Skip to content

S
swiftshader
Commit 419a5806 by Nicolas Capens Committed by Nicolas Capens
Options

Refactor surface buffer size calculation.

This eliminates the duplication between pitchB() and size(), and ensures that the latter is the full allocation size. Bug chromium:835299 Change-Id: Icf555ad497fb3b92fd00e9a3e6ced6810b2d310d Reviewed-on: https://swiftshader-review.googlesource.com/18789Tested-by: 's avatarNicolas Capens <nicolascapens@google.com> Reviewed-by: 's avatarAlexis Hétu <sugoi@google.com>
parent efdf103a
Showing with 31 additions and 87 deletions
src/Common/Math.hpp
......@@ -364,7 +364,14 @@ namespace sw
uint64_t FNV_1a(const unsigned char *data, int size); // Fowler-Noll-Vo hash function
// Round up to the next multiple of alignment
inline unsigned int align(unsigned int value, unsigned int alignment)
template<typename T>
inline T align(T value, unsigned int alignment)
{
return ((value + alignment - 1) / alignment) * alignment;
}
template<unsigned int alignment, typename T>
inline T align(T value)
{
return ((value + alignment - 1) / alignment) * alignment;
}
......
src/Renderer/Sampler.cpp
......@@ -228,7 +228,7 @@ namespace sw
{
unsigned int YStride = pitchP;
unsigned int YSize = YStride * height;
unsigned int CStride = align(YStride / 2, 16);
unsigned int CStride = align<16>(YStride / 2);
unsigned int CSize = CStride * height / 2;
mipmap.buffer[1] = (byte*)mipmap.buffer[0] + YSize;
......
src/Renderer/Surface.cpp
......@@ -1709,9 +1709,10 @@ namespace sw
{
width += 2 * border;
// Render targets require 2x2 quads
if(target || isDepth(format) || isStencil(format))
{
width = align(width, 2);
width = align<2>(width);
}
switch(format)
......@@ -1773,7 +1774,7 @@ namespace sw
case FORMAT_YV12_BT601:
case FORMAT_YV12_BT709:
case FORMAT_YV12_JFIF:
return align(width, 16);
return align<16>(width);
default:
return bytes(format) * width;
}
......@@ -1790,9 +1791,10 @@ namespace sw
{
height += 2 * border;
// Render targets require 2x2 quads
if(target || isDepth(format) || isStencil(format))
{
height = ((height + 1) & ~1);
height = align<2>(height);
}
switch(format)
......@@ -1847,6 +1849,7 @@ namespace sw
return pitchB(width, border, format, target) * ((height + 11) / 12); // Pitch computed per 12 rows
case FORMAT_ATI1:
case FORMAT_ATI2:
return pitchB(width, border, format, target) * align<4>(height); // Pitch computed per row
default:
return pitchB(width, border, format, target) * height; // Pitch computed per row
}
......@@ -2643,89 +2646,30 @@ namespace sw
{
}
unsigned int Surface::size(int width, int height, int depth, int border, int samples, Format format)
size_t Surface::size(int width, int height, int depth, int border, int samples, Format format)
{
width += 2 * border;
height += 2 * border;
// Dimensions rounded up to multiples of 4, used for compressed formats
int width4 = align(width, 4);
int height4 = align(height, 4);
switch(format)
{
case FORMAT_DXT1:
case FORMAT_ATI1:
case FORMAT_ETC1:
case FORMAT_R11_EAC:
case FORMAT_SIGNED_R11_EAC:
case FORMAT_RGB8_ETC2:
case FORMAT_SRGB8_ETC2:
case FORMAT_RGB8_PUNCHTHROUGH_ALPHA1_ETC2:
case FORMAT_SRGB8_PUNCHTHROUGH_ALPHA1_ETC2:
return width4 * height4 * depth / 2;
case FORMAT_DXT3:
case FORMAT_DXT5:
case FORMAT_ATI2:
case FORMAT_RG11_EAC:
case FORMAT_SIGNED_RG11_EAC:
case FORMAT_RGBA8_ETC2_EAC:
case FORMAT_SRGB8_ALPHA8_ETC2_EAC:
case FORMAT_RGBA_ASTC_4x4_KHR:
case FORMAT_SRGB8_ALPHA8_ASTC_4x4_KHR:
return width4 * height4 * depth;
case FORMAT_RGBA_ASTC_5x4_KHR:
case FORMAT_SRGB8_ALPHA8_ASTC_5x4_KHR:
return align(width, 5) * height4 * depth;
case FORMAT_RGBA_ASTC_5x5_KHR:
case FORMAT_SRGB8_ALPHA8_ASTC_5x5_KHR:
return align(width, 5) * align(height, 5) * depth;
case FORMAT_RGBA_ASTC_6x5_KHR:
case FORMAT_SRGB8_ALPHA8_ASTC_6x5_KHR:
return align(width, 6) * align(height, 5) * depth;
case FORMAT_RGBA_ASTC_6x6_KHR:
case FORMAT_SRGB8_ALPHA8_ASTC_6x6_KHR:
return align(width, 6) * align(height, 6) * depth;
case FORMAT_RGBA_ASTC_8x5_KHR:
case FORMAT_SRGB8_ALPHA8_ASTC_8x5_KHR:
return align(width, 8) * align(height, 5) * depth;
case FORMAT_RGBA_ASTC_8x6_KHR:
case FORMAT_SRGB8_ALPHA8_ASTC_8x6_KHR:
return align(width, 8) * align(height, 6) * depth;
case FORMAT_RGBA_ASTC_8x8_KHR:
case FORMAT_SRGB8_ALPHA8_ASTC_8x8_KHR:
return align(width, 8) * align(height, 8) * depth;
case FORMAT_RGBA_ASTC_10x5_KHR:
case FORMAT_SRGB8_ALPHA8_ASTC_10x5_KHR:
return align(width, 10) * align(height, 5) * depth;
case FORMAT_RGBA_ASTC_10x6_KHR:
case FORMAT_SRGB8_ALPHA8_ASTC_10x6_KHR:
return align(width, 10) * align(height, 6) * depth;
case FORMAT_RGBA_ASTC_10x8_KHR:
case FORMAT_SRGB8_ALPHA8_ASTC_10x8_KHR:
return align(width, 10) * align(height, 8) * depth;
case FORMAT_RGBA_ASTC_10x10_KHR:
case FORMAT_SRGB8_ALPHA8_ASTC_10x10_KHR:
return align(width, 10) * align(height, 10) * depth;
case FORMAT_RGBA_ASTC_12x10_KHR:
case FORMAT_SRGB8_ALPHA8_ASTC_12x10_KHR:
return align(width, 12) * align(height, 10) * depth;
case FORMAT_RGBA_ASTC_12x12_KHR:
case FORMAT_SRGB8_ALPHA8_ASTC_12x12_KHR:
return align(width, 12) * align(height, 12) * depth;
default:
// FIXME: Unpacking byte4 to short4 in the sampler currently involves reading 8 bytes,
// and stencil operations also read 8 bytes per four 8-bit stencil values,
// so we have to allocate 4 extra bytes to avoid buffer overruns.
return (size_t)sliceB(width, height, border, format, true) * depth * samples + 4;
case FORMAT_YV12_BT601:
case FORMAT_YV12_BT709:
case FORMAT_YV12_JFIF:
{
unsigned int YStride = align(width, 16);
unsigned int YSize = YStride * height;
unsigned int CStride = align(YStride / 2, 16);
unsigned int CSize = CStride * height / 2;
width += 2 * border;
height += 2 * border;
size_t YStride = align<16>(width);
size_t YSize = YStride * height;
size_t CStride = align<16>(YStride / 2);
size_t CSize = CStride * height / 2;
return YSize + 2 * CSize;
}
default:
return bytes(format) * width * height * depth * samples;
}
}
......@@ -3260,14 +3204,7 @@ namespace sw
void *Surface::allocateBuffer(int width, int height, int depth, int border, int samples, Format format)
{
// Render targets require 2x2 quads
int width2 = (width + 1) & ~1;
int height2 = (height + 1) & ~1;
// FIXME: Unpacking byte4 to short4 in the sampler currently involves reading 8 bytes,
// and stencil operations also read 8 bytes per four 8-bit stencil values,
// so we have to allocate 4 extra bytes to avoid buffer overruns.
return allocate(size(width2, height2, depth, border, samples, format) + 4);
return allocate(size(width, height, depth, border, samples, format));
}
void Surface::memfill4(void *buffer, int pattern, int bytes)
......
src/Renderer/Surface.hpp
......@@ -362,7 +362,7 @@ namespace sw
static int pitchP(int width, int border, Format format, bool target);
static int sliceB(int width, int height, int border, Format format, bool target);
static int sliceP(int width, int height, int border, Format format, bool target);
static unsigned int size(int width, int height, int depth, int border, int samples, Format format); // FIXME: slice * depth
static size_t size(int width, int height, int depth, int border, int samples, Format format);
static bool isStencil(Format format);
static bool isDepth(Format format);
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment