Commit Graph

272 Commits

Author SHA1 Message Date
Hans-Kristian Arntzen
c9d4f9cd74 MSL: Add a workaround path to force native arrays for everything. 2020-02-24 12:47:14 +01:00
Chip Davis
ae6c05f6f4 MSL: Move inline uniform blocks to the end of the argument buffer.
Limit inline blocks to one per descriptor set.

This should avoid the need for complicated code to calculate the
argument buffer ID stride of an inline uniform block. If there's demand
for more inline blocks, we can revisit this.
2020-01-25 13:40:51 -06:00
Chip Davis
fedbc35315 MSL: Support inline uniform blocks in argument buffers.
Here, the inline uniform block is explicit: we instantiate the buffer
block itself in the argument buffer, instead of a pointer to the buffer.
I just hope this will work with the `MTLArgumentDescriptor` API...

Note that Metal recursively assigns individual members of embedded
structs IDs. This means for automatic assignment that we have to
calculate the binding stride for a given buffer block. For MoltenVK,
we'll simply increment the ID by the size of the inline uniform block.
Then the later IDs will never conflict with the inline uniform block. We
can get away with this because Metal doesn't require that IDs be
contiguous, only monotonically increasing.
2020-01-24 18:51:24 -06:00
Hans-Kristian Arntzen
f9818f0804 Update license headers to 2020. 2020-01-16 15:24:37 +01:00
Hans-Kristian Arntzen
c3bd136df1 MSL: Add support for force-activating IAB resources.
Important for ABI compatibility on MSL in certain cases.
2020-01-16 11:12:06 +01:00
Hans-Kristian Arntzen
cc153f8d7f HLSL: Add a resource remapping API similar to MSL.
Allows more flexibility of how resources are assigned without having to
remap decorations.
2020-01-09 12:41:06 +01:00
Hans-Kristian Arntzen
93f3265fe0 MSL: Deal with packing vectors for vertex input/fragment output. 2020-01-07 14:14:31 +01:00
Hans-Kristian Arntzen
123fa906ec Merge branch 'ue4-merge' 2019-10-28 15:14:42 +01:00
Hans-Kristian Arntzen
39bd5f1834 Run format_all.sh. 2019-10-28 12:55:14 +01:00
Hans-Kristian Arntzen
1d5eba67f8 MSL: Remove hacky workaround for patch constant passing.
There was a hack to workaround a bug in DXC where control point -> patch
constant phase was passed in Function storage, but we have to use
Workgroup here. We will not support these kinds of hacks for invalid
SPIR-V, so hack the reference files to use the "proper" fix and remove
the hack for time being.
2019-10-28 12:52:28 +01:00
Hans-Kristian Arntzen
27d6d45671 MSL: Rewrite tessellation_access_chain.
To support loading array of array properly in tessellation, we need a
rewrite of how tessellation access chains are handled.

The major change is to remove the implicit unflatten step inside
access_chain which does not take into account the case where you load
directly from a control point array variable.

We defer unflatten step until OpLoad time instead.
This fixes cases where we load array of {array,matrix,struct}.

Removes the hacky path for MSL access chain index workaround.
2019-10-26 16:10:12 +02:00
Hans-Kristian Arntzen
8066d13599 MSL: Rewrite propagated depth comparison state handling.
Far cleaner, and more correct to run the traversal twice.
Fixes a case where we propagate depth state through multiple functions.
2019-10-26 16:10:11 +02:00
Bill Hollings
4b5c6c188c MSL: Support option for treating 1D textures as 2D textures of height 1.
Add CompilerMSL::Options::texture_1D_as_2D.

Metal imposes significant restrictions on 1D textures, including not being
renderable, clearable, or permitting mipmaps. This option allows SPIR-V 1D
textures to be treated as 2D textures to permit this additional behaviour.
App must of course supply the textures to Metal as 2D textures.
2019-10-24 12:31:20 -04:00
Hans-Kristian Arntzen
4ac12594c9 MSL: Be a little clearer how needs_base_vertex_idx is implemented.
There is an implicit tristate with {-1, 0, +1} values, but it was not
obvious how this was supposed to work before studying the implementation,
so refactor into a tristate enum class.
2019-10-24 12:41:37 +02:00
Hans-Kristian Arntzen
6ca3fe22c3 MSL: Remove some dead code w.r.t. vertex/instance_idx. 2019-10-24 12:36:01 +02:00
Hans-Kristian Arntzen
6edbf0c9e9 MSL: Minor cleanups for texture atomic emulation.
Storing pointers to internal objects is generally not done, IDs are
preferred.
2019-10-24 11:30:20 +02:00
Lukas Hermanns
b0d616aa6d Removed 'argument_buffer_offset' and fixed packed matrix Metal output. 2019-10-23 16:28:32 -04:00
Lukas Hermanns
6673a675ba Simplified overriding of 'access_chain_internal' function in CompilerMSL. 2019-10-22 11:06:16 -04:00
Lukas Hermanns
84351d3aed Merge remote-tracking branch 'upstream/master' 2019-10-21 18:55:36 -04:00
Lukas Hermanns
e1b161b54b Removed bounds checks in favor of SPIRV-Tools pass '--graphics-robust-access' 2019-10-21 16:39:53 -04:00
Hans-Kristian Arntzen
4bb673a626 MSL: Add opt-in support for huge IABs.
If there are enough members in an IAB, we cannot use the constant
address space as MSL compiler complains about there being too many
members. Support emitting the device address space instead.
2019-10-14 16:20:34 +02:00
Lukas Hermanns
ffbd801853 Added '--msl-invariant-float-math' option and new test case for it. 2019-10-09 14:03:06 -04:00
Lukas Hermanns
f3a6d28a1d Further updates for pull request #1162; also added two test cases for spvCubemapTo2DArrayFace function and added '--msl-framebuffer-fetch'/ '--msl-emulate-cube-array' compiler options. 2019-09-27 15:49:54 -04:00
Lukas Hermanns
7ad0a84778 Updates for pull request #1162 2019-09-24 14:35:25 -04:00
Lukas Hermanns
37df74035b Merge branch 'ue4_dev' 2019-09-20 09:42:42 -04:00
Lukas Hermanns
9f9276f5ce Fixed false-positive optimization of builtin variables (may happen when 'spvOut' is emitted). 2019-09-19 14:44:30 -04:00
Lukas Hermanns
50ac6862ac Rearranged all 'UE Change' comments to match to project's coding style. 2019-09-18 14:03:54 -04:00
Lukas Hermanns
51be601922 Avoid emitting 'spvUnsafeArray<>', 'spvFMul*', and 'spvFAdd' custom functions if they are not needed. 2019-09-17 15:10:39 -04:00
Lukas Hermanns
36eab88b23 Further adjustments to make Metal backend work again in UE4 on Mac. 2019-09-17 11:40:01 -04:00
Lukas Hermanns
7cf5d4f7a1 Added a new 'emulate_cube_array' option to SPIRV-Cross to cope with translating TextureCubeArray into texture2d_array for iOS where this type is not available. (Original Author: Mark Satterthwaite) 2019-09-13 17:24:27 -04:00
Lukas Hermanns
a9f3c981d9 Adjustments after rebase of ue4_dev branch. 2019-09-13 14:03:02 -04:00
Mark Satterthwaite
69b703f1da Add an option to SPIRV-Cross to enforce invariant floating point math to prevent different depth calculation between prepass & basepass when running on Metal 2.0 and earlier. 2019-09-12 08:35:15 -04:00
Mark Satterthwaite
842e8c48c6 Provide the Metal bindings as part of the options structure as that is more convenient. 2019-09-10 13:24:59 -04:00
Mark Satterthwaite
9e54a8dd7b Slight modifications to IAB support for Metal output, so that the caller can specify an offset for the IAB start index, as for HLSL shaders UAVs need to occupy slots 0-7. The runtime support for SSBO robustness is also much simpler if the buffer size block is at index 0. Change made in two parts. 1. Allow the caller to specify the Metal translation should use argument buffers. 2. Move this to the front of IABs for convenience of the runtime. 2019-09-10 13:09:49 -04:00
Mark Satterthwaite
d9f3576305 Metal doesn't automatically enforce robust access to buffers unlike other APIs, so for storage-buffers, which become raw T* buffers in Metal, we need to fetch the buffer size and clamp the access to a valid index within the buffer ourselves. This is essential for shaders converted from HLSL which expects all resource access to be robust, though this implementation is technically different to the HLSL specification of return-0 for OOB reads, ignore OOB writes. 2019-09-10 12:32:32 -04:00
Mark Satterthwaite
9ce3158193 When compiling from HLSL which pads and aligns float[]/float2[] within structures to float4[] we need to unpack the original type in Metal from the float4. 2019-09-10 11:21:43 -04:00
Mark Satterthwaite
42b8a62870 Fixes to the generation of Metal tessellation shaders from SPIRV so that it works correctly in more complicated cases.
First, when generating from HLSL before invoking the code that comes from the HLSL patch-function a control-flow and full memory-barrier are required to ensure that all the temporary values in thread-local storage for the patch are available.
    Second, the inputs to control and evaluation shaders must be properly forwarded from the global variables in SPIRV to the member variables in the relevant input structure.
    Finally when arrays of interpolators are used for input or output we need to add an extra level of array indirection because Metal works at a different granularity than SPIRV.

    Five parts.
    1. Fix tessellation patch function processing.
    2. Fix loads from tessellation control inputs not being forwarded to the gl_in structure array.
    3. Fix loads from tessellation evaluation inputs not being forwarded to the stage_in structure array.
    4. Workaround SPIRV losing an array indirection in tessellation shaders - not the best solution but enough to keep things progressing.
    5. Apparently gl_TessLevelInner/Outer is special and needs to not be placed into the input array.
2019-09-10 10:37:07 -04:00
Mark Satterthwaite
de6441af88 Work-around HLSL using zero-based InstanceID and VertexID variables, but SPIRV, like Metal, includes BaseInstance & BaseVertex. Until this can be fixed in DXC, which is really the proper place to solve this, we can decrement InstanceID & VertexID when the source is HLSL. Made in two parts. 1. Handle HLSL-style 0-based vertex/instance index. 2. We zero-base the InstanceID & VertexID variables for HLSL emulation elsewhere, so don't do it twice. 2019-09-09 16:55:59 -04:00
Mark Satterthwaite
97a66ff906 On iOS sub-passes can be implemented using the frame-buffer fetch API which is much more efficient than binding the textures. Change was made in three parts. 1. Use Metal's native frame-buffer fetch API for subpass inputs. 2. Make sure that frame-buffer-fetch is only available on iOS. 3. Default to using Metal's native frame-buffer fetch for subpass inputs on iOS. 2019-09-09 15:02:11 -04:00
Mark Satterthwaite
32557e9093 SPIRV doesn't distinguish depth textures from regular textures, but Metal does, so if we've ever seen a depth comparison operation we must ensure that the texture is specified as a depth-texture. 2019-09-06 16:58:27 -04:00
Hans-Kristian Arntzen
333980ae91 Refactor into stronger types in public API.
Some fallout where internal functions are using stronger types.
Overkill to move everything over to strong types right now, but perhaps
move over to it slowly over time.
2019-09-06 12:29:47 +02:00
Chip Davis
cb35934248 MSL: Support dynamic offsets for buffers in argument buffers.
Vulkan has two types of buffer descriptors,
`VK_DESCRIPTOR_TYPE_UNIFORM_BUFFER_DYNAMIC` and
`VK_DESCRIPTOR_TYPE_STORAGE_BUFFER_DYNAMIC`, which allow the client to
offset the buffers by an amount given when the descriptor set is bound
to a pipeline. Metal provides no direct support for this when the buffer
in question is in an argument buffer, so once again we're on our own.
These offsets cannot be stored or associated in any way with the
argument buffer itself, because they are set at bind time.  Different
pipelines may have different offsets set. Therefore, we must use a
separate buffer, not in any argument buffer, to hold these offsets. Then
the shader must manually offset the buffer pointer.

This change fully supports arrays, including arrays of arrays, even
though Vulkan forbids them. It does not, however, support runtime
arrays. Perhaps later.
2019-09-05 23:29:00 -05:00
Mark Satterthwaite
5e8590a23d Emulate texture atomics in Metal by binding the underlying buffer that backs the resource to a separate binding point and using that for Metal's atomic operations. This will work with texture_buffer and texture2d created from an MTLBuffer, so is perfect for emulating HLSL atomics on RWBuffer and sufficient, but not ideal, for RWTexture2D with some restrictions (limited format support and can't be used for render-targets). 2019-09-05 15:13:28 -04:00
Mark Satterthwaite
239e04762b Support Metal 2.1's texture_buffer type which is the equivalent to HLSL's Buffer/RWBuffer, so doesn't require modifying buffer sizes to match alignments. 2019-09-05 14:46:15 -04:00
Mark Satterthwaite
8596bf5ee2 In order to use Metal shader libraries properly you have to ensure that you have no duplicated global symbol names for different entities, otherwise 'metallib' won't be able to combine multiple shaders into a single library. This is broken into two parts. 1. Constant arrays of non-primitive types (i.e. matrices) won't link properly into Metal libraries. 2. Metal helper functions must be static force-inline otherwise they will cause problems when linked together in a single Metallib. 2019-09-05 14:39:06 -04:00
Mark Satterthwaite
d50659af92 Rework the way arrays are handled in Metal to remove the array copies as they are unnecessary from Metal 1.2. There were cases where copies were not being inserted and others appeared unncessary, using the template type should allow the 'metal' compiler to do the best possible optimisation. The changes are broken into three stages. 1. Allow Metal to use the array<T> template to make arrays a value type. 2. Force the use of C style array declaration for some cases which cannot be wrapped with a template. 3. Threadgroup arrays can't have a wrapper type. 4. Tweak the code to use unsafe_array in a few more places so that we can handle passing arrays of resources into the shader and then through shaders into sub-functions. 5. Handle packed matrix types inside arrays within structs. 6. Make sure that builtin arguments still retain their array qualifiers when used in leaf functions. 7. Fix declaration of array-of-array constants for Metal so we can use the array<T> template. 2019-09-05 12:39:44 -04:00
Chip Davis
103817009c MSL: Force storage images on iOS to use discrete descriptors.
Writable textures cannot use argument buffers on iOS. They must be
passed as arguments directly to the shader function. Since we won't know
if a given storage image will have the `NonWritable` decoration at the
time we encode the argument buffer, we must therefore pass all storage
images as discrete arguments. Previously, we were throwing an error if
we encountered an argument buffer with a writable texture in it on iOS.
2019-09-05 11:01:05 -05:00
Chip Davis
39dce88d3b MSL: Add support for sampler Y'CbCr conversion.
This change introduces functions and in one case, a class, to support
the `VK_KHR_sampler_ycbcr_conversion` extension. Except in the case of
GBGR8 and BGRG8 formats, for which Metal natively supports implicit
chroma reconstruction, we're on our own here. We have to do everything
ourselves. Much of the complexity comes from the need to support
multiple planes, which must now be passed to functions that use the
corresponding combined image-samplers. The rest is from the actual
Y'CbCr conversion itself, which requires additional post-processing of
the sample retrieved from the image.

Passing sampled images to a function was a particular problem. To
support this, I've added a new class which is emitted to MSL shaders
that pass sampled images with Y'CbCr conversions attached around. It
can handle sampled images with or without Y'CbCr conversion. This is an
awful abomination that should not exist, but I'm worried that there's
some shader out there which does this. This support requires Metal 2.0
to work properly, because it uses default-constructed texture objects,
which were only added in MSL 2. I'm not even going to get into arrays of
combined image-samplers--that's a whole other can of worms.  They are
deliberately unsupported in this change.

I've taken the liberty of refactoring the support for texture swizzling
while I'm at it. It's now treated as a post-processing step similar to
Y'CbCr conversion. I'd like to think this is cleaner than having
everything in `to_function_name()`/`to_function_args()`. It still looks
really hairy, though. I did, however, get rid of the explicit type
arguments to `spvGatherSwizzle()`/`spvGatherCompareSwizzle()`.

Update the C API. In addition to supporting this new functionality, add
some compiler options that I added in previous changes, but for which I
neglected to update the C API.
2019-09-01 18:35:53 -05:00
Hans-Kristian Arntzen
3ccfbce264 Run format_all.sh. 2019-08-28 14:25:26 +02:00
Hans-Kristian Arntzen
9436cd3036 MSL: Deal with array copies from and to threadgroup. 2019-08-27 13:18:01 +02:00