CI : support IOT device (IQ9) (#22987)

* update test scripts

* align CI behavior between linux and android

* remove automatically cancel in 15min

* enable cancel-in-progress

* fix ty check issue

* update and fix pylint issue

* update runner such that we are not restricted by the 15min limit rule

* fix flake8 lint issue

* update runner according to review feedback

* code update according to review feedback

* switch from llama-cli to llama-completion binary with -no-cnv flag

scripts/snapdragon/qdc/tests/run_backend_ops_posix.py

@@ -1,8 +1,9 @@
 """
 On-device test-backend-ops runner for llama.cpp (HTP0 backend).
 
-Executed by QDC's Appium test framework on the QDC runner.
+On Android: executed by QDC's Appium test framework on the QDC runner.
 The runner has ADB access to the allocated device.
+On Linux: runs test-backend-ops directly via run_linux.sh (BASH framework).
 """
 
 import os
@@ -10,7 +11,12 @@ import sys
 
 import pytest
 
-from utils import BIN_PATH, CMD_PREFIX, push_bundle_if_needed, run_adb_command, write_qdc_log
+from utils import (
+    BIN_PATH,
+    push_bundle_if_needed,
+    run_script,
+    write_qdc_log,
+)
 
 
 @pytest.fixture(scope="session", autouse=True)
@@ -20,17 +26,21 @@ def install(driver):
 
 @pytest.mark.parametrize("type_a", ["mxfp4", "fp16", "q4_0"])
 def test_backend_ops_htp0(type_a):
-    cmd = f"{CMD_PREFIX} GGML_HEXAGON_HOSTBUF=0 GGML_HEXAGON_EXPERIMENTAL=1 {BIN_PATH}/test-backend-ops -b HTP0 -o MUL_MAT"
     if type_a == "q4_0":
-        cmd += r' -p "^(?=.*type_a=q4_0)(?!.*type_b=f32,m=576,n=512,k=576).*$"'
+        pattern = r'^(?=.*type_a=q4_0)(?!.*type_b=f32,m=576,n=512,k=576).*$'
     else:
-        cmd += f" -p type_a={type_a}"
-    result = run_adb_command(
-        cmd,
-        check=False,
+        pattern = f"type_a={type_a}"
+
+    quoted_pattern = f'"{pattern}"' if type_a == "q4_0" else pattern
+    result = run_script(
+        "run-tool.sh",
+        extra_env={"HB": "0"},
+        extra_args=["test-backend-ops", "-b", "HTP0", "-o", "MUL_MAT", "-p", quoted_pattern],
     )
     write_qdc_log(f"backend_ops_{type_a}.log", result.stdout or "")
-    assert result.returncode == 0, f"test-backend-ops type_a={type_a} failed (exit {result.returncode})"
+    assert result.returncode == 0, (
+        f"test-backend-ops type_a={type_a} failed (exit {result.returncode})"
+    )
 
 
 if __name__ == "__main__":