wined3d: Render target locking fixes.

This commit is contained in:
Stefan Dösinger 2006-06-04 16:42:57 +02:00 committed by Alexandre Julliard
parent 473ad39dcb
commit 2c3adec595
3 changed files with 172 additions and 153 deletions

View File

@ -2069,7 +2069,8 @@ void drawPrimitive(IWineD3DDevice *iface,
BOOL usePixelShaderFunction = FALSE;
BOOL isLightingOn = FALSE;
WineDirect3DVertexStridedData *dataLocations;
int useHW = FALSE;
IWineD3DSwapChainImpl *swapchain;
int useHW = FALSE, i;
if (This->stateBlock->vertexShader != NULL && wined3d_settings.vs_mode != VS_NONE
&&((IWineD3DVertexShaderImpl *)This->stateBlock->vertexShader)->baseShader.function != NULL
@ -2093,6 +2094,15 @@ void drawPrimitive(IWineD3DDevice *iface,
TRACE("(%p) : using vertex declaration %p\n", iface, This->stateBlock->vertexDecl);
}
/* Invalidate the back buffer memory so LockRect will read it the next time */
for(i = 0; i < IWineD3DDevice_GetNumberOfSwapChains(iface); i++) {
IWineD3DDevice_GetSwapChain(iface, i, (IWineD3DSwapChain **) &swapchain);
if(swapchain) {
if(swapchain->backBuffer) ((IWineD3DSurfaceImpl *) swapchain->backBuffer)->Flags |= SFLAG_GLDIRTY;
IWineD3DSwapChain_Release( (IWineD3DSwapChain *) swapchain);
}
}
/* Ok, we will be updating the screen from here onwards so grab the lock */
ENTER_GL();

View File

@ -480,170 +480,173 @@ HRESULT WINAPI IWineD3DSurfaceImpl_LockRect(IWineD3DSurface *iface, WINED3DLOCKE
}
}
} else if (WINED3DUSAGE_RENDERTARGET & This->resource.usage && !(Flags&WINED3DLOCK_DISCARD)) { /* render surfaces */
GLint prev_store;
GLint prev_read;
BOOL notInContext = FALSE;
IWineD3DSwapChainImpl *targetSwapChain = NULL;
} else if (WINED3DUSAGE_RENDERTARGET & This->resource.usage ){ /* render surfaces */
if((!(Flags&WINED3DLOCK_DISCARD) && (This->Flags & SFLAG_GLDIRTY)) || !This->resource.allocatedMemory) {
GLint prev_store;
GLint prev_read;
BOOL notInContext = FALSE;
IWineD3DSwapChainImpl *targetSwapChain = NULL;
ENTER_GL();
ENTER_GL();
/**
* for render->surface copy begin to begin of allocatedMemory
* unlock can be more easy
*/
/**
* for render->surface copy begin to begin of allocatedMemory
* unlock can be more easy
*/
TRACE("locking a render target\n");
TRACE("locking a render target\n");
if (This->resource.allocatedMemory == NULL)
This->resource.allocatedMemory = HeapAlloc(GetProcessHeap() ,0 ,This->resource.size);
if (This->resource.allocatedMemory == NULL)
This->resource.allocatedMemory = HeapAlloc(GetProcessHeap() ,0 ,This->resource.size);
This->Flags |= SFLAG_ACTIVELOCK; /*When this flag is set to true, loading the surface again won't free THis->resource.allocatedMemory*/
pLockedRect->pBits = This->resource.allocatedMemory;
This->Flags |= SFLAG_ACTIVELOCK; /*When this flag is set to true, loading the surface again won't free THis->resource.allocatedMemory*/
pLockedRect->pBits = This->resource.allocatedMemory;
glFlush();
vcheckGLcall("glFlush");
glGetIntegerv(GL_READ_BUFFER, &prev_read);
vcheckGLcall("glIntegerv");
glGetIntegerv(GL_PACK_SWAP_BYTES, &prev_store);
vcheckGLcall("glIntegerv");
glFlush();
vcheckGLcall("glFlush");
glGetIntegerv(GL_READ_BUFFER, &prev_read);
vcheckGLcall("glIntegerv");
glGetIntegerv(GL_PACK_SWAP_BYTES, &prev_store);
vcheckGLcall("glIntegerv");
/* Here's what we have to do:
See if the swapchain has the same context as the renderTarget or the surface is the render target.
Otherwise, see if were sharing a context with the implicit swapchain (because we're using a shared context model!)
and use the front back buffer as required.
if not, we need to switch contexts and then switchback at the end.
*/
IWineD3DSurface_GetContainer(iface, &IID_IWineD3DSwapChain, (void **)&swapchain);
IWineD3DSurface_GetContainer(myDevice->renderTarget, &IID_IWineD3DSwapChain, (void **)&targetSwapChain);
/* Here's what we have to do:
See if the swapchain has the same context as the renderTarget or the surface is the render target.
Otherwise, see if were sharing a context with the implicit swapchain (because we're using a shared context model!)
and use the front back buffer as required.
if not, we need to switch contexts and then switchback at the end.
*/
IWineD3DSurface_GetContainer(iface, &IID_IWineD3DSwapChain, (void **)&swapchain);
IWineD3DSurface_GetContainer(myDevice->renderTarget, &IID_IWineD3DSwapChain, (void **)&targetSwapChain);
/* NOTE: In a shared context environment the renderTarget will use the same context as the implicit swapchain (we're not in a shared environment yet! */
if ((swapchain == targetSwapChain && targetSwapChain != NULL) || iface == myDevice->renderTarget) {
if (iface == myDevice->renderTarget || iface == swapchain->backBuffer) {
TRACE("locking back buffer\n");
glReadBuffer(GL_BACK);
} else if (iface == swapchain->frontBuffer) {
TRACE("locking front\n");
glReadBuffer(GL_FRONT);
} else if (iface == myDevice->depthStencilBuffer) {
FIXME("Stencil Buffer lock unsupported for now\n");
} else {
FIXME("(%p) Shouldn't have got here!\n", This);
glReadBuffer(GL_BACK);
}
} else if (swapchain != NULL) {
IWineD3DSwapChainImpl *implSwapChain;
IWineD3DDevice_GetSwapChain((IWineD3DDevice *)myDevice, 0, (IWineD3DSwapChain **)&implSwapChain);
if (swapchain->glCtx == implSwapChain->render_ctx && swapchain->drawable == implSwapChain->win) {
/* This will fail for the implicit swapchain, which is why there needs to be a context manager */
if (iface == swapchain->backBuffer) {
glReadBuffer(GL_BACK);
/* NOTE: In a shared context environment the renderTarget will use the same context as the implicit swapchain (we're not in a shared environment yet! */
if ((swapchain == targetSwapChain && targetSwapChain != NULL) || iface == myDevice->renderTarget) {
if (iface == myDevice->renderTarget || iface == swapchain->backBuffer) {
TRACE("locking back buffer\n");
glReadBuffer(GL_BACK);
} else if (iface == swapchain->frontBuffer) {
glReadBuffer(GL_FRONT);
TRACE("locking front\n");
glReadBuffer(GL_FRONT);
} else if (iface == myDevice->depthStencilBuffer) {
FIXME("Stencil Buffer lock unsupported for now\n");
} else {
FIXME("Should have got here!\n");
glReadBuffer(GL_BACK);
FIXME("(%p) Shouldn't have got here!\n", This);
glReadBuffer(GL_BACK);
}
} else {
/* We need to switch contexts to be able to read the buffer!!! */
FIXME("The buffer requested isn't in the current openGL context\n");
notInContext = TRUE;
/* TODO: check the contexts, to see if were shared with the current context */
}
IWineD3DSwapChain_Release((IWineD3DSwapChain *)implSwapChain);
}
if (swapchain != NULL) IWineD3DSwapChain_Release((IWineD3DSwapChain *)swapchain);
if (targetSwapChain != NULL) IWineD3DSwapChain_Release((IWineD3DSwapChain *)targetSwapChain);
/** the depth stencil in openGL has a format of GL_FLOAT
* which should be good for WINED3DFMT_D16_LOCKABLE
* and WINED3DFMT_D16
* it is unclear what format the stencil buffer is in except.
* 'Each index is converted to fixed point...
* If GL_MAP_STENCIL is GL_TRUE, indices are replaced by their
* mappings in the table GL_PIXEL_MAP_S_TO_S.
* glReadPixels(This->lockedRect.left,
* This->lockedRect.bottom - j - 1,
* This->lockedRect.right - This->lockedRect.left,
* 1,
* GL_DEPTH_COMPONENT,
* type,
* (char *)pLockedRect->pBits + (pLockedRect->Pitch * (j-This->lockedRect.top)));
*****************************************/
if (!notInContext) { /* Only read the buffer if it's in the current context */
long j;
#if 0
/* Bizarly it takes 120 millseconds to get an 800x600 region a line at a time, but only 10 to get the whole lot every time,
* This is on an ATI9600, and may be format dependent, anyhow this hack makes this demo dx9_2d_demo_game
* run ten times faster!
* ************************************/
BOOL ati_performance_hack = FALSE;
ati_performance_hack = (This->lockedRect.bottom - This->lockedRect.top > 10) || (This->lockedRect.right - This->lockedRect.left > 10)? TRUE: FALSE;
#endif
if ((This->lockedRect.left == 0 && This->lockedRect.top == 0 &&
This->lockedRect.right == This->currentDesc.Width
&& This->lockedRect.bottom == This->currentDesc.Height)) {
BYTE *row, *top, *bottom;
int i;
glReadPixels(0, 0,
This->currentDesc.Width,
This->currentDesc.Height,
This->glDescription.glFormat,
This->glDescription.glType,
(char *)pLockedRect->pBits);
/* glReadPixels returns the image upside down, and there is no way to prevent this.
Flip the lines in software*/
row = HeapAlloc(GetProcessHeap(), 0, pLockedRect->Pitch);
if(!row) {
ERR("Out of memory\n");
return E_OUTOFMEMORY;
}
top = This->resource.allocatedMemory;
bottom = ( (BYTE *) This->resource.allocatedMemory) + pLockedRect->Pitch * ( This->currentDesc.Height - 1);
for(i = 0; i < This->currentDesc.Height / 2; i++) {
memcpy(row, top, pLockedRect->Pitch);
memcpy(top, bottom, pLockedRect->Pitch);
memcpy(bottom, row, pLockedRect->Pitch);
top += pLockedRect->Pitch;
bottom -= pLockedRect->Pitch;
}
HeapFree(GetProcessHeap(), 0, row);
} else if (This->lockedRect.left == 0 && This->lockedRect.right == This->currentDesc.Width) {
glReadPixels(0,
This->lockedRect.top,
This->currentDesc.Width,
This->currentDesc.Height,
This->glDescription.glFormat,
This->glDescription.glType,
(char *)pLockedRect->pBits);
} else{
for (j = This->lockedRect.top; j < This->lockedRect.bottom - This->lockedRect.top; ++j) {
glReadPixels(This->lockedRect.left,
This->lockedRect.bottom - j - 1,
This->lockedRect.right - This->lockedRect.left,
1,
This->glDescription.glFormat,
This->glDescription.glType,
(char *)pLockedRect->pBits + (pLockedRect->Pitch * (j-This->lockedRect.top)));
} else if (swapchain != NULL) {
IWineD3DSwapChainImpl *implSwapChain;
IWineD3DDevice_GetSwapChain((IWineD3DDevice *)myDevice, 0, (IWineD3DSwapChain **)&implSwapChain);
if (swapchain->glCtx == implSwapChain->render_ctx && swapchain->drawable == implSwapChain->win) {
/* This will fail for the implicit swapchain, which is why there needs to be a context manager */
if (iface == swapchain->backBuffer) {
glReadBuffer(GL_BACK);
} else if (iface == swapchain->frontBuffer) {
glReadBuffer(GL_FRONT);
} else if (iface == myDevice->depthStencilBuffer) {
FIXME("Stencil Buffer lock unsupported for now\n");
} else {
FIXME("Should have got here!\n");
glReadBuffer(GL_BACK);
}
} else {
/* We need to switch contexts to be able to read the buffer!!! */
FIXME("The buffer requested isn't in the current openGL context\n");
notInContext = TRUE;
/* TODO: check the contexts, to see if were shared with the current context */
}
IWineD3DSwapChain_Release((IWineD3DSwapChain *)implSwapChain);
}
vcheckGLcall("glReadPixels");
TRACE("Resetting buffer\n");
glReadBuffer(prev_read);
vcheckGLcall("glReadBuffer");
}
LEAVE_GL();
if (swapchain != NULL) IWineD3DSwapChain_Release((IWineD3DSwapChain *)swapchain);
if (targetSwapChain != NULL) IWineD3DSwapChain_Release((IWineD3DSwapChain *)targetSwapChain);
/** the depth stencil in openGL has a format of GL_FLOAT
* which should be good for WINED3DFMT_D16_LOCKABLE
* and WINED3DFMT_D16
* it is unclear what format the stencil buffer is in except.
* 'Each index is converted to fixed point...
* If GL_MAP_STENCIL is GL_TRUE, indices are replaced by their
* mappings in the table GL_PIXEL_MAP_S_TO_S.
* glReadPixels(This->lockedRect.left,
* This->lockedRect.bottom - j - 1,
* This->lockedRect.right - This->lockedRect.left,
* 1,
* GL_DEPTH_COMPONENT,
* type,
* (char *)pLockedRect->pBits + (pLockedRect->Pitch * (j-This->lockedRect.top)));
*****************************************/
if (!notInContext) { /* Only read the buffer if it's in the current context */
long j;
#if 0
/* Bizarly it takes 120 millseconds to get an 800x600 region a line at a time, but only 10 to get the whole lot every time,
* This is on an ATI9600, and may be format dependent, anyhow this hack makes this demo dx9_2d_demo_game
* run ten times faster!
* ************************************/
BOOL ati_performance_hack = FALSE;
ati_performance_hack = (This->lockedRect.bottom - This->lockedRect.top > 10) || (This->lockedRect.right - This->lockedRect.left > 10)? TRUE: FALSE;
#endif
if ((This->lockedRect.left == 0 && This->lockedRect.top == 0 &&
This->lockedRect.right == This->currentDesc.Width
&& This->lockedRect.bottom == This->currentDesc.Height)) {
BYTE *row, *top, *bottom;
int i;
glReadPixels(0, 0,
This->currentDesc.Width,
This->currentDesc.Height,
This->glDescription.glFormat,
This->glDescription.glType,
(char *)pLockedRect->pBits);
/* glReadPixels returns the image upside down, and there is no way to prevent this.
Flip the lines in software*/
row = HeapAlloc(GetProcessHeap(), 0, pLockedRect->Pitch);
if(!row) {
ERR("Out of memory\n");
return E_OUTOFMEMORY;
}
top = This->resource.allocatedMemory;
bottom = ( (BYTE *) This->resource.allocatedMemory) + pLockedRect->Pitch * ( This->currentDesc.Height - 1);
for(i = 0; i < This->currentDesc.Height / 2; i++) {
memcpy(row, top, pLockedRect->Pitch);
memcpy(top, bottom, pLockedRect->Pitch);
memcpy(bottom, row, pLockedRect->Pitch);
top += pLockedRect->Pitch;
bottom -= pLockedRect->Pitch;
}
HeapFree(GetProcessHeap(), 0, row);
This->Flags &= ~SFLAG_GLDIRTY;
} else if (This->lockedRect.left == 0 && This->lockedRect.right == This->currentDesc.Width) {
glReadPixels(0,
This->lockedRect.top,
This->currentDesc.Width,
This->currentDesc.Height,
This->glDescription.glFormat,
This->glDescription.glType,
(char *)pLockedRect->pBits);
} else{
for (j = This->lockedRect.top; j < This->lockedRect.bottom - This->lockedRect.top; ++j) {
glReadPixels(This->lockedRect.left,
This->lockedRect.bottom - j - 1,
This->lockedRect.right - This->lockedRect.left,
1,
This->glDescription.glFormat,
This->glDescription.glType,
(char *)pLockedRect->pBits + (pLockedRect->Pitch * (j-This->lockedRect.top)));
}
}
vcheckGLcall("glReadPixels");
TRACE("Resetting buffer\n");
glReadBuffer(prev_read);
vcheckGLcall("glReadBuffer");
}
LEAVE_GL();
}
} else if (WINED3DUSAGE_DEPTHSTENCIL & This->resource.usage) { /* stencil surfaces */
if (!messages & 1) {
@ -784,14 +787,14 @@ HRESULT WINAPI IWineD3DSurfaceImpl_UnlockRect(IWineD3DSurface *iface) {
per drawprim (and leave set - it will sort itself out due to last_was_rhw */
d3ddevice_set_ortho(This->resource.wineD3DDevice);
if (iface == implSwapChain->backBuffer || iface == myDevice->renderTarget) {
glDrawBuffer(GL_BACK);
} else if (iface == implSwapChain->frontBuffer) {
if (iface == implSwapChain->frontBuffer) {
glDrawBuffer(GL_FRONT);
checkGLcall("glDrawBuffer GL_FRONT");
} else if (iface == implSwapChain->backBuffer || iface == myDevice->renderTarget) {
glDrawBuffer(GL_BACK);
checkGLcall("glDrawBuffer GL_BACK");
}
vcheckGLcall("glDrawBuffer");
/* If not fullscreen, we need to skip a number of bytes to find the next row of data */
glGetIntegerv(GL_UNPACK_ROW_LENGTH, &skipBytes);
glPixelStorei(GL_UNPACK_ROW_LENGTH, This->currentDesc.Width);
@ -2244,6 +2247,9 @@ HRESULT WINAPI IWineD3DSurfaceImpl_BltOverride(IWineD3DSurfaceImpl *This, RECT *
LEAVE_GL();
/* TODO: If the surface is locked often, perform the Blt in software on the memory instead */
This->Flags |= SFLAG_GLDIRTY;
return WINED3D_OK;
}

View File

@ -329,6 +329,9 @@ HRESULT WINAPI IWineD3DSwapChainImpl_Present(IWineD3DSwapChain *iface, CONST REC
IWineD3DDevice_Clear((IWineD3DDevice*)This->wineD3DDevice, 0, NULL, D3DCLEAR_STENCIL|D3DCLEAR_ZBUFFER, 0x00, 1.0, 0);
}
((IWineD3DSurfaceImpl *) This->frontBuffer)->Flags |= SFLAG_GLDIRTY;
((IWineD3DSurfaceImpl *) This->backBuffer)->Flags |= SFLAG_GLDIRTY;
TRACE("returning\n");
return WINED3D_OK;
}