Commit Graph

597 Commits

Author SHA1 Message Date
74767beaf6 vkd3d-shader/hlsl: Absorb hlsl_ir_constant deref offsets into const_offset. 2023-10-31 21:59:18 +01:00
1520f327e5 vkd3d-shader/hlsl: Express deref->offset in whole registers.
This is required to use SM4 relative addressing, because it is limited
to whole-register granularity.
2023-10-31 21:59:16 +01:00
61a17643a2 vkd3d-shader/hlsl: Split deref-offset into a node and a constant uint.
This uint will be used for the following:

- Since SM4's relative addressing (the capability of passing a register
  as an index to another register) only has whole-register granularity,
  we will need to make the offset node express the offset in
  whole-registers and specify the register component in this uint,
  otherwise we would have to add additional / and % operations in the
  output binary.

- If, after we apply constant folding and copy propagation, we determine
  that the offset is a single constant node, we can store all the offset
  in this uint constant, and remove the offset src.

  This allows DCE to remove a good bunch of the nodes previously required
  only for the offset constants, which makes the output more liteweight
  and readable, and simplifies the implementation of relative addressing
  when writing tpf in the following patches.

In dump_deref(), we use "c" to indicate components instead of whole
registers. Since now both the offset node and the offset uint are in
components a lowered deref would look like:

    var[@42c + 2c]

But, once we express the offset node in whole registers we will remove
the "c" from the node part:

    var[@22 + 3c]
2023-10-31 21:59:14 +01:00
81be47c00b vkd3d-shader/hlsl: Introduce hlsl_deref_is_lowered() helper.
Some functions work with dereferences and need to know if they are
lowered yet.

This can be known checking if deref->offset.node is NULL or
deref->data_type is NULL. I am using the latter since it keeps working
even after the following patches that split deref->offset into
constant and variable parts.
2023-10-31 21:59:12 +01:00
e93568f290 vkd3d-shader/hlsl: Clean-up instruction block for offset node creation. 2023-10-31 21:59:11 +01:00
68c14079a6 vkd3d-shader/hlsl: Add a pass to normalize switch cases blocks.
Signed-off-by: Nikolay Sivov <nsivov@codeweavers.com>
2023-10-31 21:59:04 +01:00
c84d4e3571 vkd3d-shader/hlsl: Add a pass to remove unreachable code.
Signed-off-by: Nikolay Sivov <nsivov@codeweavers.com>
2023-10-31 21:59:03 +01:00
a4fa323e6c vkd3d-shader/hlsl: Add copy propagation logic for switches.
Signed-off-by: Nikolay Sivov <nsivov@codeweavers.com>
2023-10-31 21:59:02 +01:00
ec8dfa467f vkd3d-shader/hlsl: Add initial support for parsing 'switch' statements.
Signed-off-by: Nikolay Sivov <nsivov@codeweavers.com>
2023-10-31 21:58:57 +01:00
7960836551 vkd3d-shader/hlsl: Remove enum hlsl_error_level (clangd).
It is only used once for calling hlsl_note(), and it expects an enum
vkd3d_shader_log_level values instead.
2023-10-12 23:27:22 +02:00
8479ceedfc vkd3d-shader/hlsl: Propagate structure fields modifiers when copying shader inputs.
Signed-off-by: Nikolay Sivov <nsivov@codeweavers.com>
2023-10-09 21:58:29 +02:00
7c378cc6f9 vkd3d-shader/hlsl: Remove conditional branching when condition is a compile time constant.
Signed-off-by: Nikolay Sivov <nsivov@codeweavers.com>
2023-10-05 16:16:09 +02:00
4ab6572be7 vkd3d-shader/hlsl: Replace hlsl_type_get_regset() uses with hlsl_deref_get_regset(). 2023-10-05 16:15:37 +02:00
a214b7374b vkd3d-shader/hlsl: Avoid hlsl_type_get_regset() in allocate_register_reservations(). 2023-10-05 16:15:34 +02:00
13f62e60e1 vkd3d-shader/tpf: Remove sm4_src_register.swizzle_type. 2023-10-03 21:27:47 +02:00
fcda20a8c3 vkd3d-shader/hlsl: Use lower_ir() for lower_sqrt(). 2023-09-25 22:07:23 +02:00
496a3a2093 vkd3d-shader/hlsl: Use lower_ir() for lower_division(). 2023-09-25 22:07:22 +02:00
ecd781e809 vkd3d-shader/hlsl: Use lower_ir() for lower_int_abs(). 2023-09-25 22:07:21 +02:00
7944ee9bed vkd3d-shader/hlsl: Use lower_ir() for lower_casts_to_bool(). 2023-09-25 22:07:20 +02:00
65bf6e997c vkd3d-shader/hlsl: Use lower_ir() for more passes. 2023-09-25 22:07:18 +02:00
6d1ba83856 vkd3d-shader/hlsl: Use conditional moves for arithmetic operators instead of branching.
Signed-off-by: Nikolay Sivov <nsivov@codeweavers.com>
2023-09-22 11:06:22 +02:00
39563aa5b3 vkd3d-shader/hlsl: Lower matrix swizzles. 2023-09-13 23:10:38 +02:00
1002a6b357 vkd3d-shader/tpf: Use 'movc' to implement ternary operator.
Signed-off-by: Nikolay Sivov <nsivov@codeweavers.com>
2023-09-07 19:15:25 +02:00
63e056512d vkd3d-shader/hlsl: Introduce an hlsl_sprintf_alloc() helper. 2023-08-30 22:48:55 +02:00
926575a6f3 vkd3d-shader/hlsl: Force sm1 inputs to be 4-component only for vertex shaders.
Pixel shaders still have an appropriate writemask.
2023-08-24 21:43:44 +02:00