Fix emissive volumes generates unexpected fireflies around intersections

Discard the whole volume stack on the last bounce (but keep
world volume if present).

Volumes are expected to be closed manifol meshes, meaning if
ray entered the volume there should be an intersection event
of ray exisintg the volume. Case when ray hit nothing and
there are still non-world volumes in the stack can happen in
either of cases.

1. Mesh is not closed manifold.

Such configurations are not really supported anyway and should
not be used.

Previous code would have consider the infinite length of the
ray to sample across, so render result wasn't really correct
anyway.

2. Exit intersection is more far away than the camera far
   clip distance.

This case also will behave differently now, but previously it
wasn't really correct either, so it's not like we're breaking
something which was working as expected.

3. We missed exit event due to intersection precision issues.

This is exact the case which this patch fixes and avoid
fireflies.

4. Volume has Camera only visibility (all the rest visibility
is set to off)

This is what could be considered a regression but could be
solved quite easily by checking volume stack's objects flags
and keep entries which doesn't have Volume Scatter visibility
(or even better: ensure Volume Scatter visibility for objects
with volume closure),

Fixes T46108: Cycles - Overlapping emissive volumes generates unexpected bright hotspots around the intersection
Also fixes fireflies appearing on the edges of cube with
emissive volue.

Reviewers: juicyfruit, brecht

Reviewed By: brecht

Maniphest Tasks: T46108

Differential Revision: https://developer.blender.org/D2212

intern/cycles/kernel/kernel_path.h

@@ -141,6 +141,10 @@ ccl_device void kernel_path_indirect(KernelGlobals *kg,
 #endif  /* __LAMP_MIS__ */
 
 #ifdef __VOLUME__
+		/* Sanitize volume stack. */
+		if(!hit) {
+			kernel_volume_clean_stack(kg, state->volume_stack);
+		}
 		/* volume attenuation, emission, scatter */
 		if(state->volume_stack[0].shader != SHADER_NONE) {
 			Ray volume_ray = *ray;
@@ -658,6 +662,10 @@ ccl_device_inline float4 kernel_path_integrate(KernelGlobals *kg,
 #endif  /* __LAMP_MIS__ */
 
 #ifdef __VOLUME__
+		/* Sanitize volume stack. */
+		if(!hit) {
+			kernel_volume_clean_stack(kg, state.volume_stack);
+		}
 		/* volume attenuation, emission, scatter */
 		if(state.volume_stack[0].shader != SHADER_NONE) {
 			Ray volume_ray = ray;

intern/cycles/kernel/kernel_path_branched.h

@@ -294,6 +294,10 @@ ccl_device float4 kernel_branched_path_integrate(KernelGlobals *kg, RNG *rng, in
 #endif  /* __KERNEL_DEBUG__ */
 
 #ifdef __VOLUME__
+		/* Sanitize volume stack. */
+		if(!hit) {
+			kernel_volume_clean_stack(kg, state.volume_stack);
+		}
 		/* volume attenuation, emission, scatter */
 		if(state.volume_stack[0].shader != SHADER_NONE) {
 			Ray volume_ray = ray;

intern/cycles/kernel/kernel_volume.h

@@ -1262,4 +1262,30 @@ ccl_device void kernel_volume_stack_update_for_subsurface(KernelGlobals *kg,
 }
 #endif
 
+/* Clean stack after the last bounce.
+ *
+ * It is expected that all volumes are closed manifolds, so at the time when ray
+ * hits nothing (for example, it is a last bounce which goes to environment) the
+ * only expected volume in the stack is the world's one. All the rest volume
+ * entries should have been exited already.
+ *
+ * This isn't always true because of ray intersection precision issues, which
+ * could lead us to an infinite non-world volume in the stack, causing render
+ * artifacts.
+ *
+ * Use this function after the last bounce to get rid of all volumes apart from
+ * the world's one after the last bounce to avoid render artifacts.
+ */
+ccl_device_inline void kernel_volume_clean_stack(KernelGlobals *kg,
+                                                 VolumeStack *volume_stack)
+{
+	if(kernel_data.background.volume_shader != SHADER_NONE) {
+		/* Keep the world's volume in stack. */
+		volume_stack[1].shader = SHADER_NONE;
+	}
+	else {
+		volume_stack[0].shader = SHADER_NONE;
+	}
+}
+
 CCL_NAMESPACE_END