diff options
| -rw-r--r-- | src/common/alignment.h | 9 | ||||
| -rw-r--r-- | src/video_core/texture_cache/accelerated_swizzle.cpp | 4 | ||||
| -rw-r--r-- | src/video_core/texture_cache/util.cpp | 10 | ||||
| -rw-r--r-- | src/video_core/textures/decoders.cpp | 8 | 
4 files changed, 15 insertions, 16 deletions
| diff --git a/src/common/alignment.h b/src/common/alignment.h index 5040043de..2127dc186 100644 --- a/src/common/alignment.h +++ b/src/common/alignment.h @@ -17,15 +17,14 @@ template <typename T>  }  template <typename T> -[[nodiscard]] constexpr T AlignDown(T value, std::size_t size) { -    static_assert(std::is_unsigned_v<T>, "T must be an unsigned value."); -    return static_cast<T>(value - value % size); +requires std::is_unsigned_v<T>[[nodiscard]] constexpr T AlignUpLog2(T value, size_t align_log2) { +    return static_cast<T>((value + ((1ULL << align_log2) - 1)) >> align_log2 << align_log2);  }  template <typename T> -[[nodiscard]] constexpr T AlignBits(T value, std::size_t align) { +[[nodiscard]] constexpr T AlignDown(T value, std::size_t size) {      static_assert(std::is_unsigned_v<T>, "T must be an unsigned value."); -    return static_cast<T>((value + ((1ULL << align) - 1)) >> align << align); +    return static_cast<T>(value - value % size);  }  template <typename T> diff --git a/src/video_core/texture_cache/accelerated_swizzle.cpp b/src/video_core/texture_cache/accelerated_swizzle.cpp index a4fc1184b..15585caeb 100644 --- a/src/video_core/texture_cache/accelerated_swizzle.cpp +++ b/src/video_core/texture_cache/accelerated_swizzle.cpp @@ -27,7 +27,7 @@ BlockLinearSwizzle2DParams MakeBlockLinearSwizzle2DParams(const SwizzleParameter      const Extent3D num_tiles = swizzle.num_tiles;      const u32 bytes_per_block = BytesPerBlock(info.format);      const u32 stride_alignment = CalculateLevelStrideAlignment(info, swizzle.level); -    const u32 stride = Common::AlignBits(num_tiles.width, stride_alignment) * bytes_per_block; +    const u32 stride = Common::AlignUpLog2(num_tiles.width, stride_alignment) * bytes_per_block;      const u32 gobs_in_x = Common::DivCeilLog2(stride, GOB_SIZE_X_SHIFT);      return BlockLinearSwizzle2DParams{          .origin{0, 0, 0}, @@ -47,7 +47,7 @@ BlockLinearSwizzle3DParams MakeBlockLinearSwizzle3DParams(const SwizzleParameter      const Extent3D num_tiles = swizzle.num_tiles;      const u32 bytes_per_block = BytesPerBlock(info.format);      const u32 stride_alignment = CalculateLevelStrideAlignment(info, swizzle.level); -    const u32 stride = Common::AlignBits(num_tiles.width, stride_alignment) * bytes_per_block; +    const u32 stride = Common::AlignUpLog2(num_tiles.width, stride_alignment) * bytes_per_block;      const u32 gobs_in_x = (stride + GOB_SIZE_X - 1) >> GOB_SIZE_X_SHIFT;      const u32 block_size = gobs_in_x << (GOB_SIZE_SHIFT + block.height + block.depth); diff --git a/src/video_core/texture_cache/util.cpp b/src/video_core/texture_cache/util.cpp index 279932778..ce8fcfe0a 100644 --- a/src/video_core/texture_cache/util.cpp +++ b/src/video_core/texture_cache/util.cpp @@ -279,7 +279,7 @@ template <u32 GOB_EXTENT>      const bool is_small = IsSmallerThanGobSize(blocks, gob, info.block.depth);      const u32 alignment = is_small ? 0 : info.tile_width_spacing;      return Extent2D{ -        .width = Common::AlignBits(gobs.width, alignment), +        .width = Common::AlignUpLog2(gobs.width, alignment),          .height = gobs.height,      };  } @@ -352,7 +352,7 @@ template <u32 GOB_EXTENT>      // https://github.com/Ryujinx/Ryujinx/blob/1c9aba6de1520aea5480c032e0ff5664ac1bb36f/Ryujinx.Graphics.Texture/SizeCalculator.cs#L134      if (tile_width_spacing > 0) {          const u32 alignment_log2 = GOB_SIZE_SHIFT + tile_width_spacing + block.height + block.depth; -        return Common::AlignBits(size_bytes, alignment_log2); +        return Common::AlignUpLog2(size_bytes, alignment_log2);      }      const u32 aligned_height = Common::AlignUp(size.height, tile_size_y);      while (block.height != 0 && aligned_height <= (1U << (block.height - 1)) * GOB_SIZE_Y) { @@ -528,9 +528,9 @@ template <u32 GOB_EXTENT>      const u32 alignment = StrideAlignment(num_tiles, info.block, bpp_log2, info.tile_width_spacing);      const Extent3D mip_block = AdjustMipBlockSize(num_tiles, info.block, 0);      return Extent3D{ -        .width = Common::AlignBits(num_tiles.width, alignment), -        .height = Common::AlignBits(num_tiles.height, GOB_SIZE_Y_SHIFT + mip_block.height), -        .depth = Common::AlignBits(num_tiles.depth, GOB_SIZE_Z_SHIFT + mip_block.depth), +        .width = Common::AlignUpLog2(num_tiles.width, alignment), +        .height = Common::AlignUpLog2(num_tiles.height, GOB_SIZE_Y_SHIFT + mip_block.height), +        .depth = Common::AlignUpLog2(num_tiles.depth, GOB_SIZE_Z_SHIFT + mip_block.depth),      };  } diff --git a/src/video_core/textures/decoders.cpp b/src/video_core/textures/decoders.cpp index 9f5181318..62685a183 100644 --- a/src/video_core/textures/decoders.cpp +++ b/src/video_core/textures/decoders.cpp @@ -49,7 +49,7 @@ void Swizzle(std::span<u8> output, std::span<const u8> input, u32 bytes_per_pixe      // We can configure here a custom pitch      // As it's not exposed 'width * bpp' will be the expected pitch.      const u32 pitch = width * bytes_per_pixel; -    const u32 stride = Common::AlignBits(width, stride_alignment) * bytes_per_pixel; +    const u32 stride = Common::AlignUpLog2(width, stride_alignment) * bytes_per_pixel;      const u32 gobs_in_x = Common::DivCeilLog2(stride, GOB_SIZE_X_SHIFT);      const u32 block_size = gobs_in_x << (GOB_SIZE_SHIFT + block_height + block_depth); @@ -217,9 +217,9 @@ void SwizzleKepler(const u32 width, const u32 height, const u32 dst_x, const u32  std::size_t CalculateSize(bool tiled, u32 bytes_per_pixel, u32 width, u32 height, u32 depth,                            u32 block_height, u32 block_depth) {      if (tiled) { -        const u32 aligned_width = Common::AlignBits(width * bytes_per_pixel, GOB_SIZE_X_SHIFT); -        const u32 aligned_height = Common::AlignBits(height, GOB_SIZE_Y_SHIFT + block_height); -        const u32 aligned_depth = Common::AlignBits(depth, GOB_SIZE_Z_SHIFT + block_depth); +        const u32 aligned_width = Common::AlignUpLog2(width * bytes_per_pixel, GOB_SIZE_X_SHIFT); +        const u32 aligned_height = Common::AlignUpLog2(height, GOB_SIZE_Y_SHIFT + block_height); +        const u32 aligned_depth = Common::AlignUpLog2(depth, GOB_SIZE_Z_SHIFT + block_depth);          return aligned_width * aligned_height * aligned_depth;      } else {          return width * height * depth * bytes_per_pixel; | 
