-
Notifications
You must be signed in to change notification settings - Fork 0
/
taskmaster.h
449 lines (362 loc) · 15.5 KB
/
taskmaster.h
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
#pragma once
/* ***** BEGIN LICENSE BLOCK *****
* Version: MPL 1.1/GPL 2.0/LGPL 2.1
*
* The contents of this file are subject to the Mozilla Public License Version
* 1.1 (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
* http://www.mozilla.org/MPL/
*
* Software distributed under the License is distributed on an "AS IS" basis,
* WITHOUT WARRANTY OF ANY KIND, either express or implied. See the License
* for the specific language governing rights and limitations under the
* License.
*
* The Original Code is COID/comm module.
*
* The Initial Developer of the Original Code is
* Outerra.
* Portions created by the Initial Developer are Copyright (C) 2017-2023
* the Initial Developer. All Rights Reserved.
*
* Contributor(s):
* Brano Kemen
*
* Alternatively, the contents of this file may be used under the terms of
* either the GNU General Public License Version 2 or later (the "GPL"), or
* the GNU Lesser General Public License Version 2.1 or later (the "LGPL"),
* in which case the provisions of the GPL or the LGPL are applicable instead
* of those above. If you wish to allow use of your version of this file only
* under the terms of either the GPL or the LGPL, and not to allow others to
* use your version of this file under the terms of the MPL, indicate your
* decision by deleting the provisions above and replace them with the notice
* and other provisions required by the GPL or the LGPL. If you do not delete
* the provisions above, a recipient may use your version of this file under
* the terms of any one of the MPL, the GPL or the LGPL.
*
* ***** END LICENSE BLOCK ***** */
#include "namespace.h"
#include "trait.h"
#include "alloc/slotalloc.h"
#include "bitrange.h"
#include "sync/queue.h"
#include "sync/mutex.h"
#include "sync/condition_variable.h"
#include "pthreadx.h"
#include "log/logger.h"
COID_NAMESPACE_BEGIN
/**
Taskmaster runs a set of worker threads and a queue of tasks that are processed by the worker threads.
Tasks with higher priorities are processed before tasks with lower priorities. LOW priority tasks can
run on a limited number of worker threads. Each job can be associated with a signal and user can wait
for this signal. Several jobs can share single signal, but single job can not trigger multiple signals.
When a thread is waiting for a signal it processes other tasks in queue.
Basic usage:
coid::taskmaster::signal_handle signal;
for (int i = 0; i < 10; ++i) {
taskmaster->push(coid::taskmaster::EPriority::NORMAL, &signal, [i](){ foo(i); });
}
taskmaster->wait(signal);
**/
class taskmaster
{
public:
struct critical_section
{
volatile int32 value = 0;
};
using wait_counter = std::atomic<uint>;
enum class EPriority {
HIGH,
NORMAL,
LOW, //< for limited "long run" threads
COUNT
};
/// @param nthreads total number of job threads to spawn
/// @param nlong_threads number of low-prio job threads (<= nthreads)
taskmaster(uint nthreads, uint nlowprio_threads);
~taskmaster();
uints get_workers_count() const { return _threads.size(); }
///Run fn(index) in parallel in task level 0
/// @param first begin index value
/// @param last end index value
/// @param fn function(index) to run
template <typename Index, typename Fn>
void parallel_for(Index first, Index last, const Fn& fn) {
wait_counter counter;
for (; first != last; ++first) {
push(EPriority::HIGH, &counter, fn, first);
}
wait(counter);
}
///Push task (functor, e.g. lamda) into queue for processing by worker threads
/// @param priority task priority, higher priority tasks are processed before lower priority
/// @param signal signal to trigger when the task finishes
/// @param fn functor to run
template <typename Fn>
void push_functor(EPriority priority, wait_counter* wait_counter_ptr, const Fn& fn)
{
push(priority, wait_counter_ptr, [](const Fn& fn) {
fn();
}, fn);
}
///Push task (function and its arguments) into queue for processing by worker threads
/// @param priority task priority, higher priority tasks are processed before lower priority
/// @param signal signal to trigger when the task finishes
/// @param fn function to run
/// @param args arguments needed to invoke the function
template <typename Fn, typename ...Args>
void push(EPriority priority, wait_counter* wait_counter_ptr, const Fn& fn, Args&& ...args)
{
using callfn = invoker<Fn, Args...>;
{
//lock to access allocator and semaphore
comm_mutex_guard<comm_mutex> lock(_task_sync);
granule* p = alloc_data(sizeof(callfn));
if (wait_counter_ptr)
{
wait_counter_ptr->fetch_add(1, std::memory_order_relaxed);
}
auto task = new(p) callfn(wait_counter_ptr, fn, std::forward<Args>(args)...);
_ready_jobs[(int)priority].push_front(task);
++_qsize;
if (priority != EPriority::LOW) ++_hqsize;
}
_cv.notify_one();
if (priority != EPriority::LOW) {
_hcv.notify_one();
}
}
///Push task (function and its arguments) into queue for processing by worker threads
/// @param priority task priority, higher priority tasks are processed before lower priority
/// @param signal signal to trigger when the task finishes
/// @param fn member function to run
/// @param obj object pointer to run the member function on
/// @param args arguments needed to invoke the function
template <typename Fn, typename C, typename ...Args>
void push_memberfn(EPriority priority, wait_counter* wait_counter_ptr, Fn fn, C* obj, Args&& ...args)
{
static_assert(std::is_member_function_pointer<Fn>::value, "fn must be a function that can be invoked as ((*obj).*fn)(args)");
using callfn = invoker_memberfn<Fn, C*, Args...>;
{
//lock to access allocator and semaphore
comm_mutex_guard<comm_mutex> lock(_task_sync);
granule* p = alloc_data(sizeof(callfn));
if (wait_counter_ptr)
{
wait_counter_ptr->fetch_add(1, std::memory_order_relaxed);
}
auto task = new(p) callfn(wait_counter_ptr, fn, obj, std::forward<Args>(args)...);
_ready_jobs[(int)priority].push_front(task);
++_qsize;
if (priority != EPriority::LOW) ++_hqsize;
}
_cv.notify_one();
if (priority != EPriority::LOW) {
_hcv.notify_one();
}
}
///Push task (function and its arguments) into queue for processing by worker threads
/// @param priority task priority, higher priority tasks are processed before lower priority
/// @param signal signal to trigger when the task finishes
/// @param fn member function to run
/// @param obj object reference to run the member function on. Can be a smart ptr type which resolves to the object with * operator
/// @param args arguments needed to invoke the function
template <typename Fn, typename C, typename ...Args>
void push_memberfn(EPriority priority, wait_counter* wait_counter_ptr, Fn fn, const C& obj, Args&& ...args)
{
static_assert(std::is_member_function_pointer<Fn>::value, "fn must be a function that can be invoked as ((*obj).*fn)(args)");
using callfn = invoker_memberfn<Fn, C, Args...>;
{
//lock to access allocator and semaphore
comm_mutex_guard<comm_mutex> lock(_task_sync);
granule* p = alloc_data(sizeof(callfn));
if (wait_counter_ptr)
{
wait_counter_ptr->fetch_add(1, std::memory_order_relaxed);
}
auto task = new(p) callfn(wait_counter_ptr, fn, obj, std::forward<Args>(args)...);
_ready_jobs[(int)priority].push_front(task);
++_qsize;
if (priority != EPriority::LOW) ++_hqsize;
}
_cv.notify_one();
if (priority != EPriority::LOW) {
_hcv.notify_one();
}
}
/// Enter critical section; no two threads can be in the same critical section at the same time
/// other threads process other tasks while waiting to enter critical section
/// @param spin_count number of spins before trying to process other tasks
/// @note never call enter(A) enter(B) exit(A) exit(B) in that order, since it can cause
// deadlock thanks to taskmaster's nature
void enter_critical_section(critical_section& critical_section, int spin_count = 1024);
/// Leave critical section
/// @note only thread which entered the critical section can leave it
void leave_critical_section(critical_section& critical_section);
///Wait for signal to become signaled
/// @param signal signal to wait for
/// @note each time a task is pushed to queue and has a signal associated, it increments the signal's counter.
// When the task finishes it decrements the counter. Once the counter == 0, the signal is in signaled state.
// Multiple tasks can use the same signal.
void wait(const std::atomic<uint32>& wait_counter);
///Terminate all task threads
/// @param empty_queue if true, wait until the task queue empties, false finish only currently processed tasks
void terminate(bool empty_queue);
protected:
///
struct threadinfo
{
thread tid;
taskmaster* master;
int order;
threadinfo() : master(0), order(-1)
{}
};
///Unit of allocation for tasks
struct granule
{
uint8 dummy[8 * sizeof(void*)];
};
static int& get_order()
{
static thread_local int order = -1;
return order;
}
granule* alloc_data(uints size)
{
uints n = align_to_chunks(size, sizeof(granule));
granule* p = _taskdata.add_contiguous_range(n);
//coidlog_devdbg("taskmaster", "pushed task id " << _taskdata.get_item_id(p));
return p;
}
///
struct invoker_base {
virtual void invoke() = 0;
virtual size_t size() const = 0;
invoker_base(wait_counter* wait_counter_ptr)
: _wait_counter_ptr(wait_counter_ptr)
, _tid(thread::self())
{}
std::atomic<uint>* get_wait_counter() const {
return _wait_counter_ptr;
}
protected:
std::atomic<uint32>* _wait_counter_ptr;
thread_t _tid;
};
template <typename Fn, typename ...Args>
struct invoker_common : invoker_base
{
invoker_common(wait_counter* wait_counter_ptr, const Fn& fn, Args&& ...args)
: invoker_base(wait_counter_ptr)
, _fn(fn)
, _tuple(std::forward<Args>(args)...)
{}
protected:
template <size_t ...I>
void invoke_fn(index_sequence<I...>) {
_fn(std::get<I>(_tuple)...);
}
template <class C, size_t ...I>
void invoke_memberfn(C& obj, index_sequence<I...>) {
(obj.*_fn)(std::get<I>(_tuple)...);
}
private:
typedef std::tuple<std::remove_reference_t<Args>...> tuple_t;
Fn _fn;
tuple_t _tuple;
};
///invoker for lambdas and functions
template <typename Fn, typename ...Args>
struct invoker : invoker_common<Fn, Args...>
{
invoker(wait_counter* wait_counter_ptr, const Fn& fn, Args&& ...args)
: invoker_common<Fn, Args...>(wait_counter_ptr, fn, std::forward<Args>(args)...)
{}
void invoke() override final {
this->invoke_fn(make_index_sequence<sizeof...(Args)>());
}
size_t size() const override final {
return sizeof(*this);
}
};
///invoker for member functions (on copied objects)
template <typename Fn, typename C, typename ...Args>
struct invoker_memberfn : invoker_common<Fn, Args...>
{
invoker_memberfn(wait_counter* wait_counter_ptr, Fn fn, const C& obj, Args&&... args)
: invoker_common<Fn, Args...>(wait_counter_ptr, fn, std::forward<Args>(args)...)
, _obj(obj)
{}
invoker_memberfn(wait_counter* wait_counter_ptr, Fn fn, C&& obj, Args&&... args)
: invoker_common<Fn, Args...>(wait_counter_ptr, fn, std::forward<Args>(args)...)
, _obj(std::forward<C>(obj))
{}
void invoke() override final {
this->invoke_memberfn(_obj, make_index_sequence<sizeof...(Args)>());
}
size_t size() const override final {
return sizeof(*this);
}
private:
C _obj;
};
///invoker for member functions (on pointers)
template <typename Fn, typename C, typename ...Args>
struct invoker_memberfn<Fn, C*, Args...> : invoker_common<Fn, Args...>
{
invoker_memberfn(wait_counter* wait_counter_ptr, Fn fn, C* obj, Args&&... args)
: invoker_common<Fn, Args...>(wait_counter_ptr, fn, std::forward<Args>(args)...)
, _obj(obj)
{}
void invoke() override final {
this->invoke_memberfn(*_obj, make_index_sequence<sizeof...(Args)>());
}
size_t size() const override final {
return sizeof(*this);
}
private:
C* _obj;
};
///invoker for member functions (on irefs)
template <typename Fn, typename C, typename ...Args>
struct invoker_memberfn<Fn, iref<C>, Args...> : invoker_common<Fn, Args...>
{
invoker_memberfn(wait_counter* wait_counter_ptr, Fn fn, const iref<C>& obj, Args&&... args)
: invoker_common<Fn, Args...>(wait_counter_ptr, fn, std::forward<Args>(args)...)
, _obj(obj)
{}
void invoke() override final {
this->invoke_memberfn(*_obj, make_index_sequence<sizeof...(Args)>());
}
size_t size() const override final {
return sizeof(*this);
}
private:
iref<C> _obj;
};
private:
taskmaster(const taskmaster&);
static void* threadfunc(void* data) {
threadinfo* ti = (threadinfo*)data;
return ti->master->threadfunc(ti->order);
}
void* threadfunc(int order);
void run_task(invoker_base* task, bool waiter);
void notify_all();
void wait_internal();
private:
comm_mutex _task_sync; //< mutex for queuing and dequeuing the tasks from queue
comm_mutex _wait_sync; //< mutex for waiting on condition variable
condition_variable _cv; //< for threads which can process low prio tasks
condition_variable _hcv; //< for threads which can not process low prio tasks
std::atomic_int _qsize; //< current queue size, used also as a semaphore
std::atomic_int _hqsize; //< current queue size without low prio tasks
volatile bool _quitting;
slotalloc_atomic<granule> _taskdata;
dynarray<threadinfo> _threads;
volatile int _nlowprio_threads;
queue<invoker_base*> _ready_jobs[(int)EPriority::COUNT];
};
COID_NAMESPACE_END