Cycles OpenCL: a few fixes to get things compiling after kernel changes,

for Apple OpenCL on OS X 10.8 and simple AO render.

Also environment variable CYCLES_OPENCL_TEST can now be set to CPU, GPU,
ACCELERATOR, DEFAULT or ALL values to test particuler devices.

6 files changed, 49 insertions, 26 deletions

diff --git a/intern/cycles/device/device_opencl.cpp b/intern/cycles/device/device_opencl.cpp
index 3ab5d9ee69b..a9b062474f7 100644
--- a/intern/cycles/device/device_opencl.cpp
+++ b/intern/cycles/device/device_opencl.cpp
@@ -40,6 +40,26 @@ CCL_NAMESPACE_BEGIN
 
 #define CL_MEM_PTR(p) ((cl_mem)(unsigned long)(p))
 
+static cl_device_type opencl_device_type()
+{
+	char *device = getenv("CYCLES_OPENCL_TEST");
+
+	if(device) {
+		if(strcmp(device, "ALL") == 0)
+			return CL_DEVICE_TYPE_ALL;
+		else if(strcmp(device, "DEFAULT") == 0)
+			return CL_DEVICE_TYPE_DEFAULT;
+		else if(strcmp(device, "CPU") == 0)
+			return CL_DEVICE_TYPE_CPU;
+		else if(strcmp(device, "GPU") == 0)
+			return CL_DEVICE_TYPE_GPU;
+		else if(strcmp(device, "ACCELERATOR") == 0)
+			return CL_DEVICE_TYPE_ACCELERATOR;
+	}
+
+	return CL_DEVICE_TYPE_GPU|CL_DEVICE_TYPE_ACCELERATOR;
+}
+
 class OpenCLDevice : public Device
 {
 public:
@@ -181,7 +201,7 @@ public:
 		vector<cl_device_id> device_ids;
 		cl_uint num_devices;
 
-		if(opencl_error(clGetDeviceIDs(cpPlatform, CL_DEVICE_TYPE_GPU|CL_DEVICE_TYPE_ACCELERATOR, 0, NULL, &num_devices)))
+		if(opencl_error(clGetDeviceIDs(cpPlatform, opencl_device_type(), 0, NULL, &num_devices)))
 			return;
 
 		if(info.num > num_devices) {
@@ -194,7 +214,7 @@ public:
 
 		device_ids.resize(num_devices);
 		
-		if(opencl_error(clGetDeviceIDs(cpPlatform, CL_DEVICE_TYPE_GPU|CL_DEVICE_TYPE_ACCELERATOR, num_devices, &device_ids[0], NULL)))
+		if(opencl_error(clGetDeviceIDs(cpPlatform, opencl_device_type(), num_devices, &device_ids[0], NULL)))
 			return;
 
 		cdDevice = device_ids[info.num];
@@ -306,10 +326,10 @@ public:
 			build_options += "-D__KERNEL_OPENCL_NVIDIA__ -cl-nv-maxrregcount=24 -cl-nv-verbose ";
 
 		else if(platform_name == "Apple")
-			build_options += "-D__CL_NO_FLOAT3__ -D__KERNEL_OPENCL_APPLE__ ";
+			build_options += "-D__KERNEL_OPENCL_APPLE__ -Wno-missing-prototypes";
 
 		else if(platform_name == "AMD Accelerated Parallel Processing")
-			build_options += "-D__CL_NO_FLOAT3__ -D__KERNEL_OPENCL_AMD__ ";
+			build_options += "-D__KERNEL_OPENCL_AMD__ ";
 
 		return build_options;
 	}
@@ -754,12 +774,12 @@ void device_opencl_info(vector<DeviceInfo>& devices)
 	if(clGetPlatformIDs(num_platforms, &platform_ids[0], NULL) != CL_SUCCESS)
 		return;
 
-	if(clGetDeviceIDs(platform_ids[0], CL_DEVICE_TYPE_GPU|CL_DEVICE_TYPE_ACCELERATOR, 0, NULL, &num_devices) != CL_SUCCESS || num_devices == 0)
+	if(clGetDeviceIDs(platform_ids[0], opencl_device_type(), 0, NULL, &num_devices) != CL_SUCCESS || num_devices == 0)
 		return;
 	
 	device_ids.resize(num_devices);
 
-	if(clGetDeviceIDs(platform_ids[0], CL_DEVICE_TYPE_GPU|CL_DEVICE_TYPE_ACCELERATOR, num_devices, &device_ids[0], NULL) != CL_SUCCESS)
+	if(clGetDeviceIDs(platform_ids[0], opencl_device_type(), num_devices, &device_ids[0], NULL) != CL_SUCCESS)
 		return;
 	
 	/* add devices */
diff --git a/intern/cycles/kernel/kernel_compat_opencl.h b/intern/cycles/kernel/kernel_compat_opencl.h
index 999820891b2..6c41bfa5521 100644
--- a/intern/cycles/kernel/kernel_compat_opencl.h
+++ b/intern/cycles/kernel/kernel_compat_opencl.h
@@ -26,6 +26,10 @@
 #define CCL_NAMESPACE_BEGIN
 #define CCL_NAMESPACE_END
 
+#ifdef __KERNEL_OPENCL_AMD__
+#define __CL_NO_FLOAT3__
+#endif
+
 #ifdef __CL_NO_FLOAT3__
 #define float3 float4
 #endif
diff --git a/intern/cycles/kernel/kernel_emission.h b/intern/cycles/kernel/kernel_emission.h
index 6f7a3e71d27..bae1b4743e4 100644
--- a/intern/cycles/kernel/kernel_emission.h
+++ b/intern/cycles/kernel/kernel_emission.h
@@ -52,7 +52,7 @@ __device_noinline float3 direct_emissive_eval(KernelGlobals *kg, float rando,
 			shader_setup_from_sample(kg, &sd, ls->P, ls->Ng, I, ls->shader, ls->object, ls->prim, u, v, t, time, ls->prim);
 		else
 #endif
-			shader_setup_from_sample(kg, &sd, ls->P, ls->Ng, I, ls->shader, ls->object, ls->prim, u, v, t, time);
+			shader_setup_from_sample(kg, &sd, ls->P, ls->Ng, I, ls->shader, ls->object, ls->prim, u, v, t, time, ~0);
 
 		ls->Ng = sd.Ng;
 
diff --git a/intern/cycles/kernel/kernel_light.h b/intern/cycles/kernel/kernel_light.h
index f6fbd3599ad..2eab293963a 100644
--- a/intern/cycles/kernel/kernel_light.h
+++ b/intern/cycles/kernel/kernel_light.h
@@ -430,11 +430,9 @@ __device void object_transform_light_sample(KernelGlobals *kg, LightSample *ls,
 	/* instance transform */
 	if(object >= 0) {
 #ifdef __OBJECT_MOTION__
-		Transform itfm;
-		Transform tfm = object_fetch_transform_motion_test(kg, object, time, &itfm);
+		Transform tfm = object_fetch_transform_motion_test(kg, object, time, NULL);
 #else
 		Transform tfm = object_fetch_transform(kg, object, OBJECT_TRANSFORM);
-		Transform itfm = object_fetch_transform(kg, object, OBJECT_INVERSE_TRANSFORM);
 #endif
 
 		ls->P = transform_point(&tfm, ls->P);
diff --git a/intern/cycles/kernel/kernel_shader.h b/intern/cycles/kernel/kernel_shader.h
index 21152efff4c..908c5a91768 100644
--- a/intern/cycles/kernel/kernel_shader.h
+++ b/intern/cycles/kernel/kernel_shader.h
@@ -262,7 +262,7 @@ __device_inline void shader_setup_from_subsurface(KernelGlobals *kg, ShaderData
 
 __device_noinline void shader_setup_from_sample(KernelGlobals *kg, ShaderData *sd,
 	const float3 P, const float3 Ng, const float3 I,
-	int shader, int object, int prim, float u, float v, float t, float time, int segment = ~0)
+	int shader, int object, int prim, float u, float v, float t, float time, int segment)
 {
 	/* vectors */
 	sd->P = P;
@@ -393,7 +393,7 @@ __device void shader_setup_from_displace(KernelGlobals *kg, ShaderData *sd,
 
 	/* watch out: no instance transform currently */
 
-	shader_setup_from_sample(kg, sd, P, Ng, I, shader, object, prim, u, v, 0.0f, TIME_INVALID);
+	shader_setup_from_sample(kg, sd, P, Ng, I, shader, object, prim, u, v, 0.0f, TIME_INVALID, ~0);
 }
 
 /* ShaderData setup from ray into background */
@@ -769,8 +769,9 @@ __device void shader_eval_surface(KernelGlobals *kg, ShaderData *sd,
 #ifdef __SVM__
 		svm_eval_nodes(kg, sd, SHADER_TYPE_SURFACE, randb, path_flag);
 #else
-		bsdf_diffuse_setup(&sd->closure);
 		sd->closure.weight = make_float3(0.8f, 0.8f, 0.8f);
+		sd->closure.N = sd->N;
+		sd->flag |= bsdf_diffuse_setup(&sd->closure);
 #endif
 	}
 }
diff --git a/intern/cycles/util/util_transform.h b/intern/cycles/util/util_transform.h
index 10120ed5fdb..33761827dd7 100644
--- a/intern/cycles/util/util_transform.h
+++ b/intern/cycles/util/util_transform.h
@@ -109,19 +109,6 @@ __device_inline Transform transform_transpose(const Transform a)
 	return t;
 }
 
-__device_inline Transform operator*(const Transform a, const Transform b)
-{
-	Transform c = transform_transpose(b);
-	Transform t;
-
-	t.x = make_float4(dot(a.x, c.x), dot(a.x, c.y), dot(a.x, c.z), dot(a.x, c.w));
-	t.y = make_float4(dot(a.y, c.x), dot(a.y, c.y), dot(a.y, c.z), dot(a.y, c.w));
-	t.z = make_float4(dot(a.z, c.x), dot(a.z, c.y), dot(a.z, c.z), dot(a.z, c.w));
-	t.w = make_float4(dot(a.w, c.x), dot(a.w, c.y), dot(a.w, c.z), dot(a.w, c.w));
-
-	return t;
-}
-
 __device_inline Transform make_transform(float a, float b, float c, float d,
 									float e, float f, float g, float h,
 									float i, float j, float k, float l,
@@ -139,6 +126,19 @@ __device_inline Transform make_transform(float a, float b, float c, float d,
 
 #ifndef __KERNEL_GPU__
 
+__device_inline Transform operator*(const Transform a, const Transform b)
+{
+	Transform c = transform_transpose(b);
+	Transform t;
+
+	t.x = make_float4(dot(a.x, c.x), dot(a.x, c.y), dot(a.x, c.z), dot(a.x, c.w));
+	t.y = make_float4(dot(a.y, c.x), dot(a.y, c.y), dot(a.y, c.z), dot(a.y, c.w));
+	t.z = make_float4(dot(a.z, c.x), dot(a.z, c.y), dot(a.z, c.z), dot(a.z, c.w));
+	t.w = make_float4(dot(a.w, c.x), dot(a.w, c.y), dot(a.w, c.z), dot(a.w, c.w));
+
+	return t;
+}
+
 __device_inline void print_transform(const char *label, const Transform& t)
 {
 	print_float4(label, t.x);