xfail the cute dsl tests for l=1 (#1868)

<!-- .github/pull_request_template.md -->

## 📌 Description

With the latest version of nvidia-cutlass-dsl, `mark_layout_dynamic` may
throw errors like

```
>       self._dltensor_wrapper.mark_layout_dynamic(leading_dim)
E       RuntimeError: Expected strides[leading_dim] == 1, but got 7340032.
```

when calling `cutlass.torch.cute_tensor_like` and `l = 1` in gemm
problem size. This issue has been reported to
[cutlass#2673](NVIDIA/cutlass#2673).

So this PR marks the cute dsl blockscaled gemm tests where `l = 1`, due
to the issue of nvidia-cutlass-dsl.

<!-- What does this PR do? Briefly describe the changes and why they’re
needed. -->

## 🔍 Related Issues

<!-- Link any related issues here -->

## 🚀 Pull Request Checklist

Thank you for contributing to FlashInfer! Before we review your pull
request, please make sure the following items are complete.

### ✅ Pre-commit Checks

- [ ] I have installed `pre-commit` by running `pip install pre-commit`
(or used your preferred method).
- [ ] I have installed the hooks with `pre-commit install`.
- [ ] I have run the hooks manually with `pre-commit run --all-files`
and fixed any reported issues.

> If you are unsure about how to set up `pre-commit`, see [the
pre-commit documentation](https://pre-commit.com/).

## 🧪 Tests

- [ ] Tests have been added or updated as needed.
- [ ] All tests are passing (`unittest`, etc.).

## Reviewer Notes

<!-- Optional: anything you'd like reviewers to focus on, concerns, etc.
-->

Author: cyx-6

tests/gemm/test_cute_dsl_blockscaled_gemm.py

@@ -19,6 +19,7 @@
 )
 from flashinfer.cute_dsl.utils import (
     get_cutlass_dtype,
+    get_num_sm,
     is_cute_dsl_available,
 )
 
@@ -56,7 +57,6 @@
 @pytest.mark.parametrize("alpha_dtype", ["float32"])
 @pytest.mark.parametrize("mma_tiler_mn", [(128, 128)])
 @pytest.mark.parametrize("cluster_shape_mn", [(1, 1)])
-@pytest.mark.parametrize("sm_count", [132, None])
 @pytest.mark.parametrize("tolerance", [1e-01])
 @pytest.mark.parametrize("iterations", [3])
 @pytest.mark.parametrize("enable_dst_signals", [False, True])
@@ -74,7 +74,6 @@ def test_blockscaled_gemm_python_interface(
     alpha_dtype: cutlass.dtype,
     mma_tiler_mn: Tuple[int, int],
     cluster_shape_mn: Tuple[int, int],
-    sm_count: int,
     tolerance: float,
     iterations: int,
     enable_dst_signals: int,
@@ -85,11 +84,13 @@ def test_blockscaled_gemm_python_interface(
 
     if not (major == 10 and minor == 0):
         pytest.skip("Cute-dsl backend is only supported on SM100.")
-    if enable_dst_signals and (sm_count is None):
-        pytest.skip("dst_signals require sm_count")
 
     l, m = lm
     k, n = kn
+    if l == 1:
+        pytest.xfail("nvidia-cutlass-dsl has issue when l=1")
+
+    sm_count = get_num_sm(device) if enable_dst_signals else None
 
     print(f"device: {device}")