-
Notifications
You must be signed in to change notification settings - Fork 23
/
Copy pathgf.cc
1296 lines (1116 loc) · 40.1 KB
/
gf.cc
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
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
#include "hedley.h"
#ifdef HEDLEY_GCC_VERSION
# pragma GCC diagnostic push
# pragma GCC diagnostic ignored "-Wcast-function-type"
#endif
#include <node.h>
#include <node_buffer.h>
#include <node_version.h>
#ifdef HEDLEY_GCC_VERSION
# pragma GCC diagnostic pop
#endif
#include <v8.h>
#include <stdlib.h>
#include <string.h>
#include <uv.h>
#include <node_object_wrap.h>
#if defined(_MSC_VER)
#include <malloc.h>
#endif
#include "../gf16/controller.h"
#include "../gf16/controller_cpu.h"
#include "../gf16/controller_ocl.h"
#include "../gf16/threadqueue.h"
#include "../hasher/hasher.h"
using namespace v8;
/*******************************************/
#if NODE_VERSION_AT_LEAST(0, 11, 0)
// for node 0.12.x
#define FUNC(name) static void name(const FunctionCallbackInfo<Value>& args)
#define HANDLE_SCOPE HandleScope scope(isolate)
#define FUNC_START \
Isolate* isolate = args.GetIsolate(); \
HANDLE_SCOPE
# if NODE_VERSION_AT_LEAST(8, 0, 0)
# define NEW_STRING(s) String::NewFromOneByte(isolate, (const uint8_t*)(s), NewStringType::kNormal).ToLocalChecked()
# define RETURN_ERROR(e) { isolate->ThrowException(Exception::Error(String::NewFromOneByte(isolate, (const uint8_t*)(e), NewStringType::kNormal).ToLocalChecked())); return; }
# define ARG_TO_NUM(t, a) (a).As<t>()->Value()
# define ARG_TO_OBJ(a) (a).As<Object>()
# else
# define NEW_STRING(s) String::NewFromUtf8(isolate, s)
# define RETURN_ERROR(e) { isolate->ThrowException(Exception::Error(String::NewFromUtf8(isolate, e))); return; }
# define ARG_TO_NUM(t, a) (a)->To##t()->Value()
# define ARG_TO_OBJ(a) (a)->ToObject()
# endif
#define RETURN_VAL(v) args.GetReturnValue().Set(v)
#define RETURN_UNDEF return;
#define RETURN_BUFFER RETURN_VAL
#define ISOLATE isolate,
#define NEW_OBJ(o) o::New(isolate)
#define BUFFER_TYPE Local<Object>
#define PERSIST_VALUE(p, v) p.Reset(isolate, v)
#define PERSIST_CLEAR(p) p.Reset()
#else
// for node 0.10.x
#define FUNC(name) static Handle<Value> name(const Arguments& args)
#define HANDLE_SCOPE HandleScope scope
#define FUNC_START HANDLE_SCOPE
#define NEW_STRING String::New
#define ARG_TO_NUM(t, a) (a)->To##t()->Value()
#define ARG_TO_OBJ(a) (a)->ToObject()
#define RETURN_ERROR(e) \
return ThrowException(Exception::Error( \
String::New(e)) \
)
#define RETURN_VAL(v) return scope.Close(v)
#define RETURN_BUFFER(v) RETURN_VAL(v->handle_)
#define RETURN_UNDEF RETURN_VAL( Undefined() );
#define ISOLATE
#define NEW_OBJ(o) o::New()
#define BUFFER_TYPE node::Buffer*
#define PERSIST_VALUE(p, v) p = Persistent<Value>::New(v)
#define PERSIST_CLEAR(p) p.Dispose()
#endif
#if NODE_VERSION_AT_LEAST(3, 0, 0) // iojs3
#define BUFFER_NEW(...) node::Buffer::New(ISOLATE __VA_ARGS__).ToLocalChecked()
#else
#define BUFFER_NEW(...) node::Buffer::New(ISOLATE __VA_ARGS__)
#endif
#if NODE_VERSION_AT_LEAST(12, 0, 0)
# define OBJ_HAS(obj, key) (obj)->Has(isolate->GetCurrentContext(), NEW_STRING(key)).ToChecked()
# define GET_OBJ(obj, key) (obj)->Get(isolate->GetCurrentContext(), NEW_STRING(key)).ToLocalChecked()
# define SET_OBJ(obj, key, val) (obj)->Set(isolate->GetCurrentContext(), NEW_STRING(key), val).Check()
# define GET_ARR(obj, idx) (obj)->Get(isolate->GetCurrentContext(), idx).ToLocalChecked()
# define SET_ARR(obj, idx, val) (obj)->Set(isolate->GetCurrentContext(), idx, val).Check()
# define SET_OBJ_FUNC(obj, key, f) (obj)->Set(isolate->GetCurrentContext(), NEW_STRING(key), f->GetFunction(isolate->GetCurrentContext()).ToLocalChecked()).Check()
#else
# define OBJ_HAS(obj, key) (obj)->Has(NEW_STRING(key))
# define GET_OBJ(obj, key) (obj)->Get(NEW_STRING(key))
# define SET_OBJ(obj, key, val) (obj)->Set(NEW_STRING(key), val)
# define GET_ARR(obj, idx) (obj)->Get(idx)
# define SET_ARR(obj, idx, val) (obj)->Set(idx, val)
# define SET_OBJ_FUNC(obj, key, f) (obj)->Set(NEW_STRING(key), f->GetFunction())
#endif
#if NODE_VERSION_AT_LEAST(0, 11, 0)
// copied from node::GetCurrentEventLoop [https://github.com/nodejs/node/pull/17109]
static inline uv_loop_t* getCurrentLoop(Isolate* isolate, int) {
/* -- don't have access to node::Environment :()
Local<Context> context = isolate->GetCurrentContext();
if(context.IsEmpty())
return uv_default_loop();
return node::Environment::GetCurrent(context)->event_loop();
*/
# if NODE_VERSION_AT_LEAST(9, 3, 0) || (NODE_MAJOR_VERSION == 8 && NODE_MINOR_VERSION >= 10) || (NODE_MAJOR_VERSION == 6 && NODE_MINOR_VERSION >= 14)
return node::GetCurrentEventLoop(isolate);
# endif
return uv_default_loop();
}
#else
static inline uv_loop_t* getCurrentLoop(int) {
return uv_default_loop();
}
#endif
struct CallbackWrapper {
CallbackWrapper() : hasCallback(false) {}
#if NODE_VERSION_AT_LEAST(0, 11, 0)
Isolate* isolate;
// use BaseObject / AsyncWrap instead? meh
explicit CallbackWrapper(Isolate* _isolate, const Local<Value>& callback) : isolate(_isolate) {
attachCallback(_isolate, callback);
}
#else
CallbackWrapper(const Local<Value>& callback) {
attachCallback(callback);
}
#endif
~CallbackWrapper() {
PERSIST_CLEAR(value);
detachCallback();
};
Persistent<Object> obj_;
// persist copy of buffer for the duration of the job
Persistent<Value> value;
bool hasCallback;
#if NODE_VERSION_AT_LEAST(0, 11, 0)
void attachCallback(Isolate* _isolate, const Local<Value>& callback) {
isolate = _isolate;
HANDLE_SCOPE;
Local<Object> obj = NEW_OBJ(Object);
SET_OBJ(obj, "ondone", callback);
obj_.Reset(ISOLATE obj);
//if (env->in_domain())
// obj_->Set(env->domain_string(), env->domain_array()->Get(0));
hasCallback = true;
}
#else
void attachCallback(const Local<Value>& callback) {
HANDLE_SCOPE;
obj_ = Persistent<Object>::New(NEW_OBJ(Object));
obj_->Set(NEW_STRING("ondone"), callback);
//SetActiveDomain(obj_); // never set in node_zlib.cc - perhaps domains aren't that important?
hasCallback = true;
}
#endif
void detachCallback() {
#if NODE_VERSION_AT_LEAST(0, 11, 0)
obj_.Reset();
#else
//if (obj_.IsEmpty()) return;
obj_.Dispose();
obj_.Clear(); // TODO: why this line?
#endif
hasCallback = false;
}
void attachValue(const Local<Value>& val) {
PERSIST_VALUE(value, val);
}
void _call(int argc, Local<Value>* argv, const HandleScope&) {
#if NODE_VERSION_AT_LEAST(0, 11, 0)
Local<Object> obj = Local<Object>::New(isolate, obj_);
# if NODE_VERSION_AT_LEAST(10, 0, 0)
// TODO: properly support async_context
node::async_context ac;
memset(&ac, 0, sizeof(ac));
node::MakeCallback(isolate, obj, "ondone", argc, argv, ac);
# else
node::MakeCallback(isolate, obj, "ondone", argc, argv);
# endif
#else
node::MakeCallback(obj_, "ondone", argc, argv);
#endif
}
inline void call() {
HANDLE_SCOPE;
_call(0, nullptr, scope);
}
inline void call(const HandleScope& scope) {
_call(0, nullptr, scope);
}
inline void call(const HandleScope& scope, std::initializer_list<Local<Value>> args) {
std::vector<Local<Value>> argList(args);
_call(argList.size(), argList.data(), scope);
}
};
struct GfOclSpec {
int platformId, deviceId;
size_t sliceOffset, sliceSize;
Galois16OCLMethods method;
Galois16Methods cksumMethod;
unsigned inputGrouping, inputMinGrouping, targetIters, targetGrouping;
};
static bool load_ocl() {
static bool oclLoaded = false;
if(!oclLoaded) {
if(PAR2ProcOCL::load_runtime()) {
return false;
}
oclLoaded = true;
}
return true;
}
class GfProc : public node::ObjectWrap {
public:
static inline void AttachMethods(Local<FunctionTemplate>& t) {
t->InstanceTemplate()->SetInternalFieldCount(1); // necessary for node::Object::Wrap
NODE_SET_PROTOTYPE_METHOD(t, "close", Close);
NODE_SET_PROTOTYPE_METHOD(t, "freeMem", FreeMem);
NODE_SET_PROTOTYPE_METHOD(t, "setRecoverySlices", SetRecoverySlices);
NODE_SET_PROTOTYPE_METHOD(t, "setCurrentSliceSize", SetCurrentSliceSize);
NODE_SET_PROTOTYPE_METHOD(t, "setNumThreads", SetNumThreads);
NODE_SET_PROTOTYPE_METHOD(t, "setProgressCb", SetProgressCb);
NODE_SET_PROTOTYPE_METHOD(t, "info", GetInfo);
NODE_SET_PROTOTYPE_METHOD(t, "add", AddSlice);
NODE_SET_PROTOTYPE_METHOD(t, "end", EndInput);
NODE_SET_PROTOTYPE_METHOD(t, "get", GetOutputSlice);
}
FUNC(New) {
FUNC_START;
if(!args.IsConstructCall())
RETURN_ERROR("Class must be constructed with 'new'");
if(args.Length() < 1)
RETURN_ERROR("Slice size required");
size_t sliceSize = (size_t)ARG_TO_NUM(Integer, args[0]);
if(sliceSize < 2 || sliceSize & 1)
RETURN_ERROR("Slice size is invalid");
bool useCpu = true;
int stagingAreas = 2;
int cpuMethod = GF16_AUTO;
unsigned cpuInputGrouping = 0, cpuInputMinGrouping = 0;
size_t cpuChunkLen = 0;
size_t cpuOffset = 0, cpuSliceSize = sliceSize;
#define ASSIGN_INT_VAL(prop, key, var, type) \
if(OBJ_HAS(prop, key)) { \
Local<Value> v = GET_OBJ(prop, key); \
if(!v->IsUndefined() && !v->IsNull()) \
var = ARG_TO_NUM(type, v); \
}
if(args.Length() >= 2) { // CPU processing props
if(args[1]->IsNull()) {
useCpu = false;
cpuSliceSize = 0;
} else {
Local<Object> prop = ARG_TO_OBJ(args[1]);
ASSIGN_INT_VAL(prop, "method", cpuMethod, Int32)
// TODO: check validity of cpuMethod
ASSIGN_INT_VAL(prop, "input_batchsize", cpuInputGrouping, Uint32)
if(cpuInputGrouping > 32768)
RETURN_ERROR("Input batchsize is invalid");
ASSIGN_INT_VAL(prop, "input_minbatchsize", cpuInputMinGrouping, Uint32)
ASSIGN_INT_VAL(prop, "chunk_size", cpuChunkLen, Uint32)
ASSIGN_INT_VAL(prop, "slice_offset", cpuOffset, Integer)
if(cpuOffset & 1 || cpuOffset > sliceSize)
RETURN_ERROR("Invalid CPU slice offset");
ASSIGN_INT_VAL(prop, "slice_size", cpuSliceSize, Integer)
if(cpuSliceSize < 2 || cpuSliceSize & 1)
RETURN_ERROR("CPU slice size must be a multiple of 2");
if(cpuOffset+cpuSliceSize > sliceSize)
RETURN_ERROR("CPU slice offset+size cannot exceed the slice size");
}
}
std::vector<struct GfOclSpec> useOcl;
if(args.Length() >= 3 && args[2]->IsArray()) { // OpenCL processing props
Local<Array> props = Local<Array>::Cast(args[2]);
for(unsigned i=0; i<props->Length(); i++) {
Local<Object> prop = ARG_TO_OBJ(GET_ARR(props, i));
struct GfOclSpec spec{-1, -1, 0, 0, GF16OCL_AUTO, GF16_AUTO, 0, 0, 0, 0};
// TODO: validate platform/device
ASSIGN_INT_VAL(prop, "platform", spec.platformId, Int32)
ASSIGN_INT_VAL(prop, "device", spec.deviceId, Int32)
ASSIGN_INT_VAL(prop, "slice_size", spec.sliceSize, Integer)
ASSIGN_INT_VAL(prop, "slice_offset", spec.sliceOffset, Integer)
int method = 0;
ASSIGN_INT_VAL(prop, "method", method, Int32)
if(method) spec.method = (Galois16OCLMethods)method;
// TODO: check validity of method
spec.cksumMethod = (Galois16Methods)cpuMethod;
method = 0;
ASSIGN_INT_VAL(prop, "cksum_method", method, Int32)
if(method) spec.cksumMethod = (Galois16Methods)method;
ASSIGN_INT_VAL(prop, "input_batchsize", spec.inputGrouping, Int32)
if(spec.inputGrouping > 32768)
RETURN_ERROR("OpenCL input batchsize is invalid");
ASSIGN_INT_VAL(prop, "input_minbatchsize", spec.inputMinGrouping, Int32)
ASSIGN_INT_VAL(prop, "target_iters", spec.targetIters, Uint32)
ASSIGN_INT_VAL(prop, "target_grouping", spec.targetGrouping, Uint32)
if(spec.targetGrouping > 65535)
RETURN_ERROR("OpenCL target grouping is invalid");
useOcl.push_back(spec);
}
}
if(args.Length() >= 4 && !args[3]->IsUndefined() && !args[3]->IsNull()) {
stagingAreas = ARG_TO_NUM(Int32, args[3]);
if(stagingAreas < 1 || stagingAreas > 32768)
RETURN_ERROR("Staging area count is invalid");
}
#undef ASSIGN_INT_VAL
if(cpuInputGrouping * stagingAreas > 65536)
RETURN_ERROR("Staging area too large");
if(!useOcl.empty()) {
if(!load_ocl()) {
RETURN_ERROR("Could not load OpenCL runtime");
}
} else if(!useCpu) {
RETURN_ERROR("At least the CPU or one OpenCL device must be enabled");
}
GfProc *self = new GfProc(sliceSize, stagingAreas, cpuOffset, cpuSliceSize, useOcl, getCurrentLoop(ISOLATE 0));
size_t usedSliceSize = 0;
if(useCpu && !self->init_cpu((Galois16Methods)cpuMethod, cpuInputGrouping, cpuChunkLen)) {
delete self;
RETURN_ERROR("Failed to allocate memory");
}
if(useCpu) self->par2cpu->setMinInputBatchSize(cpuInputMinGrouping);
int oclI = 0;
for(const auto& oclSpec : useOcl) {
usedSliceSize += oclSpec.sliceSize;
if(oclSpec.sliceSize == 0 || (oclSpec.sliceSize & 1)) {
delete self;
RETURN_ERROR("Invalid slice size allocated to OpenCL device");
}
if(oclSpec.sliceOffset & 1 || oclSpec.sliceOffset > sliceSize) {
delete self;
RETURN_ERROR("Invalid slice offset allocated to OpenCL device");
}
if(oclSpec.sliceOffset + oclSpec.sliceSize > sliceSize) {
delete self;
RETURN_ERROR("Invalid slice offset+size allocated to OpenCL device");
}
if(!self->init_ocl(oclI, oclSpec.method, oclSpec.inputGrouping, oclSpec.targetIters, oclSpec.targetGrouping, oclSpec.cksumMethod)) {
delete self;
RETURN_ERROR("Failed to initialise OpenCL device"); // TODO: add device info
}
self->par2ocl[oclI]->setMinInputBatchSize(oclSpec.inputMinGrouping);
oclI++;
}
if((useCpu && usedSliceSize >= sliceSize) || (!useCpu && usedSliceSize != sliceSize)) {
delete self;
RETURN_ERROR("Slice portions allocated to OpenCL devices is invalid");
}
self->Wrap(args.This());
RETURN_UNDEF;
}
private:
bool isRunning;
bool isClosed;
bool pendingDiscardOutput;
bool hasOutput;
CallbackWrapper progressCb;
PAR2Proc par2;
std::unique_ptr<PAR2ProcCPU> par2cpu;
std::vector<std::unique_ptr<PAR2ProcOCL>> par2ocl;
// disable copy constructor
GfProc(const GfProc&);
GfProc& operator=(const GfProc&);
protected:
FUNC(Close) {
FUNC_START;
GfProc* self = node::ObjectWrap::Unwrap<GfProc>(args.This());
if(self->isRunning)
RETURN_ERROR("Cannot close whilst running");
if(self->isClosed)
RETURN_ERROR("Already closed");
if(args.Length() >= 1 && !args[0]->IsUndefined() && !args[1]->IsNull()) {
if(!args[0]->IsFunction())
RETURN_ERROR("First argument must be a callback");
CallbackWrapper* cb = new CallbackWrapper(ISOLATE Local<Function>::Cast(args[0]));
self->par2.deinit([cb]() {
cb->call();
delete cb;
});
} else {
self->par2.deinit();
}
self->isClosed = true;
RETURN_UNDEF;
}
FUNC(FreeMem) {
FUNC_START;
GfProc* self = node::ObjectWrap::Unwrap<GfProc>(args.This());
if(self->isRunning)
RETURN_ERROR("Cannot free memory whilst running");
if(self->isClosed)
RETURN_ERROR("Already closed");
self->par2.freeProcessingMem();
self->hasOutput = false;
RETURN_UNDEF;
}
FUNC(SetCurrentSliceSize) {
FUNC_START;
GfProc* self = node::ObjectWrap::Unwrap<GfProc>(args.This());
if(self->isRunning)
RETURN_ERROR("Cannot change params whilst running");
if(self->isClosed)
RETURN_ERROR("Already closed");
if(args.Length() < 1)
RETURN_ERROR("Argument required");
size_t sliceSize = (size_t)ARG_TO_NUM(Integer, args[0]);
if(sliceSize < 2 || sliceSize & 1)
RETURN_ERROR("Slice size is invalid");
// TODO: get slice allocation from input
self->hasOutput = false;
if(!self->par2.setCurrentSliceSize(sliceSize))
RETURN_ERROR("Failed to allocate memory");
RETURN_UNDEF;
}
FUNC(SetRecoverySlices) {
FUNC_START;
GfProc* self = node::ObjectWrap::Unwrap<GfProc>(args.This());
if(self->isRunning)
RETURN_ERROR("Cannot change params whilst running");
if(self->isClosed)
RETURN_ERROR("Already closed");
if(args.Length() < 1 || !args[0]->IsArray())
RETURN_ERROR("List of recovery indicies required");
auto argOutputs = Local<Array>::Cast(args[0]);
int numOutputs = (int)argOutputs->Length();
if(numOutputs > 65534)
RETURN_ERROR("Too many recovery indicies specified");
if(numOutputs < 1)
RETURN_ERROR("At least one recovery index must be supplied");
std::vector<uint16_t> outputs(numOutputs);
for(int i=0; i<numOutputs; i++) {
Local<Value> output = GET_ARR(argOutputs, i);
outputs[i] = ARG_TO_NUM(Uint32, output);
if(outputs[i] > 65534)
RETURN_ERROR("Invalid recovery index supplied");
}
self->hasOutput = false; // probably can be retained, but we'll pretend not for consistency's sake
if(!self->par2.setRecoverySlices(outputs))
RETURN_ERROR("Failed to allocate memory");
RETURN_UNDEF;
}
FUNC(SetNumThreads) {
FUNC_START;
GfProc* self = node::ObjectWrap::Unwrap<GfProc>(args.This());
if(self->isRunning)
RETURN_ERROR("Cannot change params whilst running");
if(self->isClosed)
RETURN_ERROR("Already closed");
if(!self->par2cpu.get())
RETURN_ERROR("CPU processing not enabled");
if(args.Length() < 1)
RETURN_ERROR("Integer required");
self->par2cpu->setNumThreads(ARG_TO_NUM(Int32, args[0]));
RETURN_VAL(Integer::New(ISOLATE self->par2cpu->getNumThreads()));
}
FUNC(SetProgressCb) {
FUNC_START;
GfProc* self = node::ObjectWrap::Unwrap<GfProc>(args.This());
if(self->isClosed)
RETURN_ERROR("Already closed");
if(args.Length() >= 1) {
if(!args[0]->IsFunction())
RETURN_ERROR("Callback required");
self->progressCb.attachCallback(ISOLATE args[0]);
} else {
self->progressCb.detachCallback();
}
RETURN_UNDEF;
}
FUNC(GetInfo) {
// num threads, method name
FUNC_START;
GfProc* self = node::ObjectWrap::Unwrap<GfProc>(args.This());
if(self->isClosed)
RETURN_ERROR("Already closed");
if(!self->par2.getNumRecoverySlices())
RETURN_ERROR("setRecoverySlices not yet called");
Local<Object> ret = NEW_OBJ(Object);
if(self->par2cpu.get()) {
SET_OBJ(ret, "threads", Integer::New(ISOLATE self->par2cpu->getNumThreads()));
SET_OBJ(ret, "method_desc", NEW_STRING(self->par2cpu->getMethodName()));
SET_OBJ(ret, "chunk_size", Number::New(ISOLATE self->par2cpu->getChunkLen()));
SET_OBJ(ret, "staging_count", Integer::New(ISOLATE self->par2cpu->getStagingAreas()));
SET_OBJ(ret, "staging_size", Integer::New(ISOLATE self->par2cpu->getInputBatchSize()));
SET_OBJ(ret, "alignment", Integer::New(ISOLATE self->par2cpu->getAlignment()));
SET_OBJ(ret, "stride", Integer::New(ISOLATE self->par2cpu->getStride()));
SET_OBJ(ret, "slice_mem", Number::New(ISOLATE self->par2cpu->getAllocSliceSize()));
SET_OBJ(ret, "num_output_slices", Integer::New(ISOLATE self->par2cpu->getNumRecoverySlices()));
}
if(!self->par2ocl.empty()) {
Local<Array> oclDevInfo = Array::New(ISOLATE self->par2ocl.size());
int i = 0;
for(const auto& proc : self->par2ocl) {
const auto devInfo = proc->deviceInfo();
Local<Object> oclInfo = NEW_OBJ(Object);
SET_OBJ(oclInfo, "device_name", NEW_STRING(devInfo.name.c_str()));
SET_OBJ(oclInfo, "method_desc", NEW_STRING(proc->getMethodName()));
SET_OBJ(oclInfo, "staging_count", Integer::New(ISOLATE proc->getStagingAreas()));
SET_OBJ(oclInfo, "staging_size", Integer::New(ISOLATE proc->getInputBatchSize()));
SET_OBJ(oclInfo, "chunk_size", Number::New(ISOLATE proc->getChunkLen()));
SET_OBJ(oclInfo, "output_chunks", Integer::New(ISOLATE proc->getOutputGrouping()));
SET_OBJ(oclInfo, "slice_mem", Number::New(ISOLATE proc->getAllocSliceSize()));
SET_OBJ(oclInfo, "num_output_slices", Integer::New(ISOLATE proc->getNumRecoverySlices()));
SET_ARR(oclDevInfo, i++, oclInfo);
}
SET_OBJ(ret, "opencl_devices", oclDevInfo);
}
RETURN_VAL(ret);
}
FUNC(AddSlice) {
FUNC_START;
GfProc* self = node::ObjectWrap::Unwrap<GfProc>(args.This());
if(self->isClosed)
RETURN_ERROR("Already closed");
if(!self->par2.getNumRecoverySlices())
RETURN_ERROR("setRecoverySlices not yet called");
if(args.Length() < 3)
RETURN_ERROR("Requires 3 arguments");
if(!node::Buffer::HasInstance(args[1]))
RETURN_ERROR("Input buffer required");
if(!args[2]->IsFunction())
RETURN_ERROR("Callback required");
int idx = ARG_TO_NUM(Int32, args[0]);
if(idx < 0 || idx > 32767)
RETURN_ERROR("Input index not valid");
if(node::Buffer::Length(args[1]) > self->par2.getCurrentSliceSize())
RETURN_ERROR("Input buffer too large");
CallbackWrapper* cb = new CallbackWrapper(ISOLATE Local<Function>::Cast(args[2]));
cb->attachValue(args[1]);
self->isRunning = true;
self->hasOutput = false;
if(self->pendingDiscardOutput) {
self->pendingDiscardOutput = false;
self->par2.discardOutput();
}
bool added = self->par2.addInput(
node::Buffer::Data(args[1]), node::Buffer::Length(args[1]),
idx, false, [ISOLATE cb, idx]() {
HANDLE_SCOPE;
#if NODE_VERSION_AT_LEAST(0, 11, 0)
Local<Value> buffer = Local<Value>::New(cb->isolate, cb->value);
cb->call(scope, { Integer::New(cb->isolate, idx), buffer });
#else
Local<Value> buffer = Local<Value>::New(cb->value);
cb->call(scope, { Integer::New(idx), buffer });
#endif
delete cb;
}
);
if(!added) {
delete cb;
}
RETURN_VAL(Boolean::New(ISOLATE added));
}
FUNC(EndInput) {
FUNC_START;
GfProc* self = node::ObjectWrap::Unwrap<GfProc>(args.This());
// NOTE: it's possible to end without adding anything, so don't require !self->isRunning
if(self->isClosed)
RETURN_ERROR("Already closed");
if(args.Length() < 1 || !args[0]->IsFunction())
RETURN_ERROR("Callback required");
if(!self->par2.getNumRecoverySlices())
RETURN_ERROR("setRecoverySlices not yet called");
if(self->pendingDiscardOutput)
self->par2.discardOutput();
CallbackWrapper* cb = new CallbackWrapper(ISOLATE Local<Function>::Cast(args[0]));
self->par2.endInput([cb, self]() {
self->isRunning = false;
self->hasOutput = true;
self->pendingDiscardOutput = true;
cb->call();
delete cb;
});
RETURN_UNDEF;
}
FUNC(GetOutputSlice) {
FUNC_START;
GfProc* self = node::ObjectWrap::Unwrap<GfProc>(args.This());
if(self->isRunning)
RETURN_ERROR("Cannot get output whilst running");
if(self->isClosed)
RETURN_ERROR("Already closed");
if(!self->hasOutput)
RETURN_ERROR("No finalized output to retrieve");
if(args.Length() < 3)
RETURN_ERROR("Requires 3 arguments");
if(!node::Buffer::HasInstance(args[1]))
RETURN_ERROR("Output buffer required");
if(!args[2]->IsFunction())
RETURN_ERROR("Callback required");
if(node::Buffer::Length(args[1]) < self->par2.getCurrentSliceSize())
RETURN_ERROR("Output buffer too small");
int idx = ARG_TO_NUM(Int32, args[0]);
if(idx < 0 || idx >= self->par2.getNumRecoverySlices())
RETURN_ERROR("Recovery index is not valid");
CallbackWrapper* cb = new CallbackWrapper(ISOLATE Local<Function>::Cast(args[2]));
cb->attachValue(args[1]);
self->par2.getOutput(
idx,
node::Buffer::Data(args[1]),
[ISOLATE cb, idx](bool cksumValid) {
HANDLE_SCOPE;
#if NODE_VERSION_AT_LEAST(0, 11, 0)
Local<Value> buffer = Local<Value>::New(cb->isolate, cb->value);
cb->call(scope, { Integer::New(cb->isolate, idx), Boolean::New(cb->isolate, cksumValid), buffer });
#else
Local<Value> buffer = Local<Value>::New(cb->value);
Local<Value> _idx = Local<Value>::New(Integer::New(idx));
Local<Value> _cksumValid = Local<Value>::New(Boolean::New(cksumValid));
cb->call(scope, { _idx, _cksumValid, buffer });
#endif
delete cb;
}
);
RETURN_UNDEF;
}
explicit GfProc(size_t sliceSize, int stagingAreas, size_t cpuOffset, size_t cpuSliceSize, std::vector<struct GfOclSpec> useOcl, uv_loop_t* loop)
: ObjectWrap(), isRunning(false), isClosed(false), pendingDiscardOutput(true), hasOutput(false) {
std::vector<struct PAR2ProcBackendAlloc> procs;
for(const auto& spec : useOcl) {
auto proc = new PAR2ProcOCL(loop, spec.platformId, spec.deviceId, stagingAreas);
par2ocl.push_back(std::unique_ptr<PAR2ProcOCL>(proc));
procs.push_back({static_cast<IPAR2ProcBackend*>(proc), spec.sliceOffset, spec.sliceSize});
}
if(cpuSliceSize) {
par2cpu.reset(new PAR2ProcCPU(loop, stagingAreas));
procs.push_back({static_cast<IPAR2ProcBackend*>(par2cpu.get()), cpuOffset, cpuSliceSize});
}
// TODO: handle init returning false (currently, it can't)
par2.init(sliceSize, procs, [&](unsigned numInputs) {
if(progressCb.hasCallback) {
#if NODE_VERSION_AT_LEAST(0, 11, 0)
HandleScope scope(progressCb.isolate);
progressCb.call(scope, { Integer::New(progressCb.isolate, numInputs) });
#else
HandleScope scope;
progressCb.call(scope, { Integer::New(numInputs) });
#endif
}
});
}
bool init_cpu(Galois16Methods method, unsigned inputGrouping, size_t chunkLen) {
return par2cpu->init(method, inputGrouping, chunkLen);
}
bool init_ocl(int idx, Galois16OCLMethods method, unsigned inputGrouping, unsigned targetIters, unsigned targetGrouping, Galois16Methods cksumMethod) {
return par2ocl[idx]->init(method, inputGrouping, targetIters, targetGrouping, cksumMethod);
}
~GfProc() {
par2.deinit();
}
};
FUNC(GfInfo) {
FUNC_START;
// get method
Galois16Methods method = args.Length() >= 1 && !args[0]->IsUndefined() && !args[0]->IsNull() ? (Galois16Methods)ARG_TO_NUM(Int32, args[0]) : GF16_AUTO;
if(method == GF16_AUTO) {
// TODO: accept hints
method = PAR2ProcCPU::default_method();
}
auto info = PAR2ProcCPU::info(method);
Local<Object> ret = NEW_OBJ(Object);
SET_OBJ(ret, "id", Integer::New(ISOLATE info.id));
SET_OBJ(ret, "name", NEW_STRING(info.name));
SET_OBJ(ret, "alignment", Integer::New(ISOLATE info.alignment));
SET_OBJ(ret, "stride", Integer::New(ISOLATE info.stride));
SET_OBJ(ret, "target_chunk", Integer::New(ISOLATE info.idealChunkSize));
SET_OBJ(ret, "target_grouping", Integer::New(ISOLATE info.idealInputMultiple));
RETURN_VAL(ret);
}
static void OclDeviceToJS(
#if NODE_VERSION_AT_LEAST(0, 11, 0)
Isolate* isolate,
#endif
Local<Object>& dev, const GF16OCL_DeviceInfo& device) {
dev = NEW_OBJ(Object);
SET_OBJ(dev, "id", Integer::New(ISOLATE device.id));
SET_OBJ(dev, "name", NEW_STRING(device.name.c_str()));
SET_OBJ(dev, "vendor_id", Integer::New(ISOLATE device.vendorId));
SET_OBJ(dev, "available", Boolean::New(ISOLATE device.available));
SET_OBJ(dev, "supported", Boolean::New(ISOLATE device.supported));
switch(device.type) {
case CL_DEVICE_TYPE_DEFAULT: SET_OBJ(dev, "type", NEW_STRING("Default")); break;
case CL_DEVICE_TYPE_CPU: SET_OBJ(dev, "type", NEW_STRING("CPU")); break;
case CL_DEVICE_TYPE_GPU: SET_OBJ(dev, "type", NEW_STRING("GPU")); break;
case CL_DEVICE_TYPE_ACCELERATOR: SET_OBJ(dev, "type", NEW_STRING("Accelerator")); break;
case CL_DEVICE_TYPE_CUSTOM: SET_OBJ(dev, "type", NEW_STRING("Custom")); break;
default: SET_OBJ(dev, "type", NEW_STRING("Unknown")); // TODO: support multi-types?
}
SET_OBJ(dev, "memory_global", Number::New(ISOLATE device.memory));
SET_OBJ(dev, "cache_global", Number::New(ISOLATE device.globalCache));
SET_OBJ(dev, "memory_unified", Boolean::New(ISOLATE device.unifiedMemory));
SET_OBJ(dev, "memory_constant", Number::New(ISOLATE device.constantMemory));
if(!device.localMemoryIsGlobal || device.memory != device.localMemory)
SET_OBJ(dev, "memory_local", Number::New(ISOLATE device.localMemory));
SET_OBJ(dev, "memory_max_alloc", Number::New(ISOLATE device.maxAllocation));
SET_OBJ(dev, "workgroup_limit", Integer::New(ISOLATE device.maxWorkGroup));
SET_OBJ(dev, "workgroup_multiple", Integer::New(ISOLATE device.workGroupMultiple));
SET_OBJ(dev, "compute_units", Integer::New(ISOLATE device.computeUnits));
}
FUNC(OclDevices) {
FUNC_START;
if(!load_ocl()) {
RETURN_ERROR("Could not load OpenCL runtime");
}
const auto platforms = PAR2ProcOCL::getPlatforms();
Local<Array> ret = Array::New(ISOLATE platforms.size());
for(unsigned pf=0; pf<platforms.size(); pf++) {
const auto devices = PAR2ProcOCL::getDevices(pf);
Local<Array> retDevs = Array::New(ISOLATE devices.size());
for(unsigned dv=0; dv<devices.size(); dv++) {
Local<Object> dev;
OclDeviceToJS(ISOLATE dev, devices[dv]);
SET_ARR(retDevs, dv, dev);
}
Local<Object> retPlat = NEW_OBJ(Object);
SET_OBJ(retPlat, "id", Integer::New(ISOLATE pf));
SET_OBJ(retPlat, "name", NEW_STRING(platforms[pf].c_str()));
SET_OBJ(retPlat, "devices", retDevs);
SET_ARR(ret, pf, retPlat);
}
RETURN_VAL(ret);
}
FUNC(OclDeviceInfo) {
FUNC_START;
if(args.Length() < 2)
RETURN_ERROR("Requires 2 arguments");
int platformId = ARG_TO_NUM(Int32, args[0]);
int deviceId = ARG_TO_NUM(Int32, args[1]);
if(!load_ocl()) {
RETURN_ERROR("Could not load OpenCL runtime");
}
if(platformId == -1) platformId = PAR2ProcOCL::defaultPlatformId();
if(platformId < 0)
RETURN_UNDEF;
const auto device = PAR2ProcOCL::getDevice(platformId, deviceId);
if(!device.vendorId) // invalid device
RETURN_UNDEF;
Local<Object> dev;
OclDeviceToJS(ISOLATE dev, device);
SET_OBJ(dev, "platform_id", Integer::New(ISOLATE platformId));
RETURN_VAL(dev);
}
class HasherInput;
static std::vector<MessageThread*> HasherInputThreadPool;
struct input_blockHash {
uint64_t size;
uint64_t pos;
int count;
char* ptr;
};
struct input_work_data {
IHasherInput* hasher;
const void* buffer;
size_t len;
struct input_blockHash* bh;
CallbackWrapper* cb;
HasherInput* self;
};
class HasherInput : public node::ObjectWrap {
public:
static inline void AttachMethods(Local<FunctionTemplate>& t) {
t->InstanceTemplate()->SetInternalFieldCount(1);
NODE_SET_PROTOTYPE_METHOD(t, "update", Update);
NODE_SET_PROTOTYPE_METHOD(t, "end", End);
NODE_SET_PROTOTYPE_METHOD(t, "reset", Reset);
}
FUNC(New) {
FUNC_START;
if(!args.IsConstructCall())
RETURN_ERROR("Class must be constructed with 'new'");
if(args.Length() < 2 || !node::Buffer::HasInstance(args[1]))
RETURN_ERROR("Requires a size and buffer");
// grab slice size + buffer to write hashes into
double sliceSize = 0; // double ensures enough range even if int is 32-bit
#if NODE_VERSION_AT_LEAST(8, 0, 0)
sliceSize = args[0].As<Number>()->Value();
#else
sliceSize = args[0]->NumberValue();
#endif
HasherInput *self = new HasherInput(getCurrentLoop(ISOLATE 0));
self->bh.size = (uint64_t)sliceSize;
self->bh.pos = 0;
self->bh.count = node::Buffer::Length(args[1]) / 20;
self->bh.ptr = node::Buffer::Data(args[1]);
PERSIST_VALUE(self->ifscData, args[1]);
self->Wrap(args.This());
RETURN_UNDEF;
}
private:
IHasherInput* hasher;
uv_loop_t* loop;
int queueCount;
std::unique_ptr<MessageThread> thread;
std::unique_ptr<uv_async_t> threadSignal;
ThreadMessageQueue<struct input_work_data*> hashesDone;
struct input_blockHash bh;
Persistent<Value> ifscData;
// disable copy constructor
HasherInput(const HasherInput&);
HasherInput& operator=(const HasherInput&);
protected:
FUNC(Reset) {
FUNC_START;
HasherInput* self = node::ObjectWrap::Unwrap<HasherInput>(args.This());
if(self->queueCount)
RETURN_ERROR("Cannot reset whilst running");
self->hasher->reset();
RETURN_UNDEF;
}
static void thread_func(ThreadMessageQueue<void*>& q) {
struct input_work_data* data;
while((data = static_cast<struct input_work_data*>(q.pop())) != NULL) {
char* src_ = (char*)data->buffer;
size_t len = data->len;
// feed initial part
uint64_t blockLeft = data->bh->size - data->bh->pos;
while(len >= blockLeft) {
data->hasher->update(src_, blockLeft);
src_ += blockLeft;
len -= blockLeft;
blockLeft = data->bh->size;
data->bh->pos = 0;
if(data->bh->count) {
data->hasher->getBlock(data->bh->ptr, 0);
data->bh->ptr += 20;
data->bh->count--;
} // else there's an overflow
}
if(len) data->hasher->update(src_, len);
data->bh->pos += len;
// signal main thread that hashing has completed
data->self->hashesDone.push(data);
uv_async_send(data->self->threadSignal.get());
}
}
void after_process() {
auto doneData = hashesDone.popall();
for(struct input_work_data* data : doneData) {
static_cast<HasherInput*>(data->self)->queueCount--;
data->cb->call();
delete data->cb;
delete data;
}
}
inline void thread_send(struct input_work_data* data) {
if(thread == nullptr) {
if(HasherInputThreadPool.empty()) {
thread.reset(new MessageThread(thread_func));
thread->name = "par2_hash_input";
} else {
thread.reset(HasherInputThreadPool.back());
HasherInputThreadPool.pop_back();
}
}
thread->send(data);
}
FUNC(Update) {
FUNC_START;
HasherInput* self = node::ObjectWrap::Unwrap<HasherInput>(args.This());
// TODO: consider queueing mechanism; for now, require JS to do the queueing
if(!self->hasher)
RETURN_ERROR("Process already ended");
if(args.Length() < 2 || !node::Buffer::HasInstance(args[0]) || !args[1]->IsFunction())
RETURN_ERROR("Requires a buffer and callback");
CallbackWrapper* cb = new CallbackWrapper(ISOLATE Local<Function>::Cast(args[1]));
cb->attachValue(args[0]);
self->queueCount++;
struct input_work_data* data = new struct input_work_data;
data->cb = cb;
data->hasher = self->hasher;
data->buffer = node::Buffer::Data(args[0]);
data->len = node::Buffer::Length(args[0]);