loom

Loom is a high performance, cross platform fiber event loop library for C++20.

This abstraction serves the main purpose: To not think about IO readiness events, but just IO completion. What do I mean by this? The io_uring and epoll interface are quite different. io_uring takes the SQE, together with a callback, and the kernel itself will perform the read for e.g. epoll on the other hand, only notifies you of the readiness, and you would need to perform the read yourself.

This should be abstracted away. The user using this framework should not have to think about IO readiness events, and the underlying implementation of the asynchronous interface.

Usage

Loom requires that you hook your main method (e.g the same watch Catch2 does) in order to intercept system calls before glibc functions are invoked. This has various purposes, mainly for proper scheduling behaviour and telemetry. For example, we don't want a sleep in the fiber thread to block the entire thread.

Note

Note that all macros in loom start with $.

#include <loom/loom.h>

$LOOM_MAIN(int argc, char* argv[]) {
	int ret = loom::init(loom::EVENT_ENGINE_KQUEUE);
	LOOM_ASSERT(ret == 0, "Failed to initialize loom");
	defer(loom::fini());
}

Features

• Cross platform: kqueue, epoll
• glibc system call hooks
• Fast context switches (powered by Boost.Context)
• (Eventually) No runtime allocations- Intrusive pointers
• Supports unix domain sockets
• Support for CPU affinity and pinning threads to cores

Resources

• Tiger Style
• ⭐Fibers
• Writing high-performance servers
• kqueue documentation
• epoll documentation

Roadmap

• Inject tracy profiling before functions during build step using AST transformations
• Use eventfd for shared mem OB?