[Feat] (DirectVulkan): GPU transform feedback capture via VK_EXT_transform_feedback

Second stage of GL 3.0 transform feedback: captured draws now write real
data.

- Device setup enables the VK_EXT_transform_feedback feature when present
  and loads the bind/begin/end entry points.
- Captured draws compile an XfbCapture program variant whose last
  vertex-processing stage gets XfbBuffer/XfbStride/Offset decorations from
  the program's resolved varyings (a new spirv-opt pass). A captured
  gl_Position is mirrored into a dedicated output written before every
  OpReturn - or before every OpEmitVertex in a geometry stage - ahead of
  the position fixup, so the captured value is the shader's own pre-remap
  position.
- DrawArrays/DrawElements wrap the draw in Begin/EndTransformFeedbackEXT;
  a small counter buffer resumes the append position across draws within
  one glBeginTransformFeedback (fresh Begin starts at the bound offsets).
- Capture targets are promoted to persistently-mapped host-coherent GPU
  storage (persistent-map storage now also carries the transform feedback
  usage), so MapBuffer/GetBufferSubData read the captured bytes after the
  fence wait glEndTransformFeedback now performs.
- Draw-mode/feedback-mode validation defers to the geometry shader's
  output primitive when one is present, and glGetBooleanv reports
  GL_TRANSFORM_FEEDBACK_ACTIVE/PAUSED so dEQP's per-case state reset can
  unwind an active capture.

KHR-GL33: transform_feedback capture_vertex_*/capture_geometry_*/
discard_*/draw_xfb and clip_distance.coverage now pass; queries
(PRIMITIVES_WRITTEN) and gl_ClipDistance capture remain.

MobileGL/MG_Impl/GLImpl/Drawing/GL_Drawing.cpp

@@ -67,8 +67,12 @@ namespace MobileGL::MG_Impl::GLImpl {
         }
 
         // While transform feedback is active the draw's primitive type must match
-        // the feedback primitive mode (GL 3.3 core 13.2.2).
-        if (MG_State::pGLContext->IsTransformFeedbackActive()) {
+        // the feedback primitive mode (GL 3.3 core 13.2.2). With a geometry shader
+        // the constraint moves to the shader's output primitive type instead, so
+        // the draw mode itself is unconstrained here.
+        if (MG_State::pGLContext->IsTransformFeedbackActive() &&
+            !(MG_State::pGLContext->GetTransformFeedbackProgram() &&
+              MG_State::pGLContext->GetTransformFeedbackProgram()->GetShaderIndexByStage(ShaderStage::Geometry) >= 0)) {
             const GLenum feedbackMode = MG_State::pGLContext->GetTransformFeedbackPrimitiveMode();
             Bool compatible = false;
             switch (feedbackMode) {
@@ -514,6 +518,18 @@ namespace MobileGL::MG_Impl::GLImpl {
             return;
         }
         MG_State::pGLContext->EndTransformFeedback();
+        // Captured results must be visible to MapBuffer/GetBufferSubData after
+        // End; the capture targets are host-coherent GPU memory, so completing
+        // the GPU work is all that is required.
+        auto& backendGL = MG_Backend::gBackendFunctionsTable.GL;
+        if (backendGL.FenceSync && backendGL.ClientWaitSync) {
+            if (auto sync = backendGL.FenceSync()) {
+                backendGL.ClientWaitSync(sync, GL_SYNC_FLUSH_COMMANDS_BIT, ~0ull);
+                if (backendGL.DeleteSync) {
+                    backendGL.DeleteSync(sync);
+                }
+            }
+        }
     }
 
 } // namespace MobileGL::MG_Impl::GLImpl