In order to distinguish between resources that exist purely in CPU memory and
resources that exist in GPU memory but can be mapped.
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
wined3d_load_buffer() might invalidate the current state.
Signed-off-by: Józef Kucia <jkucia@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
This fixes a performance issue exposed by commit
afb2aa2512.
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
In order to avoid invalidating system memory for "converted" surfaces.
Fixes a regression introduced by commit 56128d18d4 ("wined3d: Drop the
special case for "converted" surfaces in wined3d_surface_blt().").
Signed-off-by: Józef Kucia <jkucia@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
We can return WINED3DOK_NOAUTOGEN only when all other resource usages
are supported.
Fixes 25a8773d21 ("wined3d: Advertise AUTOGENMIPMAP only for
renderable and filterable formats."). The problem was present before
25a8773d21, but the mentioned commit made it much worse. Previously,
it was an issue only for OpenGL implementations which do not support
the GL_SGIS_generate_mipmap extension.
Signed-off-by: Józef Kucia <jkucia@codeweavers.com>
Signed-off-by: Matteo Bruni <mbruni@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Eliminates possible GL errors when core contexts are used.
Signed-off-by: Józef Kucia <jkucia@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Effectively, this will enable Direct3D 11 by default on AMD and Intel
hardware.
Yes, this patch is awful.
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
The context_acquire() function does that.
Signed-off-by: Józef Kucia <jkucia@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
For consistency with wined3d_device_init_3d().
Signed-off-by: Józef Kucia <jkucia@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Relative addressing doesn't necessarily work for SM3.
Signed-off-by: Józef Kucia <jkucia@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
We do not want to automagically disable depth test for D3D10+.
Signed-off-by: Józef Kucia <jkucia@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Direct3D seems to define exact depth bias scale factors per format.
In order to make depth bias work reliably across OpenGL drivers we need
to slightly adjust depth bias values passed to glPolygonOffset().
Signed-off-by: Józef Kucia <jkucia@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
A plane isn't transformed as simply as a vector.
Signed-off-by: Józef Kucia <jkucia@codeweavers.com>
Signed-off-by: Matteo Bruni <mbruni@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
In D3D10+ all clip distances declared in a shader are always enabled.
In other words, enablement of clip distances is a property of a shader.
Signed-off-by: Józef Kucia <jkucia@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Many thanks to Matias N. Goldberg and Philip Rebohle for investigating and
solving the bug!
Signed-off-by: Fabian Maurer <dark.shadow4@web.de>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
GL_CLIP_PLANE0 and GL_CLIP_DISTANCE0 has the same value.
Signed-off-by: Józef Kucia <jkucia@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
We never supported, and we should never need to support more than 8 user
clip planes in wined3d.
OpenGL drivers do not expose more than 8 clip distances. Similarly,
other APIs, D3D10, D3D11 and D3D12 provide exactly 8 clip distances.
Signed-off-by: Józef Kucia <jkucia@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
World of Warplanes calls SetWindowLong(window, GWL_STYLE, WS_POPUP)
before calling reset, effectively hiding the window. It does not call
SetWindowPos(SWP_FRAMECHANGED), so the effect isn't visible and the
game displays fine after the reset. However, after the next mouse
click WINPOS_WindowFromPoint skips the window because it doesn't have
WS_VISIBLE and returns the desktop window. This in turn triggers a
focus loss and d3d9 minimizes the game.
Signed-off-by: Stefan Dösinger <stefan@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Geometry shader outputs can be freely split when there is no pixel shader.
Signed-off-by: Józef Kucia <jkucia@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Allows us to split shader output variables when rasterization
is disabled.
Signed-off-by: Józef Kucia <jkucia@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Based on a patch by Michael Müller.
Signed-off-by: Józef Kucia <jkucia@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
In Direct3D a different format is used for sampling depth textures, but
this is not the case in OpenGL. We still need to create a texture view
when a texture type is different than a view type, or when a shader
resource view is being created for a subset of sub-resources.
Signed-off-by: Józef Kucia <jkucia@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
In order to disable feature level 11_0 when indirect draws
are not supported.
Signed-off-by: Józef Kucia <jkucia@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
We do not support index buffer offsets for indirect draws.
Signed-off-by: Józef Kucia <jkucia@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
We have a more efficient codepath for SM4+ interpolation modifiers
if GLSL 4.40 is available.
Signed-off-by: Józef Kucia <jkucia@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
math.h is included in wined3d_private.h
Signed-off-by: Józef Kucia <jkucia@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
This should at least avoid trashing other stream output elements.
Signed-off-by: Józef Kucia <jkucia@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
In order to add interpolation qualifiers per shader input.
Hull shader outputs are not unrolled because they can be accessed
through the indexable vocp register.
Signed-off-by: Józef Kucia <jkucia@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
In order to generate correct GLSL for image stores.
Signed-off-by: Józef Kucia <jkucia@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
This way we avoid queueing up CS packets on each
wined3d_query_get_data() call, which has been seen in some degenerate
cases to starve other client threads that are trying to submit actual
work into the CS (sort-of a livelock).
Signed-off-by: Matteo Bruni <mbruni@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Without an explicit volatile, aarch64 gcc will copy queue->tail into a
register outside the loop and then continuously compare two registers
against each other. The *(volatile LONG *)& forces gcc to re-read the
memory every iteration. Therefore, queue_is_empty as it is written will
only work from the CS thread.
Signed-off-by: Stefan Dösinger <stefan@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
ARB_pipeline_statistics_query is included in OpenGL 4.6.
Signed-off-by: Józef Kucia <jkucia@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Makes sure that we do not overwrite source parameters for consecutive
GLSL function calls when the instruction is translated into multiple
lines.
Signed-off-by: Józef Kucia <jkucia@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
We still want to avoid invalidating the system memory copy for these surfaces,
but we've become a lot better at preserving the current location for
resources, to the point that the special case actually hurts. In particular,
if the resource wasn't already current on the CPU, using the CPU blitter would
cause us to read it back, which is what we're trying to avoid.
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Since that would imply reading them back, with associated format conversion.
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
The issue isn't so much stencil clears as clearing either depth or stencil on
formats that contain both.
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Use fences for the APPLESYNC codepath, and implement event queries
on top of fences.
Signed-off-by: Józef Kucia <jkucia@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Fixes a regression introduced by commit
be20ddc38b.
Signed-off-by: Józef Kucia <jkucia@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Fixes a regression introduced in ee17d7ba1b.
Signed-off-by: Sebastian Lackner <sebastian@fds-team.de>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Fixes a regression introduced in cad4badbcf.
Signed-off-by: Sebastian Lackner <sebastian@fds-team.de>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Fixes a regression introduced in 62ca4f3826.
Signed-off-by: Sebastian Lackner <sebastian@fds-team.de>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
This is a workaround for a bug in Mesa GLSL linker. The bug is triggered
by GLSL generated for SM3 shaders when shader inputs are not read and
interface blocks are used for shader inputs/outputs (core profile code
path).
Declarations for unused shader inputs are not generated for pixel
shaders. We still declare shader outputs in the previous shader stage,
but it's allowed by the GLSL 1.50 spec:
"Only output variables that are read by the subsequent pipeline stage
need to be written; it is allowed to have superfluous declarations of
output variables."
The table in section 4.3.4 (Input Variables) of the GLSL 4.20 spec even
more explicitly allows shader output variables with no matching input
variables declared in the consuming shader.
Signed-off-by: Józef Kucia <jkucia@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
The existing checks should be sufficient, but the GL_VERSION one is
potentially more reliable.
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
For D3D10+.
When WINED3D_RTYPE_NONE is passed the wined3d_check_device_format()
function checks if a format is supported for all reasonable resource
types for a given usage. For now render target and depth stencil support
can be queried using this method.
Signed-off-by: Józef Kucia <jkucia@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
The WINED3DFMT_FLAG_RENDERTARGET flag is cleared in
init_format_fbo_compat_info().
Signed-off-by: Józef Kucia <jkucia@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Synchronization is only needed when output variables are read.
Signed-off-by: Józef Kucia <jkucia@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Only increase it if the sub-resource wasn't already exclusively in system
memory.
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
A format that contains 3 floating point values with a 9-bit mantissa
each and a shared 5-bit exponent.
Maps directly onto the RGB9_E5_EXT format available through the
GL extension EXT_texture_shared_exponent.
Signed-off-by: Sven Hesse <drmccoy@drmccoy.de>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Some applications reset state by setting all render states to 0. That's
questionable on behalf of the application, but not an issue as far as wined3d
is concerned.
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
This should never happen. Note that the previous code would return a GPU with
64 TiB of VRAM.
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
A combination of gather4_c and gather4_po.
Signed-off-by: Sven Hesse <drmccoy@drmccoy.de>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
A context associated with a query may be destroyed together
with a swapchain.
Signed-off-by: Józef Kucia <jkucia@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Apparently, dcl_input_* instructions do not use the same values as
sysval semantics in shader signatures.
Signed-off-by: Józef Kucia <jkucia@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
This fixes a regression introduced by commit
dbd8c89ea8. When using ORM_BACKBUFFER,
off-screen surfaces may be loaded into WINED3D_LOCATION_DRAWABLE. They
shouldn't be flipped in that case.
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
When using ORM_BACKBUFFER, formats with fixups may need to be loaded into
WINED3D_LOCATION_DRAWABLE. Before commit
cad4badbcf, surface_load_drawable() would always
use the device blitter, without checking whether the blitter actually supports
the blit. The FBO blitter would be able to handle these blits, but is
unavailable when using ORM_BACKBUFFER.
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>