Manage GPU_matrix stacks per GPUContext

Previously, we had one global `GPU_matrix` stack, so the API was not
thread safe. This patch makes the stack be per `GPUContext`, effectively
making it local per thread (`GPUContext` is located in thread local
storage).

Reviewed By: brecht
Differential Revision: https://developer.blender.org/D5405

source/blender/gpu/intern/gpu_context.cpp

@@ -36,6 +36,7 @@
 
 #include "gpu_batch_private.h"
 #include "gpu_context_private.h"
+#include "gpu_matrix_private.h"
 
 #include <vector>
 #include <string.h>
@@ -71,6 +72,7 @@ struct GPUContext {
   std::unordered_set<GPUFrameBuffer *>
       framebuffers; /* Framebuffers that have FBO from this context */
 #endif
+  struct GPUMatrixState *matrix_state;
   std::vector<GLuint> orphaned_vertarray_ids;
   std::vector<GLuint> orphaned_framebuffer_ids;
   std::mutex orphans_mutex; /* todo: try spinlock instead */
@@ -139,6 +141,7 @@ GPUContext *GPU_context_create(GLuint default_framebuffer)
   GPUContext *ctx = new GPUContext;
   glGenVertexArrays(1, &ctx->default_vao);
   ctx->default_framebuffer = default_framebuffer;
+  ctx->matrix_state = GPU_matrix_state_create();
   GPU_context_active_set(ctx);
   return ctx;
 }
@@ -159,6 +162,7 @@ void GPU_context_discard(GPUContext *ctx)
     /* this removes the array entry */
     GPU_batch_vao_cache_clear(*ctx->batches.begin());
   }
+  GPU_matrix_state_discard(ctx->matrix_state);
   glDeleteVertexArrays(1, &ctx->default_vao);
   delete ctx;
   active_ctx = NULL;
@@ -333,3 +337,9 @@ GPUFrameBuffer *gpu_context_active_framebuffer_get(GPUContext *ctx)
 {
   return ctx->current_fbo;
 }
+
+struct GPUMatrixState *gpu_context_active_matrix_state_get()
+{
+  BLI_assert(active_ctx);
+  return active_ctx->matrix_state;
+}