Plumb port::Status through the internal synchronous memcopy routines.

Now, at least for the public APIs that return port::Status, they can grab the
port::Status that the implementation would like to return and use its
additional information in reporting to the user.
Change: 144741667

2 files changed, 18 insertions, 16 deletions

diff --git a/tensorflow/stream_executor/host/host_gpu_executor.cc b/tensorflow/stream_executor/host/host_gpu_executor.cc
index ff07432bb7..830bc9a681 100644
--- a/tensorflow/stream_executor/host/host_gpu_executor.cc
+++ b/tensorflow/stream_executor/host/host_gpu_executor.cc
@@ -129,23 +129,24 @@ bool HostExecutor::Memset32(Stream *stream, DeviceMemoryBase *location,
   return true;
 }
 
-bool HostExecutor::SynchronousMemcpy(DeviceMemoryBase *gpu_dst,
-                                     const void *host_src, uint64 size) {
+port::Status HostExecutor::SynchronousMemcpy(DeviceMemoryBase *gpu_dst,
+                                             const void *host_src,
+                                             uint64 size) {
   memcpy(gpu_dst->opaque(), host_src, size);
-  return true;
+  return port::Status::OK();
 }
 
-bool HostExecutor::SynchronousMemcpy(void *host_dst,
-                                     const DeviceMemoryBase &gpu_src,
-                                     uint64 size) {
+port::Status HostExecutor::SynchronousMemcpy(void *host_dst,
+                                             const DeviceMemoryBase &gpu_src,
+                                             uint64 size) {
   memcpy(host_dst, gpu_src.opaque(), size);
-  return true;
+  return port::Status::OK();
 }
 
-bool HostExecutor::SynchronousMemcpyDeviceToDevice(
+port::Status HostExecutor::SynchronousMemcpyDeviceToDevice(
     DeviceMemoryBase *gpu_dst, const DeviceMemoryBase &gpu_src, uint64 size) {
   memcpy(gpu_dst->opaque(), gpu_src.opaque(), size);
-  return true;
+  return port::Status::OK();
 }
 
 bool HostExecutor::HostCallback(Stream *stream,
diff --git a/tensorflow/stream_executor/host/host_gpu_executor.h b/tensorflow/stream_executor/host/host_gpu_executor.h
index f217f7947f..77b07e4a57 100644
--- a/tensorflow/stream_executor/host/host_gpu_executor.h
+++ b/tensorflow/stream_executor/host/host_gpu_executor.h
@@ -95,13 +95,14 @@ class HostExecutor : public internal::StreamExecutorInterface {
   bool SynchronousMemSet(DeviceMemoryBase *location, int value,
                          uint64 size) override;
 
-  bool SynchronousMemcpy(DeviceMemoryBase *gpu_dst, const void *host_src,
-                         uint64 size) override;
-  bool SynchronousMemcpy(void *host_dst, const DeviceMemoryBase &gpu_src,
-                         uint64 size) override;
-  bool SynchronousMemcpyDeviceToDevice(DeviceMemoryBase *gpu_dst,
-                                       const DeviceMemoryBase &gpu_src,
-                                       uint64 size) override;
+  port::Status SynchronousMemcpy(DeviceMemoryBase *gpu_dst,
+                                 const void *host_src, uint64 size) override;
+  port::Status SynchronousMemcpy(void *host_dst,
+                                 const DeviceMemoryBase &gpu_src,
+                                 uint64 size) override;
+  port::Status SynchronousMemcpyDeviceToDevice(DeviceMemoryBase *gpu_dst,
+                                               const DeviceMemoryBase &gpu_src,
+                                               uint64 size) override;
 
   bool HostCallback(Stream *stream, std::function<void()> callback) override;