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
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
base / message_loop / message_pump_glib.cc [blame]
// Copyright 2012 The Chromium Authors
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file.
#include "base/message_loop/message_pump_glib.h"
#include <fcntl.h>
#include <glib.h>
#include <math.h>
#include "base/logging.h"
#include "base/memory/raw_ptr.h"
#include "base/notreached.h"
#include "base/numerics/safe_conversions.h"
#include "base/posix/eintr_wrapper.h"
#include "base/synchronization/lock.h"
#include "base/threading/platform_thread.h"
namespace base {
namespace {
// Priorities of event sources are important to let everything be processed.
// In particular, GTK event source should have the highest priority (because
// UI events come from it), then Wayland events (the ones coming from the FD
// watcher), and the lowest priority is GLib events (our base message pump).
//
// The g_source API uses ints to denote priorities, and the lower is its value,
// the higher is the priority (i.e., they are ordered backwards).
constexpr int kPriorityWork = G_PRIORITY_DEFAULT_IDLE;
constexpr int kPriorityFdWatch = G_PRIORITY_DEFAULT_IDLE - 10;
// See the explanation above.
static_assert(G_PRIORITY_DEFAULT < kPriorityFdWatch &&
kPriorityFdWatch < kPriorityWork,
"Wrong priorities are set for event sources!");
// Return a timeout suitable for the glib loop according to |next_task_time|, -1
// to block forever, 0 to return right away, or a timeout in milliseconds from
// now.
int GetTimeIntervalMilliseconds(TimeTicks next_task_time) {
if (next_task_time.is_null())
return 0;
else if (next_task_time.is_max())
return -1;
auto timeout_ms =
(next_task_time - TimeTicks::Now()).InMillisecondsRoundedUp();
return timeout_ms < 0 ? 0 : saturated_cast<int>(timeout_ms);
}
bool RunningOnMainThread() {
auto pid = getpid();
auto tid = PlatformThread::CurrentId();
return pid > 0 && tid > 0 && pid == tid;
}
// A brief refresher on GLib:
// GLib sources have four callbacks: Prepare, Check, Dispatch and Finalize.
// On each iteration of the GLib pump, it calls each source's Prepare function.
// This function should return TRUE if it wants GLib to call its Dispatch, and
// FALSE otherwise. It can also set a timeout in this case for the next time
// Prepare should be called again (it may be called sooner).
// After the Prepare calls, GLib does a poll to check for events from the
// system. File descriptors can be attached to the sources. The poll may block
// if none of the Prepare calls returned TRUE. It will block indefinitely, or
// by the minimum time returned by a source in Prepare.
// After the poll, GLib calls Check for each source that returned FALSE
// from Prepare. The return value of Check has the same meaning as for Prepare,
// making Check a second chance to tell GLib we are ready for Dispatch.
// Finally, GLib calls Dispatch for each source that is ready. If Dispatch
// returns FALSE, GLib will destroy the source. Dispatch calls may be recursive
// (i.e., you can call Run from them), but Prepare and Check cannot.
// Finalize is called when the source is destroyed.
// NOTE: It is common for subsystems to want to process pending events while
// doing intensive work, for example the flash plugin. They usually use the
// following pattern (recommended by the GTK docs):
// while (gtk_events_pending()) {
// gtk_main_iteration();
// }
//
// gtk_events_pending just calls g_main_context_pending, which does the
// following:
// - Call prepare on all the sources.
// - Do the poll with a timeout of 0 (not blocking).
// - Call check on all the sources.
// - *Does not* call dispatch on the sources.
// - Return true if any of prepare() or check() returned true.
//
// gtk_main_iteration just calls g_main_context_iteration, which does the whole
// thing, respecting the timeout for the poll (and block, although it is to if
// gtk_events_pending returned true), and call dispatch.
//
// Thus it is important to only return true from prepare or check if we
// actually have events or work to do. We also need to make sure we keep
// internal state consistent so that if prepare/check return true when called
// from gtk_events_pending, they will still return true when called right
// after, from gtk_main_iteration.
//
// For the GLib pump we try to follow the Windows UI pump model:
// - Whenever we receive a wakeup event or the timer for delayed work expires,
// we run DoWork. That part will also run in the other event pumps.
// - We also run DoWork, and possibly DoIdleWork, in the main loop,
// around event handling.
//
// ---------------------------------------------------------------------------
//
// An overview on the way that we track work items:
//
// ScopedDoWorkItems are used by this pump to track native work. They are
// stored by value in |state_| and are set/cleared as the pump runs. Their
// setting and clearing is done in the functions
// {Set,Clear,EnsureSet,EnsureCleared}ScopedWorkItem. Control flow in GLib is
// quite non-obvious because chrome is not notified when a nested loop is
// entered/exited. To detect nested loops, MessagePumpGlib uses
// |state_->do_work_depth| which is incremented when DoWork is entered, and a
// GLib library function, g_main_depth(), which indicates the current number of
// Dispatch() calls on the stack. To react to them, two separate
// ScopedDoWorkItems are used (a standard one used for all native work, and a
// second one used exclusively for forcing nesting when there is a native loop
// spinning). Note that `ThreadController` flags all nesting as
// `Phase::kNested` so separating native and application work while nested isn't
// supported nor a goal.
//
// It should also be noted that a second GSource has been added to GLib,
// referred to as the "observer" source. It is used because in the case where
// native work occurs on wakeup that is higher priority than Chrome (all of
// GTK), chrome won't even get notified that the pump is awake.
//
// There are several cases to consider wrt. nesting level and order. In
// order, we have:
// A. [root] -> MessagePump::Run() -> native event -> g_main_context_iteration
// B. [root] -> MessagePump::Run() -> DoWork -> g_main_context_iteration
// C. [root] -> native -> DoWork -> MessagePump -> [...]
// The second two cases are identical for our purposes, and the last one turns
// out to be handled without any extra headache.
//
// Consider nesting case A, where native work is called from
// |g_main_context_iteration()| from the pump, and that native work spins up a
// loop. For our purposes, this is a nested loop, because control is not
// returned to the pump once one iteration of the pump is complete. In this
// case, the pump needs to enter nesting without DoWork being involved at
// all. This is accomplished using |MessagePumpGlib::NestIfRequired()|, which is
// called during the Prepare() phase of GLib. As the pump records state on entry
// and exit from GLib using |OnEntryToGlib| and |OnExitFromGlib|, we can compare
// |g_main_depth| at |HandlePrepare| with the one before we entered
// |g_main_context_iteration|. If it is higher, there is a native loop being
// spun, and |RegisterNesting| is called, forcing nesting by initializing two
// work items at once. These are destroyed after the exit from
// |g_main_context_iteration| using |OnExitFromGlib|.
//
// Then, considering nesting case B, |state_->do_work_depth| is incremented
// during any Chrome work, to allow the pump to detect re-entrancy during a
// chrome work item. This is required because `g_main_depth` is not incremented
// in any `DoWork` call not occuring during `Dispatch()` (i.e. during
// `MessagePumpGlib::Run()`). In this case, a nested loop is recorded, and the
// pump sets-and-clears scoped work items during Prepare, Check, and Dispatch. A
// work item can never be active when control flow returns to GLib (i.e. on
// return) during a nested loop, because the nested loop could exit at any
// point. This is fine because TimeKeeper is only concerned with the fact that a
// nested loop is in progress, as opposed to the various phases of the nested
// loop.
//
// Finally, consider nesting case C, where a native loop is spinning
// entirely outside of Chrome, such as inside a signal handler, the pump might
// create and destroy DoWorkItems during Prepare() and Check(), but these work
// items will always get cleared during Dispatch(), before the pump enters a
// DoWork(), leading to the pump showing non-nested native work without the
// thread controller being active, the correct situation (which won't occur
// outside of startup or shutdown). Once Dispatch() is called, the pump's
// nesting tracking works correctly, as state_->do_work_depth is increased, and
// upon re-entrancy we detect the nested loop, which is correct, as this is the
// only point at which the loop actually becomes "nested".
//
// -----------------------------------------------------------------------------
//
// As an overview of the steps taken by MessagePumpGLib to ensure that nested
// loops are detected adequately during each phase of the GLib loop:
//
// 0: Before entering GLib:
// 0.1: Record state about current state of GLib (g_main_depth()) for
// case 1.1.2.
//
// 1: Prepare.
// 1.1: Detection of nested loops
// 1.1.1: If |state_->do_work_depth| > 0, we are in nesting case B detailed
// above. A work item must be newly created during this function to
// trigger nesting, and is destroyed to ensure proper destruction order
// in the case where GLib quits after Prepare().
//
// 1.1.2: Otherwise, check if we are in nesting case A above. If yes, trigger
// nesting using ScopedDoWorkItems. The nesting will be cleared at exit
// from GLib.
//
// This check occurs only in |HandleObserverPrepare|, not in
// |HandlePrepare|.
//
// A third party is running a glib message loop. Since Chrome work is
// registered with GLib at |G_PRIORITY_DEFAULT_IDLE|, a relatively low
// priority, sources of default-or-higher priority will be Dispatch()ed
// first. Since only one source is Dispatched per loop iteration,
// |HandlePrepare| can get called several times in a row in the case that
// there are any other events in the queue. A ScopedDoWorkItem is created
// and destroyed to record this. That work item triggers nesting.
//
// 1.2: Other considerations
// 1.2.1: Sleep occurs between Prepare() and Check(). If Chrome will pass a
// nonzero poll time to GLib, the inner ScopedDoWorkItem is cleared and
// BeforeWait() is called. In nesting case A, the nesting work item will
// not be cleared. A nested loop will typically not block.
//
// Since Prepare() is called before Check() in all cases, the bulk of
// nesting detection is done in Prepare().
//
// 2: Check.
// 2.1: Detection of nested loops:
// 2.1.1: In nesting case B, |ClearScopedWorkItem()| on exit. A third party is
// running a glib message loop. It is possible that at any point the
// nested message loop will quit. In this case, we don't want to leave a
// nested DoWorkItem on the stack.
//
// 2.2: Other considerations
// 2.2.1: A ScopedDoWorkItem may be created (if it was not already present) at
// the entry to Check() to record a wakeup in the case that the pump
// slept. It is important to note that this occurs both in
// |HandleObserverCheck| and |HandleCheck| to ensure that at every point
// as the pump enters the Dispatch phase it is awake. In the case it is
// already awake, this is a very cheap operation.
//
// 3: Dispatch
// 3.1 Detection of nested loops
// 3.1.1: |state_->do_work_depth| is incremented on entry and decremented on
// exit. This is used to detect nesting case B.
//
// 3.1.2: Nested loops can be quit at any point, and so ScopedDoWorkItems can't
// be left on the stack for the same reasons as in 1.1.1/2.1.1.
//
// 3.2 Other considerations
// 3.2.1: Since DoWork creates its own work items, ScopedDoWorkItems are not
// used as this would trigger nesting in all cases.
//
// 4: Post GLib
// 4.1: Detection of nested loops
// 4.1.1: |state_->do_work_depth| is also increased during the DoWork in Run()
// as nesting in that case [calling glib from third party code] needs to
// clear all work items after return to avoid improper destruction order.
//
// 4.2: Other considerations:
// 4.2.1: DoWork uses its own work item, so no ScopedDoWorkItems are active in
// this case.
struct WorkSource : public GSource {
raw_ptr<MessagePumpGlib> pump;
};
gboolean WorkSourcePrepare(GSource* source, gint* timeout_ms) {
*timeout_ms = static_cast<WorkSource*>(source)->pump->HandlePrepare();
// We always return FALSE, so that our timeout is honored. If we were
// to return TRUE, the timeout would be considered to be 0 and the poll
// would never block. Once the poll is finished, Check will be called.
return FALSE;
}
gboolean WorkSourceCheck(GSource* source) {
// Only return TRUE if Dispatch should be called.
return static_cast<WorkSource*>(source)->pump->HandleCheck();
}
gboolean WorkSourceDispatch(GSource* source,
GSourceFunc unused_func,
gpointer unused_data) {
static_cast<WorkSource*>(source)->pump->HandleDispatch();
// Always return TRUE so our source stays registered.
return TRUE;
}
void WorkSourceFinalize(GSource* source) {
// Since the WorkSource object memory is managed by glib, WorkSource implicit
// destructor is never called, and thus WorkSource's raw_ptr never release
// its internal reference on the pump pointer. This leads to adding pressure
// to the BRP quarantine.
static_cast<WorkSource*>(source)->pump = nullptr;
}
// I wish these could be const, but g_source_new wants non-const.
GSourceFuncs g_work_source_funcs = {WorkSourcePrepare, WorkSourceCheck,
WorkSourceDispatch, WorkSourceFinalize};
struct ObserverSource : public GSource {
raw_ptr<MessagePumpGlib> pump;
};
gboolean ObserverPrepare(GSource* gsource, gint* timeout_ms) {
auto* source = static_cast<ObserverSource*>(gsource);
source->pump->HandleObserverPrepare();
*timeout_ms = -1;
// We always want to poll.
return FALSE;
}
gboolean ObserverCheck(GSource* gsource) {
auto* source = static_cast<ObserverSource*>(gsource);
return source->pump->HandleObserverCheck();
}
void ObserverFinalize(GSource* source) {
// Read the comment in `WorkSourceFinalize`, the issue is exactly the same.
static_cast<ObserverSource*>(source)->pump = nullptr;
}
GSourceFuncs g_observer_funcs = {ObserverPrepare, ObserverCheck, nullptr,
ObserverFinalize};
struct FdWatchSource : public GSource {
raw_ptr<MessagePumpGlib> pump;
raw_ptr<MessagePumpGlib::FdWatchController> controller;
};
gboolean FdWatchSourcePrepare(GSource* source, gint* timeout_ms) {
*timeout_ms = -1;
return FALSE;
}
gboolean FdWatchSourceCheck(GSource* gsource) {
auto* source = static_cast<FdWatchSource*>(gsource);
return source->pump->HandleFdWatchCheck(source->controller) ? TRUE : FALSE;
}
gboolean FdWatchSourceDispatch(GSource* gsource,
GSourceFunc unused_func,
gpointer unused_data) {
auto* source = static_cast<FdWatchSource*>(gsource);
source->pump->HandleFdWatchDispatch(source->controller);
return TRUE;
}
void FdWatchSourceFinalize(GSource* gsource) {
// Read the comment in `WorkSourceFinalize`, the issue is exactly the same.
auto* source = static_cast<FdWatchSource*>(gsource);
source->pump = nullptr;
source->controller = nullptr;
}
GSourceFuncs g_fd_watch_source_funcs = {
FdWatchSourcePrepare, FdWatchSourceCheck, FdWatchSourceDispatch,
FdWatchSourceFinalize};
} // namespace
struct MessagePumpGlib::RunState {
explicit RunState(Delegate* delegate) : delegate(delegate) {
CHECK(delegate);
}
const raw_ptr<Delegate> delegate;
// Used to flag that the current Run() invocation should return ASAP.
bool should_quit = false;
// Keeps track of the number of calls to DoWork() on the stack for the current
// Run() invocation. Used to detect reentrancy from DoWork in order to make
// decisions about tracking nested work.
int do_work_depth = 0;
// Value of g_main_depth() captured before the call to
// g_main_context_iteration() in Run(). nullopt if Run() is not calling
// g_main_context_iteration(). Used to track whether the pump has forced a
// nested state due to a native pump.
std::optional<int> g_depth_on_iteration;
// Used to keep track of the native event work items processed by the message
// pump.
Delegate::ScopedDoWorkItem scoped_do_work_item;
// Used to force the pump into a nested state when a native runloop was
// dispatched from main.
Delegate::ScopedDoWorkItem native_loop_do_work_item;
// The information of the next task available at this run-level. Stored in
// RunState because different set of tasks can be accessible at various
// run-levels (e.g. non-nestable tasks).
Delegate::NextWorkInfo next_work_info;
};
MessagePumpGlib::MessagePumpGlib()
: state_(nullptr), wakeup_gpollfd_(std::make_unique<GPollFD>()) {
DCHECK(!g_main_context_get_thread_default());
if (RunningOnMainThread()) {
context_ = g_main_context_default();
} else {
owned_context_ = std::unique_ptr<GMainContext, GMainContextDeleter>(
g_main_context_new());
context_ = owned_context_.get();
g_main_context_push_thread_default(context_);
}
// Create our wakeup pipe, which is used to flag when work was scheduled.
int fds[2];
[[maybe_unused]] int ret = pipe2(fds, O_CLOEXEC);
DCHECK_EQ(ret, 0);
wakeup_pipe_read_ = fds[0];
wakeup_pipe_write_ = fds[1];
wakeup_gpollfd_->fd = wakeup_pipe_read_;
wakeup_gpollfd_->events = G_IO_IN;
observer_source_ = std::unique_ptr<GSource, GSourceDeleter>(
g_source_new(&g_observer_funcs, sizeof(ObserverSource)));
static_cast<ObserverSource*>(observer_source_.get())->pump = this;
g_source_attach(observer_source_.get(), context_);
work_source_ = std::unique_ptr<GSource, GSourceDeleter>(
g_source_new(&g_work_source_funcs, sizeof(WorkSource)));
static_cast<WorkSource*>(work_source_.get())->pump = this;
g_source_add_poll(work_source_.get(), wakeup_gpollfd_.get());
g_source_set_priority(work_source_.get(), kPriorityWork);
// This is needed to allow Run calls inside Dispatch.
g_source_set_can_recurse(work_source_.get(), TRUE);
g_source_attach(work_source_.get(), context_);
}
MessagePumpGlib::~MessagePumpGlib() {
work_source_.reset();
close(wakeup_pipe_read_);
close(wakeup_pipe_write_);
context_ = nullptr;
owned_context_.reset();
}
MessagePumpGlib::FdWatchController::FdWatchController(const Location& location)
: FdWatchControllerInterface(location) {}
MessagePumpGlib::FdWatchController::~FdWatchController() {
if (IsInitialized()) {
auto* source = static_cast<FdWatchSource*>(source_);
source->controller = nullptr;
CHECK(StopWatchingFileDescriptor());
}
if (was_destroyed_) {
DCHECK(!*was_destroyed_);
*was_destroyed_ = true;
}
}
bool MessagePumpGlib::FdWatchController::StopWatchingFileDescriptor() {
if (!IsInitialized())
return false;
g_source_destroy(source_);
g_source_unref(source_.ExtractAsDangling());
watcher_ = nullptr;
return true;
}
bool MessagePumpGlib::FdWatchController::IsInitialized() const {
return !!source_;
}
bool MessagePumpGlib::FdWatchController::InitOrUpdate(int fd,
int mode,
FdWatcher* watcher) {
gushort event_flags = 0;
if (mode & WATCH_READ) {
event_flags |= G_IO_IN;
}
if (mode & WATCH_WRITE) {
event_flags |= G_IO_OUT;
}
if (!IsInitialized()) {
poll_fd_ = std::make_unique<GPollFD>();
poll_fd_->fd = fd;
} else {
if (poll_fd_->fd != fd)
return false;
// Combine old/new event masks.
event_flags |= poll_fd_->events;
// Destroy previous source
bool stopped = StopWatchingFileDescriptor();
DCHECK(stopped);
}
poll_fd_->events = event_flags;
poll_fd_->revents = 0;
source_ = g_source_new(&g_fd_watch_source_funcs, sizeof(FdWatchSource));
DCHECK(source_);
g_source_add_poll(source_, poll_fd_.get());
g_source_set_can_recurse(source_, TRUE);
g_source_set_callback(source_, nullptr, nullptr, nullptr);
g_source_set_priority(source_, kPriorityFdWatch);
watcher_ = watcher;
return true;
}
bool MessagePumpGlib::FdWatchController::Attach(MessagePumpGlib* pump) {
DCHECK(pump);
if (!IsInitialized()) {
return false;
}
auto* source = static_cast<FdWatchSource*>(source_);
source->controller = this;
source->pump = pump;
g_source_attach(source_, pump->context_);
return true;
}
void MessagePumpGlib::FdWatchController::NotifyCanRead() {
if (!watcher_)
return;
DCHECK(poll_fd_);
watcher_->OnFileCanReadWithoutBlocking(poll_fd_->fd);
}
void MessagePumpGlib::FdWatchController::NotifyCanWrite() {
if (!watcher_)
return;
DCHECK(poll_fd_);
watcher_->OnFileCanWriteWithoutBlocking(poll_fd_->fd);
}
bool MessagePumpGlib::WatchFileDescriptor(int fd,
bool persistent,
int mode,
FdWatchController* controller,
FdWatcher* watcher) {
DCHECK_GE(fd, 0);
DCHECK(controller);
DCHECK(watcher);
DCHECK(mode == WATCH_READ || mode == WATCH_WRITE || mode == WATCH_READ_WRITE);
// WatchFileDescriptor should be called on the pump thread. It is not
// threadsafe, so the watcher may never be registered.
DCHECK_CALLED_ON_VALID_THREAD(watch_fd_caller_checker_);
if (!controller->InitOrUpdate(fd, mode, watcher)) {
DPLOG(ERROR) << "FdWatchController init failed (fd=" << fd << ")";
return false;
}
return controller->Attach(this);
}
void MessagePumpGlib::HandleObserverPrepare() {
// |state_| may be null during tests.
if (!state_) {
return;
}
if (state_->do_work_depth > 0) {
// Contingency 1.1.1 detailed above
SetScopedWorkItem();
ClearScopedWorkItem();
} else {
// Contingency 1.1.2 detailed above
NestIfRequired();
}
return;
}
bool MessagePumpGlib::HandleObserverCheck() {
// |state_| may be null in tests.
if (!state_) {
return FALSE;
}
// Make sure we record the fact that we're awake. Chrome won't get Check()ed
// if a higher priority work item returns TRUE from Check().
EnsureSetScopedWorkItem();
if (state_->do_work_depth > 0) {
// Contingency 2.1.1
ClearScopedWorkItem();
}
// The observer never needs to run anything.
return FALSE;
}
// Return the timeout we want passed to poll.
int MessagePumpGlib::HandlePrepare() {
// |state_| may be null during tests.
if (!state_)
return 0;
const int next_wakeup_millis =
GetTimeIntervalMilliseconds(state_->next_work_info.delayed_run_time);
if (next_wakeup_millis != 0) {
// When this is called, it is not possible to know for sure if a
// ScopedWorkItem is on the stack, because HandleObserverCheck may have set
// it during an iteration of the pump where a high priority native work item
// executed.
EnsureClearedScopedWorkItem();
state_->delegate->BeforeWait();
}
return next_wakeup_millis;
}
bool MessagePumpGlib::HandleCheck() {
if (!state_) // state_ may be null during tests.
return false;
// Ensure pump is awake.
EnsureSetScopedWorkItem();
if (state_->do_work_depth > 0) {
// Contingency 2.1.1
ClearScopedWorkItem();
}
// We usually have a single message on the wakeup pipe, since we are only
// signaled when the queue went from empty to non-empty, but there can be
// two messages if a task posted a task, hence we read at most two bytes.
// The glib poll will tell us whether there was data, so this read
// shouldn't block.
if (wakeup_gpollfd_->revents & G_IO_IN) {
char msg[2];
const long num_bytes = HANDLE_EINTR(read(wakeup_pipe_read_, msg, 2));
if (num_bytes < 1) {
NOTREACHED() << "Error reading from the wakeup pipe.";
}
DCHECK((num_bytes == 1 && msg[0] == '!') ||
(num_bytes == 2 && msg[0] == '!' && msg[1] == '!'));
// Since we ate the message, we need to record that we have immediate work,
// because HandleCheck() may be called without HandleDispatch being called
// afterwards.
state_->next_work_info = {TimeTicks()};
return true;
}
// As described in the summary at the top : Check is a second-chance to
// Prepare, verify whether we have work ready again.
if (GetTimeIntervalMilliseconds(state_->next_work_info.delayed_run_time) ==
0) {
return true;
}
return false;
}
void MessagePumpGlib::HandleDispatch() {
// Contingency 3.2.1
EnsureClearedScopedWorkItem();
// Contingency 3.1.1
++state_->do_work_depth;
state_->next_work_info = state_->delegate->DoWork();
--state_->do_work_depth;
if (state_ && state_->do_work_depth > 0) {
// Contingency 3.1.2
EnsureClearedScopedWorkItem();
}
}
void MessagePumpGlib::Run(Delegate* delegate) {
RunState state(delegate);
RunState* previous_state = state_;
state_ = &state;
// We really only do a single task for each iteration of the loop. If we
// have done something, assume there is likely something more to do. This
// will mean that we don't block on the message pump until there was nothing
// more to do. We also set this to true to make sure not to block on the
// first iteration of the loop, so RunUntilIdle() works correctly.
bool more_work_is_plausible = true;
// We run our own loop instead of using g_main_loop_quit in one of the
// callbacks. This is so we only quit our own loops, and we don't quit
// nested loops run by others. TODO(deanm): Is this what we want?
for (;;) {
// ScopedWorkItem to account for any native work until the runloop starts
// running chrome work.
SetScopedWorkItem();
// Don't block if we think we have more work to do.
bool block = !more_work_is_plausible;
OnEntryToGlib();
more_work_is_plausible = g_main_context_iteration(context_, block);
OnExitFromGlib();
if (state_->should_quit)
break;
// Contingency 4.2.1
EnsureClearedScopedWorkItem();
// Contingency 4.1.1
++state_->do_work_depth;
state_->next_work_info = state_->delegate->DoWork();
--state_->do_work_depth;
more_work_is_plausible |= state_->next_work_info.is_immediate();
if (state_->should_quit)
break;
if (more_work_is_plausible)
continue;
state_->delegate->DoIdleWork();
if (state_->should_quit)
break;
}
state_ = previous_state;
}
void MessagePumpGlib::Quit() {
if (state_) {
state_->should_quit = true;
} else {
NOTREACHED() << "Quit called outside Run!";
}
}
void MessagePumpGlib::ScheduleWork() {
// This can be called on any thread, so we don't want to touch any state
// variables as we would then need locks all over. This ensures that if
// we are sleeping in a poll that we will wake up.
char msg = '!';
if (HANDLE_EINTR(write(wakeup_pipe_write_, &msg, 1)) != 1) {
NOTREACHED() << "Could not write to the UI message loop wakeup pipe!";
}
}
void MessagePumpGlib::ScheduleDelayedWork(
const Delegate::NextWorkInfo& next_work_info) {
// We need to wake up the loop in case the poll timeout needs to be
// adjusted. This will cause us to try to do work, but that's OK.
ScheduleWork();
}
bool MessagePumpGlib::HandleFdWatchCheck(FdWatchController* controller) {
DCHECK(controller);
gushort flags = controller->poll_fd_->revents;
return (flags & G_IO_IN) || (flags & G_IO_OUT);
}
void MessagePumpGlib::HandleFdWatchDispatch(FdWatchController* controller) {
DCHECK(controller);
DCHECK(controller->poll_fd_);
gushort flags = controller->poll_fd_->revents;
if ((flags & G_IO_IN) && (flags & G_IO_OUT)) {
// Both callbacks will be called. It is necessary to check that
// |controller| is not destroyed.
bool controller_was_destroyed = false;
controller->was_destroyed_ = &controller_was_destroyed;
controller->NotifyCanWrite();
if (!controller_was_destroyed)
controller->NotifyCanRead();
if (!controller_was_destroyed)
controller->was_destroyed_ = nullptr;
} else if (flags & G_IO_IN) {
controller->NotifyCanRead();
} else if (flags & G_IO_OUT) {
controller->NotifyCanWrite();
}
}
bool MessagePumpGlib::ShouldQuit() const {
CHECK(state_);
return state_->should_quit;
}
void MessagePumpGlib::SetScopedWorkItem() {
// |state_| can be null during tests
if (!state_) {
return;
}
// If there exists a ScopedDoWorkItem in the current RunState, it cannot be
// overwritten.
CHECK(state_->scoped_do_work_item.IsNull());
// In the case that we're more than two work items deep, don't bother tracking
// individual native events anymore. Note that this won't cause out-of-order
// end work items, because the work item is cleared before entering the second
// DoWork().
if (state_->do_work_depth < 2) {
state_->scoped_do_work_item = state_->delegate->BeginWorkItem();
}
}
void MessagePumpGlib::ClearScopedWorkItem() {
// |state_| can be null during tests
if (!state_) {
return;
}
CHECK(!state_->scoped_do_work_item.IsNull());
// See identical check in SetScopedWorkItem
if (state_->do_work_depth < 2) {
state_->scoped_do_work_item = Delegate::ScopedDoWorkItem();
}
}
void MessagePumpGlib::EnsureSetScopedWorkItem() {
// |state_| can be null during tests
if (!state_) {
return;
}
if (state_->scoped_do_work_item.IsNull()) {
SetScopedWorkItem();
}
}
void MessagePumpGlib::EnsureClearedScopedWorkItem() {
// |state_| can be null during tests
if (!state_) {
return;
}
if (!state_->scoped_do_work_item.IsNull()) {
ClearScopedWorkItem();
}
}
void MessagePumpGlib::RegisterNested() {
// |state_| can be null during tests
if (!state_) {
return;
}
CHECK(state_->native_loop_do_work_item.IsNull());
// Transfer `scoped_do_work_item` to `native_do_work_item`, and so the
// ephemeral `scoped_do_work_item` will be coming in and out of existence on
// top of `native_do_work_item`, whose state hasn't been deleted.
if (state_->scoped_do_work_item.IsNull()) {
state_->native_loop_do_work_item = state_->delegate->BeginWorkItem();
} else {
// This clears state_->scoped_do_work_item.
state_->native_loop_do_work_item = std::move(state_->scoped_do_work_item);
}
SetScopedWorkItem();
ClearScopedWorkItem();
}
void MessagePumpGlib::UnregisterNested() {
// |state_| can be null during tests
if (!state_) {
return;
}
CHECK(!state_->native_loop_do_work_item.IsNull());
EnsureClearedScopedWorkItem();
// Nesting exits here.
state_->native_loop_do_work_item = Delegate::ScopedDoWorkItem();
}
void MessagePumpGlib::NestIfRequired() {
// |state_| can be null during tests
if (!state_) {
return;
}
if (state_->native_loop_do_work_item.IsNull() &&
state_->g_depth_on_iteration.has_value() &&
g_main_depth() != state_->g_depth_on_iteration.value()) {
RegisterNested();
}
}
void MessagePumpGlib::UnnestIfRequired() {
// |state_| can be null during tests
if (!state_) {
return;
}
if (!state_->native_loop_do_work_item.IsNull()) {
UnregisterNested();
}
}
void MessagePumpGlib::OnEntryToGlib() {
// |state_| can be null during tests
if (!state_) {
return;
}
CHECK(!state_->g_depth_on_iteration.has_value());
state_->g_depth_on_iteration.emplace(g_main_depth());
}
void MessagePumpGlib::OnExitFromGlib() {
// |state_| can be null during tests
if (!state_) {
return;
}
state_->g_depth_on_iteration.reset();
UnnestIfRequired();
}
} // namespace base