[Bf-blender-cvs] [4df66dabb88] master: Fix T65212: Cycles SSS failing on 32 bit
Brecht Van Lommel
noreply at git.blender.org
Tue May 28 18:10:03 CEST 2019
Commit: 4df66dabb8887cf2ff281946ad7af10b165d09ae
Author: Brecht Van Lommel
Date: Tue May 28 11:52:26 2019 +0200
Branches: master
https://developer.blender.org/rB4df66dabb8887cf2ff281946ad7af10b165d09ae
Fix T65212: Cycles SSS failing on 32 bit
===================================================================
M intern/cycles/kernel/closure/bsdf_ashikhmin_velvet.h
M intern/cycles/kernel/closure/bsdf_diffuse.h
M intern/cycles/kernel/closure/bsdf_diffuse_ramp.h
M intern/cycles/kernel/closure/bsdf_hair.h
M intern/cycles/kernel/closure/bsdf_microfacet.h
M intern/cycles/kernel/closure/bsdf_oren_nayar.h
M intern/cycles/kernel/closure/bsdf_phong_ramp.h
M intern/cycles/kernel/closure/bsdf_principled_diffuse.h
M intern/cycles/kernel/closure/bsdf_principled_sheen.h
M intern/cycles/kernel/closure/bsdf_toon.h
M intern/cycles/kernel/closure/bssrdf.h
M intern/cycles/kernel/closure/volume.h
M intern/cycles/kernel/kernel_types.h
===================================================================
diff --git a/intern/cycles/kernel/closure/bsdf_ashikhmin_velvet.h b/intern/cycles/kernel/closure/bsdf_ashikhmin_velvet.h
index 8122bcc1424..3d3f20edab3 100644
--- a/intern/cycles/kernel/closure/bsdf_ashikhmin_velvet.h
+++ b/intern/cycles/kernel/closure/bsdf_ashikhmin_velvet.h
@@ -42,6 +42,8 @@ typedef ccl_addr_space struct VelvetBsdf {
float invsigma2;
} VelvetBsdf;
+static_assert(sizeof(ShaderClosure) >= sizeof(VelvetBsdf), "VelvetBsdf is too large!");
+
ccl_device int bsdf_ashikhmin_velvet_setup(VelvetBsdf *bsdf)
{
float sigma = fmaxf(bsdf->sigma, 0.01f);
diff --git a/intern/cycles/kernel/closure/bsdf_diffuse.h b/intern/cycles/kernel/closure/bsdf_diffuse.h
index 76b50548455..ea604ed0311 100644
--- a/intern/cycles/kernel/closure/bsdf_diffuse.h
+++ b/intern/cycles/kernel/closure/bsdf_diffuse.h
@@ -39,6 +39,8 @@ typedef ccl_addr_space struct DiffuseBsdf {
SHADER_CLOSURE_BASE;
} DiffuseBsdf;
+static_assert(sizeof(ShaderClosure) >= sizeof(DiffuseBsdf), "DiffuseBsdf is too large!");
+
/* DIFFUSE */
ccl_device int bsdf_diffuse_setup(DiffuseBsdf *bsdf)
diff --git a/intern/cycles/kernel/closure/bsdf_diffuse_ramp.h b/intern/cycles/kernel/closure/bsdf_diffuse_ramp.h
index 9d13eb8d4e0..aa62c1c7ceb 100644
--- a/intern/cycles/kernel/closure/bsdf_diffuse_ramp.h
+++ b/intern/cycles/kernel/closure/bsdf_diffuse_ramp.h
@@ -43,6 +43,8 @@ typedef ccl_addr_space struct DiffuseRampBsdf {
float3 *colors;
} DiffuseRampBsdf;
+static_assert(sizeof(ShaderClosure) >= sizeof(DiffuseRampBsdf), "DiffuseRampBsdf is too large!");
+
ccl_device float3 bsdf_diffuse_ramp_get_color(const float3 colors[8], float pos)
{
int MAXCOLORS = 8;
diff --git a/intern/cycles/kernel/closure/bsdf_hair.h b/intern/cycles/kernel/closure/bsdf_hair.h
index 4b6f5b3b439..7ca9424b815 100644
--- a/intern/cycles/kernel/closure/bsdf_hair.h
+++ b/intern/cycles/kernel/closure/bsdf_hair.h
@@ -44,6 +44,8 @@ typedef ccl_addr_space struct HairBsdf {
float offset;
} HairBsdf;
+static_assert(sizeof(ShaderClosure) >= sizeof(HairBsdf), "HairBsdf is too large!");
+
ccl_device int bsdf_hair_reflection_setup(HairBsdf *bsdf)
{
bsdf->type = CLOSURE_BSDF_HAIR_REFLECTION_ID;
diff --git a/intern/cycles/kernel/closure/bsdf_microfacet.h b/intern/cycles/kernel/closure/bsdf_microfacet.h
index b4da3123f28..7d7ccfa7774 100644
--- a/intern/cycles/kernel/closure/bsdf_microfacet.h
+++ b/intern/cycles/kernel/closure/bsdf_microfacet.h
@@ -48,6 +48,8 @@ typedef ccl_addr_space struct MicrofacetBsdf {
float3 T;
} MicrofacetBsdf;
+static_assert(sizeof(ShaderClosure) >= sizeof(MicrofacetBsdf), "MicrofacetBsdf is too large!");
+
/* Beckmann and GGX microfacet importance sampling. */
ccl_device_inline void microfacet_beckmann_sample_slopes(KernelGlobals *kg,
diff --git a/intern/cycles/kernel/closure/bsdf_oren_nayar.h b/intern/cycles/kernel/closure/bsdf_oren_nayar.h
index 104ed5b2818..41e5736bf49 100644
--- a/intern/cycles/kernel/closure/bsdf_oren_nayar.h
+++ b/intern/cycles/kernel/closure/bsdf_oren_nayar.h
@@ -27,6 +27,8 @@ typedef ccl_addr_space struct OrenNayarBsdf {
float b;
} OrenNayarBsdf;
+static_assert(sizeof(ShaderClosure) >= sizeof(OrenNayarBsdf), "OrenNayarBsdf is too large!");
+
ccl_device float3 bsdf_oren_nayar_get_intensity(const ShaderClosure *sc,
float3 n,
float3 v,
diff --git a/intern/cycles/kernel/closure/bsdf_phong_ramp.h b/intern/cycles/kernel/closure/bsdf_phong_ramp.h
index b6fd0e68681..cf5484383f2 100644
--- a/intern/cycles/kernel/closure/bsdf_phong_ramp.h
+++ b/intern/cycles/kernel/closure/bsdf_phong_ramp.h
@@ -44,6 +44,8 @@ typedef ccl_addr_space struct PhongRampBsdf {
float3 *colors;
} PhongRampBsdf;
+static_assert(sizeof(ShaderClosure) >= sizeof(PhongRampBsdf), "PhongRampBsdf is too large!");
+
ccl_device float3 bsdf_phong_ramp_get_color(const float3 colors[8], float pos)
{
int MAXCOLORS = 8;
diff --git a/intern/cycles/kernel/closure/bsdf_principled_diffuse.h b/intern/cycles/kernel/closure/bsdf_principled_diffuse.h
index d7795974ef5..43646aaeb5b 100644
--- a/intern/cycles/kernel/closure/bsdf_principled_diffuse.h
+++ b/intern/cycles/kernel/closure/bsdf_principled_diffuse.h
@@ -30,6 +30,9 @@ typedef ccl_addr_space struct PrincipledDiffuseBsdf {
float roughness;
} PrincipledDiffuseBsdf;
+static_assert(sizeof(ShaderClosure) >= sizeof(PrincipledDiffuseBsdf),
+ "PrincipledDiffuseBsdf is too large!");
+
ccl_device float3 calculate_principled_diffuse_brdf(
const PrincipledDiffuseBsdf *bsdf, float3 N, float3 V, float3 L, float3 H, float *pdf)
{
diff --git a/intern/cycles/kernel/closure/bsdf_principled_sheen.h b/intern/cycles/kernel/closure/bsdf_principled_sheen.h
index bc522095b3b..1326d23668a 100644
--- a/intern/cycles/kernel/closure/bsdf_principled_sheen.h
+++ b/intern/cycles/kernel/closure/bsdf_principled_sheen.h
@@ -28,6 +28,9 @@ typedef ccl_addr_space struct PrincipledSheenBsdf {
SHADER_CLOSURE_BASE;
} PrincipledSheenBsdf;
+static_assert(sizeof(ShaderClosure) >= sizeof(PrincipledSheenBsdf),
+ "PrincipledSheenBsdf is too large!");
+
ccl_device float3 calculate_principled_sheen_brdf(
const PrincipledSheenBsdf *bsdf, float3 N, float3 V, float3 L, float3 H, float *pdf)
{
diff --git a/intern/cycles/kernel/closure/bsdf_toon.h b/intern/cycles/kernel/closure/bsdf_toon.h
index f37fd228087..cc5de21ed0e 100644
--- a/intern/cycles/kernel/closure/bsdf_toon.h
+++ b/intern/cycles/kernel/closure/bsdf_toon.h
@@ -42,6 +42,8 @@ typedef ccl_addr_space struct ToonBsdf {
float smooth;
} ToonBsdf;
+static_assert(sizeof(ShaderClosure) >= sizeof(ToonBsdf), "ToonBsdf is too large!");
+
/* DIFFUSE TOON */
ccl_device int bsdf_diffuse_toon_setup(ToonBsdf *bsdf)
diff --git a/intern/cycles/kernel/closure/bssrdf.h b/intern/cycles/kernel/closure/bssrdf.h
index a7d9f90b443..125ca4dd830 100644
--- a/intern/cycles/kernel/closure/bssrdf.h
+++ b/intern/cycles/kernel/closure/bssrdf.h
@@ -30,6 +30,8 @@ typedef ccl_addr_space struct Bssrdf {
float channels;
} Bssrdf;
+static_assert(sizeof(ShaderClosure) >= sizeof(Bssrdf), "Bssrdf is too large!");
+
/* Planar Truncated Gaussian
*
* Note how this is different from the typical gaussian, this one integrates
diff --git a/intern/cycles/kernel/closure/volume.h b/intern/cycles/kernel/closure/volume.h
index 473bc0e8a82..1430f712701 100644
--- a/intern/cycles/kernel/closure/volume.h
+++ b/intern/cycles/kernel/closure/volume.h
@@ -40,6 +40,9 @@ typedef ccl_addr_space struct HenyeyGreensteinVolume {
float g;
} HenyeyGreensteinVolume;
+static_assert(sizeof(ShaderClosure) >= sizeof(HenyeyGreensteinVolume),
+ "HenyeyGreensteinVolume is too large!");
+
/* Given cosine between rays, return probability density that a photon bounces
* to that direction. The g parameter controls how different it is from the
* uniform sphere. g=0 uniform diffuse-like, g=1 close to sharp single ray. */
diff --git a/intern/cycles/kernel/kernel_types.h b/intern/cycles/kernel/kernel_types.h
index 18d2a216f72..08c0c199af6 100644
--- a/intern/cycles/kernel/kernel_types.h
+++ b/intern/cycles/kernel/kernel_types.h
@@ -804,8 +804,9 @@ typedef struct AttributeDescriptor {
* ShaderClosure has a fixed size, and any extra space must be allocated
* with closure_alloc_extra().
*
- * We pad the struct to 80 bytes and ensure it is aligned to 16 bytes, which
- * we assume to be the maximum required alignment for any struct. */
+ * We pad the struct to align to 16 bytes. All shader closures are assumed
+ * to fit in this struct size. CPU sizes are a bit larger because float3 is
+ * padded to be 16 bytes, while it's only 12 bytes on the GPU. */
#define SHADER_CLOSURE_BASE \
float3 weight; \
@@ -817,7 +818,10 @@ typedef ccl_addr_space struct ccl_align(16) ShaderClosure
{
SHADER_CLOSURE_BASE;
- float data[10]; /* pad to 80 bytes */
+#ifdef __KERNEL_CPU__
+ float pad[2];
+#endif
+ float data[10];
}
ShaderClosure;
More information about the Bf-blender-cvs
mailing list