summaryrefslogtreecommitdiff
path: root/src/mem/request.hh
blob: 50944932e78de12b9c70fe74d61f293ca43c7f2d (plain)
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
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
/*
 * Copyright (c) 2012-2013,2017-2018 ARM Limited
 * All rights reserved
 *
 * The license below extends only to copyright in the software and shall
 * not be construed as granting a license to any other intellectual
 * property including but not limited to intellectual property relating
 * to a hardware implementation of the functionality of the software
 * licensed hereunder.  You may use the software subject to the license
 * terms below provided that you ensure that this notice is replicated
 * unmodified and in its entirety in all distributions of the software,
 * modified or unmodified, in source code or in binary form.
 *
 * Copyright (c) 2002-2005 The Regents of The University of Michigan
 * Copyright (c) 2010,2015 Advanced Micro Devices, Inc.
 * All rights reserved.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions are
 * met: redistributions of source code must retain the above copyright
 * notice, this list of conditions and the following disclaimer;
 * redistributions in binary form must reproduce the above copyright
 * notice, this list of conditions and the following disclaimer in the
 * documentation and/or other materials provided with the distribution;
 * neither the name of the copyright holders nor the names of its
 * contributors may be used to endorse or promote products derived from
 * this software without specific prior written permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
 * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
 * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
 * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
 * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
 * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
 * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 *
 * Authors: Ron Dreslinski
 *          Steve Reinhardt
 *          Ali Saidi
 */

/**
 * @file
 * Declaration of a request, the overall memory request consisting of
 the parts of the request that are persistent throughout the transaction.
 */

#ifndef __MEM_REQUEST_HH__
#define __MEM_REQUEST_HH__

#include <cassert>
#include <climits>

#include "base/flags.hh"
#include "base/logging.hh"
#include "base/types.hh"
#include "cpu/inst_seq.hh"
#include "sim/core.hh"

/**
 * Special TaskIds that are used for per-context-switch stats dumps
 * and Cache Occupancy. Having too many tasks seems to be a problem
 * with vector stats. 1024 seems to be a reasonable number that
 * doesn't cause a problem with stats and is large enough to realistic
 * benchmarks (Linux/Android boot, BBench, etc.)
 */

namespace ContextSwitchTaskId {
    enum TaskId {
        MaxNormalTaskId = 1021, /* Maximum number of normal tasks */
        Prefetcher = 1022, /* For cache lines brought in by prefetcher */
        DMA = 1023, /* Mostly Table Walker */
        Unknown = 1024,
        NumTaskId
    };
}

class Request;

typedef std::shared_ptr<Request> RequestPtr;
typedef uint16_t MasterID;

class Request
{
  public:
    typedef uint64_t FlagsType;
    typedef uint8_t ArchFlagsType;
    typedef ::Flags<FlagsType> Flags;

    enum : FlagsType {
        /**
         * Architecture specific flags.
         *
         * These bits int the flag field are reserved for
         * architecture-specific code. For example, SPARC uses them to
         * represent ASIs.
         */
        ARCH_BITS                   = 0x000000FF,
        /** The request was an instruction fetch. */
        INST_FETCH                  = 0x00000100,
        /** The virtual address is also the physical address. */
        PHYSICAL                    = 0x00000200,
        /**
         * The request is to an uncacheable address.
         *
         * @note Uncacheable accesses may be reordered by CPU models. The
         * STRICT_ORDER flag should be set if such reordering is
         * undesirable.
         */
        UNCACHEABLE                = 0x00000400,
        /**
         * The request is required to be strictly ordered by <i>CPU
         * models</i> and is non-speculative.
         *
         * A strictly ordered request is guaranteed to never be
         * re-ordered or executed speculatively by a CPU model. The
         * memory system may still reorder requests in caches unless
         * the UNCACHEABLE flag is set as well.
         */
        STRICT_ORDER                = 0x00000800,
        /** This request is to a memory mapped register. */
        MMAPPED_IPR                 = 0x00002000,
        /** This request is made in privileged mode. */
        PRIVILEGED                  = 0x00008000,

        /**
         * This is a write that is targeted and zeroing an entire
         * cache block.  There is no need for a read/modify/write
         */
        CACHE_BLOCK_ZERO            = 0x00010000,

        /** The request should not cause a memory access. */
        NO_ACCESS                   = 0x00080000,
        /**
         * This request will lock or unlock the accessed memory. When
         * used with a load, the access locks the particular chunk of
         * memory. When used with a store, it unlocks. The rule is
         * that locked accesses have to be made up of a locked load,
         * some operation on the data, and then a locked store.
         */
        LOCKED_RMW                  = 0x00100000,
        /** The request is a Load locked/store conditional. */
        LLSC                        = 0x00200000,
        /** This request is for a memory swap. */
        MEM_SWAP                    = 0x00400000,
        MEM_SWAP_COND               = 0x00800000,

        /** The request is a prefetch. */
        PREFETCH                    = 0x01000000,
        /** The request should be prefetched into the exclusive state. */
        PF_EXCLUSIVE                = 0x02000000,
        /** The request should be marked as LRU. */
        EVICT_NEXT                  = 0x04000000,
        /** The request should be marked with ACQUIRE. */
        ACQUIRE                     = 0x00020000,
        /** The request should be marked with RELEASE. */
        RELEASE                     = 0x00040000,

        /** The request is an atomic that returns data. */
        ATOMIC_RETURN_OP            = 0x40000000,
        /** The request is an atomic that does not return data. */
        ATOMIC_NO_RETURN_OP         = 0x80000000,

        /** The request should be marked with KERNEL.
          * Used to indicate the synchronization associated with a GPU kernel
          * launch or completion.
          */
        KERNEL                      = 0x00001000,

        /**
         * The request should be handled by the generic IPR code (only
         * valid together with MMAPPED_IPR)
         */
        GENERIC_IPR                 = 0x08000000,

        /** The request targets the secure memory space. */
        SECURE                      = 0x10000000,
        /** The request is a page table walk */
        PT_WALK                     = 0x20000000,

        /** The request invalidates a memory location */
        INVALIDATE                  = 0x0000000100000000,
        /** The request cleans a memory location */
        CLEAN                       = 0x0000000200000000,

        /** The request targets the point of unification */
        DST_POU                     = 0x0000001000000000,

        /** The request targets the point of coherence */
        DST_POC                     = 0x0000002000000000,

        /** Bits to define the destination of a request */
        DST_BITS                    = 0x0000003000000000,

        /**
         * These flags are *not* cleared when a Request object is
         * reused (assigned a new address).
         */
        STICKY_FLAGS = INST_FETCH
    };
    static const FlagsType STORE_NO_DATA = CACHE_BLOCK_ZERO |
        CLEAN | INVALIDATE;

    /** Master Ids that are statically allocated
     * @{*/
    enum : MasterID {
        /** This master id is used for writeback requests by the caches */
        wbMasterId = 0,
        /**
         * This master id is used for functional requests that
         * don't come from a particular device
         */
        funcMasterId = 1,
        /** This master id is used for message signaled interrupts */
        intMasterId = 2,
        /**
         * Invalid master id for assertion checking only. It is
         * invalid behavior to ever send this id as part of a request.
         */
        invldMasterId = std::numeric_limits<MasterID>::max()
    };
    /** @} */

    typedef uint32_t MemSpaceConfigFlagsType;
    typedef ::Flags<MemSpaceConfigFlagsType> MemSpaceConfigFlags;

    enum : MemSpaceConfigFlagsType {
        /** Has a synchronization scope been set? */
        SCOPE_VALID            = 0x00000001,
        /** Access has Wavefront scope visibility */
        WAVEFRONT_SCOPE        = 0x00000002,
        /** Access has Workgroup scope visibility */
        WORKGROUP_SCOPE        = 0x00000004,
        /** Access has Device (e.g., GPU) scope visibility */
        DEVICE_SCOPE           = 0x00000008,
        /** Access has System (e.g., CPU + GPU) scope visibility */
        SYSTEM_SCOPE           = 0x00000010,

        /** Global Segment */
        GLOBAL_SEGMENT         = 0x00000020,
        /** Group Segment */
        GROUP_SEGMENT          = 0x00000040,
        /** Private Segment */
        PRIVATE_SEGMENT        = 0x00000080,
        /** Kergarg Segment */
        KERNARG_SEGMENT        = 0x00000100,
        /** Readonly Segment */
        READONLY_SEGMENT       = 0x00000200,
        /** Spill Segment */
        SPILL_SEGMENT          = 0x00000400,
        /** Arg Segment */
        ARG_SEGMENT            = 0x00000800,
    };

  private:
    typedef uint16_t PrivateFlagsType;
    typedef ::Flags<PrivateFlagsType> PrivateFlags;

    enum : PrivateFlagsType {
        /** Whether or not the size is valid. */
        VALID_SIZE           = 0x00000001,
        /** Whether or not paddr is valid (has been written yet). */
        VALID_PADDR          = 0x00000002,
        /** Whether or not the vaddr & asid are valid. */
        VALID_VADDR          = 0x00000004,
        /** Whether or not the instruction sequence number is valid. */
        VALID_INST_SEQ_NUM   = 0x00000008,
        /** Whether or not the pc is valid. */
        VALID_PC             = 0x00000010,
        /** Whether or not the context ID is valid. */
        VALID_CONTEXT_ID     = 0x00000020,
        /** Whether or not the sc result is valid. */
        VALID_EXTRA_DATA     = 0x00000080,
        /** Whether or not the stream ID and substream ID is valid. */
        VALID_STREAM_ID      = 0x00000100,
        VALID_SUBSTREAM_ID   = 0x00000200,
        /**
         * These flags are *not* cleared when a Request object is reused
         * (assigned a new address).
         */
        STICKY_PRIVATE_FLAGS = VALID_CONTEXT_ID
    };

  private:

    /**
     * Set up a physical (e.g. device) request in a previously
     * allocated Request object.
     */
    void
    setPhys(Addr paddr, unsigned size, Flags flags, MasterID mid, Tick time)
    {
        _paddr = paddr;
        _size = size;
        _time = time;
        _masterId = mid;
        _flags.clear(~STICKY_FLAGS);
        _flags.set(flags);
        privateFlags.clear(~STICKY_PRIVATE_FLAGS);
        privateFlags.set(VALID_PADDR|VALID_SIZE);
        depth = 0;
        accessDelta = 0;
        //translateDelta = 0;
    }

    /**
     * The physical address of the request. Valid only if validPaddr
     * is set.
     */
    Addr _paddr;

    /**
     * The size of the request. This field must be set when vaddr or
     * paddr is written via setVirt() or setPhys(), so it is always
     * valid as long as one of the address fields is valid.
     */
    unsigned _size;

    /** Byte-enable mask for writes. */
    std::vector<bool> _byteEnable;

    /** The requestor ID which is unique in the system for all ports
     * that are capable of issuing a transaction
     */
    MasterID _masterId;

    /** Flag structure for the request. */
    Flags _flags;

    /** Memory space configuraiton flag structure for the request. */
    MemSpaceConfigFlags _memSpaceConfigFlags;

    /** Private flags for field validity checking. */
    PrivateFlags privateFlags;

    /**
     * The time this request was started. Used to calculate
     * latencies. This field is set to curTick() any time paddr or vaddr
     * is written.
     */
    Tick _time;

    /**
     * The task id associated with this request
     */
    uint32_t _taskId;

    union {
        struct {
            /**
             * The stream ID uniquely identifies a device behind the
             * SMMU/IOMMU Each transaction arriving at the SMMU/IOMMU is
             * associated with exactly one stream ID.
             */
            uint32_t  _streamId;

            /**
             * The substream ID identifies an "execution context" within a
             * device behind an SMMU/IOMMU. It's intended to map 1-to-1 to
             * PCIe PASID (Process Address Space ID). The presence of a
             * substream ID is optional.
             */
            uint32_t _substreamId;
        };

        /** The address space ID. */
        uint64_t _asid;
    };

    /** The virtual address of the request. */
    Addr _vaddr;

    /**
     * Extra data for the request, such as the return value of
     * store conditional or the compare value for a CAS. */
    uint64_t _extraData;

    /** The context ID (for statistics, locks, and wakeups). */
    ContextID _contextId;

    /** program counter of initiating access; for tracing/debugging */
    Addr _pc;

    /** Sequence number of the instruction that creates the request */
    InstSeqNum _reqInstSeqNum;

    /** A pointer to an atomic operation */
    AtomicOpFunctorPtr atomicOpFunctor;

  public:

    /**
     * Minimal constructor. No fields are initialized. (Note that
     *  _flags and privateFlags are cleared by Flags default
     *  constructor.)
     */
    Request()
        : _paddr(0), _size(0), _masterId(invldMasterId), _time(0),
          _taskId(ContextSwitchTaskId::Unknown), _asid(0), _vaddr(0),
          _extraData(0), _contextId(0), _pc(0),
          _reqInstSeqNum(0), atomicOpFunctor(nullptr), translateDelta(0),
          accessDelta(0), depth(0)
    {}

    Request(Addr paddr, unsigned size, Flags flags, MasterID mid,
            InstSeqNum seq_num, ContextID cid)
        : _paddr(0), _size(0), _masterId(invldMasterId), _time(0),
          _taskId(ContextSwitchTaskId::Unknown), _asid(0), _vaddr(0),
          _extraData(0), _contextId(0), _pc(0),
          _reqInstSeqNum(seq_num), atomicOpFunctor(nullptr), translateDelta(0),
          accessDelta(0), depth(0)
    {
        setPhys(paddr, size, flags, mid, curTick());
        setContext(cid);
        privateFlags.set(VALID_INST_SEQ_NUM);
    }

    /**
     * Constructor for physical (e.g. device) requests.  Initializes
     * just physical address, size, flags, and timestamp (to curTick()).
     * These fields are adequate to perform a request.
     */
    Request(Addr paddr, unsigned size, Flags flags, MasterID mid)
        : _paddr(0), _size(0), _masterId(invldMasterId), _time(0),
          _taskId(ContextSwitchTaskId::Unknown), _asid(0), _vaddr(0),
          _extraData(0), _contextId(0), _pc(0),
          _reqInstSeqNum(0), atomicOpFunctor(nullptr), translateDelta(0),
          accessDelta(0), depth(0)
    {
        setPhys(paddr, size, flags, mid, curTick());
    }

    Request(Addr paddr, unsigned size, Flags flags, MasterID mid, Tick time)
        : _paddr(0), _size(0), _masterId(invldMasterId), _time(0),
          _taskId(ContextSwitchTaskId::Unknown), _asid(0), _vaddr(0),
          _extraData(0), _contextId(0), _pc(0),
          _reqInstSeqNum(0), atomicOpFunctor(nullptr), translateDelta(0),
          accessDelta(0), depth(0)
    {
        setPhys(paddr, size, flags, mid, time);
    }

    Request(Addr paddr, unsigned size, Flags flags, MasterID mid, Tick time,
            Addr pc)
        : _paddr(0), _size(0), _masterId(invldMasterId), _time(0),
          _taskId(ContextSwitchTaskId::Unknown), _asid(0), _vaddr(0),
          _extraData(0), _contextId(0), _pc(pc),
          _reqInstSeqNum(0), atomicOpFunctor(nullptr), translateDelta(0),
          accessDelta(0), depth(0)
    {
        setPhys(paddr, size, flags, mid, time);
        privateFlags.set(VALID_PC);
    }

    Request(uint64_t asid, Addr vaddr, unsigned size, Flags flags,
            MasterID mid, Addr pc, ContextID cid)
        : _paddr(0), _size(0), _masterId(invldMasterId), _time(0),
          _taskId(ContextSwitchTaskId::Unknown), _asid(0), _vaddr(0),
          _extraData(0), _contextId(0), _pc(0),
          _reqInstSeqNum(0), atomicOpFunctor(nullptr), translateDelta(0),
          accessDelta(0), depth(0)
    {
        setVirt(asid, vaddr, size, flags, mid, pc);
        setContext(cid);
    }

    Request(uint64_t asid, Addr vaddr, unsigned size, Flags flags,
            MasterID mid, Addr pc, ContextID cid,
            AtomicOpFunctorPtr atomic_op)
    {
        setVirt(asid, vaddr, size, flags, mid, pc, std::move(atomic_op));
        setContext(cid);
    }

    Request(const Request& other)
        : _paddr(other._paddr), _size(other._size),
          _masterId(other._masterId),
          _flags(other._flags),
          _memSpaceConfigFlags(other._memSpaceConfigFlags),
          privateFlags(other.privateFlags),
          _time(other._time),
          _taskId(other._taskId), _asid(other._asid), _vaddr(other._vaddr),
          _extraData(other._extraData), _contextId(other._contextId),
          _pc(other._pc), _reqInstSeqNum(other._reqInstSeqNum),
          translateDelta(other.translateDelta),
          accessDelta(other.accessDelta), depth(other.depth)
    {

        atomicOpFunctor.reset(other.atomicOpFunctor ?
                                other.atomicOpFunctor->clone() : nullptr);
    }

    ~Request() {}

    /**
     * Set up Context numbers.
     */
    void
    setContext(ContextID context_id)
    {
        _contextId = context_id;
        privateFlags.set(VALID_CONTEXT_ID);
    }

    void
    setStreamId(uint32_t sid)
    {
        _streamId = sid;
        privateFlags.set(VALID_STREAM_ID);
    }

    void
    setSubStreamId(uint32_t ssid)
    {
        assert(privateFlags.isSet(VALID_STREAM_ID));
        _substreamId = ssid;
        privateFlags.set(VALID_SUBSTREAM_ID);
    }

    /**
     * Set up a virtual (e.g., CPU) request in a previously
     * allocated Request object.
     */
    void
    setVirt(uint64_t asid, Addr vaddr, unsigned size, Flags flags,
            MasterID mid, Addr pc, AtomicOpFunctorPtr amo_op = nullptr)
    {
        _asid = asid;
        _vaddr = vaddr;
        _size = size;
        _masterId = mid;
        _pc = pc;
        _time = curTick();

        _flags.clear(~STICKY_FLAGS);
        _flags.set(flags);
        privateFlags.clear(~STICKY_PRIVATE_FLAGS);
        privateFlags.set(VALID_VADDR|VALID_SIZE|VALID_PC);
        depth = 0;
        accessDelta = 0;
        translateDelta = 0;
        atomicOpFunctor = std::move(amo_op);
    }

    /**
     * Set just the physical address.  This usually used to record the
     * result of a translation. However, when using virtualized CPUs
     * setPhys() is sometimes called to finalize a physical address
     * without a virtual address, so we can't check if the virtual
     * address is valid.
     */
    void
    setPaddr(Addr paddr)
    {
        _paddr = paddr;
        privateFlags.set(VALID_PADDR);
    }

    /**
     * Generate two requests as if this request had been split into two
     * pieces. The original request can't have been translated already.
     */
    // TODO: this function is still required by TimingSimpleCPU - should be
    // removed once TimingSimpleCPU will support arbitrarily long multi-line
    // mem. accesses
    void splitOnVaddr(Addr split_addr, RequestPtr &req1, RequestPtr &req2)
    {
        assert(privateFlags.isSet(VALID_VADDR));
        assert(privateFlags.noneSet(VALID_PADDR));
        assert(split_addr > _vaddr && split_addr < _vaddr + _size);
        req1 = std::make_shared<Request>(*this);
        req2 = std::make_shared<Request>(*this);
        req1->_size = split_addr - _vaddr;
        req2->_vaddr = split_addr;
        req2->_size = _size - req1->_size;
        if (!_byteEnable.empty()) {
            req1->_byteEnable = std::vector<bool>(
                _byteEnable.begin(),
                _byteEnable.begin() + req1->_size);
            req2->_byteEnable = std::vector<bool>(
                _byteEnable.begin() + req1->_size,
                _byteEnable.end());
        }
    }

    /**
     * Accessor for paddr.
     */
    bool
    hasPaddr() const
    {
        return privateFlags.isSet(VALID_PADDR);
    }

    Addr
    getPaddr() const
    {
        assert(privateFlags.isSet(VALID_PADDR));
        return _paddr;
    }

    /**
     * Time for the TLB/table walker to successfully translate this request.
     */
    Tick translateDelta;

    /**
     * Access latency to complete this memory transaction not including
     * translation time.
     */
    Tick accessDelta;

    /**
     * Level of the cache hierachy where this request was responded to
     * (e.g. 0 = L1; 1 = L2).
     */
    mutable int depth;

    /**
     *  Accessor for size.
     */
    bool
    hasSize() const
    {
        return privateFlags.isSet(VALID_SIZE);
    }

    unsigned
    getSize() const
    {
        assert(privateFlags.isSet(VALID_SIZE));
        return _size;
    }

    const std::vector<bool>&
    getByteEnable() const
    {
        return _byteEnable;
    }

    void
    setByteEnable(const std::vector<bool>& be)
    {
        assert(be.empty() || be.size() == _size);
        _byteEnable = be;
    }

    /** Accessor for time. */
    Tick
    time() const
    {
        assert(privateFlags.isSet(VALID_PADDR|VALID_VADDR));
        return _time;
    }

    /**
     * Accessor for atomic-op functor.
     */
    bool
    hasAtomicOpFunctor()
    {
        return (bool)atomicOpFunctor;
    }

    AtomicOpFunctor *
    getAtomicOpFunctor()
    {
        assert(atomicOpFunctor);
        return atomicOpFunctor.get();
    }

    /** Accessor for flags. */
    Flags
    getFlags()
    {
        assert(privateFlags.isSet(VALID_PADDR|VALID_VADDR));
        return _flags;
    }

    /** Note that unlike other accessors, this function sets *specific
        flags* (ORs them in); it does not assign its argument to the
        _flags field.  Thus this method should rightly be called
        setFlags() and not just flags(). */
    void
    setFlags(Flags flags)
    {
        assert(privateFlags.isSet(VALID_PADDR|VALID_VADDR));
        _flags.set(flags);
    }

    void
    setMemSpaceConfigFlags(MemSpaceConfigFlags extraFlags)
    {
        assert(privateFlags.isSet(VALID_PADDR | VALID_VADDR));
        _memSpaceConfigFlags.set(extraFlags);
    }

    /** Accessor function for vaddr.*/
    bool
    hasVaddr() const
    {
        return privateFlags.isSet(VALID_VADDR);
    }

    Addr
    getVaddr() const
    {
        assert(privateFlags.isSet(VALID_VADDR));
        return _vaddr;
    }

    /** Accesssor for the requestor id. */
    MasterID
    masterId() const
    {
        return _masterId;
    }

    uint32_t
    taskId() const
    {
        return _taskId;
    }

    void
    taskId(uint32_t id) {
        _taskId = id;
    }

    /** Accessor function for asid.*/
    uint64_t
    getAsid() const
    {
        assert(privateFlags.isSet(VALID_VADDR));
        return _asid;
    }

    /** Accessor function for asid.*/
    void
    setAsid(uint64_t asid)
    {
        _asid = asid;
    }

    /** Accessor function for architecture-specific flags.*/
    ArchFlagsType
    getArchFlags() const
    {
        assert(privateFlags.isSet(VALID_PADDR|VALID_VADDR));
        return _flags & ARCH_BITS;
    }

    /** Accessor function to check if sc result is valid. */
    bool
    extraDataValid() const
    {
        return privateFlags.isSet(VALID_EXTRA_DATA);
    }

    /** Accessor function for store conditional return value.*/
    uint64_t
    getExtraData() const
    {
        assert(privateFlags.isSet(VALID_EXTRA_DATA));
        return _extraData;
    }

    /** Accessor function for store conditional return value.*/
    void
    setExtraData(uint64_t extraData)
    {
        _extraData = extraData;
        privateFlags.set(VALID_EXTRA_DATA);
    }

    bool
    hasContextId() const
    {
        return privateFlags.isSet(VALID_CONTEXT_ID);
    }

    /** Accessor function for context ID.*/
    ContextID
    contextId() const
    {
        assert(privateFlags.isSet(VALID_CONTEXT_ID));
        return _contextId;
    }

    uint32_t
    streamId() const
    {
        assert(privateFlags.isSet(VALID_STREAM_ID));
        return _streamId;
    }

    bool
    hasSubstreamId() const
    {
        return privateFlags.isSet(VALID_SUBSTREAM_ID);
    }

    uint32_t
    substreamId() const
    {
        assert(privateFlags.isSet(VALID_SUBSTREAM_ID));
        return _substreamId;
    }

    void
    setPC(Addr pc)
    {
        privateFlags.set(VALID_PC);
        _pc = pc;
    }

    bool
    hasPC() const
    {
        return privateFlags.isSet(VALID_PC);
    }

    /** Accessor function for pc.*/
    Addr
    getPC() const
    {
        assert(privateFlags.isSet(VALID_PC));
        return _pc;
    }

    /**
     * Increment/Get the depth at which this request is responded to.
     * This currently happens when the request misses in any cache level.
     */
    void incAccessDepth() const { depth++; }
    int getAccessDepth() const { return depth; }

    /**
     * Set/Get the time taken for this request to be successfully translated.
     */
    void setTranslateLatency() { translateDelta = curTick() - _time; }
    Tick getTranslateLatency() const { return translateDelta; }

    /**
     * Set/Get the time taken to complete this request's access, not including
     *  the time to successfully translate the request.
     */
    void setAccessLatency() { accessDelta = curTick() - _time - translateDelta; }
    Tick getAccessLatency() const { return accessDelta; }

    /**
     * Accessor for the sequence number of instruction that creates the
     * request.
     */
    bool
    hasInstSeqNum() const
    {
        return privateFlags.isSet(VALID_INST_SEQ_NUM);
    }

    InstSeqNum
    getReqInstSeqNum() const
    {
        assert(privateFlags.isSet(VALID_INST_SEQ_NUM));
        return _reqInstSeqNum;
    }

    void
    setReqInstSeqNum(const InstSeqNum seq_num)
    {
        privateFlags.set(VALID_INST_SEQ_NUM);
        _reqInstSeqNum = seq_num;
    }

    /** Accessor functions for flags. Note that these are for testing
        only; setting flags should be done via setFlags(). */
    bool isUncacheable() const { return _flags.isSet(UNCACHEABLE); }
    bool isStrictlyOrdered() const { return _flags.isSet(STRICT_ORDER); }
    bool isInstFetch() const { return _flags.isSet(INST_FETCH); }
    bool isPrefetch() const { return (_flags.isSet(PREFETCH) ||
                                      _flags.isSet(PF_EXCLUSIVE)); }
    bool isPrefetchEx() const { return _flags.isSet(PF_EXCLUSIVE); }
    bool isLLSC() const { return _flags.isSet(LLSC); }
    bool isPriv() const { return _flags.isSet(PRIVILEGED); }
    bool isLockedRMW() const { return _flags.isSet(LOCKED_RMW); }
    bool isSwap() const { return _flags.isSet(MEM_SWAP|MEM_SWAP_COND); }
    bool isCondSwap() const { return _flags.isSet(MEM_SWAP_COND); }
    bool isMmappedIpr() const { return _flags.isSet(MMAPPED_IPR); }
    bool isSecure() const { return _flags.isSet(SECURE); }
    bool isPTWalk() const { return _flags.isSet(PT_WALK); }
    bool isAcquire() const { return _flags.isSet(ACQUIRE); }
    bool isRelease() const { return _flags.isSet(RELEASE); }
    bool isKernel() const { return _flags.isSet(KERNEL); }
    bool isAtomicReturn() const { return _flags.isSet(ATOMIC_RETURN_OP); }
    bool isAtomicNoReturn() const { return _flags.isSet(ATOMIC_NO_RETURN_OP); }

    bool
    isAtomic() const
    {
        return _flags.isSet(ATOMIC_RETURN_OP) ||
               _flags.isSet(ATOMIC_NO_RETURN_OP);
    }

    /**
     * Accessor functions for the destination of a memory request. The
     * destination flag can specify a point of reference for the
     * operation (e.g. a cache block clean to the the point of
     * unification). At the moment the destination is only used by the
     * cache maintenance operations.
     */
    bool isToPOU() const { return _flags.isSet(DST_POU); }
    bool isToPOC() const { return _flags.isSet(DST_POC); }
    Flags getDest() const { return _flags & DST_BITS; }

    /**
     * Accessor functions for the memory space configuration flags and used by
     * GPU ISAs such as the Heterogeneous System Architecture (HSA). Note that
     * these are for testing only; setting extraFlags should be done via
     * setMemSpaceConfigFlags().
     */
    bool isScoped() const { return _memSpaceConfigFlags.isSet(SCOPE_VALID); }

    bool
    isWavefrontScope() const
    {
        assert(isScoped());
        return _memSpaceConfigFlags.isSet(WAVEFRONT_SCOPE);
    }

    bool
    isWorkgroupScope() const
    {
        assert(isScoped());
        return _memSpaceConfigFlags.isSet(WORKGROUP_SCOPE);
    }

    bool
    isDeviceScope() const
    {
        assert(isScoped());
        return _memSpaceConfigFlags.isSet(DEVICE_SCOPE);
    }

    bool
    isSystemScope() const
    {
        assert(isScoped());
        return _memSpaceConfigFlags.isSet(SYSTEM_SCOPE);
    }

    bool
    isGlobalSegment() const
    {
        return _memSpaceConfigFlags.isSet(GLOBAL_SEGMENT) ||
               (!isGroupSegment() && !isPrivateSegment() &&
                !isKernargSegment() && !isReadonlySegment() &&
                !isSpillSegment() && !isArgSegment());
    }

    bool
    isGroupSegment() const
    {
        return _memSpaceConfigFlags.isSet(GROUP_SEGMENT);
    }

    bool
    isPrivateSegment() const
    {
        return _memSpaceConfigFlags.isSet(PRIVATE_SEGMENT);
    }

    bool
    isKernargSegment() const
    {
        return _memSpaceConfigFlags.isSet(KERNARG_SEGMENT);
    }

    bool
    isReadonlySegment() const
    {
        return _memSpaceConfigFlags.isSet(READONLY_SEGMENT);
    }

    bool
    isSpillSegment() const
    {
        return _memSpaceConfigFlags.isSet(SPILL_SEGMENT);
    }

    bool
    isArgSegment() const
    {
        return _memSpaceConfigFlags.isSet(ARG_SEGMENT);
    }

    /**
     * Accessor functions to determine whether this request is part of
     * a cache maintenance operation. At the moment three operations
     * are supported:

     * 1) A cache clean operation updates all copies of a memory
     * location to the point of reference,
     * 2) A cache invalidate operation invalidates all copies of the
     * specified block in the memory above the point of reference,
     * 3) A clean and invalidate operation is a combination of the two
     * operations.
     * @{ */
    bool isCacheClean() const { return _flags.isSet(CLEAN); }
    bool isCacheInvalidate() const { return _flags.isSet(INVALIDATE); }
    bool isCacheMaintenance() const { return _flags.isSet(CLEAN|INVALIDATE); }
    /** @} */
};

#endif // __MEM_REQUEST_HH__