- Wine-Commits - list.winehq.org

Conor McCarthy : vkd3d-shader/dxil: Handle inverse trigonometric functions in sm6_parser_emit_dx_unary().
by Alexandre Julliard Feb. 6, 2024

Feb. 6, 2024

Module: vkd3d Branch: master Commit: 19d23fa696523ec4f46c5cde75009a5d76f533b3 URL: https://gitlab.winehq.org/wine/vkd3d/-/commit/19d23fa696523ec4f46c5cde75009… Author: Conor McCarthy <cmccarthy(a)codeweavers.com> Date: Mon Jan 22 15:26:16 2024 +1000 vkd3d-shader/dxil: Handle inverse trigonometric functions in sm6_parser_emit_dx_unary(). --- libs/vkd3d-shader/d3d_asm.c | 3 +++ libs/vkd3d-shader/dxil.c | 12 ++++++++++++ libs/vkd3d-shader/vkd3d_shader_private.h | 3 +++ 3 files changed, 18 insertions(+) diff --git a/libs/vkd3d-shader/d3d_asm.c b/libs/vkd3d-shader/d3d_asm.c index dd96b7fa..3ae736a8 100644 --- a/libs/vkd3d-shader/d3d_asm.c +++ b/libs/vkd3d-shader/d3d_asm.c @@ -30,8 +30,11 @@ static const char * const shader_opcode_names[] = { [VKD3DSIH_ABS ] = "abs", + [VKD3DSIH_ACOS ] = "acos", [VKD3DSIH_ADD ] = "add", [VKD3DSIH_AND ] = "and", + [VKD3DSIH_ASIN ] = "asin", + [VKD3DSIH_ATAN ] = "atan", [VKD3DSIH_ATOMIC_AND ] = "atomic_and", [VKD3DSIH_ATOMIC_CMP_STORE ] = "atomic_cmp_store", [VKD3DSIH_ATOMIC_IADD ] = "atomic_iadd", diff --git a/libs/vkd3d-shader/dxil.c b/libs/vkd3d-shader/dxil.c index 291012d1..e718ed84 100644 --- a/libs/vkd3d-shader/dxil.c +++ b/libs/vkd3d-shader/dxil.c @@ -339,6 +339,9 @@ enum dx_intrinsic_opcode DX_COS = 12, DX_SIN = 13, DX_TAN = 14, + DX_ACOS = 15, + DX_ASIN = 16, + DX_ATAN = 17, DX_EXP = 21, DX_FRC = 22, DX_LOG = 23, @@ -3625,6 +3628,12 @@ static enum vkd3d_shader_opcode map_dx_unary_op(enum dx_intrinsic_opcode op) return VKD3DSIH_ISFINITE; case DX_TAN: return VKD3DSIH_TAN; + case DX_ACOS: + return VKD3DSIH_ACOS; + case DX_ASIN: + return VKD3DSIH_ASIN; + case DX_ATAN: + return VKD3DSIH_ATAN; case DX_EXP: return VKD3DSIH_EXP; case DX_FRC: @@ -4127,6 +4136,9 @@ struct sm6_dx_opcode_info */ static const struct sm6_dx_opcode_info sm6_dx_op_table[] = { + [DX_ACOS ] = {"g", "R", sm6_parser_emit_dx_unary}, + [DX_ASIN ] = {"g", "R", sm6_parser_emit_dx_unary}, + [DX_ATAN ] = {"g", "R", sm6_parser_emit_dx_unary}, [DX_BFREV ] = {"m", "R", sm6_parser_emit_dx_unary}, [DX_BUFFER_LOAD ] = {"o", "Hii", sm6_parser_emit_dx_buffer_load}, [DX_CBUFFER_LOAD_LEGACY ] = {"o", "Hi", sm6_parser_emit_dx_cbuffer_load}, diff --git a/libs/vkd3d-shader/vkd3d_shader_private.h b/libs/vkd3d-shader/vkd3d_shader_private.h index 51320ad0..3d043c24 100644 --- a/libs/vkd3d-shader/vkd3d_shader_private.h +++ b/libs/vkd3d-shader/vkd3d_shader_private.h @@ -225,8 +225,11 @@ enum vkd3d_shader_error enum vkd3d_shader_opcode { VKD3DSIH_ABS, + VKD3DSIH_ACOS, VKD3DSIH_ADD, VKD3DSIH_AND, + VKD3DSIH_ASIN, + VKD3DSIH_ATAN, VKD3DSIH_ATOMIC_AND, VKD3DSIH_ATOMIC_CMP_STORE, VKD3DSIH_ATOMIC_IADD,

1 0

Henri Verbeet : vkd3d-common: Get rid of InterlockedIncrement().
by Alexandre Julliard Feb. 6, 2024

Feb. 6, 2024

Module: vkd3d Branch: master Commit: c64921e79b5e26bf0e39cf2c188979db85f8d3b3 URL: https://gitlab.winehq.org/wine/vkd3d/-/commit/c64921e79b5e26bf0e39cf2c18897… Author: Henri Verbeet <hverbeet(a)codeweavers.com> Date: Tue Jan 30 12:54:54 2024 +0100 vkd3d-common: Get rid of InterlockedIncrement(). --- include/private/vkd3d_common.h | 5 ----- include/private/vkd3d_test.h | 40 +++++++++++++++++++--------------------- tests/vkd3d_api.c | 6 +++--- 3 files changed, 22 insertions(+), 29 deletions(-) diff --git a/include/private/vkd3d_common.h b/include/private/vkd3d_common.h index 01e219b3..6f878cce 100644 --- a/include/private/vkd3d_common.h +++ b/include/private/vkd3d_common.h @@ -305,11 +305,6 @@ static inline uint32_t vkd3d_atomic_increment_u32(uint32_t volatile *x) } #ifndef _WIN32 -static inline LONG InterlockedIncrement(LONG volatile *x) -{ - return vkd3d_atomic_increment_u32((uint32_t *)x); -} - static inline LONG InterlockedDecrement(LONG volatile *x) { return vkd3d_atomic_decrement_u32((uint32_t *)x); diff --git a/include/private/vkd3d_test.h b/include/private/vkd3d_test.h index 081443c4..a337ac07 100644 --- a/include/private/vkd3d_test.h +++ b/include/private/vkd3d_test.h @@ -101,12 +101,12 @@ static void vkd3d_test_end_todo(void); struct vkd3d_test_state { - LONG success_count; - LONG failure_count; - LONG skip_count; - LONG todo_count; - LONG todo_success_count; - LONG bug_count; + unsigned int success_count; + unsigned int failure_count; + unsigned int skip_count; + unsigned int todo_count; + unsigned int todo_success_count; + unsigned int bug_count; unsigned int debug_level; @@ -150,13 +150,13 @@ vkd3d_test_check_assert_that(unsigned int line, bool result, const char *fmt, va { if (result) { - InterlockedIncrement(&vkd3d_test_state.success_count); + vkd3d_atomic_increment_u32(&vkd3d_test_state.success_count); if (vkd3d_test_state.debug_level > 1) vkd3d_test_printf(line, "Test succeeded.\n"); } else { - InterlockedIncrement(&vkd3d_test_state.failure_count); + vkd3d_atomic_increment_u32(&vkd3d_test_state.failure_count); vkd3d_test_printf(line, "Test failed: "); vprintf(fmt, args); } @@ -180,7 +180,7 @@ vkd3d_test_check_ok(unsigned int line, bool result, const char *fmt, va_list arg if (is_bug && vkd3d_test_state.bug_enabled) { - InterlockedIncrement(&vkd3d_test_state.bug_count); + vkd3d_atomic_increment_u32(&vkd3d_test_state.bug_count); if (is_todo) result = !result; if (result) @@ -193,12 +193,12 @@ vkd3d_test_check_ok(unsigned int line, bool result, const char *fmt, va_list arg { if (result) { - InterlockedIncrement(&vkd3d_test_state.todo_success_count); + vkd3d_atomic_increment_u32(&vkd3d_test_state.todo_success_count); vkd3d_test_printf(line, "Todo succeeded: "); } else { - InterlockedIncrement(&vkd3d_test_state.todo_count); + vkd3d_atomic_increment_u32(&vkd3d_test_state.todo_count); vkd3d_test_printf(line, "Todo: "); } vprintf(fmt, args); @@ -227,7 +227,7 @@ vkd3d_test_skip(unsigned int line, const char *fmt, ...) vkd3d_test_printf(line, "Test skipped: "); vprintf(fmt, args); va_end(args); - InterlockedIncrement(&vkd3d_test_state.skip_count); + vkd3d_atomic_increment_u32(&vkd3d_test_state.skip_count); } static void VKD3D_PRINTF_FUNC(2, 3) VKD3D_UNUSED @@ -293,16 +293,14 @@ int main(int argc, char **argv) vkd3d_test_main(argc, argv); - printf("%s: %lu tests executed (%lu failures, %lu skipped, %lu todo, %lu bugs).\n", + printf("%s: %u tests executed (%u failures, %u skipped, %u todo, %u bugs).\n", vkd3d_test_name, - (unsigned long)(vkd3d_test_state.success_count - + vkd3d_test_state.failure_count + vkd3d_test_state.todo_count - + vkd3d_test_state.todo_success_count), - (unsigned long)(vkd3d_test_state.failure_count - + vkd3d_test_state.todo_success_count), - (unsigned long)vkd3d_test_state.skip_count, - (unsigned long)vkd3d_test_state.todo_count, - (unsigned long)vkd3d_test_state.bug_count); + vkd3d_test_state.success_count + vkd3d_test_state.failure_count + + vkd3d_test_state.todo_count + vkd3d_test_state.todo_success_count, + vkd3d_test_state.failure_count + vkd3d_test_state.todo_success_count, + vkd3d_test_state.skip_count, + vkd3d_test_state.todo_count, + vkd3d_test_state.bug_count); if (test_platform) free(test_platform); diff --git a/tests/vkd3d_api.c b/tests/vkd3d_api.c index 5a585fec..29948783 100644 --- a/tests/vkd3d_api.c +++ b/tests/vkd3d_api.c @@ -625,7 +625,7 @@ static void test_adapter_luid(void) struct parent { IUnknown IUnknown_iface; - LONG refcount; + unsigned int refcount; }; static struct parent *parent_from_IUnknown(IUnknown *iface) @@ -651,14 +651,14 @@ static ULONG STDMETHODCALLTYPE parent_AddRef(IUnknown *iface) { struct parent *parent = parent_from_IUnknown(iface); - return InterlockedIncrement(&parent->refcount); + return vkd3d_atomic_increment_u32(&parent->refcount); } static ULONG STDMETHODCALLTYPE parent_Release(IUnknown *iface) { struct parent *parent = parent_from_IUnknown(iface); - return InterlockedDecrement(&parent->refcount); + return InterlockedDecrement((LONG *)&parent->refcount); } static const struct IUnknownVtbl parent_vtbl =

1 0

Henri Verbeet : vkd3d-common: Get rid of InterlockedDecrement().
by Alexandre Julliard Feb. 6, 2024

Feb. 6, 2024

Module: vkd3d Branch: master Commit: 532e902a5630c8ac606c06bd25b1bcbab14f9dc3 URL: https://gitlab.winehq.org/wine/vkd3d/-/commit/532e902a5630c8ac606c06bd25b1b… Author: Henri Verbeet <hverbeet(a)codeweavers.com> Date: Tue Jan 30 12:59:02 2024 +0100 vkd3d-common: Get rid of InterlockedDecrement(). --- include/private/vkd3d_common.h | 7 ------- tests/vkd3d_api.c | 2 +- 2 files changed, 1 insertion(+), 8 deletions(-) diff --git a/include/private/vkd3d_common.h b/include/private/vkd3d_common.h index 6f878cce..8d6e7b37 100644 --- a/include/private/vkd3d_common.h +++ b/include/private/vkd3d_common.h @@ -304,13 +304,6 @@ static inline uint32_t vkd3d_atomic_increment_u32(uint32_t volatile *x) return vkd3d_atomic_add_fetch_u32(x, 1); } -#ifndef _WIN32 -static inline LONG InterlockedDecrement(LONG volatile *x) -{ - return vkd3d_atomic_decrement_u32((uint32_t *)x); -} -#endif /* _WIN32 */ - static inline void vkd3d_parse_version(const char *version, int *major, int *minor) { *major = atoi(version); diff --git a/tests/vkd3d_api.c b/tests/vkd3d_api.c index 29948783..62e30b47 100644 --- a/tests/vkd3d_api.c +++ b/tests/vkd3d_api.c @@ -658,7 +658,7 @@ static ULONG STDMETHODCALLTYPE parent_Release(IUnknown *iface) { struct parent *parent = parent_from_IUnknown(iface); - return InterlockedDecrement((LONG *)&parent->refcount); + return vkd3d_atomic_decrement_u32(&parent->refcount); } static const struct IUnknownVtbl parent_vtbl =

1 0

Petrichor Park : tests/shader-runner: Add tests for acos and asin trig intrinsics.
by Alexandre Julliard Feb. 6, 2024

Feb. 6, 2024

Module: vkd3d Branch: master Commit: 92044d5e4114ae65b77b232fba442f347f21069d URL: https://gitlab.winehq.org/wine/vkd3d/-/commit/92044d5e4114ae65b77b232fba442… Author: Petrichor Park <ppark(a)codeweavers.com> Date: Wed Sep 6 15:14:12 2023 -0500 tests/shader-runner: Add tests for acos and asin trig intrinsics. Extracted by Conor McCarthy from an HLSL patch, and modified to include SM 6 variations. --- Makefile.am | 1 + tests/hlsl/inverse-trig.shader_test | 92 +++++++++++++++++++++++++++++++++++++ 2 files changed, 93 insertions(+) diff --git a/Makefile.am b/Makefile.am index 90e7dcfc..1687d9a0 100644 --- a/Makefile.am +++ b/Makefile.am @@ -121,6 +121,7 @@ vkd3d_shader_tests = \ tests/hlsl/initializer-struct.shader_test \ tests/hlsl/intrinsic-override.shader_test \ tests/hlsl/invalid.shader_test \ + tests/hlsl/inverse-trig.shader_test \ tests/hlsl/is-front-face.shader_test \ tests/hlsl/ldexp.shader_test \ tests/hlsl/length.shader_test \ diff --git a/tests/hlsl/inverse-trig.shader_test b/tests/hlsl/inverse-trig.shader_test new file mode 100644 index 00000000..6077b4b5 --- /dev/null +++ b/tests/hlsl/inverse-trig.shader_test @@ -0,0 +1,92 @@ +% TPF does not define instructions for inverse trig; these intrinsics are +% decomposed into other instructions. FXC emits code which may vary wrt other +% implementations. DXIL defines intrinsics for inverse trig, to be implemented +% by the backend. + +[pixel shader todo] +uniform float4 a; + +float4 main() : sv_target +{ + return float4(acos(a.x), 0.0, 0.0, 0.0); +} + +[test] +uniform 0 float4 -1.0 0.0 0.0 0.0 +todo draw quad +probe all rgba (3.14159274, 0.0, 0.0, 0.0) 128 + +uniform 0 float4 -0.5 0.0 0.0 0.0 +todo draw quad +probe all rgba (2.094441441, 0.0, 0.0, 0.0) 256 + +uniform 0 float4 0.0 0.0 0.0 0.0 +todo draw quad +probe all rgba (1.57072878, 0.0, 0.0, 0.0) 1024 + +uniform 0 float4 0.5 0.0 0.0 0.0 +todo draw quad +probe all rgba (1.04715133, 0.0, 0.0, 0.0) 512 + +uniform 0 float4 1.0 0.0 0.0 0.0 +todo draw quad +probe all rgba (0.0, 0.0, 0.0, 0.0) 128 + +[pixel shader todo] +uniform float4 a; + +float4 main() : sv_target +{ + float4 result = float4(asin(a.x), 0.0, 0.0, 0.0); + // Quantize to cover implementation variations, which are quite large for asin(). + return round(result * 20000.0); +} + +[test] +uniform 0 float4 -1.0 0.0 0.0 0.0 +todo draw quad +probe all rgba (-31416.0, 0.0, 0.0, 0.0) + +[require] +shader model < 6.0 + +[test] +uniform 0 float4 -0.5 0.0 0.0 0.0 +todo draw quad +probe all rgba (-10473.0, 0.0, 0.0, 0.0) + +uniform 0 float4 0.0 0.0 0.0 0.0 +todo draw quad +probe all rgba (1.0, 0.0, 0.0, 0.0) + +uniform 0 float4 0.5 0.0 0.0 0.0 +todo draw quad +probe all rgba (10473.0, 0.0, 0.0, 0.0) + +[require] +shader model >= 6.0 + +% We implement SM 6.0 inverse trig instructions using the native equivalents +% available in the backend. The values below are from the AMD Windows drivers, +% which are very close to those from Ubuntu's calculator app. Results from +% RADV are a bit lower than these, hence the large max ulp difference. +[test] +uniform 0 float4 -0.5 0.0 0.0 0.0 +todo draw quad +probe all rgba (-10472.0, 0.0, 0.0, 0.0) 4096 + +uniform 0 float4 0.0 0.0 0.0 0.0 +todo draw quad +probe all rgba (0.0, 0.0, 0.0, 0.0) + +uniform 0 float4 0.5 0.0 0.0 0.0 +todo draw quad +probe all rgba (10472.0, 0.0, 0.0, 0.0) 4096 + +[require] +% reset requirements + +[test] +uniform 0 float4 1.0 0.0 0.0 0.0 +todo draw quad +probe all rgba (31416.0, 0.0, 0.0, 0.0)

1 0

Henri Verbeet : vkd3d: Introduce a debug helper for GPU descriptor handles.
by Alexandre Julliard Feb. 6, 2024

Feb. 6, 2024

Module: vkd3d Branch: master Commit: 51cdddb96141587a99e17237370194adc14290fe URL: https://gitlab.winehq.org/wine/vkd3d/-/commit/51cdddb96141587a99e1723737019… Author: Henri Verbeet <hverbeet(a)codeweavers.com> Date: Tue Jan 30 15:12:56 2024 +0100 vkd3d: Introduce a debug helper for GPU descriptor handles. --- libs/vkd3d/command.c | 16 ++++++++-------- libs/vkd3d/utils.c | 5 +++++ libs/vkd3d/vkd3d_private.h | 1 + 3 files changed, 14 insertions(+), 8 deletions(-) diff --git a/libs/vkd3d/command.c b/libs/vkd3d/command.c index f9940b3d..609300d3 100644 --- a/libs/vkd3d/command.c +++ b/libs/vkd3d/command.c @@ -4499,8 +4499,8 @@ static void STDMETHODCALLTYPE d3d12_command_list_SetComputeRootDescriptorTable(I { struct d3d12_command_list *list = impl_from_ID3D12GraphicsCommandList5(iface); - TRACE("iface %p, root_parameter_index %u, base_descriptor %#"PRIx64".\n", - iface, root_parameter_index, base_descriptor.ptr); + TRACE("iface %p, root_parameter_index %u, base_descriptor %s.\n", + iface, root_parameter_index, debug_gpu_handle(base_descriptor)); d3d12_command_list_set_descriptor_table(list, VKD3D_PIPELINE_BIND_POINT_COMPUTE, root_parameter_index, base_descriptor); @@ -4511,8 +4511,8 @@ static void STDMETHODCALLTYPE d3d12_command_list_SetGraphicsRootDescriptorTable( { struct d3d12_command_list *list = impl_from_ID3D12GraphicsCommandList5(iface); - TRACE("iface %p, root_parameter_index %u, base_descriptor %#"PRIx64".\n", - iface, root_parameter_index, base_descriptor.ptr); + TRACE("iface %p, root_parameter_index %u, base_descriptor %s.\n", + iface, root_parameter_index, debug_gpu_handle(base_descriptor)); d3d12_command_list_set_descriptor_table(list, VKD3D_PIPELINE_BIND_POINT_GRAPHICS, root_parameter_index, base_descriptor); @@ -5432,8 +5432,8 @@ static void STDMETHODCALLTYPE d3d12_command_list_ClearUnorderedAccessViewUint(ID struct d3d12_resource *resource_impl; VkClearColorValue colour; - TRACE("iface %p, gpu_handle %#"PRIx64", cpu_handle %s, resource %p, values %p, rect_count %u, rects %p.\n", - iface, gpu_handle.ptr, debug_cpu_handle(cpu_handle), resource, values, rect_count, rects); + TRACE("iface %p, gpu_handle %s, cpu_handle %s, resource %p, values %p, rect_count %u, rects %p.\n", + iface, debug_gpu_handle(gpu_handle), debug_cpu_handle(cpu_handle), resource, values, rect_count, rects); resource_impl = unsafe_impl_from_ID3D12Resource(resource); if (!(descriptor = d3d12_desc_from_cpu_handle(cpu_handle)->s.u.view)) @@ -5496,8 +5496,8 @@ static void STDMETHODCALLTYPE d3d12_command_list_ClearUnorderedAccessViewFloat(I VkClearColorValue colour; struct vkd3d_view *view; - TRACE("iface %p, gpu_handle %#"PRIx64", cpu_handle %s, resource %p, values %p, rect_count %u, rects %p.\n", - iface, gpu_handle.ptr, debug_cpu_handle(cpu_handle), resource, values, rect_count, rects); + TRACE("iface %p, gpu_handle %s, cpu_handle %s, resource %p, values %p, rect_count %u, rects %p.\n", + iface, debug_gpu_handle(gpu_handle), debug_cpu_handle(cpu_handle), resource, values, rect_count, rects); resource_impl = unsafe_impl_from_ID3D12Resource(resource); if (!(view = d3d12_desc_from_cpu_handle(cpu_handle)->s.u.view)) diff --git a/libs/vkd3d/utils.c b/libs/vkd3d/utils.c index 28913a83..ac79ae5d 100644 --- a/libs/vkd3d/utils.c +++ b/libs/vkd3d/utils.c @@ -678,6 +678,11 @@ const char *debug_d3d12_shader_component_mapping(unsigned int mapping) debug_d3d12_shader_component(D3D12_DECODE_SHADER_4_COMPONENT_MAPPING(3, mapping))); } +const char *debug_gpu_handle(D3D12_GPU_DESCRIPTOR_HANDLE handle) +{ + return vkd3d_dbg_sprintf("{%#"PRIx64"}", handle.ptr); +} + const char *debug_vk_extent_3d(VkExtent3D extent) { return vkd3d_dbg_sprintf("(%u, %u, %u)", diff --git a/libs/vkd3d/vkd3d_private.h b/libs/vkd3d/vkd3d_private.h index bb78ba77..5955e616 100644 --- a/libs/vkd3d/vkd3d_private.h +++ b/libs/vkd3d/vkd3d_private.h @@ -1963,6 +1963,7 @@ HRESULT return_interface(void *iface, REFIID iface_iid, REFIID requested_iid, vo const char *debug_cpu_handle(D3D12_CPU_DESCRIPTOR_HANDLE handle); const char *debug_d3d12_box(const D3D12_BOX *box); const char *debug_d3d12_shader_component_mapping(unsigned int mapping); +const char *debug_gpu_handle(D3D12_GPU_DESCRIPTOR_HANDLE handle); const char *debug_vk_extent_3d(VkExtent3D extent); const char *debug_vk_memory_heap_flags(VkMemoryHeapFlags flags); const char *debug_vk_memory_property_flags(VkMemoryPropertyFlags flags);

1 0

Petrichor Park : tests/shader-runner: Add tests for atan and atan2 trig intrinsics.
by Alexandre Julliard Feb. 6, 2024

Feb. 6, 2024

Module: vkd3d Branch: master Commit: e6746cbbdc649d8322c230ae6055cc5a44966551 URL: https://gitlab.winehq.org/wine/vkd3d/-/commit/e6746cbbdc649d8322c230ae6055c… Author: Petrichor Park <ppark(a)codeweavers.com> Date: Thu Dec 7 11:07:42 2023 -0600 tests/shader-runner: Add tests for atan and atan2 trig intrinsics. Extracted by Conor McCarthy from an HLSL patch, with ulp values doubled in some cases to cover SM 6 results. --- tests/hlsl/inverse-trig.shader_test | 104 ++++++++++++++++++++++++++++++++++++ 1 file changed, 104 insertions(+) diff --git a/tests/hlsl/inverse-trig.shader_test b/tests/hlsl/inverse-trig.shader_test index 6077b4b5..dd069fef 100644 --- a/tests/hlsl/inverse-trig.shader_test +++ b/tests/hlsl/inverse-trig.shader_test @@ -90,3 +90,107 @@ probe all rgba (10472.0, 0.0, 0.0, 0.0) 4096 uniform 0 float4 1.0 0.0 0.0 0.0 todo draw quad probe all rgba (31416.0, 0.0, 0.0, 0.0) + + +[pixel shader todo] +uniform float4 a; + +float4 main() : sv_target +{ + return float4(atan(a.x), 0.0, 0.0, 0.0); +} + +[test] +uniform 0 float4 -1.0 0.0 0.0 0.0 +todo draw quad +probe all rgba (-0.785409629, 0.0, 0.0, 0.0) 512 + +uniform 0 float4 -0.5 0.0 0.0 0.0 +todo draw quad +probe all rgba (-0.4636476, 0.0, 0.0, 0.0) 256 + +uniform 0 float4 0.0 0.0 0.0 0.0 +todo draw quad +probe all rgba (0.0, 0.0, 0.0, 0.0) 256 + +uniform 0 float4 0.5 0.0 0.0 0.0 +todo draw quad +probe all rgba (0.4636476, 0.0, 0.0, 0.0) 256 + +uniform 0 float4 1.0 0.0 0.0 0.0 +todo draw quad +probe all rgba (0.785409629, 0.0, 0.0, 0.0) 512 + +[pixel shader todo] +uniform float4 a; + +float4 main() : sv_target +{ + // The argument order is (y,x), and test case inputs are (y,x) also. + return float4(atan2(a.x, a.y), 0.0, 0.0, 0.0); +} + +[test] +% Non-degenerate cases +uniform 0 float4 1.0 1.0 0.0 0.0 +todo draw quad +probe all rgba (0.785385, 0.0, 0.0, 0.0) 512 + +uniform 0 float4 5.0 -5.0 0.0 0.0 +todo draw quad +probe all rgba (2.356194, 0.0, 0.0, 0.0) 256 + +uniform 0 float4 -3.0 -3.0 0.0 0.0 +todo draw quad +probe all rgba (-2.356194, 0.0, 0.0, 0.0) 256 + +uniform 0 float4 1.0 0.0 0.0 0.0 +todo draw quad +probe all rgba (1.570796, 0.0, 0.0, 0.0) 256 + +uniform 0 float4 -1.0 0.0 0.0 0.0 +todo draw quad +probe all rgba (-1.570796, 0.0, 0.0, 0.0) 256 + +uniform 0 float4 0.0 1.0 0.0 0.0 +todo draw quad +probe all rgba (0.0, 0.0, 0.0, 0.0) 256 + +uniform 0 float4 0.0 -1.0 0.0 0.0 +todo draw quad +probe all rgba (3.1415927, 0.0, 0.0, 0.0) 256 + +% Degenerate cases +uniform 0 float4 0.00001 0.00002 0.0 0.0 +todo draw quad +probe all rgba (0.463647, 0.0, 0.0, 0.0) 256 + +uniform 0 float4 0.00001 -0.00002 0.0 0.0 +todo draw quad +probe all rgba (2.677945, 0.0, 0.0, 0.0) 256 + +uniform 0 float4 -0.00001 100000.0 0.0 0.0 +todo draw quad +probe all rgba (-0.000000000099986595, 0.0, 0.0, 0.0) 2048 + +uniform 0 float4 10000000.0 0.00000001 0.0 0.0 +todo draw quad +probe all rgba (1.570796, 0.0, 0.0, 0.0) 256 + +% Negative zero behavior should be to treat it the +% same as normal zero. +uniform 0 float4 1000000000.0 0.0 0.0 0.0 +todo draw quad +probe all rgba (1.570796, 0.0, 0.0, 0.0) 256 + +uniform 0 float4 1000000000.0 -0.0 0.0 0.0 +todo draw quad +probe all rgba (1.570796, 0.0, 0.0, 0.0) 256 + +uniform 0 float4 0.0 -1.0 0.0 0.0 +todo draw quad +probe all rgba (3.1415927, 0.0, 0.0, 0.0) 256 + +uniform 0 float4 -0.0 -1.0 0.0 0.0 +todo draw quad +probe all rgba (3.1415927, 0.0, 0.0, 0.0) 256

1 0

Henri Verbeet : vkd3d-utils: Use PRIuPTR for SIZE_T variables in debug traces.
by Alexandre Julliard Feb. 6, 2024

Feb. 6, 2024

Module: vkd3d Branch: master Commit: 2f9906ee9029e5be22be9a8608817676a8c2a8e1 URL: https://gitlab.winehq.org/wine/vkd3d/-/commit/2f9906ee9029e5be22be9a8608817… Author: Henri Verbeet <hverbeet(a)codeweavers.com> Date: Tue Jan 30 16:48:55 2024 +0100 vkd3d-utils: Use PRIuPTR for SIZE_T variables in debug traces. --- libs/vkd3d-utils/reflection.c | 3 ++- libs/vkd3d-utils/vkd3d_utils_main.c | 49 +++++++++++++++++----------------- libs/vkd3d-utils/vkd3d_utils_private.h | 2 ++ 3 files changed, 29 insertions(+), 25 deletions(-) diff --git a/libs/vkd3d-utils/reflection.c b/libs/vkd3d-utils/reflection.c index 2f604864..1973bda7 100644 --- a/libs/vkd3d-utils/reflection.c +++ b/libs/vkd3d-utils/reflection.c @@ -309,7 +309,8 @@ HRESULT WINAPI D3DReflect(const void *data, SIZE_T data_size, REFIID iid, void * struct d3d12_reflection *object; HRESULT hr; - TRACE("data %p, data_size %lu, iid %s, reflection %p.\n", data, data_size, debugstr_guid(iid), reflection); + TRACE("data %p, data_size %"PRIuPTR", iid %s, reflection %p.\n", + data, (uintptr_t)data_size, debugstr_guid(iid), reflection); if (!IsEqualGUID(iid, &IID_ID3D12ShaderReflection)) { diff --git a/libs/vkd3d-utils/vkd3d_utils_main.c b/libs/vkd3d-utils/vkd3d_utils_main.c index f847b250..ae197682 100644 --- a/libs/vkd3d-utils/vkd3d_utils_main.c +++ b/libs/vkd3d-utils/vkd3d_utils_main.c @@ -129,8 +129,8 @@ VKD3D_UTILS_API HRESULT WINAPI D3D12CreateDevice(IUnknown *adapter, HRESULT WINAPI D3D12CreateRootSignatureDeserializer(const void *data, SIZE_T data_size, REFIID iid, void **deserializer) { - TRACE("data %p, data_size %lu, iid %s, deserializer %p.\n", - data, data_size, debugstr_guid(iid), deserializer); + TRACE("data %p, data_size %"PRIuPTR", iid %s, deserializer %p.\n", + data, (uintptr_t)data_size, debugstr_guid(iid), deserializer); return vkd3d_create_root_signature_deserializer(data, data_size, iid, deserializer); } @@ -138,8 +138,8 @@ HRESULT WINAPI D3D12CreateRootSignatureDeserializer(const void *data, SIZE_T dat HRESULT WINAPI D3D12CreateVersionedRootSignatureDeserializer(const void *data, SIZE_T data_size, REFIID iid,void **deserializer) { - TRACE("data %p, data_size %lu, iid %s, deserializer %p.\n", - data, data_size, debugstr_guid(iid), deserializer); + TRACE("data %p, data_size %"PRIuPTR", iid %s, deserializer %p.\n", + data, (uintptr_t)data_size, debugstr_guid(iid), deserializer); return vkd3d_create_versioned_root_signature_deserializer(data, data_size, iid, deserializer); } @@ -253,12 +253,12 @@ HRESULT WINAPI D3DCompile2(const void *data, SIZE_T data_size, const char *filen HRESULT hr; int ret; - TRACE("data %p, data_size %lu, filename %s, macros %p, include %p, entry_point %s, " + TRACE("data %p, data_size %"PRIuPTR", filename %s, macros %p, include %p, entry_point %s, " "profile %s, flags %#x, effect_flags %#x, secondary_flags %#x, secondary_data %p, " - "secondary_data_size %lu, shader_blob %p, messages_blob %p.\n", - data, data_size, debugstr_a(filename), macros, include, debugstr_a(entry_point), + "secondary_data_size %"PRIuPTR", shader_blob %p, messages_blob %p.\n", + data, (uintptr_t)data_size, debugstr_a(filename), macros, include, debugstr_a(entry_point), debugstr_a(profile), flags, effect_flags, secondary_flags, secondary_data, - secondary_data_size, shader_blob, messages_blob); + (uintptr_t)secondary_data_size, shader_blob, messages_blob); if (flags & ~(D3DCOMPILE_DEBUG | D3DCOMPILE_PACK_MATRIX_ROW_MAJOR | D3DCOMPILE_PACK_MATRIX_COLUMN_MAJOR)) FIXME("Ignoring flags %#x.\n", flags); @@ -374,9 +374,9 @@ HRESULT WINAPI D3DCompile(const void *data, SIZE_T data_size, const char *filena const D3D_SHADER_MACRO *macros, ID3DInclude *include, const char *entrypoint, const char *profile, UINT flags, UINT effect_flags, ID3DBlob **shader, ID3DBlob **error_messages) { - TRACE("data %p, data_size %lu, filename %s, macros %p, include %p, entrypoint %s, " + TRACE("data %p, data_size %"PRIuPTR", filename %s, macros %p, include %p, entrypoint %s, " "profile %s, flags %#x, effect_flags %#x, shader %p, error_messages %p.\n", - data, data_size, debugstr_a(filename), macros, include, debugstr_a(entrypoint), + data, (uintptr_t)data_size, debugstr_a(filename), macros, include, debugstr_a(entrypoint), debugstr_a(profile), flags, effect_flags, shader, error_messages); return D3DCompile2(data, data_size, filename, macros, include, entrypoint, profile, flags, @@ -400,8 +400,8 @@ HRESULT WINAPI D3DPreprocess(const void *data, SIZE_T size, const char *filename {VKD3D_SHADER_COMPILE_OPTION_API_VERSION, VKD3D_SHADER_API_VERSION_1_10}, }; - TRACE("data %p, size %lu, filename %s, macros %p, include %p, preprocessed_blob %p, messages_blob %p.\n", - data, size, debugstr_a(filename), macros, include, preprocessed_blob, messages_blob); + TRACE("data %p, size %"PRIuPTR", filename %s, macros %p, include %p, preprocessed_blob %p, messages_blob %p.\n", + data, (uintptr_t)size, debugstr_a(filename), macros, include, preprocessed_blob, messages_blob); if (messages_blob) *messages_blob = NULL; @@ -606,7 +606,7 @@ HRESULT WINAPI D3DCreateBlob(SIZE_T data_size, ID3DBlob **blob) HRESULT hr; void *data; - TRACE("data_size %lu, blob %p.\n", data_size, blob); + TRACE("data_size %"PRIuPTR", blob %p.\n", (uintptr_t)data_size, blob); if (!blob) { @@ -699,7 +699,8 @@ static HRESULT get_blob_part(const void *data, SIZE_T data_size, if (!data || !data_size || flags || !blob) { - WARN("Invalid arguments: data %p, data_size %lu, flags %#x, blob %p.\n", data, data_size, flags, blob); + WARN("Invalid arguments: data %p, data_size %"PRIuPTR", flags %#x, blob %p.\n", + data, (uintptr_t)data_size, flags, blob); return D3DERR_INVALIDCALL; } @@ -795,36 +796,36 @@ done: HRESULT WINAPI D3DGetBlobPart(const void *data, SIZE_T data_size, D3D_BLOB_PART part, UINT flags, ID3DBlob **blob) { - TRACE("data %p, data_size %lu, part %s, flags %#x, blob %p.\n", data, - data_size, debug_d3d_blob_part(part), flags, blob); + TRACE("data %p, data_size %"PRIuPTR", part %s, flags %#x, blob %p.\n", + data, (uintptr_t)data_size, debug_d3d_blob_part(part), flags, blob); return get_blob_part(data, data_size, part, flags, blob); } HRESULT WINAPI D3DGetDebugInfo(const void *data, SIZE_T data_size, ID3DBlob **blob) { - TRACE("data %p, data_size %lu, blob %p.\n", data, data_size, blob); + TRACE("data %p, data_size %"PRIuPTR", blob %p.\n", data, (uintptr_t)data_size, blob); return get_blob_part(data, data_size, D3D_BLOB_DEBUG_INFO, 0, blob); } HRESULT WINAPI D3DGetInputAndOutputSignatureBlob(const void *data, SIZE_T data_size, ID3DBlob **blob) { - TRACE("data %p, data_size %lu, blob %p.\n", data, data_size, blob); + TRACE("data %p, data_size %"PRIuPTR", blob %p.\n", data, (uintptr_t)data_size, blob); return get_blob_part(data, data_size, D3D_BLOB_INPUT_AND_OUTPUT_SIGNATURE_BLOB, 0, blob); } HRESULT WINAPI D3DGetInputSignatureBlob(const void *data, SIZE_T data_size, ID3DBlob **blob) { - TRACE("data %p, data_size %lu, blob %p.\n", data, data_size, blob); + TRACE("data %p, data_size %"PRIuPTR", blob %p.\n", data, (uintptr_t)data_size, blob); return get_blob_part(data, data_size, D3D_BLOB_INPUT_SIGNATURE_BLOB, 0, blob); } HRESULT WINAPI D3DGetOutputSignatureBlob(const void *data, SIZE_T data_size, ID3DBlob **blob) { - TRACE("data %p, data_size %lu, blob %p.\n", data, data_size, blob); + TRACE("data %p, data_size %"PRIuPTR", blob %p.\n", data, (uintptr_t)data_size, blob); return get_blob_part(data, data_size, D3D_BLOB_OUTPUT_SIGNATURE_BLOB, 0, blob); } @@ -865,7 +866,7 @@ HRESULT WINAPI D3DStripShader(const void *data, SIZE_T data_size, UINT flags, ID HRESULT hr; int ret; - TRACE("data %p, data_size %lu, flags %#x, blob %p.\n", data, data_size, flags, blob); + TRACE("data %p, data_size %"PRIuPTR", flags %#x, blob %p.\n", data, (uintptr_t)data_size, flags, blob); if (!blob) { @@ -875,7 +876,7 @@ HRESULT WINAPI D3DStripShader(const void *data, SIZE_T data_size, UINT flags, ID if (!data || !data_size) { - WARN("Invalid arguments: data %p, data_size %lu.\n", data, data_size); + WARN("Invalid arguments: data %p, data_size %"PRIuPTR".\n", data, (uintptr_t)data_size); return D3DERR_INVALIDCALL; } @@ -945,8 +946,8 @@ HRESULT WINAPI D3DDisassemble(const void *data, SIZE_T data_size, {VKD3D_SHADER_COMPILE_OPTION_API_VERSION, VKD3D_SHADER_API_VERSION_1_10}, }; - TRACE("data %p, data_size %lu, flags %#x, comments %p, blob %p.\n", - data, data_size, flags, comments, blob); + TRACE("data %p, data_size %"PRIuPTR", flags %#x, comments %p, blob %p.\n", + data, (uintptr_t)data_size, flags, comments, blob); if (flags) FIXME("Ignoring flags %#x.\n", flags); diff --git a/libs/vkd3d-utils/vkd3d_utils_private.h b/libs/vkd3d-utils/vkd3d_utils_private.h index e5fb1933..7fece3d6 100644 --- a/libs/vkd3d-utils/vkd3d_utils_private.h +++ b/libs/vkd3d-utils/vkd3d_utils_private.h @@ -32,6 +32,8 @@ #include "vkd3d_memory.h" #include <vkd3d_utils.h> +#include <inttypes.h> + #ifndef D3DERR_INVALIDCALL #define D3DERR_INVALIDCALL _HRESULT_TYPEDEF_(0x8876086c) #endif

1 0

Giovanni Mascellani : vkd3d-shader/ir: Introduce a simple control flow graph structurizer.
by Alexandre Julliard Feb. 6, 2024

Feb. 6, 2024

Module: vkd3d Branch: master Commit: 51f13391e6117592a4c77bbb457302c8c3bfa606 URL: https://gitlab.winehq.org/wine/vkd3d/-/commit/51f13391e6117592a4c77bbb45730… Author: Giovanni Mascellani <gmascellani(a)codeweavers.com> Date: Tue Jan 16 23:56:43 2024 +0100 vkd3d-shader/ir: Introduce a simple control flow graph structurizer. The structurizer is implemented along the lines of what is usually called the "structured program theorem": the control flow is completely virtualized by mean of an additional TEMP register which stores the block index which is currently running. The whole program is then converted to a huge switch construction enclosed in a loop, executing at each iteration the appropriate block and updating the register depending on block jump instruction. The algorithm's generality is also its major weakness: it accepts any input program, even if its CFG is not reducible, but the output program lacks any useful convergence information. It satisfies the letter of the SPIR-V requirements, but it is expected that it will be very inefficient to run on a GPU (unless a downstream compiler is able to devirtualize the control flow and do a proper convergence analysis pass). The algorithm is however very simple, and good enough to at least pass tests, enabling further development. A better alternative is expected to be upstreamed incrementally. Side note: the structured program theorem is often called the Böhm-Jacopini theorem; Böhm and Jacopini did indeed prove a variation of it, but their algorithm is different from what is commontly attributed to them and implemented here, so I opted for not using their name. --- libs/vkd3d-shader/ir.c | 147 ++++++++++++++++++++++++++++++++- tests/hlsl/conditional.shader_test | 2 +- tests/hlsl/for.shader_test | 2 +- tests/hlsl/function-return.shader_test | 44 +++++----- tests/hlsl/loop.shader_test | 6 +- tests/hlsl/return.shader_test | 32 +++---- tests/hlsl/sm6-ternary.shader_test | 4 +- tests/hlsl/switch.shader_test | 40 ++++----- 8 files changed, 209 insertions(+), 68 deletions(-)

1 0

Giovanni Mascellani : vkd3d-shader/ir: Handle PHI nodes when materializing SSA registers.
by Alexandre Julliard Feb. 6, 2024

Feb. 6, 2024

Module: vkd3d Branch: master Commit: 19aef21369c2e20d4239dd77c9c0c4e27b5dd18d URL: https://gitlab.winehq.org/wine/vkd3d/-/commit/19aef21369c2e20d4239dd77c9c0c… Author: Giovanni Mascellani <gmascellani(a)codeweavers.com> Date: Tue Jan 9 15:56:45 2024 +0100 vkd3d-shader/ir: Handle PHI nodes when materializing SSA registers. PHI nodes cannot be used with TEMP registers, so they have to be converted to MOV/MOVC nodes and moved before the BRANCH node. --- libs/vkd3d-shader/ir.c | 170 +++++++++++++++++++++++++++++++++++++++++++++++-- 1 file changed, 163 insertions(+), 7 deletions(-) diff --git a/libs/vkd3d-shader/ir.c b/libs/vkd3d-shader/ir.c index b92a04b2..1f14f702 100644 --- a/libs/vkd3d-shader/ir.c +++ b/libs/vkd3d-shader/ir.c @@ -2666,27 +2666,110 @@ static void materialize_ssas_to_temps_process_src_param(struct vkd3d_shader_pars materialize_ssas_to_temps_process_reg(parser, &src->reg); } +static const struct vkd3d_shader_src_param *materialize_ssas_to_temps_compute_source(struct vkd3d_shader_instruction *ins, + unsigned int label) +{ + unsigned int i; + + assert(ins->handler_idx == VKD3DSIH_PHI); + + for (i = 0; i < ins->src_count; i += 2) + { + if (label_from_src_param(&ins->src[i + 1]) == label) + return &ins->src[i]; + } + + vkd3d_unreachable(); +} + +static bool materialize_ssas_to_temps_synthesize_mov(struct vkd3d_shader_parser *parser, + struct vkd3d_shader_instruction *instruction, const struct vkd3d_shader_location *loc, + const struct vkd3d_shader_dst_param *dest, const struct vkd3d_shader_src_param *cond, + const struct vkd3d_shader_src_param *source, bool invert) +{ + struct vkd3d_shader_src_param *src; + struct vkd3d_shader_dst_param *dst; + + if (!vsir_instruction_init_with_params(&parser->program, instruction, loc, + cond ? VKD3DSIH_MOVC : VKD3DSIH_MOV, 1, cond ? 3 : 1)) + return false; + + dst = instruction->dst; + src = instruction->src; + + dst[0] = *dest; + materialize_ssas_to_temps_process_dst_param(parser, &dst[0]); + + assert(dst[0].write_mask == VKD3DSP_WRITEMASK_0); + assert(dst[0].modifiers == 0); + assert(dst[0].shift == 0); + + if (cond) + { + src[0] = *cond; + src[1 + invert] = *source; + memset(&src[2 - invert], 0, sizeof(src[2 - invert])); + src[2 - invert].reg = dst[0].reg; + materialize_ssas_to_temps_process_src_param(parser, &src[1]); + materialize_ssas_to_temps_process_src_param(parser, &src[2]); + } + else + { + src[0] = *source; + materialize_ssas_to_temps_process_src_param(parser, &src[0]); + } + + return true; +} + static enum vkd3d_result materialize_ssas_to_temps(struct vkd3d_shader_parser *parser) { struct vkd3d_shader_instruction *instructions = NULL; + struct materialize_ssas_to_temps_block_data + { + size_t phi_begin; + size_t phi_count; + } *block_index = NULL; size_t ins_capacity = 0, ins_count = 0, i; + unsigned int current_label = 0; if (!reserve_instructions(&instructions, &ins_capacity, parser->program.instructions.count)) goto fail; + if (!(block_index = vkd3d_calloc(parser->program.block_count, sizeof(*block_index)))) + { + ERR("Failed to allocate block index.\n"); + goto fail; + } + for (i = 0; i < parser->program.instructions.count; ++i) { struct vkd3d_shader_instruction *ins = &parser->program.instructions.elements[i]; - size_t j; - if (ins->handler_idx == VKD3DSIH_PHI) + switch (ins->handler_idx) { - WARN("Unhandled PHI when materializing SSA registers.\n"); - vkd3d_shader_parser_error(parser, VKD3D_SHADER_ERROR_VSIR_NOT_IMPLEMENTED, - "Unhandled PHI when materializing SSA registers."); - vkd3d_free(instructions); - return VKD3D_ERROR_NOT_IMPLEMENTED; + case VKD3DSIH_LABEL: + current_label = label_from_src_param(&ins->src[0]); + break; + + case VKD3DSIH_PHI: + assert(current_label != 0); + assert(i != 0); + if (block_index[current_label - 1].phi_begin == 0) + block_index[current_label - 1].phi_begin = i; + block_index[current_label - 1].phi_count += 1; + break; + + default: + current_label = 0; + break; } + } + + for (i = 0; i < parser->program.instructions.count; ++i) + { + struct vkd3d_shader_instruction *ins = &parser->program.instructions.elements[i]; + size_t j; for (j = 0; j < ins->dst_count; ++j) materialize_ssas_to_temps_process_dst_param(parser, &ins->dst[j]); @@ -2694,6 +2777,77 @@ static enum vkd3d_result materialize_ssas_to_temps(struct vkd3d_shader_parser *p for (j = 0; j < ins->src_count; ++j) materialize_ssas_to_temps_process_src_param(parser, &ins->src[j]); + switch (ins->handler_idx) + { + case VKD3DSIH_LABEL: + current_label = label_from_src_param(&ins->src[0]); + break; + + case VKD3DSIH_BRANCH: + { + if (vsir_register_is_label(&ins->src[0].reg)) + { + const struct materialize_ssas_to_temps_block_data *data = &block_index[label_from_src_param(&ins->src[0]) - 1]; + + if (!reserve_instructions(&instructions, &ins_capacity, ins_count + data->phi_count)) + goto fail; + + for (j = data->phi_begin; j < data->phi_begin + data->phi_count; ++j) + { + const struct vkd3d_shader_src_param *source; + + source = materialize_ssas_to_temps_compute_source(&parser->program.instructions.elements[j], current_label); + if (!materialize_ssas_to_temps_synthesize_mov(parser, &instructions[ins_count], &ins->location, + &parser->program.instructions.elements[j].dst[0], NULL, source, false)) + goto fail; + + ++ins_count; + } + } + else + { + struct materialize_ssas_to_temps_block_data *data_true = &block_index[label_from_src_param(&ins->src[1]) - 1], + *data_false = &block_index[label_from_src_param(&ins->src[2]) - 1]; + const struct vkd3d_shader_src_param *cond = &ins->src[0]; + + if (!reserve_instructions(&instructions, &ins_capacity, + ins_count + data_true->phi_count + data_false->phi_count)) + goto fail; + + for (j = data_true->phi_begin; j < data_true->phi_begin + data_true->phi_count; ++j) + { + const struct vkd3d_shader_src_param *source; + + source = materialize_ssas_to_temps_compute_source(&parser->program.instructions.elements[j], current_label); + if (!materialize_ssas_to_temps_synthesize_mov(parser, &instructions[ins_count], &ins->location, + &parser->program.instructions.elements[j].dst[0], cond, source, false)) + goto fail; + + ++ins_count; + } + + for (j = data_false->phi_begin; j < data_false->phi_begin + data_false->phi_count; ++j) + { + const struct vkd3d_shader_src_param *source; + + source = materialize_ssas_to_temps_compute_source(&parser->program.instructions.elements[j], current_label); + if (!materialize_ssas_to_temps_synthesize_mov(parser, &instructions[ins_count], &ins->location, + &parser->program.instructions.elements[j].dst[0], cond, source, true)) + goto fail; + + ++ins_count; + } + } + break; + } + + case VKD3DSIH_PHI: + continue; + + default: + break; + } + if (!reserve_instructions(&instructions, &ins_capacity, ins_count + 1)) goto fail; @@ -2701,6 +2855,7 @@ static enum vkd3d_result materialize_ssas_to_temps(struct vkd3d_shader_parser *p } vkd3d_free(parser->program.instructions.elements); + vkd3d_free(block_index); parser->program.instructions.elements = instructions; parser->program.instructions.capacity = ins_capacity; parser->program.instructions.count = ins_count; @@ -2711,6 +2866,7 @@ static enum vkd3d_result materialize_ssas_to_temps(struct vkd3d_shader_parser *p fail: vkd3d_free(instructions); + vkd3d_free(block_index); return VKD3D_ERROR_OUT_OF_MEMORY; }

1 0

Giovanni Mascellani : vkd3d-shader/ir: Materialize SSA registers to temporaries.
by Alexandre Julliard Feb. 6, 2024

Feb. 6, 2024

Module: vkd3d Branch: master Commit: e0d3e9c376a9fc4f80bf9dc79201766030b191a7 URL: https://gitlab.winehq.org/wine/vkd3d/-/commit/e0d3e9c376a9fc4f80bf9dc792017… Author: Giovanni Mascellani <gmascellani(a)codeweavers.com> Date: Tue Jan 9 15:56:45 2024 +0100 vkd3d-shader/ir: Materialize SSA registers to temporaries. For simplicity PHI nodes are not currently handled. The goal for this pass is to make the CFG structurizer simpler, because it doesn't have to care about the more rigid rules SSA registers have to satisfy than TEMP registers. It is likely that the generated code will be harder for downstream compilers to optimize and execute efficiently, so once a complete structurizer is in place this pass should be removed, or at least greatly reduced in scope. --- libs/vkd3d-shader/ir.c | 81 ++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 81 insertions(+) diff --git a/libs/vkd3d-shader/ir.c b/libs/vkd3d-shader/ir.c index a0d6212c..b92a04b2 100644 --- a/libs/vkd3d-shader/ir.c +++ b/libs/vkd3d-shader/ir.c @@ -2637,6 +2637,84 @@ fail: return VKD3D_ERROR_OUT_OF_MEMORY; } +static void materialize_ssas_to_temps_process_src_param(struct vkd3d_shader_parser *parser, struct vkd3d_shader_src_param *src); + +/* This is idempotent: it can be safely applied more than once on the + * same register. */ +static void materialize_ssas_to_temps_process_reg(struct vkd3d_shader_parser *parser, struct vkd3d_shader_register *reg) +{ + unsigned int i; + + if (reg->type == VKD3DSPR_SSA) + { + reg->type = VKD3DSPR_TEMP; + reg->idx[0].offset += parser->program.temp_count; + } + + for (i = 0; i < reg->idx_count; ++i) + if (reg->idx[i].rel_addr) + materialize_ssas_to_temps_process_src_param(parser, reg->idx[i].rel_addr); +} + +static void materialize_ssas_to_temps_process_dst_param(struct vkd3d_shader_parser *parser, struct vkd3d_shader_dst_param *dst) +{ + materialize_ssas_to_temps_process_reg(parser, &dst->reg); +} + +static void materialize_ssas_to_temps_process_src_param(struct vkd3d_shader_parser *parser, struct vkd3d_shader_src_param *src) +{ + materialize_ssas_to_temps_process_reg(parser, &src->reg); +} + +static enum vkd3d_result materialize_ssas_to_temps(struct vkd3d_shader_parser *parser) +{ + struct vkd3d_shader_instruction *instructions = NULL; + size_t ins_capacity = 0, ins_count = 0, i; + + if (!reserve_instructions(&instructions, &ins_capacity, parser->program.instructions.count)) + goto fail; + + for (i = 0; i < parser->program.instructions.count; ++i) + { + struct vkd3d_shader_instruction *ins = &parser->program.instructions.elements[i]; + size_t j; + + if (ins->handler_idx == VKD3DSIH_PHI) + { + WARN("Unhandled PHI when materializing SSA registers.\n"); + vkd3d_shader_parser_error(parser, VKD3D_SHADER_ERROR_VSIR_NOT_IMPLEMENTED, + "Unhandled PHI when materializing SSA registers."); + vkd3d_free(instructions); + return VKD3D_ERROR_NOT_IMPLEMENTED; + } + + for (j = 0; j < ins->dst_count; ++j) + materialize_ssas_to_temps_process_dst_param(parser, &ins->dst[j]); + + for (j = 0; j < ins->src_count; ++j) + materialize_ssas_to_temps_process_src_param(parser, &ins->src[j]); + + if (!reserve_instructions(&instructions, &ins_capacity, ins_count + 1)) + goto fail; + + instructions[ins_count++] = *ins; + } + + vkd3d_free(parser->program.instructions.elements); + parser->program.instructions.elements = instructions; + parser->program.instructions.capacity = ins_capacity; + parser->program.instructions.count = ins_count; + parser->program.temp_count += parser->program.ssa_count; + parser->program.ssa_count = 0; + + return VKD3D_OK; + +fail: + vkd3d_free(instructions); + + return VKD3D_ERROR_OUT_OF_MEMORY; +} + enum vkd3d_result vkd3d_shader_normalise(struct vkd3d_shader_parser *parser, const struct vkd3d_shader_compile_info *compile_info) { @@ -2652,6 +2730,9 @@ enum vkd3d_result vkd3d_shader_normalise(struct vkd3d_shader_parser *parser, { if ((result = lower_switch_to_if_ladder(&parser->program)) < 0) return result; + + if ((result = materialize_ssas_to_temps(parser)) < 0) + return result; } else {

1 0