This "quiescent state based reclamation" module provides support for
the qp-trie module in dns/qp. It is a replacement for liburcu, written
without reference to the urcu source code, and in fact it works in a
significantly different way.
A few specifics of BIND make this variant of QSBR somewhat simpler:
* We can require that wait-free access to a qp-trie only happens in
an isc_loop callback. The loop provides a natural quiescent state,
after the callbacks are done, when no qp-trie access occurs.
* We can dispense with any API like rcu_synchronize(). In practice,
it takes far too long to wait for a grace period to elapse for each
write to a data structure.
* We use the idea of "phases" (aka epochs or eras) from EBR to
reduce the amount of bookkeeping needed to track memory that is no
longer needed, knowing that the qp-trie does most of that work
already.
I considered hazard pointers for safe memory reclamation. They have
more read-side overhead (updating the hazard pointers) and it wasn't
clear to me how to nicely schedule the cleanup work. Another
alternative, epoch-based reclamation, is designed for fine-grained
lock-free updates, so it needs some rethinking to work well with the
heavily read-biased design of the qp-trie. QSBR has the fastest read
side of the basic SMR algorithms (with no barriers), and fits well
into a libuv loop. More recent hybrid SMR algorithms do not appear to
have enough benefits to justify the extra complexity.
the isc_glob module was originally needed to support posix-style glob
processing on Windows, but is now just an unnecessary wrapper around
glob(3). this commit removes it.
Add a singly-linked stack that supports lock-free prepend and drain (to
empty the list and clean up its elements). Intended for use with QSBR
to collect objects that need safe memory reclamation, or any other user
that works with adding objects to the stack and then draining them in
one go like various work queues.
In <isc/atomic.h>, add an `atomic_ptr()` macro to make type
declarations a little less abominable, and clean up a duplicate
definition of `atomic_compare_exchange_strong_acq_rel()`
removed references in code comments, doc/dev documentation, etc, to
isc_task, isc_timer_reset(), and isc_timertype_inactive. also removed a
coccinelle patch related to isc_timer_reset() that was no longer needed.
as there is no further use of isc_task in BIND, this commit removes
it, along with isc_taskmgr, isc_event, and all other related types.
functions that accepted taskmgr as a parameter have been cleaned up.
as a result of this change, some functions can no longer fail, so
they've been changed to type void, and their callers have been
updated accordingly.
the tasks table has been removed from the statistics channel and
the stats version has been updated. dns_dyndbctx has been changed
to reference the loopmgr instead of taskmgr, and DNS_DYNDB_VERSION
has been udpated as well.
change functions using isc_taskmgr_beginexclusive() to use
isc_loopmgr_pause() instead.
also, removed an unnecessary use of exclusive mode in
named_server_tcptimeouts().
most functions that were implemented as task events because they needed
to be running in a task to use exclusive mode have now been changed
into loop callbacks instead. (the exception is catz, which is being
changed in a separate commit because it's a particularly complex change.)
This changes the internal isc_rwlock implementation to:
Irina Calciu, Dave Dice, Yossi Lev, Victor Luchangco, Virendra
J. Marathe, and Nir Shavit. 2013. NUMA-aware reader-writer locks.
SIGPLAN Not. 48, 8 (August 2013), 157–166.
DOI:https://doi.org/10.1145/2517327.24425
(The full article available from:
http://mcg.cs.tau.ac.il/papers/ppopp2013-rwlocks.pdf)
The implementation is based on the The Writer-Preference Lock (C-RW-WP)
variant (see the 3.4 section of the paper for the rationale).
The implemented algorithm has been modified for simplicity and for usage
patterns in rbtdb.c.
The changes compared to the original algorithm:
* We haven't implemented the cohort locks because that would require a
knowledge of NUMA nodes, instead a simple atomic_bool is used as
synchronization point for writer lock.
* The per-thread reader counters are not being used - this would
require the internal thread id (isc_tid_v) to be always initialized,
even in the utilities; the change has a slight performance penalty,
so we might revisit this change in the future. However, this change
also saves a lot of memory, because cache-line aligned counters were
used, so on 32-core machine, the rwlock would be 4096+ bytes big.
* The readers use a writer_barrier that will raise after a while when
readers lock can't be acquired to prevent readers starvation.
* Separate ingress and egress readers counters queues to reduce both
inter and intra-thread contention.
Unfortunately, C still lacks a standard function for pause (x86,
sparc) or yeild (arm) instructions, for use in spin lock or CAS loops.
BIND has its own based on vendor intrinsics or inline asm.
Previously, it was buried in the `isc_rwlock` implementation. This
commit renames `isc_rwlock_pause()` to `isc_pause()` and moves
it into <isc/pause.h>.
This commit also fixes the configure script so that it detects ARM
yield support on systems that identify as `aarch*` instead of `arm*`.
On 64-bit ARM systems we now use the ISB (instruction synchronization
barrier) instruction in preference to yield. The ISB instruction
pauses the CPU for longer, several nanoseconds, which is more like the
x86 pause instruction. There are more details in a Rust pull request,
which also refers to MySQL making the same change:
https://github.com/rust-lang/rust/pull/84725
removed some functions that are no longer used and unlikely to
be resurrected, and also some that were only used to support Windows
and can now be replaced with generic versions.
isc_bind9 was a global bool used to indicate whether the library
was being used internally by BIND or by an external caller. external
use is no longer supported, but the variable was retained for use
by dyndb, which needed it only when being built without libtool.
building without libtool is *also* no longer supported, so the variable
can go away.
the rate limter now uses loop callbacks rather than task events.
the API for isc_ratelimiter_enqueue() has been changed; we now pass
in a loop, a callback function and a callback argument, and
receive back a rate limiter event object (isc_rlevent_t). it
is no longer necessary for the caller to allocate the event.
the callback argument needs to include a pointer to the rlevent
object so that it can be freed using isc_rlevent_free(), or by
dequeueing.
The total memory counter had again little or no meaning when we removed
the internal memory allocator. It was just a monotonic counter that
would count add the allocation sizes but never subtracted anything, so
it would be just a "big number".
The maxinuse memory counter indicated the highest amount of
memory allocated in the past. Checking and updating this high-
water mark value every time memory was allocated had an impact
on server performance, so it has been removed. Memory size can
be monitored more efficiently via an external tool logging RSS.
The malloced and maxmalloced memory counters were mostly useless since
we removed the internal allocator blocks - it would only differ from
inuse by the memory context size itself.
Return 'isc_result_t' type value instead of 'bool' to indicate
the actual failure. Rename the function to something not suggesting
a boolean type result. Make changes in the places where the API
function is being used to check for the result code instead of
a boolean value.
As this code is on hot path (NSEC3) this introduces an additional
optimization of the EVP_MD API - instead of calling EVP_MD_CTX_new() on
every call to isc_iterated_hash(), we create two thread_local objects
for each thread - a basectx and mdctx, initialize basectx once and then
use EVP_MD_CTX_copy_ex() to flip the initialized state into mdctx. This
saves us couple more valuable microseconds from the isc_iterated_hash()
call.
The implicit algorithm fetch causes a lock contention and significant
slowdown for small input buffers. For more details, see:
https://github.com/openssl/openssl/issues/19612
Instead of using EVP_DigestInit_ex() initialize empty MD_CTX objects for
each algorithm and use EVP_MD_CTX_copy_ex() to initialize MD_CTX from a
static copy. Additionally avoid implicit algorithm fetching by using
EVP_MD_fetch() for OpenSSL 3.0.
Prefer the pthread_barrier implementation on platforms where it is
available over uv_barrier implementation. This also solves the problem
with thread sanitizer builds on macOS that doesn't have pthread barrier.
Always track the per-worker sockets in the .active_sockets field in the
isc__networker_t struct and always track the per-socket handles in the
.active_handles field ian the isc_nmsocket_t struct.
DSCP has not been fully working since the network manager was
introduced in 9.16, and has been completely broken since 9.18.
This seems to have caused very few difficulties for anyone,
so we have now marked it as obsolete and removed the
implementation.
To ensure that old config files don't fail, the code to parse
dscp key-value pairs is still present, but a warning is logged
that the feature is obsolete and should not be used. Nothing is
done with configured values, and there is no longer any
range checking.
Additionally to renaming, it changes the function definition so that
it accepts a pointer to pointer instead of returning a pointer to the
new object.
It is mostly done to make it in line with other functions in the
module.
Additionally to renaming, it changes the function definition so that
it accepts a pointer to pointer instead of returning a pointer to the
new object.
It is mostly done to make it in line with other functions in the
module.
This commit unties generic TLS code (isc_nm_tlssocket) from DoH, so
that it will be available regardless of the fact if BIND was built
with DNS over HTTP support or not.
This commit adds ability to turn the Nagle's algorithm on or off via
connections handle. It adds the isc_nmhandle_set_tcp_nodelay()
function as the public interface for this functionality.
This commit adds an initial implementation of isc_nm_streamdnssocket
transport: a unified transport for DNS over stream protocols messages,
which is capable of replacing both TCP DNS and TLS DNS
transports. Currently, the interface it provides is a unified set of
interfaces provided by both of the transports it attempts to replace.
The transport is built around "isc_dnsbuffer_t" and
"isc_dnsstream_assembler_t" objects and attempts to minimise both the
number of memory allocations during network transfers as well as
memory usage.
This commit adds the implementation for an "isc_dnsstream_assembler_t"
object. The object is built on top of "isc_dnsbuffer_t" and is
intended to encapsulate the state machine used for handling DNS
messages received in the format used for messages transmitted over
TCP.
The idea is that the object accepts the input data received from a
socket, tries to assemble DNS messages from the incoming data and
calls the callback which contains the status of the incoming data as
well as a pointer to the memory region referencing the data of the
assembled message. It is capable of assembling DNS messages no matter
how torn apart they are when sent over network.
The following statuses might be passed to the callback:
* ISC_R_SUCCESS - a message has been successfully assembled;
* ISC_R_NOMORE - not enough data has been processed to assemble a
message;
* ISC_R_RANGE - there was an attempt to process a zero-sized DNS
message (someone attempts to send us junk data).
One could say that the object replaces the implementation of
"isc__nm_*_processbuffer()" functions used by the old TCP DNS and TLS
DNS transports with a better defined state machine completely
decoupled from the networking code itself.
Such a design makes it trivial to write unit tests for it, leading to
better verification of its correctness.
Another important difference is directly related to the fact that it
is built on top of "isc_dnsbuffer_t", which tries to manage memory in
a smart way. In particular:
* It tries to use a static buffer for smaller messages, reducing
pressure on the memory manager (hot path);
* When allocating dynamic memory for larger messages, it tries to
allocate memory conservatively (generic path).
These characteristics is a significant upgrade over the older logic
where a 64KB(+2 bytes) buffer was allocated from dynamic memory
regardless of the fact if we need a buffer this large or not. That is,
lesser memory usage is expected in a generic case for DNS transports
built on top of "isc_dnsstream_assembler_t."
This commit adds "isc_dnsbuffer_t" object implementation, a thin
wrapper on top of "isc_buffer_t" which has the following
characteristics:
* provides interface specifically atuned for handling/generating DNS
messages, especially in the format used for DNS messages over TCP;
* avoids allocating dynamic memory when handling small DNS messages,
while transparently switching to using dynamic memory when handling
larger messages. This approach significantly reduces pressure on the
memory allocator, as most of the DNS messages are small.
The purpose of this commit is to aid compiler in generating better
code when working with `isc_buffer_t` objects by using restricted
pointers (and, to a lesser extent, 'const' modifier for read-only
arguments).
This way we, basically, instruct the compiler that the members of
structured passed by pointers into the functions can be treated as
local variables in the scope of a function. That should reduce the
number of load/store operations emitted by compilers when accessing
objects (e.g. 'isc_buffer_t') via pointers.
Add two extra functions needed by StreamDNS:
1. isc_buffer_setmctx() sets the buffer internal memory context, so we
can use isc_buffer_reserve() on the buffer. For this, we also need
to track whether the .base was dynamically allocated or not. This
needs to be called after isc_buffer_init() and before first
isc_buffer_reserve() call.
2. isc_buffer_clearmctx() clears the buffer internal memory context, and
frees any dynamically allocated buffer. This needs to be called
after the last isc_buffer_reserve() call and before calling the
isc_buffer_invalidate()
