The __builtin_expect() can be used to provide the compiler with branch
prediction information. The Gcc manual says[1] on the subject:
In general, you should prefer to use actual profile feedback for
this (-fprofile-arcs), as programmers are notoriously bad at
predicting how their programs actually perform.
Stop using __builtin_expect() and ISC_LIKELY() and ISC_UNLIKELY() macros
to provide the branch prediction information as the performance testing
shows that named performs better when the __builtin_expect() is not
being used.
1. https://gcc.gnu.org/onlinedocs/gcc/Other-Builtins.html#index-_005f_005fbuiltin_005fexpect
Unify the header guard style and replace the inconsistent include guards
with #pragma once.
The #pragma once is widely and very well supported in all compilers that
BIND 9 supports, and #pragma once was already in use in several new or
refactored headers.
Using simpler method will also allow us to automate header guard checks
as this is simpler to programatically check.
For reference, here are the reasons for the change taken from
Wikipedia[1]:
> In the C and C++ programming languages, #pragma once is a non-standard
> but widely supported preprocessor directive designed to cause the
> current source file to be included only once in a single compilation.
>
> Thus, #pragma once serves the same purpose as include guards, but with
> several advantages, including: less code, avoidance of name clashes,
> and sometimes improvement in compilation speed. On the other hand,
> #pragma once is not necessarily available in all compilers and its
> implementation is tricky and might not always be reliable.
1. https://en.wikipedia.org/wiki/Pragma_once
With isc_mem_get() and dns_name_dup() no longer being able to fail, some
functions can now only return ISC_R_SUCCESS. Change the return type to
void for the following function(s):
* dns_zone_setprimaries()
* dns_zone_setparentals()
* dns_zone_setparentals()
* dns_zone_setalsonotify()
With isc_mem_get() and dns_name_dup() no longer being able to fail, some
functions can now only return ISC_R_SUCCESS. Change the return type to
void for the following function(s):
* dns_view_adddelegationonly()
* dns_view_excludedelegationonly()
With isc_mem_get() and dns_name_dup() no longer being able to fail, some
functions can now only return ISC_R_SUCCESS. Change the return type to
void for the following function(s):
* dns_ssutable_addrule()
* dns_ssutable_create()
* dns_ssutable_createdlz()
With isc_mem_get() and dns_name_dup() no longer being able to fail, some
functions can now only return ISC_R_SUCCESS. Change the return type to
void for the following function(s):
* dns_resolver_addalternate()
With isc_mem_get() and dns_name_dup() no longer being able to fail, some
functions can now only return ISC_R_SUCCESS. Change the return type to
void for the following function(s):
* name_duporclone()
With isc_mem_get() and dns_name_dup() no longer being able to fail, some
functions can now only return ISC_R_SUCCESS. Change the return type to
void for the following function(s):
* build_event()
With isc_mem_get() and dns_name_dup() no longer being able to fail, some
functions can now only return ISC_R_SUCCESS. Change the return type to
void for the following function(s):
* dns_catz_options_copy()
* dns_catz_options_setdefault()
* dns_catz_entry_new()
* dns_catz_entry_copy()
Replace some "master/slave" terminology in the code with the preferred
"primary/secondary" keywords. This also changes user output such as
log messages, and fixes a typo ("seconary") in cfg_test.c.
There are still some references to "master" and "slave" for various
reasons:
- The old syntax can still be used as a synonym.
- The master syntax is kept when it refers to master files and formats.
- This commit replaces mainly keywords that are local. If "master" or
"slave" is used in for example a structure that is all over the
place, it is considered out of scope for the moment.
Originally, the hash table used in RBT database would be resized when it
reached certain number of elements (defined by overcommit). This was
causing resolution brownouts for busy resolvers, because the rehashing
could take several seconds to complete. This was mitigated by
pre-allocating the hash table in the RBT database used for caching to be
large-enough as determined by max-cache-size. The downside of this
solution was that the pre-allocated hash table could take a significant
chunk of the memory even when the resolver cache would be otherwise
empty because the default value for max-cache-size is 90% of available
memory.
Implement incremental resizing[1] to perform the rehashing gradually:
1. During the resize, allocate the new hash table, but keep the old
table unchanged.
2. In each lookup or delete 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.
The upside of this approach is that it's no longer necessary to
pre-allocate large hash table, because the RBT hash table rehashing
doesn't cause resolution brownouts anymore and thus we can use the
memory as needed.
1. https://en.m.wikipedia.org/wiki/Hash_table#Dynamic_resizing
Remove the dynamic registration of result codes. Convert isc_result_t
from unsigned + #defines into 32-bit enum type in grand unified
<isc/result.h> header. Keep the existing values of the result codes
even at the expense of the description and identifier tables being
unnecessary large.
Additionally, add couple of:
switch (result) {
[...]
default:
break;
}
statements where compiler now complains about missing enum values in the
switch statement.
There's value mismatch between the return type of dns_rrl() that's
dns_rrl_result_t and ISC_R_SUCCESS which belongs to isc_result_t. This
works incidentally, because DNS_RRL_RESULT_OK == ISC_R_SUCCESS.
This would break when we change isc_result_t to be static enum in
consecutive commit. Change the value to match the type.
Renamed some functions for clarity and readability:
- dns_dispatch_addresponse() -> dns_dispatch_add()
- dns_dispatch_removeresponse() -> dns_dispatch_done()
The dns_dispatch_cancel() function now calls dns_dispatch_done()
directly, so it is no longer ever necessary to call both functions.
dns_dispatch_cancel() is used to terminate dispatch connections
that are still pending, while dns_dispatch_done() is used when they
are complete.
This commit make the code handling incoming zone transfers to verify
if they are allowed to be done over the underlying connections. As a
result the check ensures that the "dot" ALPN token has been negotiated
over the underlying connection.
as libdns is no longer exported, it's not necessary to have
init and shutdown functions. the only purpose they served
was to create a private mctx and run dst_lib_init(), which
can be called directly instead.
- Prevent shutdown races: attach/detach to dns_resolver in dns_fetch_t
and fctx_t; delay destruction of fctx when finds are still active;
reference the fctx while canceling; reverse the order of
fctx_destroy() and empty_bucket().
- Don't resend queries if fetches have been canceled.
- It's possible for fctx_doshutdown() to run before a TCP connection has
completed. if the query is not on the queries list, then it is not
canceled, but the adbaddrinfo is freed. when tcp_connected() runs
later, the query is in an inconstent state. to fix this, we add the
query to queries before running dns_dispatch_connect(), instead of in
the connect callback.
- Combined the five fctx_cleanup* functions into a single one.
- Added comments and changed some names to make this code easier to
understand.
udp_recv() will call dispatch_getnext() if the message received is
invalid or doesn't match; we need to reduce the timeout each time this
happens so we can't be starved forever by someone sending garbage
packets.
- disp_connected() has been split into two functions,
udp_connected() (which takes 'resp' as an argument) and
tcp_connected() (which takes 'disp', and calls the connect callbacks
for all pending resps).
- In dns_dispatch_connect(), if a connection is already open, we need to
detach the dispentry immediately because we won't be running
tcp_connected().
- dns_disptach_cancel() also now calls the connect callbacks for pending
TCP responses, and the response callbacks for open TCP connections
waiting on read.
- If udp_connected() runs after dns_dispatch_cancel() has been called,
ensure that the caller's connect callback is run.
- If a UDP connection fails with EADDRINUSE, we try again up to five
times with a different local port number before giving up.
- If a TCP connection is canceled while still pending connection, the
connect timeout may still fire. we attach the dispatch before
connecting to ensure that it won't be detached too soon in this case.
- The dispentry is no longer removed from the pending list when
deactivating, so that the connect callback can still be run if
dns_dispatch_removeresponse() was run while the connecting was
pending.
- Rewrote dns_dispatch_gettcp() to avoid a data race.
- startrecv() and dispatch_getnext() can be called with a NULL resp when
using TCP.
- Refactored udp_recv() and tcp_recv() and added result logging.
- EOF is now treated the same as CANCELED in response callbacks.
- ISC_R_SHUTTINGDOWN is sent to the reponse callbacks for all resps if
tcp_recv() is triggered by a netmgr shutdown. (response callbacks
are *not* sent by udp_recv() in this case.)
- startrecv() and getnext() have been rewritten.
- Don't set TCP flag when connecting a UDP dispatch.
- Prevent TCP connections from trying to connect twice.
- dns_dispatch_gettcp() can now find a matching TCP dispatch that has
not yet fully connected, and attach to it. when the connection is
completed, the connect callbacks are run for all of the pending
entries.
- An atomic 'state' variable is now used for connection state instead of
attributes.
- When dns_dispatch_cancel() is called on a TCP dispatch entry, only
that one entry is canceled. the dispatch itself should not be shut
down until there are no dispatch entries left associated with it.
- Other incidental cleanup, including removing DNS_DISPATCHATTR_IPV4 and
_IPV6 (they were being set in the dispatch attributes but never used),
cleaning up dns_requestmgr_create(), and renaming dns_dispatch_read()
to the more descriptive dns_dispatch_resume().
- It is no longer necessary to pass a 'timeout' callback to
dns_dispatch_addresponse(); timeouts are handled directly by the
'response' callback instead.
- The netmgr handle is no longer passed to dispatch callbacks, since
they don't (and can't) use it. instead, dispatch_cb_t now takes a
result code, region, and argument.
- Cleaned up timeout-related tests in dispatch_test.c
- Responses received by the dispatch are no longer sent to the caller
via a task event, but via a netmgr-style recv callback. the 'action'
parameter to dns_dispatch_addresponse() is now called 'response' and
is called directly from udp_recv() or tcp_recv() when a valid response
has been received.
- All references to isc_task and isc_taskmgr have been removed from
dispatch functions.
- All references to dns_dispatchevent_t have been removed and the type
has been deleted.
- Added a task to the resolver response context, to be used for fctx
events.
- When the caller cancels an operation, the response handler will be
called with ISC_R_CANCELED; it can abort immediately since the caller
will presumably have taken care of cleanup already.
- Cleaned up attach/detach in resquery and request.
Remove the debugging printfs. (leaving this as a separate commit rather
than squashing it so we can restore it in the future if we ever need it
again.)
Since every dispsock was associated with a dispentry anyway (though not
always vice versa), the members of dispsock have been combined into
dispentry, which is now reference-counted. dispentry objects are now
attached before connecting and detached afterward to prevent races
between the connect callback and dns_dispatch_removeresponse().
Dispatch and dispatchmgr objects are now reference counted as well, and
the shutdown process has been simplified. reference counting of
resquery and request objects has also been cleaned up significantly.
dns_dispatch_cancel() now flags a dispentry as having been canceled, so
that if the connect callback runs after cancellation, it will not
initiate a read.
The isblackholed() function has been simplified.
- The `timeout_action` parameter to dns_dispatch_addresponse() been
replaced with a netmgr callback that is called when a dispatch read
times out. this callback may optionally reset the read timer and
resume reading.
- Added a function to convert isc_interval to milliseconds; this is used
to translate fctx->interval into a value that can be passed to
dns_dispatch_addresponse() as the timeout.
- Note that netmgr timeouts are accurate to the millisecond, so code to
check whether a timeout has been reached cannot rely on microsecond
accuracy.
- If serve-stale is configured, then a timeout received by the resolver
may trigger it to return stale data, and then resume waiting for the
read timeout. this is no longer based on a separate stale timer.
- The code for canceling requests in request.c has been altered so that
it can run asynchronously.
- TCP timeout events apply to the dispatch, which may be shared by
multiple queries. since in the event of a timeout we have no query ID
to use to identify the resp we wanted, we now just send the timeout to
the oldest query that was pending.
- There was some additional refactoring in the resolver: combining
fctx_join() and fctx_try_events() into one function to reduce code
duplication, and using fixednames in fetchctx and fetchevent.
- Incidental fix: new_adbaddrinfo() can't return NULL anymore, so the
code can be simplified.
The flow of operations in dispatch is changing and will now be similar
for both UDP and TCP queries:
1) Call dns_dispatch_addresponse() to assign a query ID and register
that we'll be listening for a response with that ID soon. the
parameters for this function include callback functions to inform the
caller when the socket is connected and when the message has been
sent, as well as a task action that will be sent when the response
arrives. (later this could become a netmgr callback, but at this
stage to minimize disruption to the calling code, we continue to use
isc_task for the response event.) on successful completion of this
function, a dispatch entry object will be instantiated.
2) Call dns_dispatch_connect() on the dispatch entry. this runs
isc_nm_udpconnect() or isc_nm_tcpdnsconnect(), as needed, and begins
listening for responses. the caller is informed via a callback
function when the connection is established.
3) Call dns_dispatch_send() on the dispatch entry. this runs
isc_nm_send() to send a request.
4) Call dns_dispatch_removeresponse() to terminate listening and close
the connection.
Implementation comments below:
- As we will be using netmgr buffers now. code to send the length in
TCP queries has also been removed as that is handled by the netmgr.
- TCP dispatches can be used by multiple simultaneous queries, so
dns_dispatch_connect() now checks whether the dispatch is already
connected before calling isc_nm_tcpdnsconnect() again.
- Running dns_dispatch_getnext() from a non-network thread caused a
crash due to assertions in the netmgr read functions that appear to be
unnecessary now. the assertions have been removed.
- fctx->nqueries was formerly incremented when the connection was
successful, but is now incremented when the query is started and
decremented if the connection fails.
- It's no longer necessary for each dispatch to have a pool of tasks, so
there's now a single task per dispatch.
- Dispatch code to avoid UDP ports already in use has been removed.
- dns_resolver and dns_request have been modified to use netmgr callback
functions instead of task events. some additional changes were needed
to handle shutdown processing correctly.
- Timeout processing is not yet fully converted to use netmgr timeouts.
- Fixed a lock order cycle reported by TSAN (view -> zone-> adb -> view)
by by calling dns_zt functions without holding the view lock.
We now use dns_dispatch_cancel() for this purpose. NOTE: The caller
still has to track whether there are pending send or connect events in
the dispatch or dispatch entry; later this should be moved into the
dispatch module as well.
Also removed some public dns_dispatch_*() API calls that are no longer
used outside dispatch itself.
dns_dispatch_connect() connects a dispatch socket (for TCP) or a
dispatch entry socket (for UDP). This is the next step in moving all
uses of the isc_socket code into the dispatch module.
This API is temporary; it needs to be cleaned up further so that it can
be called the same way for both TCP and UDP.
Continuing the effort to move all uses of the isc_socket API into
dispatch.c, this commit removes the dns_tcpmsg module entirely, as
dispatch was its only caller, and moves the parts of its functionality
that were being used into the dispatch module.
This code will be removed when we switch to using netmgr TCPDNS.
Previously, creation of TCP dispatches differed from UDP in that a TCP
dispatch was created to attach to an existing socket, whereas a UDP
dispatch would be created in a vacuum and sockets would be opened on
demand when a transaction was initiated.
We are moving as much socket code as possible into the dispatch module,
so that it can be replaced with a netmgr version as easily as
possible. (This will also have the side effect of making TCP and UDP
dispatches more similar.)
As a step in that direction, this commit changes
dns_dispatch_createtcp() so that it creates the TCP socket.
- Many dispatch attributes can be set implicitly instead of being passed
in. we can infer whether to set DNS_DISPATCHATTR_TCP or _UDP from
whether we're calling dns_dispatch_createtcp() or _createudp(). we
can also infer DNS_DISPATCHATTR_IPV4 or _IPV6 from the addresses or
the socket that were passed in.
- We no longer use dup'd sockets in UDP dispatches, so the 'dup_socket'
parameter has been removed from dns_dispatch_createudp(), along with
the code implementing it. also removed isc_socket_dup() since it no
longer has any callers.
- The 'buffersize' parameter was ignored and has now been removed;
buffersize is now fixed at 4096.
- Maxbuffers and maxrequests don't need to be passed in on every call to
dns_dispatch_createtcp() and _createudp().
In all current uses, the value for mgr->maxbuffers will either be
raised once from its default of 20000 to 32768, or else left
alone. (passing in a value lower than 20000 does not lower it.) there
isn't enough difference between these values for there to be any need
to configure this.
The value for disp->maxrequests controls both the quota of concurrent
requests for a dispatch and also the size of the dispatch socket
memory pool. it's not clear that this quota is necessary at all. the
memory pool size currently starts at 32768, but is sometimes lowered
to 4096, which is definitely unnecessary.
This commit sets both values permanently to 32768.
- Previously TCP dispatches allocated their own separate QID table,
which didn't incorporate a port table. this commit removes
per-dispatch QID tables and shares the same table between all
dispatches. since dispatches are created for each TCP socket, this may
speed up the dispatch allocation process. there may be a slight
increase in lock contention since all dispatches are sharing a single
QID table, but since TCP sockets are used less often than UDP
sockets (which were already sharing a QID table), it should not be a
substantial change.
- The dispatch port table was being used to determine whether a port was
already in use; if so, then a UDP socket would be bound with
REUSEADDR. this commit removes the port table, and always binds UDP
sockets that way.
Currently the netmgr doesn't support unconnected, shared UDP sockets, so
there's no reason to retain that functionality in the dispatcher prior
to porting to the netmgr.
In this commit, the DNS_DISPATCHATTR_EXCLUSIVE attribute has been
removed as it is now non-optional; UDP dispatches are alwasy exclusive.
Code implementing non-exclusive UDP dispatches has been removed.
dns_dispatch_getentrysocket() now always returns the dispsocket for UDP
dispatches and the dispatch socket for TCP dispatches.
There is no longer any need to search for existing dispatches from
dns_dispatch_getudp(), so the 'mask' option has been removed, and the
function renamed to the more descriptive dns_dispatch_createudp().
- style cleanup
- removed NULL checks in places where they are not currently needed
- use isc_refcount for dispatch reference counting
- revised code flow for readability
- remove some #ifdefs that are no longer relevant
- remove unused struct members
- removed unnecessary function parameters
- use C99 struct initialization
The DNS_REQUESTOPT_SHARE flag was added when client-side pipelining of
TCP queries was implemented. there was no need to make it optional;
forcing it to be in effect for all requests simplfiies the code.
- UDP buffersize is now established when creating dispatch manager
and is always set to 4096.
- Set up the default port range in dispatchmgr before setting the magic
number.
- Magic is not set until dispatchmgr is fully created.
- DNS_DISPATCHATTR_CANREUSE was never set. the code that implements it
has been removed.
- DNS_DISPATCHOPT_FIXEDID and DNS_DISPATCHATTR_FIXEDID were both
defined, but only the DISPATCHOPT was ever set; it appears the
DISPATCHATTR was added accidentally.
- DNS_DISPATCHATTR_NOLISTEN was set but never used.
Previously a missing/deleted zone which was referenced by a catalog
zone was causing a crash when doing a reload.
This commit will make `named` to ignore the fact that the zone is
missing, and make sure to restore it later on.
