fixed fia pad logic in graph mode. (#7144)

### What this PR does / why we need it?
related to vllm PR #34043 this pr delete func
‘relax_for_mixed_batch_cudagraphs’, num_reqs no longer equals the actual
number of requests, due to fia operator requires that
query_start_loc[-1] equals the total number of computed tokens, so this
func delete cause the ifa error.
In full graph mode, set num_reqs_paded = num_reqs to fix the error
### Does this PR introduce _any_ user-facing change?

### How was this patch tested?

- vLLM version: v0.16.0
- vLLM main:
4034c3d32e

---------

Signed-off-by: wangxiaoxin-sherie <wangxiaoxin7@huawei.com>
Co-authored-by: wangxiaoxin-sherie <wangxiaoxin7@huawei.com>

tests/e2e/multicard/2-cards/test_full_graph_mode.py

@@ -18,7 +18,6 @@
 #
 import os
 
-import pytest
 from vllm import SamplingParams
 
 from tests.e2e.conftest import VllmRunner
@@ -68,7 +67,6 @@ def test_qwen3_moe_full_decode_only_tp2():
     )
 
 
-@pytest.mark.skip(reason="CANN8.5 failed with this test, fix me")
 def test_qwen3_moe_full_graph_tp2():
     if "HCCL_OP_EXPANSION_MODE" in os.environ:
         del os.environ["HCCL_OP_EXPANSION_MODE"]