summaryrefslogtreecommitdiffstats
path: root/src/video_core/gpu.h
blob: 56a2032759abf3b509aa3a096a090531b521005e (plain) (blame)
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
// Copyright 2018 yuzu Emulator Project
// Licensed under GPLv2 or any later version
// Refer to the license.txt file included.

#pragma once

#include <array>
#include <memory>
#include "common/common_types.h"
#include "core/hle/service/nvflinger/buffer_queue.h"
#include "video_core/dma_pusher.h"
#include "video_core/memory_manager.h"

namespace Core {
class System;
}

namespace VideoCore {
class RendererBase;
} // namespace VideoCore

namespace Tegra {

enum class RenderTargetFormat : u32 {
    NONE = 0x0,
    RGBA32_FLOAT = 0xC0,
    RGBA32_UINT = 0xC2,
    RGBA16_UNORM = 0xC6,
    RGBA16_UINT = 0xC9,
    RGBA16_FLOAT = 0xCA,
    RG32_FLOAT = 0xCB,
    RG32_UINT = 0xCD,
    BGRA8_UNORM = 0xCF,
    BGRA8_SRGB = 0xD0,
    RGB10_A2_UNORM = 0xD1,
    RGBA8_UNORM = 0xD5,
    RGBA8_SRGB = 0xD6,
    RGBA8_SNORM = 0xD7,
    RGBA8_UINT = 0xD9,
    RG16_UNORM = 0xDA,
    RG16_SNORM = 0xDB,
    RG16_SINT = 0xDC,
    RG16_UINT = 0xDD,
    RG16_FLOAT = 0xDE,
    R11G11B10_FLOAT = 0xE0,
    R32_UINT = 0xE4,
    R32_FLOAT = 0xE5,
    B5G6R5_UNORM = 0xE8,
    BGR5A1_UNORM = 0xE9,
    RG8_UNORM = 0xEA,
    RG8_SNORM = 0xEB,
    R16_UNORM = 0xEE,
    R16_SNORM = 0xEF,
    R16_SINT = 0xF0,
    R16_UINT = 0xF1,
    R16_FLOAT = 0xF2,
    R8_UNORM = 0xF3,
    R8_UINT = 0xF6,
};

enum class DepthFormat : u32 {
    Z32_FLOAT = 0xA,
    Z16_UNORM = 0x13,
    S8_Z24_UNORM = 0x14,
    Z24_X8_UNORM = 0x15,
    Z24_S8_UNORM = 0x16,
    Z24_C8_UNORM = 0x18,
    Z32_S8_X24_FLOAT = 0x19,
};

/// Returns the number of bytes per pixel of each rendertarget format.
u32 RenderTargetBytesPerPixel(RenderTargetFormat format);

/// Returns the number of bytes per pixel of each depth format.
u32 DepthFormatBytesPerPixel(DepthFormat format);

struct CommandListHeader;
class DebugContext;

/**
 * Struct describing framebuffer configuration
 */
struct FramebufferConfig {
    enum class PixelFormat : u32 {
        ABGR8 = 1,
        BGRA8 = 5,
    };

    /**
     * Returns the number of bytes per pixel.
     */
    static u32 BytesPerPixel(PixelFormat format);

    VAddr address;
    u32 offset;
    u32 width;
    u32 height;
    u32 stride;
    PixelFormat pixel_format;

    using TransformFlags = Service::NVFlinger::BufferQueue::BufferTransformFlags;
    TransformFlags transform_flags;
    Common::Rectangle<int> crop_rect;
};

namespace Engines {
class Fermi2D;
class Maxwell3D;
class MaxwellDMA;
class KeplerCompute;
class KeplerMemory;
} // namespace Engines

enum class EngineID {
    FERMI_TWOD_A = 0x902D, // 2D Engine
    MAXWELL_B = 0xB197,    // 3D Engine
    KEPLER_COMPUTE_B = 0xB1C0,
    KEPLER_INLINE_TO_MEMORY_B = 0xA140,
    MAXWELL_DMA_COPY_A = 0xB0B5,
};

class GPU {
public:
    explicit GPU(Core::System& system, VideoCore::RendererBase& renderer);

    virtual ~GPU();

    struct MethodCall {
        u32 method{};
        u32 argument{};
        u32 subchannel{};
        u32 method_count{};

        bool IsLastCall() const {
            return method_count <= 1;
        }

        MethodCall(u32 method, u32 argument, u32 subchannel = 0, u32 method_count = 0)
            : method(method), argument(argument), subchannel(subchannel),
              method_count(method_count) {}
    };

    /// Calls a GPU method.
    void CallMethod(const MethodCall& method_call);

    /// Returns a reference to the Maxwell3D GPU engine.
    Engines::Maxwell3D& Maxwell3D();

    /// Returns a const reference to the Maxwell3D GPU engine.
    const Engines::Maxwell3D& Maxwell3D() const;

    /// Returns a reference to the GPU memory manager.
    Tegra::MemoryManager& MemoryManager();

    /// Returns a const reference to the GPU memory manager.
    const Tegra::MemoryManager& MemoryManager() const;

    /// Returns a reference to the GPU DMA pusher.
    Tegra::DmaPusher& DmaPusher();

    /// Returns a const reference to the GPU DMA pusher.
    const Tegra::DmaPusher& DmaPusher() const;

    struct Regs {
        static constexpr size_t NUM_REGS = 0x100;

        union {
            struct {
                INSERT_PADDING_WORDS(0x4);
                struct {
                    u32 address_high;
                    u32 address_low;

                    GPUVAddr SmaphoreAddress() const {
                        return static_cast<GPUVAddr>((static_cast<GPUVAddr>(address_high) << 32) |
                                                     address_low);
                    }
                } smaphore_address;

                u32 semaphore_sequence;
                u32 semaphore_trigger;
                INSERT_PADDING_WORDS(0xC);

                // The puser and the puller share the reference counter, the pusher only has read
                // access
                u32 reference_count;
                INSERT_PADDING_WORDS(0x5);

                u32 semaphore_acquire;
                u32 semaphore_release;
                INSERT_PADDING_WORDS(0xE4);

                // Puller state
                u32 acquire_mode;
                u32 acquire_source;
                u32 acquire_active;
                u32 acquire_timeout;
                u32 acquire_value;
            };
            std::array<u32, NUM_REGS> reg_array;
        };
    } regs{};

    /// Push GPU command entries to be processed
    virtual void PushGPUEntries(Tegra::CommandList&& entries) = 0;

    /// Swap buffers (render frame)
    virtual void SwapBuffers(
        std::optional<std::reference_wrapper<const Tegra::FramebufferConfig>> framebuffer) = 0;

    /// Notify rasterizer that any caches of the specified region should be flushed to Switch memory
    virtual void FlushRegion(VAddr addr, u64 size) = 0;

    /// Notify rasterizer that any caches of the specified region should be invalidated
    virtual void InvalidateRegion(VAddr addr, u64 size) = 0;

    /// Notify rasterizer that any caches of the specified region should be flushed and invalidated
    virtual void FlushAndInvalidateRegion(VAddr addr, u64 size) = 0;

private:
    void ProcessBindMethod(const MethodCall& method_call);
    void ProcessSemaphoreTriggerMethod();
    void ProcessSemaphoreRelease();
    void ProcessSemaphoreAcquire();

    /// Calls a GPU puller method.
    void CallPullerMethod(const MethodCall& method_call);

    /// Calls a GPU engine method.
    void CallEngineMethod(const MethodCall& method_call);

    /// Determines where the method should be executed.
    bool ExecuteMethodOnEngine(const MethodCall& method_call);

protected:
    std::unique_ptr<Tegra::DmaPusher> dma_pusher;
    VideoCore::RendererBase& renderer;

private:
    std::unique_ptr<Tegra::MemoryManager> memory_manager;

    /// Mapping of command subchannels to their bound engine ids.
    std::array<EngineID, 8> bound_engines = {};

    /// 3D engine
    std::unique_ptr<Engines::Maxwell3D> maxwell_3d;
    /// 2D engine
    std::unique_ptr<Engines::Fermi2D> fermi_2d;
    /// Compute engine
    std::unique_ptr<Engines::KeplerCompute> kepler_compute;
    /// DMA engine
    std::unique_ptr<Engines::MaxwellDMA> maxwell_dma;
    /// Inline memory engine
    std::unique_ptr<Engines::KeplerMemory> kepler_memory;
};

#define ASSERT_REG_POSITION(field_name, position)                                                  \
    static_assert(offsetof(GPU::Regs, field_name) == position * 4,                                 \
                  "Field " #field_name " has invalid position")

ASSERT_REG_POSITION(smaphore_address, 0x4);
ASSERT_REG_POSITION(semaphore_sequence, 0x6);
ASSERT_REG_POSITION(semaphore_trigger, 0x7);
ASSERT_REG_POSITION(reference_count, 0x14);
ASSERT_REG_POSITION(semaphore_acquire, 0x1A);
ASSERT_REG_POSITION(semaphore_release, 0x1B);

ASSERT_REG_POSITION(acquire_mode, 0x100);
ASSERT_REG_POSITION(acquire_source, 0x101);
ASSERT_REG_POSITION(acquire_active, 0x102);
ASSERT_REG_POSITION(acquire_timeout, 0x103);
ASSERT_REG_POSITION(acquire_value, 0x104);

#undef ASSERT_REG_POSITION

} // namespace Tegra