Remove unnessary uniform barriers in Vulkan.

src/gpu/vk/GrVkGpu.cpp

Index: src/gpu/vk/GrVkGpu.cpp
diff --git a/src/gpu/vk/GrVkGpu.cpp b/src/gpu/vk/GrVkGpu.cpp
index 4d0ecbf5ea14b096ff164e4d75438caf0f0b8fed..830e66f5fb95d92d0210cf52deb903edeefb6ee9 100644
--- a/src/gpu/vk/GrVkGpu.cpp
+++ b/src/gpu/vk/GrVkGpu.cpp
@@ -1496,7 +1496,6 @@ void GrVkGpu::onDraw(const GrPipeline& pipeline,
     const GrVkRenderPass* renderPass = vkRT->simpleRenderPass();
     SkASSERT(renderPass);
 
-
     GrPrimitiveType primitiveType = meshes[0].primitiveType();
     sk_sp<GrVkPipelineState> pipelineState = this->prepareDrawState(pipeline,
                                                                     primProc,
@@ -1553,10 +1552,6 @@ void GrVkGpu::onDraw(const GrPipeline& pipeline,
                 pipelineState->freeTempResources(this);
                 SkDEBUGCODE(pipelineState = nullptr);
                 primitiveType = nonIdxMesh->primitiveType();
-                // It is illegal for us to have the necessary memory barriers for when we write and
-                // update the uniform buffers in prepareDrawState while in an active render pass.
-                // Thus we must end the current one and then start it up again.
-                fCurrentCmdBuffer->endRenderPass(this);
                 pipelineState = this->prepareDrawState(pipeline,
                                                        primProc,
                                                        primitiveType,
@@ -1564,7 +1559,6 @@ void GrVkGpu::onDraw(const GrPipeline& pipeline,
                 if (!pipelineState) {
                     return;
                 }
-                fCurrentCmdBuffer->beginRenderPass(this, renderPass, *vkRT);
             }
             SkASSERT(pipelineState);
             this->bindGeometry(primProc, *nonIdxMesh);