## Summary

When `litellm.google_genai.generate_content()` routes through the adapter path (i.e., `generate_content_provider_config is None`), the success logging handler crashes with:

``` ValueError: Google GenAI Generate Content: httpx_response is None ```

The actual LLM call succeeds (HTTP 200 is returned to the caller), but the background logging worker fails because the logging handler assumes a raw `httpx.Response` is always available, which is not true for the adapter path.

## Reproduction

### Setup

Define an OpenAI-compatible provider with a Gemini alias in your proxy config:

```yaml model_list: - model_name: "sensenova-6.7-flash-lite" litellm_params: model: openai/sensenova-6.7-flash-lite api_base: https://token.sensenova.cn/v1

router_settings: model_group_alias: gemini-3.5-flash-lite: sensenova-6.7-flash-lite gemini-2.5-flash-lite: sensenova-6.7-flash-lite gemini-3.1-flash-lite: sensenova-6.7-flash-lite flash-lite: sensenova-6.7-flash-lite

general_settings: master_key: sk-1234

litellm_settings: drop_params: True ```

### Call

```bash curl -X POST http://localhost:4000/v1beta/models/gemini-3.1-flash-lite%3AgenerateContent \…