How GNU Radio Core Works -- An A…

I find a paper about GNURADIO CORE and think it is a good reference. I will use Chinese to introduce it.

我会用中文主要介绍一下这个GNURADIO CORE WORK主要讲什么东西。

Contents

1 How the GNU Radio scheduler is called and what it does      1

2 How a thread of each block works                            5

   内容

1.  GNU Radio 调度器怎样调用 和 怎么工作

2.  每个模块的线程怎么工作

1 How the GNU Radio scheduler is called and what it does

As we become quite familiar with GNU Radio Python codes now, it is essential to figure out

what's going on behind the plain Python codes; that is, how GNU radio core works. A typical

example of GNU Radio 3 is as follows. (There is an old dial-tone example using flow_graph(),

which is obsolete! )

上面的话没太多意思，就是让我们从dial-tone开始去探索GNU Radio怎么工作

Besides some routine connections for blocks, the GNU Radio running thread starts at

my_top_block().run()

We are going to figure out what's going on after this code.  Since Python and C++ classes are 1-1 corresponding via SWIG, We have to go deeper  and to find the run() function in C++ class  gr_top_block.

这里告诉你，PYTHON和C++通过SWIG对应起来，我们要了解 python里 run()这个函数，就要深入C++去了解。

In file:  gr top block.cc

void 

gr_top_block::run()

{

  start();

  wait();

}

Then, go to start();

先找到run(),看看代码，发现start()，去看start()

In file: gr top block.cc

void 

gr_top_block::start()

{

    d_impl->start();

}

d_impl is a member that points to the class gr_top_block_impl with the following codes:

看到start() 其实是调用了d_impl->start()，去class gr_top_block_impl 再去看

In file : gr top block impl.cc

void

gr_top_block_impl::start()

{

    gruel::scoped_lock l(d_mutex);

    if (d_state != IDLE)

      throw std::runtime_error("top_block::start: top block already running or wait() not called after previous stop()");

    if (d_lock_count > 0)

      throw std::runtime_error("top_block::start: can't start with flow graph locked");

    // Create new flat flow graph by flattening hierarchy

    d_ffg = d_owner->flatten();

    // Validate new simple flow graph and wire it up

    d_ffg->validate();

    d_ffg->setup_connections();

    d_scheduler = make_scheduler(d_ffg);

    d_state = RUNNING;

}

The codes do some sanity check and then create the GNU Radio Scheduler by calling

d_scheduler = make_scheduler(d_ffg);

Let go to function make_scheduler() as follows

上面的代码通过调用 d_scheduler = make_scheduler(d_ffg); 很严谨的检查并且创建了GNU Radio的调度器。让我们去make_scheduler(d_ffg)看看

In file : gr top block impl.cc

static gr_scheduler_sptr

make_scheduler(gr_flat_flowgraph_sptr ffg)

{

    static scheduler_maker   factory = 0;

    if (factory == 0){

      char *v = getenv("GR_SCHEDULER");

      if (!v)

        factory = scheduler_table[0].f; // use default

      else {

        for (size_t i = 0; i < sizeof(scheduler_table)/sizeof(scheduler_table[0]); i++){

if (strcmp(v, scheduler_table[i].name) == 0){

  factory = scheduler_table[i].f;

  break;

}

        }

        if (factory == 0){

std::cerr << "warning: Invalid GR_SCHEDULER environment variable value \""

  << v << "\".   Using \"" << scheduler_table[0].name << "\"\n";

factory = scheduler_table[0].f;

        }

      }

    }

    return factory(ffg);

}

In the above program, what is the variable static scheduler_maker factory

we can look into the file and find,

在上面的程序中，scheduler_maker factory 是什么玩意？我们在文件中找找

typedef gr_scheduler_sptr (*scheduler_maker) (gr_flat_flowgraph_sptr ffg);

Well, factory is a function pointer! Where does it points to? We can find

factory 是一个函数指针，指向哪里？找找

factory = schedulertable[i].f

OK. Let us find what is in scheduler_table:

好了，让我们继续找找什么是scheduler_table

static struct scheduler_table {

    const char       *name;

    scheduler_maker f;

} scheduler_table[] = {

    { "TPB", gr_scheduler_tpb::make }, // first entry is default

    { "STS", gr_scheduler_sts::make }

};

Great! It points to a member function, make, in a scheduler's class. Then the program 

is very easy to understand: it checks whether there exists a Linux environment variable: 

GR_SCHEDULER. If no, use the default scheduler; otherwise, use user's choice. And we do not 

have so many choices on the scheduler:

好耶！它指向一个在scheduler类里边的成员函数，make。简单理解：检查是否存在Linux 环境变量：GR_SCHEDULER。如果没有，就是用默认的调度器；否则，是用用户选择的。其实我们也没多少选择不是么，就下边两个：

1. TPB (default): multi-threaded scheduler.

2. STS: single-threaded scheduler.

By default, gr_scheduler_tpb::make will be called.

In file: gr scheduler_tpb.cc

gr_scheduler_sptr

gr_scheduler_tpb::make(gr_flat_flowgraph_sptr ffg)

{

    return gr_scheduler_sptr(new gr_scheduler_tpb(ffg));

}

The constructor of gr_scheduler_tpb is called.

In file : gr scheduler tpb.cc

gr_scheduler_tpb::gr_scheduler_tpb(gr_flat_flowgraph_sptr ffg)

    : gr_scheduler(ffg)

{

    // Get a topologically sorted vector of all the blocks in use.

    // Being topologically sorted probably isn't going to matter, but

    // there's a non-zero chance it might help...

    gr_basic_block_vector_t used_blocks = ffg->calc_used_blocks();

    used_blocks = ffg->topological_sort(used_blocks);

    gr_block_vector_t blocks = gr_flat_flowgraph::make_block_vector(used_blocks);

    // Ensure that the done flag is clear on all blocks

    for (size_t i = 0; i < blocks.size(); i++){

      blocks[i]->detail()->set_done(false);

    }

    // Fire off a thead for each block

    for (size_t i = 0; i < blocks.size(); i++){

      std::stringstream name;

      name << "thread-per-block[" << i << "]: " << blocks[i];

      d_threads.create_thread(

        gruel::thread_body_wrapper<tpb_container>(tpb_container(blocks[i]), name.str()));

    }

}

Nothing strange here, the only thing needs to mention is

 d_threads.create_thread(

        gruel::thread_body_wrapper<tpb_container>(tpb_container(blocks[i]), name.str()));

   

thread_body_wrapper wraps the main thread with the block name. Then, the thread 

begins from thread_body_wrapper(). Let us see part of the program to find what happens.

thread_body_wrapper 把每个block的主要线程都包住了。然后，线程从thread_body_wrapper()开始。让我们看看到底这部分程序发生了什么。

In file: thread_body_wrapper.h

  

namespace gruel 

{

  void mask_signals();

  template <class F>

  class thread_body_wrapper

  {

    F d_f;

    std::string d_name;

  public:

    explicit thread_body_wrapper(F f, const std::string &name="")

      : d_f(f), d_name(name) {}

    void operator()()

    {

      mask_signals();

      try {

d_f();

      }

      catch(boost::thread_interrupted const &)

      {

      }

      catch(std::exception const &e)

      {

std::cerr << "thread[" << d_name << "]: "

 << e.what() << std::endl;

      }

      catch(...)

      {

std::cerr << "thread[" << d_name << "]: "

 << "caught unrecognized exception\n";

      }

    }

  };

}

 

See the overloading of operate(), actually, d_f() is called, which is explicitly linked to

tpb_container(). So go to the code of tpb_container:

看到重载operate()，实际上调用了d_f()，这个函数清晰的链接到tpb_container()。

In file : gr_scheduler_tpb.cc

class tpb_container

{

    gr_block_sptr d_block;

   

public:

    tpb_container(gr_block_sptr block) : d_block(block) {}

    void operator()()

    {

      gr_tpb_thread_body body(d_block);

    }

};

Well. the overloading of operate() just constructs another class, gr_tpb_thread_body, 

with the block pointer. From here, the scheduler's work is done.

Let's briefly summarize what is going on with the GNU Radio scheduler.

 

1. Analyze used blocks in gr_top_block.

2. Default scheduler is TPB, which creates multi-threads for blocks.

3. The scheduler creates one concurrent thread for each block.

4. For each block, the thread's entry is gr_tpb_thread_body body(d_block).

总结GNU RADIO 调度器作用：
    1. 分析在 gr_top_block中每个block

2.默认调度器是TPB，用来创造每个blocks里边的多线程

3.调度器为每个block创建线程

4.对于每个block， 线程入口都在gr_tpb_thread_body 里边

2 How a thread of each block works

As we discussed, the TPB scheduler generates a thread for each block whose entry starts from

the constructor of class gr_tpb_thread_body. Let us go over the constructor:

正如我们讨论的，TPB调度器为每个block产生一个线程，这些线程的入口来自一个gr_tpb_thread_body类的构造器。

In file : gr_tpb_thread_body.cc

gr_tpb_thread_body::gr_tpb_thread_body(gr_block_sptr block)

    : d_exec(block)

{

    // std::cerr << "gr_tpb_thread_body: " << block << std::endl;

    gr_block_detail *d = block->detail().get();

    gr_block_executor::state s;

    pmt_t msg;

//Here starts the main loop of the thread.

//开始线程的循环

    while (1){

//First, the thread processes all signals

//首先，线程处理所有信号

      boost::this_thread::interruption_point();

 

      // handle any queued up messages

      while ((msg = d->d_tpb.delete_head_nowait()))

        block->handle_msg(msg);

      d->d_tpb.clear_changed();

      s = d_exec.run_one_iteration();

      switch(s){

      case gr_block_executor::READY: // Tell neighbors we made progress.

        d->d_tpb.notify_neighbors(d);

        break;

      case gr_block_executor::READY_NO_OUTPUT: // Notify upstream only

        d->d_tpb.notify_upstream(d);

        break;

      case gr_block_executor::DONE: // Game over.

        d->d_tpb.notify_neighbors(d);

        return;

      case gr_block_executor::BLKD_IN: // Wait for input.

        {

gruel::scoped_lock guard(d->d_tpb.mutex);

while (!d->d_tpb.input_changed){

 

  // wait for input or message

  while(!d->d_tpb.input_changed && d->d_tpb.empty_p())

    d->d_tpb.input_cond.wait(guard);

  // handle all pending messages

  while ((msg = d->d_tpb.delete_head_nowait_already_holding_mutex())){

    guard.unlock(); // release lock while processing msg

    block->handle_msg(msg);

    guard.lock();

  }

}

        }

        break;

       

      case gr_block_executor::BLKD_OUT: // Wait for output buffer space.

        {

gruel::scoped_lock guard(d->d_tpb.mutex);

while (!d->d_tpb.output_changed){

 

  // wait for output room or message

  while(!d->d_tpb.output_changed && d->d_tpb.empty_p())

    d->d_tpb.output_cond.wait(guard);

  // handle all pending messages

  while ((msg = d->d_tpb.delete_head_nowait_already_holding_mutex())){

    guard.unlock(); // release lock while processing msg

    block->handle_msg(msg);

    guard.lock();

  }

}

        }

        break;

      default:

        assert(0);

      }

    }

}

So far so good. We can see run_one_iteration() is the key in the whole thread, it 

includes the major functionality of the block. Go to its source. Woo! it is a little long.

很好，我们看到run_one_iteration()，它是整个线程的核心，它包括block的主要功能。去看看那些源代码吧！哇！！有一点点长噢。

In file : gr_block_executor.cc

gr_block_executor::state 

gr_block_executor::run_one_iteration()

{

  ...

// Do the actual work of the block int n = m->general_work (noutput_items, d_ninput_items, d_input_items, d_output_items); LOG(*d_log << " general_work: noutput_items = " << noutput_items << " result = " << n << std::endl);

  ...

}

Overall, the code first checks

1. whether there exists 

sufficient 

data space for output. No → return BLKD_OUT

2. whether there are 

sufficient 

input data available, No → 

 return BLKD_IN

总体上，这些代码首先检查：

1. 是否存在足够的输出数据空间

2. 是否存在足够的可靠输入数据

If there are sufficient input data and sufficient output space, the code runs the actual work

of the block: general_work().

So up to now, we can know how each thread in the GNU radio core works. Let us briefly

如果这里有足够的输入数据和足够的输出空间，代码就会执行block里边的work咯: gnerral_work()

现在，我们可以知道在GNU Radio里的每个线程都怎么工作了噢，所以短暂的总结一下：

summarize

1. A thread for each block has a while (1) loop.

2. The loop processes signals and run the key function run_one_iteration().

3. run_one_iteration() checks if there are sufficient data for input and sufficient available 

space for output for the block.

4. If yes, call general_work() to run the main functionality of the block. Otherwise, return 

BLKD_OUT, BLKD_IN, or others.

1. 每个block的线程都有一个while(1)循环

2. 循环处理信号并且运行主要函数 run_one_iteration()

3. run_one_iteration() 检查 是否有足够的数据输入和为这个block输出足够的可靠的空间

4. 如果都可以，调用general_work() 去跑这个block主要的功能。否则，返回BLKD_OUT, BLKD_IN,或者其他。