mirror of
https://github.com/RPCSX/rpcsx.git
synced 2026-04-04 14:08:37 +00:00
rsx: Texture fixes
- gl/vk: Fix subresource copy/blit - gl/vk: Fix default_component_map reading - vk: Reimplement cell readback path and improve software channel decoder - Properly name the subresource layout field - its in blocks not bytes! - Implement d24s8 upload from memory correctly - Do not ignore DEPTH_FLOAT textures - they are depth textures and abide by the depth compare rules - NOTE: Redirection of 16-bit textures is not implemented yet
This commit is contained in:
kd-11
kd-11
parent
63d9cb37ec
commit
a42b00488d
15 changed files with 591 additions and 557 deletions
|
|
@ -82,7 +82,7 @@ namespace rsx
|
|||
* N - Sample count
|
||||
*/
|
||||
template <typename T, typename U>
|
||||
void scale_image_fallback_impl(T* dst, const U* src, u16 src_width, u16 src_height, u16 dst_pitch, u16 src_pitch, u8 pixel_size, u8 samples_u, u8 samples_v)
|
||||
void scale_image_fallback_impl(T* dst, const U* src, u16 src_width, u16 src_height, u16 dst_pitch, u16 src_pitch, u8 element_size, u8 samples_u, u8 samples_v)
|
||||
{
|
||||
u32 dst_offset = 0;
|
||||
u32 src_offset = 0;
|
||||
|
|
@ -112,51 +112,39 @@ namespace rsx
|
|||
}
|
||||
}
|
||||
|
||||
void scale_image_fallback(void* dst, const void* src, u16 src_width, u16 src_height, u16 dst_pitch, u16 src_pitch, u8 pixel_size, u8 samples_u, u8 samples_v)
|
||||
void scale_image_fallback(void* dst, const void* src, u16 src_width, u16 src_height, u16 dst_pitch, u16 src_pitch, u8 element_size, u8 samples_u, u8 samples_v)
|
||||
{
|
||||
switch (pixel_size)
|
||||
switch (element_size)
|
||||
{
|
||||
case 1:
|
||||
scale_image_fallback_impl<u8, u8>((u8*)dst, (const u8*)src, src_width, src_height, dst_pitch, src_pitch, pixel_size, samples_u, samples_v);
|
||||
scale_image_fallback_impl<u8, u8>((u8*)dst, (const u8*)src, src_width, src_height, dst_pitch, src_pitch, element_size, samples_u, samples_v);
|
||||
break;
|
||||
case 2:
|
||||
scale_image_fallback_impl<u16, u16>((u16*)dst, (const u16*)src, src_width, src_height, dst_pitch, src_pitch, pixel_size, samples_u, samples_v);
|
||||
scale_image_fallback_impl<u16, u16>((u16*)dst, (const u16*)src, src_width, src_height, dst_pitch, src_pitch, element_size, samples_u, samples_v);
|
||||
break;
|
||||
case 4:
|
||||
scale_image_fallback_impl<u32, u32>((u32*)dst, (const u32*)src, src_width, src_height, dst_pitch, src_pitch, pixel_size, samples_u, samples_v);
|
||||
break;
|
||||
case 8:
|
||||
scale_image_fallback_impl<u64, u64>((u64*)dst, (const u64*)src, src_width, src_height, dst_pitch, src_pitch, pixel_size, samples_u, samples_v);
|
||||
break;
|
||||
case 16:
|
||||
scale_image_fallback_impl<u128, u128>((u128*)dst, (const u128*)src, src_width, src_height, dst_pitch, src_pitch, pixel_size, samples_u, samples_v);
|
||||
scale_image_fallback_impl<u32, u32>((u32*)dst, (const u32*)src, src_width, src_height, dst_pitch, src_pitch, element_size, samples_u, samples_v);
|
||||
break;
|
||||
default:
|
||||
fmt::throw_exception("unsupported pixel size %d" HERE, pixel_size);
|
||||
fmt::throw_exception("unsupported element size %d" HERE, element_size);
|
||||
}
|
||||
}
|
||||
|
||||
void scale_image_fallback_with_byte_swap(void* dst, const void* src, u16 src_width, u16 src_height, u16 dst_pitch, u16 src_pitch, u8 pixel_size, u8 samples_u, u8 samples_v)
|
||||
void scale_image_fallback_with_byte_swap(void* dst, const void* src, u16 src_width, u16 src_height, u16 dst_pitch, u16 src_pitch, u8 element_size, u8 samples_u, u8 samples_v)
|
||||
{
|
||||
switch (pixel_size)
|
||||
switch (element_size)
|
||||
{
|
||||
case 1:
|
||||
scale_image_fallback_impl<u8, u8>((u8*)dst, (const u8*)src, src_width, src_height, dst_pitch, src_pitch, pixel_size, samples_u, samples_v);
|
||||
scale_image_fallback_impl<u8, u8>((u8*)dst, (const u8*)src, src_width, src_height, dst_pitch, src_pitch, element_size, samples_u, samples_v);
|
||||
break;
|
||||
case 2:
|
||||
scale_image_fallback_impl<u16, be_t<u16>>((u16*)dst, (const be_t<u16>*)src, src_width, src_height, dst_pitch, src_pitch, pixel_size, samples_u, samples_v);
|
||||
scale_image_fallback_impl<u16, be_t<u16>>((u16*)dst, (const be_t<u16>*)src, src_width, src_height, dst_pitch, src_pitch, element_size, samples_u, samples_v);
|
||||
break;
|
||||
case 4:
|
||||
scale_image_fallback_impl<u32, be_t<u32>>((u32*)dst, (const be_t<u32>*)src, src_width, src_height, dst_pitch, src_pitch, pixel_size, samples_u, samples_v);
|
||||
break;
|
||||
case 8:
|
||||
scale_image_fallback_impl<u64, be_t<u64>>((u64*)dst, (const be_t<u64>*)src, src_width, src_height, dst_pitch, src_pitch, pixel_size, samples_u, samples_v);
|
||||
break;
|
||||
case 16:
|
||||
scale_image_fallback_impl<u128, be_t<u128>>((u128*)dst, (const be_t<u128>*)src, src_width, src_height, dst_pitch, src_pitch, pixel_size, samples_u, samples_v);
|
||||
scale_image_fallback_impl<u32, be_t<u32>>((u32*)dst, (const be_t<u32>*)src, src_width, src_height, dst_pitch, src_pitch, element_size, samples_u, samples_v);
|
||||
break;
|
||||
default:
|
||||
fmt::throw_exception("unsupported pixel size %d" HERE, pixel_size);
|
||||
fmt::throw_exception("unsupported element size %d" HERE, element_size);
|
||||
}
|
||||
}
|
||||
|
||||
|
|
@ -185,9 +173,9 @@ namespace rsx
|
|||
}
|
||||
|
||||
template <int N>
|
||||
void scale_image_fast(void *dst, const void *src, u8 pixel_size, u16 src_width, u16 src_height, u16 padding)
|
||||
void scale_image_fast(void *dst, const void *src, u8 element_size, u16 src_width, u16 src_height, u16 padding)
|
||||
{
|
||||
switch (pixel_size)
|
||||
switch (element_size)
|
||||
{
|
||||
case 1:
|
||||
scale_image_impl<u8, u8, N>((u8*)dst, (const u8*)src, src_width, src_height, padding);
|
||||
|
|
@ -202,14 +190,14 @@ namespace rsx
|
|||
scale_image_impl<u64, u64, N>((u64*)dst, (const u64*)src, src_width, src_height, padding);
|
||||
break;
|
||||
default:
|
||||
fmt::throw_exception("unsupported pixel size %d" HERE, pixel_size);
|
||||
fmt::throw_exception("unsupported pixel size %d" HERE, element_size);
|
||||
}
|
||||
}
|
||||
|
||||
template <int N>
|
||||
void scale_image_fast_with_byte_swap(void *dst, const void *src, u8 pixel_size, u16 src_width, u16 src_height, u16 padding)
|
||||
void scale_image_fast_with_byte_swap(void *dst, const void *src, u8 element_size, u16 src_width, u16 src_height, u16 padding)
|
||||
{
|
||||
switch (pixel_size)
|
||||
switch (element_size)
|
||||
{
|
||||
case 1:
|
||||
scale_image_impl<u8, u8, N>((u8*)dst, (const u8*)src, src_width, src_height, padding);
|
||||
|
|
@ -224,17 +212,17 @@ namespace rsx
|
|||
scale_image_impl<u64, be_t<u64>, N>((u64*)dst, (const be_t<u64>*)src, src_width, src_height, padding);
|
||||
break;
|
||||
default:
|
||||
fmt::throw_exception("unsupported pixel size %d" HERE, pixel_size);
|
||||
fmt::throw_exception("unsupported pixel size %d" HERE, element_size);
|
||||
}
|
||||
}
|
||||
|
||||
void scale_image_nearest(void* dst, const void* src, u16 src_width, u16 src_height, u16 dst_pitch, u16 src_pitch, u8 pixel_size, u8 samples_u, u8 samples_v, bool swap_bytes)
|
||||
void scale_image_nearest(void* dst, const void* src, u16 src_width, u16 src_height, u16 dst_pitch, u16 src_pitch, u8 element_size, u8 samples_u, u8 samples_v, bool swap_bytes)
|
||||
{
|
||||
//Scale this image by repeating pixel data n times
|
||||
//n = expected_pitch / real_pitch
|
||||
//Use of fixed argument templates for performance reasons
|
||||
|
||||
const u16 dst_width = dst_pitch / pixel_size;
|
||||
const u16 dst_width = dst_pitch / element_size;
|
||||
const u16 padding = dst_width - (src_width * samples_u);
|
||||
|
||||
if (!swap_bytes)
|
||||
|
|
@ -244,30 +232,30 @@ namespace rsx
|
|||
switch (samples_u)
|
||||
{
|
||||
case 1:
|
||||
scale_image_fast<1>(dst, src, pixel_size, src_width, src_height, padding);
|
||||
scale_image_fast<1>(dst, src, element_size, src_width, src_height, padding);
|
||||
break;
|
||||
case 2:
|
||||
scale_image_fast<2>(dst, src, pixel_size, src_width, src_height, padding);
|
||||
scale_image_fast<2>(dst, src, element_size, src_width, src_height, padding);
|
||||
break;
|
||||
case 3:
|
||||
scale_image_fast<3>(dst, src, pixel_size, src_width, src_height, padding);
|
||||
scale_image_fast<3>(dst, src, element_size, src_width, src_height, padding);
|
||||
break;
|
||||
case 4:
|
||||
scale_image_fast<4>(dst, src, pixel_size, src_width, src_height, padding);
|
||||
scale_image_fast<4>(dst, src, element_size, src_width, src_height, padding);
|
||||
break;
|
||||
case 8:
|
||||
scale_image_fast<8>(dst, src, pixel_size, src_width, src_height, padding);
|
||||
scale_image_fast<8>(dst, src, element_size, src_width, src_height, padding);
|
||||
break;
|
||||
case 16:
|
||||
scale_image_fast<16>(dst, src, pixel_size, src_width, src_height, padding);
|
||||
scale_image_fast<16>(dst, src, element_size, src_width, src_height, padding);
|
||||
break;
|
||||
default:
|
||||
scale_image_fallback(dst, src, src_width, src_height, dst_pitch, src_pitch, pixel_size, samples_u, 1);
|
||||
scale_image_fallback(dst, src, src_width, src_height, dst_pitch, src_pitch, element_size, samples_u, 1);
|
||||
}
|
||||
}
|
||||
else
|
||||
{
|
||||
scale_image_fallback(dst, src, src_width, src_height, dst_pitch, src_pitch, pixel_size, samples_u, samples_v);
|
||||
scale_image_fallback(dst, src, src_width, src_height, dst_pitch, src_pitch, element_size, samples_u, samples_v);
|
||||
}
|
||||
}
|
||||
else
|
||||
|
|
@ -277,30 +265,30 @@ namespace rsx
|
|||
switch (samples_u)
|
||||
{
|
||||
case 1:
|
||||
scale_image_fast_with_byte_swap<1>(dst, src, pixel_size, src_width, src_height, padding);
|
||||
scale_image_fast_with_byte_swap<1>(dst, src, element_size, src_width, src_height, padding);
|
||||
break;
|
||||
case 2:
|
||||
scale_image_fast_with_byte_swap<2>(dst, src, pixel_size, src_width, src_height, padding);
|
||||
scale_image_fast_with_byte_swap<2>(dst, src, element_size, src_width, src_height, padding);
|
||||
break;
|
||||
case 3:
|
||||
scale_image_fast_with_byte_swap<3>(dst, src, pixel_size, src_width, src_height, padding);
|
||||
scale_image_fast_with_byte_swap<3>(dst, src, element_size, src_width, src_height, padding);
|
||||
break;
|
||||
case 4:
|
||||
scale_image_fast_with_byte_swap<4>(dst, src, pixel_size, src_width, src_height, padding);
|
||||
scale_image_fast_with_byte_swap<4>(dst, src, element_size, src_width, src_height, padding);
|
||||
break;
|
||||
case 8:
|
||||
scale_image_fast_with_byte_swap<8>(dst, src, pixel_size, src_width, src_height, padding);
|
||||
scale_image_fast_with_byte_swap<8>(dst, src, element_size, src_width, src_height, padding);
|
||||
break;
|
||||
case 16:
|
||||
scale_image_fast_with_byte_swap<16>(dst, src, pixel_size, src_width, src_height, padding);
|
||||
scale_image_fast_with_byte_swap<16>(dst, src, element_size, src_width, src_height, padding);
|
||||
break;
|
||||
default:
|
||||
scale_image_fallback_with_byte_swap(dst, src, src_width, src_height, dst_pitch, src_pitch, pixel_size, samples_u, 1);
|
||||
scale_image_fallback_with_byte_swap(dst, src, src_width, src_height, dst_pitch, src_pitch, element_size, samples_u, 1);
|
||||
}
|
||||
}
|
||||
else
|
||||
{
|
||||
scale_image_fallback_with_byte_swap(dst, src, src_width, src_height, dst_pitch, src_pitch, pixel_size, samples_u, samples_v);
|
||||
scale_image_fallback_with_byte_swap(dst, src, src_width, src_height, dst_pitch, src_pitch, element_size, samples_u, samples_v);
|
||||
}
|
||||
}
|
||||
}
|
||||
|
|
@ -413,4 +401,26 @@ namespace rsx
|
|||
++src_ptr;
|
||||
}
|
||||
}
|
||||
|
||||
void convert_le_d24x8_to_le_f32(void *dst, void *src, u32 row_length_in_texels, u32 num_rows)
|
||||
{
|
||||
const u32 num_pixels = row_length_in_texels * num_rows;
|
||||
verify(HERE), (num_pixels & 3) == 0;
|
||||
|
||||
const auto num_iterations = (num_pixels >> 2);
|
||||
|
||||
__m128i* dst_ptr = (__m128i*)dst;
|
||||
__m128i* src_ptr = (__m128i*)src;
|
||||
|
||||
const __m128 scale_vector = _mm_set1_ps(1.f / 16777214.f);
|
||||
const __m128i mask = _mm_set1_epi32(0x00FFFFFF);
|
||||
for (u32 n = 0; n < num_iterations; ++n)
|
||||
{
|
||||
const __m128 src_vector = _mm_cvtepi32_ps(_mm_and_si128(mask, _mm_loadu_si128(src_ptr)));
|
||||
const __m128 normalized_vector = _mm_mul_ps(src_vector, scale_vector);
|
||||
_mm_stream_si128(dst_ptr, (__m128i&)normalized_vector);
|
||||
++dst_ptr;
|
||||
++src_ptr;
|
||||
}
|
||||
}
|
||||
}
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue