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
|
// Copyright 2018 yuzu Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.
#include "common/assert.h"
#include "core/core.h"
#include "core/core_timing.h"
#include "core/memory.h"
#include "core/settings.h"
#include "video_core/engines/fermi_2d.h"
#include "video_core/engines/kepler_compute.h"
#include "video_core/engines/kepler_memory.h"
#include "video_core/engines/maxwell_3d.h"
#include "video_core/engines/maxwell_dma.h"
#include "video_core/gpu.h"
#include "video_core/gpu_thread.h"
#include "video_core/renderer_base.h"
namespace Tegra {
u32 FramebufferConfig::BytesPerPixel(PixelFormat format) {
switch (format) {
case PixelFormat::ABGR8:
case PixelFormat::BGRA8:
return 4;
default:
return 4;
}
UNREACHABLE();
}
GPU::GPU(Core::System& system, VideoCore::RendererBase& renderer) : renderer{renderer} {
auto& rasterizer{renderer.Rasterizer()};
memory_manager = std::make_unique<Tegra::MemoryManager>();
dma_pusher = std::make_unique<Tegra::DmaPusher>(*this);
maxwell_3d = std::make_unique<Engines::Maxwell3D>(system, rasterizer, *memory_manager);
fermi_2d = std::make_unique<Engines::Fermi2D>(rasterizer, *memory_manager);
kepler_compute = std::make_unique<Engines::KeplerCompute>(*memory_manager);
maxwell_dma = std::make_unique<Engines::MaxwellDMA>(system, rasterizer, *memory_manager);
kepler_memory = std::make_unique<Engines::KeplerMemory>(system, rasterizer, *memory_manager);
if (Settings::values.use_asynchronous_gpu_emulation) {
gpu_thread = std::make_unique<VideoCommon::GPUThread::ThreadManager>(renderer, *dma_pusher);
}
}
GPU::~GPU() = default;
Engines::Maxwell3D& GPU::Maxwell3D() {
return *maxwell_3d;
}
const Engines::Maxwell3D& GPU::Maxwell3D() const {
return *maxwell_3d;
}
MemoryManager& GPU::MemoryManager() {
return *memory_manager;
}
const MemoryManager& GPU::MemoryManager() const {
return *memory_manager;
}
DmaPusher& GPU::DmaPusher() {
return *dma_pusher;
}
const DmaPusher& GPU::DmaPusher() const {
return *dma_pusher;
}
void GPU::PushGPUEntries(Tegra::CommandList&& entries) {
if (Settings::values.use_asynchronous_gpu_emulation) {
gpu_thread->SubmitList(std::move(entries));
} else {
dma_pusher->Push(std::move(entries));
dma_pusher->DispatchCalls();
}
}
void GPU::SwapBuffers(
std::optional<std::reference_wrapper<const Tegra::FramebufferConfig>> framebuffer) {
if (Settings::values.use_asynchronous_gpu_emulation) {
gpu_thread->SwapBuffers(std::move(framebuffer));
} else {
renderer.SwapBuffers(std::move(framebuffer));
}
}
void GPU::FlushRegion(VAddr addr, u64 size) {
if (Settings::values.use_asynchronous_gpu_emulation) {
gpu_thread->FlushRegion(addr, size);
} else {
renderer.Rasterizer().FlushRegion(addr, size);
}
}
void GPU::InvalidateRegion(VAddr addr, u64 size) {
if (Settings::values.use_asynchronous_gpu_emulation) {
gpu_thread->InvalidateRegion(addr, size);
} else {
renderer.Rasterizer().InvalidateRegion(addr, size);
}
}
void GPU::FlushAndInvalidateRegion(VAddr addr, u64 size) {
if (Settings::values.use_asynchronous_gpu_emulation) {
gpu_thread->FlushAndInvalidateRegion(addr, size);
} else {
renderer.Rasterizer().FlushAndInvalidateRegion(addr, size);
}
}
u32 RenderTargetBytesPerPixel(RenderTargetFormat format) {
ASSERT(format != RenderTargetFormat::NONE);
switch (format) {
case RenderTargetFormat::RGBA32_FLOAT:
case RenderTargetFormat::RGBA32_UINT:
return 16;
case RenderTargetFormat::RGBA16_UINT:
case RenderTargetFormat::RGBA16_UNORM:
case RenderTargetFormat::RGBA16_FLOAT:
case RenderTargetFormat::RG32_FLOAT:
case RenderTargetFormat::RG32_UINT:
return 8;
case RenderTargetFormat::RGBA8_UNORM:
case RenderTargetFormat::RGBA8_SNORM:
case RenderTargetFormat::RGBA8_SRGB:
case RenderTargetFormat::RGBA8_UINT:
case RenderTargetFormat::RGB10_A2_UNORM:
case RenderTargetFormat::BGRA8_UNORM:
case RenderTargetFormat::BGRA8_SRGB:
case RenderTargetFormat::RG16_UNORM:
case RenderTargetFormat::RG16_SNORM:
case RenderTargetFormat::RG16_UINT:
case RenderTargetFormat::RG16_SINT:
case RenderTargetFormat::RG16_FLOAT:
case RenderTargetFormat::R32_FLOAT:
case RenderTargetFormat::R11G11B10_FLOAT:
case RenderTargetFormat::R32_UINT:
return 4;
case RenderTargetFormat::R16_UNORM:
case RenderTargetFormat::R16_SNORM:
case RenderTargetFormat::R16_UINT:
case RenderTargetFormat::R16_SINT:
case RenderTargetFormat::R16_FLOAT:
case RenderTargetFormat::RG8_UNORM:
case RenderTargetFormat::RG8_SNORM:
return 2;
case RenderTargetFormat::R8_UNORM:
case RenderTargetFormat::R8_UINT:
return 1;
default:
UNIMPLEMENTED_MSG("Unimplemented render target format {}", static_cast<u32>(format));
return 1;
}
}
u32 DepthFormatBytesPerPixel(DepthFormat format) {
switch (format) {
case DepthFormat::Z32_S8_X24_FLOAT:
return 8;
case DepthFormat::Z32_FLOAT:
case DepthFormat::S8_Z24_UNORM:
case DepthFormat::Z24_X8_UNORM:
case DepthFormat::Z24_S8_UNORM:
case DepthFormat::Z24_C8_UNORM:
return 4;
case DepthFormat::Z16_UNORM:
return 2;
default:
UNIMPLEMENTED_MSG("Unimplemented Depth format {}", static_cast<u32>(format));
return 1;
}
}
// Note that, traditionally, methods are treated as 4-byte addressable locations, and hence
// their numbers are written down multiplied by 4 in Docs. Here we are not multiply by 4.
// So the values you see in docs might be multiplied by 4.
enum class BufferMethods {
BindObject = 0x0,
Nop = 0x2,
SemaphoreAddressHigh = 0x4,
SemaphoreAddressLow = 0x5,
SemaphoreSequence = 0x6,
SemaphoreTrigger = 0x7,
NotifyIntr = 0x8,
WrcacheFlush = 0x9,
Unk28 = 0xA,
Unk2c = 0xB,
RefCnt = 0x14,
SemaphoreAcquire = 0x1A,
SemaphoreRelease = 0x1B,
Unk70 = 0x1C,
Unk74 = 0x1D,
Unk78 = 0x1E,
Unk7c = 0x1F,
Yield = 0x20,
NonPullerMethods = 0x40,
};
enum class GpuSemaphoreOperation {
AcquireEqual = 0x1,
WriteLong = 0x2,
AcquireGequal = 0x4,
AcquireMask = 0x8,
};
void GPU::CallMethod(const MethodCall& method_call) {
LOG_TRACE(HW_GPU, "Processing method {:08X} on subchannel {}", method_call.method,
method_call.subchannel);
ASSERT(method_call.subchannel < bound_engines.size());
if (ExecuteMethodOnEngine(method_call)) {
CallEngineMethod(method_call);
} else {
CallPullerMethod(method_call);
}
}
bool GPU::ExecuteMethodOnEngine(const MethodCall& method_call) {
const auto method = static_cast<BufferMethods>(method_call.method);
return method >= BufferMethods::NonPullerMethods;
}
void GPU::CallPullerMethod(const MethodCall& method_call) {
regs.reg_array[method_call.method] = method_call.argument;
const auto method = static_cast<BufferMethods>(method_call.method);
switch (method) {
case BufferMethods::BindObject: {
ProcessBindMethod(method_call);
break;
}
case BufferMethods::Nop:
case BufferMethods::SemaphoreAddressHigh:
case BufferMethods::SemaphoreAddressLow:
case BufferMethods::SemaphoreSequence:
case BufferMethods::RefCnt:
break;
case BufferMethods::SemaphoreTrigger: {
ProcessSemaphoreTriggerMethod();
break;
}
case BufferMethods::NotifyIntr: {
// TODO(Kmather73): Research and implement this method.
LOG_ERROR(HW_GPU, "Special puller engine method NotifyIntr not implemented");
break;
}
case BufferMethods::WrcacheFlush: {
// TODO(Kmather73): Research and implement this method.
LOG_ERROR(HW_GPU, "Special puller engine method WrcacheFlush not implemented");
break;
}
case BufferMethods::Unk28: {
// TODO(Kmather73): Research and implement this method.
LOG_ERROR(HW_GPU, "Special puller engine method Unk28 not implemented");
break;
}
case BufferMethods::Unk2c: {
// TODO(Kmather73): Research and implement this method.
LOG_ERROR(HW_GPU, "Special puller engine method Unk2c not implemented");
break;
}
case BufferMethods::SemaphoreAcquire: {
ProcessSemaphoreAcquire();
break;
}
case BufferMethods::SemaphoreRelease: {
ProcessSemaphoreRelease();
break;
}
case BufferMethods::Yield: {
// TODO(Kmather73): Research and implement this method.
LOG_ERROR(HW_GPU, "Special puller engine method Yield not implemented");
break;
}
default:
LOG_ERROR(HW_GPU, "Special puller engine method {:X} not implemented",
static_cast<u32>(method));
break;
}
}
void GPU::CallEngineMethod(const MethodCall& method_call) {
const EngineID engine = bound_engines[method_call.subchannel];
switch (engine) {
case EngineID::FERMI_TWOD_A:
fermi_2d->CallMethod(method_call);
break;
case EngineID::MAXWELL_B:
maxwell_3d->CallMethod(method_call);
break;
case EngineID::KEPLER_COMPUTE_B:
kepler_compute->CallMethod(method_call);
break;
case EngineID::MAXWELL_DMA_COPY_A:
maxwell_dma->CallMethod(method_call);
break;
case EngineID::KEPLER_INLINE_TO_MEMORY_B:
kepler_memory->CallMethod(method_call);
break;
default:
UNIMPLEMENTED_MSG("Unimplemented engine");
}
}
void GPU::ProcessBindMethod(const MethodCall& method_call) {
// Bind the current subchannel to the desired engine id.
LOG_DEBUG(HW_GPU, "Binding subchannel {} to engine {}", method_call.subchannel,
method_call.argument);
bound_engines[method_call.subchannel] = static_cast<EngineID>(method_call.argument);
}
void GPU::ProcessSemaphoreTriggerMethod() {
const auto semaphoreOperationMask = 0xF;
const auto op =
static_cast<GpuSemaphoreOperation>(regs.semaphore_trigger & semaphoreOperationMask);
if (op == GpuSemaphoreOperation::WriteLong) {
auto address = memory_manager->GpuToCpuAddress(regs.smaphore_address.SmaphoreAddress());
struct Block {
u32 sequence;
u32 zeros = 0;
u64 timestamp;
};
Block block{};
block.sequence = regs.semaphore_sequence;
// TODO(Kmather73): Generate a real GPU timestamp and write it here instead of
// CoreTiming
block.timestamp = Core::System::GetInstance().CoreTiming().GetTicks();
Memory::WriteBlock(*address, &block, sizeof(block));
} else {
const auto address =
memory_manager->GpuToCpuAddress(regs.smaphore_address.SmaphoreAddress());
const u32 word = Memory::Read32(*address);
if ((op == GpuSemaphoreOperation::AcquireEqual && word == regs.semaphore_sequence) ||
(op == GpuSemaphoreOperation::AcquireGequal &&
static_cast<s32>(word - regs.semaphore_sequence) > 0) ||
(op == GpuSemaphoreOperation::AcquireMask && (word & regs.semaphore_sequence))) {
// Nothing to do in this case
} else {
regs.acquire_source = true;
regs.acquire_value = regs.semaphore_sequence;
if (op == GpuSemaphoreOperation::AcquireEqual) {
regs.acquire_active = true;
regs.acquire_mode = false;
} else if (op == GpuSemaphoreOperation::AcquireGequal) {
regs.acquire_active = true;
regs.acquire_mode = true;
} else if (op == GpuSemaphoreOperation::AcquireMask) {
// TODO(kemathe) The acquire mask operation waits for a value that, ANDed with
// semaphore_sequence, gives a non-0 result
LOG_ERROR(HW_GPU, "Invalid semaphore operation AcquireMask not implemented");
} else {
LOG_ERROR(HW_GPU, "Invalid semaphore operation");
}
}
}
}
void GPU::ProcessSemaphoreRelease() {
const auto address = memory_manager->GpuToCpuAddress(regs.smaphore_address.SmaphoreAddress());
Memory::Write32(*address, regs.semaphore_release);
}
void GPU::ProcessSemaphoreAcquire() {
const auto address = memory_manager->GpuToCpuAddress(regs.smaphore_address.SmaphoreAddress());
const u32 word = Memory::Read32(*address);
const auto value = regs.semaphore_acquire;
if (word != value) {
regs.acquire_active = true;
regs.acquire_value = value;
// TODO(kemathe73) figure out how to do the acquire_timeout
regs.acquire_mode = false;
regs.acquire_source = false;
}
}
} // namespace Tegra
|