griefith/test2
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

Fix #143218: Uninitialized VBO data using multiple UV tangents

Browse Source 
Caused by b19696c0b8.

I misunderstood the meaning of the vertex buffer allocation size.
Even though the attributes aren't interleaved, the fact that there
are multiple attributes is still included in the "element size" that's
multiplied with the size argument to `GPU_vertbuf_data_alloc`.

Also switch to spans and indices rather than incrementing raw pointers,
which would have made this much faster to debug.
This commit is contained in:
Hans Goudey
2025-07-25 17:00:41 -04:00
parent 9d993b4f74
commit 33b430892c
1 changed files with 16 additions and 14 deletions
Download Patch File Download Diff File Expand all files Collapse all files

30
source/blender/draw/intern/mesh_extractors/extract_mesh_vbo_tan.cc
View File

@@ -7,6 +7,7 @@
*/
#include <climits>
#include <iostream>
#include "BLI_string.h"
@@ -156,32 +157,35 @@ gpu::VertBufPtr extract_tangents(const MeshRenderData &mr,
GPUVertFormat format = {0};
const Array<Array<float4>> tangents = extract_tan_init_common(mr, cache, &format, gpu_attr_type);
const int vbo_size = tangents.size() * mr.corners_num;
gpu::VertBufPtr vbo = gpu::VertBufPtr(GPU_vertbuf_create_with_format(format));
GPU_vertbuf_data_alloc(*vbo, vbo_size);
GPU_vertbuf_data_alloc(*vbo, mr.corners_num);
if (use_hq) {
short4 *tan_data = vbo->data<short4>().data();
MutableSpan tan_data = vbo->data<short4>();
int vbo_index = 0;
for (const int i : tangents.index_range()) {
const Span<float4> layer_data = tangents[i];
for (int corner = 0; corner < mr.corners_num; corner++) {
*tan_data = gpu::convert_normal<short4>(float3(layer_data[corner]));
(*tan_data)[3] = (layer_data[corner][3] > 0.0f) ? SHRT_MAX : SHRT_MIN;
tan_data++;
tan_data[vbo_index] = gpu::convert_normal<short4>(float3(layer_data[corner]));
tan_data[vbo_index].w = (layer_data[corner][3] > 0.0f) ? SHRT_MAX : SHRT_MIN;
vbo_index++;
}
}
BLI_assert(vbo_index == tan_data.size());
}
else {
gpu::PackedNormal *tan_data = vbo->data<gpu::PackedNormal>().data();
MutableSpan tan_data = vbo->data<gpu::PackedNormal>();
int vbo_index = 0;
vbo->data<gpu::PackedNormal>().fill(gpu::PackedNormal{101, 102, 103, -2});
for (const int i : tangents.index_range()) {
const Span<float4> layer_data = tangents[i];
for (int corner = 0; corner < mr.corners_num; corner++) {
*tan_data = gpu::convert_normal<gpu::PackedNormal>(float3(layer_data[corner]));
tan_data->w = (layer_data[corner][3] > 0.0f) ? 1 : -2;
tan_data++;
tan_data[vbo_index] = gpu::convert_normal<gpu::PackedNormal>(float3(layer_data[corner]));
tan_data[vbo_index].w = (layer_data[corner][3] > 0.0f) ? 1 : -2;
vbo_index++;
}
}
BLI_assert(vbo_index == tan_data.size());
}
return vbo;
@@ -202,15 +206,13 @@ gpu::VertBufPtr extract_tangents_subdiv(const MeshRenderData &mr,
GPUVertFormat format = {0};
const Array<Array<float4>> tangents = extract_tan_init_common(mr, cache, &format, gpu_attr_type);
const int coarse_vbo_size = tangents.size() * mr.corners_num;
gpu::VertBufPtr vbo = gpu::VertBufPtr(
GPU_vertbuf_create_on_device(format, subdiv_cache.num_subdiv_loops));
gpu::VertBuf *coarse_vbo = GPU_vertbuf_calloc();
/* Dynamic as we upload and interpolate layers one at a time. */
GPU_vertbuf_init_with_format_ex(*coarse_vbo, get_coarse_tan_format(), GPU_USAGE_DYNAMIC);
GPU_vertbuf_data_alloc(*coarse_vbo, coarse_vbo_size);
GPU_vertbuf_data_alloc(*coarse_vbo, mr.corners_num);
/* Index of the tangent layer in the compact buffer. Used layers are stored in a single buffer.
*/