MySQL 8.1.0
Source Code Documentation
s_mysql_reference_caching_channel Struct Reference

A reference caching channel service. More...

#include <reference_caching.h>

Public Attributes

mysql_service_status_t(* create )(const char *service_names[], reference_caching_channel *out_channel)
 Creates a channel and returns a handle for it. More...
mysql_service_status_t(* destroy )(reference_caching_channel channel)
 Destroys a channel handle. More...
mysql_service_status_t(* invalidate )(reference_caching_channel channel)
 Invalidate a channel. More...

Detailed Description

A reference caching channel service.

The goal of this service is to ensure that event producers spend a MINIMAL amount of time in the emitting code when there are no consumers of the produced events.


An event consumer is an implementation of a component service. An event producer is a piece of code that wants to inform all currently registered event consumers about an event. A channel is a device that serves as a singleton of all reference caches for all threads that are to produce events. A reference cache is something each thread producing events must maintain for its lifetime and use it to find the references to event consumers it needs to call.

Typical lifetime of the event consumers

At init time the event consumer will register implementations of any of the services it's interested in receiving notifications for.

Then optionally it might force a channel invalidation to make sure all existing event producers will start sending notifications to it immediately. Or it can just sit and wait for the natural producer's reference cache refresh cycles to kick in.

Now it is receiving notifications from all event producers as they come.

When it wishes to no longer receive notifications it needs to mark itself as invisible for reference cache fill-ins. In this way all reference cache flushes will not pick this implementation up even if it's still registered in the registry.

Eventually all active references will be removed by the natural producers flush cycles. The consumer may choose to expedite this by triggering a channel invalidation.

As with all service implementations, when all references to the services are released it can unload.

Typical lifetime of the event producers

An event producer will usually start at init time by creating channel(s).

Then, for each thread wishing to produce events, a reference cache must be created and maintained until the thread will no longer be producing events.

Now the thread can produce events using the reference cache. This is done by calling the get method and then iterating over the resulting set of references and calling each one in turn as one would normally do for registry service references. It is assumed that the references are owned by the cache and thus they should not be released.

With some cyclical (e.g. at the end of each statement or something) the event producing thread needs to flush the cache. This is to ensure that references to event consumers are not held for very long and that new event consumers are picked up. However flushing the cache is a relatively expensive operation and thus a balance between the number of events produced and the cache being flushed must be achieved.

General remarks

Channels are something each event producer must have to produce events. Channels are to be created by a single thread before the first event is ever produced. And, once created they are to be kept until after the last event is produced.

Channels serve as singletons for caches and you only need one channel instance per event producer component. There usually will be multiple caches (one per event producing thread) per channel.

Creating and destroying a channel is a relatively "expensive" operation that might involve some synchronization and should not be done frequently.

Channels exist to allow a non-related thread to trigger invalidation of all currently active caches on that channel. This is necessary when for example event consumers are registered and are about to be removed.

Invalidating a channel is a thread-safe operation that can be invoked without synchronization at any time.

Each channel is associated with a specific set of service names.

It is a set of service names, not implementation names !

The names are stored and used in event caches to handle all implementations of that particular service.

Member Data Documentation

◆ create

mysql_service_status_t(* s_mysql_reference_caching_channel::create) (const char *service_names[], reference_caching_channel *out_channel)

Creates a channel and returns a handle for it.

The handle created by this method must be destroyed by the invalidate method otherwise there might be leaks.

The channel should be created before any events are to be produced on it.

service_namesa list of service names that this channel will operate on. Terminated by a NULL pointer
[out]out_channelplaceholder for the newly created handle.
Return values

◆ destroy

mysql_service_status_t(* s_mysql_reference_caching_channel::destroy) (reference_caching_channel channel)

Destroys a channel handle.

Should make sure no other thread is using the handle and no caches are allocated on it.

channelthe handle to destroy
Return values

◆ invalidate

mysql_service_status_t(* s_mysql_reference_caching_channel::invalidate) (reference_caching_channel channel)

Invalidate a channel.

This is thread safe to call without synchronization and relatively fast.

Forces an asynchronous flush on all caches that are allocated on that channel when they're next accessed.

channelthe handle to destroy
Return values

The documentation for this struct was generated from the following file: