Fix CUDA adapter formatting

Author: callumfare

sycl/plugins/cuda/pi_cuda.hpp

@@ -27,13 +27,13 @@
 
 #include <ur/adapters/cuda/context.hpp>
 #include <ur/adapters/cuda/device.hpp>
+#include <ur/adapters/cuda/event.hpp>
 #include <ur/adapters/cuda/kernel.hpp>
+#include <ur/adapters/cuda/memory.hpp>
 #include <ur/adapters/cuda/platform.hpp>
 #include <ur/adapters/cuda/program.hpp>
-#include <ur/adapters/cuda/event.hpp>
 #include <ur/adapters/cuda/queue.hpp>
 #include <ur/adapters/cuda/sampler.hpp>
-#include <ur/adapters/cuda/memory.hpp>
 
 // Share code between the PI Plugin and UR Adapter
 #include <pi2ur.hpp>
@@ -52,7 +52,7 @@ struct _pi_context : ur_context_handle_t_ {
   using ur_context_handle_t_::ur_context_handle_t_;
 };
 
-struct _pi_mem  : ur_mem_handle_t_ {
+struct _pi_mem : ur_mem_handle_t_ {
   using ur_mem_handle_t_::ur_mem_handle_t_;
 };
 

sycl/plugins/unified_runtime/ur/adapters/cuda/common.cpp

@@ -86,7 +86,6 @@ void sycl::detail::ur::cuPrint(const char *Message) {
   std::cerr << "ur_print: " << Message << std::endl;
 }
 
-
 // Global variables for ZER_EXT_RESULT_ADAPTER_SPECIFIC_ERROR
 thread_local ur_result_t ErrorMessageCode = UR_RESULT_SUCCESS;
 thread_local char ErrorMessage[MaxMessageSize];

sycl/plugins/unified_runtime/ur/adapters/cuda/device.cpp

@@ -974,7 +974,8 @@ UR_APIEXPORT ur_result_t UR_APICALL urDeviceGetInfo(ur_device_handle_t device,
         cuDeviceGetPCIBusId(AddressBuffer, AddressBufferSize, device->get()) ==
         CUDA_SUCCESS);
     // CUDA API (8.x - 12.1) guarantees 12 bytes + \0 are written
-    sycl::detail::ur::assertion(strnlen(AddressBuffer, AddressBufferSize) == 12);
+    sycl::detail::ur::assertion(strnlen(AddressBuffer, AddressBufferSize) ==
+                                12);
     return ReturnValue(AddressBuffer,
                        strnlen(AddressBuffer, AddressBufferSize - 1) + 1);
   }

sycl/plugins/unified_runtime/ur/adapters/cuda/event.cpp

@@ -46,8 +46,8 @@ ur_event_handle_t_::ur_event_handle_t_(ur_context_handle_t context,
     : commandType_{UR_COMMAND_EVENTS_WAIT}, refCount_{1}, has_ownership_{false},
       hasBeenWaitedOn_{false}, isRecorded_{false}, isStarted_{false},
       streamToken_{std::numeric_limits<uint32_t>::max()}, evEnd_{eventNative},
-      evStart_{nullptr}, evQueued_{nullptr}, queue_{nullptr}, context_{
-                                                                  context} {
+      evStart_{nullptr}, evQueued_{nullptr}, queue_{nullptr},
+      context_{context} {
   urContextRetain(context_);
 }
 

sycl/plugins/unified_runtime/ur/adapters/cuda/memory.hpp

@@ -7,9 +7,9 @@
 //===-----------------------------------------------------------------===//
 #pragma once
 
+#include <cassert>
 #include <cuda.h>
 #include <ur_api.h>
-#include <cassert>
 
 #include "common.hpp"
 

sycl/plugins/unified_runtime/ur/adapters/cuda/platform.cpp

@@ -57,9 +57,9 @@ UR_DLLEXPORT ur_result_t UR_APICALL urPlatformGetInfo(
 /// However because multiple devices in a context is not currently supported,
 /// place each device in a separate platform.
 ///
-UR_DLLEXPORT ur_result_t UR_APICALL urPlatformGet(uint32_t NumEntries,
-                          ur_platform_handle_t *phPlatforms,
-                          uint32_t *pNumPlatforms) {
+UR_DLLEXPORT ur_result_t UR_APICALL
+urPlatformGet(uint32_t NumEntries, ur_platform_handle_t *phPlatforms,
+              uint32_t *pNumPlatforms) {
 
   try {
     static std::once_flag initFlag;

sycl/plugins/unified_runtime/ur/adapters/cuda/program.cpp

@@ -377,7 +377,7 @@ urProgramRelease(ur_program_handle_t program) {
       // status.
       if (program->buildStatus_ == UR_PROGRAM_BUILD_STATUS_SUCCESS) {
         result = UR_CHECK_ERROR(cuModuleUnload(cuModule));
-      } else if(program->buildStatus_ == UR_PROGRAM_BUILD_STATUS_NONE) {
+      } else if (program->buildStatus_ == UR_PROGRAM_BUILD_STATUS_NONE) {
         // Nothing to free.
         result = UR_RESULT_SUCCESS;
       }

sycl/plugins/unified_runtime/ur/adapters/cuda/queue.hpp

@@ -69,8 +69,8 @@ struct ur_queue_handle_t_ {
         device_{device}, refCount_{1}, eventCount_{0}, compute_stream_idx_{0},
         transfer_stream_idx_{0}, num_compute_streams_{0},
         num_transfer_streams_{0}, last_sync_compute_streams_{0},
-        last_sync_transfer_streams_{0}, flags_(flags),
-        ur_flags_(ur_flags), has_ownership_{backend_owns} {
+        last_sync_transfer_streams_{0}, flags_(flags), ur_flags_(ur_flags),
+        has_ownership_{backend_owns} {
     urContextRetain(context_);
     urDeviceRetain(device_);
   }

sycl/plugins/unified_runtime/ur/adapters/cuda/ur_interface_loader.cpp

@@ -148,7 +148,8 @@ urGetMemProcAddrTable(ur_api_version_t version, ur_mem_dditable_t *pDdiTable) {
   }
   pDdiTable->pfnBufferCreate = urMemBufferCreate;
   pDdiTable->pfnBufferPartition = urMemBufferPartition;
-  pDdiTable->pfnBufferCreateWithNativeHandle = urMemBufferCreateWithNativeHandle;
+  pDdiTable->pfnBufferCreateWithNativeHandle =
+      urMemBufferCreateWithNativeHandle;
   pDdiTable->pfnImageCreateWithNativeHandle = urMemImageCreateWithNativeHandle;
   pDdiTable->pfnGetInfo = urMemGetInfo;
   pDdiTable->pfnGetNativeHandle = urMemGetNativeHandle;