Upgrade to new vllm commit (#3719)

### What this PR does / why we need it?
Upgrade to new vllm commit:
vllm-project/vllm@c9461e0

- Fix many imports, caused by
vllm-project/vllm#26908
- Fix import ```sha256```, caused by
vllm-project/vllm#27169
- Remove ```SchedulerConfig.send_delta_data```, caused by
vllm-project/vllm#27142
- Fix ```FusedMoE``` because of dual stream execution, caused by
vllm-project/vllm#26440

### Does this PR introduce _any_ user-facing change?
N/A

### How was this patch tested?
CI passed with new added/existing test.


- vLLM version: v0.11.0rc3
- vLLM main:
vllm-project/vllm@17c540a

---------

Signed-off-by: MengqingCao <cmq0113@163.com>
Signed-off-by: Icey <1790571317@qq.com>
Co-authored-by: MengqingCao <cmq0113@163.com>

Author: wxsIcey

.github/workflows/format_pr_body.yaml

@@ -36,7 +36,7 @@ jobs:
 
       - name: Get vLLM version
         run: |
-          VLLM_COMMIT=17c540a993af88204ad1b78345c8a865cf58ce44
+          VLLM_COMMIT=c9461e05a4ed3557cfbf4b15ded1e26761cc39ca
           echo "VLLM_COMMIT=https://github.com/vllm-project/vllm/commit/$VLLM_COMMIT" >> $GITHUB_ENV
 
       - name: Checkout repository

.github/workflows/vllm_ascend_test.yaml

@@ -42,7 +42,7 @@ jobs:
   lint:
     uses: ./.github/workflows/pre-commit.yml
     with:
-      vllm: 17c540a993af88204ad1b78345c8a865cf58ce44
+      vllm: c9461e05a4ed3557cfbf4b15ded1e26761cc39ca
 
   changes:
     runs-on: ubuntu-latest
@@ -83,7 +83,7 @@ jobs:
         VLLM_USE_MODELSCOPE: True
     strategy:
       matrix:
-        vllm_version: [17c540a993af88204ad1b78345c8a865cf58ce44, v0.11.0]
+        vllm_version: [c9461e05a4ed3557cfbf4b15ded1e26761cc39ca, v0.11.0]
     steps:
       - name: Install packages
         run: |
@@ -140,7 +140,7 @@ jobs:
     name: e2e-light
     strategy:
       matrix:
-        vllm_version: [17c540a993af88204ad1b78345c8a865cf58ce44, v0.11.0]
+        vllm_version: [c9461e05a4ed3557cfbf4b15ded1e26761cc39ca, v0.11.0]
     # Note (yikun): If CI resource are limited we can split job into two chain jobs
     needs: [lint, changes]
     # only trigger e2e test after lint passed and the change is e2e related with pull request.

.github/workflows/vllm_ascend_test_full.yaml

@@ -69,7 +69,7 @@ jobs:
     name: e2e-full
     strategy:
       matrix:
-        vllm_version: [17c540a993af88204ad1b78345c8a865cf58ce44, v0.11.0]
+        vllm_version: [c9461e05a4ed3557cfbf4b15ded1e26761cc39ca, v0.11.0]
     needs: [changes]
     if: ${{ needs.changes.outputs.e2e_tracker == 'true' }}
     uses: ./.github/workflows/_e2e_test.yaml

examples/offline_data_parallel.py

@@ -63,7 +63,11 @@
 from vllm import LLM, SamplingParams
 from vllm.distributed.parallel_state import (  # noqa E402
     destroy_distributed_environment, destroy_model_parallel)
-from vllm.utils import get_open_port
+from vllm_ascend.utils import vllm_version_is
+if vllm_version_is("0.11.0"):
+    from vllm.utils import get_open_port
+else:
+    from vllm.utils.network_utils import get_open_port
 
 os.environ["VLLM_USE_MODELSCOPE"] = "True"
 os.environ["VLLM_WORKER_MULTIPROC_METHOD"] = "spawn"

examples/offline_external_launcher.py

@@ -65,9 +65,15 @@
 import torch
 from vllm import LLM, SamplingParams
 from vllm.distributed.parallel_state import (  # noqa E402
-    destroy_distributed_environment, destroy_model_parallel, get_tp_group)
-from vllm.utils import get_open_port, GiB_bytes
+    destroy_distributed_environment, destroy_model_parallel, get_tp_group) 
 from safetensors.torch import load_file
+from vllm_ascend.utils import vllm_version_is
+if vllm_version_is("0.11.0"):
+    from vllm.utils import GiB_bytes, get_open_port
+    
+else:
+    from vllm.utils.mem_constants import GiB_bytes
+    from vllm.utils.network_utils import get_open_port
 
 os.environ["VLLM_USE_MODELSCOPE"] = "True"
 os.environ["VLLM_WORKER_MULTIPROC_METHOD"] = "spawn"

examples/offline_inference_sleep_mode_npu.py

@@ -20,7 +20,11 @@
 
 import torch
 from vllm import LLM, SamplingParams
-from vllm.utils import GiB_bytes
+from vllm_ascend.utils import vllm_version_is
+if vllm_version_is("0.11.0"):
+    from vllm.utils import GiB_bytes
+else:
+    from vllm.utils.mem_constants import GiB_bytes
 
 os.environ["VLLM_USE_MODELSCOPE"] = "True"
 os.environ["VLLM_WORKER_MULTIPROC_METHOD"] = "spawn"

examples/offline_weight_load.py

@@ -66,8 +66,14 @@
 from vllm import LLM, SamplingParams
 from vllm.distributed.parallel_state import (  # noqa E402
     destroy_distributed_environment, destroy_model_parallel, get_tp_group)
-from vllm.utils import get_open_port, GiB_bytes
 from safetensors.torch import load_file
+from vllm_ascend.utils import vllm_version_is
+if vllm_version_is("0.11.0"):
+    from vllm.utils import GiB_bytes, get_open_port
+    
+else:
+    from vllm.utils.mem_constants import GiB_bytes
+    from vllm.utils.network_utils import get_open_port
 
 os.environ["VLLM_USE_MODELSCOPE"] = "True"
 os.environ["VLLM_WORKER_MULTIPROC_METHOD"] = "spawn"

tests/e2e/conftest.py

@@ -45,7 +45,6 @@
 from vllm.outputs import RequestOutput
 from vllm.platforms import current_platform
 from vllm.transformers_utils.utils import maybe_model_redirect
-from vllm.utils import get_open_port
 
 from tests.e2e.model_utils import (TokensTextLogprobs,
                                    TokensTextLogprobsPromptLogprobs)
@@ -55,6 +54,12 @@
 # we not explicitly patch here, some of them might be effectiveless
 # in pytest scenario
 from vllm_ascend.utils import adapt_patch  # noqa E402
+from vllm_ascend.utils import vllm_version_is
+
+if vllm_version_is("0.11.0"):
+    from vllm.utils import get_open_port
+else:
+    from vllm.utils.network_utils import get_open_port
 
 adapt_patch(True)
 adapt_patch(False)

tests/e2e/multicard/test_single_request_aclgraph.py

@@ -19,9 +19,14 @@
 
 import openai
 import pytest
-from vllm.utils import get_open_port
 
 from tests.e2e.conftest import RemoteOpenAIServer
+from vllm_ascend.utils import vllm_version_is
+
+if vllm_version_is("0.11.0"):
+    from vllm.utils import get_open_port
+else:
+    from vllm.utils.network_utils import get_open_port
 
 MODELS = [
     "Qwen/Qwen3-30B-A3B",

tests/e2e/nightly/models/test_qwen3_32b.py

@@ -18,10 +18,15 @@
 
 import openai
 import pytest
-from vllm.utils import get_open_port
 
 from tests.e2e.conftest import RemoteOpenAIServer
 from tools.aisbench import run_aisbench_cases
+from vllm_ascend.utils import vllm_version_is
+
+if vllm_version_is("0.11.0"):
+    from vllm.utils import get_open_port
+else:
+    from vllm.utils.network_utils import get_open_port
 
 MODELS = [
     "Qwen/Qwen3-32B",