Previously, tasks could be created either unbound or bound to a specific
thread (worker loop). The unbound tasks would be assigned to a random
thread every time isc_task_send() was called. Because there's no logic
that would assign the task to the least busy worker, this just creates
unpredictability. Instead of random assignment, bind all the previously
unbound tasks to worker 0, which is guaranteed to exist.
This commit separates TLS context creation code from xfrin_start() as
it has become too large and hard to follow into a new
function (similarly how it is done in dighost.c)
The dead code has been removed from the cleanup section of the TLS
creation code:
* there is no way 'tlsctx' can equal 'found';
* there is no way 'sess_cache' can be non-NULL in the cleanup section.
Also, it fixes a bug in the older version of the code, where TLS
client session context fetched from the cache would not get passed to
isc_nm_tlsdnsconnect().
This commit extends TLS context cache with TLS client session cache so
that an associated session cache can be stored alongside the TLS
context within the context cache.
This commit adds an implementation of a client TLS session cache. TLS
client session cache is an object which allows efficient storing and
retrieval of previously saved TLS sessions so that they can be
resumed. This object is supposed to be a foundation for implementing
TLS session resumption - a standard technique to reduce the cost of
re-establishing a connection to the remote server endpoint.
OpenSSL does server-side TLS session caching transparently by
default. However, on the client-side, a TLS session to resume must be
manually specified when establishing the TLS connection. The TLS
client session cache is precisely the foundation for that.
Clang added support for the gcc-style fallthrough
attribute (i.e. __attribute__((fallthrough))) in version 10. However,
__has_attribute(fallthrough) will return 1 in C mode in older versions,
even though they only support the C++11 fallthrough attribute. At best,
the unsupported attribute is simply ignored; at worst, it causes errors.
The C2x fallthrough attribute has the advantages of being supported in
the broadest range of clang versions (added in version 9) and being easy
to check for support. Use C2x [[fallthrough]] attribute if possible, and
fall back to not using an attribute for clang versions that don't have
it.
Courtesy of Joshua Root
Since the fctx hash table is now self-resizing, and resolver tasks are
selected to match the thread that created the fetch context, there
shouldn't be any significant advantage to having multiple tasks per CPU;
a single task per thread should be sufficient.
Additionally, the fetch context is always pinned to the calling netmgr
thread to minimize the contention just to coalesced fetches - if two
threads starts the same fetch, it will be pinned to the first one to get
the bucket.
After removing the isc_task_onshutdown(), the isc_task_shutdown() and
isc_task_destroy() became obsolete.
Remove calls to isc_task_shutdown() and replace the calls to
isc_task_destroy() with isc_task_detach().
Simplify the internal logic to destroy the task when the last reference
is removed.
The isc_task_onshutdown() was used to post event that should be run when
the task is being shutdown. This could happen explicitly in the
isc_test_shutdown() call or implicitly when we detach the last reference
to the task and there are no more events posted on the task.
This whole task onshutdown mechanism just makes things more complicated,
and it's easier to post the "shutdown" events when we are shutting down
explicitly and the existing code already always knows when it should
shutdown the task that's being used to execute the onshutdown events.
Replace the isc_task_onshutdown() calls with explicit calls to execute
the shutdown tasks.
As we are going to use libuv outside of the netmgr, we need the shims to
be readily available for the rest of the codebase.
Move the "netmgr/uv-compat.h" to <isc/uv.h> and netmgr/uv-compat.c to
uv.c, and as a rule of thumb, the users of libuv should include
<isc/uv.h> instead of <uv.h> directly.
Additionally, merge netmgr/uverr2result.c into uv.c and rename the
single function from isc__nm_uverr2result() to isc_uverr2result().
It used to require two 32-bit integer divisions to get a random number
less than some limit. Now we use Daniel Lemire's "nearly-divisionless"
algorithm for unbiased bounded random numbers, which requires one
64-bit integer multiply in the usual case, and one 32-bit integer
division in rare slow cases. Even the slow cases are faster than
before; there are also fewer branches.
I think this algorithm is exceptionally beautiful. It also has more
clever tricks than lines of code, so I have done my best to explain
how it works.
When TASKMGR_TRACE=1 is defined, the task and event objects have
detailed tracing information about function, file, line, and
backtrace (to the extent tracked by gcc) where it was created.
At exit, when there are unfinished tasks, they will be printed along
with the detailed information.
The only place where isc_task_sendto() was used was in dns_resolver
unit, where the "sendto" part was actually no-op, because dns_resolver
uses bound tasks. Remove the isc_task_sendto() and
isc_task_sendtoanddetach() functions in favor of using bound tasks
create with isc_task_create_bound().
Additionally, cache the number of running netmgr threads (nworkers)
locally to reduce the number of function calls.
For some applications, it's useful to not listen on full battery of
threads. Add workers argument to all isc_nm_listen*() functions and
convenience ISC_NM_LISTEN_ONE and ISC_NM_LISTEN_ALL macros.
This commit adds isc_nmsocket_set_tlsctx() - an asynchronous function
that replaces the TLS context within a given TLS-enabled listener
socket object. It is based on the newly added reference counting
functionality.
The intention of adding this function is to add functionality to
replace a TLS context without recreating the whole socket object,
including the underlying TCP listener socket, as a BIND process might
not have enough permissions to re-create it fully on reconfiguration.
The implementation is done on top of the reference counting
functionality found in OpenSSL/LibreSSL, which allows for avoiding
wrapping the object.
Adding this function allows using reference counting for TLS contexts
in BIND 9's codebase.
Previously, HAVE_SO_REUSEPORT_LB has been defined only in the private
netmgr-int.h header file, making the configuration of load balanced
sockets inoperable.
Move the missing HAVE_SO_REUSEPORT_LB define the isc/netmgr.h and add
missing isc_nm_getloadbalancesockets() implementation.
Previously, the option to enable kernel load balancing of the sockets
was always enabled when supported by the operating system (SO_REUSEPORT
on Linux and SO_REUSEPORT_LB on FreeBSD).
It was reported that in scenarios where the networking threads are also
responsible for processing long-running tasks (like RPZ processing, CATZ
processing or large zone transfers), this could lead to intermitten
brownouts for some clients, because the thread assigned by the operating
system might be busy. In such scenarious, the overall performance would
be better served by threads competing over the sockets because the idle
threads can pick up the incoming traffic.
Add new configuration option (`load-balance-sockets`) to allow enabling
or disabling the load balancing of the sockets.
Previously, the RPZ updates ran quantized on the main nm_worker loops.
As the quantum was set to 1024, this might lead to service
interruptions when large RPZ update was processed.
Change the RPZ update process to run as the offloaded work. The update
and cleanup loops were refactored to do as little locking of the
maintenance lock as possible for the shortest periods of time and the db
iterator is being paused for every iteration, so we don't hold the rbtdb
tree lock for prolonged periods of time.
The reference counting and isc_timer_attach()/isc_timer_detach()
semantic are actually misleading because it cannot be used under normal
conditions. The usual conditions under which is timer used uses the
object where timer is used as argument to the "timer" itself. This
means that when the caller is using `isc_timer_detach()` it needs the
timer to stop and the isc_timer_detach() does that only if this would be
the last reference. Unfortunately, this also means that if the timer is
attached elsewhere and the timer is fired it will most likely be
use-after-free, because the object used in the timer no longer exists.
Remove the reference counting from the isc_timer unit, remove
isc_timer_attach() function and rename isc_timer_detach() to
isc_timer_destroy() to better reflect how the API needs to be used.
The only caveat is that the already executed event must be destroyed
before the isc_timer_destroy() is called because the timer is no longet
attached to .ev_destroy_arg.
Previously, the task privileged mode has been used only when the named
was starting up and loading the zones from the disk as the "first" thing
to do. The privileged task was setup with quantum == 2, which made the
taskmgr/netmgr spin around the privileged queue processing two events at
the time.
The same effect can be achieved by setting the quantum to UINT_MAX (e.g.
practically unlimited) for the loadzone task, hence the privileged task
mode was removed in favor of just processing all the events on the
loadzone task in a single task_run().
After switching to per-thread resources in the zonemgr, the performance
was decreased because the memory context, zonetask and loadtask was
picked from the pool at random.
Pin the zone to single threadid (.tid) and align the memory context,
zonetask and loadtask to be the same, this sets the hard affinity of the
zone to the netmgr thread.
Previously, the zonemgr created 1 task per 100 zones and 1 memory
context per 1000 zones (with minimum 10 tasks and 2 memory contexts) to
reduce the contention between threads.
Instead of reducing the contention by having many resources, create a
per-nm_thread memory context, loadtask and zonetask and spread the zones
between just per-thread resources.
Note: this commit alone does decrease performance when loading the zone
by couple seconds (in case of 1M zone) and thus there's more work in
this whole MR fixing the performance.
Add isc_task_setquantum() function that modifies quantum for the future
isc_task_run() invocations.
NOTE: The current isc_task_run() caches the task->quantum into a local
variable and therefore the current event loop is not affected by any
quantum change.
The isc_task_purge() and isc_task_purgerange() were now unused, so sweep
the task.c file. Additionally remove unused ISC_EVENTATTR_NOPURGE event
attribute.
The isc_task_purgerange() was walking through all events on the task to
find a matching task. Instead use the ISC_LINK_LINKED to find whether
the event is active.
Cleanup the related isc_task_unsend() and isc_task_unsendrange()
functions that were not used anywhere.
Adding extra val & 0xffff in the isc_hash_bits32() macros in the hotpath
has significantly reduced the performance. Turn the macro into static
inline function matching the previous hash_32() function used to compute
hashval matching the hashtable->bits.
Previously, the isc_ht API would always take the key as a literal input
to the hashing function. Change the isc_ht_init() function to take an
'options' argument, in which ISC_HT_CASE_SENSITIVE or _INSENSITIVE can
be specified, to determine whether to use case-sensitive hashing in
isc_hash32() when hashing the key.
Fibonacci hashing was implemented in four separate places (rbt.c,
rbtdb.c, resolver.c, zone.c). This commit combines them into a single
implementation. The hash_32() function is now replaced with
isc_hash_bits32().
This commit adds support for ISC_R_TLSBADPEERCERT error code, which is
supposed to be used to signal for TLS peer certificates verification
in dig and other code.
The support for this error code is added to our TLS and TLS DNS
implementations.
This commit also adds isc_nm_verify_tls_peer_result_string() function
which is supposed to be used to get a textual description of the
reason for getting a ISC_R_TLSBADPEERCERT error.
This commit adds support for keeping CA certificates stores associated
with TLS contexts. The intention is to keep one reusable store per a
set of related TLS contexts.
This commit adds a set of functions that can be used to implement
Strict and Mutual TLS:
* isc_tlsctx_load_client_ca_names();
* isc_tlsctx_load_certificate();
* isc_tls_verify_peer_result_string();
* isc_tlsctx_enable_peer_verification().
This commit adds a set of high-level utility functions to manipulate
the certificate stores. The stores are needed to implement TLS
certificates verification efficiently.
Previously, it was possible to assign a bit of memory space in the
nmhandle to store the client data. This was complicated and prevents
further refactoring of isc_nmhandle_t caching (future work).
Instead of caching the data in the nmhandle, allocate the hot-path
ns_client_t objects from per-thread clientmgr memory context and just
assign it to the isc_nmhandle_t via isc_nmhandle_set().
C11 has builtin support for _Noreturn function specifier with
convenience noreturn macro defined in <stdnoreturn.h> header.
Replace ISC_NORETURN macro by C11 noreturn with fallback to
__attribute__((noreturn)) if the C11 support is not complete.
Previously, the unreachable code paths would have to be tagged with:
INSIST(0);
ISC_UNREACHABLE();
There was also older parts of the code that used comment annotation:
/* NOTREACHED */
Unify the handling of unreachable code paths to just use:
UNREACHABLE();
The UNREACHABLE() macro now asserts when reached and also uses
__builtin_unreachable(); when such builtin is available in the compiler.
Gcc 7+ and Clang 10+ have implemented __attribute__((fallthrough)) which
is explicit version of the /* FALLTHROUGH */ comment we are currently
using.
Add and apply FALLTHROUGH macro that uses the attribute if available,
but does nothing on older compilers.
In one case (lib/dns/zone.c), using the macro revealed that we were
using the /* FALLTHROUGH */ comment in wrong place, remove that comment.
Instead of passing the "workers" variable back and forth along with
passing the single isc_nm_t instance, add isc_nm_getnworkers() function
that returns the number of netmgr threads are running.
Change the ns_interfacemgr and ns_taskmgr to utilize the newly acquired
knowledge.
Previously, an incremental hash table resizing was implemented for the
dns_rbt_t hash table implementation. Using that as a base, also
implement the incremental hash table resizing also for isc_ht API
hashtables:
1. During the resize, allocate the new hash table, but keep the old
table unchanged.
2. In each lookup, delete, or iterator operation, check both tables.
3. Perform insertion operations only in the new table.
4. At each insertion also move <r> elements from the old table to
the new table.
5. When all elements are removed from the old table, deallocate it.
To ensure that the old table is completely copied over before the new
table itself needs to be enlarged, it is necessary to increase the
size of the table by a factor of at least (<r> + 1)/<r> during resizing.
In our implementation <r> is equal to 1.
The downside of this approach is that the old table and the new table
could stay in memory for longer when there are no new insertions into
the hash table for prolonged periods of time as the incremental
rehashing happens only during the insertions.
Change the isc_interval_t implementation from separate data type and
separate implementation to be shim implementation on top of isc_time_t.
The distinction between isc_interval_t and isc_time_t has been kept
because they are semantically different - isc_interval_t is relative and
isc_time_t is absolute, but this allows isc_time_t and isc_interval_t to
be freely interchangeable, f.e. this:
isc_time_t *t1;
isc_interval_t *interval;
isc_time_t *t2;
isc_interval_set(interval, isc_time_seconds(t2), isc_time_nanoseconds(t2);;
isc_time_subtract(t1, interval, t2);
isc_interval_set(interval, isc_time_seconds(t2), isc_time_nanoseconds(t2));
to just:
isc_time_t *t1;
isc_interval_t *interval;
isc_time_t *t2;
isc_time_subtract(t1, t2, interval);
without introducing a whole set of new functions.
The isc_timer_reset() now works only with intervals for once timers.
This makes the API almost 1:1 compatible with the libuv timers making
the further refactoring possible.
The isc_timer_create() function was a bit conflated. It could have been
used to create a timer and start it at the same time. As there was a
single place where this was done before (see the previous commit for
nta.c), this was cleaned up and the isc_timer_create() function was
changed to only create new timer.
