/* vim: set ts=8 sw=8 noexpandtab: */
//  qcms
//  Copyright (C) 2009 Mozilla Corporation
//  Copyright (C) 1998-2007 Marti Maria
//
// 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, sublicense, 
// 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 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 
// NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS 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.

#include <stdlib.h>
#include <math.h>
#include <assert.h>
#include <string.h> //memcpy
#include "qcmsint.h"
#include "chain.h"
#include "halffloat.h"
#include "matrix.h"
#include "transform_util.h"

/* for MSVC, GCC, Intel, and Sun compilers */
#if defined(_M_IX86) || defined(__i386__) || defined(__i386) || defined(_M_AMD64) || defined(__x86_64__) || defined(__x86_64)
#define X86
#endif /* _M_IX86 || __i386__ || __i386 || _M_AMD64 || __x86_64__ || __x86_64 */

// Build a White point, primary chromas transfer matrix from RGB to CIE XYZ
// This is just an approximation, I am not handling all the non-linear
// aspects of the RGB to XYZ process, and assumming that the gamma correction
// has transitive property in the tranformation chain.
//
// the alghoritm:
//
//            - First I build the absolute conversion matrix using
//              primaries in XYZ. This matrix is next inverted
//            - Then I eval the source white point across this matrix
//              obtaining the coeficients of the transformation
//            - Then, I apply these coeficients to the original matrix
static struct matrix build_RGB_to_XYZ_transfer_matrix(qcms_CIE_xyY white, qcms_CIE_xyYTRIPLE primrs)
{
	struct matrix primaries;
	struct matrix primaries_invert;
	struct matrix result;
	struct vector white_point;
	struct vector coefs;

	double xn, yn;
	double xr, yr;
	double xg, yg;
	double xb, yb;

	xn = white.x;
	yn = white.y;

	if (yn == 0.0)
		return matrix_invalid();

	xr = primrs.red.x;
	yr = primrs.red.y;
	xg = primrs.green.x;
	yg = primrs.green.y;
	xb = primrs.blue.x;
	yb = primrs.blue.y;

	primaries.m[0][0] = xr;
	primaries.m[0][1] = xg;
	primaries.m[0][2] = xb;

	primaries.m[1][0] = yr;
	primaries.m[1][1] = yg;
	primaries.m[1][2] = yb;

	primaries.m[2][0] = 1 - xr - yr;
	primaries.m[2][1] = 1 - xg - yg;
	primaries.m[2][2] = 1 - xb - yb;
	primaries.invalid = false;

	white_point.v[0] = xn/yn;
	white_point.v[1] = 1.;
	white_point.v[2] = (1.0-xn-yn)/yn;

	primaries_invert = matrix_invert(primaries);

	coefs = matrix_eval(primaries_invert, white_point);

	result.m[0][0] = coefs.v[0]*xr;
	result.m[0][1] = coefs.v[1]*xg;
	result.m[0][2] = coefs.v[2]*xb;

	result.m[1][0] = coefs.v[0]*yr;
	result.m[1][1] = coefs.v[1]*yg;
	result.m[1][2] = coefs.v[2]*yb;

	result.m[2][0] = coefs.v[0]*(1.-xr-yr);
	result.m[2][1] = coefs.v[1]*(1.-xg-yg);
	result.m[2][2] = coefs.v[2]*(1.-xb-yb);
	result.invalid = primaries_invert.invalid;

	return result;
}

struct CIE_XYZ {
	double X;
	double Y;
	double Z;
};

/* CIE Illuminant D50 */
static const struct CIE_XYZ D50_XYZ = {
	0.9642,
	1.0000,
	0.8249
};

/* from lcms: xyY2XYZ()
 * corresponds to argyll: icmYxy2XYZ() */
static struct CIE_XYZ xyY2XYZ(qcms_CIE_xyY source)
{
	struct CIE_XYZ dest;
	dest.X = (source.x / source.y) * source.Y;
	dest.Y = source.Y;
	dest.Z = ((1 - source.x - source.y) / source.y) * source.Y;
	return dest;
}

/* from lcms: ComputeChromaticAdaption */
// Compute chromatic adaption matrix using chad as cone matrix
static struct matrix
compute_chromatic_adaption(struct CIE_XYZ source_white_point,
                           struct CIE_XYZ dest_white_point,
                           struct matrix chad)
{
	struct matrix chad_inv;
	struct vector cone_source_XYZ, cone_source_rgb;
	struct vector cone_dest_XYZ, cone_dest_rgb;
	struct matrix cone, tmp;

	tmp = chad;
	chad_inv = matrix_invert(tmp);

	cone_source_XYZ.v[0] = source_white_point.X;
	cone_source_XYZ.v[1] = source_white_point.Y;
	cone_source_XYZ.v[2] = source_white_point.Z;

	cone_dest_XYZ.v[0] = dest_white_point.X;
	cone_dest_XYZ.v[1] = dest_white_point.Y;
	cone_dest_XYZ.v[2] = dest_white_point.Z;

	cone_source_rgb = matrix_eval(chad, cone_source_XYZ);
	cone_dest_rgb   = matrix_eval(chad, cone_dest_XYZ);

	cone.m[0][0] = cone_dest_rgb.v[0]/cone_source_rgb.v[0];
	cone.m[0][1] = 0;
	cone.m[0][2] = 0;
	cone.m[1][0] = 0;
	cone.m[1][1] = cone_dest_rgb.v[1]/cone_source_rgb.v[1];
	cone.m[1][2] = 0;
	cone.m[2][0] = 0;
	cone.m[2][1] = 0;
	cone.m[2][2] = cone_dest_rgb.v[2]/cone_source_rgb.v[2];
	cone.invalid = false;

	// Normalize
	return matrix_multiply(chad_inv, matrix_multiply(cone, chad));
}

/* from lcms: cmsAdaptionMatrix */
// Returns the final chrmatic adaptation from illuminant FromIll to Illuminant ToIll
// Bradford is assumed
static struct matrix
adaption_matrix(struct CIE_XYZ source_illumination, struct CIE_XYZ target_illumination)
{
#if defined (_MSC_VER)
#pragma warning(push)
/* Disable double to float truncation warning 4305 */
#pragma warning(disable:4305)
#endif
	struct matrix lam_rigg = {{ // Bradford matrix
	                         {  0.8951,  0.2664, -0.1614 },
	                         { -0.7502,  1.7135,  0.0367 },
	                         {  0.0389, -0.0685,  1.0296 }
	                         }};
#if defined (_MSC_VER)
/* Restore warnings */
#pragma warning(pop)
#endif
	return compute_chromatic_adaption(source_illumination, target_illumination, lam_rigg);
}

/* from lcms: cmsAdaptMatrixToD50 */
static struct matrix adapt_matrix_to_D50(struct matrix r, qcms_CIE_xyY source_white_pt)
{
	struct CIE_XYZ Dn;
	struct matrix Bradford;

	if (source_white_pt.y == 0.0)
		return matrix_invalid();

	Dn = xyY2XYZ(source_white_pt);

	Bradford = adaption_matrix(Dn, D50_XYZ);
	return matrix_multiply(Bradford, r);
}

qcms_bool set_rgb_colorants(qcms_profile *profile, qcms_CIE_xyY white_point, qcms_CIE_xyYTRIPLE primaries)
{
	struct matrix colorants;
	colorants = build_RGB_to_XYZ_transfer_matrix(white_point, primaries);
	colorants = adapt_matrix_to_D50(colorants, white_point);

	if (colorants.invalid)
		return false;

	/* note: there's a transpose type of operation going on here */
	profile->redColorant.X = double_to_s15Fixed16Number(colorants.m[0][0]);
	profile->redColorant.Y = double_to_s15Fixed16Number(colorants.m[1][0]);
	profile->redColorant.Z = double_to_s15Fixed16Number(colorants.m[2][0]);

	profile->greenColorant.X = double_to_s15Fixed16Number(colorants.m[0][1]);
	profile->greenColorant.Y = double_to_s15Fixed16Number(colorants.m[1][1]);
	profile->greenColorant.Z = double_to_s15Fixed16Number(colorants.m[2][1]);

	profile->blueColorant.X = double_to_s15Fixed16Number(colorants.m[0][2]);
	profile->blueColorant.Y = double_to_s15Fixed16Number(colorants.m[1][2]);
	profile->blueColorant.Z = double_to_s15Fixed16Number(colorants.m[2][2]);

	return true;
}

#if 0
static void qcms_transform_data_rgb_out_pow(qcms_transform *transform, unsigned char *src, unsigned char *dest, size_t length, qcms_format_type output_format)
{
	const int r_out = output_format.r;
	const int b_out = output_format.b;

	int i;
	float (*mat)[4] = transform->matrix;
	for (i=0; i<length; i++) {
		unsigned char device_r = *src++;
		unsigned char device_g = *src++;
		unsigned char device_b = *src++;

		float linear_r = transform->input_gamma_table_r[device_r];
		float linear_g = transform->input_gamma_table_g[device_g];
		float linear_b = transform->input_gamma_table_b[device_b];

		float out_linear_r = mat[0][0]*linear_r + mat[1][0]*linear_g + mat[2][0]*linear_b;
		float out_linear_g = mat[0][1]*linear_r + mat[1][1]*linear_g + mat[2][1]*linear_b;
		float out_linear_b = mat[0][2]*linear_r + mat[1][2]*linear_g + mat[2][2]*linear_b;

		float out_device_r = pow(out_linear_r, transform->out_gamma_r);
		float out_device_g = pow(out_linear_g, transform->out_gamma_g);
		float out_device_b = pow(out_linear_b, transform->out_gamma_b);

		dest[r_out] = clamp_u8(out_device_r*255);
		dest[1]     = clamp_u8(out_device_g*255);
		dest[b_out] = clamp_u8(out_device_b*255);
		dest += 3;
	}
}
#endif

static void qcms_transform_data_gray_out_lut(qcms_transform *transform, unsigned char *src, unsigned char *dest, size_t length, qcms_format_type output_format)
{
	const int r_out = output_format.r;
	const int b_out = output_format.b;

	unsigned int i;
	for (i = 0; i < length; i++) {
		float out_device_r, out_device_g, out_device_b;
		unsigned char device = *src++;

		float linear = transform->input_gamma_table_gray[device];

		out_device_r = lut_interp_linear(linear, transform->output_gamma_lut_r, transform->output_gamma_lut_r_length);
		out_device_g = lut_interp_linear(linear, transform->output_gamma_lut_g, transform->output_gamma_lut_g_length);
		out_device_b = lut_interp_linear(linear, transform->output_gamma_lut_b, transform->output_gamma_lut_b_length);

		dest[r_out] = clamp_u8(out_device_r*255);
		dest[1]     = clamp_u8(out_device_g*255);
		dest[b_out] = clamp_u8(out_device_b*255);
		dest += 3;
	}
}

/* Alpha is not corrected.
   A rationale for this is found in Alvy Ray's "Should Alpha Be Nonlinear If
   RGB Is?" Tech Memo 17 (December 14, 1998).
	See: ftp://ftp.alvyray.com/Acrobat/17_Nonln.pdf
*/

static void qcms_transform_data_graya_out_lut(qcms_transform *transform, unsigned char *src, unsigned char *dest, size_t length, qcms_format_type output_format)
{
	const int r_out = output_format.r;
	const int b_out = output_format.b;

	unsigned int i;
	for (i = 0; i < length; i++) {
		float out_device_r, out_device_g, out_device_b;
		unsigned char device = *src++;
		unsigned char alpha = *src++;

		float linear = transform->input_gamma_table_gray[device];

		out_device_r = lut_interp_linear(linear, transform->output_gamma_lut_r, transform->output_gamma_lut_r_length);
		out_device_g = lut_interp_linear(linear, transform->output_gamma_lut_g, transform->output_gamma_lut_g_length);
		out_device_b = lut_interp_linear(linear, transform->output_gamma_lut_b, transform->output_gamma_lut_b_length);

		dest[r_out] = clamp_u8(out_device_r*255);
		dest[1]     = clamp_u8(out_device_g*255);
		dest[b_out] = clamp_u8(out_device_b*255);
		dest[3]     = alpha;
		dest += 4;
	}
}


static void qcms_transform_data_gray_out_precache(qcms_transform *transform, unsigned char *src, unsigned char *dest, size_t length, qcms_format_type output_format)
{
	const int r_out = output_format.r;
	const int b_out = output_format.b;

	unsigned int i;
	for (i = 0; i < length; i++) {
		unsigned char device = *src++;
		uint16_t gray;

		float linear = transform->input_gamma_table_gray[device];

		/* we could round here... */
		gray = linear * PRECACHE_OUTPUT_MAX;

		dest[r_out] = transform->output_table_r->data[gray];
		dest[1]     = transform->output_table_g->data[gray];
		dest[b_out] = transform->output_table_b->data[gray];
		dest += 3;
	}
}


static void qcms_transform_data_graya_out_precache(qcms_transform *transform, unsigned char *src, unsigned char *dest, size_t length, qcms_format_type output_format)
{
	const int r_out = output_format.r;
	const int b_out = output_format.b;

	unsigned int i;
	for (i = 0; i < length; i++) {
		unsigned char device = *src++;
		unsigned char alpha = *src++;
		uint16_t gray;

		float linear = transform->input_gamma_table_gray[device];

		/* we could round here... */
		gray = linear * PRECACHE_OUTPUT_MAX;

		dest[r_out] = transform->output_table_r->data[gray];
		dest[1]     = transform->output_table_g->data[gray];
		dest[b_out] = transform->output_table_b->data[gray];
		dest[3]     = alpha;
		dest += 4;
	}
}

static void qcms_transform_data_rgb_out_lut_precache(qcms_transform *transform, unsigned char *src, unsigned char *dest, size_t length, qcms_format_type output_format)
{
	const int r_out = output_format.r;
	const int b_out = output_format.b;

	unsigned int i;
	float (*mat)[4] = transform->matrix;
	for (i = 0; i < length; i++) {
		unsigned char device_r = *src++;
		unsigned char device_g = *src++;
		unsigned char device_b = *src++;
		uint16_t r, g, b;

		float linear_r = transform->input_gamma_table_r[device_r];
		float linear_g = transform->input_gamma_table_g[device_g];
		float linear_b = transform->input_gamma_table_b[device_b];

		float out_linear_r = mat[0][0]*linear_r + mat[1][0]*linear_g + mat[2][0]*linear_b;
		float out_linear_g = mat[0][1]*linear_r + mat[1][1]*linear_g + mat[2][1]*linear_b;
		float out_linear_b = mat[0][2]*linear_r + mat[1][2]*linear_g + mat[2][2]*linear_b;

		out_linear_r = clamp_float(out_linear_r);
		out_linear_g = clamp_float(out_linear_g);
		out_linear_b = clamp_float(out_linear_b);

		/* we could round here... */
		r = out_linear_r * PRECACHE_OUTPUT_MAX;
		g = out_linear_g * PRECACHE_OUTPUT_MAX;
		b = out_linear_b * PRECACHE_OUTPUT_MAX;

		dest[r_out] = transform->output_table_r->data[r];
		dest[1]     = transform->output_table_g->data[g];
		dest[b_out] = transform->output_table_b->data[b];
		dest += 3;
	}
}

static void qcms_transform_data_rgba_out_lut_precache(qcms_transform *transform, unsigned char *src, unsigned char *dest, size_t length, qcms_format_type output_format)
{
	const int r_out = output_format.r;
	const int b_out = output_format.b;

	unsigned int i;
	float (*mat)[4] = transform->matrix;
	for (i = 0; i < length; i++) {
		unsigned char device_r = *src++;
		unsigned char device_g = *src++;
		unsigned char device_b = *src++;
		unsigned char alpha = *src++;
		uint16_t r, g, b;

		float linear_r = transform->input_gamma_table_r[device_r];
		float linear_g = transform->input_gamma_table_g[device_g];
		float linear_b = transform->input_gamma_table_b[device_b];

		float out_linear_r = mat[0][0]*linear_r + mat[1][0]*linear_g + mat[2][0]*linear_b;
		float out_linear_g = mat[0][1]*linear_r + mat[1][1]*linear_g + mat[2][1]*linear_b;
		float out_linear_b = mat[0][2]*linear_r + mat[1][2]*linear_g + mat[2][2]*linear_b;

		out_linear_r = clamp_float(out_linear_r);
		out_linear_g = clamp_float(out_linear_g);
		out_linear_b = clamp_float(out_linear_b);

		/* we could round here... */
		r = out_linear_r * PRECACHE_OUTPUT_MAX;
		g = out_linear_g * PRECACHE_OUTPUT_MAX;
		b = out_linear_b * PRECACHE_OUTPUT_MAX;

		dest[r_out] = transform->output_table_r->data[r];
		dest[1]     = transform->output_table_g->data[g];
		dest[b_out] = transform->output_table_b->data[b];
		dest[3]     = alpha;
		dest += 4;
	}
}

// Not used
/* 
static void qcms_transform_data_clut(qcms_transform *transform, unsigned char *src, unsigned char *dest, size_t length, qcms_format_type output_format)
{
	const int r_out = output_format.r;
	const int b_out = output_format.b;

	unsigned int i;
	int xy_len = 1;
	int x_len = transform->grid_size;
	int len = x_len * x_len;
	float* r_table = transform->r_clut;
	float* g_table = transform->g_clut;
	float* b_table = transform->b_clut;
  
	for (i = 0; i < length; i++) {
		unsigned char in_r = *src++;
		unsigned char in_g = *src++;
		unsigned char in_b = *src++;
		float linear_r = in_r/255.0f, linear_g=in_g/255.0f, linear_b = in_b/255.0f;

		int x = floor(linear_r * (transform->grid_size-1));
		int y = floor(linear_g * (transform->grid_size-1));
		int z = floor(linear_b * (transform->grid_size-1));
		int x_n = ceil(linear_r * (transform->grid_size-1));
		int y_n = ceil(linear_g * (transform->grid_size-1));
		int z_n = ceil(linear_b * (transform->grid_size-1));
		float x_d = linear_r * (transform->grid_size-1) - x; 
		float y_d = linear_g * (transform->grid_size-1) - y;
		float z_d = linear_b * (transform->grid_size-1) - z; 

		float r_x1 = lerp(CLU(r_table,x,y,z), CLU(r_table,x_n,y,z), x_d);
		float r_x2 = lerp(CLU(r_table,x,y_n,z), CLU(r_table,x_n,y_n,z), x_d);
		float r_y1 = lerp(r_x1, r_x2, y_d);
		float r_x3 = lerp(CLU(r_table,x,y,z_n), CLU(r_table,x_n,y,z_n), x_d);
		float r_x4 = lerp(CLU(r_table,x,y_n,z_n), CLU(r_table,x_n,y_n,z_n), x_d);
		float r_y2 = lerp(r_x3, r_x4, y_d);
		float clut_r = lerp(r_y1, r_y2, z_d);

		float g_x1 = lerp(CLU(g_table,x,y,z), CLU(g_table,x_n,y,z), x_d);
		float g_x2 = lerp(CLU(g_table,x,y_n,z), CLU(g_table,x_n,y_n,z), x_d);
		float g_y1 = lerp(g_x1, g_x2, y_d);
		float g_x3 = lerp(CLU(g_table,x,y,z_n), CLU(g_table,x_n,y,z_n), x_d);
		float g_x4 = lerp(CLU(g_table,x,y_n,z_n), CLU(g_table,x_n,y_n,z_n), x_d);
		float g_y2 = lerp(g_x3, g_x4, y_d);
		float clut_g = lerp(g_y1, g_y2, z_d);

		float b_x1 = lerp(CLU(b_table,x,y,z), CLU(b_table,x_n,y,z), x_d);
		float b_x2 = lerp(CLU(b_table,x,y_n,z), CLU(b_table,x_n,y_n,z), x_d);
		float b_y1 = lerp(b_x1, b_x2, y_d);
		float b_x3 = lerp(CLU(b_table,x,y,z_n), CLU(b_table,x_n,y,z_n), x_d);
		float b_x4 = lerp(CLU(b_table,x,y_n,z_n), CLU(b_table,x_n,y_n,z_n), x_d);
		float b_y2 = lerp(b_x3, b_x4, y_d);
		float clut_b = lerp(b_y1, b_y2, z_d);

		dest[r_out] = clamp_u8(clut_r*255.0f);
		dest[1]     = clamp_u8(clut_g*255.0f);
		dest[b_out] = clamp_u8(clut_b*255.0f);
		dest += 3;
	}
}
*/

// Using lcms' tetra interpolation algorithm.
static void qcms_transform_data_tetra_clut_rgba(qcms_transform *transform, unsigned char *src, unsigned char *dest, size_t length, qcms_format_type output_format)
{
	const int r_out = output_format.r;
	const int b_out = output_format.b;

	unsigned int i;
	int xy_len = 1;
	int x_len = transform->grid_size;
	int len = x_len * x_len;
	float* r_table = transform->r_clut;
	float* g_table = transform->g_clut;
	float* b_table = transform->b_clut;
	float c0_r, c1_r, c2_r, c3_r;
	float c0_g, c1_g, c2_g, c3_g;
	float c0_b, c1_b, c2_b, c3_b;
	float clut_r, clut_g, clut_b;
	for (i = 0; i < length; i++) {
		unsigned char in_r = *src++;
		unsigned char in_g = *src++;
		unsigned char in_b = *src++;
		unsigned char in_a = *src++;
		float linear_r = in_r/255.0f, linear_g=in_g/255.0f, linear_b = in_b/255.0f;

		int x = floor(linear_r * (transform->grid_size-1));
		int y = floor(linear_g * (transform->grid_size-1));
		int z = floor(linear_b * (transform->grid_size-1));
		int x_n = ceil(linear_r * (transform->grid_size-1));
		int y_n = ceil(linear_g * (transform->grid_size-1));
		int z_n = ceil(linear_b * (transform->grid_size-1));
		float rx = linear_r * (transform->grid_size-1) - x; 
		float ry = linear_g * (transform->grid_size-1) - y;
		float rz = linear_b * (transform->grid_size-1) - z; 

		c0_r = CLU(r_table, x, y, z);
		c0_g = CLU(g_table, x, y, z);
		c0_b = CLU(b_table, x, y, z);

		if( rx >= ry ) {
			if (ry >= rz) { //rx >= ry && ry >= rz
				c1_r = CLU(r_table, x_n, y, z) - c0_r;
				c2_r = CLU(r_table, x_n, y_n, z) - CLU(r_table, x_n, y, z);
				c3_r = CLU(r_table, x_n, y_n, z_n) - CLU(r_table, x_n, y_n, z);
				c1_g = CLU(g_table, x_n, y, z) - c0_g;
				c2_g = CLU(g_table, x_n, y_n, z) - CLU(g_table, x_n, y, z);
				c3_g = CLU(g_table, x_n, y_n, z_n) - CLU(g_table, x_n, y_n, z);
				c1_b = CLU(b_table, x_n, y, z) - c0_b;
				c2_b = CLU(b_table, x_n, y_n, z) - CLU(b_table, x_n, y, z);
				c3_b = CLU(b_table, x_n, y_n, z_n) - CLU(b_table, x_n, y_n, z);
			} else { 
				if (rx >= rz) { //rx >= rz && rz >= ry
					c1_r = CLU(r_table, x_n, y, z) - c0_r;
					c2_r = CLU(r_table, x_n, y_n, z_n) - CLU(r_table, x_n, y, z_n);
					c3_r = CLU(r_table, x_n, y, z_n) - CLU(r_table, x_n, y, z);
					c1_g = CLU(g_table, x_n, y, z) - c0_g;
					c2_g = CLU(g_table, x_n, y_n, z_n) - CLU(g_table, x_n, y, z_n);
					c3_g = CLU(g_table, x_n, y, z_n) - CLU(g_table, x_n, y, z);
					c1_b = CLU(b_table, x_n, y, z) - c0_b;
					c2_b = CLU(b_table, x_n, y_n, z_n) - CLU(b_table, x_n, y, z_n);
					c3_b = CLU(b_table, x_n, y, z_n) - CLU(b_table, x_n, y, z);
				} else { //rz > rx && rx >= ry
					c1_r = CLU(r_table, x_n, y, z_n) - CLU(r_table, x, y, z_n);
					c2_r = CLU(r_table, x_n, y_n, z_n) - CLU(r_table, x_n, y, z_n);
					c3_r = CLU(r_table, x, y, z_n) - c0_r;
					c1_g = CLU(g_table, x_n, y, z_n) - CLU(g_table, x, y, z_n);
					c2_g = CLU(g_table, x_n, y_n, z_n) - CLU(g_table, x_n, y, z_n);
					c3_g = CLU(g_table, x, y, z_n) - c0_g;
					c1_b = CLU(b_table, x_n, y, z_n) - CLU(b_table, x, y, z_n);
					c2_b = CLU(b_table, x_n, y_n, z_n) - CLU(b_table, x_n, y, z_n);
					c3_b = CLU(b_table, x, y, z_n) - c0_b;
				}
			}
		} else {
			if (rx >= rz) { //ry > rx && rx >= rz
				c1_r = CLU(r_table, x_n, y_n, z) - CLU(r_table, x, y_n, z);
				c2_r = CLU(r_table, x, y_n, z) - c0_r;
				c3_r = CLU(r_table, x_n, y_n, z_n) - CLU(r_table, x_n, y_n, z);
				c1_g = CLU(g_table, x_n, y_n, z) - CLU(g_table, x, y_n, z);
				c2_g = CLU(g_table, x, y_n, z) - c0_g;
				c3_g = CLU(g_table, x_n, y_n, z_n) - CLU(g_table, x_n, y_n, z);
				c1_b = CLU(b_table, x_n, y_n, z) - CLU(b_table, x, y_n, z);
				c2_b = CLU(b_table, x, y_n, z) - c0_b;
				c3_b = CLU(b_table, x_n, y_n, z_n) - CLU(b_table, x_n, y_n, z);
			} else {
				if (ry >= rz) { //ry >= rz && rz > rx 
					c1_r = CLU(r_table, x_n, y_n, z_n) - CLU(r_table, x, y_n, z_n);
					c2_r = CLU(r_table, x, y_n, z) - c0_r;
					c3_r = CLU(r_table, x, y_n, z_n) - CLU(r_table, x, y_n, z);
					c1_g = CLU(g_table, x_n, y_n, z_n) - CLU(g_table, x, y_n, z_n);
					c2_g = CLU(g_table, x, y_n, z) - c0_g;
					c3_g = CLU(g_table, x, y_n, z_n) - CLU(g_table, x, y_n, z);
					c1_b = CLU(b_table, x_n, y_n, z_n) - CLU(b_table, x, y_n, z_n);
					c2_b = CLU(b_table, x, y_n, z) - c0_b;
					c3_b = CLU(b_table, x, y_n, z_n) - CLU(b_table, x, y_n, z);
				} else { //rz > ry && ry > rx
					c1_r = CLU(r_table, x_n, y_n, z_n) - CLU(r_table, x, y_n, z_n);
					c2_r = CLU(r_table, x, y_n, z_n) - CLU(r_table, x, y, z_n);
					c3_r = CLU(r_table, x, y, z_n) - c0_r;
					c1_g = CLU(g_table, x_n, y_n, z_n) - CLU(g_table, x, y_n, z_n);
					c2_g = CLU(g_table, x, y_n, z_n) - CLU(g_table, x, y, z_n);
					c3_g = CLU(g_table, x, y, z_n) - c0_g;
					c1_b = CLU(b_table, x_n, y_n, z_n) - CLU(b_table, x, y_n, z_n);
					c2_b = CLU(b_table, x, y_n, z_n) - CLU(b_table, x, y, z_n);
					c3_b = CLU(b_table, x, y, z_n) - c0_b;
				}
			}
		}
				
		clut_r = c0_r + c1_r*rx + c2_r*ry + c3_r*rz;
		clut_g = c0_g + c1_g*rx + c2_g*ry + c3_g*rz;
		clut_b = c0_b + c1_b*rx + c2_b*ry + c3_b*rz;

		dest[r_out] = clamp_u8(clut_r*255.0f);
		dest[1]     = clamp_u8(clut_g*255.0f);
		dest[b_out] = clamp_u8(clut_b*255.0f);
		dest[3]     = in_a;
		dest += 4;
	}
}

// Using lcms' tetra interpolation code.
static void qcms_transform_data_tetra_clut(qcms_transform *transform, unsigned char *src, unsigned char *dest, size_t length, qcms_format_type output_format)
{
	const int r_out = output_format.r;
	const int b_out = output_format.b;

	unsigned int i;
	int xy_len = 1;
	int x_len = transform->grid_size;
	int len = x_len * x_len;
	float* r_table = transform->r_clut;
	float* g_table = transform->g_clut;
	float* b_table = transform->b_clut;
	float c0_r, c1_r, c2_r, c3_r;
	float c0_g, c1_g, c2_g, c3_g;
	float c0_b, c1_b, c2_b, c3_b;
	float clut_r, clut_g, clut_b;
	for (i = 0; i < length; i++) {
		unsigned char in_r = *src++;
		unsigned char in_g = *src++;
		unsigned char in_b = *src++;
		float linear_r = in_r/255.0f, linear_g=in_g/255.0f, linear_b = in_b/255.0f;

		int x = floor(linear_r * (transform->grid_size-1));
		int y = floor(linear_g * (transform->grid_size-1));
		int z = floor(linear_b * (transform->grid_size-1));
		int x_n = ceil(linear_r * (transform->grid_size-1));
		int y_n = ceil(linear_g * (transform->grid_size-1));
		int z_n = ceil(linear_b * (transform->grid_size-1));
		float rx = linear_r * (transform->grid_size-1) - x; 
		float ry = linear_g * (transform->grid_size-1) - y;
		float rz = linear_b * (transform->grid_size-1) - z; 

		c0_r = CLU(r_table, x, y, z);
		c0_g = CLU(g_table, x, y, z);
		c0_b = CLU(b_table, x, y, z);

		if( rx >= ry ) {
			if (ry >= rz) { //rx >= ry && ry >= rz
				c1_r = CLU(r_table, x_n, y, z) - c0_r;
				c2_r = CLU(r_table, x_n, y_n, z) - CLU(r_table, x_n, y, z);
				c3_r = CLU(r_table, x_n, y_n, z_n) - CLU(r_table, x_n, y_n, z);
				c1_g = CLU(g_table, x_n, y, z) - c0_g;
				c2_g = CLU(g_table, x_n, y_n, z) - CLU(g_table, x_n, y, z);
				c3_g = CLU(g_table, x_n, y_n, z_n) - CLU(g_table, x_n, y_n, z);
				c1_b = CLU(b_table, x_n, y, z) - c0_b;
				c2_b = CLU(b_table, x_n, y_n, z) - CLU(b_table, x_n, y, z);
				c3_b = CLU(b_table, x_n, y_n, z_n) - CLU(b_table, x_n, y_n, z);
			} else { 
				if (rx >= rz) { //rx >= rz && rz >= ry
					c1_r = CLU(r_table, x_n, y, z) - c0_r;
					c2_r = CLU(r_table, x_n, y_n, z_n) - CLU(r_table, x_n, y, z_n);
					c3_r = CLU(r_table, x_n, y, z_n) - CLU(r_table, x_n, y, z);
					c1_g = CLU(g_table, x_n, y, z) - c0_g;
					c2_g = CLU(g_table, x_n, y_n, z_n) - CLU(g_table, x_n, y, z_n);
					c3_g = CLU(g_table, x_n, y, z_n) - CLU(g_table, x_n, y, z);
					c1_b = CLU(b_table, x_n, y, z) - c0_b;
					c2_b = CLU(b_table, x_n, y_n, z_n) - CLU(b_table, x_n, y, z_n);
					c3_b = CLU(b_table, x_n, y, z_n) - CLU(b_table, x_n, y, z);
				} else { //rz > rx && rx >= ry
					c1_r = CLU(r_table, x_n, y, z_n) - CLU(r_table, x, y, z_n);
					c2_r = CLU(r_table, x_n, y_n, z_n) - CLU(r_table, x_n, y, z_n);
					c3_r = CLU(r_table, x, y, z_n) - c0_r;
					c1_g = CLU(g_table, x_n, y, z_n) - CLU(g_table, x, y, z_n);
					c2_g = CLU(g_table, x_n, y_n, z_n) - CLU(g_table, x_n, y, z_n);
					c3_g = CLU(g_table, x, y, z_n) - c0_g;
					c1_b = CLU(b_table, x_n, y, z_n) - CLU(b_table, x, y, z_n);
					c2_b = CLU(b_table, x_n, y_n, z_n) - CLU(b_table, x_n, y, z_n);
					c3_b = CLU(b_table, x, y, z_n) - c0_b;
				}
			}
		} else {
			if (rx >= rz) { //ry > rx && rx >= rz
				c1_r = CLU(r_table, x_n, y_n, z) - CLU(r_table, x, y_n, z);
				c2_r = CLU(r_table, x, y_n, z) - c0_r;
				c3_r = CLU(r_table, x_n, y_n, z_n) - CLU(r_table, x_n, y_n, z);
				c1_g = CLU(g_table, x_n, y_n, z) - CLU(g_table, x, y_n, z);
				c2_g = CLU(g_table, x, y_n, z) - c0_g;
				c3_g = CLU(g_table, x_n, y_n, z_n) - CLU(g_table, x_n, y_n, z);
				c1_b = CLU(b_table, x_n, y_n, z) - CLU(b_table, x, y_n, z);
				c2_b = CLU(b_table, x, y_n, z) - c0_b;
				c3_b = CLU(b_table, x_n, y_n, z_n) - CLU(b_table, x_n, y_n, z);
			} else {
				if (ry >= rz) { //ry >= rz && rz > rx 
					c1_r = CLU(r_table, x_n, y_n, z_n) - CLU(r_table, x, y_n, z_n);
					c2_r = CLU(r_table, x, y_n, z) - c0_r;
					c3_r = CLU(r_table, x, y_n, z_n) - CLU(r_table, x, y_n, z);
					c1_g = CLU(g_table, x_n, y_n, z_n) - CLU(g_table, x, y_n, z_n);
					c2_g = CLU(g_table, x, y_n, z) - c0_g;
					c3_g = CLU(g_table, x, y_n, z_n) - CLU(g_table, x, y_n, z);
					c1_b = CLU(b_table, x_n, y_n, z_n) - CLU(b_table, x, y_n, z_n);
					c2_b = CLU(b_table, x, y_n, z) - c0_b;
					c3_b = CLU(b_table, x, y_n, z_n) - CLU(b_table, x, y_n, z);
				} else { //rz > ry && ry > rx
					c1_r = CLU(r_table, x_n, y_n, z_n) - CLU(r_table, x, y_n, z_n);
					c2_r = CLU(r_table, x, y_n, z_n) - CLU(r_table, x, y, z_n);
					c3_r = CLU(r_table, x, y, z_n) - c0_r;
					c1_g = CLU(g_table, x_n, y_n, z_n) - CLU(g_table, x, y_n, z_n);
					c2_g = CLU(g_table, x, y_n, z_n) - CLU(g_table, x, y, z_n);
					c3_g = CLU(g_table, x, y, z_n) - c0_g;
					c1_b = CLU(b_table, x_n, y_n, z_n) - CLU(b_table, x, y_n, z_n);
					c2_b = CLU(b_table, x, y_n, z_n) - CLU(b_table, x, y, z_n);
					c3_b = CLU(b_table, x, y, z_n) - c0_b;
				}
			}
		}
				
		clut_r = c0_r + c1_r*rx + c2_r*ry + c3_r*rz;
		clut_g = c0_g + c1_g*rx + c2_g*ry + c3_g*rz;
		clut_b = c0_b + c1_b*rx + c2_b*ry + c3_b*rz;

		dest[r_out] = clamp_u8(clut_r*255.0f);
		dest[1]     = clamp_u8(clut_g*255.0f);
		dest[b_out] = clamp_u8(clut_b*255.0f);
		dest += 3;
	}
}

static void qcms_transform_data_rgb_out_lut(qcms_transform *transform, unsigned char *src, unsigned char *dest, size_t length, qcms_format_type output_format)
{
	const int r_out = output_format.r;
	const int b_out = output_format.b;

	unsigned int i;
	float (*mat)[4] = transform->matrix;
	for (i = 0; i < length; i++) {
		unsigned char device_r = *src++;
		unsigned char device_g = *src++;
		unsigned char device_b = *src++;
		float out_device_r, out_device_g, out_device_b;

		float linear_r = transform->input_gamma_table_r[device_r];
		float linear_g = transform->input_gamma_table_g[device_g];
		float linear_b = transform->input_gamma_table_b[device_b];

		float out_linear_r = mat[0][0]*linear_r + mat[1][0]*linear_g + mat[2][0]*linear_b;
		float out_linear_g = mat[0][1]*linear_r + mat[1][1]*linear_g + mat[2][1]*linear_b;
		float out_linear_b = mat[0][2]*linear_r + mat[1][2]*linear_g + mat[2][2]*linear_b;

		out_linear_r = clamp_float(out_linear_r);
		out_linear_g = clamp_float(out_linear_g);
		out_linear_b = clamp_float(out_linear_b);

		out_device_r = lut_interp_linear(out_linear_r, 
				transform->output_gamma_lut_r, transform->output_gamma_lut_r_length);
		out_device_g = lut_interp_linear(out_linear_g, 
				transform->output_gamma_lut_g, transform->output_gamma_lut_g_length);
		out_device_b = lut_interp_linear(out_linear_b, 
				transform->output_gamma_lut_b, transform->output_gamma_lut_b_length);

		dest[r_out] = clamp_u8(out_device_r*255);
		dest[1]     = clamp_u8(out_device_g*255);
		dest[b_out] = clamp_u8(out_device_b*255);
		dest += 3;
	}
}

static void qcms_transform_data_rgba_out_lut(qcms_transform *transform, unsigned char *src, unsigned char *dest, size_t length, qcms_format_type output_format)
{
	const int r_out = output_format.r;
	const int b_out = output_format.b;

	unsigned int i;
	float (*mat)[4] = transform->matrix;
	for (i = 0; i < length; i++) {
		unsigned char device_r = *src++;
		unsigned char device_g = *src++;
		unsigned char device_b = *src++;
		unsigned char alpha = *src++;
		float out_device_r, out_device_g, out_device_b;

		float linear_r = transform->input_gamma_table_r[device_r];
		float linear_g = transform->input_gamma_table_g[device_g];
		float linear_b = transform->input_gamma_table_b[device_b];

		float out_linear_r = mat[0][0]*linear_r + mat[1][0]*linear_g + mat[2][0]*linear_b;
		float out_linear_g = mat[0][1]*linear_r + mat[1][1]*linear_g + mat[2][1]*linear_b;
		float out_linear_b = mat[0][2]*linear_r + mat[1][2]*linear_g + mat[2][2]*linear_b;

		out_linear_r = clamp_float(out_linear_r);
		out_linear_g = clamp_float(out_linear_g);
		out_linear_b = clamp_float(out_linear_b);

		out_device_r = lut_interp_linear(out_linear_r, 
				transform->output_gamma_lut_r, transform->output_gamma_lut_r_length);
		out_device_g = lut_interp_linear(out_linear_g, 
				transform->output_gamma_lut_g, transform->output_gamma_lut_g_length);
		out_device_b = lut_interp_linear(out_linear_b, 
				transform->output_gamma_lut_b, transform->output_gamma_lut_b_length);

		dest[r_out] = clamp_u8(out_device_r*255);
		dest[1]     = clamp_u8(out_device_g*255);
		dest[b_out] = clamp_u8(out_device_b*255);
		dest[3]     = alpha;
		dest += 4;
	}
}

#if 0
static void qcms_transform_data_rgb_out_linear(qcms_transform *transform, unsigned char *src, unsigned char *dest, size_t length, qcms_format_type output_format)
{
	const int r_out = output_format.r;
	const int b_out = output_format.b;

	int i;
	float (*mat)[4] = transform->matrix;
	for (i = 0; i < length; i++) {
		unsigned char device_r = *src++;
		unsigned char device_g = *src++;
		unsigned char device_b = *src++;

		float linear_r = transform->input_gamma_table_r[device_r];
		float linear_g = transform->input_gamma_table_g[device_g];
		float linear_b = transform->input_gamma_table_b[device_b];

		float out_linear_r = mat[0][0]*linear_r + mat[1][0]*linear_g + mat[2][0]*linear_b;
		float out_linear_g = mat[0][1]*linear_r + mat[1][1]*linear_g + mat[2][1]*linear_b;
		float out_linear_b = mat[0][2]*linear_r + mat[1][2]*linear_g + mat[2][2]*linear_b;

		dest[r_out] = clamp_u8(out_linear_r*255);
		dest[1]     = clamp_u8(out_linear_g*255);
		dest[b_out] = clamp_u8(out_linear_b*255);
		dest += 3;
	}
}
#endif

/*
 * If users create and destroy objects on different threads, even if the same
 * objects aren't used on different threads at the same time, we can still run
 * in to trouble with refcounts if they aren't atomic.
 *
 * This can lead to us prematurely deleting the precache if threads get unlucky
 * and write the wrong value to the ref count.
 */
static struct precache_output *precache_reference(struct precache_output *p)
{
	qcms_atomic_increment(p->ref_count);
	return p;
}

static struct precache_output *precache_create()
{
	struct precache_output *p = malloc(sizeof(struct precache_output));
	if (p)
		p->ref_count = 1;
	return p;
}

void precache_release(struct precache_output *p)
{
	if (qcms_atomic_decrement(p->ref_count) == 0) {
		free(p);
	}
}

#ifdef HAVE_POSIX_MEMALIGN
static qcms_transform *transform_alloc(void)
{
	qcms_transform *t;
	if (!posix_memalign(&t, 16, sizeof(*t))) {
		return t;
	} else {
		return NULL;
	}
}
static void transform_free(qcms_transform *t)
{
	free(t);
}
#else
static qcms_transform *transform_alloc(void)
{
	/* transform needs to be aligned on a 16byte boundrary */
	char *original_block = calloc(sizeof(qcms_transform) + sizeof(void*) + 16, 1);
	/* make room for a pointer to the block returned by calloc */
	void *transform_start = original_block + sizeof(void*);
	/* align transform_start */
	qcms_transform *transform_aligned = (qcms_transform*)(((uintptr_t)transform_start + 15) & ~0xf);

	/* store a pointer to the block returned by calloc so that we can free it later */
	void **(original_block_ptr) = (void**)transform_aligned;
	if (!original_block)
		return NULL;
	original_block_ptr--;
	*original_block_ptr = original_block;

	return transform_aligned;
}
static void transform_free(qcms_transform *t)
{
	/* get at the pointer to the unaligned block returned by calloc */
	void **p = (void**)t;
	p--;
	free(*p);
}
#endif

void qcms_transform_release(qcms_transform *t)
{
	/* ensure we only free the gamma tables once even if there are
	 * multiple references to the same data */

	if (t->output_table_r)
		precache_release(t->output_table_r);
	if (t->output_table_g)
		precache_release(t->output_table_g);
	if (t->output_table_b)
		precache_release(t->output_table_b);

	free(t->input_gamma_table_r);
	if (t->input_gamma_table_g != t->input_gamma_table_r)
		free(t->input_gamma_table_g);
	if (t->input_gamma_table_g != t->input_gamma_table_r &&
	    t->input_gamma_table_g != t->input_gamma_table_b)
		free(t->input_gamma_table_b);

	free(t->input_gamma_table_gray);

	free(t->output_gamma_lut_r);
	free(t->output_gamma_lut_g);
	free(t->output_gamma_lut_b);

	transform_free(t);
}

#ifdef X86
// Determine if we can build with SSE2 (this was partly copied from jmorecfg.h in
// mozilla/jpeg)
 // -------------------------------------------------------------------------
#if defined(_M_IX86) && defined(_MSC_VER)
#define HAS_CPUID
/* Get us a CPUID function. Avoid clobbering EBX because sometimes it's the PIC
   register - I'm not sure if that ever happens on windows, but cpuid isn't
   on the critical path so we just preserve the register to be safe and to be
   consistent with the non-windows version. */
static void cpuid(uint32_t fxn, uint32_t *a, uint32_t *b, uint32_t *c, uint32_t *d) {
       uint32_t a_, b_, c_, d_;
       __asm {
              xchg   ebx, esi
              mov    eax, fxn
              cpuid
              mov    a_, eax
              mov    b_, ebx
              mov    c_, ecx
              mov    d_, edx
              xchg   ebx, esi
       }
       *a = a_;
       *b = b_;
       *c = c_;
       *d = d_;
}
#elif (defined(__GNUC__) || defined(__SUNPRO_C)) && (defined(__i386__) || defined(__i386))
#define HAS_CPUID
/* Get us a CPUID function. We can't use ebx because it's the PIC register on
   some platforms, so we use ESI instead and save ebx to avoid clobbering it. */
static void cpuid(uint32_t fxn, uint32_t *a, uint32_t *b, uint32_t *c, uint32_t *d) {

	uint32_t a_, b_, c_, d_;
       __asm__ __volatile__ ("xchgl %%ebx, %%esi; cpuid; xchgl %%ebx, %%esi;" 
                             : "=a" (a_), "=S" (b_), "=c" (c_), "=d" (d_) : "a" (fxn));
	   *a = a_;
	   *b = b_;
	   *c = c_;
	   *d = d_;
}
#endif

// -------------------------Runtime SSEx Detection-----------------------------

/* MMX is always supported per
 *  Gecko v1.9.1 minimum CPU requirements */
#define SSE1_EDX_MASK (1UL << 25)
#define SSE2_EDX_MASK (1UL << 26)
#define SSE3_ECX_MASK (1UL <<  0)

static int sse_version_available(void)
{
#if defined(__x86_64__) || defined(__x86_64) || defined(_M_AMD64)
	/* we know at build time that 64-bit CPUs always have SSE2
	 * this tells the compiler that non-SSE2 branches will never be
	 * taken (i.e. OK to optimze away the SSE1 and non-SIMD code */
	return 2;
#elif defined(HAS_CPUID)
	static int sse_version = -1;
	uint32_t a, b, c, d;
	uint32_t function = 0x00000001;

	if (sse_version == -1) {
		sse_version = 0;
		cpuid(function, &a, &b, &c, &d);
		if (c & SSE3_ECX_MASK)
			sse_version = 3;
		else if (d & SSE2_EDX_MASK)
			sse_version = 2;
		else if (d & SSE1_EDX_MASK)
			sse_version = 1;
	}

	return sse_version;
#else
	return 0;
#endif
}
#endif

static const struct matrix bradford_matrix = {{	{ 0.8951f, 0.2664f,-0.1614f},
						{-0.7502f, 1.7135f, 0.0367f},
						{ 0.0389f,-0.0685f, 1.0296f}}, 
						false};

static const struct matrix bradford_matrix_inv = {{ { 0.9869929f,-0.1470543f, 0.1599627f},
						    { 0.4323053f, 0.5183603f, 0.0492912f},
						    {-0.0085287f, 0.0400428f, 0.9684867f}}, 
						    false};

// See ICCv4 E.3
struct matrix compute_whitepoint_adaption(float X, float Y, float Z) {
	float p = (0.96422f*bradford_matrix.m[0][0] + 1.000f*bradford_matrix.m[1][0] + 0.82521f*bradford_matrix.m[2][0]) /
		  (X*bradford_matrix.m[0][0]      + Y*bradford_matrix.m[1][0]      + Z*bradford_matrix.m[2][0]     );
	float y = (0.96422f*bradford_matrix.m[0][1] + 1.000f*bradford_matrix.m[1][1] + 0.82521f*bradford_matrix.m[2][1]) /
		  (X*bradford_matrix.m[0][1]      + Y*bradford_matrix.m[1][1]      + Z*bradford_matrix.m[2][1]     );
	float b = (0.96422f*bradford_matrix.m[0][2] + 1.000f*bradford_matrix.m[1][2] + 0.82521f*bradford_matrix.m[2][2]) /
		  (X*bradford_matrix.m[0][2]      + Y*bradford_matrix.m[1][2]      + Z*bradford_matrix.m[2][2]     );
	struct matrix white_adaption = {{ {p,0,0}, {0,y,0}, {0,0,b}}, false};
	return matrix_multiply( bradford_matrix_inv, matrix_multiply(white_adaption, bradford_matrix) );
}

void qcms_profile_precache_output_transform(qcms_profile *profile)
{
	/* we only support precaching on rgb profiles */
	if (profile->color_space != RGB_SIGNATURE)
		return;

	if (qcms_supports_iccv4) {
		/* don't precache since we will use the B2A LUT */
		if (profile->B2A0)
			return;

		/* don't precache since we will use the mBA LUT */
		if (profile->mBA)
			return;
	}

	/* don't precache if we do not have the TRC curves */
	if (!profile->redTRC || !profile->greenTRC || !profile->blueTRC)
		return;

	if (!profile->output_table_r) {
		profile->output_table_r = precache_create();
		if (profile->output_table_r &&
				!compute_precache(profile->redTRC, profile->output_table_r->data)) {
			precache_release(profile->output_table_r);
			profile->output_table_r = NULL;
		}
	}
	if (!profile->output_table_g) {
		profile->output_table_g = precache_create();
		if (profile->output_table_g &&
				!compute_precache(profile->greenTRC, profile->output_table_g->data)) {
			precache_release(profile->output_table_g);
			profile->output_table_g = NULL;
		}
	}
	if (!profile->output_table_b) {
		profile->output_table_b = precache_create();
		if (profile->output_table_b &&
				!compute_precache(profile->blueTRC, profile->output_table_b->data)) {
			precache_release(profile->output_table_b);
			profile->output_table_b = NULL;
		}
	}
}

/* Replace the current transformation with a LUT transformation using a given number of sample points */
qcms_transform* qcms_transform_precacheLUT_float(qcms_transform *transform, qcms_profile *in, qcms_profile *out, 
                                                 int samples, qcms_data_type in_type)
{
	/* The range between which 2 consecutive sample points can be used to interpolate */
	uint16_t x,y,z;
	uint32_t l;
	uint32_t lutSize = 3 * samples * samples * samples;
	float* src = NULL;
	float* dest = NULL;
	float* lut = NULL;
	float inverse;

	src = malloc(lutSize*sizeof(float));
	dest = malloc(lutSize*sizeof(float));

	if (src && dest) {
		/* Prepare a list of points we want to sample: x, y, z order */
		l = 0;
		inverse = 1 / (float)(samples-1);
		for (x = 0; x < samples; x++) {
			for (y = 0; y < samples; y++) {
				for (z = 0; z < samples; z++) {
					src[l++] = x * inverse; // r
					src[l++] = y * inverse; // g
					src[l++] = z * inverse; // b
				}
			}
		}

		lut = qcms_chain_transform(in, out, src, dest, lutSize);

		if (lut) {
			transform->r_clut = &lut[0]; // r
			transform->g_clut = &lut[1]; // g
			transform->b_clut = &lut[2]; // b
			transform->grid_size = samples;
			if (in_type == QCMS_DATA_RGBA_8) {
				transform->transform_fn = qcms_transform_data_tetra_clut_rgba;
			} else {
				transform->transform_fn = qcms_transform_data_tetra_clut;
			}
		}
	}

	// XXX: qcms_modular_transform_data may return the lut in either the src or the
	// dest buffer. If so, it must not be free-ed.
	if (src && lut != src) {
		free(src);
	}
	if (dest && lut != dest) {
		free(dest);
	}

	if (lut == NULL) {
		return NULL;
	}
	return transform;
}

/* Create a transform LUT using the given number of sample points. The transform LUT data is stored
   in the output (cube) in bgra format in zyx sample order. */
qcms_bool qcms_transform_create_LUT_zyx_bgra(qcms_profile *in, qcms_profile *out, qcms_intent intent,
                                             int samples, unsigned char* cube)
{
	uint16_t z,y,x;
	uint32_t l,index;
	uint32_t lutSize = 3 * samples * samples * samples;

	float* src = NULL;
	float* dest = NULL;
	float* lut = NULL;
	float inverse;

	src = malloc(lutSize*sizeof(float));
	dest = malloc(lutSize*sizeof(float));

	if (src && dest) {
		/* Prepare a list of points we want to sample: z, y, x order */
		l = 0;
		inverse = 1 / (float)(samples-1);
		for (z = 0; z < samples; z++) {
			for (y = 0; y < samples; y++) {
				for (x = 0; x < samples; x++) {
					src[l++] = x * inverse; // r
					src[l++] = y * inverse; // g
					src[l++] = z * inverse; // b
				}
			}
		}

		lut = qcms_chain_transform(in, out, src, dest, lutSize);

		if (lut) {
			index = l = 0;
			for (z = 0; z < samples; z++) {
				for (y = 0; y < samples; y++) {
					for (x = 0; x < samples; x++) {
						cube[index++] = (int)floorf(lut[l + 2] * 255.0f + 0.5f); // b
						cube[index++] = (int)floorf(lut[l + 1] * 255.0f + 0.5f); // g
						cube[index++] = (int)floorf(lut[l + 0] * 255.0f + 0.5f); // r
						cube[index++] = 255;                                     // a
						l += 3;
					}
				}
			}
		}
	}

	// XXX: qcms_modular_transform_data may return the lut data in either the src or
	// dest buffer so free src, dest, and lut with care.

	if (src && lut != src)
		free(src);
	if (dest && lut != dest)
		free(dest);

	if (lut) {
		free(lut);
		return true;
	}

	return false;
}

#define NO_MEM_TRANSFORM NULL

qcms_transform* qcms_transform_create(
		qcms_profile *in, qcms_data_type in_type,
		qcms_profile *out, qcms_data_type out_type,
		qcms_intent intent)
{
	bool precache = false;

        qcms_transform *transform = transform_alloc();
        if (!transform) {
		return NULL;
	}
	if (out_type != QCMS_DATA_RGB_8 &&
                out_type != QCMS_DATA_RGBA_8) {
            assert(0 && "output type");
	    transform_free(transform);
            return NULL;
        }

	transform->transform_flags = 0;

	if (out->output_table_r &&
			out->output_table_g &&
			out->output_table_b) {
		precache = true;
	}

	if (qcms_supports_iccv4 && (in->A2B0 || out->B2A0 || in->mAB || out->mAB)) {
		// Precache the transformation to a CLUT 33x33x33 in size.
		// 33 is used by many profiles and works well in pratice. 
		// This evenly divides 256 into blocks of 8x8x8.
		// TODO For transforming small data sets of about 200x200 or less
		// precaching should be avoided.
		qcms_transform *result = qcms_transform_precacheLUT_float(transform, in, out, 33, in_type);
		if (!result) {
            		assert(0 && "precacheLUT failed");
			transform_free(transform);
			return NULL;
		}
		return result;
	}

	if (precache) {
		transform->output_table_r = precache_reference(out->output_table_r);
		transform->output_table_g = precache_reference(out->output_table_g);
		transform->output_table_b = precache_reference(out->output_table_b);
	} else {
		if (!out->redTRC || !out->greenTRC || !out->blueTRC) {
			qcms_transform_release(transform);
			return NO_MEM_TRANSFORM;
		}
		build_output_lut(out->redTRC, &transform->output_gamma_lut_r, &transform->output_gamma_lut_r_length);
		build_output_lut(out->greenTRC, &transform->output_gamma_lut_g, &transform->output_gamma_lut_g_length);
		build_output_lut(out->blueTRC, &transform->output_gamma_lut_b, &transform->output_gamma_lut_b_length);
		if (!transform->output_gamma_lut_r || !transform->output_gamma_lut_g || !transform->output_gamma_lut_b) {
			qcms_transform_release(transform);
			return NO_MEM_TRANSFORM;
		}
	}

        if (in->color_space == RGB_SIGNATURE) {
		struct matrix in_matrix, out_matrix, result;

		if (in_type != QCMS_DATA_RGB_8 &&
                    in_type != QCMS_DATA_RGBA_8){
                	assert(0 && "input type");
			transform_free(transform);
                	return NULL;
            	}
		if (precache) {
#if defined(SSE2_ENABLE)
		    if (sse_version_available() >= 2) {
			    if (in_type == QCMS_DATA_RGB_8)
				    transform->transform_fn = qcms_transform_data_rgb_out_lut_sse2;
			    else
				    transform->transform_fn = qcms_transform_data_rgba_out_lut_sse2;
		    } else
#endif
			{
				if (in_type == QCMS_DATA_RGB_8)
					transform->transform_fn = qcms_transform_data_rgb_out_lut_precache;
				else
					transform->transform_fn = qcms_transform_data_rgba_out_lut_precache;
			}
		} else {
			if (in_type == QCMS_DATA_RGB_8)
				transform->transform_fn = qcms_transform_data_rgb_out_lut;
			else
				transform->transform_fn = qcms_transform_data_rgba_out_lut;
		}

		//XXX: avoid duplicating tables if we can
		transform->input_gamma_table_r = build_input_gamma_table(in->redTRC);
		transform->input_gamma_table_g = build_input_gamma_table(in->greenTRC);
		transform->input_gamma_table_b = build_input_gamma_table(in->blueTRC);
		if (!transform->input_gamma_table_r || !transform->input_gamma_table_g || !transform->input_gamma_table_b) {
			qcms_transform_release(transform);
			return NO_MEM_TRANSFORM;
		}


		/* build combined colorant matrix */
		in_matrix = build_colorant_matrix(in);
		out_matrix = build_colorant_matrix(out);
		out_matrix = matrix_invert(out_matrix);
		if (out_matrix.invalid) {
			qcms_transform_release(transform);
			return NULL;
		}
		result = matrix_multiply(out_matrix, in_matrix);

		/* store the results in column major mode
		 * this makes doing the multiplication with sse easier */
		transform->matrix[0][0] = result.m[0][0];
		transform->matrix[1][0] = result.m[0][1];
		transform->matrix[2][0] = result.m[0][2];
		transform->matrix[0][1] = result.m[1][0];
		transform->matrix[1][1] = result.m[1][1];
		transform->matrix[2][1] = result.m[1][2];
		transform->matrix[0][2] = result.m[2][0];
		transform->matrix[1][2] = result.m[2][1];
		transform->matrix[2][2] = result.m[2][2];

		/* Flag transform as matrix. */
		transform->transform_flags |= TRANSFORM_FLAG_MATRIX;

	} else if (in->color_space == GRAY_SIGNATURE) {
		if (in_type != QCMS_DATA_GRAY_8 &&
				in_type != QCMS_DATA_GRAYA_8){
			assert(0 && "input type");
			transform_free(transform);
			return NULL;
		}

		transform->input_gamma_table_gray = build_input_gamma_table(in->grayTRC);
		if (!transform->input_gamma_table_gray) {
			qcms_transform_release(transform);
			return NO_MEM_TRANSFORM;
		}

		if (precache) {
			if (in_type == QCMS_DATA_GRAY_8) {
				transform->transform_fn = qcms_transform_data_gray_out_precache;
			} else {
				transform->transform_fn = qcms_transform_data_graya_out_precache;
			}
		} else {
			if (in_type == QCMS_DATA_GRAY_8) {
				transform->transform_fn = qcms_transform_data_gray_out_lut;
			} else {
				transform->transform_fn = qcms_transform_data_graya_out_lut;
			}
		}
	} else {
		assert(0 && "unexpected colorspace");
		transform_free(transform);
		return NULL;
	}
	return transform;
}

/* __force_align_arg_pointer__ is an x86-only attribute, and gcc/clang warns on unused
 * attributes. Don't use this on ARM or AMD64. __has_attribute can detect the presence
 * of the attribute but is currently only supported by clang */
#if defined(__has_attribute)
#define HAS_FORCE_ALIGN_ARG_POINTER __has_attribute(__force_align_arg_pointer__)
#elif defined(__GNUC__) && defined(__i386__)
#define HAS_FORCE_ALIGN_ARG_POINTER 1
#else
#define HAS_FORCE_ALIGN_ARG_POINTER 0
#endif

#if HAS_FORCE_ALIGN_ARG_POINTER
/* we need this to avoid crashes when gcc assumes the stack is 128bit aligned */
__attribute__((__force_align_arg_pointer__))
#endif
void qcms_transform_data(qcms_transform *transform, void *src, void *dest, size_t length)
{
	static const struct _qcms_format_type output_rgbx = { 0, 2 };

	transform->transform_fn(transform, src, dest, length, output_rgbx);
}

void qcms_transform_data_type(qcms_transform *transform, void *src, void *dest, size_t length, qcms_output_type type)
{
	static const struct _qcms_format_type output_rgbx = { 0, 2 };
	static const struct _qcms_format_type output_bgrx = { 2, 0 };

	transform->transform_fn(transform, src, dest, length, type == QCMS_OUTPUT_BGRX ? output_bgrx : output_rgbx);
}

qcms_bool qcms_supports_iccv4;

void qcms_enable_iccv4()
{
	qcms_supports_iccv4 = true;
}

static inline qcms_bool transform_is_matrix(qcms_transform *t)
{
	return (t->transform_flags & TRANSFORM_FLAG_MATRIX) ? true : false;
}

qcms_bool qcms_transform_is_matrix(qcms_transform *t)
{
	return transform_is_matrix(t);
}

float qcms_transform_get_matrix(qcms_transform *t, unsigned i, unsigned j)
{
	assert(transform_is_matrix(t) && i < 3 && j < 3);

	// Return transform matrix element in row major order (permute i and j)

	return t->matrix[j][i];
}

static inline qcms_bool supported_trc_type(qcms_trc_type type)
{
	return type == QCMS_TRC_HALF_FLOAT;
}

const uint16_t half_float_one = 0x3c00;

size_t qcms_transform_get_input_trc_rgba(qcms_transform *t, qcms_profile *in, qcms_trc_type type, unsigned short *data)
{
	const size_t size = 256; // The input gamma tables always have 256 entries.

	size_t i;

	if (in->color_space != RGB_SIGNATURE || !supported_trc_type(type))
		return 0;

	// qcms_profile *in is assumed to be the profile on the input-side of the color transform t.
	// When a transform is created, the input gamma curve data is stored in the transform ...

	if (!t->input_gamma_table_r || !t->input_gamma_table_g || !t->input_gamma_table_b)
		return 0;

	// Report the size if no output data is requested. This allows callers to first work out the
	// the curve size, then provide allocated memory sufficient to store the curve rgba data.

	if (!data)
		return size;

	for (i = 0; i < size; ++i) {
		*data++ = float_to_half_float(t->input_gamma_table_r[i]); // r
		*data++ = float_to_half_float(t->input_gamma_table_g[i]); // g
		*data++ = float_to_half_float(t->input_gamma_table_b[i]); // b
		*data++ = half_float_one;                                 // a
	}

	return size;
}

const float inverse65535 = (float) (1.0 / 65535.0);

size_t qcms_transform_get_output_trc_rgba(qcms_transform *t, qcms_profile *out, qcms_trc_type type, unsigned short *data)
{
	size_t size, i;

	if (out->color_space != RGB_SIGNATURE || !supported_trc_type(type))
		return 0;

	// qcms_profile *out is assumed to be the profile on the output-side of the transform t.
	// If the transform output gamma curves need building, do that. They're usually built when
	// the transform was created, but sometimes not due to the output gamma precache ...

	if (!out->redTRC || !out->greenTRC || !out->blueTRC)
		return 0;
	if (!t->output_gamma_lut_r)
		build_output_lut(out->redTRC, &t->output_gamma_lut_r, &t->output_gamma_lut_r_length);
	if (!t->output_gamma_lut_g)
		build_output_lut(out->greenTRC, &t->output_gamma_lut_g, &t->output_gamma_lut_g_length);
	if (!t->output_gamma_lut_b)
		build_output_lut(out->blueTRC, &t->output_gamma_lut_b, &t->output_gamma_lut_b_length);

	if (!t->output_gamma_lut_r || !t->output_gamma_lut_g || !t->output_gamma_lut_b)
		return 0;

	// Output gamma tables should have the same size and should have 4096 entries at most (the
	// minimum is 256). Larger tables are rare and ignored here: fail by returning 0.

	size = t->output_gamma_lut_r_length;
	if (size != t->output_gamma_lut_g_length)
		return 0;
	if (size != t->output_gamma_lut_b_length)
		return 0;
	if (size < 256 || size > 4096)
		return 0;

	// Report the size if no output data is requested. This allows callers to first work out the
	// the curve size, then provide allocated memory sufficient to store the curve rgba data.

	if (!data)
		return size;

	for (i = 0; i < size; ++i) {
		*data++ = float_to_half_float(t->output_gamma_lut_r[i] * inverse65535); // r
		*data++ = float_to_half_float(t->output_gamma_lut_g[i] * inverse65535); // g
		*data++ = float_to_half_float(t->output_gamma_lut_b[i] * inverse65535); // b
		*data++ = half_float_one;                                               // a
	}

	return size;
}