Multicore

Detailed Description

Support for exploiting multiple cores on a server.

Seastar supports multicore servers by using sharding. Each logical core (lcore) runs a separate event loop, with its own memory allocator, TCP/IP stack, and other services. Shards communicate by explicit message passing, rather than using locks and condition variables as with traditional threaded programming.

Classes
class	seastar::sharded< Service >
class	seastar::async_sharded_service< T >
class	seastar::peering_sharded_service< Service >
	Provide a sharded service with access to its peers. More...
class	seastar::no_sharded_instance_exception
	Exception thrown when a sharded object does not exist. More...
class	seastar::sharded_parameter< Func, Params >
	Helper to pass a parameter to a sharded<> object that depends on the shard. It is evaluated on the shard, just before being passed to the local instance. It is useful when passing parameters to sharded::start(). More...
class	seastar::foreign_ptr< PtrType >

Functions
template<typename T >
foreign_ptr< T >	make_foreign (T ptr)

Function Documentation

make_foreign()

template<typename T >

foreign_ptr< T > make_foreign ( T  ptr )

related

Wraps a raw or smart pointer object in a foreign_ptr<>.