Commit Graph

323 Commits

Author SHA1 Message Date
f3df3a7727 PixelShaderGen: Clamp texture layer when using manual texture sampling with stereoscopic 3D
Otherwise, texelFetch() will use an out-of-bounds layer for game textures (that have 1 layer; EFB copies have 2 layers in stereoscopic 3D mode), which is undefined behavior (often resulting in a black image). The fast texture sampling path uses texture(), which always clamps (see https://www.khronos.org/opengl/wiki/Array_Texture#Access_in_shaders), so it was unaffected by this difference.
2022-12-27 13:45:13 -08:00
9aece1810c Merge pull request #10836 from iwubcode/d3d_uint_fix
VideoCommon: fix uint shader compiler error when using d3d
2022-10-18 12:12:08 -04:00
5ef8a7973e BPMemory: Make TevKSel more clear
It stores both the konst selection value for alpha and color channels (for two tev stages per ksel), and half of a swap table row (there are 4 total swap tables, which can be used for swizzling the rasterized color and the texture color, and indices selecting which tables to use are stored per tev stage in the alpha combiner).  Since these are indexed very differently, the old code was hard to follow.
2022-08-29 11:10:05 -07:00
05135b4f43 VideoCommon: fix uint shader compiler error in specialized shaders. This error is in renderers that use uint for their color output (for logic ops). Remove D3D check for uint output since other backends could use uint output as well. 2022-08-06 16:08:58 -05:00
a5ef9dfd53 VideoBackends:Metal: Use DriverDetails for bugs 2022-07-21 20:44:19 -05:00
70b0b03c3c Merge pull request #10747 from tellowkrinkle/LateUIDFixup
Add a post-cache shader UID fixup pass
2022-07-17 00:43:16 -04:00
6ab24e6c17 VideoCommon: Better driver bug handling
Adds a pass to process driver deficiencies between UID caching and use, allowing a full view of the whole pipeline, since some bugs/workarounds involve interactions between blend modes and the pixel shader
2022-07-13 21:51:24 -05:00
6bd0fc86ba VideoCommon: Properly mask fbfetch logic op emulation 2022-07-13 02:27:45 -05:00
c2d3b7e7f5 VideoCommon: update pixel shader output to match old D3D code 2022-06-24 18:09:53 -05:00
cad1d6ce90 VideoCommon: fix support of stereoscopic rendering after moving d3d to SPIRV generation 2022-06-24 18:09:53 -05:00
5dd2704416 D3D / VideoCommon: generate HLSL from SPIRV 2022-06-24 18:09:53 -05:00
26529a31ab VideoCommon: Fix SSBO layout and remove associated "bug" 2022-06-16 20:26:11 -05:00
c42392c565 Merge pull request #10290 from OatmealDome/m1-earlyz-bug
DriverDetails: Add broken discard with early-Z bug on Apple Silicon GPUs
2022-04-24 13:30:04 -04:00
04fdadd9d5 VideoCommon: Rename norm0/norm1/norm2 to normal/tangent/binormal 2022-04-22 16:54:36 -07:00
259a5fc7c0 DriverDetails: Add broken discard with early-Z bug on Apple Silicon GPUs 2022-04-20 14:56:34 -04:00
c1d87db6fa PixelShaderGen: Add support for non-dual source shader blending 2022-04-19 10:55:26 -04:00
abffa93a72 MoltenVK: Fix pixel shader typo 2022-04-10 20:51:20 +02:00
bbb64ff993 Shadergen: Use real_ocol0 workaround for shader logic ops
Previously we were using this workaround when using framebuffer fetch
to emulate dual source blending, but it seems like we also need to use
it when using framebuffer fetch to emulate logic ops, otherwise some
Adreno devices get a crash when compiling OpenGL ES ubershaders.

Using the workaround in specialized shaders doesn't seem to be
necessary, but I've made the same change there for consistency.

This gets us closer to fixing https://bugs.dolphin-emu.org/issues/12791
but doesn't actually fix it.
2022-02-28 18:32:19 +01:00
608799f1e4 PixelShaderGen: Check uid_data for use_shader_logic_op
On devices which have hardware support for dual source blending
but not logic ops, this lets us skip performing the framebuffer
fetch in situations where the game isn't actually using logic ops.
2022-02-28 00:04:31 +01:00
444f6fd0cb Treat alpha as 0 if alpha is 1 for blending
This removes the white box in fortune street again, without causing Mario Kart Wii to regress.
2022-02-08 15:15:15 -08:00
0327e6acb4 Use the same logic for lerp bias for color and alpha
It doesn't make sense for alpha to add the bias ONLY when dividing by 2, while color doesn't apply the bias for divide by 2 only; hardware testing indicates that alpha should have the bias.

This fixes the menus in Mario Kart Wii (https://bugs.dolphin-emu.org/issues/11909) but reintroduces the white rectangle in Fortune Street.

This reverts commit 5aaa5141ed (and several other matching changes elsewhere).
2022-02-08 15:15:15 -08:00
29d6dd609c Fix non-constexpr format strings 2022-01-13 11:11:08 -08:00
fc0d958e26 PixelShaderGen: Fix invalid use of int3(0)
This syntax is allowed by GLSL, but HLSL doesn't allow it.  This meant that games using R8 comparisons in equal mode would produce shaders that failed to compile.  Super Mario Galaxy's water levels were affected by this.
2022-01-12 21:50:34 -08:00
a96cfe2531 GLES: Fix LOD bias int/float mismatch
Another simple GLES shader compilation error.
2021-12-28 19:19:22 +01:00
056613ecc5 PixelShaderGen: Add LOD bias to texture() call on systems that don't support it in the sampler 2021-12-28 02:10:45 -05:00
b1f79d9ecf Merge pull request #10215 from OatmealDome/shader-logic-ops
VideoCommon: Support shader logic ops on Metal (Apple GPUs) and OpenGL ES
2021-12-22 16:39:54 -05:00
380b333387 PixelShaderGen: Convert to EnumMap 2021-12-18 12:51:55 -08:00
18b2f6953d PixelShaderGen: Add shader logic ops support on OpenGL ES
To do this, I had to decouple framebuffer fetch from shader blending. We need to be able to access framebuffer fetch input when using shader logic ops.
2021-12-06 22:36:40 -05:00
e0837cb847 PixelShaderGen: Add shader logic ops support on Metal 2021-12-06 22:36:40 -05:00
c12b9b013b PixelShaderGen: Add logic ops to pixel_shader_uid_data 2021-12-06 22:36:34 -05:00
04ec02c06b ConstantManager: Add logic ops to PixelShaderConstants 2021-11-24 17:52:26 -05:00
1adff1c467 VideoCommon: Skip textureQueryLevels if it doesn't exist 2021-11-17 21:28:39 -08:00
bdcfb31187 VideoCommon: Handle custom texture sizes correctly
Specifically, when using Manual Texture Sampling, if textures sizes don't match the size the game specifies, things previously broke.  That can happen with custom textures, and also with scaled EFB copies at non-native IRs.  It breaks most obviously by not scaling the texture coordinates (so only part of the texture shows up), but the hardware wrapping functionality also assumes texture sizes are a power of 2 (or else it will behave weirdly in a way that matches how hardware behaves weirdly).  The fix is to provide alternative texture wrapping logic when custom texture sizes are possible.
2021-11-17 21:28:36 -08:00
93eea7cb13 VideoCommon: Add option to use old behavior (Fast Texture Sampling)
Co-authored-by: JosJuice <josjuice@gmail.com>
2021-11-17 21:27:32 -08:00
ee80298ca4 VideoCommon: Implement diagonal LOD
Note that both GLSL and HLSL provide a fwidth (fragment width) function defined as `fwidth(p) = abs(dFdx(p)) + abs(dFdy(p))`.  However, it's easy enough to implement this ourselves (and it makes the code a bit more obvious).
2021-11-17 20:04:34 -08:00
51e3334526 VideoCommon: Use coarse derivatives for Manual Texture Sampling if possible 2021-11-17 20:04:34 -08:00
ddf2691395 VideoCommon: Manually handle texture wrapping and sampling 2021-11-17 20:04:34 -08:00
4a9b26de86 VideoCommon: Expose SamplerState to shaders
The benefit to exposing this over the raw BP state is that adjustments Dolphin makes, such as LOD biases from arbitrary mipmap detection, will work properly.
2021-11-17 20:04:34 -08:00
9ef228503a VideoCommon: Provide raw texdims to shaders 2021-11-17 20:04:34 -08:00
555a93057c VideoCommon: Allow BitfieldExtract in specialized shaders 2021-11-17 20:04:33 -08:00
a372a5947b VideoCommon: Fix color channel logic when per-pixel lighting is in use
This was broken in #10012 (specifically by 06579e4d53 and c3dec34391).
2021-10-13 20:43:32 -07:00
1161af8059 VideoCommon: Abstract bounding box
This moves much of the duplicated bounding box code into VideoCommon,
leaving only the specific buffer implementations in each backend.
2021-10-04 15:51:24 -04:00
92fc4f1eca PixelShaderGen: Fix OpenGL ES bounding box compilation error 2021-07-20 11:39:32 +02:00
5e0520b6e0 VideoCommon: Expand vector comparisons instead of overloading any()
For whatever stupid reason, Mali drivers do not allow overloading
built-in functions.
2021-07-20 05:24:47 -04:00
a379456f38 Fix indentation for alphabump 2021-07-08 15:48:14 -07:00
2feced2e33 Fix indirect textures when format is not ITF_8 2021-07-08 15:48:14 -07:00
e149ad4f0a treewide: convert GPLv2+ license info to SPDX tags
SPDX standardizes how source code conveys its copyright and licensing
information. See https://spdx.github.io/spdx-spec/1-rationale/ . SPDX
tags are adopted in many large projects, including things like the Linux
kernel.
2021-07-05 04:35:56 +02:00
3da0976a81 VideoCommon: Define scalar any() and all() functions in GLSL
This fixes bounding box shaders failing to compile under Vulkan, due to
differences between GLSL and HLSL in the return value of vector
comparisons and what types these functions accept. I included all() for
the sake of completeness.
2021-06-13 04:50:35 -04:00
0c6e00ce0c Merge pull request #9801 from Techjar/bbox-rounding-hack
VideoCommon: Only include centered pixels in bounding box
2021-06-11 00:41:11 -04:00
b267f54259 VideoCommon: Only include centered pixels in bounding box
At higher resolutions, our bounding box dimensions end up being
slightly larger than original hardware in some cases. This is not
necessarily wrong, it's just an artifact of rendering at a higher
resolution, due to bringing out detail that wouldn't have appeared on
original hardware. It causes a texel to fall partially on what would
have been a single pixel at native resolution, resulting in the
coordinates getting bumped up to the next valid value. In many cases,
these slightly larger bounding boxes are perfectly fine, as games don't
hard-code expected dimensions. It is problematic in Paper Mario TTYD
though, for a somewhat complicated reason.

Paper Mario TTYD frequently uses EFB copies to pre-render a bunch of
animation frames for a character sprite (especially in Chapter 2), so
that it can then render 100 or more of them without bringing the
GameCube to its knees. Based on my observation, the game seems to set
aside a region of memory to store these EFB copies. This region is
obviously fairly small, as the GameCube only has 24MB of RAM. There are
2 rooms in Chapter 2 where you fight a horde of as many as 100 Jabbies,
which are also rendered using EFB copies, so in this room the game ends
up making 130(!) EFB copies just for Puni and Jabbi sprites. This seems
to nearly fill the region of memory it set aside for them.
Unfortunately, our slightly larger bounding boxes at higher resolutions
results in overflowing this memory, causing very strange behavior. Some
EFB copies partially overlap game state, resulting in reading it as a
garbage RGB5A3 texture that constantly changes. Others apparently
somehow trigger a corner case in our persistent buffer mapping, causing
them to partially overwrite earlier EFB copies.

What this change does is only include the screen coordinates that align
with the equivalent native resolution pixel centers, which generally
results in the bounding boxes being more in line with original
hardware. It isn't perfect, but it's enough to fix Paper Mario TTYD's
Jabbi rooms by avoiding the buffer overflow. Notably, it is more
accurate at odd resolutions than at even resolutions. Native resolution
is completely unaffected by this change, as should be the case. This
change may also have a small positive impact on shader performance at
higher resolutions, as there will be less atomic operations performed.
2021-06-10 21:49:52 -04:00