gallium: Move sRGB <-> RGB handling to libmesautil

Signed-off-by: Jason Ekstrand <[email protected]>
Reviewed-by: Marek Olšák <[email protected]>

5 files changed, 312 insertions, 2 deletions

diff --git a/src/util/Makefile.am b/src/util/Makefile.am
index a2aeafc703b..e738ab08d2f 100644
--- a/src/util/Makefile.am
+++ b/src/util/Makefile.am
@@ -23,6 +23,10 @@ SUBDIRS = . tests
 
 include Makefile.sources
 
+BUILT_SOURCES = $(MESA_UTIL_GENERATED_FILES)
+
+CLEANFILES = $(BUILT_SOURCES)
+
 noinst_LTLIBRARIES = libmesautil.la
 
 libmesautil_la_CPPFLAGS = \
@@ -30,7 +34,12 @@ libmesautil_la_CPPFLAGS = \
 	-I$(top_srcdir)/include \
 	$(VISIBILITY_CFLAGS)
 
-libmesautil_la_SOURCES = $(MESA_UTIL_FILES)
+libmesautil_la_SOURCES = \
+	$(MESA_UTIL_FILES) \
+        $(MESA_UTIL_GENERATED_FILES)
+
+format_srgb.c: $(srcdir)/format_srgb.py
+	$(AM_V_GEN) $(PYTHON2) $< > $@
 
 ### Tests:
 
diff --git a/src/util/Makefile.sources b/src/util/Makefile.sources
index 5daa731ce1d..b99aa25e86e 100644
--- a/src/util/Makefile.sources
+++ b/src/util/Makefile.sources
@@ -1,3 +1,6 @@
 MESA_UTIL_FILES :=	\
 	hash_table.c	\
 	ralloc.c
+
+MESA_UTIL_GENERATED_FILES = \
+	format_srgb.c
diff --git a/src/util/SConscript b/src/util/SConscript
index 8afeedaf5e9..84803c016d9 100644
--- a/src/util/SConscript
+++ b/src/util/SConscript
@@ -11,10 +11,20 @@ env.Prepend(CPPPATH = [
     '#src/util',
 ])
 
+env.CodeGenerate(
+    target = 'format_srgb.c',
+    script = 'format_srgb.py',
+    source = [],
+    command = python_cmd + ' $SCRIPT > $TARGET'
+)
+
 # parse Makefile.sources
 source_lists = env.ParseSourceList('Makefile.sources')
 
-mesautil_sources = source_lists['MESA_UTIL_FILES']
+mesautil_sources = (
+    source_lists['MESA_UTIL_FILES'] +
+    source_lists['MESA_UTIL_GENERATED_FILES']
+)
 
 mesautil = env.ConvenienceLibrary(
     target = 'mesautil',
diff --git a/src/util/format_srgb.h b/src/util/format_srgb.h
new file mode 100644
index 00000000000..8e92c38f461
--- /dev/null
+++ b/src/util/format_srgb.h
@@ -0,0 +1,133 @@
+/**************************************************************************
+ *
+ * Copyright 2010 VMware, Inc.
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
+ * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
+ * USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ **************************************************************************/
+
+/**
+ * @file
+ * SRGB translation.
+ *
+ * @author Brian Paul <[email protected]>
+ * @author Michal Krol <[email protected]>
+ * @author Jose Fonseca <[email protected]>
+ */
+
+#ifndef U_FORMAT_SRGB_H_
+#define U_FORMAT_SRGB_H_
+
+#include <stdint.h>
+
+extern const float
+util_format_srgb_8unorm_to_linear_float_table[256];
+
+extern const uint8_t
+util_format_srgb_to_linear_8unorm_table[256];
+
+extern const uint8_t
+util_format_linear_to_srgb_8unorm_table[256];
+
+extern const unsigned
+util_format_linear_to_srgb_helper_table[104];
+
+
+/**
+ * Convert a unclamped linear float to srgb value in the [0,255].
+ */
+static inline uint8_t
+util_format_linear_float_to_srgb_8unorm(float x)
+{
+   /*
+    * This is taken from https://gist.github.com/rygorous/2203834
+    * Use LUT and do linear interpolation.
+    */
+   union {
+      uint32_t ui;
+      float f;
+   } almostone, minval, f;
+   unsigned tab, bias, scale, t;
+
+   almostone.ui = 0x3f7fffff;
+   minval.ui = (127-13) << 23;
+
+   /*
+    * Clamp to [2^(-13), 1-eps]; these two values map to 0 and 1, respectively.
+    * The tests are carefully written so that NaNs map to 0, same as in the
+    * reference implementation.
+    */
+   if (!(x > minval.f))
+      x = minval.f;
+   if (x > almostone.f)
+      x = almostone.f;
+
+   /* Do the table lookup and unpack bias, scale */
+   f.f = x;
+   tab = util_format_linear_to_srgb_helper_table[(f.ui - minval.ui) >> 20];
+   bias = (tab >> 16) << 9;
+   scale = tab & 0xffff;
+
+   /* Grab next-highest mantissa bits and perform linear interpolation */
+   t = (f.ui >> 12) & 0xff;
+   return (uint8_t) ((bias + scale*t) >> 16);
+}
+
+
+/**
+ * Convert an 8-bit sRGB value from non-linear space to a
+ * linear RGB value in [0, 1].
+ * Implemented with a 256-entry lookup table.
+ */
+static inline float
+util_format_srgb_8unorm_to_linear_float(uint8_t x)
+{
+   return util_format_srgb_8unorm_to_linear_float_table[x];
+}
+
+
+/*
+ * XXX These 2 functions probably don't make a lot of sense (but lots
+ * of potential callers which most likely all don't make sense neither)
+ */
+
+/**
+ * Convert a 8bit normalized value from linear to srgb.
+ */
+static inline uint8_t
+util_format_linear_to_srgb_8unorm(uint8_t x)
+{
+   return util_format_linear_to_srgb_8unorm_table[x];
+}
+
+
+/**
+ * Convert a 8bit normalized value from srgb to linear.
+ */
+static inline uint8_t
+util_format_srgb_to_linear_8unorm(uint8_t x)
+{
+   return util_format_srgb_to_linear_8unorm_table[x];
+}
+
+
+#endif /* U_FORMAT_SRGB_H_ */
diff --git a/src/util/format_srgb.py b/src/util/format_srgb.py
new file mode 100644
index 00000000000..d5cbcf7646b
--- /dev/null
+++ b/src/util/format_srgb.py
@@ -0,0 +1,155 @@
+#!/usr/bin/env python
+
+CopyRight = '''
+/**************************************************************************
+ *
+ * Copyright 2010 VMware, Inc.
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL VMWARE AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ **************************************************************************/
+
+/**
+ * @file
+ * SRGB translation.
+ *
+ * @author Brian Paul <[email protected]>
+ * @author Michal Krol <[email protected]>
+ * @author Jose Fonseca <[email protected]>
+ */
+'''
+
+
+import math
+import struct
+
+
+def srgb_to_linear(x):
+    if x <= 0.04045:
+        return x / 12.92
+    else:
+        return math.pow((x + 0.055) / 1.055, 2.4)
+
+
+def linear_to_srgb(x):
+    if x >= 0.0031308:
+        return 1.055 * math.pow(x, 0.41666666) - 0.055
+    else:
+        return 12.92 * x
+
+
+def generate_srgb_tables():
+    print 'const float'
+    print 'util_format_srgb_8unorm_to_linear_float_table[256] = {'
+    for j in range(0, 256, 4):
+        print '   ',
+        for i in range(j, j + 4):
+            print '%.7e,' % (srgb_to_linear(i / 255.0),),
+        print
+    print '};'
+    print
+    print 'const uint8_t'
+    print 'util_format_srgb_to_linear_8unorm_table[256] = {'
+    for j in range(0, 256, 16):
+        print '   ',
+        for i in range(j, j + 16):
+            print '%3u,' % (int(srgb_to_linear(i / 255.0) * 255.0 + 0.5),),
+        print
+    print '};'
+    print
+    print 'const uint8_t'
+    print 'util_format_linear_to_srgb_8unorm_table[256] = {'
+    for j in range(0, 256, 16):
+        print '   ',
+        for i in range(j, j + 16):
+            print '%3u,' % (int(linear_to_srgb(i / 255.0) * 255.0 + 0.5),),
+        print
+    print '};'
+    print
+
+# calculate the table interpolation values used in float linear to unorm8 srgb
+    numexp = 13
+    mantissa_msb = 3
+# stepshift is just used to only use every x-th float to make things faster,
+# 5 is largest value which still gives exact same table as 0
+    stepshift = 5
+    nbuckets = numexp << mantissa_msb
+    bucketsize = (1 << (23 - mantissa_msb)) >> stepshift
+    mantshift = 12
+    valtable = []
+    sum_aa = float(bucketsize)
+    sum_ab = 0.0
+    sum_bb = 0.0
+    for i in range(0, bucketsize):
+        j = (i << stepshift) >> mantshift
+        sum_ab += j
+        sum_bb += j*j
+    inv_det = 1.0 / (sum_aa * sum_bb - sum_ab * sum_ab)
+
+    for bucket in range(0, nbuckets):
+        start = ((127 - numexp) << 23) + bucket*(bucketsize << stepshift)
+        sum_a = 0.0
+        sum_b = 0.0
+
+        for i in range(0, bucketsize):
+            j = (i << stepshift) >> mantshift
+            fint = start + (i << stepshift)
+            ffloat = struct.unpack('f', struct.pack('I', fint))[0]
+            val = linear_to_srgb(ffloat) * 255.0 + 0.5
+            sum_a += val
+            sum_b += j*val
+
+        solved_a = inv_det * (sum_bb*sum_a - sum_ab*sum_b)
+        solved_b = inv_det * (sum_aa*sum_b - sum_ab*sum_a)
+
+        scaled_a = solved_a * 65536.0 / 512.0
+        scaled_b = solved_b * 65536.0
+
+        int_a = int(scaled_a + 0.5)
+        int_b = int(scaled_b + 0.5)
+
+        valtable.append((int_a << 16) + int_b)
+
+    print 'const unsigned'
+    print 'util_format_linear_to_srgb_helper_table[104] = {'
+
+    for j in range(0, nbuckets, 4):
+        print '   ',
+        for i in range(j, j + 4):
+            print '0x%08x,' % (valtable[i],),
+        print
+    print '};'
+    print
+
+def main():
+    print '/* This file is autogenerated by u_format_srgb.py. Do not edit directly. */'
+    print
+    # This will print the copyright message on the top of this file
+    print CopyRight.strip()
+    print
+    print '#include "format_srgb.h"'
+    print
+    generate_srgb_tables()
+
+
+if __name__ == '__main__':
+    main()