[SYCL][E2E] Fix lit for arch selection and l0v2 adapter (#20728)

Fix llvm-lit to enable the simultaneous use of _v1 and _v2 to select the
l0 adapter and the use of arch to select the GPU, eg.:
`llvm-lit --param "sycl_devices=level_zero_v2:arch-intel_gpu_mtl_u"
sycl/test-e2e`
Necessary to run tests on integrated GPU on system with another GPU
Includes fixed setting of ZE_AFFINITY_MASK in multi-gpu systems

---------

Signed-off-by: Mateusz P. Nowak <mateusz.p.nowak@intel.com>

Author: mateuszpn

sycl/test-e2e/AddressSanitizer/lit.local.cfg

@@ -31,4 +31,5 @@ unsupported_san_flags = [
 if any(flag in config.cxx_flags for flag in unsupported_san_flags):
     config.unsupported=True
 
-config.environment["ZE_AFFINITY_MASK"] = "0"
+if config.ze_affinity_mask is not None:
+    config.environment["ZE_AFFINITY_MASK"] = config.ze_affinity_mask

sycl/test-e2e/MemorySanitizer/lit.local.cfg

@@ -39,4 +39,5 @@ unsupported_san_flags = [
 if any(flag in config.cxx_flags for flag in unsupported_san_flags):
     config.unsupported=True
 
-config.environment["ZE_AFFINITY_MASK"] = "0"
+if config.ze_affinity_mask is not None:
+    config.environment["ZE_AFFINITY_MASK"] = config.ze_affinity_mask

sycl/test-e2e/ThreadSanitizer/lit.local.cfg

@@ -34,4 +34,5 @@ unsupported_san_flags = [
 if any(flag in config.cxx_flags for flag in unsupported_san_flags):
     config.unsupported=True
 
-config.environment["ZE_AFFINITY_MASK"] = "0"
+if config.ze_affinity_mask is not None:
+    config.environment["ZE_AFFINITY_MASK"] = config.ze_affinity_mask

sycl/test-e2e/format.py

@@ -283,6 +283,15 @@ def get_extra_env(sycl_devices):
             # so that device might still be accessible to some of the tests yet
             # we won't set the environment variable below for such scenario.
             extra_env = []
+
+            # Include ZE_AFFINITY_MASK if set by lit.local.cfg (e.g., for sanitizer tests)
+            if "ZE_AFFINITY_MASK" in test.config.environment:
+                extra_env.append(
+                    "ZE_AFFINITY_MASK={}".format(
+                        test.config.environment["ZE_AFFINITY_MASK"]
+                    )
+                )
+
             if "level_zero:gpu" in sycl_devices and litConfig.params.get("ur_l0_debug"):
                 extra_env.append("UR_L0_DEBUG={}".format(test.config.ur_l0_debug))
 
@@ -342,8 +351,14 @@ def get_extra_env(sycl_devices):
                 elif "level_zero_v1" in full_dev_name:
                     expanded += " env UR_LOADER_USE_LEVEL_ZERO_V2=0"
 
+                # If ZE_AFFINITY_MASK is set in local config, it filters devices so we should use :0
+                device_selector = parsed_dev_name
+                if test.config.ze_affinity_mask is not None:
+                    backend, _ = parsed_dev_name.split(":", 1)
+                    device_selector = f"{backend}:0"
+
                 expanded += " ONEAPI_DEVICE_SELECTOR={} {}".format(
-                    parsed_dev_name, test.config.run_launcher
+                    device_selector, test.config.run_launcher
                 )
                 cmd = directive.command.replace("%{run}", expanded)
                 # Expand device-specific condtions (%if ... %{ ... %}).
@@ -353,6 +368,11 @@ def get_extra_env(sycl_devices):
                     "linux",
                     "windows",
                     "preview-breaking-changes-supported",
+                    # the following entries are used by architecture-based filtering
+                    # (:arch- device, not :gpu or :cpu)
+                    "cpu",
+                    "gpu",
+                    "accelerator",
                 ]:
                     if cond_features in test.config.available_features:
                         conditions[cond_features] = True

sycl/test-e2e/lit.cfg.py

@@ -96,6 +96,7 @@
     "LIBCLANG_NOTHREADS",
     "LIBCLANG_RESOURCE_USAGE",
     "LIBCLANG_CODE_COMPLETION_LOGGING",
+    "ZE_AFFINITY_MASK",
 ]
 
 # Names of the Release and Debug versions of the XPTIFW library
@@ -917,12 +918,14 @@ def get_sycl_ls_verbose(sycl_device, env):
 
     env = copy.copy(llvm_config.config.environment)
 
+    backend_for_selector = backend.replace("_v2", "").replace("_v1", "")
+
     # Find all available devices under the backend
-    env["ONEAPI_DEVICE_SELECTOR"] = backend + ":*"
+    env["ONEAPI_DEVICE_SELECTOR"] = backend_for_selector + ":*"
 
     detected_architectures = []
 
-    platform_devices = remove_level_zero_suffix(backend + ":*")
+    platform_devices = backend_for_selector + ":*"
 
     for line in get_sycl_ls_verbose(platform_devices, env).stdout.splitlines():
         if re.match(r" *Architecture:", line):
@@ -948,6 +951,15 @@ def get_sycl_ls_verbose(sycl_device, env):
 
 config.sycl_devices = filtered_sycl_devices
 
+# Determine ZE_AFFINITY_MASK for Level Zero devices.
+# Sanitizer tests need to set ZE_AFFINITY_MASK to a single device index
+config.ze_affinity_mask = None
+for sycl_device in remove_level_zero_suffix(config.sycl_devices):
+    be, dev = sycl_device.split(":")
+    if be == "level_zero" and dev.isdigit():
+        config.ze_affinity_mask = dev
+        break
+
 for sycl_device in remove_level_zero_suffix(config.sycl_devices):
     be, dev = sycl_device.split(":")
     config.available_features.add("any-device-is-" + dev)
@@ -1114,6 +1126,13 @@ def get_sycl_ls_verbose(sycl_device, env):
     features.update(device_family)
 
     be, dev = sycl_device.split(":")
+    if dev.isdigit():
+        backend_devices = available_devices.get(be, "gpu")
+        if isinstance(backend_devices, tuple):
+            # arch-selection is typically used to select gpu device
+            dev = "gpu"
+        else:
+            dev = backend_devices
     features.add(dev.replace("fpga", "accelerator"))
     if "level_zero_v2" in full_name:
         features.add("level_zero_v2_adapter")
@@ -1146,6 +1165,13 @@ def get_sycl_ls_verbose(sycl_device, env):
     else:
         config.intel_driver_ver[full_name] = {}
 
+# When running with a single device, merge device-specific features into global
+# available_features so that %if conditionals work correctly, even if device is not :gpu or :cpu
+if len(config.sycl_devices) == 1:
+    single_device = list(config.sycl_dev_features.keys())[0]
+    device_features = config.sycl_dev_features[single_device]
+    config.available_features.update(device_features)
+
 if lit_config.params.get("compatibility_testing", "False") != "False":
     config.substitutions.append(("%clangxx", " true "))
     config.substitutions.append(("%clang", " true "))