in the "refactor tcpquota and pipeline refs" commit, the counting
of active interfaces was tightened in such a way that named could
fail to listen on an interface if there were more interfaces than
tcp-clients. when checking the quota to start accepting on an
interface, if the number of active clients was above zero, then
it was presumed that some other client was able to handle accepting
new connections. this, however, ignored the fact that the current client
could be included in that count, so if the quota was already exceeded
before all the interfaces were listening, some interfaces would never
listen.
we now check whether the current client has been marked active; if so,
then the number of active clients on the interface must be greater
than 1, not 0.
(cherry picked from commit 02365b87ea0b1ea5ea8b17376f6734c811c95e61)
- if the TCP quota has been exceeded but there are no clients listening
for new connections on the interface, we can now force attachment to the
quota using isc_quota_force(), instead of carrying on with the quota not
attached.
- the TCP client quota is now referenced via a reference-counted
'ns_tcpconn' object, one of which is created whenever a client begins
listening for new connections, and attached to by members of that
client's pipeline group. when the last reference to the tcpconn
object is detached, it is freed and the TCP quota slot is released.
- reduce code duplication by adding mark_tcp_active() function
- convert counters to stdatomic
(cherry picked from commit a8dd133d270873b736c1be9bf50ebaa074f5b38f)
- ensure that tcpactive is cleaned up correctly when accept() fails.
- set 'client->tcpattached' when the client is attached to the tcpquota.
carry this value on to new clients sharing the same pipeline group.
don't call isc_quota_detach() on the tcpquota unless tcpattached is
set. this way clients that were allowed to accept TCP connections
despite being over quota (and therefore, were never attached to the
quota) will not inadvertently detach from it and mess up the
accounting.
- simplify the code for tcpquota disconnection by using a new function
tcpquota_disconnect().
- before deciding whether to reject a new connection due to quota
exhaustion, check to see whether there are at least two active
clients. previously, this was "at least one", but that could be
insufficient if there was one other client in READING state (waiting
for messages on an open connection) but none in READY (listening
for new connections).
- before deciding whether a TCP client object can to go inactive, we
must ensure there are enough other clients to maintain service
afterward -- both accepting new connections and reading/processing new
queries. A TCP client can't shut down unless at least one
client is accepting new connections and (in the case of pipelined
clients) at least one additional client is waiting to read.
(cherry picked from commit 427a2fb4d17bc04ca3262f58a9dcf5c93fc6d33e)
the TCP client quota could still be ineffective under some
circumstances. this change:
- improves quota accounting to ensure that TCP clients are
properly limited, while still guaranteeing that at least one client
is always available to serve TCP connections on each interface.
- uses more descriptive names and removes one (ntcptarget) that
was no longer needed
- adds comments
(cherry picked from commit 9e74969f85329fe26df2fad390468715215e2edd)
tcp-clients settings could be exceeded in some cases by
creating more and more active TCP clients that are over
the set quota limit, which in the end could lead to a
DoS attack by e.g. exhaustion of file descriptors.
If TCP client we're closing went over the quota (so it's
not attached to a quota) mark it as mortal - so that it
will be destroyed and not set up to listen for new
connections - unless it's the last client for a specific
interface.
(cherry picked from commit eafcff07c25bdbe038ae1e4b6660602a080b9395)
- Always set is_zonep in query_getdb; previously it was only set if
result was ISC_R_SUCCESS or ISC_R_NOTFOUND.
- Don't reset is_zone for redirect.
- Style cleanup.
(cherry picked from commit a85cc641d7a4c66cbde03cc4e31edc038a24df46)
In dns_rpz_update_from_db we call setup_update which creates the db
iterator and calls dns_dbiterator_first. This unpauses the iterator and
might cause db->tree_lock to be acquired. We then do isc_task_send(...)
on an event to do quantum_update, which (correctly) after each iteration
calls dns_dbiterator_pause, and re-isc_task_sends itself.
That's an obvious bug, as we're holding a lock over an async task send -
if a task requesting write (e.g. prune_tree) is scheduled on the same
workers queue as update_quantum but before it, it will wait for the
write lock indefinitely, resulting in a deadlock.
To fix it we have to pause dbiterator in setup_update.
(cherry picked from commit 06021b3529)
When parsing message with DNS_MESSAGE_BESTEFFORT (used exclusively in
tools, never in named itself) if we hit an invalid SIG(0) in wrong
place we continue parsing the message, and put the sig0 in msg->sig0.
If we then hit another sig0 in a proper place we see that msg->sig0
is already 'taken' and we don't free name and rdataset, and we don't
set seen_problem. This causes an assertion failure.
This fixes that issue by setting seen_problem if we hit second sig0,
tsig or opt, which causes name and rdataset to be always freed.
(cherry picked from commit 51a55ddbb7)
This changes dns_dtdata struct to not expose data types from dnstap.pb-c.h to
prevent the need for including this header where not really needed.
(cherry picked from commit 8ccce7e24b)
Change to cmocka broken initialization of TZ environment. This time,
commit 1cf1254051 is not soon enough. Has
to be moved more forward, before any other tests. It library is not full
reinitialized on each test.
(cherry picked from commit 71c4fad592)
When sending an udp query (resquery_send) we first issue an asynchronous
isc_socket_connect and increment query->connects, then isc_socket_sendto2
and increment query->sends.
If we happen to cancel this query (fctx_cancelquery) we need to cancel
all operations we might have issued on this socket. If we are under very high
load the callback from isc_socket_connect (resquery_udpconnected) might have
not yet been fired. In this case we only cancel the CONNECT event on socket,
and ignore the SEND that's waiting there (as there is an `else if`).
Then we call dns_dispatch_removeresponse which kills the dispatcher socket
and calls isc_socket_close - but if system is under very high load, the send
we issued earlier might still not be complete - which triggers an assertion
because we're trying to close a socket that's still in use.
The fix is to always check if we have incomplete sends on the socket and cancel
them if we do.
(cherry picked from commit 56183a3917)
Make delv honor the operating system's preferred ephemeral port range
instead of always using the default 1024-65535 range for outgoing
messages.
(cherry picked from commit ada6846a10)
Use them in structs for various rdata types where they are missing.
This doesn't change the structs since we are replacing explicit
uint8_t field types with aliases for uint8_t.
Use dns_dsdigest_t in library function arguments.
Improve dnssec-cds with these more specific types.
(cherry picked from commit 0f219714e1)
During server reconfiguration, plugin instances set up for the old views
are unloaded very close to the end of the whole process, after new
plugin instances are set up. As the log message announcing plugin
unloading is emitted at the default "info" level, the user might be
misled into thinking that it is the new plugin instances that are being
unloaded for some reason, particularly because all other messages logged
at the "info" level around the same time inform about setting things up
rather than tearing them down. Since no distinction is currently made
between destroying a view due to reconfiguration and due to a shutdown
in progress, there is no easy way to vary the contents of the log
message depending on circumstances. Since this message is not a
particularly critical one, demote it to debug level to prevent
confusion.
(cherry picked from commit af4b81f944)
When the "library" part of a "plugin" configuration stanza does not
contain at least one path separator, treat it as a filename and assume
it is a name of a shared object present in the named plugin installation
directory. Absolute and relative paths can still be used and will be
used verbatim. Get the full path to a plugin before attempting to
check/register it so that all relevant log messages include the same
plugin path (apart from the one logged when the full path cannot be
determined).
(cherry picked from commit 1a9fc624ca)
Implement a helper function which, given an input string:
- copies it verbatim if it contains at least one path separator,
- prepends the named plugin installation directory to it otherwise.
This function will allow configuration parsing code to conveniently
determine the full path to a plugin module given either a path or a
filename.
While other, simpler ways exist for making sure filenames passed to
dlopen() cause the latter to look for shared objects in a specific
directory, they are very platform-specific. Using full paths is thus
likely the most portable and reliable solution.
Also added unit tests for ns_plugin_expandpath() to ensure it behaves
as expected for absolute paths, relative paths, and filenames, for
various target buffer sizes.
(Note: plugins share a directory with named on Windows; there is no
default plugin path. Therefore the source path is copied to the
destination path with no modification.)
(cherry picked from commit d181c28c60)
in query_respond_any(), the assumption had previously been made that it
was impossible to get past iterating the node with a return value of
ISC_R_NOMORE but not have found any records, unless we were searching
for RRSIG or SIG. however, it is possible for other types to exist but
be hidden, such as when the zone is transitioning from insecure to
secure and DNSSEC types are encountered, and this situation could
trigger an assertion. removed the assertion and reorganized the code.
(cherry picked from commit 3e74c7e5ff)
In case when a zone fails to load because the file does not exist
or is malformed, we should not run the callback that updates the
zone database when the load is done. This is achieved by
unregistering the callbacks if at zone load end if the result
indicates something else than success.
As pointed out in !813 db_registered is sort of redundant. It is
set to `true` only in `dns_zone_rpz_enable_db()` right before the
`dns_rpz_dbupdate_callback()` callback is registered. It is only
required in that callback and it is the only place that the callback
is registered. Therefore there is no path that that `REQUIRE` can
fail.
The `db_registered` variable is only set to `false` in
`dns_rpz_new_zone`, so it is not like the variable is unset again
later.
The only other place where `db_registered` is checked is in
`rpz_detach()`. If `true`, it will call
`dns_db_updatenotify_unregister()`. However if that happens, the
`db_registered` is not set back to `false` thus this implies that
this may happen multiple times. If called a second time, most
likely the unregister function will return `ISC_R_NOTFOUND`, but
the return value is not checked anyway. So it can do without the
`db_registered` check.
This may happen when loading an RPZ failed and the code path skips
calling dns_db_endload(). The dns_rpz_zone_t object is still kept
marked as having registered db. So when this object is finally
destroyed in rpz_detach(), this code will incorrectly call
`dns_db_updatenotify_unregister()`:
if (rpz->db_registered)
dns_db_updatenotify_unregister(rpz->db,
dns_rpz_dbupdate_callback, rpz);
and trigger this assertion failure:
REQUIRE(db != NULL);
To fix this, only call `dns_db_updatenotify_unregister()` when
`rpz->db` is not NULL.
If `dns_dnssec_keyfromrdata` failed we don't need to call
`dst_key_free` because no `dstkey` was created. Doing so
nevertheless will result in an assertion failure.
This can happen if the key uses an unsupported algorithm.
