[SYCL][CUDA] Implement sycl_ext_oneapi_peer_access extension (#8303)

This implements the current extension doc from
intel/llvm#6104 in the CUDA backend only.

Fixes intel/llvm#7543.
Fixes intel/llvm#6749.

---------

Signed-off-by: JackAKirk <[email protected]>
Co-authored-by: Nicolas Miller <[email protected]>
Co-authored-by: JackAKirk <[email protected]>
Co-authored-by: Steffen Larsen <[email protected]>

Author: 4 people

common.cpp

@@ -98,6 +98,20 @@ thread_local char ErrorMessage[MaxMessageSize];
   ErrorMessageCode = ErrorCode;
 }
 
+void setPluginSpecificMessage(CUresult cu_res) {
+  const char *error_string;
+  const char *error_name;
+  cuGetErrorName(cu_res, &error_name);
+  cuGetErrorString(cu_res, &error_string);
+  char *message = (char *)malloc(strlen(error_string) + strlen(error_name) + 2);
+  strcpy(message, error_name);
+  strcat(message, "\n");
+  strcat(message, error_string);
+
+  setErrorMessage(message, UR_RESULT_ERROR_ADAPTER_SPECIFIC);
+  free(message);
+}
+
 // Returns plugin specific error and warning messages; common implementation
 // that can be shared between adapters
 ur_result_t urGetLastResult(ur_platform_handle_t, const char **ppMessage) {

common.hpp

@@ -36,6 +36,8 @@ extern thread_local char ErrorMessage[MaxMessageSize];
 [[maybe_unused]] void setErrorMessage(const char *pMessage,
                                       ur_result_t ErrorCode);
 
+void setPluginSpecificMessage(CUresult cu_res);
+
 /// ------ Error handling, matching OpenCL plugin semantics.
 namespace sycl {
 __SYCL_INLINE_VER_NAMESPACE(_V1) {

ur_interface_loader.cpp

@@ -288,6 +288,19 @@ UR_DLLEXPORT ur_result_t UR_APICALL urGetCommandBufferExpProcAddrTable(
   return retVal;
 }
 
+UR_DLLEXPORT ur_result_t UR_APICALL urGetUsmP2PExpProcAddrTable(
+    ur_api_version_t version, ur_usm_p2p_exp_dditable_t *pDdiTable) {
+  auto retVal = validateProcInputs(version, pDdiTable);
+  if (UR_RESULT_SUCCESS != retVal) {
+    return retVal;
+  }
+  pDdiTable->pfnEnablePeerAccessExp = urUsmP2PEnablePeerAccessExp;
+  pDdiTable->pfnDisablePeerAccessExp = urUsmP2PDisablePeerAccessExp;
+  pDdiTable->pfnPeerAccessGetInfoExp = urUsmP2PPeerAccessGetInfoExp;
+
+  return retVal;
+}
+
 #if defined(__cplusplus)
 } // extern "C"
 #endif

usm_p2p.cpp

@@ -0,0 +1,69 @@
+//===--------- usm_p2p.cpp - CUDA Adapter---------------------------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===---------------------------------------------------------------===//
+
+#include "common.hpp"
+#include "context.hpp"
+
+UR_APIEXPORT ur_result_t UR_APICALL urUsmP2PEnablePeerAccessExp(
+    ur_device_handle_t commandDevice, ur_device_handle_t peerDevice) {
+
+  ur_result_t result = UR_RESULT_SUCCESS;
+  try {
+    ScopedContext active(commandDevice->getContext());
+    UR_CHECK_ERROR(cuCtxEnablePeerAccess(peerDevice->getContext(), 0));
+  } catch (ur_result_t err) {
+    result = err;
+  }
+  return result;
+}
+
+UR_APIEXPORT ur_result_t UR_APICALL urUsmP2PDisablePeerAccessExp(
+    ur_device_handle_t commandDevice, ur_device_handle_t peerDevice) {
+
+  ur_result_t result = UR_RESULT_SUCCESS;
+  try {
+    ScopedContext active(commandDevice->getContext());
+    UR_CHECK_ERROR(cuCtxDisablePeerAccess(peerDevice->getContext()));
+  } catch (ur_result_t err) {
+    result = err;
+  }
+  return result;
+}
+
+UR_APIEXPORT ur_result_t UR_APICALL urUsmP2PPeerAccessGetInfoExp(
+    ur_device_handle_t commandDevice, ur_device_handle_t peerDevice,
+    ur_exp_peer_info_t propName, size_t propSize, void *pPropValue,
+    size_t *pPropSizeRet) {
+
+  UrReturnHelper ReturnValue(propSize, pPropValue, pPropSizeRet);
+
+  int value;
+  CUdevice_P2PAttribute cu_attr;
+  try {
+    ScopedContext active(commandDevice->getContext());
+    switch (propName) {
+    case UR_EXP_PEER_INFO_UR_PEER_ACCESS_SUPPORTED: {
+      cu_attr = CU_DEVICE_P2P_ATTRIBUTE_ACCESS_SUPPORTED;
+      break;
+    }
+    case UR_EXP_PEER_INFO_UR_PEER_ATOMICS_SUPPORTED: {
+      cu_attr = CU_DEVICE_P2P_ATTRIBUTE_NATIVE_ATOMIC_SUPPORTED;
+      break;
+    }
+    default: {
+      return UR_RESULT_ERROR_INVALID_ENUMERATION;
+    }
+    }
+
+    UR_CHECK_ERROR(cuDeviceGetP2PAttribute(
+        &value, cu_attr, commandDevice->get(), peerDevice->get()));
+  } catch (ur_result_t err) {
+    return err;
+  }
+  return ReturnValue(value);
+}