From 0522439ac2083f8ee5bee66437548cec12cf4643 Mon Sep 17 00:00:00 2001 From: "Bill(Shuzhou) Liu" Date: Mon, 31 Jul 2023 08:35:24 -0500 Subject: [PATCH] Crash when ecc count sysfile cannot be read Replace assert with error handling code. Change-Id: I6500ae4d38a8caea87828aa7d76373d20c8354c7 --- src/rocm_smi.cc | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/src/rocm_smi.cc b/src/rocm_smi.cc index 4851e8e398..f06ecfac0b 100755 --- a/src/rocm_smi.cc +++ b/src/rocm_smi.cc @@ -682,6 +682,7 @@ rsmi_dev_ecc_count_get(uint32_t dv_ind, rsmi_gpu_block_t block, DEVICE_MUTEX ret = GetDevValueVec(type, dv_ind, &val_vec); + if (val_vec.size() != 2 ) ret = RSMI_STATUS_FILE_ERROR; if (ret == RSMI_STATUS_FILE_ERROR) { ss << __PRETTY_FUNCTION__ << " | ======= end =======" @@ -698,8 +699,6 @@ rsmi_dev_ecc_count_get(uint32_t dv_ind, rsmi_gpu_block_t block, return ret; } - assert(val_vec.size() == 2); - std::string junk; std::istringstream fs1(val_vec[0]);