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
|
// SPDX-FileCopyrightText: Copyright 2025 Eden Emulator Project
// SPDX-License-Identifier: GPL-3.0-or-later
// SPDX-FileCopyrightText: Copyright 2019 yuzu Emulator Project
// SPDX-License-Identifier: GPL-2.0-or-later
#include <cstring>
#include "common/algorithm.h"
#include "common/assert.h"
#include "video_core/engines/engine_upload.h"
#include "video_core/guest_memory.h"
#include "video_core/memory_manager.h"
#include "video_core/rasterizer_interface.h"
#include "video_core/textures/decoders.h"
namespace Tegra::Engines::Upload {
State::State(MemoryManager& memory_manager_, Registers& regs_)
: regs{regs_}, memory_manager{memory_manager_} {}
State::~State() = default;
void State::BindRasterizer(VideoCore::RasterizerInterface* rasterizer_) {
rasterizer = rasterizer_;
}
void State::ProcessExec(const bool is_linear_) {
write_offset = 0;
flushed_offset = 0;
copy_size = regs.line_length_in * regs.line_count;
inner_buffer.resize_destructive(copy_size);
is_linear = is_linear_;
}
void State::ProcessData(const u32 data, const bool is_last_call) {
const u32 sub_copy_size = (std::min)(4U, copy_size - write_offset);
std::memcpy(&inner_buffer[write_offset], &data, sub_copy_size);
write_offset += sub_copy_size;
if (!is_last_call) {
return;
}
if (native_partial && is_linear && regs.line_count == 1) {
const u32 count = write_offset - flushed_offset;
if (count) {
rasterizer->AccelerateInlineToMemory(regs.dest.Address() + flushed_offset, count,
std::span<const u8>{inner_buffer.data() + flushed_offset, count});
flushed_offset = write_offset;
}
return;
}
ProcessData(inner_buffer);
}
void State::ProcessData(const u32* data, size_t num_data) {
if (native_partial && is_linear && regs.line_count == 1) {
const auto count = std::min<size_t>(num_data * sizeof(u32), copy_size - write_offset);
if (count) {
rasterizer->AccelerateInlineToMemory(regs.dest.Address() + write_offset, count,
std::span<const u8>{reinterpret_cast<const u8*>(data), count});
write_offset += static_cast<u32>(count);
flushed_offset = write_offset;
}
return;
}
std::span<const u8> read_buffer(reinterpret_cast<const u8*>(data), num_data * sizeof(u32));
ProcessData(read_buffer);
}
void State::ProcessData(std::span<const u8> read_buffer) {
const GPUVAddr address{regs.dest.Address()};
if (is_linear) {
for (size_t line = 0; line < regs.line_count; ++line) {
const GPUVAddr dest_line = address + line * regs.dest.pitch;
std::span<const u8> buffer(read_buffer.data() + line * regs.line_length_in,
regs.line_length_in);
rasterizer->AccelerateInlineToMemory(dest_line, regs.line_length_in, buffer);
}
} else {
u32 width = regs.dest.width;
u32 x_elements = regs.line_length_in;
u32 x_offset = regs.dest.x;
const u32 bpp_shift = Common::FoldRight(
4U, [](u32 x, u32 y) { return (std::min)(x, static_cast<u32>(std::countr_zero(y))); },
width, x_elements, x_offset, static_cast<u32>(address));
width >>= bpp_shift;
x_elements >>= bpp_shift;
x_offset >>= bpp_shift;
const u32 bytes_per_pixel = 1U << bpp_shift;
const std::size_t dst_size = Tegra::Texture::CalculateSize(
true, bytes_per_pixel, width, regs.dest.height, regs.dest.depth,
regs.dest.BlockHeight(), regs.dest.BlockDepth());
Tegra::Memory::GpuGuestMemoryScoped<u8,
Tegra::Memory::GuestMemoryFlags::SafeReadCachedWrite>
tmp(memory_manager, address, dst_size, &tmp_buffer);
Tegra::Texture::SwizzleSubrect(tmp, read_buffer, bytes_per_pixel, width, regs.dest.height,
regs.dest.depth, x_offset, regs.dest.y, x_elements,
regs.line_count, regs.dest.BlockHeight(),
regs.dest.BlockDepth(), regs.line_length_in);
}
}
} // namespace Tegra::Engines::Upload
|
