numam-spdk/doc/concurrency.md

Message Passing and Concurrency

Theory

One of the primary aims of SPDK is to scale linearly with the addition of hardware. This can mean a number of things in practice. For instance, moving from one SSD to two should double the number of I/O's per second. Or doubling the number of CPU cores should double the amount of computation possible. Or even doubling the number of NICs should double the network throughput. To achieve this, the software must be designed such that threads of execution are independent from one another as much as possible. In practice, that means avoiding software locks and even atomic instructions.

Traditionally, software achieves concurrency by placing some shared data onto the heap, protecting it with a lock, and then having all threads of execution acquire the lock only when that shared data needs to be accessed. This model has a number of great properties:

• It's relatively easy to convert single-threaded programs to multi-threaded programs because you don't have to change the data model from the single-threaded version. You just add a lock around the data.
• You can write your program as a synchronous, imperative list of statements that you read from top to bottom.
• Your threads can be interrupted and put to sleep by the operating system scheduler behind the scenes, allowing for efficient time-sharing of CPU resources.

Unfortunately, as the number of threads scales up, contention on the lock around the shared data does too. More granular locking helps, but then also greatly increases the complexity of the program. Even then, beyond a certain number highly contended locks, threads will spend most of their time attempting to acquire the locks and the program will not benefit from any additional CPU cores.

SPDK takes a different approach altogether. Instead of placing shared data in a global location that all threads access after acquiring a lock, SPDK will often assign that data to a single thread. When other threads want to access the data, they pass a message to the owning thread to perform the operation on their behalf. This strategy, of course, is not at all new. For instance, it is one of the core design principles of Erlang and is the main concurrency mechanism in Go. A message in SPDK typically consists of a function pointer and a pointer to some context, and is passed between threads using a lockless ring. Message passing is often much faster than most software developer's intuition leads them to believe, primarily due to caching effects. If a single core is consistently accessing the same data (on behalf of all of the other cores), then that data is far more likely to be in a cache closer to that core. It's often most efficient to have each core work on a relatively small set of data sitting in its local cache and then hand off a small message to the next core when done.

In more extreme cases where even message passing may be too costly, a copy of the data will be made for each thread. The thread will then only reference its local copy. To mutate the data, threads will send a message to each other thread telling them to perform the update on their local copy. This is great when the data isn't mutated very often, but may be read very frequently, and is often employed in the I/O path. This of course trades memory size for computational efficiency, so it's use is limited to only the most critical code paths.

Message Passing Infrastructure

SPDK provides several layers of message passing infrastructure. The most fundamental libraries in SPDK, for instance, don't do any message passing on their own and instead enumerate rules about when functions may be called in their documentation (e.g. @ref nvme). Most libraries, however, depend on SPDK's thread abstraction, located in libspdk_thread.a. The thread abstraction provides a basic message passing framework and defines a few key primitives.

First, spdk_thread is an abstraction for a thread of execution and spdk_poller is an abstraction for a function that should be periodically called on the given thread. On each system thread that the user wishes to use with SPDK, they must first call spdk_allocate_thread(). This function takes three function pointers - one that will be called to pass a message to this thread, one that will be called to request that a poller be started on this thread, and finally one to request that a poller be stopped. The implementation of these functions is not provided by this library. Many applications already have facilities for passing messages, so to ease integration with existing code bases we've left the implementation up to the user. However, for users starting from scratch, see the following section on the event framework for an SPDK-provided implementation.

The library also defines two other abstractions: spdk_io_device and spdk_io_channel. In the course of implementing SPDK we noticed the same pattern emerging in a number of different libraries. In order to implement a message passing strategy, the code would describe some object with global state and also some per-thread context associated with that object that was accessed in the I/O path to avoid locking on the global state. The pattern was clearest in the lowest layers where I/O was being submitted to block devices. These devices often expose multiple queues that can be assigned to threads and then accessed without a lock to submit I/O. To abstract that, we generalized the device to spdk_io_device and the thread-specific queue to spdk_io_channel. Over time, however, the pattern has appeared in a huge number of places that don't fit quite so nicely with the names we originally chose. In today's code spdk_io_device is any pointer, whose uniqueness is predicated only on its memory address, and spdk_io_channel is the per-thread context associated with a particular spdk_io_device.

The threading abstraction provides functions to send a message to any other thread, to send a message to all threads one by one, and to send a message to all threads for which there is an io_channel for a given io_device.

The event Framework

As the number of example applications in SPDK grew, it became clear that a large portion of the code in each was implementing the basic message passing infrastructure required to call spdk_allocate_thread(). This includes spawning one thread per core, pinning each thread to a unique core, and allocating lockless rings between the threads for message passing. Instead of re-implementing that infrastructure for each example application, SPDK provides the SPDK @ref event. This library handles setting up all of the message passing infrastructure, installing signal handlers to cleanly shutdown, implements periodic pollers, and does basic command line parsing. When started through spdk_app_start(), the library automatically spawns all of the threads requested, pins them, and calls spdk_allocate_thread() with appropriate function pointers for each one. This makes it much easier to implement a brand new SPDK application and is the recommended method for those starting out. Only established applications with sufficient message passing infrastructure should consider directly integrating the lower level libraries.

Limitations of the C Language

Message passing is efficient, but it results in asynchronous code. Unfortunately, asynchronous code is a challenge in C. It's often implemented by passing function pointers that are called when an operation completes. This chops up the code so that it isn't easy to follow, especially through logic branches. The best solution is to use a language with support for futures and promises, such as C++, Rust, Go, or almost any other higher level language. However, SPDK is a low level library and requires very wide compatibility and portability, so we've elected to stay with plain old C.

We do have a few recommendations to share, though. For simple callback chains, it's easiest if you write the functions from bottom to top. By that we mean if function foo performs some asynchronous operation and when that completes function bar is called, then function bar performs some operation that calls function baz on completion, a good way to write it is as such:

void baz(void *ctx) {
        ...
}

void bar(void *ctx) {
        async_op(baz, ctx);
}

void foo(void *ctx) {
        async_op(bar, ctx);
}

Don't split these functions up - keep them as a nice unit that can be read from bottom to top.

For more complex callback chains, especially ones that have logical branches or loops, it's best to write out a state machine. It turns out that higher level languages that support futures and promises are just generating state machines at compile time, so even though we don't have the ability to generate them in C we can still write them out by hand. As an example, here's a callback chain that performs foo 5 times and then calls bar - effectively an asynchronous for loop.

enum states {
        FOO_START = 0,
        FOO_END,
        BAR_START,
        BAR_END
};

struct state_machine {
        enum states state;

        int count;
};

static void
foo_complete(void *ctx)
{
    struct state_machine *sm = ctx;

    sm->state = FOO_END;
    run_state_machine(sm);
}

static void
foo(struct state_machine *sm)
{
    do_async_op(foo_complete, sm);
}

static void
bar_complete(void *ctx)
{
    struct state_machine *sm = ctx;

    sm->state = BAR_END;
    run_state_machine(sm);
}

static void
bar(struct state_machine *sm)
{
    do_async_op(bar_complete, sm);
}

static void
run_state_machine(struct state_machine *sm)
{
    enum states prev_state;

    do {
        prev_state = sm->state;

        switch (sm->state) {
            case FOO_START:
                foo(sm);
                break;
            case FOO_END:
                /* This is the loop condition */
                if (sm->count++ < 5) {
                    sm->state = FOO_START;
                } else {
                    sm->state = BAR_START;
                }
                break;
            case BAR_START:
                bar(sm);
                break;
            case BAR_END:
                break;
        }
    } while (prev_state != sm->state);
}

void do_async_for(void)
{
        struct state_machine *sm;

        sm = malloc(sizeof(*sm));
        sm->state = FOO_START;
        sm->count = 0;

        run_state_machine(sm);
}

This is complex, of course, but the run_state_machine function can be read from top to bottom to get a clear overview of what's happening in the code without having to chase through each of the callbacks.