Files
LookingGlass/idd/LGIdd/capture/CSoftwareFrameProcessor.cpp
2026-08-15 02:28:26 +10:00

363 lines
12 KiB
C++

/**
* Looking Glass
* Copyright © 2017-2026 The Looking Glass Authors
* https://looking-glass.io
*
* This program 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 2 of the License, or (at your option)
* any later version.
*
* This program 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 this program; if not, write to the Free Software Foundation, Inc., 59
* Temple Place, Suite 330, Boston, MA 02111-1307 USA
*/
#include "capture/CSoftwareFrameProcessor.h"
#include "capture/CFrameProcessorUtil.h"
#include "transport/IFrameTransport.h"
#include "CSRWLock.h"
#include "CDebug.h"
#include <utility>
CSoftwareFrameProcessor::CSoftwareFrameProcessor(
IFrameTransport * transport, std::shared_ptr<CD3D12Device> dx12,
CPostProcessor postProcessors[CAPTURE_PIPELINE_SLOTS],
CSRWLock * pipelineLock, HANDLE terminateEvent) :
CFrameProcessor(transport, std::move(dx12), postProcessors,
pipelineLock, terminateEvent)
{
}
void CSoftwareFrameProcessor::CompletionFunction(
CD3D12CommandSlot * slot, bool result, void * param1, void * param2)
{
auto processor = static_cast<CSoftwareFrameProcessor *>(param1);
auto context = static_cast<CopyContext *>(param2);
const FrameCopyBatch batch = context->batch;
CFrameBufferResource * fbRes =
batch.accepted == 1U && batch.prepared.count == 1 ?
batch.resources[0] : nullptr;
if (fbRes)
fbRes->MarkCompletion();
const bool succeeded = result && fbRes;
if (succeeded)
{
uint64_t stagedCopyTime = 0;
if (fbRes->GetMap())
{
const uint64_t stagedCopyStart = CFrameScheduler::Nanotime();
if (fbRes->IsFullCopy())
processor->m_transport->WriteFrameTarget(batch.prepared.token,
0, fbRes->GetMap(), 0, fbRes->GetFrameSize(), false);
else
{
const unsigned pitch = fbRes->GetCopyPitch();
const unsigned bytesPerPixel = fbRes->GetCopyBytesPerPixel();
const RECT * dirtyRects = fbRes->GetCopyDirtyRects();
const unsigned count = fbRes->GetCopyDirtyRectCount();
for (const RECT * rect = dirtyRects;
rect < dirtyRects + count; ++rect)
{
const size_t rowOffset =
(size_t)rect->top * pitch +
(size_t)rect->left * bytesPerPixel;
const size_t rowBytes =
(size_t)(rect->right - rect->left) * bytesPerPixel;
processor->m_transport->WriteFrameTargetRows(
batch.prepared.token, 0, fbRes->GetMap(), rowOffset,
rowBytes, pitch, (unsigned)(rect->bottom - rect->top));
}
}
stagedCopyTime = CFrameScheduler::Nanotime() - stagedCopyStart;
}
uint64_t gpuStart = 0;
uint64_t gpuEnd = 0;
const uint64_t copyReady = CFrameScheduler::Nanotime();
const bool gpuTimingValid = slot->GetGPUTimes(gpuStart, gpuEnd);
processor->m_transport->FinalizeFrameTarget(
batch.prepared.token, 0);
const uint64_t publishedAt = CFrameScheduler::Nanotime();
const uint64_t postProcessStart = fbRes->GetPostProcessStart();
const uint64_t copyStart = fbRes->GetCopyStart();
uint64_t postProcessTime = copyStart >= postProcessStart ?
copyStart - postProcessStart : 0;
uint64_t copyTime = copyReady >= copyStart ?
copyReady - copyStart : 0;
if (gpuTimingValid && gpuStart >= postProcessStart &&
gpuEnd >= gpuStart && gpuEnd <= copyReady)
{
postProcessTime = gpuStart - postProcessStart;
copyTime = gpuEnd - gpuStart + stagedCopyTime;
}
const uint64_t elapsed = publishedAt >= postProcessStart ?
publishedAt - postProcessStart : 0;
const uint64_t measured = postProcessTime + copyTime;
const uint64_t readyTime = elapsed > measured ?
elapsed - measured : 0;
processor->m_transport->SetFrameTargetTiming(batch.prepared.token, 0,
fbRes->GetCaptureTime(), postProcessTime, copyTime, readyTime, 0,
publishedAt);
processor->m_transport->TryRecordFrameTiming(
batch.prepared.token, 0, publishedAt - copyStart);
if (fbRes->GetTimingToken() && context->timingStart &&
copyReady >= context->timingStart)
processor->m_postProcessors[0].RecordTiming(
fbRes->GetTimingEffectIndex(), fbRes->GetTimingToken(),
fbRes->IsFullCopy(), copyReady - context->timingStart);
processor->m_transport->CompleteFrameTarget(
batch.prepared.token, 0, true);
}
else
{
if (batch.accepted)
processor->m_transport->FailFrameBatch(batch.prepared.token);
processor->SetFullDamage();
processor->m_transport->ForceFrame();
}
context->completionSucceeded.store(
succeeded, std::memory_order_relaxed);
context->completionDone.store(true, std::memory_order_release);
}
bool CSoftwareFrameProcessor::Submit(const FrameSubmission& submission)
{
if (submission.noImageUpdate && !HasPendingDamage())
return true;
CSRWSharedLock pipelineLock(*m_pipelineLock);
CPostProcessor& postProcessor = m_postProcessors[0];
const D12FrameFormat& dstFormat = postProcessor.GetOutputFormat();
const unsigned pitch = postProcessor.GetOutputPitch();
const size_t frameSize = postProcessor.GetOutputSize();
if (!pitch || !frameSize ||
frameSize > m_transport->GetMaxFrameSize())
{
DEBUG_ERROR("Software frame does not fit in primary frame memory");
SetFullDamage();
return false;
}
enum class BackpressureResult
{
RETRY,
COMPLETE,
FAILED,
};
const auto handleBackpressure = [this, &submission]()
{
if (!submission.noImageUpdate)
{
m_transport->FrameSuperseded();
return BackpressureResult::COMPLETE;
}
const DWORD result = WaitForSingleObject(m_terminateEvent, 1);
if (result == WAIT_TIMEOUT)
return BackpressureResult::RETRY;
if (result == WAIT_OBJECT_0)
return BackpressureResult::COMPLETE;
DEBUG_ERROR_HR(HRESULT_FROM_WIN32(GetLastError()),
"Failed while waiting to retry software frame publication");
return BackpressureResult::FAILED;
};
uint64_t contentSerial = 0;
for (;;)
{
FramePlan plan = {};
if (!m_transport->GetImmediatePrimaryFramePlan(
CFrameScheduler::Nanotime(), plan))
{
const BackpressureResult result = handleBackpressure();
if (result == BackpressureResult::RETRY)
continue;
return result == BackpressureResult::COMPLETE;
}
if (plan.count != 1 || !plan.targets[0].primary)
{
DEBUG_ERROR("Software capture received an invalid primary frame plan");
SetFullDamage();
return false;
}
CD3D12CommandSlot * copySlot = m_dx12->GetCopySlot();
if (!copySlot)
{
const BackpressureResult result = handleBackpressure();
if (result == BackpressureResult::RETRY)
continue;
return result == BackpressureResult::COMPLETE;
}
if (!contentSerial)
contentSerial = m_transport->NextContentSerial();
RECT currentDirtyRects[LG_MAX_DIRTY_RECTS] = {};
unsigned nbDirtyRects = 0;
const bool hasDamage =
TakePendingDamage(currentDirtyRects, &nbDirtyRects);
CFrameProcessorUtil::ClipDirtyRects(currentDirtyRects,
&nbDirtyRects, dstFormat.width, dstFormat.height);
RECT previousDirtyRects[LG_MAX_DIRTY_RECTS] = {};
unsigned nbPreviousDirtyRects = 0;
GetPreviousDamage(previousDirtyRects, &nbPreviousDirtyRects);
PreparedFrameBatch prepared = {};
if (!m_transport->PrepareFrameBatch(plan, contentSerial,
pitch, frameSize,
submission.sourceFormat, dstFormat,
currentDirtyRects, nbDirtyRects, submission.noImageUpdate,
prepared))
{
copySlot->Cancel();
RestorePendingDamage(currentDirtyRects, nbDirtyRects, hasDamage);
const BackpressureResult result = handleBackpressure();
if (result == BackpressureResult::RETRY)
continue;
return result == BackpressureResult::COMPLETE;
}
if (prepared.count != 1 || !prepared.targets[0].direct)
{
copySlot->Cancel();
m_transport->AbortFrameBatch(prepared.token);
RestorePendingDamage(currentDirtyRects, nbDirtyRects, hasDamage);
DEBUG_ERROR("Software capture received an invalid primary frame target");
SetFullDamage();
return false;
}
CFrameBufferResource * fbRes =
m_frameBuffers.Get(prepared.targets[0], frameSize);
if (!fbRes)
{
copySlot->Cancel();
m_transport->AbortFrameBatch(prepared.token);
RestorePendingDamage(currentDirtyRects, nbDirtyRects, hasDamage);
DEBUG_ERROR("Failed to get the software frame target resource");
SetFullDamage();
return false;
}
if (!submission.source->Signal() ||
!submission.source->Sync(*copySlot))
{
copySlot->Cancel();
m_transport->AbortFrameBatch(prepared.token);
RestorePendingDamage(currentDirtyRects, nbDirtyRects, hasDamage);
SetFullDamage();
return false;
}
const uint64_t timingStart = submission.timingToken ?
CFrameScheduler::Nanotime() : 0;
RECT copyDirtyRects[LG_MAX_DIRTY_RECTS * 2] = {};
unsigned nbCopyDirtyRects = 0;
const bool fullCopy = CFrameProcessorUtil::BuildCopyDamage(
postProcessor, prepared.targets[0].fullCopy,
previousDirtyRects, nbPreviousDirtyRects,
currentDirtyRects, nbDirtyRects,
dstFormat.width, dstFormat.height,
copyDirtyRects, &nbCopyDirtyRects);
const unsigned bytesPerPixel =
dstFormat.format == FRAME_TYPE_RGBA16F ? 8 : 4;
const uint64_t copyStart = CFrameScheduler::Nanotime();
fbRes->SetTiming(
submission.captureTime, submission.postProcessStart, copyStart);
fbRes->SetPostProcessSample(
submission.timingEffectIndex, submission.timingToken, fullCopy);
fbRes->SetCopyDamage(copyDirtyRects, nbCopyDirtyRects,
fullCopy, pitch, bytesPerPixel);
fbRes->ResetCompletion();
copySlot->BeginTiming();
postProcessor.CopyToFrameBuffer(copySlot->GetGfxList(),
fbRes->Get().Get(), submission.source->GetRes().Get(),
copyDirtyRects, nbCopyDirtyRects, fullCopy);
copySlot->EndTiming();
const uint32_t accepted =
m_transport->PublishFrameBatch(prepared.token);
if (accepted != 1U)
{
copySlot->Cancel();
if (accepted)
m_transport->FailFrameBatch(prepared.token);
RestorePendingDamage(currentDirtyRects, nbDirtyRects, hasDamage);
const BackpressureResult result = handleBackpressure();
if (result == BackpressureResult::RETRY)
continue;
return result == BackpressureResult::COMPLETE;
}
m_copy.batch = {};
m_copy.batch.prepared = prepared;
m_copy.batch.resources[0] = fbRes;
m_copy.batch.accepted = accepted;
m_copy.timingStart = timingStart;
m_copy.completionSucceeded.store(false, std::memory_order_relaxed);
m_copy.completionDone.store(false, std::memory_order_release);
copySlot->SetCompletionCallback(
&CompletionFunction, this, &m_copy);
const FrameBatchToken token = prepared.token;
const bool executed = copySlot->Execute();
const bool submittedWork =
!executed && copySlot->HasSubmittedWork();
const bool completionDone =
m_copy.completionDone.load(std::memory_order_acquire);
const bool completionSucceeded = completionDone &&
m_copy.completionSucceeded.load(std::memory_order_relaxed);
if (!executed)
{
if (completionSucceeded ||
(!completionDone && submittedWork))
{
m_transport->CommitFrameBatch(token);
CommitDamage(currentDirtyRects, nbDirtyRects);
RestorePendingDamage(
currentDirtyRects, nbDirtyRects, hasDamage);
}
else
{
if (!completionDone)
m_transport->FailFrameBatch(token);
RestorePendingDamage(
currentDirtyRects, nbDirtyRects, hasDamage);
if (!completionDone)
{
SetFullDamage();
m_transport->ForceFrame();
}
}
return false;
}
if (completionDone && !completionSucceeded)
{
RestorePendingDamage(currentDirtyRects, nbDirtyRects, hasDamage);
return false;
}
m_transport->CommitFrameBatch(token);
CommitDamage(currentDirtyRects, nbDirtyRects);
return true;
}
}