4559 Commits

Author SHA1 Message Date
Francisco Casas
b84f560bdf vkd3d-shader/hlsl: Set writemasks correctly for SM1 scalar and vector types.
Because of the change introduced in

f21693b2 vkd3d-shader/hlsl: Use reg_size as component count when allocating a single register.

SM1 scalars and vectors were not longer getting the correct writemask
when they are allocated.

This happened because they have to reserve the whole register even if
they only use some of its components, so their reg_size may differ from
the number of components.

This commit fixes that.
2023-02-08 22:09:39 +01:00
Francisco Casas
6b4a4c4c6d vkd3d-shader/hlsl: Expect component count in allocate_register().
This in order to set the correct writemasks for SM1 registers.
2023-02-08 22:09:38 +01:00
Francisco Casas
0ecac994d0 vkd3d-shader/hlsl: Rename 'component_count' arguments to 'reg_size'.
component_count will be used in the next patch for the actual number of
components of the type and not its register size.
2023-02-08 22:09:36 +01:00
Zebediah Figura
ab601d65cd vkd3d-shader: Initialize the parameter allocator head.
This fixes a memory leak.

Fixes: 007f894b9493eb80091899dab971e369f1673a4f
2023-02-07 22:15:08 +01:00
Nikolay Sivov
91a70c7dda vkd3d-shader/hlsl: Handle saturate() for SM1.
Signed-off-by: Nikolay Sivov <nsivov@codeweavers.com>
2023-02-07 22:15:06 +01:00
Zebediah Figura
77d716c811 vkd3d-shader/hlsl: Add functions to the global list inside the func_prototype rule. 2023-02-07 22:15:06 +01:00
Zebediah Figura
4c46075d86 vkd3d-shader/hlsl: Use the original hlsl_ir_function_decl struct rather than allocating a new one for each definition.
We need to make sure every invocation points to the same hlsl_ir_function_decl
and the same parameters.

This fixes some invalid memory accesses.
2023-02-07 22:15:06 +01:00
Zebediah Figura
25d49b518d vkd3d-shader/hlsl: Put synthetic variables into a dummy scope.
Prevent them from being ever looked up.

Our naming scheme for synthetic variables already effectively prevents this, but
this is better for clarity. We also will need to be able to move some named
variables into a dummy scope to account for complexities around function
definition and declarations.
2023-02-07 22:15:06 +01:00
Zebediah Figura
8755a92196 vkd3d-shader/hlsl: Add a hlsl_cleanup_semantic() helper. 2023-02-07 22:15:06 +01:00
Zebediah Figura
cb2c89a589 vkd3d-shader/hlsl: Store function parameters in an array. 2023-02-07 22:15:06 +01:00
Zebediah Figura
898fc9e198 vkd3d: Fix checking for failure from SleepConditionVariableCS().
Fixes: 552926cfca64db45e9731f675c65a7214bfa6441
2023-02-07 22:15:06 +01:00
Matteo Bruni
2e074ebce7 vkd3d: Initialize image aspect for NULL SRVs. 2023-02-07 22:08:00 +01:00
Giovanni Mascellani
552926cfca vkd3d: Do not allow synchronization primitives to fail.
In practice they never fail. If they fail, it means that there
is some underlying platform problem and there is little we can do
anyway. Under pthreads function prototypes allow returning failure,
but that's only used for "error checking" mutexes, which we
don't use.

On the other hand, error handling in vkd3d is rather inconsistent:
sometimes the errors are ignored, sometimes logged, sometimes
passed to the caller. It's hard to handle failures appropriately
if you can't even keep your state consistent, so I think it's
better to avoid trying, assume that synchronization primitives do
not fail and at least have consistent logging if something goes
wrong.
2023-02-02 20:51:27 +01:00
Zebediah Figura
a66fe31fe5 vkd3d: Do not write the point size for SPIR-V shaders.
We disable shaderTessellationAndGeometryPointSize.
2023-02-02 20:51:19 +01:00
Zebediah Figura
8fc8d34ea0 vkd3d-shader/spirv: Introduce an option to control whether point size is written. 2023-02-02 20:51:18 +01:00
Nikolay Sivov
06f300ec59 vkd3d-shader/hlsl: Support dot() for SM1. 2023-02-02 20:51:12 +01:00
Zebediah Figura
aa3badbd9d vkd3d-shader/hlsl: Emit a hlsl_fixme() for HLSL_IR_CALL instructions.
Avoid outputting invalid shaders.
2023-01-31 22:27:00 +01:00
Philip Rebohle
f9e7cb6345 include: Fix incorrect UpdateTileMappings declaration.
Signed-off-by: Philip Rebohle <philip.rebohle@tu-dortmund.de>
2023-01-26 21:52:39 +01:00
Nikolay Sivov
8bb023e5af vkd3d-shader/hlsl: Support exp() intrinsic. 2023-01-26 21:52:28 +01:00
Nikolay Sivov
8ca2f65bc1 vkd3d-shader/hlsl: Support exp2() intrinsic. 2023-01-26 21:52:26 +01:00
Nikolay Sivov
fa971f32bc vkd3d-shader/hlsl: Write 'exp' instructions for SM1. 2023-01-26 21:52:25 +01:00
Nikolay Sivov
335f741630 vkd3d-shader/hlsl: Add a helper to write per-component unary instructions. 2023-01-26 21:52:24 +01:00
Francisco Casas
f33ca836d7 vkd3d-shader/hlsl: Make single-component swizzles retrieve a scalar. 2023-01-26 21:52:18 +01:00
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
Zebediah Figura
653cc02f4c vkd3d-shader/hlsl: Write SM4 thread ID registers. 2023-01-25 22:47:46 +01:00