164 Commits

Author SHA1 Message Date
Giovanni Mascellani
d2f8a576a8 vkd3d-shader/hlsl: Avoid infinite loop and invalid derefs in copy-prop.
Co-authored-by: Francisco Casas <fcasas@codeweavers.com>
Co-authored-by: Zebediah Figura <zfigura@codeweavers.com>

Because copy_propagation_transform_object_load() replaces a deref
instead of an instruction, it is currently prone to two problems:

1- It can replace a deref with the same deref, returning true every
time and getting the compilation stuck in an endless loop of
copy-propagation iterations.

2- When performed multiple times in the same deref, the second time it
can replace the deref with a deref from a temp that is only valid in
another point of the program execution, resulting in an incorrect value.

This patch preempts this by avoiding replacing derefs when the new deref
doesn't point to a uniform variable. Because, uniform variables cannot
be written to.
2023-01-26 21:52:07 +01:00
Nikolay Sivov
3c23e1713c vkd3d-shader/hlsl: Implement sqrt() for SM1. 2023-01-25 22:10:15 +01:00
Francisco Casas
6b82ba9488 vkd3d-shader/hlsl: Fold swizzle chains. 2023-01-24 18:10:53 +01:00
Zebediah Figura
b7d34e8307 vkd3d-shader/hlsl: Apply copy propagation to swizzled loads. 2023-01-24 18:10:50 +01:00
Francisco Casas
c2a7a40d3a vkd3d-shader/hlsl: Replace loads with constants in copy prop.
If a hlsl_ir_load loads a variable whose components are stored from different
instructions, copy propagation doesn't replace it.

But if all these instructions are constants (which currently is the case
for value constructors), the load could be replaced with a constant value.
Which is expected in some other instructions, e.g. texel_offsets when
using aoffimmi modifiers.

For instance, this shader:

```
sampler s;
Texture2D t;

float4 main() : sv_target
{
    return t.Gather(s, float2(0.6, 0.6), int2(0, 0));
}
```

results in the following IR before applying the patch:
```
  float | 6.00000024e-01
  float | 6.00000024e-01
   uint | 0
        | = (<constructor-2>[@4].x @2)
   uint | 1
        | = (<constructor-2>[@6].x @3)
 float2 | <constructor-2>
    int | 0
    int | 0
   uint | 0
        | = (<constructor-5>[@11].x @9)
   uint | 1
        | = (<constructor-5>[@13].x @10)
   int2 | <constructor-5>
 float4 | gather_red(resource = t, sampler = s, coords = @8, offset = @15)
        | return
        | = (<output-sv_target0> @16)
```

and this IR afterwards:
```
 float2 | {6.00000024e-01 6.00000024e-01 }
   int2 | {0 0 }
 float4 | gather_red(resource = t, sampler = s, coords = @2, offset = @3)
        | return
        | = (<output-sv_target0> @4)
```
2023-01-24 18:10:45 +01:00
Zebediah Figura
8c2b8ff245 vkd3d-shader/hlsl: Synthesize the swizzle and replace the instruction inside of copy_propagation_compute_replacement().
Rename it to copy_propagation_replace_with_single_instr() accordingly.

The idea is to introduce a constant vector replacement pass which will do the
same thing.
2023-01-24 18:10:41 +01:00
Zebediah Figura
5d34790402 vkd3d-shader/hlsl: Call copy_propagation_get_value() directly in copy_propagation_transform_object_load().
copy_propagation_compute_replacement() is not doing very much for us, and
conceptually is a bit of an odd fit anyway, since it's meant to deal with
multi-component types.
2023-01-24 18:10:40 +01:00
Zebediah Figura
8fd30aa87d vkd3d-shader/hlsl: Add some swizzle manipulation definitions. 2023-01-24 18:10:39 +01:00
Zebediah Figura
9c817e5e6d vkd3d-shader/hlsl: Forbid recursive calls. 2023-01-19 19:16:27 +01:00
Francisco Casas
17d6a4411e vkd3d-shader/hlsl: Validate that non-uniform objects are not referenced.
Note that in the future we should call
validate_static_object_references() after DCE and pruning branches,
because shaders such as these compile (at least in more modern versions
of the native compiler):

Branch pruning:
```
static RWTexture2D<float> tex;

float4 main() : sv_target
{
    if (0)
    {
        tex[int2(0, 0)] = 2;
    }
    return 0;
}
```

DCE:
```
static Texture2D tex;
uniform uint i;

float4 main() : sv_target
{
    float4 unused = tex.Load(int3(0, 1, 2));

    return 0;
}
```

These are "todo" tests in hlsl-static-initializer.shader_test
that depend on this.
2023-01-19 12:29:39 +01:00
Zebediah Figura
b29d3489de vkd3d-shader/hlsl: Generate IR for user-defined function calls. 2023-01-13 17:32:42 +01:00
Zebediah Figura
30550c0831 vkd3d-shader/hlsl: Avoid assuming that expressions have at least one argument. 2023-01-13 17:32:40 +01:00
Francisco Casas
cc811dc3c2 vkd3d-shader/hlsl: Rename hlsl_struct_field.modifiers to "storage_modifiers". 2023-01-11 16:03:47 +01:00
Francisco Casas
4dbbb8beb4 vkd3d-shader/hlsl: Rename hlsl_ir_var.modifiers to "storage_modifiers". 2023-01-11 16:03:45 +01:00
Francisco Casas
f100f5b726 vkd3d-shader/hlsl: Check for non-static object references on resource stores. 2022-11-18 22:29:04 +01:00
Francisco Casas
f21693b284 vkd3d-shader/hlsl: Use reg_size as component count when allocating a single register.
Otherwise, for instance, the added test results in:

debug_hlsl_writemask: Assertion `!(writemask & ~VKD3DSP_WRITEMASK_ALL)' failed.

Which happens in allocate_variable_temp_register() when the variable's
type reg_size is <= 4 but its component count is larger, which may
happen if it contains objects.
2022-11-10 22:48:26 +01:00
Francisco Casas
3153ce3145 vkd3d-shader/hlsl: Don't allocate object types as constant registers. 2022-11-10 22:48:16 +01:00
Zebediah Figura
718c79b823 vkd3d-shader/hlsl: Parse the numthreads attribute. 2022-11-08 20:53:04 +01:00
Zebediah Figura
d6799bd5d3 vkd3d-shader/hlsl: Parse function attributes. 2022-11-08 20:53:03 +01:00
Zebediah Figura
c416627e64 vkd3d-shader/hlsl: Propagate copies for resource store instructions. 2022-10-31 22:07:47 +01:00
Zebediah Figura
03f9d16047 vkd3d-shader/hlsl: Parse UAV stores. 2022-10-31 22:07:44 +01:00
Zebediah Figura
2ec67e0f10 vkd3d-shader/hlsl: Allocate UAVs. 2022-10-31 22:07:40 +01:00
Giovanni Mascellani
0a07ac6f88 vkd3d-shader/hlsl: Lower float modulus.
Signed-off-by: Giovanni Mascellani <gmascellani@codeweavers.com>
2022-10-19 21:59:17 +02:00
Giovanni Mascellani
eb119878f7 vkd3d-shader/hlsl: Lower int modulus.
Signed-off-by: Giovanni Mascellani <gmascellani@codeweavers.com>
2022-10-19 21:59:15 +02:00
Giovanni Mascellani
eb7b594002 vkd3d-shader/hlsl: Lower int division. 2022-10-17 17:58:17 +02:00