Batch GPU uploads to eliminate per-upload fence waits (stutter fix)

Every uploadBuffer/VkTexture::upload called immediateSubmit which did a
separate vkQueueSubmit + vkWaitForFences. Loading a single creature model
with textures caused 4-8+ fence waits; terrain chunks caused 80+ per batch.

Added beginUploadBatch/endUploadBatch to VkContext: records all upload
commands into a single command buffer, submits once with one fence wait.
Staging buffers are deferred for cleanup after the batch completes.

Wrapped in batch mode:
- CharacterRenderer::loadModel (creature VB/IB + textures)
- M2Renderer::loadModel (doodad VB/IB + textures)
- TerrainRenderer::loadTerrain/loadTerrainIncremental (chunk geometry + textures)
- TerrainRenderer::uploadPreloadedTextures
- WMORenderer::loadModel (group geometry + textures)

src/rendering/m2_renderer.cpp

@@ -1185,6 +1185,10 @@ bool M2Renderer::loadModel(const pipeline::M2Model& model, uint32_t modelId) {
         }
     }
 
+    // Batch all GPU uploads (VB, IB, textures) into a single command buffer
+    // submission with one fence wait, instead of one fence wait per upload.
+    vkCtx_->beginUploadBatch();
+
     if (hasGeometry) {
         // Create VBO with interleaved vertex data
         // Format: position (3), normal (3), texcoord0 (2), texcoord1 (2), boneWeights (4), boneIndices (4 as float)
@@ -1536,6 +1540,8 @@ bool M2Renderer::loadModel(const pipeline::M2Model& model, uint32_t modelId) {
         }
     }
 
+    vkCtx_->endUploadBatch();
+
     // Allocate Vulkan descriptor sets and UBOs for each batch
     for (auto& bgpu : gpuModel.batches) {
         // Create combined UBO for M2Params (binding 1) + M2Material (binding 2)