forked from melonDS-emu/melonDS
-
Notifications
You must be signed in to change notification settings - Fork 0
/
ARMJIT_Memory.cpp
1344 lines (1205 loc) · 41.1 KB
/
ARMJIT_Memory.cpp
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
/*
Copyright 2016-2022 melonDS team
This file is part of melonDS.
melonDS is free software: you can redistribute it and/or modify it under
the terms of the GNU General Public License as published by the Free
Software Foundation, either version 3 of the License, or (at your option)
any later version.
melonDS is distributed in the hope that it will be useful, but WITHOUT ANY
WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
You should have received a copy of the GNU General Public License along
with melonDS. If not, see http://www.gnu.org/licenses/.
*/
#if defined(__SWITCH__)
#include <switch.h>
#elif defined(_WIN32)
#include <windows.h>
#else
#include <sys/mman.h>
#include <sys/stat.h>
#include <fcntl.h>
#include <unistd.h>
#include <signal.h>
#endif
#if defined(__ANDROID__)
#include <dlfcn.h>
#include <linux/ashmem.h>
#include <sys/ioctl.h>
#endif
#include "ARMJIT_Memory.h"
#include "ARMJIT_Internal.h"
#include "ARMJIT_Compiler.h"
#include "DSi.h"
#include "GPU.h"
#include "GPU3D.h"
#include "Wifi.h"
#include "NDSCart.h"
#include "SPU.h"
#include <stdlib.h>
/*
We're handling fastmem here.
Basically we're repurposing a big piece of virtual memory
and map the memory regions as they're structured on the DS
in it.
On most systems you have a single piece of main ram,
maybe some video ram and faster cache RAM and that's about it.
Here we have not only a lot more different memory regions,
but also two address spaces. Not only that but they all have
mirrors (the worst case is 16kb SWRAM which is mirrored 1024x).
We handle this by only mapping those regions which are actually
used and by praying the games don't go wild.
Beware, this file is full of platform specific code and copied
from Dolphin, so enjoy the copied comments!
*/
namespace ARMJIT_Memory
{
struct FaultDescription
{
u32 EmulatedFaultAddr;
u8* FaultPC;
};
bool FaultHandler(FaultDescription& faultDesc);
}
// Yes I know this looks messy, but better here than somewhere else in the code
#if defined(__x86_64__)
#if defined(_WIN32)
#define CONTEXT_PC Rip
#elif defined(__linux__)
#define CONTEXT_PC uc_mcontext.gregs[REG_RIP]
#elif defined(__APPLE__)
#define CONTEXT_PC uc_mcontext->__ss.__rip
#elif defined(__FreeBSD__)
#define CONTEXT_PC uc_mcontext.mc_rip
#elif defined(__NetBSD__)
#define CONTEXT_PC uc_mcontext.__gregs[_REG_RIP]
#endif
#elif defined(__aarch64__)
#if defined(_WIN32)
#define CONTEXT_PC Pc
#elif defined(__linux__)
#define CONTEXT_PC uc_mcontext.pc
#elif defined(__APPLE__)
#define CONTEXT_PC uc_mcontext->__ss.__pc
#elif defined(__FreeBSD__)
#define CONTEXT_PC uc_mcontext.mc_gpregs.gp_elr
#elif defined(__NetBSD__)
#define CONTEXT_PC uc_mcontext.__gregs[_REG_PC]
#endif
#endif
#if defined(__ANDROID__)
#define ASHMEM_DEVICE "/dev/ashmem"
#endif
#if defined(__SWITCH__)
// with LTO the symbols seem to be not properly overriden
// if they're somewhere else
void HandleFault(u64 pc, u64 lr, u64 fp, u64 faultAddr, u32 desc);
extern "C"
{
void ARM_RestoreContext(u64* registers) __attribute__((noreturn));
extern char __start__;
extern char __rodata_start;
alignas(16) u8 __nx_exception_stack[0x8000];
u64 __nx_exception_stack_size = 0x8000;
void __libnx_exception_handler(ThreadExceptionDump* ctx)
{
ARMJIT_Memory::FaultDescription desc;
u8* curArea = (u8*)(NDS::CurCPU == 0 ? ARMJIT_Memory::FastMem9Start : ARMJIT_Memory::FastMem7Start);
desc.EmulatedFaultAddr = (u8*)ctx->far.x - curArea;
desc.FaultPC = (u8*)ctx->pc.x;
u64 integerRegisters[33];
memcpy(integerRegisters, &ctx->cpu_gprs[0].x, 8*29);
integerRegisters[29] = ctx->fp.x;
integerRegisters[30] = ctx->lr.x;
integerRegisters[31] = ctx->sp.x;
integerRegisters[32] = ctx->pc.x;
if (ARMJIT_Memory::FaultHandler(desc))
{
integerRegisters[32] = (u64)desc.FaultPC;
ARM_RestoreContext(integerRegisters);
}
HandleFault(ctx->pc.x, ctx->lr.x, ctx->fp.x, ctx->far.x, ctx->error_desc);
}
}
#elif defined(_WIN32)
static LONG ExceptionHandler(EXCEPTION_POINTERS* exceptionInfo)
{
if (exceptionInfo->ExceptionRecord->ExceptionCode != EXCEPTION_ACCESS_VIOLATION)
{
return EXCEPTION_CONTINUE_SEARCH;
}
ARMJIT_Memory::FaultDescription desc;
u8* curArea = (u8*)(NDS::CurCPU == 0 ? ARMJIT_Memory::FastMem9Start : ARMJIT_Memory::FastMem7Start);
desc.EmulatedFaultAddr = (u8*)exceptionInfo->ExceptionRecord->ExceptionInformation[1] - curArea;
desc.FaultPC = (u8*)exceptionInfo->ContextRecord->CONTEXT_PC;
if (ARMJIT_Memory::FaultHandler(desc))
{
exceptionInfo->ContextRecord->CONTEXT_PC = (u64)desc.FaultPC;
return EXCEPTION_CONTINUE_EXECUTION;
}
return EXCEPTION_CONTINUE_SEARCH;
}
#else
static struct sigaction OldSaSegv;
static struct sigaction OldSaBus;
static void SigsegvHandler(int sig, siginfo_t* info, void* rawContext)
{
if (sig != SIGSEGV && sig != SIGBUS)
{
// We are not interested in other signals - handle it as usual.
return;
}
if (info->si_code != SEGV_MAPERR && info->si_code != SEGV_ACCERR)
{
// Huh? Return.
return;
}
ucontext_t* context = (ucontext_t*)rawContext;
ARMJIT_Memory::FaultDescription desc;
u8* curArea = (u8*)(NDS::CurCPU == 0 ? ARMJIT_Memory::FastMem9Start : ARMJIT_Memory::FastMem7Start);
desc.EmulatedFaultAddr = (u8*)info->si_addr - curArea;
desc.FaultPC = (u8*)context->CONTEXT_PC;
if (ARMJIT_Memory::FaultHandler(desc))
{
context->CONTEXT_PC = (u64)desc.FaultPC;
return;
}
struct sigaction* oldSa;
if (sig == SIGSEGV)
oldSa = &OldSaSegv;
else
oldSa = &OldSaBus;
if (oldSa->sa_flags & SA_SIGINFO)
{
oldSa->sa_sigaction(sig, info, rawContext);
return;
}
if (oldSa->sa_handler == SIG_DFL)
{
signal(sig, SIG_DFL);
return;
}
if (oldSa->sa_handler == SIG_IGN)
{
// Ignore signal
return;
}
oldSa->sa_handler(sig);
}
#endif
namespace ARMJIT_Memory
{
void* FastMem9Start, *FastMem7Start;
#ifdef _WIN32
inline u32 RoundUp(u32 size)
{
return (size + 0xFFFF) & ~0xFFFF;
}
#else
inline u32 RoundUp(u32 size)
{
return size;
}
#endif
const u32 MemBlockMainRAMOffset = 0;
const u32 MemBlockSWRAMOffset = RoundUp(NDS::MainRAMMaxSize);
const u32 MemBlockARM7WRAMOffset = MemBlockSWRAMOffset + RoundUp(NDS::SharedWRAMSize);
const u32 MemBlockDTCMOffset = MemBlockARM7WRAMOffset + RoundUp(NDS::ARM7WRAMSize);
const u32 MemBlockNWRAM_AOffset = MemBlockDTCMOffset + RoundUp(DTCMPhysicalSize);
const u32 MemBlockNWRAM_BOffset = MemBlockNWRAM_AOffset + RoundUp(DSi::NWRAMSize);
const u32 MemBlockNWRAM_COffset = MemBlockNWRAM_BOffset + RoundUp(DSi::NWRAMSize);
const u32 MemoryTotalSize = MemBlockNWRAM_COffset + RoundUp(DSi::NWRAMSize);
const u32 OffsetsPerRegion[memregions_Count] =
{
UINT32_MAX,
UINT32_MAX,
MemBlockDTCMOffset,
UINT32_MAX,
MemBlockMainRAMOffset,
MemBlockSWRAMOffset,
UINT32_MAX,
UINT32_MAX,
UINT32_MAX,
MemBlockARM7WRAMOffset,
UINT32_MAX,
UINT32_MAX,
UINT32_MAX,
UINT32_MAX,
UINT32_MAX,
MemBlockNWRAM_AOffset,
MemBlockNWRAM_BOffset,
MemBlockNWRAM_COffset
};
enum
{
memstate_Unmapped,
memstate_MappedRW,
// on Switch this is unmapped as well
memstate_MappedProtected,
};
u8 MappingStatus9[1 << (32-12)];
u8 MappingStatus7[1 << (32-12)];
#if defined(__SWITCH__)
VirtmemReservation* FastMem9Reservation, *FastMem7Reservation;
u8* MemoryBase;
u8* MemoryBaseCodeMem;
#elif defined(_WIN32)
u8* MemoryBase;
HANDLE MemoryFile;
LPVOID ExceptionHandlerHandle;
#else
u8* MemoryBase;
int MemoryFile;
#endif
bool MapIntoRange(u32 addr, u32 num, u32 offset, u32 size)
{
u8* dst = (u8*)(num == 0 ? FastMem9Start : FastMem7Start) + addr;
#ifdef __SWITCH__
Result r = (svcMapProcessMemory(dst, envGetOwnProcessHandle(),
(u64)(MemoryBaseCodeMem + offset), size));
return R_SUCCEEDED(r);
#elif defined(_WIN32)
bool r = MapViewOfFileEx(MemoryFile, FILE_MAP_READ | FILE_MAP_WRITE, 0, offset, size, dst) == dst;
return r;
#else
return mmap(dst, size, PROT_READ | PROT_WRITE, MAP_SHARED | MAP_FIXED, MemoryFile, offset) != MAP_FAILED;
#endif
}
bool UnmapFromRange(u32 addr, u32 num, u32 offset, u32 size)
{
u8* dst = (u8*)(num == 0 ? FastMem9Start : FastMem7Start) + addr;
#ifdef __SWITCH__
Result r = svcUnmapProcessMemory(dst, envGetOwnProcessHandle(),
(u64)(MemoryBaseCodeMem + offset), size);
return R_SUCCEEDED(r);
#elif defined(_WIN32)
return UnmapViewOfFile(dst);
#else
return munmap(dst, size) == 0;
#endif
}
#ifndef __SWITCH__
void SetCodeProtectionRange(u32 addr, u32 size, u32 num, int protection)
{
u8* dst = (u8*)(num == 0 ? FastMem9Start : FastMem7Start) + addr;
#if defined(_WIN32)
DWORD winProtection, oldProtection;
if (protection == 0)
winProtection = PAGE_NOACCESS;
else if (protection == 1)
winProtection = PAGE_READONLY;
else
winProtection = PAGE_READWRITE;
bool success = VirtualProtect(dst, size, winProtection, &oldProtection);
assert(success);
#else
int posixProt;
if (protection == 0)
posixProt = PROT_NONE;
else if (protection == 1)
posixProt = PROT_READ;
else
posixProt = PROT_READ | PROT_WRITE;
mprotect(dst, size, posixProt);
#endif
}
#endif
struct Mapping
{
u32 Addr;
u32 Size, LocalOffset;
u32 Num;
void Unmap(int region)
{
u32 dtcmStart = NDS::ARM9->DTCMBase;
u32 dtcmSize = ~NDS::ARM9->DTCMMask + 1;
bool skipDTCM = Num == 0 && region != memregion_DTCM;
u8* statuses = Num == 0 ? MappingStatus9 : MappingStatus7;
u32 offset = 0;
while (offset < Size)
{
if (skipDTCM && Addr + offset == dtcmStart)
{
offset += dtcmSize;
}
else
{
u32 segmentOffset = offset;
u8 status = statuses[(Addr + offset) >> 12];
while (statuses[(Addr + offset) >> 12] == status
&& offset < Size
&& (!skipDTCM || Addr + offset != dtcmStart))
{
assert(statuses[(Addr + offset) >> 12] != memstate_Unmapped);
statuses[(Addr + offset) >> 12] = memstate_Unmapped;
offset += 0x1000;
}
#ifdef __SWITCH__
if (status == memstate_MappedRW)
{
u32 segmentSize = offset - segmentOffset;
printf("unmapping %x %x %x %x\n", Addr + segmentOffset, Num, segmentOffset + LocalOffset + OffsetsPerRegion[region], segmentSize);
bool success = UnmapFromRange(Addr + segmentOffset, Num, segmentOffset + LocalOffset + OffsetsPerRegion[region], segmentSize);
assert(success);
}
#endif
}
}
#ifndef __SWITCH__
#ifndef _WIN32
u32 dtcmEnd = dtcmStart + dtcmSize;
if (Num == 0
&& dtcmEnd >= Addr
&& dtcmStart < Addr + Size)
{
bool success;
if (dtcmStart > Addr)
{
success = UnmapFromRange(Addr, 0, OffsetsPerRegion[region] + LocalOffset, dtcmStart - Addr);
assert(success);
}
if (dtcmEnd < Addr + Size)
{
u32 offset = dtcmStart - Addr + dtcmSize;
success = UnmapFromRange(dtcmEnd, 0, OffsetsPerRegion[region] + LocalOffset + offset, Size - offset);
assert(success);
}
}
else
#endif
{
bool succeded = UnmapFromRange(Addr, Num, OffsetsPerRegion[region] + LocalOffset, Size);
assert(succeded);
}
#endif
}
};
ARMJIT::TinyVector<Mapping> Mappings[memregions_Count];
void SetCodeProtection(int region, u32 offset, bool protect)
{
offset &= ~0xFFF;
//printf("set code protection %d %x %d\n", region, offset, protect);
for (int i = 0; i < Mappings[region].Length; i++)
{
Mapping& mapping = Mappings[region][i];
if (offset < mapping.LocalOffset || offset >= mapping.LocalOffset + mapping.Size)
continue;
u32 effectiveAddr = mapping.Addr + (offset - mapping.LocalOffset);
if (mapping.Num == 0
&& region != memregion_DTCM
&& (effectiveAddr & NDS::ARM9->DTCMMask) == NDS::ARM9->DTCMBase)
continue;
u8* states = (u8*)(mapping.Num == 0 ? MappingStatus9 : MappingStatus7);
//printf("%x %d %x %x %x %d\n", effectiveAddr, mapping.Num, mapping.Addr, mapping.LocalOffset, mapping.Size, states[effectiveAddr >> 12]);
assert(states[effectiveAddr >> 12] == (protect ? memstate_MappedRW : memstate_MappedProtected));
states[effectiveAddr >> 12] = protect ? memstate_MappedProtected : memstate_MappedRW;
#if defined(__SWITCH__)
bool success;
if (protect)
success = UnmapFromRange(effectiveAddr, mapping.Num, OffsetsPerRegion[region] + offset, 0x1000);
else
success = MapIntoRange(effectiveAddr, mapping.Num, OffsetsPerRegion[region] + offset, 0x1000);
assert(success);
#else
SetCodeProtectionRange(effectiveAddr, 0x1000, mapping.Num, protect ? 1 : 2);
#endif
}
}
void RemapDTCM(u32 newBase, u32 newSize)
{
// this first part could be made more efficient
// by unmapping DTCM first and then map the holes
u32 oldDTCMBase = NDS::ARM9->DTCMBase;
u32 oldDTCMSize = ~NDS::ARM9->DTCMMask + 1;
u32 oldDTCMEnd = oldDTCMBase + NDS::ARM9->DTCMMask;
u32 newEnd = newBase + newSize;
printf("remapping DTCM %x %x %x %x\n", newBase, newEnd, oldDTCMBase, oldDTCMEnd);
// unmap all regions containing the old or the current DTCM mapping
for (int region = 0; region < memregions_Count; region++)
{
if (region == memregion_DTCM)
continue;
for (int i = 0; i < Mappings[region].Length;)
{
Mapping& mapping = Mappings[region][i];
u32 start = mapping.Addr;
u32 end = mapping.Addr + mapping.Size;
printf("unmapping %d %x %x %x %x\n", region, mapping.Addr, mapping.Size, mapping.Num, mapping.LocalOffset);
bool overlap = (oldDTCMSize > 0 && oldDTCMBase < end && oldDTCMEnd > start)
|| (newSize > 0 && newBase < end && newEnd > start);
if (mapping.Num == 0 && overlap)
{
mapping.Unmap(region);
Mappings[region].Remove(i);
}
else
{
i++;
}
}
}
for (int i = 0; i < Mappings[memregion_DTCM].Length; i++)
{
Mappings[memregion_DTCM][i].Unmap(memregion_DTCM);
}
Mappings[memregion_DTCM].Clear();
}
void RemapNWRAM(int num)
{
for (int i = 0; i < Mappings[memregion_SharedWRAM].Length;)
{
Mapping& mapping = Mappings[memregion_SharedWRAM][i];
if (DSi::NWRAMStart[mapping.Num][num] < mapping.Addr + mapping.Size
&& DSi::NWRAMEnd[mapping.Num][num] > mapping.Addr)
{
mapping.Unmap(memregion_SharedWRAM);
Mappings[memregion_SharedWRAM].Remove(i);
}
else
{
i++;
}
}
for (int i = 0; i < Mappings[memregion_NewSharedWRAM_A + num].Length; i++)
{
Mappings[memregion_NewSharedWRAM_A + num][i].Unmap(memregion_NewSharedWRAM_A + num);
}
Mappings[memregion_NewSharedWRAM_A + num].Clear();
}
void RemapSWRAM()
{
printf("remapping SWRAM\n");
for (int i = 0; i < Mappings[memregion_WRAM7].Length;)
{
Mapping& mapping = Mappings[memregion_WRAM7][i];
if (mapping.Addr + mapping.Size <= 0x03800000)
{
mapping.Unmap(memregion_WRAM7);
Mappings[memregion_WRAM7].Remove(i);
}
else
i++;
}
for (int i = 0; i < Mappings[memregion_SharedWRAM].Length; i++)
{
Mappings[memregion_SharedWRAM][i].Unmap(memregion_SharedWRAM);
}
Mappings[memregion_SharedWRAM].Clear();
}
bool MapAtAddress(u32 addr)
{
u32 num = NDS::CurCPU;
int region = num == 0
? ClassifyAddress9(addr)
: ClassifyAddress7(addr);
if (!IsFastmemCompatible(region))
return false;
u32 mirrorStart, mirrorSize, memoryOffset;
bool isMapped = GetMirrorLocation(region, num, addr, memoryOffset, mirrorStart, mirrorSize);
if (!isMapped)
return false;
u8* states = num == 0 ? MappingStatus9 : MappingStatus7;
//printf("mapping mirror %x, %x %x %d %d\n", mirrorStart, mirrorSize, memoryOffset, region, num);
bool isExecutable = ARMJIT::CodeMemRegions[region];
u32 dtcmStart = NDS::ARM9->DTCMBase;
u32 dtcmSize = ~NDS::ARM9->DTCMMask + 1;
u32 dtcmEnd = dtcmStart + dtcmSize;
#ifndef __SWITCH__
#ifndef _WIN32
if (num == 0
&& dtcmEnd >= mirrorStart
&& dtcmStart < mirrorStart + mirrorSize)
{
bool success;
if (dtcmStart > mirrorStart)
{
success = MapIntoRange(mirrorStart, 0, OffsetsPerRegion[region] + memoryOffset, dtcmStart - mirrorStart);
assert(success);
}
if (dtcmEnd < mirrorStart + mirrorSize)
{
u32 offset = dtcmStart - mirrorStart + dtcmSize;
success = MapIntoRange(dtcmEnd, 0, OffsetsPerRegion[region] + memoryOffset + offset, mirrorSize - offset);
assert(success);
}
}
else
#endif
{
bool succeded = MapIntoRange(mirrorStart, num, OffsetsPerRegion[region] + memoryOffset, mirrorSize);
assert(succeded);
}
#endif
ARMJIT::AddressRange* range = ARMJIT::CodeMemRegions[region] + memoryOffset / 512;
// this overcomplicated piece of code basically just finds whole pieces of code memory
// which can be mapped/protected
u32 offset = 0;
bool skipDTCM = num == 0 && region != memregion_DTCM;
while (offset < mirrorSize)
{
if (skipDTCM && mirrorStart + offset == dtcmStart)
{
#ifdef _WIN32
SetCodeProtectionRange(dtcmStart, dtcmSize, 0, 0);
#endif
offset += dtcmSize;
}
else
{
u32 sectionOffset = offset;
bool hasCode = isExecutable && ARMJIT::PageContainsCode(&range[offset / 512]);
while (offset < mirrorSize
&& (!isExecutable || ARMJIT::PageContainsCode(&range[offset / 512]) == hasCode)
&& (!skipDTCM || mirrorStart + offset != NDS::ARM9->DTCMBase))
{
assert(states[(mirrorStart + offset) >> 12] == memstate_Unmapped);
states[(mirrorStart + offset) >> 12] = hasCode ? memstate_MappedProtected : memstate_MappedRW;
offset += 0x1000;
}
u32 sectionSize = offset - sectionOffset;
#if defined(__SWITCH__)
if (!hasCode)
{
//printf("trying to map %x (size: %x) from %x\n", mirrorStart + sectionOffset, sectionSize, sectionOffset + memoryOffset + OffsetsPerRegion[region]);
bool succeded = MapIntoRange(mirrorStart + sectionOffset, num, sectionOffset + memoryOffset + OffsetsPerRegion[region], sectionSize);
assert(succeded);
}
#else
if (hasCode)
{
SetCodeProtectionRange(mirrorStart + sectionOffset, sectionSize, num, 1);
}
#endif
}
}
assert(num == 0 || num == 1);
Mapping mapping{mirrorStart, mirrorSize, memoryOffset, num};
Mappings[region].Add(mapping);
//printf("mapped mirror at %08x-%08x\n", mirrorStart, mirrorStart + mirrorSize - 1);
return true;
}
bool FaultHandler(FaultDescription& faultDesc)
{
if (ARMJIT::JITCompiler->IsJITFault(faultDesc.FaultPC))
{
bool rewriteToSlowPath = true;
u8* memStatus = NDS::CurCPU == 0 ? MappingStatus9 : MappingStatus7;
if (memStatus[faultDesc.EmulatedFaultAddr >> 12] == memstate_Unmapped)
rewriteToSlowPath = !MapAtAddress(faultDesc.EmulatedFaultAddr);
if (rewriteToSlowPath)
faultDesc.FaultPC = ARMJIT::JITCompiler->RewriteMemAccess(faultDesc.FaultPC);
return true;
}
return false;
}
const u64 AddrSpaceSize = 0x100000000;
void Init()
{
#if defined(__SWITCH__)
MemoryBase = (u8*)aligned_alloc(0x1000, MemoryTotalSize);
virtmemLock();
MemoryBaseCodeMem = (u8*)virtmemFindCodeMemory(MemoryTotalSize, 0x1000);
bool succeded = R_SUCCEEDED(svcMapProcessCodeMemory(envGetOwnProcessHandle(), (u64)MemoryBaseCodeMem,
(u64)MemoryBase, MemoryTotalSize));
assert(succeded);
succeded = R_SUCCEEDED(svcSetProcessMemoryPermission(envGetOwnProcessHandle(), (u64)MemoryBaseCodeMem,
MemoryTotalSize, Perm_Rw));
assert(succeded);
// 8 GB of address space, just don't ask...
FastMem9Start = virtmemFindAslr(AddrSpaceSize, 0x1000);
assert(FastMem9Start);
FastMem7Start = virtmemFindAslr(AddrSpaceSize, 0x1000);
assert(FastMem7Start);
FastMem9Reservation = virtmemAddReservation(FastMem9Start, AddrSpaceSize);
FastMem7Reservation = virtmemAddReservation(FastMem7Start, AddrSpaceSize);
virtmemUnlock();
u8* basePtr = MemoryBaseCodeMem;
#elif defined(_WIN32)
ExceptionHandlerHandle = AddVectoredExceptionHandler(1, ExceptionHandler);
MemoryFile = CreateFileMapping(INVALID_HANDLE_VALUE, NULL, PAGE_READWRITE, 0, MemoryTotalSize, NULL);
MemoryBase = (u8*)VirtualAlloc(NULL, AddrSpaceSize*4, MEM_RESERVE, PAGE_READWRITE);
VirtualFree(MemoryBase, 0, MEM_RELEASE);
// this is incredible hacky
// but someone else is trying to go into our address space!
// Windows will very likely give them virtual memory starting at the same address
// as it is giving us now.
// That's why we don't use this address, but instead 4gb inwards
// I know this is terrible
FastMem9Start = MemoryBase + AddrSpaceSize;
FastMem7Start = MemoryBase + AddrSpaceSize*2;
MemoryBase = MemoryBase + AddrSpaceSize*3;
MapViewOfFileEx(MemoryFile, FILE_MAP_READ | FILE_MAP_WRITE, 0, 0, MemoryTotalSize, MemoryBase);
u8* basePtr = MemoryBase;
#else
// this used to be allocated with three different mmaps
// The idea was to give the OS more freedom where to position the buffers,
// but something was bad about this so instead we take this vmem eating monster
// which seems to work better.
MemoryBase = (u8*)mmap(NULL, AddrSpaceSize*4, PROT_NONE, MAP_ANON | MAP_PRIVATE, -1, 0);
munmap(MemoryBase, AddrSpaceSize*4);
FastMem9Start = MemoryBase;
FastMem7Start = MemoryBase + AddrSpaceSize;
MemoryBase = MemoryBase + AddrSpaceSize*2;
#if defined(__ANDROID__)
static void* libandroid = dlopen("libandroid.so", RTLD_LAZY | RTLD_LOCAL);
using type_ASharedMemory_create = int(*)(const char* name, size_t size);
static void* symbol = dlsym(libandroid, "ASharedMemory_create");
static auto shared_memory_create = reinterpret_cast<type_ASharedMemory_create>(symbol);
if (shared_memory_create)
{
MemoryFile = shared_memory_create("melondsfastmem", MemoryTotalSize);
}
else
{
int fd = open(ASHMEM_DEVICE, O_RDWR);
ioctl(fd, ASHMEM_SET_NAME, "melondsfastmem");
ioctl(fd, ASHMEM_SET_SIZE, MemoryTotalSize);
MemoryFile = fd;
}
#else
char fastmemPidName[snprintf(NULL, 0, "/melondsfastmem%d", getpid()) + 1];
sprintf(fastmemPidName, "/melondsfastmem%d", getpid());
MemoryFile = shm_open(fastmemPidName, O_RDWR | O_CREAT | O_EXCL, 0600);
if (MemoryFile == -1)
{
printf("Failed to open memory using shm_open!");
}
shm_unlink(fastmemPidName);
#endif
if (ftruncate(MemoryFile, MemoryTotalSize) < 0)
{
printf("Failed to allocate memory using ftruncate!");
}
struct sigaction sa;
sa.sa_handler = nullptr;
sa.sa_sigaction = &SigsegvHandler;
sa.sa_flags = SA_SIGINFO;
sigemptyset(&sa.sa_mask);
sigaction(SIGSEGV, &sa, &OldSaSegv);
#ifdef __APPLE__
sigaction(SIGBUS, &sa, &OldSaBus);
#endif
mmap(MemoryBase, MemoryTotalSize, PROT_READ | PROT_WRITE, MAP_SHARED | MAP_FIXED, MemoryFile, 0);
u8* basePtr = MemoryBase;
#endif
NDS::MainRAM = basePtr + MemBlockMainRAMOffset;
NDS::SharedWRAM = basePtr + MemBlockSWRAMOffset;
NDS::ARM7WRAM = basePtr + MemBlockARM7WRAMOffset;
NDS::ARM9->DTCM = basePtr + MemBlockDTCMOffset;
DSi::NWRAM_A = basePtr + MemBlockNWRAM_AOffset;
DSi::NWRAM_B = basePtr + MemBlockNWRAM_BOffset;
DSi::NWRAM_C = basePtr + MemBlockNWRAM_COffset;
}
void DeInit()
{
#if defined(__SWITCH__)
virtmemLock();
virtmemRemoveReservation(FastMem9Reservation);
virtmemRemoveReservation(FastMem7Reservation);
virtmemUnlock();
svcUnmapProcessCodeMemory(envGetOwnProcessHandle(), (u64)MemoryBaseCodeMem, (u64)MemoryBase, MemoryTotalSize);
free(MemoryBase);
#elif defined(_WIN32)
assert(UnmapViewOfFile(MemoryBase));
CloseHandle(MemoryFile);
RemoveVectoredExceptionHandler(ExceptionHandlerHandle);
#else
sigaction(SIGSEGV, &OldSaSegv, nullptr);
#ifdef __APPLE__
sigaction(SIGBUS, &OldSaBus, nullptr);
#endif
munmap(MemoryBase, MemoryTotalSize);
close(MemoryFile);
#endif
}
void Reset()
{
for (int region = 0; region < memregions_Count; region++)
{
for (int i = 0; i < Mappings[region].Length; i++)
Mappings[region][i].Unmap(region);
Mappings[region].Clear();
}
for (size_t i = 0; i < sizeof(MappingStatus9); i++)
{
assert(MappingStatus9[i] == memstate_Unmapped);
assert(MappingStatus7[i] == memstate_Unmapped);
}
printf("done resetting jit mem\n");
}
bool IsFastmemCompatible(int region)
{
#ifdef _WIN32
/*
TODO: with some hacks, the smaller shared WRAM regions
could be mapped in some occaisons as well
*/
if (region == memregion_DTCM
|| region == memregion_SharedWRAM
|| region == memregion_NewSharedWRAM_B
|| region == memregion_NewSharedWRAM_C)
return false;
#endif
return OffsetsPerRegion[region] != UINT32_MAX;
}
bool GetMirrorLocation(int region, u32 num, u32 addr, u32& memoryOffset, u32& mirrorStart, u32& mirrorSize)
{
memoryOffset = 0;
switch (region)
{
case memregion_ITCM:
if (num == 0)
{
mirrorStart = addr & ~(ITCMPhysicalSize - 1);
mirrorSize = ITCMPhysicalSize;
return true;
}
return false;
case memregion_DTCM:
if (num == 0)
{
mirrorStart = addr & ~(DTCMPhysicalSize - 1);
mirrorSize = DTCMPhysicalSize;
return true;
}
return false;
case memregion_MainRAM:
mirrorStart = addr & ~NDS::MainRAMMask;
mirrorSize = NDS::MainRAMMask + 1;
return true;
case memregion_BIOS9:
if (num == 0)
{
mirrorStart = addr & ~0xFFF;
mirrorSize = 0x1000;
return true;
}
return false;
case memregion_BIOS7:
if (num == 1)
{
mirrorStart = 0;
mirrorSize = 0x4000;
return true;
}
return false;
case memregion_SharedWRAM:
if (num == 0 && NDS::SWRAM_ARM9.Mem)
{
mirrorStart = addr & ~NDS::SWRAM_ARM9.Mask;
mirrorSize = NDS::SWRAM_ARM9.Mask + 1;
memoryOffset = NDS::SWRAM_ARM9.Mem - NDS::SharedWRAM;
return true;
}
else if (num == 1 && NDS::SWRAM_ARM7.Mem)
{
mirrorStart = addr & ~NDS::SWRAM_ARM7.Mask;
mirrorSize = NDS::SWRAM_ARM7.Mask + 1;
memoryOffset = NDS::SWRAM_ARM7.Mem - NDS::SharedWRAM;
return true;
}
return false;
case memregion_WRAM7:
if (num == 1)
{
mirrorStart = addr & ~(NDS::ARM7WRAMSize - 1);
mirrorSize = NDS::ARM7WRAMSize;
return true;
}
return false;
case memregion_VRAM:
if (num == 0)
{
mirrorStart = addr & ~0xFFFFF;
mirrorSize = 0x100000;
return true;
}
return false;
case memregion_VWRAM:
if (num == 1)
{
mirrorStart = addr & ~0x3FFFF;
mirrorSize = 0x40000;
return true;
}
return false;
case memregion_NewSharedWRAM_A:
{
u8* ptr = DSi::NWRAMMap_A[num][(addr >> 16) & DSi::NWRAMMask[num][0]];
if (ptr)
{
memoryOffset = ptr - DSi::NWRAM_A;
mirrorStart = addr & ~0xFFFF;
mirrorSize = 0x10000;
return true;
}
return false; // zero filled memory
}
case memregion_NewSharedWRAM_B:
{
u8* ptr = DSi::NWRAMMap_B[num][(addr >> 15) & DSi::NWRAMMask[num][1]];
if (ptr)
{
memoryOffset = ptr - DSi::NWRAM_B;
mirrorStart = addr & ~0x7FFF;
mirrorSize = 0x8000;
return true;
}
return false; // zero filled memory
}
case memregion_NewSharedWRAM_C:
{
u8* ptr = DSi::NWRAMMap_C[num][(addr >> 15) & DSi::NWRAMMask[num][2]];
if (ptr)
{
memoryOffset = ptr - DSi::NWRAM_C;
mirrorStart = addr & ~0x7FFF;
mirrorSize = 0x8000;
return true;
}
return false; // zero filled memory
}
case memregion_BIOS9DSi:
if (num == 0)
{
mirrorStart = addr & ~0xFFFF;
mirrorSize = DSi::SCFG_BIOS & (1<<0) ? 0x8000 : 0x10000;
return true;
}
return false;
case memregion_BIOS7DSi:
if (num == 1)
{
mirrorStart = addr & ~0xFFFF;
mirrorSize = DSi::SCFG_BIOS & (1<<8) ? 0x8000 : 0x10000;
return true;
}
return false;
default: