Optimize INV_PREMUL

Our inverse premultiply is rather unoptimized. It's major weakness is
using three divide operation which are the slowest integer operations.

This patch reduces the three divisions to just one but using an
approximation that is accurate for all uchar values.

The patch also adds a general short-cut for alpha==255.

Together these improvements makes it 2 to 16 times faster depending on
how many 0 and 255 alpha values are encountered.

Change-Id: I96d7098a1bc320793b0d0526637acd1fdb5a43eb
Reviewed-by: Gunnar Sletta <gunnar.sletta@jollamobile.com>

2 files changed, 11 insertions, 7 deletions

diff --git a/src/gui/painting/qdrawhelper.cpp b/src/gui/painting/qdrawhelper.cpp
index 141688b826..8aef64c780 100644
--- a/src/gui/painting/qdrawhelper.cpp
+++ b/src/gui/painting/qdrawhelper.cpp
@@ -274,7 +274,7 @@ static const uint *QT_FASTCALL convertFromARGB32PM(uint *buffer, const uint *src
 
     if (!layout->premultiplied) {
         for (int i = 0; i < count; ++i)
-            buffer[i] = qAlpha(src[i]) == 255 ? src[i] : INV_PREMUL(src[i]);
+            buffer[i] = INV_PREMUL(src[i]);
         src = buffer;
     }
     for (int i = 0; i < count; ++i) {
diff --git a/src/gui/painting/qdrawhelper_p.h b/src/gui/painting/qdrawhelper_p.h
index 6af4aa4943..4b11a8e92b 100644
--- a/src/gui/painting/qdrawhelper_p.h
+++ b/src/gui/painting/qdrawhelper_p.h
@@ -691,12 +691,16 @@ static Q_ALWAYS_INLINE uint BYTE_MUL_RGB16_32(uint x, uint a) {
     return t;
 }
 
-#define INV_PREMUL(p)                                   \
-    (qAlpha(p) == 0 ? 0 :                               \
-    ((qAlpha(p) << 24)                                  \
-     | (((255*qRed(p))/ qAlpha(p)) << 16)               \
-     | (((255*qGreen(p)) / qAlpha(p))  << 8)            \
-     | ((255*qBlue(p)) / qAlpha(p))))
+static Q_ALWAYS_INLINE uint INV_PREMUL(uint p) {
+    const uint alpha = qAlpha(p);
+    if (alpha == 255)
+        return p;
+    if (alpha == 0)
+        return 0;
+    // (p*(0x00ff00ff/alpha)) >> 16 == (p*255)/alpha for all p and alpha <= 256.
+    const uint invAlpha = 0x00ff00ffU / alpha;
+    return qRgba((qRed(p)*invAlpha)>>16, (qGreen(p)*invAlpha)>>16, (qBlue(p)*invAlpha)>>16, alpha);
+}
 
 struct quint24 {
     quint24(uint value);