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 29b7efdd9f, 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
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.
As part of resquery_response() refactoring [1], a goto statement was
replaced [2] with a call to a new function - originally called
rctx_delegation(), now folded into rctx_answer_none() - extracted from
existing code. However, one call site of that refactored function does
not reset the "result" variable, causing a referral with a non-empty
ANSWER section to be inadvertently treated as an error, which prevents
resolution of names reliant on servers sending such responses. Fix by
resetting the "result" variable to ISC_R_SUCCESS when a response
containing a non-empty ANSWER section can be treated as a delegation.
[1] see RT #45362
[2] see commit e1380a16741a3b4a57e54d7a9ce09dd12691522f
- added new 'validate-except' option, which configures an NTA with
expiry of 0xffffffff. NTAs with that value in the expiry field do not
expire, are are not written out when saving the NTA table and are not
dumped by rndc secroots
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().
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.
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.
Changes introduced by the previous two commits make the parts of
query_delegation() and query_zone_delegation() which prepare a
delegation response functionally equivalent. Extract this code into a
separate function, query_prepare_delegation_response(), and then call
the latter from both query_delegation() and query_zone_delegation() in
order to reduce code duplication. Add a comment describing the purpose
of the extracted code. Fix coding style issues.
When query processing hits a delegation from a locally configured zone,
an attempt may be made to look for a better answer in the cache. In
such a case, the zone-sourced delegation data is set aside and the
lookup is retried using the cache database. When that lookup is
completed, a decision is made whether the answer found in the cache is
better than the answer found in the zone.
Currently, if the zone-sourced answer turns out to be better than the
one found in the cache:
- qctx->zdb is not restored into qctx->db,
- qctx->node, holding the zone database node found, is not even saved.
Thus, in such a case both qctx->db and qctx->node will point at cache
data. This is not an issue for BIND versions which do not support
mirror zones because in these versions non-recursive queries always
cause the zone-sourced delegation to be returned and thus the
non-recursive part of query_delegation() is never reached if the
delegation is coming from a zone. With mirror zones, however,
non-recursive queries may cause cache lookups even after a zone
delegation is found. Leaving qctx->db assigned to the cache database
when query_delegation() determines that the zone-sourced delegation is
the best answer to the client's query prevents DS records from being
added to delegations coming from mirror zones. Fix this issue by
keeping the zone database and zone node in qctx while the cache is
searched for an answer and then restoring them into qctx->db and
qctx->node, respectively, if the zone-sourced delegation turns out to be
the best answer. Since this change means that qctx->zdb cannot be used
as the glue database any more as it will be reset to NULL by RESTORE(),
ensure that qctx->db is not a cache database before attaching it to
qctx->client->query.gluedb.
Furthermore, current code contains a conditional statement which
prevents a mirror zone from being used as a source of glue records.
Said statement was added to prevent assertion failures caused by
attempting to use a zone database's glue cache for finding glue for an
NS RRset coming from a cache database. However, that check is overly
strict since it completely prevents glue from being added to delegations
coming from mirror zones. With the changes described above in place,
the scenario this check was preventing can no longer happen, so remove
the aforementioned check.
If qctx->zdb is not NULL, qctx->zfname will also not be NULL;
qctx->zsigrdataset may be NULL in such a case, but query_putrdataset()
handles pointers to NULL pointers gracefully. Remove redundant
conditional expressions to make the cleanup code in query_freedata()
match the corresponding sequences of SAVE() / RESTORE() macros more
closely.
