From b55d343f3089a579bde76d310cb7995f7455acf4 Mon Sep 17 00:00:00 2001
From: krishung5 <krish@nvidia.com>
Date: Tue, 7 May 2024 09:16:11 -0700
Subject: [PATCH] Update comments

---
 src/shared_memory_manager.cc | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/src/shared_memory_manager.cc b/src/shared_memory_manager.cc
index 52b5fc9d6f..aafe092e81 100644
--- a/src/shared_memory_manager.cc
+++ b/src/shared_memory_manager.cc
@@ -255,7 +255,7 @@ OpenCudaIPCRegion(
   cudaError_t err = cudaIpcOpenMemHandle(
       data_ptr, *cuda_shm_handle, cudaIpcMemLazyEnablePeerAccess);
   if (err != cudaSuccess) {
-    // Should not pass the detailed error message back to the client.
+    // Log detailed error message
     LOG_ERROR << "failed to open CUDA IPC handle: " << cudaGetErrorString(err);
     return TRITONSERVER_ErrorNew(
         TRITONSERVER_ERROR_INVALID_ARG,
@@ -296,7 +296,7 @@ CheckCudaSharedMemoryRegionSize(
   // User-provided offset and byte_size should not go out-of-bounds.
   if (err != nullptr || byte_size > shm_region_size) {
     if (err != nullptr) {
-      // Should not pass the detailed error message back to the client.
+      // Log detailed error message
       LOG_ERROR << TRITONSERVER_ErrorMessage(err);
       TRITONSERVER_ErrorDelete(err);
     }