commit 909240df86af614c92df1a69d2d4b208394adf05
parent 9204ac856314243af7b7bc2c88b0cf96fbc4e02c
Author: Lee Salzman <lsalzman@mozilla.com>
Date: Tue, 7 Oct 2025 19:07:18 +0000
Bug 1993075 - Cleanup SWGL dithering code. r=aosmond
Differential Revision: https://phabricator.services.mozilla.com/D267852
Diffstat:
3 files changed, 73 insertions(+), 91 deletions(-)
diff --git a/gfx/wr/glsl-to-cxx/src/hir.rs b/gfx/wr/glsl-to-cxx/src/hir.rs
@@ -3970,7 +3970,7 @@ pub fn ast_to_hir(state: &mut State, tu: &syntax::TranslationUnit) -> Translatio
None,
Type::new(Void),
vec![Type::new(Sampler2D), Type::new(Int), Type::new(Float), Type::new(Bool), Type::new(Bool),
- Type::new(Vec2), Type::new(Vec2), Type::new(Float), Type::new(Vec4)],
+ Type::new(Vec2), Type::new(Vec2), Type::new(Float)],
);
declare_function(
state,
@@ -3986,7 +3986,7 @@ pub fn ast_to_hir(state: &mut State, tu: &syntax::TranslationUnit) -> Translatio
None,
Type::new(Void),
vec![Type::new(Sampler2D), Type::new(Int), Type::new(Float), Type::new(Bool), Type::new(Vec2),
- Type::new(Float), Type::new(Vec4)],
+ Type::new(Float)],
);
declare_function(
state,
@@ -4002,7 +4002,7 @@ pub fn ast_to_hir(state: &mut State, tu: &syntax::TranslationUnit) -> Translatio
None,
Type::new(Void),
vec![Type::new(Sampler2D), Type::new(Int), Type::new(Int), Type::new(Float), Type::new(Bool), Type::new(Vec2),
- Type::new(Float), Type::new(Vec4)],
+ Type::new(Float)],
);
declare_function(
state,
diff --git a/gfx/wr/glsl-to-cxx/src/lib.rs b/gfx/wr/glsl-to-cxx/src/lib.rs
@@ -1865,6 +1865,9 @@ pub fn show_hir_expr_inner(state: &OutputState, expr: &hir::Expr, top_level: boo
);
return;
}
+ } else if sym.name.starts_with("swgl_commitDithered") {
+ state.used_fragcoord.set(
+ state.used_fragcoord.get() | 1 | 2);
}
show_sym(state, name)
}
diff --git a/gfx/wr/swgl/src/swgl_ext.h b/gfx/wr/swgl/src/swgl_ext.h
@@ -8,6 +8,7 @@
// efficiently process it and modulate it with color without incurring blend
// stage overheads.
#include <cstdint>
+
template <typename P, typename C>
static void commit_masked_solid_span(P* buf, C color, int len) {
override_clip_mask();
@@ -1388,7 +1389,7 @@ static inline WideRGBA8 sampleGradient(sampler2D sampler, int address,
// TODO: These should be updated for parity with the shader dither
// implementation once something more final exists there. Right now, these are
// very close but slightly off.
-static const VectorType<int16_t, 16> ditherNoise[64] = {
+static const WideRGBA8 ditherNoise[64] = {
{2, 2, 2, 128, 194, 194, 194, 128, 50, 50, 50, 128, 242, 242, 242, 128},
{194, 194, 194, 128, 50, 50, 50, 128, 242, 242, 242, 128, 14, 14, 14, 128},
{50, 50, 50, 128, 242, 242, 242, 128, 14, 14, 14, 128, 206, 206, 206, 128},
@@ -1483,13 +1484,15 @@ static const VectorType<int16_t, 16> ditherNoise[64] = {
{150, 150, 150, 128, 86, 86, 86, 128, 2, 2, 2, 128, 194, 194, 194, 128},
{86, 86, 86, 128, 2, 2, 2, 128, 194, 194, 194, 128, 50, 50, 50, 128}};
-// Values in color should be in the 0x00FF..0xFF00 range so that dithering has
+static ALWAYS_INLINE const WideRGBA8* getDitherNoise(int32_t fragCoordY) {
+ return &ditherNoise[(fragCoordY & 7) * 8];
+}
+
+// Values in color should be in the 0..0xFF00 range so that dithering has
// enough overhead to avoid overflow and underflow.
-static inline VectorType<uint16_t, 4 * 4> dither(
- VectorType<uint16_t, 4 * 4> color, ivec4_scalar fragCoord,
- const VectorType<int16_t, 16>* ditherNoiseYIndexed) {
- color += ditherNoiseYIndexed[fragCoord.x & 7];
- return color;
+static ALWAYS_INLINE WideRGBA8 dither(WideRGBA8 color, int32_t fragCoordX,
+ const WideRGBA8* ditherNoiseYIndexed) {
+ return color + ditherNoiseYIndexed[fragCoordX & 7];
}
// Samples an entire span of a linear gradient by crawling the gradient table
@@ -1499,7 +1502,8 @@ template <bool BLEND, bool DITHER>
static bool commitLinearGradient(sampler2D sampler, int address, float size,
bool tileRepeat, bool gradientRepeat, vec2 pos,
const vec2_scalar& scaleDir, float startOffset,
- uint32_t* buf, int span, vec4 fragCoord) {
+ uint32_t* buf, int span,
+ vec4 fragCoord = vec4()) {
assert(sampler->format == TextureFormat::RGBA32F);
assert(address >= 0 && address < int(sampler->height * sampler->stride));
GradientStops* stops = (GradientStops*)&sampler->buf[address];
@@ -1513,13 +1517,9 @@ static bool commitLinearGradient(sampler2D sampler, int address, float size,
}
// Only incremented in the case of dithering
- ivec4_scalar currentFragCoord =
- ivec4_scalar(fragCoord.x.x, fragCoord.y.x, fragCoord.z.x, fragCoord.w.x);
-
- const VectorType<int16_t, 16>* ditherNoiseYIndexed;
- if (DITHER) {
- ditherNoiseYIndexed = &ditherNoise[((uint32_t)fragCoord.y.x & 7) * 8];
- }
+ int32_t currentFragCoordX = int32_t(fragCoord.x.x);
+ const auto* ditherNoiseYIndexed =
+ DITHER ? getDitherNoise(int32_t(fragCoord.y.x)) : nullptr;
// If we have a repeating brush, then the position will be modulo the [0,1)
// interval. Compute coefficients that can be used to quickly evaluate the
@@ -1619,20 +1619,18 @@ static bool commitLinearGradient(sampler2D sampler, int address, float size,
if (chunks >= 1.0f) {
int inside = int(chunks);
// Sample the start color from the min entry and the end color from the
- // max entry of the merged gradient. For non-dithered, these are scaled to
- // a range of 0..0xFF00, as that is the largest shifted value that can fit
- // in a U16. For dithered, these are scaled to a range of 0x00FF-0xFF00 to
- // allow room to avoid overflow and underflow when applying the dither
- // pattern. Since we are only doing addition with the step value, we can
- // still represent negative step values without having to use an explicit
- // sign bit, as the result will still come out the same, allowing us to
- // gain an extra bit of precision. We will later shift these into 8 bit
- // output range while committing the span, but stepping with higher
- // precision to avoid banding. We convert from RGBA to BGRA here to avoid
- // doing this in the inner loop.
+ // max entry of the merged gradient. These are scaled to a range of
+ // 0..0xFF00, as that is the largest shifted value that can fit in a U16.
+ // For dithering, this allows room to avoid overflow and underflow
+ // when applying the dither pattern. Since we are only doing addition with
+ // the step value, we can still represent negative step values without
+ // having to use an explicit sign bit, as the result will still come out
+ // the same, allowing us to gain an extra bit of precision. We will later
+ // shift these into 8 bit output range while committing the span, but
+ // stepping with higher precision to avoid banding. We convert from RGBA
+ // to BGRA here to avoid doing this in the inner loop.
auto minColorF = stops[minIndex].startColor.zyxw * float(0xFF00);
auto maxColorF = stops[maxIndex].end_color().zyxw * float(0xFF00);
-
// Get the color range of the merged gradient, normalized to its size.
auto colorRangeF =
(maxColorF - minColorF) * (1.0f / (maxIndex + 1 - minIndex));
@@ -1661,11 +1659,10 @@ static bool commitLinearGradient(sampler2D sampler, int address, float size,
if (DITHER) {
commit_blend_span<BLEND>(
buf,
- bit_cast<WideRGBA8>(
- dither(color, currentFragCoord, ditherNoiseYIndexed) >> 8));
- currentFragCoord.x += 4;
+ dither(color, currentFragCoordX, ditherNoiseYIndexed) >> 8);
+ currentFragCoordX += 4;
} else {
- commit_blend_span<BLEND>(buf, bit_cast<WideRGBA8>(color >> 8));
+ commit_blend_span<BLEND>(buf, color >> 8);
}
color += deltaColor;
}
@@ -1700,17 +1697,13 @@ static bool commitLinearGradient(sampler2D sampler, int address, float size,
// have different table entries.
Float entry = clamp(offset * size + 1.0f, 0.0f, 1.0f + size);
if (DITHER) {
- auto gradientSample = static_cast<VectorType<uint16_t, 4 * 4>>(
- sampleGradient(sampler, address, entry))
- << 8;
+ auto gradientSample = sampleGradient(sampler, address, entry) << 8;
commit_blend_span<BLEND>(
- buf, static_cast<WideRGBA8>(dither(gradientSample, currentFragCoord,
- ditherNoiseYIndexed) >>
- 8));
- currentFragCoord.x += 4;
+ buf,
+ dither(gradientSample, currentFragCoordX, ditherNoiseYIndexed) >> 8);
+ currentFragCoordX += 4;
} else {
- commit_blend_span<BLEND>(
- buf, static_cast<WideRGBA8>(sampleGradient(sampler, address, entry)));
+ commit_blend_span<BLEND>(buf, sampleGradient(sampler, address, entry));
}
span -= 4;
buf += 4;
@@ -1733,11 +1726,11 @@ static bool commitLinearGradient(sampler2D sampler, int address, float size,
if (blend_key) { \
drawn = commitLinearGradient<true, false>( \
sampler, address, size, tileRepeat, gradientRepeat, pos, scaleDir, \
- startOffset, swgl_OutRGBA8, swgl_SpanLength, make_vec4(0.f)); \
+ startOffset, swgl_OutRGBA8, swgl_SpanLength); \
} else { \
drawn = commitLinearGradient<false, false>( \
sampler, address, size, tileRepeat, gradientRepeat, pos, scaleDir, \
- startOffset, swgl_OutRGBA8, swgl_SpanLength, make_vec4(0.f)); \
+ startOffset, swgl_OutRGBA8, swgl_SpanLength); \
} \
if (drawn) { \
swgl_OutRGBA8 += swgl_SpanLength; \
@@ -1745,19 +1738,19 @@ static bool commitLinearGradient(sampler2D sampler, int address, float size,
} \
} while (0)
-#define swgl_commitDitheredLinearGradientRGBA8( \
- sampler, address, size, tileRepeat, gradientRepeat, pos, scaleDir, \
- startOffset, fragCoord) \
+#define swgl_commitDitheredLinearGradientRGBA8(sampler, address, size, \
+ tileRepeat, gradientRepeat, \
+ pos, scaleDir, startOffset) \
do { \
bool drawn = false; \
if (blend_key) { \
drawn = commitLinearGradient<true, true>( \
sampler, address, size, tileRepeat, gradientRepeat, pos, scaleDir, \
- startOffset, swgl_OutRGBA8, swgl_SpanLength, fragCoord); \
+ startOffset, swgl_OutRGBA8, swgl_SpanLength, gl_FragCoord); \
} else { \
drawn = commitLinearGradient<false, true>( \
sampler, address, size, tileRepeat, gradientRepeat, pos, scaleDir, \
- startOffset, swgl_OutRGBA8, swgl_SpanLength, fragCoord); \
+ startOffset, swgl_OutRGBA8, swgl_SpanLength, gl_FragCoord); \
} \
if (drawn) { \
swgl_OutRGBA8 += swgl_SpanLength; \
@@ -1790,7 +1783,8 @@ static ALWAYS_INLINE auto fastLength(V v) {
template <bool BLEND, bool DITHER>
static bool commitRadialGradient(sampler2D sampler, int address, float size,
bool repeat, vec2 pos, float radius,
- uint32_t* buf, int span, vec4 fragCoord) {
+ uint32_t* buf, int span,
+ vec4 fragCoord = vec4()) {
assert(sampler->format == TextureFormat::RGBA32F);
assert(address >= 0 && address < int(sampler->height * sampler->stride));
GradientStops* stops = (GradientStops*)&sampler->buf[address];
@@ -1822,13 +1816,9 @@ static bool commitRadialGradient(sampler2D sampler, int address, float size,
}
// Only incremented in the case of dithering
- ivec4_scalar currentFragCoord =
- ivec4_scalar(fragCoord.x.x, fragCoord.y.x, fragCoord.z.x, fragCoord.w.x);
-
- const VectorType<int16_t, 16>* ditherNoiseYIndexed;
- if (DITHER) {
- ditherNoiseYIndexed = &ditherNoise[((uint32_t)fragCoord.y.x & 7) * 8];
- }
+ int32_t currentFragCoordX = int32_t(fragCoord.x.x);
+ const auto* ditherNoiseYIndexed =
+ DITHER ? getDitherNoise(int32_t(fragCoord.y.x)) : nullptr;
float invDelta, middleT, middleB;
if (deltaDelta > 0) {
@@ -1950,7 +1940,7 @@ static bool commitRadialGradient(sampler2D sampler, int address, float size,
// Figure out how many chunks are actually inside the merged gradient.
if (t + 4.0f <= endT) {
int inside = int(endT - t) & ~3;
- // Convert start and end colors to BGRA and scale to 0x00FF..0xFF00 range
+ // Convert start and end colors to BGRA and scale to 0..0xFF00 range
// (for dithered) or 0..255 (for non-dithered) later.
auto minColorF =
stops[minIndex].startColor.zyxw * (DITHER ? float(0xFF00) : 255.0f);
@@ -1976,10 +1966,8 @@ static bool commitRadialGradient(sampler2D sampler, int address, float size,
CONVERT(round_pixel(colorF + deltaColorF * offsetG.z, 1), U16),
CONVERT(round_pixel(colorF + deltaColorF * offsetG.w, 1), U16));
commit_blend_span<BLEND>(
- buf,
- static_cast<WideRGBA8>(
- dither(color, currentFragCoord, ditherNoiseYIndexed) >> 8));
- currentFragCoord.x += 4;
+ buf, dither(color, currentFragCoordX, ditherNoiseYIndexed) >> 8);
+ currentFragCoordX += 4;
} else {
auto color = combine(
packRGBA8(round_pixel(colorF + deltaColorF * offsetG.x, 1),
@@ -2100,7 +2088,7 @@ static bool commitRadialGradientFromStops(sampler2D sampler, int offsetsAddress,
int colorsAddress, float stopCount,
bool repeat, vec2 pos,
float startRadius, uint32_t* buf,
- int span, vec4 fragCoord) {
+ int span, vec4 fragCoord = vec4()) {
assert(sampler->format == TextureFormat::RGBA32F);
// Stop offsets are expected to be stored just after the colors.
assert(colorsAddress >= 0 && colorsAddress < offsetsAddress);
@@ -2136,13 +2124,9 @@ static bool commitRadialGradientFromStops(sampler2D sampler, int offsetsAddress,
}
// Only incremented in the case of dithering
- ivec4_scalar currentFragCoord =
- ivec4_scalar(fragCoord.x.x, fragCoord.y.x, fragCoord.z.x, fragCoord.w.x);
-
- const VectorType<int16_t, 16>* ditherNoiseYIndexed;
- if (DITHER) {
- ditherNoiseYIndexed = &ditherNoise[((uint32_t)fragCoord.y.x & 7) * 8];
- }
+ int32_t currentFragCoordX = int32_t(fragCoord.x.x);
+ const auto* ditherNoiseYIndexed =
+ DITHER ? getDitherNoise(int32_t(fragCoord.y.x)) : nullptr;
float invDelta, middleT, middleB;
if (deltaDelta > 0) {
@@ -2252,7 +2236,7 @@ static bool commitRadialGradientFromStops(sampler2D sampler, int offsetsAddress,
// Figure out how many pixels belonging to whole chunks are inside the gradient
// stop pair.
int inside = int(endT - t) & ~3;
- // Convert start and end colors to BGRA and scale to 0x00FF..0xFF00 range
+ // Convert start and end colors to BGRA and scale to 0..0xFF00 range
// (for dithered) and 0.255 range (for non-dithered).
auto minColorF =
stopColors[stopIndex].zyxw * (DITHER ? float(0xFF00) : 255.0f);
@@ -2286,10 +2270,8 @@ static bool commitRadialGradientFromStops(sampler2D sampler, int offsetsAddress,
CONVERT(round_pixel(colorF + deltaColorF * offsetG.z, 1), U16),
CONVERT(round_pixel(colorF + deltaColorF * offsetG.w, 1), U16));
commit_blend_span<BLEND>(
- buf,
- static_cast<WideRGBA8>(
- dither(color, currentFragCoord, ditherNoiseYIndexed) >> 8));
- currentFragCoord.x += 4;
+ buf, dither(color, currentFragCoordX, ditherNoiseYIndexed) >> 8);
+ currentFragCoordX += 4;
} else {
auto color = combine(
packRGBA8(round_pixel(colorF + deltaColorF * offsetG.x, 1),
@@ -2323,11 +2305,9 @@ static bool commitRadialGradientFromStops(sampler2D sampler, int offsetsAddress,
CONVERT(round_pixel(colorF + deltaColorF * offsetG.z, 1), U16),
CONVERT(round_pixel(colorF + deltaColorF * offsetG.w, 1), U16));
commit_blend_span<BLEND>(
- buf,
- static_cast<WideRGBA8>(
- dither(color, currentFragCoord, ditherNoiseYIndexed) >> 8),
+ buf, dither(color, currentFragCoordX, ditherNoiseYIndexed) >> 8,
remainder);
- currentFragCoord.x += 4;
+ currentFragCoordX += 4;
} else {
auto color = combine(
packRGBA8(round_pixel(colorF + deltaColorF * offsetG.x, 1),
@@ -2382,11 +2362,11 @@ static bool commitRadialGradientFromStops(sampler2D sampler, int offsetsAddress,
if (blend_key) { \
drawn = commitRadialGradient<true, false>( \
sampler, address, size, repeat, pos, radius, swgl_OutRGBA8, \
- swgl_SpanLength, make_vec4(0.f)); \
+ swgl_SpanLength); \
} else { \
drawn = commitRadialGradient<false, false>( \
sampler, address, size, repeat, pos, radius, swgl_OutRGBA8, \
- swgl_SpanLength, make_vec4(0.f)); \
+ swgl_SpanLength); \
} \
if (drawn) { \
swgl_OutRGBA8 += swgl_SpanLength; \
@@ -2395,17 +2375,17 @@ static bool commitRadialGradientFromStops(sampler2D sampler, int offsetsAddress,
} while (0)
#define swgl_commitDitheredRadialGradientRGBA8(sampler, address, size, repeat, \
- pos, radius, fragCoord) \
+ pos, radius) \
do { \
bool drawn = false; \
if (blend_key) { \
drawn = commitRadialGradient<true, true>(sampler, address, size, repeat, \
pos, radius, swgl_OutRGBA8, \
- swgl_SpanLength, fragCoord); \
+ swgl_SpanLength, gl_FragCoord); \
} else { \
drawn = commitRadialGradient<false, true>( \
sampler, address, size, repeat, pos, radius, swgl_OutRGBA8, \
- swgl_SpanLength, fragCoord); \
+ swgl_SpanLength, gl_FragCoord); \
} \
if (drawn) { \
swgl_OutRGBA8 += swgl_SpanLength; \
@@ -2424,11 +2404,11 @@ static bool commitRadialGradientFromStops(sampler2D sampler, int offsetsAddress,
if (blend_key) { \
drawn = commitRadialGradientFromStops<true, false>( \
sampler, offsetsAddress, colorsAddress, size, repeat, pos, \
- startRadius, swgl_OutRGBA8, swgl_SpanLength, make_vec4(0.f)); \
+ startRadius, swgl_OutRGBA8, swgl_SpanLength); \
} else { \
drawn = commitRadialGradientFromStops<false, false>( \
sampler, offsetsAddress, colorsAddress, size, repeat, pos, \
- startRadius, swgl_OutRGBA8, swgl_SpanLength, make_vec4(0.f)); \
+ startRadius, swgl_OutRGBA8, swgl_SpanLength); \
} \
if (drawn) { \
swgl_OutRGBA8 += swgl_SpanLength; \
@@ -2437,18 +2417,17 @@ static bool commitRadialGradientFromStops(sampler2D sampler, int offsetsAddress,
} while (0)
#define swgl_commitDitheredRadialGradientFromStopsRGBA8( \
- sampler, offsetsAddress, colorsAddress, size, repeat, pos, startRadius, \
- fragCoord) \
+ sampler, offsetsAddress, colorsAddress, size, repeat, pos, startRadius) \
do { \
bool drawn = false; \
if (blend_key) { \
drawn = commitRadialGradientFromStops<true, true>( \
sampler, offsetsAddress, colorsAddress, size, repeat, pos, \
- startRadius, swgl_OutRGBA8, swgl_SpanLength, fragCoord); \
+ startRadius, swgl_OutRGBA8, swgl_SpanLength, gl_FragCoord); \
} else { \
drawn = commitRadialGradientFromStops<false, true>( \
sampler, offsetsAddress, colorsAddress, size, repeat, pos, \
- startRadius, swgl_OutRGBA8, swgl_SpanLength, fragCoord); \
+ startRadius, swgl_OutRGBA8, swgl_SpanLength, gl_FragCoord); \
} \
if (drawn) { \
swgl_OutRGBA8 += swgl_SpanLength; \
