Fix missing 64bit casts when calculating Cycles render buffer offset

Found those missing casts while looking into a crash report made in
the Blender Chat. Was unable to reproduce the crash, but the casts
should totally be there to avoid integer overflow.

3 files changed, 11 insertions, 9 deletions

diff --git a/intern/cycles/kernel/device/gpu/kernel.h b/intern/cycles/kernel/device/gpu/kernel.h
index 82b51843864..328c58e7905 100644
--- a/intern/cycles/kernel/device/gpu/kernel.h
+++ b/intern/cycles/kernel/device/gpu/kernel.h
@@ -647,8 +647,9 @@ ccl_device_inline void kernel_gpu_film_convert_half_write(ccl_global uchar4 *rgb
     const int x = render_pixel_index % width; \
     const int y = render_pixel_index / width; \
 \
-    ccl_global const float *buffer = render_buffer + offset + x * kfilm_convert.pass_stride + \
-                                     y * stride * kfilm_convert.pass_stride; \
+    const uint64_t buffer_pixel_index = x + y * stride; \
+    ccl_global const float *buffer = render_buffer + offset + \
+                                     buffer_pixel_index * kfilm_convert.pass_stride; \
 \
     ccl_global float *pixel = pixels + \
                               (render_pixel_index + rgba_offset) * kfilm_convert.pixel_stride; \
@@ -677,8 +678,9 @@ ccl_device_inline void kernel_gpu_film_convert_half_write(ccl_global uchar4 *rgb
     const int x = render_pixel_index % width; \
     const int y = render_pixel_index / width; \
 \
-    ccl_global const float *buffer = render_buffer + offset + x * kfilm_convert.pass_stride + \
-                                     y * stride * kfilm_convert.pass_stride; \
+    const uint64_t buffer_pixel_index = x + y * stride; \
+    ccl_global const float *buffer = render_buffer + offset + \
+                                     buffer_pixel_index * kfilm_convert.pass_stride; \
 \
     float pixel[4]; \
     film_get_pass_pixel_##variant(&kfilm_convert, buffer, pixel); \
diff --git a/intern/cycles/kernel/film/adaptive_sampling.h b/intern/cycles/kernel/film/adaptive_sampling.h
index ad9b3b08ac5..16867c39d99 100644
--- a/intern/cycles/kernel/film/adaptive_sampling.h
+++ b/intern/cycles/kernel/film/adaptive_sampling.h
@@ -91,13 +91,13 @@ ccl_device void kernel_adaptive_sampling_filter_x(KernelGlobals kg,
   bool prev = false;
   for (int x = start_x; x < start_x + width; ++x) {
     int index = offset + x + y * stride;
-    ccl_global float *buffer = render_buffer + index * kernel_data.film.pass_stride;
+    ccl_global float *buffer = render_buffer + (uint64_t)index * kernel_data.film.pass_stride;
     const uint aux_w_offset = kernel_data.film.pass_adaptive_aux_buffer + 3;
 
     if (buffer[aux_w_offset] == 0.0f) {
       if (x > start_x && !prev) {
         index = index - 1;
-        buffer = render_buffer + index * kernel_data.film.pass_stride;
+        buffer = render_buffer + (uint64_t)index * kernel_data.film.pass_stride;
         buffer[aux_w_offset] = 0.0f;
       }
       prev = true;
@@ -124,13 +124,13 @@ ccl_device void kernel_adaptive_sampling_filter_y(KernelGlobals kg,
   bool prev = false;
   for (int y = start_y; y < start_y + height; ++y) {
     int index = offset + x + y * stride;
-    ccl_global float *buffer = render_buffer + index * kernel_data.film.pass_stride;
+    ccl_global float *buffer = render_buffer + (uint64_t)index * kernel_data.film.pass_stride;
     const uint aux_w_offset = kernel_data.film.pass_adaptive_aux_buffer + 3;
 
     if (buffer[aux_w_offset] == 0.0f) {
       if (y > start_y && !prev) {
         index = index - stride;
-        buffer = render_buffer + index * kernel_data.film.pass_stride;
+        buffer = render_buffer + (uint64_t)index * kernel_data.film.pass_stride;
         buffer[aux_w_offset] = 0.0f;
       }
       prev = true;
diff --git a/intern/cycles/kernel/integrator/init_from_bake.h b/intern/cycles/kernel/integrator/init_from_bake.h
index 3772db845a8..293c1d243f8 100644
--- a/intern/cycles/kernel/integrator/init_from_bake.h
+++ b/intern/cycles/kernel/integrator/init_from_bake.h
@@ -102,7 +102,7 @@ ccl_device bool integrator_init_from_bake(KernelGlobals kg,
   /* Setup render buffers. */
   const int index = INTEGRATOR_STATE(state, path, render_pixel_index);
   const int pass_stride = kernel_data.film.pass_stride;
-  ccl_global float *buffer = render_buffer + index * pass_stride;
+  ccl_global float *buffer = render_buffer + (uint64_t)index * pass_stride;
 
   ccl_global float *primitive = buffer + kernel_data.film.pass_bake_primitive;
   ccl_global float *differential = buffer + kernel_data.film.pass_bake_differential;