression determining which flags should be used for the
dns_journal_rollforward() call so that DNS_JOURNALOPT_RESIGN is set when
zone_postload() is called for the signed version of an inline-signed
zone.
Extend bin/tests/system/stop.pl so that it can use "rndc halt" instead
of "rndc stop" as the former allows master file flushing upon shutdown
to be suppressed.
(cherry picked from commit 8db550c42f)
dst__openssl_toresult3() first calls toresult() and subsequently uses
ERR_get_error_line_data() in a loop. Given this, it is a mistake to use
ERR_get_error() in toresult() because it causes the retrieved error to
be removed from the OpenSSL error queue, thus preventing it from being
retrieved by the subsequent ERR_get_error_line_data() calls. Fix by
using ERR_peek_error() instead of ERR_get_error() in toresult().
(cherry picked from commit 36436268b5)
When two or more absolute, two-label names are added to a completely
empty RBT, an extra, empty node for the root name will be created due to
node splitting. check_tree() expects that, but the extra node will not
be created when just one name is added to a completely empty RBT. This
problem could be handled inside check_tree(), but that would introduce
unnecessary complexity into it since adding a single name will result in
a different node count for a completely empty RBT (node count will be 1)
and an RBT containing only an empty node for the root name, created due
to prior node splitting (node count will be 2). Thus, first explicitly
create a node for the root name to prevent rare check_tree() failures
caused by a single name being added in the first iteration of the
insert/remove loop.
(cherry picked from commit 13fe763798)
If "rndc signing -nsec3param ..." is ran for a zone which has not yet
been loaded or transferred (i.e. its "db" field is NULL), it will be
silently ignored by named despite rndc logging an "nsec3param request
queued" message, which is misleading. Prevent this by keeping a
per-zone queue of NSEC3PARAM change requests which arrive before a zone
is loaded or transferred and processing that queue once the raw version
of an inline-signed zone becomes available.
(cherry picked from commit cb40c5229a)
In order to decrease code duplication, express the logic contained in
all zone logging functions using dns_zone_logv() calls.
(cherry picked from commit 5c03cd339e)
Add a new libdns function, dns_zone_logv(), which takes a single va_list
argument rather than a variable number of arguments and can be used as a
base for implementing more specific zone logging functions.
(cherry picked from commit bb2dfb3f49)
- use RADIX_V4, RADIX_V6, RADIX_V4_ECS, and RADIX_V6_ECS as array
indices instead of 0 through 3.
- remove some unused macros
(cherry picked from commit f7f20b1202)
Certain isc_buffer_*() functions might call memmove() with the second
argument (source) set to NULL and the third argument (length) set to 0.
While harmless, it triggers an ubsan warning:
runtime error: null pointer passed as argument 2, which is declared to never be null
Modify all memmove() call sites in lib/isc/include/isc/buffer.h and
lib/isc/buffer.c which may potentially use NULL as the second argument
(source) so that memmove() is only called if the third argument (length)
is non-zero.
(cherry picked from commit 6ddbca6f2b)
compare_rdata() was meant to be used as a qsort() callback. Meanwhile,
dns_rdataslab_merge() calls compare_rdata() for a pair of dns_rdata_t
structures rather than a pair of struct xrdata structures, which is
harmless, but triggers an ubsan warning:
rdataslab.c:84:33: runtime error: member access within address <address> with insufficient space for an object of type 'const struct xrdata'
Use dns_rdata_compare() instead of compare_rdata() to prevent the
warning from being triggered.
(cherry picked from commit 9bc6ba0be9)
Add some basic test cases ensuring dns__zone_updatesigs() behaves as
expected.
(cherry picked from commit 8b9d2c27b4)
(cherry picked from commit 3e93e4bb62)
Add a new ATF test, sigs_test, containing everything required to start
defining test cases for dns__zone_updatesigs(). The framework is
written in a way which ensures that changes to zone database applied by
any dns__zone_updatesigs() invocation are preserved between subsequent
checks.
(cherry picked from commit 1f10186476)
(cherry picked from commit 3dde7c42db)
Rename find_zone_keys() to dns__zone_findkeys() and move it to
lib/dns/zone_p.h, so that it can be used in unit tests. Add a comment
describing the purpose of this function.
(cherry picked from commit d7143986b1)
(cherry picked from commit fc0e99c7d7)
Rename update_sigs() to dns__zone_updatesigs() and move it to
lib/dns/zone_p.h, so that it can be unit tested. Add a comment
describing the purpose of this function.
(cherry picked from commit b1947cee82)
(cherry picked from commit 4d06f50ba8)
Rename zonediff_t to dns__zonediff_t and move it to lib/dns/zone_p.h, so
that unit tests can be written for functions taking pointers to
structures of this type as arguments.
(cherry picked from commit ace465a9f9)
(cherry picked from commit ea15c54d8a)
Add a new private header file, lib/dns/zone_p.h, which will hold type
definitions and function prototypes not meant to be exported by libdns,
but required by zone-related unit tests.
(cherry picked from commit c1bc3be806)
(cherry picked from commit 2b0add6d1a)
Implement dns_test_difffromchanges(), a function which enables preparing
a dns_diff_t structure from a mostly-textual representation of zone
database changes to be applied. This will improve readability of test
case definitions by allowing contents of a dns_diff_t structure, passed
e.g. to update_sigs(), to be represented in a human-friendly manner.
(cherry picked from commit 3c22af0d35)
(cherry picked from commit d4c603eb8a)
Remove the underscore from "rdata_fromstring" so that all helper
functions for libdns tests use a common naming covention.
(cherry picked from commit 2980cbd55f)
(cherry picked from commit 107102d333)
The dns_test_makezone() helper function always assigns the created zone
to some view, which is not always necessary and complicates cleanup of
non-managed zones as they are required not to be assigned to any view.
Rework dns_test_makezone() in order to make it easier to use in unit
tests operating on non-managed zones. Use dns_name_fromstring() instead
of dns_name_fromtext() to simplify code. Do not use the CHECK() macro
and add comments to make code flow simpler to follow. Use
dns_test_makeview() instead of dns_view_create().
Adjust existing unit tests using this function so that they still pass.
(cherry picked from commit bfbeef3609)
(cherry picked from commit f70c02d2c2)
When inline signing is enabled for a zone without creating signing keys
for it, changes subsequently applied to the raw zone will not be
reflected in the secure zone due to the dns_update_signaturesinc() call
inside receive_secure_serial() failing. Given that an inline zone will
be served (without any signatures) even with no associated signing keys
being present, keep applying raw zone deltas to the secure zone until
keys become available in an attempt to follow the principle of least
astonishment.
(cherry picked from commit 6acf326969)
(cherry picked from commit 8a58a60772)
If a raw zone is modified, but the dns_update_signaturesinc() call in
receive_secure_serial() fails, the corresponding secure zone's database
will not be modified, even though by that time a message containing the
bumped signed serial will already have been logged. This creates
confusion, because a different secure zone version will be served than
the one announced in the logs. Move the relevant dns_zone_log() call so
that it is only performed if the secure zone's database is modified.
(cherry picked from commit cfbc8e264d)
(cherry picked from commit cdc7ab42b1)
- IMHO we should consider removing dnsconf.c and deprecating the
/etc/dns.conf file, though, as I don't think it's likely anyone
is using it
(cherry picked from commit a08ba418ef)
Replace dns_fixedname_init() calls followed by dns_fixedname_name()
calls with calls to dns_fixedname_initname() where it is possible
without affecting current behavior and/or performance.
This patch was mostly prepared using Coccinelle and the following
semantic patch:
@@
expression fixedname, name;
@@
- dns_fixedname_init(&fixedname);
...
- name = dns_fixedname_name(&fixedname);
+ name = dns_fixedname_initname(&fixedname);
The resulting set of changes was then manually reviewed to exclude false
positives and apply minor tweaks.
It is likely that more occurrences of this pattern can be refactored in
an identical way. This commit only takes care of the low-hanging fruit.
(cherry picked from commit 4df4a8e731)
(cherry picked from commit 0041aeb751)
The "sortlist-v4.conf" unit test for irs_resconf_load() is always run
twice due to a duplicate entry in the "tests" table. Remove one of them
to prevent this.
(cherry picked from commit 6c09f305ae)
irs_resconf_load() stores the value returned by add_search() into ret
without consulting its current value first. This causes any previous
errors raised while parsing resolv.conf to be ignored as long as any
"domain" or "search" statement is present in the file.
Prevent this by returning early in case an error is detected while
parsing resolv.conf. Ensure that "searchlist" and "magic" members of
the created irs_resconf_t structure are always initialized before
isc_resconf_destroy() is called.
(cherry picked from commit 1f400b68a8)
generated into builddir. If out-of-tree build is used, make unit
will always fail. Kyuafiles and testdata still have to be copied
manually into the builddir.
(cherry picked from commit 95cde3608a)
Remove a block of code which dates back to commit 8a2ab2b920, when
dns_zone_asyncload() did not yet check DNS_ZONEFLG_LOADPENDING.
Currently, no race in accessing DNS_ZONEFLG_LOADPENDING is possible any
more, because:
- dns_zone_asyncload() is still the only function which may queue
zone_asyncload(),
- dns_zone_asyncload() accesses DNS_ZONEFLG_LOADPENDING under a lock
(and potentially queues an event under the same lock),
- DNS_ZONEFLG_LOADPENDING is not cleared until the load actually
completes.
Thus, the rechecking code can be safely removed from zone_asyncload().
Note that this also brings zone_asyncload() to a state in which the
completion callback is always invoked. This is required to prevent
leaking memory in case something goes wrong in zone_asyncload() and a
zone table the zone belongs to is indefinitely left with a positive
reference count.
Code handling cancellation of asynchronous zone load events was likely
copied over from other functions when asynchronous zone loading was
first implemented in commit 8a2ab2b920. However, unlike those other
functions, asynchronous zone loading events currently have no way of
getting canceled once they get posted, which means the aforementioned
code is effectively dead. Remove it to prevent confusion.
zone_load() is not always synchronous, it may only initiate an
asynchronous load and return DNS_R_CONTINUE, which means zone loading
has not yet been completed. In such a case, zone_asyncload() must not
clear DNS_ZONEFLG_LOADPENDING immediately and leave that up to
zone_postload().
While this is not an issue in named, which only calls
dns_zone_asyncload() from task-exclusive mode, this function is exported
by libdns and thus may in theory be concurrently called for the same
zone by multiple threads. It also does not hurt to be consistent
locking-wise with other DNS_ZONEFLG_LOADPENDING accesses.
