Files
Xenia-Canary/src/xenia/gpu/texture_info.cc
chss95cs@gmail.com 8f7f7dc6ad fixed wine crash from use of NtSetEventPriorityBoost
add xe::clear_lowest_bit, use it in place of shift-andnot in some bit iteration code
make is_allocated_ and is_enabled_ volatile in xma_context
preallocate avpacket buffer in XMAContext::Setup, the reallocations of the buffer in ffmpeg were showing up on profiles
check is_enabled and is_allocated BEFORE locking an xmacontext. XMA worker was spending most of its time locking and unlocking contexts
Removed XeDMAC, dma:: namespace. It was a bad idea and I couldn't make it work in the end. Kept vastcpy and moved it to the memory namespace instead
Made the rest of global_critical_region's members static. They never needed an instance.
Removed ifdef'ed out code from ring_buffer.h
Added EventInfo struct to threading, added Event::Query to aid with implementing NtQueryEvent.
Removed vector from WaitMultiple, instead use a fixed array of 64 handles that we populate. WaitForMultipleObjects cannot handle more than 64 objects.
Remove XE_MSVC_OPTIMIZE_SMALL() use in x64_sequences, x64 backend is now always size optimized because of premake
Make global_critical_region_ static constexpr in shared_memory.h to get rid of wasteage of 8 bytes (empty class=1byte, +alignment for next member=8)
Move trace-related data to the tail of SharedMemory to keep more important data together
In IssueDraw build an array of fetch constant addresses/sizes, then pre-lock the global lock before doing requestrange for each instead of individually locking within requestrange for each of them
Consistent access specifier protected for pm4_command_processor_declare
Devirtualize WriteOneRegisterFromRing.
Move ExecutePacket and ExecutePrimaryBuffer to pm4_command_buffer_x
Remove many redundant header inclusions access xenia-gpu
Minor microoptimization of ExecutePacketType0

Add TextureCache::RequestTextures for batch invocation of LoadTexturesData

Add TextureCache::LoadTexturesData for reducing the number of times we release and reacquire the global lock.
Ideally you should hold the global lock for as little time as possible, but if you are constantly acquiring and releasing it you are actually more likely to have contention
Add already_locked param to ObjectTable::LookupObject to help with reducing lock acquire/release pairs
Add missing checks to XAudioRegisterRenderDriverClient_entry. this is unlikely to fix anything, it was just an easy thing to do
Add NtQueryEvent system call implementation. I don't actually know of any games that need it.
Instead of using std::vector + push_back in KeWaitForMultipleObjects and xeNtWaitForMultipleObjectsEx use a fixed size array of 64 and track the count. More than 64 objects is not permitted by the kernel. The repeated reallocations from push_back were appearing unusually high on the profiler, but were masked until now by waitformultipleobjects natural overhead
Pre-lock the global lock before looking up each handle for xeNtWaitForMultipleObjectsEx and KeWaitForMultipleObjects.
Pre-lock before looking up the signal and waiter in NtSignalAndWaitForSingleObjectEx
add missing checks to NtWaitForMultipleObjectsEx
Support pre-locking in XObject::GetNativeObject
2022-10-08 09:55:17 -07:00

228 lines
7.0 KiB
C++

/**
******************************************************************************
* Xenia : Xbox 360 Emulator Research Project *
******************************************************************************
* Copyright 2014 Ben Vanik. All rights reserved. *
* Released under the BSD license - see LICENSE in the root for more details. *
******************************************************************************
*/
#include "xenia/gpu/texture_info.h"
#include "xenia/base/logging.h"
#include "xenia/base/xxhash.h"
namespace xe {
namespace gpu {
using namespace xe::gpu::xenos;
bool TextureInfo::Prepare(const xe_gpu_texture_fetch_t& fetch,
TextureInfo* out_info) {
// https://msdn.microsoft.com/en-us/library/windows/desktop/cc308051(v=vs.85).aspx
// a2xx_sq_surfaceformat
std::memset(out_info, 0, sizeof(TextureInfo));
auto& info = *out_info;
info.format = fetch.format;
info.endianness = fetch.endianness;
info.dimension = fetch.dimension;
info.width = info.height = info.depth = 0;
info.is_stacked = false;
switch (info.dimension) {
case xenos::DataDimension::k1D:
// we treat 1D textures as 2D
info.dimension = DataDimension::k2DOrStacked;
info.width = fetch.size_1d.width;
assert_true(!fetch.stacked);
break;
case xenos::DataDimension::k2DOrStacked:
info.width = fetch.size_2d.width;
info.height = fetch.size_2d.height;
if (fetch.stacked) {
info.depth = fetch.size_2d.stack_depth;
info.is_stacked = true;
}
break;
case xenos::DataDimension::k3D:
info.width = fetch.size_3d.width;
info.height = fetch.size_3d.height;
info.depth = fetch.size_3d.depth;
assert_true(!fetch.stacked);
break;
case xenos::DataDimension::kCube:
info.width = fetch.size_2d.width;
info.height = fetch.size_2d.height;
assert_true(fetch.size_2d.stack_depth == 5);
info.depth = fetch.size_2d.stack_depth;
assert_true(!fetch.stacked);
break;
default:
assert_unhandled_case(info.dimension);
break;
}
info.pitch = fetch.pitch << 5;
info.mip_min_level = fetch.mip_min_level;
info.mip_max_level =
std::max(uint32_t(fetch.mip_min_level), uint32_t(fetch.mip_max_level));
info.is_tiled = fetch.tiled;
info.has_packed_mips = fetch.packed_mips;
info.extent = TextureExtent::Calculate(out_info, true);
info.SetupMemoryInfo(fetch.base_address << 12, fetch.mip_address << 12);
// We've gotten this far and mip_address is zero, assume no extra mips.
if (info.mip_max_level > 0 && !info.memory.mip_address) {
info.mip_max_level = 0;
}
assert_true(info.mip_min_level <= info.mip_max_level);
return true;
}
bool TextureInfo::PrepareResolve(uint32_t physical_address,
xenos::TextureFormat format,
xenos::Endian endian, uint32_t pitch,
uint32_t width, uint32_t height,
uint32_t depth, TextureInfo* out_info) {
assert_true(width > 0);
assert_true(height > 0);
std::memset(out_info, 0, sizeof(TextureInfo));
auto& info = *out_info;
info.format = format;
info.endianness = endian;
info.dimension = xenos::DataDimension::k2DOrStacked;
info.width = width - 1;
info.height = height - 1;
info.depth = depth - 1;
info.pitch = pitch;
info.mip_min_level = 0;
info.mip_max_level = 0;
info.is_tiled = true;
info.has_packed_mips = false;
info.extent = TextureExtent::Calculate(out_info, true);
info.SetupMemoryInfo(physical_address, 0);
return true;
}
uint32_t TextureInfo::GetMaxMipLevels() const {
return 1 + xe::log2_floor(std::max({width + 1, height + 1, depth + 1}));
}
const TextureExtent TextureInfo::GetMipExtent(uint32_t mip,
bool is_guest) const {
if (mip == 0) {
return extent;
}
uint32_t mip_width, mip_height;
if (is_guest) {
mip_width = xe::next_pow2(width + 1) >> mip;
mip_height = xe::next_pow2(height + 1) >> mip;
} else {
mip_width = std::max(1u, (width + 1) >> mip);
mip_height = std::max(1u, (height + 1) >> mip);
}
return TextureExtent::Calculate(format_info(), mip_width, mip_height,
depth + 1, is_tiled, is_guest);
}
void TextureInfo::GetMipSize(uint32_t mip, uint32_t* out_width,
uint32_t* out_height) const {
assert_not_null(out_width);
assert_not_null(out_height);
if (mip == 0) {
*out_width = width + 1;
*out_height = height + 1;
return;
}
uint32_t width_pow2 = xe::next_pow2(width + 1);
uint32_t height_pow2 = xe::next_pow2(height + 1);
*out_width = std::max(width_pow2 >> mip, 1u);
*out_height = std::max(height_pow2 >> mip, 1u);
}
uint64_t TextureInfo::hash() const {
return XXH3_64bits(this, sizeof(TextureInfo));
}
void TextureInfo::SetupMemoryInfo(uint32_t base_address, uint32_t mip_address) {
uint32_t bytes_per_block = format_info()->bytes_per_block();
memory.base_address = 0;
memory.base_size = 0;
memory.mip_address = 0;
memory.mip_size = 0;
if (mip_min_level == 0 && base_address) {
// There is a base mip level.
memory.base_address = base_address;
memory.base_size = GetMipExtent(0, true).visible_blocks() * bytes_per_block;
}
if (mip_min_level == 0 && mip_max_level == 0) {
// Sort circuit. Only one mip.
return;
}
if (mip_min_level == 0 && base_address == mip_address) {
// TODO(gibbed): This doesn't actually make any sense. Force only one mip.
// Offending title issues: #26, #45
return;
}
if (mip_min_level > 0) {
if ((base_address && !mip_address) || (base_address == mip_address)) {
// Mip data is actually at base address?
mip_address = base_address;
base_address = 0;
} else if (!base_address && mip_address) {
// Nothing needs to be done.
} else {
// WTF?
assert_always();
}
}
memory.mip_address = mip_address;
if (!has_packed_mips) {
for (uint32_t mip = std::max(1u, mip_min_level); mip < mip_max_level;
mip++) {
memory.mip_size += GetMipExtent(mip, true).all_blocks() * bytes_per_block;
}
memory.mip_size +=
GetMipExtent(mip_max_level, true).visible_blocks() * bytes_per_block;
return;
}
uint32_t width_pow2 = xe::next_pow2(width + 1);
uint32_t height_pow2 = xe::next_pow2(height + 1);
// Walk forward to find the address of the mip.
uint32_t packed_mip_base = std::max(1u, mip_min_level);
for (uint32_t mip = packed_mip_base; mip < mip_max_level;
mip++, packed_mip_base++) {
uint32_t mip_width = std::max(width_pow2 >> mip, 1u);
uint32_t mip_height = std::max(height_pow2 >> mip, 1u);
if (std::min(mip_width, mip_height) <= 16) {
// We've reached the point where the mips are packed into a single tile.
break;
}
memory.mip_size += GetMipExtent(mip, true).all_blocks() * bytes_per_block;
}
}
} // namespace gpu
} // namespace xe