Index: src/gpu/gl/GrGLBuffer.cpp |
diff --git a/src/gpu/gl/GrGLBuffer.cpp b/src/gpu/gl/GrGLBuffer.cpp |
deleted file mode 100644 |
index ad56ceb729f47ece80db3969e295c847e6652e3d..0000000000000000000000000000000000000000 |
--- a/src/gpu/gl/GrGLBuffer.cpp |
+++ /dev/null |
@@ -1,336 +0,0 @@ |
-/* |
- * Copyright 2016 Google Inc. |
- * |
- * Use of this source code is governed by a BSD-style license that can be |
- * found in the LICENSE file. |
- */ |
- |
-#include "GrGLBuffer.h" |
-#include "GrGLGpu.h" |
-#include "SkTraceMemoryDump.h" |
- |
-#define GL_CALL(X) GR_GL_CALL(this->glGpu()->glInterface(), X) |
-#define GL_CALL_RET(RET, X) GR_GL_CALL_RET(this->glGpu()->glInterface(), RET, X) |
- |
-#if GR_GL_CHECK_ALLOC_WITH_GET_ERROR |
- #define CLEAR_ERROR_BEFORE_ALLOC(iface) GrGLClearErr(iface) |
- #define GL_ALLOC_CALL(iface, call) GR_GL_CALL_NOERRCHECK(iface, call) |
- #define CHECK_ALLOC_ERROR(iface) GR_GL_GET_ERROR(iface) |
-#else |
- #define CLEAR_ERROR_BEFORE_ALLOC(iface) |
- #define GL_ALLOC_CALL(iface, call) GR_GL_CALL(iface, call) |
- #define CHECK_ALLOC_ERROR(iface) GR_GL_NO_ERROR |
-#endif |
- |
-#ifdef SK_DEBUG |
-#define VALIDATE() this->validate() |
-#else |
-#define VALIDATE() do {} while(false) |
-#endif |
- |
-GrGLBuffer* GrGLBuffer::Create(GrGLGpu* gpu, GrBufferType type, size_t size, |
- GrAccessPattern accessPattern) { |
- static const int kIsVertexOrIndex = (1 << kVertex_GrBufferType) | (1 << kIndex_GrBufferType); |
- bool cpuBacked = gpu->glCaps().useNonVBOVertexAndIndexDynamicData() && |
- kDynamic_GrAccessPattern == accessPattern && |
- ((kIsVertexOrIndex >> type) & 1); |
- SkAutoTUnref<GrGLBuffer> buffer(new GrGLBuffer(gpu, type, size, accessPattern, cpuBacked)); |
- if (!cpuBacked && 0 == buffer->fBufferID) { |
- return nullptr; |
- } |
- return buffer.release(); |
-} |
- |
-// GL_STREAM_DRAW triggers an optimization in Chromium's GPU process where a client's vertex buffer |
-// objects are implemented as client-side-arrays on tile-deferred architectures. |
-#define DYNAMIC_DRAW_PARAM GR_GL_STREAM_DRAW |
- |
-inline static void get_target_and_usage(GrBufferType type, GrAccessPattern accessPattern, |
- const GrGLCaps& caps, GrGLenum* target, GrGLenum* usage) { |
- static const GrGLenum nonXferTargets[] = { |
- GR_GL_ARRAY_BUFFER, |
- GR_GL_ELEMENT_ARRAY_BUFFER |
- }; |
- GR_STATIC_ASSERT(0 == kVertex_GrBufferType); |
- GR_STATIC_ASSERT(1 == kIndex_GrBufferType); |
- |
- static const GrGLenum drawUsages[] = { |
- DYNAMIC_DRAW_PARAM, // TODO: Do we really want to use STREAM_DRAW here on non-Chromium? |
- GR_GL_STATIC_DRAW, |
- GR_GL_STREAM_DRAW |
- }; |
- static const GrGLenum readUsages[] = { |
- GR_GL_DYNAMIC_READ, |
- GR_GL_STATIC_READ, |
- GR_GL_STREAM_READ |
- }; |
- GR_STATIC_ASSERT(0 == kDynamic_GrAccessPattern); |
- GR_STATIC_ASSERT(1 == kStatic_GrAccessPattern); |
- GR_STATIC_ASSERT(2 == kStream_GrAccessPattern); |
- GR_STATIC_ASSERT(SK_ARRAY_COUNT(drawUsages) == 1 + kLast_GrAccessPattern); |
- GR_STATIC_ASSERT(SK_ARRAY_COUNT(readUsages) == 1 + kLast_GrAccessPattern); |
- |
- SkASSERT(accessPattern >= 0 && accessPattern <= kLast_GrAccessPattern); |
- |
- switch (type) { |
- case kVertex_GrBufferType: |
- case kIndex_GrBufferType: |
- *target = nonXferTargets[type]; |
- *usage = drawUsages[accessPattern]; |
- break; |
- case kXferCpuToGpu_GrBufferType: |
- if (GrGLCaps::kChromium_TransferBufferType == caps.transferBufferType()) { |
- *target = GR_GL_PIXEL_UNPACK_TRANSFER_BUFFER_CHROMIUM; |
- } else { |
- SkASSERT(GrGLCaps::kPBO_TransferBufferType == caps.transferBufferType()); |
- *target = GR_GL_PIXEL_UNPACK_BUFFER; |
- } |
- *usage = drawUsages[accessPattern]; |
- break; |
- case kXferGpuToCpu_GrBufferType: |
- if (GrGLCaps::kChromium_TransferBufferType == caps.transferBufferType()) { |
- *target = GR_GL_PIXEL_PACK_TRANSFER_BUFFER_CHROMIUM; |
- } else { |
- SkASSERT(GrGLCaps::kPBO_TransferBufferType == caps.transferBufferType()); |
- *target = GR_GL_PIXEL_PACK_BUFFER; |
- } |
- *usage = readUsages[accessPattern]; |
- break; |
- default: |
- SkFAIL("Unexpected buffer type."); |
- break; |
- } |
-} |
- |
-GrGLBuffer::GrGLBuffer(GrGLGpu* gpu, GrBufferType type, size_t size, GrAccessPattern accessPattern, |
- bool cpuBacked) |
- : INHERITED(gpu, type, size, accessPattern, cpuBacked), |
- fCPUData(nullptr), |
- fTarget(0), |
- fBufferID(0), |
- fSizeInBytes(size), |
- fUsage(0), |
- fGLSizeInBytes(0) { |
- if (cpuBacked) { |
- if (gpu->caps()->mustClearUploadedBufferData()) { |
- fCPUData = sk_calloc_throw(fSizeInBytes); |
- } else { |
- fCPUData = sk_malloc_flags(fSizeInBytes, SK_MALLOC_THROW); |
- } |
- } else { |
- GL_CALL(GenBuffers(1, &fBufferID)); |
- fSizeInBytes = size; |
- get_target_and_usage(type, accessPattern, gpu->glCaps(), &fTarget, &fUsage); |
- if (fBufferID) { |
- gpu->bindBuffer(fBufferID, fTarget); |
- CLEAR_ERROR_BEFORE_ALLOC(gpu->glInterface()); |
- // make sure driver can allocate memory for this buffer |
- GL_ALLOC_CALL(gpu->glInterface(), BufferData(fTarget, |
- (GrGLsizeiptr) fSizeInBytes, |
- nullptr, // data ptr |
- fUsage)); |
- if (CHECK_ALLOC_ERROR(gpu->glInterface()) != GR_GL_NO_ERROR) { |
- gpu->releaseBuffer(fBufferID, fTarget); |
- fBufferID = 0; |
- } else { |
- fGLSizeInBytes = fSizeInBytes; |
- } |
- } |
- } |
- VALIDATE(); |
- this->registerWithCache(); |
-} |
- |
-inline GrGLGpu* GrGLBuffer::glGpu() const { |
- SkASSERT(!this->wasDestroyed()); |
- return static_cast<GrGLGpu*>(this->getGpu()); |
-} |
- |
-inline const GrGLCaps& GrGLBuffer::glCaps() const { |
- return this->glGpu()->glCaps(); |
-} |
- |
-void GrGLBuffer::onRelease() { |
- if (!this->wasDestroyed()) { |
- VALIDATE(); |
- // make sure we've not been abandoned or already released |
- if (fCPUData) { |
- SkASSERT(!fBufferID); |
- sk_free(fCPUData); |
- fCPUData = nullptr; |
- } else if (fBufferID) { |
- this->glGpu()->releaseBuffer(fBufferID, fTarget); |
- fBufferID = 0; |
- fGLSizeInBytes = 0; |
- } |
- fMapPtr = nullptr; |
- VALIDATE(); |
- } |
- |
- INHERITED::onRelease(); |
-} |
- |
-void GrGLBuffer::onAbandon() { |
- fBufferID = 0; |
- fGLSizeInBytes = 0; |
- fMapPtr = nullptr; |
- sk_free(fCPUData); |
- fCPUData = nullptr; |
- VALIDATE(); |
- INHERITED::onAbandon(); |
-} |
- |
-void GrGLBuffer::onMap() { |
- if (this->wasDestroyed()) { |
- return; |
- } |
- |
- VALIDATE(); |
- SkASSERT(!this->isMapped()); |
- |
- if (0 == fBufferID) { |
- fMapPtr = fCPUData; |
- VALIDATE(); |
- return; |
- } |
- |
- bool readOnly = (kXferGpuToCpu_GrBufferType == this->type()); |
- |
- // Handling dirty context is done in the bindBuffer call |
- switch (this->glCaps().mapBufferType()) { |
- case GrGLCaps::kNone_MapBufferType: |
- break; |
- case GrGLCaps::kMapBuffer_MapBufferType: |
- this->glGpu()->bindBuffer(fBufferID, fTarget); |
- // Let driver know it can discard the old data |
- if (GR_GL_USE_BUFFER_DATA_NULL_HINT || fGLSizeInBytes != fSizeInBytes) { |
- GL_CALL(BufferData(fTarget, fSizeInBytes, nullptr, fUsage)); |
- } |
- GL_CALL_RET(fMapPtr, MapBuffer(fTarget, readOnly ? GR_GL_READ_ONLY : GR_GL_WRITE_ONLY)); |
- break; |
- case GrGLCaps::kMapBufferRange_MapBufferType: { |
- this->glGpu()->bindBuffer(fBufferID, fTarget); |
- // Make sure the GL buffer size agrees with fDesc before mapping. |
- if (fGLSizeInBytes != fSizeInBytes) { |
- GL_CALL(BufferData(fTarget, fSizeInBytes, nullptr, fUsage)); |
- } |
- GrGLbitfield writeAccess = GR_GL_MAP_WRITE_BIT; |
- // TODO: allow the client to specify invalidation in the transfer buffer case. |
- if (kXferCpuToGpu_GrBufferType != this->type()) { |
- writeAccess |= GR_GL_MAP_INVALIDATE_BUFFER_BIT; |
- } |
- GL_CALL_RET(fMapPtr, MapBufferRange(fTarget, 0, fSizeInBytes, |
- readOnly ? GR_GL_MAP_READ_BIT : writeAccess)); |
- break; |
- } |
- case GrGLCaps::kChromium_MapBufferType: |
- this->glGpu()->bindBuffer(fBufferID, fTarget); |
- // Make sure the GL buffer size agrees with fDesc before mapping. |
- if (fGLSizeInBytes != fSizeInBytes) { |
- GL_CALL(BufferData(fTarget, fSizeInBytes, nullptr, fUsage)); |
- } |
- GL_CALL_RET(fMapPtr, MapBufferSubData(fTarget, 0, fSizeInBytes, |
- readOnly ? GR_GL_READ_ONLY : GR_GL_WRITE_ONLY)); |
- break; |
- } |
- fGLSizeInBytes = fSizeInBytes; |
- VALIDATE(); |
-} |
- |
-void GrGLBuffer::onUnmap() { |
- if (this->wasDestroyed()) { |
- return; |
- } |
- |
- VALIDATE(); |
- SkASSERT(this->isMapped()); |
- if (0 == fBufferID) { |
- fMapPtr = nullptr; |
- return; |
- } |
- // bind buffer handles the dirty context |
- switch (this->glCaps().mapBufferType()) { |
- case GrGLCaps::kNone_MapBufferType: |
- SkDEBUGFAIL("Shouldn't get here."); |
- return; |
- case GrGLCaps::kMapBuffer_MapBufferType: // fall through |
- case GrGLCaps::kMapBufferRange_MapBufferType: |
- this->glGpu()->bindBuffer(fBufferID, fTarget); |
- GL_CALL(UnmapBuffer(fTarget)); |
- break; |
- case GrGLCaps::kChromium_MapBufferType: |
- this->glGpu()->bindBuffer(fBufferID, fTarget); |
- GL_CALL(UnmapBufferSubData(fMapPtr)); |
- break; |
- } |
- fMapPtr = nullptr; |
-} |
- |
-bool GrGLBuffer::onUpdateData(const void* src, size_t srcSizeInBytes) { |
- if (this->wasDestroyed()) { |
- return false; |
- } |
- |
- SkASSERT(!this->isMapped()); |
- SkASSERT(GR_GL_ARRAY_BUFFER == fTarget || GR_GL_ELEMENT_ARRAY_BUFFER == fTarget); |
- VALIDATE(); |
- if (srcSizeInBytes > fSizeInBytes) { |
- return false; |
- } |
- if (0 == fBufferID) { |
- memcpy(fCPUData, src, srcSizeInBytes); |
- return true; |
- } |
- SkASSERT(srcSizeInBytes <= fSizeInBytes); |
- // bindbuffer handles dirty context |
- this->glGpu()->bindBuffer(fBufferID, fTarget); |
- |
-#if GR_GL_USE_BUFFER_DATA_NULL_HINT |
- if (fSizeInBytes == srcSizeInBytes) { |
- GL_CALL(BufferData(fTarget, (GrGLsizeiptr) srcSizeInBytes, src, fUsage)); |
- } else { |
- // Before we call glBufferSubData we give the driver a hint using |
- // glBufferData with nullptr. This makes the old buffer contents |
- // inaccessible to future draws. The GPU may still be processing |
- // draws that reference the old contents. With this hint it can |
- // assign a different allocation for the new contents to avoid |
- // flushing the gpu past draws consuming the old contents. |
- // TODO I think we actually want to try calling bufferData here |
- GL_CALL(BufferData(fTarget, fSizeInBytes, nullptr, fUsage)); |
- GL_CALL(BufferSubData(fTarget, 0, (GrGLsizeiptr) srcSizeInBytes, src)); |
- } |
- fGLSizeInBytes = fSizeInBytes; |
-#else |
- // Note that we're cheating on the size here. Currently no methods |
- // allow a partial update that preserves contents of non-updated |
- // portions of the buffer (map() does a glBufferData(..size, nullptr..)) |
- GL_CALL(BufferData(fTarget, srcSizeInBytes, src, fUsage)); |
- fGLSizeInBytes = srcSizeInBytes; |
-#endif |
- VALIDATE(); |
- return true; |
-} |
- |
-void GrGLBuffer::setMemoryBacking(SkTraceMemoryDump* traceMemoryDump, |
- const SkString& dumpName) const { |
- SkString buffer_id; |
- buffer_id.appendU32(this->bufferID()); |
- traceMemoryDump->setMemoryBacking(dumpName.c_str(), "gl_buffer", |
- buffer_id.c_str()); |
-} |
- |
-#ifdef SK_DEBUG |
- |
-void GrGLBuffer::validate() const { |
- SkASSERT(GR_GL_ARRAY_BUFFER == fTarget || GR_GL_ELEMENT_ARRAY_BUFFER == fTarget || |
- GR_GL_PIXEL_PACK_BUFFER == fTarget || GR_GL_PIXEL_UNPACK_BUFFER == fTarget || |
- GR_GL_PIXEL_PACK_TRANSFER_BUFFER_CHROMIUM == fTarget || |
- GR_GL_PIXEL_UNPACK_TRANSFER_BUFFER_CHROMIUM == fTarget); |
- // The following assert isn't valid when the buffer has been abandoned: |
- // SkASSERT((0 == fDesc.fID) == (fCPUData)); |
- SkASSERT(0 != fBufferID || 0 == fGLSizeInBytes); |
- SkASSERT(nullptr == fMapPtr || fCPUData || fGLSizeInBytes <= fSizeInBytes); |
- SkASSERT(nullptr == fCPUData || nullptr == fMapPtr || fCPUData == fMapPtr); |
-} |
- |
-#endif |